text
stringlengths 66
377k
|
|---|
Intraductal biliary polypectomy performed with a nasogastroscope
# Introduction !
A 79-year-old man was admitted to emergency for septic shock due to angiocholitis, with acute renal failure, hyperlactatemia, and thrombopenia. His medical history included a cholecystectomy for acute cholecystitis and removal of choledocholithiasis that occurred more than 10 years ago.
Case report ! Computed tomography confirmed the recurrence of cholelithiasis in the common bile duct (CBD) and in the dilated right hepatic duct associated with atrophy of S6 -7 hepatic segments (- " [fig_ref] !Figure 1: Abdominal computed tomography scan showing the dilated right intrahepatic ducts with intraductal... [/fig_ref]. Multiple stones and pus were extracted from the CBD during endoscopic retrograde cholangiopancreatography, which revealed a tight stricture, which could not be bypassed using several guide wires and upstream dilation, and intraductal stones in the right hepatic duct. The duodenoscope was therefore replaced by a nasogastroscope (GIF-N180, Olympus) introduced with a 50 cm overtube for stability, to allow visualization of the site of stenosis and removal of the stone in the right hepatic duct (- " [fig_ref] Figure 2: Fluoroscopy showing the position of the nasogastroscope and the biopsy forceps in... [/fig_ref]. A 6 mm polyp (Paris 0 -1 sp) (- " [fig_ref] Figure 3: Cholangioscopic view of the polyp below the duct stricture [/fig_ref] was visualized below the stricture and was removed using a diathermy snare (SD-221L-25, Olympus) (- " [fig_ref] Figure 4: Cholangioscopic view of the post-polypectomy site [/fig_ref]. This was followed by several targeted biopsies of the biliary stricture. Scopes were exchanged over a 0.035 Jagwire to dilate the stricture using a 6 mm × 4 cm Hurricane balloon, to extract pus and stones, and to place a 7-Fr plastic biliary stent. Histology showed no malignant cells, but revealed an inflammatory infiltrate in the biopsies and a fibroinflammatory polyp. During follow-up that lasted almost two years, the patient's progress was favorable, the stent was left in place for more than one year, and he experienced no recurrent cholangitis.
Druez Anne et al. Polypectomy with a nasogastroscope … Endoscopy International Open 2014; 02: E124-E125 Abstract: We report the first case of intraductal biliary polypectomy performed with a nasogastroscope.
# Discussion
## !
This case report illustrates the feasibility of biliary intraductal polypectomy during cholangioscopy performed with a nasogastroscope.
## Abbreviations
[fig] !Figure 1: Abdominal computed tomography scan showing the dilated right intrahepatic ducts with intraductal stones. [/fig]
[fig] Figure 2: Fluoroscopy showing the position of the nasogastroscope and the biopsy forceps in the right hepatic duct. [/fig]
[fig] Figure 3: Cholangioscopic view of the polyp below the duct stricture. [/fig]
[fig] Figure 4: Cholangioscopic view of the post-polypectomy site. [/fig]
|
Inhibitory Effects of Respiration Inhibitors on Aflatoxin Production
Aflatoxin production inhibitors, which do not inhibit the growth of aflatoxigenic fungi, may be used to control aflatoxin without incurring a rapid spread of resistant strains. A respiration inhibitor that inhibits aflatoxin production was identified during a screening process for natural, aflatoxin-production inhibitors. This prompted us to evaluate respiration inhibitors as potential aflatoxin control agents. The inhibitory activities of four natural inhibitors, seven synthetic miticides, and nine synthetic fungicides were evaluated on aflatoxin production in Aspergillus parasiticus. All of the natural inhibitors (rotenone, siccanin, aptenin A5, and antimycin A) inhibited fungal aflatoxin production with IC 50 values around 10 µM. Among the synthetic miticides, pyridaben, fluacrypyrim, and tolfenpyrad exhibited strong inhibitory activities with IC 50 values less than 0.2 µM, whereas cyflumetofen did not show significant inhibitory activity. Of the synthetic fungicides, boscalid, pyribencarb, azoxystrobin, pyraclostrobin, and kresoxim-methyl demonstrated strong inhibitory activities, with IC 50 values less than 0.5 µM. Fungal growth was not significantly affected by any of the inhibitors tested at concentrations used. There was no correlation observed between the targets of respiration inhibitors (complexes I, II, and III)OPEN ACCESSToxins 2014, 6 1194 and their IC 50 values for aflatoxin-production inhibitory activity. This study suggests that respiration inhibitors, including commonly used pesticides, are useful for aflatoxin control.
# Introduction
Contamination of agricultural products by mycotoxins, toxic secondary metabolites of fungi, is a worldwide human and livestock health concern that has the potential for drastic economic consequences. Among the mycotoxins, aflatoxins produced by some Aspergillus sp. generate the most concern because of their potent toxicity, carcinogenicity [bib_ref] Global burden of aflatoxin-induced hepatocellular carcinoma: A risk assessment, Liu [/bib_ref] , and high contamination in a wide variety of food and feed commodities. Aflatoxin contamination in agricultural products is a serious problem, but it is difficult to resolve due to the lack of an effective method to control aflatoxin production.
Antifungal agents are typically applied for addressing mycotoxin contamination, however, their use can induce the rapid spread of antifungal-resistant strains [bib_ref] Possible environmental origin of resistance of Aspergillus fumigatus to medical triazoles, Snelders [/bib_ref]. A few fungicides exist that are effective against aflatoxigenic fungi in the field [bib_ref] Ecology of Aspergillus flavus, regulation of aflatoxin production, and management strategies to..., Abbas [/bib_ref]. Additionally, specific aflatoxin-production inhibitors, which do not significantly affect fungal growth, may be useful for the control and prevention of aflatoxin contamination in food and feed without incurring a rapid spread of resistant strains. To date, some pesticides [bib_ref] Inhibition of aflatoxin biosynthesis by organophosphorous compounds, Dutton [/bib_ref] , microbial metabolites, and plant constituents [bib_ref] Diverse inhibitors of aflatoxin biosynthesis, Holmes [/bib_ref] [bib_ref] Isolation of methyl syringate as a specific aflatoxin production inhibitor from the..., Jermnak [/bib_ref] have been shown to be specific aflatoxin-production inhibitors.
We screened a natural products library (RIKEN Natural Products Depositor) and found that siccanin, a respiration inhibitor, inhibits aflatoxin production in Aspergillus parasiticus. Since the relationship between respiration inhibitory activity and aflatoxin-production inhibitory activity has not been reported, the aim of this study was to examine the aflatoxin-production inhibitory activity of natural and synthetic respiration inhibitors.
# Results and discussion
## Inhibitory activity of natural respiration inhibitors on aflatoxin production
All of the four natural respiration inhibitors with different targets (complexes I, II, and III in the mitochondrial respiratory chain) inhibited aflatoxin production of A. parasiticus in a dose-dependent manner . The IC 50 value required for each compound to inhibit aflatoxin production of A. parasiticus is shown in . Rotenone (a complex I inhibitor), siccanin and atpenin A5 (complex II inhibitors), and antimycin A (a complex III inhibitor) had similar activities with IC 50 values around 10 µM. None of the four inhibitors significantly reduced fungal mycelial weight at the concentrations tested. This indicates that they have a high selectivity for aflatoxin production.
All four of the inhibitors tested are known antifungal agents. However, the aflatoxigenic fungus, A. parasiticus, showed high resistance against them. Atpenin A5 and siccanin strongly inhibit growth of some pathogenic fungi, such as Trichophyton mentagrophytes [bib_ref] Atpenins, new antifungal antibiotics produced by Penicillium sp, Ōmura [/bib_ref] [bib_ref] Studies on antibiotics from Helminthosporium sp. fungi. VII. Siccanin, a new antifungal..., Ishibashi [/bib_ref]. Siccanin strongly inhibits succinate dehydrogenase of complex II of T. mentagrophytes, which suggests that complex II may be the primary target for the fungicidal action of siccanin [bib_ref] Siccanin rediscovered as a species-selective succinate dehydrogenase inhibitor, Mogi [/bib_ref]. Atpenin A5 inhibits the mammalian complex II more strongly than siccanin [bib_ref] Atpenins, potent and specific inhibitors of mitochondrial complex II (succinate-ubiquinone oxidoreductase), Miyadera [/bib_ref] , but there is no information on the inhibitory activities of atpenin A5 and siccanin on complex II of aflatoxigenic fungi and it is not clear if the inhibitory activities on the fungal complex II are parallel to their aflatoxin production inhibitory activities. . Effects of natural respiration inhibitors, rotenone (a); siccanin (b); atpenin A5 (c); and antimycin A (d), on aflatoxin (total aflatoxins B 1 and G 1 ) production (gray bars) and mycelial weight (black triangles) of A. parasiticus. n = 4-5, ** p < 0.01; * p < 0.05, vs. control. . Aflatoxin-production inhibitory activity of respiration inhibitors. Note: * For production of total aflatoxin (aflatoxin B 1 and aflatoxin G 1 ).
## Inhibitory activities of synthetic pesticides with respiration inhibitory activity on aflatoxin production
Aflatoxin-production inhibitory activities of seven commercially available miticides and nine fungicides with respiration inhibitory activity were examined [fig_ref] Figure 2: Effects of synthetic miticides, pyridaben [/fig_ref]. The IC 50 values obtained are listed in . Among the miticides, pyridaben and tolfenpyrad (complex I inhibitors) [bib_ref] Complex I inhibitors as insecticides and acaricides, Lummen [/bib_ref] and fluacrypyrim and acequinocyl (complex III inhibitors) demonstrated stronger inhibitory activities than the natural inhibitors. Specifically, pyridaben and fluacrypyrim had IC 50 values less than 0.1 µM. Mepronil (a complex II inhibitor) and bifenazate (a complex III inhibitor) exhibited weak activities, and cyflumetofen (a complex II inhibitor) displayed very weak activity .
It has been shown that cyflumetofen strongly inhibits the mitochondrial complex II of the spider mite, but it does not inhibit the mitochondrial complex II of insects, crustaceans, or mammals [bib_ref] Development of a new acaricide, Ikemi [/bib_ref]. Although it is not clear if cyflumetofen inhibits complex II of fungus, its high selectivity for inhibiting the spider mite complex II might be related to its weak aflatoxin-production inhibitory activity. We did not observe a significant reduction of fungal mycelial weight by any of the miticides tested at the concentrations tested [fig_ref] Figure 2: Effects of synthetic miticides, pyridaben [/fig_ref]. This finding indicates that some miticides, such as pyridaben and fluacrypyrim, can inhibit aflatoxin production by the aflatoxigenic fungus with high selectivity. ; cyazofamid (c); pyraclostrobin (d); kresoxym-methyl (e); azoxystrobin (f); trifloxystrobin (g); picoxystrobin (h); and metominostrobin (i), on aflatoxin (total aflatoxins B 1 and G 1 ) production (blue bars) and mycelial weight (black triangles) of A. parasiticus. n = 4-5; ** p < 0.01; * p < 0.05, vs. control.
All fungicides tested showed strong aflatoxin-production inhibitory activity [fig_ref] Figure 3: Effects of synthetic fungicides, boscalid [/fig_ref]. Among them, boscalid (a complex II inhibitor) [bib_ref] Progress in understanding molecular mechanisms and evolution of resistance to succinate dedydrogenase..., Avenot [/bib_ref] and pyribencarb, kresoxim-methyl, azoxystrobin, and pyraclostrobin (complex III inhibitors) [bib_ref] Review of strobilurin fungicide chemicals, Balba [/bib_ref] inhibited aflatoxin production strongly with IC 50 values comparable to those of pyridaben and fluacrypyrim mentioned above . Since none of the fungicides significantly reduced fungal mycelial weight at the concentrations tested [fig_ref] Figure 3: Effects of synthetic fungicides, boscalid [/fig_ref] , these fungicides also show high selectivity for inhibiting aflatoxin production. Salicylaldehyde was previously shown to enhance the anti-fungal activity of antimycin A and kresoxim-methyl against aflatoxigenic fungi [bib_ref] Chemosensitization of aflatoxigenic fingi to antimycin A and strobilurin using salicylaldehyde, a..., Kim [/bib_ref] , but aflatoxin-production inhibitory activities of antimycin A and kresoxim-methyl were not reported.
Overall, the current study examined inhibitory activities of 20 compounds on aflatoxin production. From the results summarized in , it is difficult to identify a correlation between the targets of the respiration inhibitors (complexes I, II, and III) and their IC 50 values for aflatoxin-production inhibitory activity, suggesting that respiration inhibitors with a variety of targets may have a potential for inhibiting aflatoxin production. Work that investigates the mode of action of respiration inhibitors for inhibition of aflatoxin production is currently in progress.
## Experimental section
## Strains, chemicals, and culture conditions
Aspergillus parasiticus NRRL 2999 was used as a producer of aflatoxins B 1 and G 1 throughout the study [bib_ref] Inhibitory activity of blasticidin A, a strong aflatoxin production inhibitor, on protein..., Yoshinari [/bib_ref]. Aflatoxins B 1 and G 1 are the main aflatoxins produced by the NRRL 2999 strain. NRRL 2999 was maintained on potato dextrose (PD) agar (Difco, MD) and subcultured monthly. A spore suspension prepared from a week-old culture at a concentration of 2.7 × 10 3 cells/µL was used as the inoculum. The spore suspension (30 µL/well) was inoculated into PD liquid media in 24-well microplates (1 mL/well). All test compounds were dissolved in dimethyl sulfoxide and added to the wells (final concentration of dimethyl sulfoxide was 0.1% v/v). The plates were incubated undisturbed at 27.5 °C for three days.
Siccanin and atpenin A are fungal metabolites obtained from the natural products library of the Kitasato Institute for Life Sciences. Rotenone and antimycin A were purchased from MP Biomedicals, LLC, Illkirch, France and Sigma-Aldrich, St. Louis, MO, USA, respectively. Cyflumetofen was gifted from Dr. Ikemi of Otsuka Chemical Co., Ltd. Pesticides except for cyflumetofen were purchased from Wako Pure Chemical Industries, Ltd., Osaka, Japan.
# Analysis of aflatoxin
After three days of incubation, the NRRL 2999 culture broth of each well was centrifuged to obtain the mycelia and the culture supernatant. The mycelia were washed with 1 mL of distilled water two times and collected by centrifugation in a 1.5 mL microtube. After freeze-drying the mycelia, the mycelial weight was calculated by subtracting the weight of an empty 1.5 mL microtube from the total weight. Aflatoxins B 1 and G 1 in the culture supernatant were analyzed in the following manner. The supernatant (0.7 mL) was extracted with 200 µL of chloroform, and the chloroform solution was distilled off by air-drying. The remaining residue was dissolved in 0.1 mL of a 90% aqueous acetonitrile solution. The dissolved mixture was subjected to reverse-phase HPLC on a 250 mm × 4.6 mm i.d. Capcell pak C 18 UG120 column by an isocratic elution of acetonitrile:methanol:water (10:30:60) over 20 min at a flow rate of 1.0 mL with detection at 365 nm to quantify aflatoxins B 1 (retention time: 8.3 min) and G 1 (retention time: 11.1 min).
# Data analysis
The NRRL 2999 strain was cultured in PD liquid medium in a well of microplate, with or without a sample, at 27.5 °C for three days. Amounts of aflatoxins B 1 and G 1 in the culture supernatant and mycelial weight were measured according to the methods mentioned above. This experiment was repeated four or five times (n = 4-5). The aflatoxin amounts and mycelial weight of each well were normalized to those of control well. The normalized values were used for statistical analysis. Data are presented as the mean ± SE. Differences between groups were assessed by Dunnett's test.
# Conclusions
The results obtained in this study indicate that respiration inhibitors can control aflatoxin production, and a microbe that produces a respiration inhibitor, such as antimycin A, has the potential to be a bio-control agent. There are a number of commonly used synthetic pesticides that exhibit respiration inhibitory activity, including those used in this study. This study suggests that those pesticides may be also useful as aflatoxin control agents.
[fig] Figure 2: Effects of synthetic miticides, pyridaben (a); tolfenpyrad (b); mepronil (c); fluacrypyrim (d); acequinocyl (e); and bifenazate (f), on aflatoxin (total aflatoxins B 1 and G 1 ) production (gray bars) and mycelial weight (black triangles) of A. parasiticus. n = 4; * p < 0.05, vs. control. [/fig]
[fig] Figure 3: Effects of synthetic fungicides, boscalid (a); pyribencarb (b) [/fig]
|
Lack of beta-arrestin signaling in the absence of active G proteins
Supplementary Figure 1. Biochemical and functional characterization of ΔGsix cells. (a-e) Western blots of wild-type, ΔGq/11/12/13 and ΔGsix HEK293 cell lysates probed with antibodies against Gα q/11/14 (a), Gα 12 (b), Gα 13 (c), Gα s/olf (d) and Gα i (e). Uncropped blots are available in Supplementary Fig. 17. (f) cAMP levels after stimulation of vasopressin V2 receptor (V2R) or mock transfected wild-type and ΔGsix cells with the indicated compounds or vehicle for 10 minutes. (g,h) DMR traces of PGD 2 -stimulated DP2 receptor in ΔGsix HEK293 cells in absence (g) and presence (h) of PTX. (i) Concentration-response-curve of peak value data from (g) and (h). (j,k) Real-time whole cell response of ΔGsix (j) and wild-type (k) HEK293 cells upon stimulation with 30 nM epidermal growth factor (EGF). (a-e) Representative blots of 2 independent experiments. (f,i) Data are mean +/± SEM of 3 independent experiments. (g,hj,k) Shown are representative traces (mean + SEM) of 3 independent experiments, each performed in triplicate.
## Supplementary
. G protein-dependence of ERK1/2 phosphorylation for the GPR17 receptor. 10 µM MDL-induced kinetic pERK1/2 and total ERK profile of GPR17 in ΔGsix cells (+/-PTX/FR). Data are mean ±/+ SEM of 3 independent experiments, each measured in triplicates. Fiob, fold increase over basal. . Comparative analysis of ERK1/2 phosphorylation and cell surface abundance of DP2, GPR17 and FFA2 in parental HEK293 and Δβarr1/2 cells. (a-c) pERK1/2 and total ERK1/2 kinetic profile of DP2 (a), GPR17 (b), and FFA2 (c) in wild-type and Δβarr1/2 cells stimulated with 1 µM PGD 2 (a), 10 µM MDL (b), and 100 µM C3 (c). (d-f) Concentration-effect-curves of pERK1/2 levels after 5 minutes stimulation of WT and ∆βarr1/2 cells expressing DP2 (d, pEC 50 : 9.78±0.32 (WT); 9.0±0.35 (∆βarr1/2)), GPR17 (e, pEC 50 : 6.64±0.7 (WT); 5.86±0.14 (∆βarr1/2)) or FFA2 (f, pEC 50 : 4.0±0.1 (WT); 4.0±0.12 (∆βarr1/2)) with their cognate agonist. (g-i) Quantification of receptor expression in wild-type and Δβarr1/2 cells of DP2 (g), GPR17 (h), and FFA2eYFP (i). Data are mean +/± SEM of 3 independent experiments (3 technical replicates). For statistical analysis, two-sample paired Wilcoxon test was applied to paired points at different times (a-c) or at different concentrations (d-f) and two-tailed paired t-test (g-i) were performed. *, P<0.05; **, P<0.01; not significant where no asterisk. Fiob, fold increase over basal.
## Supplementary
## Supplementary
[fig] Supplementary, Figure 2: Forskolin-mediated DMR in the absence and presence of signal transduction inhibitors PTX and FR. (a-c) Forskolin (fsk)-induced DMR traces of DP2 (a), GPR17 (b), or FFA2 (c) wild-type HEK293 cells in absence and presence of G protein inhibitors. (d-f) Quantification of fsk-mediated DMR responses from conditions in (a-c). (a-c) Shown are representative traces (mean + SEM) of 3 independent experiments, performed in triplicate wells for each condition. (d-f) AUC of DMR traces over the entire measurement time was analyzed and presented as mean values + SEM of 3 independent experiments (3 technical replicates). [/fig]
[fig] Figure 5 Supplementary, Figure 8: Comparative analysis of ERK1/2 phosphorylation for ligand-stimulated β2AR in parental HEK293 and Δβarr1/2 cells. (a,b) pERK1/2 kinetic profiles of β2AR in wild-type parental and Δβarr1/2 HEK293 cells stimulated with 10 µM isoproterenol (Iso) (a) and 10 µM carvedilol (Carv) (b). Data are mean ± SEM of 4 independent experiments (3 technical replicates each). For statistical analysis, two-sample paired Wilcoxon test was applied to paired points at different times; not significant where no asterisk.SupplementaryFigure 6. Cell surface abundance of β2-adrenergic receptor (β2AR) in different HEK293 cell lines. Surface ELISA detecting N-terminally HA-tagged β2AR in wild-type, ΔGs and Δβarr1/2 HEK293 cells. Data are mean + SEM of 3 independent experiments, performed in triplicate.SupplementaryFigure 7. Comparative analysis of ERK1/2 phosphorylation for ligand-stimulated AT1R and V2R in parental HEK293 and Δβarr1/2 cells. (a-c) pERK1/2 kinetic profiles of AT1R (a,b) and V2R (c) in wildtype parental and Δβarr1/2 HEK293 cells stimulated with 100 nM angiotensin II (AngII) (a), 30 µM [Sar 1 , Ile 4 , Ile 8 ]AngII (SII) (b), and 1 µM arginine-vasopressin (AVP) (c). Data are mean ± SEM of 4 (AT1R in WT), 5 (AT1R in Δβarr1/2), and 3 (V2R) independent experiments (3 technical replicates each). For statistical analysis, two-sample paired Wilcoxon test was applied to paired points at different times. *, P<0.05; not significant where no asterisk. Impact of increasing β-arrestin2 amounts on β2AR-mediated pERK1/2 kinetics. (a-c) Kinetic ERK1/2 phosphorylation and total ERK profile of 10 µM Isoproterenol-stimulated Δβarr1/2 cells stably transfected with β2AR and transiently transfected with either empty vector (pcDNA3.1) or a range of β-arrestin2-GFP (βArr2) amounts.(d-f) Fluorescence imaging of β-arrestin2 transfected cells for each condition in a-c. (a-c) Data are mean ± SEM of 4 independent experiments (3 technical replicates). For statistical analysis, twosample paired Wilcoxon test was applied to paired points at different times. *, P<0.05; not significant where no asterisk. (d-f) Representative images of 4 independent experiments. Scale bar is 100 µm. Supplementary Figure 9. Cell surface abundance and functionality of M3D-βArr in wild-type HEK293 cells. (a) Surface amounts of N-terminally HA-tagged DREADD constructs (M3D-WT, M3D-Gq, M3D-βArr) in wildtype (WT) HEK293 cells. (b) CNO-induced β-arrestin2 recruitment of the M3D-βArr receptor in absence (black) and presence (grey) of G protein inhibitors (PTX/FR). Data are mean + SEM of 3 independent experiments, each measured in triplicates. Supplementary Figure 10. Cell surface DREADD abundance and β-arrestin2 recruitment in the absence of G proteins. (a), Surface ELISA of N-terminally HA-tagged DREADD constructs (M3D-WT, M3D-Gq, M3D-βArr) in HEK293 cells collectively lacking G proteins (ΔGsix + PTX). (b) β-arrestin2 recruitment by M3D-WT and M3D-βArr receptors upon CNO stimulation in cells either partially (ΔGsix, black) or collectively (ΔGsix + PTX, grey) depleted of G proteins. Data are mean + SEM of 3 independent experiments (3 technical replicates). Supplementary Figure 11. DREADD cell surface abundance in Δβarr1/2 cells. Surface ELISA of N-terminally HA-tagged DREADD constructs (M3D-WT, M3D-Gq, M3D-βArr) in Δβarr1/2 cells, normalized to surface abundance of receptors in WT HEK293 cells (100%). Data are mean + SEM of 3 independent experiments, performed in triplicate. For statistical analysis, one sample t-test was performed. *, P<0.05. Supplementary Figure 12. Concentration-effect curves of DMR recordings for CNO-activated DREADDs in HEK293 cells. Quantification of CNO-induced DMR response in wild-type HEK293 cells expressing the M3D-WT (pEC 50 : 6.68) (a), the M3D-Gq (pEC 50 : 8.12) (b), or the M3D-βArr (c) receptor. AUC of DMR traces over the entire measurement time was analyzed and presented as mean values ± SEM of 3 independent experiments (3 technical replicates). Supplementary Figure 14. DMR analysis of different HEK293 cell clones. Quantification of DMR response towards multiple stimuli in wild-type (a), wild-type (parental) (b), Δβarr1/2 clone #2 (c) and Δβarr1/2 clone #3 (d). (e-h) AUC of DMR traces over the entire measurement time was analyzed and presented as mean values + SEM of 3 independent experiments, each measured in triplicates. Supplementary Figure 15. Agonist-mediated internalization in wild-type (parental) and Δβarr1/2 HEK293 clones #1, #2, and #3. Quantification of cell surface receptor abundance of FLAG-tagged Vasopressin V2 receptor (FLAG-V2R) (a) and FLAG-tagged β2-adrenergic receptor (FLAG-β2AR) (b) in parental HEK293 cells (Parent) and Δβarr1/2 clones #1, #2, and #3, either transfected with vector (Mock) or β-arrestin1 (βarr1) or β-arrestin2 (βarr2) and stimulated with 100 nM AVP (a) or 10 µM Isoproterenol (Iso) (b). Data are mean + SEM of 4 independent experiments (3 technical replicates). Supplementary Figure 16. FFA2 receptor internalization in the absence of active G proteins. Structuredillumination micrographs of YFP-tagged FFA2 receptors in wild-type (a) or ΔG12/13 (b) cells, imaged 30 minutes after buffer treatment (left panels) or stimulation with 100 µM C3 (right panels). Scale represents 10 µm; representative images of 3 independent experiments. a b [/fig]
|
Compartmentalized Neuronal Culture for Viral Transport Research
Neuron-invading viruses usually enter via the peripheral organs/tissues of their mammalian hosts and are transported to the neurons. Virus trafficking is critical for transport or spread within the nervous system. Primary culture of neurons is a valuable and indispensable method for neurobiological research, allowing researchers to investigate basic mechanisms of diverse neuronal functions as well as retrograde and anterograde virus transport in neuronal axons. Primary ganglion sensory neurons from mice can be cultured in a compartmentalized culture device, which allows spatial fluidic separation of cell bodies and distal axons. These neurons serve as an important model for investigating the transport of viruses between the neuronal soma and distal axons. Alphaherpesviruses are fascinating and important human and animal pathogens, they replicate and establish lifelong latent infection in the peripheral nervous system, the mechanism of the viral transport along the axon is the key to understand the virus spread in the nervous system. In this review, we briefly introduce and evaluate the most frequently used compartmentalization tools in viral transport research, with particular emphasis on alphaherpesviruses.
# Introduction
Studying viruses that infect the nervous system remains a challenge, especially determining how viruses spread within the nervous system. Most viruses begin infecting the nervous system from periphery tissue, often at epithelial or endothelial cell surfaces, invading the peripheral nervous system (PNS) or the central nervous system (CNS) along the axon of the neuron. The CNS is well-protected by complex multilayer barriers and immune responses. After virus infection, viruses invade the CNS via two main circuits: the blood supply and peripheral nerves. Neurons have long axons, and transport along the axon is necessary for the spread of some viruses; therefore, studying the mechanism of virus transport along axons is useful for understanding virus spread in the nervous system. Isolated primary neurons such as dorsal root ganglia (DRG) neurons and superior cervical ganglia (SCG)are widely used for experimental investigation of relevant questions in neurobiology and neurodevelopment and have served as an ideal model for understanding mechanisms of virus spread in the PNS. DRG including the neuronal cell bodies of sensory neurons and glial cells and are located in close proximity to the dorsal root of the spinal cord. DRG neurons are first-order neurons essential for the afferent transport of sensory information through long axons from the limbs and trunk into the CNS. DRG have pseudo-unipolarized morphology with well-defined cellular compartments, extremely long axons, and soma. The differentiation of sensory neurons from induced pluripotent stem cells (IPSCs) or embryonic stem cells (ESCs) which share the same properties with DRG neurons and also become ideal cell type to investigate the virus spread in the neurons. Conventional neuron culture methods do not allow manipulation and control of the culture environment between the soma and axon with enough spatial and temporal resolution to explore virus infection. Several new methods and techniques, consisting of compartmentalized neuronal culture tools, have therefore been developed for exploring virus spread in neurons through the axon. We summarize currently used methods in the field of compartmentalized platforms for understanding viral infection and transport in the nervous system.
## Alphaherpesviruses
Alphaherpesviruses (αHV) belong the family Herpesviridae comprising 41 species. Most are neurotropic viruses [such as herpes simplex virus (HSV), pseudorabies virus (PRV), and varicella-zoster virus (VZV)] and invade the nervous system when infecting their natural host. Herpesviruses infect epithelial cells, resulting in invasion of the PNS. However, the mechanism of invasion and transport of αHV to neurons remains elusive.
HSV and VZV are prevalent and important human pathogens with seroprevalence of ∼80% in adults. HSV-1 causes cold sores; it replicates in mucosal epithelia and is transmitted to nerve endings. Transmission is retrograde along the axon to the peripheral ganglia, where the virus can establish long-lasting latent infection; Reactivation of HSV can be triggered by a variety of internal or external stimuli to result in the production of infectious virus, and release of it to the mucosal epithelia by anterograde transport. VZV, causative agent of chickenpox, can also establish a latent infection in the nervous system. Its reactivation in a dorsal root ganglion can lead to herpes zoster with debilitating chronic pain.
PRV is not a human pathogen, but its genome structure and function are similar to those of VZV and HSV. Its natural hosts are pigs, but it can infect a broad range of animals. PRV is easy to culture in the laboratory, with rodent and chicken embryo models widely used to study mechanisms of PRV spread and replication in neurons. All these features make PRV a perfect model for herpesvirus research. Bidirectional transport along the axon enables PRV invasion, replication, and spread in PNS neurons.
αHV usually infect the mucosal surface of the host and then invade the nervous system by retrograde transport. Subsequently, the virus spreads to nerve endings to egress or enter the CNS by anterograde transport.
Understanding how viruses are transmitted from the mucosal surface to the nervous system can help avoid viral infection of neurons. It is therefore critical and necessary to understand the actively transport and spread of viruses in the neurons. By learning how viruses infect the nervous system and travel in neurons, we can mimic this process to study and treat related neurological diseases.
## Compartmentalized neuronal culture
The most frequently used compartmentalized neuronal culture systems are the Campenot chamber and microfluidic cell culture. These tools enable manipulation and separation of the soma and axons.
## Campenot chamber
In 1977, the compartmentalized culture system known as the "Campenot chamber" was introduced by. This three-chamber compartmentalized culture system was used to investigate nerve growth factor (NGF) control over the outgrowth of neurites from sympathetic neurons in vitro. It allows neurons to be seeded in one chamber and neurites to extend to other chambers, resulting in spatial separation of the neurons and axon. The fluidic separation of the chamber meets the requirements for NGF treatment of cell bodies or axons separately.
Briefly, silicone grease is used to seal the chamber to a tissue culture dish pre-coated with collagen and scratched to form parallel tracks. This separates the three chambers while allowing axonal growth underneath the watertight silicone grease into a separate compartment. This method has been extensively used to study various neuronal functions, permitting analysis of RNA, protein, and lipids in both neuronal soma and distal axons separately during neuronal culture. This has helped in studying neurotrophin signaling in the transport in the neurons. Campenot chambers have been used by the Enquest laboratory to investigate viral transport. The unique design of the Teflon ring makes it easy to infect the neurons distally or proximally. In this tri-chamber system, neuronal soma and their long axons are separately maintained in different chambers, allowing establishment of fluidic milieu isolated over each segment of the neuron, fluids in three chambers are separated and do not mix. Neuroinvasion viruses can therefore be inoculated into either the chamber containing the neuronal soma or the chamber with distal axons. If infection begins at the neuronal soma, viral particles will be transmitted anterogradely to distal axons along microtubules; in contrast, viral particles may display retrograde transport along microtubules to the soma from distal axons.
## Compartmentalized microfluidic culture
Compartmentalized microfluidic technology has provided advanced platforms for neuroscience research over several decades. It combines microfluidic, microfabrication, and surface micropatterning techniques to generate a neuronal culture system in a multicompartment device. Application of microfluidics in the field of cell biology is becoming a powerful tool, due to its ability to precisely control, manipulate, and monitor sub-millimeter volume cellular culture environments. Initially, microfluidic devices were designed for chemical and biomolecular analyses with higher sensitivity than traditional methods; more recently, they have been applied in cell biology.
Polydimethylsiloxane (PDMS) is the most popular material for fabricating microfluidic chambers using soft lithography techniques, with a plasma or non-plasma bonding methodused to seal the chamber to a glass substrate. The chambers are divided by a barrier into which micron-size grooves are engraved, allowing neurites past the chambers.
## Advantages and disadvantages of campenot chambers and compartmentalized microfluidic culture systems
Compartmentalized culture is becoming a valuable and powerful tool for neuroscience studies. It is difficult to manipulate and control the axonal growth of neurons in tissue culture plates with isotropic environment because establishment of neuronal polarity involves spontaneous and randomly orientated outgrowth. Compartmentalized culture systems confer a major advantage over single-chamber cell cultures by spatially separating the cellular compartment, the soma, and extension of the axon.
The Campenot chamber culture method has several disadvantages: (1) The device is difficult to assemble for the early users; (2) mechanical disturbance of the chamber may lead to leak owing to the silicon grease and disrupt the growth of neurites; (3) it is difficult to adapt the technique for sophisticated microscopy, owing to the structure of inner barrier and unoptical material, such as confocal microscopy and live cell imaging during the viral transport along the axon.
Microfluidics utilizes PDMS, which is easy to fabricate, low cost, and has optical properties compatible with highresolution microscopy, including confocal microcopy. It is also non-toxic and permeable to gas, providing alternatively solution for compartmentalized neuronal cell culture. it offers optical transparency, biocompatibility, and spatial and temporal control for manipulating cellular microenvironments.
The disadvantages of microfluidic devices are as following: (1) The PDMS surface can absorb small hydrophobic molecules; (2) limited number of axons entering the microfluidic grooves; (3) small volumes of media and constant requirements for fluidic flow to maintain diffusion of small molecules; (4) frequency of channel "blockage" during construction and reuse.
Most of these disadvantages of two methods can be surmounted by skilled operation or redesign the model for fabricating the microfluidic device. All these properties make Campenot chamber and microfluidic culture are ideal and best tool for neuron growth and transport research.
There are still some other compartmentalized devices which are modified based on the "Campenot chamber" or microfluidics, such as two chambers deviceand redesigned microfluidic culture device.
## Application of compartmentalized devices in neuronal transport models in vitro
Neuroinvasive viruses spread by infecting nerve terminals and transporting to the soma or other neurons, and by egressing from the soma to the nerve terminals utilizing cellular anterograde and retrograde transport.
Neurons are polarized cells with very long axons, which generally arise from the cell soma and synaptic contacts with other neurons or neurites. Axonal transport is an essential cellular process responsible for trafficking cargo such as mRNA granules, proteins, lipids, mitochondria, synaptic vesicles, autophagosomes, lysosomes, and organelles to their destinations along a polarized microtubule network using kinesin and dynein motor proteins. Anterograde transport is movement of newly biosynthesized proteins and lipids, and organelles from the soma to the synapse or cell membrane of distal axons utilizing the motor protein kinesin. In the opposite direction, retrograde transport is involved in maintaining homeostasis by shuttling organelles and aging proteins from axon termini to the cell soma for degradation and recycling of componentsutilizing the cytoplasmic protein dynein.
Several studies have identified viruses that are transported along the axon. When αHV enters into axons after membrane fusion, the capsid tegument protein VP1/2 recruits dynein motor proteins for retrograde transport to the cell soma; however, kinesin motor proteins can also bind purified capsid tegument complexes, suggesting that VP1/2 participates in multiple intracellular axonal transport processes during virus replication. During αHV egress and spread, the viral membrane glycoprotein gE/gI and US9 protein of PRV and HSV participate in anterograde transport to direct progeny virions either into peripheral tissues or into the CNS. Interactions between viral proteins and cellular factors are complex, requiring further study using these techniques.
## Conclusions and future prospects
In conclusion, compartmentalized culture methods can be used to investigate viral transport. The most common and popular are the "Campenot chamber" and microfluidic culture system, and these methods provide the broadest picture of viral transport mechanisms.
Compartmentalized neuronal culture devices overcome the restrictions of conventional culture methods, making it easy and possible to understand the roles of viral proteins and neuronal proteins in axonal transport. This method can also be applied to co-culture epithelial or glial cells to study the role of these cells during the transport of the virus along the axon. By understanding how viruses move in axons, we can obtain more insight into the process of viral transmission and take measures to control it, especially the threat of latency of herpesviruses.
However, many questions still need to be addressed, such as how transport is coordinated by viral and cellular factors; how virion assembly affects the direction of transport; and how viral proteins and motor proteins are assembled.
# Author contributions
YW and YS designed the concept of the review article. YW, SW, HW, and YS wrote the manuscript. XL, JM, LW, HQ, MK, and AR helped with revision of the manuscript. All authors read and approved the final manuscript. |
Ethical dilemmas experienced by nurses while caring for patients during the COVID‐19 pandemic: An integrative review of qualitative studies
Aim: This study aimed to identify ethical dilemmas faced by nurses while caring for patients during the COVID-19 pandemic.Background: Nurses express several concerns during disease outbreaks, some of which are related to ethical dilemmas.Evaluation: It is an integrative review in which four databases were searched. Critical appraisal tools and PRISMA guidelines were used. Content analysis was performed to analyse the obtained data.Key issues: A total of 14 studies were identified. The results are presented into four categories: concerns with beneficence-nonmaleficence; awareness of need for autonomy; challenges to justice; and coping with ethical dilemmas.Conclusion: While caring for patients during the COVID-19 pandemic, nurses often put their own health and that of their families at risk. The ethical dilemmas faced by nurses are mainly caused by the lack of Protective Personal Equipment (PPE), shortages of medical supplies and personnel and the uncertainties that permeate an environment threatened by a new and highly contagious disease such as COVID-19.
# | introduction
The advent of the COVID-19 pandemic has had a significant impact on humanity; although several fields have been affected by the pandemic, the most impacted was undoubtedly the health care sector [bib_ref] Nurses on the frontline against the COVID-19 pandemic: An integrative review, Al Thobaity [/bib_ref]. Since its emergence in late December 2019, the disease has infected and caused the death of millions of people (World Health Organization , 2022), and still today, more than 2 years later, it continues to challenge scientists and health professionals with the emergence of mutations and variants [bib_ref] Emerging sars-cov-2 variants: A review of its mutations, its implications and vaccine..., Ramesh [/bib_ref].
Lack of knowledge about the disease, its contagion and treatment; fear of getting infected; fear for loved ones; discriminatory acts; and shortage of human and medical resources are some of the challenges faced by health professionals since the beginning of the pandemic (al Thobaity & Alshammari, 2020; [bib_ref] Fear of nurses during COVID-19 pandemic in Saudi Arabia: A cross-sectional assessment, Moussa [/bib_ref] [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref]. Especially nurses, who are health professionals present at all times on the front line of the battle against the COVID-19, faced and still face several difficulties due to the pandemic. Rationing of limited resources, restrictions on the freedom and autonomy of patients and their families and the distinction between groups, in choosing who should receive care when patients are many and nurses are few, are among the many ethical problems faced by nurses. Nurses face yet another huge ethical dilemma, as they have an obligation to care for patients and for themselves and their families [bib_ref] Ethical rationing of personal protective equipment to minimize moral residue during the..., Binkley [/bib_ref] [bib_ref] Self-care as an ethical obligation for nurses, Linton [/bib_ref]. When providing care to patients with nurses have to deal with all these ethical dilemmas, which, in addition to putting pressure on them, interfere with the quality of care [bib_ref] Nurses on the frontline against the COVID-19 pandemic: An integrative review, Al Thobaity [/bib_ref]. Thus, it is important to examine the ethical dilemmas faced by nurses during the COVID-19 pandemic.
# | background
Nurses in their daily practices have to protect their patients and their families without neglecting self-care; these ethical problems are obviously exacerbated in periods of crisis [bib_ref] Self-care as an ethical obligation for nurses, Linton [/bib_ref]. As in previous outbreaks, nurses are currently facing excessive workload, shortages of medical supplies and human resources, lack of knowledge and skills and fear of getting infected and infecting loved ones [bib_ref] Deciding to work during the Ebola outbreak: The voices and experiences of..., Kollie [/bib_ref] , which can trigger ethical dilemmas due to the risk of harming themselves and others. During the decision-making process, nurses are guided by ethical principles such as respect for autonomy, beneficence, nonmaleficence and justice [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref].
Whereas the ethical principle of beneficence is related to doing good, the principle of nonmaleficence is about not causing harm to the patient [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref]. During the pandemic, both principles were threatened since nursing shortages, lack of knowledge about the disease, treatment limitations, lack of resources, such as Protective Personal Equipment (PPE), and other medical materials prevented some patients from receiving assistance promptly, in addition to threatening the health of caregivers [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref].
The ethical principle regarding autonomy refers to the rights of individuals to dignity, to be informed about their health and to be able to make choices without suffering external pressure [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref] ; patients are also entitled to confidentiality regarding their health status and treatment [bib_ref] Privacy and patient confidentiality in times of Covid-19, Shekhawat [/bib_ref]. Holistic and humanized care is centred on the individuals, including attention to their values, preferences and needs; patients must be free to choose their treatment, therefore having the right to be informed about their illness and the entire assistance process, including nursing care [bib_ref] Where does patient autonomy live, Fontes [/bib_ref]. Nurses are trained to provide this type of care to patients. However, in periods of crisis, lack of knowledge and uncertainties arise and profound changes in the world health scenario give rise to situations that threaten the patient's autonomy over his/her own life, and nurses are also affected by these changes, facing major ethical problems [bib_ref] Where does patient autonomy live, Fontes [/bib_ref]. Still, on the ethical principle of autonomy, it is necessary to emphasize that nurses also have the right to make autonomous decisions about their obligations to serve others when their lives are threatened by PPE shortages [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref].
Justice implies equity, fairness and proportionality; thus, the concept of justice in the field of health also refers to the elimination of unequal access to health services, guaranteeing access to quality health for all [bib_ref] Choosing which COVID-19 patient to save? The ethical triage and rationing dilemma, Jaziri [/bib_ref]. If before the pandemic for some countries guaranteeing quality health care for all was a huge challenge, with the emergence of the COVID-19, the situation has worsened worldwide [bib_ref] Choosing which COVID-19 patient to save? The ethical triage and rationing dilemma, Jaziri [/bib_ref] [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref]. It is still necessary to consider the situation of health professionals, especially nurses, who, due to an unequal distribution of PPE, faced great risk when providing care to patients with COVID-19 [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref].
Health institutions must be well organised to provide care during times of crisis. Pandemics lead to the rapid spread of disease affecting the ability of these institutions to provide assistance to the population. Health care institution administrators and also nurse managers must envision this possibility and must be prepared for such events before they occur [bib_ref] Hospital preparedness assessment against COVID-19 pandemic: A case study in Turkish tertiary..., Gul [/bib_ref]. It is also important to ensure organisational fairness so that everyone who is part of the health team feels valued and indispensable [bib_ref] A qualitative study on the experiences of the first nurses assigned to..., Yildirim [/bib_ref]. Efficient and effective management of human resources, equipment, materials and information is essential for the control of pandemics [bib_ref] Hospital preparedness assessment against COVID-19 pandemic: A case study in Turkish tertiary..., Gul [/bib_ref]. In addition, nurse managers play an important role in supporting their team so that frontline nurses can develop emotional and professional competence to respond to emergencies [bib_ref] Experiences of clinical first-line nurses treating patients with COVID-19: A qualitative study, Tan [/bib_ref].
Health professionals must develop their functions based on scientific knowledge, technical and communication skills associated with ethical and professional values [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref]. However, during the COVID-19 pandemic, ethical principles are being threatened in several dimensions, putting at risk not only the quality of care but also the physical and mental health of nurses and other health care workers [bib_ref] Self-care as an ethical obligation for nurses, Linton [/bib_ref] [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref]. Thus, organisational support is pivotal for nurses to cope with ethical dilemmas (American Nurses Association [ANA], 2020). Importantly, the focus of this review is to examine the ethical dilemmas experienced by nurses while caring for patients with COVID-19.
## | aim
This study aimed to identify ethical dilemmas faced by nurses while caring for patients during the COVID-19 pandemic.
# | methods
## | design
This integrative review of qualitative studies was conducted using the [bib_ref] Shortage of personal protective equipment endangering health workers worldwide, Whittemore [/bib_ref] framework. As qualitative research allows exploring people's lived experiences, the selection of studies in which this method was used was considered better suited to identify ethical dilemmas experienced by nurses during the COVID-19 pandemic.
## | search strategy
The search for original primary qualitative research articles on ethical dilemmas experienced by nurses while caring for patients during the COVID-19 pandemic was carried out in December 2021. The descriptors used for the database searches were 'Covid' AND 'ethical dilemmas' AND 'nurses'. The electronic databases searched were PubMed, Google Scholar, MEDLINE and Scopus.
## | inclusion and exclusion criteria
Original primary qualitative research articles reporting ethical dilemmas of nurses caring for patients with COVID-19, whose full texts were available on the Internet in English, were included. Articles that included other participants besides staff nurses (other health care workers, nurse managers, nursing assistants), those that despite addressing experiences did not assess ethical dilemmas perceived by nurses while caring for patients during the COVID-19 and nonqualitative studies were excluded.
## | search outcome
The Preferred Reporting Items for Systematic Reviews and Metaanalyses (PRISMA) was used to guide the selection of articles . Through the electronic databases, 918 articles were found, with the inclusion of four more articles found through other sources, a total of 922 articles were identified, then 19 duplicate articles were excluded, and the result decreased to 903. Titles and abstracts were read, and 889 articles were excluded for not matching the inclusion criteria of the review. The remaining 14 articles were read and reread, and all were included in the quality appraisal.
## | quality appraisal
The articles were appraised using the Critical Appraisal Skills Programme (CASP) . The CASP is a checklist with 10 main questions used to evaluate qualitative studies . Each main question was rated as 'Yes' (2 points), 'Unclear' (1 point) or 'No'
(0 points). The included studies scored between 17 and 20 points.
Regarding the methodological quality of the studies, the most common weaknesses found were a lack of information on the relationship between researcher and participants and geographic limitations. No articles were excluded based on critical appraisal; all 14 evaluated articles were included in this review [fig_ref] ,: four articles [/fig_ref].
## | data extraction
General information was extracted from the selected studies and organised through an evidence-based instrument developed and pilot tested on the first two included articles by the author. The form contains the following headings: reference, title, journal, country, aim, design, participants, main results and limitations. Relevant data are included in quality appraisal [fig_ref] ,: four articles [/fig_ref].
## | data synthesis
Data analysis involved a long process in which comparisons of information were conducted resulting in codes organised into categories [bib_ref] Shortage of personal protective equipment endangering health workers worldwide, Whittemore [/bib_ref]. Data were thoroughly and repeatedly analysed by the author. Grammatical methods were used for coding similar information contained in the studies included in this review. The codes originated by comparing the data of the reviewed studies were organised into four categories: concerns with beneficence-nonmaleficence; awareness of need for autonomy; challenges to justice; and coping with ethical dilemmas [fig_ref] ,: four articles [/fig_ref].
# | results
The 14 selected articles are from 10 different journals. Thirteen (92.8%) of the studies were carried out in 2021. The number of participants varied from 10 to 43 nurses. Studies were conducted in Iran (n = 4), China (n = 2), Turkey (n = 2), the United States (n = 2), Canada (n = 1), Jordan (n = 1), Korea (n = 1) and Sweden (n = 1). In nine studies, the majority of participants were female; in two studies, all participants were female [bib_ref] Visitor restrictions, palliative care, and epistemic agency: A qualitative study of nurses..., Mcmillan [/bib_ref] ; in two studies, gender was not mentioned [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] ; and in one study, the majority of participants (60%) were male [bib_ref] Nurses' ethics in the care of patients during the COVID-19 pandemic, Alloubani [/bib_ref].
A total of seven articles had the main objective related to topics on nurses' ethics during the COVID-19 pandemic [bib_ref] Nurses' ethics in the care of patients during the COVID-19 pandemic, Alloubani [/bib_ref] [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] [bib_ref] Visitor restrictions, palliative care, and epistemic agency: A qualitative study of nurses..., Mcmillan [/bib_ref] [bib_ref] Nurses perception of ethical challenges in caring for patients with covid-19: A..., Rezaee [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref]
## | concerns with beneficence-nonmaleficence
According to the analysed literature, nurses questioned themselves at various moments due to different circumstances, having doubts if they were doing what was good and right for patients during the COVID-19 pandemic. Nurses reported that at times, they no longer knew what was ethical and what was not, that many decisions were made without nurses being consulted, and they were the last ones to know for example about the placement of patients [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref]. Nurses faced dilemmas as they had to protect themselves, fight for their rights and, at the same time, be beneficial to [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] ; thus, they feared that they were not doing their job correctly [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Experiences of nurses caring for patients with COVID-19 in Turkey: A phenomenological..., Muz [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] and that they were harming the patient due to inexperience or lack of necessary training [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref]. In addition, due to nursing shortages and long shifts, nurses were not able to stay at patients' bedsides at the time of need [bib_ref] Nurses perception of ethical challenges in caring for patients with covid-19: A..., Rezaee [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref].
Nurses also experienced moral distress because they felt unable to provide the necessary support to patients who suffered from the loss of loved ones or because they felt lonely in isolated rooms [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref]. Nurses faced dilemmas due to the prohibition of visits to patients during the COVID-19 pandemic, as they had doubts whether this measure was more beneficial or harmful to patients and their families [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref].
Nurses pointed out that family visits are not only important to provide emotional support but also to give information about patients [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref]. Especially in palliative care services, visitor restrictions were pointed out as a huge ethical dilemma for nurses, because contact with family is considered a palliative intervention [bib_ref] Visitor restrictions, palliative care, and epistemic agency: A qualitative study of nurses..., Mcmillan [/bib_ref]. Nurses also emphasized the lack of support for families after the patient's discharge or death [bib_ref] Nurses perception of ethical challenges in caring for patients with covid-19: A..., Rezaee [/bib_ref]. Patients were in dire need of spiritual care, which unfortunately could not be provided at that time, so nurses witnessed the loss of spiritual vitality of patients [bib_ref] Nurses perception of ethical challenges in caring for patients with covid-19: A..., Rezaee [/bib_ref].
Due to the chaotic situation characteristic of a pandemic period, confusion, denial of reality and indecision were perceived by nurses who cared for patients with COVID-19 [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] , facts that can affect the nursing care. Nurses experienced role confusion as some of them had to perform practices that doctors should do [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] or provide the support that should be given by a psychologist [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref]. The low sense of responsibility in the nursing units and the insufficient assistance to the emergency can result in problems related to professional ethics [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref]. Nurses reported that some doctors were monitoring patients by video or telephone, avoiding entering the wards [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref]. In addition, the fear of becoming infected can make health professionals act slower than usual while providing care to patients or entering COVID-19 wards [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref].
Even if nurses were afraid of caring for patients with COVID-19, they were aware that it was their duty as nurses to care for patients regardless of their illnesses [bib_ref] Nurses' ethics in the care of patients during the COVID-19 pandemic, Alloubani [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] [bib_ref] Experiences of nurses caring for patients with COVID-19 in Turkey: A phenomenological..., Muz [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] ; they were concerned with providing safe care to patients updating themselves [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] , and they were reminding 10
[formula] Abbasinia et al. (2021) Y Y Y Y Y U Y Y Y Y 19 Alloubani et al. (2021) Y Y Y Y Y Y Y Y Y Y 20 Jia et al. (2021) Y Y Y U Y N Y Y Y Y 17 Karaca and Aydin Ozkan (2021) Y Y Y Y Y N Y Y Y Y 18 Kelley et al. (2021) Y Y Y Y Y U Y Y Y Y 19 Kwon and Choi (2021) Y Y Y Y Y N Y Y Y Y 18 Liu et al. (2021) Y Y Y Y Y U Y Y Y Y 19 McMillan et al. (2021) Y Y Y Y Y Y Y Y Y Y 20 Moghaddam-Tabrizi and Sodeify (2021) Y Y Y Y Y U Y Y Y Y 19 Mohammadi et al. (2021) Y Y Y Y Y N Y Y Y Y 18 Muz and Erdogan Yuce (2021) Y Y Y Y Y N Y Y Y Y 18 Rezaee et al. (2020) Y Y Y Y Y N Y Y Y Y 18 Silverman et al. (2021) Y Y Y Y Y Y Y Y Y Y 20 Stenlund and Strandberg (2021) Y Y Y Y Y N Y Y Y YY = 'Yes' (2 points); U = 'Unclear' (1 point); N = 'No' (0 points). [/formula]
## T a b l e 2 reviewed articles and summary of results
Author (
## | awareness of need for autonomy
The studies identified that in some situations, the patient's autonomy and self-determination were not maintained, which led to the emergence of ethical dilemmas in nurses. Patients' rights were neglected, and some patients could not even choose their treatment or care plans because they could not communicate [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] and were forced to give consent as they had no other options [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref].
Some studies pointed out that patients' opinions, perspectives, values and beliefs were not taken into account as visits restrictions were mandatory [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] [bib_ref] Visitor restrictions, palliative care, and epistemic agency: A qualitative study of nurses..., Mcmillan [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref].
Nurses emphasized that because visits were not allowed, patients could not choose to be close to a loved one at the time of death, and many families were unable to see the patient for the last time [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref] [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref].
Some issues addressed in the studies imply both patients' autonomy, concerning the right to receive information about his/hers health status and treatment, and nurses' duty to act with professionalism.
Nurses reported that information such as oxygen saturation was omitted to protect the patient's mental health [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref]. Relatives did not have full access to information about patients as communication was done by telephone, which limited the interaction between nurses and family members [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref]. Decisions on not to resuscitate elderly patients were made without the patient or family being able to choose [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] , and some patients and family members were deciding about maintaining treatments that, according to nurses, were useless [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref].
Nurses faced ethical dilemmas due to violation of patient privacy and dignity; as there were too many patients to care for in the same unit, patient privacy was neglected [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref] [bib_ref] Intensive care nurses experiences of Covid-19 care: A practical and ethical challenge-A..., Stenlund [/bib_ref]. Nurses also pointed out as an ethical dilemma, the importance of keeping individual patient confidentiality and providing the necessary information to the authorities Aydin Ozkan, 2021).
## | challenges to justice
Nurses need to be impartial and fair when giving care. The scientific literature reported many challenges faced by nurses regarding justice during the COVID-19 pandemic. Nurses faced dilemmas especially in the first days of the pandemic due to limited medical resources [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] , which makes it difficult to provide equal care. Nurses also reported the difficulty of choosing among patients who needed more care [bib_ref] Intensive care nurses ethical challenges caring for people with covid-19: A qualitative..., Karaca [/bib_ref]. Young patients with wives and children waiting for them desired to live too strong; on the other hand, older patients would refuse the treatment, but nurses knew that all patients should receive care [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref].
It is difficult to provide equal care when patients are admitted according to hospital capacity, not based on their medical needs [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref]. Inequalities in caring and visitor policies were identified [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref]. Nurses reported that there were so many COVID-19 patients to care for that they were unable to provide the necessary care for all of them and had to choose between patients [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref]. Unequal care was also noticed when nurses emphasized that whereas in some institutions patients did not receive any visits, in others, the rules of time and scheduling of visits varied [bib_ref] Visitor restrictions, palliative care, and epistemic agency: A qualitative study of nurses..., Mcmillan [/bib_ref].
The principle of justice was mentioned by nurses who became infected with COVID-19, according to them, while they were patients their rights were not respected and they did not receive equal treatment [bib_ref] Workplace challenges and nurses recovered from COVID-19, Mohammadi [/bib_ref]. Nurses also emphasized the lack of professionalism that generated inequality between nursing and other professions [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Workplace challenges and nurses recovered from COVID-19, Mohammadi [/bib_ref] ; nurses stayed at patients' bedsides at all times, whereas doctors avoided entering the wards [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref]. Even cleaning staff avoided entering the patients' rooms, and cleaning was often done by nurses [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref]. Nurses felt as if their lives were less important than the lives of other professionals [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] , and they emphasized that nurse managers should advocate for frontline nurses [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref]. On the other hand, some nurse participants pointed out that due to the pandemic period, the community looked at them with different eyes, and the nursing profession was valued [bib_ref] Experiences of nurses caring for patients with COVID-19 in Turkey: A phenomenological..., Muz [/bib_ref].
## | coping with ethical dilemmas
Studies identified ways to cope with ethical dilemmas. Study and discussion groups with the presence of doctors and nurses [bib_ref] Ethical dilemmas faced by frontline support nurses fighting COVID-19, Liu [/bib_ref] and planning, control, support, catharsis, focus [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] , talking to loved ones, journaling, exercising [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref] , learning specialized skills in nursing, scientific research and management were the means used by nurses to solve problems Nurses pointed out that working in such a critical period has improved interpersonal relationships at the workplace and that intra-and inter-professional relationships were important for solving their dilemmas [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] ; they tried to see the situation as an opportunity to serve the community and the profession (Moghaddam-Tabrizi & Sodeify, 2021). Nurses sought support from the government, nursing association and the community (Moghaddam-Tabrizi & Sodeify, 2021), and they also emphasized that support from administrators of health institutions [bib_ref] Nurses' ethics in the care of patients during the COVID-19 pandemic, Alloubani [/bib_ref] [bib_ref] Nurses ethical challenges caring for people with COVID-19: A qualitative study, Jia [/bib_ref] [bib_ref] Nurses perception of ethical challenges in caring for patients with covid-19: A..., Rezaee [/bib_ref] , and especially, nurse managers (Moghaddam- [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref]. Also, continuing education and counseling programmes were reported as important measures not only to afford nurses knowledge but also to provide a safe environment for patients [bib_ref] Lived experiences of nurses in the care of patients with COVID-19: A..., Moghaddam-Tabrizi [/bib_ref] [bib_ref] Experiences of nurses caring for patients with COVID-19 in Turkey: A phenomenological..., Muz [/bib_ref]. However, some nurses pointed out a lack of further organisational support [bib_ref] United States nurses experiences during the COVID-19 pandemic: A grounded theory, Kelley [/bib_ref] [bib_ref] Moral distress in nurses caring for patients with Covid-19, Silverman [/bib_ref]. The ethical principles of beneficence (do good) and nonmaleficence (do no harm) mean providing benefits to people [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref] and must be followed by nurses in the exercise of their functions. During periods of crisis, nurses often do not have much choice and have to adopt attitudes less than perfect that end up generating ethical dilemmas [bib_ref] Ethical dilemmas due to the Covid-19 pandemic, Robert [/bib_ref]. Nursing is a science and an art that, to be practised efficiently and effectively, requires specialized knowledge and skills, which, in turn, are acquired through years of study and experience [bib_ref] Blending the art and science of nursing, Vega [/bib_ref]. Nurses must be wellprepared for providing care, but with the emergence of the pandemic, many of them had to work in COVID-19 wards and use medical equipment they had never used before without receiving adequate training [bib_ref] Covid-19: Ethical challenges for nurses, Morley [/bib_ref] [bib_ref] NurSes challenges, concerns and unfair requirements during the COVID-19 outbreak, Sperling [/bib_ref]. Similar situations were described by nurses who cared for patients with Middle East Respiratory Syndrome (MERS) [bib_ref] Nurses experiences of care for patients with Middle East respiratory syndrome-coronavirus in..., Kim [/bib_ref] and Ebola [bib_ref] Health workers experiences of coping with the Ebola epidemic in sierra Leones..., Raven [/bib_ref].
# | discussion
Furthermore, fear of COVID-19 can undermine nursing care, and the fear of becoming infected is mentioned by nurses in several studies carried out during the COVID-19 pandemic [bib_ref] Fear of nurses during COVID-19 pandemic in Saudi Arabia: A cross-sectional assessment, Moussa [/bib_ref] [bib_ref] NurSes challenges, concerns and unfair requirements during the COVID-19 outbreak, Sperling [/bib_ref] ; despite being afraid of becoming ill, nurses must be aware of their obligation to care for their patients [bib_ref] A nurse's obligatIons to patients with Ebola, Casey [/bib_ref]. On the other hand, self-care is more than an ethical obligation because nurses need to be physically, mentally and socially healthy to provide quality care to patients and the community [bib_ref] Self-care as an ethical obligation for nurses, Linton [/bib_ref] [bib_ref] Enfermagem Brasileira na linha de frente contra o novo Coronavírus: Quem cuidará..., Souza E Souza [/bib_ref]. Thus, nurse managers need to understand the impact of the COVID-19 pandemic on frontline nurses and support them by providing training and promoting a safe environment for staff and patients [bib_ref] Experiences of clinical first-line nurses treating patients with COVID-19: A qualitative study, Tan [/bib_ref] [bib_ref] A qualitative study on the experiences of the first nurses assigned to..., Yildirim [/bib_ref]. Also, ethics consultants can help nurses to develop the necessary skills to deal with ethical dilemmasarising from the COVID-19 pandemic.
The rapid shift from patient-centred care to public health-centred care has also shifted the focus of ethics in nursing, and nurses had to adapt to this new reality immediately, which caused ethical problems [bib_ref] Self-care strategies in response to nurses moral injury during COVID-19 pandemic, Hossain [/bib_ref]. Nurses are trained to provide holistic patient care; however, during the COVID-19 pandemic, mainly due to the lack of human and medical resources, this care was not always possible [bib_ref] Self-care strategies in response to nurses moral injury during COVID-19 pandemic, Hossain [/bib_ref] [bib_ref] Covid-19: Ethical challenges for nurses, Morley [/bib_ref]. Rights, such as autonomy, trust, minimizing harm and proportionality must be considered when adopting strategies concerning hospital and nursing home visitors during the COVID-19 pandemic [bib_ref] Visitor restrictions during the COVID-19 pandemic: An ethical case study, Hartigan [/bib_ref].
Autonomy is a person's right to make choices based on his/her own values and beliefs [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref] so patients have the right to be fully informed about their illness and its treatment and may, therefore, accept or not medical procedures and nursing care; thus, before any health care worker can provide assistance, the patient's consent is required [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref]. The individual has a right to privacy, which is the right to self-determination in which the person has moral authority over his/hers personal characteristics [bib_ref] Protection of privacy and confidentiality as a patient right: Physicians' and nurses'..., Demirsoy [/bib_ref]. In crises such as the COVID-19 pandemic, patient privacy can be threatened [bib_ref] Privacy and patient confidentiality in times of Covid-19, Shekhawat [/bib_ref].
Due to the COVID-19 pandemic, measures were taken to protect the community, and the individual interests were pushed; thus, nurses faced ethical problems because of such changes [bib_ref] Where does patient autonomy live, Fontes [/bib_ref]. It is important to highlight that the healthy nursepatient relationship is pivotal for the success of caring, and patients must trust in nurses; ethical problems experienced during the COVID-19 pandemic should not harm the long history of trust between nurses and patients [bib_ref] Covid-19: Ethical challenges for nurses, Morley [/bib_ref]. Therefore, nurses must be guided by leaders who are transparent in their communication, thus being role models to support the nursing team with adequate knowledge and judgement for ethical decision-making [bib_ref] Cultivating ethical leadership in the recovery of COVID-19, Markey [/bib_ref].
According to the ethical principle of justice, nursing care must be fair, equitable and adequate [bib_ref] Principles of clinical ethics and their application to practice, Varkey [/bib_ref] [bib_ref] Between a rock and a hard place: Ethics, nurses safety, and the..., Mulaudzi [/bib_ref] ; this decision becomes even more difficult if the distribution of PPE is not equal [bib_ref] Challenges experienced by ICU nurses throughout the provision of care for COVID-19..., Moradi [/bib_ref]. Nurses in risk groups or who do not feel safe due to lack of necessary PPE should not care for patients with COVID-19
(ANA, 2020).
Nurses' professional values [bib_ref] NurSes challenges, concerns and unfair requirements during the COVID-19 outbreak, Sperling [/bib_ref] and moral resilience [bib_ref] Self-care strategies in response to nurses moral injury during COVID-19 pandemic, Hossain [/bib_ref] , hospitals, institutions, administrators (ANA, 2020) and especially nurse managers' support are pointed out as important factors in helping nurses resolve ethical dilemmas during the COVID-19 pandemic [bib_ref] Cultivating ethical leadership in the recovery of COVID-19, Markey [/bib_ref]. The approach of nurse managers, being role models, respecting team members and patients, thus developing a work environment supported by ethical principles, is of fundamental importance for nurses to be able to make appropriate decisions and resolve their ethical dilemmas [bib_ref] Cultivating ethical leadership in the recovery of COVID-19, Markey [/bib_ref] Zhou & Zhang, 2021).
# | limitations
Using a search approach that specifically looked for the term 'ethical dilemma' may have overlooked sources discussing ethical conflicts without using this exact descriptor. In addition, only studies published in English, whose full texts were available on the Internet, were included; thus, relevant studies may have been excluded. It is important to emphasize that because the COVID-19 pandemic remains ongoing, different nursing ethical dilemmas may be noticed in the future.
# | conclusions
Nurses are facing huge ethical dilemmas during the COVID-19 pandemic, mainly because in providing the care, they often put their own health and that of their families at risk. Ethical nursing dilemmas regarding beneficence, nonmaleficence, autonomy and justice were identified in this literature review. The reasons for these ethical problems are usually related to the lack of PPE, shortages of medical supplies and personnel and the uncertainties that permeate an environment threatened by a new and highly contagious disease such as COVID-19.
Knowing about the various ethical dilemmas faced by nurses during the COVID-19 pandemic provides information for support programmes to be developed in health institutions to minimize the problems faced by these professionals whose roles are fundamental for the control of the pandemic. Further research carried out using broader search criteria should be done to identify and describe additional sources of nursing ethical conflicts during the pandemic.
## | implications for nursing management
This review provides information that can inspire nurse managers working during the COVID-19 pandemic to support and empower nurses to act in accordance with ethical principles, which is important in order for nurses to protect themselves while providing efficient and effective care. It is known that nurses need, in addition to training and knowledge, to feel safe to provide quality care. The nurse manager must develop an appropriate and secure working environment in which nurses are well supported to make decisions based on ethical principles.
The support of nurse managers is widely cited by nurses as being fundamental for solving problems during routine nursing practices; in cases of crisis such as the COVID-19 pandemic, the role of the nurse manager as a model for the nursing team becomes even more important.
## Acknowledgement
This research received no specific grant from any funding agency in the public commercial or not-for-profit sector.
## Conflict of interest
The author declares that there is no conflict of interest with respect to the research, authorship and/or publication of this article.
# Ethical approval
No ethical approval was required for this integrative review.
# Data availability statement
Data sharing is not applicable to this article as no new data were created or analysed in this study.
## Orcid
Ana Luiza Ferreira Aydogdu https://orcid.org/0000-0002-0411-0886
[fig] F: patients and the community(Alloubani et al., 2021;Kelley et al., 2021;Liu et al., 2021;Muz & Erdogan Yuce, 2021;Silverman et al., 2021).Nurses faced clinical dilemmas due to fighting an unknown virus and, consequently, lack of information about the illness and its treatment(Abbasinia et al., 2021;Kelley et al., 2021; Moghaddam-Tabrizi & Sodeify, 2021; Muz & Erdogan Yuce, 2021; Rezaee et al., 2020; Silverman et al., 2021). Ethical dilemmas in nursing management were identified as important information was not shared with frontline nurses in time (Kelley et al., 2021; Muz & Erdogan Yuce, 2021), equipment was not equally distributed and nurses had to work with limited PPE (Kelley et al., 2021; Liu et al., 2021; Moghaddam-Tabrizi & Sodeify, 2021). Some nurses reported a lack of time (Muz & Erdogan Yuce, 2021; Silverman et al., 2021; Stenlund & Strandberg, 2021), human resources (Liu et al., 2021; Moghaddam-Tabrizi & Sodeify, 2021; Muz & Erdogan Yuce, 2021) and medical supplies (Jia et al., 2021; Liu et al., 2021; Silverman et al., 2021; Stenlund & Strandberg, 2021) to provide the necessary care. Not being able to provide holistic patient care was another fact pointed out by nurses as a trigger for ethical dilemmas. Difficulties in providing psychological (Jia et al., 2021; Muz & Erdogan Yuce, 2021), physical (Jia et al., 2021; Karaca & Aydin Ozkan, 2021; Liu et al., 2021; Muz & Erdogan Yuce, 2021) and social care (Jia et al., 2021;Kwon & Choi, 2021; McMillan et al., 2021) and lack of (Rezaee et al., 2020), comfort care (Kelley et al., 2021; Silverman et al., 2021), end of life care (Kelley et al., 2021; McMillan et al., 2021) and family-centred care (Rezaee et al., 2020) were identified by nurses during the COVID-19 pandemic. These dilemmas emerged because nurses had a lack of knowledge and skills to work in COVID-19 wards [/fig]
[table] ,: four articles(Kelley et al., 2021;Kwon & Choi, 2021;Moghaddam- Tabrizi & Sodeify, 2021;Muz & Erdogan Yuce, 2021) had nurses' experiences in care of patients with COVID-19 as main objective, one article (Silverman et al., 2021) was about moral distress in nurses caring for patients with COVID-19, and the remaining article (Mohammadi et al., 2021) was conducted with nurses who were infected with COVID-19. The participants' perceptions of ethical dilemmas are presented in the results of all 14 articles. The results of this review are presented into four categories: beneficence-nonmaleficence; autonomy; justice; and coping with ethical dilemmas. [/table]
|
PedMap: a pediatric diseases map generated from clinical big data from Hangzhou, China
epidemiological knowledge of pediatric diseases may improve professionals' understanding of the pathophysiology of and risk factors for diseases and is also crucial for decision making related to workforce and resource planning in pediatric departments. in this study, a pediatric disease epidemiology knowledgebase called pedMap (http://pedmap.nbscn.org) was constructed from the clinical data from 5 447 202 outpatient visits of 2 189 868 unique patients at a children's hospital (Hangzhou, China) from 2013 to 2016. The top 100 most-reported pediatric diseases were identified and visualized. These common pediatric diseases were clustered into 4 age groups and 4 seasons. The prevalence, age distribution and co-occurrence diseases for each disease were also visualized. furthermore, an online prediction tool based on Gaussian regression models was developed to predict pediatric disease incidence based on weather information. PedMap is the first comprehensive epidemiological resource to show the full view of age-related, seasonal, climate-related variations in and co-occurrence patterns of pediatric diseases.Many countries experience significant shortages in pediatric specialists 1-3 . China has the largest population of children in the world. According to a report from UNICEF (the United Nations International Children's Emergency Fund), approximately 271 million children between 0 and 17 years were in China in 2015 4 . At the same time, there are only 118 000 registered pediatricians in China 5 . The shortage of pediatricians and specialized children's hospitals in China is a well-known challenge 2,6 . With such a pediatric resource shortage, comprehensive epidemiological knowledge of pediatric diseases may improve professionals' understanding of the pathophysiology and risk factors of disease and is also crucial for decision making related to workforce and resource planning in pediatric departments. The goal of this study is to identify the most common pediatric diseases and their age distributions, seasonal variations and co-occurrence patterns.Age plays an important role in pediatric medicine and is also an important factor when considering phenotypic changes in health and diseases states 7 . Some researchers have described age-related patterns of diseases 8 . However, there is no report on how to classify common pediatric diseases into age groups thus far. The recognition of the seasonality of disease occurrence has been longstanding, dating at least to the time of Hippocrates (∼380 BC). Seasonal variations in pediatric diseases have has been well described in many infectious diseases (e.g., HFMD 9 , pneumonia and gastroenteritis 10 ), several chronic diseases (e.g., asthma 11 and diabetes 12 ) and other diseases (e.g., dermatoses 13 and epistaxis 14 ). However, the seasonality of diseases is not always consistent around the world, given the diverse meteorological environment and developing/developed statuses 15 ; some studies have even shown conflicting results and age-related differentiation 14 . Seasons are divisions of the year marked by weather changes. Therefore, the seasonality of diseases is a comprehensive reflection of correlations among various meteorological features and diseases. Local long-term meteorological data and large cohort studies always have more accuracy in revealing these patterns than other types of studies. Several studies on co-occurrence diseases focus on depression 16 , diseases of immune dysfunction 17 and chronic diseases among older adults 18 . Recently, an eHDN (epidemiological human disease network) was published to show the interconnections among human diseases 19 . However, a network that shows the relationships among pediatric diseases is still not available.Traditional epidemiological studies often require long-term data collection efforts. The accumulated clinical practice big data in recent years provides an opportunity to glean new actionable knowledge, especially
knowledge concerning children's health [bib_ref] The inevitable application of big data to health care, Murdoch [/bib_ref] [bib_ref] Data Science for Child Health, Bennett [/bib_ref]. To explore the age distribution and seasonality and co-occurrence patterns in pediatrics, we analyzed all first outpatient visits (n = 5 447 202) to our children's hospital over a . Based on data mining, clustering and association analysis with meteorological data, an epidemiological knowledgebase of pediatric diseases called PedMap was constructed and published at http:// pedmap.nbscn.org.
## Methods pediatric disease data. this retrospective study was approved by the institutional review board/ethics
Committee of the Children's Hospital of Zhejiang University School of Medicine (Hangzhou, China). All research was performed in accordance with relevant guidelines and regulations. Written informed consent was waived by the Institutional Review Board/Ethics Committee, as the utilization of anonymized retrospective data does not require patient consent under the local legislation. According to a government statistical report, at the end of 2018, there were about 12.2 million children (<14 years old) in Hangzhou, the capital and most populous city of Zhejiang Province in East China with an area of 34,585 km 2 . Given its status as the center of children's healthcare in Zhejiang Province and the National Clinical Research Center for Child Health in China, the number of daily outpatient visits in this hospital often exceeds 10 000 visits. Because patients do not require a referral to visit children's hospitals as outpatients in China, the disease spectrum of outpatients is an adequate reflection of the disease spectrum of the population of children in this area. We collected the diagnosis of all patients who first visited our children's hospital for a certain condition between January 1, 2013, and December 31, 2016. During this period, there were 5 447 202 outpatient visits involving 2 189 868 unique patients with 6 433 different diagnosis terms. We selected the 100 most common pediatric diseases based on the monthly average incidence of diseases. Then, we manually coded the 100 most common pediatric diseases with ICD-10 (the 10 th revision of the International Statistical Classification of Diseases and Related Health Problems) for further analysis. As physicians use different levels of diagnosis granularity in practice, some codes can be very similar, such as those for diarrhea (K52.916) and acute enteritis (K52.904), but most diagnoses have with only 4-length codes, such as pneumonia (J18.9). If two diagnoses are coded with the same ICD-10 code, the incidence data of the two diagnoses will be combined. The ages in months when the patients visited were also collected for each outpatient visit. Then, the means, standard deviations and median ages for the 100 most common pediatric diseases were calculated. For each disease, the histograms for age in intervals of 1 month were also generated by normalizing the distribution with the maximum value. The disease incidences for each disease and each month was normalized as a z-scores, which is the signed fractional numbers of standard deviations by which the incidence is above or below the overall mean across the 48 months.
Age and seasonal pattern recognition. Both the normalized age histograms over 216 months and incidence z-scores over 48 months are type of longitudinal data. We used the kml package in R (version 3.4.0) to cluster the longitudinal data based on k-means clustering 22 , as it is not possible to know beforehand the optimum number of clusters. The kml package provides a graphical interface for choosing the 'best' number of clusters. The age group and seasonal knowledge were also used to choose the best number of patterns. More details about the data preprocessing procedure and clustering of the longitudinal data are described in the supplemental methods section.
climate data and regression model. and amount of rainfall (mm). The monthly averages of these climate features were calculated and combined with the normalized incidence of each disease to support the Pearson correlation analysis and Gaussian regression analysis using the cor() and glm() R functions, respectively. Because there is always a time lag between weather changes and patient visits, we also created weekly climate features and disease incidence datasets to perform a cross-correlation analysis and identify the lag relationships among them 23 .
Co-occurrence odds ratios and co-occurrence network. Among the top 100 common pediatric diseases, the odds ratio (OR) of one disease cooccurring with another disease, which means that the same patient visited the center for both of these two diseases during the study period, but not necessarity in the same visit, were calculated by performing Fisher's exact test in R (version 3.4.0). Diseases with a log(OR) greater than 2 were used to construct a pediatric disease co-occurrence network.
other methods used in this study. The distances between pairs of ICD-10 codes of the 100 most common pediatric diseases were calculated to build a distance matrix (details about this calculation are described in the supplemental methods section). Then, this distance matrix was mapped to 2D space using multidimensional scaling (cmdscale in R version 3.4.0) 24 .
# Results
Top 100 most common pediatric diseases. Based on the average incidence per month during the four years of the study, the top 100 most common pediatric diseases were selected. These diseases accounted for more than 78% of all the outpatient visits. As shown in Supplemental, the top 100 diseases were plotted by size depending on the average incidence per month and were colored by disease categories in the ICD-10 coding system. The most enhanced ICD-10 category is J (diseases of the respiratory system), and acute upper respiratory infection is the top disease, occurring approximately 20 240 times per month on average. In total, 19 diseases of the respiratory system, such as bronchitis, pneumonia and acute pharyngitis, are included in the top 100 most www.nature.com/scientificreports www.nature.com/scientificreports/ common pediatric diseases. The next most detailed ICD-10 category is R (symptoms, signs and abnormal clinical and laboratory findings), which includes the most common symptoms in children, such as fever, abdominal pain, cough and vomiting. The third popular category is K (diseases of the digestive system), such as dyspepsia and enteritis. In addition, the top 100 most common pediatric diseases covered most of the ICD-10 categories, such as L30 (dermatitis), N47 (diseases of the genitourinary system, especially phimosis), F90-98 (behavioral and emotional disorders with onset usually occurring in childhood and adolescence, especially developmental disorders and hyperkinetic disorders), E (endocrine, nutritional and metabolic diseases, such as hypothyroidism and disorders of puberty), H (diseases of the eye and ear, such as conjunctivitis, ametropia and otitis media), B (viral infections, especially herpangina), Q (congenital malformations and deformations, such as ankyloglossia and concealed penises), G (diseases of the nervous system, such as epilepsy), P07 (preterm infants), S09 (injuries of the head), D18 (hemangiomas) and Z01 (dental examinations). However, the distribution of incidence among the diseases is not balanced: the top 10 diseases account for more than 41% of all the outpatient visits and more than half of the outpatient visits for the top 100 diseases. A summary of the seasonal and age patterns of common pediatric diseases is shown in Supplemental and is explained in the following sections.
The age distribution of the top 100 most common pediatric diseases. The age distribution of all outpatient visits is shown in Supplemental [fig_ref] Figure 2: Seasonality of the top 100 pediatric diseases [/fig_ref]. Overall, there is a downward trend with increasing age, but there are two peaks at approximately 6 months and 40 months. The means and standard deviations of age for the top 100 pediatric diseases were calculated and shown in Supplemental [fig_ref] Figure 3: Correlation of weather features with disease incidence [/fig_ref]. Several diseases affect only small infants, such as P59.9 (neonatal jaundice), J21.9 (pediatric bronchiolitis) and H04.4 (chronic dacryocystitis). There are also several diseases that affect only older children, especially diseases from ICD-10 category E (endocrine, nutritional and metabolic diseases), such as E66.9 (obesity) and E34.3 (short stature). The mean age for most diseases ranges from 20 to 80 months, and the mean age for high-incidence pediatric diseases is approximately 3 years old. The standard deviation ranged from 4.556 to 64.242 months and showed very different age distributions for different diseases. The standard deviation is positively correlated with the mean age of onset for the disease (R = 0.485), meaning that some of the diseases affect only certain children at younger ages. the age patterns for common pediatric diseases. The normalized age distributions of the top 100 diseases are shown in Supplemental [fig_ref] Figure 4: The pediatric disease co-occurrence network [/fig_ref]. At first glance, there are different types of distributions. We used the kml package to cluster these 100 age distributions using a k value ranging from 2 to 10 and reran the model 5 times with different initial conditions. The theoretically "best" k value for clustering was 2 for imbalanced distributions, as shown in Supplemental . However, we chose to use k = 4 for clustering because it could provide more detailed clustering results and relatively good clustering quality (based on the Calinski & Harabatz index [bib_ref] A Dendrite Method Foe Cluster Analysis, Caliñski [/bib_ref].
As shown in, Cluster A contains 32 diseases that mostly affect young infants, and Cluster B contains 30 diseases that mainly affect infants who are approximately 1 year old. The diseases in Clusters C and D, which have wider peaks, primarily affect the preschool children and schoolchildren, respectively. There are noticeable spikes of the mean trajectories because some of the birthdate registered to the system do not have month granularity and are set to the year.
The distribution of different disease categories in the 4 age clusters is shown in. Cluster A (young infants) is characterized by conditions originating in the perinatal period, congenital malformations and diseases of the circulation system. Cluster B (infants) includes injuries, skin diseases, and digestive system diseases. Cluster C (preschool children) was mainly affected by genitourinary and respiratory diseases. Cluster D (schoolchildren) was primarily affected by mental and behavioral disorders and endocrine, nutritional and metabolic diseases. The incidence of disease in Cluster B (infants) was the highest among the 4 clusters.
The season features of 100 common pediatric diseases. As shown in [fig_ref] Figure 2: Seasonality of the top 100 pediatric diseases [/fig_ref] , the seasonal variation in pediatric diseases is clear, and some of the diseases are well known to pediatricians in practice. To identify the four seasons associated with pediatric diseases, k = 4 was also used to cluster the 100 common pediatric diseases, as shown in [fig_ref] Figure 2: Seasonality of the top 100 pediatric diseases [/fig_ref]. Cluster A, which contains 31 diseases, shows a spring peak, and Cluster C, which contains 26 diseases, shows a summer peak. Only 14 diseases included in Cluster D were common in autumn. The most common diseases in the winter were classified into Cluster B. The distribution of different disease categories in the 4 seasonal patterns is shown in [fig_ref] Figure 2: Seasonality of the top 100 pediatric diseases [/fig_ref]. This figure shows that the most common viral infections, such as tonsillitis, pharyngitis and herpangina, occurred in the spring (Cluster A), and diseases of the circulatory, nervous, and genitourinary systems and endocrine, nutritional and metabolic diseases are common in the summer (Cluster C). In the autumn (Cluster D), diseases of the digestive system, such as dyspepsia, diarrhea and abdominal pain, were frequent. Diseases in the winter (Cluster B), which have a high average incidence, included diseases of the respiratory system, such as AURIs, bronchitis, fever and pneumonia. the associations among disease incidence and weather features. Seasons are divisions of the year marked by changes in the weather, ecology, and the amount of daylight. Therefore, the seasonality of pediatric diseases is a comprehensive reflection of correlations among various meteorological features and diseases. The univariate correlations among monthly weather features and disease incidence were calculated and plotted as a heatmap in [fig_ref] Figure 3: Correlation of weather features with disease incidence [/fig_ref].
Many diseases were highly correlated with weather features (the light red and green represent the strong negative and positive correlation, the detail of the correlation coefficient and corresponding p-value were shown in Supplemental , especially temperature and temperature-related features (dew point and sea-level pressure). The diseases of the respiratory systems were mainly clustered on the left side of the heatmap, which indicates a negative correlation with temperature and humidity. The amount of precipitation and visibility in the summer are always higher than those in the other seasons. Therefore, they also show correlation patterns similar to those related to temperature. The correlations among humidity, wind and disease incidence were relatively weak, and only a few diseases, such as papular urticaria and eczema, are correlated with high humidity and low humidity, respectively.
The temporal relationships among different weather features and the incidence of common pediatric diseases were analyzed based on cross-correlation analysis across ±20 weeks. The cross-correlation absolute R values larger than 0.5 or the top 10 correlated diseases of a weather feature are shown in [fig_ref] Figure 3: Correlation of weather features with disease incidence [/fig_ref]. As the dew point and sea-level pressure are highly correlated with temperature, diseases correlated with these characteristics were ignored in [fig_ref] Figure 3: Correlation of weather features with disease incidence [/fig_ref]. Diseases of the respiratory system, such as pediatric bronchiolitis (J21.9), pneumonia (J18.9), bronchitis (J40.x00) and asthmatic bronchitis (J45.9), showed a negative correlation with temperature, and the incidence peaked slightly before the temperature valley (the lowest point of temperature). The peak for vomiting (R11.x) usually occurred 6 weeks before the temperature decreased to the valley. The diseases that were positively correlated with temperature are more widely distributed and diverse, including epistaxis (R04.0), thrush (B37.9), hydroceles (N43.3) and some summer-vacation-related patient visits, such as those for precocity (E30.1), obesity (E66.9), phimosis (N47.x00) and growth retardation (M89.2). Some injuries were also correlated with temperature due to less protection by clothes and more outdoor activities in the summer, such as skin traumas (T14.0), fractures (M48.x) and head injuries (S09.9). Dyspepsia (K30.x) usually occurs 7 weeks after the temperature www.nature.com/scientificreports www.nature.com/scientificreports/ diseases were correlated with multiple weather features, such as enteritis (K52.9), which was correlated with temperature, visibility, humidity and even wind speed at different temporal lags. Several long time-lag correlations also emerged but have no discernible clinical meaning.
Regression model. Based on the finding that most pediatric diseases were correlated with seasons and various weather factors, a regression analysis that included the temperature, dew point, sea-level pressure, humidity, visibility, wind speed and amount of rainfall was conducted for 200 diseases (we extend the analysis to more diseases in this step). All of these regression models using the monthly average climate feature as predictors can output the monthly disease incidence z-score. The top 10 best regression models, which were evaluated based on the residual deviance, are shown in Supplemental . The prediction value fit the true incidence very well in these diseases.
The pediatric disease co-occurrence network. A pediatric disease co-occurrence network was generated as shown in [fig_ref] Figure 4: The pediatric disease co-occurrence network [/fig_ref]. There are some well-known co-occurrence patterns, such as obesity with concealed www.nature.com/scientificreports www.nature.com/scientificreports/ penises, hydroceles and inguinal hernias, Zinc deficiency and anorexia. However, there are also many interesting novel connections between diseases, such as between tic disorders and conjunctivitis and between dental caries and ametropia. The most closely related diseases involved allergic conjunctivitis, which is connected to many other allergic diseases, such as asthma, allergic rhinitis and inflammation of different parts of the body. It has been suggested that the progression of clinical signs of allergic eczema, asthma, and allergic rhinitis follow each other in time and may cooccur [bib_ref] Allergic Rhinitis and its Impact on Asthma (ARIA) guidelines-2016 revision, Brożek [/bib_ref]. This comorbidity has been attributed to a common mechanism of altered immunologic mechanisms favoring a systemic response of type 2 helper T-cell cytokines to environmental allergens. Such a disease co-occurrence network will inspire researchers to explore disease relationships and their common pathogenesis. For this reason, we also published a low confidence (log(OR) >1.5) pediatric disease co-occurrence network on PedMap, which will be described below.
pedMap. Using the results described above, an online pediatric disease map called PedMap was developed and published at http://pedmap.nbscn.org. On PedMap, users select different map styles to explore the age-related, seasonal and climate-related variations in and co-occurrence network for common pediatric diseases. As PedMap was based on a dynamic chart that supports the zooming in and zooming out, users can click on a disease in the map to open a new webpage that provides detailed information about the disease, which includes basic incidence information and incidence z-score curves of the four years accompanied by corresponding weather data, age distributions and comorbidities. The details of PedMap are described in the Supplemental materials "Introduction to PedMap".
The coefficients of the regression model of the top 100 pediatric diseases were used to develop an online incidence prediction tool that accepts customized climate information to predict the incidence z-score of common pediatric diseases. As different prediction models are characterized by different goodness of fit values, we use the k-means clustering to classify three different prediction confidence levels [bib_ref] A k-means clustering algorithm, Hartigan [/bib_ref]. The confidence level was labeled with the disease name in the online tools.
# Discussion
PedMap provides an overview of common pediatric diseases, as shown in Supplemental . Which clearly shows which diseases will affect children at different stages and in different seasons. The diseases that affect younger children mainly occur in the spring and winter, and diseases that affect older children mostly occur in the summer. The autumn is the safest season for children of all ages. Furthermore, a pediatric disease co-occurrence network was first generated to cover the most common pediatric diseases. Therefore, this knowledge resource and prediction tool could improve the accuracy of the workforce and resource planning of the pediatric department. In China, the government has required medical institutions at all levels to open isolated enteric disease clinics from May to October each year to control enteric infections, such as cholera. However, in this study, the data show that the incidence peak of diarrhea always occurs after October (Supplemental . Diarrhea is also classified as an autumn disease instead of a summer disease, as is traditionally assumed [bib_ref] Seasonality of clustering of fever and diarrhea in Beijing, Li [/bib_ref]. Therefore, it would be wise to keep the enteric disease clinic at the local children's hospital open until December.
The pediatric disease co-occurrence network contains many connections that have yet been well explained. This network will inspire researchers to further explore the relationships between pediatric diseases and their common pathogenesis. The temporal pattern of co-occurrence will provide more information about the occurrence, development and outcome of diseases.
As PedMap was built based on a local dataset, its applicability to other regions and countries is unknown. Furthermore, in addition to real diseases and climate associations, many social behaviors are factors influencing outpatient visits. For example, summer vacation was the best times to treat phimosis and other chronic conditions, such as obesity, enuresis and short stature in schoolchildren. However, our prediction model did not consider these social factors. Finally, the disease data were obtained from the outpatient department and do not include lower-severity conditions that may not lead to individuals to seek care.
conclusion PedMap provides a convenient and comprehensive view of age-related, seasonal and climate-related variations in and co-occurrence of common pediatric diseases. The prediction tool can help with decision making related to workforce and resource planning in pediatric departments based on weather forecasts. The visualization of epidemiological knowledge of pediatric diseases also helps in the development of evidence-based children's healthcare programs.
## Data availability
All data generated or analyzed during this study are included in this published article and supplementary information files and on the PedMap website (http://pedmpa.nbscn.org).
[fig] Figure 1: (A) The clustering results of the age distributions for 100 common pediatric diseases. The top section shows the original age distribution trajectories of the 100 diseases. The lower section shows the trajectories of the means of the 4 clusters, which are based on k-means clustering for longitudinal data: Cluster A (young infants), Cluster B (infant), Cluster C (preschool children) and Cluster D (schoolchildren). (B) The disease distribution in the 4 age clusters. The heatmap shows the percentage of different disease categories distributed in the 4 age clusters as shown in (A). The bar chart on the top shows the incidence for these age clusters. (2019) 9:17867 | https://doi.org/10.1038/s41598-019-54439-wwww.nature.com/scientificreports www.nature.com/scientificreports/ peak. Skin diseases, such as dermatitis (L30.900), eczema (L30.902) and urticaria (L50.9), are weather-sensitive but have different features: Dermatitis (L30.900) is positively correlated with temperature at lag 5, but eczema (L30.902) is negatively correlated with temperature at lag −1. Urticaria (L50.9) was positively correlated with humidity but negatively correlated with visibility and wind speed. Correlations among pediatric diseases and other weather features, such as humidity, visibility, wind speed and amount of rainfall, were relatively weak. Only some correlations exceeded 0.5, such as those for asthmatic bronchitis (J45.9) and adenoid vegetation (J35.2), which were negatively correlated with visibility. Epistaxis (R04.0) was positively correlated with visibility. Some [/fig]
[fig] Figure 2: Seasonality of the top 100 pediatric diseases. (A) The heatmap of the normalized incidence of the 100 most common pediatric diseases across 4 years in different months. (High incidence z-scores is green, and low incidence z-score are red). (B) The clustering results of the seasonal pattern of the 100 most common pediatric diseases. The top section shows the original disease incidence trajectories of the 100 diseases over 48 months. The lower section shows the trajectories of the means of the 4 clusters based on k-means clustering for longitudinal data: Cluster A (spring), Cluster B (winter), Cluster C (summer), Cluster D (autumn). (C) The heatmap shows the percentage of different disease categories distributed in season clusters as shown in (B). The bar chart on the top shows the incidence for these season clusters. Scientific RepoRtS | (2019) 9:17867 | https://doi.org/10.1038/s41598-019-54439-w [/fig]
[fig] Figure 3: Correlation of weather features with disease incidence. (A) The heatmap of the correlation between weather features and the incidence of the 100 most common pediatric diseases. Red and green indicate negative and positive correlation, respectively, and the black indicates the no correlations. (B) The cross-correlation analysis identified the correlation coefficients and time lags between weather features and disease incidence. Scientific RepoRtS | (2019) 9:17867 | https://doi.org/10.1038/s41598-019-54439-w [/fig]
[fig] Figure 4: The pediatric disease co-occurrence network. The connected nodes show a cooccurring disease relationships with a log(OR) greater than 2. [/fig]
|
Comparison of metaheuristics to measure gene effects on phylogenetic supports and topologies
Background: A huge and continuous increase in the number of completely sequenced chloroplast genomes, available for evolutionary and functional studies in plants, has been observed during the past years. Consequently, it appears possible to build large-scale phylogenetic trees of plant species. However, building such a tree that is well-supported can be a difficult task, even when a subset of close plant species is considered. Usually, the difficulty raises from a few core genes disturbing the phylogenetic information, due for example from problems of homoplasy. Fortunately, a reliable phylogenetic tree can be obtained once these problematic genes are identified and removed from the analysis.Therefore, in this paper we address the problem of finding the largest subset of core genomes which allows to build the best supported tree. Results: As an exhaustive study of all core genes combination is untractable in practice, since the combinatorics of the situation made it computationally infeasible, we investigate three well-known metaheuristics to solve this optimization problem. More precisely, we design and compare distributed approaches using genetic algorithm, particle swarm optimization, and simulated annealing. The latter approach is a new contribution and therefore is described in details, whereas the two former ones have been already studied in previous works. They have been designed de novo in a new platform, and new experiments have been achieved on a larger set of chloroplasts, to compare together these three metaheuristics.Conclusions:The ways genes affect both tree topology and supports are assessed using statistical tools like Lasso or dummy logistic regression, in an hybrid approach of the genetic algorithm. By doing so, we are able to provide the most supported trees based on the largest subsets of core genes.
# Background
These last years the investigation of the evolutionary relationship between different plants has benefited from the multiplication of the available chloroplast sequences. Indeed, thanks to various tools it is possible to process *Correspondence: [email protected] 1 FEMTO-ST Institute, UMR 6174 CNRS, DISC Computer Science Department, Univ. Bourgogne Franche-Comté (UBFC), [bib_ref] Particle swarm optimization: developments, applications and resources, Eberhart [/bib_ref] Route de Gray, 25000 Besançon, France Full list of author information is available at the end of the article these sequences in order to build a phylogenetic tree that accurately characterizes the evolutionary lineages among the chloroplasts. Efficient coding sequence prediction and annotation tools have been developed to deal specifically with chloroplasts, for example DOGMA [bib_ref] Automatic annotation of organellar genomes with dogma, Wyman [/bib_ref] , and there is also a great choice for the alignment of sequences. Moreover, given a set of sequences or characters, many wellestablished bioinformatics programs based on Bayesian inference or maximum likelihood, like BEAST or RAxML [bib_ref] Raxml version 8: A tool for phylogenetic analysis and post-analysis of large..., Stamatakis [/bib_ref] , can be used to reconstruct a phylogenetic tree. The objective is usually to obtain a reliable phylogeny on which biological investigations can be applied. For instance, in, a comparative analysis of expressed rice gene homologues in 48 other diverse eukaryotic species has been performed, and a phylogenetic tree of life based on 98 of these genes conserved across the species has been computed using such software. It has been used to estimate more accurately the divergence time among a large number of species pairs. However, the genome of rice plant has approximately 40,000 genes, and their tree has been computed using less than 1% of such genes.
Several methods can be used to estimate the robustness of the produced tree, the most widely used are the bootstrap and the decay (or Bremer) analyses [bib_ref] Raxml-iii: a fast program for maximum likelihood-based inference of large phylogenetic trees, Stamatakis [/bib_ref] [bib_ref] Mrbayes 3: Bayesian phylogenetic inference under mixed models, Ronquist [/bib_ref]. Obviously, a first condition to be able to build a phylogenetic tree for a given set of close plant species is to identify as precisely as possible the corresponding core genome(the set of genes in common). However, even if the core genome is large and accurate, the resulting phylogeny is not necessarily well-supported [bib_ref] Finding optimal finite biological sequences over finite alphabets: the optifin toolbox, Garnier [/bib_ref]. In fact, the core genome genes are not constrained through evolution in a similar way. On the one hand some evolve under strong evolutionary constraints and thus reflect the story of the species while, on the other hand, other genes evolve more freely due to a lower role in the survival and adaptability of a species. The latter tell their own history and thus disturb the phylogenetic information. Furthermore, the way the robustness and accuracy of the obtained phylogenetic tree are altered by the amount of used data for the reconstruction process is not completed understood. Nevertheless, if we consider a set of species reduced to lists of gene sequences, an obvious dependence between the chosen subset of sequences and the obtained tree (topology, branch length, and/or robustness) can be observed. This dependence is usually regarded by the mean of gene trees merged in a phylogenetic network. In fact, phylogenetic networks are necessary to represent events like horizontal gene transfers, but statistical methods to infer such networks are still limited and under development.
In this article, we consider the situation from a dual point of view, that consists in starting with the complete core genome and then to remove the genes responsible for inconsistent phylogenetic signal. In other words, the objective is to find the largest part of the core genome that produces a phylogenetic tree as supported as possible, and which therefore gives the fairest view of the relationships between most of the sequences under consideration. Searching the problematic genes by exhaustively testing the combinations of core genome genes is nonsense due their huge number. Therefore, to speed up the finding of a satisfactory combination we rather consider metaheuristics. The first one, introduced in a previous work [bib_ref] Hybrid genetic algorithm and lasso test approach for inferring well supported phylogenetic..., Alkindy [/bib_ref] , is an ad hoc Genetic Algorithm (GA) which in some cases is not able to converge towards a suitable solution. Next, a Binary Particle Swarm Optimization (PSO) approach has been published in the the CIBB proceedings book [bib_ref] Well-supported phylogenies using largest subsets of core-genes by discrete particle swarm optimization, Alsrraj [/bib_ref]. Finally, in this article, which extends and improves the two former ones, we study the relevance of the Simulated Annealing (SA) algorithm to fulfill the optimization task. Also notice that the different metaheuristics have been executed in a distributed manner using supercomputing facilities. To sum up, the contribution of this article is threefold: first, it proposes a new simulated annealing approach, second a new version of the PSO, and third a comparison of the three metaheuristics on a large number of new groups of species. Compared to usual articles studying the tree of life like, our approach is diametrically opposed: instead of using existing phylogenetic software on a small collection of core genes, in order to provide new discoveries on some aspects of the Evolution, we propose a pipeline of 3 metaheuristics, to find the largest subset of core genes leading to the most supported tree.
# Methods
## Problem description
Let us introduce the problem of determining a phylogeny (evolution tree) for a given set of species by considering a set of chloroplast genomes that have been annotated using DOGMA [bib_ref] Automatic annotation of organellar genomes with dogma, Wyman [/bib_ref] (the approach we applied is detailed in "Results" section). To start we need to pick one or several genes on which the phylogeny will be based. Therefore we use the restricted core genome [bib_ref] Finding the core-genes of chloroplasts, Alkindy [/bib_ref] [bib_ref] Gene similarity-based approaches for determining core-genes of chloroplasts, Alkindy [/bib_ref] , which consists of conserved genes present everywhere, whose size is larger than one hundred genes when the species are close enough. Then multiple sequence alignments are performed using muscle [bib_ref] Muscle: multiple sequence alignment with high accuracy and high throughput, Edgar [/bib_ref] and finally a phylogenetic tree is inferred thanks to the maximum-likelihood tree builder RAxML [bib_ref] Raxml version 8: A tool for phylogenetic analysis and post-analysis of large..., Stamatakis [/bib_ref].
The relevance of the obtained tree is then assessed by its bootstrap values: if these ones are all above 95 the tree is well-supported, in which case we can reasonably estimate that the phylogeny of these species is solved. Bootstrapping is a random sampling technique commonly used to estimate the significance of branches of a phylogenetic tree. It consists to randomly select columns in the aligned DNA core sequences to be neglected during the tree building process and to check whether the same nodes are recovered. A large number of bootstrap repetitions, usually between 50 and 1000, are used to assess the tree reliability. As an illustration, a node which appears 95 times out of 100 by dropping a column means that the node is well-supported. Conversely, a low support value claims that a reduced part of the alignment supports the node, since by removing columns the node is reconstructed in different ways.
When such a well-supported tree is not built, but rather a tree having some branches exhibiting low supports, some genes of the core genome can be responsible of this lack of support. The objective is then to identify the most supported tree using the largest subset of core genes, a typical optimization problem. Obviously, the optimization problem we face cannot be solved by a brute force approach checking all possible combination of genes, due to the resulting combinatorial explosion. Indeed, for a core genome of n genes there would be 2 n trees to infer and that is clearly intractable in practice. To overcome such a combinatorial situation, a typical choice is to use a metaheuristic method.
In [bib_ref] Hybrid genetic algorithm and lasso test approach for inferring well supported phylogenetic..., Alkindy [/bib_ref] , we have first investigated the mixing of a genetic algorithm with Lasso tests to find problematic genes. Unfortunately, thorough and careful experimental investigations have led to results, recalled in [fig_ref] Table 1: Results of genetic algorithm approach on various families [/fig_ref] , showing that this proposal is not able to predict the phylogeny of some particular plant orders. As can be seen, the lowest bootstrap value (or bootstrap score) obtained for 15 group of species is below 95 (column b in the table). The relevance of binary particle swarm optimization to find the largest subset of core genes has been studied in [bib_ref] Well-supported phylogenies using largest subsets of core-genes by discrete particle swarm optimization, Alsrraj [/bib_ref] , producing slightly better bootstrap scores than GA with Lasso. In this paper we introduce a third well-known metaheuristic method, namely simulated annealing, and we compare the three approaches considering new sets of species. Like the two former ones, the computations with SA algorithm will be done in a distributed manner. Multiple algorithm instances will be launched using a same cooling schedule and at the end of each Markov chain, for a same temperature, a centralized communication scheme will take place.
To sum up, gives an overview of the proposed pipeline to obtain the ancestral history of a set of species.
## Phylogenetic predictions using metaheuristics genetic algorithm approach
To make this article self-contained, we summarize hereafter the main steps of the genetic algorithm combined with Lasso test proposed in [bib_ref] Hybrid genetic algorithm and lasso test approach for inferring well supported phylogenetic..., Alkindy [/bib_ref] aiming at finding problematic genes in core genome.
The n core genes are sorted alphabetically, and at each subset we associate a binary word of length n: its ith character is 1 if and only if the i-th core gene is in the considered subset. In the proposed GA, a first stage to initialize the GA population (1) computes the set of n-length binary words containing the word having only 1's (the whole core genome which is composed of n genes), (2) all words having exactly one 0 (all but 1 gene) further denoted as systematic mode, and (3) 200 words having between 2 and 10 0's randomly located. Each of these words is associated with the score b+p 2 where b is the lowest bootstrap of the reconstructed phylogenetic tree and p is the percentage of considered core genes.
More precisely, the population is initialized with the 50 best words. Then, the GA iterates until discovering a word whose score is larger than 95, or at most for 200 iterations. Each iteration, which produces a new population, consists of the following steps:
1. Repeat 5 times a random pickup of a pair (w 1 , w 2 ) of words and mix them using a crossover approach. In this step, indexes {1, . . . , n} are partitioned into k, k ≤ n 2 , subsets I 1 , . . . I k . A new word w is then defined by w i = w 1 i if i belongs to some I j where j is odd; otherwise w i = w 2 i . The obtained words are added to the population P, resulting in population P c . 2. Mutate 5 words of the population P c . More precisely, for each of these words, k randomly selected binary values of w are switched leading to a new word. The mutated words are added to P c leading to population P m . 3. Produce population P r by adding 5 new random binary words having less than 10% of 0's to P m . 4. Select the 50 best words in population P r to form the new population P.
The aforementioned GA may not produce wellsupported trees. Nevertheless, the whole set of produced words with their associated scores contains valuable information about which gene breaks supports. The idea is to focus on each topology having a frequency of occurrence larger than 10%. Then for each best word of these best topologies, and for each problematic bootstrap in its associated tree, we apply a Lasso test [bib_ref] Regression shrinkage and selection via the lasso, Tibshirani [/bib_ref] , which is recalled hereafter.
Let W be a m × n matrix where each line W i = (X i1 , . . . , X ij , . . . , X in ), 1 ≤ i ≤ m, is a word. For each W i , let Y i be the real positive support value for each problematic bootstrap b per topology and per gene. The Lasso test β = (β 1 , . . . , β i , . . . , β n ) is thus defined by:
[formula] β = argmin ⎧ ⎪ ⎨ ⎪ ⎩ m i=1 ⎛ ⎝ Y i − n j=1 β j X ij ⎞ ⎠ 2 + λ n j=1 |β j | ⎫ ⎪ ⎬ ⎪ ⎭ .(1) [/formula]
It is not hard to see that the sign of β j is positive (resp. negative) if the bootstrap support increases (resp. decreases) with respect to j.
This test allows thus to remove problematic genes, i.e., genes j, 1 ≤ j ≤ n, such that β j is negative. Finally, a last genetic algorithm phase is launched on the updated population, in order to mix these promising words.
## Binary particle swarm optimization approach
Particle Swarm Optimization [bib_ref] Particle swarm optimization, Kennedy [/bib_ref] is a stochastic metaheuristic which has been successfully applied on artificial neural network training, fuzzy system control. . . In this scheme, an emergent behavior enables individual swarm Occ provides the number of genomes within the group while taxa is for the number of species. c and b respectively correspond to the percentage of core genes and the lowest bootstrap of the solution produced by the GA, while Likelihood is the likelihood of the best tree. Finally, Terminus specifies at which stage the GA stopped, while Outgroup is the considered outgroup Overview of the proposed pipeline members, particles, to take advantage from neighboring particles, which are closer numerically to the optimal solution. In the case of the standard Binary PSO (BPSO) model [bib_ref] A novel binary particle swarm optimization, Khanesar [/bib_ref] , each particle position is a n-length vector of binary values. A score function associates a real number to such kind of vector w.r.t. the optimization problem. BPSO aims at moving the particles in the n dimensions in order to obtain the optimal position with respect to this function. More precisely, each particle i is represented by a nlength vector X i of binary values, which has the same meaning than binary words in GA, indicating the gene contents of the associated core subset. Again, the j-th coordinate of X i is 1 if and only if the associated j-th parameter is selected. A swarm of L particles is a list of position vectors (X 1 , X 2 , . . . , X L ) together with their associated velocities (V 1 , V 2 , ..., V L ). Each V i is a n-length vector of real numbers between 0 and 1. Each velocity vector V i is updated as follows:
[formula] V i (t+1) = wV i (t)+φ 1 P best i − X i +φ 2 P best g − X i , (2) [/formula]
where w, φ 1 , and φ 2 are weighted parameters setting the level of each three trends for the particle, which are respectively to continue with the same inertia, to follow the direction of its own best neighbouring position P best i , or to follow the one of the global best known solution P best g . Each position X ij of the particle i is updated as follows:
[formula] X ij (t + 1) = 1 if r ij ≤ 1 1+e −V ij (t+1) , 0 otherwise,(3) [/formula]
where the scalar r ij depends on both the particle i and the parameter j. It is not hard to recognize a choice guided by a threshold r ij and a sigmoid [bib_ref] A novel binary particle swarm optimization, Khanesar [/bib_ref] function. Let us now recall how this BPSO optimization scheme has been parameterized to solve our phylogeny problem [bib_ref] Well-supported phylogenies using largest subsets of core-genes by discrete particle swarm optimization, Alsrraj [/bib_ref].
Each vector x i ∈ {0, 1} n corresponds to a subset of core genes. It is associated with the deduced following data: the percentage p of considered core genes, the lowest bootstrap b of its induced phylogenetic tree, and, finally, the score b+p 2 . The approach was to construct phylogenies based on neighbouring core genes: this leads to trees with similar topologies and with close bootstrap values with a high probability. During BPSO initialization, the L particles are randomly distributed among subsets of core genes (binary words) with a high percentage of 1's. Further iterations move these particles in such a way that they will converge to an optimal node.
As in [bib_ref] Hybrid pso and ga for global maximization, Premalatha [/bib_ref] , at each iteration, the particle velocity is updated as in Eq. (2) where φ 1 and φ 2 are random numbers belonging to [0.1,0.5], and w linearly decreases between the first iteration and the last one from w max = 0.9 to w min = 0.4, as suggested in [bib_ref] Particle swarm optimization: developments, applications and resources, Eberhart [/bib_ref]. A large inertia weight indeed facilitates a global search, while a small inertia weight tends more to a local investigation.
A distributed version of BPSO algorithm has been proposed to minimize the execution time: each particle is executed in a worker core in order to compute its fitness value and obtained results are centralized by a supervisor master core. More precisely, the master initializes the particles of the swarm, distributes them to the workers, and waits until all of them have finished their task. It determines then the position of the particle that has the best fitness value as the global best position, and sends this information to the workers that update their respective particle velocity and position. This mechanism is repeated until a particle achieves a fitness value larger than or equal to 95 with a large set of included genes. In the following, two distributed versions of the BPSO are considered.
The former, further denoted as PSO version I, updates the velocity as follows:
[formula] V i (t + 1) = x·[ V i (t) + C 1 (P best i − X i ) + C 2 (P best g − X i )] (4) [/formula]
where x, C 1 , and C 2 are weighted parameters setting the level of each three trends for the particle. The default values of these parameters are C 1 = C 2 = 2.05, while x, which represents the constriction coefficient, is computed according to formula [bib_ref] Particle swarm optimization methods, taxonomy and applications, Sedighizadeh [/bib_ref] [bib_ref] The swarm and the queen: towards a deterministic and adaptive particle swarm..., Clerc [/bib_ref] :
[formula] x = 2 × k |2 − C − ( √ C × (C − 4))| ,( 5 ) [/formula]
where k is a random value in [0,1] and C = C 1 + C 2 , C ≥ 4. According to Clerc [bib_ref] The swarm and the queen: towards a deterministic and adaptive particle swarm..., Clerc [/bib_ref] , using a constriction coefficient results in particle convergence over time. This latter, denoted as PSO version II, updates the velocity as formalized in Eq. (2).
## A simulated annealing approach
General presentation The original Simulated Annealing (SA) method is a local search based threshold class algorithm. Basically, a threshold algorithm is a loop in which a move is either done or not, according to a given criterion and until reaching a freeze. Specifically, after an initialization step, this loop is composed by (a) a move in the neighborhood of the current solution, (b) an evaluation of this new position by a real-valued scoring function, then (c) a test, given a well chosen criterion, to store this position as the new best one. Various criteria can be considered. For instance, if a position is evaluated as a better solution than the best existing one, it becomes the reference solution for next iterations when the acceptation criterion is "only if best cost (score)" algorithm, which is a variant of a classical greedy local search. The "all is accepted" algorithm produces, for its part, a random walk. Finally, between these two extremes situations, an acceptation criterion allows to store sometimes too positions with poorer scores than the best solution, which is an upward move via a stochastic component to avoid local minima. Such a stochastic approach facilitates theoretical analysis of asymptotic convergence. As such algorithms can be successfully used for a broad range of optimization problems, SA has been largely covered in the literature during the last decades [bib_ref] Simulated annealing-an annotated bibliography, Collins [/bib_ref] , for both empirical [bib_ref] Optimization by simulated annealing: an experimental evaluation; part i, graph partitioning, Johnson [/bib_ref] [bib_ref] Optimization by simulated annealing: an experimental evaluation; part ii, graph coloring and..., Johnson [/bib_ref] -typically on NP-hard problems -and theoretical perspectives [bib_ref] A theoretical framework for simulated annealing, Romeo [/bib_ref]. In simulated annealing, the criterion is inspired by the Metropolis-Hastings statistical (Markov chain Monte Carlo) thermodynamics algorithm. SA simulates the cooling of a material in a heat bath until a steady (frozen or thermodynamic equilibrium) state. When the solid material is heated over its melting point, its solidification rate induces its structural properties. Two major antagonistic strategies are commonly used. On the one hand, after a fast cooling (quenching), the steady state is constituted by different thermodynamic free level areas. This corresponds to a local minimum for a local search, when considering energy as a score. On the other hand, after a slow cooling (annealing), almost one sole thermostatic level is expected, which corresponds to a global minimum. As feasible solutions of SA are system states, the structural proximity of the latter leads to the concept of solution neighborhood.
Thermodynamic laws show that at temperature t, the probability to increase in energy of the value δE is given by p(δE) = exp(−δE/kt) with k equal to the Boltzmann's constant. Metropolis simulations [bib_ref] Equation of state calculations by fast computing machines, Metropolis [/bib_ref] consist in the generation of a state perturbation, in the evaluation of energy modification, and finally in the decision to reject or not the new state according to the probability p(δE). That is, the probability to keep a better (lower) level of energy is 1, while the one to keep an infinitely worst level of energy is equal to 0. Or, in other words, the likelihood to save a given state decreases as the energy level increases. A criterion to increase the probability to reach convergence is the so-called logarithmic fading of control parameter (i.e., temperature). The simplest choice is t n+1 = C · t n , where C ∈ [0, 1] is a constant. A best global solution is reached by searching series of equilibria. Each equilibrium is obtained by series of Metropolis thresholds. The stop condition is typically an arbitrary duration or a number of loop iterations. Then the temperature is decreased and the last obtained equilibrium becomes the starting state for a new series of thresholds. The final stop is triggered if no improvement has been found since an arbitrary number of equilibria. Let us finally notice that, as a large set of temperature cooling schedules (decreasing function [bib_ref] Cooling schedules for optimal annealing, Hajek [/bib_ref] [bib_ref] Simulated annealing: searching for an optimal temperature schedule, Cohn [/bib_ref] , of moving functions, of criteria, of strategies regarding initial values, of improvements on score function, of stop criteria, and even of theoretical modeling [bib_ref] Simulated annealing: a proof of convergence, Granville [/bib_ref] [bib_ref] Convergence and finite-time behavior of simulated annealing, Mitra [/bib_ref] [bib_ref] Convergence of an annealing algorithm, Lundy [/bib_ref] have been proposed in the literature [bib_ref] Convergence and finite-time behavior of simulated annealing, Mitra [/bib_ref] , simulated annealing should be regarded more as a large family of algorithms than as a single one. Some members of the family including Basin Hopping [bib_ref] Basin hopping as a general and versatile optimization framework for the characterization..., Olson [/bib_ref] are themselves described as frameworks for ad-hoc global optimization algorithms.
A general overview of our proposal can be found in [fig_ref] Figure 2: Simulated annealing as a threshold class algorithm [/fig_ref] , while algorithm details are provided hereafter.
Designing SA for phylogenetic studies The objective is now to apply the simulated annealing method to find the largest subset of core genes that leads to the most supported phylogenetic tree. Intermediate computations of subsets will help to understand, using regressions, the effects of given genes on both topology and supports. However, SA is complex to set up in practice, and finding new optima in finite time cannot be guaranteed, as reported by Aarts, Korst, and van Laarhoven. To enlarge the probability of success, we targeted the following requirements during our experiments:
- a cooling schedule fitting with complexity, time, convergence, and precision considerations (cf. the temperature scheduling paragraph below); - a concise representation for the problem under consideration (detailed in "About a relevant configuration of SA according to the state space"); - a moving function adapted to the state solution space (same paragraph as above); - and, similarly, an acceptation function adapted to the state space (see the "Proposed SA optimization").
These four requirements are discussed hereafter.
Temperature scheduling. A criterion to increase the probability to reach convergence is the so-called logarithmic fading of control parameter (i.e., temperature). The most simple choice is t n+1 = Cṫ n , where C ∈[ 0, 1] is a constant. However, according to our experiments, such a solution is not able to produce relevant results in the phylogenetic problem under consideration. This is why the control parameter has been updated following a tiered approach, leading to an inhomogeneous Markov model: the temperature decreases only after the end of its associated Markov chain. Additionally, near an equilibrium, the Markov chain length must increase when the control parameter decreases. But, as above, at low temperature the computation time may become prohibitive without any synchronisation between the control parameter and the Markov chain characteristics. To solve such an issue, various schedule solutions proposed in the literature link these two parameters. After having tested classical benchmarking functions like the well known three-hump camel, Levi, and Booth, we finally have chosen:
[formula] t n+1 = ⎛ ⎝ t f t 1 nm−1 i ⎞ ⎠ × t n [/formula]
where t is the control parameter, t i and t f are respectively the maximum (initial) and minimum (final) of allowed control parameter values for the SA computation, while n m is the maximal number of Markov chains (equal to the temperature steps) allowed during computation.
About a relevant configuration of SA according to the state space. As in the other methods, the state space is constituted by Boolean vectors X i of the form (X i1 , . . . , X in ), where n is the number of core genes. X ij is equal to 1 if and only if gene number j in alphabetic order is in the alignment provided to the phylogenetic tool. We thus navigate again on the n-cube on which each node (that is, each state) corresponds to a subset of core genes and has additionally a labeled value provided by the subset scoring function which is again the average between the lowest bootstrap and the number of selected core genes. We can easily define a distance between two points inside this cube, like an Hamming distance between Boolean vectors, and the node score can be considered as the altitude of the current position.
To sum up, there is a topology on the state space, with neighborhood notion between two states, while the altitude (the score of a subset of genes, which is related to the SA energy) is varying between two locations. Both the density and the form of energy peaks are varying through the landscape. Neighborhoods and moves, acceptation probability, temperature scheduling functions, and their related initial values are dependant on the characteristics, or the topology, of this state space. Obviously, there is no general way to set up the parameters of the simulated annealing in this situation, as usually with such heuristics. Even choosing close configurations of closely related problems like similar chloroplasts is not a guarantee of success.
Having these considerations in mind, we have stated some hypotheses at the basis of the neighboring notion. First of all, we assume that a solution is better if it is closer to the whole core genome, so improving the number of 1's in the Boolean vector is a desired trend. Secondly, we assume no correlation between genes, and so removing (or adding) one gene cannot modify so much the scoring function. As a consequence, the next investigated state should be near the previous one, in terms of Hamming distance, and most likely with a similar or larger number of active genes. In particular, moves in the state space cannot be randomized as what occurs in the original SA algorithm. Furthermore, the starting state must be the Boolean vector constituted by 1's (that is, the whole core genome), while the scoring function must preferably tend to add genes in the considered subset (if possible). With such requirements, the neighborhood function has been designed as follows:
- A number between 1 and move_distance max (a parameter to set) is randomly chosen, following a Gaussian law. It corresponds to the number of coordinates that may possibly change. - A subset of distinct coordinates are chosen accordingly, defining this move. - For each Boolean coordinate, if the associated gene is inactive (0), it is activated (1). Otherwise, the gene is inactivated with a probability equal to nz nc × α, where nz is the number of inactivated genes in the best current solution, nc is the total number of core genes in the problem, and α is a user-defined parameter.
Proposed SA optimization. Scores in this proposal are obtained using RAxML [bib_ref] Raxml-iii: a fast program for maximum likelihood-based inference of large phylogenetic trees, Stamatakis [/bib_ref]. As an inference of a bootstrapped and rooted phylogenetic tree may take times, and as we need to compute several trees, each calculated state is tagged so that it is never recomputed without an explicit user demand. Associated and detailed results are buffered on disk. Then a simple, reliable, and not really space-characteristics dependent solution is the synchronization of some SAs after the end of a Markov chain. In order to do so, a batch of SAs is launched with the same configuration. After a chain, each running SA shares its own best known solution to a server. Then, it demands to this server if a better state has been found before starting the next chain. Finally, each SA halts after n local non optimizing chains. So a stopped SA is not restarted, even if a better solution is found elsewhere (i.e., the proposed SA stops as soon as possible). Acceptance function is also selected to take advantage of previous moves, to allow some (not too large) jumps. This is an adaptation of the so-called Tsallis acceptance probabilitieswith a control parameter normalization:
[formula] 1 − (1 − q) * ¯ * t 1 1−q , [/formula]
where is the score difference between the previous and current states,¯ their mean, t is a control parameter, and q is a user-defined factor.
How to stop the SA. To fix a predefined control (temperature) value needs to know some state space characteristics, so we choose an end criterion related to the absence of progression in scores. In other words, the proposed simulated annealing algorithm stops after n consecutive Markov chains without any score improvement. As SA is very slow on low temperatures, the choice has been to choose a small value for n. Then, a greedy local search can be launched on SA best states.
# Results
## Data generation genomes recovery and annotations
Seven hundred eighty complete genomes of chloroplasts have been downloaded from the NCBI, constituting the set of all available complete chloroplastic genomes at the date of the beginning of our study [bib_ref] Well-supported phylogenies using largest subsets of core-genes by discrete particle swarm optimization, Alsrraj [/bib_ref]. Various gene prediction methods have been previously tested, in order to translate these complete genomes in lists of annotated coding sequences. These methods encompass the single use of NCBI annotated genomes, the use of automatic annotation tools specific to organelles like DOGMA [bib_ref] Automatic annotation of organellar genomes with dogma, Wyman [/bib_ref] , and the mix of both.
Indeed, annotations from NCBI website are of very variable quality: humanly well-curated genomes go together with genomes having a lot of annotation errors, concerning either the gene names (classification or spelling errors) or DNA sequences (start and stop position, length). As the number of well annotated genomes was not enough to constitute a testing set for our experiments, we are then left to find an acceptable way to annotate the whole 780 complete genomes. As stated above, we tested various ways to annotate the genomes, and we evaluated them by checking their ability to recover the annotations (sequence positions and gene names) of the subset of humanly, well-curated genomes.
According to our experiments, there was no way to improve enough the quality of NCBI annotations [bib_ref] Finding the core-genes of chloroplasts, Alkindy [/bib_ref]. Neither by cross-validating them using automatic annotation tools, nor by trying to correct errors in gene names and positions with these tools and some edit distances [bib_ref] Gene similaritybased approaches for determining core-genes of chloroplasts, Alkindy [/bib_ref] [bib_ref] Improved core genes prediction for constructing well-supported phylogenetic trees in large sets..., Alkindy [/bib_ref]. Furthermore, to cluster the whole NCBI DNA sequences fail in separating well annotated genes in well separated clusters, due to junk DNA in the NCBI sequences. The large number of obvious errors in the NCBI annotated complete chloroplastic genomes can be explained by the large variety of annotation tools used during sequence submission, most of them being not specific to this kind of genomes (unlike DOGMA), to a misuse of these tools, or due to errors in manual annotations. The absence of a clear norm in the gene naming process adds difficulties, so that the sole method to provide accurate annotations to these 780 complete genomes was to constitute a basis of knowledge, with a subset of well curated genomes that represent well the plant diversity. And, to blast each genome against the basis, which is indeed what is done by DOGMA.
We finally have written a script that automatically send requests to the DOGMA web service, and recovers the annotated genomes. Due to this automatic process, the gene name spelling issue is resolved, and we can recover the clusters of homologous coding sequences by simply considering gene names. By applying the same tool for coding sequence prediction and naming process, we have resolved the problem of quality variability in annotations. And as DOGMA has been specifically designed for chloroplasts, errors in sequence positions have been reduced as possible. At this stage, and using our script on DOGMA web service, we have then a collection of 780 complete and "well" annotated chloroplastic genomes, from which gene names can be used to recover core and pan genomes of any subset of genomes.
## Extracting subsets of genomes for simulations
To test the ability, for the three proposed metaheuristics methods, to find the largest subset of core genes that leads to the most supported trees, we needed to extract, from the set of annotated genomes, various distinct subsets that are such that:
- Using the whole core genome in the alignment, we cannot obtain a well supported tree. - The time to compute this tree is reasonable, as we want to compute a lot of trees using a lot of subsets of core genes. For a given subset of core genes, this computation time encompasses:
1. the multi-alignment of each core gene using Muscle [bib_ref] Muscle: multiple sequence alignment with high accuracy and high throughput, Edgar [/bib_ref] , 2. the concatenation of each aligned sequence to reconstruct the "sub" genome of each considered species (i.e., the part corresponding to the considered subset of core genes), 3. the computation of the best phylogenetic tree corresponding to this alignment (with RAxML [bib_ref] Raxml-iii: a fast program for maximum likelihood-based inference of large phylogenetic trees, Stamatakis [/bib_ref] , 4. the addition of bootstrap supports to this best tree using RAxML again, 5. and finally the verification that one of these supports is lower than 95 at least. If so, this tree is considered as not well supported.
Given a subset of genomes, the multi-alignment of each core gene can be computed only once, prior to the research of the best subset of core genes leading to the most supported tree. So we do not have to consider the alignment stage when searching subsets of genomes with: (a) problematic phylogenies and (b) a time to infer their tree as low as possible. We stopped the process above before Stage 4 and we randomly pick another subset of species if the time to find their best phylogenetic trees using their whole core genome (i.e., Stage 3) exceeds 10 seconds. If this computation time is below this threshold, we then compute 50 bootstraps and we check if the best bootstrapped tree has a problem of supports. If so, we have found a convenient subset of annotated genomes, on which we can test the three metaheuristics.
## A simple comparison in small dimensions
After having executed the three metaheuristics previously described, we have validated them on test examples. We have first performed a 1D/2D comparison of the three proposals, to obtain an easy-to-understand representation of the convergence of the optimization algorithms. [fig_ref] Figure 3: Illustration of output provided by simulated annealing approach [/fig_ref] represents the output evolution of the simulated annealing, with the consecutive ends of the Markov chains and the evolution of acceptation density. From the results, we can deduce that the desired convergence behavior is well obtained, and that the comparison seems fair: no algorithm seems to underperform the other ones, and the general evolution of the energy seems to be comparable for the three algorithms. Such results allow us to further investigate simulated annealing, particle swarm optimization, and genetic algorithm for their ability to find the largest subset of core genes that leads to the most supported tree.
## Experimenting the heuristics on small collections of genomes
We first focus on small sets of species with unresolved phylogenies, for computational reasons and because small trees are easier to compare. Even in such small sets, as the core genome contains more than 100 genes, the number of combinations to test is far from what is tractable using a brute force approach. We will see that it is easy to obtain various opposed but very well supported trees using large subsets of core genes, leading to the necessity to optimize both parameters.
## A first family of algae
We have first considered the family listed in [fig_ref] Table 2: Family number 1 [/fig_ref] This family is constituted by 6 genomes, of length ranging from 120,144 to 166,111 nucleotides. The number of detected genes, for its part, ranges from 138 to 271, with a core genome of 122. The phylogeny with the alignment of these core genes leads to a small weakness in one branch (bootstrap of 94), as depicted in . Indeed, inside this bacillariophyta phylum (eukaryotic algae), C.closterium, and F.solaris are naturally in the same clade, being both in the same class of bacillariophyceae, while the three other species are in three different classes inside this phylum.
To wonder whether some genes may be responsible of such weak uncertainty, we have firstly launched the genetic algorithm: its systematic mode (in population initialization stage) indeed first tries to remove each core gene separately. This GA has stopped after 29 iterations, in systematic mode, leading to 2 topologies:
- Topology 0, depicted in [fig_ref] Figure 5: Obtained topologies with the first family [/fig_ref] , has occurred 27 times.
The best obtained tree has a lowest bootstrap of 96, while in average the lowest bootstrap is equal to 86. - Topology 1, for its part (see [fig_ref] Figure 5: Obtained topologies with the first family [/fig_ref] , has occurred twice, with a non supported branch of 64 in its best tree.
As during these experiments, we have not leaved the initialization phase, it is useless to detail here the parameters set to configure the GA. The PSO, for its part, has been configured as follows: 3 particles, a fitness lower than 0.05 to freeze the runs, and all constants that define the velocity Phylogeny of family Number 1 with the whole core genome equal to 1. This heuristics has rapidly found a first well supported phylogenetic tree in a third different topology, and with all supports equal to 100, see [fig_ref] Figure 5: Obtained topologies with the first family [/fig_ref]. However, the PSO has used only 47.5% of the core genes to reach such a tree. According to our stop criterion, this tree has not been returned by the algorithm. Indeed, this example illustrates the ability of the particle swarm optimization algorithm to more globally visit the whole space at the beginning, in order to discover regions of interest.
If we compare for instance the behavior of the PSO during the same time than the one required to finish the GA (29 iterations), we discovered 5 topologies, two of them having all their supports equal to 100 (Topologies 0 and 2 in [fig_ref] Figure 5: Obtained topologies with the first family [/fig_ref] , occurring respectively 17 and 7 times). They however used only between 44.26% and 48.36% at this starting point in the PSO. Bit by bit, over iterations, the percentage of core genes is enlarging, and the swarms tend to prefer the Topology 0. Finally, after 350 computed trees (which was the stopping condition), this topology has been obtained in 53.42% of the cases, and its best tree has a lowest bootstrap of 100 using 66.39% of core genes. The number of occurrences of the other topologies has growth more slowly and, even if all the bootstraps of their best representatives exceed the value of 98, the latter fails in the attempt to significantly increase the number of considered core genes in these representatives (always lower than 55.8%).
The simulated annealing, for its part, raised 3 topologies, exactly the ones depicted in [fig_ref] Figure 5: Obtained topologies with the first family [/fig_ref]. It has been launched with an initial temperature equal to 100, a final one of 1e-10, and an optimal exponential temperature function. Acceptation function was the Tsallis normalized one, with a q factor of 0.25, and initial (resp. final) acceptance of 0.7 (resp. 1e-05). A remarkable element is that these 3 topologies have the whole bootstraps equal to 100. Furthermore, Topology 2 appears as the best one according to the produced result (it was Topology 0 according to the GA, while PSO has not succeeded in separating these two topologies). With details, the SA has stopped after 364 computed trees with 6 occurrences of Topology 0, 43 of Topo. 1, and 315 for the Topology 2. Similarly, the percentage of core genes leading to the best representative in each topology is respectively of 56.56% (Topo. 0), 74.59% (Topo. 1), and 94.98% (Topo. 2), which thus outperforms the other ones according to these simulations.
Obviously, both PSO and SA have converged to local minima that are not global ones if we consider that both minimum bootstraps and proportion of core genes must be maximized. Launching them again with other initial values and parameters may select other optimal positions in the cube. The genetic algorithm with this family is emblematic, as during its initial population generation it has returned Topology 0 that is totally supported with 99.18% of the core genome. This topology seems to be an acceptable representation of the phylogenetic relationship between these chloroplasts. But it is remarkable that, using the same large proportion of core gene, we can break in the sister relationship between L.danicus and C.daemon. Indeed, this behavior has been obtained frequently with various collections of data, which will be illustrated below.
Up to now, we only have considered one problematic bootstrap, which may be easy to resolve when removing genes. New difficulties are added when there are at least two problems in the list of bootstraps, as improving the first one may lead to a decrease in the second value. We have investigated this point in the second tested family.
## A second family with two problematic bootstraps
The second small set of genomes is constituted by 4 Bacillariophyta plus an Alveolata as outgroup, as listed in [fig_ref] Table 3: Family number 2 [/fig_ref]. Taxonomic details are provided hereafter, while the phylogenetic tree based on the alignment of the core genome is provided in [fig_ref] Figure 6: Obtained topologies with the second family [/fig_ref] The phylogenetic tree is not well-supported, having two bootstrap values of 86. Furthermore, T.oceanica and L.danicus are not sisters in this tree, while they belong in the Coscinodiscophyceae class of diatom. More seriously, the two other species belong to the Bacillariaceae family, which is in contradiction with this tree. It is not a necessity to recover exactly the known taxonomy, as we focus on chloroplasts, but this tree is at least suspicious if we consider both supports and taxonomy. This example illustrates the fact that to use the largest common subset of sequences is not sufficient enough to guarantee a well conducted phylogenetic study. Conversely, and obviously, to have good supports is not enough, as all best trees in the different topologies of the previous family are well supported in the SA case: the largest number of core genes must be thus coupled with the research of the best supports.
Once again, the genetic algorithm has stopped rapidly, in the systematic mode. The 22 first genes have been tested (i.e., removed) before finding Topology 0 of [fig_ref] Figure 6: Obtained topologies with the second family [/fig_ref] with a lowest bootstrap equal to 96 (and 99.18% of the genes), thus stopping the GA, while a new topology (Topology 1, see [fig_ref] Figure 6: Obtained topologies with the second family [/fig_ref] has occurred three times (best tree having twice 94 as bootstraps). Compared with the first family, the genetic algorithm stops here before succeeding to reinforce the confidence put in Topology 0, which justifies to test the two other approaches.
PSO heuristics produces the same two topologies after 1165 computed trees, with all supports equal to 100, and approximately the same number of trees (632 for Topo. 0 and 533 for Topo. 1) and of genes (70.49% versus 74.59%). We stopped the swarm manually, as these two scores have not been improved during the last 500 iterations. Obviously, the 3 particles have been blocked in two local extrema, and the way we configured their velocity (0.9 and 0.8 for φ 1 and φ 2 ) does not allow them to leave these optima. So we still cannot choose definitively the topology number 0. Finally, the simulated annealing has produced 400 trees before convergence. They all belong to the two topologies detailed above. However, produced results show that Topology number 1 must be preferred, according to the SA, and this latter is neither the one obtained with the whole core genome, nor the best one according to GA. Indeed, after convergence, all bootstraps here are equal to 100 in the best tree found inside each topology. But topology of [fig_ref] Figure 6: Obtained topologies with the second family [/fig_ref] has been obtained in 88.5% of the cases. More significantly, best tree in Topology 1 is obtained using 96.72% of the core genome, while for Topology 0, the best tree uses 90.98% of it. Remark that using the ninetenths of the core genome, you can obtain a first topology with all supports equal to 100, while using more than 96% you can find a different topology with again all supports equal to 100. And, if we consider the average between the lowest bootstrap and the proportion of core genes as a score, the best topology according to GA has a score of 97.59/100, while it is of 98.36 for Topology 1 found by the SA.
# Discussion
We will now further investigate the simulated annealing convergence process, before studying more deeply the two other algorithms in a next section.
## Early analysis on sa computed problem: an illustration
An example of a SA batch run (three clients on the first family described previously) is depicted in [fig_ref] Figure 7: Illustration of clade analysis with a 3-parallelized SA [/fig_ref]. For easy understanding, only some outputs have been reported in the figure.
On the lower part, all moves of the simulated annealing are reported with their nature : synchronized move in yellow (i.e., copy, from a shared memory, of the best known solution found in the three SAs), move with an accepted status in orange, and rejected moves in black. Active genes are filled squares and not selected ones are white squares. Other important data for analysis are reported, such as: temperature, accepted score of other SAs (green and purple), and Hamming distance between two consecutive positions (moving behavior indicator).
On the upper part, a graph of accepted scores from the three SAs is provided, with the temperature variations due to move iterations (a lower score is a better one). As we represented the first run on a new collection of genomes, no previous configurations were available to set up the parameters. Consequently, a broad range of temperatures has been considered. The Markov chains are short, in order to reduce the computation time. From this beginning of an experiment, it can be deduced that:
- the temperature ranges well, allowing further experiments on the same set of data;
- even with a poor configuration, SAs have found a score "not so bad", which is associated to a topology that other heuristics have considered as a good one.
Another SA evolution is provided in [fig_ref] Figure 8: Illustration of convergence on 3-parallelized SA [/fig_ref] , in which the three main curves do not represent moves, but "moves of locally selected moves", which are stabilized over time.
## A further comparison of the distributed versions of ga and bpso performance
During the experiments of the previous section, it was impossible to evaluate in practice the behavior of the genetic algorithm, as this latter found an optimum during the initialization stage. Similarly, BPSO has underperformed the two other algorithms, while SA always produced interesting results. This is why we decided, after having studied the SA evolution on the first family, to further investigate both BPSO (with its two velocity versions) and GA in large collections of experiments, distributed in a supercomputer facilities. To do so, 12 groups of plant genomes have been extracted from our set of annotated genomes. They have been applied on our two swarm versions, and results have been compared to the genetic algorithm ones.
Comparisons are provided in [fig_ref] Table 4: Groups from BPSO version I [/fig_ref]. In these tables, Topo. column stands for the number of topologies, NbTrees is the total number of obtained trees using 10 swarms, b is the minimum bootstrap value of selected w, 100 − p is the number of missing genes in w and Occ. is the number of occurrences of the best obtained topology from 10 swarms. As can be seen in these tables, the two versions of BPSO did not provide the same kind of results:
- In the case of Chlorophyta, Pinus, and Bambusoideae, the second version of the BPSO has outperformed the first one, as the minimum bootstrap b of the best tree is finally larger for at least one swarm. - In the Ericales case, the first version has produced the best result.
We can also remark that Malpighiales has better b in GA than the two versions of BPSO. Pinus data set has got maximum bootstrap b larger than what has been obtained using the genetic algorithm, while Picea and Trebouxiophyceae have got the same values of b than with genetic algorithm. Further comparison results between GA and both versions of BPSOs are provided in [fig_ref] Figure 9: BPSO with 10 and 15 particles vs [/fig_ref].
According to this figure, we can conclude that the two approaches lead to quite equivalent bootstrap values in most data sets, while on particular subgroups obtained results are complementary. In particular, BPSO often produces better bootstraps than GA (see Magnoliidae or on Bambusoideae), but with a larger number of removed genes. Finally, using 15 particles instead of 10 does not improve so much the obtained results (see [fig_ref] Figure 9: BPSO with 10 and 15 particles vs [/fig_ref] and [fig_ref] Table 6: Comparison between the two versions of binary particule swarm optimization [/fig_ref].
# Conclusion
This article has presented three metaheuristics to produce a well supported phylogenetic tree based on the largest possible subset of core genes. These methods are, namely, genetic algorithm, binary particle swarm optimization, and simulated annealing. They have been evaluated on various sets of chloroplast species and deployed on a supercomputer facilities. Given the average between the percentage of core genes and the lowest bootstrap as scoring function, we have shown on simple examples that, given a set of species, various global optima with contradictory topologies can be reached. These first experiments emphasize that sometimes the phylogeny of chloroplasts cannot perfectly be resolved using a tree: a phylogenetic network may be more close to the reality, branches within this network being as strong as the associated tree topology is frequent. Phylogenetic networks can be obtained by merging gene trees. In future work, we will propose a way to obtain such networks with large subsets of random core genes, and will show that such ways reinforce the stability and the confidence of the network. We intend to provide too criteria to decide if either a tree or a network is preferable for a given set of DNA sequences. We will measure the impact of this choice and of the coexistence of different well-supported topologies on works like ancestral genome reconstruction. Finally, the various ways to set up the metaheuristics proposed here will be systematically investigated, to find the best manner to configure these ones when targeting the largest subset of core genes leading to the most supported tree or network. Abbreviations BPSO: Binary particle swarm optimization; GA: Genetic algorithm; PSO: Particle swarm optimization; SA: Simulated annealing
[fig] Figure 2: Simulated annealing as a threshold class algorithm. a Generic threshold class algorithm. b Metropolis algorithm. c Simulated annealing algorithm [/fig]
[fig] Figure 3: Illustration of output provided by simulated annealing approach: three-hump camel function, one instance of paralleled SA with final greedy local descent (iteration time on the abscissa) [/fig]
[fig] Figure 5: Obtained topologies with the first family. a Topology 0. b Topology 1. c Topology 2 [/fig]
[fig] Figure 6: Obtained topologies with the second family. a Topology 0 with the whole core genome. b Topology 1 obtained by GA [/fig]
[fig] Figure 7: Illustration of clade analysis with a 3-parallelized SA [/fig]
[fig] Figure 8: Illustration of convergence on 3-parallelized SA [/fig]
[fig] Figure 9: BPSO with 10 and 15 particles vs. GA. a BPSO with 15 particles vs. GA. b BPSO with 10 particles vs. GA [/fig]
[table] Table 1: Results of genetic algorithm approach on various families [/table]
[table] Table 2: Family number 1 (Pelargonium cotyledonis as outgroup) [/table]
[table] Table 3: Family number 2 (Chromera velia as outgroup) [/table]
[table] Table 4: Groups from BPSO version I [/table]
[table] Table 5: Groups from PSO version II [/table]
[table] Table 6: Comparison between the two versions of binary particule swarm optimization (with 10 or 15 particles) and the genetic algorithm [/table]
|
Dynamic Contacts of U2, RES, Cwc25, Prp8 and Prp45 Proteins with the Pre-mRNA Branch-Site and 3' Splice Site during Catalytic Activation and Step 1 Catalysis in Yeast Spliceosomes
Little is known about contacts in the spliceosome between proteins and intron nucleotides surrounding the pre-mRNA branch-site and their dynamics during splicing. We investigated protein-pre-mRNA interactions by UV-induced crosslinking of purified yeast B act spliceosomes formed on site-specifically labeled pre-mRNA, and analyzed their changes after conversion to catalytically-activated B* and step 1 C complexes, using a purified splicing system. Contacts between nucleotides upstream and downstream of the branch-site and the U2 SF3a/b proteins Prp9, Prp11, Hsh49, Cus1 and Hsh155 were detected, demonstrating that these interactions are evolutionarily conserved. The RES proteins Pml1 and Bud13 were shown to contact the intron downstream of the branch-site. A comparison of the B act crosslinking pattern versus that of B* and C complexes revealed that U2 and RES protein interactions with the intron are dynamic. Upon step 1 catalysis, Cwc25 contacts with the branch-site region, and enhanced crosslinks of Prp8 and Prp45 with nucleotides surrounding the branch-site were observed. Cwc25's step 1 promoting activity was not dependent on its interaction with pre-mRNA, indicating it acts via protein-protein interactions. These studies provide important insights into the spliceosome's protein-pre-mRNA network and reveal novel RNP remodeling events during the catalytic activation of the spliceosome and step 1 of splicing.Author SummaryThe spliceosome is a highly dynamic RNP machine that during the catalytic cycle undergoes many changes in composition and conformation. The pre-catalytic B act spliceosome contains the U2, U6 and U5 snRNAs and~40 proteins, which are evolutionarily conserved between budding yeast and metazoans. The B act spliceosome is converted to a PLOS Genetics | catalytically-activated B Ã spliceosome and following recruitment of the Cwc25 protein, step 1 of splicing is catalyzed and the C spliceosome is generated. The U2 snRNP plays an essential role in branch-site selection and pre-mRNA splicing catalysis. During the B act to B Ã transition the affinity of several U2 SF3a/b proteins for the spliceosome is significantly reduced. Whether this is due to remodeling events affecting U2 snRNP contacts with the pre-mRNA is not known. Information about conserved spliceosomal protein-pre-mRNA contacts and their dynamics during splicing remains limited. Here we investigated pre-mRNA-protein contact sites in yeast B act spliceosomes by UV-induced crosslinking. We detected contacts of nucleotides surrounding the branch-site with several of the U2 SF3a/b proteins, and we show that these interactions are evolutionarily conserved. We carried out a similar investigation with B Ã and C spliceosomes and provide important insights into the dynamics of pre-mRNA-protein interactions involving the essential U2, RES, Cwc25, Prp8 and Prp45 proteins.The removal of introns from nuclear pre-mRNAs proceeds by way of two phosphoester transfer reactions and is catalyzed by the spliceosome, a large ribonucleoprotein (RNP) complex composed of the snRNPs U1, U2, U4/U6 and U5 and several proteins [1]. The spliceosome is a highly dynamic RNP machine that undergoes many changes in composition and conformation during its work cycle[2].Initially, the U1 snRNP recognizes the 5' splice site (5' SS) and U2 snRNP recognizes the branch-site (BS) of the pre-mRNA, resulting in the formation of the pre-spliceosome or A complex. The pre-formed U4/U6.U5 tri-snRNP is then recruited, generating the B complex, which does not yet have an active site. Subsequent activation of the spliceosome (leading to the B act complex) involves major rearrangement of the spliceosomal RNA-RNA interaction network. This rearrangement is catalyzed by the combined action of the RNA helicases Prp28 and Brr2 and leads to the displacement of the U1 and U4 snRNAs and the formation of new basepair interactions between the U2 and U6 snRNAs and the 5' SS [3]. Thus, a web of RNA-RNA interactions holds the 5' SS and the BS together for step 1 catalysis[4].The B act complex, which contains U2, U6 and U5 and~40 proteins in the yeast S. cerevisiae[5], is converted by the DEAH-box NTPase Prp2, in co-operation with the G-patch protein Spp2, into a catalytically activated complex (B Ã )[6][7][8]. Following the recruitment of the splicing factor Cwc25[7,9], step 1 is catalyzed, whereby the 2'-OH of the BS adenosine attacks the 5' SS of the pre-mRNA generating the cleaved 5' exon and intron 3' exon; concomitantly the C complex is formed. This then catalyses step 2, in which the 3' SS is cleaved, resulting in the excision of the intron and ligation of the 5' and 3' exons, after which the mRNA product is released. The excised intron lariat remains associated with U2, U5 and U6 snRNPs, which then dissociate and take part in subsequent rounds of splicing.During the transformation of complex B into B act , not only is the spliceosome's RNA network radically rearranged, but also its protein composition changes significantly; as a result, several proteins are released, while twelve B act proteins are recruited. At the same time the NTC (nineteen complex) and the NTC-related proteins are stably integrated into the B act complex[5]. Likewise, the three proteins comprising the RES (pre-mRNA retention and splicing) complex [10] are also stably integrated into the B act complex.Despite the substantial restructuring that the spliceosome has undergone at this point, it does not yet have a functional active site. Previous studies showed that the binding affinity of Protein-Pre-mRNA Contacts in Yeast Spliceosomes PLOS Genetics |
Introduction several proteins is significantly changed during the Prp2-mediated transition of B act spliceosomes to catalytically activated B Ã spliceosomes [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref] [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. During this step, the essential splicing factor Cwc24 is quantitatively displaced from the B Ã complex. The U2-associated SF3a and SF3b proteins Prp11 and Cus1 and the RES protein Bud13 all remain bound to the B Ã spliceosome under near-physiological conditions, but their binding is reduced at high salt concentrations [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. The destabilization of these proteins' binding by Prp2 and Spp2 indicates that the structure of the catalytic core of the spliceosome near the BS is remodeled. This could lead to a proper 5'SS and BS configuration for nucleophilic attack on the 5' SS phosphodiester bond during step 1 catalysis [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref] [bib_ref] Stevens SW Release of SF3 from the intron branchpoint activates the first..., Lardelli [/bib_ref]. However, while it is clear that the affinity of the U2 and RES proteins for the spliceosome is significantly reduced during catalytic activation, it is not known whether this implies remodeling events involving contact of U2 and RES proteins with the pre-mRNA. Likewise, information about the set of spliceosomal protein-pre-mRNA contacts and their dynamics during splicing remains limited, but is crucial for unraveling potential functions of spliceosomal proteins for the formation and maintenance of the spliceosome's RNA-RNA network during catalysis.
The U2 snRNA/BS interaction is established in the A complex and is thought to make the BS adenosine bulge out for nucleophilic attack on the 5' SS during step 1 catalysis [bib_ref] Sharp PA Branch nucleophile selection in pre-mRNA splicing: evidence for the bulged..., Query [/bib_ref]. In human pre-spliceosomes and spliceosomes the U2 SF3a/b proteins help to recruit the U2 snRNP to the BS, and all of them except SF3b130 can be crosslinked, in a sequence-independent manner, to a region upstream of the BS (the so-called "anchoring site"), to the BS itself and to a region downstream of it [bib_ref] Reed R A potential role for U2AF-SAP 155 interactions in recruiting U2..., Gozani [/bib_ref] [bib_ref] Reed R Evidence that sequence-independent binding of highly conserved U2 snRNP proteins..., Gozani [/bib_ref]. The BS sequence is highly conserved in yeast but only weakly conserved in metazoans. Given the short length of the BS sequence, and its degeneracy in metazoans, it has been suggested that spliceosomal proteins function together with the U2 snRNA/BS duplex to tether the U2 snRNP to pre-mRNA during spliceosome assembly. In yeast there is perfect complementarity between the BS sequence and the U2 RNA. Thus the anchoring/stabilization of the U2 snRNP to the BS sequence in yeast could be different from that in human, and it may not depend critically on protein-RNA interactions. Although most of the U2 proteins in the yeast S. cerevisiae are evolutionarily conserved [bib_ref] U2 fulfils a commitment, Hodges [/bib_ref] [bib_ref] Rymond BC Interactions of the yeast SF3b splicing factor, Wang [/bib_ref] , it is not known whether they interact in a similar way with the BS region in yeast spliceosomes, or whether a site equivalent to the human anchoring site exists in yeast pre-mRNAs. So far, only Hsh155, the yeast homologue of human SF3b155, has been shown to crosslink to pre-mRNA between the BS and the 3'SS [bib_ref] Spatial organization of protein-RNA interactions in the branch site-3' splice site region..., Mcpheeters [/bib_ref]. Furthermore, it was recently shown that the RES subunit Snu17 is in contact with the pre-mRNA downstream of the BS in proximity of U2-Hsh155 [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. However, is it currently not known whether additional components of the yeast U2 SF3a/ b and RES subunits make direct contact with the pre-mRNA in spliceosomal complexes.
To address these questions, we have investigated protein-pre-mRNA interactions by UVinduced crosslinking of purified yeast spliceosomes stalled at the B act assembly stage or after conversion of B act to B Ã and C complexes, using a purified yeast splicing system [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref]. Results of these studies revealed contacts in B act complexes between pre-mRNA nucleotides directly upstream of the BS and the yeast U2 proteins Prp9, Prp11, Hsh49, Cus1 and Hsh155; the latter were also in contact with the intron further downstream of the BS. Thus, these interactions are evolutionarily conserved between yeast and man. Consistent with previous results demonstrating a Snu17-pre-mRNA crosslink [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref] , we observed that also the RES components Pml1 and Bud13 crosslinked to the intron downstream of the BS in B act complexes. Subsequent UV crosslinking with purified spliceosomes that had been stalled after catalytic activation by Prp2/ Spp2 and consecutive step 1 catalysis by Cwc25 revealed remodeling events involving contacts between U2 SF3a/b proteins upstream of the BS and the RES proteins downstream of it. Finally, concomitantly with these remodeling events, enhanced contacts of Cwc25, Prp8 and the NTC-related protein Prp45 with the BS and/or 3'SS regions were observed. These studies thus provide novel insights into the extensive protein-pre-mRNA interactions and their dynamics within and surrounding the pre-mRNA BS and 3'SS regions during step 1 of splicing in yeast spliceosomes.
# Results
UV crosslinking of affinity-purified B act spliceosomes and identification of protein-pre-mRNA crosslinks by 2D gel electrophoresis To obtain insights into the nature and number of proteins that are in direct contact with the region at the 3' end of the intron in purified yeast spliceosomes, we constructed a pre-mRNA which was body-labeled with 32 P-UTP during transcription in the 3' third of the intron, including exon 2 and 47 nucleotides (nts) upstream of the BS (termed hereinafter "3'-region-labeled pre-mRNA"; [fig_ref] Fig 1: Fig 1 [/fig_ref]. The experimental strategy used to produce the 3'-region-labeled pre-mRNA is outlined in S1A , the 3' fragment was obtained by T7 transcription. For this purpose, a T7 promoter was added by PCR and the PCR product was transcribed in vitro with an excess of GMP to ensure the presence of a monophosphate at the 5' end and with α-32 P UTP to randomly trace-label the entire RNA transcript (see S1 Text for details). To produce the 5' fragment, unlabeled actin pre-mRNA, prepared by transcription in vitro, was specifically cleaved between nucleotides 425 and 426 by a DNA enzyme based on the "8-17" catalytic motif [bib_ref] A DNA enzyme that cleaves RNA, Breaker [/bib_ref] (S1A . The 5' cleavage fragment was dephosphorylated, gel purified and ligated to the T7-transcribed 3' fragment by DNA splint directed RNA ligation [bib_ref] Site-specific modification of pre-mRNA: the 2'-hydroxyl groups at the splice sites, Moore [/bib_ref]. The 3'-region-labeled pre-mRNA allows the analysis by UV crosslinking of protein-pre-mRNA interactions at the BS site, the region directly upstream of the BS as well as around the 3'SS.
Protein-pre-mRNA interactions were analyzed initially in purified B act spliceosomes, which were assembled in vitro by incubating heat-inactivated splicing extracts from a temperaturesensitive prp2-1 yeast strain with the 3'-region-labeled pre-mRNA that also contained an MS2 binding site at its 5' end [bib_ref] The evolutionarily conserved core design of the catalytic activation step of the..., Fabrizio [/bib_ref] [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref]. B act spliceosomes were purified by glycerol-gradient centrifugation and MS2-MBP affinity chromatography and then were irradiated with UV light at 254 nm, and digested under denaturing conditions with a mixture of RNases T1, A and I. The entire protein mixture was then separated by two-dimensional (2D) gel electrophoresis as described for human spliceosomal complexes [bib_ref] Semiquantitative Proteomic Analysis of the Human Spliceosome via a Novel Two-Dimensional Gel..., Agafonov [/bib_ref]. Our 2D gel electrophoresis method is based on charge-driven separation of proteins under denaturing conditions at acidic pH in the first dimension and further separated by molecular weight though SDS gradient PAGE in the second. In contrast to the commonly used isoelectric focusing (IEF), this system prevents proteins from reaching zero charge and allows separation without in-gel precipitation over a wide range of isoelectric points (IEPs) and with masses greater than 300 kDa [bib_ref] Semiquantitative Proteomic Analysis of the Human Spliceosome via a Novel Two-Dimensional Gel..., Agafonov [/bib_ref]. [fig_ref] Fig 1: Fig 1 [/fig_ref] a RuBPS-stained 2D gel (S1 Text) of the total proteins isolated from non-irradiated B act complexes. Individual protein spots were cut out of the gel and peptides were identified by mass spectrometry (MS). Only a few contaminant proteins were found, such as Xnrn1/ Kem1 and Hrb1/Tom34, which are usually present in small amounts in preparations of yeast spliceosomes [bib_ref] The evolutionarily conserved core design of the catalytic activation step of the..., Fabrizio [/bib_ref] [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref]. All the previously identified B act complex proteins were observed [bib_ref] The evolutionarily conserved core design of the catalytic activation step of the..., Fabrizio [/bib_ref] ; these included nearly all of the U2 SF3a/b proteins (i.e. SF3b: Rse1, Hsh155, Cus1 and Hsh49; SF3a: Prp9, Prp11 and Prp21), which could be well separated from each other. The low-MW U2 proteins Msl1, Rds3 and Ysf3 and the Sm proteins D2, E, F and G ran out of the gel but could be identified in 2D gels which were modified to improve the resolution of smaller proteins [bib_ref] Semiquantitative Proteomic Analysis of the Human Spliceosome via a Novel Two-Dimensional Gel..., Agafonov [/bib_ref]. A subset of U5 proteins (Prp8, Brr2 and Snu114), and most proteins of the NTC complex and NTC-related proteins, were also located as single, distinct spots. Proteins of the RES complex (Ist3/Snu17, Pml1 and Bud13) [bib_ref] Proteomic analysis identifies a new complex required for nuclear pre-mRNA retention and..., Dziembowski [/bib_ref] were also identified. A comparison of previous MS analyses 2D gel electrophoresis of affinity-purified yeast spliceosomal B act ΔPrp2 complexes. (A) Schematic representation of the 3'-region-labeled actin pre-mRNA which was used to assemble spliceosomal complexes for 2D gel electrophoresis. The 3' portion of the pre-mRNA body-labeled with 32 P-UTP is shown in black and includes the guanosine at position 426 of the intron up to the end of the 3' exon. The unlabeled pre-mRNA is shown in gray. (B) B act complexes were purified according to protocol 1 (S1 Text). Total proteins of purified, non-irradiated (-UV) and RNase-digested B act complexes were separated electrophoretically by 2D gel electrophoresis and then stained with RuBPS [bib_ref] Semiquantitative Proteomic Analysis of the Human Spliceosome via a Novel Two-Dimensional Gel..., Agafonov [/bib_ref]. The directions of the first-dimension and second-dimension electrophoreses are shown at the top and on the left. In the first-dimension of gel electrophoresis, proteins are separated by charge while in the seconddimension they are separated according to their molecular weight. The proteins observed in the predominant spots were cut from the gel and analyzed by mass spectrometry, and the proteins identified are indicated. Two predominant spots corresponding to the contaminant proteins Xrn1 and Hrb1 are indicated; additional predominant spots corresponding to contaminant proteins are labeled with asterisks, from top to bottom: Prp5, Scp160, Kre33, Sup35, Bfr1 and Nop1. Those corresponding to RNases and MS2-MBP are also indicated. (C) Autoradiography of the 2D gel comprising total proteins of purified, UVirradiated (+UV) and RNase digested B act complexes. The circles indicate the position of the RuBPS stained spots shown in (B). Radioactive spots corresponding to proteins that were not further characterized in this work are indicated by a dot, from top to bottom: Cwc22, Cwc2/Cwc27, Isy1. of purified yeast B act spliceosomal complexes [bib_ref] The evolutionarily conserved core design of the catalytic activation step of the..., Fabrizio [/bib_ref] [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref] [bib_ref] Dissection of the factor requirements for spliceosome disassembly and the elucidation of..., Fourmann [/bib_ref] with those of our 2D analysis indicates a general reliability of this method for separating and identifying proteins that co-purify with yeast spliceosomal complexes [bib_ref] Semiquantitative Proteomic Analysis of the Human Spliceosome via a Novel Two-Dimensional Gel..., Agafonov [/bib_ref]. [fig_ref] Fig 1: Fig 1 [/fig_ref] an autoradiography of the 2D gel performed as described above but with UV-irradiated B act complexes. Exposure to 254-nm UV light is known to induce direct (zerolength) crosslinks between nitrogenous bases of nucleic acids and amino-acid side chains when they are in a favorable configuration. We observed prominent 32 P-labeled spots of U2-Hsh155 and the NTC-related protein Prp46, both of which could be superimposed on the RuBPSstained spots (indicated by circles in [fig_ref] Fig 1: Fig 1 [/fig_ref]. This indicates that the covalent attachment of a few RNA nts to proteins larger than 50 kDa, after UV-irradiation, did not alter significantly their migration behavior. A predominant crosslink in the middle of the autoradiogram was due to the contaminating poly(A)-binding protein Hrb1/Tom34, while other contaminant proteins were crosslinked to pre-mRNA at very low levels or not at all (marked by asterisks in [fig_ref] Fig 1: Fig 1 [/fig_ref]. Prominent radioactive spots were also observed for smaller U2 proteins (MW < 50 kDa), such as U2-Hsh49 and also two proteins of the RES complex (Pml1 and Snu17). The covalent attachment of RNA nts to smaller proteins led to a shift of their crosslinked species to the acidic region (i.e. left side) of the gel in the first dimension; in addition, they were separated into several spots and did not co-localize with the RuBPS stained spot, which were located slightly below [fig_ref] Fig 1: Fig 1 [/fig_ref]. Nevertheless, in all three cases crosslinked species migrated to the left side of the gel, where no other co-migrating proteins were visible in the RuBPS stained gel, indicating that the crosslinked proteins of interest were not contaminated with other proteins. We recently showed that the RES complex subunit Snu17 crosslinks in the B act complex to a 14-ntlong region of the pre-mRNA intron downstream of the BS, as shown after digestion with RNase T1 [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. Thus, the presence of several spots in the 2D gel may indicate that crosslinked pre-mRNA-protein species included shorter digestion products of the 14-nt-long RNA fragment (note that treatment with three different RNases was performed for 2D gel analysis). The same is likely to be true for RES-Pml1 and U2-Hsh49, whose crosslinked species showed a similar separation behavior [fig_ref] Fig 1: Fig 1 [/fig_ref]. The unequivocal identification of these proteins will be demonstrated below.
We also observed that low levels of additional proteins crosslinked to the 3' part of the intron, such as the U5 proteins Prp8 and Brr2, the U2 proteins Prp9, Prp11, Cus1, the RES protein Bud13 and a few B act -specific proteins (i.e. Cwc22, Cwc27/Cwc2 and Isy1, indicated by dots; [fig_ref] Fig 1: Fig 1 [/fig_ref]. These proteins crosslinked much less strongly to the 3' part of the intron in the B act spliceosome than those described above, suggesting that they are in contact with the 3' region of the pre-mRNA but are in a conformation that does not favor the formation of UVinduced crosslinks. Alternatively, digestion with a mixture of three different RNases before 2D gel electrophoresis may lead to partial digestion of the crosslinking site.
## Identification of crosslinks between u2 and res proteins and defined regions of the intron within b act spliceosomes
Next, we focused on the characterization of the crosslinked U2 SF3a/b and RES complex proteins with two major objectives: (i) identification of the candidate proteins crosslinked to pre-mRNA by pull-down of their tagged version, and (ii) identification of the region/position of crosslinks within the intron. Therefore, we generated yeast strains carrying a C-terminally TAP-tagged version of most of the U2 SF3a/b proteins and the two RES subunits Pml1 and Bud13. In addition, to localize protein-pre-mRNA interactions to well-defined short regions in the RNA stretch directly upstream of the BS or around and downstream of it, we synthesized site-specifically labeled pre-mRNA (S1B We prepared eight different actin-pre-mRNA constructs, each of which harbored a single 32 P label directly 5' of a distinct guanosine residue in the neighborhood of the BS (G452-G516, summarized in [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. For this purpose, full-length non-32 P-labeled pre-mRNAs were cleaved into two pieces at a specific position by using a distinct DNA enzyme; after 5' 32 P-labeling of the 3' piece, the two fragments were ligated by using the DNA splint-directed RNA-ligation method of Moore and Sharp. In this way, full-length pre-mRNAs were recreated, each containing a 32 P-label at the desired position [bib_ref] Site-specific modification of pre-mRNA: the 2'-hydroxyl groups at the splice sites, Moore [/bib_ref] [bib_ref] Use of deoxyribozymes in RNA research, Silverman [/bib_ref] [bib_ref] Uhlenbeck OC 3'-Phosphatase activity in T4 polynucleotide kinase, Cameron [/bib_ref] (S1B .
Extracts from the prp2-1 strain harboring the TAP-tagged versions of proteins were then used for assembly of yeast B act complexes on each of the site-specifically 32 P-labeled pre-mRNAs. The purity of B act complexes was determined by analyzing the composition of their associated snRNAs and pre-mRNA (i.e. for the presence of U2, U5L, U5S and U6 snRNA and the absence of splicing intermediates of the pre-mRNA; [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. Each purified B act complex was irradiated with UV light at 254 nm and disrupted by incubating at 70°C in 3% SDS. After complete digestion with RNase T1 (which cleaves 3' of guanosine residues), the RNA fragments shown in [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] obtained, each of which contained a single radioactive phosphate 5' of the terminal guanosine residue. We then immunoprecipitated TAP-tagged crosslinked proteins with IgG Sepharose beads and analyzed the immunoprecipitates by western blotting using the PAP complex (peroxidase-anti-peroxidase) [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] , upper panels, western blot). Autoradiography of the membrane revealed the 32 P-labeled RNA fragment crosslinked to each precipitated protein (lower panels). Thus, we were able to assign a well-defined pre-mRNA region crosslinked to a known protein and could map the entire intron area spanning from nts 447-516.
We first analyzed the U2 SF3a/b proteins and initially focused on the region upstream and across the BS [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. For the RES complex proteins we focused on the region downstream of the BS [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] because our earlier results showed that Snu17 is in direct contact with this region [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. Western blotting confirmed that U2 proteins were immunoprecipitated either before irradiation (-UV) or after it (+UV); however, when UV irradiation was omitted, 32 Plabeled fragments were not precipitated with the proteins [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] lower panels, autoradiography, lanes 1-4). After UV irradiation, the U2 proteins Prp9, Cus1, Prp11 and Hsh49 were found crosslinked to the pre-mRNA fragments 32 P-labeled at the G452 and G460 positions (i.e. fragments 447-452 and 453-460; [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] lower panels, lanes 5 and 6). None of the U2 proteins analyzed crosslinked to the downstream fragments 461-467 and 468-478, with the exception of Hsh155, which crosslinked to the RNA region 461-467 and with lower intensity to the BS region 468-478 [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. The U2 proteins Rse1 and Prp21 and the two small proteins Rds3 and Ysf3 did not crosslink to pre-mRNA. This is consistent with earlier reports that the putative human homologue of Rse1, SAP130, could not be crosslinked to pre-mRNA [bib_ref] Reed R Evidence that sequence-independent binding of highly conserved U2 snRNP proteins..., Gozani [/bib_ref] and Rds3 did not bind RNA in vitro [bib_ref] Solution structure of the U2 snRNP protein Rds3p reveals a knotted zinc-finger..., Van Roon [/bib_ref]. Taken together, these results indicate that the U2 proteins Prp9, Cus1, Prp11, Hsh49 and Hsh155 are in direct contact with the pre-mRNA in the B act complex, and their interaction is confined to a 14-nt-long region of the intron upstream of the BS (447-460), with the exception of Hsh155, which interacts in addition with the intron downstream of the BS [bib_ref] Spatial organization of protein-RNA interactions in the branch site-3' splice site region..., Mcpheeters [/bib_ref] (see also below). [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] shows a similar pull-down experiment performed with purified, irradiated and RNAse-T1-digested B act complexes carrying TAP-tagged RES Pml1 and Bud13 proteins. Western blotting showed that Pml1-TAP and Bud13-TAP were immunoprecipitated both before and after UV-irradiation. Pml1-TAP crosslinked the most strongly to the pre-mRNA fragments 483-496 and 500-511, while Bud13-TAP crosslinked to the fragment 500-511 [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] lower panels). Thus, Pml1 interacts directly with the 14-nt-long region of the intron downstream of the BS (483-496), and both Pml1 and Bud13 interact further downstream than Snu17 [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref].
Dynamic contacts of U2 SF3 proteins with the pre-mRNA upstream of the branch-site upon catalytic activation and step 1 catalysis Next, we expanded our analysis to the dynamics of protein-RNA interactions during catalytic activation and step 1 catalysis by using a purified yeast splicing system to investigate changes of . Schematic representation of site-specifically labeled pre-mRNAs carrying a single 32 P-labeled phosphate 5' of the guanosines shown in green. The RNA fragments remaining after digestion with RNase T1 are indicated by a box below the sequence. Spliceosomes were assembled on site-specifically labeled pre-mRNAs in splicing extracts of a yeast prp2-1 strain carrying proteins tagged with the TAP-tag at their C termini. B act ΔPrp2 complexes were purified according to protocol 2 (S1 Text). (B). Purified B act ΔPrp2 complexes containing TAP-tagged U2 SF3a/b proteins were UV irradiated (+ lanes) or non-irradiated (-lanes). All samples were then digested with RNase T1 and subjected to immunoprecipitation with IgG Sepharose. Immunoprecipitates were analyzed by SDS-PAGE and subsequent western blotting with peroxidase anti-peroxidase (PAP) complex antibody (upper panel). The western blot shows bands of the expected size of the U2 proteins indicated (note that the TAP-tag increases the size of a protein by ca. 21 kDa). The autoradiography of the membrane is shown in the lower panel. (C) As in (B), except that purified B actΔPrp2 complexes containing TAP-tagged RES complex proteins were used. The arrows indicate 32 P-labeled RNA fragments crosslinked to the respective proteins. UV crosslinking intensities in purified yeast spliceosomes stalled at specific assembly stages after B act , such as the B Ã and C complex stages [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref]. B act spliceosomes were prepared as above by incubating distinct site-specifically labeled actin pre-mRNAs [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] with prp2-1 heat-inactivated splicing extracts in which proteins were untagged. Each B act spliceosome was affinitypurified as above. One portion of B act spliceosomes was complemented with ATP and recombinant Prp2 and Spp2 (whereby transformation of complex B act to B Ã occurs) and one portion was complemented with ATP plus Prp2, Spp2 and Cwc25 (whereby transformation of complex B Ã to C occurs and step 1 is catalyzed). Spliceosomes were then further purified by glycerol-gradient sedimentation. The actual conversion from B act to B Ã to C was analyzed for each purified complex by gel electrophoresis [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. The presence of U2, U5 and U6 snRNA, and the total (or, for B Ã , nearly total) absence of splicing intermediates of the pre-mRNA confirmed the identity of the B act and B Ã complexes, while the presence of step 1 products confirmed the identity of the C complex (S3C and S4C Figs). In addition, the efficiency of conversion of B act to B Ã was determined by western-blot analysis, which revealed the nearly complete dissociation of the splicing factor Cwc24 from the B Ã complex during catalytic activation, as previously shown by MS and dual-color fluorescence cross-correlation spectroscopy (dcFCCS) [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref] [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. The efficiency of conversion of B Ã to C complexes was monitored from the formation of step 1 splicing products, analyzed by 8% denaturing RNA PAGE and quantified by phosphorimager. The % of step 1 products (compared to the total RNA in a lane) was calculated to be~40% (S3D and S4D Figs).
Peak fractions of purified B act , B Ã and C complexes were irradiated with 254-nm UV light and-after denaturation and digestion with RNase T1 -the crosslinked 32 P-labeled proteins were analyzed by SDS-PAGE. The gel was subjected to autoradiography [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. Each sitespecifically labeled pre-mRNA showed a distinct crosslinking pattern, revealing bands of different intensities and masses. [fig_ref] Fig 1: Fig 1 [/fig_ref] shows different degrees of crosslinking of four proteins with sizes consistent with the apparent molecular masses of untagged Prp9, Cus1, Prp11 and Hsh49, which crosslinked to the pre-mRNA fragments 447-452 and 453-460 in the B act complex. To ascertain that the four untagged crosslinked proteins corresponded to Prp9, Cus1, Prp11 and Hsh49 as shown in [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] we compared untagged and tagged proteins in parallel experiments. The molecular masses of tagged proteins are increased by a predicted 21kDa, along with the complete disappearance of the untagged version. [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] shows the patterns of untagged Hsh49, Prp11, Cus1 and Prp9 crosslinked to the pre-mRNA fragment 32 Plabeled at G460 in the B act complex. When the crosslinked proteins were compared with their tagged versions, we observed that Hsh49 shifted from 25kDa to~50kDa (compare lanes 2 and 5, red arrow), Prp9 shifted from 60kDa to~90kDa (compare lanes 2 and 3, yellow arrow), and Prp11 and Cus1 showed the expected size-shifts consistent with the addition of the TAP-tag (compare lane 2 with lanes 4 and 6; green and blue arrows, respectively). Similar comparisons were performed for the identification of Hsh155 and the RES proteins (S5B and S5C . Taken together, these data allow assignment of the radioactive bands shown in The pattern in lane 1 of [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] shows that Hsh49 crosslinked in highest yield to the~6-ntlong region 447-452, whereas Prp9 and Cus1 crosslinked at low levels to the same fragment. Prp11 was not clearly distinguishable from Hsh49; however, as shown by a light exposure of the gel in [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] its crosslinking yield was very weak. Although the chemistry of the different sites in the RNA and proteins may affect the intensity of the crosslinks, these results do suggest that Prp9, Cus1 and Prp11 make no close contacts with this particular RNA region. Remarkably, however, the intensities of Prp9 and Cus1 crosslinks were much stronger in the~8-ntlong region 32 P-labeled further downstream (i.e. at G460) in the B act complex, whereas crosslinks of Hsh49 (and Prp11) remained unchanged in this downstream~8-nt-long region [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] ; see also Figs 3B and S5A for the crosslinking intensity of Prp11).
Intriguingly, upon conversion of the B act to the B Ã and C complexes, crosslinks of Hsh49 to the~6-nt-long region (447-452) were greatly reduced, as shown by light exposure of the gel in [fig_ref] Fig 1: Fig 1 [/fig_ref]. Quantification of the intensity of Hsh49 crosslinks indicated that it was reduced by 40% and 80% in the B Ã and C complexes, respectively, relative to the B act complex (S6A This indicates that remodeling of the spliceosome leads to a reduced interaction of Hsh49 with the~6-nt-long region of the intron. Quantification of Prp11 crosslinks was difficult, as its weak signals did not resolve well from the strong signals of Hsh49. Interestingly, reduced crosslinking yield of Hsh49 to the downstream~8-nt-long fragment (453-460) was also observed during spliceosome remodeling. However, the yield of crosslinks of Hsh49 to this region were reduced to a less significant extent in the B Ã and C complexes compared with the B act complex (20% and 40%, respectively; S6A . This indicates that Hsh49 maintains a relatively strong interaction with the~8-nt-long region of the intron during catalytic activation and step 1 catalysis, compared to the upstream~6-nt-long region. Similarly, during spliceosome remodeling the levels of crosslinking of Prp9 to the~8-nt-long fragment (453-460) decreased by~40% in both B Ã and C complexes compared with the B act complex, indicating that also the binding site of Prp9 is destabilized (Figs 3A and 3B, lanes 4-6, and S6A). Although Cus1 was difficult to quantify, the pattern of its crosslinking seemed reduced to a similar extent (see Figs 3B, lanes 4-6, and S6A). Taken together, these results revealed that contacts of Hsh49, Prp11, Cus1 and Prp9 with adjacent regions of the intron were reduced to various degrees during the spliceosome's conformational changes, indicating remodeling of the binding sites of these proteins.
To obtain independent evidence of the decrease over time of Hsh49 crosslinks, we performed 2D gel electrophoresis with UV-irradiated and RNase-digested B Ã and C complexes assembled on the 3'-region-labeled pre-mRNA, and compared the intensities of the signal from their crosslinked species with those observed in corresponding experiments with the B act complex [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. Consistently with the data shown in [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] the crosslinking level of Hsh49 to the 3'-region-labeled pre-mRNA was high in the B act complex. However, the crosslinking level of Hsh49 decreased by~10% in the B Ã complex and by more than 40% in the C complex relative to the B act complex [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , as determined by quantification of the radioactive spots (S6B of Hsh49 in 2D gels). Thus, these results confirmed that contacts of Hsh49 with the 3' regions of the intron were reduced during the spliceosome's conformational changes, indicating remodeling of its binding sites.
A protein with the molecular mass of~110 kDa, which was identified as Hsh155 (S5B and [fig_ref] Fig 1: Fig 1 [/fig_ref]. These results indicate that Hsh155 is in contact with a large 50-nt-long region of the intron and it spans the BS; however, its pattern There were also additional crosslinks that were not compatible with any obvious U2 snRNP protein equivalent [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. This suggests that there are additional proteins that also contribute to the protein-pre-mRNA interaction network in this region. Of interest is a 25 kDa protein that crosslinked with low intensity to the 11-nt-long BS fragment in the C complex (see below for the characterization of this protein). In addition, ã 15 kDa protein was observed that crosslinked to the BS fragment in B act , B Ã and C complexes; the identity of this protein could not be determined either by 2D gel electrophoresis or by tagging the small U2 proteins Rds3 or Ysf3. Furthermore, crosslinking of Prp8 and Prp45 was identified (see below for a detailed description).
Dynamic contacts of all three RES subunits with the pre-mRNA downstream of the branch-site upon spliceosome remodeling
To shed some light on the dynamic interactions between the RES complex proteins and the intron, we investigated possible changes of their crosslinking pattern as described above for the U2 proteins. A protein of~20 kDa was efficiently crosslinked to the pre-mRNA fragment 483-496 [fig_ref] Fig 1: Fig 1 [/fig_ref]. This protein, which was identified as Snu17 (see S5B and S5C [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , crosslinked strongest in the B act complex; however, the intensity of crosslinking decreased by~70% in the B Ã and C complexes relative to the B act complex [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , indicating that the interaction of Snu17 with the 14-nt-long region 483-496 is weakened after activation of the spliceosome by Prp2 [fig_ref] Fig 1: Fig 1 [/fig_ref] independent experiment). Analysis of Snu17 crosslinks by 2D gel electrophoresis, confirmed that the binding of Snu17 to the 3' region of the pre-mRNA is drastically reduced during spliceosome remodeling (Figs 3C and S6B for quantification of Snu17 in 2D gels)
The intensity of Pml1 crosslinking was much lower than that of Snu17 in the same region of the intron, suggesting that Pml1 either makes no close contacts with this region or simply binds in a manner unsuitable for forming crosslinks. Nonetheless, during transition of the spliceosome from B act to B Ã the crosslinking intensity of Pml1 was reduced by~50% and by another 10% from B Ã to C [fig_ref] Fig 1: Fig 1 [/fig_ref]. A similar decrease in crosslinking intensity was confirmed by 2D gel electrophoresis (Figs 3C and S6B for quantification of Pml1 in 2D gels). Likewise, the levels of Pml1 and Bud13 crosslinks to the region further downstream (500-511) decreased by~50-60% in the B Ã and C complexes relative to B act (Figs 3A, lanes 19-21, and S6C). The intensity of Snu17 crosslinking was dramatically reduced in the 500-511 region of the intron, as compared with that upstream (483-496), indicating that the closest interaction of Snu17 with the intron is with the 14-nt-long region 483-496. This is consistent with results of our earlier pull-down experiments, which showed that Snu17 interacts directly with this region [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. Taken together, these data indicate remodeling events involving RES protein contacts with the intron region downstream of the BS upon spliceosome conformational changes.
Enhanced contacts of Prp8 with the branch-site and 3' splice site regions of the intron upon catalytic activation of the spliceosome by Prp2
A large protein with a molecular mass of~250 kDa, which is the expected size for Prp8, crosslinked in very low yield to regions 461-467 and 468-478 of the intron [fig_ref] Fig 1: Fig 1 [/fig_ref] indicated by arrowheads). Quantification of these crosslinked species revealed that Prp8 crosslinked to the region 461-467 of the intron: first in the B Ã complex and with increased level (bỹ 30%) in the post-step 1 spliceosome [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. Interestingly, Prp8 crosslinked weakly also to the BS sequence 468-478 in the C complex, indicating that during/after step 1 catalysis, Prp8 is favorably positioned for interaction with the BS [fig_ref] Fig 1: Fig 1 [/fig_ref]. A stronger Prp8 crosslink was observed further downstream, to the 14-nt-long region 483-496 in the B Ã complex, the intensity of which was enhanced in the C complex [fig_ref] Fig 1: Fig 1 [/fig_ref] ; see also S7 . Thus, our results indicate that Prp8 is favorably positioned for its interaction with the BS upon catalytic activation of the spliceosome by Prp2/Spp2, and with the 3'SS region upon subsequent step 1 catalysis by Cwc25.
Again independent evidence of the temporal increase of Prp8 crosslinks was obtained by 2D gel electrophoresis performed with UV-irradiated and RNase-digested B Ã and C complexes assembled on the 3'-region-labeled pre-mRNA, and the intensities of their crosslinked species was compared with those observed in the B act complex [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. Consistent with the data shown in [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] the crosslinking level of Prp8 to the 3'-region-labeled pre-mRNA was very low in the B act complex, indicating that Prp8 makes no close contacts with the 3' region of the intron before catalytic activation by Prp2/Spp2. However, the crosslinking level of Prp8 increased more than 60% in the B Ã complex and even more than 90% in the C complex relative to the B act complex [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref]. Taken together, these results indicate that contacts of Prp8 with the BS and 3'SS regions begin during/after the catalytic activation by Prp2/Spp2 and the interaction with the 3'SS is enhanced after step 1 catalysis promoted by Cwc25, and are consistent with previous results that showed contacts of Prp8 with the 3'SS subsequent to step 1 catalysis in yeast extracts [bib_ref] Extensive interactions of PRP8 protein with the 5' and 3' splice sites..., Teigelkamp [/bib_ref] [bib_ref] Beggs JD Interaction of the yeast splicing factor PRP8 with substrate RNA..., Teigelkamp [/bib_ref] [bib_ref] A novel role for a U5 snRNP protein in 3' splice site..., Umen [/bib_ref] [bib_ref] Prp8p interact with the 3' splice site in two distinct stages during..., Umen [/bib_ref].
Cwc25 is in contact with the branch-site region of the pre-mRNA in the C complex
To determine the identity of the 25kDa protein, which crosslinked to the BS fragment 468-478 in the post-step 1 spliceosome, we used recombinant full-length Cwc25 and truncated variants thereof in reconstitution of the C complex. Reconstituted C complexes were UV-irradiated and RNAse-T1-digested as above. [fig_ref] Fig 4: Cwc25 crosslinks within the branch-site region in the post-step 1 spliceosome [/fig_ref] that recombinant Cwc25 crosslinked to the BS fragment in the C complex (lane 6). A truncated variant of Cwc25 (residues 1-168), lacking 11 amino acids at the C-terminus, showed a similar crosslinking yield (lane 5), In contrast, the two variants Cwc25 1-102 and 1-125, lacking 77 and 54 amino acids at their C-termini, did not crosslink to this region (lanes 3 and 4). Consistently with previous observations [bib_ref] Cheng SC Link of NTR-mediated spliceosome disassembly with DEAH-box ATPases Prp2, Prp16,..., Chen [/bib_ref] , this experiment demonstrates that Cwc25 is in contact with the BS sequence and that at least the N-terminal 168 amino acids of Cwc25 are needed for this.
Intriguingly, the addition of the truncated version of Cwc25 1-168 [fig_ref] Fig 4: Cwc25 crosslinks within the branch-site region in the post-step 1 spliceosome [/fig_ref] to B Ã spliceosomes promoted step 1 catalysis, which was even more efficient than that observed with the full-length version [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. Surprisingly, Cwc25 1-125 promoted step 1 catalysis even in the absence of RNA crosslinking [fig_ref] Fig 4: Cwc25 crosslinks within the branch-site region in the post-step 1 spliceosome [/fig_ref] , indicating that Cwc25's activity in promoting step 1 can be uncoupled from its activity in RNA-binding/ crosslinking. This result suggests that Cwc25 1-125 may still interact with one or more proteins in the neighborhood of the BS and thus render the microenvironment of the catalytic center favorable for step 1 catalysis. Candidate proteins for interaction with Cwc25 are Prp8 and Hsh155, which are shown here to crosslink to the BS region concomitantly with Cwc25 [fig_ref] Fig 4: Cwc25 crosslinks within the branch-site region in the post-step 1 spliceosome [/fig_ref]. Furthermore, Yju2 may also interact with Cwc25, as it was previously shown to be involved in recruiting Cwc25 to the spliceosome [bib_ref] Cwc25 is a novel splicing factor required after Prp2 and Yju2 to..., Chiu [/bib_ref].
The NTC-related proteins Prp45 and Prp46 contact the region near the 3' splice site
In addition to the NTC, two splicing factors, namely Prp45 and Prp46 [bib_ref] Beggs JD Identification and characterization of Prp45p and Prp46p, essential pre-mRNA splicing..., Albers [/bib_ref] that interact with components of the NTC, and whose function is related to NTC in human and yeast, were also observed in the 2D gel carried out with B act spliceosomes [fig_ref] Fig 1: Fig 1 [/fig_ref]. We observed that Prp45 did not crosslink in B act complexes assembled on the 3'-region-labeled pre-mRNA after irradiation with UV light [fig_ref] Fig 1: Fig 1 [/fig_ref] , in contrast, Prp46, which was previously shown to interact with Prp45 in vitro and in vivo [bib_ref] Beggs JD Identification and characterization of Prp45p and Prp46p, essential pre-mRNA splicing..., Albers [/bib_ref] , crosslinked in high yield to the 3' end of the intron in B act spliceosomes [fig_ref] Fig 1: Fig 1 [/fig_ref]. To determine whether this crosslink was retained during spliceosome remodeling, we prepared 2D gels from crosslinked, RNase-digested B Ã and C complexes assembled on the 3'-region-labeled pre-mRNA. [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref] that the crosslink of Prp46 was preserved with a similar yield in the B Ã complex but it increased by~20% in the C complex (S6B This indicates that Prp46 remains in contact with the 3' end of the intron during remodeling of the B act to B Ã and to C complexes. To map more precisely the RNA interaction site of Prp46, we performed UV crosslinking of B act spliceosomes assembled on site-specifically labeled pre-mRNAs in Prp46-TAP extract [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref]. After pull-down, we observed that Prp46-TAP crosslinked weakly to both RNA fragments labeled at G511 and G516 [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref]. Despite the strong crosslink of Prp46 observed in 2D gels obtained from the B act complex [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref] , we detected low levels of Prp46 crosslinks in the B act complex assembled on each of the two site-specifically labeled pre-mRNAs (lanes 7 and 8). Taken together, these results suggest that the prominent crosslink of Prp46 in the 2D gel may be due either (i) to an additional crosslinked protein co-migrating with Prp46 or (ii) to interaction with a region located further upstream than the region 479-482. Alternatively, or additionally, the TAP-tag fused to Prp46 may prevent efficient crosslinking of Prp46 to the intron [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref].
Although Prp45 did not crosslink in B act complexes assembled on the 3'-region-labeled pre-mRNA after UV irradiation [fig_ref] Fig 1: Fig 1 [/fig_ref] , we nonetheless observed a protein with the expected size of Prp45 (i.e.~42kDa), which crosslinked in low yield to the fragment 483-496, in C complexes (Figs 3A, lane 18 marked by a dot, and S7). The identity of Prp45 was determined by pull-down of crosslinked and T1-digested complexes containing Prp45-TAP, assembled on pre-mRNA site-specifically labeled at G496 [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref]. Prp45-TAP crosslinked (with low intensity) only in the C complex (lane 6), indicating that Prp45 makes contact with the region of the intron 483-496 after step 1 catalysis. Independent evidence that Prp45 contacts the pre-mRNA upon step 1 catalysis was obtained again from the analysis of 2D gels obtained from crosslinked B Ã and C complexes that were assembled on the 3'-region-labeled pre-mRNA [fig_ref] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during... [/fig_ref]. Prp45 crosslinked in the B Ã complex at low levels, yet the intensity of this crosslinked species increased more than 80% in the C complex (Figs 5A, S6B and S7). This result suggests a temporal interaction of Prp45 with the intron's region near the 3' SS upstream of Prp46, and indicates that Prp45 contacts the 3' end of the intron after/during step 1 catalysis.
# Discussion
Here we have investigated pre-mRNA-protein contact sites in affinity-purified yeast B act spliceosomes by UV crosslinking. A number of crosslinked proteins of the U2 snRNP, including the SF3a subunits Prp9 and Prp11 and the SF3b proteins Cus1, Hsh49 and Hsh155, as well as RES complex proteins and their contact sites on the pre-mRNA intron, could be precisely assigned by performing crosslinking followed by 2D gel electrophoresis and immunoprecipitation. Taken together, the results indicate that the branch-site region is contacted at several positions, apparently over its entire length, by proteins. A similar investigation was carried out with affinity-purified spliceosomal B Ã and C complexes. The results presented here provide muchneeded information regarding the spliceosomal pre-mRNA-protein network, and they show for the first time that also yeast U2 SF3a/b proteins, as their human counterpart, are tightly anchored around the BS region. They also provide insight into the dynamics of pre-mRNAprotein interactions involving Cwc25, Prp8 and Prp45 within the spliceosome upon its conversion into the B Ã (i.e., catalytically activated) complex and during the subsequent conversion of the latter into the C complex (i.e., the step 1 spliceosome). [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref] B actΔPrp2 spliceosomes were assembled on pre-mRNAs site-specifically labeled 5' of the different G nucleotides as shown, in splicing extracts of a yeast prp2-1 strain carrying Prp46 tagged with the TAP-tag and were purified according to protocol 2 (S1 Text). Spliceosomes were then UV irradiated, digested with RNase T1, and subjected to immunoprecipitation with IgG Sepharose. Immunoprecipitates were analyzed by SDS-PAGE and subsequent western blotting as above (upper panel). The western blot shows a band of the expected size of Prp46-TAP. The autoradiography of the membrane is shown in the lower panel and the RNA fragments crosslinked to Prp46-TAP are marked by an arrow. (C) as in (B) B act spliceosomes were assembled on the pre-mRNA site-specifically labeled 5' of the G nucleotide at position 496, in splicing extracts of a yeast prp2-1 strain carrying Prp45 tagged with the TAP-tag. Upon addition of Prp2/Spp2 and Cwc25, B* and C complexes were obtained. The western blot shows a band of the expected size of Prp45-TAP. The RNA fragment crosslinked to Prp45-TAP is marked by an arrow.
## U2 sf3a and sf3b interactions with the intron region surrounding the bs are conserved in evolution
In the human system, the U2 protein-pre-mRNA interactions are already established in the early A complex but remain in the rearranged, activated spliceosome [bib_ref] Reed R Evidence that sequence-independent binding of highly conserved U2 snRNP proteins..., Gozani [/bib_ref] [bib_ref] Direct interactions between pre-mRNA and six U2 small nuclear ribonucleoproteins during spliceosome..., Staknis [/bib_ref]. Here, we analyzed U2 protein-pre-mRNA interactions initially in purified B act complexes, likely our data obtained with B act complexes apply also to earlier complexes (i.e. A and B complexes), which for practical reasons were not analyzed here.
Using a combination of UV crosslinking and immunoprecipitation of TAP-tagged proteins, we were able to assign a number of U2 snRNP proteins crosslinked to specific sites using pre-mRNAs that were labeled at specific positions by a combination of DNA enzymes cleavage and splint-directed ligation [bib_ref] Site-specific modification of pre-mRNA: the 2'-hydroxyl groups at the splice sites, Moore [/bib_ref]. Site-specific labeling of the RNA with 32 P is a much more promising approach for UV crosslinking studies, because the RNA-protein interaction site can be precisely mapped on the RNA. For the first time, we were able to use purified yeast spliceosomes to perform a comprehensive protein-pre-mRNA interaction analysis and thus to assign a welldefined RNA region crosslinked to a known protein. In this way we were able to map an extensive area, spanning a 70-nt-long region of the intron.
Consistent with previous studies with human spliceosomal complexes, crosslinking sites involving the yeast U2 SF3a proteins Prp9 and Prp11, as well as SF3b proteins Cus1 and Hsh49, were observed within a 14-nt-long region upstream of the BS of affinity-purified B act complexes [fig_ref] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the... [/fig_ref]. Furthermore, consistent with previous results obtained in yeast [bib_ref] Spatial organization of protein-RNA interactions in the branch site-3' splice site region..., Mcpheeters [/bib_ref] and human [bib_ref] Reed R A potential role for U2AF-SAP 155 interactions in recruiting U2..., Gozani [/bib_ref] spliceosomes, immunoprecipitation revealed contacts between a region located further downstream (surrounding the BS) and SF3b Hsh155. These results provide evidence that the region directly upstream of the BS, and surrounding it, is the main interaction platform of the yeast U2 snRNP proteins. Likewise all human U2 snRNP-associated SAPs, except for SAP130, were found in direct contact with a 20-nt-long region upstream of the BS in the isolated spliceosomal complexes A, B, and C [bib_ref] Direct interactions between pre-mRNA and six U2 small nuclear ribonucleoproteins during spliceosome..., Staknis [/bib_ref] [bib_ref] Reed R A novel set of spliceosome-associated proteins and the essential splicing..., Gozani [/bib_ref] and SF3b155 was also found to bind to a site downstream of the BS [bib_ref] Reed R A potential role for U2AF-SAP 155 interactions in recruiting U2..., Gozani [/bib_ref]. Thus, our data furthermore suggest that U2 protein-pre-mRNA interactions with the regions upstream and downstream of the BS are conserved between yeast and human (S1 . Furthermore, consistent with earlier findings [bib_ref] Reed R Evidence that sequence-independent binding of highly conserved U2 snRNP proteins..., Gozani [/bib_ref] , an oligoribonucleotide complementary to the 14-nt-long region upstream of the BS inhibits formation of the yeast A, B and B act complexes (S8 Thus, interactions of SF3a and SF3b with the pre-mRNA appear to be a prerequisite for pre-spliceosome formation also in yeast, indicating that the perfect complementarity between the BS sequence and the U2 snRNA is not sufficient to anchor the U2 snRNP to the BS sequence, and that stable binding of U2 is largely dependent on U2 protein-pre-mRNA interactions.
## The res proteins pml1 and bud13 make weak contacts with the intron between the branch-site and the 3' splice site
The RES complex is a conserved, spliceosome-associated module that has been shown to enhance splicing of a subset of transcripts and to promote the nuclear retention of unspliced pre-mRNAs in yeast [bib_ref] Proteomic analysis identifies a new complex required for nuclear pre-mRNA retention and..., Dziembowski [/bib_ref]. Furthermore, it was shown to be required for efficient splicing of TAN1 pre-mRNA, and the intron sequence between the 5'SS and the BS was necessary and sufficient to mediate dependence upon RES [bib_ref] The pre-mRNA retention and splicing complex controls tRNA maturation by promoting TAN1..., Zhou [/bib_ref]. Here, we identified low-yield crosslinks between the BS and the 3'SS of both Pml1 and Bud13. Consistent with results from our earlier studies [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref] , we show that the other RES complex protein, Snu17, is in direct contact with the pre-mRNA in the region between the BS and the 3'SS within a 14-nt-long RNA stretch upstream of the G nucleotide at position 496. Our data do not conflict with the result observed with the TAN1 pre-mRNA because the requirement of TAN1 intron nts upstream of the BS for RES dependence could be transient, and an interaction may occur earlier during spliceosome assembly.
Our observation of direct contact between Snu17 and the intron is in agreement with earlier reports showing that Snu17 consists primarily of a RRM, which is probably involved in contacting the RNA. It was also reported that the RRM of Snu17 is atypical and acts as a central binding platform that provides two separate interaction surfaces, which interact with disordered parts of Bud13 and Pml1 at the same time [bib_ref] A novel protein-protein interaction in the RES (REtention and Splicing) complex, Tripsianes [/bib_ref] [bib_ref] Crystal structure of the Pml1p subunit of the yeast precursor mRNA retention..., Trowitzsch [/bib_ref] [bib_ref] An unusual RNA recognition motif acts as a scaffold for multiple proteins..., Trowitzsch [/bib_ref]. While Bud13 and Pml1 do not harbor typical RNA-binding domains, Bud13 contains a conserved lysine-rich region that might bind RNA. In addition, Pml1p or U2 proteins interacting with RES in the spliceosome (such as the SF3b Hsh155) might facilitate the recognition of RNA by RES. A recent NMR solution structure of the core of the RES complex revealed that complex formation leads to intricate folding of the three components that stabilize the RRM fold of Snu17 upon binding of Bud13 and Pml1, while RNA binding efficiency is increased [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. Taken together, our results indicate that Snu17 crosslinks directly to the intron between the BS and the 3'SS in the B act complex, while Pml1 and Bud13 may make contact with the intron through their elaborated interconnection with Snu17, but they may be in a conformation that does not favor the formation of UV-induced crosslinks (see [fig_ref] Fig 6: Fig 6 [/fig_ref] for a summary and S2 .
As Hsh155 is in contact with nucleotides of the pre-mRNA between the BS and 3'SS [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , which are also in contact with all three components of RES, this indicates that Hsh155 and the RES proteins are in close proximity to one another in the B act spliceosome. This is consistent with earlier studies showing by a yeast two-hybrid screen and co-immunoprecipitation experiments that Snu17 interacts with U2 SF3b proteins [bib_ref] Rymond BC Interactions of the yeast SF3b splicing factor, Wang [/bib_ref]. Furthermore, the RES complex subunit Snu17 was shown to bind to the U2 snRNP [bib_ref] A Novel Yeast U2 snRNP Protein, Snu17p, Is Required for the First..., Gottschalk [/bib_ref]. Taken together, all these studies indicate that there is a direct interconnection between RES, the U2 SF3b proteins and the pre-mRNA downstream of the BS.
Evidence for subtle changes of the U2 and RES protein contacts with the pre-mRNA upon spliceosome remodeling by Prp2 and subsequent step 1 catalysis Examination of U2 protein-pre-mRNA interactions in purified spliceosomal complexes stalled after catalytic activation by Prp2/Spp2 (B Ã complex) and subsequent step 1 catalysis by Cwc25 (to form the C complex), revealed that the spliceosome structure involving the region of the intron upstream of the BS and the SF3 proteins Prp9, Hsh49 and Cus1 undergoes a conformational change during spliceosome activation and subsequent step 1 catalysis. That is, crosslinks of Prp9, Hsh49 and Cus1 were significantly reduced in both spliceosomal complexes compared with those observed with the B act complex, indicating that binding to pre-mRNA of these proteins is destabilized after ATP hydrolysis by Prp2. This is consistent with the remodeling of the structure of the catalytic core of the spliceosome near the BS upon nucleophile attack on the 5' SS phosphodiester bond during step 1 catalysis. That is, alterations in U2 protein binding are probably due to conformational changes that destabilize the interactions of these proteins with the pre-mRNA upstream of the BS concomitant with step 1. Intriguingly, our data reveal that contacts of Hsh49, Cus1 and Prp9 (and to a lesser extent Prp11) with two adjacent short regions of the intron upstream of the BS were reduced, to different degrees, during the remodeling of the spliceosome. This indicates that SF3 proteins remain in contact with the intron upstream of the BS even after step 1 catalysis, yet their binding affinity to the pre-mRNA is significantly reduced at a certain site and partially abolished at another. Our results further indicate that the complete set of U2 proteins remains in contact with the U2 snRNA via protein-RNA or protein-protein interaction. Indeed, it was shown that the U2 snRNP is released from the intron-lariat spliceosome in vitro as an integral snRNP, indicating that it remains intact during the entire splicing cycle and that none of its proteins are lost under physiological conditions in vitro [bib_ref] Dissection of the factor requirements for spliceosome disassembly and the elucidation of..., Fourmann [/bib_ref]. Furthermore, as suggested by their decreased efficiency of crosslinking during spliceosome remodeling [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , the binding of the RES complex as a whole is reduced. Remodeling events involving the RES complex proteins are intriguing because the same stretch of the intron is also bound by Prp2 and is essential for Prp2-and Spp2-mediated catalytic activation [bib_ref] The G-patch protein Spp2 couples the spliceosome-stimulated ATPase activity of the DEAH-box..., Warkocki [/bib_ref] [bib_ref] The interaction of Prp2 with a defined region of the intron is..., Liu [/bib_ref]. Indeed, Prp2 was crosslinked to the same nucleotides of the intron as the RES proteins [bib_ref] The G-patch protein Spp2 couples the spliceosome-stimulated ATPase activity of the DEAH-box..., Warkocki [/bib_ref] ; thus, it may be possible that Prp2 recognizes this stretch of RNA "productively" only when it is in contact with the RES proteins. The RES complex could be recognized as an entry point or primary target by Prp2/Spp2 to initiate translocations along the intron [bib_ref] The interaction of Prp2 with a defined region of the intron is..., Liu [/bib_ref] , thereby destabilizing RNAbound proteins and acting as a classical RNPase. Alternatively, it was recently suggested that Prp2, in addition to binding the intron, is probably involved in several protein-protein interactions in the spliceosome [bib_ref] The G-patch protein Spp2 couples the spliceosome-stimulated ATPase activity of the DEAH-box..., Warkocki [/bib_ref]. This would lead to the formation of a relay system that could transmit a power stroke within the motor module of Prp2/ATP, through the various anchor points that Prp2 shares with other components of the spliceosome [bib_ref] The G-patch protein Spp2 couples the spliceosome-stimulated ATPase activity of the DEAH-box..., Warkocki [/bib_ref]. Thus, the RES-the binding of which is destabilized upon Prp2/Spp2-mediated B Ã formation [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref] (and this work)could be an important primary element of this communication system. Importantly, it was recently reported that a prp2 mutant was suppressed by deletion of PML1, indicating that Pml1 stabilizes an interaction that Prp2 destabilizes [bib_ref] The DExD/H-box ATPase Prp2p destabilizes and proofreads the catalytic RNA core of..., Wlodaver [/bib_ref].
Prp8 is favorably positioned for its interaction with the branch-site and the 3' splice-site regions of the intron after catalytic activation of the spliceosome by Prp2
To date, Prp8 is the only spliceosomal protein that has been shown to crosslink to all the three regions in pre-mRNA that are required for splicing (5'SS, 3'SS, and BS), as well as to U5 and U6 snRNAs [bib_ref] Prp8 protein: at the heart of the spliceosome, Grainger [/bib_ref] [bib_ref] The second catalytic step of pre-mRNA splicing, Umen [/bib_ref]. Here, by 2D gel electrophoresis of affinity-purified B act complexes, we provide evidence that Prp8 -although already stably associated with B act -makes no close contacts with the 3' region of the intron before catalytic activation by Prp2/Spp2. Using affinity-purified spliceosomes stalled at the B Ã and C stages, we show that conformational changes leading to step 1 catalysis bring Prp8 to a position near the BS and the 3'SS (Figs 3 and S7) [bib_ref] Extensive interactions of PRP8 protein with the 5' and 3' splice sites..., Teigelkamp [/bib_ref] [bib_ref] Beggs JD Interaction of the yeast splicing factor PRP8 with substrate RNA..., Teigelkamp [/bib_ref] [bib_ref] A novel role for a U5 snRNP protein in 3' splice site..., Umen [/bib_ref] [bib_ref] Prp8p interact with the 3' splice site in two distinct stages during..., Umen [/bib_ref]. That is, remodeling at the catalytic core of the spliceosome accompanies stabilization of Prp8-pre-mRNA contacts. Intriguingly, previous work showed that high-affinity binding sites are created in the B Ã complex-also for additional factors required for step 1 catalysis such as Yju2 and Cwc25 -during catalytic activation [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. Thus, the ATP-dependent Prp2-driven activation of the spliceosome leads not only to reduced contacts with the pre-mRNA of U2 and RES proteins, but also to stabilization of other protein-pre-mRNA interactions by promoting direct contact with the pre-mRNA. Taken together, these results provide new insight into the dynamics of protein-pre-mRNA interactions (simultaneous reduction of some and enhancement of others) within the spliceosome during its catalytic activation and catalysis. Summary of site-specific UV crosslinking of proteins to the intron region around the branch-site in the B act complex and their dynamics following the conversion of B act into the B* complex (i.e., catalytically activated) and during the subsequent conversion of the latter into the C complex (i.e., the step 1 spliceosome). Schematic representation of the secondary-structure model of U2/U6/pre-mRNA interactions in the B act complex. The branch point A is indicated by a red bold letter. Sites in the pre-mRNA's intron crosslinked to the U2-SF3a (grey), U2-SF3b (green), RES complex proteins [(shades of purple, represented schematically according to [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref] [bib_ref] Crystal structure of the Pml1p subunit of the yeast precursor mRNA retention..., Trowitzsch [/bib_ref] [bib_ref] Structure of the yeast Pml1 splicing factor and its integration into the..., Brooks [/bib_ref] ], Prp8, Cwc25 and Prp45 are indicated by the number of the site-specifically labeled guanosines. The regions of site-specific UV-crosslinking are also summarized on the Tables on the right. Changes in crosslinking yields upon conversion of B act to B* to C are highlighted by changes in color intensities.
doi:10.1371/journal.pgen.1005539.g006
The step-1-promoting and RNA-binding activities of Cwc25 are independent
Step 1 catalysis cannot occur efficiently without Cwc25 [bib_ref] Reconstitution of both steps of Saccharomyces cerevisiae splicing with purified spliceosomal components, Warkocki [/bib_ref] [bib_ref] Cwc25 is a novel splicing factor required after Prp2 and Yju2 to..., Chiu [/bib_ref]. After Prp2-mediated catalytic activation of the spliceosome, a strong binding site is created on the B Ã spliceosome for the step 1 factor Cwc25. While Cwc25 only shows background binding to complex B act , its binding to complex B Ã has a K d value in the subnanomolar range [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. Consistent with the enhanced binding of Cwc25 upon the action of Prp2, we show here that Cwc25 crosslinks to the 11-ntlong BS fragment. These data are in agreement with earlier reports showing that Cwc25 crosslinks to the intron sequence three bases downstream of the BS [bib_ref] Cheng SC Link of NTR-mediated spliceosome disassembly with DEAH-box ATPases Prp2, Prp16,..., Chen [/bib_ref]. Interestingly, using truncated versions of recombinant Cwc25 for reconstitution of C complexes, we observed that Cwc25 1-125 (lacking 54 amino acids at its C-terminus) promoted step 1 catalysis even in the absence of RNA crosslinking, indicating that Cwc25's step-1-promoting activity is not coupled to its pre-mRNA interacting activity. This indicates that contacts of Cwc25 1-125 would theoretically occur with one or more proteins in the proximity of the BS, thus making the microenvironment of the catalytic center suitable for step 1 catalysis. This would be consistent with Cwc25 being one of the intrinsically disordered proteins, which are highly connected or "promiscuous" proteins that undergo several simultaneous or sequential interactions and use regions of disorder as a scaffold for assembling an interacting group of proteins [bib_ref] Malleable ribonucleoprotein machine: protein intrinsic disorder in the Saccharomyces cerevisiae spliceosome, Coelho Ribeiro Mde [/bib_ref]. Thus, Cwc25 might act as an important hub in the catalytic center of the spliceosome. Indeed, we observed Cwc25's contacts in the BS region of C complexes concomitant with enhanced crosslinking of Prp8 and Prp45 to the same or a slightly downstream region [fig_ref] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site [/fig_ref] , suggesting that Cwc25 co-ordinates the catalytic center through protein-protein interaction.
Prp45 is favorably positioned for its interaction with the 3' splice site regions of the intron after step 1 catalysis
We showed that Prp45 contacts the pre-mRNA only after step 1 catalysis (Figs 5C and S7), although it is already associated with the spliceosome at the B act stage [fig_ref] Fig 1: Fig 1 [/fig_ref]. Earlier results showed that, in addition to their interaction in two-hybrid screens, Prp45 and Prp46 interact in vitro, most probably through direct protein-protein contact [bib_ref] Beggs JD Identification and characterization of Prp45p and Prp46p, essential pre-mRNA splicing..., Albers [/bib_ref]. Here, Prp45 crosslinked to the pre-mRNA region of the intron 483-496 and Prp46 to the region immediately downstream (i.e. 500-516), indicating that the two proteins are also in close contact during spliceosome remodeling. Interestingly, Prp45 crosslinked during or after step 1 catalysis to the same pre-mRNA region of the intron (i.e. 483-496) where Prp8 was also found to crosslink with highest yield indicating that profound remodeling events involving this region occur. Indeed, a simultaneous reduction of Snu17 and Pml1 contacts was also observed (Figs 6 and S7 for a summary).
The contact of Prp45 to this region is consistent with earlier work that showed that Prp45 interacts with Prp22, a DEAH-box RNA helicase involved in spliceosome disassembly [bib_ref] Beggs JD Identification and characterization of Prp45p and Prp46p, essential pre-mRNA splicing..., Albers [/bib_ref] [bib_ref] Prp45 affects Prp22 partition in spliceosomal complexes and splicing efficiency of non-consensus..., Gahura [/bib_ref] , which was also crosslinked to the 3'SS in post-step 1 spliceosomes [bib_ref] Muhlenkamp P Interaction of the yeast DExH-box RNA helicase prp22p with the..., Mcpheeters [/bib_ref]. In addition, a temperature-sensitive allele of Prp45 was shown to be synthetically lethal with alleles of several secondstep splicing factors (Slu7, Prp17, Prp18 and Prp22) and with several NTC components. Thus, Prp45 may be required for Prp22 as well as the recruitment or stabilization of additional step 2 factors, and the positioning of Prp45 close to the 3'SS after step 1 may determine the timing of this event. The timing of the direct interaction of Prp45 with the pre-mRNA may explain the contribution of this protein to step 2 catalysis: this could be effected either (i) by participating in maintaining the step 2 conformation, or (ii) by binding and regulating the Prp22 ATPase/ translocase activity [bib_ref] Prp45 affects Prp22 partition in spliceosomal complexes and splicing efficiency of non-consensus..., Gahura [/bib_ref].
# Materials and methods
Site-specific cleavage of RNA with DNA enzymes In a first step, actin pre-mRNA was prepared by transcription in vitro with T7 RNA polymerase (S1 Text). The transcription reaction was not gel-purified, but instead was precipitated with ethanol. After washing twice with 70% ethanol, the precipitated RNA was dried, dissolved in 50μl CE buffer (10 mM cacodylic acid-KOH, pH 7.0, 0.2 mM EDTA-KOH, pH 8) or water and applied to a G 50 spin column (GE Healthcare). The eluted RNA was then subjected to DNA enzyme cleavage, essentially as described previously [bib_ref] Use of deoxyribozymes in RNA research, Silverman [/bib_ref]. First, a threefold molar excess of the DNA enzyme over the pre-mRNA was added to the reaction mixture. The solution was then adjusted to 15 mM NaCl and 5 mM TRIS-HCl, pH 7.7. After denaturation at 70°C for 2 min the mixture was kept at room temperature for 5 min. Finally, 150 mM NaCl, 50 mM Tris-HCl, pH 7.7 and 2 mM of both MgCl 2 and MnCl 2 were added and the mixture was incubated at 30°C for 3 hrs. To remove the cyclic phosphate produced at the 3' end of the 5' fragment by the DNA enzyme, the intrinsic 3'-phosphatase activity at low ATP concentration of T4 polynucleotide kinase (T4 PNK) was used [bib_ref] Uhlenbeck OC 3'-Phosphatase activity in T4 polynucleotide kinase, Cameron [/bib_ref]. The reaction was supplemented with 2 unites/μl T4 PNK, PNK buffer and 0.4 mM ATP and incubated for 1 h at 37°C [bib_ref] Uhlenbeck OC 3'-Phosphatase activity in T4 polynucleotide kinase, Cameron [/bib_ref]. The RNA digestion fragments were gel-purified as described for in vitro transcriptions (S1 Text).
## 5' 32 p labeling with t4 pnk
For the production of site-specifically labeled pre-mRNAs, the purified 3' pre-mRNA fragment created by DNA enzyme cleavage was 5'-phosphorylated with 2 μM γ-32 P ATP, T4-PNK buffer, 2 units/μl RNAsin and 1 unit/μl T4-PNK in a total volume of 20 μl or more, depending on the experiment. The reaction mixture was incubated for 1 h at 37°C and then purified by using a G 50 spin column, followed by phenol-chloroform-isoamyl alcohol (PCI) extraction and ethanol precipitation.
DNA splint-directed Moore and Sharp RNA ligation RNA fragments were ligated by DNA splint-directed RNA ligation [bib_ref] Site-specific modification of pre-mRNA: the 2'-hydroxyl groups at the splice sites, Moore [/bib_ref]. The 5'-ligation fragments were prepared by DNA enzyme cleavage followed by 3'-dephosphorylation as described above. For site-specific labeling, the 3' ligation fragment was labeled at the 5' end with γ-32 P ATP as described above. For region-specific labeling, the 3' fragment was produced by radioactive in vitro transcription using GMP as a starting nucleotide (S1 Text). The 5' ligation fragment, the DNA splint and the 3' ligation fragment were mixed in a 1.4:1.2:1 ratio. After addition of T4 DNA ligase buffer and water, the reaction was denatured for 2 min at 70°C and the sample was then cooled to 30°C at 6°C per min. Thereafter 1 mM ATP, 2 units/μl RNAsin and 3 units/μl T4 DNA ligase were added and the reaction was incubated for 3 hrs at 30°C. Finally, the ligation product was gel-purified. The efficiency of ligation was~30-60%. Proteins isolated from the B act , B Ã and C complexes, before and after crosslinking, were analyzed by SDS-PAGE. (C). The RNA isolated from these complexes was analyzed by denaturing PAGE and stained with silver. The presence of U2, U5L, U5S and U6 snRNA, and the nearly total absence of splicing intermediates of the pre-mRNA confirmed B act and B Ã complexes identity. The presence of step 1 products established the identity of the C complex. Species on the gel were quantified using the ImageQuant software. The efficiency of step 1 was determined by the formula: (intron-3' exon + 5' exon) / (intron-3' exon + 5' exon + pre-mRNA) x 100, and was calculated to be~40%. The asterisk indicates the presence of a small amount of U1 snRNA. (D) Western-blot analysis showing that the transformation from B act to B Ã was efficient, as revealed by the almost complete dissociation of Cwc24 from the B Ã complex during catalytic activation [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. (TIF) S4 of B act , B Ã and C complexes assembled on site-specifically labeled pre-mRNAs carrying a single 32 P-labeled phosphate 5' of the guanosines G482, 496, 511 and 516. (A) Schematic representation of site-specifically labeled pre-mRNAs as described in S2 Proteins isolated from the B act , B Ã and C complexes, before and after crosslinking, were analyzed by SDS-PAGE. (C). The RNA isolated from these complexes was analyzed by denaturing PAGE and stained with silver. The presence of U2, U5L, U5S and U6 snRNA, and the nearly total absence of splicing intermediates of the pre-mRNA confirmed B act and B Ã complexes identity. The presence of step 1 products established the identity of the C complex. Species on the gel were quantified using the ImageQuant software. The efficiency of step 1 was determined by the formula: (intron-3' exon + 5' exon) / (intron-3' exon + 5' exon + pre-mRNA) x 100, and was calculated to be~40%. The asterisk indicates the presence of a small amount of U1 snRNA. (D) Western-blot analysis showing that the transformation from B act to B Ã was efficient, as revealed by the almost complete dissociation of Cwc24 from the B Ã complex during catalytic activation [bib_ref] Prp2-mediated protein rearrangements at the catalytic core of the spliceosome as revealed..., Ohrt [/bib_ref]. (TIF) S5 of proteins crosslinked to site-specifically labeled pre-mRNAs in purified B act complexes. (A-C, upper panels) B act complexes were assembled on the site-specifically labeled pre-mRNA shown using yeast extracts containing proteins-highlighted with colors-tagged with the TAP-tag as indicated or with no tag ('Untagged', lanes 1 and 2). Peak fractions of purified B act complexes were UV-irradiated, digested with RNase T1 and then separated on SDS PAGE gels. After transfer to the nitrocellulose membrane, samples were visualized by autoradiography (upper panels) or western blotting, using the PAP antibody complex (lower panels). The addition of the 21 kDa TAP-tag to proteins resulted in an increase in their apparent molecular masses, as shown. (C) Note that UV-irradiation of the B act complex carrying Snu17-TAP led to a shift of Snu17, but also to the disappearance/shift of Pml1 (lane 3), probably owing to their intricate folding [bib_ref] Cooperative structure of the heterotrimeric pre-mRNA retention and splicing complex, Wysoczanski [/bib_ref]. The sizes in kilodaltons of the protein molecular-mass markers are shown to the right of the autoradiography or to the left of the western blot. Asterisks: uncharacterized degradation products of Hsh155. Wild-type actin pre-mRNA was incubated with a 100-fold molar excess of a 14-nt-long 2'-O-Methyl RNA oligonucleotide complementary to the pre-mRNA sequence nucleotides 447-460 (Anchoring Site Oligo) or with a control oligonucleotide complementary to a sequence of the intron more upstream (nucleotides 415-428), in 20 mM HEPES-KOH pH 7.9. The reaction mixture was incubated at 70°C for 2 min and then cooled to 4°C at 10°C/minute. In vitro splicing and B actΔPrp2 complex assembly were performed as described in the S1 Text. (Left panel) Glycerol-gradient sedimentation profiles of B act ΔPrp2 spliceosomes (formed on body-32 P-labeled wild-type actin pre-mRNA). 10-30% (v/ v) glycerol gradients containing 75 mM KCl were centrifuged for 2 h at 60000 rpm in a TH660 rotor (Sorvall). The radioactivity contained in each fraction was determined by Cherenkov counting and plotted. (Right panel) Glycerol-gradient fractions were digested with Proteinase K and analyzed by denaturing gel electrophoresis and silver staining. (TIF) S1
## Supporting information
[fig] Fig 1: Fig 1. 2D gel electrophoresis of affinity-purified yeast spliceosomal B act ΔPrp2 complexes. (A) Schematic representation of the 3'-region-labeled actin pre-mRNA which was used to assemble spliceosomal complexes for 2D gel electrophoresis. The 3' portion of the pre-mRNA body-labeled with 32 P-UTP is shown in black and includes the guanosine at position 426 of the intron up to the end of the 3' exon. The unlabeled pre-mRNA is shown in gray. (B) B act complexes were purified according to protocol 1 (S1 Text). Total proteins of purified, non-irradiated (-UV) and RNase-digested B act complexes were separated electrophoretically by 2D gel electrophoresis and then stained with RuBPS [23]. The directions of the first-dimension and second-dimension electrophoreses are shown at the top and on the left. In the first-dimension of gel electrophoresis, proteins are separated by charge while in the seconddimension they are separated according to their molecular weight. The proteins observed in the predominant spots were cut from the gel and analyzed by mass spectrometry, and the proteins identified are indicated. Two predominant spots corresponding to the contaminant proteins Xrn1 and Hrb1 are indicated; additional predominant spots corresponding to contaminant proteins are labeled with asterisks, from top to bottom: Prp5, Scp160, Kre33, Sup35, Bfr1 and Nop1. Those corresponding to RNases and MS2-MBP are also indicated. (C) Autoradiography of the 2D gel comprising total proteins of purified, UVirradiated (+UV) and RNase digested B act complexes. The circles indicate the position of the RuBPS stained spots shown in (B). Radioactive spots corresponding to proteins that were not further characterized in this work are indicated by a dot, from top to bottom: Cwc22, Cwc2/Cwc27, Isy1. [/fig]
[fig] Fig 2: Site-specific UV crosslinking of U2 SF3a/b and RES complex proteins to the intron region around the branch-site in the yeast spliceosomal B act ΔPrp2 complex. (A) [/fig]
[fig] Fig 3: Dynamics of protein-pre-mRNA interactions around the branch-site. Schematic representation of site-specifically labeled pre-mRNAs as shown inFig 2.B actΔPrp2 complexes were purified according to protocol 2 (S1 Text). Purified B actΔPrp2 and reconstituted B* and C complexes were UV irradiated, digested with RNase T1, and analyzed by SDS-PAGE electrophoresis. The amounts and molarity of the eluted spliceosomes were calculated on the basis of the specific activity of the pre-mRNA and equal molar quantity of B act , B* and C complexes were loaded onto the gel.The autoradiography of the gel is shown.Question marks indicate uncharacterized crosslinked proteins. The arrowheads point to crosslinked Prp8 and the dot indicates crosslinked Prp45. (B) Lighter exposure of the bottom half of the gel shown in panel (A). (C) Autoradiography of the bottom part of 2D gels comprising total proteins of purified, UVirradiated and RNase-digested B act , B* and C complexes, respectively, which show the crosslinking intensities of Pml1, Hsh49 and Snu17 in each of the complexes. (D) As in C, only that the upper part of 2D gels is shown. The asterisk indicates a contaminant crosslinked protein. doi:10.1371/journal.pgen.1005539.g003of interaction with the intron does not seem to change significantly during remodeling of the B act to B Ã and C complexes, with a decrease in the yield of crosslinking of only~20% (S6AFig). [/fig]
[fig] Fig 4: Cwc25 crosslinks within the branch-site region in the post-step 1 spliceosome. (A) Detail of site-specifically labeled pre-mRNAs carrying a single 32 P-labeled phosphate 5' of the guanosine in position 478 of the intron. The RNA fragment remaining after digestion with RNase T1 is indicated. Spliceosomes were assembled on the site-specifically labeled pre-mRNA, in extracts of a yeast prp2-1 strain. B actΔPrp2 complexes were purified according to protocol 2 (S1 Text). Purified B actΔPrp2 complexes were not irradiated (lane 1) or UV-irradiated (lane 2) or complemented with recombinant Prp2, Spp2 and full-length Cwc25 or truncated variants thereof, in the presence of ATP, to generate C complexes. All samples were UV-irradiated and then digested with RNase T1, and analyzed by SDS-PAGE electrophoresis. The autoradiography of the gel is shown. The question mark indicates an uncharacterized crosslinked protein. (B) The RNA isolated from the complexes shown in (A) before UV irradiation, was analyzed by denaturing gel and stained with silver. The presence of U2, U5L, U5S and U6 snRNA, and the absence of splicing intermediates of the pre-mRNA confirmed B act complexes identity (lane 1). The presence of step 1 products established the identity of the C complex. The asterisk indicates the presence of a small amount of U1 snRNA. Symbols for pre-mRNA, splicing intermediates and products are indicated on the left: the 5' exon in black, intron as a thin black line, 3' exon in white.doi:10.1371/journal.pgen.1005539.g004 [/fig]
[fig] Fig 5: Interaction of Prp46 and Prp45 with the region around the 3'SS during catalytic activation and step 1 catalysis. (A) Autoradiography of portions of 2D gels comprising total proteins of purified, UV-irradiated and RNase digested B act , B* and C complexes, respectively. The asterisks indicate contaminant crosslinked proteins. (B) Schematic representation of site-specifically labeled pre-mRNAs as shown in [/fig]
[fig] Fig 6: Fig 6. Summary of site-specific UV crosslinking of proteins to the intron region around the branch-site in the B act complex and their dynamics following the conversion of B act into the B* complex (i.e., catalytically activated) and during the subsequent conversion of the latter into the C complex (i.e., the step 1 spliceosome). Schematic representation of the secondary-structure model of U2/U6/pre-mRNA interactions in the B act complex. The branch point A is indicated by a red bold letter. Sites in the pre-mRNA's intron crosslinked to the U2-SF3a (grey), U2-SF3b (green), RES complex proteins [(shades of purple, represented schematically according to [20,38,40]], Prp8, Cwc25 and Prp45 are indicated by the number of the site-specifically labeled guanosines. The regions of site-specific UV-crosslinking are also summarized on the Tables on the right. Changes in crosslinking yields upon conversion of B act to B* to C are highlighted by changes in color intensities. [/fig]
[fig] S1: Fig. Synthesis of region-and site-specifically labeled pre-mRNA. (A) Experimental strategy used for the production of region-specifically labeled pre-mRNA. Upper panels. Left: Reaction mechanism of RNA cleavage by DNA enzymes (adapted from Silverman and Baum [25]). Right: Representation of the 8-17 deoxyribozyme. Base pairing between the target RNA and the recognition arms of the DNA enzymes and the supposed structure and sequence of the catalytic DNA loop are represented schematically (adapted from Silverman and Baum [25]). The pre-mRNA is represented schematically by lines indicating the intron and by rectangles indicating the exons. Radioactively labeled stretches are shown in green. (B) Experimental strategy used for the production of site-specifically labeled pre-mRNA. The pre-mRNA is represented schematically by lines indicating the intron and by rectangles indicating the exons. The 32 P introduced by this procedure is shown in green. (TIF) S2 Fig. Characterization of B act complexes carrying U2 proteins tagged with the TAP-tag and assembled on site-specifically labeled pre-mRNAs. (A) Schematic representation of sitespecifically labeled pre-mRNAs carrying a single 32 P-labeled phosphate 5' of the guanosines shown in green. The RNA fragments theoretically remaining after digestion with RNase T1 are indicated by a bar below the sequence. (B) The RNA isolated from the B act complexes carrying the U2 proteins indicated, tagged with the TAP-tag, was analyzed on a denaturing gel with silver-staining of the RNA and autoradiography. The presence of U2, U5L, U5S and U6 snRNA, and the absence of splicing intermediates of the pre-mRNA confirmed B act complex identity. Asterisks indicate the presence of small amount of U1, U4 and ribosomal RNAs, respectively. (TIF) S3 Fig. Characterization of B act , B Ã and C complexes assembled on site-specifically labeled pre-mRNAs carrying a single 32 P-labeled phosphate 5' of the guanosines G452, 460, 467 and 478. (A) Schematic representation of site-specifically labeled pre-mRNAs as described in S2 [/fig]
[table] Table: Human and yeast SF3a/SF3b proteins. (DOCX) S2 Table. Human and yeast RES complex proteins. (DOCX) S1 Text. Supporting Protocols. (DOCX) [/table]
|
All-oral combination of oral vinorelbine and capecitabine as first-line chemotherapy in HER2-negative metastatic breast cancer: an International Phase II Trial
BACKGROUND: This multicentre, international phase II trial evaluated the efficacy and safety profile of a first-line combination of oral vinorelbine plus capecitabine for women with metastatic breast cancer (MBC). METHODS: Patients with measurable, HER2-negative disease received, as a first line in metastatic setting, 3-weekly cycles of oral vinorelbine 80 mg m À2 (after a first cycle at 60) on day 1 and day 8, plus capecitabine 1000 mg m À2 (750 if X65 years of age) twice daily, on days 1 -14. Treatment was continued until progression or unacceptable toxicity. RESULTS: A total of 55 patients were enrolled and 54 were treated (median age: 58.5 years). Most (78%) had visceral involvement and 63% had received earlier (neo)adjuvant chemotherapy. The objective response rate (RECIST) in 49 evaluable patients was 51% (95% confidence interval (CI), 36 -66), including complete response in 4%. The clinical benefit rate (response or stable disease for X6 months) was 63% (95% CI, 48 -77). The median duration of response was 7.2 months (95% CI, 6.4 -10.2). After a median follow-up of 41 months, median progression-free survival was 8.4 months (95% CI, 5.8 -9.7) and median overall survival was 29.2 months (95% CI, 18.2 -40.1). Treatment-related adverse events were manageable, the main grade 3 -4 toxicity was neutropaenia (49%); two patients experienced febrile neutropaenia and three patients had a neutropaenic infection (including one septic death). A particularly low rate of alopaecia was observed. CONCLUSION: These results show that the all-oral combination of oral vinorelbine and capecitabine is an effective and well-tolerated first-line regimen for MBC.
Breast cancer is among the most common cancers in Western countries. Advanced breast cancer is the leading cause of death in women aged from 40 to 54 years. About one in eight Western women will develop breast cancer if they live up to an age of 85 years. Despite improving locoregional and adjuvant treatment, many patients still develop recurrent and/or metastatic breast cancer (MBC) within 10 years and will subsequently die of the disease.
There is no single standard of care for patients with MBC, as treatment plans require an individualised approach based on multiple factors, including tumour biology, growth rate of disease, presence of visceral metastases, history of earlier therapy and response, risk for toxicity and patient preference [bib_ref] Extending survival with chemotherapy in metastatic breast cancer, O'shaughnessy [/bib_ref]. Preferred first-line single agents in advanced breast cancer are anthracyclines, taxanes, capecitabine, gemcitabine and vinorelbine . More recently, bevacizumab has been evaluated in combination with taxanes with interesting clinical results. With anthracyclines and taxanes being used increasingly in early-stage setting, there is an even greater need for other active options in advanced setting to improve outcomes and/or quality of life. The development of oral chemotherapy formulations should allow a higher efficiency, by providing consistent efficacy and reduced patient constraints.
A phase II trial has evaluated oral vinorelbine (Navelbine Oral) single agent as first-line chemotherapy for locally advanced or metastatic breast cancer, showing an overall response rate of 31% and a median progression-free survival (PFS) of 17.4 weeks [bib_ref] Phase II study of oral vinorelbine in first-line advanced breast cancer chemotherapy, Freyer [/bib_ref]. Capecitabine (Xeloda) as a single agent has been largely studied in pre-treated metastatic patients, with a median response rate of 28% and a median time to progression of 4.7 months [bib_ref] Capecitabine monotherapy: safe and effective treatment for metastatic breast cancer, Ershler [/bib_ref]. Both agents have been included among commonly used compounds for the treatment of MBC in ESMO Clinical Recommendations [bib_ref] Locally recurrent or metastatic breast cancer: ESMO Clinical Recommendations for diagnosis, treatment..., Kataja [/bib_ref].
After establishing both oral vinorelbine and capecitabine as a standard of care in MBC, the development of the combination of both agents has a very strong rationale. Moreover, the two agents have different mechanisms of action, different and acceptable safety profiles and synergistic antitumour activity in preclinical models [bib_ref] Antitumour activity of combination therapy with capecitabine plus vinorelbine, and capecitabine plus..., Sawada [/bib_ref]. The combination of intravenous vinorelbine (Navelbine) and capecitabine has shown promising efficacy in MBC [bib_ref] Capecitabine and vinorelbine in elderly patients (X65 years) with metastatic breast cancer:..., Hess [/bib_ref] [bib_ref] Phase II trial of capecitabine and vinorelbine as first-line chemotherapy for metastatic..., Ghosn [/bib_ref] [bib_ref] Phase I/II study of capecitabine and vinorelbine in pretreated patients with metastatic..., Welt [/bib_ref] [bib_ref] Activity and safety of vinorelbine and capecitabine as first-line treatment in patients..., Palumbo [/bib_ref]. By combining oral vinorelbine and capecitabine, similar outcomes might be achieved without the burden of intravenous infusion.
Three phase I trials have evaluated the combination of oral vinorelbine and capecitabine in MBC [bib_ref] A phase I study of an all oral combination of vinorelbine/capecitabine in..., Kellokumpu-Lehtinen [/bib_ref] [bib_ref] Dose-finding and pharmacokinetic study of an all-oral combination regimen of oral vinorelbine..., Nole [/bib_ref] [bib_ref] Phase I-II study of oral vinorelbine and capecitabine in metastatic breast cancer:..., Anton [/bib_ref]. In all these studies, the recommended dose of capecitabine for phase II trials was 1000 mg m À2 twice daily, on days 1 -14. As special caution is needed with capecitabine in combination for elderly patients, the doses administered in this trial were reduced to 750 mg m À2 twice daily, on days 1 -14 if age was X65 years. Regarding oral vinorelbine, different schedules have been evaluated and, among these options, we evaluated in our trial the schedule of 60 mg m À2 on days 1 and 8 for cycle 1, with a dose escalation to 80 mg m À2 for subsequent cycles in the absence of grade 3 or 4 haematological toxicity.
This international, open-label, phase II trial was designed to evaluate the activity and safety of an all-oral combination of oral vinorelbine and capecitabine as first-line therapy for patients with HER2-negative MBC.
# Materials and methods
## Eligibility criteria
Eligible patients were female, X18 years, with documented metastatic breast adenocarcinoma untreated by chemotherapy. Other inclusion criteria included HER2-negative disease (IHC 0 -1 or IHC 2 þ confirmed as FISH negative), Karnofsky performance status X70%, at least one measurable lesion according to RECIST criteria [bib_ref] New guidelines to evaluate the response to treatment in solid tumors, Therasse [/bib_ref] and a life expectancy X16 weeks. Adjuvant or neoadjuvant chemotherapy containing an anthracycline and/or a taxane was allowed if X6 months had elapsed between the last dose of chemotherapy and documentation of relapse. Earlier hormone therapy for advanced disease was allowed. Patients were required to have adequate bone marrow and hepatic and renal functions, indicated by haemoglobin X10 g per 100 ml, absolute neutrophil count X2 Â 10 9 per l, platelet count X100 Â 10 9 per l, total serum bilirubin p1.5 Â upper normal limit (UNL), AST/ALT p2.5 Â UNL, (p3.5 Â UNL in case of liver metastases), alkaline phosphatase p2.5 UNL (or p5 UNL for bone metastases) and creatinine clearance 450 ml min À1 (calculated using the Cockroft and Gault formula). Patients were required to give written informed consent before study-specific procedures were performed and to comply with protocol for the duration of the study.
Patients were ineligible if they had only local relapse, earlier chemotherapy in a metastatic setting, previous exposure to a vinca-alkaloid or capecitabine, serious illness or medical conditions such as cardiac disease, unstable diabetes, uncontrolled hypercalcaemia, severe peripheral neuropathy, active infection or previous organ allograft. Patients were also excluded if they were pregnant or lactating, required a concurrent use of the antiviral sorivudine or a chemically related analogue such as brivudine, had clinical central nervous system (CNS) or leptomeningeal metastases, had a malabsorption disease that may affect absorption or oral chemotherapy, had possible hypersensitivity to fluoropyrimidine therapy, had participated in another clinical trial with any investigational drug within 30 days before study inclusion or had a history of another malignancy except cured basal-cell carcinoma of the skin or excised carcinoma in situ of the cervix.
## Primary and secondary end points
The primary end point of the study was overall response. Overall response was defined as the best confirmed response recorded from the date of registration until the end of study period. Secondary objectives included the evaluation of safety, duration of response, PFS and overall survival.
## Treatment plan
Treatment was provided in 3-weekly cycles. Oral vinorelbine was administered at 60 mg m À2 on days 1 and 8 of the first cycle and escalated to 80 mg m À2 at cycle 2 and subsequent cycles in the absence of grade 3 or 4 haematological toxicity.
Capecitabine was administered at a dose of 1000 mg m À2 twice daily, on days 1 -14 (750 mg m À2 twice daily for patients X65 years).
Prophylactic oral antiemetic medication with a 5-HT 3 antagonist was recommended before each oral vinorelbine administration.
Treatment was continued until disease progression, unacceptable toxicity or patient's refusal.
## Dose modifications
Dose adjustments and/or treatment delays could be made in the event of dose-limiting haematological or non-haematological toxicities. If study treatment could not be administered after two delays (meaning 2 weeks) of the theoretical day 1 because of any toxicity, it had to be permanently discontinued. Thus, the maximum interval between the start of one cycle and the next was 5 weeks. If one of the agents had to be permanently discontinued, the patient was withdrawn from the study.
Oral vinorelbine was not administered if patients had grade X2 neutropaenia, and capecitabine was interrupted in case of grade X3 neutropaenia. After one episode of grade 3 or 4 neutropaenia, the dose of oral vinorelbine was permanently decreased to 60 mg m À2 for subsequent cycles. Patients experiencing grade 3 or 4 neutropaenia, with or without fever, were allowed to receive a granulocyte colony-stimulating factor (G-CSF) in subsequent cycles at the investigator's discretion. If AST/ALT/alkaline phosphatase increased to 45.0 UNL, or if bilirubin increased to 41.5 UNL, both agents were not administered and a reassessment was carried out a week later. If grade X2 diarrhoea or hand -foot syndrome occurred, administration of capecitabine was interrupted until it resolved to grade 0 or 1 and doses were decreased by 25% (if grade 3) or 50% (if grade 4) for subsequent cycles.
## Study assessments
Every patient who entered the study underwent baseline assessments, including medical history, physical examination, performance status, HER2 testing (either on the primary tumour or on a metastatic site), electrocardiogram and chest X-ray. Tumour measurements by imaging were determined as clinically indicated by computed tomography (CT), abdominal ultrasound, bone scan and/or brain CT scan (if suspicion of CNS involvement). Complete blood cell counts were carried out within 2 days before each oral vinorelbine administration.
Responses were assessed every two cycles until disease progression, or more frequently if early progression was suspected. The best overall response achieved, according to RECIST criteria, was reported for each patient. A complete response (CR) required a complete disappearance of all lesions, and a partial response (PR) required at least a 30% decrease of the sum of the longest diameters of target lesions. Both CR and PR had to be confirmed at least 4 weeks later. Stable disease (SD) was defined as neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for progressive disease (PD). Progressive disease was defined as at least a 20% increase in the sum of longest diameters of target lesions and/or the appearance of new lesions. Clinical benefit was defined as patients achieving CR, PR or SD, maintained for a minimum of 6 months.
Adverse events and medical history were recorded throughout the study. The severity of adverse events was graded according to NCI common toxicity criteria version 2.0.
# Statistical analysis
The one-sample multiple testing procedure for phase II clinical trials as described by [bib_ref] One sample multiple testing procedure for phase II clinical trials, Fleming [/bib_ref] was used. This procedure uses the standard single-stage test procedure at the last one of two pre-specified tests, while allowing for early termination (should extreme results be seen) and essentially preserving the size and power of the single-stage procedure.
On the basis of a type I error rate of 5% and a 95% power to reject the null hypothesis of a 25% objective response rate (complete or partial), a sample size of 55 patients was needed.
All treated patients were included in the intent-to-treat (ITT) analysis and were analysed for safety. The evaluable population was defined as all patients eligible for the trial who underwent a full evaluation of target and non-target lesions and had received at least two cycles of study treatment (including patients with PD documented before the second cycle).
Response rate and clinical benefit were tabulated together with 95% confidence interval (CI), following the exact method. The Kaplan -Meier method was applied to overall survival, PFS and duration of response. Subset analysis (according to baseline characteristics) was performed for response rate.
# Results
## Patient characteristics
Between March 2004 and June 2005, 55 patients with MBC were enrolled from 13 sites in six countries. One patient who entered into the study did not receive study treatment because of the presence of a major exclusion criterion (low creatinine clearance). Five of the remaining 54 patients were not evaluable for response, but were included in the ITT analysis: one patient was not eligible for the study (no measurable disease as defined in the protocol), two patients were not assessable as a result of premature study discontinuation (one patient withdrew from the study after the first therapy administration owing to abdominal pain, the second patient died during the second cycle from sepsis) and two patients having received three and six cycles did not undergo a full evaluation of all target and non-target lesions. Therefore, 49 patients with measurable disease were assessable for disease response per protocol.
The patients' characteristics are described in [fig_ref] Table 1: Patient characteristics [/fig_ref]. Patients typically had visceral metastases and a good performance status. All had HER2-negative disease. Sixty-three percent of patients had received earlier (neo)adjuvant chemotherapy, including an anthracycline in the majority and a taxane in some cases.
## Clinical efficacy
The objective response rate was 51% among the 49 evaluable patients (95% CI, 36.3 -65.6), including CRs in 4.1%. Clinical benefit (CR þ PR þ SD X6 months) was 63.3% (95% CI, 48.3 -76.6). In the ITT population, the objective response rate was 46.3% (95% CI, 32.6 -60.4). Median time to response was 3.1 months (range 1.3 -6.7) and median duration of response was 7.2 months (95% CI, 6.4 -10.2) [fig_ref] Table 2: Response to treatment CI ¼ confidence interval [/fig_ref].
A subanalysis of responses according to patients' characteristics is shown in [fig_ref] Table 3: Subanalysis of responses according to patients' characteristics [/fig_ref]. It is noteworthy that the response rate in patients with liver involvement (52.2%) was similar to the rate in the overall population.
After a median follow-up of 41 months, 20 of the 54 patients treated in the study were still alive. Median PFS was 8.4 months (95% CI, 5.8 -9.7 - [fig_ref] Figure 1: Progression-free survival [/fig_ref]. Median PFS in the ER-positive (n ¼ 39) and ER-negative (n ¼ 12) population was 8.9 (95% CI, 6.7 -11.0) and 4.0 (95% CI, 1.4 -8.2) months, respectively. Median overall survival was 29.2 months (95% CI, 18.2 -40.1 - [fig_ref] Figure 2: Overall survival [/fig_ref]. The majority of patients received further lines of treatment after discontinuation of study therapy: chemotherapy (85% of patients), hormone therapy (54%) and radiotherapy (33%).
## Drug delivery
The median number of cycles was seven (range: 1 -58). Thirteen patients received study treatment for more than 40 weeks, and eight patients were treated for more than 1 year. The median relative dose intensity of oral vinorelbine and capecitabine was 86.8 and 86.7%, respectively. Dose escalation of oral vinorelbine was achieved in more than 90% of the patients. In 22 patients (41%), the dose of one or both of the agents was reduced. Every oral vinorelbine intake scheduled at day 1 could be administered and only 6.2% of the doses planned at day 8 had to be omitted.
In eight patients (14.8%), G-CSF was administered to manage neutropaenia with a curative intent. Two of these patients also received G-CSF with a preventive purpose. In all, G-CSF was administered in 14 cycles (2.8%). [fig_ref] Table 4: Treatment-related adverse events Per patient [/fig_ref] shows the incidence of the most common grade 3 -4 side effects related to treatment. The most frequent haematological toxicity was neutropaenia, with grade 3 -4 neutropaenia being observed in 49% of patients. Neutropaenia was usually brief and infrequently associated with infections. Only two patients (3.8%) experienced febrile neutropaenia. Three additional patients (5.6%) had documented infection associated with grade 3 -4 neutropaenia (one fatal septicaemia, one pneumonia and one urinary tract infection). The septic death occurred in a 65-year-old patient with a medical history of non-insulin-dependent diabetes mellitus and a mitral valve replacement who presented, during the second cycle, a positive blood culture in a context of grade 4 neutropaenia and grade 3 diarrhoea.
# Treatment-related toxicity
Non-haematological toxicities were mild. The main grade 3 gastrointestinal adverse event was vomiting, which was seen in 9.3% of patients and in 1% of cycles. The incidence of grade 3 -4 diarrhoea was observed in 3.7% of patients and in 0.6% of cycles without any grade 4. Grade 2 alopaecia was observed in 9.3% and no grade 3 -4 neuropathy was observed.
# Discussion
This multicentre study confirms the efficacy of oral vinorelbine with capecitabine previously seen in other studies, with a consistent response rate of more than 50% [bib_ref] A phase I/II study of capecitabine combined with oral vinorelbine as first..., Delcambre [/bib_ref] [bib_ref] A phase II trial of oral vinorelbine and capecitabine in anthracycline pretreated..., Finek [/bib_ref] [bib_ref] Phase II study of an all-oral regimen combining oral vinorelbine with capecitabine..., Nole [/bib_ref]. The response rate in our trial is comparable with the activity observed with intravenous combinations. Moreover, a similar response rate was achieved in the critical segment of patients having visceral metastases. Disease control was obtained in two-thirds of the patients.
Secondary efficacy end points were also very encouraging: median PFS was 8.4 months and median overall survival was 29.2 months.
There is little evidence from trials reported that major differences exist between many commonly used chemotherapy regimens [bib_ref] Chemotherapy in metastatic breast cancer: a summary of all randomised trials reported, Wilcken [/bib_ref]. Among them, the combinations of a taxane with an antimetabolite have been evaluated in two major phase III trials [bib_ref] Superior survival with capecitabine plus docetaxel combination therapy in anthracycline-pretreated patients with..., O'shaughnessy [/bib_ref] [bib_ref] Gemcitabine plus paclitaxel vs paclitaxel monotherapy in patients with metastatic breast cancer..., Albain [/bib_ref]. The activity observed with the combination of oral vinorelbine and capecitabine in our trial compares favourably with any of these regimens and provides a strong argument for its development in randomised, comparative trials with any of them. The results of a randomised phase III trial evaluating, as a firstline treatment for HER2-negative MBC patients, the combination of paclitaxel and bevacizumab vs paclitaxel alone have been published recently [bib_ref] Paclitaxel plus bevacizumab vs paclitaxel alone for metastatic breast cancer, Miller [/bib_ref]. The patient populations entered in this study and in our trial present some similar characteristics (around one-third of patients without any previous chemotherapy and about 10% of previous anthracycline and taxane in early-stage setting). There is no major difference in the outcome of patients in terms of overall response rate and median overall survival between our trial and the paclitaxel -bevacizumab regimen. Moreover, the selection of patients who may benefit from a bevacizumab-based regimen is not clear yet, this fact being a critical issue because of its high cost.
The toxicities associated with oral vinorelbine and capecitabine were predictable and manageable. As in other phase II studies evaluating this regimen, significant haematological and nonhaematological adverse events were not frequently observed. Nevertheless, as oral chemotherapy is largely taken at home by patients, it is recommended to provide them with adequate information about the measures to be taken in case of gastrointestinal adverse events and to manage them appropriately. The rate of alopaecia in patients receiving this combination was particularly low.
The good tolerability of the combination is apparent from the long duration of therapy in our study. Chemotherapy could be continued until disease progression, with few patients requiring a cessation of treatment because of intolerable adverse effects. The median number of cycles given was seven, and some patients received treatment for more than 1 year. Two meta-analyses comparing longer vs shorter chemotherapy duration in MBC patients have shown that a longer duration was linked with a survival benefit [bib_ref] A meta-analysis of chemotherapy duration in metastatic breast cancer, Gennari [/bib_ref].
Several surveys have shown that, provided efficacy and tolerability are not compromised, most patients prefer oral to intravenous chemotherapy [bib_ref] Patient preferences for oral vs intravenous palliative chemotherapy, Liu [/bib_ref] [bib_ref] Effective oral chemotherapy for breast cancer: pillars of strength, Findlay [/bib_ref]. Oral chemotherapy may reduce anxiety in patients who are afraid of injections. Whether it is administered at home or at the hospital, it reduces treatment-related constraints by requiring fewer and shorter hospital visits and has a smaller impact on daily activities [bib_ref] Effective oral chemotherapy for breast cancer: pillars of strength, Findlay [/bib_ref]. Therefore, with a similar efficacy and fewer constraints, this all-oral combination could be considered as being more efficient than intravenous treatments.
In conclusion, the oral vinorelbine/capecitabine combination evaluated in this study is effective and well tolerated. This, together with all the benefits of an oral treatment, the possibility of a longer duration of chemotherapy and of a prolonged infusion-free survival, makes it an attractive regimen that could be proposed as a valid option in first line for HER2-negative, MBC patients. 2008 and San Antonio Breast Cancer Symposium, San Antonio, Texas, USA, 2008.
[fig] Figure 1: Progression-free survival (months) intent-to-treat analysis. [/fig]
[fig] Figure 2: Overall survival (months) intent-to-treat analysis. [/fig]
[table] Table 1: Patient characteristics [/table]
[table] Table 2: Response to treatment CI ¼ confidence interval; CR ¼ complete response; PR ¼ partial response; SD ¼ stable disease. [/table]
[table] Table 3: Subanalysis of responses according to patients' characteristics [/table]
[table] Table 4: Treatment-related adverse events Per patient (%) N ¼ 53 a Per cycle (%) N ¼ 496 One patient was not evaluable for haematological adverse events. vinorelbine/capecitabine in MBC N Tubiana-Mathieu et al [/table]
|
Fetal Exposure to Perfluorinated Compounds and Attention Deficit Hyperactivity Disorder in Childhood
Background: The association between exposure to perfluorinated compounds (PFCs) and attention deficit hyperactivity disorder (ADHD) diagnosis has been sparsely investigated in humans and the findings are inconsistent.Objectives: A matched case-control study was conducted to investigate the association between fetal exposure to PFCs and ADHD diagnosis in childhood.Methods:The study base comprised children born in Malmö , Sweden, between 1978 and 2000 that were followed up until 2005. Children with ADHD (n = 206) were identified at the Department of Child and Adolescent Psychiatry. Controls (n = 206) were selected from the study base and were matched for year of birth and maternal country of birth. PFC concentrations were measured in umbilical cord serum samples. The differences of the PFC concentrations between cases and controls were investigated using Wilcoxon's paired test. Possible threshold effects (above the upper quartile for perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) and above limit of detection [LOD] for perfluorononanoic acid (PFNA)) were evaluated by conditional logistic regression.Results: The median umbilical cord serum concentrations of PFOS were 6.92 ng/ml in the cases and 6.77 ng/ml in the controls. The corresponding concentrations of PFOA were 1.80 and 1.83 ng/ml. No associations between PFCs and ADHD were observed. Odds ratios adjusted for smoking status, parity, and gestational age were 0.81 (95% confidence interval [CI] 0.50 to 1.32) for PFOS, 1.07 (95% CI 0.67 to 1.7) for PFOA, and 1.1 (95% CI 0.75 to 1.7) for PFNA.Conclusions:The current study revealed no support for an association between fetal exposure to PFOS, PFOA, or PFNA and ADHD. Citation: Ode A, Källén K, Gustafsson P, Rylander L, Jö nsson BAG, et al. (2014) Fetal Exposure to Perfluorinated Compounds and Attention Deficit Hyperactivity Disorder in Childhood. PLoS ONE 9(4): e95891.
# Introduction
Emission of pollutants from densely populated areas and industries is a growing environmental problem. Contaminants present in the environment can have a negative impact on both human health and environment. Perfluorinated compounds (PFCs) are extremely stable and persistent man-made organic chemicals that have been identified as environmental pollutants. The unique properties of PFCs have made them highly useful in numerous industrial and consumer applications such as lubricants, firefighting foams, cleaning agents, and in surface coating for paper, food packaging, textiles, furniture, carpets and cookware [bib_ref] Perfluorinated compounds-exposure assessment for the general population in Western countries, Fromme [/bib_ref] [bib_ref] Perspective on fluorocarbon chemistry, Lemal [/bib_ref].
PFCs, particularly perfluorooctane sulfonate (PFOS) and perfluorooctanic acid (PFOA) have been widely detected in the environment, wildlife, and humans [bib_ref] Levels and trends of poly-and perfluorinated compounds in the arctic environment, Butt [/bib_ref] [bib_ref] Global distribution of perfluorooctane sulfonate in wildlife, Giesy [/bib_ref] [bib_ref] Perfluorochemical surfactants in the environment, Giesy [/bib_ref] [bib_ref] Perfluorooctanesulfonate and Related Fluorochemicals in Human Blood from Several Countries, Kannan [/bib_ref] [bib_ref] Historical Comparison of Perfluorooctanesulfonate, Perfluorooctanoate, and Other Fluorochemicals in Human Blood, Olsen [/bib_ref]. Humans are exposed to PFCs through consumer products as well as contaminated air, water, and food [bib_ref] Perfluorinated compounds-exposure assessment for the general population in Western countries, Fromme [/bib_ref]. In recent years, studies have revealed that PFCs cross the placenta and accumulate in the fetus [bib_ref] Isomer profiles of perfluorochemicals in matched maternal, cord, and house dust samples:..., Beesoon [/bib_ref] [bib_ref] Perfluorooctane Sulfonate (PFOS) and Related Perfluorinated Compounds in Human Maternal and Cord..., Inoue [/bib_ref] [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref]. The fetal brain is immature and is therefore susceptible to injury caused by toxic agents [bib_ref] Developmental neurotoxicity of industrial chemicals, Grandjean [/bib_ref]. Animal data have indicated that PFCs accumulate in the brain both before and after the blood-brain barrier is formed [bib_ref] Neuroendocrine Effects of Perfluorooctane Sulfonate in Rats, Austin [/bib_ref] [bib_ref] Gestational and lactational exposure to potassium perfluorooctanesulfonate (K+ PFOS) in rats: developmental..., Butenhoff [/bib_ref] [bib_ref] Studies on the toxicological effects of PFOA and PFOS on rats using..., Cui [/bib_ref] [bib_ref] Brain region distribution and patterns of bioaccumulative perfluoroalkyl carboxylates and sulfonates in..., Greaves [/bib_ref].
Animal studies have shown that neonatal exposure to low doses of PFCs induced irreversible neurotoxic effects in adult mice and caused changes in behavior and habituation by altering the dopaminergic and cholinergic system [bib_ref] Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural..., Johansson [/bib_ref] [bib_ref] Developmental neurotoxicity of perfluorinated chemicals modeled in vitro, Slotkin [/bib_ref]. PFCs also alter levels of neural proteins that are important for the formation and growth of the synapses [bib_ref] Neonatal exposure to PFOS and PFOA in mice results in changes in..., Johansson [/bib_ref]. Defects in the dopamine transporters and receptors have been suggested to be the most significant neurobiological problem in attention deficit hyperactivity disorder (ADHD) [bib_ref] Molecular Genetics of Attention-Deficit/Hyperactivity Disorder, Faraone [/bib_ref] [bib_ref] Candidate gene studies of ADHD: a meta-analytic review, Gizer [/bib_ref].
ADHD is a neurodevelopmental disorder defined by inattention, hyperactivity and impulsivity [bib_ref] Candidate gene studies of attention-deficit/ hyperactivity disorder, Faraone [/bib_ref] [bib_ref] Associations between cerebral blood-flow measured by single photon emission computed tomography (SPECT),..., Gustafsson [/bib_ref]. The disorder has its onset in childhood, and persists into adolescence and into adulthood in some cases [bib_ref] The age-dependent decline of attention deficit hyperactivity disorder: a meta-analysis of follow-up..., Faraone [/bib_ref] [bib_ref] Childhood predictors of adult attention-deficit/hyperactivity disorder: results from the World Health Organization..., Lara [/bib_ref]. The genetic factor is believed to play the major role in the development of ADHD [bib_ref] Candidate gene studies of attention-deficit/ hyperactivity disorder, Faraone [/bib_ref] [bib_ref] Environmental risk factors for attention-deficit hyperactivity disorder, Banerjee [/bib_ref] [bib_ref] What causes attention deficit hyperactivity disorder?, Thapar [/bib_ref]. In addition, exposure to environmental toxins, such as lead, mercury, and persistent chlorinated biphenyls, has also been related to ADHD [bib_ref] Environmental risk factors for attention-deficit hyperactivity disorder, Banerjee [/bib_ref] [bib_ref] Exposures to environmental toxicants and attention deficit hyperactivity disorder in U.S. children, Braun [/bib_ref] [bib_ref] Lead and PCBs as risk factors for attention deficit/hyperactivity disorder, Eubig [/bib_ref].
Two cross-sectional studies based on parent-reported ADHD diagnosis have investigated the potential association between PFC levels in school-age children and ADHD [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] [bib_ref] Serum perfluorinated compound concentration and attention deficit/hyperactivity disorder in children 5-18 years..., Stein [/bib_ref]. The study by Hoffman et al. [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] found a positive relationship between ADHD and PFC levels in the blood of children between 12 and 15 years, whereas an association with only perfluorohexane sulfonate (PFHxS) was found in the study by Stein and Savitz [bib_ref] Serum perfluorinated compound concentration and attention deficit/hyperactivity disorder in children 5-18 years..., Stein [/bib_ref]. In another cross-sectional study, PFC exposure was associated with impulsivity in children [bib_ref] Perfluorochemical (PFC) exposure in children: associations with impaired response inhibition, Gump [/bib_ref]. Other studies based on questionnaires investigated whether behavioral health and motor coordination as well as motor and mental developmental milestones were associated with maternal PFCs during pregnancy and found no such associations except for PFOS which was associated with delayed motor development in the first two years of life [bib_ref] Prenatal exposure to perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS) and maternally reported developmental..., Fei [/bib_ref] [bib_ref] Prenatal exposure to perfluorinated chemicals and behavioral or coordination problems at age..., Fei [/bib_ref].
The frequency of children receiving an ADHD diagnosis has increased in recent years [bib_ref] Diagnosed attention deficit hyperactivity disorder and learning disability: United States, Pastor [/bib_ref]. Improved diagnostic criteria might be responsible for the increased detection of ADHD cases. Increased exposure to environmental pollutants might also contribute to the high prevalence of ADHD. Since the human brain is susceptible to disturbance by environmental pollutants during the fetal period, it is of importance to investigate the association between exposure to these pollutants during the sensitive period of fetal development and ADHD.
The objective of this study is to investigate the association between fetal exposure to PFCs and ADHD diagnosis in childhood. Unlike previous studies, this case-control study is based on clinical ADHD diagnosis and PFCs are measured in umbilical cord serum samples which reflect the PFC concentrations in the fetus. The study is a part of the Fetal Environment and Neurodevelopment Disorders in Epidemiological Research project (the FENDER project).
# Material and methods
## Participants
The selection procedure of the children with ADHD diagnosis has been previously described by Gustafsson and Kallen [bib_ref] Perinatal, maternal, and fetal characteristics of children diagnosed with attention-deficit-hyperactivity disorder: results..., Gustafsson [/bib_ref]. Briefly, at the Department of Child and Adolescent Psychiatry in the city of Malmö, 419 children born and living in Malmö between 1978 and 2000 with ADHD diagnosis were identified and were followed up until 2005. During the study period, the children with ADHD were diagnosed by one of ten experienced clinicians at the department using the Diagnostic and Statistical Manual of Mental Disorders (DSM). A child with suspected attention difficulties, hyperactivity and/or difficulties with impulse control is usually assessed to the child and adolescent psychiatry by a special teacher and a school psychologist or by the parents. The assessment begins with gathering information about the child's general medical health condition and the child's development from birth until the present time. The school psychologist or the psychologist at the psychiatric clinic performs a cognitive testing with the Wechsler Intelligence Scale (WISC). The parents and the teacher are asked to fill in questionnaires like SNAP-IV, Conner's questionnaire or the 5-15 questionnaire which all cover the symptoms of ADHD. Parents are usually asked to fill in the BRIEF-questionnaire concerning the child's executive functions in every-day life. Sometimes a member of the team at the clinic observes the child at school. The child's ability to concentrate is tested with TEA-Ch or with a computerized test of attention such as QB-Tech or IVA+. The child psychiatrist performs a paediatric examination with assessment of neurological soft-signs. The child's behaviour in different test situations and at the visits at the clinic is observed and registered. When all parts of the assessment have been performed, a team consisting of doctor, psychologist and sometimes a social worker meet and discuss the findings to come to a consensus decision concerning the diagnosis using DSM criteria. The DSM criteria DSM-III-R 11 and DSM-IV 12 were used before 1994 and from 1994 and onwards, respectively. Age at the time of diagnosis varied between 5 and 17 years, with most children being diagnosed between the ages of 8 and 12 years.
Using the personal identification numbers, children with ADHD were linked to the Swedish Medical Birth Register (SMBR) which contains demographic and obstetric information on nearly all (99%) the mothers and the infants in Sweden. Umbilical cord serum samples for children with ADHD were collected from the Malmö Maternity Unit Serum Biobank (MMUSB) using the personal identification numbers. Nearly all deliveries in Malmö take place at the Malmö University Hospital Maternity Unit, where blood samples from the mother and from the umbilical cord of the newborn have been collected at the time of delivery and stored at 220uC at the MMUSB since 1969. Controls were selected into two phases. In the first phase, for each ADHD case with an available umbilical blood sample in the biobank, the next-baby-born with serum sample of the same sex was selected as a control. However, a new publication by Gustafsson and Kallen [bib_ref] Perinatal, maternal, and fetal characteristics of children diagnosed with attention-deficit-hyperactivity disorder: results..., Gustafsson [/bib_ref] revealed the impact of maternal country of birth on the diagnosis of ADHD. Thus, the benefit of matching for the maternal country of birth completely overrode that from matching for the infant's sex. Therefore, in the second phase, a pool of ten eligible controls per ADHD case were collected from the SMBR and were matched to the cases for year of birth (612 months) and country of birth of the mother. The sample of the next-baby-born from the first phase was used if no newborn in that eligible pool of controls had an available umbilical blood sample in the biobank. The selection procedure for cases and controls is presented in [fig_ref] Figure 1: Flowchart for the selection procedure of the children with attention deficit hyperactivity... [/fig_ref].
# Ethics statements
At the Maternity Unit, the women were informed that the umbilical cord serum sample collected could be used for research purposes in the future and those who accepted gave their verbal informed consent that was documented in the medical records. During the study period only verbal informed consent was obtained. The written informed consent has been implemented in 2005 and therefore could not be considered for the current study. The data were analysed anonymously. The study protocol followed the requirements of the Declaration of Helsinki and the study, together with the consent procedure, was approved by the Research Ethics Committee at Lund University, Sweden.
## Analysis of perfluorinated compounds and cotinine in umbilical cord serum
The analyses of PFHxS, PFOS, PFOA, PFNA, perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnDA), perfluorododecanoic acid (PFDoDA), and cotinine were performed as previously described [bib_ref] Blood serum concentrations of perfluorinated compounds in men from Greenlandic Inuit and..., Lindh [/bib_ref]. Briefly, aliquots of 100 mL sera were added with isotopically labeled internal standards, the proteins were precipitated by acetonitrile and centrifugation, and analysis was then performed using a hybrid triple quadrupole linear iontrap mass spectrometer (LC/MS/MS; UFLCXR, Shimadzu Corporation, Kyoto, Japan; QTRAP 5500; Sciex, Framingham, MA, USA). The limits of detections for the detected PFCs and cotinine were 0.2 ng/ml. To increase the accuracy, the result reported is the average of two measurements from the same sample worked up and analyzed on different days. In all sample batches, the quality of the measurements was controlled by analyzing chemical blanks and in-house quality control (QC) samples. The reproducibility, determined as the relative standard deviation, between measured duplicate samples was 11% for PFOS, 12% for PFOA, 12% for PFNA, and 9% for cotinine. The reproducibility in QC samples was 8% for PFOS, 11% for PFOA, 8% for PFNA, and 5% for cotinine. Usually we are able to analyze several more PFCs with the method but some factor, probably during the storage of the samples, resulted in a high background noise in the chromatograms making detection impossible. Thus, PFHxS, PFDA, PFUnDA, and PFDoDA could not be detected in the samples due to this effect. On the other hand, due to the high correlation between PFOS and other PFCs often only PFOS and PFOA are reported in studies of PFCs. Although contamination of samples during collection is believed to be minimal, field blanks could not be provided to control for eventual contamination of the samples with PFCs. The analyses of PFOS and PFOA are part of the round robin intercomparison program (Professor Dr. med. Hans Drexler, Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine, University of Erlangen-Nuremberg, Germany) with results within the tolerance limits.
## Statistical analyses
The Wilcoxon's paired test was used to compare the PFC concentrations between ADHD cases and controls. Conditional logistic regression analysis was used to assess the association between fetal exposure to PFCs and ADHD. The odds ratio was calculated for both 1 unit increase (nanogram per milliliter) in the concentrations of PFOS and PFOA and for comparisons between concentrations above and below the 75 th percentile for the control group. For PFNA the concentrations above the limit of detection (LOD) were compared to those below LOD (0.2 ng/ml). The potential confounding variables that were considered in the present study were smoking during pregnancy, parity, and gestational age at birth, since they have been found to be associated with both PFC exposure and ADHD [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref] [bib_ref] Determinants of fetal exposure to polyfluoroalkyl compounds in, Apelberg [/bib_ref] [bib_ref] Sibship size, birth order, family structure and childhood mental disorders, Carballo [/bib_ref] [bib_ref] The relationship between attention deficit hyperactivity disorder and premature infants in Taiwanese:..., Chu [/bib_ref] [bib_ref] Associations between prenatal cigarette smoke exposure and externalized behaviors at school age..., Desrosiers [/bib_ref] [bib_ref] Perfluorinated chemicals and fetal growth: a study within the Danish National Birth..., Fei [/bib_ref] [bib_ref] Maternal smoking and hyperactivity in 8-year-old children, Kotimaa [/bib_ref] [bib_ref] Effects of low birth weight, maternal smoking in pregnancy and social class..., Langley [/bib_ref] [bib_ref] Correlations between prenatal exposure to perfluorinated chemicals and reduced fetal growth, Washino [/bib_ref].
Smoking during pregnancy was determined by cotinine levels in umbilical cord serum. Cotinine levels below the LOD (0.2 ng/ml) were related to nonsmoking pregnant women, cotinine levels higher than 15 ng/ml were related to active smokers and levels between 0.2 and 15 ng/ml were related to second-hand smokers [bib_ref] Self-reported nicotine exposure and plasma levels of cotinine in early and late..., George [/bib_ref]. Parity was divided into three groups according to number of previously born children (0 [i.e. nulliparous], 1, or $2 children). Gestational age was entered in the analyses as class variable divided into three groups; ,37, 37-42, and .42 weeks of pregnancy.
The odds ratios were calculated in paired samples (n = 202) using Egret for Windows 2.0 (Cytel Software Corporation). The rest of the analyses were performed in IBM SPSS Statistics version 20 (IBM Corporation 1989, 2011).
The power calculation was based on 206 cases and matched controls. With the current setting, we had an 80% chance of detecting a difference in the levels of 0.20 standard deviations, with a value of 0.05, between cases and controls. For the analysis of the threshold effect, with a value of 0.05 and b value of 0.80, the lowest detectable odds ratio was 1.8.
# Results
PFOS and PFOA concentrations were above the LOD in 98% of the samples, whereas for PFNA about 12% were above the LOD. PFOS and PFOA concentrations below the LOD in individual samples (n = 2 for each) were replaced with 0.2 ng/ml.
The demographic characteristics and the umbilical cord PFC concentrations of the study population are presented in [fig_ref] Table 1: Median concentration [/fig_ref]. [fig_ref] Figure 2: Boxplot of the umbilical cord concentrations of perfluorooctane sulfonate [/fig_ref] shows the distribution of PFOS and PFOA levels in the ADHD cases and the controls. The median concentrations of PFNA above LOD for cases and controls were 0.31 and 0.28 ng/ ml, respectively. Wilcoxon's paired test revealed no differences in cord serum PFC concentrations between children with ADHD diagnosis and controls (p = 0.72, 0.44, and 0.48 for PFOS, PFOA, and PFNA respectively).
Conditional logistic regression analyses revealed no significant associations between umbilical cord concentrations of PFCs and ADHD [fig_ref] Table 2: The crude and adjusted odds ratio with 95% confidence interval of attention... [/fig_ref]. The result did not change after adjusting for smoking during pregnancy, parity, and gestational age at birth.
# Discussion
The present study found no statistically significant associations between exposure to PFCs during pregnancy and ADHD diagnosis during childhood, although the measured umbilical cord concentrations of PFOS were among the highest in Europe [bib_ref] Perfluorinated chemicals and fetal growth: a study within the Danish National Birth..., Fei [/bib_ref] [bib_ref] Placental transfer of perfluorinated compounds is selective-a Norwegian Mother and Child sub-cohort..., Gutzkow [/bib_ref] [bib_ref] Partition of environmental chemicals between maternal and fetal blood and tissues, Needham [/bib_ref] and even among the highest in the world [bib_ref] Isomer profiles of perfluorochemicals in matched maternal, cord, and house dust samples:..., Beesoon [/bib_ref] [bib_ref] Trans-placental transfer of thirteen perfluorinated compounds and relations with fetal thyroid hormones, Kim [/bib_ref] [bib_ref] Distribution of perfluorochemicals between sera and milk from the same mothers and..., Kim [/bib_ref] [bib_ref] Comparison on gestation and lactation exposure of perfluorinated compounds for newborns, Liu [/bib_ref]. For PFOA, the levels were higher than those measured in Norway and other non-European countries but lower than those in Danmark and Faraoe Islands [bib_ref] Isomer profiles of perfluorochemicals in matched maternal, cord, and house dust samples:..., Beesoon [/bib_ref] [bib_ref] Perfluorinated chemicals and fetal growth: a study within the Danish National Birth..., Fei [/bib_ref] [bib_ref] Placental transfer of perfluorinated compounds is selective-a Norwegian Mother and Child sub-cohort..., Gutzkow [/bib_ref] [bib_ref] Partition of environmental chemicals between maternal and fetal blood and tissues, Needham [/bib_ref].
Animal data revealed that neonatal mice that were exposed to high doses of PFOS and PFOA showed behavioral defects which ranged from slight effects at the anxiety level [bib_ref] Behavioral effects in adult mice exposed to perfluorooctane sulfonate (PFOS), Fuentes [/bib_ref] to reduced habituation and hyperactivity in adult mice [bib_ref] Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural..., Johansson [/bib_ref]. It has been suggested that PFOS and PFOA act as developmental neurotoxicants that mediate their effects on normal brain development, with consequences for cognitive and behavioral functions, through different mechanisms. Examples of those mechanisms are alteration in the dopaminergic system [bib_ref] Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural..., Johansson [/bib_ref] [bib_ref] Developmental neurotoxicity of perfluorinated chemicals modeled in vitro, Slotkin [/bib_ref] , elevated levels of proteins important for normal neuronal survival, growth and synaptogenesis, such as CaMKII, GAP-43, synaptophysin and tau, in the brain [bib_ref] Neonatal exposure to PFOS and PFOA in mice results in changes in..., Johansson [/bib_ref] , and induction of apoptosis of neuronal cells [bib_ref] Perfluorooctane sulfonate induces apoptosis of cerebellar granule cells via a ROS-dependent protein..., Lee [/bib_ref]. Although most of these findings were obtained from experiments on mice or rat derived cell lines that were exposed to extremely high levels of PFOS and PFOA compared to the low levels measured in the present study, other studies found that PFCs were detrimental to neurodevelopment at levels comparable to those observed in humans [bib_ref] Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural..., Johansson [/bib_ref] [bib_ref] Neonatal exposure to PFOS and PFOA in mice results in changes in..., Johansson [/bib_ref].
Our study is primarily comparable to the study by Fei and Olsen [bib_ref] Prenatal exposure to perfluorinated chemicals and behavioral or coordination problems at age..., Fei [/bib_ref] because both studies used measures of prenatal rather than postnatal exposure to PFCs. Fei and Olsen [bib_ref] Perfluorochemical (PFC) exposure in children: associations with impaired response inhibition, Gump [/bib_ref] found higher levels of PFOS and PFOA compared to those seen among pregnant women in other countries including the Nordic countries [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref] [bib_ref] Placental transfer of perfluorinated compounds is selective-a Norwegian Mother and Child sub-cohort..., Gutzkow [/bib_ref]. Consistent with that study, our results provide further indication that fetal exposure to PFCs at the present levels do not play a major role in having ADHD diagnosis at later age. Hoffman et al. [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] found an association between PFC serum concentrations and ADHD in children aged 12 to 15 years. Another study by Stein and Savitz [bib_ref] Serum perfluorinated compound concentration and attention deficit/hyperactivity disorder in children 5-18 years..., Stein [/bib_ref] on the relationship between self-reported ADHD and PFC levels in children in the same age range as for those in the study of Hoffman et al. [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] showed an association with PFHxS but not with the other PFCs even though both ADHD prevalence and exposure levels for PFCs were higher in the study by Stein and Savitz. Exposure to PFCs tends to be higher among newborns, toddlers, and children due to high uptake via food consumption, hand-to-mouth transfer of the PFCs from carpets, and through ingestion of dust [bib_ref] Estimating consumer exposure to PFOS and PFOA, Trudel [/bib_ref]. If the positive association between PFC exposure and self-reported ADHD found in the study by Hoffman et al. [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] was not due to a chance finding, that might indicate that postnatal exposure to PFCs, rather than prenatal exposure, is associated with ADHD. The present study has some limitations. Unfortunately, while 419 children with ADHD diagnosis were identified at the Department of Child and Adolescent Psychiatry, there were significant losses to get to the final study sample. The study was restricted to children born in Malmö with available obstetric and demographic information from the SMBR and stored cord blood samples in the biobank. About 50% of the identified children met these two inclusion criteria and were included in the study. The second limitation of the current study is the small number of ADHD cases. Although we would be able to detect an odd ratio of 1.8 and a difference in PFC levels of at least 0.20 standard deviations between cases and controls, it should be emphasized that the statistical power was not high enough to detect some minor associations and we could not accordingly rule out small effects. In addition, we lack information about other exposures that are significant for ADHD, such as exposure to mercury and lead. The PFC levels measured here might also be too low to trigger undesirable effects on the brain development.
During our study period, the clinical diagnostic criteria for ADHD were changed from the definition in DSM-III-R to the definition in DSM-IV, where DSM-IV is regarded as more inclusive. Thus, DSM-IV criteria yield a higher prevalence of ADHD [bib_ref] Variations in prevalence of attention deficit hyperactivity disorder worldwide, Skounti [/bib_ref]. Most individuals (93%-97.5%) who fulfill a diagnosis of ADHD according to DSM-III-R also fulfill the diagnostic criteria according to DSM-IV [bib_ref] DSM-IV field trials for attention deficit hyperactivity disorder in children and adolescents, Lahey [/bib_ref] [bib_ref] Correspondence between DSM-III-R and DSM-IV attention-deficit/hyperactivity disorder, Biederman [/bib_ref] [bib_ref] Comparison of diagnostic criteria for attention deficit disorders in a German elementary..., Baumgaertel [/bib_ref]. Individuals with ADHD according to DSM-IV that also fulfilled diagnostic criteria according to DSM-III-R were 85% [bib_ref] DSM-IV field trials for attention deficit hyperactivity disorder in children and adolescents, Lahey [/bib_ref] and 60% [bib_ref] Comparison of diagnostic criteria for attention deficit disorders in a German elementary..., Baumgaertel [/bib_ref]. Thus the overlap between ADHD diagnoses according to DSM-III-R and DSM-IV is considerable. The more inclusive diagnosis used in the latter part of the study probably includes some less severe cases which might slightly weaken possible statistical associations between exposure to PFCs and having an ADHD diagnosis.
The study has also several important strengths. First, unlike most of the previous studies on the associations between PFC levels and ADHD, our prospective study design is more reliable in the sense that it is based on clinical diagnosis of ADHD made at the Department of Child and Adolescent Psychiatry. Children were diagnosed at the same psychiatric clinic through the whole study period.
Second, the present study is based on analyzed blood samples from the fetal period, which we believe is the most susceptible exposure window, whereas in other studies, which were of a crosssectional nature, blood samples were collected from school-age children [bib_ref] Exposure to polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-15..., Hoffman [/bib_ref] [bib_ref] Serum perfluorinated compound concentration and attention deficit/hyperactivity disorder in children 5-18 years..., Stein [/bib_ref].
Third, we were able to account for important covariates; smoking during pregnancy, parity, and gestational age at birth that have been found to be associated with both PFC levels in pregnant women or infants [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref] [bib_ref] Determinants of fetal exposure to polyfluoroalkyl compounds in, Apelberg [/bib_ref] [bib_ref] Perfluorinated chemicals and fetal growth: a study within the Danish National Birth..., Fei [/bib_ref] [bib_ref] Correlations between prenatal exposure to perfluorinated chemicals and reduced fetal growth, Washino [/bib_ref] and ADHD diagnosis or symptoms [bib_ref] Sibship size, birth order, family structure and childhood mental disorders, Carballo [/bib_ref] [bib_ref] The relationship between attention deficit hyperactivity disorder and premature infants in Taiwanese:..., Chu [/bib_ref] [bib_ref] Associations between prenatal cigarette smoke exposure and externalized behaviors at school age..., Desrosiers [/bib_ref] [bib_ref] Maternal smoking and hyperactivity in 8-year-old children, Kotimaa [/bib_ref] [bib_ref] Effects of low birth weight, maternal smoking in pregnancy and social class..., Langley [/bib_ref]. Epidemiological findings have shown that prevalence of ADHD is higher among males [bib_ref] Molecular Genetics of Attention-Deficit/Hyperactivity Disorder, Faraone [/bib_ref]. It has also previously been shown that infant sex has no effect on the concentrations of PFCs [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref] [bib_ref] Perfluorinated chemicals and fetal growth: a study within the Danish National Birth..., Fei [/bib_ref] [bib_ref] Maternal exposure to perfluorinated acids and fetal growth, Hamm [/bib_ref]. Thus, infant sex was not considered as a potential confounder in the current data set.
In a previous study, we found that fetuses of mothers originating from a country other than Sweden, especially those from Middle East and sub-Saharan Africa, had lower PFC levels in the cord blood than fetuses of native Swedish mothers [bib_ref] Determinants of maternal and fetal exposure and temporal trends of perfluorinated compounds, Ode [/bib_ref]. Another study found higher odds of having an ADHD diagnosis for native Swedish children compared to children of mothers born outside Sweden [bib_ref] Perinatal, maternal, and fetal characteristics of children diagnosed with attention-deficit-hyperactivity disorder: results..., Gustafsson [/bib_ref]. Since the proportion of immigrants is relatively high in Malmö, no matching for the maternal country of birth might result in a false positive relationship between PFCs and ADHD. Human serum PFC levels showed an increasing pattern from the early 1970s through the late 1990s, followed by leveling out and a decreasing trend right after the phase-out of the production of PFOS and PFOS related compounds in 2002 [bib_ref] Historical Comparison of Perfluorooctanesulfonate, Perfluorooctanoate, and Other Fluorochemicals in Human Blood, Olsen [/bib_ref] [bib_ref] Polyfluoroalkyl chemicals in the U.S. population: data from the National Health and..., Calafat [/bib_ref] [bib_ref] Historical and geographical aspects of the increasing perfluorooctanoate and perfluorooctane sulfonate contamination..., Harada [/bib_ref] [bib_ref] Time trends and the influence of age and gender on serum concentrations..., Haug [/bib_ref] [bib_ref] Analysis of a homologous series of perfluorocarboxylates from American Red Cross adult..., Olsen [/bib_ref]. Diagnosis criteria for ADHD have been changed during the study period. We matched for the year of delivery to remove the effect of those differences in PFC levels and diagnosis on the results.
According to our findings, fetal exposure to PFOS, PFOA and PFNA was not associated with ADHD diagnosis in childhood.
[fig] Figure 1: Flowchart for the selection procedure of the children with attention deficit hyperactivity disorder and controls. doi:10.1371/journal.pone.0095891.g001 [/fig]
[fig] Figure 2: Boxplot of the umbilical cord concentrations of perfluorooctane sulfonate (PFOS) (a) and perfluorooctanoic acid (PFOA) (b) in cases having attention deficit hyperactivity disorder and controls. The extreme values of perfluorooctanoic acid for the ADHD cases, 48 and 36 ng/ml, and for the controls, 66, 49, 31, and 23 ng/ml, are not presented in the boxplot. doi:10.1371/journal.pone.0095891.g002 [/fig]
[table] Table 1: Median concentration (in nanograms/milliliters) of perfluorinated compounds by the maternal and infant demographic characteristics. [/table]
[table] Table 2: The crude and adjusted odds ratio with 95% confidence interval of attention deficit hyperactivity disorder and exposure to perfluorinated compounds.Abbreviations: ADHD, attention deficit hyperactivity disorder; PFOS, perfluorooctane sulfonate; PFOA, perfluorooctanoic acid; PFNA, perfluorononanoic acid; LOD, limit of detection.Adjusted for maternal active smoking, parity, and gestational age at birth. b Odds ratio is calculated for 1 ng/ml increase in umbilical cord serum concentration.Odds ratio is calculated for PFOS and PFOA concentrations at or above the 75 th percentile (75 th percentile for PFOS and PFOA were 9,1 ng/ml and 2,4 ng/ml, respectively). d Odds ratio is calculated for PFNA concentrations at or above the LOD (0.2 ng/ml). doi:10.1371/journal.pone.0095891.t002 [/table]
|
PTPN2 regulates the activation of KRAS and plays a critical role in proliferation and survival of KRAS-driven cancer cells
Edited by Alex TokerRAS genes are the most commonly mutated in human cancers and play critical roles in tumor initiation, progression, and drug resistance. Identification of targets that block RAS signaling is pivotal to develop therapies for RAS-related cancer. As RAS translocation to the plasma membrane (PM) is essential for its effective signal transduction, we devised a high-content screening assay to search for genes regulating KRAS membrane association. We found that the tyrosine phosphatase PTPN2 regulates the plasma membrane localization of KRAS. Knockdown of PTPN2 reduced the proliferation and promoted apoptosis in KRAS-dependent cancer cells, but not in KRAS-independent cells. Mechanistically, PTPN2 negatively regulates tyrosine phosphorylation of KRAS, which, in turn, affects the activation KRAS and its downstream signaling. Consistently, analysis of the TCGA database demonstrates that high expression of PTPN2 is significantly associated with poor prognosis of patients with KRAS-mutant pancreatic adenocarcinoma. These results indicate that PTPN2 is a key regulator of KRAS and may serve as a new target for therapy of KRAS-driven cancer.
RAS proteins are small GTPases that regulate diverse cellular processes, including proliferation, differentiation, migration, apoptosis, and senescence. Mammalian cells mainly express three RAS genes that encode four highly homologous proteins: HRAS, NRAS, KRAS4A, and KRAS4B. KRAS4A and KRAS4B result from an alternative splicing at the C terminus of the KRAS gene. Because KRAS4B is the predominant splice variant of KRAS, it is referred to as KRAS hereafter.
RAS genes are the most frequently mutated oncogene in cancer, nearly 20-30% of human malignancies carry RAS gene mutations. Among the RAS gene family, KRAS is the most commonly mutated, which occurs in 71% of pancreatic, 29% of colorectal, and 18.6% of lung carcinomas. It has been shown that mutated KRAS not only plays pivotal roles in cancer initiation, but also contribute to several hallmarks of human cancer. Moreover, inhibition of activated KRAS could delay tumor progression both in vitro and in vivo. These observations prompted many groups to target either mutant KRAS directly or downstream effectors.
Thus far, directly targeting oncogenic KRAS only succeeded in one certain form, KRAS G12C, which comprises only 12% of KRAS mutations in all human cancers, so it is still needed to develop new target molecules for other oncogenic KRAS. Inhibiting protein-protein interactions and KRAS localization are novel approaches to target mutant KRAS and block oncogenic KRAS signaling, although the efficacy in clinical is still unknown. Targeting the KRAS effector signaling pathways could also prove efficacious in treating tumors with KRAS mutations, as the inhibitors have entered clinical trials demonstrating promising clinical activity in KRAS mutant tumor. However, toxicities associated with their sustained inhibition, variable responses rates, and acquired adaptive resistance due to the activation of other kinases are limiting the efficacy and the clinical progression of these compounds as monotherapy treatment. Hence it remains an urgent need to identify new target strategies to block oncogenic KRAS signaling. KRAS interacts with downstream effectors only when it associates with the plasma membrane (PM), so inhibition of KRAS localization is a valid therapeutic approach to block signal transmission by oncogenic KRAS.
KRAS is synthesized as cytosolic proteins and gains affinity for the PM through post-translational modification of its carboxyl terminal CAAX motif (CVIM sequence) by farnesyltransferases (FTase) or geranylgeranyltransferase (GGTase), followed by cleavage of the VIM residues by RAS converting enzyme 1 (RCE1) and methyl esterification of the farnesylated cysteine residue by isoprenylcysteine carboxyl methyltransferase (ICMT). Given the critical role of prenylation for KRAS membrane association and neoplastic transformation, and farnesylation of KRAS by FTase is the first step in the KRAS post-translational modification, FTase is the ideal target for KRAS-driven cancers. However, FTase inhibitors (FTIs) failed to show high efficacy as expected in clinic. The main reason is that KRAS protein can undergo alternative prenylation by GGTase in the presence of FTIs. Inhibitors targeting FTase and GGTase in combination have been proved too toxic to be clinically useful.
Despite the clinical failure of FTIs, inhibition of KRAS-PM interactions remains an attractive therapeutic approach to abrogate the KRAS oncogenic activitysiRNA screening to identify key molecules required for oncogenic KRAS plasma membrane association. We show that, for the first time, protein-tyrosine phosphatase non-receptor type 2 (PTPN2) regulates the KRAS plasma membrane association and plays an important role in KRAS-dependent cancer cell proliferation and survival. Mechanistic studies demonstrate that PTPN2 negatively regulates tyrosine phosphorylation of KRAS, which, in turn, affects the activation KRAS and its downstream signaling. Our data suggest that PTPN2 could serve as a potential therapeutic target for KRAS-driven cancer.
# Results
Identification of PTPN2 as a regulator for KRAS plasma membrane association by an siRNA screening
To identify genes required for KRAS membrane trafficking, we developed an image-based screening assay that monitors the degree of KRAS G12D membrane association . To this end, we established a GFP-fused KRAS G12D (GFP-KRAS G12D ) human embryonic kidney (HEK) 293T cell line stably expressing GFP-KRAS G12D . We found that the majority of GFP-KRAS G12D proteins are localized to the PM . We then screened a 147-siRNA sublibrary enriched with targets that regulate protein translocation, using an Opera Phenix High-Content Screening System (the siRNA-target genes and sequence was listed in . The PDE6D siRNA was used as a positive control. After cells were treated with arrayed siR-NAs for 72 h in 96-well-plates, one siRNA, PTPN2 siRNA, significantly changed the ratio of PM-bound GFP-KRAS G12D over cytosolic GFP-KRAS G12D .
We used additional PTPN2 siRNA, as well as KRAS siRNA, from GenePharma to confirm the screening result. Two different PTPN2 siRNAs (PTPN2 siRNA-432 and PTPN2 siRNA-1161) could effectively decrease the level of PTPN2 in GFP-KRAS G12D HEK 293T cells . We then checked the level of GFP-KRAS G12D on the PM using a membrane protein extraction kit, and found that silencing PTPN2 reduced the translocation of KRAS G12D to the PM by approximately 30% compared with the negative control (NC) . Meanwhile, the level of KRAS G12D in the cytoplasm was slightly increased . These data suggest that PTPN2 plays an important role in regulating the plasma membrane association of KRAS G12D .
## Ptpn2 is required for kras-dependent tumor cell growth
We next investigated the effect of PTPN2 on cancer cell growth, using pooled siRNAs to knockdown PTPN2 in a panel of five mutant KRAS-harboring human cancer cell lines (H460, PaTu8988T, HCT-116, A549, and DLD-1) and two KRAS wildtype human cancer cell lines (HT-1080 and SK-MEL-30). Pooled siRNAs to knockdown KRAS were used as a positive control. Cells were treated with siRNAs for 72 h, followed by the CellTiter-Glo ® luminescent cell viability assay. As shown in, silencing PTPN2 with pooled siRNAs markedly attenuated the proliferation of H460 (lung), PaTu8988T (pancreatic), HCT-116 (colon), and HT-1080 (fibrosarcoma) cells, and slight but significantly growth inhibition of A549 (lung) and SK-MEL-30 (skin). However, it is not the case in DLD-1 (colon) cells. Interestingly, the growth inhibitory effect of knocking down PTPN2 was consistent with that of knocking down KRAS, suggesting that PTPN2 plays a critical role in KRAS oncogenic signaling. Accordingly, DLD-1, which is KRAS-independent, was less sensitive to the knockdown of either PTPN2 or KRAS. Both two pooled siRNAs effectively decreased the expression of PTPN2 and KRAS in all seven cell lines.
We further confirmed the effect of PTPN2 on the three mutant KRAS-dependent cell lines by using two different siRNAs, respectively. Both siRNAs were found to reduce the expression of PTPN2 and suppress the proliferation of the cell lines, C and D).
## Ptpn2 is required for kras-dependent tumor cell survival
For more accurate detection of cell proliferation, we examined the effect of PTPN2 and KRAS knocking down on cell proliferation by a BrdU assay. Consistent with the results of Cell-Titer-Glo ® luminescent cell viability assay, the fraction of BrdU-positive cells was significantly decreased in all these cell lines described above except for DLD-1. It is notable that the effect of PTPN2 knockdown is less than that of KRAS knockdown in H460, PaTu8988T, and A549, suggesting that PTPN2 has limited activity on the proliferation signaling of KRAS in these cells.
We also examined the effect of PTPN2 on cell survival. All these tumor cell lines described above were treated with PTPN2, KRAS, or control siRNAs for 72 h, followed by staining with Annexin V-APC and propidium iodide. Flow cytometric analysis showed that the percentage of apoptosis in these KRAS-dependent tumor cell lines was significantly higher than that of controls. The effect of PTPN2 knockdown is no less than that of KRAS knockdown in KRAS-dependent tumor cell lines, suggesting that PTPN2 is required for the cell survival signaling of KRAS.
## Ptpn2 regulates the activation of oncogenic kras and its downstream signaling
To gain insights into the mechanism by which PTPN2 regulates the proliferation and survival of KRAS-dependent tumor cells, we first checked whether PTPN2 affects the KRAS activation, using an RAS-GTP pulldown assay. As shown in, the level of GTP-KRAS G12D was significantly decreased in GFP-KRAS G12D expressing 293T cells treated with PTPN2 siRNA, either pooled or two different single siRNAs, compared with that in GFP-KRAS G12D expressing 293T cells treated with scrambled siRNA. Similarly, the level of GTP-bound KRAS in H460 cells was also decreased by treating with pooled PTPN2 siRNAs.
We next tested whether PTPN2 affects the activation of KRAS downstream signaling pathway. As shown in, both KRAS and PTPN2 knockdown could dramatically decrease the phosphorylation levels of MEK and ERK in KRASdependent tumor cell lines HCT-116, PaTu8988T, and H460. Collectively, these data demonstrate that PTPN2 is required for the full activation of oncogenic KRAS and its downstream signaling pathway.
We also examined the protein stability of GFP-KRAS G12D in HEK 293T cells in the presence of protein synthesis inhibitor cycloheximide (CHX, 50 mg/ml). As shown in, the endogenous KRAS protein was markedly susceptible to PTPN2 knockdown, whereas only a marginal change was noted for the exogenous GFP-KRAS G12D protein. The different effects of PTPN2 on the stability between endogenous and exogenous KRAS may be partly explained by the exogenous KRAS is the mutant form, whereas the endogenous KRAS is the wildtype form. We also examined the effect of PTPN2 knockdown on half-life of the endogenous KRAS protein in the KRAS mutant cell line. We transfected cancer cells carrying KRAS-Q61K (H460) mutant with either control or PTPN2 siRNA. As shown in, C and D, mutant KRAS half-life changed marginally in PTPN2 knockdown cells. Taken together, these data demonstrate that PTPN2 is not involved in maintaining mutant KRAS protein stability.
## Ptpn2 regulates the level of tyrosine phosphorylation of kras
It has been shown that tyrosine phosphorylation of KRAS reduces its signal transduction activity, whereas tyrosine-phosphorylated KRAS can be dephosphorylated by SHP2 (encoded . PTPN2 regulates KRAS membrane association. A, schematic representation of the screening. Arrayed siRNAs were transfected into GFP-KRAS G12D 293T using RNAiMAX. After the 72-h incubation, images were acquired for 17 fields with a 340 water immersion objective using the GFP-confocal mode in an Opera Phenix High-Content Screening System. B, PTPN2 siRNA induced GFP-KRAS G12D mislocalization. Representative images (GFP channel only) of GFP-KRAS G12D HEK 293T cells transfected with PTPN2 smartpool ON-TARGETplus siRNAs. Scale bar: 50 mm. C, Western blot analysis of the knockdown efficiency of PTPN2 in GFP-KRAS G12D HEK 293T after transfection with the indicated siRNAs. The immunoblot data are representative of at least three independent experiments. D, GFP-KRAS G12D HEK 293T cells were treated with pooled PTPN2 siRNAs or NC siRNA for 72 h. To detect exogenous GFP-fused KRAS proteins, membrane proteins, and cytoplasmic proteins were isolated, followed by Western blot analysis with an anti-GFP antibody. (Na 1 -K 1 )-ATPase was used as the input of membrane proteins control, and GAPDH was used as the input of cytoplasmic proteins control. E, quantitation of data in C are shown. Signal intensity was quantified for GFP-KRAS G12D in the membrane fraction and cytoplasmic fraction (n = 4 independent biological experiments depicted as normalized intensity for siPTPN2 over the NC control; mean 6 S.D.; two-tailed t test. *, p , 0.05).
by PTPN11). PTPN2 is a member of the PTP family and has been shown to dephosphorylate several tyrosine kinases. We tested whether PTPN2 regulates KRAS signaling through dephosphorylating KRAS. Myc-KRAS G12D HEK 293T cell lysates were immunoprecipitated with anti-Myc antibody followed by anti-pTyr immunoblot. We found the level of phosphorylated KRAS in Myc-KRAS G12D HEK 293T cells was significantly increased after knockdown of PTPN2. These results indicate PTPN2 is also a tyrosine phosphatase for KRAS. To confirm that the observed effect on KRAS phosphorylation was attributable specifically to PTPN2 knockdown, we generated two siRNA-resistant wildtype PTPN2 isoforms (PTPN2-1-R and PTPN2-2-R) and a phosphatase-inactive mutant (PTPN2-D182A) (Asp-182 changed to Ala) construct for rescue experiments. Transfection of PTPN2-1-R and PTPN2-2-R constructs into PTPN2 siRNA (1161)-treated cells markedly increased PTPN2 expression. More importantly, overexpression of PTPN2-2-Res but not PTPN2-2-D182A-Res, nor PTPN2-1-Res, decreased the level of tyrosine phosphorylation of KRAS. This result indicates that PTPN2 regulates tyrosine phosphorylation of KRAS through its tyrosine phosphatase activity and that PTPN2-2-Res is the primary isoform to carry out this function.
High level PTPN2 expression is associated with poor prognosis of pancreatic adenocarcinoma Our data show that PTPN2 plays an important role in the maintenance of KRAS-dependent tumor cells. To determine the impact of PTPN2 expression in KRAS-related cancers, we analyzed clinical databases. We separated KRAS mutation, KRAS expression, PTPN2 expression, and clinical information of KRAS-related cancers from TCGA cohorts . We first performed overall survival analysis to compare the survival of patients harboring mutant or wildtype KRAS in pancreatic adenocarcinoma (PAAD), lung adenocarcinoma (LUAD), and colorectal adenocarcinoma (COAD). As shown in, KRAS mutation is associated with poor prognosis of PAAD, but not that of LUAD or COAD. We then analyzed the impact of PTPN2 on survival of patients containing KRAS mutations. As shown in, high expression of PTPN2 is significantly associated with poor prognosis in KRAS-mutant PAAD patients (p = 0.0058), but not in patients with LUAD or COAD. In KRAS-WT patients, the high expression of KRAS mRNA is significantly associated with poor prognosis in PAAD (p = 0.0112) and LUAD (p = 0.0105) but not in COAD (p = 0.1145). However, PTPN2 expression levels are not significantly associated with survival in KRAS high expression patients or KRAS low expression patients with PAAD (p = 0.9121 and p = 0.2106, respectively). These clinical data support our conclusion that PTPN2 plays an important role in KRAS-mutant-dependent tumors.
# Discussion
It has been known that 20-30% human cancers, including a high percentage of pancreatic, lung, and colorectal cancers, are driven by mutations in KRAS. Genetically engineered mouse models have shown that Kras mutation could faithfully recapitulate the progression of the human cancer, and a mutated form of Kras is sufficient to initiate pancreatic and lung cancer. Furthermore, silencing Kras with siRNA in these genetically engineered mouse models significantly prolonged the survival of mice. Although scientists have made great progresses in the last 3 decades toward understanding the KRAS signaling pathways, drug development in blocking KRAS function has been ineffective. Based on the fact that KRAS regulates signaling pathways for cell growth only when associated with the PM, targeting KRAS membrane translocation has become an alternative approach. In this study, we screened a custom-designed siRNA library using an image based highcontent screening. We found that PTPN2 is required for the effective PM localization of KRAS, and PTPN2 is required for KRAS-dependent cell survival and, to a lesser extent, proliferation. Mechanistically, PTPN2-negative regulated KRAS tyrosyl phosphorylation. In addition, analysis of the data from TCGA, we found that high PTPN2 expression is associated with poor prognosis of KRAS-mutant PAAD patients.
PTPN2, also known as T cell protein-tyrosine phosphatase (TCPTP), is a non-receptor phosphatase that is ubiquitously expressed, and plays critical roles in T cell-mediated immunity and inflammation. The role of PTPN2 in tumors has just begun to be investigated in recent years. It has been reported that PTPN2 is frequently mutated and repressed in T-ALL, and further experiments reveal that PTPN2 negatively regulates the JAK/STAT signaling. Other studies have found that PTPN2 can act as an oncogene. PTPN2 played
## Ptpn2 is important for kras-driven cancer
tumor-promoting functions in B-cell lymphomas, and Ptpn2 depletion decreased murine B-cell lymphoma cell proliferation and completely abolished the cancer in vivo. It was also shown that PTPN2 expression levels are strongly associated with prognosis in patients with glioma and glioblastoma. Patients with a high expression of PTPN2 tend to have a poor prognosis, suggesting that PTPN2 can promote tumor development. Here, we show for the first time that PTPN2 plays a tumor-promoting function in KRAS-driven cancer.
It has been proven that KRAS is phosphorylated via Src, which alters the conformation of switch I and II regions, profoundly reducing the KRAS' binding affinity to RAF. In contrast, SHP2 dephosphorylates KRAS and thereby restores the affinity to RAF, which enhances the KRAS-mediated MAPK pathway. Inhibition of SHP2 promotes accumulation of phosphorylated KRAS and subsequently suppressing KRAS-RAF-MAPK signaling. Here, PTPN2 is identified as another phosphatase that dephosphorylates KRAS and regulates the activation of KRAS and its downstream signaling.
It has been suggested that the type of mutations in KRAS may have an influence on its ability to transform and the drug responses of cancer patients. SHP2 inhibitor RMC-4550, for example, has shown a potent inhibitory effect on the cancer cell lines bearing missense mutations in KRAS at Gly-12, but not Gly-13, and Gln-61. In this study, we found it is interesting that the requirement of PTPN2 for KRAS activation is independent of the mutant KRAS isoforms, which included G12V (PaTu8988T), G13D (HCT-116), and Q61K (H460). Both PTPN2 and SHP2 can dephosphorylate KRAS, but their negative regulatory signal molecules are somewhat different. The underlying mechanism requires subsequent experiments to explore.
Recently, studies have shown that Kras-mutant-bearing cancers display features with reduced T helper 1 cells as well as reduced infiltration of cytotoxic cells via recruiting the myeloid-derived suppressor cells, which made Kras-mutant tumor resistant to immune checkpoint blockade (ICB) therapy. One mechanism is that Kras-driven cancers could down- PTPN2 is important for KRAS-driven cancer regulate the interferon-g (IFN-g) and IFN-a responses, which are crucial in anti-PD-1 therapy in patients with cancer. It is interesting that loss of Ptpn2 results in an increase in number and activation of CD81 T cells, and enhancing of IFN-g-mediated effects on antigen presentation. Therefore Ptpn2 has recently been identified as a novel cancer immunotherapy target in a CRISPR screening in vivo, where deletion of this gene increased the efficacy of immune checkpoint blockade therapy in melanoma. Moreover, one study has demon-strated that myeloid cell-specific loss of Ptpn2 promotes inflammasome activation, resulting in protection from colorectal cancer. Thus, inhibition of PTPN2 could suppress KRAS cancer, whereas enhancing tumor immunity.
In summary, we identified PTPN2 is a key regulator of KRAS activation and signaling transduction. The results indicate that PTPN2 may be a novel therapeutic target for KRAS-driven cancers.
# Materials and methods
## Cell culture
Human lung (H460 and A549), pancreatic (PaTu8988T), colon (HCT-116 and DLD-1), fibrosarcoma (HT1080), and skin (SK-MEL-30) cancer cell lines and the HEK 293T cell line were obtained from the American Type Culture Collection (ATCC) and authenticated by ATCC and DSMZ using Short Tandem Repeat profiling analysis. These cells were detected periodically to ensure mycoplasma-free cells. HCT-116 and DLD-1 cells were cultured in RPMI-1640 medium (BasalMedia, Shanghai, China); HEK 293T, H460, PaTu8988T, and A549 cells were cultured in Dulbecco's modified Eagle's medium (BasalMedia, Shanghai, China), HT-1080 and SK-MEL-30 were cultured in minimal essential medium (BasalMedia, Shanghai, China). All media were supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA), 13 penicillin/streptomycin, and the cells were cultured in a humidified 5% CO 2 incubator at 37°C.
## Dna constructs
GFP-fused KRAS G12D , Myc-fused KRAS G12D , PTPN2 siRNA-1161-resistant (PTPN2-Res), and catalytically inactive mutation (PTPN2-D182A) plasmids were generated as previously described. All constructs were confirmed by DNA sequencing before use.
## Hek 293t cell line stably expressing gfp-kras g12d
HEK 293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C, 5% CO 2 . Transfection of GFP-KRAS G12D was carried out using Lipofectamine 3000 reagent (Invitrogen) per the manufacturer's protocol. GFP-positive cells were sorted twice on a fluorescent-activated cell sorting (FACS) machine, and further placed on a 96-well flat-bottom tissue culture plate, one GFP positive cell/well. Cells were further cultured for 2 weeks, detached with trypsin-EDTA, and then cultured in a 6-wellplate. The GFP-KRAS G12D expressing cell lines were used for further experiments. then screened using an Opera Phenix High-Content Screening System (PerkinElmer). Images were acquired for 17 fields with a 340 water immersion objective using the GFP-confocal mode. The siRNA was listed in .
# Western blot analysis
To prepare whole-cell lysates, cells were washed twice with ice-cold phosphate-buffered saline (PBS) and lysed in SDS lysis buffer (100 mM Tris-HCl, 2% SDS, 10% glycerol, 50 mM dithiothreitol, 0.1% bromophenol blue, pH 6.8) supplemented with protease inhibitor cocktail and 2 mM phenylmethylsulfonyl fluoride. Proteins from the lysates were separated by SDSpolyacrylamide gel electrophoresis and Western blotted with the following antibodies:
## Flow cytometry
The thymidine analog 5-bromo-2-deoxyuridine (BrdU) incorporation assay was performed according to the standard protocol of the manufacturer (BD Pharmingen TM BrdU Flow Kits; catalog number 559619). Cells were transfected with siRNA via reverse transfection using RNAiMAX reagent (Invitrogen, catalog number 13778) at a final concentration of 50 nM and seeded in 6-well-plates for 72 h. BrdU (10 mM) was added and incubated further for 3 h. Then, cells were washed three times with 13 PBS and fixed and permeabilized the cells with BD Cytofix/Cytoperm Buffer. Cell were further permeabilizated with Cytoperm Permeabilization Buffer Plus for 10 min on ice and re-fixed with Cytofix/Cytoperm Buffer for 10 min on ice. After incubation with DNase for 1 h at 37°C, cells were stained with FITC-conjugated anti-BrdU antibody for 30 min at room temperature. Before detecting by flow cytometry, cells were stained with nucleolus dye 7-amino-actinomycin D for 15 min. The percentage of BrdU-positive cells was counted and reckoned by using the GraphPad Prism 5.
Apoptosis was detected and quantified using the Annexin V Apoptosis Detection Kit APC (eBioscience, Waltham, MA, USA, catalog number 88800772). Briefly, after treatment with siRNA for 72 h, cells were harvested and washed with ice-cold PBS, and then suspended in Annexin binding buffer. Subsequently, cells were incubated with Annexin V-APC and propidium iodide for 15 min at room temperature in the dark and immediately analyzed using a BD LSRFortessa flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). Data were analyzed with the FlowJo (Tree Star) software.
## Sirna-mediated knockdown of kras and ptpn2
Cells were transfected with siRNA via reverse transfection using Dharmafect 1 Transfection Reagent (Dharmacon, catalog number T2001-01) or RNAiMAX reagent (Invitrogen, catalog number 13778) at a final concentration of 50 nM and seeded in 6-well-plates for Western blotting or apoptosis assays and 96-well-plates for cell viability assays. Transfected cells were collected 72 h after transfection for Western blot analysis as described above, and after 72 h for the CellTiter-Glo ® Luminescent Cell Viability Assay (Promega) as described below. The siRNAs were obtained from GenePharma Corporation (Shanghai, China) and the sequences used for KRAS siRNA and PTPN2 siRNA experiments are as follows: negative control, sense UUCUCCGAACGUGUCACGUTT and antisense ACG-UGACACGUUCGGAGAATT; KRAS siRNA-244, sense GCC-UUGACGAUACAGCUAATT and antisense UUAGCUGUA-UCGUCAAGGCTT; KRAS siRNA-585, sense GGACUUAG-CAAGAAGUUAUTT and antisense AUAACUUCUUGCUAA-GUCCTT; KRAS siRNA-643, sense GGUGUUGAUGAUGC-CUUCUTT and antisense AGAAGGCAUCAUCAACACCTT; PTPN2 siRNA-432, Sense CAAAGGAGUUACAUCUUAATT and antisense UUAAGAUGUAACUCCUUUGTT; and PTPN2 siRNA-1161, sense GCCUUUGAUCAUUCACCAATT and antisense UUGGUGAAUGAUCAAAGGCTT.
## Cell viability assay
Cell viability assays were carried out using the CellTiter-Glo ® Luminescent Cell Viability Assay as previously described. H460, PaTu8988T, HCT-116, A549, DLD-1, HT-1080, and SK-MEL-30 cells were transfected with siRNA via reverse transfection using Dharmafect 1 or RNAiMAX at a final concentration of 50 nM and seeded in regular 96-well-plates at a density of 5 3 10 3 to 1 3 10 4 cells/well. After 72 h incubation, cell viability was measured using the CellTiter-Glo reagent. The luminescence was detected using an Envision plate reader (PerkinElmer).
## Kras activity assay
KRAS activity was determined using RAS activation assay kit (EMD Millipore, 17-218, Burlington, MA, USA) according to the manufacturer's protocol with minor modifications. Briefly, cells were lysed in Pierce IP Lysis Buffer (Thermo Scientific, catalog number 87788) supplemented with protease and phosphatase inhibitors (Roche Applied Science), and lysates were further incubated with 5 mg of RAF-1 RBD-agarose beads for 8 h at 4°C. After washing the agarose beads three times with Pierce IP Lysis Buffer, the activated KRAS (GTP-RAS) bound to RAF-1 RBD-agarose beads was released by the addition of SDS lysis buffer. Finally, samples were subjected to Western PTPN2 is important for KRAS-driven cancer blotting analysis as previously described and blots were probed using an anti-KRAS antibody (Sigma, 1:1000).
## Immunoprecipitation and immunoblotting
HEK 293T cells were transfected with siRNA via reverse transfection using Dharmafect 1 at a final concentration of 50 nM and seeded in regular 6-well-plates at a density of 1 3 10 6 cells/well. After 24 h incubation, cells were transiently transfected the Myc-tagged KRAS G12D plasmid using the Lipo6000 Transfection Reagent (Beyotime, Shanghai, China, catalog number C0526) per the manufacturer's protocol and further cultured for 48 h. Cells were lysed in Pierce IP Lysis Buffer supplemented with protease and phosphatase inhibitors (Roche), and lysates were further incubated with 10 mg of anti-Myc tag antibody (agarose) (Abcam, Cambridge, England, UK, catalog number ab1253) or IgG-agarose beads (Abcam, catalog number ab104155) for 8 h at 4°C. After washing the agarose beads three times with Pierce IP Lysis Buffer, the Myc-tagged-KRAS bound to agarose beads was released by the addition of SDS lysis buffer. Finally, samples were subjected to Western blotting analysis as previously described and blots were probed using an anti-phosphotyrosine antibody (Cell Signaling Technology, catalog number 9411, 1:1000) or anti-Myc antibody (Cell Signaling Technology, 1:1000).
## Kras protein stability
GFP-KRAS G12D HEK 293T cells and H460 were transfected with siRNA via reverse transfection using RNAiMAX at a final concentration of 50 nM and seeded in regular 12-well-plates at a density of 2.5 3 10 5 cells/well. Forty-eight h later fresh medium containing cycloheximide (50 mg/ml) was added. At the indicated time intervals, cells were washed twice with ice-cold PBS and lysed in SDS lysis buffer supplemented with protease inhibitor cocktail and 2 mM PMSF. Finally, samples were subjected to Western blotting analysis as previously described.
Analyses of the association of the KRAS mutation, KRAS expression, and PTPN2 expression in PAAD, LUAD, or COAD clinical outcome
The KRAS mutation, PTPN2 expression, and clinical information of KRAS-related cancers were extracted from the TCGA database. This dataset contains survival data with clinical information, KRAS mutations, KRAS mRNA expression counts, and PTPN2 mRNA expression counts. The groups were separated by the mean expression level of the group. Overall survival stratified by expression levels of the gene of interest was evaluated using Kaplan-Meier analysis, and comparisons between groups were evaluated using log-rank tests. p , 0.05 was considered statistically significant.
# Statistical analysis
Data of continuous variables are presented as mean 6 SD. Comparisons between treatments were analyzed by one-way ANOVA followed by Tukey's test using GraphPad Prism 5. p , 0.05 was deemed statistically significant.
## Data availability
All data relevant to this study are included within this manuscript. |
Cholesteatoma of external auditory canal: a case report
The authors present a case of cholesteatoma of external auditory canal (CEAC) with extensive invasion of mastoid; ossicle chain and tympanic membrane remained intact. The only symptom was chronic otorrhea. Diagnosis was based on clinical elements and CT scan was used to measure pathology and program surgery. Treatment was modified radical mastoidectomy associated with meatoplasty. Due to the insidious character of CEAC and the proximity with important structures of the external auditory canal, it must be always considered in differential diagnosis for lesions of external auditory canal. This case report intended to review clinical and surgical aspects of treatment of CEAC and present our approach in a case with severe lesions. Rev Bras Otorrinolaringol. V.71, n.1, 91-3, jan./feb. 2005
The authors present a case of cholesteatoma of external auditory canal (CEAC) with extensive invasion of mastoid; ossicle chain and tympanic membrane remained intact. The only symptom was chronic otorrhea. Diagnosis was based on clinical elements and CT scan was used to measure pathology and program surgery. Treatment was modified radical mastoidectomy associated with meatoplasty. Due to the insidious character of CEAC and the proximity with important structures of the external auditory canal, it must be always considered in differential diagnosis for lesions of external auditory canal. This case report intended to review clinical and surgical aspects of treatment of CEAC and present our approach in a case with severe lesions.
# Introduction
Cholesteatoma of the external auditory canal (CEAC) is a rare condition affecting especially elderly people. The progress of the disease is slow and the symptoms are not too evident. This leads to late diagnosis, with progressive bone lysis affecting important circumjacent structures 1,2 .
The authors present a case of CEAC with extensive invasion of the mastoid, exposing the lateral sinus and the dura. Hearing acuity and structures of the tympanic cavity were unaffected.
We discuss the possible causes and treatment of choice.
## Literature review
In clinical practice of ear diseases, cholesteatoma occurs on the tympanum-mastoid segment. It rarely originates from the external auditory canal (EAC), whose incidence is about 0.1-0.5% in new patients with ear problems 1-3 .
In 1850 2,4 , Toynbee was the first author describing that cholesteatoma originates from the external auditory canal as "epidermal sheets". Up to 1980, CEAC and keratosis obturans (KO) were considered different presentations of the same disease. Pipergerdes et al.described CEAC and keratosis obturans (KO) as two distinct clinical-pathological processes: KO as keratin accumulation in the EAC; and CEAC as bone erosion resulting from squamous tissue on a specific spot of the EAC 2 .
In CEAC, lateral epithelium desquamation is affected. Keratin debris are blocked, causing local erosion and bone lysis, which can be severe 2,5 . Holt 2,3,6,7 listed situations in which the disease can be diagnosed: postoperative period of ear surgery; EAC trauma; EAC obstruction -due to osteoma or stenosis, for instance; idiopathic -possibly, in these cases, local periostitis may originate the process of hyperkeratosis.
Diagnosis is made through physical examination. There is otalgia, otorrhea, hearing acuity is unaffected; otoscopy shows an intact tympanic membrane and erosion on a distinct spot of the external auditory canal [bib_ref] Cholesteatoma of External Auditory Canal, Coelho [/bib_ref] [bib_ref] Primary Cholesteatoma of the External Auditory Meatus, Segal [/bib_ref] [bib_ref] Ear Canal Cholesteatoma, Holt [/bib_ref].
Otorhinolaryngologists must be aware that cholesteatoma may affect circumjacent structures (lateral sinus, facial nerve, posterior cranial fossa). Therefore, a CT scan is recommended for all patients [bib_ref] CT Appearances of External Ear Canal Cholesteatoma, Malcom [/bib_ref] [bib_ref] Cholesteatoma of External Auditory Canal, Coelho [/bib_ref].
Differential diagnosis for CEAC should include necrotizing external otitis, tumors and KO. The latter is more frequent in young adults presenting severe otalgia and bilateral sensorial hearing loss. Differently from CEAC, which usually affects older patients, KO is unilateral, there is otorrhea, chronic pain and hearing acuity is preserved. In KO, the EAC has a keratin plug; otoscopy, following keratin plug removal, shows stenosis, redness and granulation tissue. In CEAC, there is an epidermal diverticulum on the lowest part of the canal, with epidermal debris and otorrhea; the other parts of the EAC are normal [bib_ref] CT Appearances of External Ear Canal Cholesteatoma, Malcom [/bib_ref] [bib_ref] External Ear Canal Cholesteatoma, Birzgalis [/bib_ref].
The treatment of CEAC may be clinical or surgical. The former is carried out through local flushing and topic antibiotics. The latter is through resection of the cholesteatoma and necrotic bone [bib_ref] Primary Cholesteatoma of the External Auditory Meatus, Segal [/bib_ref] [bib_ref] External Canal Cholesteatoma, Vrabec [/bib_ref] [bib_ref] Cholesteatoma of the External Ear Canal, Farrior [/bib_ref] [bib_ref] Ear Canal Cholesteatoma, Holt [/bib_ref].
Clinical treatment is indicated whenever the lesion is limited and there is no otalgia, in cases that surgery is contraindicated, in patients refusing surgical procedures and, when observation without treatment is an option to evaluate progression 2,6 .
Surgical treatment is indicated in the following situations: chronic pain (despite treatment), frequent infection (there is risk for bacterial resistance), facial palsy or chronic dizziness, progression during follow-up, involvement of hypotympanum, jugular or mastoid (showed in CT scans), diabetes mellitus or immunosuppression (predisposition to necrotizing external otitis) 2,4,6,9 .
## Case report
Patient A. F., male, 59 years old. Visited the otorhinolaryngological outpatient facility showing otorrhea, otalgia and ear fullness on the right for ten (10) years. Otoscopy on his right ear showed epidermal debris in the canal. Following debris removal, the tympanic membrane was intact, and there was severe erosion of the posteroinferior quadrant of the EAC. Otoscopy of the left ear was normal. Bilateral pure tone audiometry was normal.
We suspected of CEAC and requested a CT scan of temporal bones. CT scan showed erosion of the posterior wall of the EAC, mastoid involvement and exposure of the lateral sinus and meninges [fig_ref] Figure 1: Temporal bone axial CT scan showing contact of the cholesteatoma's capsule with... [/fig_ref].
Transoperative observation revealed erosion of the entire posterior wall of the EAC, except for a thin bone band close to the tympanic ring [fig_ref] Figure 2: Transoperative picture of modified radical mastoidectomy showing intact TM and CEAC affecting... [/fig_ref]. The tympanic membrane, ossicle chain and facial nerve canal were preserved.
We conducted modified radical mastoidectomy type I plus meatoplasty. The patient showed good outcomes and normal hearing, without recurrences, in a postoperative period of two years.
# Discussion
CEAC, due to its slow and insidious progression, may severely affect structures of the middle ear at the time of presentation. In our case report, the lesion was severe at first presentation, and we were not able to visualize its limits through otoscopy. CT studies were important to determine the severity of the disease and involvement of other structures; therefore, CT scan is indicated whenever CEAC is suspected.
The disease had possibly appeared spontaneously, since the patient did not have any predisposing factor to develop CEAC.
Mastoid invasion is one of the indications for surgery. After the patient agreed to undergo surgery, we performed modified radical mastoidectomy and wide meatoplasty; thus, we could achieve good control of the cavity during the postoperative period.
We were able to preserve the tympanic cavity and its structures; therefore, the surgical procedure was able to cure and preserve the patient's hearing acuity. It is worth mentioning that, in spite of the increased progression of the disease, there was no involvement of the facial nerve, inner ear or nerves emerging from the jugular foramen.
## Closing remarks
Otorhinolaryngologists must suspect of CEAC in cases of chronic otorrhea and intact tympanic membrane. After diagnosis of CEAC, CT scans of the temporal bone must be performed in all patients to evaluate the real severity of the disease and involvement of circumjacent structures (jugular bulb, cranial nerves and structures of the tympanic cavity). In most cases, the treatment of choice is surgery, whose main purpose is to eradicate the lesion and, if possible, preserve the patient's hearing acuity. The modified radical mastoidectomy is the most common surgical procedure carried out in these cases.
[fig] Figure 2: Transoperative picture of modified radical mastoidectomy showing intact TM and CEAC affecting the mastoid. [/fig]
[fig] Figure 1: Temporal bone axial CT scan showing contact of the cholesteatoma's capsule with sigmoid sinus and meninges. [/fig]
|
Galileo Galilei: Science vs. faith
# Introduction
Galileo , professor of mathematics at the University of Padua from 1592 to 1610, was a pillar in the history of our University and a symbol of freedom for research and teaching, well stated in the university motto ''Universa Universis Patavina Libertas'' (Total freedom in Padua, open to all the world). [bib_ref] Origin and development of modern medicine at the University of Padua and..., Zampieri [/bib_ref] He invented the experimental method, based on evidence and calculation (''science is measure'')and was able, by using the telescope, to confirm the Copernican heliocentric theory, a challenge to the Bible. Bertrand Russell , in his book ''The Problems of Philosophy'' stated: ''Almost everything that distinguishes modern world from earlier centuries is attributable to science, which achieved the most spectacular triumphs in the seventeenth century. Together with Harvey, Newton and Keplero, Galileo was a protagonist of this scientific revolution in the late Renaissance''.His life was a continuous struggle to defend science from the influence of religious prejudices. He was catholic, forced by the Inquisition to deny his views, and was condemned to home arrest for the rest of his life. Here is the history of his life, a pendulum between science and religious beliefs.
## The early period in pisa
Galileo was born in Pisa from Giulia Ammannati (1538-1620) and Vincenzo Galilei (1520-1591), a musician, on February 15, 1564 (the same month and year when Michelangelo Buonarroti died). On November 5, 1581, he was registered in the faculty of Artists at the University of Pisa as a student of Medicine. In Florence in 1583 he met the mathematician Ostilio Ricci (1540-1603) and took up Mathematics, leaving Medicine, but never graduated. On 1589, he became Lecturer of Maths at the University of Pisa, with a salary of only 60 ducats/year. It was in 1590 that he carried out the famous experiment of falling spheres from the Tower of Pisa.
## The patavian period
On September 25, 1592, at the early age of 28 years [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] he was appointed at the University of Padua to take over the vacant chair of Mathematics with a better salary (180 florins/year), still very little when compared to the salary of the anatomist and surgeon Fabricius ab Acquapendente (1533-1619) (1200 florins/year). In Padua, he spent 18 years, ''the best years of my life''.He invented the compass [fig_ref] Figure 2: The geometric and military compass invented by Galileo [/fig_ref] , he wrote books on mechanics and cosmography for students. The teaching activity was hard, since the University council was very demanding.He invented a telescope [fig_ref] Figure 3: The ''Eye'' of Galileo's telescope [/fig_ref] , with a 18-20 fold magnification,which made it possible to establish that the Milky Way consisted of a multitude of stars; to see the inequalities of the surface of the moon [fig_ref] Figure 4: Drawings made by Galileo during his observations of the Moon [/fig_ref] , the ''rings'' of Saturn, sun spots, the phases of Venus, and, last but not least, moons rotating around Jupiter, thus confirming the heliocentric theory of Nicolaus Copernicus (1473-1543).The latter observation was made during the cold shiny nights of 7-13th January 1610, when he observed four objects moving around Jupiter, which he called ''Medicea Siderea'' (he was the tutor in Florence of the young Cosimo II de' Medici (1590-1621) [fig_ref] Figure 5: Cosimo II de' Medici, Grand Duke of Tuscany [/fig_ref]. He wrote and published quickly the ''Sidereus Nuncius'' (''celestial announcement''), a book of only 55 pages, dedicated to Cosimo II, and full of notes and drawings proving his observations.On August 24, 1609, he had had the opportunity to present the telescope to the Doge Leonardo Donato and to the Council of Ten of the Serenissima Republic of Venice, watching at distance ships entering the Laguna from the top of the bell tower in St. Mark's Square. They immediately realized the importance of the tool for military purposes and increased the salary of Galileo up to 1000 florins. Nevertheless, on June 10, 1610, he resigned from the chair in Padua, having accepted from Cosimo II de' Medici the offer of a chair at the University of Pisa, free from teaching obligations and including Philosophy besides Maths. On June 15, 1610, Galileo left Padua to move to Florence. His friend Giovanni Francesco Sagredo (1571-1620) wrote a letter to him, which was like a presentiment ''Where you will find the freedom and monarchy of yourself you had in Venice?''The figure of Leonardo Donato (1536-1612), Doge in Venice from 1606-1610, deserves a note. He was the main supporter of Galileo at the time of his call to the University of Padua in 1592. In 1594, he was the one of the three reformers that had the rule of the University, and promoted, together with Fabricius ab Acquapendente and Paolo Sarpi (1552-1623), the building of the famous Anatomical Theatre, the first research laboratory in the history of Medicine.
Five years later William Harvey (1578-1657) arrived in Padua to study medicine and had the opportunity to dissect cadavers and to see valves in the veins. This observation inspired him to put forward the theory of blood circulation. At the time of Doge Leonardo Donato, the Republic of Venice was excommunicated (''Interdetto'') by the Pope Paolo V (Camillo Borghese (1552-1621)), for having condemned two criminal friars, instead of entrusting them to the church. Paolo Sarpi, a friar who defended the position of the Republic of Venice, was wounded during a subsequent assassination attempt.
Oddly enough, years earlier -in 1592 -Leonardo Donato was ambassador of the Venice Republic at the Vatican in Rome and met Camillo Borghese (the future Paul V), who threatened him by saying ''If I will be Pope, at the first occasion I will excommunicate you'' and Donato replied ''If I will be Doge, I will not care of it''.
## The early florentine period
On July 10, 1610 Galileo was nominated as ''Primary Mathematician'' at the University of Pisa and Philosopher by the Tuscan Grand Duke Cosimo II, with a salary of 1000 ducats/year (more than the salary offered by the Doge to remain in Padua). It is the period of triumph for Galileo, following the sidereal discoveries.
On April 1611, he visited Rome to meet Pope Paul V and the Roman College of Jesuits, the chief of which, Christopher Clavius (1538-1612) [fig_ref] Figure 6: Christopher Clavius, Jesuit of the Roman College [/fig_ref] , fascinated by the observation of Jupiter's moons, reserved him a friendly welcome.On April 26, 1611 he was enrolled as a member of the Lincei Academy, founded in 1603 by the young Federico Cesi (1585-1630) [fig_ref] Figure 7: Federico Cesi, founder of the Lincei Academy in 1603 [/fig_ref] , who named the telescope ''Galileo spectacle''. Unfortunately, this honeymoon period soon ended.
The Dominican friars (''Domini canes'' = the dogs of God) started attacking Galileo, because of the ideas behind Copernican theory, in which the sun, not the earth, was at He said, ''It is publicly well known that Mister Galileo keeps these two beliefs, namely the earth rotates around itself and the sun is motionless. These sentences reject the divine scriptures as dictated by the Holy Father and as consequence repel the faith, which teaches that we must consider true what the scriptures contain.''He referred to the book of Joshua 10, 12-15 where Joshua said:
''Sun, stand still over Gibeon, and you, moon, over the Valley of Aijalon. So the sun stood still, and the moon stopped, till the nation avenged itself on its enemies''.Moreover, in the Ecclesiastes, 1, 4 there is this sentence:
''The earth is fixed and the sun arises on the east and sets on the west''.The position of the Dominicans, who had the leadership of the Holy Office, was not universally shared by other representatives of the Church. Cardinal Cesare Baronio (1538-1607) in 1615 was quoted in a correspondence between Galileo and Cristina Lorena (1565-1637), mother of Archduke Cosimo II, as saying ''The intention of the Holy Spirit is to teach how to go to heaven and not how the heaven is going''.
Cardinal Roberto Bellarmino (1542-1621), a Jesuit, who had graduated at the University of Padua [fig_ref] Figure 8: Cardinal Roberto Bellarmino graduated at the University of Padua [/fig_ref] and was the chief of the Holy Office (Inquisition), who had condemned Giordano Bruno (1548-1600) to the stake for heresy, because he questioned the trinity and believed in the existence of an infinite universe, wrote to Father Paolo Antionio Foscarini (1565-1616) on April 12, 1615: ''Whenever the motionless of the sun and the movement of the earth around the sun will be proven, then we have to be prudent in interpreting the Scriptures which appear opposite and say that we do not understand rather than that is false what is demonstrated.'' 14
## The late florentine period
On February 25, 1616, the Holy Office (Inquisition) officially condemned the Copernican Theory. Cardinal Bellarmino met Galileo and gave him a friendly warning not to sustain, teach, or refer to Copernican theory, otherwise the risk was to undergo Inquisition. 15 [fig_ref] Figure 9: Maffeo Barberini, Pope Urban VIII [/fig_ref] , adrmirer of Galileo, was elected Pope with the name of Urban VIII. Galileo believed that, with the new Pope as a friend, the atmosphere might change and he started writing the ''Dialogue'' [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] , where his theses were reported. The book received the ''nihil obstat'' authorisation by the Holy Office in Florence, and was published in 1632. The language was not Latin, but Italian, thus allowing it to be read by lay people.
Unexpectedly, Pope Urban VIII changed his mind and submitted Galileo to the Inquisition for having published the book without his authorisation (''imprimatur''). Galileo was reluctant to undergo the trial, claiming health problems. Thereafter, the Pope sent a committee of physicians to Florence to check whether the impossibility of moving to Rome was true. He was found to be affected by ''pulsus irregularis et inegualis'' (atrial fibrillation?), kidney stones and melancholy. Nevertheless he was regarded healthy and forced to travel to Rome.
He was put into jail and in June 1633, brought to trial in the Church of Santa Maria Minerva [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref].He was charged with heresy for believing false doctrines, contrary to Sacred and Divine Scriptures, namely that the sun and not the earth is the centre of the universe. He was condemned to deny, curse, and hate his mistakes (''abiura''). ''With sincere feeling and faith I abjure, swear and abhor my mistakes and heresies contrary to Catholic and Apostolic Roman Church.''He was forced not to treat orally or to write any more on the issue, and condemned to home arrest for the rest of his life. The book was prohibited forever.
The Committee of the Holy Office Inquisition was made up mostly of priests and friars of the Dominican and Theatine orders. The leader of the Holy Office was a Dominican, Father Vincenzo Maculan (1578-1667). The day when the sentence of condemnation was read, Cardinal Francesco Barberini (1597-1679) [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] , friend of Galileo, nephew of Urban VIII, and member of the Holy Office, deserted the meeting. He disagreed and did not want to share this dreadful decision.
## The decline and fatal outcome
Galileo was confined to his house in Arcetri, not so far from the cloister where his daughter Virginia (1600-1634) (Sister Maria Celeste) was living and where she died prematurely in 1634.In 1637 Galileo became blind bilaterally [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] , most probably as a final complication of a long-standing reactive arthritis, [bib_ref] Galileo as patient, Thiene [/bib_ref] [bib_ref] Atti e memorie dell'accademia galileiana di scienze lettere ed arti in Padova, Thiene [/bib_ref] [bib_ref] New interpretation of Galileo's arthritis and blindness, Zanatta [/bib_ref] and was assisted by his pupils Vincenzo Viviani (1622-1703) and Evangelista Torricelli (1608-1647).
In 1638 he published with Elzevier a new version of Dialogue, with a different title ''Discourses and Mathematical Demonstrations Relating Two New Sciences'', confirming his views. On January 8, 1642, on the same day of his Jupiter observations by telescope, he died at Arcetri and was buried in a hidden Chapel of Santa Croce Church in Florence, without any official ceremony because he was still considered a heretic.
In 1737, a mausoleum was built in the same church of Santa Croce and his remains were transferred together with those of his daughter Virginia. On that occasion, the exhumation gave the opportunity to remove the fifth lumbar vertebra [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] , which is now on display at the University of Padua. 21
## Four centuries later
The persecutions Galileo suffered during his life by the Roman Church were a source of longstanding controversy and diaspora between the Vatican and the University of Padua. Galileo was considered a victim of intolerance against science and intellectual freedom.
Four centuries later, the University of Padua celebrated a special year with meetings and honoris causa gradations. The Aula Magna (Great Hall) was named Galileo Galilei. Meanwhile Pope John Paul II (1920-2005) set up a committee of cardinals and theologists to re-examine the ''Tolemaic-Copernican controversy of the XVI-XVII century''.
The committee, led by Cardinal Paul Poupard (1930 -), worked hard for a few years and eventually gave a report to the pope, which is well summarised in its final part: Could it be that God asked Galileo to leave aside the presumption of science, just as God asked Abraham to sacrifice his son Isaac to test his loyalty [fig_ref] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano'' [/fig_ref] ?
## Conclusion and message to the young
The origin of science is an instance of struggle for intellectual liberty. When you read a book for your education and for improving your profession, look for the freedom and not only for the truth.
[fig] Figure 1: Portrait of the young Galileo by Domenico Cresti ''Il Passignano''. [/fig]
[fig] Figure 2: The geometric and military compass invented by Galileo. [/fig]
[fig] Figure 3: The ''Eye'' of Galileo's telescope.Zanatta et al. GCSP 2017:10 [/fig]
[fig] Figure 4: Drawings made by Galileo during his observations of the Moon. [/fig]
[fig] Figure 5: Cosimo II de' Medici, Grand Duke of Tuscany. [/fig]
[fig] Figure 6: Christopher Clavius, Jesuit of the Roman College. [/fig]
[fig] Figure 7: Federico Cesi, founder of the Lincei Academy in 1603. the centre of the universe, thus contradicting the Bible. The Dominican friar Tommaso Caccini (1574-1648), on December 21, 1614, violently blamed Galileo from the pulpit of Santa Maria Novella in Florence, strongly defending the literal scriptures, and clearly accusing Galileo of heresy. [/fig]
[fig] Figure 8: Cardinal Roberto Bellarmino graduated at the University of Padua. On March 5, 1616, the book ''De revolutionibus orbium coelestium'' by Nicolaus Copernico was prohibited and the heliocentric theory was refused, being considered opposite to the Scriptures. Unfortunately, Cardinal Bellarmino and Cosimo II died in 1621, and with them Galileo lost his supporters. On August 6, 1623, Cardinal Maffeo Barberini (1568-1644) [/fig]
[fig] Figure 9: Maffeo Barberini, Pope Urban VIII. [/fig]
|
Trehalose Is a Key Determinant of the Quiescent Metabolic State That Fuels Cell Cycle Progression upon Return to Growth
[fig] Figure S3: WT cells grown in medium lacking trehalose show fewer dense/quiescent cells at later time points following long-term batch culture growth. WT yeast were grown in YPD, YPGal and YPGal-T media starting from log phase and the density of individual cells was resolved as described. Time (h) of collection is indicated at the top right corner of each tube and optical density of the culture (OD 600 ) at the bottom. Density tubes from the three different media conditions are aligned according to the relative growth stage for comparison. Note that there are fewer dense/quiescent cells in the YPGal-T sample starting at 72 h of growth. [/fig]
[fig] Figure S4: Dynamics of NTH1 mRNA and protein levels during the yeast metabolic cycle. (A) NTH1 mRNA level peaks during the RC, quiescent phase of the yeast metabolic cycle (data obtained from Tu et al., Science 2005). (B) Nth1p protein levels do not change significantly during the yeast metabolic cycle. Cells expressing Nth1p-FLAG were collected at 12 time points through one yeast metabolic cycle as shown. The lysates of the 12 samples were subjected to western blot analysis using a FLAG antibody. [/fig]
|
Utilization of dental care: An Indian outlook
Oral health has a significant impact on the quality of life, appearance, and self-esteem of the people. Preventive dental visits help in the early detection and treatment of oral diseases. Dental care utilization can be defined as the percentage of the population who access dental services over a specified period of time. There are reports that dental patients only visit the dentist when in pain and never bother to return for follow-up in most cases. To improve oral health outcomes an adequate knowledge of the way the individuals use health services and the factors predictive of this behavior is essential. The interest in developing models explaining the utilization of dental services has increased; issues like dental anxiety, price, income, the distance a person had to travel to get care, and preference for preservation of teeth are treated as barriers in regular dental care. Published materials which pertain to the use of dental services by Indian population have been reviewed and analyzed in depth in the present study. Dental surgeons and dental health workers have to play an adequate role in facilitating public enlightenment that people may appreciate the need for regular dental care and make adequate and proper use of the available dental care facilities.
# Introduction
Oral health is a critical but an overlooked component of overall health and well-being among children and adults. Oral health problems such as dental caries, periodontitis, and oral cancers are a global health problem in both industrialized and especially in developing countries. Dental disease restricts activities in school, work, and home and often significantly diminishes the quality of life for many children and adults, especially those who are low-income or uninsured. Huge differences exist in health status including oral health between urban and rural population in India and other developing countries. [bib_ref] Targeting poor health: Improving oral health for the poor and the underserved, Singh [/bib_ref] Although there have been impressive advances in both dental technology and in the scientific understanding of Review Article oral diseases, significant disparities remain in both the rates of dental disease and access to dental care among sub-groups of the population.
India has approximately 289 dental colleges with around 25,000 graduates each year [ [fig_ref] Figure 1: Geographical distribution of dental colleges in India [/fig_ref] ]. Even with such a large work force, most of the people in India do not have access to basic oral health care. [bib_ref] Dental education in India, Sivapathasundharam [/bib_ref] The dentist to population ratio is 1:10,000 in urban areas whereas it drastically falls to 1:150,000 in rural areas. [bib_ref] Challenges to the oral health workforce in India, Tandon [/bib_ref] Although, dental care is a part of primary health care in India, dental care services are available in very few states at the primary health care level. Patients are not covered under any type of insurance, and generally pay out of their pockets to get treatment from both public and private dentists. Utilization is the actual attendance by the members of the public at oral health care facilities to receive care. In regions where adequate dental manpower is available yet the utilization of oral health care services is low thereby widening the oral health differences across the social economic classes. [bib_ref] Targeting poor health: Improving oral health for the poor and the underserved, Singh [/bib_ref] Various factors like demographic, behavioral, socio-economic, cultural, and epidemiogical, etc., contribute to people's decision to either forgo care or seek professional assistance for dental problems []. [bib_ref] Behavioral and socioeconomic correlates of dental problem experience and patterns of health..., Cohen [/bib_ref] [bib_ref] Revisiting the behavioral model and access to medical care: Does it matter?, Andersen [/bib_ref] The present paper focuses on the availability of dental care and the pattern of utilization of dental care facilities by the Indian population residing in different parts of the country.
# Methods
A thorough review of literature was done which engaged most of the articles published in peer-reviewed journals relating to the subject of utilization of dental care among Indian population. The review itself began with the search of relevant key words linked with the dental care like utilization, access, barriers, dental care, India, etc., in various search engines including PubMed. Reports published only in English language were included in the review. The spotlight of the present review would not only be on the pattern of utilization of dental services by the Indian population but also on various hurdles that come across utilization. The search also targeted various socio-demographic characteristics and anxiety levels of the subjects that can influence the rate of dental service attenders. The present review also highlights important measures that can be undertaken to improve access for effective utilization of dental services.
## Oral health care system in india
Oral health care in India is delivered mainly by the following establishments:- Majority of dental services in India is being provided by the private dental practitioners, followed by non-governmental organizations. Various nation-wide surveys have conducted to study the pattern of utilization of dental services by Indian population. The main objective behind these surveys was to evaluate the various factors that contributed towards utilization of dental services by the people residing in varied geographical regions of the country and factors predictive of this behavior.
## Studies conducted in northern india
A retrospective study was conducted to evaluate the type of patients, disease pattern, and services rendered in dental outreach programs in rural areas of Haryana, India. [bib_ref] Utilization of services rendered in dental outreach programs in rural areas of..., Vashisth [/bib_ref] A total of 1371 individuals attended the outreach program seeking the treatment. The results of the study indicated that utilization of dental services was found to be more in females than in males. The utilization of dental services was found to be influenced by the socio-demographic characteristics of the population like age, education, occupation, etc., The study concluded that there was need to motivate people giving them information but paying attention to the individual reasons which restricted their behavior.
A cross-sectional study was carried out in Chandigarh in the year 2008, over a period of 8 months, which had two components: Community survey and the Health Facility survey. [bib_ref] Access to public dental care facilities in Chandigarh, Verma [/bib_ref] The main objective of the study was to determine the level of dental health care assess and associated factors, at various public health facilities of Chandigarh. The community survey included interviews of adult respondents at their homes and the health facility survey was initiated to interview the dentists at public health facilities to analyze the records in their clinics. In the community survey, a total of 203 persons were interviewed, 101 in urban areas and 102 in rural areas. Out of all the respondents who were having dental problems at the time of survey, 40% in the urban areas and 57.7% in the rural areas preferred to visit a dentist in the government set up for their problem. Giving less importance to dental problems, lack of time, and self-medicating were other reasons cited for not consulting the dentist. Time taken to reach a dentist was more for rural when compared with urban respondents. Therefore, specific efforts targeted to increase awareness toward oral health are required. The main objectives of the study were to identify the oral health practices and patterns of utilization of dental services, to assess oral health status and treatment needs of the elderly population, and to test alternate strategies for controlling oral health problems among the elderly. [bib_ref] Oral healthcare for elderly: Identifying the needs and feasible strategies for service..., Goel [/bib_ref] The rural areas of Delhi were included in the present study and a two stage sampling technique was adopted. Most of the subjects (80%) reported availability of dental services in their area, of which a major proportion was being provided by the private sector. One-fifth of the subjects reported having suffered from dental problems and 60% of these visited a dentist to avail dental care. Reasons given by the subjects as barriers to accessing oral health care were related to lack of priority for oral health (attitudes) and their dependent status (non-ambulatory/disabled elderly). Therefore it was emphasized to change patient perception on oral health through health education and incorporate domiciliary dental care in gerontology.
A study was conducted in a group of six villages in the district of Lucknow, Uttar Pradesh. A total of 227 individuals aged 50 years or above were interviewed and clinically examined. [bib_ref] Increasing the prosthodontic awareness of an aging Indian rural population, Parlani [/bib_ref] An educational and motivational program to increase prosthodontic awareness was organized and results were evaluated before and after the program. Certain myths that proved to be a hurdle in utilization of dental services prevailed in the study population like tooth loss is an extension of old age, eating tobacco prevented caries, dental diseases can be cured by medicines alone, tooth extraction leads to loss of vision, and oral prophylaxis causes loosening of teeth.
## Studies conducted in western india
A survey was conducted in Udaipur city in 2008, which is located in south-eastern zone of Rajasthan state. [bib_ref] A relation between dental anxiety, the parental family and regularity of dental..., Kulkarni [/bib_ref] Dental anxiety is often reported as a cause of irregular dental attendance, delay in seeking dental care or even avoidance of dental care. [bib_ref] Social sciences and dentistry: A critical bibliography, Cohen [/bib_ref] [bib_ref] Dental attendance in 1998 and implications for the future, Nuttall [/bib_ref] Therefore, the aim of the study was to examine relationship between regularity of dental attendance and other variables like dental anxiety, dental behavior of parents, the dental upbringing of the respondents, education, socio-economic status, and sex. Many people all over the world are dentally anxious, but different studies show considerable results. According to the results of the present study, dentally anxious subjects are more irregular dental attendees than non-anxious people. Non-anxious who are regular dental attendees comprise 14.7%. Education, dental upbringing, regular dental attendance, socio-economic status, and interaction between education and anxiety were found to be importance for the prediction of regularity of dental attendance.
Another study which co-related anxiety level of the subjects with socio-demographic characteristics was conducted in Gujarat. [bib_ref] Prevalence and socio-demographic correlates of dental anxiety among a group of adult..., Malvania [/bib_ref] A total of 150 patients waiting in the outpatient Department of Oral Diagnosis of a Dental College in Vadodara were included in the study. Results of the study indicated that prevalence of dental anxiety among the study population was 46%. Females were found to be significantly more anxious than the males. Subjects residing in villages were more anxious when compared with the subjects residing in the city. Subjects with traumatic negative dental experience in the past showed higher anxiety scores. This can lead to the development of negative attitude toward dentist or dental treatment and consequently non-utilization of dental services. [bib_ref] Onset of and patterns of change in dental anxiety in adolescence and..., Locker [/bib_ref] It was emphasized to include behavior sciences in dental education and the integration of ethical considerations in the academic dental curriculum could help to improve the situation.
A descriptive cross-sectional study was conducted in Jaipur, Rajasthan to determine the association between socio-demographic factors and dental services use among patients visiting a dental college and hospital. [bib_ref] Association between socio-demographic factors and dental service utilization among people visiting a..., Devaraj [/bib_ref] The study sample included 180 people, aged 15-65 years visiting the outpatient department of the hospital in a 5-day period.
According to the results of the study, place of residence and income/month were significantly associated with dental service utilization as people residing in urban areas and economically sound visited the dentist more often when compared with people residing in rural areas and belonging to low-income groups. However, there was no significant difference between age, gender, and education level with dental service utilization. It is cited that this could be due to the fact that the dental college hospitals and most of the private dental clinics are situated within the city limits and very less or virtually no dental care services are available in the rural areas.
A cross-sectional survey was carried out among 427 randomly selected individuals in Udaipur in 2009 using a pre-tested questionnaire. [bib_ref] Barriers to the utilization of dental services in udaipur, Kakatkar [/bib_ref] The objective of this study was to determine the barriers in regular dental care and home care and to assess their association with age, sex, education, and income. Results of the survey showed that the male group had more dental visits, but females experienced higher dental fear. The younger age group had more visits within 1 year in comparison to the older group. Increase in education decreases the barriers for regular dental care. Income had a significantly negative correlation with dental visit. This study also revealed the patient's perceived need that people visited the dentist only if they had symptoms which may be due to their belief that dental conditions are not serious or life threatening. It is suggested that to motivate people successfully, one not only has to give them information but also has to pay attention to the individual reasons which restrict their behavior.
Results of another cross-sectional survey [bib_ref] Barriers to oral health care amongst different social classes in India, Garcha [/bib_ref] conducted to investigate and compare the influence of social and cultural factors as access barriers to oral health care amongst people from various social classes in Pimpri, Gujarat indicated that irrespective of the social class difference, 88% participants wished to seek only expert/professional advice for the dental treatment. Unavailability of services on Sunday, going to dentist only when in pain, trying self-care or home remedy, inadequate government policies, and budgetary constraints were among the major access barriers which proved to be an obstacle in utilization of dental care.
## Surveys conducted on south indian population
A house-to-house survey was conducted in the field practice in Mangalore, Karnataka where dental services are provided free of cost. The main objective of the survey was to determine the factors related to the utilization of dental services. [bib_ref] Utilization of dental services in a field practice area in Mangalore, Karnataka, Poudyal [/bib_ref] The study sample comprised 195 adults to whom questionnaires were distributed. The results of the study depicted that nearly 30% of the study population had never visited a dentist although 44% of them had dental problems at the start of the study and majority of them were aware that free preventive dental procedures were provided nearby. Not having any problems with their teeth and lack of time were the two major barriers for dental visit reported by the study population. It is recommended that awareness of the people have to be improved and people be motivated to use the services available so that they can lead a socially and economically productive life.
A community-based cross-sectional study was conducted among 300 people aged 60 years and above in villages around Manipal, South India in 2008. [bib_ref] Barriers to seeking dental care among elderly in a rural South Indian..., Thomas [/bib_ref] The objective behind the survey was to identify the various barriers to avail dental health services. A house-to-house interview was conducted on 300 individuals who comprised the study population. The available dental care agencies were private clinics and government hospitals. Among them 90% who utilized dental agency utilized private dental care. The remaining utilized traditional medicines for dental problems. A positive correlation was found between socio-economic status and readiness to avail free dental services. Age was cited as an important barrier to avail dental services even if services were given free of cost. As age increased, utilization of dental services decreased. Anxiety and fear of dental treatment was more common in women especially those of low-socio-economic status. Imparting preventive dental education and strengthening of the primary health centers can go long way in reducing these barriers.
A cross-sectional study was conducted among 11-12 year school children in Bangalore city. The aim of this study was to assess knowledge, attitude, and practice towards oral health. [bib_ref] Oral health-related KAP among 11-to 12-year-old school children in a government-aided missionary..., Harikiran [/bib_ref] The study group comprised 212 children (males and females). The survey found that pain and discomfort from teeth were common whereas dental visits were infrequent. Fear of the dentist was the main cause of irregular visit in 46% of the study participants. Findings of this study also show that utilization of dental service is mainly for pain relief with the mother being the prime person involved in the utilization of dental services. It is suggested that systematic community-oriented oral health promotion programs are needed to target lifestyles and needs of school children.
A study was conducted among the municipal employees of Mysore city in 2004 to assess the prevalence of dental caries, periodontal diseases, oral pre-malignant, and malignant lesions in relation to socio-economic factors. [bib_ref] Oral health status in relation to socioeconomic factors among the municipal employees..., Shekar [/bib_ref] According to the findings of the study, subjects who had caries were higher in the persons with lower socio-economic status. This can be attributed towards poor utilization of dental services which can be related to the cost and lack of awareness on the etiological factors for oral diseases. [bib_ref] Is there equity in oral healthcare utilization: Experience after achieving Universal coverage, Somkotra [/bib_ref] [bib_ref] Prevalence of dental caries among the population of Gwalior (India) in relation..., Khan [/bib_ref] During any dental program planning, priority should be given to lower class people having higher prevalence of diseases and unmet treatment needs.
## Other studies
A study was conducted in Majuli, Assam to find out various medicinal plants used for dental care either in flowering or fruiting stage by common people. [bib_ref] Traditional practice for oral hygiene and hazard by common people: A case..., Bora [/bib_ref] During the survey a total number of 23 plant species belonging to15 family were recorded and use of plant parts are different to different localities. The traditional method of treatments and cares are still prevalent within different tribes of Majuli, Assam. The present trend of urbanization of the study areas also indicate that inspite of establishment of small health centers in the area, uses of plants and traditional practices will continue to play a significant role in the socio-cultural life of these village communities. The use of medicinal plants for curing oral health problems could be a major contributor for not utilizing available dental services by these people.
An oral care medicinal plants survey was conducted in different districts of Tamil Nadu during the period of 2000-2004 used by village people and ethic tribes of Tamil Nadu. [bib_ref] Traditional oral care medicinal plants survey of Tamil Nadu, Ganesan [/bib_ref] A total of 114 plant species were identified which were used to relieve toothache, used as toothbrush, mouthwash/gargle, and treat gum disorders. All these practices are a major barrier towards utilization of dental care services by these people.Overall, 28% of respondents reported oral health problems in India. West Bengal (42%) has the highest proportion of respondents with oral health problems. Respondents treated for oral health problems ranges between 21% and 28%, except West Bengal. Prevalence of oral health problems does not systematically vary by residence, insurance status, and by income quintiles.Of those who were diagnosed with oral health problems, 51% have been treated. The percent of respondents treated for oral health problems is highest in Karnataka (72%) and lowest in Assam (26%). Prevalence of oral health problems is higher among females than in males. However, the percentage who received treatment for oral health problems do not vary much by sexes. A higher percentage of urban and higher income quintile respondents received treatment for oral health problems.Role of dental insurance in dental care utilization
Unlike most western countries, specific dental insurance plans are not common in India. Indian Dental Association has been striving to bring out a new all-inclusive oral and dental health care insurance scheme. However, it has been unable to achieve anything substantial in this front. We, as oral health care workers, are capable to reach every class and village across the country. Dental health insurance can also bring about dental health care awareness percolating at the gross root levels. It would serve as a good motivation to the people to regularly visit the dentist and this in turn serves as an effective preventive measure. If we have to create awareness and pass on the benefits of longevity of teeth across the society, dental profession should impress on to the policy makers to have beneficial dental insurance schemes for the masses. [bib_ref] Dental insurance! Are we ready?, Toor [/bib_ref] CONCLUSION AND RECOMMENDATIONS Dental disease is a serious public health problem with universal distribution and affecting all age groups. However, despite this universal distribution, only a few seek dental care. Thus a wide gap is created between the actual dental needs of the population and the demand for dental care which is quite understandable from the cited literature. In India, people encounter various obstacles in utilization of dental services. These barriers can be removed by motivating people and making them aware about the oral health problems that remove anxiety and fear so that they develop positive attitude towards dental treatment. It is suggested that mobile dental clinics, dental camps, and dental outreach programs could be solutions to spread awareness and disseminate treatment. There is a need for reasonably priced, rural oral health centers to make dental care available to rural strata of the population. Unmet treatment needs of the people belonging to lower class should be addressed during conduction of dental programs. School-based screening and motivation programs significantly improve the percentage of children who seek free dental treatment at a dental school. [bib_ref] Does school-based dental screening for children increase follow-up treatment at dental school..., Hebbal [/bib_ref] These programs can also target lifestyles and needs of the school children.
Studies regarding the utilization dental services by north-east Indian population are almost non-existent. Therefore it is the responsibility of the health sector to
[fig] Figure 1: Geographical distribution of dental colleges in India [/fig]
[fig] A: three-phase survey was conducted in Delhi in 2003 by Maulana zad Dental College and Hospital and supported by the Government of India WHO Collaborative Program. [/fig]
|
Recommendations for the Establishment and Operation of Human Milk Banks in Europe: A Consensus Statement From the European Milk Bank Association (EMBA)
Objectives: To develop recommendations from the European Milk Bank Association (EMBA) for the establishment and operation of human milk banks (HMB) in Europe.Method: A working group comprising members of the EMBA was convened in 2015 to develop Europe-wide recommendations for milk banks. Each member had experience of guideline development and/or milk banking operations. An initial survey was agreed using collated published global recommendations. A total of 108 potential recommendations were included in the survey; responders noted which were included in their national guidelines. The responses were collated, compared, and discussed and the group determined where there was consensus and where substantial or minor differences were identified. Where there was consensus or robust published evidence on which to base recommendations these were included. When there was no consensus and no clear evidence base, a statement of explanation based on collective expert opinion was agreed.Results: Published, internationally available guidelines with recommendations for human milk banks from France, Italy, and the UK, together with guidelines and Switzerland were included as source materials. These covered: General recommendations; Donor recruitment and screening; Expression, handling, and storage of donor human milk (DHM); Pooling of DHM; Milk screening; Milk treatment (pasteurization); Delivery of DHM to recipients.The reviewer MG declared a past co-authorship with one of the authors AG to the handling Editor.
Conclusions: Evidence based recommendations and consensus statements from the EMBA will now be published on the EMBA website to assist in the safe establishment and operation of HMBs throughout Europe. These have also been used to inform the chapter on human milk to be included in the 2019 edition of the Guide to the quality and safety of tissues and cells for human application, published by the European Directorate for the Quality of Medicines & HealthCare (EDQM).
Keywords: donor human milk (DHM), Human Milk Bank (HMB), breastfeeding, pasteurization, donor screening, bacteriology testing BACKGROUND Human milk (HM) is the preferred nutrition for preterm infants (1), but not all mothers are able to provide their child with enough milk. There are specific beneficial effects of breastfeeding in these infants as HM feeding reduces the risk of short term and long term complications related to prematurity. Donor HM (DHM) can supplement the supply of maternal breastmilk when it is insufficient or provide the preferred alternative when the mother is not breastfeeding. DHM is beneficial for the health of preterm infants, first because significant properties are preserved following holder pasteurization (2) and secondly because it prevents feeding these infants with a preterm formula. According to the World Health Organization (1) the American Academy of Pediatrics (3) and the European Society of Pediatric Gastroenterology, Hepatology, and Nutrition (4) the feeding of preterm infants with mother's own milk is recommended as the first choice and if this is not available pasteurized DHM from an established milk bank should be the next alternative.
The main benefits for preterm infants that receive DHM instead of formula are faster gastric emptying, faster attainment of full enteral feedings, improved gut growth and maturation, decreased risk of necrotizing enterocolitis and late onset sepsis, improved neurodevelopmental outcomes, less retinopathy of prematurity and improved visual development [bib_ref] Does breastmilk influence the development of bronchopulmonary dysplasia?, Spiegler [/bib_ref] [bib_ref] Breast milk and neonatal necrotising enterocolitis, Lucas [/bib_ref] [bib_ref] Association of human milk feedings with a reduction in retinopathy of prematurity..., Hylander [/bib_ref].
HMBs have existed since 1909 in Europe and countries that supported the development of a significant number of HMBs developed national guidelines to harmonize practices. The European Milk Bank Association 1 (EMBA) was officially launched in October 2010. According to the definitions agreed by the EMBA Board of Directors:
- Human milk banks (HMB) collect, screen, store, process, and distribute DHM. - DHM is breastmilk that has been expressed by a mother and provided freely to a HMB to be fed to another mother's child.
The EMBA constitution is available to view on the EMBA website 1 . The Association welcomes membership from HMBs, milk bank associations and individuals who support the aims of the Association. Membership affords the opportunity to access conference presentations from the past 6 years as well as additional materials of interest to those who bank or use DHM.
[formula] 1 http://europeanmilkbanking.com [/formula]
The EMBA is a non-profit organization that was established to promote milk banking and to encourage international cooperation between the HMBs of the countries of Europe as well as globally. The association's headquarters are in Milan, Italy but association meetings and conferences are held in different locations within Europe. The primary aim of the EMBA is to support and promote exclusive breastfeeding, which is universally acknowledged to be the optimal method of feeding all infants. The use of DHM from HMBs for preterm and sick infants temporarily without access to their own mother's milk has been widely shown to be linked to increased rates of exclusive breastfeeding on discharge from neonatal units [bib_ref] Presence of human milk bank is associated with elevated rate of exclusive..., Arslanoglu [/bib_ref] [bib_ref] Impact of donor milk availability on breast milk use and necrotizing enterocolitis..., Kantorowska [/bib_ref].
The EMBA was also established to support human milk banking in Europe by promoting safe practices and fostering sustainable operations. Hence a further aim of the association, as listed in the EMBA Constitution 1 is to "Prepare international and regularly revised guidelines to set standards for the practice of milk banking." The usual methodology for establishing guidelines is to convene appropriate multidisciplinary expertise, consider all of the stages of a process and conduct a review of the evidence before making recommendations linked to the strength of the evidence. In the absence of such a published guideline, aimed at milk banks throughout Europe, it is expected that the development of Europe-wide recommendations by EMBA for HMBs will promote best practice, optimize quality, and safety within the HMBs that are currently operational as well as offer valuable assistance to those who are establishing a new HMB. Evidence and expert opinion based recommendations should be the standard throughout Europe to promote consistent practices both within and between countries.
There are currently 226 HMBs operating in 28 countries in Europe. These include the first milk banks in Estonia, Lithuania, Poland, Portugal, and Russia. A further 16 HMBs are planned. The current status of HMBs in Europe is maintained on the EMBA website 1 .
The 226 HMBs in Europe incorporate newly established organizations with a limited history of operational expertise as well as ones that have links to the very early milk banks founded over a century ago. They include national and regionalized services as well as small operations serving a single neonatal unit. The volumes of DHM that are tested and processed and provided mainly to sick and preterm infants vary widely from country to country 1 as well as within a single countryas do the criteria for receipt of DHM. An EMBA coordinated survey of all the HMBs in Europe will be carried out in 2019. This will gather up to date information about the extent of milk banking activity including the number of donors recruited, volumes of milk donated and processed, volumes of milk provided to hospitals, and the number of recipients. It will also collate further information about milk banking practices.
There are currently no published Europe-wide guidelines however internationally published guidelines for HMBs are freely available from France 2 , Italy [bib_ref] Guidelines for the establishment and operation of a donor human milk bank, Arslanoglu [/bib_ref] , and the UK. Countries with nationally agreed and recognized guidelines include Austria, Denmark, Germany, Norway, Slovakia, Spain, Sweden, and Switzerland. English translations of recently updated (2017-2018) national guidelines from Sweden and Spain can be downloaded via links from the EMBA website. Differences exist between national guidelines as a result of variations in practices, regulation, and organization of HMBs in each country. Differences in practices are due to the lack of evidence for some points related to the operation of a HMB. DHM status globally is variable; it is either considered a food, a health product food or a tissue; the latter ones being subject to stricter regulation. Organization of HMBs is different when collecting and delivering DHM only for infants hospitalized in the same building, or over a more or less wide area including several health care institutions.
The HMBs of Europe also have very different historical backgrounds 1 and in most cases have operated independently of each other although regular meetings take place between milk banking practitioners in Norway and Sweden and between those of Austria, Germany, and Switzerland.
A major role of the EMBA has, since its inception, been the provision of education, largely via the organization of international congresses. Presentations at these congresses have revealed similarities between the milk banking practices throughout Europe and have also revealed fundamental differences. These have substantiated the results of surveys of HMBs carried out both before and after the establishment of EMBA in 2007 and in 2012. In some cases the official national or local recommendations from various countries lead to significant differences in practice. For example most published guidelines recommend that all DHM should be heat treated [using the standard holder pasteurization method 2 [bib_ref] Guidelines for the establishment and operation of a donor human milk bank, Arslanoglu [/bib_ref] ]. However, in Norway mainly raw (non-heat treated) DHM is provided by the country's 12 milk banks. This is considered possible and reasonable because (1) the demand can be met due to very high breastfeeding rates, even with extremely strict donor screening, (2) there is complete traceability of milk from donor to consumer, (3) there is a small population with very low HIV and hepatitis rates, (4) extremely rigorous and regularly repeated testing of donors (every 3 months) is practiced [bib_ref] Characteristics of breast milk and serology of women donating breast milk to..., Lindemann [/bib_ref] [bib_ref] Donor milk banking and breastfeeding in Norway, Grøvslien [/bib_ref]. However, in almost half of the 20 official HMBs in Germany both raw and pasteurized DHM is available and used in accordance with local clinical judgment based on the gestation and health of the recipient.
## Objective
One of the clear objectives of the EMBA 1 has been to develop Europe-wide recommendations for HMBs. This is especially needed by countries without experience of human milk banking or the use of DHM. Such recommendations should optimize safety and draw upon the available evidence. The processes and practices within human milk banking do not lend themselves to randomized controlled trials and there are few meta-analyses or systematic reviews available to refer to. In the absence of these, expert opinion is required and where this differs, a consensus and in some cases compromise is sought. The EMBA Guideline Working Group was formed in 2015 to undertake this task. EMBA members from 13 countries contributed to the work of the group. Throughout its 3 year existence, members of the Guideline Working Group have included representatives from Austria, France, Germany, Italy, Norway, Poland, Portugal, Serbia, Slovenia, Slovakia, Spain, Switzerland, and the UK.
# Method
The three main tasks of the group were: 1) to complete a comprehensive survey of current practices within their national milk banks (See Appendix for details of the 108 survey questions) 2) to use the known guidelines to assess where consensus could be achieved 3) identify where research evidence is available to support recommendations.
A useful tool in developing the structure for the survey was the global implementation frameworkpublished by the NGO PATH 3 . The framework includes a compilation of practices in HMBs from donor recruitment to delivery of DHM to the recipient. Each of the group members completed the survey in accordance with their local/regional or national guidelines and, if not included in the guidelines, in accordance with local practice. Once completed, the survey responses were collated and highlighted according to whether international consensus was apparent, whether there was near consensus and where practices and recommendations differed significantly. The next step was to note where no published evidence was available. A list of agreed recommendations that were evidence based and that could form the basis of EMBA recommendations was drawn up and presented to the wider EMBA membership at further meetings. Where there was no consensus and no evidence upon which to determine a recommendation, expert opinion within the group was used to decide whether to include a recommendation or to provide an explanation of why no clear recommendation could be made at that time.
# Results
The recommendations that were found to be included in all of the published guidelines or accepted as best practice according to the available evidence or expert opinion and have therefore been agreed by the Working Group are presented below. Many of these are underpinned by principles of good manufacturing process (GMP) especially those related to consistency and quality control.
## Emba recommendations for the establishment and operation of human milk banks in europe
General Recommendations 1. A robust quality assurance plan (e.g., HACCP-Hazard
Analysis Quality Control Points) should be in place to ensure the safe operation of the HMB [bib_ref] Guidelines for the establishment and operation of a donor human milk bank, Arslanoglu [/bib_ref]. 2. All equipment should be maintained in accordance with manufacturers' instructions and checked and qualified annually to ensure conformity with recommendations (GMP). 3. Containers should be closed or sealed in accordance with manufacturers' recommendations (GMP). 4. Containers should not be overfilled as DHM will expand when frozen (GMP). 5. Containers of human milk should at all times be labeled with the donor's name or unique ID and the date the milk was expressed. In addition, depending on the stage in the process, DHM should also be labeled with the name of the HMB, whether the milk is raw or processed, the milk's expiry date and if the milk is ready (cleared) to use. 6. HMBs should at all times minimize exposure of the human milk to sunlight and/or phototherapy lights. 7. HMB staff should have health checks when employed and be immunized in accordance with local hospital or national health service protocols. 8. Staff should undertake training by an experienced member of staff (or in accordance with national accredited training programs where available) before undertaking unsupervised work in a HMB. 9. DHM should be handled hygienically and HMB staff should wash their hands in accordance with local protocols prior to entering clean areas, handling DHM or equipment used in the collection, storage, testing and processing of DHM (GMP). 10. HMB staff should at all times consider the ethical implications of their work with donors, parents, carers, and infants. For example donors should not be encouraged or coerced to donate more milk than may be optimal for themselves or their infants. The use of DHM should not undermine or interfere with the mother's provision of her own milk or of breastfeeding unless there are concerns about its safety. 11. Records should be kept of all donors and their donations of milk including volumes, dates and any additional relevant information (traceability). 12. Prioritization of recipients should be locally determined. [bib_ref] Guidelines for the establishment and operation of a donor human milk bank, Arslanoglu [/bib_ref]. Steps should be taken to prevent the temperature of DHM rising during transport e.g., by the use of insulated transport containers (boxes or bags) and thermal (ice) packs or the use of dry ice where necessary. 14. The use of tamper evident transport containers is recommended and the temperature of the interior of the container should be monitored during transport using data loggers or checked on arrival. A record of this should be maintained as part of the HMB records (GMP). 15. Transport boxes/bags in which containers of DHM are carried should be insulated and easy to clean. [bib_ref] Donor milk banking and breastfeeding in Norway, Grøvslien [/bib_ref]. The transport container should be decontaminated between batches of milk. Keep raw milk from different mothers separate. Avoid using the same transport container to transport raw and pasteurized DHM. 17. Containers of DHM should not be placed directly into the transport container-use of clear polythene bags helps the HMB identify the contents and check that they are correctly labeled without the need for unnecessary handling.
Donor Recruitment and Screening 11. Monitor and record refrigeration and freezing equipment continuously or at least every 24 h. 12. Store raw and pasteurized human milk in separate clearly labeled refrigerators and freezers or if not possible in separate clearly labeled fridge and freezer compartments. 13. Thaw frozen raw human milk in a refrigerator to prevent its temperature rising above 8 - C or if impractical because of time constraints on a counter where it can be monitored and transferred to a refrigerator immediately once thawed. 14. Pooling of DHM.
14.1 The pooling of DHM from the same donor is acceptable prior to any heat processing 14.2 Maintain records of all pooling including the names/ID numbers of the donors, dates the milk was expressed and any medications taken.
## Milk screening; pre-and post-pasteurization testing
Within Europe, as well as globally, there is no consensus for recommendations for the microbiological testing of DHM either before or after pasteurization (see [fig_ref] TABLE 1 |: Published microbiological screening criteria for acceptance of donor human milk prior to... [/fig_ref]. Throughout Europe, local and national guidelines vary both in the timing and frequency of testing and in the acceptance criteria. There is a lack of published evidence to inform decision making and the working group members concluded that to maximize safety of vulnerable immune-compromised recipients, best practice suggests:
- Before pasteurization -All pools of milk be tested prior to pasteurization -Each batch of milk be tested after pasteurization -Acceptance criteria: 10 5 cfu/ml or less of non-pathogenic organisms and no pathogens for each pool of milk tested prior to pasteurization. Discard all samples of milk from a pool that does not meet this standard.
- After pasteurization -Discard the batch if there is any microbial growth detected in a random sample taken after pasteurization.
It is recognized that most published guidelines include less stringent screening criteria [fig_ref] TABLE 1 |: Published microbiological screening criteria for acceptance of donor human milk prior to... [/fig_ref] as part of the overall recommendations. The EMBA therefore recommend that where milk banks follow nationally or locally agreed guidance, alternative microbiological screening criteria may be adopted if done so as part of the overall adherence to the guideline being followed.
## Milk treatment (pasteurization)
1. Current recommended heat treatment/pasteurization temperature and time is 62.5 - C for 30 min followed by rapid cooling to at least 10 - C and preferably 4 - C prior to transfer to a freezer (2). 2. Monitor the process and record temperatures throughout the treatment (2). 3. Pasteurized milk that has been freeze dried and vacuum packed if performed in a milk bank in accordance with GMP and all relevant safety procedures may be used.
## Delivery of dhm to recipients
1. Never use a microwave oven to defrost or warm milk. 2. All donor milk and containers should be labeled at each stage to ensure traceability and tracking of the milk. 3. The receiving hospital should record/document how donor milk is used including in the infant's hospital notes. 4. Before administration of donor milk, informed consent is required from recipient's mother/parent/carer in accordance with local protocols. 5. It is recognized that some religious beliefs and practices influence the acceptability of anonymized donor human milk and these should be taken into account when discussing donor milk with donors and with parents and carers and when drawing up local protocols.
# Discussion
The Working Group was able to arrive at a consensus for recommendations covering most of the major processes involved in recruiting and screening donors, storing, handling, and transporting DHM and in its testing and processing. In these cases, the evidence, as referenced in the relevant national guidelines was checked and noted. Where published evidence was not available, expert opinion was the basis for the recommendation. The group also highlighted several areas for which consensus could not be achieved and where there was no clear evidence to inform a recommendation. In these instances the group provided a statement of explanation and suggested best practice. The group also acknowledged that not all countries that have milk banks were represented within the guideline group.
A further consideration was that HMBs are funded in different ways and have varying resources. It was agreed that to make a single recommendation that may undermine a milk bank's ability to continue to operate would not be in the best interests of the overall population of recipients. The process by which DHM is tested to provide microbiological safety is an example of where no consensus could be found. Very divergent practices are currently in place throughout Europe and there is no agreement within the published guidelines as to the optimal testing regime either before or after pasteurization or both. There is also no clear evidence base on which to determine the recommendations. However, it is possible to make a recommendation as to which practice will offer optimal safety for all recipients, including the most vulnerable neonates, and through which no untested milk reaches the recipients and the strictest published acceptance criteria are used. Recommended testing regimes exist in Europe that include less frequent sampling [e.g., only testing the first donation and randomly testing subsequent donations as in the Italian guidelines [bib_ref] Guidelines for the establishment and operation of a donor human milk bank, Arslanoglu [/bib_ref] ]. There are also recommendations in which the acceptance criteria are less strict (e.g., pre-pasteurization criteria of 10 5 cfu/ml or less of any organisms and 10 4 cfu/ml or less of Staphylococcus aureus and Enterobacteriaceae and the presence of <10 cfu/ml of any organism post-pasteurization as per the UK's NICE guideline. These are acceptable alternatives but do not provide the same assurance of safety. It should be noted however that less strict testing regimes will lead to more processed milk being available. An additional consideration is that if DHM is fed without any heat treatment (as occurs in some parts of Europe), stricter local microbiological testing protocols should be adopted to increase the safety of the recipients.
The tracking of DHM and the traceability of the milk throughout all the HMB processes is an essential component of safety and quality assurance. This is enhanced by the availability of customized and purpose developed barcode tracking systems and the use of the internationally agreed coding system, ISBT128. Regular strict quality control of pasteurization equipment is necessary to maintain optimum safety and quality of DHM [bib_ref] The effect of Holder pasteurization on nutrients and biologically-active components in donor..., Peila [/bib_ref] [bib_ref] Air and water processes do not produce the same high-quality pasteurization of..., Buffin [/bib_ref].
Whilst safety is at the forefront of all recommendations pertaining to HMBs and the use of DHM it is important to ensure that the ethical considerations of banking and sharing human milk and duties of care to the donors as well as the recipients are not overlooked [bib_ref] Ensuring safety in donor human milk banking in neonatal intensive care, Hartmann [/bib_ref].
A limitation of this consensus statement is that no recommendations were made regarding the analysis of HM. HM analysis is a means of assessing protein and energy contents and to facilitate targeted fortification of HM (i.e., to fortify milk according to its individual composition to reach a target composition as a means of helping to cover the theoretical nutritional needs of each infant). Some HMBs include the assessment of HM composition in their practice. This is done either to perform targeted fortification or to select HM with the highest nutrient content for the smallest babies. However, there is no consensus about the best strategy for fortification. Only adjustable fortification, which does not require HM composition assessment, has been shown to improve postnatal growth.
# Conclusions
There are no Europe-wide guidelines covering practices within HMBs. Historically national and/or local guidelines have been developed and although the recommendations include many similarities, clear differences also exist. Developing recommendations for HMBs operating throughout Europe was a challenging task because of the diversity within European countries. However, with the input of leading milk banking experts from a wide range of countries and with additional help from the EMBA Board it was possible to agree a pragmatic approach to where differences could not be resolved through reference to published research. The resultant guidance will now be freely available on the EMBA website to assist in the safe establishment and operation of HMBs throughout Europe. The recommendations have also been used to help inform the chapter on human milk to be included in the 2019 edition of the Guide to the quality and safety of tissues and cells for human application, published by the European Directorate for the Quality of Medicines & Health Care (EDQM).
# Author contributions
GW, EB, CG, AnnG, RM-M, SA, DB, C-YB, RB, AntG, GM, AW, and J-CP contributed to the working group and/or to the discussions of the findings of the working group and all offered amendments to the previous drafts and/or agreed the manuscript.
[table] TABLE 1 |: Published microbiological screening criteria for acceptance of donor human milk prior to and after pasteurization. a Human Milk Bank Association of North America. http://hmbana.org. [/table]
|
The multi-allelic APRR2 gene is associated with fruit pigment accumulation in melon and watermelon
Color and pigment contents are important aspects of fruit quality and consumer acceptance of cucurbit crops. Here, we describe the independent mapping and cloning of a common causative APRR2 gene regulating pigment accumulation in melon and watermelon. We initially show that the APRR2 transcription factor is causative for the qualitative difference between dark and light green rind in both crops. Further analyses establish the link between sequence or expression level variations in the CmAPRR2 gene and pigment content in the rind and flesh of mature melon fruits. A genome-wide association study (GWAS) of young fruit rind color in a panel composed of 177 diverse melon accessions did not result in any significant association, leading to an earlier assumption that multiple genes are involved in shaping the overall phenotypic variation in this trait. Through resequencing of 25 representative accessions and allelism tests between light rind accessions, we show that multiple independent single nucleotide polymorphisms in the CmAPRR2 gene are causative of the light rind phenotype. The multi-haplotypic nature of this gene explains the lack of detection power obtained through genotyping by sequencing-based GWAS and confirms the pivotal role of this gene in shaping fruit color variation in melon. This study demonstrates the power of combining bi-and multi-allelic designs with deep sequencing, to resolve lack of power due to high haplotypic diversity and low allele frequencies. Due to its central role and broad effect on pigment accumulation in fruits, the APRR2 gene is an attractive target for carotenoid bio-fortification of cucurbit crops.
# Introduction
Flesh and rind pigmentation are key components affecting the nutritional value and consumer preference of the major cucurbits crops, melon and watermelon. Both crops exhibit extreme diversity in fruit traits, including size, shape, color, texture, aroma, and sugar content. Regulation of rind color in cucurbits initiates early in fruit development and is expressed as green color intensity at the young fruit stage, reflecting chlorophyll concentrations . Most watermelons remain green at maturity, with chlorophyll being their main rind pigment, and therefore rind color variation in watermelon is mostly expressed as green pigment intensity in uniform or striped patterns. Conversely, melon rind color transforms during development, leading to extensive variation in mature fruit pigment profiles that include different combinations of carotenoids, flavonoids, and chlorophylls. The genetic basis of this variation is only partly resolved. Several external fruit color quantitative trait loci have been mapped in populations derived from a cross between a Piel de Sapo line and PI161375. It has been previously reported that the mature yellow rind color of yellow casaba melon accessions (Cucumis melo var. inodorous) is caused by the accumulation of naringenin chalcone, a yellow flavonoid pigment . A Kelch domaincontaining F-box protein-coding gene (CmKFB) on chromosome 10 was identified as causative for the naringenin chalcone accumulation in melon fruit rind. While it is logical to assume that young and mature fruit color intensity are correlated, and that common genetic factors may be involved, thus far such genes have not been reported in melon.
Three major flesh color categories are defined in melon: green, white, and orange, with β-carotene and chlorophyll being the predominant pigments of the orange and green phenotypes, respectively . The major locus qualitatively differentiating between orange and non-orange flesh is green flesh (gf), located on chromosome 9. gf was recently shown to be the CmOr gene, which governs carotenoid accumulation and orange flesh color. A second qualitative flesh color locus, white flesh (wf), which is associated with the difference between white and green flesh, has been previously described and mapped to chromosome 8. Another layer of quantitative variation in flesh pigment content and color intensity exists within those color classes, as defined by several QTL mapping studies. These include the recent fine-mapping to a candidate causative gene level of flesh carotenoid QTLs using a recombinant inbred line (RIL) population. Thus far, however, causative genes governing this quantitative variation have not been shown.
In recent years, a few transcription factors involved in regulation and synchronization of chlorophyll and carotenoid accumulation were identified in plants. Among these are Golden2-like (GLK2) transcription factors, which regulate chloroplast development . Allelic or expression variation in the GLK2 gene was shown to be associated with levels of chlorophyll and carotenoids in Arabidopsis, tomato, and pepper. A related but distinct transcription factor gene, the ARABIDOPSIS PSEUDO-RESPONSE REGULATOR2-LIKE gene (APRR2) with phenotypic effects comparable with those of GLK2, was identified and shown to regulate pigment accumulation in tomato and pepper, and overexpression of the APRR2 gene in tomato increased the number of plastids and the color intensity. Recently, the APRR2 gene was also shown to be causative of the white immature rind color (w) in cucumber (Cucumis sativus) , a close relative of melon (Cucumis melo). The white rind phenotype of immature cucumbers in this study was shown to be associated with reduced chloroplast number and chlorophyll content.
In the current study, we used bi-parental populations to map and identify the APRR2 gene as a common causative regulator of pigment accumulation in both melon and watermelon. We show that the effect of this transcription factor on pigment accumulation is initially observed in the rind of young fruits (chlorophylls) and extends to rind and flesh of mature melon fruits (chlorophylls and carotenoids). Through further analysis of wider genetic variation in melon, we revealed a unique multiallelic pattern that inhibited our ability to detect a significant signal through genotyping by sequencing (GBS)-based genomewide association study (GWAS). By zooming in on this allelic series, we confirmed the central role of this gene in shaping the color variation of young fruit rind across melon diversity.
# Materials and methods
## Plant materials and field trials
The germplasm used in this study included four sets. (i) TAD×DUL RILs (F 7 )-bi-parental segregating population derived from the cross of the dark rind line, 'Dulce' (DUL; C. melo var. reticulatus) with the light rind line, 'Tam Dew' (TAD; C. melo var. inodorous). One hundred and sixty-four F 7 RILs were developed through single-seeddescent. All RILs, F 1 , and the parental lines were grown in a randomized block design (RCBD) in an open field at Newe Ya'ar Research Center, in the spring-summer seasons of 2016 and 2017. Each line was represented by two replicates of five plants per plot. (ii) NA×DUL F 3 and F 3:4 -an F 3 population from this cross was grown in two repetitions in a greenhouse at Beit Elazari, Israel in 2013 as previously described. This population is derived from the cross between the light rind line 'Noy-Amid' (NA; C. melo var. inodorous) and the common dark rind parent, DUL. One hundred and fourteen F 3:4 families from F 3 genotyped plants alongside the parental lines and their F 1 were grown in an RCBD in an open-field trial at Newe Ya'ar Research Center in the spring-summer season of 2017 in two replicates of six plants per plot. (iii) Melo180 GWAS panel-a Newe-Ya'ar melon collection used in this study comprised 177 diverse accessions that represent the two melon subspecies (ssp. agrestis and ssp. melo) and 11 taxonomic groups. Each accession was represented by three plots of five plants each in an RCBD in the open field at Newe-Ya'ar in summer 2015 . (iv) NY0016×EMB F 2:3 -for mapping the light rind trait in watermelon, the light rind inbred accession NY0016 was crossed with the canary yellow accession Early Moon Beam (EMB) to produce 87 F 2:3 families. During the summer of 2016 and 2017, 10 plants per F 2:3 family and two plots of 10 plants from the parents and F 1 were sown in the open field at Newe-Ya'ar. All the populations used in this study were grown under standard horticultural conditions in open fields at Newe Ya'ar Research Center, northern Israel (32°43′05.4″N, 35°10′47.7″E), soil type was grumusol, and the plants were drip-irrigated and drip-fertilized.
## Fruit color phenotyping
In the melon populations, fruit images were taken with a digital camera on developing fruits of all accessions throughout the season, from anthesis to harvest. Rind color of young fruits was scored in the field at 10-15 d after anthesis and confirmed based on fruit images from the same developmental stage. Mature rind and flesh color were measured on ripe fruits, which were harvested based on abscission in climacteric fruits, or days after anthesis, rind color, and TSS (total soluble solids) in non-climacteric fruits. Five mature fruits per plot were photographed externally, then cut along the longitudinal section and scanned for internal imaging, using a standard document scanner (Canon, Lide120) as described previously . Scanned images were analyzed using the Tomato Analyzer softwarefor color (L, A, B, Chroma, and Hue) and morphological features. Rind and flesh tissues were sampled into 50 ml tubes from at least three fruits per plot, immediately frozen in liquid nitrogen, and then stored at -80 °C for further analyses. For the watermelon mapping experiment (NY0016×EMB F 2:3 ), 10 F 3 individuals per F 2:3 family were harvested at maturity (~70 d post-sowing), imaged, and phenotyped for rind color as above.
## Quantification of carotenoids and chlorophyll
Carotenoids were extracted from 0.5 mg ground tissue samples in a mixture of hexane:acetone:ethanol (2:1:1, v/v/v) as described previously, and separated using a Waters 2695 HPLC apparatus equipped with a Waters 996 PDA detector (Milford, MA, USA). Carotenoids were identified by their characteristic absorption spectra, distinctive retention time, and comparison with authentic standards. Quantification was performed by integrating the peak areas with standard curves of authentic standards with the Waters millennium chromatography software. Lutein and β-carotene were relatively quantified at 450 nm and 270 nm, respectively, by integrating their peak areas and calculating their percentage from total integrated peak areas. Tissues for chlorophyll determination were sampled as explained for carotenoid analysis. Chlorophyll extraction was performed in dimmed light to avoid possible photodegradation of chlorophyll. Chlorophyll was extracted by adding 5 ml of DMSO to 0.5 g, vortexing, and incubating in the dark at room temperature for 24 h. The extract was analyzed for absorbance in the wavelengths of 663 nm and 645 nm using a Cary50Bio spectrophotometer (Varian). Chlorophyll concentration was calculated as described by.
## Genotyping
DNA isolations were performed using the GenElute™ Plant Genomic Miniprep Kit (Sigma-Aldrich, St. Louis, MO, USA). DNA quality and quantification were determined using a Nanodrop ND-1000 (Nanodrop Technologies, Wilmington, DE, USA) spectrophotometer, electrophoresis on an agarose gel (1.0%), and Qubit ® dsDNA BR Assay Kit (Life Technologies, Eugene, OR, USA).
## Gbs analysis, snp calling, and map construction
TAD×DUL RILs DNA from 164 F 7 individuals was processed by Novogene (Novogene Bioinformatics Institute, Beijing, China) for GBS analysis. An aliquot of 0.3-0.6 μg of genomic DNA from each sample was digested with the ApeKI restriction enzyme, based on the in silico evaluation results, and the obtained fragments were ligated with two barcoded adaptors at each end of the digested fragment. Following several rounds of PCR amplification, all the samples were pooled and size-selected for the required fragments to complete the library construction. Samples were diluted to 1 ng µl -1 and the insert size was assessed using the Agilent ® 2100 bioanalyzer; quantitative PCR was performed to detect the effective concentration of each library. Libraries with a concentration >2 nM were sequenced on an Illumina HiSeq 2000/2500 platform as 144 bp, pairedend reads and mapped to the C. melo reference genome DHL92 v3.5.1; available at http://melonomics.cragenomica.es/ files/Genome/Melon_genome_v3.5.1/). Over 570 million reads were produced covering nearly 21% of the genome across >35 million tags at an average read depth of nine reads per site. Single nucleotide polymorphism (SNP) calling was carried out using Broad Institute's genome analysis toolkit (GATK), resulting in 1 205 528 raw SNPs. Sites with a depth of less than three reads per site or >50% missing data were filtered out using TASSEL v5.2.43. Data were then imputed using the full-sib families linkage disequilibrium (LD) algorithmfollowed by the removal of individuals with excess heterozygosity. The genotypic data set was phased to ABH format consisting of 89 343 SNPs across 146 lines. Binning was performed using SNPbinnerwith a minimum ratio between cross-points set at 0.001 and minimum bin size of 1000 bp. Bin statistics and genetic distance were calculated using an in-house script developed in python, based on the Kosambi mapping function. The final set included 2853 recombination bins across 146 lines. Evaluation of genotypic data quality was done by accurately mapping flesh color to a 55 kb interval spanning the previously published CmOr gene (Melo3C05449).
NA×DUL F 3:4 DNA from 140 F 3 individuals was processed by NRGene LTD (Nes Ziyyona, Israel) for restriction site-associated DNA sequencing (RADseq), SNP calling was performed following similar methods, and the initial marker set included 43 975 SNPs across 140 individuals with an average depth of 16 reads per site. Further filtering and imputation were performed as described above, and the final set for binning was composed of 19 015 SNPs across 134 individuals. Binning and genetic map construction were carried out using the same parameters as those used for the TAD×DUL population, yielding 1321 bins across 134 individuals.
## Gwas180
Genotyping of this diversity panel was performed using GBS, as described by. The final SNP set included 23 931 informative SNPs [at a minor allele frequency (MAF) of >5% across 177 accessions.
## Ny0016×emb f 2:3
The watermelon mapping population NY0016×EMB was genotyped by GBS. Library construction, sequencing, and SNP calling were performed at the Genomic Diversity Facility at Cornell University (Ithaca, NY, USA) as described by. Sequences in this project were aligned to the Charleston Gray genome, version 1 (available at http://cucurbitgenomics.org/organism/4 ).
## Bulk segregant analysis by sequencing (bsa-seq) of the watermelon population
DNA samples from 35 F 2 plants (from the NY0016×EMB cross) homozygote for the rind color trait (based on the F 3 family's phenotypes) were prepared in two bulks (light rind, 19 F 2 samples; and dark rind, 16 F 2 samples). These samples, as well as DNA samples of the parental lines (EMB and NY0016), were used for whole-genome resequencing (WGS) performed at the DNA Services Center at the University of Illinois, Urbana-Champaign. Six shotgun genomic libraries were prepared with the Hyper Library construction kit from Kapa Biosystems (Roche) with no PCR amplification. The libraries were quantitated by qPCR and sequenced on one lane for 151 cycles from each end of the fragments on a HiSeq 4000 using a HiSeq 4000 sequencing kit version1. Fastq files were generated and demultiplexed with the bcl2fastq v2.17.1.14 Conversion Software (Illumina). Average output per library was 44 million reads of 150 bp. All raw reads were mapped to the Charleston Gray reference genome using the Burrows-Wheeler Aligner (BWA), producing analysis-ready BAM files for variant discovery with Broad Institute's GATK. Homozygous SNPs between the two parents were extracted from the vcf file that was further filtered to a total depth of >20 reads per site. The read depth information for the homozygous SNPs in the 'light' and 'dark' pools was obtained to calculate the SNP-index. For each site, we then calculated for each bulk the ratio of the number of 'reference' reads to the total number of reads, which represented the SNP index of that site. The difference between the SNP-index of two pools was calculated as ΔSNP-index. The sliding window method was used to perform the whole-genome scan and identify the trait locus confidence interval on chromosome 9.
## Wgs of 25 representative diverse melon accessions
DNA of the 25 core accessions was shipped to the Genomic Diversity Facility at Cornell University (Ithaca, NY) for WGS to an estimated 30× depth.
Validation of rare alleles at the APRR2 genes The causative variants at the different APRR2 alleles in melon and watermelon, which were discovered based on next-generation sequencing (NGS) of genomic DNA, were confirmed on the parental lines and relevant segregants through Sanger sequencing of genomic DNA in melon and cDNA from mRNA that was extracted from fruits in watermelon.
# Qrt-pcr analysis
Rind samples were peeled from fruits harvested throughout development, from anthesis to maturity, and immediately frozen in liquid nitrogen. Three fruits were sampled from each genotype at each developmental stage. A 100-150 mg aliquot of frozen rind tissue per sample was used for RNA extraction using a Plant/Fungi Total RNA Purification Kit (NORGEN Biotek Corp., Canada). First-strand cDNA was synthesized using a cDNA Reverse Transcription Kit (Applied Biosystems, USA). The 10 μl qPCR volume included 1 μl of cDNA template, 0.2 μl of each primer (10 μM), 5 μl of Fast SYBR Green Master Mix (Applied Biosystems, USA), and RNase-free water to a final volume of 10 μl. qRT-PCR, with an annealing temperature of 60 °C, was performed in triplicate on a 96-well plate in the Step-One Plus Real-Time PCR system (Applied Biosystems, USA). The melon Cyclophilin A gene (Melo3C013375) was used as a control to normalize the qRT-PCR values across different samples. Primers are listed in at JXB online
# Data analysis
Trait mapping Genome-wide linkage analysis for young fruit rind color in the bi-parental populations was performed in R/qtlusing the Composite Interval Mapping (CIM) function and confirmed using single marker analyses in TASSEL utilizing a generalized linear model (GLM) and in the JMP V13.1 software package (SAS Institute, Cary, NC, USA). A GWAS of the melon diversity collection was performed by a mixed linear model (MLM) analysis in TASSEL, using both the population structure (Q matrix) and relatedness [kinship (k) matrix] as covariates to control for population structure. Multiple comparison corrections to significance thresholds were performed using the FDR (false discoovery rate) approach. All further statistical analyses (correlations and ANOVA) were performed using the JMP V13.1 software package.
## Population structure, kinship, and ld analysis
Relatedness between the melon accessions in the diverse collection was estimated in TASSEL software v5.2.43 using the pairwise kinship matrix (k matrix) through the Centered IBS method. LD between intrachromosomal pairs of sites was done on chromosome 4 using the full matrix option in TASSEL.
## Sequence analyses
Sequence alignments and comparison of APRR2 alleles were performed using the BioEdit software packageand the integrative genomics viewer (IGV) package. Comparative analysis of haplotype diversity across 2200 genes on melon chromosome 4 was performed following these steps: (i) a VCF file containing ~4 000 000 high quality SNPs across the core set of 25 melon lines (MAF >0.1 and <10% missing data per SNP) was created based on alignments to the melon genome version 3.5.1; (ii) the corresponding gene annotations file was used to create a subset of exonic SNPs on all annotated genes on chromosome 4; and (iii) the number of exonic SNP haplotypes per gene was calculated.
# Results
## Gwas of young fruit rind color in melon
Most melons can be visually classified into two distinct young fruit [~10 days post-anthesis (DPA)] rind colors; dark or light green, reflecting qualitative variation in chlorophyll content. Light immature rind color was previously reported to display a recessive single-gene inheritance in a bi-parental segregating population. In the current study, young fruit rind color was visually scored on a previously described diverse melon collection composed of 177 accessions . The collection was genotyped genome-wide with 23 931 informative, GBS-derived SNP markers, and was shown to be an effective resource for mapping simple traits in melon . Here, we used a subset composed of 120 accessions with a clearly defined dark or light rind phenotype (see the example infor GWAS. Accessions with prominent non-uniform rind color (stripes or dots) were excluded from this analysis. We also excluded Charentais lines, as their dominant grayish light rind is exceptional and phenotypically distinct from the common light rind in other melon types. While the dark and light phenotypes were distributed uniformly across the genetic variation and were represented in balanced proportions across this set (41% and 59%, respectively;, a genome-wide population structurecorrected analysis did not result in any significant marker-trait association. This result has led to the assumption that while this highly heritable trait may show simple inheritance in a specific bi-allelic cross, it is possibly more complex and explained by multiple loci across a multi-allelic diverse collection.
## Mapping and identification of the young fruit light rind gene in melon
In order to dissect this trait further using a simpler genetic design, we analyzed two segregating bi-parental populations: the first is composed of 164 RILs (F 7 ) from a cross between a light rind honeydew parent (TAD) and a dark rind reticulatus parent (DUL;. The second population is composed of 114 F 3:4 families derived from a cross of DUL with another light rind accession, a yellow casaba inodorous melon (NA;. These segregating populations were visually phenotyped for young fruit rind color over two seasons, and a consistent single gene (Mendelian) ratio was observed in dark:light phenotypes. The populations were then genotyped through GBS, and 89 343 (TAD×DUL RILs) and 43 975 (NA×DUL F 3:4 ) informative SNP markers were identified and used for mapping. Whole-genome linkage analysis using the four data sets (two populations over two growing seasons) resulted in the identification of a single highly significant consistent trait locus on chromosome 4. The common confidence interval for this trait locus spans a 290 kb region (chromosme 4: 640-930 kb;on the melon reference genome (Garcia-Mas et al., 2012; http://cucurbitgenomics.org/ organism/3), as confirmed also through substitution mapping using recombinants within this interval in the TAD×DUL RILs population. Annotation of the genomic sequence at this interval revealed 33 putative genes including a strong candidate, Melo3C003375, which is annotated as an ARABIDOPSIS PSEUDO-RESPONSE REGULATOR2-LIKE (APRR2) gene, the melon homolog of a recently reported causative gene of the recessive white rind (w) mutation in cucumber .
We then compared the CmAPRR2 gene (Melo3C003375) sequence between the parental lines of the mapping populations. Genomic and mRNA sequencing revealed multiple polymorphisms, including two different exonic polymorphisms causing independent stop codons in each of the light rind parents compared with the common dark parent (DUL). G to T substitution in exon 8 in TAD compared with DUL led to a premature stop codon and a predicted aberrant protein of 292 amino acids compared with the normal 527 amino acid protein of DUL. A 13 bp insertion in exon 9 of NA result in a frameshift leading to a different premature stop codon in this line and a predicted protein of 430 amino acids.
Furthermore, we crossed TAD and NA with each other and with DUL (as a reference testcross) and phenotyped the F 1 s for young fruit rind color. Both testcrosses with DUL resulted, as expected, in a dark rind in the F 1 . However, the F 1 of TAD×NA had a clear light rind and confirmed the allelic nature of these recessive phenotypes. This further corroborates that these independent predicted causative mutations in the CmAPRR2 gene are indeed allelic.
## Expression pattern of the cmaprr2 gene in melon fruit
Fruits from the light (TAD) and dark (DUL) parental lines were sampled during development from anthesis to maturity, and mRNA levels of the CmAPRR2 gene were analyzed by qRT-PCR. We show here that CmAPRR2 has a higher expression level in fruit compared with leaves, as also shown in cucumber and pepper, and in agreement with the Melonet-DB gene expression atlas. In accordance with these studies, we also show that the CmAPRR2 peak expression in fruit rind occurs at ~15 DPA, before the initiation of ripening and color change. Comparison between the parental lines of the mapping population also showed significantly lower levels of CmAPRR2 expression in light rind fruits throughout fruit development. Both light and dark lines have reduced CmAPRR2 expression levels at the mature fruit stage and were not significantly different from each other at that stage.
## Multiple independent predicted causative mutations in the cmaprr2 gene across melon diversity
In light of the conflict between the clear identification of a single causative gene through linkage mapping in two different bi-parental crosses on the one hand, and the absence of any significant genome-wide signal from the GWAS analysis on the other hand, we re-sequenced and compared the genomic sequence of the CmAPRR2 gene across a core panel of 25 diverse melon lines. This core panel was selected to represent the different groups and overall diversity in our collection as described previously;. Nineteen lines from the panel, showing a clear dark or light young fruit rind phenotype, were used for the sequence comparison. Seventeen SNPs and InDels within exons in the CmAPRR2 gene were identified across this panel. Eight of these SNPs were either synonymous or did not show a distinct allelic state between dark and light accessions. The remaining nine polymorphisms are independently inherited (not in LD with each other) and display low frequency alleles (0.05-0.15) which are unique to the light rind accessions. Four of these polymorphisms (2, 6, 13, and 14) are SNPs that change a single amino acid. Three are InDels (4, 7, and 12) that cause frameshifts leading to major modification in predicted protein sequence, and the remaining two (9 and 11) are the causative polymorphisms described above, leading to premature stop codons as in TAD and NA. These five major polymorphisms (4, 7, 9, 11, and 12) explain the light rind phenotype in eight of the 11 light rind accessions in the core panel. The nonsynonymous polymorphisms 2, 6, and 13 are potentially causative of the light rind phenotype in accessions BAHC and QME. The only light rind accession that could not be explained by non-synonymous variation within the CmAPRR2 coding sequence is SAS. However, the low mRNA expression of CmAPRR2 in young fruit (5 DPA) rinds of SAS, which was similar to the expression level in TAD, and significantly lower compared with DUL, suggests expression level variation as a possible causative element for the light rind phenotype in this line. To test whether all these 'light' accessions are indeed allelic and caused by different mutations in the CmAPRR2 gene, we performed allelism tests where all the 'light' accessions (n=11) were intercrossed and the resulting 55 F 1 s were phenotypically evaluated for young fruit rind color. As a reference, these 'light' accessions were crossed with two 'dark' testers (DUL and Ananas Yoqne'am; AY).shows that all 55 'light'×'light' F 1 hybrids displayed light immature fruit rinds, while all 22 'light'×'dark' testcrosses displayed dark rinds. These results confirm the allelism between the 11 'light' lines, including the light rind phenotype of SAS. Combined interpretation of the sequence variation and allelism tests across this representative core panel indicates that most of the young fruit rind color variation can be explained by multiple independent polymorphisms related to the CmAPRR2 gene. Extrapolation of the results from this core set suggests that each of these causative variants is most probably also present at low allele frequency, across the wider diversity panel, leading to the nonsignificant associations observed in our GBS-based GWAS experiment. It is worth noting that these independent mutations are not in LD with each other, leading to the high haplotype diversity in this gene. In a comparative analysis of haplotype diversity based on exonic SNPs across 2200 genes on melon chromosome 4, we found that CmAPRR2 is indeed the second most diverse gene, irrespective of the number of SNPs and transcript length (seeand the Materials and methods). Analysis of the LD pattern in the genomic region surrounding the CmAPRR2 locus confirmed the low LD between SNPs in this region. The fact that these allelic polymorphisms are not in LD with each other allows us aggregate them into a theoretical unified functional polymorphism, resulting in increased frequency of the aberrant CmAPRR2 allele (0.52;. This analysis, in turn, produces a significant association between the CmAPRR2 gene and the light rind trait.
## Mapping and identification of the light rind color gene in watermelon
In parallel with melon, we also studied the genetics of rind color in watermelon. The main difference is that in watermelon, due to its non-climacteric fruit ripening, chlorophylls are the main rind pigments also during fruit maturity, and therefore light and dark green were visually scored on mature fruits. Our light rind source in this study was an heirloom accession named NY0016that was crossed on different lines, varying in their rind color, to produce F 1 s and F 2 s. All the F 1 hybrids had dark rind fruits (irrespective of the stripe pattern in the 'dark' parents), proving the recessive nature of the light rind phenotype of NY0016. All F 2 populations were phenotyped for rind color, and a consistent 3:1 Mendelian ratio was observed for dark and light rinds, respectively . The cross between the light rind accession (NY0016) and a striped (dark) parent (EMB) was selected for linkage mapping, and this population was advanced to F 3 to perform F 2:3 analysis. GBS of the F 2 population (n=87) resulted in a final high-quality set of 3160 filtered SNPsthat was used for genetic mapping of the light rind phenotype. Seven to ten fruits per F 3 family were visually scored, and each family was classified into a defined category: fixed for light rind, fixed for dark rind, or segregating for rind color. The observed 1:2:1 frequencies of light, segregating, and dark across the F 3 families supported a single gene inheritance for this trait. Whole-genome linkage analysis resulted in the identification of a single significant trait locus on chromosome 9 (R 2 =0.62, P=2.9×10 −18 ;. The confidence interval for this locus spanned 1.7 Mb with 80 predicted genes on the watermelon reference genome (Charleston Gray, http:// cucurbitgenomics.org/organism/4). In order to narrow down this genomic interval, we performed mapping by sequencing of DNA bulks (BSA-Seq) from 35 selected F 2 segregants that were homozygous for light or dark rind (on the basis of fixed F 3 phenotypes). The ~30× WGS resulted in the identification of 400 000 high-quality SNPs differentiating between parental lines. Comparison of allele frequencies between the light and dark bulks across the 400 000 SNPs (ΔSNP-index analysis) confirmed the trait locus on chromosome 9 and allowed us to narrow down the genomic confidence interval to a 900 kb region with 30 predicted genes; Supplementary . Review of the list of genes within the confidence interval revealed a strong candidate, CICG09G012330, the watermelon homolog ClAPRR2 gene, highly similar to the causative melon (Melo3C003375) and cucumber (Csa3G904140.3;genes. Comparison of the genomic sequence of CICG09G012330 between the mapping population parental lines revealed several SNPs, none of them within exons. The only putative causative SNP at that point was at the intron 6-exon 7 junction. Parent and segregant mRNA sequence comparison revealed an alternative splicing in the intron 6-exon 7 junction, leading to a 16 bp deletion at the mRNA of the light rind parent and corresponding segregants carrying the 'light' allele. This 16 bp deletion, which created a frameshift leading to a premature stop codon and a predicted aberrant protein, is most probably causative for the light rind phenotype of NY0016.
## Allelic variation in the cmaprr2 gene is associated with mature fruit rind and flesh pigmentation in melon
Earlier analyses of rinds from TAD, DUL, and selected F 3 families from their cross demonstrated the correlation between young fruit chlorophyll content and mature fruit carotenoids. To test whether the CmAPRR2 gene is also associated with mature fruit rind pigmentation in melon, we analyzed mature fruits from the TAD×DUL RIL population for rind color and carotenoid content. We harvested 10 mature fruits per line, external images of the fruits were taken for color scoring, and the rinds were sampled for carotenoid profiling. Color intensity variation at the young (green) stage in this population is qualitative and could be visually classified into two distinct classes (dark or light green) on a single-fruit basis. However, at the mature stage, any effect of the CmAPRR2 gene on rind color is visually of a quantitative nature, as also shown in tomato. Rind netting, which segregates in this population, further masked rind color and complicated visual scoring and carotenoid quantification. For the parental lines, TAD, with the light green rind at the young fruit stage, has a creamyellowish rind at maturity, and DUL, with the dark green rind at the young stage, has an orange rind masked by dense rind netting. Visual observation of standardized external images of selected mildly netted mature fruits, representing both alleles in the CmAPRR2 gene, suggested a possible effect of this gene on rind color intensity, such that on average the 'dark' allele is associated with a deeper orange color. We confirmed this effect through analysis of the carotenoid content in rinds of 50 selected mildly netted segregants (25 RILs carrying each allele in the CmAPRR2 gene). The 'light' allele was significantly associated with a >10-fold reduction in total carotenoids in the rind, which is consistent with the reduced chlorophyll levels observed in young fruits of this group. Allelic variation in this gene explains 37% of the variation in lutein content (P=2×10 −6 ), 34% of the variation in β-carotene content (P=8×10 −6 ), and 37% of total carotenoids (P=3×10 −6 ) in fruit rind in this population.
To examine whether the effect of the CmAPRR2 gene also extends to mature fruit flesh color, we phenotyped the TAD×DUL RILs for flesh color intensity using longitudinal fruit section scanning and quantitative image analysis (n=145 lines×10 fruits per line). This population is segregating for the main flesh color gene in melon, CmOr, discriminating between orange and non-orange flesh, and it is therefore composed of orange flesh lines (48%) and green flesh lines (49%); the remaining 3% of the lines are segregating due to residual heterozygosity. In order to test the association of the CmAPRR2 gene with flesh color, we analyzed the orange and green fruits separately. A significant association between the CmAPRR2 allelic segregation and color intensity was found in the orange-flesh group, which accumulated β-carotene as the main flesh pigment (R 2 =0.25, P=4.7×10 −5 ;, E) and, as expected, the 'dark' allele was associated with stronger pigmentation and higher predicted β-carotene content (R 2 =0.62, P=0.0053; see the Materials and methods and. Significant association of this gene with flesh color was also found in the green flesh group, which accumulated chlorophyll as the main flesh pigment (R 2 =0.31, P=2×10 −6 ) as the 'dark' allele was significantly associated with higher green Chroma, G) reflecting higher chlorophyll content.
## Expression level of the cmaprr2 gene is associated with mature flesh color in melon
In order to test whether the expression level of the CmAPRR2 gene is associated with mature fruit pigmentation, we analyzed RNA-Seq data and mature fruit flesh carotenoids on a different RIL population derived from a cross between DUL and an Indian phut snapmelon (Momordica group), PI414723 (hereafter called 414). While 414 has a spotted rind (and not a clear light or dark phenotype), we assume based on testcrosses with some of the light rind accessions that, like DUL, it also carries a 'dark' allele of the CmAPRR2 gene. This assumption is supported by the fact that it does not show any of the predicted 'light' non-synonymous polymorphisms found at the CmAPRR2 gene. DUL and 414 are genetically and phenotypically distant and differ in their mature fruit flesh color and carotenoid content. DUL has dark orange flesh while 414 has light (salmon-colored) orange flesh and, accordingly, the RIL population segregates for these traits. RNA-Seq was previously performed on mature fruit flesh of 96 RILs from this crossand allowed us to execute a genomewide eQTL analysis for the CmAPRR2 gene (Melo3C003375). A single, highly significant, cis-eQTL was mapped to chromosome 4 and defined by a 270 kb interval flanking this locus . This result confirmed the heritable variation in the CmAPRR2 expression level in this population and that the expression of this gene is mostly regulated by cis-acting sequence variants. We then tested the correlation between the CmAPRR2 expression level and flesh β-carotene content across the 96 RILs, and found a significant positive correlation (R=0.38, P=0.0008;. Since both parental lines of this population carry the 'dark' allele based on the coding sequence of the CmAPRR2 gene, this result provides a quantitative support for the possible relationship between the expression level of this gene and pigment accumulation in melon. It is important to note that this population is segregating at additional QTLs that affect carotenoid content in mature fruit flesh, including a major QTL on chromosome 8, which was recently mapped to a candidate gene level. This variation further masked the specific effect of the CmAPRR2 expression level on flesh carotenoid content. We also assume that the observed correlation is an underestimation, as gene expression in this experiment was measured on mature fruits, whereas the peak of expression of the CmAPRR2 gene is much earlier, before fruit ripening (~15 DPA).
## Expression of cmaprr2 in melon is correlated with plastid development-related genes
To characterize co-expression patterns associated with the CmAPRR2 gene, we calculated the correlations between the expression of Melo3C003375 and all annotated melon genes (n=27 557), using RNA-Seq data from mature fruits of the 414×DUL RIL population (n=96). Fourteen thousand genes expressed in mature fruit flesh were used for this correlation analysis. Gene Ontology (GO) enrichment analysis was performed on 200 genes that had the strongest correlations with Melo3C003375 (R>0.43, FDR-adjusted P<0.001). The four most significant functional groups that were enriched are related to photosynthesis, light reaction, and plastid organization, and the 15 most enriched components are related to plastids and chloroplasts (Supplementary . These results support the predicted involvement of the CmAPRR2 gene in the regulation of chloroplast and chromoplast development.
# Discussion
## Color variation in immature fruit rind in melon
External fruit color is an important attribute in melons as it is a key factor defining consumers' preference. Melon rind color transforms during fruit development and ripening, mainly by shifting from the green rind of immature fruits, where chlorophyll is the main pigment, to variable rind colors composed of different combinations of chlorophylls, carotenoids, and flavonoids . Inheritance of external color of immature fruit has been previously described in two different studies: the white color of immature fruits was reported byto be dominant to green immature fruits and controlled by a single gene named Wi.described a recessive gene for light immature exterior color in a cross between an American muskmelon (Reticulatus group) and an American honeydew-type melon (Inodorous group). Recently, a major gene for external color of immature fruit was mapped in a cross between 'Védrantais', a Charentais line from the Cantalupensis group, and 'Piel de Sapo', from the Inodorous group. The dominant light rind from the 'Védrantais' parent, that most likely correspond to the Wi gene, was mapped to an ~1.6 Mb interval on chromosome 7. These results confirm our observation that the dominant light grayish rind of Charentais accessions is phenotypically and genetically distinct and controlled by a different gene from the one we identified in the current study, which correspond to the recessive gene described by.
## Aprr2-like transcription factors are key regulators of fruit pigmentation
The results of the current study support the pivotal role of APRR2-like genes in regulation of pigmentation in fruits.showed that overexpression of an APRR2 gene in tomato resulted in increased chlorophyll content in immature fruits and a higher level of carotenoids in ripe tomatoes. Both effects resulted from an increase in plastid number. They also provided evidence for association between a null mutation in an APRR2 gene and external fruit color intensity in green peppers. These results are complementary to reports on the role of a related transcription factors group, GLKs, which were shown to be associated with levels of chlorophyll and carotenoids in Arabidopsis, tomato, and pepper. Recently,reported that an APRR2 gene is causative for the white rind (w) mutation in cucumber, expressed as reduced chloroplast density and chlorophyll content in young cucumber fruits. In all these crop plant species (tomato, pepper, and cucumber), there is also a correlation between expression levels of either APRR2 or GLK genes and pigment intensity. In the current study, we performed high-resolution NGS-based mapping in segregating populations and found that null mutations in the CmAPRR2 and ClAPRR2 genes are associated with light rind color in melon and watermelon, respectively. We also showed that expression of the CmAPRR2 gene is correlated with pigment intensity in melon (Figs 2D, 6D) and that, as in pepperand cucumber , these transcription factors show their strongest expression in fruit and reach their peak expression at ~10-20 DPA and before fruit ripening. Our results expand the extent of experimental data that demonstrate the conserved function of APRR2-like genes in regulating fruit pigmentation as shown by the comparable expression profiles and analogous phenotypes associated with variation in these genes.
## Cmaprr2 is associated with pigment accumulation across fruit developmental stages and tissues
Variation in ripe fruit color is substantially wider compared with the variation in the immature stage. While in the immature stage, color variation mostly reflects chlorophyll concentrations, in the mature stage, biosynthetic pathways of additional pigments (i.e. carotenoids and flavonoids) are involved, leading to extended complexity of the genetic architecture. This complexity is also expressed by the independent genetic control of flesh and rind colors in melon, as best demonstrated by the ability to combine different rind and flesh colors through classical breeding. Rind and flesh color QTLs were mapped in multiple melon populationsand a few causative color genes were identified. However, so far, a common regulator that affects pigmentation throughout the different developmental stages and fruit tissues (rind and flesh) has not been described in melon. In the current study, we showed that the CmAPRR2 gene is such a key regulator, associated with pigment concentrations during the course of fruit development and across fruit tissues. Furthermore, the mapping population in this study (TAD×DUL RILs), which independently segregated for both CmOr and CmAPRR2 genes, allowed us to show that the CmAPRR2 effect is also independent of the type of pigment accumulated in the flesh, as it was associated with variation in both chlorophyll and carotenoid concentrations. We also showed here, using a different segregating population (414×DUL RILs) that was previously subjected to mature fruit RNA-Seq and carotenoid analysis, that the cis-regulated CmAPRR2 expression variation is correlated with flesh β-carotene content in mature fruits. Since both parental lines of this population carry a predicted 'dark' allele based on the coding sequence of the CmAPRR2 gene, we assume that this experiment provided another piece of evidence for the relationship between the expression level of CmAPRR2 and flesh pigment content. The proposed involvement of this transcription factor in regulation of plastid development in fruitssupports the broad effect of CmAPRR2 that was described here.
## Multi-allelic nature of the cmaprr2 gene in melon
The genetic architecture that describes a phenotypic trait is strongly dependent on the type and structure of the germplasm used for study. For quantitative polygenic traits, QTL segregation and detection will not necessarily overlap across different mapping populations. The same can apply to simple traits, where independent mutations in different genes, which are involved in a common biological process, lead to the same discrete phenotype. While bi-parental populations will comprise only part of the picture in such cases, diverse collections or multi-parental segregating populations are more effective in comprehensively characterizing this architecture. In the current study, we tried to genetically characterize the light immature rind phenotype in melon using a diverse collection, assuming that it is under a simple genetic control as previously described. In GWAS, lack of detection power can result from low heritability, low frequency of the phenotype under investigation, strong confounding effect of population structure, or insufficient markers density. While none of these factors seemed to apply in our case, we did not obtain any significant GWA signal, which led to the intuitive assumption that multiple genes are associated with the light immature rind phenotype in our collection. The identification of two independent allelic nonsense mutations in the CmAPRR2 gene, through linkage analyses, indicated that we might be looking at a different scenario. Through complementary resequencing of a diverse core panel and comprehensive allelism testing, we were able to demonstrate that this trait is a unique case of simple genetic architecture. On the functional level, it seems to be controlled by a single gene that segregates in a Mendelian manner in bi-parental crosses, but the multi-allelic pattern at the CmAPRR2 gene drove reduced power through the GWAS, which masked this simplicity and created the observed contradiction between the different mapping strategies. These results provide a thought-provoking example of another possible inherent complexity that can arise in GWAS-independent low-frequency causative variants within a common gene. In the current scenario, even wholegenome deep resequencing of the GWAS panel, which would target each of the variants in the CmAPRR2 gene, would not necessarily resolve the lack of detection power, as each of these independent variants remains at low frequency. Genetic mapping studies are rapidly shifting towards sequencing-based genotyping and, in most cases, marker density is no longer a bottleneck in GWAS.
A key challenge remains in prioritizing GWAS signals and improving weak signals obtained from low-frequency causative variants. The availability of wholegenome assemblies and corresponding protein-coding gene annotations, alongside additional layers of information, such as expression profiles from RNA-Seq experiments, now facilitate the integration of multiple data layers to improve GWAS results. Our example of the CmAPRR2 gene suggests that adding functional annotation prediction to GWAS SNPs and treating predicted genes as integral functional units could potentially be used as an informative layer that can boost the signal of causative weak associations.
In summary, we have identified the CmAPRR2 gene as a common regulator of fruit pigmentation in melon and watermelon. The conserved and broad effect of this gene across species, fruit tissues, developmental stages, and different types of pigment accumulated suggests its potential as a useful target for carotenoid bio-fortification of cucurbits and other fruits.
## Supplementary data
Supplementary data are available at JXB online.. cDNA sequence comparison between the mapping population parental lines DUL, TAD, and NA.. Number of annotated transcript haplotypes versus number of SNPs per base pair (a) and transcript length (b) across 2200 genes on chromosome 4.. CmAPRR2 gene is located in a low LD region.. Segregation and scoring of rind color of 87 F 3 families from the NY0016×EMB cross.. Analysis of rind chlorophyll and carotenoids during fruit development on TAD, DUL, and selected F 3 families from their cross.. Prediction of flesh β-carotene based on fruit-section image analyses in the TAD×DUL RILs. . Melon young fruit light rind QTL interval: annotations and positions of genes. . Segregation of light rind in four F 2 watermelon populations. . Light rind QTL interval in watermelon: annotations and positions of genes. . Gene Ontology (GO) enrichment analysis for 200 genes correlated with expression of the APRR2 gene (Melo3C003375) in melon fruit. . List of primers used for RT-qPCR. |
Does HIV Cause Cardiovascular Disease?
## Perspectives
November 2006 | Volume 3 | Issue 11 | e496
## Hiv, antiretroviral therapy, and cardiovascular disease
Cardiovascular disease is an increasing cause of morbidity in HIV-infected adults receiving antiretroviral therapy (ART). ART, particularly protease inhibitors and to a lesser extent nucleoside analogue reverse transcriptase inhibitors, can adversely affect lipid and glucose metabolism [bib_ref] Cardiovascular risk and body fat abnormalities in HIV-infected adults, Grinspoon [/bib_ref]. Moreover, there is a strong correlation between ART duration and the risk of myocardial infarction, an association in part linked to higher plasma levels of total cholesterol and triglyceride and to lower levels of high-density lipoprotein (HDL) cholesterol [bib_ref] Combination antiretroviral therapy and the risk of myocardial infarction, Friis-Møller [/bib_ref]. Paradoxically, interruption of ART also appears to be associated with an increased short-term risk of cardiovascular disease. These fi ndings suggest that HIV itself may also increase cardiovascular risk, and that control of HIV replication might reduce this risk.
HDL cholesterol levels are reduced in untreated HIV infection [bib_ref] Lipids, lipoproteins, triglyceride clearance, and cytokines in human immunodefi ciency virus infection..., Grunfeld [/bib_ref] and in healthy volunteers exposed for a short term to the HIV protease inhibitors ritonavir-boosted atazanavir and lopinavir [bib_ref] The effect of low-dose ritonavir monotherapy on fasting serum lipid concentrations, Shafran [/bib_ref]. In HIV-infected individuals who start ART that effectively suppresses HIV replication, HDL cholesterol levels increase, regardless of whether a protease inhibitor is used [bib_ref] A oncedaily lopinavir/ritonavir-based regimen provides noninferior antiviral activity compared with a twice-daily..., Johnson [/bib_ref] [bib_ref] Nevirapinecontaining antiretroviral therapy in HIV-1 infected patients results in an anti-atherogenic lipid..., Van Der Valk [/bib_ref] , implying that the HIV effect on HDL cholesterol levels is greater than the ART effect.
## How does hiv lower hdl cholesterol levels?
The pathogenesis of low HDL cholesterol levels in untreated HIV infection is unknown. In a study recently published in PLoS Biology, Mujawar et al. investigated whether these lower levels might be due to impaired cholesterol effl ux from macrophages [bib_ref] Human immunodefi ciency virus impairs reverse cholesterol transport from macrophages, Mujawar [/bib_ref] , a process mediated by the ABCA1 cell-surface cholesterol transporter. ABCA1 lipidates apoA-I, the major apolipoprotein in HDL, and thus plays a central role in formation of nascent HDL. Mutations in ABCA1 cause Tangier disease, which is associated with low HDL cholesterol and accelerated atherosclerosis [bib_ref] ABCA1: The gatekeeper for eliminating excess tissue cholesterol, Oram [/bib_ref].
The researchers found that expression of HIV nef, a protein that enhances HIV replication and infectivity, specifi cally inhibited ABCA1-dependent cholesterol effl ux from macrophages (the precursors of foam cells in atherosclerotic plaque) and re-localized ABCA1 to an exclusive plasma membrane distribution. While nef-transfected cells exhibited increased apoA-I binding, apoA-I internalization was blocked, suggesting that nef at the plasma membrane may prevent ABCA1 internalization and subsequent apoA-I lipidation. As cholesterol is required for HIV replication [bib_ref] Lipid rafts and HIV pathogenesis: virion-associated cholesterol is required for fusion and..., Liao [/bib_ref] , Mujawar et al. next examined whether enhancing cholesterol effl ux from HIV-infected macrophages affected HIV replication. Using a liver X receptor (LXR) agonist to transcriptionally upregulate ABCA1, cholesterol effl ux was augmented, signifi cantly reducing virion-associated cholesterol and infectivity.
## What do the findings mean?
Mujawar et al. propose that HIV nef redirects cholesterol from an ABCA1mediated effl ux pathway to viruscontrolled cholesterol transport in order to ensure suffi cient cholesterol for virion assembly. The interruption of a host cholesterol traffi cking pathway by an intracellular pathogen is also employed by Toxoplasma gondii, in which endocytosed low-density lipoprotein cholesterol is diverted to a specialized vacuole to support parasite growth and replication [bib_ref] Toxoplasma gondii exploits host low-density lipoprotein receptor-mediated endocytosis for cholesterol acquisition, Coppens [/bib_ref]. Given that impaired ABCA1 function results in decreased HDL cholesterol and accelerated atherosclerosis, the fi ndings of Mujawar et al. provide a possible mechanism to explain low HDL cholesterol in HIV infection and increased cardiovascular risk in HIVinfected adults.
It remains to be determined, however, if this mechanism is the greatest contributor to these low levels. If so, one would expect that HDL cholesterol levels would fall rapidly after primary HIV infection and return to pre-infection levels with effective ART that did not affect lipid metabolism. In addition, the observed reductions might be inversely proportionate to plasma HIV RNA levels. Alternatively, low HDL cholesterol levels may also be caused by ART-induced lipodystrophy. This ART side effect is characterized by peripheral lipoatrophy and relative central adiposity and is strongly associated with low HDL cholesterol levels as well as insulin resistance and
## Does hiv cause cardiovascular disease? andrew carr * , daniel ory
There is a strong correlation between ART duration and the risk of myocardial infarction.
The Perspectives section is for experts to discuss the clinical practice or public health implications of a published article that is freely available online. hypertriglyceridemia, all features of congenital lipodystrophies in which abnormalities of ABCA1 have not been identifi ed [bib_ref] Combination antiretroviral therapy and the risk of myocardial infarction, Friis-Møller [/bib_ref] [bib_ref] An objective case defi nition of lipodystrophy in HIV-infected adults: A casecontrol..., Carr [/bib_ref] [bib_ref] Acquired and inherited lipodystrophies, Garg [/bib_ref].
## The strengths and limitations of the study
The fi ndings of Mujawar et al. raise several important questions. Is the apparent nef-stimulated decrease in ABCA1 protein levels due to ABCA1 degradation, or possibly due to altered partitioning of ABCA1 into detergent-resistant lipid domains? Does nef physically interact with ABCA1? Use of the myristoylation-defi cient nef in this study only suggests that nef must be targeted to the plasma membrane-likely to cholesterol rafts [bib_ref] The Nef protein of HIV-1 associates with rafts and primes T cells..., Wang [/bib_ref]. Furthermore, experiments involving immunoprecipitation of raftassociated proteins must be interpreted with caution, and should be followed up with more rigorous studies (e.g., identifi cation of interaction domains and fl uorescence resonance energy transfer) that provide support for a specifi c ABCA1-nef interaction at the plasma membrane. Finally, might LXR ligands activate target genes, other than ABCA1, that could attenuate virionassociated cholesterol? Treatment of HIV-infected ABCA1-defi cient macrophages with LXR ligands should establish a defi nitive role for ABCA1 in countering HIV replication.
## Where to from here?
The present data provide a clue as to why HIV infection, as well as ART, might accelerate atherosclerosis. The data support the shift away from a paradigm of delaying or stopping ART to reduce the risk of cardiovascular disease and raise the possibility that antiretroviral drugs without direct metabolic effects may actually reduce cardiovascular risk. The data also emphasize the need for clinicians to consider all the metabolic effects (in particular the ratio of total cholesterol to HDL) of each antiretroviral drug, not just total cholesterol levels, in the management of cardiovascular risk in HIV-infected adults.
The study by Mujawar et al. also has important implications for developing new approaches for suppression of HIV replication. Selective LXR modulators under active development for treatment of atherosclerosis may prove useful as ART.
The long-term effects of each antiretroviral drug on HDL cholesterol, as well as other lipid and glycemic parameters, need to be more completely analyzed in randomized trials, as well as in shorter healthy volunteer studies. It remains to be determined whether new antiretroviral drug classes, such as HIV fusion, CCR5, and integrase inhibitors, will be free of these metabolic complications. |
Centenarians and COVID-19: Is There a Link between Longevity and Better Immune Defense?
[bib_ref] Morbidity profiles of centenarians: survivors, delayers, and escapers, Evert [/bib_ref] [bib_ref] Inflammaging: a new immune -metabolic viewpoint for age-related diseases, Franceschi [/bib_ref] [bib_ref] SARS CoV2 infection the longevity study perspectives, Lio [/bib_ref] [bib_ref] Editorial: COVID-19 spiraling of frailty in older Italian patients, Abbatecola [/bib_ref] [bib_ref] Editorial: COVID-19 spiraling of frailty in older Italian patients, Abbatecola [/bib_ref] [bib_ref] Genetic cartography of longevity in humans and mice: current landscape and horizons, Hook [/bib_ref] [bib_ref] On the road to accurate biomarkers for cardiometabolic diseases by integrating precision..., Scola [/bib_ref] [bib_ref] Anti-inflamm-ageing and/or anti-age-related disease emerging treatments: a historical alchemy or revolutionary effective..., Balistreri [/bib_ref] [bib_ref] Centenarians in nursing homes during the COVID-19 pandemic, Couderc [/bib_ref] [bib_ref] COVID-19 mortality in lombardy: the vulnerability of the oldest old and the..., Marcon [/bib_ref]
## Conflict of interest
The author declares no competing financial or nonfinancial interests.
# Funding sources
P.K. was supported by a special internal fund from the DRI&P, IAR&FR PAS, in Olsztyn (3/FBW/2020).
# Author contributions
P.K. had the conceptual idea of this comment and wrote the letter. |
Genome-Wide Analysis Indicates Lineage-Specific Gene Loss during Papilionoideae Evolution
Gene loss is the driving force for changes in genome and morphology; however, this particular evolutionary event has been poorly investigated in leguminous plants. Legumes (Fabaceae) have some lineage-specific and diagnostic characteristics that are distinct from other angiosperms. To understand the potential role of gene loss in the evolution of legumes, we compared six genomesequenced legume species of Papilionoideae, the largest representative clade of Fabaceae, such as Glycine max, with 34 nonlegume plant species, such as Arabidopsis thaliana. The results showed that the putative orthologs of the 34 Arabidopsis genes belonging to 29 gene families were absent in these legume species but these were conserved in the sequenced nonlegume angiosperm lineages. Further evolutionary analyses indicated that the orthologs of these genes were almost completely lost in the Papillionoideae ancestors, thus designated as the legume lost genes (LLGs), and these underwent purifying selection in nonlegume plants. Most LLGs were functionally unknown. In Arabidopsis, two LLGs were well-known genes that played a role in plant immunity such as HARMLESS TO OZONE LAYER 1 and HOPZ-ACTIVATED RESISTANCE 1, and 16 additional LLGs were predicted to participate in plant-pathogen interactions in in silico expression and protein-protein interaction network analyses. Most of these LLGs' orthologs in various plants were also found to be associated with biotic stress response, indicating the conserved role of these genes in plant defense. The evolutionary implication of LLGs during the development of the ability of symbiotic nitrogen fixation involving plant and bacterial interactions, which is a well-known characteristic of most legumes, is also discussed. Our work sheds light on the evolutionary implication of gene loss events in Papilionoideae evolution, as well as provides new insights into crop design to improve nitrogen fixation capacity.
# Introduction
Genomic changes such as gene gain and loss events frequently occur during genome evolution [bib_ref] An evolutionary analysis of orphan genes in Drosophila, Domazet-Loso [/bib_ref] [bib_ref] Gene loss, protein sequence divergence, gene dispensability, expression level, and interactivity are..., Krylov [/bib_ref]. Genes specifically added to the genome of a species lineage are defined as "taxonomically restricted" genes (TRGs) or orphan genes and these have no significant sequence similarity to genes of other species lineages [bib_ref] Orphans as taxonomically restricted and ecologically important genes, Wilson [/bib_ref]. Comparative genomics has now demonstrated that TRGs are a universal feature of any genome [bib_ref] More than just orphans: are taxonomically-restricted genes important in evolution?, Khalturin [/bib_ref] [bib_ref] The evolutionary origin of orphan genes, Tautz [/bib_ref] [bib_ref] New gene evolution: little did we know, Long [/bib_ref] [bib_ref] Coming of age: orphan genes in plants, Arendsee [/bib_ref] and play essential roles in the species evolution. Drosophila TRGs are involved in the evolution of lineage-specific ecological adaptations [bib_ref] An evolutionary analysis of orphan genes in Drosophila, Domazet-Loso [/bib_ref] , and Hydra TRGs play a role in the creation of phylum-specific novelties and in the innate defense system, and are thus also involved in species-specific adaptive processes [bib_ref] More than just orphans: are taxonomically-restricted genes important in evolution?, Khalturin [/bib_ref]. Lineage-or speciesspecific TRGs have also been identified in various plants such as in Arabidopsis thaliana [bib_ref] Comparative analyses reveal distinct sets of lineagespecific genes within Arabidopsis thaliana, Lin [/bib_ref] , Oryza sativa [bib_ref] Identification and characterization of lineagespecific genes within the Poaceae, Campbell [/bib_ref] , Solanum spp. [bib_ref] Comparative analyses of six solanaceous transcriptomes reveal a high degree of sequence..., Rensink [/bib_ref] , and legumes [bib_ref] Computational identification and characterization of novel genes from legumes, Graham [/bib_ref] [bib_ref] Genome sequence of the palaeopolyploid soybean, Schmutz [/bib_ref]. Some TRGs are preferentially expressed in A. thaliana [bib_ref] Evolutionary origins of Brassicaceae specific genes in Arabidopsis thaliana, Donoghue [/bib_ref] and rice [bib_ref] Significant comparative characteristics between orphan and nonorphan genes in the rice (Oryza..., Guo [/bib_ref] in reaction to abiotic stresses, whereas O. sativa defense-responsive gene 10 (OsDR10), a rice tribe-specific gene, negatively regulates resistance to a broad spectrum of Xanthomonas oryzae pv. oryzae strains [bib_ref] A rice gene of de novo origin negatively regulates pathogen-induced defense response, Xiao [/bib_ref]. Therefore, TRGs are involved in the response to various stresses, thus contributing to adaptive evolution.
Genes can also be deleted from the genome during evolution. Some members in one gene family are often lost in certain lineages, but, in extreme cases, an entire gene family may be deleted from the genomes of certain lineages [bib_ref] Lineage-specific loss and divergence of functionally linked genes in eukaryotes, Aravind [/bib_ref] [bib_ref] The life and death of gene families, Demuth [/bib_ref] , creating lineage-specific lost genes. Gene loss and pseudogenization can lead to immediate loss of gene function, thus severely affecting major physiological processes of organisms. However, it may also open new developmental opportunities, confer a selective advantage, and serve as an engine for evolutionary change in bacterium and animals [bib_ref] When less is more: gene loss as an engine of evolutionary change, Olson [/bib_ref] [bib_ref] Less is more: selective advantages can explain the prevalent loss of biosynthetic..., Souza [/bib_ref] [bib_ref] Fungal evolutionary genomics provides insight into the mechanisms of adaptive divergence in..., Gladieux [/bib_ref]. Loss of superfluous genes contributes to bacteria fitness [bib_ref] Selection-driven gene loss in bacteria, Koskiniemi [/bib_ref] and is the driving force in the adaptation of parasites to eukaryotic cells [bib_ref] Massive comparative genomic analysis reveals convergent evolution of specialized bacteria, Merhej [/bib_ref] [bib_ref] Comparative genomics suggests that the human pathogenic fungus Pneumocystis jirovecii acquired obligate..., Cisse [/bib_ref] [bib_ref] Gene loss rather than gene gain is associated with a host jump..., Sharma [/bib_ref]. The loss of a penicillin-binding protein may contribute to resistance to the cephalosporin drug, ceftazidime, in Burkholderia pseudomallei [bib_ref] Bacterial gene loss as a mechanism for gain of antimicrobial resistance, Torok [/bib_ref]. Human-specific loss of a myosin heavy chain isoform expressed in the masticatory muscles has been linked to the weakening of human jaw muscles, which has been suggested to increase cranial capacity in humans [bib_ref] Myosin gene mutation correlates with anatomical changes in the human lineage, Stedman [/bib_ref].
Gene losses are also involved in the evolutionary divergence of floral morphology in plants. The loss of an anthocyanin pathway enzyme is associated with the transition from blue to red floral pigmentation, thus resulting in phenotypic differences among species of the Andean Iochroma of the Solanaceae [bib_ref] Gene loss and parallel evolution contribute to species difference in flower color, Smith [/bib_ref]. The loss of lineage-specific MADSbox genes such as GLOBOSA or DEFICIENS is potentially associated with the evolutionary divergence of floral morphology during the radiation of the Euasterids I within core eudicots [bib_ref] Gene duplication and loss in a MADS box gene transcription factor circuit, Lee [/bib_ref]. Heterotopic expression of a MADS-box gene 2like from Physalis floridana (MPF2-like) is required for the fruiting calyx inflation trait called "Chinese lantern," yet is physiologically known as inflated calyx syndrome (ICS) in Physalis [bib_ref] Heterotopic expression of MPF2 is the key to the evolution of the..., He [/bib_ref] , whereas loss of a copy of MPF2-like genes is involved in the loss of ICS in Tubocapsicum [bib_ref] MPF2-like-A MADS-box genes control the inflated calyx syndrome in Withania (Solanaceae): roles..., Khan [/bib_ref]. Therefore, the evolution of lineage-specific gene loss can result in new morphological traits among species or genera.
The legume (Fabaceae) consisting of three clades, Papilionoideae, Caesalpinioideae, and Mimosaceae, includes important grain, pasture, and agroforestry species and is characterized by unusual flower structure, podded fruit, and the ability of most species to form nodules with rhizobia [bib_ref] Occurrence of nodulation in the leguminosae, De Faria [/bib_ref]. However, genetic variations that could distinguish legumes from nonlegumes have not been identified. In particular, nodule formation is a developmental process connecting plant and bacterial cell differentiation [bib_ref] An integrated analysis of plant and bacterial gene expression in symbiotic root..., Roux [/bib_ref]. Much of this process remains a mystery although a few symbiosisrelated genes have been identified in legumes [bib_ref] A plant regulator controlling development of symbiotic root nodules, Schauser [/bib_ref] [bib_ref] Four genes of Medicago truncatula controlling components of a nod factor transduction..., Catoira [/bib_ref] [bib_ref] LysM domain receptor kinases regulating rhizobial Nod factor-induced infection, Limpens [/bib_ref]. Genome sequencing of Glycine max and its comparison with distantly related species such as Populus trichocarpa has revealed specific gene gains in legumes [bib_ref] Genome sequence of the palaeopolyploid soybean, Schmutz [/bib_ref]. However, gene loss has not been evaluated in relation to the evolution of legumes. No genome of Caesalpinioideae and Mimosaceae has yet been sequenced, but the whole-genome sequencing of five additional legume species in Papilionoideae, the largest and most widely distributed clade of Fabaceae such as Lotus japonicus [bib_ref] Genome structure of the legume, Lotus japonicus, Sato [/bib_ref] , Medicago truncatula [bib_ref] The Medicago genome provides insight into the evolution of rhizobial symbioses, Young [/bib_ref] , Cajanus cajan [bib_ref] Draft genome sequence of pigeonpea (Cajanus cajan), an orphan legume crop of..., Varshney [/bib_ref] , Cicer arietinum [bib_ref] Draft genome sequence of chickpea (Cicer arietinum) provides a resource for trait..., Varshney [/bib_ref] , and Phaseolus vulgaris [bib_ref] A reference genome for common bean and genome-wide analysis of dual domestications, Schmutz [/bib_ref] has allowed investigations on lineage-specific losses in Papilionoideae, thereby gaining insights into the adaptive role of gene loss in the entire legume family. In this study, we identified the legume lost genes (LLGs) through genome-wide comparative analyses of legume and nonlegume species. Thirty-four Arabidopsis genes had orthologs in nonlegume species but were not detected in legumes. Eighteen LLGs were directly or indirectly inferred to function in the plant-pathogen interaction in nonlegumes. Therefore, the loss of these genes might have partially contributed to genomic changes that were related to the evolution of symbiotic nitrogen fixation in legumes.
# Materials and methods
## Sequence availability
Whole genome-wide primary transcript sequences of G. max, P. vulgaris, M. truncatula, and 34 nonlegume species were downloaded from Phytozome v10 (http://www.phytozome. net, last accessed March 1, 2015). Sequences of L. japonicus were obtained from Kazusa DNA Research Institute (http:// www.kazusa.or.jp/lotus, last accessed March 1, 2015), and those of C. cajan and C. arietinum were downloaded from the International Crops Research Institute for the Semi-Arid Tropics (http://www.icrisat.org, last accessed March 1, 2015). The version of each database is summarized in supplementary table S1, Supplementary Material online.
## Identification of llgs
To identify possible LLGs, we used all coding sequences (CDS) of A. thaliana to conduct Basic Local Alignment Search Tool (BLAST) analysis of sequences of earlier described legume species. The identification steps are presented in supplementary figure S1, Supplementary Material online. Putative LLGs were selected under the BLAST results, with an E-value cutoff of 1 Â 10 À5 . To rule out Arabidopsis-specific genes, the putative LLGs were searched in three selected genomes of Vitis vinifera, Prunus persica, and P. trichocarpa (E-value = 1 Â 10 À10 ). Arabidopsis genes without any hits in six legume species but showing homologous sequences in all three nonlegume species were further verified at the protein level in the aforementioned nine species using bidirectional BLASTP (E-value = 1 Â 10 À4 ), and bidirectional best hits were defined as putative orthologs. When the best hit in a species of one putative LLG in Arabidopsis was also the best match for another putative LLG in Arabidopsis, the putative LLG's ortholog (bidirectional best hit) was not considered to be lost from this species. When proteins with bidirectional best hits in three nonlegume species had hits in legumes with an E-value of <1 Â 10 À4 but did not show bidirectional best hits, these were defined as Group 1 LLGs, indicating that the legume species had lost orthologs of the LLGs. Protein sequences without any hits in the legume species but with bidirectional best hits in the three nonlegume species were defined as Group 2 LLGs. These BLAST results were further verified by orthoMCL v1.4 analysis [bib_ref] OrthoMCL: identification of ortholog groups for eukaryotic genomes, Li [/bib_ref]. Phylogenetic analyses were performed whenever a conflicting signal was observed among bidirectional BLAST and orthoMCL. Nucleotide sequences were aligned using the Clustal X v2.1 program with default parameters [bib_ref] Clustal W and Clustal X version 2.0, Larkin [/bib_ref]. Alignments were optimized via manual adjustment, and partial sequences with poor alignment were excluded. Substitution saturation was tested using DAMBE v6.0.1 before phylogenetic analysis [bib_ref] DAMBE5: a comprehensive software package for data analysis in molecular biology and..., Xia [/bib_ref]. Unrooted maximumlikelihood trees were constructed using the PhyML v3.1 program using a generalized time-reversible model with 100 bootstrap resamplings [bib_ref] New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of..., Guindon [/bib_ref].
The LLGs' orthologs were further characterized in 29 other angiosperm species and Selaginella moellendorffii by bidirectional BLASTP using protein sequences (E-value = 1 Â 10 À4 ). The phylogeny of the involved plant species was derived from Cogepedia (http://genomevolution.org/wiki/index.php/ Sequenced_plant_genomes, last accessed March 1, 2015) and APG III (Angiosperm Phylogeny Group 2009). Ancestral character state reconstruction was performed with the Markov k-state 1 parameter model (Mk1) in Mesquite 3.03 (http://mesquiteproject.org, last accessed March 1, 2015). Information on the gene families of the LLGs was generated from the Phytozome v10 clusters at the angiosperm node. Gene ontology (GO) annotations of LLGs were derived from Blast2GO [bib_ref] Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics..., Conesa [/bib_ref] using the National Center for Biotechnology Information nonredundant database .
## Identification of conserved genes in angiosperms
The genes conserved in A. thaliana, V. vinifera, P. trichocarpa, P. persica, and all six legumes were used as controls throughout the work. To identify conserved genes, the primary CDS sequences of A. thaliana were subjected to BLAST analysis using the aforementioned three nonlegume and six legume species (E-value = 1 Â 10 À5 ), and then Arabidopsis genes showing homologs in the aforementioned nine species were subjected to BLASTP. The resulting proteinencoding genes showing bidirectional best hits in all nine species (E-value = 1 Â 10 À4 ) were designated as conserved genes.
Gene Structure Analysis CDS length, intron number, and intron length covering the CDS of the identified LLGs in nonlegume species were obtained from gff3 profiles downloaded from Phytozome v10 and TAIR10 (http://www.arabidopsis.org, last accessed March 1, 2015).
## In silico expression prediction
The expression data of roots, seedlings, expanding leaves, stems, vegetative shoot meristems, whole inflorescences, flowers, and fruits of Arabidopsis were obtained from a previous work [bib_ref] At-TAX: a whole genome tiling array resource for developmental expression analysis and..., Laubinger [/bib_ref]. Relative gene expression levels (Zscores) in different tissues were calculated as previously described [bib_ref] A gene expression atlas of the model legume Medicago truncatula, Benedito [/bib_ref]. When the Z-score value of a given gene in a tissue was not <1.5, the gene was considered highly expressed in the tissue. Gene expression of the identified LLGs under various hormonal treatments and biotic stresses was based on a previous study [bib_ref] Integration of Arabidopsis thaliana stress-related transcript profiles, promoter structures, and cell-specific expression, Ma [/bib_ref]. The expression data of the identified LLGs' orthologs in response to biotic stresses in tomato were taken from the Tomato Functional Genomics database (http://ted.bti.cornell.edu/, last accessed March 1, 2015), and related data of rice and grape were extracted from the Gene Expression Omnibus database (http://www.ncbi.nlm.nih. gov/projects/geo, last accessed March 1, 2015). Heat map of gene expression was performed using MeV 4.9.0 (http://www. tm4.org/mev.html, last accessed March 1, 2015).
## Gene coexpression and enrichment analyses
The genes coexpressed with the LLGs in A. thaliana were identified using the MAS5 algorithm in CressExpress v3.2 (http:// cressexpress.org, last accessed March 1, 2015). The employed parameters were as follows: cutoff value for Kolmogorov-Smirnov quality-control statistic was 0.15, and R 2 threshold for pathway-level coexpression was set as 0.36. Genes coexpressed with each LLG were sorted by correlation index, and the top 50 (if <50, then all genes were used) were used in enrichment analysis. Protein sequences of genes coexpressed with each LLG were submitted to KOBAS 2.0 (http://kobas.cbi.pku.edu. cn, last accessed March 1, 2015) for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Biocyc enrichment analyses. GO enrichment analysis was performed on AGRIGO v1.2 (http://bioinfo.cau.edu.cn/agriGO, last accessed March 1, 2015), with the reference genome locus obtained from TAIR10. The P value of all enrichment analyses was set as 0.05.
## Protein-protein interaction prediction
The identified LLGs in Arabidopsis were submitted to the A. thaliana section of STRING 10 (http://string-db.org, last accessed March 1, 2015) to generate protein-protein interaction (PPI) networks.
# Microcolinearity analysis
To identify a syntenic block harboring a LLG between a legume species and a nonlegume species, we first determined the respectively 10 genes upstream and downstream of each LLG in the genome of Arabidopsis. OrthoMCL v1.4 [bib_ref] OrthoMCL: identification of ortholog groups for eukaryotic genomes, Li [/bib_ref] was used to construct orthologous groups around LLGs across multiple plant taxa with default parameters. The schematic diagram of local genomic synteny was drawn manually.
## Selection test
Gene selection was evaluated by using the o (o = dN/dS; dN, nonsynonymous substitution rates; dS, synonymous substitution rates) that was calculated by the codeml program of PAML v4.8 under Models 0 [bib_ref] PAML 4: phylogenetic analysis by maximum likelihood, Yang [/bib_ref]. The protein sequences were aligned using Clustal-omega v1.2.0, with default parameters [bib_ref] Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega, Sievers [/bib_ref] , and then the nucleotide sequences were aligned using a Perl module derived from ParaAT [bib_ref] ParaAT: a parallel tool for constructing multiple protein-coding DNA alignments, Zhang [/bib_ref]. Poorly aligned regions were removed from each alignment using Gblocks v0.91 b [bib_ref] Selection of conserved blocks from multiple alignments for their use in phylogenetic..., Castresana [/bib_ref] with the following parameters: type of sequence = codons, maximum number of contiguous nonconserved positions = 8, minimum length of a block = 10, and gap positions excluded from all sequences. After performing Gblocks, matrixes with nucleotide sites < 50 were discarded. The significance in o difference was evaluated using the Kolmogorov-Smirnov test.
# Results
## Llgs survey
To identify LLGs, we selected four nonlegume species and six legume species to initiate genome-wide comparisons (supplementary , Supplementary Material online). We first used the Arabidopsis genome to probe the genomes of nine other plant species at the nucleotide level, and determined that 70 Arabidopsis genes had homologous sequences in three other nonlegume species, whereas their homologous sequences were not detected in the six legume species (supplementary table S2, Supplementary Material online), indicating a potential loss of these genes in legumes. We then verified these identified putative LLGs at the protein level. Using BLASTP, 34 of these Arabidopsis genes were found to have putative orthologous proteins in all three nonlegume species (V. vinifera, P. trichocarpa, and P. persica), whereas no putative orthologous proteins were detected in the six legume species, Supplementary Material online). These genes were designated as LLGs and were further divided into two groups. In the legume species, 26 of the LLGs lost their orthologous genes but had homologous sequences and were thus designated as Group 1 LLGs. On the other hand, Group 2 LLGs included eight genes without any homologous sequences in various legume species.
We further performed orthoMCL analyses in the legume species, and 33 orthologs of the aforementioned LLGs were not detected, thus supporting the BLAST results. Nevertheless, a few inconsistencies were observed. No orthologs of AT1G68940 were detected in the legume species using BLAST, whereas in the orthoMCL analyses, AT1G68940 was determined to have possible orthologs in the legume species, and AT3G61210, AT2G43910, and AT2G43920 seemed to be specific to Arabidopsis. To assess these inconsistencies, phylogenetic analyses of related genes were performed (supplementary [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref] , Supplementary Material online), and the results suggested that the orthologs of these genes were present in nonlegumes but absent from legumes. Thus, we ultimately identified 34 LLGs that belonged to 29 gene families in A. thaliana (supplementary table S3, Supplementary Material online). Group 2 LLGs (8 genes) belonged to seven families, and six LLGs in this group formed six single-copy gene family, except for AT2G43910 and AT2G43920 being homologs. In Group 1 LLGs (26 genes), ten were from single-copy gene families in Arabidopsis
## Evidence of gene loss from microsynteny analyses
To verify LLGs, we assessed for genomic microsynteny around LLGs in nonlegume species in relation to that of legumes. Although these plant species have increasing taxonomic distance and complicated genome structure due to duplications, losses, and segmental reshuffling during evolution, the stretches of syntenic chromosomal segments could be still identified. For example, in nonlegumes (A. thaliana, V. vinifera, P. trichocarpa, and P. persica), 1-3 HOL genes were clustered, and their downstream and upstream regions shared a few conserved genes, which was indicative of synteny [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref]. However, no HOL homologous sequences were detected in four legume species (G. max, P. vulgaris, C. cajan, and C. arietinum), although these were conserved in nonlegumes [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref] , thus indicating loss of HOL genes in these legumes. Altogether, 20 LLGs were found to maintain a relatively good local synteny among nonlegume species compared with legumes (supplementary table S4, Supplementary Material online). Thus, well maintenance of microsynteny verified the LLGs.
To understand the evolutionary implications of LLGs, we next investigated their evolution history in nonlegume plants.
## The evolution of llgs in nonlegume angiosperms
## Selection pressure
We evaluated the selection pressure of these LLGs in the aforementioned four nonlegume plant species through calculating dN/dS (o). We found that the o values of LLGs was <0.35 (0.08 < o < 0.33) suggesting that these LLGs might have undergone purifying selection during nonlegume evolution. Moreover, there was no difference in selection pressures between Groups 1 and 2 LLGs (P = 0.15). We further identified 5,935 Arabidopsis genes having reciprocal best hits in the genomes of three nonlegumes and six legumes as conserved genes in angiosperms (supplementary table S5, Supplementary Material online). The o of these genes in legumes ranged from 0.00010 to 0.65 (supplementary table S5, Supplementary Material online), while it ranged from 0.00064 to 0.48 in the four nonlegume species (indicated by black columns, , indicating these conserved genes underwent purifying selection in both nonlegumes and legumes. However, the o distribution of these conserved genes in nonlegumes was significantly different from that of LLGs (P = 7.81eÀ07; . These results indicated that LLGs were generally conserved during nonlegume evolution, yet might have undergone a relatively relaxed purifying selection compared with conserved genes.
## Gene loss patterns
We next investigated the loss pattern of LLGs, including additional 30 genome-sequenced plants that included 29 angiosperm species and S. moellendorffii (supplementary table S1, y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT5G04840.1 (1) y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT5G10830.1 (1) y (2) y (1) y (2) n (0) n (0) n (0) n (0) n (0) n (0) AT5G12460.1 (1) y (1) y (2) y (2) n (0) n (0) n (0) n (0) n (0) n (0) AT5G66160.1 (1) y (1) y (1) y (1) n (0) n (0) n (0) n (0) n (0) n (0) RMR1 Vv, Vitis vinifera; Pt, Populus trichocarpa; Pp, Prunus persica; Gm, Glycine max; Pv, Phaseolus vulgaris; Cc, Cajanus cajan; Mt, Medicago truncatula; Ca, Cicer arietinum; Lj, Lotus japonicus. The symbol y indicates that the LLG had putative orthologs in this species, and n represents that the LLG had homologous protein sequences but no putative ortholog in this species. N indicates that the LLG had no homologous sequence in this species. Genes in bold indicate the Group 2 LLGs. The numbers in parenthesis represent number of genes in this species within the orthoMCL group containing LLG. a HOL1 and HOL2 are in one orthoMCL group; and HOL1, HOL2, and HOL3 belong to the HOL family. b AT2G18520 and AT4G36680 are in one orthoMCL group. Supplementary Material online). The orthologs of LLGs in these plant species were identified (supplementary table S6, Supplementary Material online), and the presence-absence pattern of the orthologs of each LLG was mapped to the phylogenetic tree of the involved plants [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref]. The presence of the ortholog was indicated in blue. When no ortholog was detected, the presence of putative homologs was indicated in orange. In extreme cases, the absence of homologs was highlighted in gray. The number of species with LLG orthologs in nonlegumes ranged from 17 to 33, and on average, around 28 nonlegume species harbored LLG orthologs. These results again indicated that LLGs were conserved in most plants, whereas these were lost in legumes. Moreover, the absence of LLGs in nonlegumes seemed to be independent of their phylogeny [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref] , that is, Cucurbitales and Rosales are closely related to legumes, and the orthologs of 15 LLGs were not detected in C. sativus, whereas these were detected in P. persica, indicating that LLGs might have been randomly lost in a few nonlegume species during evolution. This is consistent with observations made in Poaceae (Poales) and Malpighiales. Poaceae are very different from legumes, and only six LLGs (AT1G09195, AT4G22160, AT2G39100, AT5G12460, AT1G71120, and AT3G50950) were absent from all investigated species of Poaceae [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref] , hinting that these might be lost in Poaceae, which was supported by ancestral state reconstruction (supplementary [fig_ref] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis [/fig_ref] , Supplementary Material online). A few LLGs were absent from Euphorbiaceae (Malpighiales) but existed in other species within Malpighiales. Some LLGs, such as AT5G44010, AT4G14970, and AT5G49110, may have undergone multiple, independent gene loss events throughout Angiospermae [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref]. Although the systematic loss of a few LLGs in other lineages could not be excluded currently (supplementary [fig_ref] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis [/fig_ref] , Supplementary Material online), the gene loss pattern of these LLGs in nonlegumes was similar to that of the previously identified 5,935 conserved genes in angiosperms (supplementary [fig_ref] FIG. 5: -The largest PPI network associated with LLGs [/fig_ref] , Supplementary Material online) but different from the evolutionary pattern of legume LLGs that was apparently specific and systematic [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref]. Ancestral state reconstruction further showed that the ancestors of Papillionoideae might have lost most of the LLG orthologs, whereas these LLGs' homologs were found in the ancestors of angiosperms with a few exclusions such as AT1G35340 [fig_ref] FIG. 3: -LLG evolution in sequenced angiosperms [/fig_ref] , Supplementary Material online).
[formula] AT1G09195.2 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT1G35340.1 (1) y (1) y (1) y (1) N (0) N (0) N (0) n (0) n (0) N (0) AT1G64385.1 (1) y (1) y (2) y (1) N (0) n (0) N (0) n (0) N (0) N (0) AT2G43210.1 (1) y (1) y (2) y (1) n (0) n (0) n (0) N (0) N (0) n (0) AT2G43910.2 (2) a y (0) y (0) y (0) N (0) N (0) N (0) N (0) N (0) N (0) HOL1 AT2G43920.1 (2) a y (0) y (0) y (0) N (0) N (0) N (0) N (0) N (0) N (0) HOL2 AT2G43940.1 (1) a y (2) y (1) y (1) n (0) n (0) N (0) N (0) n (0) N (0) HOL3 AT4G14970.1 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT4G22160.2 (1) y (2) y (3) y (1) N (0) N (0) N (0) n (0) n (0) N (0) AT4G29560.1 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT5G44010.1 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT5G49110.2 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT5G65740.2 (1) y (1) y (1) y (1) N (0) N (0) N (0) N (0) N (0) N (0) AT1G13630.1 (1) y (1) y (1) y (2) n (0) n (0) n (0) n (0) n (0) n (0) AT1G55580.1 (1) y (1) y (3) y (1) n (0) n (0) n (0) n (0) n (0) n (0) LAS AT1G55590.1 (1) y (1) y (1) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT1G68940.3 (3) d y (2) y (4) y (2) n (2) n (1) n (1) n (1) n (1) n (1) AT1G71120.1 (1) y (1) y (1) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT2G05810.1 (1) y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT2G18520.1 (2) b y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT4G36680.1 (2) b y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT2G39100.1 (1) y (1) y (1) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT2G45530.1 (1) y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT3G24515.1 (1) y (1) y (2) y (1) n (0) n (0) n (0) n (0) n (0) n (0) AT3G50950.2 (1) y (1) y (1) y (2) n (0) n (0) n (0) n (0) n (0) n (0)ZAR1 [/formula]
Comparison of structural features (i.e., protein length, intron number, and average intron length) of conserved genes and LLGs in Arabidopsis (supplementary table S7, Supplementary Material online), as well as assessment of the evolution of LLG structural characteristics in sequenced species (supplementary figs. S6-S8, Supplementary Material online) did not indicate any distinct structural variations.
## Functional clues of the llgs in nonlegumes
We further explored the evolutionary implications of these LLGs by investigating the functional roles of these genes in nonlegumes. Literature search revealed that only six LLGs were functionally inferred, which included HOL1, HOL2, HOL3, LATERAL SUPPRESSOR (LAS), receptor homology-transmembrane-ring H2 domain Protein 1 (RMR1), and HOPZ-ACTIVATED RESISTANCE 1 (ZAR1), and GO annotation revealed that LLGs may involve different biological processes such as those that participate in DNA repair (GO: 0006281) of AT4G14970, AT5G49110, and AT5G65740 (supplementary table S2, Supplementary Material online). Some LLGs have been found to be associated with stress response such as HOL1, HOL2, and ZAR1 (GO: 0006952). AT5G10830 was annotated for respiratory burst involved in defense response (GO: 0002679), and AT3G61210 was annotated for response to ethylene and salt stress (GO: 0009723 and GO: 0009651). The functions of unknown LLGs in nonlegumes were next envisioned through in silico expression and PPI analyses in Arabidopsis because information on these LLGs in other angiosperms is limited. Forty plant species whose genomes have been sequenced are included, and their phylogeny was deduced from Cogepedia and APG III. Selaginella moellendorffii was used as outgroup.
## Expression of llgs in arabidopsis
We investigated the expression of LLGs in different tissues of the plant model, Arabidopsis (see Materials and Methods). Besides lacking AT1G09195 expression, the remaining LLGs were differentially expressed in eight tissues, whereas some LLGs showed distinct tissue-specific expression patterns (supplementary . Eleven LLGs were highly expressed in roots such as LAS, HOL3, AT4G24340, and AT5G10830, whereas LLGs such as AT1G35340, AT3G50950, HOL2, AT4G24350, and AT1G71120 were highly expressed in expanding leaves, and LLGs AT1G64385, AT2G39100, and AT5G04840 were preferentially expressed in fruits (indicated by the red boxes, supplementary . On the other hand, no LLGs were highly expressed in Arabidopsis flowers (Z-score < 0.94).
We also investigated LLG expression in response to various biotic stimuli (e.g., hormones, elicitors, and pathogens). The transcript profiles of 30 LLGs were detected, whereas those of four LLGs such as AT1G09195, AT5G65740, AT1G64385, and HOL3 were not detected. In addition, 14 LLGs were involved in response to biotic treatments, including the two function-known genes, HOL1 and ZAR1 [fig_ref] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis [/fig_ref]. The expression of HOL1 and its close homolog, HOL2, responded to hormone treatments such as abscisic acid (ABA), 1-aminocyclopropane-1-carboxylic acid (ACC), methyl jasmonate (MeJA), as well as to elicitors such as hairpin z (hrpz). In addition, challenging with bacterial pathogens such as Pseudomonas syringae pv. tomato DC3000 (PstDC3000), P. syringae pv. tomato avrRPM1 (Pstavrrpm1), and Botrytis cinerea (B. cinerea) resulted in a significant downregulation in expression of the two HOL genes [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref] , Supplementary Material online). ZAR1 also responded to these hormonal treatments and elicitors such as hrpz, and its expression was downregulated during PstDC3000 and Pstavrrpm1 treatments, but upregulated during P. syringae pv. tomato DC3000 hrcC (Psthrcc) and P. syringae pv. phaseolicola (Pstpsph) treatments [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref] , Supplementary Material online). Besides these three genes, Arabidopsis orthologs of 11 functionally unknown LLGs also showed differential expression (both upregulation and downregulation) in response to these treatments, which included AT1G35340, AT2G05810, AT2G18520, AT2G39100, AT3G61210, AT4G24340, AT4G24350, AT4G36680, AT5G01015, AT5G10830, and AT5G44010 [fig_ref] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis [/fig_ref] , indicating that these genes are probably also involved in plant defense response in Arabidopsis. We further investigated the expression of these LLG orthologs in other plants whose transcriptomic variations challenging its own bacterial pathogens are publically available (supplementary table S8, Supplementary Material online), which showed that the orthologs of fourteen LLGs, such as AT1G35340 in rice, tomato, and grape also responded to various biotic stresses (table 2) suggesting that the role of each LLG may be conserved in other nonlegumes.
## Enrichment analysis of genes coexpressed with llgs
Genes in the same pathway and genes that have related functions often exhibit similar expression patterns, which is why analysis of gene coexpression networks is a useful way of developing functional annotation [bib_ref] Co-expression tools for plant biology: opportunities for hypothesis generation and caveats, Usadel [/bib_ref] [bib_ref] Comparative analyses reveal distinct sets of lineagespecific genes within Arabidopsis thaliana, Lin [/bib_ref] [bib_ref] Gene coexpression network analysis as a source of functional annotation for rice..., Childs [/bib_ref]. To further explore the function of these LLGs, we performed coexpression analyses. A total of 21 LLG coexpressed gene sets were detected (R 2 > 0.36), which in turn were further subjected to functional enrichment analysis. The three coexpressed LLG groups included AT1G35340/ AT5G01015, LAS/AT2G05810/AT4G29560, and AT2G18520/AT4G36680 , Supplementary Material online). Moreover, the genes coexpressed with LLGs were putatively involved in multiple fundamental biological processes such as DNA replication, ribosome biogenesis, protein processing, and secondary metabolites (supplementary table S10, Supplementary Material online). DNA replication (ath03030) was significantly enriched in the coexpressed genes of AT4G14970 (supplementary table S10, Supplementary Material online). Ribosome biogenesis (ath03008) and pyrimidine ribonucleotide biosynthetic process (GO: 0009220) were significantly enriched in the coexpression genes of AT2G18520 and AT4G36680, whereas protein processing in the endoplasmic reticulum (ath04141) and purine transport (GO: 0006863) were simultaneously enriched in the coexpressed genes of AT2G05810 and LAS (supplementary table S10, Supplementary Material online). The enriched KEGG pathway associated with secondary metabolites such as phenylpropanoid biosynthesis (ath00940) was significantly enriched in coexpression genes of AT5G10830 (supplementary table S10, Supplementary Material online). Furthermore, three LLGs (ZAR1, AT4G24340, and AT4G24350) were associated with plant defense (supplementary table S10, Supplementary Material online). For example, defense-response and incompatible interaction (GO: 0009814) and plant-pathogen interaction (ath04626) were enriched in the gene set that was coexpressed with ZAR1, whereas plant-type hypersensitive response (GO: 0010363) and jasmonic acid biosynthesis (PWY-735) were enriched in the gene set that was coexpressed with AT4G24340/ AT4G24350 (supplementary table S10, Supplementary Material online).
## Ppis associated with llgs
We further predicted PPI networks associated with the identified LLGs. Among all the LLGs examined, four PPI networks were detected (supplementary [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref] , Supplementary Material online). Three networks involved in each of the two LLGs such as AT2G43210/AT2G05810, AT2G18520/ AT4G36680, and AT4G24340/AT4G24350, whereas the largest PPI network was associated with four LLGs, AT3G24515, AT4G14970, AT5G49110, and AT5G65740 (supplementary [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref] , Supplementary Material online), suggesting that these played a role in protein ubiquitination because AT3G24515 encoded the putative ubiquitinconjugating enzyme 37 (UBC37).
We also observed that LLGs or LLG PPI networks interacted with various non-LLG proteins that were involved in a wide range of biological processes such as DNA repair, cell division, protein processing, and plant defense (supplementary tables S11 and S12, Supplementary Material online). The LLG (AT2G43210)-interacting proteins included Arabidopsis Cell Division Cycle 48B (AtCDC48B), AtCDC48C, AtCDC48, Radiation-Sensitive 23B (RAD23B), RAD23C, and RAD23D (supplementary table S11, Supplementary Material online), which was suggestive of its involvement in cell division. On the other hand, LLG AT5G10830 appeared to be associated with factors such as BONZAI association Protein 1 ; supplementary table S11, Supplementary Material online), which suggested that it might play a role in plant immunity. Interestingly, the largest PPI network of the four LLGs interacted with ten non-LLG proteins [fig_ref] FIG. 5: -The largest PPI network associated with LLGs [/fig_ref]. Eight of these proteins such as ataxia-telangiectasia mutated (ATM), ataxia telangiectasia-mutated and Rad3-related protein (ATR), Nijmegen breakage Syndrome 1 (NBS1), breast cancer 2-like B (BRCA2B), BRCA2 (IV), meiotic recombination 11 (MRE11), ultraviolet-hypersensitive 1 (UVH1), and Fanconi/Fancd2-associated Nuclease I (FAN1) participated in DNA repair, whereas Other possible developmental roles of these LLGs genes could not be ruled out in plants but extensive data mining suggested that LLGs and their orthologs in nonlegumes are associated primarily with plant defense response [fig_ref] Table 2: Summary of LLGs Involved in Plant Defense Response [/fig_ref].
# Discussion
Gene loss has been investigated in the past decade [bib_ref] Lineage-specific loss and divergence of functionally linked genes in eukaryotes, Aravind [/bib_ref] [bib_ref] Microbial minimalism: genome reduction in bacterial pathogens, Moran [/bib_ref] [bib_ref] Gene loss, protein sequence divergence, gene dispensability, expression level, and interactivity are..., Krylov [/bib_ref] [bib_ref] Gene losses during human origins, Wang [/bib_ref] ; however, its importance has only recently attracted attention. Gene loss probably affects organisms to a greater extent than do most amino acid substitutions, thus serving as one of the main drivers in the evolution of gene families, morphological diversity, and adaptation [bib_ref] Gene duplication and loss in a MADS box gene transcription factor circuit, Lee [/bib_ref] [bib_ref] Gene loss and parallel evolution contribute to species difference in flower color, Smith [/bib_ref] [bib_ref] Selection-driven gene loss in bacteria, Koskiniemi [/bib_ref] [bib_ref] Convergent gene loss following gene and genome duplications creates single-copy families in..., Smet [/bib_ref] [bib_ref] The loss of adipokine genes in the chicken genome and implications for..., Dakovic [/bib_ref] , as well as in organogenesis and speciation [bib_ref] Multiple rounds of speciation associated with reciprocal gene loss in polyploid yeasts, Scannell [/bib_ref] [bib_ref] Recurrent gene loss correlates with the evolution of stomach phenotypes in gnathostome..., Castro [/bib_ref]. However, gene loss during legume evolution has not been extensively investigated. In this study, we evaluated gene loss events that might have occurred during the evolution of Papilionoideae at the genomelevel, and identified 34 LLGs that were lost in a legume-specific manner [fig_ref] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species [/fig_ref]. Altogether 21 LLGs and orthologs in nonlegume species were determined to be associated with plant defense systems [fig_ref] Table 2: Summary of LLGs Involved in Plant Defense Response [/fig_ref]. Therefore, adaptive evolution of Papilionoideae might be implicated in the evolution of these LLGs.
## Llgs are largely involved in plant defense response in nonlegumes
The identified LLGs belonged to multiple gene families, and most of these were not functionally inferred. Based on literature search, gene expression analysis, and PPI prediction, we determined that LLGs might have played diverse roles in nonlegumes. LLG AT5G66160, which encodes the receptor homology region transmembrane domain ring H2 motif Protein 1 (AtRMR1), functions as the sorting receptor of phaseolin, thus facilitating in trafficking protein molecules to its corresponding storage vacuole [bib_ref] AtRMR1 functions as a cargo receptor for protein trafficking to the protein..., Park [/bib_ref]. LLG AT1G55580, which encodes LAS, plays a key regulatory role in the formation of lateral shoots during the vegetative development of tomato [bib_ref] The Lateral suppressor (Ls) gene of tomato encodes a new member of..., Schumacher [/bib_ref] , Arabidopsis [bib_ref] Molecular analysis of the LATERAL SUPPRESSOR gene in Arabidopsis reveals a conserved..., Greb [/bib_ref] , and cucumber [bib_ref] The Cucumber Lateral Suppressor gene (CLS) is functionally associated with axillary meristem..., Yuan [/bib_ref]. On the other hand, LLG AT2G43210 is possibly involved in cell division because most of its protein partners play essential roles in the cell cycle (supplementary table S11, Supplementary Material online; [bib_ref] In planta analysis of the cell cycledependent localization of AtCDC48A and its..., Park [/bib_ref] [bib_ref] The RAD23 family provides an essential connection between the 26S proteasome and..., Farmer [/bib_ref]. We also determined that a substantial amount of LLGs were involved in biotic stress responses [fig_ref] Table 2: Summary of LLGs Involved in Plant Defense Response [/fig_ref]. HOL1 is involved in the defense response to pathogens in Arabidopsis (Nagatoshi and Nakamura 2009), whereas ZAR1 is responsible for the recognition of the P. syringae Type III secreted effector HopZ1a, which attenuates HopZ1a virulence [bib_ref] Allele-specific virulence attenuation of the Pseudomonas syringae HopZ1a type III effector via..., Lewis [/bib_ref]. Similar to the two well-known defense-response genes, 12 LLGs were determined to respond to various biotic stresses in Arabidopsis [fig_ref] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis [/fig_ref] , and ten of the 14 LLGs were also detected in either rice, tomato, or grape (table 2), indicating that these might also participate in the defense response. Moreover, genes sharing a common role or function in a particular pathway were coexpressed. We observed that genes coexpressed with LLGs ZAR1, AT4G24340, and AT4G24350 were enriched in the regulation of plant defense response (supplementary table S10, Supplementary Material online).
The proteins in a PPI network are also probably involved in the same functional pathway [bib_ref] Comparative genomics reveals new candidate genes involved in selenium metabolism in prokaryotes, Lin [/bib_ref] [bib_ref] Novel genes affecting blood pressure detected via gene-based association analysis, Zhang [/bib_ref]. Our PPI network prediction provides substantial informative functional clues for some LLGs. Notably, the largest PPI network that associated four LLGs AT3G24515 (UBC37), AT5G49110, AT5G65740, and AT4G14970 interacted with ATM, ATR, NBS1, BRCA2B, BRCA2(IV), MRE11, UVH1, and FAN1. Arabidopsis ATM, ATR, and NBS1 were involved in double-strand breaks of meiosis [bib_ref] AtATM is essential for meiosis and the somatic response to DNA damage..., Garcia [/bib_ref] [bib_ref] NBS1 is involved in DNA repair and plays a synergistic role with..., Waterworth [/bib_ref] [bib_ref] Both ATM and ATR promote the efficient and accurate processing of programmed..., Culligan [/bib_ref] , and BRCA2B and BRCA2(IV) are important for both DNA break
[formula] AT1G35340.1 Y Y Y Y AT1G64385.1 Y AT2G43910.2 (HOL1) Y Y AT2G43920.1 (HOL2) Y AT4G14970.1 Y AT4G29560.1 Y AT5G44010.1 Y Y AT5G49110.2 Y AT5G65740.2 Y AT2G05810.1 Y Y AT2G18520.1 Y Y AT4G36680.1 Y Y Y AT2G39100.1 Y AT3G24515.1 (UBC37) YY AT3G50950.2 (ZAR1) Y Y Y Y AT3G61210.1 Y Y Y AT4G11670.1 Y AT4G24340.1 Y Y Y AT4G24350.1 Y Y Y AT5G01015.1 Y AT5G10830.1 Y Y Y Y [/formula]
Y indicates that evidence for LLG involvement in defense response was detected. PPI, protein-protein interaction.
repair and homologous recombination in somatic or meiotic cells [bib_ref] Inefficient double-strand DNA break repair is associated with increased fasciation in Arabidopsis..., Abe [/bib_ref] [bib_ref] BRCA2 is a mediator of RAD51-and DMC1-facilitated homologous recombination in Arabidopsis thaliana, Seeliger [/bib_ref]. MRE11 (AT5G54260) plays a role in the early stages of MRE [bib_ref] Meiotic defects in the Arabidopsis rad50 mutant point to conservation of the..., Bleuyard [/bib_ref] [bib_ref] Mre11 deficiency in Arabidopsis is associated with chromosomal instability in somatic cells..., Puizina [/bib_ref]. UVH1 (AT5G41150), also known as AtRAD1, is a homolog of the yeast repair endonuclease RAD1, and is involved in nucleotide excision repair and telomere stability [bib_ref] The Arabidopsis UVH1 gene is a homolog of the yeast repair endonuclease..., Fidantsef [/bib_ref] [bib_ref] ERCC1/XPF protects short telomeres from homologous recombination in Arabidopsis thaliana, Vannier [/bib_ref]. FAN1 (AT1G48360) is involved in DNA crosslink repair [bib_ref] The nuclease FAN1 is involved in DNA crosslink repair in Arabidopsis thaliana..., Herrmann [/bib_ref]. These observations indicate that these LLGs were presumably involved in DNA repair. Increased somatic recombination was observed in plants subjected to pathogen stress [bib_ref] Pathogen stress increases somatic recombination frequency in Arabidopsis, Lucht [/bib_ref]. Furthermore, the DNA damage repair proteins, BRCA2 and RAD51, are involved in the regulation of plant defense gene expression [bib_ref] A comparison of the effects of DNA-damaging agents and biotic elicitors on..., Choi [/bib_ref] [bib_ref] Arabidopsis SNI1 and RAD51D regulate both gene transcription and DNA recombination during..., Durrant [/bib_ref] [bib_ref] Arabidopsis BRCA2 and RAD51 proteins are specifically involved in defense gene transcription..., Wang [/bib_ref] [bib_ref] DNA repair proteins are directly involved in regulation of gene expression during..., Song [/bib_ref] , indicating that the LLGs associated with DNA repair might also be involved in plant responses to microbial pathogens. Furthermore, AT4G38600 (UPL3) and AT5G02880 (UPL4), which were observed in the largest LLGs' PPI network, exert a role in ubiquitination system [bib_ref] The HECT ubiquitin-protein ligase (UPL) family in Arabidopsis: UPL3 has a specific..., Downes [/bib_ref] , and ubiquitination is required in plant immunity for the degradation of invading proteins [bib_ref] Ubiquitination in plant immunity, Trujillo [/bib_ref]. Therefore, our multiple lines of evidence suggest that around 21 LLGs and orthologs are directly or indirectly involved in plant defense responses.
## The evolution of llgs in angiosperms
A gene that is continuously maintained in the genome indicates that it plays an essential role in viability [bib_ref] Gene loss, protein sequence divergence, gene dispensability, expression level, and interactivity are..., Krylov [/bib_ref]. In contrast, a gene without extensive and essential biological functions could be lost during evolution. Therefore, selection could be a significant driving force of gene loss [bib_ref] Selection-driven gene loss in bacteria, Koskiniemi [/bib_ref]. Because there are no wholegenome sequences available for the other two clades of Fabaceae, this study determined that LLGs specifically originated from Papilionoideae during legume evolution. On the other hand, nonlegume angiosperm species apparently had conserved these LLGs in its genome, which subsequently underwent negative selection. Single-copy genes often exhibit higher sequence conservation than nonsingle copy genes [bib_ref] Convergent gene loss following gene and genome duplications creates single-copy families in..., Smet [/bib_ref]. In line with this observation, 16 LLGs were single-copy families in Arabidopsis, whereas these maintained a low number of copies in most nonlegume species (supplementary table S3, Supplementary Material online). Therefore, LLGs might have originated from a direct gene deletion from the genome of a legume ancestor, instead of sequence divergence that mainly occurred when a gene is subjected to positive selection. The protein length and exon number of orphan genes, also called TGSs, are significantly different from those of nonorphan genes [bib_ref] An evolutionary analysis of orphan genes in Drosophila, Domazet-Loso [/bib_ref]. However, this study determined that the structural evolution of LLGs, lineage-specific lost genes, did not play a role in the emergence of these LLGs during plant evolution. Therefore, the major factors that drove the lineage-specific loss of these LLGs remain unclear. Nevertheless, most LLGs originated from ancestral legumes. Gene loss has been considered as a common and advantageous response during the genome evolution of living organisms [bib_ref] Gene losses during human origins, Wang [/bib_ref]. However, the role of LLGs in legume evolution requires more extensive investigations. Nonetheless, in the light of the putative role of LLGs in response to biotic stresses, we speculate that the loss of these genes plays a beneficial and adaptive role in the evolution of legumes.
## Evolutionary implication of llgs
Root nodule is specialized organ of legumes, and nodule formation is initiated through the molecular cross-talk between a bacterium and a plant, thus involving a complex and precise interplay between host and symbiont, and shifting the intracellular signaling from defense response to symbiosis [bib_ref] The Sinorhizobium meliloti MsbA2 protein is essential for the legume symbiosis, Beck [/bib_ref] [bib_ref] From defense to symbiosis: limited alterations in the kinase domain of LysM..., Nakagawa [/bib_ref]. Several symbiosis-related genes have been identified in legumes, and their mutants show various defects in the nodule formation [bib_ref] A plant regulator controlling development of symbiotic root nodules, Schauser [/bib_ref] [bib_ref] Four genes of Medicago truncatula controlling components of a nod factor transduction..., Catoira [/bib_ref] [bib_ref] LysM domain receptor kinases regulating rhizobial Nod factor-induced infection, Limpens [/bib_ref]. These symbiosis-related genes have orthologs in nonlegumes [bib_ref] Tracing nonlegume orthologs of legume genes required for nodulation and arbuscular mycorrhizal..., Zhu [/bib_ref] [bib_ref] Evolutionary genomics of LysM genes in land plants, Zhang [/bib_ref]. Therefore, the lineage-specific gain or loss of certain genes is likely required in the development of the legume nodulation pathway.
Legume-specific gene families have been identified [bib_ref] Novel paralogous gene families with potential function in legume nodules and seeds, Silverstein [/bib_ref] [bib_ref] Genome sequence of the palaeopolyploid soybean, Schmutz [/bib_ref] , and some of these show root-and/or nodule-specific expression [bib_ref] RNA-Seq Atlas of Glycine max: a guide to the soybean transcriptome, Severin [/bib_ref] , thus indicating the potential role of lineage-specific gene gain in the formation or maintenance of symbiosis. As compensation, we exploited the possible role of lineagespecific gene loss in the evolution of legumes. Eighteen LLGs in Arabidopsis were determined to participate in defense response (table 2) such as HOL1, HOL2, and ZAR1. In particular, four LLGs in a PPI network (AT3G24515, AT4G14970, AT5G49110, and AT5G65740) were apparently involved in plant-bacterial interactions. Moreover, some LLGs associated with defense responses were highly expressed in roots such as AT2G18520, AT4G36680, AT4G24340, AT3G24515 (UBC37), and AT5G10830 [fig_ref] Table 2: Summary of LLGs Involved in Plant Defense Response [/fig_ref]. LAS was also upregulated in roots but was apparently not associated with nodulation, whereas LAS was predicted to interact with carotenoid cleavage Dioxygenase 7 (supplementary table S11, Supplementary Material online), which is the ortholog in L. japonicus that controls determinate nodulation [bib_ref] CAROTENOID CLEAVAGE DIOXYGENASE 7 modulates plant growth, reproduction, senescence, and determinate nodulation..., Liu [/bib_ref] , thus suggesting that the root-expressed LAS might also be involved in nodulation. Therefore, LLGs could largely contribute to the improvement of compatibility between legume and rhizobia, thereby facilitating the establishment of reciprocal symbiosis.
In summary, through a genome-wide comparison, we identified a set of LLGs. The mechanisms and driving forces of LLG losses remain elusive; nonetheless, evolutionary loss of certain genes that are involved in plant immunity may provide new insights into elucidating the mechanisms underlying symbiotic nitrogen fixation. This work, for the first time, sheds light on the evolutionary implications of gene loss events in the evolution of Papilionoideae. Whether these findings can be generalized across the entire legume family requires further investigations. Engineering nitrogen-fixed genes in crops is essential for sustainable food production, and the results of this study thus also suggest that knocking out certain LLGs should also be considered in such kind of crop design.
[fig] FIG. 1: -Local synteny around the HOL genes in legumes and nonlegume species. Species names are provided on the left. The block arrow on the horizontal line represents one open reading frame of a gene and its orientation. The yellow blocks are HOL family genes. The same color block arrows connected with the same color lines indicate the putative orthologs in different species. The gray block arrows are nonhomologous genes. The nomenclature of each putative gene is shown above the corresponding block arrow. FIG. 2.-Selection analyses of LLGs in nonlegumes. Selection was evaluated in Arabidopsis thaliana, Prunus persica, Populus trichocarpa, and Vitis vinifera. Yellow lines and cyan lines, respectively, indicate the dN/dS value of the Group 1 LLGs and Group 2 LLGs. The dN/dS distribution of the 5,935 conserved genes in these four plants is presented in black columns. [/fig]
[fig] FIG. 3: -LLG evolution in sequenced angiosperms. Gray represents species without any homologous sequence to LLGs. Orange indicates species with LLG homologs but not orthologs. Blue represents species with putative LLG orthologs. The black and gray stars indicate the ancestors of Papillionoideae and Angiospermae, respectively, and their ancestral states are represented at the bottom (for details, see supplementary fig. S4, Supplementary Material online). [/fig]
[fig] FIG. 4: -Heat map of LLG expression under different stresses in Arabidopsis. Hormone treatments include ABA, ACC, and MeJA. Elicitors include CaCl 2 (Ca), glutathione S-transferase (GST), hrpz, GST-necrosis-inducing phytophtora Protein 1 (npp1), flagellin (flg), and lipopolysaccharide (lps). Bacterial stresses include Pseudomonas syringae pv. tomato DC3000 (PstDC3000), P. syringae pv. tomato avrRPM1 (Pstavrrpm1), P. syringae pv. tomato DC3000 hrcC (Psthrcc), P. syringae pv. phaseolicola (Pstpsph), Botrytis cinerea (B. cinerea), Erysiphe orontii (E. orontii), and Phytophthora infestans (P. infestans). The column above each treatment represents gene expression of different time points after treatment (for details, see supplementary fig. S10, Supplementary Material online).The color scale indicates the log2 values of expression change (treatments/control). Yellow indicates upregulation under treatments, and blue indicates downregulation under treatments. Green boxes indicate the log2 values of fold changes (treatments/control) > 1.0, and red boxes indicate the log2 values of fold changes (treatments/control) < À1.0. Gene names in red are Group 2 LLGs. AT2G43910/20 represents AT2G43910 (HOL1) and AT2G43920 (HOL2), and AT4G24340/50 represents AT4G24340 and AT4G24350 because one probe ID could detect the two closely related genes. two proteins were from the HECT ubiquitin-protein ligase (UPL) family of proteins such as AT4G38600 (UPL3) and AT5G02880 (UPL4), thus indicating that these LLG proteins is likely involved in ubiquitination (fig. 5and supplementary table S12, Supplementary Material online). [/fig]
[fig] FIG. 5: -The largest PPI network associated with LLGs. Each node represents a protein. The four nodes covered by a green circle indicate the four LLGs. The colorful connecting lines represent the types of evidence supporting each association: coexpression (dark brown), experiments (pink), databases (cyan), homology (violet), and text mining (light green). [/fig]
[table] Table 2: Summary of LLGs Involved in Plant Defense Response [/table]
|
Reoccurrence of Bleeding of a Chronic Subdural Haematoma Following a Fall
The case of a 60-year-old patient who presented with an acute-on-chronic subdural haematoma is reported. Chronic haematoma usually remains asymptomatic, and this is considered to be an unusual course of events. Trivial or minor injury may cause the cortical bridge veins and fragile vessels in the former haematoma to rupture with concomitant reoccurrence of bleeding. Old age, repeated traumatic brain injuries, brain atrophy, antiplatelet agents and oral anticoagulants such as warfarin are considered to be the underlying conditions to cause the reoccurrence of bleeding.However, our patient did not have any of those conditions.
# Introduction
Although uncommon, acute-on-chronic subdural haematoma (ACSDH) is not rare. [bib_ref] Natural history of chronic subdural haematoma, Lee [/bib_ref] found that 8% of chronic subdural haematomas (CSDH) were actually ACSDHs, [bib_ref] Natural history of chronic subdural haematoma, Lee [/bib_ref] [bib_ref] Acute-on-Chronic Subdural Hematoma: Not Uncommon Events, Lee [/bib_ref] although there are few cases reported in medical literature. Chronic haematomas develop by the pooling of blood caused by tears in the bridging vessels in the subdural space. Old age, repeated trauma and brain atrophy are the main risk factors. Unlike acute subdural haematomas, which may be life-threatening, chronic haemorrhages have a better prognosis. Some chronic subdural haematomas remain asymptomatic, but a mild injury or brain trauma may lead to the subdural reoccurrence of bleeding.
## Case report
A 60-year-old man was admitted to hospital following a mild traumatic head injury. He was lethargic and stuporous.
His wife reported that six months before, he had accidentally fallen and hit his head and had been admitted to hospital.
At that visit, he had no neurological symptoms, except for headache and nausea. No lower or upper limbs paresis were present. A brain CT was performed, and the definitive diagnosis was an acute left frontoparietal subdural haematoma. Given it was small in size and the lack of severe symptoms after 48 hours of observation, the supervising clinicians decided to manage the case with careful monitoring over time. In accordance with this policy, a second brain CT scan was carried out several weeks later, which did not show any changes compared to the previous CT scan [fig_ref] Figure 1: A [/fig_ref]. This, together with the fact that the patient had been asymptomatic since the initial admission to hospital, led to a diagnosis of chronic subdural haematoma.
At the time of his second admission to hospital, he was assessed in the Emergency Medicine Department. His relatives claimed that the trauma to his head had been mild, explaining that he had taken a nap and had rolled over and fallen off the couch. After a few minutes, he showed decreased alertness and was unable to follow simple commands or speak.
On physical examination, his pupils were normal, but right hemiparesis was present. The patient was not on any drug treatment, nor was he taking any vitamin K antagonists or antiplatelet agents.
The patient then underwent a brain CT scan, which showed acute reoccurrence of bleeding of the previous chronic subdural haematoma [fig_ref] Figure 1: A [/fig_ref]. The patient was referred to the Neurosurgery Department to undergo surgical drainage, after which he fully recovered.
# Discussion
Chronic haematomas are a mixture of liquefied blood and a semisolid clot pooling between the dura mater and the arachnoid mater [bib_ref] Natural history of chronic subdural haematoma, Lee [/bib_ref] [bib_ref] Acute-on-chronic subdural hematoma, Kloss [/bib_ref]. Although the pathophysiology of ACSDH is complex and incompletely understood, it appears that rotational forces accompanying movement of the skull exert tensile strain and rupture of bridging veins leading to acute haemorrhaging into the subdural space [bib_ref] Symptomatic Acute-on-Chronic Subdural Hematoma: A Clinicopathological Study, Castellani [/bib_ref]. When acute onset occurs, subdural haematomas can be life-threatening though some chronic haematomas remain asymptomatic and have a good prognosis. Brain atrophy after a traumatic brain injury is considered to be one of the underlying conditions that leads to the development of chronic haematomas. Elderly and alcoholic patients, with a history of repeated trauma or receiving treatment such as antiplatelet agents or warfarin, are more likely to develop this kind of haematoma. A chronic subdural haematoma leads to an excessive local activation of coagulation and fibrinolytic events, and a new onset haematoma would not necessarily form a solid clot [bib_ref] Acute-on-Chronic Subdural Hematoma: Not Uncommon Events, Lee [/bib_ref].
Nevertheless, acute-on-chronic haematomas are not very common, representing about 8% of chronic subdural haematomas [bib_ref] Acute on chronic subdural hematoma in a female boxer: a case report, Miele [/bib_ref] [bib_ref] Acute-on-chronic subdural hematoma by spinal anesthesia in a patient with undiagnosed chronic..., Park [/bib_ref]. Removal of the lesion may re-lieve the pressure on the brain and resolve the midline shift.
# Conclusions
Underlying conditions leading to a chronic subdural haematoma are old age, repeated trauma and brain atrophy. However, repetitive trauma can cause a reoccurrence of bleeding of a chronic haematoma even after a mild traumatic brain injury.
This case highlights the fact that acute bleeding over a chronic subdural haematoma may be identified by the onset of acute symptoms.
## Conflicts of interest
The author has no competing interests in the manuscript. . B: The current CT scan showed an acute crescent-shaped haematoma, 21 mm thick, causing mass-effect with raised intracranial pressure and midline shift to the right (12 mm). Chronic haematoma is hypodense, whereas acute (or acute-onchronic) haematoma is hyperdense.
[fig] Figure 1: A: Axial slide of a brain CT scan, showing a chronic subdural haematoma (white arrow), 13 mm thick and mild midline shift to the right [/fig]
|
A rare benign intrathoracic mass in a patient with history of rocket explosion
# Introduction
Thoracic splenosis is a rare condition that follows diaphragmatic injury leading to autotransplantation of splenic tissue into the pleural cavity. Trauma appears to the most common etiology with as many as sixty percent of patients endorsing a clear history of a traumatic event. It is mostly asymptomatic and incidentally diagnosed, which is why there is a delay in its diagnosis. Therapy is not indicated unless patient is symptomatic. Considering the wide differential of thoracic splenosis, majority of patients undergo extensive workups and invasive procedures which can be clearly prevented and complications avoided. We present this case of thoracic splenosis in an elderly male with a past history of traumatic event. Through this case we want to make the physicians and pulmonologists cognizant of this condition preventing unnecessary workup and patient morbidity.
## Case presentation
A sixty year old white male with a past medical history of Type 2 Diabetes Mellitus with neuropathy, hypertension, and kidney stones, who presented with nasal congestion and cough productive of white sputum. Patient denied any shortness of breath, recent weight loss, night sweats, or increasing fatigue. He had a sinus infection for more than 1 week, and was previously treated with 10 days of moxifloxacin. A chest X-ray done was concerning for a lung nodule. Subsequently, a Computer Tomography of the chest was done which showed multiple pleural based noncalcified nodular densities along the base of the left hemithorax [fig_ref] Figure 1: Multiple pleural based noncalcified nodular densities along the base of the left... [/fig_ref]. Pulmonology was consulted for further workup of lung nodule. After the scars on his chest and abdomen were seen on exam, further inquiry revealed a history of remote injury involving a rocket explosion with shrapnel causing severe throcoabdominal injuries. He had to have a splenectomy and rib cage repair in a MASH (Mobile Army Surgical Hospital) unit. A colloid liver spleen scan [fig_ref] Figure 2: Explanted splenic tissue in the left hemithorax [/fig_ref] was performed which confirmed the presence of explanted splenic tissue in the left hemithorax.
# Discussion
Thoracic splenosis is a rare, benign condition which involves autotransplantation of splenic tissue into the pleural cavity secondary to trauma or surgery. It is parenthetically detected, asymptomatic, and treatment is not often indicated. The first case of thoracic splenosis was reported in 1937 by Shaw and Shafi in a 20-year old Egyptian man, and ever since, less than 50 new cases have been reported in the literature [bib_ref] Traumatic autoplastic transplantation of splenic tissue in man with observations on the..., Shaw [/bib_ref]. It involves 16%e67% of patients with past splenic trauma and or past splenectomy [bib_ref] Thoracic splenosis after blunt trauma: frequency and imaging findings, Normand [/bib_ref]. Pathogenesis of thoracic splenosis is depicted in [fig_ref] Figure 3: Pathogenesis of thoracic splenosis [/fig_ref] [bib_ref] Thoracic splenosis: a case report and the importance of clinical history, Kim [/bib_ref].
Autotransplanted spleens have no hilum and the arterial supply can pass through any site in the capsule; however, accessory spleens have hilum where the arteries enter [bib_ref] Splenosis mimicking neoplasm in the perirenal space: CT characteristics, Darling [/bib_ref]. Splenosis is microscopically identical to normal spleen with both having thick capsule, trabeculae, and white and red pulp [bib_ref] Splenosis mimicking neoplasm in the perirenal space: CT characteristics, Darling [/bib_ref] [bib_ref] Splenosis: autotransplantation of splenic tissue, Fleming [/bib_ref]. Although it is usually asymptomatic and diagnosed incidentally; it can occasionally present as hemoptysis and pleuritic chest pain [bib_ref] Intrathoracic splenosis presenting as persistent chest pain, Shinichi [/bib_ref]. Diagnosis can be challenging without knowledge of preceding splenic injury, often leading to the use of biopsy, video-assisted thoracoscopic surgery (VATS) and even thoracotomy for diagnosis, causing significant morbidity and mortality among patient population [bib_ref] Thoracic splenosis diagnosed by fine needle aspiration cytology:a case report, Syed [/bib_ref] [bib_ref] Thoracic splenosis: know iteavoid unnecessary investigations, interventions, and thoracotomy, Khan [/bib_ref].
There is a wide list of differentials for thoracic splenosis which include low grade lymphoma, thymoma, primary lung carcinoma, mesothelioma, thoracic endometriosis, mediastinal tumor, neurogenic tumors and metastatic lesions. It may present as soliatary (25% cases) or multiple nodules (75% of cases) on CT scans [bib_ref] Thoracic splenosis: know iteavoid unnecessary investigations, interventions, and thoracotomy, Khan [/bib_ref]. Scintigraphy performed with heat-damaged 99Tc-labelled red blood cells is considered the most sensitive and specific imaging modality for the diagnosis of splenosis [9e11] and can demonstrate splenic tissue even when minimally present. This is because splenic tissue takes up more than 90% of damaged red blood cells [bib_ref] Splenosis: a review, Fremont [/bib_ref] [bib_ref] Scintigraphic assessment of ectopic splenic tissue localization and function following splenectomy for..., Zwas [/bib_ref].
Removal of thoracic splenic tissue is inadvisable especially in patients without functional abdominal splenic tissue may render the patient a splenic, increasing the risk of infection, although this notion is still debatable [bib_ref] The born-again spleen: return of splenic function after splenectomy for trauma, Pearson [/bib_ref]. Surgical removal is considered in symptomatic patients and patients with hematological disease [bib_ref] Thoracic splenosis: a case report and the importance of clinical history, Kim [/bib_ref] [bib_ref] Thoracic splenosis: know iteavoid unnecessary investigations, interventions, and thoracotomy, Khan [/bib_ref].
In conclusion, if a patient has an appropriate history of splenic injury and multiple, asymptomatic, left-side pleural lesions, intrathoracic splenosis should be considered in the differential diagnosis.
[fig] Figure 1: Multiple pleural based noncalcified nodular densities along the base of the left hemithorax. [/fig]
[fig] Figure 2: Explanted splenic tissue in the left hemithorax. [/fig]
[fig] Figure 3: Pathogenesis of thoracic splenosis. [/fig]
|
Intraoperative ventilation strategies for obese patients undergoing bariatric surgery: systematic review and meta-analysis
Background: Obesity is a global epidemic, and it is widely known that increased Body mass index (BMI) is associated with alterations in respiratory mechanics. Bariatric surgery is established as an effective treatment for this condition. Objective: To assess the safety and effectiveness of different ventilation strategies in obese patients undergoing bariatric surgery. Methods: A systematic review of randomized clinical trials aimed at evaluating ventilation strategies for obese patients was carried out. Primary outcomes: in-hospital mortality, adequacy of gas exchange, and respiration mechanics alterations. Results: Fourteen clinical trials with 574 participants were included. When recruitment maneuvers (RM) vs Positive endexpiratory pressure (PEEP) were compared, RM resulted in better oxygenation p = 0.03 (MD 79.93), higher plateau pressure p < 0.00001 (MD 7.30), higher mean airway pressure p < 0.00001 (MD 6.61), and higher compliance p < 0.00001 (MD 21.00); when comparing RM + Zero end-expiratory pressure (ZEEP) vs RM + PEEP 5 or 10 cmH2O, RM associated with PEEP led to better oxygenation p = 0.001 (MD 167.00); when comparing Continuous Positive Airway Pressure (CPAP) 40 cmH2O + PEEP 10 cmH2O vs CPAP 40 cmH2O + PEEP 15 cmH2O, CPAP 40 + PEEP 15 achieved better gas exchange p = 0.003 (MD 36.00) and compliance p = 0.0003 (MD 3.00). Conclusion: There is some evidence that the alveolar recruitment maneuvers associated with PEEP lead to better oxygenation and higher compliance. There is no evidence of differences between pressure control ventilation (PCV) and Volume control ventilation (VCV).
# Background
Obesity is a global epidemic that causes major economic, social and psychological impacts. Body mass index (BMI) values above 30 Kg/mcan result in a reduction in life expectancy similar to that caused by smoking. Bariatric surgery is an effective intervention against weight gain and the majority of people who undergo such surgery show an improvement in, or the resolution of, conditions such as diabetes, dyslipidemia, hypertension and obstructive sleep apnaea.
The growing number of bariatric surgeries highlights the importance of invasive ventilator support. Anesthetic induction in obese patients can result in a significant reduction in respiratory compliance and increase resistance and pressure in the airway. A correlation has also been found between a high BMI and an increase in breathing effort and a reduction in oxygenation levels, which may lead to atelectasis and slower weaning from mechanical ventilation.
To date, no standard ventilation strategy has been established for obese patients, although there is some evidence that recruitment maneuvers (RM) combined with Positive End-Expiratory Pressure (PEEP) improves oxygenation and compliance in comparison with other strategies. A systematic review can therefore make a significant contribution to the decision-making process of healthcare professionals, particularly surgeons and anesthesiologists, when choosing the best ventilation strategy during the surgery and anesthesia of obese patients, with the aim of reducing complications, costs and mortality.
## Objectives
To assess the effectiveness and safety of different ventilation strategies for obese participants undergoing bariatric surgery under general anesthesia.
# Methods
The methodology described by the Cochrane Collaboration was employed in this systematic review. This research was approved by the ethics committee of the federal university of São Paulo -Unifesp -CAAE: 57099216.0.0000.5505.
## Criteria for considering studies for this review
Randomized controlled trials (RCTs) that evaluated different ventilation strategies for obese patients undergoing bariatric surgery, under general anesthesia, regardless of age and gender, were included.
Obesity was defined as BMI greater than 30 Kg/m 2. Primary outcomes: in hospital mortality, adequacy of intra-operative gas exchange, pulmonary mechanics (plateau pressure, mean airway pressures, lung compliance and lung resistance) alteration.
Secondary outcomes: Intraoperative and postoperative respiratory complications such barotrauma, hemodynamic instability, pneumonia, atelectasis, reintubation, self-extubation and the need for noninvasive mechanical ventilation measured in hours or days; cardiovascular responses; need for hospitalization in the intensive care unit (ICU) and length of stay (LOS) in the post-anesthesia care unit (PACU).
## Search methods for identification of studies
Searches (see attachment) were performed in the Cochrane Central Register of Controlled Trials; MEDLINE via Ovid (1966 to present); old MEDLINE (1951 to present); and EMBASE via Ovid (January 1990 to present), without language or location restrictions. The highly sensitive Cochrane filter for randomized controlled trials was applied to the MEDLINE and EMBASE searches. Trial registers such as www.clinicaltrials.gov and the Current Controlled Clinical Trials Website (http://www.controlled-trials.com/) were also searched for ongoing trials.
## Data collection and analysis
Two authors (GMCS and SAZ) independently screened all the potential studies identified and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. The full-text reports/publications were then retrieved and two authors independently screened the full text and identified the studies for inclusion. Disagreements were resolved through discussion or if required consultation with a third author. Duplicates were excluded and multiple reports of the same study were collated so that each study, rather than report, is the unit of interest in the review. The selection process was recorded in appropriate detail, as set out in the complete PRISMA flow diagram.
The authors were contacted and additional details were requested. Disagreements were resolved by consensus or by involving a third author.
## Assessment of risk of bias in the included studies
Risk of bias was assessed at study level using Cochrane's 'Risk of Bias' tool. Two review authors (GMCS and SAZ) independently assessed the methodologic quality of each study included and resolved their disagreements by discussion.
## Data synthesis
To consider the measures of treatment effect for dichotomous outcomes, the total number of events within each randomized group were entered and the risk ratios with 95% confidence intervals (CI) were calculated. For data presented in other forms, such as odds or hazard ratios, the generic variance option was used, although different effect measures (odds, risk or hazard ratios) were not combined in the same model. Mean differences were calculated for continuous outcomes measured on the same scale in different studies.
## Assessment of heterogeneity
Statistical heterogeneity was evaluated by assessing forest plots and examining the I 2 value, which describes the proportion of total variation across studies caused by heterogeneity rather than chance. An I 2 value greater than 50% was considered as the cut-off point to identify the presence of considerable heterogeneity.
# Results
The initial search identified 1018 citations through database searches and manual searches. After screening by title and abstract, full-text articles of 40 studies that were potentially eligible for inclusion in the review were obtained. A total of 25 of these were excluded due to not being randomized, presenting data in graphs, did not present data for extraction or did not respond to the PICO of this review. Following this process, fourteen studies were included in the review.
## Risk of bias in included studies
Random sequence generation and allocation concealment were correctly described in seven studies. In the blinding of participants and personnel domain all the studies were classified as high risk as the personnel could not be blinded. Four studiesadequately described the blinding of outcome assessment. Eleven studies did not describe losses or exclusions which could cause imbalance between the groups. Only two studiesemployed selective reporting and while two studiespresented other sources of bias .
## Effects of interventions alveolar recruitment maneuvers versus peep only
Three studiescompared alveolar recruitment maneuvers (RM) versus PEEP to evaluate intra operative gas exchange, with the mean PaO 2 /FiO 2 ratio found to be greater in the groups that underwent RM, p = 0.03, (MD 79.93, 95% CI 8.83 to 151.04; participants = 121; studies = 5; I 2 = 80%,). shows the comparison of three different studies, separate in four subgroups, where the best results were in favor of RM .
Three studiesevaluated mean airway pressures by comparing RM with progressive PEEP of 10, 15 and 20 cmH 2 O versus Peep of 4 or 5 cmH 2 O only and found that the use of PEEP without RM led to lower airway pressure, p < 0.001 (MD 9.29, 95% CI 5.05 to 13.53; participants = 98; studies = 4; I 2 = 89%). shows the comparison of three different studies, separate in two subgroups, where the best results were in favor of PEEP when airway pressure was measured. .
Two studiesevaluated compliance by comparing RM with PEEP. The study by Reinius et al.compared
## Pressure control ventilation versus volume control ventilation
Of the two included studies that evaluated this comparison, only the study by Cadi et al.evaluated the PaO 2 /FiO 2 ratio, finding that the PCV mode achieved greater oxygenation than the VCV mode, p = 0.007, (MD 82.00, 95% CI 21.90 to 142.10; participants = 36; studies = 1; I2 = 0%). The study by De Baerdemaeker et al.did not identify differences in the variables analyzed. No differences were found between the VCV and PCV modes in the evaluation of mean airway pressure, plateau pressure, lung compliance, lung resistance and arterial pressure. When the two comparisons were pooled the shortest LOS in the PACU was found in the group that received RM plus PEEP, p = 0.01, (MD -15.89, 95% CI − 28.68 to − 3.10; participants = 77; studies = 2; I 2 = 0%).
The study by Talab et al.compared RM with CPAP 40 cmH 2 O plus ZEEP versus RM with CPAP 40 cmH 2 O plus PEEP 10 cmH 2 O and found fewer patients with lamellar atelectasis in the group that received RM plus ZEEP, p = 0.007, (RR 5.22, 95% CI 1.33 to 20.55; participants = 39; The study by evaluated the I:E 1:1 ratio versus the I:E 1:2 ratio and found that the I:E 1: 1 ratio group achieved greater lung compliance than the group with a I:E 1:2 ratio, p = 0.01, (MD 4.67, 95% CI 1.06 to 8.28; participants = 30; studies = 1; I 2 = 0%, low-quality evidence).
# Discussion
The present systematic review evaluated different ventilatory strategies for obese patients undergoing bariatric surgery, such as: comparison between PCV and VCV; comparison of different forms of RM, different PEEP levels and comparison between I:E 1:1 ratio and I:E 2:2. Fourteen studies with a total of 574 participants were included.
Significant variability in interventions were found. This demonstrates the lack of consensus on how to ventilate obese patients undergoing surgery, corroborating a review published by Aldenkortt et al.The main finding of the present study is the evidence that obese patients receiving mechanical ventilation benefit from RM, especially when combined with PEEP, as evidenced by improvements in oxygenation and respiratory compliance. While it was observed in this systematic review that the isolated use of PEEP was more effective when higher values were used, however the best result was the combination of the RM with higher levels of PEEP. In addition to these findings, no difference was found between VCV and PCV modes of ventilation in all analyzed outcomes, corroborating another study by Aldenkortt et al.. No respiratory complications or major adverse events were reported in the studies included in this review. Such findings are similar to those found by Aldenkortt et al.. and Hu et al.. Recent guidelines regarding mechanical ventilation of patients with acute respiratory distress syndrome (ARDS) have shown that the incidence of complications associated with diferent mechanical ventilation strategies is low.
There is insufficient evidence to support differences between VCV and PCV in the evaluated outcomes. While the study by Cadi et al.. Showed that the pressure controlled mode led to a higher PaO 2 /FiO 2 ratio than the volume controlled mode, the study by De Baerdemaeker et al.. Did not identify a difference in PaO 2 between the two modes.
Three studiesincluded in this review describe the performance of more than one alveolar recruitment maneuver. There is no consensus on the ideal number of alveolar recruitment maneuvers regarding frequency and repetitions, however, the use of various maneuvers with patients with ARDS is associated with decreased pulmonary shunt and increased compliance.
Despite the wide variety of interventions and outcomes evaluated, the present review provides some evidence that the use of PEEP effectively improves oxygenation and compliance of the respiratory system. Better results seem to be achieved, however, when it is combined with alveolar recruitment maneuvers, and the absence of adverse effects shows that it is an effective and safe strategy for obese patients undergoing bariatric surgery. Briel et al.published a systematic review and metaanalysis comparing the use of high versus low PEEP values for ARDS patients, and concluded that the use of high levels of PEEP was associated with lower hospital mortality in this group of patients. The American Thoracic Society also currently recommends the use of high levels of PEEP for patients with ARDS.
One study compared I:E 1:1 ratio with I:E 1:2 ratio and found that only the 1:1 ratio only improved lung compliance. Few studies evaluated the use of the I:E 1:1 ratio, while some studies evaluated different inverted ratios in patients with ARDS, with conflicting results regarding its effectiveness.
Limitations Important methodological limitations of the studies included reduced the certainty of the evidence offered by most of the included trials. Many of the trials were small and included different outcome measures, and selective outcome reporting was occasionally an issue.
The paucity of long-term follow-up data, the small sample sizes, and the heterogeneous nature of the measured outcomes limit the generalizability of the results.
# Conclusions
There is evidence that alveolar recruitment maneuvers plus PEEP improve gas exchange with an increase in respiratory system compliance. The quality of such evidence is low, however.
There is no evidence to support that there is a difference between the volume and pressure controlled modes.
The various interventions assessed were shown to be safe with no major adverse events reported. |
Dose Intraoperative Fluoroscopy Precisely Predict Catheter Tip Location via Superior Vena Cava Route?
Adequate catheter tip location is crucial for functional intravenous port and central venous catheter. Numerous complications were reported because of catheter migration that caused by inadequate tip location. Different guidelines recommend different ideal locations without consensus. Another debate is actual movement of intravascular portion of implanted catheter. From literature review, the catheter migrated peripherally an average of 20 mm on the erect chest radiographs. In this study, we want to verify the actual presentation of catheter movement within a vessel and try to find a quantitative catheter length model to recommend.From March 2012 to March 2013, 346 patients were included into this prospective cohort study. We collect clinical data from medical record and utilized picture archiving and communication system to measure all image parameters. Statistical analysis was utilized to identify the risk factors for catheter migration.The nonmigration group had 221 patients (63.9%); 67 (19.4%) patients were classified into the peripheral migration group; and 58 (16.8%) patients were classified into the central migration group. Patients with short height (P ¼ 0.03), larger superior vena cava (SVC) diameters at the brachiocephalic vein confluence site (P ¼ 0.02), and longer implanted catheter length (P ¼ 0.0004) had greater risks for central migration. We utilized regression curve for further analysis and height (centimeters)/10 had moderate correlation distances from the entry vessel to the carina.Although intravascular movement of catheter was exist in implanted catheter, the intraoperative fluoroscopy could provide accurate catheter tip location in 63.9% patients. Additional length of catheter implantation seems unnecessary in 80.6% patients. Patients with short height, larger SVC diameters at the brachiocephalic vein confluence site had greater risk for catheter central movement. Height/10 may be consider as reference length of implantation for inexperience surgeon and precise implantation length could be adjust under guidance of fluoroscopy. (Medicine 94(49):e2199) Abbreviations: CT = computed tomography, IVC = inferior vena cava, RA = right atrium, SD = standard deviation, SVC = superior vena cava.
Abstract: Adequate catheter tip location is crucial for functional intravenous port and central venous catheter. Numerous complications were reported because of catheter migration that caused by inadequate tip location. Different guidelines recommend different ideal locations without consensus. Another debate is actual movement of intravascular portion of implanted catheter. From literature review, the catheter migrated peripherally an average of 20 mm on the erect chest radiographs. In this study, we want to verify the actual presentation of catheter movement within a vessel and try to find a quantitative catheter length model to recommend.
From March 2012 to March 2013, 346 patients were included into this prospective cohort study. We collect clinical data from medical record and utilized picture archiving and communication system to measure all image parameters. Statistical analysis was utilized to identify the risk factors for catheter migration.
The nonmigration group had 221 patients (63.9%); 67 (19.4%) patients were classified into the peripheral migration group; and 58 (16.8%) patients were classified into the central migration group. Patients with short height (P ¼ 0.03), larger superior vena cava (SVC) diameters at the brachiocephalic vein confluence site (P ¼ 0.02), and longer implanted catheter length (P ¼ 0.0004) had greater risks for central migration. We utilized regression curve for further analysis and height (centimeters)/10 had moderate correlation distances from the entry vessel to the carina.
Although intravascular movement of catheter was exist in implanted catheter, the intraoperative fluoroscopy could provide accurate catheter tip location in 63.9% patients. Additional length of catheter implantation seems unnecessary in 80.6% patients. Patients with short height, larger SVC diameters at the brachiocephalic vein confluence site had greater risk for catheter central movement. Height/10 may be consider as reference length of implantation for inexperience surgeon and precise implantation length could be adjust under guidance of fluoroscopy.
# Background
A dequate catheter tip location is crucial for functional intravenous port and central venous catheter. From literature review, catheter with shallow tip location were associated to catheter migration. [bib_ref] Risk factors and possible mechanisms of intravenous port catheter migration, Wu [/bib_ref] Numerous complications were reported because of catheter migration that caused by inadequate tip location. Too deep catheter tip may lead arrhythmias, 2,3 heart injury, 4 and vessel perforation. [bib_ref] Central venous access catheters: radiological management of complications, Teichgräber [/bib_ref] Too shallow tip location may cause migrated catheter to nearby vein and cause various clinical symptoms, including back pain, 5 brachial plexopathy, [bib_ref] Vertebral vein migration of a long-term central venous access catheter: a cause..., Winston [/bib_ref] neurologic deficit, [bib_ref] Vertebral vein migration of a long-term central venous access catheter: a cause..., Winston [/bib_ref] and cortical vein thrombosis. [bib_ref] Unsuspected cerebral perfusion. A complication of the use of a central venous..., Klein [/bib_ref] [bib_ref] Spreading cortical venous thrombosis due to infusion of hyperosmolar solution into the..., Souter [/bib_ref] In order to minimize the migration risk, adequate tip location is important for clinical practice. From literature review, the consensus of tip location is located between distal third of superior vena cava (SVC) and junction site to right atrium (RA). [bib_ref] Relationship between chest port catheter tip position and port malfunction after interventional..., Schutz [/bib_ref] These structure could not be visualized in chest plain film and several landmarks were recommended as a reference, including T5 to T6 intervertebral space, right tracheobronchial angle, and carina. [bib_ref] Relationship between chest port catheter tip position and port malfunction after interventional..., Schutz [/bib_ref] [bib_ref] The carina as a radiological landmark for central venous catheter tip position, Stonelake [/bib_ref] [bib_ref] Cardiac tamponade from central venous catheters, Defalque [/bib_ref] [bib_ref] Depth of central venous catheterization: an audit of practice in a cardiac..., Rutherford [/bib_ref] However, the interpretation variation still existed because of the difference of clinical experience. In our previous study, we identified the location that 0.68 cm below carina had minimal risk of catheter migration. [bib_ref] Analysis of chest X-ray plain film images of intravenous ports inserted via..., Fu [/bib_ref] However, all these studies based on the measurement of postoperation chest plain film instead of actual intraoperative fluoroscopy. In addition, the catheter tip may travels during image acquiring 9 and posture changes. [bib_ref] Relationship between chest port catheter tip position and port malfunction after interventional..., Schutz [/bib_ref] [bib_ref] Changes in tunneled catheter tip position when a patient is upright, Nazarian [/bib_ref] [bib_ref] How much do arm movements displace cubital central venous catheters?, Kalso [/bib_ref] The relationship between intraoperative fluoroscopy and postprocedure chest plain film remain unknown. The aim of this study is tried to identify the difference these 2 image tools and quantify a recommended length of implanted catheter via SVC route in order to minimize the technical errors.
# Materials and methods
## Patients
From March 2012 to March 2013, 475 oncology patients received intravenous port implantation for chemotherapy preparations. All these patients were recruited into this study. Patients who could stand received standing chest posterior-anterior plain film after intravenous port implantation. Not only does this confirm catheter tip location and measure the distance to the carina, but also makes sure that the nut-catheter angle is obtuse. Chest or whole body computed tomography (CT) was arranged by medical oncologist as a comparison standard for all patients because they were planned to received palliative therapy due to advance disease. Exclusion criteria were as follows: intravenous port implantation via internal jugular vein (28 patients) and inferior vena cava (IVC) route (3 patients; femoral or greater saphenous vein), without chest CT (82 patients) and standing chest posterior-anterior plain film (16 patients). After which 346 patients were included and further analyze in this study. All these selected patients were free from operation related complication, such as migration, facture, and malfunction. The study was approved by Chang Gung Medical Foundation Institutional Review Board and the IRB number is 100-4193A3.
## Principle for choosing an entry vessel and implantation method
According to our own study, we standardized the entry vessel principle and the entry vessel that drained to SVC was before those drained to IVC. [bib_ref] Single-center study of vascular access sites for intravenous ports, Wu [/bib_ref] [bib_ref] The treatment results of a standard algorithm for choosing the best entry..., Wei [/bib_ref] The cephalic vein was the first choice for exploration due to having the most superficial position. If there was no cephalic vein or it could not be accessed by a metallic guide-wire, the thoracoacromial vein would be used for catheter implantation. [bib_ref] Deltoid branch of thoracoacromial vein: a safe alternative entry vessel for intravenous..., Su [/bib_ref] If this failed, the ipsilateral internal jugular vein may be the next choice. No more subclavian punctures are made in order to avoid an iatrogenic pneumohemothorax. For patients who were identified cephalic or thoracoacromial vein with adequate size, vessel cutdown method was tried first. Target vessel underwent distal ligation and catheter was implanted via the venostomy under fluoroscopy. Additional stay suture would be placed at proximal end of venostomy site order to avoid bleeding. For patients who were identified cephalic or thoracoacromial vein with adequate size but met resistance in manual implantation, a metallic wire (V-18 Control Wire, 0.018 in., 200 cm, Boston Scientific, Natick, MA, USA or Guide Wire M 0.035 in., 150 cm, Terumo Cooperation, Tokyo, Japan), was utilized prior catheter implantation in order to establish entry route. Catheter was implanted by sliding over the wire under fluoroscopy. For patients who were identified cephalic or thoracoacromial vein with inadequate size, a metallic wire was utilized to establish entry route and peel-apart sheath were used sliding over the wire to create subcutaneous tunnel that permit catheter implantation. For those without accessible cephalic and thoracoacromial vein, echo-guide puncture for internal jugular vein access at thyroid cartilage level (ie, high neck puncture). After a catheter was implanted, an additional subcutaneous tunnel would be created between entry site and pocket site in order to allow catheter embedding in subcutaneous tissue.
## Implanted device, surveillance, and image parameter collection
Only the B'Braun Fr 6.5 (B'Braun Medical, Chasseneuil, France) single lumen intravenous port was utilized in this study. We utilized carina as the reference of landmark because of its relative immobile position and easily visualization under fluoroscopy. All catheter was implanted to location that 1 cm below carina, which was relative immobile and easily be seen intraoperation fluoroscopy. Actual implanted catheter length and the intraoperative tip location that revealed from intraoperative fluoroscopy were recorded. All patients had to receive chest plain film to confirm tip location. Medical records were reviewed and picture archiving and communication system (General Electric Co., Fairfield, CT, USA) was used to measure all clinical parameters.
## Definition of measurement and subgrouping
We measure the nut-catheter angle, catheter tip distance to the carina from chest plain that shown in. In addition, we measured short axis of the SVC's diameter at the brachiocephalic vein confluence site and the azygos vein confluence site and calculate the vertical distance between the lower margin of the SVC's confluence site with the brachiocephalic vein and the upper margin of the SVC's confluence site with the azygos vein in order to describe the surrounding configuration around catheter tip. All parameters except nut-catheter angle were recorded in centimeters. Nut-catheter angles were recorded in degrees. The nonmovement group was defined as difference of tip location between intraoperation fluoroscopy and postoperation chest plain film was less than 1 cm. The central movement group was identified the tip location caudal migration greater than 1 cm in postoperation chest plain film compared with intraoperation fluoroscopy. The peripheral movement group was identified the tip location cephalad migration greater than 1 cmin postoperation chest plain film compared with intraoperation fluoroscopy.
## Statistics
Categorical variables were compared using x 2 , Fisher exact tests, or ANOVA. The t test was used to compare continuous variables with normal distribution. A P-value of less than 0.05 was considered to indicate statistical significance, and all tests were 2-tailed. Descriptive statistics and logistic regression were performed using SAS statistical packages, version 9 (SAS Institute, Cary, NC).
# Results
There were 346 patients were included and further analyze in this study. Two hundred thirty-one patients (66.7%) were male. The mean age of these patients was 59.6 years. The mean body weight and height were 61.3 kg and 161.4 cm, respectively. Mean body mass index (BMI) of these patients was 23.4. Majority (92.4%) patients received catheter from right side entry vessel. All clinical characteristics of these patients are shown in [fig_ref] TABLE 1: Clinical Characteristics of All Patients BMI ¼ body mass index, CT ¼... [/fig_ref]. We further divided these patients into 3 groups. Two hundred twenty-one patients (63.9%) was identified as nonmovement group because of the difference between intraoperative fluoroscopy and chest plain film were less than 1 cm. In addition, 67 (19.4%) patients were classified into the peripheral movement group and 58 (16.8%) patients were classified into the central movement group [fig_ref] TABLE 2: Clinical Characteristics of Nonmovement, Central Movement, and Peripheral Movement Groups [/fig_ref]. All parameters of the 3 groups were similar and had no statistical significance except for implanted catheter length (P ¼ 0.01).
We further analyzed the difference between the nonmovement and central movement group and found that all parameters were similar except for implanted catheter length (P ¼ 0.02; [fig_ref] TABLE 3: Clinical Characteristics of Nonmovement and Central Movement GroupsVariables [/fig_ref]. In addition, we also found that the SVC's diameter at the brachiocephalic vein confluence site may relate to central movement (P ¼ 0.06; [fig_ref] TABLE 3: Clinical Characteristics of Nonmovement and Central Movement GroupsVariables [/fig_ref]. We further used multivariate analysis [fig_ref] TABLE 4: Multivariate Analysis of Risk Factors of Central Movement [/fig_ref] and identified as body height had a negative correlation with central movement (P ¼ 0.03) and implanted catheter length (P ¼ 0.0004) and the SVC's diameter at the brachiocephalic vein confluence site (P ¼ 0.02) positively correlated with central movement. However, there was no definite risk factor for peripheral movement.
Because there is no quantified model for recommendation of catheter implantation, we further utilized regression curve to further analysis and identified the relationship between the distance from the entry vessel to the carina. Height (in centimeters) divided by 10 was identified moderate correlation between the distance from the entry vessel to the carina for all patients , and this correction was also revealed in the nonmigration group. , correlation coefficient: 0.482).
# Discussion
Inadequate tip location may lead complications after intravenous port implantation. For those with deep tip location, arrhythmias, and myocardium injury were reported. [bib_ref] Central venous access catheters: radiological management of complications, Teichgräber [/bib_ref] [bib_ref] Sinus arrest and asystole caused by a peripherally inserted central catheter, Nazinitsky [/bib_ref] [bib_ref] Cardiac tamponade from misplaced central venous line in pericardiophrenic vein, Van Haeften [/bib_ref] For those with shallow tip location, complications such as catheter-related thrombosis, catheter migration, brachial plexopathy, neurologic deficit, and cortical vein thrombosis were reported because of slow blood flow velocity could not dilute the infused mecication. 1,5-9 Adequate tip location could minimize the risk of catheter-related complication but the adequate tip location were remain debates because of the intravascular portion of catheter were free mobile within vessel. There were no consensus in ideal catheter tip location between different guidelines Some favor the right atrial-SVC junction and others prefer a tip position within the RA as the optimal position. [bib_ref] Central venous access catheters: radiological management of complications, Teichgräber [/bib_ref] [bib_ref] Relationship between chest port catheter tip position and port malfunction after interventional..., Schutz [/bib_ref] [bib_ref] Quality improvement guidelines for central venous access, Lewis [/bib_ref] [bib_ref] Central venous access: a primer for the diagnostic radiologist, Funaki [/bib_ref] However, these locations were invisible in image tools and several landmarks are recommended as references to help localize ideal tip location, including T5 to T6 intervertebral space, [bib_ref] Cardiac tamponade from central venous catheters, Defalque [/bib_ref] right mainbronchuscriterion, [bib_ref] Depth of central venous catheterization: an audit of practice in a cardiac..., Rutherford [/bib_ref] and carina. [bib_ref] The carina as a radiological landmark for central venous catheter tip position, Stonelake [/bib_ref] In this study, carina was the chosen landmark not only its immobility and visibility but also minimal migration risk that conformed in our previous study. [bib_ref] Analysis of chest X-ray plain film images of intravenous ports inserted via..., Fu [/bib_ref] In this study, 221 patients (63.9%) was identified catheter tip movement less than 1 cm between intraoperation fluoroscopy and postoperation chest plain film. This is totally different compared with previous literature. This is because gravity not only pull down the abdominal organs, but also pull intravascular portion of intravenous port downward. In addition, we fixed the port over fascia of pectoralis major muscle that minimal the catheter movement. Furthermore, additional 2 to 3 cm of catheter length implantation is not necessary in clinical practice and this may cause too deep catheter tip location. Even securing the implanted length of the catheter as best as we could, still 36.1% of patients were identified to have catheter movement within the vessel greater than 1 cm. For peripheral movement group, we did not identify any intrinsic risk factor through either univariate or multivariate analysis. This may be caused by patients' personnel intrinsic factors, including confluence angle and patterns of SVC, brachiocephalic veins, and azygos veins, and further investigation is warranted. For central movement group, we identified that shorter body height (P ¼ 0.03), longer implanted catheter length (P ¼ 0.0004), and larger SVC diameter at the brachiocephalic vein confluence site (P ¼ 0.02) were risk factors for catheter central movement. The implanted catheter length was a more predominant risk factor for central migration and may be caused by variations in fluoroscopy interpretation among surgeons. For patients with shorter body height, a shorter length for the vascular route that lead patients at risk for catheter central movement. For patients with larger SVC diameters, a larger intravascular space which What's the reliable quantified implanted catheter length from entry site to carina? From view of surface anatomy, the implanted catheter had to cross the width of body and then pass downward to carina via SVC. From view of real anatomy, the catheter was implanted via entry vessel, such as cephalic vein, thoracoacromial vein, or internal jugular vein, to subclavian vein and low third of SVC. The former could be calculated by measurement, the later would have more variation that the vessel configuration was 3 dimension that would not calculated. From literature review, Peres [bib_ref] Positioning central venous catheters: a prospective survey, Peres [/bib_ref] and Czepizak et al [bib_ref] Evaluation of formulas for optimal positioning of central venous catheters, Czepizak [/bib_ref] were recommended that right infraclavicular subclavian catheters should be inserted to height/10-2 cm, right internal or external jugular catheters to height/10 cm and left external jugular catheters to height/10 þ 4 cm. However, the implantation method that utilized in these studies was puncture method but not vessel cutdown. In addition, these authors did not explain why they chosen height (in centimeter)/10 as a reference standard. In this study, we identified height/10 to have moderate correlation to the distance from the entry vessel to the carina. The moderate correlation may be caused by the 3-dimensional configuration of entry vessel. The reference length from entry vessel to carina from this study was height/10, and actual length could be further decided by surgeon under intraoperation fluoroscopy surveillance. This recommendation is easily to follow and could place catheter tip more precisely.
However, there were still limitations in our study. First, this is a retrospectively study and majority patients (92.8%) received intravenous port implantation via right side entry vessel. Our result may be more coincidence in implantation via right entry vessel. The recommended length from left side entry vessel could be longer because that would be cross midline via left brachiocephalic vein and the consideration to minimized the catheter impingement to SVC lateral wall. However, by the aid of intraoperative fluoroscopy, we could visualized the carina and placed catheter more precisely. Second, we used fluoroscopy instead of ECG as intraoperation surveillance. From literature review, the ECG monitoring system could precisely predict catheter tip location of peripherally inserted central catheter up to 97.7% under echo-guidance puncture. [bib_ref] Evaluation of the Sherlock 3CG Tip Confirmation System on peripherally inserted central..., Johnston [/bib_ref] In addition, no more need for radiological image tool. However, in central vein catheter implantation the ECG monitoring still has high variation in tip location and malposition rate. [bib_ref] Analysis of the Sherlock II tip location system for inserting peripherally inserted..., Lelkes [/bib_ref] In this study, we use vessel cutdown as implantation method. The method not only minimized tissue trauma (ie, no more additional wound and needle puncture) but also ensure high successful rate (100%) in single operation. Fluoroscopy could provide real-time visualization of entry vessel and provide 3-dimensional configuration of entry vessel. We could utilize metallic wire to establish entry route prior catheter implantation to overcome the anatomic variation and inadequate vessel size. Third, we did not identify any risk factors for peripheral movement group, which may need further investigation. Fourth, we could not conduct a quantified formula for calculation for each entry points because of limited cases and further investigation was warranted. Even the limitations remains, our study still has useful clinical information as follows. The intravenous movement of catheter was also exist in intravenous port catheter. The intraoperative fluoroscopy provides accurate catheter tip location in 63.9% patients. Additional length of catheter implantation seems unnecessary in 80.6% patients, that is, those who presented as nonmovement and central migration group. Patients with short height, larger SVC diameters at the brachiocephalic vein confluence site had greater risk for catheter central movement. Height/10 may be consider as reference length of implantation for inexperience surgeon and precise implantation length could be adjust under guidance of fluoroscopy.
# Conclusion
The intravenous movement of catheter was also exist in intravenous port catheter. The intraoperative fluoroscopy provides accurate catheter tip location in 63.9% patients. Additional length of catheter implantation seems unnecessary in 80.6% patients, that is, those who presented as nonmovement and central migration group. Patients with short height, larger SVC diameters at the brachiocephalic vein confluence site had greater risk for catheter central movement. Height/10 may be consider as reference length of implantation for inexperience surgeon and precise implantation length could be adjust under guidance of fluoroscopy.
[fig] FIGURE 1: (A) Nut-catheter angle 9,10 : The angle between the locking nut toward the ring center and the proximal end of the outflow of the catheter. (B) Catheter tip location (distance to carina) 9,10 : The distance between the catheter tip and the carina. (C) Short axis of the SVC's diameter at the brachiocephalic vein confluence site. (D) Short axis of the SVC's diameter at the azygos vein confluence site. [/fig]
[table] TABLE 1: Clinical Characteristics of All Patients BMI ¼ body mass index, CT ¼ computed tomography, SVC ¼ supersuperior vena cava. [/table]
[table] TABLE 2: Clinical Characteristics of Nonmovement, Central Movement, and Peripheral Movement Groups [/table]
[table] TABLE 3: Clinical Characteristics of Nonmovement and Central Movement GroupsVariables (Mean AE SD) BMI ¼ body mass index, CT ¼ computed tomography, SD ¼ standard deviation, SVC ¼ superior vena cava. [/table]
[table] TABLE 4: Multivariate Analysis of Risk Factors of Central Movement [/table]
|
Exploring the complex free-energy landscape of the simplest glass by rheology
For amorphous solids, it has been intensely debated whether the traditional view on solids, in terms of the ground state and harmonic low energy excitations on top of it, such as phonons, is still valid. Recent theoretical developments of amorphous solids revealed the possibility of unexpectedly complex free-energy landscapes where the simple harmonic picture breaks down. Here we demonstrate that standard rheological techniques can be used as powerful tools to examine nontrivial consequences of such complex free-energy landscapes. By extensive numerical simulations on a hard sphere glass under quasistatic shear at finite temperatures, we show that above the so-called Gardner transition density, the elasticity breaks down, the stress relaxation exhibits slow, and ageing dynamics and the apparent shear modulus becomes protocol-dependent. Being designed to be reproducible in laboratories, our approach may trigger explorations of the complex free-energy landscapes of a large variety of amorphous materials.
A morphous and crystalline solids have very different behaviours under external perturbations, especially rheological properties under shear deformations [bib_ref] The nature of boson peak in Raman scattering in glasses, Malinovsky [/bib_ref] [bib_ref] Relationships of shear-induced changes in the potential energy landscape to the mechanical..., Malandro [/bib_ref] [bib_ref] Strain versus stress in a model granular material: a devil's staircase, Combe [/bib_ref] [bib_ref] Nonlinear stress and fluctuation dynamics of sheared disordered wet foam, Pratt [/bib_ref] [bib_ref] A nanoindentation study of serrated flow in bulk metallic glasses, Schuh [/bib_ref] [bib_ref] Amorphous systems in athermal, quasistatic shear, Maloney [/bib_ref] [bib_ref] Do athermal amorphous solids exist?, Hentschel [/bib_ref] [bib_ref] Modeling the mechanics of amorphous solids at different length scale and time..., Rodney [/bib_ref] [bib_ref] Replica theory of the rigidity of structural glasses, Yoshino [/bib_ref] [bib_ref] Shear modulus of glasses: results from the full replica-symmetry-breaking solution, Yoshino [/bib_ref] [bib_ref] Avalanche contribution to shear modulus of granular materials, Otsuki [/bib_ref] [bib_ref] Marginal stability in structural, spin, and electron glasses, Müller [/bib_ref] [bib_ref] Protocol-dependent shear modulus of amorphous solids, Nakayama [/bib_ref] [bib_ref] Breakdown of elasticity in amorphous solids, Biroli [/bib_ref] [bib_ref] Universality of slip avalanches in flowing granular matter, Denisov [/bib_ref] [bib_ref] Breakdown of nonlinear elasticity in amorphous solids at finite temperatures, Procaccia [/bib_ref]. It is well known that by increasing the shear strain, a crystal displays a linear elastic response, followed by plastic deformation and yielding. However, experiments and numerical simulations show that this picture breaks down for amorphous solids, such as glasses [bib_ref] Relationships of shear-induced changes in the potential energy landscape to the mechanical..., Malandro [/bib_ref] [bib_ref] A nanoindentation study of serrated flow in bulk metallic glasses, Schuh [/bib_ref] [bib_ref] Amorphous systems in athermal, quasistatic shear, Maloney [/bib_ref] [bib_ref] Do athermal amorphous solids exist?, Hentschel [/bib_ref] [bib_ref] Modeling the mechanics of amorphous solids at different length scale and time..., Rodney [/bib_ref] [bib_ref] Protocol-dependent shear modulus of amorphous solids, Nakayama [/bib_ref] [bib_ref] Breakdown of nonlinear elasticity in amorphous solids at finite temperatures, Procaccia [/bib_ref] [bib_ref] Elasticity in amorphous solids: nonlinear or piecewise linear?, Dubey [/bib_ref] , granular matter [bib_ref] Strain versus stress in a model granular material: a devil's staircase, Combe [/bib_ref] [bib_ref] Avalanche contribution to shear modulus of granular materials, Otsuki [/bib_ref] [bib_ref] Universality of slip avalanches in flowing granular matter, Denisov [/bib_ref] and foams [bib_ref] Nonlinear stress and fluctuation dynamics of sheared disordered wet foam, Pratt [/bib_ref] , where the elastic behaviour is mixed with plastic events. Such plastic events cause sudden drops in stress-strain curves, and are sometimes referred to as crackling noise [bib_ref] Crackling noise, Sethna [/bib_ref] , due to their similarities to avalanches in earthquakes. An apparent shear modulus or rigidity m, which is the ratio between the stress and strain, can be nevertheless defined and measured. Experiments on glassy emulsion systems [bib_ref] Elasticity of compressed emulsions, Mason [/bib_ref] [bib_ref] Osmotic pressure and viscoelastic shear moduli of concentrated emulsions, Mason [/bib_ref] show that m scales linearly mBP with the pressure P both below and above the jamming density, while harmonic treatments predict mBP 1.5 (below) [bib_ref] On the rigidity of a hard-sphere glass near random close packing, Brito [/bib_ref] and mBP 0.5 (above) [bib_ref] Jamming at zero temperature and zero applied stress: the epitome of disorder, O'hern [/bib_ref] , respectively. These contradictions reveal that amorphous solids can be strikingly softer than purely harmonic solids like crystals, even at sufficiently low temperatures where the harmonic expansion was conventionally expected to be valid.
On the theoretical side, the mean-field theory based on the exact solution in the large dimensional limit of the hard sphere glass has brought a more accurate and comprehensive picture beyond the harmonic description [bib_ref] Shear modulus of glasses: results from the full replica-symmetry-breaking solution, Yoshino [/bib_ref] [bib_ref] Breakdown of elasticity in amorphous solids, Biroli [/bib_ref] [bib_ref] Exact theory of dense amorphous hard spheres in high dimension. I. The..., Kurchan [/bib_ref] [bib_ref] Exact theory of dense amorphous hard spheres in high dimension. II. The..., Kurchan [/bib_ref] [bib_ref] Exact theory of dense amorphous hard spheres in high dimension. III. The..., Charbonneau [/bib_ref] [bib_ref] Fractal free energy landscapes in structural glasses, Charbonneau [/bib_ref] [bib_ref] Following the evolution of hard sphere glasses in infinite dimensions under external..., Rainone [/bib_ref] [bib_ref] Following the evolution of glassy states under external perturbations: the full replica..., Rainone [/bib_ref]. The main outcome is the prediction of a Gardner transition (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] that divides the classical amorphous phase into two: in the stable phase (or normal phase), the state is confined in one of the simple smooth basins on the free-energy landscape; once the system is compressed above the Gardner transition density j G (or is cooled down below the Gardner transition temperature T G ), the simple glass basin splits into a fractal hierarchy of subbasins and the glass state becomes marginally stable. Although similar ideas of complex energy landscapes have been conceived phenomenologically in earlier works (see ref. [bib_ref] Exploring the potential energy landscape of glass-forming systems: from inherent structures via..., Heuer [/bib_ref] and references therein), the mean-field theory gives a firmer first principle ground for such a picture, with falsifiable predictions. In particular, the theory predicts that the elastic anomalies and nontrivial rheology should only appear in the marginally stable phase (or Gardner phase) 10,14 that lies deep inside the glassy phase. However, the mean-field theory is exact only in the large dimensional limit, and its relevance in real systems is far from obvious. Here we test the theoretical proposal of the nontrivial rheology in physically relevant dimensions d ¼ 2 and 3 and compare quantitatively the theoretical predictions with our numerical data.
We design laboratory-reproducible rheological protocols to examine the signatures of the intriguing complex free-energy landscape. Our protocols are applied on densely packed hard spheres, a simple and representative glass-forming model. Our result shows the anticipated anomalous rheology emerging at the Gardner transition that turns out to be strikingly similar to the dynamical responses of spin glasses to an external magnetic field. The evidence of a complex free-energy landscape in the Gardner phase is consistent with a previous numerical study [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref] where particles' vibrational dynamics is analysed. That approach has been used in a recent experiment of an agitated granular system [bib_ref] Experimental evidences of the Gardner phase in a granular glass, Seguin [/bib_ref]. However, generalizing the method to other systems, such as molecular glasses, may not be easy due to the difficulty of tracking trajectories of individual particles. The approach proposed in this study overcomes this problem, since it requires no microscopic information, but only the standard macroscopic rheological measurements (the shear stress and strain) that are well accessible in many experimental systems, including molecular and metallic glasses, polymers and colloids. In the present paper, however, we do not attempt to judge whether the Gardner transition survives in finite dimensional systems as a sharp phase transition or becomes a crossover (in the thermodynamic limit), but rather we aim to explore the possibilities to observe its nontrivial signatures in experimentally feasible length/timescales.
# Results
Preparation of stable glasses. To avoid crystallization, we work on a polydisperse mixture of hard spheres whose diameters are distributed according to a probability distribution P(D)BD À 3 , for D min rDoD min /0. (see Supplementary Note 1). A glass is typically obtained by a slow compression (or cooling) annealing from a dilute state, where it falls out of equilibrium at the compression (or cooling) rate-dependent glass transition density j g (or glass transition temperature T g ). Since we choose hard spheres as our working system, the density is the control parameter.
We design a numerical protocol to mimic a simple shear experiment of deeply annealed glasses (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. Our protocol includes three steps. We first use the swap algorithm [bib_ref] Experimental evidences of the Gardner phase in a granular glass, Seguin [/bib_ref] [bib_ref] Equilibrium sampling of hard spheres up to the jamming density and beyond, Berthier [/bib_ref] to prepare a well-equilibrated, supercooled-liquid configuration at various densities j g (see Supplementary Methods and [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. The algorithm combines the Lubachevsky-Stillinger algorithm 37 that consists of standard event-driven molecular dynamics (MD) and slow compression, with Monte Carlo swaps of particle diameters. The MD time is expressed in units of ffiffiffiffiffiffiffiffiffiffiffiffi bm D 2 p , where the particle mass m and mean diameter D, as well as the inverse temperature b, are all set to unity. In other words, a particle travels over a distance of the order of the diameter within a unit MD time. From the thermodynamic point of view, the system is still in the liquid but we work at density j g sufficiently above the mode-coupling theory (MCT) crossover density j d . Then, once we switch off the particle swapping and return to the natural dynamics simulated by MD, the a-relaxation time has become much larger than our MD simulation timescales so that the system behaves essentially as a solid. This glass is thus ultrastable, in a sense similar to those obtained by vapour deposition experiments [bib_ref] Suppression of tunneling two-level systems in ultrastable glasses of indomethacin, Pérez-Castañeda [/bib_ref] [bib_ref] Hydrogen-free amorphous silicon with no tunneling states, Liu [/bib_ref] [bib_ref] Suppression of b relaxation in vapor-deposited ultrastable glasses, Yu [/bib_ref]. At a given density j g , we prepare many of such equilibrated configurations that are statistically independent from each other, and we call them samples in the following.
Second, subsequently the equilibrated configuration is compressed up to a target density j with a compression rate d g ¼ 10 À 3 . From a single sample that is a starting equilibrated configuration at j g , we generate an ensemble of compressed glasses at j, obtained by choosing statistically independent initial particle velocities drawn from the Maxwell-Boltzmann distribution. We call each of such compressed glasses as a realization in the following. These realizations are out of equilibrium, since they no longer follow the liquid equation of state (EOS), but we consider that they remain in restricted equilibrium [bib_ref] Following the evolution of hard sphere glasses in infinite dimensions under external..., Rainone [/bib_ref] for joj G , that is, they are equilibrated within the given glass state determined by the sample. The MD preserves the kinetic energy so that the system remains at the unit temperature throughout our simulations. The typical scale of the vibrations of the particles within the glass states depends on j. For instance, it varies from 10 À 1 to 10 À 2 for j ¼ 0.645 to j ¼ 0.688 (for j g ¼ 0.643, see [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref] of ref. [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref] , so that particles make 10 to 10 2 collisions within a unit MD time.
Third, for a given realization, a simple shear is applied. The simple shear is modelled by an affine deformation of the x-coordinates of all particles, x i -x i þ gz i , under the Lees-Edwards boundary condition 41 with fixed system volume. The shear strain is increased quasistatically with a small constant shear rate _ g¼10 À 4 , such that the shear rate dependence is negligible in the regime joj G (see for a discussion on the _ g dependence), and the shear stress S is measured at different g. The shear stress S and the pressure P are both calculated from interparticle interactions due to collisions between hard sphere particles. For convenience, we introduce reduced pressure p ¼ bP/r and reduced stress s ¼ bS/r, where r is the number density of the particles (see Supplementary Note 1). Note that as the pressure, the shear stress is entirely due to momentum exchanges between the particles so that the rigidity is purely entropic in hard sphere systems. Furthermore, because shear stress and pressure have the same physical dimension, it is convenient to introduce a rescaled stresss¼s=p.
Breakdown of elasticity. [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] shows the phase diagram for our polydisperse hard sphere model, and typical stress-strain curves of individual realizations in different density regimes. In the stable glass phase j g ojoj G [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] , the stress-strain curve shows a smooth linear (harmonic) response regime at small g, followed by a sharp drop of the stress s, signalling the yielding of the system. At yielding, a system-wide shear band emerges (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] , and the system is driven out of a free-energy metastable glass basin. After yielding, the system enters a steady flow state, similar to those observed in athermal amorphous solids under quasistatic shear [bib_ref] Amorphous systems in athermal, quasistatic shear, Maloney [/bib_ref] [bib_ref] Statistical physics of elastoplastic steady states in amorphous solids: Finite temperatures and..., Karmakar [/bib_ref]. In the Gardner phase j G ojoj J , where j J is the jamming density, the harmonic response is punctuated by mesoscopic plastic events (MPEs) that can happen at very small g (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. These MPEs correspond to sudden avalanche-like heterogeneous rearrangements of particle positions without formation of band-like patterns (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. Similar MPEs have been observed in quasistatic shear simulations at zero temperatures 2,6 , but our simulations are performed at finite temperatures. Note that the details of the plastic events, including the locations of yielding, jamming and MPEs, depend on the samples (see and realizations (see . For the behaviour of the stress-strain curves averaged over many realizations and samples, see Supplementary Figs 2-4, as well as Supplementary Notes 2 and 3.
For large j, the stress s grows dramatically at large g, and appears to diverge (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. This shear jamming phenomenon is due to the dilatancy effect of hard sphere glasses under shear: the pressure p increases with g when the system volume is fixed. Note that if p is kept as a constant when g is increased, then the volume expands due to the dilatancy effect. In that case, shear jamming does not appear and shear yielding is recovered (see . While the switching from shear yielding to shear jamming with increasing j is not a consequence of the Gardner transition, it implies that the system is trapped more deeply in the metastable basin, and that the activated barrier crossing between metastable basins becomes forbidden. However, Using the swap algorithm we first prepare equilibrium samples at various densities j g (green squares) whose pressure obeys the Carnahan-Stirling empirical liquid EOS 34 . Next we switch off the swap algorithm, and perform compression annealing from j g to jamming (blue triangles), producing realizations of compressed glasses at various densities j. The system is now out of equilibrium and the pressure follows the glass EOSs pp1/(j J À j) (black dotted lines) [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref]. The Gardner transition j G (red circles and line) separates the stable (light yellow regime) and the marginally stable (light blue regime) glass phases. The insets show schematic depictions of free-energy landscapes in these two different phases. As an example, an equilibrium configuration is prepared at j g ¼ 0.643, and compressed (solid black line) up to j J ¼ 0.690(1). We show typical stress-strain curves under quasistatic shear with increasing g, using a single realization of the compressed glass of N ¼ 1,000 particles, at (b) j ¼ 0.670 (pink cross) and (d) j ¼ 0.688 (pink plus) that are below and above j G ¼ 0.684(1) respectively. Curves in (b,d) are zoomed in (insets) for gr0.025, to show the different small-g behaviours in the two cases. The real-space vector fields of particle displacements are visualized in (c) for a yielding event (between the two red circles in (b)), and (e) for a MPE (between the two red circles in (d)), where each sphere is located at the equilibrium position before yielding/ MPE, and each vector represents the displacement during yielding/MPE. We have subtracted the affine part caused by shear from the displacements, and only show top 20% particles with large displacements. A shear band around the middle of the z-axis is observed in (c). The sizes of particles are reduced by a factor of 0.4, and the vectors are amplified in length by a factor of 2 in (c) and a factor of 15 in (e). The colour represents the magnitude of displacement.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14935 ARTICLE the emergence of subbasins in the Gardner phase [bib_ref] Fractal free energy landscapes in structural glasses, Charbonneau [/bib_ref] [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref] implies that even though the usual relaxation (a-relaxation) is frozen, an additional slow dynamics may appear. This aspect is explored below.
Ageing and slow dynamics. We next show that in the Gardner phase, the relaxation of shear stress becomes complicated, accompanied by ageing and a slow dynamics. Due to the similarity between the Gardner transition and the spin glass transition, here it is very useful to first recall what happens in spin glasses that are essentially disordered and highly frustrated magnets [bib_ref] Spin glasses: experimental facts, theoretical concepts, and open questions, Binder [/bib_ref]. The mean-field spin glass theory has suggested complex free-energy landscapes of spin glasses manifested as continuous replica symmetry breaking, much as what happens in the Gardner phase of hard sphere glasses [bib_ref] Exact theory of dense amorphous hard spheres in high dimension. III. The..., Charbonneau [/bib_ref] [bib_ref] Fractal free energy landscapes in structural glasses, Charbonneau [/bib_ref]. Remarkably, this feature is predicted to have a reflection in the dynamics, resulting in nontrivial dynamical responses to external magnetic field, and ageing effects in the relaxation of magnetization [bib_ref] Analytical solution of the off-equilibrium dynamics of a long-range spin-glass model, Cugliandolo [/bib_ref] [bib_ref] On the out-of-equilibrium relaxation of the sherrington-kirkpatrick model, Cugliandolo [/bib_ref] [bib_ref] Measuring equilibrium properties in aging systems, Franz [/bib_ref].
In experiments, the simplest approach to examine the intriguing features of the dynamics is a combination of the so-called zero-field cooling (zfc) and field cooling (fc) protocols.
In the zfc protocol, one cools a spin glass sample from a high temperature in the paramagnetic phase down to a target temperature T, where a magnetic field h is switched on and one measures the increase of the magnetization. In the fc protocol, one first switches on the magnetic field h, and then subsequently cools the system down to the target temperature T and measures the remanent magnetization. The key point is that, in the two protocols, the order of cooling and switching on of the magnetic field is reversed. In such experiments [bib_ref] Low-dc-field susceptibility of Cu Mn spin glass, Nagata [/bib_ref] [bib_ref] Experimental study of the de Almeida-Thouless line by using typical Ising spin-glass..., Katori [/bib_ref] , the magnetizations observed in the zfc/fc protocols are the same if the working temperature T is higher than the spin glass transition temperature, while the fc magnetization becomes larger than the zfc magnetization if T is lower than the spin glass transition temperature. The anomaly, that is, the difference between the zfc and fc magnetizations, is naturally explained by the mean-field theory. Furthermore, examinations of the ageing effects by these protocols give detailed information about the complex free-energy landscape [bib_ref] Analytical solution of the off-equilibrium dynamics of a long-range spin-glass model, Cugliandolo [/bib_ref] [bib_ref] On the out-of-equilibrium relaxation of the sherrington-kirkpatrick model, Cugliandolo [/bib_ref] [bib_ref] Measuring equilibrium properties in aging systems, Franz [/bib_ref].
It has been pointed out theoretically that the shear on structural glasses plays a very similar role as the magnetic field on spin glasses [bib_ref] Replica theory of the rigidity of structural glasses, Yoshino [/bib_ref] [bib_ref] Emergence of rigidity at the structural glass transition: a first-principles computation, Yoshino [/bib_ref] , and that the relaxation of the shear stress should also reflect the complex free-energy landscape encoded by the continuous replica symmetry breaking solution in the . We show results for several different waiting time t w , under an instantaneous increment of shear strain g ¼ 10 À 3 . Data are averaged over many realizations of compressed glasses obtained from a single equilibrated sample at j g ¼ 0.643 with N ¼ 1,000 particles. Here the rescaled remanent stresss 0 is measured in the ZFC protocol at j, after the longest waiting time t w ¼ 1,000 and before the shear strain is applied. The differences ZFC t; t w ð ÞÀs FC t; t w ð Þ quickly vanishes and does not show significant t w dependence at (c) j ¼ 0.670, while it decays much slower and shows a strong t w -dependent ageing effect at (d) j ¼ 0.688. Note that by definition,s FC ðtÞ is a one variable function, but we plot it here ass FC t; t w ð Þto compare it withs ZFC t; t w ð Þ. The pressure p is independent of time and protocol, in both cases (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. The error bars denote the s.e.m.
Gardner phase (see , and [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. The shear strain and stress in structural glasses correspond to the magnetic field and magnetization in spin glasses, respectively. Furthermore, apparently compression in hard sphere glasses corresponds to cooling in spin glasses. Therefore, inspired by the zfc/fc experiments in spin glasses, we design two distinct protocols that are combinations of compression and shear exerted in reversed orders. In the zero-field compression (ZFC) protocol, we first compress the configuration from j g to j, and set the time to zero. We then wait for time t w before a shear strain g is applied instantaneously (see Supplementary Methods), and measure the relaxation of the stress s ZFC (t, t w ) as a function of the time t ¼ t À t w elapsed after switching on the strain. On the other hand, in the field compression (FC) protocol, we first apply an instantaneous increment of shear strain at the initial density j g , compress the configuration to j and set the time to zero. Then, we measure the relaxation of the stress s FC (t) as the function of the elapsed time t.
For joj G , no ageing effect is observed, and the dynamics is fast. The s ZFC (t, t w ) is stationary or time translationally invariant, that is, s ZFC (t, t w ) ¼ s ZFC (t), depending only on the time difference t ¼ t À t w but not on the waiting time t w (see [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref]. After a timescale t b corresponding to the ballistic motions of particles [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref] , the ZFC stress s ZFC (t, t w ) converges quickly to s FC (t) that is almost a constant in time.
In contrast, for j4j G , s ZFC (t, t w ) displays strong t w -dependent ageing effects manifesting the out-of-equilibrium nature of the system, as well as a slow dynamics. In such a situation, different large time limits can emerge depending on the order of t-N and t w -N (ref.. An important feature that can be seen in [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref] is that s ZFC (t, t w ) exhibits a plateau suggesting the existence of a large time limit s ZFC lim t-N lim tw -N s ZFC (t, t w ) where t w -N is taken before t-N. On the other hand, s FC (t) is again essentially constant in time t (for t4t b ) and we shall denote it as s FC . In the reversed order of the large time limits, we expect that the ZFC shear stress decays to the FC one, lim tw -N lim t-N s ZFC (t, t w ) ¼ s FC . However, the convergence becomes slower as t w increases, and its corresponding timescale could be beyond the simulation time window, as shown in the case of [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref].
Apparently s ZFC is larger than s FC when j4j G , implying the ergodicity breaking. The ageing effect and the slowing down of dynamics show the similarities between the Gardner transition and the liquid-glass transition, demonstrating that the Gardner transition could be considered as a 'glass transition within the glass phase' (see also Supplementary Note 3). In a sharp contrast, because the Gardner transition is absent in a crystal, its shear stress relaxes faster when j increases, and no ageing is present.
Protocol-dependent shear modulus. The above observation suggests that the linear shear moduli measured by the two protocols should be distinct in the Gardner phase. We determine the apparent shear modulus m as m ZFC ¼ (s ZFC À s 0 )/g and m FC ¼ (s FC À s 0 )/g, where s 0 is the remanent shear stress at j before g is applied. The shear strain is increased quasistatically with rate _ g¼10 À 4 up to a predetermined small target g. The shear stress is measured at t ¼ 1 after waiting for t w ¼ 10. Details on the time and j g dependences of the shear modulus are discussed in [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref] shows that while m ZFC and m FC are indistinguishable in the stable glass phase joj G (or pop G ), they become clearly distinct in the Gardner phase j4j G (or p4p G ). For a similar result of a two-dimensional bidisperse hard disk model, see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. This behaviour of shear modulus is a consequence of the time dynamics of the shear stress illustrated in [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref] : at the timescales used to measure the shear modulus (t ¼ 1 and t w ¼ 10), the two shear stresses s ZFC and s FC have converged to the same value for joj G , but remain different for j4j G . The bifurcation point determines the Gardner transition threshold j G (or p G ). Within the numerical accuracy, the j G determined from this approach is fully consistent with the previous estimate based on particles' vibrational motions and caging order parameters [bib_ref] Growing timescales and lengthscales characterizing vibrations of amorphous solids, Berthier [/bib_ref]. To further test this result, we perform detailed analysis on its dependence on the number of particles N and the shear strain g, as discussed below.
We find no appreciable finite size effects for m FC (see [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref] that is in contrast to the observation in nonequilibrated systems, where m FC decreases to zero in the thermodynamic limit [bib_ref] Protocol-dependent shear modulus of amorphous solids, Nakayama [/bib_ref]. It suggests that preparing deeply equilibrium configurations is the key to observe the nonvanishing m FC . While the shear moduli measured around the Gardner transition, and therefore the determination of j G , are N independent, stronger finite size effects are observed for m ZFC at large p near the jamming limit: m ZFC is lower in larger systems, suggesting a stronger nonlinear effect. Nevertheless, the data of m ZFC (p), with a fixed g, appear to converge for N\2,000, confirming that m ZFC (p) and m FC (p) remain distinguishable in the thermodynamic limit, for j4j G .
Regarding the g-dependence, [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref] shows that, within the numerical accuracy, m FC is independent of g, as long as g is sufficiently small. On the other hand, for j4j G and a given N, m ZFC slightly increases with decreasing g. This result shows that in the Gardner phase, the nonlinear effect on m ZFC remains even for very small g, consistent with the observation of elasticity breakdown in [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. Such nonlinear effects are observed for any N studied (see [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] , and we expect that in the thermodynamic limit N-N, a pure linear behaviour of m ZFC can only exist in the limit g-0 (ref. [bib_ref] Statistical physics of the yielding transition in amorphous solids, Karmakar [/bib_ref]. The vanishing of the pure elastic regime distinguishes the Gardner phase from the normal glass and crystalline phases. For a more detailed discussion on how the shear moduli depend on the strain g, the particle number N, the initial density j g and the waiting time t w , see Supplementary Note 4.
For j4j g , the mean-field theory predicts two power-law scalings in the large p limit 10 : m ZFC $ p k with k ¼ 1.41574..., and m FC Bp. The first scaling has also been derived semiempirically by an independent approach 54 . We find good agreement between the theory and simulation on the scaling of m FC (see [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref]. For m ZFC , a noticeable discrepancy is observed in the limit of large N for a fixed finite g [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref] , but the discrepancy decreases when g-0 for a fixed N [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref] , or when N is decreased for a fixed g [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref]. This is because the mean-field m ZFC is obtained in the pure linear response limit g-0, while the nonlinear effect caused by MPEs would increase with g and N, as discussed above. The scaling m FC Bp is consistent with the experimental observation in emulsions [bib_ref] Elasticity of compressed emulsions, Mason [/bib_ref] [bib_ref] Osmotic pressure and viscoelastic shear moduli of concentrated emulsions, Mason [/bib_ref]. Considering the experimental system is possibly not deeply equilibrated, we expect that the relaxation of experimental m(t) is sufficiently fast, and the measurement was performed in the long time limit m(t-N)-m FC (see the discussion of [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref].
Interpretation of results. The Gardner transition is a consequence of the split of glass basins in the phase space [bib_ref] Fractal free energy landscapes in structural glasses, Charbonneau [/bib_ref] , and the split of particle cages in the real space (see [fig_ref] Figure 4 |: Illustration of protocols [/fig_ref]. The schematic plot of the free-energy F as a function of g in [fig_ref] Figure 4 |: Illustration of protocols [/fig_ref] illustrates how a glass basin splits into many subbasins once the system is compressed above j G . Here we interpret our results based on this free-energy landscape viewpoint. First, in the ZFC protocol, the system intends to remain in one of the subbasins after compression (note that different realizations may end up in different subbasins), but as g increases in a quasistatic shear procedure [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] , it may become unstable where the shear stress drops abruptly, resulting in a MPE. The MPE could be interpreted as shear-induced barrier crossing between subbasins, analogous to the barrier crossing between basins in a yielding event. Second, if g is fixed, the shear stress relaxes with time, and according to the Arrhenius law, the emergence of barriers between subbasins would result in a slowing down of the relaxation dynamics with j [fig_ref] Figure 2 |: Relaxation of shear stress [/fig_ref]. The appearance of ageing further reveals the emergence of complex structures within a basin, similar to the mechanism of ageing in the glass transition. Third, because in the FC protocol, the system can overcome the subbasin barriers, the m FC always corresponds to the second-order curvature of basins, rather than that of subbasins as in the m ZFC case. This results in m FC om ZFC for the regime j4j G as observed in [fig_ref] Figure 3 |: Protocol-dependent shear modulus [/fig_ref]. Note that according to [fig_ref] Figure 4 |: Illustration of protocols [/fig_ref] , one should obtain a shear modulus close to m ZFC if an additional strain is applied after FC, as confirmed in [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref]. On the other hand, no basin split occurs and therefore two protocols are equivalent in the stable phase joj G . A previous study 13 has shown that m ZFC ¼ m FC for crystals. The similarity between crystals and stable glasses further confirms that their free-energy basins are similarly structureless.
# Discussion
We wish to stress that our data cannot exclude that the Gardner transition becomes just a crossover in finite dimensions, such that no real phase transition exists. Yet, irrespective of the sharpness of the Gardner transition, we rationalize here in a unified framework all the observations obtained on the rheological (a) In the ZFC protocol, the system is first compressed and then sheared, while the order is reversed in the FC protocol. (b) State point (j g ,0): the schematic free-energy F as a function of the strain g at the initial density j ¼ j g before compression. We assume that the initial state point (black open circle) is located at the minimum of the parabola. To show an example of the real-space particle caging, we also plot three independent trajectories of the same tagged particle in the same two-dimensional sample (see Supplementary Note 5). (c) State point (j, 0): if the system is compressed first to j (above the Gardner transition density j G ), the free-energy basin (red dashed line) splits into many subbasins (blue line): the state point (blue solid circle) becomes trapped in one of the subbasins. The dotted blue lines represent the metastable region of the subbasins. The split of free-energy basin corresponds to the split of cage in the real space (as an example, see the independent trajectories representing three split cages). (d) State point (j g , g): on the other hand, if the system is sheared first, the state point (red solid circle) is forced to climb up the parabola of the basin. (e) State point (j, g): after both shear and compression, the state point can be located at different points in the same free-energy landscape, depending on the order of the compression and shear. In the ZFC case, the state point (blue solid circle) is forced to climb up the subbasin where it is trapped, while it can remain at lower free-energy state in the FC protocol (red solid circle). Because subbasins are metastable (dotted blue line), MPEs occur with increasing g in a quasistatic shear, and slow relaxation occurs for a fixed g (green arrow). The shear stress s is determined by s $ dF=dg (right panel), and the shear modulus by m¼ds=dg $ d 2 F=dg 2 . The stress-strain curves show that for j4j G , m ZFC (slope of blue line) is larger than m FC (slope of dashed red line).
behaviour of the simple hard sphere glass, and find quantitatively reasonable agreement between the theory and simulations. Thus, even if the Gardner transition is not sharp in the thermodynamic limit, for accessible sizes in numerical simulations, and likely for those in experiments as well [bib_ref] Experimental evidences of the Gardner phase in a granular glass, Seguin [/bib_ref] , a behaviour reminiscent of the transition can be clearly observed.
Finally, we make remarks on experimental consequences. It is an intriguing question to clarify whether the phase diagram presented in [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] is generic in a wide range of amorphous solids, ranging from different kinds of glasses to soft matter such as colloids (one can choose to change the temperature or pressure as the control parameter depending on specific systems). The crucial point is to keep track of the dynamical effects that might have been overlooked in some previous experiments, for the following two reasons. First, in reasonably stabilized dense systems, the liquid EOS (green line in [fig_ref] Figure 1 |: Typical stress responses under quasistatic shear [/fig_ref] and the Gardner line (red line) becomes separated enough, so that the liquid dynamics (a-relaxation) and the intriguing internal glassy dynamics (b-relaxation induced by the Gardner transition) can be well separated in timescales. In this respect, recently developed experimental techniques, such as the vapour deposition [bib_ref] Suppression of tunneling two-level systems in ultrastable glasses of indomethacin, Pérez-Castañeda [/bib_ref] [bib_ref] Hydrogen-free amorphous silicon with no tunneling states, Liu [/bib_ref] [bib_ref] Suppression of b relaxation in vapor-deposited ultrastable glasses, Yu [/bib_ref] and the high pressure path 55 , or the use of sufficiently old natural glasses 56 , would provide ideal settings. If such an ideal setting is not possible, one could freeze the a-relaxation out of the experimental time window, by working at sufficiently low temperatures or high densities. The second reason is that by experimentally studying the ageing effects due to the internal dynamics of the amorphous solids, the complexity of the free-energy landscape could become manifested as we demonstrated in the present paper.
Data availability. The data that support the findings of this study are available in Osaka University Knowledge Archive (OUKA) with the identifier. http://hdl.handle.net/11094/59688.
[fig] Figure 1 |: Typical stress responses under quasistatic shear. (a) Illustration of the protocol on the polydisperse hard sphere glass phase diagram (adapted from ref. 34), where k B T/P ¼ 1/(rp). The MCT dynamical crossover (yellow star) is located at j d ¼ 0.594(1) along the equilibrium liquid EOS (green line). [/fig]
[fig] Figure 2 |: Relaxation of shear stress. Relaxations of the rescaled ZFC shear stresss ZFC ¼s ZFC =p (filled symbols) and the rescaled FC shear stress s FC ¼s FC =p (open symbols) show different behaviours at (a) j ¼ 0.670 and (b) j ¼ 0.688, corresponding to the pink plus and cross in Fig. 1, respectively (the Gardner transition density j G ¼ 0.684(1) (ref. 34)) [/fig]
[fig] Figure 3 |: Protocol-dependent shear modulus. (a) The rescaled shear modulusm¼m=p, obtained from both ZFC (filled symbols) and FC (open symbols), is plotted as a function of p, for j g ¼ 0.643 and several N. The data are obtained by using g ¼ 2 Â 10 À 3 , and are averaged over N s E200 samples, and N r E100 individual realizations for each sample. The two shear moduli m ZFC and m FC coincide below p G (vertical dashed line), and become distinct above, where p G ¼ 265 (ref. 34). The data are compared with the large p scalings predicted by the mean-field theory m ZFC $ p 1:41574 and m FC Bp (black solid lines). The difference m ZFC À m FC is plotted as a function of j in the inset, where the vertical dashed line represents j G ¼ 0.684 (ref. 34). (b) Rescaled ZFC and FC shear moduli obtained from a few different g, for N ¼ 1,000 systems. The error bars denote the s.e.DOI: 10.1038/ncomms14935 | www.nature.com/naturecommunications [/fig]
[fig] Figure 4 |: Illustration of protocols. We show the evolution of the freeenergy landscape and the state point (j, g) under compression and shear. [/fig]
|
High-Incidence of Human Adenoviral Co-Infections in Taiwan
Background: Respiratory infections caused by adenovirus (HAdV) are common year round. Recently, a significant increase of adenoviral infections was observed in Taiwan.Objective:To understand the prevalence and molecular epidemiology of respiratory adenovirus circulating in Taiwan for the past decade.Study Design:One hundred and twenty-six human adenoviruses, isolated between 2002 to 2011, were characterized via DNA sequencing of the hexon and fiber genes. The nucleotide sequences were then compared by phylogenetic analysis.Results: HAdV-B3 accounted for 64.3% (81/126) and peaked almost every year, whereas the sequences of hexon and fiber genes of HAdV-B3 were highly conserved in different years. A high incidence of co-infection of adenoviruses was observed (19.0%, 24/126); HAdV-B3 co-infected with HAdV-C2 was the most common combination (58.3%, 14/24). An additional interesting finding of repeated infection was noted in 10 children, all of whom showed first infection with adenovirus species HAdV-C, followed by species HAdV-B or HAdV-E.Conclusions:HAdV-B3 was the predominant type of respiratory adenovirus circulating in Taiwan over the past ten years. This merits further attention for vaccine development. Furthermore, the observed high-incidence of adenoviral co-infections along with repeated infections found in our study provides important epidemiological insights into adenovirus infections.
# Introduction
Adenoviruses exhibit various clinical presentations: respiratory illness, gastroenteritis, ocular or urinary tract infection [bib_ref] Immunoglobulin class-specific serological responses to adenovirus in respiratory infections of young adult..., Julkunen [/bib_ref] [bib_ref] Direct identification of enteric adenovirus, a candidate new serotype, associated with infantile..., Johansson [/bib_ref] [bib_ref] Restriction endonucleases in identification of a genome type of adenovirus 19 associated..., Wadell [/bib_ref] [bib_ref] Adenoviruses in the immunocompromised host, Hierholzer [/bib_ref] [bib_ref] Prevalence of antibodies to adenovirus serotypes 4 and 7 among unimmunized US..., Ludwig [/bib_ref]. Respiratory adenoviruses usually cause mild upper but can also cause severe lower respiratory tract infection [bib_ref] Severe communityacquired pneumonia caused by adenovirus type 11 in immunocompetent adults in..., Gu [/bib_ref] [bib_ref] Epidemiology of severe pediatric adenovirus lower respiratory tract infections in Manitoba, Canada, Alharbi [/bib_ref]. Our prior report showed adenovirus accounting for about 4.0% of all respiratory infections in Taiwan during 1997-1999 and was the most common agent associated with tonsillitis and pharyngitis [bib_ref] Respiratory viral infections among pediatric inpatients and outpatients in Taiwan from 1997..., Tsai [/bib_ref].
Adenoviruses are non-enveloped DNA viruses and are divided into seven species (A to G) with more than 57 genotypes identified [bib_ref] Computational analysis of two species C human adenoviruses provides evidence of a..., Walsh [/bib_ref]. In general, respiratory diseaseassociated adenoviruses include species HAdV-B (HAdV-B3, HAdV-B7, HAdV-B11 and HAdV-B14), HAdV-C (HAdV-C1, HAdV-C2 and HAdV-C5) and HAdV-E (HAdV-E4) [bib_ref] Molecular analysis of adenovirus isolates from vaccinated and unvaccinated young adults, Blasiole [/bib_ref] [bib_ref] Molecular epidemiology and brief history of emerging adenovirus 14-associated respiratory disease in..., Kajon [/bib_ref] [bib_ref] Genome type analysis of adenovirus types 3 and 7 isolated during successive..., Kim [/bib_ref] [bib_ref] PCR analysis of egyptian respiratory adenovirus isolates, including identification of species, serotypes,..., Metzgar [/bib_ref]. Molecular epidemiology of respiratory adenoviruses, assessed by DNA sequencing of the hexon and/or fiber genes, has been reported in several countries. In the USA , Germany , Palestine and , HAdV-B3 was the most predominant type, followed by HAdV-B7 or HAdV-C2 [bib_ref] Human adenoviruses in respiratory infections: sequencing of the hexon hypervariable region reveals..., Biere [/bib_ref] [bib_ref] A molecular epidemiology survey of respiratory adenoviruses circulating in children residing in..., Qurei [/bib_ref] [bib_ref] Molecular epidemiology and clinical presentation of human adenovirus infections in Kansas City..., Selvaraju [/bib_ref] [bib_ref] Human adenovirus infection in children with acute respiratory tract disease in Guangzhou, Zou [/bib_ref]. In contrast, Israel and had HAdV-C1 and HAdV-C2 as predominant types, respectively [bib_ref] Molecular identification of adenovirus causing respiratory tract infection in pediatric patients at..., Abd-Jamil [/bib_ref] [bib_ref] Adenovirus infections in hospitalized patients in Israel: epidemiology and molecular characterization, Mandelboim [/bib_ref]. By using genomic-RFLP or PCR-RFLP analysis, previous reports have shown that HAdV-E4 and HAdV-B3 were the major types of respiratory adenovirus circulating in southern Taiwan in 2001 and 2002 respectively [bib_ref] Respiratory adenoviral infections in children: a study of hospitalized cases in southern..., Chen [/bib_ref]. HAdV-B3 was the major type circulating in northern Taiwan during 2004 and 2005 [bib_ref] A community-derived outbreak of adenovirus type 3 in children in Taiwan between, Chang [/bib_ref].
Most recently, a significant increase of adenovirus prevalence was found in Taiwan in 2011, possibly due to different emerging adenoviruses. To better understand the molecular epidemiology and evolution of adenoviruses circulating in Taiwan, we investigated the genotypes of respiratory adenoviruses circulating in Taiwan over a full decade (2002 to 2011) by DNA sequencing of hexon and fiber genes. The prevalent types of adenovirus in Taiwan in the past decade were identified. Interestingly, a high-incidence of adenoviral co-infections as well as adenoviral repeated infections was revealed in this study.
# Materials and methods
# Ethics statement
This study was conducted in Taiwan only, and Institutional Review Board (IRB) approval was obtained from National Cheng Kung University Hospital (No. A-BR-101-020). This was a retrospective study without intervention or obtaining extra clinical specimens. Human specimens were not directly used in this research and informed consent was waived. The waiving of informed consent was also approved by the Institutional Review Board of National Cheng Kung University Hospital.
## Viral culture
Viral cultures of laboratory-confirmed adenovirus infection were processed as described previously [bib_ref] Respiratory viral infections among pediatric inpatients and outpatients in Taiwan from 1997..., Tsai [/bib_ref]. This study used adenovirus isolated from either nasopharyngeal aspirate or throat swabs; and stored at -80°C at the Virology Laboratory of National Cheng Kung University Hospital before use. The study population was children from either hospitalized patients or outpatients of Cheng Kung University Hospital. Viruses used in this study were sub-cultured in A549 cells with Dulbecco's modified eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin. A549 cells in 15-ml culture tubes were infected with viral suspension, and then incubated at 35°C with 5% CO 2 . When 85% cytopathic effect was observed, cells were harvested for DNA extraction.
## Viral genome extraction and genotyping
Viral DNA was extracted by the phenol/chloroform method according to a modified procedure from a previous report [bib_ref] Assessment of genetic variability among subspecies b1 human adenoviruses for molecular epidemiology..., Kajon [/bib_ref]. Volumes of reagents used in the procedure were reduced, and DNA pellets were finally resuspended in 50 μl of ddH 2 O.
To type the adenovirus, the Loop 1 region of the hexon gene was amplified with generic primer pair HXL1F (5′-CGTGTGCAGTTYGCCCG) and HXL1R (5′-ACAGCCTGATTCCACAT). The Loop 2 region of the hexon gene was amplified with generic primer pair BL and BR (5′-CTTGTATGTGGAAAGGCAC) [fig_ref] Table 1: The co-infection isolates found in this study a [/fig_ref] [bib_ref] Phylogenetic analysis of the main neutralization and hemagglutination determinants of all human..., Madisch [/bib_ref]. To identify the hexon gene in co-infected isolates, type-specific primer pairs were designed. For the Loop 2 region, primers AdC2F and HXL2R were used for HAdV-C2; AdC5F and HXL2R for HAdV-C5; CDL and HXL2R-2 for HAdV-C6. In addition, AdBL1F and BR were used to amplify the region containing both Loop 1 and 2 for HAdV-B3. To amplify the fiber gene, primers FiBL and FiBR were used for HAdV-B3, HAdV-E4, HAdV-B7 and HAdV-B11; FiCL-2 and FiCR-2 for HAdV-C1; FiCL and FiCR for HAdV-C2; FiCL-2 and FiCR-3 for HAdV-C5. PCR mixtures consisted of 1 U of DNA polymerase (KOD plus polymerase, Toyobo), 1 mM MgSO 4 , 0.2 mM dNTP, 300 pM of each primer and 1 to 2 μl of template from the original purified DNA solution in a 50-μl reaction volume. PCR cycling consisted of initial denaturation at 95°C for 5 minutes, followed by 40 cycles of 94°C for 20 seconds, 54°C or 56°C for 20 seconds, depending on primers used, and 72°C for 40 or 80 seconds, depending on the length of PCR products. DNA sequencing of PCR products was carried out with the primers used in PCR by Sanger's method.
# Phylogenetic analysis
Phylogenetic trees were constructed by the neighbor-joining method based on PAUP software [bib_ref] Inferring evolutionary trees with PAUP*, Wilgenbusch [/bib_ref]. About 1500 base pairs of hexon gene (nt 139-1612 based on HAdV-B3) and 700 base pairs of fiber gene (nt 224-888 based on HAdV-B3) were analyzed. Sequences of reference strains were obtained from GenBank (for hexon gene, accession number for HAdV-C1
# Results
## Dna sequencing of hexon and fiber genes of respiratory adenovirus circulating between 2002 and 2011
To analyze the molecular epidemiology of respiratory adenoviruses circulating in Taiwan, adenovirus-positive isolates collected at National Cheng Kung University Hospital (NCKUH) during 2002-2011 were examined. There were 2,001 adenovirus-positive isolates associated with respiratory diseases [fig_ref] Figure 1: Numbers of respiratory adenoviral isolates collected at National Cheng Kung University Hospital... [/fig_ref]. One to two isolates per month over a decade were randomly selected for our investigation. A total of 106 adenovirus isolates were successfully subcultured. In addition, 20 isolates from repeated infections were also examined. The repeated infections were identified from our adenovirus database by computerized system according to patient identification number. In total, there were 126 isolates
## Multiple infections of adenovirus
Sequencing results showed superimposed peaks in the chromatograms in 26 isolates using the generic primers BL and BR for the hexon gene [fig_ref] Figure 3: Multiple templates revealed after nucleotide sequencing of the hexon Loop 2 region [/fig_ref]. Multiple templates indicated a mix of multiple adenoviral strains in one specimen. To confirm this probability, PCR products amplified by BL and BR for hexon Loop 2 region in two co-infected isolates were cloned and sequenced; distinct hexon genes of different types were verified (data not shown). The other co-infected isolates were confirmed by direct evidence of sequencing results of the hexon and/or fiber genes using type-specific primers developed in this study for different hexon (Loop 1 or 2) or fiber regions [fig_ref] Table 1: The co-infection isolates found in this study a [/fig_ref]. Co-infections were confirmed when two types of hexon genes and/or fiber genes from the same region were identified by sequencing [fig_ref] Table 1: The co-infection isolates found in this study a [/fig_ref]. In all, 24 samples (19.0%, 24/126) with more than one adenoviral strain in each were identified: 14 samples (58.3%) containing HAdV-C2 and HAdV-B3; 5 (20.8%) containing HAdV-C1 and HAdV-B3; 2 (8.3%) containing HAdV-B3 and HAdV-C5; 1 each (4.1%) containing HAdV-B3 and HAdV-C6; HAdV-C2 and HAdV-B11; and HAdV-C2, HAdV-B3 and HAdV-B11. Interestingly, all co-infected strains involved different species of adenovirus, including species HAdV-B and HAdV-C.
## Repeated infections of adenoviruses
From our database, we noted that there were some children with repeated adenoviral infections from 2002 to 2011. Repeated infections may be caused by a variant strain of adenovirus. Therefore, these twenty isolates from three females and seven males were examined for genotypes [fig_ref] Table 2: Repeated infections of adenovirus found in this study [/fig_ref]. Intervals between infections in each patient ranged from three months to five years. Intriguingly, all these children were first infected with species HAdV-C adenoviruses (HAdV-C1, HAdV-C2 or HAdV-C5), followed by species HAdV-B (HAdV-B3) or species HAdV-E (HAdV-E4). This is the first report tracking repeated adenoviral infections over a decade.
## Phylogenetic analysis of the hexon and fiber genes
Phylogenetic analysis of hexon and fiber genes identified 152 strains clustering with reference strains from the GenBank [fig_ref] Figure 4: Analysis of partial hexon nucleotide sequences [/fig_ref] and [fig_ref] Figure 1: Numbers of respiratory adenoviral isolates collected at National Cheng Kung University Hospital... [/fig_ref]. Strains of most genotypes from diverse years were grouped into only one cluster in both hexon and fiber genes, indicating sequences conserved throughout. Interestingly, only HAdV-C5 strains revealed two phylogenetic [fig_ref] Table 2: Repeated infections of adenovirus found in this study [/fig_ref] ; by contrast, their fiber gene regions were grouped in one cluster. Differences in the fiber gene of HAdV-C1 strains of 2.2% and 6.3% were also found in nucleotide and amino acid sequences, respectively. Overall, the results indicate higher nucleotide variability in species HAdV-C adenoviruses than in species HAdV-B and HAdV-E.
# Discussion
This molecular epidemiology study of adenovirus in Taiwan found a high incidence of adenoviral co-infection along with repeated infections. Adenoviral co-infections found in the present study were based on DNA sequencing of hexon and fiber genes, thus providing a high level of confidence in the validity of our data. Furthermore, unique infection patterns involving different species and type order were revealed in repeated adenoviral infections. Taken together, our investigation affords interesting insights into the respiratory adenoviral epidemiology.
Multiple-infection (co-infection) of adenoviruses was detected in 19.0% of clinical isolates analyzed. Adenoviral co-infection has also been reported in other countries, with a range of 1.5-19.0% [bib_ref] PCR analysis of egyptian respiratory adenovirus isolates, including identification of species, serotypes,..., Metzgar [/bib_ref] [bib_ref] Rapid identification of human adenovirus types 3 and 7 from respiratory specimens..., Lee [/bib_ref] [bib_ref] Use of PCR to demonstrate presence of adenovirus species B, C, or..., Echavarria [/bib_ref] [bib_ref] Molecular typing of clinical adenovirus specimens by an algorithm which permits detection..., Mccarthy [/bib_ref] [bib_ref] Comprehensive serotyping and epidemiology of human adenovirus isolated from the respiratory tract..., Lee [/bib_ref]. Variance in co-infection rate may arise from different communities where samples were taken or methodologies used. Primers used in PCR may also play a critical role in detecting co-infection. In this study, the phenomenon of multiple templates was detected in 26 isolates with the generic primer pair of hexon Loop 2, but only in two isolates with the generic primer pair of hexon Loop 1 (data not shown). This indicates that multiple primer sets may be required to detect co-infection.
This study discovered adenoviral co-infection via multiple PCR-sequencing templates, which were amplified using generic primers of the hexon Loop 2 region. Distinct hexon genes in co-infected isolates were then verified with typespecific primers and followed by nucleotide sequencing. In other studies, most adenoviral co-infections were determined with type-specific primers without DNA sequencing [bib_ref] PCR analysis of egyptian respiratory adenovirus isolates, including identification of species, serotypes,..., Metzgar [/bib_ref] [bib_ref] Rapid identification of human adenovirus types 3 and 7 from respiratory specimens..., Lee [/bib_ref] [bib_ref] Use of PCR to demonstrate presence of adenovirus species B, C, or..., Echavarria [/bib_ref] [bib_ref] Comprehensive serotyping and epidemiology of human adenovirus isolated from the respiratory tract..., Lee [/bib_ref] [bib_ref] Coinfections of adenovirus species in previously vaccinated patients, Vora [/bib_ref]. For detecting co-infection, the application of generic primers is more convenient than that of type-specific primers in a community with diverse adenovirus populations; nucleotide sequencing must be determined subsequently. Genotypes in co-infected isolates still must be identified with type-specific primers, since multiple templates may appear in nucleotide sequencing. Furthermore, direct sequencing of the same genomic region should be the required criterion for the determination of co-infection since recombinants between different adenoviruses may exist. In addition, our evidence for co-infections may give rise to potential recombinants which can lead to more potent HAdVs [bib_ref] Computational analysis of four human adenovirus type 4 genomes reveals molecular evolution..., Dehghan [/bib_ref].
Adenoviral co-infection has never been reported in previous epidemiology studies in Taiwan, which used genomic-RFLP or PCR-RFLP for adenovirus typing [bib_ref] Respiratory adenoviral infections in children: a study of hospitalized cases in southern..., Chen [/bib_ref] [bib_ref] A community-derived outbreak of adenovirus type 3 in children in Taiwan between, Chang [/bib_ref]. Our study revealed only dominant type (species HAdV-C) in the genomic-RFLP pattern (data not shown). Bias in titer in co-infected isolates was also found by plaque purification in another study [bib_ref] Coinfections of adenovirus species in previously vaccinated patients, Vora [/bib_ref] , which suggests that PCR-sequencing for genotyping may be period examined. Most common co-infections of adenoviruses in other nations also correlated with the predominant types in their community [bib_ref] PCR analysis of egyptian respiratory adenovirus isolates, including identification of species, serotypes,..., Metzgar [/bib_ref] [bib_ref] Rapid identification of human adenovirus types 3 and 7 from respiratory specimens..., Lee [/bib_ref] [bib_ref] Coinfections of adenovirus species in previously vaccinated patients, Vora [/bib_ref] , suggesting adenovirus coinfection is not type-dependent. Nevertheless, we noted that strains in co-infected isolates always came from different species (HAdV-B and HAdV-C), with the exception of one coinfected case with three strains. These findings suggest a requirement for different species involved in co-infection and warrant further investigation.
Notably, repeated infections of adenovirus from 10 children were found and analyzed; all 10 were first infected with species HAdV-C, followed by species HAdV-B or HAdV-E. This suggests that primary infection with species HAdV-C may be able to protect against subsequent HAdV-C infections, but not against subsequent infection by species HAdV-B or HAdV-E. This may also explain why the proportion of species HAdV-C adenovirus is lower than species HAdV-B in many countries [bib_ref] Human adenoviruses in respiratory infections: sequencing of the hexon hypervariable region reveals..., Biere [/bib_ref] [bib_ref] A molecular epidemiology survey of respiratory adenoviruses circulating in children residing in..., Qurei [/bib_ref] [bib_ref] Molecular epidemiology and clinical presentation of human adenovirus infections in Kansas City..., Selvaraju [/bib_ref] [bib_ref] Human adenovirus infection in children with acute respiratory tract disease in Guangzhou, Zou [/bib_ref]. Adenoviral infection of vaccinated persons has been reported in American military recruits [bib_ref] Coinfections of adenovirus species in previously vaccinated patients, Vora [/bib_ref]. Infections in vaccinated persons were usually caused by a variant of the vaccine-targeted strain HAdV-E4 or HAdV-B7, indicating that adaptive immunity induced by species HAdV-B or HAdV-E is insufficient to safeguard against infection by variant strains of species HAdV-B or HAdV-E. Repeated detection of adenovirus in upper respiratory infections has also been reported by Kula et al [bib_ref] Persistence of adenovirus nucleic acids in nasopharyngeal secretions: a diagnostic conundrum, Kalu [/bib_ref]. Three types of repeated infections were identified including 1) adenovirus of same strain, 2) adenovirus of same type but variant strains, 3) adenovirus of different types which were infected with species HAdV-B or HAdV-E (HAdV-B3 or HAdV-E4) and followed by species HAdV-C (HAdV-C1) of adenovirus in an 8-month or 24-month old child, respectively. The difference between this study and our results may be due to the younger age of patients and shorter interval of repeated infections in the study by Kula et al [bib_ref] Persistence of adenovirus nucleic acids in nasopharyngeal secretions: a diagnostic conundrum, Kalu [/bib_ref]. Taken together, these results suggest that further serologic study is required to understand the phenomenon of repeated infections of adenovirus. This study showed HAdV-B3 as the most common respiratory adenovirus strain circulating in Taiwan during 2002-2011, except in 2007. It was likewise the most common in America, Germany, Palestine and China during 1999-2010 [bib_ref] Human adenoviruses in respiratory infections: sequencing of the hexon hypervariable region reveals..., Biere [/bib_ref] [bib_ref] A molecular epidemiology survey of respiratory adenoviruses circulating in children residing in..., Qurei [/bib_ref] [bib_ref] Molecular epidemiology and clinical presentation of human adenovirus infections in Kansas City..., Selvaraju [/bib_ref] [bib_ref] Human adenovirus infection in children with acute respiratory tract disease in Guangzhou, Zou [/bib_ref]. This indicates that the most prevalent type of respiratory adenovirus in Taiwan is similar to that in many other countries. Sequences of hexon and fiber genes of HAdV-B3 were highly conserved between outbreaks, suggesting HAdV-B3 outbreaks may not be caused by changes in antigenicity. Conservation of the hexon gene evident in this study is consistent with a previous report that the hexon gene was conserved in most types circulating in Japan from 1988 to 2007 [bib_ref] Stability of the seven hexon hypervariable region sequences of adenovirus types 1-6..., Mizuta [/bib_ref]. This study found two phylogenic clusters of HAdV-C5 in the hexon gene. Variation within HAdV-C5 was also discovered in Germany and Japan [bib_ref] Human adenoviruses in respiratory infections: sequencing of the hexon hypervariable region reveals..., Biere [/bib_ref] [bib_ref] Stability of the seven hexon hypervariable region sequences of adenovirus types 1-6..., Mizuta [/bib_ref]. It is not clear whether different HAdV-C5 clusters are prevalent in different years, since only 10 isolates of HAdV-B5 were identified here.
In summary, this 10-year surveillance study showed that HAdV-B3 accounted for 64.3% of respiratory adenoviral infections, with strong nucleotide sequence conservation of the hexon and fiber genes of HAdV-B3. The results suggest development of HAdV-B3 vaccine is warranted. In addition, a high incidence of adenoviral co-infection was revealed. Most co-infections were HAdV-B3 co-infected with HAdV-C2. Finally, variant antigenic specificity among types of adenovirus was observed in children with repeated infections. [fig_ref] Figure 1: Numbers of respiratory adenoviral isolates collected at National Cheng Kung University Hospital... [/fig_ref]. Analysis of nucleotide sequences of a portion of the fiber gene (nucleotides 224-888/HAdV-B3).
## Supporting information
Phylogenetic tree with 1000 bootstrap replicates was constructed from selected sequences of adenoviruses isolated during 2002-2011. Reference strains were selected from the GenBank, bootstrap values greater than 80 considered as significant and indicated in the figure (Accession number for HAdV-C1: AF534906, KNIH 00/2: AY224420, Guangzhou01: DQ099432.4, NHRC 3: AY599837, HAdV-C5: AY339865, 0901HZ/ShX/CHN/2009: JF800905, Taiwan/760/2002: FJ841913). Strains of species HAdV-B from co-infected isolates are indicated by original isolate name; strains of species HAdV-C from co-infected isolates are indicated by isolate name-2 (e.g., N78/TW/03 is HAdV-B3, N78/TW/03-2 is HAdV-C2). Strains of repeated infections were not shown in this phylogenetic tree. (TIFF) [fig_ref] Table 1: The co-infection isolates found in this study a [/fig_ref]. Primers used in this study a . a Primer BL, BR, CDL, FiBL and FiBR were designed by Madich et al. [bib_ref] Phylogenetic analysis of the main neutralization and hemagglutination determinants of all human..., Madisch [/bib_ref]. b Y refers to nucleobase cytosine or thymine. c R refers to nucleobase adenine or guanine. (DOCX)
[fig] Figure 1: Numbers of respiratory adenoviral isolates collected at National Cheng Kung University Hospital from 2002 to 2011. doi: 10.1371/journal.pone.0075208.g001 clusters with 4.7% difference in the hexon region [/fig]
[fig] Figure 2: Proportional distribution of different adenoviral types from 2002 to 2011. Total amount of HAdV-C1, HAdV-C2, HAdV-B3, HAdV-E4, HAdV-C5, HAdV-C6, HAdV-B7 and HAdV-B11 strains is 15, 30, 81, 5, 10, 3, 3 and 5, respectively.an advantageous strategy for detecting co-infection events. Most common adenoviral co-infections revealed in this study was HAdV-B3 co-infected with HAdV-C2, likely because HAdV-B3 and HAdV-C2 are predominant types in Taiwan for the [/fig]
[fig] Figure 3: Multiple templates revealed after nucleotide sequencing of the hexon Loop 2 region. Partial result of the chromatogram of N8630/TW/09 is shown, sequencing carried out with the primer BL used in the PCR. Upper panel shows nucleotide sequences of HAdV-B3 (1240-1264) and HAdV-C2 (1297-1321) corresponding to the region, with consensus nucleotides marked. The phenomenon of multiple templates was also observed when nucleotide sequencing was carried out with primer BR (data not shown). doi: 10.1371/journal.pone.0075208.g003 [/fig]
[fig] Figure 4: Analysis of partial hexon nucleotide sequences (nucleotides 139-1612/HAdV-B3). Phylogenetic tree with 1000 bootstrap replicates was constructed from selected sequences of adenoviruses isolated during 2002-2011, reference strains selected from GenBank. Bootstrap values greater than 80 were considered as significant and indicated in the figure. Strains of species HAdV-B from co-infected isolates are indicated by original isolate name; strains of species HAdV-C from co-infected isolates are indicated by isolate name -2 (e.g., N78/TW/03 is HAdV-B3, N78/TW/03-2 is HAdV-C2). Strains of repeated infections and strains in which only Loop 2 sequence has been identified were not shown in this phylogenetic tree. doi: 10.1371/journal.pone.0075208.g004 [/fig]
[table] Table 2: Repeated infections of adenovirus found in this study. [/table]
[table] Table 1: The co-infection isolates found in this study a . [/table]
[table] Table S2: Pairwise comparison of nucleotide and amino acid sequences a . a About 1,500 nucleotides of hexon and 940 nucleotides of fiber genes from different types were analyzed. b NA, not available. (DOCX) [/table]
|
Early Clinical Manifestations Associated with Death from Visceral Leishmaniasis
Background: In Brazil, lethality from visceral leishmaniasis (VL) is high and few studies have addressed prognostic factors. This historical cohort study was designed to investigate the prognostic factors for death from VL in Belo Horizonte (Brazil).Methodology: The analysis was based on data of the Reportable Disease Information System-SINAN (Brazilian Ministry of Health) relating to the clinical manifestations of the disease. During the study period (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009), the SINAN changed platform from a Windows to a Net-version that differed with respect to some of the parameters collected. Multivariate logistic regression models were performed to identify variables associated with death from VL, and these were included in prognostic score.Principal Findings: Model 1 (period 2002-2009; 111 deaths from VL and 777 cured patients) included the variables present in both SINAN versions, whereas Model 2 (period 2007-2009; 49 deaths from VL and 327 cured patients) included variables common to both SINAN versions plus the additional variables included in the Net version. In Model 1, the variables significantly associated with a greater risk of death from VL were weakness (OR 2.9; 95%CI 1.3-6.4), Leishmania-HIV coinfection (OR 2.4; 95%CI 1.2-4.8) and age $60 years (OR 2.5; 95%CI 1.5-4.3). In Model 2, the variables were bleeding (OR 3.5; 95%CI 1.2-10.3), other associated infections (OR 3.2; 95%CI 1.3-7.8), jaundice (OR 10.1; 95%CI 3.7-27.2) and age $60 years (OR 3.1; 95%CI 1.4-7.1). The prognosis score was developed using the variables associated with death from VL of the latest version of the SINAN (Model 2). The predictive performance of which was evaluated by sensitivity (71.4%), specificity (73.7%), positive and negative predictive values (28.9% and 94.5%) and area under the receiver operating characteristic curve (75.6%).
# Introduction
The number of new cases of visceral leishmaniasis (VL) is estimated to be around 500,000 per year worldwide with over 50,000 deaths. The majority (.90%) of cases is concentrated in six countries, namely, Bangladesh, Brazil, Ethiopia, India, Nepal and Sudan. In Latin America, the causative agent of VL is the intracellular protozoan Leishmania infantum (syn. L. chagasi) [bib_ref] Control of visceral leishmaniasis in latin america-a systematic review, Romero [/bib_ref] , and the disease is systemic and characterized clinically by prolonged fever, weight loss, hepatomegaly, splenomegaly, hypergammaglobulinemia and pancytopenia. In the absence of treatment, the disease may have fatal consequences [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref]. Additionally, the susceptibility to VL, and consequently the epidemiology of the disease, has been influenced by the expansion of human immunodeficiency virus (HIV) in South America, Asia and Africa. Indeed, of the 88 countries that are endemic for VL,have already reported cases of Leishmania-HIV co-infection [bib_ref] Leishmania and human immunodeficiency virus coinfection: the first 10 years, Alvar [/bib_ref]. On this basis, VL is considered an extremely serious public health problem.
In Brazil since the 1980's the geographical distribution of VL has expanded, partly due to increased urbanization [bib_ref] Complexities of assessing the disease burden attributable to leishmaniasis, Bern [/bib_ref] [bib_ref] Worldwide increasing risk factors for leishmaniasis, Desjeux [/bib_ref] [bib_ref] Emergent outbreak of visceral leishmaniasis in Mato Grosso do Sul State, Oliveira [/bib_ref] [bib_ref] An urban outbreak of visceral leishmaniasis in Natal, Brazil, Jeronimo [/bib_ref] [bib_ref] The reemergence of visceral leishmaniasis in Brazil, Arias [/bib_ref] [bib_ref] Casos confirmados de Leishmaniose Visceral, Ministry [/bib_ref] [bib_ref] Letalidade por Leishmaniose Visceral. Brasil, Ministry [/bib_ref]. In the 1980s, an average of 1,500 cases was reported each year in Brazil and between 2000-2009 the average increased to 3,480 cases annually, an increase of 132% [bib_ref] Casos confirmados de Leishmaniose Visceral, Ministry [/bib_ref].
In fact, between 1990 and 2009, a total of 57,973 clinical cases of VL had been reported [bib_ref] Casos confirmados de Leishmaniose Visceral, Ministry [/bib_ref] , representing 90% of all cases notified in the American continent. These figures do not take into account the unreported cases, which are not insignificant [bib_ref] Control of visceral leishmaniasis in latin america-a systematic review, Romero [/bib_ref] [bib_ref] Complexities of assessing the disease burden attributable to leishmaniasis, Bern [/bib_ref] [bib_ref] Epidemiological aspects of human and canine visceral leishmaniasis in Venezuela, Zerpa [/bib_ref].
According to the Visceral Leishmaniais Control and Surveillance Program (VLCP) of Brazil [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] all suspected and confirmed cases of VL must be notified to the sanitary authorities and registered in the Reportable Disease Information System (SINAN, Brazilian Ministry of Health). This system not only provides a center for the collection and processing of data, but also for the dissemination of information generated by the epidemiological surveillance systems linked to the municipal, state and federal governments. Moreover, the SINAN contributes to the knowledge of morbimortality worldwide by VL, since it helps make up the consolidated data from institutions such as PAHO (Pan American Health Organization) and WHO (World Health Organization).
The control of VL in urban areas of Brazil represents, however, a difficult and continuous challenge despite the measures adopted by the Brazilian Ministry of Health, and implemented through of the VLCP [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] , which emphasizes the early diagnosis and treatment of clinical cases [bib_ref] The logic of visceral leishmaniasis control, Dye [/bib_ref]. In most countries, pentavalent antimonial drugs has been the first choice treatment for more than 70 years. Although more specific guidelines for the management of patients suffering from severe VL have been developed in Brazil [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] , the case fatality rate remains high [bib_ref] Letalidade por Leishmaniose Visceral. Brasil, Ministry [/bib_ref]. In Belo Horizonte, the capital of the State of Minas Gerais, the VLCP guidelines have been followed since 1993 but death from VL has not been reduced. In fact, during the period 2002-2009 the case fatality rate of VL in Belo Horizonte ranged from 8.2% (in 2007) to [bib_ref] Prognostic factors in meningococcal disease: development of a beside predictive model an..., Barquet [/bib_ref].0% (in 2009) with an average of 12.6%, whilst the rate in the country was much lower and varied between 5.6% (in 2008) and with an average of 7.0% [bib_ref] Letalidade por Leishmaniose Visceral. Brasil, Ministry [/bib_ref].
Since reduction of case fatality rate is one of the goals of VLCP [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] , it is important to understand the determinants of such a poor result in a metropolis presenting active transmission of L. infantum [bib_ref] Urban parasitology:-visceral leishmaniasis inBrazil, Harway [/bib_ref]. Hence, the objective of the present study was to investigate the early clinical manifestations associated with death from VL using information available from the SINAN database obtained at the moment of clinical case suspicion. Also a prognostic clinical score was proposed to identify patients at a higher risk of death.
# Methods
# Ethical statement
The study was approved by the Ethical Review Board of the Universidade Federal de Minas Gerais (No.211/09) and of the Municipality Health Service of Belo Horizonte (No.075.2008). Data were analyzed anonymously.
## Study design and population
The study was carried out in Belo Horizonte, the capital of the State of Minas Gerais, located in southeastern Brazil, an area comprising 2,375,444 inhabitants. This historical cohort study was based on secondary VL data from 2002 to 2009. Owing to the prolonged incubation period of the disease, data for 2009 were only finalized in March 2010 and, hence, the complete set of information for that year was not available at the time of the study. Data were obtained from the Reportable Disease Information System-SINAN (Brazilian Ministry of Health and Municipality Health Service of Belo Horizonte) and complemented with data of the Mortality Information System-SIM (Brazilian Ministry of Health and Municipality Health Service of Belo Horizonte). The selection criteria for inclusion in the study were: (i) the patient was resident in Belo Horizonte; (ii) the patient represented a new case of VL; (iii) the case was registered at SINAN, and, if appropriate, (iv) the primary cause of death of the patient was VL. Based on these criteria, 888 VL patients (92% of all of those registered) were selected for the study and, of these, 111 died (88% of all deaths from VL.
## Sinan database
The epidemiological surveillance system of the Brazil involved registration of the suspected VL case at SINAN using a form comprising the following entries: date of notification, health unit responsible for notification, address, age, sex, level of schooling, occupation of patient, date of the start of symptoms and clinical manifestations (signs and symptoms). Subsequently, further information was added to the records including the results of specific laboratory examinations, date of beginning of treatment, initial drug used for treatment, drug used following failure of the initial therapy, and evolution of the case.
The SINAN database changed platform during the study period from a Windows-based version to a Net version . As shown in [fig_ref] Table 1: Comparison between the variables of Windows and Net versions of the SINAN... [/fig_ref] , the Net version contained more information than the Windows version except for the field relating to co-infections, which was simplified (HIV remained but tuberculosis was removed). Of the 888 cases, 512 had been registered in the Windows version and 376 in the Net version. Data from the two versions of the SINAN database was combined in order to create a single database, and the consistency of the data contained therein evaluated. The variables analyzed in the present study were sex and age of patient, clinical manifestations (signs and symptoms), co-infections (HIV and tuberculosis), date of notification, date of the onset of symptoms, initial drug administered, duration of treatment with pentavalent antimonial drugs, other drug administered following failure of initial therapy, evolution of the case (cure or death from VL) and date of death (if applicable). Those variables exhibiting high levels of missing information, such as schooling (55.5%), ethnicity (51.8%), weight (83.8%), occupation (85.0%), date of the start of treatment (69.1%) and relapse (86%) were excluded from the study.
# Statistical analysis
Statistical analyses of the data were performed using STATA version 11.0 software (Stata Corp., College Station, TX, USA)
## Author summary
The visceral leishmaniasis (VL) is a disease potentially fatal if not diagnosed and treated opportunely. This article presents the results of the study on the manifestations identified at the time of the clinical suspicion of the VL cases. This study was conducted in Belo Horizonte, the capital of the State of Minas Gerais, located in southeastern Brazil. This study is both timely and substantive because the Belo Horizonte is an area of transmission of VL, with one of the highest VL-death proportions of Brazil. The patients with higher risk of death had at least one of the following characteristics: $60 years, weakness, HIV coinfection, bleeding, jaundice and other associated infections. During the period 2002-2009, 8% to 22% of the patients with VL progressed to death in Belo Horizonte, whilst the proportion in the country was much lower and varied between 5% and 9%. This study has identified vulnerable patients who are at higher risk of death from VL and who would benefit from early predictive evaluation of the prognostic. Hence, the knowledge regarding the factors associated with death may contribute for clinical management and for reduction of deaths from VL.
considering 111 deaths from VL and 777 cured patients. Univariate logistic regression analysis was used to evaluate the demographic and clinical variables according to the occurrence of death from VL. Variables associated with death from VL at a significance level of p,0.25, along with those previously considered in the literature to be biologically important in the occurrence of death from L. infantum, were included in multivariate logistic regression analyses. Variables with more than two categories were transformed into dummies variables.
Two multivariate logistic regression models were subsequently analyzed. With the aim of avoiding selection bias, and to allow better adjustment of the model to VL data, the category ''unknown'' was maintained for variables for which information was missing. Variables presenting collinearity were evaluated and those that better explained the occurrence of death from VL were retained in the model.
A step-by-step backward selection procedure was used to select the variables and to produce the final multivariate logistic regression models. Only adjusted variables showing a significant association (p,0.05) with the occurrence of death from VL remained in the final models. The strength of association was determined by odds ratio at a 95% confidence interval.
The predictive factors relating to death from VL that were identified by Model 2 (period 2007-2009) were used to create a prognosis score. According to the methodology described by , the regression coefficient of each of the variables was divided by the smallest coefficient and the quotient was rounded to the nearest integer in order to facilitate the clinical use of the system. The prognosis score was validated against all cases registered during the period 2007-2009. The actual evolution of each patient, defined as death from VL (yes versus no), was compared with the predictive score. For the purposes of comparison, patients attaining a predictive score in the range from 1 to 5 received an allocated score of 1, while those presenting none of the death prognosis factors received an allocated score of 0. The predictive performance of the score was determined by sensitivity, specificity, positive and negative predictive values, and the area under the receiver operating characteristic curve (ROC).
# Results
The incidence of VL in Belo Horizonte during the period 2002-2009 varied from 3.4 to 6.6/100,000 inhabitants and the case fatality rate was 13.1% . The highest levels of lethality were observed in 2009 (22.0%) followed by 2004 (18.7%). The ages of the 888 VL subjects varied between 3 months and 93 years; the ages of those that died from VL ranged from 5 months to 86 years (n = 111; median = 18 years; interquartile range = 4241 years), whereas the ages of cured patients varied from 3 months to 93 years (n = 777; median = 32 years; interquartile range = 7254 years). Case fatality rate amongst patients presenting Leishmania-HIV co-infection was 17.7% (12/68), while lethality amongst those with Leishmania-tuberculosis co-infection was 10.5% (4/38). Of the nine patients presenting both types of co-infection (Leishmania-HIV-tuberculosis), four progressed to death (44.4%). The number of deaths from VL and the case fatality rate was higher in patients $60 years (24.4%), followed by those within the age range 30-39 years (20.2%) [fig_ref] Table 3: Visceral leishmaniasis cases and fatality rate by age and sex, Belo Horizonte,Brazil,... [/fig_ref].
The results of univariate logistic regression analysis of the demographic and clinical variables with respect of deaths from VL are shown in [fig_ref] Table 4: Demographic and clinical variables according to death from visceral leishmaniasis, Belo Horizonte,... [/fig_ref]. The most significant non-adjusted variables (p,0.05) associated with death were weakness, edema, bleeding, other associated infections, jaundice, Leishmania-HIV coinfection, Leishmania-tuberculosis co-infection, treatment with amphotericin, treatment with pentavalent antimonial drug for 21 to 40 days, male gender, age range 30-39 years and $60 years.
The results of the multivariate logistic regression analysis of those variables that were associated (p,0.25) with death from VL, It is important to emphasize that inclusion of the category unknown in the multivariate logistic regression analysis had no effect on the results generated by models 1 and 2, since the odds ratio and the 95% CI values remained unchanged in either the presence or absence of this category. The use of the category unknown not only allowed a larger number of cases to be considered in the final models but also improved the adjustment of these models to the VL data as shown by the log likelihood and p values. [fig_ref] Table 7: Predictive scoring system for death from visceral leishmaniasis, Belo Horizonte, Brazil, 2007-2009 [/fig_ref] presents the predictive score, prepared from VL cases notified between 2007 and 2009, for each of the four deathassociated factors revealed by Model 2. According to this system, a score of 1 was attributed to the three variables (age $60 years, bleeding and other associated infections), while a score of 2 was attributed to the variable jaundice. The performance measures of the predictive score were sensitivity (71.4%), specificity (73.7%), positive and negative predictive values (28.9% and 94.5%) and the area under the ROC curve (75.6%).
# Discussion
In the present study, the factors associated with death from VL were weakness, Leishmania-HIV co-infection, other associated infections, bleeding, jaundice, and age $60 years. These findings are in agreement with the Brazilian Ministry of Health [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] in which factors associated with death from VL were considered to be age ,6 months or .65 years, jaundice, bleeding and comorbidities including bacterial infections.
The records analyzed in the present study were obtained from a database containing details of registered VL cases, each of which would normally have been notified on clinical suspicion of the disease. Our results reveal that it is possible to detect the factors associated with death from VL at first clinical suspicion of the disease and, hence, to identify the most vulnerable patients.
Timely application of specific and effective measures to the patients would contribute greatly to a reduction in the lethality of the disease. Several authors have suggested that knowledge regarding the laboratory and clinical profiles of patients and their association with death from VL could assist in the clinical management and reduce lethality [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] [bib_ref] Risk factors for death in children with visceral leishmaniasis, Sampaio [/bib_ref].
The case fatality rate registered in Belo Horizonte is one of the highest in Brazil. Some hypotheses could explain the variation in the rates as accessibility to the health service, suspicion of VL and delayed diagnosis, treatment opportunity, clinical management of the patient, toxicity of drugs, comorbities and Leishmania population in circulation in Belo Horizonte. Considering the case fatality rates in 2004 (18.7%) and 2009 (22.0%), it is possible that the complex combination of these factors could contribute to the unacceptable fatality rates during these years.
Factors associated with lethality from VL have been reported in the literature [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] [bib_ref] Risk factors for death in children with visceral leishmaniasis, Sampaio [/bib_ref] [bib_ref] Epidemic visceral leishmaniasis in southern Sudan: treatment of severely debilitated patients under..., Seaman [/bib_ref] [bib_ref] Conflict and kala-azar: determinants of adverse outcomes of kala-azar among patients in..., Collin [/bib_ref]. A retrospective cohort study [bib_ref] Risk factors for death in children with visceral leishmaniasis, Sampaio [/bib_ref] conducted in Recife (north-eastern Brazil) identified risk factors associated with VL death in young patients (,15 years old) as mucosal hemorrhage, jaundice, dyspnoea, bacterial infections, reduced number of neutrophils and platelets. A case-control study [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] conducted in Teresina (also in north-eastern Brazil) described the occurrence of fever for more than 60 days, diarrhea, jaundice and anemia as predictive factors for the death in VL patients. Another case-control study performed in Teresina [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] confirmed that bacterial infections and hemorrhage were the most relevant factors associated with death from VL. However, a study performed in Campo Grande (central west Brazil) identified bacterial infections as the main cause of death among VL patients. According to Seaman et al [bib_ref] Epidemic visceral leishmaniasis in southern Sudan: treatment of severely debilitated patients under..., Seaman [/bib_ref] , age $45 years, disease duration .5 months, undernutrition and intense anemia were associated with an increased risk of death from VL in Sudan. Following a study also performed in Sudan, Collin et al [bib_ref] Conflict and kala-azar: determinants of adverse outcomes of kala-azar among patients in..., Collin [/bib_ref] reported that the prognosis factors for death were age ,2 or $45 years, disease duration .5 months, undernutrition, anemia, splenomegaly and, particularly, episodes of diarrhea, vomit and bleeding.
The time interval between the onset of symptoms and time of diagnosis was estimated and this interval was not significantly associated with death from VL [fig_ref] Table 5: Clinical variables according to death from visceral leishmaniasis, Belo Horizonte, Brazil, 2002-2009 [/fig_ref]. However, other studies show the necessity to reduce delay in diagnosis and describe the [bib_ref] Risk factors for relapse of visceral leishmaniasis in Georgia, Kajaia [/bib_ref] , the factors associated with VL relapse were delay in diagnosis for .90 days, haemoglobin level ,60 g/L and age ,1 year. In Tunisia, children were evaluated and seven prognostic factors at the time of hospital admission were identified: visit delayed .56 days, fever lasting .21 days, normal or low temperature, hemorrhagic syndrome, hemoglobin rate ,5.5 g/dL, sedimentation rate ,25 mm and hypoalbuminemia ,30 g/L. In Sudan [bib_ref] Conflict and kala-azar: determinants of adverse outcomes of kala-azar among patients in..., Collin [/bib_ref] , risk factors for death among adults were age $45 years, malnutrition, anemia and duration of illness $5 months. Also in Sudan [bib_ref] Epidemic visceral leishmaniasis in southern Sudan: treatment of severely debilitated patients under..., Seaman [/bib_ref] , the risk factors in adult patients were duration of illness $5 months, age $45 years, hemoglobin level ,60 g/L, and body mass index ,12 kg/m 2 . The weakness, which is one of the earliest clinical symptoms of VL, was found to be significantly associated with death. The most likely explanation for this finding is that the study was based on clinical manifestations presented by VL-suspect patients at their first medical appointment. However, the early detection of weakness may help to identify those patients presenting a higher likelihood of an unsatisfactory evolution of their disease. A study conducted in central west Brazil involving 55 individuals that had died from VL described the occurrence of hyporexia (65.5%), asthenia (58.1%) and adynamia (29.0%). It is probable that this set of conditions formed part of the weakness reported by patients during anamnesis, thus confirming the findings of the present study.
A VL-suspected case is defined as a patient presenting fever and splenomegaly who originates from an Leishmania transmission area or, if originating from an area in which transmission is absent, where differential diagnosis has been discarded [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref]. As expected, there were high frequencies of patients in the study population presenting either fever with splenomegaly (87.3%) or fever with splenomegaly and hepatomegaly (83.0%), and the frequencies were similar in death from VL and in cured patients. The occurrence of jaundice was registered for 27 patients, 14 (51.9%) of whom died from VL. The presence of jaundice increased the chances of death from VL by a factor of 10 (ORadjusted 10.1; 95%CI 3.7-27.2). The large amplitude of the confidence interval can be explained by the small number of patients presenting jaundice. Following a study carried out in the Brazilian State of Piauí involving 12 deaths from VL and 78 cured individuals, Werneck et al. [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] reported that the death-associated factors were anemia, jaundice, fever for more than 60 days and diarrhea. These authors pointed out that the last three symptoms mentioned were identified at the first clinical examination and that the presence of jaundice increased the chances of death by 10.6- [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref]. It may be concluded that identification of jaundice at the first examination or during a later follow is a valuable predictor of VL patients likely to present a negative prognosis. The occurrence of jaundice may indicate liver damage [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] which, in some cases of VL, has been observed in the form of a massive necrotic destruction of the organ [bib_ref] Histopathological patterns of the liver involvement in visceral leishmaniasis, Duarte [/bib_ref] [bib_ref] Acute hepatitis as a presenting manifestation of kala azar, Hervas [/bib_ref]. Moderate alterations of hepatic function, together with thrombocytopenia, may give rise to serious hemorrhagic processes [bib_ref] Epidemic visceral leishmaniasis in southern Sudan: treatment of severely debilitated patients under..., Seaman [/bib_ref]. A study of clinical and laboratory data relating to 55 patients who had died from VL revealed increased levels of albumin and aspartate aminotransferase at the time of admission to hospital, and high levels of creatinine and amylase prior to death. According to Jeronimo et al. [bib_ref] An urban outbreak of visceral leishmaniasis in Natal, Brazil, Jeronimo [/bib_ref] , increased levels of liver enzymes in untreated patients presenting a profile of jaundice at the time of admission to hospital, may signal the presence of hepatitis by Leishmania infection.
In the present study, the ages of cured patients varied from 3 to 93 years old whereas the ages of those who had died from VL were between 5 months and 86 years. The lethality in older individuals ($60 years) was the highest (24.4%), a value similar to that reported by Oliveira et al.for elderly subjects. In Sudan [bib_ref] Epidemic visceral leishmaniasis in southern Sudan: treatment of severely debilitated patients under..., Seaman [/bib_ref] , age $45 years old was associated with an increased risk of death from VL. It is expected that older VL patients would present a higher risk of mortality for many reasons, e.g. cardiovascular diseases that may coexist with leishmaniasis. Additionally, although N-methyl glucamine antimoniate constitutes the first choice treatment for VL, the drug may have side effects including severe cardiac arrhythmia. This medication was prescribed for 50 of the patients aged $60 years included in the present study, and seven (14.0%) of these died. However, other studies [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] [bib_ref] Conflict and kala-azar: determinants of adverse outcomes of kala-azar among patients in..., Collin [/bib_ref] associated a higher risk of death with those of a younger age. In Uganda [bib_ref] Risk factors for in-hospital mortality of visceral leishmaniasis patients in eastern Uganda, Mueller [/bib_ref] , the main risk factors for in-hospital death identified were age ,6 years and .15 years, concomitant tuberculosis or hepatopathy, and drug-related adverse events. The case fatality rate among patients .45 years of age was strikingly high (29.0%).
Multivariate regression analysis models 1 and 2 identified age of $60 years as a factor associated with death from VL (OR 2.5 and 3.1, respectively). It is worth noting that the present study involved 86 patients aged $60 years and of these 21 died from VL accounting for 19.0% of the total deaths analyzed. However, in 2009, the Brazilian Ministry of Health mentioned that VL patients aged ,1 and .40 years were at greater risk, and more recently, Costa et al. [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] described that individuals in these two age ranges are the most vulnerable to death from VL.
In this study, 63 subjects presented other associated infections with significantly difference between the cured patients (n = 44; 5.7%) and those that died from VL (n = 19; 17.1%) Multivariate logistic regression analysis showed that the occurrence of other infections increased the chances of death by 3-fold. A recent study involving 55 hospitalized VL patients who progressed to death found that 65.5% had been diagnosed with other infections, most commonly sepsis (66.7%) and pneumonia (63.9%), at the time of admission and during hospitalization. The most likely explanation for such a high frequency of co-infection is that all of the studied patients were hospitalized and could be evaluated at the time of admission and throughout hospitalization. Patients with VL are characteristically neutropenic and, therefore, present reduced inflammatory response and are at increased risk from other established or concealed infections [bib_ref] Bacterial infections in patients with visceral leishmaniasis, Andrade [/bib_ref]. This type of physiopathological mechanism may explain the inclusion of other infections among the prognosis factors for death from VL.
Twenty-seven individuals presented history of bleeding at the first clinical examination with frequencies that were significantly different between the 16 cured patients (2.1%) and the 11 deaths from VL (9.9%). The presence of bleeding increased the chances of death by 3.5-fold. Oliveira et al.recently reported the occurrence of bleeding in 32.7% of deaths from VL. Infection and hemorrhage are classical complications of VL [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] and, in the present study, were recorded in 19 and 11 deaths, respectively. Twelve patients were affected by both complications, and of these six (50.0%) died from VL. In a case-control study, Werneck et al. [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] showed that all of the cases of death from VL presented infectious or hemorrhagic complications, while Costa et al. [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref] reported that these conditions represented the two most relevant factors associated with death from VL. Hemorrhagic phenomena are probably associated with disseminated intravascular coagulation, i.e. the activation of coagulation and fibrinolysis that is initiated as part of the inflammatory response by a mechanism similar to that established for sepsis [bib_ref] The impact of the inflammatory response on coagulation, Esmon [/bib_ref].
Following a study of Tunisian children affected from VL, it was reported that bleeding and a period of more than 56 days between the onset symptoms and the first clinical examination were among the seven most important factors associated with a negative prognosis. In the present study, however, the time between the onset of symptoms and clinical suspicion of VL showed no such association.
According to univariate analysis, Leishmania-HIV and Leishmaniatuberculosis co-infections were both significantly (p,0.05) associated with death from VL. However, in the multivariate analysis only co-infection with HIV was significantly correlated with death (ORadjusted 2.4; 95% CI 1. . Leishmania-HIV co-infection is an emerging problem that requires urgent attention since, in recent years, VL has become an opportunistic disease in HIVinfected subjects. Indeed, VL may recur many times in HIVinfected patients regardless of the provision of adequate treatment, and the outcome is often fatal [bib_ref] Predictors of Visceral Leishmaniasis Relapse in HIV Infected Patients: A Systematic Review, Cota [/bib_ref] [bib_ref] The relationship between leishmaniasis and AIDS: the second 10 years, Alvar [/bib_ref] [bib_ref] Leishmaniasis as an emerging infection, Choi [/bib_ref] [bib_ref] Impact of opportunistic disease on survival in patients with HIV infection, Chaisson [/bib_ref]. Moreover, the presence of VL accelerates the progression of HIV by promoting viral replication and aggravating the status of immunosuppression [bib_ref] Impact of opportunistic disease on survival in patients with HIV infection, Chaisson [/bib_ref]. Additionally, it has been observed in vitro that HIV induces the replication of Leishmania by reducing T cells that are able to recognize Leishmania antigens. According to Cruz et al. [bib_ref] Leishmania/ HIV co-infections in the second decade, Cruz [/bib_ref] , HIV may invade and replicate within Leishmania-bearing macrophages even though TCD4 + are the preferred cells. Thus, both pathogens can act in synergism and aggravate the condition of a co-infected patient [bib_ref] Live and killed human immunodeficiency virus type-1 increases the intracellular growth of..., Wolday [/bib_ref].
In a study conducted in north-eastern Brazil involving a chronological serie of hospital records during a period of and relating to 396 VL-patients (76 deaths and 320 cured patients), nine Leishmania-HIV co-infected individuals were identified and six of these progressed to death [bib_ref] Is severe visceral leishmaniasis a systemic inflammatory response syndrome? A case control..., Costa [/bib_ref]. The authors suggested that increased lethality with co-infection could be associated with an increased risk of the disease itself or with the occurrence of other opportunistic infections. The available evidence clearly indicates that VL patients should be tested for HIV, and co-infected individuals should receive special care [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Predictors of Visceral Leishmaniasis Relapse in HIV Infected Patients: A Systematic Review, Cota [/bib_ref].
Regarding the initial drug used in treatment [fig_ref] Table 5: Clinical variables according to death from visceral leishmaniasis, Belo Horizonte, Brazil, 2002-2009 [/fig_ref] , it can be observed that the fatality rate was 6.9% (42/612) for those patients treated with Pentavalent Antimonial. Among those who received Amphotericin, the case fatality rate was 28.9% . This difference demonstrated by univariate logistic regression analysis can be explained by the fact that in Brazil, the pentavalent antimonials are the drugs of choice for treatment of VL due to its proven therapeutic efficacy. Amphotericin is recommend as the first choice in patients six months or under and for those over 65 years old and those with severe clinical manifestation such as: malnutrition, comorbidities (which include bacterial infection), jaundice, hemorrhagic phenomena (except epistaxis), generalized edema, signs of toxemia (lethargy, poor perfusion, cyanosis, tachycardia or bradycardia, hypoventilation or hyperventilation, and hemodynamic instability) [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref].
Measures that have been applied to control the canine reservoir and the insect vector have not been successful in preventing the spread of VL in Brazil, and early diagnose and treatment of human cases remain the main approaches for reducing lethality. Guidelines published by the Brazilian Ministry of Health stress the key factors associated with death from VL with the aim of ensuring that patients who require special care can be identified and classified according to severity of risk [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref]. While the results obtained in the present study support these definitions, in order to reduce death from VL it is necessary first to identify the prognosis factors and then to adopt the correct clinical strategies. As stated by Werneck et al. [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] , the key challenge lies in making the correct medical decisions following the identification of high-risk patients.
The prognostic score generated in the present study is based on four clinical variables (age$60 years, bleeding, other associated infections and jaundice). The predictive performance of this score was: sensitivity 71.4%, specificity 73.7%, positive and negative predictive value (28.9% and 94.5%) and area under the ROC curve (75.6%). The predictive performance could be improved with addition of laboratory variables according to other studies carried out in Brazil [bib_ref] Prognostic factors for death from visceral leishmaniasis in Teresina, Werneck [/bib_ref] [bib_ref] Risk factors for death in children with visceral leishmaniasis, Sampaio [/bib_ref]. However, these laboratory variables are unavailable in SINAN. The advantage of the system developed herein is its simplicity because it was based on patient age and clinical symptoms at the clinical suspicious. Additionally, decisions regarding the clinical management of VL patients can be facilitated by the identification of either a distinct prognosis factor or a combination of factors. However, further studies that advance our understanding of the factors associated with moderate and severe forms of VL are also necessary. Furthermore, it is particularly important to validate the applicability of the score in immunosuppressed VL individuals.
The key issue, however, is to define a simple prognosis score that could be applied in basic health units and would allow the early detection of VL cases for redirection to specialized health service. For the purposes of comparison, patients attaining a predictive score in the range from 1 to 5 received an allocated score of 1, while those presenting none of the death prognosis factors received an allocated score of 0. The scoring system presented here can be used to identify patients running a higher risk of death from VL at the time of clinical suspicion. Those patients with a score between 1 and 5 should receive specialized clinical management during treatment.
Our study had limitations that deserve to be discussed. The multivariate analyses did not take into account all possible factors that could contribute to unfavorable evolution of the VL case, e.g. nutritional status, the presence of other comorbidities (not included here as autoimmune diseases, alcoholism, and other drug abuse). These variables are not collected through the SINAN system. It is noteworthy that this study was designed to assess whether the variables in the SINAN could be useful in identifying patients facing a higher risk of death from VL at moment of clinical suspicion. However, we consider of great importance further studies designed to include, in addition to the SINAN database, clinical and laboratory variables of VL patients.
Another limitation of this study is related to the validation of the prognostic scoring system which was proposed using the same patients whose data was used in its preparation (patients in the period 2007-2009). In fact, as this score included the variables of the newer version of SINAN, there is until now no distinct set of patients for validation. In the future, we intend to validate this prognostic scoring system using several random samples from the SINAN database.
In order to improve the quality of assistance, the Brazilian Ministry of Health [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] [bib_ref] Secretaria de Vigilância em Saúde, Brasil [/bib_ref] has urged local surveillance authorities to provide infrastructure to the basic health units and to encourage the training of the professionals to organize comprehensive support to the patients with efficiency. Moreover, the flow of VL patients through specialized reference service and other health units must be established and disclosed. Some researchers have further suggested that VL-lethality could be greatly reduced by speeding up laboratory diagnosis and developing less toxic alternative drugs.
In Brazil, suspected VL cases must be notified to the health authorities using the appropriate SINAN epidemiological form. All fields in the form must be completed even in the absence of information (missing code). However, some of the variables (schooling, occupation, ethnicity, weight, date of the start of treatment and relapse) could not be evaluated in the present study because the corresponding fields had high proportion of missing data. It would have been helpful to know the date of the start of treatment, since early diagnosis and treatment of VL are important in reducing the lethality of the disease. Indeed, the absence of vital information in partially completed medical records is a chronic problem in Brazil [bib_ref] Avaliação dos Prontuários Médicos de Hospitais de Ensino do Brasil, Silva [/bib_ref].
SINAN has a specific module for VL which includes the registration of the variables described in the literature associated with the disease. In Brazil, the Epidemiological Surveillance Service works in conjunction with other information systems such as Mortality Information System (SIM). So it is routine for this service to verify if deaths from reportable diseases such as VL are listed in SINAN system. Of course, there may be some underreporting due to the difficulty in defining the VL clinical case or identifying the VL as cause of death. Therefore, underreporting is minimized by comparing the two systems mentioned.
This study may provide ammunition to the debate on the usefulness and limitations of the Reportable Disease Information System (SINAN) database. According to the World Health Organization, improvement in the flow of information is one of the main challenges in the control of tropical neglected diseases.
This study has identified vulnerable patients who are at higher risk of death from VL and who would benefit from early predictive evaluation of the prognostic. In conclusion the knowledge regarding the factors associated with death may contribute for clinical management and reduction of lethality from VL.
[table] Table 1: Comparison between the variables of Windows and Net versions of the SINAN database. Leishmania-HIV co-infection and age $60 years were associated with a greater chance of death. Model 2 (period 2007-2009) revealed that variables included in the SINAN Net version, such as other associated infections, bleeding and jaundice, were significantly associated with the increased likelihood of deaths from VL. [/table]
[table] Table 3: Visceral leishmaniasis cases and fatality rate by age and sex, Belo Horizonte,Brazil, 2002 [/table]
[table] Table 4: Demographic and clinical variables according to death from visceral leishmaniasis, Belo Horizonte, Brazil, 2002-2009. The total number of individuals listed in SINAN and included in the study (Belo Horizonte, Brazil) were 888, of which 512 were registered in the Windows version and 376 in the Net version of the database. b Numerical information not applicable. [/table]
[table] Table 5: Clinical variables according to death from visceral leishmaniasis, Belo Horizonte, Brazil, 2002-2009. [/table]
[table] Table 6: Factors associated with death from visceral leishmaniasis, Belo Horizonte, Brazil, 2002-2009. [/table]
[table] Table 7: Predictive scoring system for death from visceral leishmaniasis, Belo Horizonte, Brazil, 2007-2009. [/table]
|
Dynamic interplay between CXCL levels in chronic Hepatitis C patients treated by Interferon
Background: Combined pegylated interferon-α and ribavirin therapy has sustained virological response (SVR) rates of 54% to 61%. Pretreatment predictors of SVR to interferon therapy have not been fully investigated yet. The current study assesses a group of chemokines that may predict treatment response in Egyptian patients with chronic HCV infection. Patients and methods: CXCL5, CXCL9, CXCL11, CXCL12, CXCL 13, CXCL 16 chemokines and E-Cadherin were assayed in 57 chronic HCV patients' sera using quantitative ELISA plate method. All studied patients were scheduled for combined pegylated interferon alpha and ribavirin therapy (32 patients received pegylated interferon α 2b, and 25 patients received pegylated interferon α 2a). Quantitative hepatitis C virus RNA was done by real time RT-PCR and HCV genotyping by INNOLIPAII. Results: There was no significant difference (p > 0.05) in baseline HCV RNA levels between responders and non-responders to interferon. A statistically significant difference in CXCL13 (p = 0.017) and E-Cadherin levels (P = 0.041) was reported between responders and nonresponders at week 12. Significant correlations were found between changes in the CXCL13 levels and CXCL9, CXCL16, E-cadherin levels as well as between changes in E-cadherin levels and both CXCL16 and ALT levels that were maintained during follow up. Also, significant changes have been found in the serum levels of CXCL5, CXCL13, and CXCL16 with time (before pegylated interferon α 2 a and α 2 b therapy, and at weeks 12 and 24) with no significant difference in relation to interferon type and response to treatment.Conclusion: Serum levels of CXCL13 and E-Cadherin could be used as surrogate markers to predict response of combined PEG IFN-α/RBV therapy, especially at week 12. However, an extended study including larger number of patients is needed for validation of these findings.
# Background
Hepatitis C virus (HCV) is a major cause of a wide spectrum of liver diseases ranging from mild chronic hepatitis to end stage cirrhosis and hepatocellular carcinoma. The World Health Organization (WHO) estimates that 170 million individuals worldwide are infected with HCV, However, the prevalence of HCV infection varies throughout the world. One decade ago, Frank et al., reported that Egypt had the highest number of reported HCV infections, largely attributed to the use of contaminated parenteral antischistosomal therapy [bib_ref] The role of parenteral antischistosomal therapy in the spread of hepatitis C..., Frank [/bib_ref]. The overall prevalence of HCV antibodies in the Egyptian population was 15% for age of 15 to 59 years and infection was higher among rural residents [bib_ref] Egypt: Ministry of Health, El-Zanaty and Associates, and Macro International, El-Zanaty [/bib_ref] , and about 91% of the patients were infected with HCV genotype 4 [bib_ref] Genetic epidemiology of hepatitis C virus throughtout Egypt, Ray [/bib_ref].
Combined pegylated interferon (IFN)-α and ribavirin (RBV) therapy has sustained virological response (SVR) rates reaching up to 61% in different clinical trials for HCV Genotype 1 patients [bib_ref] PEGASYS International Study Group: Peginterferon alfa-2a and ribavirin combination therapy in chronic..., Hadziyannis [/bib_ref] [bib_ref] Peginterferon alfa-2a (40kDa) and ribavirin: comparable rates of sustained virological response in..., Reddy [/bib_ref]. The therapeutic efficacy of PEG IFN-α/RBV is likely due to their antiviral and immune-modulatory properties [bib_ref] Host factors and failure of interferon-a treatment in hepatitis C virus, Gao [/bib_ref]. Factors associated with SVR to interferon treatment have not been fully investigated in HCV genotype 4 infected patients. Previous studies demonstrated that age, body weight, baseline ALT, baseline HCV RNA viral load, HCV genotype and the level of fibrosis or cirrhosis are some predictors of response [bib_ref] Predicting antiviral treatment response in chronic hepatitis C: how accurate and how..., Lee [/bib_ref] [bib_ref] Improving outcome in patients with hepatitis C virus genotype 4, Kamal [/bib_ref] [bib_ref] Roudot-Thoraval F, Observational VHC4 Study Group: Epidemiological characteristics and response to peginterferon..., Roulot [/bib_ref] [bib_ref] Response to pegylated interferon alfa-2a and ribavirin in chronic hepatitis C genotype..., El Makhzangy [/bib_ref]. However, recent studies indicated that cytokines could be used as markers for disease progression in HCV infected patients [bib_ref] Cytokine profile in Egyptian HCV genotype-4 in relation to liver disease progression, Zekri [/bib_ref].
Some recent studies demonstrate that IL-8, and CXCL10 chemokines correlate with poor response to antiviral therapy in CHC patients [bib_ref] Cytokine profile in Egyptian HCV genotype-4 in relation to liver disease progression, Zekri [/bib_ref] [bib_ref] Chemokines in hepatitis C virus infection: pathogenesis, prognosis and therapeutics, Wald [/bib_ref] [bib_ref] Association of pretreatment serum interferon c inducible protein 10 levels with sustained..., Diago [/bib_ref] or have prognostic utility as a marker of treatment outcome [bib_ref] Evidence for an antagonistform of the chemokine CXCL10 in patients chronically infected..., Casrouge [/bib_ref] [bib_ref] Soluble inflammatory markers as predictors of virological response in patients with chronic..., Moura [/bib_ref] [bib_ref] The role of chemokines as inflammatory Mediators in Chronic Hepatitis C Virus..., Zeremski [/bib_ref]. Moreover, control of HCV infection may depend in part on chemokinemediated recruitment of specific T cells to the liver [bib_ref] UH: Chemokines in innate and adaptive host defense: basic chemokinese grammar for..., Rot [/bib_ref]. In this context, Moura et al., [bib_ref] Soluble inflammatory markers as predictors of virological response in patients with chronic..., Moura [/bib_ref] evaluated the association between pretreatment plasma levels of chemokines CCL2, CCL3, CCL11, CCL24, CXCL9, CXCL10 and soluble tumor necrosis factor receptors and the virological response in CHC-treated patients. They found that pretreatment CXCL10 level predicts EVR and SVR to IFN-α and ribavirin and may therefore be useful in the evaluation of candidates for therapy.
However, no previous studies have been done in Egypt yet to assess the possible role of chemokines levels as predictors of response to interferon therapy. The current study aims to assess 1) the changes in serum levels of CXCL5, CXCL9, CXCL11, CXCL12, CXCL 13, CXCL 16 chemokines and E-Cadherin in Egyptian patients with chronic HCV genotype 4 infection who were recruited for combined PEG IFN-α/RBV therapy at baseline and after 12 and 24 weeks, and 2) to determine the utility of using these markers as predictors of the treatment response through determining the influence of therapy on these chemokines.
# Results
The current study was conducted on 57 CHC genotype-4 patient s, of which 32 (56.1%) received pegylated interferon α 2b, and 25 (30.4%) received pegylated interferon α 2a. Clinical and laboratory findings of the studied patients are shown in [fig_ref] Table 1: Clinical and Laboratory findings of the studied 57 chronic hepatitis C patients [/fig_ref]. There was no significant difference between pegintron and pegasys patients regarding histopathological and laboratory findings (p > 0.05) except for the age (p=0.04).
Baseline mean values of HCV RNA Log (EQ/ML) levels for the patients who received pegylated interferon α2b (Peg-intron) and pegylated interferon α 2a (Pegasys) were 5.4431 Log (EQ/ML) and 5.3824 Log (EQ/ML); respectively.
Different baseline data that might affect achievement of sustained virological response (SVR) such as age, body mass index, AST, ALT, and viral load were analyzed for pegylated interferon α2b and α2a therapy. There was a statistically significant difference between responders and non-responders in the AFP value for patients who received pegylated interferon α 2b (p=0.041) and α 2a (p = 0.05) therapy. Also, there was a statistically significant difference between responders and non-responders in AST value for patients who received pegylated interferon α2b (p=0.049) and α2a (0.001) therapy. On the other hand, the other baseline data did not differ significantly between the two studied groups [fig_ref] Table 2: Baseline differences of Lab findings among responders and non-responders of Pegylated interferon... [/fig_ref].
Serum levels of the studied chemokines in response to pegylated interferon α2b and α2a therapy were determined before the beginning of treatment, after 12 and 24 weeks of treatment and correlated to the clinical data and response to treatment. However, full clinical and follow up data were available for 25 patients in group I (receiving Peg-intron) and 19 patients in group II (receiving pegasys treatment) only [fig_ref] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing... [/fig_ref]. Each group was divided into two arms (responders and non-responders). As shown in [fig_ref] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing... [/fig_ref] and [fig_ref] Figure 1: CXCL-5 levels in responders and non responders to a [/fig_ref] the serum pretreatment levels of CXCL5 (1.275 ± 0.969 pg/ml) and (1.337 ± 0.792 pg/ml) were significantly higher than those assessed either at 12 (0.711 ± 0.459 pg/ml) and (0.908 ± 0.588 pg/ml) or 24 weeks (0.722 ± 0.466 pg/ml) and (0.951 ± .571 pg/ml) of pegylated interferon α2b therapy for both non-responders and responders; respectively.
The same findings were observed in patients receiving pegylated interferon α2 a therapy, since CXCL5 pretreatment levels (1.220 ± 1.022 pg/ml) and (1.423 ± .925 pg/ ml) were significantly higher than those assessed either after 12 (1.004 ± .699 pg/ml) and (0.808 ± 0.609 pg/ml) or 24 weeks (0.766 ± 0.433 pg/ml) and (0.862 ± 0.634 pg/ ml) of non-responders and responders; respectively (pvalue = 0.0001). No significant difference was found in relation to treatment type, response to treatment, timetype interaction, time-response interaction, or typeresponse interaction.
CXCL9 levels decreased in responders to pegylated interferon α2b and α2a therapy after 12 and 24 weeks. Whereas in non-responders, CXCL9 levels decreased at 12 weeks of therapy, then it increased at 24 weeks. No significant difference was found for the effect of time, treatment type, response to treatment, time-type interaction, time-response interaction, or type-response interaction [fig_ref] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing... [/fig_ref] and [fig_ref] Figure 2: CXCL-9 levels in responders and non responders to a [/fig_ref].
Serum levels of CXCL11 were elevated in responders to pegylated interferon α2b therapy at 12 and 24 weeks, while it diminished at 12 weeks then increased at 24 weeks for both responders and non-responders to pegylated interferon α2a therapy. However, the difference was not statistically significant in relation to the effect of time, treatment type, response to treatment, time-type interaction, time-response interaction, or type-response interaction [fig_ref] Table 4: Different chemokine's levels of 55 patients at 12 weeks in response to... [/fig_ref] and [fig_ref] Figure 3: CXCL-11 levels in responders and non responders to a [/fig_ref].
CXCL12 levels showed slight increase in responders to peg-intron therapy at 12 weeks and decreased in nonresponders. In patients receiving pegasys therapy, it decreased after 12 and 24 weeks, but no significant difference was found for time change, treatment type, response to treatment, time-type interaction, time-response interaction, or type-response interaction [fig_ref] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing... [/fig_ref] and [fig_ref] Figure 4: CXCL-12 levels in responders and non responders to a [/fig_ref].
Serum pretreatment levels of CXCL13 (0.405 ± 0.274 pg/ ml and 0.298 ± 0.105 pg/ml) were significantly lower than those evaluated either at 12 (0.529 ± 0.271 pg/ml and 0.341 ± 0.145 pg/ml) or 24 weeks (0.628 ± 0.455 pg/ml and 0.399 ± 0.128 pg/ml) in pegylated interferon α2a treated patients for the non-responders and responders; respectively (p = 0.002). Similarly, CXCL13 pretreatment levels for responders of pegylated interferon α2b therapy (0.315 ±0.105 pg/ml) were significantly lower than those assessed either after 12 (0.433 ± 0.246 pg/ml) or 24 weeks (0.539 ±0.377 pg/ml). For the pegylated interferon α2a nonresponders, CXCL13 levels increased at 12 weeks from 0.340 ± 0.069 to 0.504 ± 0.139 pg/ml then, it decreased at 24 weeks to 0.473 ± 0.181 pg/ml [fig_ref] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing... [/fig_ref] and [fig_ref] Figure 5: CXCL-13 levels in responders and non responders to a [/fig_ref]. However, no significant difference was found in relation to treatment type, response to treatment, time type interaction, time-response interaction, or type-response interaction.
Pretreatment levels of CXCL16 for non-responders to pegylated interferon α2b and α2a therapy; respectively (0.946 ± 0.101 and 1.146 ± 0.424 ng/ml) and responders (0.874 ± 0.244 & 1.012 ± 0.331 ng/ml) were significantly lower than those after 12 weeks for non-responders [fig_ref] Figure 6: CXCL-16 levels in responders and non responders to a [/fig_ref]. However, no significant difference was found in relation to treatment type, response to treatment, timetype interaction, time-response interaction, or typeresponse interaction.
Serum levels of E-Cadherin increased observably in non-responders and responders after 12 weeks from the beginning of therapy (either Pegylated IFNα 2a or 2b groups) and continued to increase at 24 weeks but in responders only. No significant difference was found in relation to time change, treatment type, response to treatment, time-type interaction, timeresponse interaction, or type-response interaction (Table 3 and [fig_ref] Figure 7: E-cadherin levels in responders and non responders to a [/fig_ref]. No significant correlation was found between the expression levels of any of the studied chemokines and baseline HCV RNA in our studied patients.
## Correlations between different chemokines
Before treatment initiation CXCL13 was significantly correlated to CXCL9 (r = 0.385, p = 0.003), CXCL16 (r = 0.366, p = 0.048) and E-cadherin (r = 0.511, p = 0.000). While E-cadherin was significantly correlated to CXCL16 (r = 0.628, p = 0.000). Otherwise there was no significant correlation between the others chemokine's studied or between them and the baseline AST, ALT and albumin levels except for the significant correlation between Ecadherin and ALT level (r = 0.363, p = 0.048). At 12 weeks of therapy E-cadherin was significantly correlated to CXCL5 (r = 0.358, p = 0.047), CXCL9 (r = 0.311, p = 0.021), CXCL12 (r = 0.789, p = 0.000), CXCL13 (r = 0.583, p = 0.000), CXCL16 (r = 0.810, p = 0.000) and ALT levels (r = 0.311, p = 0.021).
CXCL5 was significantly correlated to CXCL9 (r = 0.359, p = 0.007), CXCL11(r = 0.456, p = 0.000), CXCL12 (r = 0.417, p = 0.002), CXCL13 (r = 0.448, p = 0.001) and CXCL16 (r = 0.326, p = 0.015). CXCL9 was significantly correlated to CXCL12 (r = 0.640, p = 0.000), CXCL13 (r = 0.422, p = 0.001) and CXCL16 (r = 0.519, p = 0.000). CX CL12 was significantly correlated to CXCL13 (r = 0.622, p = 0.000) and CXCL16 (r = 0.827, p = 0.000). CXCL13 was significantly correlated to CXCL16 (r = 0.531, p = 0.000).
At 24 weeks of therapy CXCL5 was significantly correlated to CXCL9 (r = 0.327, p = 0.027), CXCL11 (r = 0.587,
# Discussion
Previous studies have investigated pretreatment predictors of response to interferon based therapy in HCV-associated CHC patients. Some of these studies have shown that cytokines could be used as predictors of treatment outcome [bib_ref] Soluble inflammatory markers as predictors of virological response in patients with chronic..., Moura [/bib_ref]. In the current study, different baseline data that might affect the achievement of SVR and different chemokines levels were analyzed at different time intervals from starting therapy. We were able to show a statistically significant difference between responders and nonresponders to peg-intron and pegasys therapy in relation to AFP and AST values. Assessment of the correlation between the expression levels of these chemokine's and response to treatment at baseline, week 12, and 24 revealed that the baseline serum level of CXCL5 was significantly higher than those after 12 or 24 weeks from the beginning of pegylated interferon α2a and α2b therapy for responders and non-responders but with no significant correlation with response to treatment. This could be attributed to an interaction between immune response and viral activity, which is indicated by the differences in cytokines and chemokine's expression during the course of treatment [bib_ref] Th1 and Th2 cytokines are elevated in HCV-infected SVR(+) patients treated with..., Wan [/bib_ref]. On the other hand, Dou et al., [bib_ref] Gene expression profiles on three kinds of genotype hepatitis C virus core..., Dou [/bib_ref] demonstrated that the HCV-2a core protein was shown to down-regulated CXCL5 gene in human hepatoma (Huh-7) cell line.
Our results regarding the decrease in the CXCL9 level in responders to pegylated interferon α 2 b and α 2 a therapy after 12 and 24 weeks are consistent with those of Wan et al., [bib_ref] Th1 and Th2 cytokines are elevated in HCV-infected SVR(+) patients treated with..., Wan [/bib_ref] who demonstrated that serum CXCL9 levels decreased significantly after the initiation of IFN therapy indicating a possible effect on the rate of viral clearance. On the other hand Butera et al., [bib_ref] Plasma chemokine levels correlate with the outcome of antiviral therapy in patients..., Butera [/bib_ref] found that pretreatment samples showed no consistent differences in CXCL9 level between patients who subsequently failed therapy and those who subsequently developed asustained response to antiviral therapy. They also found that CXCL9 level declined during therapy in both responders and non-responders. Among sustained responders, CXCL9 level remained low after completion of therapy while in non-responders, the drop in CXCL9 was transient. A possible explanation for the difference between our results and the later study could be our smaller sample size and the fact that all our patients were genotype 4.
Our data regarding the elevated CXCL11 level in responders of peg-intron therapy at 12 and 24 weeks, followed by a decrease at 12 weeks then an increase at 24 weeks for both responders and non-responders of pegasys therapy with no significant difference regarding time effect, treatment type, and response to treatment contrast with those of Butera et al., [bib_ref] Plasma chemokine levels correlate with the outcome of antiviral therapy in patients..., Butera [/bib_ref] who found that CXCL11 level measurement was lower in those who subsequently achieved SVR than in those who subsequently have no reduction in HCV RNA during or after antiviral therapy but with no significant differences in response. Therefore, they concluded that there was no association between pretreatment levels of CXCL11 and the outcome of treatment. The increase in CXCL11 level in the current study may be related to an increase in its receptor (CXCR3) levels in CHC patients as previously mentioned by Perney et al., [bib_ref] CXCR3 expression on peripheral CD4+ T cells as a predictive marker of..., Perney [/bib_ref]. We also reported a slightly increase in CXCL12 level was responders to PEG-IFN α2b therapy at 12 weeks and a decrease in non-responders and in all patients receiving PEG-IFN α2a after 12 and 24 weeks. Cicinnati et al., [bib_ref] Altered chemotactic response of myeloid and plasmacytoid dendritic cells from patients with..., Cicinnati [/bib_ref] have previously mentioned that IFN-α can modulate the expression of G protein coupled chemokine receptors, and some adhesion molecules, which are required for dendritic cell (DC) trafficking into secondary lymphoid tissue. Accordingly, the effect of IFN on CXCL12 levels, in the current study, could be explained by IFN-α therapyinduced migration of DC from inflamed hepatic portal areas towards secondary lymphoid tissue with increased chemotaxis to CXCL11 and CXCL12. Our results regarding the significant reduction of pretreatment CXCL13 level was with time of interferon treatment for both responders and non-responders regardless of treatment type are consistent with Sansonno et al. [bib_ref] Increased serum levels of the chemokine CXCL13 and up-regulation of its gene..., Sansonno [/bib_ref] who mentioned that up-regulation of CXCL13 gene expression is a distinctive feature of HCV-infected patients. Therefore high levels of this chemokine could be detected in the liver and skin of patients with CHC and these levels might decrease after therapy. In our patients, pretreatment CXCL16 level increased significantly with time in responders and non-responders to pegylated interferon α2b and α2a therapy with no significant difference between the two groups. This could be attributed to the effect of IFN, which improves TH-1 response [bib_ref] Pretreatment expression of the perforin gene by circulating CD8(+) T lymphocytes predicts..., Balian [/bib_ref]. IFN-α treatment has been shown to regulate the activity of other cytokines and chemokines either directly or indirectly [bib_ref] Interferons alpha and beta as immune regulators-a new look, Biron [/bib_ref].
Alteration in E-Cadherin expression has been associated with human carcinogenesis including HCC [bib_ref] Cell adhesion system and human cancer morphogenesis, Hirohashi [/bib_ref]. In our series, E-Cadherin increased in non-responders and responders to Pegylated IFNα 2a or 2b after 12 weeks from the beginning of therapy however it continued to increase at 24 weeks only in responders of both groups. This could be attributed to the effect of HCV core protein, which represses E-Cadherin expression at the transcriptional level by down-regulating its promoter activity through promoter hypermethylation [bib_ref] Hepatitis C virus core protein downregulates E-cadherin expression via activation of DNA..., Arora [/bib_ref]. The significant difference in serum levels of E-Cadherin, reported in our study, at week 12 between responders and non-responders shows that it could be used as a predictor of EVR in IFNtreated CHC patients.
No significant correlation was found in the current study between any of the studied chemokines levels and baseline HCV RNA. On the other hand, baseline data have shown significant correlations between CXCL9 and CXCL13, CXCL13 and CXCL16, CXCL13 and E-cadherin, CXCL16 and E-cadherin, as well as between ALT level and E-cadherin that were maintained in follow up results.
In other words, these significant correlations were present at pretreatment, after 12 weeks, and after 24 weeks. In a previous study done by Butera et al., [bib_ref] Plasma chemokine levels correlate with the outcome of antiviral therapy in patients..., Butera [/bib_ref] stated that there was no correlations between CXCL9 and CXCL11 chemokines levels and viral load. On the other hand, Perney et al., [bib_ref] CXCR3 expression on peripheral CD4+ T cells as a predictive marker of..., Perney [/bib_ref] found a correlation between the expression of CXCR3 on peripheral blood CD4+ T cells and AST serum levels. Cicinnati et al., [bib_ref] Altered chemotactic response of myeloid and plasmacytoid dendritic cells from patients with..., Cicinnati [/bib_ref] also found that there were no significant correlations between systemic levels of CXCL12 and gender, age, grade of hepatic inflammation, stage of fibrosis, transaminases, bilirubin, HCV genotype or viral load. Also, Sansonno et al., [bib_ref] Increased serum levels of the chemokine CXCL13 and up-regulation of its gene..., Sansonno [/bib_ref] found that there was no correlation between CXCL13 concentrations and circulating viral load or ALT levels. One of the explanations for these findings is the lack of a direct effect of the virus on the sources of CXCL13 production.
# Conclusion
In conclusion, our results have shown that serum levels of CXCL13 and E-Cadherin have the potential to be used as serological markers at week 12 to the response of combined PEG IFN-α/RBV therapy. Also, significant changes with time have been found in the serum levels of CXCL5, CXCL13, and CXCL16 before pegylated interferon α 2 a and α 2 b therapy, and at weeks 12 and 24 with no significant difference in relation to treatment type, response to treatment, time-type interaction, time-response interaction, or type-response interaction. However, there were some limitations of our study, as example, this study should be done on a large scale in order to be validated. Also, it is better to apply one type of PEG-IFN preparations, and avoiding confounding was not an easy task, when assessing the association between tested chemokines and an outcome variable which leads to an overestimate or underestimate of the true association between exposure and outcome.
# Methods
## Patients
This prospective study included 57 patients with chronic HCV (CHC) liver disease. These patients constituted a part of a randomized controlled study including 200 patients receiving either pegylated interferon α 2b or pegylated interferon α 2a in combination with ribavirin during the period from September 2006 to March 2009. A written informed consent for immunological research, in which patients were assigned to give 10 mL of venous blood at each studied point during the course of therapy, blood was obtained from all patients before enrollment in the study. Only 57 patients accepted to share in the current study (32 received pegylated interferon α 2b and 25 received pegylated interferon α 2a, at baseline and after 12 and 24 weeks). All patients were treated in Al Qahira Al Fatimia Hospital (a district hospital at the middle of Cairo) as a part of the national program for combating viral hepatitis under supervision and sponsorship of The Ministry of Health. A written consent was obtained from all patients prior to enrollment in the study and the ethical committee of Al Qahira Al Fatimia Hospital and Ministry of Health approved the protocol, which was in accordance with the ethical guidelines of the Helsinki Declaration.
Selection criteria included male or female patients with CHC, aged 18 years or older with HBsAg negative, antinuclear antibody (ANA) <1:160, positive anti-HCV antibodies and HCV RNA by RT-PCR. All patients had white blood cells (WBCs) > 4000/mm3, neutrophil count >2000/ mm3, platelets >75 000/mm3, prothrombin time <2 seconds above the upper limit of normal (ULN), direct bilirubin 0.3 mg/dL or within 20% of ULN, albumin >3.5, alpha fetoprotein <100, serum creatinine within normal limit (WNL), fasting blood sugar 115 mg/dl or within 20% ULN and if diabetic Hb A1C < 8.5% with normal T3, T4 and TSH. On the other hand, patients with decompensated liver disease or those with chronic liver disease due to causes other than HCV e.g. HBV, alpha-1 antitrypsin deficiency, Wilson's disease, haemochromatosis, alcoholic liver disease or autoimmune disease were excluded from the study. Similarly, patients with hypersensitivity to interferon or ribavirin, pregnant or breast feeding females and those with any co-morbid conditions were excluded. All patients were subjected to full history taking and clinical assessment, routine laboratory work up including complete blood picture, liver biochemical profile, serum urea and creatinine, blood glucose, pregnancy test for married females, HBsAg, ANA, TSH, AFP and anti-schistosomal Ab titer. ECG was done for males > 40 years and females >50 years old. Ocular examination and abdominal ultrasound were also done.
Follow up and monitoring response to antiviral therapy was done on week 1, 2 and 4 after initiation of therapy then every 4 weeks. During each follow up, signs and symptoms of possible adverse effects due to the drugs were evaluated by routine laboratory studies. If treatment was continued in the presence of adverse effects, either dose adjustments were considered or the particular side effect was treated or monitored without lowering dosages depending on its severity.
Quantitative HCV RNA by Real time-polymerase chain reaction (RT-PCR) was performed after 12 weeks of therapy to determine Early Virological Response (EVR) upon which the decision to continue treatment after 12 weeks was taken, in addition to the patients' tolerance to drugs and laboratory profile. Qualitative HCV viral load was also done after 24 weeks of treatment and if negative, the patient continued on the treatment. It was also done once the course of treatment was completed to document the End of Treatment Response (ETR). Sustained Virological Response (SVR) was defined as the absence of detectable HCV RNA in serum at the 24th week after the end of treatment.
Quantitative real time PCR for determination of HCV viremia RNA was extracted from patients' blood using the QIAmp Viral RNA Mini Kit (QIAGEN, Santa Clarita, U. S.A) according to the manufacturer's instructions, then quantitative Real time-polymerase chain reaction (RT-PCR) was performed using previously standardized realtime RT-PCR protocol for HCV, supplied from applied biosystems (USA) and according to Zekri et al. [bib_ref] Consensus siRNA for inhibition of HCV genotype-4 replication, Zekri [/bib_ref].
## Hcv genotyping
HCV genotyping was done using the INNOLIPA-II technique as previously described by Zekri et al. [bib_ref] TRUGENE sequencing versus INNO-LiPA for sub-genotyping of HCV genotype-4, Zekri [/bib_ref].
## Monitoring patients' response to therapy
Responses to therapy in patients with HCV were determined according to the consensus guidelines of the National Institutes of Health. Responders to therapy were defined by normalization of serum alanine aminotransferase (ALT) and absence of detectable serum HCV RNA at the end of treatment (48 weeks). Relapsed responders to therapy were defined by normalization of serum ALT and absence of detectable serum HCV RNA at the end of treatment but with an increase of the serum ALT and the presence of HCV RNA at follow-up (72 weeks). Non-responders were defined by elevated serum ALT and the presence of HCV RNA at the end of treatment.
## Assay of chemokines
Ten ml of blood were obtained from patients for serum separation before the start of therapy, after 12 and 24 weeks of therapy. CXCL5, CXCL9, CXCL11, CXCL12, CXCL 13, CXCL 16 chemokines and E-Cadherin were
## Histopathological examination
Pretreatment liver biopsies from all patients were examined to confirm the diagnosis and determine the stage of fibrosis and the grade of inflammation using Modified Knodell score [bib_ref] Formulation and application of numerical scaring system for assessing histological activity in..., Knodell [/bib_ref] , in the form of 18 points for histopathology activity index (HAI), six points for stage of fibrosis and four points for stage of steatosis [bib_ref] Classification of chronic hepatitis: diagnosis, grading and staging, Desmet [/bib_ref].
The histological activity index in the studied patients were graded into minimal (0-3), mild (4-8), moderate [bib_ref] Roudot-Thoraval F, Observational VHC4 Study Group: Epidemiological characteristics and response to peginterferon..., Roulot [/bib_ref] [bib_ref] Response to pegylated interferon alfa-2a and ribavirin in chronic hepatitis C genotype..., El Makhzangy [/bib_ref] and severe [bib_ref] Cytokine profile in Egyptian HCV genotype-4 in relation to liver disease progression, Zekri [/bib_ref] [bib_ref] Chemokines in hepatitis C virus infection: pathogenesis, prognosis and therapeutics, Wald [/bib_ref] [bib_ref] Association of pretreatment serum interferon c inducible protein 10 levels with sustained..., Diago [/bib_ref] [bib_ref] Evidence for an antagonistform of the chemokine CXCL10 in patients chronically infected..., Casrouge [/bib_ref] [bib_ref] Soluble inflammatory markers as predictors of virological response in patients with chronic..., Moura [/bib_ref] [bib_ref] The role of chemokines as inflammatory Mediators in Chronic Hepatitis C Virus..., Zeremski [/bib_ref] points. The stage of fibrosis was graded into minimal (0-1), moderate (2-4) and marked (5-6) points.
# Statistical methods
Statistical analysis was done using SPSS version 17 statistical software package for windows. Quantitative variables were expressed by mean and SD (Standard deviation), comparison of means using Student's t-test, Mann-Whitney U-test or ANOVA test were done when appropriate. Time-type interaction, time-response interaction, or typeresponse interaction were measured and analyzed statistically by repeated measures analysis of variance. In the table of multivariate tests, we took into consideration the p-values (for Wilks Lamda), and partial Eta square (according to Cohen's standard) for measuring the interactions and their effect size. MANOVA was used under the same circumstances as ANOVA but when there were multiple dependent variables as well as independent variables within the model which the researcher wishes to test. MANOVA is also considered a valid alternative to the repeated measures ANOVA when sphericity was violated. Also, Pearson correlation test was performed for correlating quantitative variables. Qualitative variables were expressed by frequency and percent, comparison of mean values and assessment of relations using chisquared test or Fisher's exact test was done when appropriate. P-value was considered significant when < 0.05 significant and highly significant when < 0.01.
## Competing interests
The authors declare that they have no competing interests.
Authors' contributions ARNZ generated the idea, helped in revision of the manuscript, was responsible for the whole practical part, AAB managed the pathology work, helped in revision of the manuscript, WSM was help in editing, revising and did the statistical part of the manuscript, HMAED carried out the ELISA for the specified chemokines, shared in editing and revising the manuscript, HS was responsible for patient's treatment and sampling wrote the draft of the manuscript, NZ managed the patients and their clinical part, DHE was help in editing, revising the manuscript and AO managed the patients and their clinical part, shared in the idea generation. All authors read and approved the final manuscript.
## Financial competing interests
The contents are solely the responsibility of the authors and do not necessarily represent the views of the funding source. The authors declare that they have no competing interests.
[fig] Figure 1: CXCL-5 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[fig] Figure 2: CXCL-9 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[fig] Figure 3: CXCL-11 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[fig] Figure 4: CXCL-12 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[fig] Figure 5: CXCL-13 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[fig] Figure 6: CXCL-16 levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. assayed using the quantitative ELISA plate method. The kits used were provided by Quantikine kits (R&D Systems, Inc. 614 McKinley Place NE Minneapolis, MN 55413 United States of America), sponsored by NCI and used according to the manufacturer's instructions. [/fig]
[fig] Figure 7: E-cadherin levels in responders and non responders to a) PEG-Interferone and, b) pegasys at different time intervals. [/fig]
[table] Table 1: Clinical and Laboratory findings of the studied 57 chronic hepatitis C patients [/table]
[table] Table 3: Different chemokine's levels of 44 patients in response to treatment p-value comparing the effect of time a , treatment type b , response to treatment c , time-type interaction d , time-response interaction e , type-response interaction f . [/table]
[table] Table 2: Baseline differences of Lab findings among responders and non-responders of Pegylated interferon α 2b and α 2a therapy [/table]
[table] Table 4: Different chemokine's levels of 55 patients at 12 weeks in response to treatment p-value comparing the effect of time a , treatment type b , response to treatment c , time-type interaction d , time-response interaction e , type-response interaction f . [/table]
|
Ethical Concerns of and Risk Mitigation Strategies for Crowdsourcing Contests and Innovation Challenges: Scoping Review
Background: Crowdsourcing contests (also called innovation challenges, innovation contests, and inducement prize contests) can be used to solicit multisectoral feedback on health programs and design public health campaigns. They consist of organizing a steering committee, soliciting contributions, engaging the community, judging contributions, recognizing a subset of contributors, and sharing with the community.Objective: This scoping review describes crowdsourcing contests by stage, examines ethical problems at each stage, and proposes potential ways of mitigating risk.Methods:Our analysis was anchored in the specific example of a crowdsourcing contest that our team organized to solicit videos promoting condom use in China. The purpose of this contest was to create compelling 1-min videos to promote condom use. We used a scoping review to examine the existing ethical literature on crowdsourcing to help identify and frame ethical concerns at each stage.Results:Crowdsourcing has a group of individuals solve a problem and then share the solution with the public. Crowdsourcing contests provide an opportunity for community engagement at each stage: organizing, soliciting, promoting, judging, recognizing, and sharing. Crowdsourcing poses several ethical concerns: organizing-potential for excluding community voices; soliciting-potential for overly narrow participation; promoting-potential for divulging confidential information; judging-potential for biased evaluation; recognizing-potential for insufficient recognition of the finalist; and sharing-potential for the solution to not be implemented or widely disseminated.Conclusions:Crowdsourcing contests can be effective and engaging public health tools but also introduce potential ethical problems. We present methods for the responsible conduct of crowdsourcing contests.Res 2018;20(3):e75)(J Med Internet
# Introduction
Crowdsourcing refers to "the practice of obtaining information or services by soliciting input from a large number of people, typically via the internet and often without offering compensation." [1] The term encompasses a wide range of practices that were originally developed to iteratively improve commercial products based on crowd input and to change the traditional relationship between a business and a client [bib_ref] Crowdsourcing applications for public health, Brabham [/bib_ref]. For example, the online encyclopedia Wikipedia allows anonymous volunteers to write, edit, and manage online encyclopedia entries. Wikipedia has rapidly grown and now has 4.9 million articles that are being edited by 70,000 active contributors. Crowdsourcing is used in the government and nonprofit sectors to generate innovative concepts and designs [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref]. Crowdsourcing can take a wide variety of forms, including online games [bib_ref] Predicting protein structures with a multiplayer online game, Cooper [/bib_ref] , distributed health system platforms [bib_ref] Mobile-phone dispatch of laypersons for CPR in out-of-hospital cardiac arrest, Ringh [/bib_ref] , and contests to solicit new ideas [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref].
Our discussion of crowdsourcing will focus on contests, also called innovation challenges, innovation contests, and inducement prize contests. Crowdsourcing contests include prize-based open contests in which individuals or teams work alone and those in which individuals work together. Contests typically include the following stages: organizing a steering committee, soliciting contributions, promoting the contest, judging contributions by experts or the crowd, recognizing excellent contributions, and sharing contributions. In the past 10 years, contests have been used to promote public health [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref]. Crowdsourcing contests have been used to develop health messages [bib_ref] Creative contributory contests to spur innovation in sexual health: 2 cases and..., Zhang [/bib_ref] , inform health policy [bib_ref] CrowdOutAIDS: crowdsourcing youth perspectives for action, Hildebrand [/bib_ref] , and improve medical diagnostics [bib_ref] Collective intelligence meets medical decision-making: the collective outperforms the best radiologist, Wolf [/bib_ref]. These kinds of contests can increase community engagement [bib_ref] Creative contributory contests to spur innovation in sexual health: 2 cases and..., Zhang [/bib_ref] [bib_ref] Hidden in plain sight: a crowdsourced public art contest to make automated..., Merchant [/bib_ref] , improve health [bib_ref] Hidden in plain sight: a crowdsourced public art contest to make automated..., Merchant [/bib_ref] [bib_ref] Crowdsourcing HIV testing: a pragmatic, non-inferiority randomized controlled trial in China, Tang [/bib_ref] , and save money [bib_ref] Crowdsourcing HIV testing: a pragmatic, non-inferiority randomized controlled trial in China, Tang [/bib_ref].
However, crowdsourcing contests introduce a number of potential ethical concerns, including not being sufficiently inclusive, only relying on the internet, and not disseminating the solution widely. Identifying and responding to these shortcomings is important for establishing crowdsourcing as a force for the public good and as a useful public health tool. These concerns have received limited attention in the public health literature on crowdsourcing to date [bib_ref] Crowdsourcing applications for public health, Brabham [/bib_ref] [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref]. This paper describes crowdsourcing contests by stage, describes common ethical challenges, and provides guidance on implementing crowdsourcing contests ethically.
# Methods
We conducted a scoping review [bib_ref] Scoping studies: advancing the methodology, Levac [/bib_ref] to synthesize literature on the ethical conduct of health-related crowdsourcing projects. This review includes applied and theoretical ethics literature related to crowdsourcing research and practice. Scoping reviews allow one to examine the literature in a structured way but are different from systematic reviews in their methodology and content. Our review focused on sources between January 1, 2005 and July 1, 2017. We examined a wide range of anthropological, ethical, social science, and related literature on crowdsourcing to promote public health. We anchored this discussion in a particular example of a single crowdsourcing contest. In addition, we examined several crowdsourcing contest failures to understand concerns and potential ethical problems.
We identified studies using keyword searches in electronic databases, including MEDLINE (OVID interface, 1946 onwards), Google Scholar, expert opinion, and Wikipedia. For database searches, we used phrases and synonymous variations of the following terms: crowdsourcing, innovation challenge, ethics, implementation ethics, and applied ethical analysis. We also identified studies based on searches of reference lists, hand-searching key journals identified from initial database inquiries, and unpublished conference abstracts. We prioritized studies that examined crowdsourcing contests in health contexts. Our search included studies that provided empirical or theoretical data on crowdsourcing contests in the past 12 years.
# Results
## Overview
Our scoping review data are organized according to the 6 stages of a crowdsourcing contest-organizing, soliciting, promoting, judging, recognizing, and sharing [bib_ref] Creative contributory contests to spur innovation in sexual health: 2 cases and..., Zhang [/bib_ref] [bib_ref] Hidden in plain sight: a crowdsourced public art contest to make automated..., Merchant [/bib_ref]. First, the contest organizers form a contest steering committee to articulate the purpose, values, and methods of the contest. Second, an open call for content (eg, concepts, images, videos, or other materials) is announced via in-person events and social media. This open call clarifies the goals and terms of the contest, the prize or incentive structure, and the nature of participation. The open call plays a key role in defining the crowd. Third, the crowd is iteratively engaged through feedback sessions, in-person events, and social media. Fourth, a group of judges evaluates each contribution based on prespecified criteria to determine finalists. In some cases, the judges are the crowd itself. Finalists and others are awarded prizes according to their rank order. The judging process aggregates crowd wisdom. Fifth, contest finalists are announced and recognized through an incentive structure. Sixth, the steering committee shares the finalist solution(s) with the community. After discussing each of these 6 stages, we review the literature on failures in crowdsourcing and discuss ethical principles of crowdsourcing contests.
## Organizing a steering committee
The first step of a crowdsourcing contest is to establish a steering committee that will decide the structure and function of the contest. The steering committee powerfully shapes the contest and provides a set of norms, expectations, and deadlines. Often contests are divided into ones that focus on engaging large numbers of the community or on resulting in a high-quality outcome [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref]. The condom video contest in China was focused on creating a high-quality video. The condom video contest steering committee was composed of youth, community health leaders, men who have sex with men, doctors, business leaders, and researchers. The group was organized by Sesh Global, an organization with experience in crowdsourcing contests. The steering committee met on a monthly basis to discuss the scope, rules, and promotion of the contest. In addition, the steering committee used email and social media to discuss contest developments.
One potential ethical problem with organizing a steering committee is the possibility of excluding community members or voices that are important to the contest. Often individuals from marginalized, vulnerable groups who lack a voice in decision making are less likely to be represented on steering committees. This problem has important implications for contests because the steering committee establishes the expectations and rules governing the entire process. For example, a contest focused on gay men and HIV ought to include gay men and people living with HIV. A committee lacking appropriate representation of key groups could undermine both effectiveness and trust in the contest.
One way to mitigate the risk of excluding important community voices is to have transparent criteria for selecting steering committee members. In addition, aligning the composition of the steering committee with the overall purpose of the contest could help ensure that community voices are represented. Given local power dynamics related to nonexpert advice, it may also be useful to have local, in-person meetings of the steering committee specifically to establish trust and align expectations.
# Soliciting contributions
Contest organizers design the open call soliciting contributions. The call for contributions is open so that anyone can contribute. Open calls can be through social media, in-person, or both. We define social media as websites or apps that allow users to create and share content or to engage in social networking . Our condom video contest call for entries in China shows how the language, format, and structure shape a crowd (Multimedia Appendix 1). The call for entries was distributed through social media and in-person events at local high schools, colleges, and community-based organizations. The choice of distribution channels encouraged young Chinese individuals to participate but allowed entries from anyone.
One potential problem with calls for entries is over-reliance on social media announcements and insufficient attention to in-person events. Most private sector contests have focused on using social media calls to solicit entries, including several calls exclusively through social media [bib_ref] May the best analyst win, Carpenter [/bib_ref] [bib_ref] Brokering knowledge in biosciences with InnoCentive. Interview by Semahat S. Demir, Hussein [/bib_ref]. There are two ethical problems with exclusively social media open calls. First, there is still a substantial digital divide between those who use social media and those who do not. Individuals who use social media tend to be from more developed regions or sectors and have higher socioeconomic status compared with those who do not use social media. An exclusively social media call would not only constrain participation among some vulnerable groups, but it could worsen some of the entrenched social inequities. Second, among those who use social media, there is a further barrier to engaging sophisticated contest platforms, such as Ideascale, a company that creates online platforms for crowdsourcing contests. A wide range of these platforms have been developed to crowdsource tasks. However, individuals who have the skills, knowledge, and experience to participate in these online platforms are a subset of the crowd, skewing its composition and unfairly excluding those without these skills but who are interested and could meaningfully contribute.
Careful attention to soliciting contributions can help to deal with these problems. In-person contest promotion events are one mechanism to broaden access to contests and diversify the crowd. These events have been used in several health contests [bib_ref] Creative contributory contests to spur innovation in sexual health: 2 cases and..., Zhang [/bib_ref] and have been found to increase participation and quality of participant entries. In-person events could take the form of classroom didactics, interactive feedback sessions, or community-led events. Capacity building sessions [bib_ref] Promoting older adults' well-being through internet training and use, Shapira [/bib_ref] could help individuals to learn about contributing on social media platforms. In addition to in-person contest promotion, ensuring multiple channels for contributing would be useful. This could include providing contributions through mail, in-person, or short text message. Contests should be as inclusive as possible relative to the intended audience. It is important to note that the goal is not for universal participation but to provide an opportunity to participate to those who would have a reasonable expectation of contributing to the contest.
## Promoting crowd engagement and contributions
Following the open call, there is an iterative process of engagement between organizers and potential participants. Our condom contest organized in-person and social media engagement activities to promote submissions. This included integration of in-person and social media activities so that they complemented each other. Approximately three-quarters of those who submitted to the contest participated in at least one engagement activity. These activities established trust in the contest, built confidence in contributing, and established social norms about how to participate in the contest. Engagement activities avoided giving examples in order to decrease cognitive fixation and increase innovation [bib_ref] Constraining effects of examples in a creative generation task, Smith [/bib_ref] [bib_ref] How examples may (and may not) constrain creativity, Marsh [/bib_ref]. However, this stage of crowdsourcing contests also raises potential ethical concerns. Disclosure of confidential information by contributors and the possibility of social media trolling are two primary concerns.
Authentic engagement in a contest allows those who contribute to draw on their own unique talents, preferences, and local social context. However, this personal process introduces the risk of private information being divulged, often unintentionally, as part of engagement and contributing. These concerns have been raised more generally in the crowdsourcing literature [bib_ref] Is there PAPA in crowd work? A literature review on ethical dimensions..., Durward [/bib_ref]. In our case, some condom videos included identifiable individuals. The contest organizers had clear guidelines establishing that all videos could be publically viewed and any individual who participated in the video gave permission to be included. This decreased the risk of unintended disclosure associated with viewing the videos. In addition, contest organizers may consider having more stringent requirements about obtaining written consent for an individual's photograph or other personal information to be included in the contribution or going back to finalists to confirm consent before video or other forms of dissemination.
In addition, social media trolling has been reported within crowdsourcing contests. The word troll comes from the Scandinavian mythology, referring to evil small creatures who disturb travelers . Today the term "trolling" refers to individuals who (usually anonymously) harass, provoke, or insult others online [bib_ref] Internet defamation as profit center: the monetization of online harassment, Bartow [/bib_ref]. Trolling has been reported in a range of social media contexts [bib_ref] VA: Association for the Advancement of Computing in Education, Shin [/bib_ref] [bib_ref] Representations of 'trolls' in mass media communication: a review of media-texts and..., Bishop [/bib_ref] , including contests. Although trolling may be largely protected in some countries by the right to free speech, organizers of crowdsourcing contests should make efforts to anticipate these harms and provide protection. This type of ethical concern can to some extent be addressed through online platform moderation and algorithms for detecting offensive words, as well as informing participants of potential risks of trolling during the consent process.
## Aggregating crowd wisdom: judging contributions
Once the crowd has engaged in the contest and submitted contributions, these contributions are judged. The condom contest videos were evaluated by a multisectoral group of local people living with HIV, youth, physicians, and public health experts. Local judging increased community ownership of the contest and increased the likelihood that local characteristics (eg, using the local dialect) would be incorporated. At the same time, the nature of judging brings up a range of potential ethical issues, including how to fairly select judges. One common approach to judging contests has been to involve the crowd in judging entries, cited as a cost-effective way to engage potential participants. However, if the crowd is exclusively defined online, it would be prone to the same problems described above in addition to bias, inconsistent judging criteria, and favoring popular opinion. Crowd evaluation may also lead to voting based on criteria not consistent with the goals of the contest. For example, the British contest to name a government research vessel resulted in the entry "Boaty McBoatface" receiving 124,109 votes, more than fourfold greater than the next entry. Organizers found themselves in the dilemma of accepting an absurd name or rejecting a crowdsourced outcome. They eventually compromised by using "Boaty McBoatface" to name a submersible carried by the research vessel dubbed the Sir David Attenborough. In addition, the contest contributor with the largest number of online followers may be more likely to receive votes in support of their contest entry. Empirical evidence from private sector contests confirms that online crowd evaluation is biased toward individuals with greater social networks compared with expert judge evaluation. Two studies found that individuals who win crowd-judged prizes are not as likely to sustain their engagement over time compared with individuals who win expert-judged prizes [bib_ref] Crowdsourcing and knowledge sharing: strategic user behavior on taskscn, Yang [/bib_ref].
When a crowdsourcing contest has a relatively low number of entries (<100), a panel of expert judges could evaluate contributions. Judges would need to be selected in a fair way that is consistent with the mission and goals of the overall contest. Elements from "fair process" procedures-which emphasize transparency, justification of rationales, opportunities to appeal decisions, and so on-could help in the constitution of the judging panel and also help guide the decisions they make. This could include evaluation of the judge panel to help ensure that a broad range of judges are represented and decrease reliance on social media. Several private sector contests demonstrate the feasibility of having a judging panel evaluate contributions.
## Celebrating crowd wisdom: recognizing contributions
Following the judging process, contributions can be recognized by prizes or incentives, acknowledgment, and retention of legal rights to products created. Incentive structures for crowdsourcing vary based on the goals and missions of the contest. Some contests have a single large prize [bib_ref] The optimal allocation of prizes in contests, Moldovanu [/bib_ref] while others recognize a number of contributions [bib_ref] Multiple-prize contests -the optimal allocation of prizes, Sisak [/bib_ref]. The condom contest included individual prizes for the top three contributors as well as participation prizes. The top contributors were announced on social media as a further form of recognition. All contributors retained the rights to their videos (those who submitted the videos could use them for any purposes), consistent with the goal of the contest to promote community agency.
An ethical challenge related to recognition in crowdsourcing contests is the potential for exploiting those who make substantial contributions. Insufficient recognition of those who contribute to contests has been noted in many online contest settings [bib_ref] Who are the crowdworkers? Shifting demographics in mechanical Turk, Ross [/bib_ref] [bib_ref] The future of crowd work, Kittur [/bib_ref] [bib_ref] Internet-based crowdsourcing and research ethics: the case for IRB review, Graber [/bib_ref]. The condom contest decreased the likelihood of this exploitation because there were several formal and informal ways of recognizing participants, alongside retention of their legal rights. The contest also shared the finalist video online in several forums.
Appropriately recognizing contributions provides a way of addressing these concerns about crowd exploitation related to the incentive structure and acknowledgment. First, clearly stating during the consent process how contributions will be recognized can mitigate exploitation to some extent. Second, incentive structures with multiple prizes (of different types) promote a broad spectrum of participation. Including special prize categories that focus on participation rather than merit have been used in some contests [bib_ref] Creative contributory contests to spur innovation in sexual health: 2 cases and..., Zhang [/bib_ref]. Third, formally acknowledging and celebrating contributions are important. Several studies have shown that intrinsic benefits of participation (such as recognition and media attention) are more important than extrinsic benefits in the context of crowdsourcing [bib_ref] Hidden in plain sight: a crowdsourced public art contest to make automated..., Merchant [/bib_ref]. Governance of ownership and permissible uses of finalist contributions may also minimize exploitation.
## Sharing and implementing the solution with the community
The final stage of a crowdsourcing contest is to share the solution more widely with the community that contributed. Henk van Ess has argued that crowdsourcing must give back to the public and share the solution more widely. In this way, crowdsourcing reciprocates in a commensurate way to what the community contributed. Other crowdsourced research has suggested that perceptions of fairness are important for those contributing to crowdsourcing projects [bib_ref] The Impact of Fairness on the Performance of Crowdsourcing: An Empirical Analysis..., Mazzola [/bib_ref]. Our condom video contest provided prizes to finalists, participation prizes, and then made the videos available on public platforms in China.
Limited sharing of the crowdsourced solution presents an important ethical concern associated with crowdsourcing contests. This also differentiates public-oriented contests from their private sector counterparts. Most private sector contests see the finalist solutions as their own intellectual property; the terms of many private contests give intellectual property rights to organizers. Limited sharing could take the form of only describing contests in articles that are inaccessible to nonsubscribers behind a paywall.
Clearly establishing a plan early in the process for sharing and prizes as part of the call for entries can mitigate the risk of insufficient sharing. While having clear sharing expectations is important, there should also be sufficient flexibility to give the steering committee ultimate authority in making final decisions. For example, the condom contest mandated sharing of the final video on regional, national, and international networks. The call for entries specified this plan, in addition to a plan to recognize excellent entries. The benchmark for "excellent" was decided by the steering committee.
## Learning from failures
Previous examples of crowdsourcing contests that were partially or incompletely effective provide guidance [fig_ref] Table 1: Implementation ethics issues and potential solutions associated with crowdsourcing contests [/fig_ref]. In 2013, the condom manufacturing company Durex invited the public to vote on which city in the world should receive a special condom rush delivery service. By the end of the contest, the nonexistent city of "Batman" in Turkey had received the most votes, and contest organizers were faced with the ethical dilemma of blatantly rejecting the clear will of the crowd or endorsing a deliberately facetious winning entry. Such instances of crowd hijacking are not uncommon [bib_ref] Crowdsourcing in a time of empowered stakeholders: lessons from crowdsourcing campaigns, Wilson [/bib_ref] , and contest organizers should prepare for scenarios where the crowd may use the contest platform to advance an agenda that deviates from that of the contest organizers'. Given the unpredictability of the crowd, it is important for organizers to clearly explain their rights to prospective participants, including the right to deem certain kinds of entries inadmissible.
Past contests have also shown that successfully developing products through crowdsourcing is not formulaic, and that breakthrough innovations are by no means guaranteed. The start-up company Quirky had managed to secure hundreds of millions of investment dollars to develop innovative household consumer products through open online contests. However, despite deep financial resources and hundreds of thousands of contest contributors, Quirky failed to produce any radically innovative products and eventually declared bankruptcy 6 years after its founding. One of the major problems was that Quirky innovators had disagreements with the company in the late stages of business development. This miscommunication could be avoided by involving community members earlier in the process of development.
Finally, a German contest solicited public input on a ban on circumcision. A political party decided to crowdsource local opinions on the topic, targeting the area of North Rhine-Westphalia. Despite having a population of 18 million individuals, the contest only received 20 submissions. This underscores the importance of having a steering committee that plans in advance and understands community interests and willingness to take part in crowdsourcing contests.
## Ethical principles in crowdsourcing contests
We identified several general reviews that broadly considered ethical principles associated with crowdsourcing contests [bib_ref] The ethical use of crowdsourcing, Standing [/bib_ref]. These highlighted theoretical concerns about privacy, accuracy of information, property, and accessibility in the context of computer science. However, this limited literature did not focus on health contests. In-person events to promote contests; multiple ways of receiving contributions Online contests limit participation to a subset of internet-using individuals 2 Soliciting
Allow contributions via email, in-person, cell phones, and other forms that do not require online access or social media Social networking sites narrow participation in contests to a subset of social media-savvy individuals Clear contest guidelines that clarify whose permission has been obtained and potentially enhanced consent process before dissemination Public contributions may include confidential or private information 3 Promoting
## Social media moderators and algorithms for detection of explicit language
Social media platforms for contributing may introduce opportunities for online harassment In-person prize announcements Online contests may not sufficiently recognize contributions Establish a formal mechanism to share or implement the solution more widely with the local community More is taken from the community than given back 6 Sharing
# Discussion
## Principal findings
Our review suggests that crowdsourcing contests for public health introduce several potential ethical concerns. These concerns can be categorized within the 6 stages of crowdsourcing contests-organizing, soliciting, promoting, judging, recognizing, and sharing. Our analysis suggests that these concerns can be minimized with appropriate planning and consultation. Our review expands the limited literature on crowdsourcing contests for public health [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref] by focusing on ethics, examining an empirical case, and including a formal scoping review.
Our data suggest that several ethical concerns associated with crowdsourcing contests can be anticipated and avoided. For example, the awkward situation of having crowds decide on a trivial name such as Boaty McBoatface can be avoided by separating the process of soliciting names and choosing names eligible to be voted on. Other risks can be mitigated through appropriate contest planning. For example, inviting a diversity of local steering group members can increase the likelihood of local community perspective representation. Other ethical concerns are related to the use of social media within contests (eg, privacy concerns) and have been well described elsewhere [bib_ref] Attitudes toward the ethics of research using social media: a systematic review, Golder [/bib_ref]. All of these types of ethical concerns underscore the need for organizers of crowdsourcing contests to include sufficient time for planning and designing a contest.
Our scoping review did not identify studies that articulated ethical principles of crowdsourcing contests for health. This may be because few studies have focused on using crowdsourcing contests to improve public health [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref]. This also may be related to the breadth of diversity of crowdsourcing contests activities, including research studies, community engagement programs, and communications strategies. However, our analysis suggests that there are several shared contest stages each of which has potential ethical concerns.
Crowdsourcing contests have implications for public health research and policy. In terms of research, further empirical study on ethical problems associated with crowdsourcing contests is necessary. Such research could help to refine the method and increase the likelihood of crowdsourcing contests achieving their goals. This research could include qualitative studies of those participating and organizing crowdsourcing contests. Such research would provide valuable input for a future ethical framework specific to crowdsourcing. In terms of policy, crowdsourcing contests could help to inform public health policy. The multisectoral, transparent, and open nature of contests establishes a strong foundation for policy making. For example, the World Health Organization (WHO) used a crowdsourcing contest [bib_ref] The HepTestContest: a global innovation contest to identify approaches to hepatitis B..., Tucker [/bib_ref] to solicit descriptions of hepatitis testing that were directly included in the 2017 WHO hepatitis testing guidelines. Given the potential for contests to inform policy, more formal principles and ethical considerations associated with crowdsourcing contests may be useful.
# Limitations
Our analysis has several limitations. First, most health contests have been single events and have yet to be serialized and formally incorporated into routine public health practice. Serial contests are likely to have different ethical challenges and may be substantially different in terms of implementation. Second, health-focused crowdsourcing contests are relatively new and while there are many examples of health-related contests, few are formally evaluated using validated metrics. Our analysis focused on a single example. Further implementation research is needed to define the most efficient and responsible use of crowdsourcing contests. Third, there is variation in the extent to which contests are driven by crowd input. Some health projects involve the community at all stages, while others have more intensive community input only at the start of the project [bib_ref] Systematic review of innovation design contests for health: spurring innovation and mass..., Pan [/bib_ref].
# Conclusions
Crowdsourcing contests may be a useful tool to develop inclusive public health programs but also pose ethical concerns at each stage. Unraveling these ethical concerns requires careful planning, consideration, and consultation. Our analysis provides several practical steps for the responsible conduct of crowdsourcing contests and identifies areas for future research.
[table] Table 1: Implementation ethics issues and potential solutions associated with crowdsourcing contests. [/table]
|
Unusual tubular cystic lesion compressing inferior vena cava
Fluid collections are important local complications of acute pancreatitis (AP). They are usually located in peri-pancreatic region but can also be found in various atypical locations like mediastinum, spleen, liver, pelvis or neck. Such pancreatic fluid collections in atypical locations can cause diagnostic dilemma and may cause delay in management. We report an interesting case of AP who presented to us with a walled-off necrosis mimicking a long cystic mass lesion surrounding the IVC.DiscussionPancreatic fluid collections at an atypical location pose a special challenge to the diagnosis as well as treatment for the treating clinician. Due to close proximity of pancreas with various
# Introduction
Acute pancreatitis (AP), especially acute necrotizing pancreatitis, is a life-threatening condition with high morbidity and mortality. Peripancreatic fluid collections are one of the most important local complications of AP. [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref] Pancreatic pseudocyst is a collection of pancreatic secretions enclosed by a well-formed fibrous wall. [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref] Although such pancreatic collections are most commonly found in the peripancreatic region, they can also be found in various atypical locations, such as the mediastinum, spleen, liver, pelvis, or neck. [bib_ref] Endoscopic management of pancreatic pseudocysts at atypical locations, Bhasin [/bib_ref] [bib_ref] Successful resolution of a mediastinal pseudocyst and pancreatic pleural effusion by endoscopic..., Bhasin [/bib_ref] Such pancreatic fluid collections in atypical locations can cause diagnostic dilemma and may cause delay in management. A total of 40-60% of pancreatic fluid collection resolves without any need of intervention. [bib_ref] Conservative treatment as an option in the management of pancreatic pseudocyst, Cheruvu [/bib_ref] However, symptomatic fluid collections require drainage. Here, we report an interesting case that presented to us with a long cystic mass lesion surrounding the inferior vena cava (IVC) after an episode of acute abdominal pain.
## Case report
A 32-year-old male presented elsewhere with abdominal pain following an alcohol binge. He was diagnosed with acute necrotizing pancreatitis based on elevated serum amylase and lipase and imaging findings. He was treated with conservative management at that center. After an asymptomatic period of 1 week, he had recurrence of abdominal pain and was referred to us for further management. Contrast-enhanced computed tomography of the abdomen demonstrated a long tubular cystic lesion measuring 7 × 3 cm around the IVC; arrows). Endoscopic ultrasound (EUS) indicated a focal area of altered echotexture in the uncinate process of pancreas; open arrows), suggesting focal pancreatitis. The inflammatory process was seen extending into the adjacent retroperitoneum; bold arrows), leading to the formation of a cystic lesion compressing the IVC. This cystic lesion had a variable amount of solid component; bold arrows), suggestive of walled-off necrosis (WON). A small duodenal cystic lesion was also noted. Because of ongoing pain, nasojejunal (NJ) tube feeding was initiated. Post-NJ feeding, the patient had marked improvement in symptoms and was pain free 2 weeks later. EUS demonstrated resolution of peri-IVC WON. Oral feeding was gradually initiated, and the patient was pain free after 8 weeks of follow up.
visceral organs and vascular structures, pancreatic fluid collections can cause various complications by either causing compression or erosion of these structures. Pancreatitis-related venous complications are quite frequent but rarely reported. [bib_ref] Venous complications of pancreatitis: a review, Hira [/bib_ref] It includes splenic or portal vein thrombosis, acute venous thrombosis causing bowel infarction, and renal or testicular vein compression causing varicocele. A pseudocyst may also be present in the mediastinum and presents with dysphagia or cough. However, a pseudocyst presenting around the IVC is rarely reported in literature and may even cause IVC thrombosis and pulmonary embolism. [bib_ref] Pancreatic pseudocyst that compressed the inferior vena cava and resulted in edema..., Browman [/bib_ref] [bib_ref] Pancreatic pseudocyst of head: compressing of vena cava inferior with lethal embolism..., Berges [/bib_ref] Management options in such an atypically located pseudocyst include either medical management or percutaneous, endoscopic, or surgical drainage. [bib_ref] Endoscopic management of pancreatic pseudocysts at atypical locations, Bhasin [/bib_ref] Medical or conservative management includes symptomatic management in the form of analgesics, antiemetics, NJ feeding, or total parenteral nutrition. [bib_ref] Impact of nasojejunal feeding on outcome of patients with walled off pancreatic..., Rana [/bib_ref] With such a conservative management, 40-60% of pseudocysts resolve without any further intervention. However, in the presence of complications such as infection, persistent pain, suspected bleeding, or compression of surrounding organs, drainage is required. Endoscopic transpapillary or transmural drainage is usually the preferred drainage route depending on the size, location, and presence of duct communication. [bib_ref] Endoscopic management of pancreatic pseudocysts at atypical locations, Bhasin [/bib_ref] [bib_ref] Endoscopic therapy for pancreatic duct leaks and disruptions, Varadarajulu [/bib_ref] Our patient did not require drainage of the collection as he responded to conservative management with NJ feeding. This case report highlights the importance of clinical suspicion and early diagnosis of such atypical locations of the pancreatic collections. EUS can be of great value in the diagnosis of such an atypical cystic lesion as well as in differentiating pseudocyst from WON.
[fig] Figure 1: (a) and (b) Contrast-enhanced computed tomography of abdomen: Long tubular cystic lesion measuring 7 × 3 cm around inferior vena cava. (c) Endoscopic ultrasound (EUS): Focal area of altered echotexture in the uncinate process of pancreas (open arrows), suggesting focal pancreatitis. (d) EUS: Cystic lesion had variable amount of solid component (bold arrows) suggestive of walled-off necrosis. A small duodenal cystic lesion is also noted (open arrows). [/fig]
|
Automatic Forest-Fire Measuring Using Ground Stations and Unmanned Aerial Systems
This paper presents a novel system for automatic forest-fire measurement using cameras distributed at ground stations and mounted on Unmanned Aerial Systems (UAS). It can obtain geometrical measurements of forest fires in real-time such as the location and shape of the fire front, flame height and rate of spread, among others. Measurement of forest fires is a challenging problem that is affected by numerous potential sources of error. The proposed system addresses them by exploiting the complementarities between infrared and visual cameras located at different ground locations together with others onboard Unmanned Aerial Systems (UAS). The system applies image processing and geo-location techniques to obtain forest-fire measurements individually from each camera and then integrates the results from all the cameras using statistical data fusion techniques. The proposed system has been extensively tested and validated in close-to-operational conditions in field fire experiments with controlled safety conditions carried out in Portugal and Spain from 2001 to 2006.
# Introduction
Wildfires destroy thousands of hectares each year and incur very high social, environmental and economic costs. Forest-fire fighting is a very dangerous activity that requires extensive resources and causes many casualties every year. In many cases, lack of precise information on the state and evolution of the fire front is one of the main causes of accidents.
Forest-fire fighting is traditionally based on estimations made by fire fighting experts from visual observations directly on the terrain or by analyzing data provided by sensors. These estimations are subject to a high degree of error due to human inaccuracy in the visual estimation and smoke hampering the observation of the fire. More recently, airborne systems have been used in order to provide a broader view of the fire, but the monitoring activities are still carried out based on visual estimations by experts.
Automatic forest-fire perception is a complex problem. Besides the general drawbacks that are present in perception in natural scenarios, such as uncontrollable and sudden changes in environmental conditions, there are others related to the particular characteristics of fire, including the harsh environment and the difficulties in predicting the spread of fire and smoke. Most research and development efforts devoted to automatic forest-fire perception have been focused on fire detection. Although there are a large variety of automatic detection systems, the number of automatic forest-fire measuring systems that have been tested and validated in field fires is very low.
This paper describes a forest-fire measurement system that automatically computes the evolution of the most important geometrical measurements of fire propagation in real-time, such as the position of the fire front, the rate of spread and the maximum height of the flames. The proposed system addresses the sources of error present in forest-fire measurement by exploiting complementarities between different types of cameras that are distributed on the terrain, so that they can obtain a complete and robust perception of the forest fire. The main sensors used are visual and infrared cameras located at distributed ground stations and mounted on UAS. As a secondary objective, 3D views of the fire are generated from the computed measurements. These views, compatible with ARCINFO Geographical Information System (GIS), are transmitted using a TCP/IP protocol allowing them to be visualized on the Internet. Fire brigades can use them in fire fighting planning to predict the potential evolution of the fire and determine the optimal location of fire fighting resources. The proposed measurement system combines cameras at fixed locations, which can already be installed as part of the forest infrastructure, together with cameras onboard UAS, which provide high flexibility suitable for highly changing environments. To the best of our knowledge, the system presented in this paper is one of the first forest-fire measurement systems that integrate results from visual and infrared cameras distributed at both fixed locations and cameras onboard UAS. It has been extensively validated in forest-fire field experiments in close-to-operational conditions in which plots of land of up to 2.5 hectares were burned. This paper is structured as follows. Section 2 describes related work in forest-fire detection and monitoring and presents the motivation and main advantages and constraints of the proposed work. Section 3 briefly describes the proposed system including the general architecture, the sensors involved and the system deployment. Section 4 focuses on the processing techniques used to extract fire measurements from each camera and Section 5 describes the multi-camera data fusion method. Section 6 shows some results obtained in field experiment. Conclusions are set out in the last section.
## Related work and motivation
A number of systems for automatic forest-fire detection have been developed based on different sensors. Some of the first automatic forest-fire detection systems were based on infrared cameras capable of detecting the radiation emitted by a fire [bib_ref] Remote Sensing. Communications and Information Technologies for Vegetation Fire Emergencies, Laurenti [/bib_ref]. These systems are often on ground stations at locations with high visibility. The stations are equipped with pan and tilt units that sweep the area under surveillance at specific intervals in order to detect fires at their early stages. These detection systems require a direct view of the radiation source. In addition, they often have a high false alarm rate which constrains their use in operational conditions. Some False Alarm Reduction (FAR) systems for forest-fire detection have been developed. The FAR system described in [bib_ref] An intelligent system for false alarm reduction in infrared forest-fire detection, Arrue [/bib_ref] combines different sources of information including infrared, visual, meteorological and terrain contextual information.
Fire detection based on visual cameras aims to detect the smoke plume produced by the fire. They can detect fires without direct vision, e.g., behind a hill. Different approaches to smoke detection with visual images have been developed based on contrast [bib_ref] Autonomous Forest Fire Detection, Den Breejen [/bib_ref] , texture [bib_ref] An Early Fire Detection Method Based on Smoke Texture Analysis and Discrimination, Cui [/bib_ref] , or motion analysis through wavelet analysis [bib_ref] Smoke Detection Using Image Processing, Gómez-Rodrí Guez [/bib_ref] , among others. Visual-based methods require daylight to operate and the accuracy in locating the fire is lower than in the case of infrared detection. In addition, detection systems based on Light Detection and Ranging (LIDAR) devices that identify the concentration of fire smoke particles have been developed [bib_ref] Lightning direction-finding systems for forest fire detection, Krider [/bib_ref] [bib_ref] Feasibility of forest-fire smoke detection using lidar, Utkin [/bib_ref] [bib_ref] Evaluation of smoke dispersion from forest fire plumes using lidar experiments and..., Lavrov [/bib_ref].
Forest-fire detection based on processing satellite images has been intensively researched, see [bib_ref] Satellite-based forest fire detection for fire control in boreal forests, Rauste [/bib_ref] [bib_ref] Detection and growth of an Alaskan forest fire using GOES-9 3.9 mum..., Hufford [/bib_ref] [bib_ref] Combining AVHRR and ATSR satellite sensor data for operational boreal forest fire..., Kelha [/bib_ref] [bib_ref] An efficient contextual algorithm to detect subsurface fires with NOAA/AVHRR data, Gautam [/bib_ref] for instance. These systems have been successfully tested in large, uniform and unpopulated regions. However, in populated areas these methods have relevant limitations. Satellite spatial and temporal resolutions involve significant detection delays that can be unsuitable in crowded areas. In addition, human activity can cause frequent false alarms that hamper the application of the systems in operational conditions. A number of automatic detection systems based on aerial platforms have been developed. The Airborne Wildfire Intelligence System (AWIS) includes wildfire detection and mapping of the fire-front and burned area [bib_ref] The airborne wildfire intelligence system: A decision support tool for wildland fire..., Campbell [/bib_ref]. The manoeuvrability of aerial vehicles such as helicopters, with hovering and vertical takeoff and landing capabilities, makes them an ideal tool to complement fixed cameras in fire monitoring and measurement. The use of UAS in forest-fire scenarios has been demonstrated in projects such as the FiRE project. In the FiRE project an ALTUS UAS, an adaptation of the Predator UAS, was demonstrated in fire detection and localisation. While the FiRE project considers one single and complex vehicle endowed with high accuracy sensors, in this paper the fire measurement system uses a team of small and simple UAS cooperating in fire perception.
In [bib_ref] Cooperative forest fire surveillance using a team of small unmanned air vehicles, Casbeer [/bib_ref] , the feasibility of the application of a team of small (low altitude, short endurance) UAS to cooperatively monitor and track the propagation of large forest fires is explored. The paper provides simulations using a six-degree of freedom dynamic model for the UAS and a numerical propagation model for the forest fire. However, results in actual fire fighting activities have still not been carried out. In addition, in [bib_ref] Unmanned Aerial Vehicle (UAV) Real-Time Video Registration for Forest Fire Monitoring, Zhou [/bib_ref] a method for the ortho-rectification of images gathered from an UAS, and their application in fire monitoring activities is presented. However, no actual fire monitoring results are described.
Recently, Wireless Sensor Networks (WSNs) have also been developed for forest-fire detection activities. In [bib_ref] Animals as mobile biological sensors for forest fire detection, Sahin [/bib_ref] , a system consisting of sensors (thermo and radiation with GPS) carried by animals living in the environment is used for forest-fire detection, although no experiments are presented. In [bib_ref] A wireless sensor network deployment for rural and forest fire detection and..., Lloret [/bib_ref] , the design and development of a WSN involving sensors and IP cameras is described. In [bib_ref] Forest fire modeling and early detection using wireless sensor networks, Hefeeda [/bib_ref] , the problem of fire detection is modelled as a k-coverage problem in WSNs. Some algorithms are proposed to solve the problem, but results are only shown in simulations. In [bib_ref] Performance of a protected wireless sensor network in a fire. Analysis of..., Antoine-Santoni [/bib_ref] , a WSN with nodes equipped with temperature and humidity sensors were used to detect a fire and extract measurements of fire spread. Although promising, this technology still has unresolved issues such as the costs of maintaining thousands of WSN nodes deployed in forest areas and their potential pollution effects.
In [bib_ref] Early Forest Fire Detection Using Radio-Acoustic Sounding System, Sahin [/bib_ref] , a fire detection method based on measuring the increase in air temperature using sonar units has been proposed. Sound wave speed is affected by air temperature. Acoustic sources distributed in the forest generate sound waves with a specific frequency. Radars at fire watchtowers continuously scan for acoustic waves. The differences in the speed of acoustic waves are used to detect fires.
Despite the large variety of forest-fire detection systems, the number of automatic forest-fire monitoring and measuring systems is still scarce.
Some work on measuring laboratory fires in controlled conditions have been proposed using different sensors. One basic method consists of the use of a set of thermocouples to detect the location of the fire front, see [bib_ref] Identification of the upward gas flow velocity and of the geometric characteristics..., Chetehouna [/bib_ref] for instance. The work in [bib_ref] Computing the rate of spread of linear flame fronts by thermal image..., Pastor [/bib_ref] describes a method based on linear transformations of infrared images to compute the positions and the rate of spread of a linear fire front propagating on a flat surface. The processing of multispectral infrared images to determine the fire perimeter, the 'active' fire line and the fire propagation direction is proposed in. The system described in [bib_ref] Laboratory fire spread analysis using visual and infrared images, Martinez De Dios [/bib_ref] combines infrared and visual cameras to obtain fire measurements in laboratory experiments. This work was developed and validated in laboratory fires in controlled conditions. The extension to forest fires is not addressed in these papers.
The method described in [bib_ref] A 3D vision system for the measurement of the rate of spread..., Rossi [/bib_ref] uses stereo-vision units with visual cameras to obtain fire geometry measurements. It is useful in laboratory fires but does not address practical problems of real forest fires, such as smoke occluding the visual images, preventing the extraction of valid measurements.
WSNs have also been proposed for this fire monitoring. In [bib_ref] A Multi-Tiered Portable Wireless System for Monitoring Weather Conditions in Wildland Fire..., Hartung [/bib_ref] , FireWxNet is described, consisting of a WSN designed to report weather conditions (humidity, temperature, wind speed) as well as images in fire environments. The paper shows real deployment of the WSN, which is evaluated in terms of battery performance, packet yield and information gathered.
Satellite-based fire monitoring systems, such as [bib_ref] Forest fire monitoring using NOAA satellite AVHRR, Flannigan [/bib_ref] [bib_ref] New Fast detection method for forest fire monitoring and application based on..., Feng [/bib_ref] [bib_ref] Remote sensing of burn severity: Experience from western Canada boreal fires, Hall [/bib_ref] , have been shown to be useful for monitoring of wildland fires. Although significant advances have been carried out, the temporal and spatial resolutions are still low for accurate monitoring in cases, such as urban-wildland interface areas, where very frequent and precise measurements of fire evolution are required [bib_ref] Wildland-Urban Interface Analyses for Fire Management Planning, Marzano [/bib_ref] [bib_ref] Potential and limitations on the operational use of remote sensing for active..., San-Miguel-Ayanz [/bib_ref].
We present here a system that allows us to obtain closer views and detailed information of relevance for fire fighting activities. It can provide fire measurements with a spatial resolution lower than one meter at rates of 2 Hz. Thus, the method is suitable for cases where intense monitoring is necessary, for instance fires in the urban-wildland interface where, often, cameras are already installed. The objective of the proposed system is not to measure large wildland fires. The number of cameras and communication infrastructure would be unaffordable for operational conditions. Its extension to large fires is the object of current research. Section 6 contains further discussion on the application of the proposed system for real fires.
By analyzing the previous approaches, the system presented in this paper intends to fill a gap in the current systems in terms of the spatial and temporal resolution of the information obtained. A system of this kind is not intended to substitute other available means but to complement the already existing tools. The use of UAS together with static cameras provides modularity and flexibility, which are suitable properties in highly changing environments [bib_ref] Multiple eyes in the skies: Architecture and perception issues in the COMETS..., Ollero [/bib_ref]. The mobility of UAS can be used to dynamically reduce uncertainty in fire perception, for instance, by taking images from viewpoints complementary to views from fixed cameras. In addition, they enable reacting to changes in the scenario, for instance, by moving to another viewpoint in the case where smoke occludes the images.
The authors believe that the main contributions of the work presented in this paper are:
-Exploitation of complementarities between visual and infrared cameras at static locations and mounted on UAS in order to improve the perception in terms of accuracy, resolution, robustness and the capability to adapt to changes in environment conditions. -Integration of measurements from the cameras available in a statistical framework that adapts the merging process dynamically and is able to continue providing fire estimations if one or more cameras stop working (due to damage by the fire, for example). -Implementation and validation of the system in close-to-operational conditions in field experiments carried out in Portugal and Spain from 2001 to 2006.
## General description
The main objective of the proposed system is to automatically obtain geometrical measurements of forest fires in real time, such as the location and shape of the fire front, the rate of spread and the fire flame height. A fire geometrical model commonly assumed in the forest-fire domain is depicted inThe main fire features in this model can be divided into fire base measurements (fire-front location l, width w) and into flame measurements (length d, height h and inclination angle θ). Two photographs of a fire taken during a field fire experiment in Serra de Gestosa (Portugal) in 2006 are shown in. For geometry measuring purposes the fire fronts can be approximated by a concatenation of triangles, see, each of them being characterised by the aforementioned features. The rate of spread is determined by the temporal evolution of the fire-front location.
## Sensors
The system uses two main types of imaging sensors with different wavelength bands: infrared cameras and visual cameras, which have interesting synergies for fire perception. Visual cameras provide images in the visible range 0.4-0.7 μm. They can be used to obtain flame measurements (height, inclination angle) applying suitable image processing techniques. They can also obtain fire base measurements (location, width). However, visual images can be occluded by smoke, whose spread is difficult to be predicted in natural outdoor scenarios.
On the other hand, infrared cameras provide images of the scene containing the radiation intensity field within the infrared band. Infrared cameras are not affected by smoke, being highly transparent compared to the high radiation levels originated in a forest fire. The radiation intensity emitted by the base of the fire is considerably higher than that of the flames [bib_ref] Infrared Measurements of Energy Release and Flame Temperatures of Forest Fires, Den Breejen [/bib_ref]. Thus, infrared images can be used to obtain measurements of the fire base but are not useful for measuring the flames. Both types of sensors are necessary to obtain the aforementioned fire geometrical model.
The measurement system can use thermal infrared cameras and also non-thermal cameras, which do not provide temperature measurements but do allow qualitative estimations of the radiation intensity. Although cameras in the mid-infrared spectral window 3-5 μm are preferred mainly due to its lower atmospheric absorption, the proposed system can indistinguishably use cameras in the mid-infrared or far-infrared windows. In fact, both types of cameras were used in most of the field experiments that have been carried out.
## System deployment
Perception using distributed cameras increases the robustness of the measurement against potential sources of errors. In some cases, distributed fixed cameras cannot cope with unexpected changes in the spread of fire and smoke. Unmanned Aerial Systems, used as 'flying cameras', are ideal platforms on which to overcome these constraints: they can be controlled to move to a suitable viewpoint to improve the observation of the fire or to complement views from fixed cameras. The system presented here exploits the complementarities between the cameras by adopting a statistical sensor fusion approach. Sensor fusion techniques can be used to merge measurements from different sensors in order to obtain overall estimations. Thus, they reduce the influence of errors in measurements and increase the overall accuracy of the system. Statistical data fusion techniques, such as Bayesian Recursive Filters, dynamically adapt the merging process taking into account current errors in the measurements from each of the sensors. The data fusion technique adopted is described in Section 5.
In a typical case, the fire measuring system requires one of more camera stations and one main processing station. Each camera station (fixed on the ground or mounted onboard an Unmanned Aerial System) can be equipped with one camera (visual or infrared) or two cameras (visual and infrared). A deployment with three ground camera stations and two UAS camera stations is depicted in . A photograph of a ground station with one infrared camera and one visual camera in a fire experiment is shown in .
## Figure 2. (a)
A typical deployment of the proposed forest-fire measuring system. (b) Picture of a ground station with one Mitsubishi IRM-300 infrared camera and one visual camera in a fire experiment in Serra de Gestosa (Portugal).
## (a) (b)
If a camera station is deployed with the camera optical axes in the main direction of the fire-front advance (we call them frontal views), visual cameras can be used to obtain measurements of the fire base and of the flames when smoke does not occlude the images. Frontal infrared cameras enable us to obtain fire base measurements. If a camera station is deployed with the camera axes perpendicular to the fire-front spread (lateral views), visual images are useful to determine flame height and inclination angle. In practice deployment constraints can arise due to the topography of the terrain. The viewpoint of cameras onboard UAS can be set to complement static cameras and can be changed dynamically, for instance, if smoke occludes the images. The forest-fire measuring system was designed and developed to be modular and flexible to a wide range of different deployments (fixed/mobile, infrared/visual cameras) suitable for the topography and conditions of each case.
Aerial images can be used to measure the fire base (location, width and shape) but cannot accurately measure flame height due to the parallax problem. The Unmanned Aerial Systems are equipped with Differential GPS (DGPS) receivers and Inertial Measurement Units (IMUs) so that their location and orientation is known with accuracy. The photograph of a helicopter UAS used in the experiments is shown in . It carries a pan and tilt device with one low-cost infrared micro-camera in the far infrared band and one visual camera.
## Forest-fire measurement processing
A diagram of the main steps in the proposed forest-fire measuring system can be observed in . The main inputs of the system are the images from each of the cameras. All the cameras have been internally calibrated using calibration patterns. Each image is associated to data regarding the type of camera, its location and orientation, timestamps and camera calibration information. In the case of cameras onboard an UAS, all the images captured are tagged locally with the composed location and orientation of the aerial vehicle and the pan and tilt unit. The proposed forest-fire measurement is carried out in two main steps: single-camera processing and multi-camera data fusion. Single-camera processing blocks apply image processing methods to compute fire measurements independently for each of the cameras deployed. This block includes image pre-processing methods to filter out spurious effects and, in the case of UAS, special software dedicated to reject camera vibrations induced by the vehicle. Electro-mechanical systems, such as gimbals, were avoided in our implementation due to the payload constraints of small aerial platforms, such as those employed in the experiments. Instead, it includes stabilization methods based on image processing. Single-camera processing blocks also include image processing methods to extract fire features from the images. Different algorithms are used depending on the type of camera-infrared/visual and fixed/mobile-providing high modularity and flexibility suitable for a wide variety of camera deployments. This block also includes methods to transform these image-plane fire features to real-world measurements. The main techniques used are summarised in the next section.
The multi-camera fusion block integrates the fire measurements computed individually from each camera to obtain overall forest-fire estimations. The data fusion method adopted is based on Kalman Filtering, see Section 5. It implements temporal and spatial filtering techniques to cancel high-frequency fluctuations and local errors. This block also generates 3D views of the fire.
## Single-camera processing
This section briefly describes the main techniques used to obtain forest-fire measurements from each camera: image pre-processing, fire feature extraction and image geo-referencing.
## Image pre-processing
The objective is to increase the quality of the images before extracting fire features. In the first step, a simple 3 × 3 median filter is applied to cancel out image noise and potential spurious effects. Changes in lighting conditions are an important source of error in outdoor computer vision applications. In the forest-fire measuring system the lighting compensation method proposed in [bib_ref] An Infrared Vision System for Field Robotics Applications, Martí Nez-De Dios [/bib_ref] is adopted. The main idea is to dynamically modify parameters of the cameras, such as gain and brightness level, in order to keep the lighting conditions in the images gathered constant despite the changes in the environment. The current illumination conditions of the images (essentially brightness and contrast) are measured and compared with the reference-desired-values. The error is used to select suitable lighting parameters of the camera, as in a feedback control scheme.
In addition, the vision-based method described in [bib_ref] Motion Analysis and Geolocation for Aerial Monitoring in the COMETS Multi-UAV System, Merino [/bib_ref] is used to cancel the vibrations induced by the UAS in the images. Assume Im t (x,y) and Im t+1 (x,y) are two consecutive images in a sequence. The image motion compensation used is applied in three steps: estimation of the motion between Im t (x,y) and Im t+1 (x,y), fitting of the motion to a model and application of the inverted motion model to each pixel in Im t+1 (x,y). The resulting motion-compensated image, Im t+1 * (x,y), has no apparent motion with respect to Im t (x,y). Motion estimation is based on feature association. Assume that the scene contains enough and sufficiently distributed features perceptible in both images. The corner detector described inis applied to both images. Then, in feature association, each feature from Im t (x,y) is the centre of a window that is used as a template for matching over Im t+1 (x,y). Features are associated based on normalized cross-correlation [bib_ref] Real-time feature matching in image sequences for non-structured environments. Applications to vehicle..., Ferruz [/bib_ref] , see [fig_ref] Figure 5: Feature extraction and association in two consecutive aerial visual images [/fig_ref]. Feature association allows us to extract the motion of the features in two consecutive images. However, not all the objects in the images may have the same movement, and features with movements different to the general scene motion originate errors in the motion estimation. These feature associations are considered disturbances and should not be used for motion estimation. A method based on Least Median of Squares is used to identify them [bib_ref] Motion Analysis and Geolocation for Aerial Monitoring in the COMETS Multi-UAV System, Merino [/bib_ref]. In the next step, valid feature associations are used to fit the motion model. The homography matrix is used to model the motion between images since it can be used to describe the transformations originated by changes in the location and orientation of the camera when the imaged scene can be approximated as a plane. Once the homography matrix has been computed, motion compensation (image warping) is performed by applying the inverse homography matrix to all pixels in Im t+1 (x,y). The performance of the image stabilization method is illustrated in [fig_ref] Figure 6: Image stabilization in three consecutive aerial images [/fig_ref]. Three consecutive aerial images are shown in [fig_ref] Figure 6: Image stabilization in three consecutive aerial images [/fig_ref] (a-c). [fig_ref] Figure 6: Image stabilization in three consecutive aerial images [/fig_ref] (d-f) shows the resulting images after vibration cancellation (the image in [fig_ref] Figure 6: Image stabilization in three consecutive aerial images [/fig_ref] is considered as a reference) and fire feature extraction (described in the next subsection). Notice that the position of the fire front after vibration cancellation is very similar in the three images.
## Fire feature extraction
Two stages can be identified in this step: fire segmentation and feature extraction. Temperature thresholding is used for segmenting images from thermal infrared cameras. Their temperature measurements depend on parameters, such as the surface emissivity of the object, which in general cannot be known with accuracy in operational fighting conditions. This lack of accuracy has a low influence on image segmentation due to the high difference between fire and background temperatures. On the other hand, non-thermal infrared cameras, such as the Raytheon 2000AS onboard the UAS, provide qualitative estimations of the radiation intensity at the scene. In these cases the thresholding algorithm described in [bib_ref] A multiresolution threshold selection method based on training, Martí Nez-De Dios [/bib_ref] is used. As mentioned previously, infrared images are used to obtain measurements of the fire base but cannot be used to measure the flames. Visual cameras are used instead. Fire segmentation in visual images is carried out by a learning-based method similar to that described in [bib_ref] Flame recognition in video, Philips [/bib_ref]. The method requires training images in which a user has determined the pixels that correspond to fire. A histogram is built with the RGB values of each pixel considered as fire in the training images.
From the segmented images it is simple to determine the fire contours. From the fire contours it is possible to determine the fire geometrical features. The extraction of features from infrared images is illustrated in [fig_ref] Figure 7: Extraction of features from infrared images [/fig_ref]. Segmentation of infrared images provides the fire base pixels, whose contour is the fire-base contour, see [fig_ref] Figure 7: Extraction of features from infrared images [/fig_ref]. The direction of fire advance can be estimated by analysing the motion of the centroid of the fire-base pixels through time, and then it is possible to distinguish between the front and the rear fire-base contours, see [fig_ref] Figure 7: Extraction of features from infrared images [/fig_ref]. Then, fire-base width measurements can be obtained, see [fig_ref] Figure 7: Extraction of features from infrared images [/fig_ref]. Similar procedures are applied to visual images. For further details refer to [bib_ref] Gómez-Rodrí guez, F. Computer vision techniques for forest fire perception, Martinez-De Dios [/bib_ref].
## Image calibration and geo-referencing
The objective is to transform the fire features extracted on the image plane to real world measurements. If the terrain can be locally approximated by a plane, the relation between the terrain plane and the image plane is a homography. A point P on the terrain is transformed to point p on the image plane of camera i using p = H i P, where H i is the homography matrix for camera i, H i = A i T i , where A i represents the internal calibration for camera i and T i is the transformation matrix that relates the real-world coordinate system to the coordinate system of camera i. The computation of the homography involves establishing correspondences between real-world points on the terrain and their image coordinates. Although the minimum number is four, several well-distributed correspondences are used to reduce the error. In the experiments described in this paper more than 12 were used. Landmarks such as trees and rocks of known size were used to find correspondences. In a real fire application, we do not consider it practical to deploy items only for image calibration purposes. In this case, objects of known location and size that are present in the image, such as fire fighter trucks, could be used. Planar terrain is not a hard constraint. Even though the terrain is not planar, in most cases this assumption is valid if the cameras are located far enough from the fire.
The homography-based method requires planar terrain and a sufficient number of correspondences. If the conditions are not met, the proposed fire measuring system uses the following projection-based method. If a Digital Terrain Model (DTM) of the environment is available, geo-referencing can also be carried out by projecting the image pixels on the DTM. Projection on the terrain requires accurate knowledge of the location and orientation of the camera. In the experiments carried out, they were measured with Differential GPS and Inertial Measurement Units both for cameras on ground stations and onboard UAS. In addition, it requires precise synchronization between images and location and orientation measurements. GPS timing is taken as the reference time for all the sensors. Time stamps are used for all the data to avoid timing confusions.
Further details on both geo-referencing methods can be found in [bib_ref] Motion Analysis and Geolocation for Aerial Monitoring in the COMETS Multi-UAV System, Merino [/bib_ref].
## Multi-camera forest-fire estimation
The objective is to integrate all the fire measurements obtained independently from cameras with different perception capabilities. Bayesian Filters provide a well-founded mathematical framework for estimating the state of the system using observations in presence of noise: sensors are modelled as uncertain sources. Decentralised schemes require a strong communication infrastructure (with sufficient bandwidth and coverage) which is often inexistent in wildland environments. In contrast, centralised schemes only require point-to-point communication between each camera station and the main processing station. Adopting a practical approach, in the proposed system we chose the latter option for its easier deployment. In the experiments carried out with up to six distributed cameras, the proposed system was capable of operating on a standard laptop at a rate not lower than 2 Hz, which can be considered real-time for the monitoring of forest fires, whose measurements vary in a clearly larger time scale.
The basic diagram of the Recursive Bayesian Filter (RBF) used is shown in [fig_ref] Figure 8: Diagram of the Recursive Bayesian Filter used [/fig_ref]. The input is z t , the set of measurements obtained individually from each of the N cameras. The output of the block is the estimation of the state of the fire front at time t, s t . The RBF requires one update model in order to perform short-term prediction of the evolution of the state and one observation model for each of the cameras used in the deployment. Inaccuracies in the prediction and noise in the observations should be taken into account. RBFs obtain an updated estimation of the state as a weighted average using the prediction of its next state as well as using a new measurement from the sensor. The purpose of this weighting is to give more trust to values with better (i.e., smaller) estimated uncertainty. This process is repeated with every new measurement. . The state of the entire fire front s t contains a complete geometrical description of the fire front. Its size is 5 × NT. It is possible to select its size by modifying the distance between the fire-front triangles, which can be useful when there are very large fire fronts. The state of two adjacent fire triangles cannot be considered independent. Thus, it is necessary to use one Bayesian Filter for the whole fire front.
## (a) (b)
Forest-fire modelling is a very complex problem. Forest-fire propagation is subject to the influence of many effects related to terrain topography, meteorology and fuel conditions, among others [bib_ref] Forest fire propagation, Viegas [/bib_ref]. A very wide variety of methods and approaches for forest-fire modelling has been researched. The main practical motivation is to use these models in forest-fire attack to predict the behaviour of a fire during the next hours or days, seefor instance. Consequently, many of the approaches aim to accurately analyse the phenomena involved in combustion in forest environments, taking into account many factors and thus, resulting in significantly complex models.
It is not the objective of the proposed method to use an exhaustive and highly accurate fire propagation model. The objective in our problem is only to allow the merging of fire measurements in a very short-term prediction-update scheme. In fact, such high accuracy is not needed in our system since the prediction-update cycles of the RBF occur at a frequency not lower than 2 Hz: (1) the prediction error is low for very short-term predictions; (2) the prediction errors are corrected in the update stage of the RBF. The recursive prediction-update scheme of RBFs has been proven to be robust to inaccuracies in the prediction stage: the update stage is capable of contrasting the predictions with the new measurements and of making the suitable corrections.
For measurement merging purposes, our system assumes that, in the very short-term, fire location can be described by a linear dynamic system plus noise in order to take into account the inaccuracies, weigh the contribution of the rate of spread at j, j − 1 and j + 1, respectively.
Many effects, such as fuel characteristics and wind conditions, are not considered and would involve important errors in long-term prediction but it is not the case in our problem: we use this prediction only for local short-term updating of the RBF and for measurement merging purposes. Under these conditions, a simple local short-term fire prediction will suffice to integrate fire measurements, as can be observed in the experimental results in Section 6.
The prediction includes temporal and spatial smoothing properties and is capable of cancelling high-frequency fluctuations and local errors in the measurements. From the above expressions, it is easy to obtain a linear representation, s t+1 = As t + w t . Many of the entries in A are zero, involving moderate computational cost. w t takes into account these inaccuracies in the prediction, which can be considered to be originated by a high number of independent effects. These errors are assumed to be Gaussian. We will denote the covariance of w t by matrix Q. The parameters α 1 , α 2 , … α 5 and Q were set to average values determined by fitting the model with real experimental data.
The observation model for camera i follows the expression z t,i = C i s t +v t,i , where v t,i is the observation uncertainty assumed to be Gaussian with zero mean and covariance matrix R i . z t,i is the vector of measurements obtained by camera i for all the fire-front triangles,
[formula] [ ] T T NT i t T i t i t z z = z , 1 , , . [/formula]
The observations of fire-front triangle j at time t obtained by camera i is
[formula] [ ] T j i t, j i t, j i t, j i t, j i t θ h fw x = z , . [/formula]
C i can be built using cc i , (4 × 5 matrix that relates j i t z , and j t ts ) and 0 4,5 (4 × 5 zero matrix). cc i depends on the type of camera. For instance, cc 1 in (1) corresponds to an infrared frontal camera that provides fire base location and width measurements while cc 2 corresponds to a lateral visual camera that measures flames height and inclination angle:
[formula] cc (1) [/formula]
C i matrices are sparse and involve moderate computational burden. R i is different for each camera and depends mainly on the camera location and orientation and on the camera type, reflecting the fact that some cameras are more affected by some disturbances, e.g., smoke occlusions, than others. It can also be computed comparing the measurements made with data obtained using photogrammetry.
In our problem, a Kalman Filter is used to implement the RBF. Kalman Filters are suitable for systems with linear prediction and observation and Gaussian uncertainties. The simple prediction and observation models assumed allow efficient implementation. Section 6.2 presents some results that illustrate the advantages of the sensor fusion method proposed.
## Field experiments
This section is divided into two parts. The first analyses the influence of the main sources of error on fire measurement. The second describes the fire experiments and presents some results.
## Sources of error
The main sources of error in the proposed fire measurement are smoke occluding the visual images, high-frequency fluctuations in the fire front and errors in image geo-referencing. Changes in lighting conditions are compensated for by the method described in [bib_ref] An Infrared Vision System for Field Robotics Applications, Martí Nez-De Dios [/bib_ref]. Image synchronization errors can be neglected since fire measurements vary in a clearly larger time scale.
Smoke occluding the visual images hampers the extraction of valid measurements in visual images. As an example,illustrates the effect when computing the location of the most advanced point of a fire-front spreading down-slope. Our system naturally addresses smoke occluding the visual images by using infrared cameras and by deploying cameras at different locations, which helps to reduce the probability of simultaneous occlusion in all the visual cameras. In addition, UAS can be dynamically controlled to move to locations with good fire perception. Besides, from a data fusion perspective, errors originated by smoke occlusions can be easily detected. Sudden changes in the measurements produce unexpectedly high errors in the update step of the Kalman Filter, e t,i = z t,i − C i s t , which can be detected by identifying when |e t,i | is higher than a threshold T, see. The value of T depends on the degree of uncertainty of the measurement. We adopted T = k√σ, where σ is the expected variance in case of no occlusion. k was experimentally chosen as k = 10. A measurement is considered valid if |e t,i | < T. The Kalman Filter only integrates measurements considered to be valid.
The proposed fire measurement system is also robust to the failure of cameras, for instance, in case of the camera being damaged by the fire. Camera malfunctioning originates unexpected changes in the measurements and high errors in the update step of the Kalman Filter. Thus, the aforementioned method also prevents the system from integrating measurements from cameras that are malfunctioning. Of course, in the case of failure, the number of cameras whose measurements are integrated by the fire measuring system decreases.
To test the influence of the geo-referencing methods several Monte Carlo analyses were conducted by deliberately introducing errors in both geo-referencing methods described in Section 4.3. The error was a Gaussian noise with zero mean and different variances. Fire features extracted in the image plane were geo-located by both error-polluted geo-referencing methods. Some results in two images of the same experiment with frontal (left) and aerial views (right) are shown in. In the frontal view, the errors were introduced in the correspondences used for the computation of the homography matrix. In the aerial view, the errors were introduced in the location of the camera. The abscissa axis represents the standard deviation of the noise introduced and the ordinate axis represents the standard deviation of the error in geo-referencing in axes x (full line inand y (dashed line). The results show that the homography-based method is rather robust to errors in the selection of correspondences, see(left). On the other hand, terrain projection is not as robust to errors in the location of the camera, see(right). To prevent these errors the location of the UAS is measured with Differential GPS (DGPS).. Results of sensitivity analysis in homography-based (left) and terrain projection (right) methods.
The homography-based method requires planar terrain and a sufficient number of correspondences. If the conditions are not met, the projection-based method is adopted. Although more sensitive to errors, it can be applied if a Digital Terrain Model is available.
# Experimental results
The proposed system has been extensively tested in fire experiments carried out in Portugal and Spain from 2001 to 2006. In the experiments, plots of dimensions of up to 160 m × 160 m were burned in controlled conditions [bib_ref] Analysis of Fire Behaviour in the Mediterranean Shrubs: The Gestosa Fire Experiments..., Viegas [/bib_ref]. Significant resources were used in these experiments including over 80 firemen, over 5 fire trucks, manned helicopters and several UAS. Three images of the Gestosa site (Portugal) can be observed in. The experimental area of Gestosa included more than 60 different plots distributed over several square kilometres. Each year different plots were burnt and different positions of the ground cameras were used. The deployment of the system including the number and type of cameras used and their location was chosen for each experiment taking into account the terrain topography. The forest-fire measuring system was designed and developed to be modular and flexible to a wide range of different camera deployments. Visual cameras are equipped with motorized zooms. Infrared cameras are equipped with different lenses, which are selected to best fit the image. The visual cameras were JAI 2060 with a resolution of 752 × 582 and were equipped with zooms with focal lengths from 15 to 90 mm. Four different infrared cameras are used: Mitsubishi IRM-300 in the mid-infrared spectral window with lenses with a horizontal field of view (hFoV) of 24°, FLIR 550 Elite in the mid-infrared window with two lenses (hFoV of 10° and 20°), FLIR Thermacam P20 in the far-infrared window with three lenses (hFoV of 12°, 24° and 45°) and Raytheon 2000AS micro-cameras in the far-infrared window with a 24° hFoV lens was used onboard the UAS.
The proposed system relies on close views of the fire from different viewpoints. Of course, the resolution of the fire measurements obtained depends on the distance from the cameras to the fire and on the lenses used. In real fires, our suggestion is to use zooms for ground stations and to complement them with images from UAS controlled to gather close views of the fires. In the experiments carried out, the system provided reliable results if the fire-front line corresponded to percentages of the image width higher than 18%. As an example, in the experiments performed in year 2002, the ground frontal station was located at approximately 2.9 km. In these experiments the FLIR 550 Elite infrared camera with 10° hFoV lenses was used. The spatial resolution of the measurements from the frontal infrared was 0.5 pixels/m. These measurements were merged with others obtained from cameras onboard UAS with close views, resulting in an increase in the resolution of the overall measurements.
In a typical deployment, the images gathered by cameras in static ground stations were transmitted to the main processing station using analogue radio links with directional antennas. For UAS the images were captured and compressed using the lossless JPEG format with onboard electronics and then were transmitted to the main processing station using a long-range radio modem with onmidirectional antennas. In a typical deployment, all the processing modules in the fire measurement system were executed in the main processing station. In some fire experiments, we tested configurations where the single camera processing modules for UAS cameras were executed onboard the UAS, and the results were transmitted to the main processing station for measurement integration. The development of a fully decentralised fire measuring system is the object of current research.
Prior to execution, the fire measuring system is configured with data from the deployment, such as the location, orientation and type of the cameras, the number of UAS and the type of the cameras onboard.shows the main window of the fire measuring tool in an experiment with frontal visual and infrared cameras and one lateral visual camera. Images from the three cameras, graphics with some fire measurements and the 3D view of the fire are displayed. A screenshot of the UAS station during an experiment is shown in. The images from the visual and infrared cameras onboard and navigation data can be seen. A video of the experiment can be observed in. The videos in [51] and show the processing of aerial visual and fixed visual images in another field experiment. The aerial images are stabilized before segmentation. The videos are generated with a low frame rate to enable the visualisation of results.; the average flame height of the fire front throughout the experiment in; the location of the most advanced point of the fire front throughout the experiment in. It should be pointed out that these results were obtained on-site, in real-time, during the fire experiment. The advantages of using the proposed Kalman Filter method to reduce the errors in fire measurement can be observed in. It shows the location of the fire front every 30 s. in an experiment with a forest fire spreading down-slope.represents the results obtained individually from: one visual camera mounted on an UAS; one frontal ground visual camera; and one frontal ground infrared camera, respectively. The direction of fire advance is represented by an arrow. It can be observed that the fire measurements obtained from the three cameras contain some errors. For instance, during certain periods, the fire produced high levels of smoke and the visual images did not generate measurements, see the dashed ellipses in. The infrared camera is only capable of obtaining valid measurements in this interval.represent results estimated by the Kalman Filter with measurements from only one camera. During the smoke occlusion intervals, the Kalman Filter relies on the measurements from the infrared camera. Red fire-front lines represent estimations obtained by merging results from at least two of the three cameras. The overall perception after integrating measurements from different distributed cameras is significantly better than any of the individual perceptions.
The proposed system has been compared to classical fire measuring methods consisting in manually timing the intervals in which fires burn through threads placed at regular distances.shows the temporal evolution of the position of the most advanced point of the fire front computed by the fire measurement system. The dots inrepresent the time at which the fire burns through the threads. The mean relative error between both results is lower than 8%. Similar results were found in other experiments that were carried out. The experiments showed that the exploitation of the synergies between different cameras located at different positions enables a better estimation of the fire parameters, in terms of accuracy and reliability; overcoming the disadvantages of classical vision-based systems such as occlusion or low resolution. In addition, the data fusion adopted uses the integration of the available information to recover from the short-term failure of cameras. The approach is even able to continue providing fire estimations if one or more cameras stop working (due to damage by the fire, for example). Of course, in this case, the estimated fire parameters will be less accurate due to the lack of this information.
The authors believe that the proposed system is suitable for measuring fires in populated or close to urban areas, where accurate and frequent fire measures are necessary. Fires close to urban areas are usually smaller than those in wildland areas: the lower detection delay and higher availability of extinction means help to prevent the fire from becoming large. In addition, in these areas cameras and communication infrastructure are often already installed and there is good accessibility that allows the quick deployment of the system. The setup makes it difficult to implement the proposed approach in large wildland fires, where the number of cameras and the required communication infrastructures would make the approach unaffordable for operational conditions. However, the main constraints for its use in wildland fires originate in the deployment of the system and not in the image processing or sensor fusion techniques themselves. Deployment is the main issue for using the proposed system in real forest fires. We do not consider that deploying ad-hoc infrastructure to be used in the event of fire is the best option in all cases: the maintenance costs can be high and we might never have a fire. We believe that the best deployment strategy is to benefit from available existing infrastructure and to deploy the required components of the system once the fire has been detected. In case of fire, the system components could be quickly transported in all-terrain trucks. These modified trucks, equipped with cameras on pan and tilt systems, computing and communication resources, could act as mobile ground stations or as processing stations. In case of uncontrolled fires, the trucks-and the ground cameras-could move to a safer location or with better visibility conditions. On the other, UAS, transported also with the trucks, can be launched manually or using catapults. Although intense implementation work is required to develop such a system, most of the processing and measuring algorithms proposed in this paper remain valid. In addition, it should be noticed that advanced command posts equipped with sensors, computing systems and communications are already being used operationally in forest-fire fighting in many countries, for instance, Spain.
In recent years we have also researched into other solutions based on transportation and self-deployment of loads with several cooperating UAS [bib_ref] Cooperative Autonomous Helicopters for Load Transportation and Environment Perception, Kondak [/bib_ref]. The load cannot be transported by only one UAS due to payload constraints and thus, the tight cooperation of several of them is required. These technologies, which have already been tested in urban scenarios [bib_ref] Ollero, A. A distributed architecture for a robotic platform with aerial sensor..., Maza [/bib_ref] , could be applied for the deployment of camera stations and communication infrastructure in areas with difficult accessibility. Another option under current research is to extend the approach to a system composed only of cooperating UAS and no ground stations, which would accelerate system deployment and allow us to cover larger fires.
# Conclusions
This paper describes an automatic forest-fire measuring system with cameras fixed at ground stations and mounted on UAS. The method can obtain geometrical forest-fire measurements such as the location and shape of the fire front, fire-base width and flame height, as well as the temporal evolution of these features including the rate of spread, in real-time. The system can provide these fire measurements with significant spatial and frequency resolutions, being suitable for fires of moderate size that require intense monitoring, for instance, those close to urban areas. The objective of the proposed system is not to measure large wildland fires.
The proposed system tackles the sources of error and unpredictable changes in scenario conditions exploiting the complementarities between different types of cameras (visual and/or infrared) fixed at different locations and mounted on UAS. The method is based on obtaining geo-referenced fire measurements individually from each camera and applying statistical data fusion techniques to merge all the measurements into unique estimates. It makes extensive use of image stabilization, fire segmentation, feature extraction, image geo-location and data fusion techniques.
The proposed system has been extensively tested and validated in close-to-operational conditions in field experiments from 2001 to 2006, in which several tenths of hectares were burned under controlled safety conditions. The promising results obtained represent a step forwards to automatic fire measurement in full fire fighting conditions. This paper opens up several topics of research. UAS have been shown to be very valuable in forest-fire perception. A fire measurement with no ground cameras and only UAS would allow a higher degree of flexibility to adapt to changes in fire and environmental conditions. Fire measurement accuracy increases with the number of cameras deployed. The use of decentralized data fusion techniques in which each agent maintains a perception of the state of the fire would improve the scalability and robustness of the system. Both topics would allow us to cover larger fires and accelerate system deployment. Future work will also explore new mechanisms for system self-deployment in areas with difficult access and without existing infrastructure. In particular, we are analysing the integration of self-deployed ad-hoc sensor networks, ground cameras and aerial robots with sensing capabilities in order to reduce the infrastructure requirements of the proposed approach, enabling it to be applied to extensive wild forest areas.
[fig] Figure 1: (a) Basic scheme of fire geometrical model. (b) Visual images taken during a field fire experiment. (c) Geometrical model of a fire front. [/fig]
[fig] Figure 3, Figure 4: (a) Main sensors and components of one helicopter used in forest-fire measurement. (b) Detailed picture of the pan and tilt device with one Raytheon 2000AS OEM infrared micro-camera and one visual camera. Diagram of the proposed forest-fire measurement system. [/fig]
[fig] Figure 5: Feature extraction and association in two consecutive aerial visual images. [/fig]
[fig] Figure 6: Image stabilization in three consecutive aerial images: (a-c) original consecutive aerial images with vibrations, (d-f) resulting images after vibration cancellation and extraction of features. [/fig]
[fig] Figure 7: Extraction of features from infrared images: (a) fire-front, (b) once the fire-base has been segmented, it is easy to identify the fire-base contour, (c) taking into account the direction of fire advance, it is simple to differentiate between the front and the rear fire-base contours, (d) the fire-base width is computed as the distance between pixels on the front and the rear contours along the direction of fire advance. [/fig]
[fig] Figure 8: Diagram of the Recursive Bayesian Filter used.As described in Section 3 fire fronts can be approximated by a concatenation of triangles, each of them characterized by the position and width of the fire-base and height and inclination angle of the flames, seeFigure 9. The state of the whole fire front is the state of a series of equallyis the state of forest-fire triangle j at time t and NT is the number of triangles. The state of a system should contain all the information from the system necessary to predict its future. Thus, j t ts should include the local rate of spread associated to triangle j, [/fig]
[fig] Figure 9: (a) Schematic geometry of a fire-front triangle. (b) Diagram representing the state of the fire front as the state of a series of equally-spaced fire-front triangles. [/fig]
[fig] ..: The fire base width is related to the rate of spread and the time required α 1 is used to weigh the contribution of j t rs . Flame height depends on the amount of combustible being burnt, which depends on the rate of spread: Also, contiguous fire-front triangles may interact with each other. We describe that interaction in terms of rate of spread, α 3 , α 4 and α 5 [/fig]
|
Linguistic effect on speech perception observed at the brainstem
Linguistic experience affects speech perception from early infancy, as previously evidenced by behavioral and brain measures. Current research focuses on whether linguistic effects on speech perception can be observed at an earlier stage in the neural processing of speech (i.e., auditory brainstem). Brainstem responses reflect rapid, automatic, and preattentive encoding of sounds. Positive experiential effects have been reported by examining the frequency-following response (FFR) component of the complex auditory brainstem response (cABR) in response to sustained high-energy periodic portions of speech sounds (vowels and lexical tones). The current study expands the existing literature by examining the cABR onset component in response to transient and low-energy portions of speech (consonants), employing simultaneous magnetoencephalography (MEG) in addition to electroencephalography (EEG), which provide complementary source information on cABR. Utilizing a cross-cultural design, we behaviorally measured perceptual responses to consonants in native Spanish-and English-speaking adults, in addition to cABR. Brain and behavioral relations were examined. Results replicated previous behavioral differences between language groups and further showed that individual consonant perception is strongly associated with EEG-cABR onset peak latency. MEG-cABR source analysis of the onset peaks complimented the EEG-cABR results by demonstrating subcortical sources for both peaks, with no group differences in peak locations. Current results demonstrate a brainstem-perception relation and show that the effects of linguistic experience on speech perception can be observed at the brainstem level.speech perception | brainstem response | linguistic experience | MEG/EEG
Linguistic experience affects speech perception from early infancy, as previously evidenced by behavioral and brain measures. Current research focuses on whether linguistic effects on speech perception can be observed at an earlier stage in the neural processing of speech (i.e., auditory brainstem). Brainstem responses reflect rapid, automatic, and preattentive encoding of sounds. Positive experiential effects have been reported by examining the frequency-following response (FFR) component of the complex auditory brainstem response (cABR) in response to sustained high-energy periodic portions of speech sounds (vowels and lexical tones). The current study expands the existing literature by examining the cABR onset component in response to transient and low-energy portions of speech (consonants), employing simultaneous magnetoencephalography (MEG) in addition to electroencephalography (EEG), which provide complementary source information on cABR. Utilizing a cross-cultural design, we behaviorally measured perceptual responses to consonants in native Spanish-and English-speaking adults, in addition to cABR. Brain and behavioral relations were examined. Results replicated previous behavioral differences between language groups and further showed that individual consonant perception is strongly associated with EEG-cABR onset peak latency. MEG-cABR source analysis of the onset peaks complimented the EEG-cABR results by demonstrating subcortical sources for both peaks, with no group differences in peak locations. Current results demonstrate a brainstem-perception relation and show that the effects of linguistic experience on speech perception can be observed at the brainstem level.
speech perception | brainstem response | linguistic experience | MEG/EEG E arly language experience profoundly affects how speech sounds are perceived throughout life. Transient consonants present the most challenging and intriguing case as their perception relies on neural encoding and processing of extremely brief, highly dynamic, and yet often weak acoustic signals. The smallest change in the signal (e.g., a few milliseconds increase in a pause) at a critical point in the stimulus can result in a drastic change in perception (e.g., from "bat" to "pat"), a phenomenon referred to as "categorical perception" (CP) for speech perception [bib_ref] Speech perception as categorization, Holt [/bib_ref]. For decades, research has repeatedly demonstrated that these critical positions for acoustic change (i.e., categorical boundaries) differ for speakers of different language backgrounds [bib_ref] Voice-timing perception in Spanish word-initial stops, Abramson [/bib_ref] [bib_ref] An effect of linguistic experience: The discrimination of [r] and [l] by..., Miyawaki [/bib_ref].
The influence of linguistic experience begins very early in development [bib_ref] Early language acquisition: Cracking the speech code, Kuhl [/bib_ref]. A rich literature has documented that by 12-mo of age, infants' ability to discriminate nonnative consonants already begins to decline while the ability to discriminate native consonants improves [bib_ref] Infants show a facilitation effect for native language phonetic perception between 6..., Kuhl [/bib_ref] [bib_ref] Cross-language speech perception: Evidence for perceptual reorganization during the first year of..., Werker [/bib_ref]. For adults, discrimination of nonnative consonants remains far from "native-like," even after periods of intensive training [bib_ref] Training Japanese listeners to identify English/r/ and/l/: A first report, Logan [/bib_ref] [bib_ref] Age constraints on second-language acquisition, Flege [/bib_ref] [bib_ref] Learning foreign sounds in an alien world: Videogame training improves non-native speech..., Lim [/bib_ref].
Research focusing on the cortical level of neural processing has found evidence to corroborate perceptual data. For example, the mismatch response (MMR), indexing neural sensitivity to speech sound changes at the cortical level, has been observed to be larger in speakers of a language in which the speech contrast is native than in speakers of a language in which the contrast was nonnative [bib_ref] Neurophysiologic correlates of cross-language phonetic perception, Sharma [/bib_ref]. In infants, MMRs to nonnative contrasts at 11 mo of age were observed to be significantly reduced compared with MMRs observed at 7 mo age [bib_ref] Brain potentials to native and nonnative speech contrasts in 7-and 11-month-old American..., Rivera-Gaxiola [/bib_ref].
The current study set out to examine whether the effects of linguistic experience on consonant perception already manifest at a deeper and earlier stage than the cortex, namely, at the level of auditory brainstem. The auditory brainstem encodes acoustic information and relays them to the auditory cortex. Its activity can be recorded noninvasively at the scalp using a simple setup of three electroencephalographic (EEG) electrodes [bib_ref] Human auditory evoked potentials: Possible brain stem components detected on the scalp, Jewett [/bib_ref]. Using extremely brief stimuli such as clicks or tone pips, the auditory brainstem response (ABR) can be elicited and has been used clinically to assess the integrity of the auditory pathway [bib_ref] Universal infant hearing screening by automated auditory brainstem response measurement, Mason [/bib_ref]. More recently, researchers began recording auditory brainstem responses to complex sounds, such as speech and musical tones [i.e., complex ABR (cABR)] with the same EEG setup [bib_ref] Auditory brain stem response to complex sounds: A tutorial, Skoe [/bib_ref]. The cABR consists of an onset component (onset) and a frequency-following response (FFR) component [bib_ref] Cortical contributions to the auditory frequency-following response revealed by MEG, Coffey [/bib_ref]. The onset reflects the encoding of transient changes in the signal (e.g., the consonant) while the FFR reflects the brainstem's tracking of sustained periodic information (i.e., fundamental frequency and higher harmonics in vowels and tones) [bib_ref] Brain stem response to speech: A biological marker of auditory processing, Johnson [/bib_ref]. Thus far, most studies have focused on the FFR to vocalic portions of speech sounds, which contain high levels of acoustic energy (e.g., lexical tones and vowels). Early language experience has been observed to enhance FFR to lexical tones and vowels Significance Early linguistic experience affects perception and cortical processing of speech, even in infants. The current study examined whether linguistic effects extend to brainstem speech encoding, where responses are rapid, automatic, and preattentive. We focused on transient consonants and measured perception behaviorally and the corresponding complex auditory brainstem response (cABR) onset using simultaneous electroencephalography/magnetoencephalography (EEG/MEG) in native Spanish and English speakers. We demonstrate that the latency of an EEG-cABR onset peak predicts consonant perception and that the perception differs according to language background. MEG-cABR analysis demonstrates deep sources for the onset, providing complimentary support for a brainstem origin. Effects of early linguistic experience on speech perception can be observed at the earlier stage of speech encoding at the brainstem. [bib_ref] Cross-domain effects of music and language experience on the representation of pitch..., Bidelman [/bib_ref] [bib_ref] Native language shapes automatic neural processing of speech, Intartaglia [/bib_ref] [bib_ref] Encoding of pitch in the human brainstem is sensitive to language experience, Krishnan [/bib_ref]. Similarly, researchers have demonstrated that early music training experience enhances adults' FFR to nonnative lexical tones [bib_ref] Musical experience shapes human brainstem encoding of linguistic pitch patterns, Wong [/bib_ref] [bib_ref] Music training for the development of auditory skills, Kraus [/bib_ref] [bib_ref] Music enrichment programs improve the neural encoding of speech in at-risk children, Kraus [/bib_ref]. To date, only one study examines directly the relations between vowel perception and the FFR component, suggesting that FFRs to vowels reflect the acoustic signal of the vowels, rather than the perception of them [bib_ref] Tracing the emergence of categorical speech perception in the human auditory system, Bidelman [/bib_ref]. However, the underlying source of FFR remains a topic of active debate [bib_ref] Cortical contributions to the auditory frequency-following response revealed by MEG, Coffey [/bib_ref] [bib_ref] Cortical correlates of the auditory frequency-following and onset responses: EEG and fMRI..., Coffey [/bib_ref].
To examine transient consonant encoding at the brainstem level, we differed from previous studies and focused on the onset component of cABR as a function of perception and linguistic experience. This approach allows us to address several important gaps in the literature and therefore can advance our theoretical understanding of the effects of linguistic experience on speech perception, reflected deep at the brainstem level. First, we increased the confidence that any effects observed in the current study are predominantly subcortical in nature through two means: (i) by examining the EEG-cABR onset to transient stop consonants (<20 ms) and (ii) by using simultaneous magnetoencephalography (MEG)/EEG recordings. The early nature of the onset response ensures a predominant subcortical source. Simultaneous MEG recordings complimented the EEG-cABR by providing source information for the onset peaks. Dipole modeling methods revealed subcortical sources for the onset peaks, corroborating previous research [bib_ref] Cortical contributions to the auditory frequency-following response revealed by MEG, Coffey [/bib_ref] and supporting a brainstem origin for the EEG-cABR onset. Second, by gathering both perception data and cABR from the same individual, the current study is able to examine the brainstem-perception relation.
Specifically, we tested two hypotheses that (i) individuals' EEG-cABR onset in response to consonants is related to their perception of the sounds, and (ii) both the EEG-cABR onset and perception are affected by listeners' language background. Further, using simultaneously measured MEG-cABR, we provide complimentary information to EEG-cABR regarding the sources of the onset peaks.
Monolingual English speaking adults (n = 29) and native Spanish speaking adults (n = 20) were recruited to first complete computer-based identification and discrimination tasks to assess their categorical perception of bilabial stop consonants varying on the voice-onset-time (VOT) continuum (for details see Materials and Methods). The VOT continuum ranged from −30 ms to +30 ms in 10-ms increments. In the identification task, participants judged whether the speech sound is /ba/ or /pa/ through speeded key presses. The percent identification of a sound on the continuum as /pa/ is calculated as the outcome measure. In the discrimination task, participants judged whether two speech sounds were the same or different. The sensitivity indices (d′) for pairs of adjacent sounds along the continuum were calculated as the outcome measure to minimize response bias. In line with previous research, monolingual English speakers demonstrated significantly different categorical perception for the VOT continuum compared with native Spanish speakers and SI Appendix, . Even though the language background in the native Spanish speaker group is much wider in range (SI Appendix, Table S1), the categorical perception was more variable in English speakers, as the VOT cue is not the only and may not be the predominant cue for the voiced-voiceless (/b/-/p/) contrast in some English speakers (e.g., aspiration). To maximize the detection of differences at the brainstem level, we selected participants whose perception results suggested strong weighting of the VOT cue for further MEG/EEG recording. The selection criteria were: (i) strong evidence of a category boundary from the identification or discrimination task and (ii) a phonetic boundary aligned with previous studies (< +10 ms for native Spanish speakers and > +10 ms for monolingual English speakers).
In the simultaneous MEG/EEG recording sessions, participants' brainstem response to the stimulus with +10 ms VOT (p10 from here on, , where maximal difference in perception between groups was observed, was recorded in two blocks (3,000 trials per block). Three EEG electrodes were placed in accordance with cABR recording conventions (CZ, ground on forehead and reference on right ear lobe). Recordings were done while participants were sitting under the MEG dewar watching silent videos. The stimuli were presented in alternating polarities through insert earphones (Materials and Methods). Individuals who were selected for MEG/EEG recordings and completed successful MEG/EEG recordings were included in the analysis: fourteen monolingual English speakers and fifteen native Spanish speakers. The two groups did not differ in their music training background (subject details in Materials and Methods).
The MEG/EEG data were preprocessed, filtered, and averaged (across trials and then blocks) to increase the signal-tonoise ratio. The cABR responses were extracted for each participant in both the EEG and the MEG channels (details in Materials and Methods). In B-D, an example from an individual's recording for both MEG and EEG is shown to demonstrate the data quality and also that the brainstem responses can be observed on multiple MEG channels. In , dipole analysis from one example individual with existing MRI shows MEG onset peaks are indeed localized to subcortical sources.
To address our research questions, two analyses were done on the EEG-cABR data: (i) the onset peaks of EEG-cABR (magnitudes and latencies) corresponding to the stop consonant (peak I for initial noise burst and peak II for voice onset, were compared between the two groups with different linguistic backgrounds. We hypothesized that because VOT is a timing cue, a group difference would likely lie in the latency of response. (ii) The onset peaks in EEG-cABR were examined in relation to individuals' behavioral results on p10 perception . In addition, source information for the peaks was provided through the equivalent current dipole (ECD) analyses on the MEG-cABR data. Dipole locations for both peaks were compared between groups . We expected deep sources for both peaks (in support of EEG brainstem response) for both groups and that the dipole source locations would not differ between groups. . Categorical perception of bilabial stop consonants on a VOT continuum in monolingual English speakers (black) and native Spanish speakers (red) from the final sample. (A) Data from the identification task. Shaded area for stimulus with +10 ms VOT (p10), highlights the range of behavior across the two groups. (B) Data from the discrimination task. Shaded area for stimulus pair (0/+10 ms VOT) highlights the range of behavior across the two groups.
## A b
# Results
Behavioral results from all participants are shown in SI Appendix, . Significant group differences on the perception of p10 were confirmed. Similarly, behavioral results from the final sample included in the study are shown in . As expected, the effects in the final sample were the same but larger due to the selection process. Native Spanish speakers identified the p10 syllable significantly more strongly as "pa" (mean pa identification = 0.85 ± 0.13) than monolingual English speakers [mean pa identification = 0.35 ± 0.24], t (1, 27) = 6.96, P < 0.001, Bonferroni correction applied, d = 2.59 , shaded area). Spanish speakers also discriminated stop consonants with 0 and +10 ms VOT more accurately (mean d′ = 2.37 ± 0.72) than English speakers [mean d′ = 0.76 ± 0.74], t (1, 27) = 5.90, P < 0.001, Bonferroni correction applied, d = 2.21 .
For the EEG-cABR data, we first extracted the magnitude and latency values for the first two peaks , peak I and peak II) in the response. Peak I is in response to the initial noise burst while peak II is in response to the voice onset of the stop consonant . Independent t tests were conducted to compare the magnitudes and latency values between the two groups.
For peak I, independent t tests revealed no significant difference in either magnitude or latency between the two groups [magnitude: t (1, 27) = −1.13, latency: t (1, 27) = 0.117] . For peak II, independent t tests revealed no significant difference in magnitude, t (1, 27) = 1.69, but a significant difference in latency, t (1, 27) = −3.98, P < 0.001, d = 1.45 . That is, and as predicted, native Spanish speakers exhibited a later peak II (latency = 30.6 ms ± 0.36 ms) than English speakers (latency = 29.69 ms ± 0.81 ms).
Further, we conducted correlation analyses between individuals' peak II latencies and their behaviorally measured perceptual data to examine the brain-behavior relation. Significant regression models were observed between peak II latency and p10 identification (R 2 = 0.323, β = 0.569, P = 0.001, , as well as between peak II latency and d′ scores measuring discrimination ability for the +10-ms/0-ms syllable pair (R 2 = 0.209, β = 0.457, P = 0.013, . That is, the longer the peak II latency in a participant's cABR onset, the more strongly the participant identified the stimulus as /pa/ and the higher the participant's discrimination of the stimulus from its neighboring stimulus on the continuum (VOT = 0 ms).
From the MEG-cABR data, we obtained source information on the onset peaks to compliment the EEG-cABR by using the ECD method (Materials and Methods). A spherical head model was used with all magnetometers for dipole fitting without MRI structural scans. Specifically, the fitted ECD coordinates for each peak were extracted and examined to address: (i) whether the sources of these two peaks suggest deep sources, approximating brainstem origins; and (ii) whether there were differences in the sources of these peaks between groups. The group averaged topographies around Peak I and Peak II time points and ECD locations for both peaks can be visualized in . The topographies demonstrate field patterns with deep sources (i.e., activities along the edge, instead of cortical sources (i.e., dipolar patterns in each hemisphere, SI Appendix, , identified in previous studies [bib_ref] Cortical contributions to the auditory frequency-following response revealed by MEG, Coffey [/bib_ref] [bib_ref] Sparsity enables estimation of both subcortical and cortical activity from MEG and..., Krishnaswamy [/bib_ref]. For peak I, no significant differences were observed for x, y, or z coordinates shaded area indicates time window for peak extraction. (C) Group averages for peak I and peak II magnitude. No statistical difference was observed between groups for either peak. (D) Group averages for peak I and peak II latency. Native Spanish speakers exhibited a significantly longer latency for peak II, but not peak I. In the current study, by examining the EEG-cABR onset to transient consonants, we minimized the effects of cortical contributions, and by simultaneously recording MEG data, we complimented the EEG-cABR results by providing source information of the onset peaks through dipole modeling methods. Adult speakers from different linguistic backgrounds (monolingual English speakers vs. native Spanish speakers) were examined both behaviorally, using classic speech perception tasks, and at the auditory brainstem level using simultaneous MEG/ EEG recordings of cABR. We tested the idea that individuals' EEG-cABR onsets in response to consonants are related to their perception, and that both onset responses and perception are clearly differentiated by the language background of the listener. Further, we provided complimentary information to EEG-cABR regarding the source of the onset peaks using MEG-cABR data.
Native Spanish speakers exhibited a significantly later peak II in the EEG-cABR onsets in response to a bilabial stop consonant (p10), compared with the monolingual English speakers. This peak corresponds to the onset of voicing in this stop consonant. Significant relations were observed between the individual latency value for peak II and listeners' perception of this stop consonant (p10). That is, across language backgrounds, the longer the latency for peak II, the more strongly the speaker identified the stop consonant as /p/, and the better the speaker discriminated the stimulus from a neighboring stop consonant on the VOT continuum (VOT = 0 ms). Simultaneous MEG recordings allowed access to the source information through ECD modeling methods. The onset peaks were localized to deep sources, complementing the EEG-cABR and providing further support for a brainstem origin. As predicted, no differences between groups in the source locations were observed for either peak.
Taken together, our current results strongly support the idea that the effects of early linguistic experience on transient consonant perception can be observed at a very early stage of auditory processing (i.e., the brainstem level). Specifically, brainstem encoding is strongly associated with the perception of the consonants measured behaviorally. Both onset responses and perception can be separated by the language background of the listener, suggesting brainstem speech encoding is experience related and behaviorally relevant.
The current study is consistent with previous research suggesting early experiential effects on the FFR component of cABR, a component that reflects the encoding of periodic information in complex sounds, such as fundamental frequency of voice or tones [bib_ref] Auditory brain stem response to complex sounds: A tutorial, Skoe [/bib_ref] [bib_ref] Brain stem response to speech: A biological marker of auditory processing, Johnson [/bib_ref]. However, the origin of the FFR has been debated [bib_ref] Cortical contributions to the auditory frequency-following response revealed by MEG, Coffey [/bib_ref] [bib_ref] Cortical correlates of the auditory frequency-following and onset responses: EEG and fMRI..., Coffey [/bib_ref]. By examining the early EEG-cABR onsets to a transient stop consonant in relation to perception, along with the ECD modeling from MEG-cABR data, the current study extends the existing literature and provides additional and stronger evidence to support the theoretical framework that the early stage of speech sound encoding at the brainstem level may be modulated by long-term experience through a descending corticofugal system [bib_ref] Music training for the development of auditory skills, Kraus [/bib_ref]. The corticofugal pathway originates from cortex and can extend to as early as the auditory periphery (cochlea) [bib_ref] Role of corticofugal feedback in hearing, Suga [/bib_ref]. For example, it has been demonstrated in animals that short-term laboratory-controlled experience can affect auditory frequency tuning curves at the auditory cortex, brainstem, midbrain, as well as in the hair cells in the cochlea [bib_ref] Multiparametric corticofugal modulation and plasticity in the auditory system, Suga [/bib_ref] [bib_ref] Modulation of cochlear hair cells by the auditory cortex in the mustached..., Xiao [/bib_ref]. In humans, at the onset of hearing around 24-25 wk of gestation, complex and dynamic speech sounds become audible [bib_ref] Development of fetal hearing, Hepper [/bib_ref] [bib_ref] The development of human fetal hearing, Birnholz [/bib_ref] , and short-term in utero auditory experience is already manifested by the time of an infant's birth [bib_ref] Language experienced in utero affects vowel perception after birth: A two-country study, Moon [/bib_ref] [bib_ref] Of human bonding: Newborns prefer their mothers' voices, Decasper [/bib_ref] [bib_ref] Learning-induced neural plasticity of speech processing before birth, Partanen [/bib_ref]. Language experience is therefore one of the longest and most profound experiences humans have, making the current results compatible with a corticofugal mechanism. At the same time, the current study provides a rare opportunity to examine a brain-behavior relation when both measurements can be transformed onto the same scale (i.e., time). While numerous studies have established that a small change in the neural signal (e.g., microvolts in EEG amplitude) can be predictive of a behavioral shift at a larger scale (e.g., perception and/or cognitive skills) (e.g., refs. 40 and 41), it is difficult to speculate the exact "amplification mechanism." In the current study, the "cross-over" point in category perception can also be calculated (i.e., where identification is 50% on a psychometric function fitted on each individual's data), indicating the stimulus position where perception changes from /ba/ to /pa/. The result shows that a range of 3.4 ms in the brainstem peak II latency across subjects corresponds to a range of 18.9 ms in their category boundaries, providing some information of this brain-to-perception amplification mechanism.
Several points should be taken into consideration regarding the current results. First, group effects on the cABR onsets remain limited to the subgroup that showed strong weighting of the VOT cue. It remains to be further examined whether cABR A B . (A) Group average topographies in the peak I and peak II time windows. (B) Group average dipole (ECD) locations for peak I and peak II. All peaks have deep sources and no statistical differences were observed between groups in the ECD coordinates. . Relations between peak II latency and perception. (A) Scatterplot between peak II latency for p10 stimulus in cABR and percent identified p10 as /p/. (B) Scatterplot between peak II latency for p10 stimulus in cABR and d′ sensitivity to discriminate between p10 and stop consonant with 0 ms VOT (+0/+10 ms VOT pair). Significant models were observed for both regressions.
## A b
onset is also related to perception in individuals who do not rely on the VOT cue in the voiced/voiceless contrast. Discovering this will help us further elucidate the mechanism through which cABR is related to perception and experience. Second, while the cABR onset peaks can be predominantly considered subcortical in nature, given the high level of repetition in stimulus presentation, other sources (e.g., auditory cortex) may contribute as a result of online adaptation and learning. More sensitive MEG recording methods (e.g., more advanced denoising methods to allow analysis with fewer trials) with individual MRIs in the future would allow distributed source localization methods and could potentially help elucidate dynamic changes in the underlying sources during the actual recording session. The current study also generates important future research directions regarding this experiential effect; for example, what is the earliest point in development that an experiential effect can be observed at the brainstem level and how does this effect interact with perceptual development as well as cortical processing? Speech categories emerge as distributional patterns in speech input and social experiences occur, and different cortical processing patterns have been observed for different speech categories [bib_ref] Infant sensitivity to distributional information can affect phonetic discrimination, Maye [/bib_ref] [bib_ref] Categorical speech representation in human superior temporal gyrus, Chang [/bib_ref]. During the "sensitive period" between 6 and 12 mo of age, infants rapidly learn to discriminate native speech contrasts while their ability to discriminate nonnative speech contrasts start to decline [bib_ref] Infants show a facilitation effect for native language phonetic perception between 6..., Kuhl [/bib_ref]. Theories posit that the neural system starts a process of "neural commitment" to the processing of native speech categories by 12 mo of age [bib_ref] Theta brain rhythms index perceptual narrowing in infant speech perception, Bosseler [/bib_ref]. To understand the extent of such neural commitment, future research is warranted to examine the relations between cortical and brainstem levels of phonetic processing in cross-linguistic adult studies as well as infants at the beginning and the end of the sensitive period for phonetic learning.
# Materials and methods
Behavioral Assessment on Categorical Perception. Participants. Monolingual English speakers (n = 29, male = 9, age = 21.48 ± 2.15) and native Spanish speakers (n = 20, male = 6, age = 25.40 ± 4.76) were first recruited to complete the behavior assessment on their categorical perception of bilabial stop consonants varying on the VOT dimension. All participants were healthy adults with no reported speech, hearing, or language disorders. All participants were right-handed (Edinburgh handedness quotient = 0.98 ± 0.04). All procedures were approved by the Institutional Review Board of the University of Washington and informed consents were obtained from all participants. Stimuli. Bilabial stop consonants with varying VOTs were synthesized by Klatt synthesizer in Praat software. The VOT values ranged from −30 ms to +30 ms with 10-ms increments. The syllable with 0 ms VOT was first synthesized with a 2-ms noise burst and vowel /a/. The duration of the syllable was 90 ms. The fundamental frequency of the vowel /a/ began at 95 Hz and ended at 90 Hz. Silent gaps or prevoicing were added after or before the initial noise burst to create syllables with positive or negative VOTs. The fundamental frequency for the prevoicing portion was 100 Hz. The waveform of the syllable p10 is shown in . Equipment and procedure. Upon arrival, all participants were consented and completed a short survey on their language and music backgrounds. Then they proceeded to a sound attenuated booth for the computer-based identification and discrimination tasks to assess categorical perception. All sounds were delivered through Sennheiser HDA 280 headphones at 72 dB SPL. Both tasks were completed on a Dell XPS13 9333 computer running Psychophysical Toolbox [bib_ref] The psychophysics toolbox, Brainard [/bib_ref] in MATLAB version 2016a (MathWorks, Inc.).
In an identification trial, a syllable was played and the participant was instructed to identify the sound as either /ba/ or /pa/ through key presses within 1 s. All seven syllables on the continuum were repeated 25 times in a randomized order. In a discrimination trial, two syllables, either identical or adjacent on the continuum, were played with a 300-ms interstimulus interval. Participants were instructed to respond whether the two sounds were the same or different through a key press within 1 s. All 24 (6 pairs × 4 sequences) possible combinations (e.g., +10/0, 0/+10, +10/+10, and 0/0) were repeated 20 times in a randomized order. Practice trials were first administered and participants had to reach 80% to proceed. One participant was excluded during this phase.
Data analysis. For the identification task, the percentage of identification of the syllable as /pa/ was calculated for each sound. For the discrimination task, d′ values for each pair of syllables (e.g., +30/+20 ms VOT) were calculated to index the sensitivity of discrimination. The d′ measure takes into consideration both hit (response of "different" when sounds were different) and false alarm (response of "different" when sounds were the same) responses, and therefore addresses the issue of response bias.
## M/eeg brainstem recording.
Participants. A subgroup of the participants who completed the behavioral assessments was further selected for M/EEG recording of their brainstem responses (monolingual English speakers = 16, native Spanish speakers = 18).
Among the participants invited back, two participants opted out for M/EEG recording and three participants' M/EEG recordings were rendered too noisy (i.e., impendence of EEG channels >10 kΩ). Final sample for group analyses (i.e., complete behavioral and good M/EEG recording) included 14 monolingual English speakers (male = 5, age = 21.21 ± 1.76) and 15 native Spanish speakers (male = 5, age = 26.00 ± 5.04). There was no difference between the two groups in their music training background [years of training: monolingual English = 3.66 ± 2.80, native Spanish = 2.12 ± 2.24, t (1, 27) = 1.64, P = 0.112]. The native Spanish speakers reported speaking at least one foreign language (English) at high proficiency (SI Appendix, ; however, none were raised as simultaneous bilinguals. The English speakers reported to be monolingual with minimal proficiency in other languages. Stimuli. Bilabial stop consonant with +10 ms VOT (p10 from here on) was selected for the M/EEG recording given the range in perception data . The waveform of the synthesized syllable is shown in . The 10-ms silent gap starts from the offset of the noise burst and ends at the onset of voicing, making the first peak of onset of voicing at 19 ms (edge of the shaded area). Equipment and procedure. Simultaneous MEG and EEG recordings were completed inside a magnetically shielded room (MSR) (IMEDCO). For EEG, a standard three-electrode setup was used: CZ electrode on a 10-20 system, ground electrode on the forehead and the reference electrode on right earlobe. Impendence of all electrodes was kept under 10 kΩ. For MEG, an Elekta Neuromag system was used with 204 planar gradiometers and 102 magnetometers. Five head-position-indicator (HPI) coils were attached to identify head positions under the MEG dewar at the beginning of each block. Three landmarks [left preauricular (LPA) and right preauricular (RPA), and nasion] and the HPI coils were digitized along with 100 additional points along the head surface (Isotrak data) with an electromagnetic 3D digitizer (Fastrak, Polhemus). The sounds were delivered from a TDT RP 2.7 (Tucker-Davis Technologies Real-Time Processor), controlled by custom Python software on a Hewlett-Packard workstation, to insert earphones (Nicolet TIP-300). The stimulus was processed such that the rms values were referenced to 0.01 and it was further resampled to 24,414 Hz for the TDT. The sounds were played at the intensity level of 80 dB with alternating polarities. The interstimulus intervals were 150 ms with jitters within a ±10-ms window. Two blocks of recordings (3,000 sounds per block) were completed at the sampling rate of 5,000 Hz. The participants listened passively and watched silent videos during recording. Data analysis. All data analysis was done using the MNE-Python software [bib_ref] MNE software for processing MEG and EEG data, Gramfort [/bib_ref] in conjunction with the Elekta Neuromag source modeling software (XFit-5.5). For EEG data, after referencing the CZ channel, the data from each block was first notch filtered at 60 Hz and at its harmonics up to 2,000 Hz to remove any power line interference. It was further band-pass filtered between 80 Hz and 2,000 Hz using a fourth order Butterworth filter. Epochs between −10 ms and 150 ms (in relation to sound onset) were extracted and averaged within the block and then across the blocks. Epochs with peak-to-peak values exceeding 100 μV were rejected. An example from one participant is shown in . Because the two peaks are visible in all participants with consistent timing , we extracted the magnitude and latency values for the two peaks related to the stop consonant for each participant by identifying the maximum value and its corresponding time in the windows between 5 and 15 ms for peak I and between 25 and 35 ms for peak II [bib_ref] Auditory brain stem response to complex sounds: A tutorial, Skoe [/bib_ref].
MEG data were first preprocessed using the oversampled temporal projection (OTP) method [bib_ref] Reducing sensor noise in MEG and EEG recordings using oversampled temporal projection, Larson [/bib_ref] and the temporally extended spatial-signalseparation (tSSS) method [bib_ref] Presentation of electromagnetic multichannel data: The signal space separation method, Taulu [/bib_ref] [bib_ref] Removal of magnetoencephalographic artifacts with temporal signal-space separation: Demonstration with single-trial auditory-evoked..., Taulu [/bib_ref] to suppress sensor noise and magnetic interference originating from outside of the MEG dewar. Both algorithms are implemented in the MNE-Python software. The preprocessed data were then subjected to the same notch filter (60 Hz and harmonics) and band-pass filters (80-2,000 Hz) as for the EEG data. The same epochs (−10-150 ms) were extracted and averaged. The epochs with peak-to-peak amplitude exceeding 4 pT/cm for gradiometers or 4.0 pT/cm for magnetometers were rejected.
[fig] Figure 2, Figure 3: [x: t (1, 27) = 0.21, y: t (1, 27) = −0.27, z: t (1, 27) = −0.06]. For peak II, no significant differences were observed either for x, y, or z coordinates [x: t (1, 27) = 0.16, y: t (1, 27) = −0.65, z: t (1, 27) = −0.86]. Discussion The current study demonstrated that the effects of early linguistic experience on transient consonant perception can be (A) Waveform of the p10 stimulus. Shaded area represents the stop consonant portion. (B) MEG channel distribution with the EEG-CZ channel. (C) cABR to the p10 stimulus from one representative MEG channel from one individual. Shaded area corresponds to the onset component to the stop consonant. (D) cABR to the same stimulus simultaneously recorded by the EEG channel in the same individual. Shaded area corresponds to the onset component to the stop consonant. (E) Dipole locations from one individual demonstrated brainstem sources for onset peaks (magenta, peak I; cyan, peak II). (A) Group average of the onset response from the EEG channel for p10 stimulus. Shaded area indicates the peaks of interest. Peak I represents brainstem response to the initial noise burst. Peak II represents brainstem response to the voice onset. (B) Individual onset responses from the final sample; [/fig]
|
The role of p21ras in CD28 signal transduction: triggering of CD28 with antibodies, but not the ligand B7-1, activates p21ras
CD28 is a 44-kD homodimer expressed on the surface of the majority of human T cells that provides an important costimulus for T cell activation. The biochemical basis of the CD28 accessory signals is poorly understood. Triggering of the T cell antigen receptor (TCR) activates the p21 r~s proteins. Here we show that ligation of CD28 by a monoclonal antibody (mAb) also stimulates p21 r~s and induces Ras-dependent events such as stimulation of the microtubule-associated protein (MAP) kinase ERK2 and hyperphosphorylation of Raf-1. One physiological ligand for CD28 is the molecule B7-1. In contrast to the effect of CD28 mAb, the present studies show that interactions between CD28 and B7-1 do not stimulate p21 r~s signaling pathways. Two substrates for TCR-regulated protein tyrosine kinases (PTKs) have been implicated in p21 r~ activation in T cells: p95 V~v and a 36-kD protein that associates with a complex of Grb2 and the Ras exchange protein Sos. Triggering CD28 with both antibodies and B7-1 activates cellular PTKs, and we have exploited the differences between antibodies and B7-1 for p21 ~' s activation in an attempt to identify critical PTK-controlled events for Ras activation in T cells. The data show that antibodies against TCR or CD28 induce tyrosine phosporylation of both Vav and p36. B7-1 also induces Vav tyrosine phosphorylation but has no apparent effect on tyrosine phosphorylation of the Grb2associated p36 protein. The intensity of the Vav tyrosine phosphorylation is greater in B7-1 than in TCR-stimulated cells. Moreover the kinetics of Vav tyrosine phosphorylation is prolonged in the B7-1-stimulated cells. These studies show that for CD28 signaling, the activation of p21 ~s correlates more closely with p36 tyrosine phosphorylation than with Vav tyrosine phosphorylation. However, the experiments demonstrate that Vav is a major substrate for B7-activated PTKs and hence could be important in CD28 signal transduction pathway.
T lymphocyte activation is initiated when the T cell encounters antigen/major histocompatibility molecules presented to the T cell on the surface of an APC. The signals generated by the TCR ensure the immune specificity of T cell activation, but for a full immune response there is a requirement for additional signals triggered by accessory molecules present on the surface of the APC [bib_ref] Roles of multiple accessory molecules in T cell activation, Van Seventer [/bib_ref]. A large number of molecules participate in the interaction between the T cell and the APC, including molecules like CD4 and CD8, as well as adhesion molecules and integrins such as CD2 and LFA-1. However, there is considerable evidence that CD28, a 44-kD homodimeric glycoprotein expressed by most mature T lymphocytes, is one of the more important receptors for the generation of the "second signal" for T cell activation [bib_ref] The role of the CD28 receptor during T cell responses to antigen, Linsley [/bib_ref]. Two physiological ligands for CD28 have been described: B7-1 (B7/BB1) and B7-2 (B70). These B7 molecules are expressed by APCs and their interactions with CD28 have 1067 been shown to costimulate T ceils in cooperation with engagement of the TCR .
Triggering of the TCR/CD3 complex initiates a cascade of biochemical events, the earliest of which is activation of intracellular protein tyrosine kinases (PTKs) t (10). This activation of PTKs appears to be absolutely required for all subsequent T cell responses (11, 12). One substrate for TCRstimulated PTKs is an inositol lipid-specific phospholipase C, PLC3'I (13). Hence PTKs enable the TCR to regulate the hydrolysis of membrane phosphoinositides (Ptdlns), releasing diacylglycerols (DAG) that activate the serine/threonine kinase, protein kinase C (PKC) (14-16), and inositol polyphosphates that regulate intracellular Ca 2+.
The TCR also stimulates the guanine nucleotide binding proteins p21TM (17, [bib_ref] The regulation and function of p21ras in T lymphocytes, Downward [/bib_ref] , which then couple the TCR to the MAP kinase ERK2 and hence to the regulation of transcriptional factors such as AP-1, NFAT, and the IL-2 gene [bib_ref] p21 r~s couples in T cell antigen receptor to extracellular signal-regulated kinase..., Izquierdo [/bib_ref] [bib_ref] Downward. 1992. p21ras mediates control of IL2 gene promoter function in T..., Rayter [/bib_ref] [bib_ref] p21ras function is important for T cell antigen receptor and protein kinase..., Woodrow [/bib_ref]. TCR activation of p21TM requires cellular PTK function and recent studies have identified two PTK substrates, Vav and a 36-kD membrane protein, as potential mediators of the TCR/p2Y as link [bib_ref] Tyrosine kinase-stimulated guanine nucleotide exchange activity ofvav in T cell activation, Gulbins [/bib_ref] [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. In fibroblasts, PTKs stimulate p2Y a~ by regulating the activity of Ras GTP/GDP exchange proteins. T cells express two potential Ras guanine nucleotide exchange proteins: Sos, a homologue of the Drosophila Son of Sevenless gene product [bib_ref] Identification of murein homologues of the Drosophila Son of Sevenless, Botwell [/bib_ref] , and Vav, which has in vitro guanine nucleotide exchange activity for Ras. It is proposed that the activity of Vav is regulated by tyrosine phosphorylation [bib_ref] Tyrosine kinase-stimulated guanine nucleotide exchange activity ofvav in T cell activation, Gulbins [/bib_ref]. In contrast, Sos regulates p21 ras when it is recruited to the cell membrane by the adaptor protein Grb2 [bib_ref] The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to..., Lowenstein [/bib_ref] [bib_ref] Association of Sos Ras exchange factor protein with Grb2 is implicated in..., Egan [/bib_ref] [bib_ref] Epidermal growth factor regulates p21ras through the formation of a complex of..., Buday [/bib_ref] [bib_ref] The SH2 and SH3 domains of mammalian Grb2 couple the EGF receptor..., Rozakis-Adcock [/bib_ref] [bib_ref] The SH2/SH3 domain containing protein GRB-2 interacts with tyrosine phosphorylated IRS-1 and..., Scolnik [/bib_ref]. Grb2 is composed of two Src homology (SH)3 domains that bind to Sos and an SH2 domain that interacts with tyrosine phosphoproteins [bib_ref] The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to..., Lowenstein [/bib_ref]. In TCR-stimulated cells, there is a rapid formation of a complex between Sos/Grb2 and a 36-kD membrane protein that is substrate for the TCRinduced PTKs [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. This complex is analogous to the Sos/Grb2 epidermal growth factor receptor (EGFR) complex that localizes Sos to the plasma membrane in fibroblasts.
Little is known about the CD28 costimulatory signals relative to the wealth of knowledge about TCR signal transduction. CD28 signal transduction does appear to be controlled by PTKs and it has been described that CD28 regulates PLC3,1 tyrosine phosphorylation, Ptdlns hydrolysis, and intracellular Ca 2+ levels [bib_ref] CD28 receptor crosslinking induces tyrosine phosphorylation ofPLC gammal, Ledbetter [/bib_ref] [bib_ref] Signalling through CD28 T-cell activation pathway involves an inositol phospholipid-specific phospholipase C..., Nun [/bib_ref] [bib_ref] Synergybetween the T3/antigen receptor complex and Tp44 in the activation of human..., Weiss [/bib_ref]. However, calcium signaling pathways in T cells are sensitive to inhibition by the immunosuppressive drugs cyclosporin A (CsA) and FKS06 and one characteristic of CD28 costimulatory signals is that they are not susceptible to inhibition by CsA [bib_ref] T-cell proliferation involving the CD28 pathway is associated with cyclosporine-resistant interleukin 2..., June [/bib_ref]. The role of Ras in CD28 signaling has never been studied, although recent studies have shown unequivocally that p2Y a~ function is essential for T cell activation. Moreover, T cells transfected with v-Ha-ras, a constitutively activated Ras mutant, are hyperresponsive to TCR ligation [bib_ref] Downward. 1992. p21ras mediates control of IL2 gene promoter function in T..., Rayter [/bib_ref] [bib_ref] p21ras function is important for T cell antigen receptor and protein kinase..., Woodrow [/bib_ref]. This latter observation in particular has prompted us to examine the effect of CD28 on the p21TM pathway.
The present data show that there is the potential for CD28 to couple to p21TM because ligation of CD28 with specific antibodies induces the accumulation of Ras/GTP complexes, the hyperphosphorylation of a MAPkinase kinase kinase (MAPKKK) Raf-1 and the activation of a MAP kinase ERK2. CD28 antibodies also induce the tyrosine phosphorylation of Vav and the Sos/Grb2-associated protein, p36. In contrast, the costimulatory interaction between CD28 and its natural ligand B7-1 did not induce the Ras signaling pathway. B7-1 was able to induce tyrosine phophorylation of Vav and the failure to activate Ras was correlated with a failure of B7-1 to induce tyrosine phosphorylation of p36. Studies of Vav tyrosine phosphorylation revealed that TCR ligation induces a relatively weak, rapid, and transient phosphorylation of Vav.
In contrast, the Vav tyrosine phosphorylation monitored in B7-1-activated cells is markedly stronger and prolonged.
# Materials and methods
Cell Lines. The subclone of theJurkat T leukemia cells, J.H6.2 was maintained in RPMI 1640 supplemented with 10% heatinactivated FCS. TheJurkat J.H6.2 cell line has been described [bib_ref] CD28 mAbs with distinct binding properties differ in their ability to induce..., Nun [/bib_ref] and this cell line does not express the other ligand for B7 molecules, CTLA-4 (data not shown), huB7 + L cells were obtained by transfection of the human B7-1 cDNA in routine L cells (A. Truneh, manuscript in preparation) and selected by G418 resistance. These cells were maintained in DMEM supplemented with 10% FCS, penicillin, streptomycin, and glutamine. For T cell stimulations, L cells were detached from the tissue culture flasks by incubation with trypsin/EDTA, washed once in PBS then once in RPMI medium, and resuspended in RPMI medium. Untransfected L cells were used as control cells.
Antibodies and Fusion Proteins. mAb UCHT1 reactive with the e chain of the human TCR/CD3 antigen complex and mAb CD28.2 or CD28.5 reactive with the human CD28 molecule (34) were purified from hybridoma supernatants by protein A-Sepharose. To stimulate cells, UCHT1 or CD28.2 were used at 10 #g/ml and CD28.5 at 50 #g/ml. The ascites of the anti-B7 mAb 104 is a kind gift ofJ. Banchereau (Schering Plough, Dardilly, France) and was used at 1 in 100.
The mAb Y13-259 was used to immunoprecipitate p21TM (17). The rabbit antiserum 66 was used to detect p74 '~q in immunoblottingand the rabbit antiserum 122 was used to immunoprecipitate and to detect p42 ~k-~. The monoclonal antiphosphotyrosine antibody 4G10 was purchased from Upstate Biotechnology Inc. (Lake Placid, NY). The monoclonal anti-Grb2 antibody was purchased from AFFINITI (Nottingham, UK). The antipeptide antiserum VAV-1 used to immunoprecipitate and detect in immunoblotting the p95 ray protein was described previously [bib_ref] Product of the vav proto-oncogene defines a new class of tyrosine protein..., Bustelo [/bib_ref].
The fusion protein encoding glutathion-s-transferase (GST) mSosl COOH terminus (residues 1135-1336) was coupled to glutathione beads and used to purify Grb2 as described [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref] [bib_ref] Association of Sos Ras exchange factor protein with Grb2 is implicated in..., Egan [/bib_ref].
Assay ofp2I ~s Activation. p21 ras proteins were immunoprecipitated with the antibody Y13-259 from cells in which guanine nucleotides were labeled biosynthetically with [32p]orthophosphate (17). Labeled guanine nucleotides bound to p21 r~ proteins were eluted, separated by TIC, and then quantitated by direct scanning for B emissions (AMBIS Systems Inc., San Diego, CA). Results are expressed as the percentage of p21 r~ proteins bound to GTP relative to total guanine nucleotides complexed to the protein.
Immunoloreciloitation and Immunoblotting. The Jurkat T cells were removed from culture, washed twice in RPMI 1640, and suspended at 4 x 106 cells/ml in RPMI medium. The cells were prewarmed at 37~ for 3 rain. Cells were stimulated at 37~ then pelleted in a microcentrifuge, and lysed in a buffer containing 50 mM Hepes (pH 7.4), 1% NP-40, 150 mM NaC1, 20 mM NaF, 20 mM iodoacetamide, 1 mM PMSF, 1 #g/ml protease inhibitors (leupeptin, pepstatin A, chymotrypsin), and 1 mM Na3VO4.
Acetone-precipitated proteins from cell lysates corresponding to 4 x 106 Jurkat cells were resolved in either 10% SDS-PAGE (for detection of ERK2) or 8% SDS-PAGE standard gels (for detection of other proteins).
For immunoprecipitation, lysates were clarified and incubated with purified anti-Vav antibody coupled to protein A-Sepharose or 5/xg of GST fusion protein immobilized on glutathione beads for 2 h at 4oC. Immunoprecipitates or precipitates were washed four times in 1 ml of lysis buffer and then boiled in reducing SDS gel sample buffer for 5 min. Samples were resolved by 8% SDS-PAGE standard gels.
For immunoblotting, membranes were blocked and probed with specific antibodies. Blots were then incubated with the appropriate second antibodies, anti-rabbit IgG or anti-mouse IgG (Amersham International, Little Chalfort, UK), all conjugated with horseradish peroxidase. Immunoreactive bands were visualized by enhanced chemiluminescence (Amersham International).
Analysis of MAP Kinase Activity in ERK2 Immunoprecipitates. ERK2 was immunoprecipitated with a specific antiserum and kinase activity of ERK2 assessed as described previously. Briefly 4 x 106 Jurkat T cells were stimulated at 37~ as indicated, then pelleted and lysed in a buffer containing 50 mM Hepes (pH 7.4), 1% NP-40, 150 mM NaC1, 20 mM NaF, 20 mM iodoacetamide, 1 mM PMSF, 1 #g/ml protease inhibitors (leupeptin, pepstatin A, chymotrypsin), and 1 mM Na3VO4. After 10 min of preclearing with protein A insoluble suspension (Sigma Chemical Co., St. Louis, MO), then 10 min of preclearing with protein A-Sepharose beads (Pharmacia LKB, Uppsala, Sweden), lysates were incubated for 2 h at 4~ with 10 #1 of the 122 antiserum precoupled to 50/zl of a 50% suspension of protein A-Sepharose beads. Immunoprecipitates were washed three times with lysis buffer and once with kinase assay buffer (30 mM Tris [pH 8], 20 mM MgC12, and 2 mM MnC12). In vitro kinase assay buffer was carried out for 30 rain at 30~ in 20/~1 of kinase assay buffer supplemented with 10 #M ATP, 5/zCi y-[32p]ATP, and 10 #g myelin basic protein (MBP) as a substrate. The kinase reaction was stopped with 20/A of 2 x SDS sample buffer and samples were run in 15% SDS-PAGE gels. Quantitation of 32p incorporated into the MBP protein band was done by/3 radiation scanning of dried gels using a scanner (AMBIS Systems Inc.).
## Ii~2 production. jurkat cells were washed several times in cul-
ture medium, then resuspended at a concentration of 0.5 x 106/ml. Duplicate cultures (200/zl) in 96-well flat-bottomed plates were set up in the presence of saturating concentrations of mAbs or PHA (10 #g/ml). Purified PHA was supplied by Industries Biologiques Fran~aise (Villeneuve-la-Garenne, France). The L cells were used at a concentration of 0.25 x 105/ml. After 18 h, supernatants were collected and stored at -80~ for IL-2 assay. IL-2 concentrations were measured by ELISA, using a commercially available kit (Immunotech S.A., Marseille, France) according to the manufacturer's instructions.
# Results
CD28 Triggering Activates p2I ras in Jurkat Cells. T cells can be activated via the CD28 pathway by triggering cells with CD28 antibodies. To determine whether the p21 ras guanine nucleotide binding cycle is regulated via the CD28 molecule, p21r~/GTP loading experiments were performed. Endogenous guanine nucleotides were metabolically labeled with [32p]orthophosphate and p21 ~ protein was immunoprecipitated from Jurkat cells exposed to antibodies against CD28 or the CD3/TCR complex, and the phorbol ester phorbol 12, 13-dibutyrate (PDBu) that directly activates PKC. The mAb used to trigger the CD3/TCR complex was UCHT1. Two CD28 mAbs were used, CD28.2, which can induce a IL-2 secretion in Jurkat cells and CD28.5, which is nonfunctional in this respect, and that apparently binds to CD28 without triggering CD28 signal transduction [bib_ref] CD28 mAbs with distinct binding properties differ in their ability to induce..., Nun [/bib_ref]. The proportion of p21 ~a~ bound to GTP (the ratio be-1069 Nun,s et al.
tween GTP and GTP plus GDP-labeled nucleotides on p21 TM) was evaluated by 3 emission scanning of TIC-resolved guanine nucleotides eluted from immunoprecipitated ras proteins. [fig_ref] Figure 1: p21 ras is activated in Jurkat cells stimulated by the CD28 mAb... [/fig_ref] A shows that PKC stimulation with PDBu increased the level of GTP bound to ras from 11% in unstimulated Jurkat cells to 40% (fourfold). CD28 triggering with CD28.2 resulted in a threefold increase in the proportion of GTP-bound ras (from 11 to 29%), this increase is similar to the increase after CD3/TCR triggering with UCHT1. The CD28 mAb, CD28.5 did not induce an increase in the proportion of GTP-bound ras. The kinetics of the increase in the proportion of GTP-bound ras are shown in [fig_ref] Figure 1: p21 ras is activated in Jurkat cells stimulated by the CD28 mAb... [/fig_ref] C. The TCR response peaked after 1 min, whereas the CD28 response was maximal at 5 min. Both TCR and CD28 responses could still be detected 30 min after CD3/TCR or CD28 stimulation.
CD28 Triggering Induces the Phosphorylation of ERK2 and RAE In many cell systems, p21 ras couples receptor stimulated PTKs to Raf-1 and hence to the MAP kinase cascade. The predominant MAP kinase in T cells is ERK2 and p21 ras plays a crucial role in the TCR/ERK2 link. The activation of ERK2 can be monitored by Western blot analysis with an ERK2-specific antiserum because the activation of this kinase requires its phosphorylation on threonine and tyrosine residues and the phosphorylated "active" ERK2 has a reduced mobility on SDS-PAGE gels compared with nonphosphorylated, "inactive" ERK2. The experiment in [fig_ref] Figure 2: CD28 /ml [/fig_ref] shows the reduced mobility of ERK2 in Jurkat cells treated with the PKC activator, PDBu, a CD3 mAb UCHT1 and CD28 mAbs compared with control unstimulated cells. Stimulation with UCHT1 or CD28.2 rapidly induced a mobility shift in the band corresponding to ERK2 that was maximal between 5 and 30 min of exposure to the antibody. PDBu also induced a rapid shift in ERK2 mobility in Jurkat cells. Previous studies have shown that Raf-1 is activated rapidly in TCR-and phorbol ester-stimulated T cells [bib_ref] Rapid activation of c-Raf-1 after stimulation of the T-cell receptor or the..., Siegd [/bib_ref]. It is also well documented that Raf-1 is hyperphosphorylated in response to T cell activation [bib_ref] Interleukin-2-triggered Raf-1 expression, phosphorylation, and associated kinase activity increase through G1 and..., Zmuidzinas [/bib_ref] [bib_ref] T cell antigen receptor engagement stimulates c-rafphosphorylation and induces c-raf kinase activity..., Siegel [/bib_ref] and this hyperphosphorylation of Raf-1 correlates with activation of the enzyme although it is apparently not directly involved in mediating the increase in Raf-1 enzymatic activity [bib_ref] Conditional transformation of cells and rapid activation of the mitogen-activated protein kinase..., Samuels [/bib_ref]. The stimulation of Raf-1 has been recently shown in peripheral blood T cells after CD28 mAb cross-linking [bib_ref] Rapid activation of c-Raf-1 after stimulation of the T-cell receptor or the..., Siegd [/bib_ref]. Hyperphosphorylated Raf-1 has reduced mobility on SDS-PAGE gels hence to determine whether Raf-1 is hyperphosphorylated in Jurkat cells, a Western blot of cell lysates prepared from cells treated with UCHT1, CD28 mAbs, or PDBu was probed with a Raf-1 antisera. The data in [fig_ref] Figure 2: CD28 /ml [/fig_ref] show the UCHT1, CD28.2, and PDBu induce a shift in Raf-1 mobility that is comparable with the ERK2 response. The CD28 mAb, CD28.5, which does not costimulate T cells, did not induce a detectable hyperphosphorylation of Raf-1 or ERK2.
## B7-1-transfected l cells induce tyrosine phosphorylation of cellular substrates and il-2 secretion, but not ras activation or raf and erk2 phosphorylations.
It has been shown that the natural ligand for CD28, B7-1, expressed on the surface of Chinese hamster ovary cells can provide costimulation for T cells, replacing the effects seen with CD28 mAbs (9). To determine whether IL-2 secretion could be induced by ligation of CD28 with B7-1 inJurkat cells, the culture supernatants from cells stimulated with P H A in the presence of L cells transfected with human B7-1 (huB7 + L cells) or CD28.2 mAb were collected and assayed for IL-2 production. A small amount of IL-2 was detected when the Jurkat cells were stimulated with PHA alone [fig_ref] Figure 3: 1I [/fig_ref]. When CD28.2 or huB7 § L cells were added to PHA-stimulated Jurkat cells, IL-2 secretion was markedly increased (20-fold). The addition of anti-B7 mAb 104 blocked the stimulation of IL-2 production by huB7 + L cells.
The experiment in [fig_ref] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of... [/fig_ref] A compares the effect of huB7 + L cells, U C H T 1 , CD28.2, and PDBu on Ras/GTP loading. The data show that PDBu, U C H T 1 , and CD28.2 induced an increase in cellular levels of active Ras/GTP complexes, whereas huB7 + L cells have no effect. In parallel experiments [fig_ref] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of... [/fig_ref] , the phosphorylation of Raf-1 and ERK2 in B7-1-stimulated T cells was monitored. UCHT1, CD28.2, or PDBu induced an increase of the apparent molecular weight of p74 raf-t or p42 erk2. However, over a 30-min time course, there was no detectable effect of huB7 + L cells on Raf-1 or ERK2 electrophoretic mobility.
To examine the consequences of CD28 ligation on ERK2 activity, MBP kinase assays were carried out on MAP kinase immunoprecipitates. The data in [fig_ref] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of... [/fig_ref] C show that UCHT1, PDBu, and CD28.2 induced an approximate two-to threefold increase in ERK2 kinase activity in Jurkat T cells, whereas huB7 § L cells have no effect.
It has been described previously that CD28 ligation activates cellular PTKs [bib_ref] Antibody and B7/BBl-mediated ligation of the CD28 receptor induces tyrosine phosphorylation in..., Vandenberghe [/bib_ref] [bib_ref] CD28-induced T cell activation. Evidence for a protein-tyrosine kinase signal transduction pathway, Lu [/bib_ref]. Accordingly, we reprobed the Raf-1 and ERK2 blots with an antiphosphotyrosine antibody. The data in [fig_ref] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of... [/fig_ref] show that a 95-kD tyrosine phosphoprotein was detected in Jurkat cells within 5 min after contact of the cells with huB7 + L cells. The results of the antiphosphotyrosine Western blot show that the failure to detect an effect of huB7 § L cells on Raf-1 or ERK2 was not due to the failure of B7-1 to trigger CD28. In these experimental conditions, we can not detect tyrosine phosphoproteins in lysates from L cells alone and the pattern of tyrosine phosphorylations in Jurkat cells was the same with paraformaldehyde-fixed huB7 + L cells or not fixed huB7 + L cells (data not shown). The 95-kD tyrosine phosphoprotein This experiment was performed three times with identical results.
was also detected after stimulation of T cells with UCHT1 or CD28.2. However, the pattern and intensity of tyrosine phosphorylation in TCR-or CD28-activated T cells was different. Some tyrosine phosphoproteins were common to TCR-and CD28-induced cells but in general the level of phosphorylation was stronger in the TCR-activated cells. There was one tyrosine phosphoprotein of approximately 65 kD that was detected in the B7-1 stimulated cells but not in TCR-stimulated Jurkat. This molecule could be a unique substrate for CD28-activated PTKs.
## Stimulation of jurkat cells through the cd28 molecule by a specific mab or the ligand b7-1 induces a tyrosine phosphorylation of vav.
One candidate for the 95-kD tyrosine phosphoprotein in the TCR-and CD28-activated cells is p95 Vav, which is known to be tyrosine-phosphorylated in response to TCR triggering [bib_ref] Product of the vav proto-oncogene defines a new class of tyrosine protein..., Bustelo [/bib_ref] [bib_ref] Tyrosine phosphorylation of vav proto-oncogene product containing SH2 domain and transcription factor..., Margolis [/bib_ref]. The stimulation of p21 ra' requires cellular PTK function, and since Vav has been described to act as a guanine nucleotide exchange protein for ras in vitro, it has been proposed that Vav tyrosine phosphorylation is an important event for Ras regulation [bib_ref] Tyrosine kinase-stimulated guanine nucleotide exchange activity ofvav in T cell activation, Gulbins [/bib_ref]. We therefore compared the tyrosine phosphorylation of Vav in TCR-or CD28activated cells. For these experiments, antiphosphotyrosine immunoblots were performed on Vav immunoprecipitates. In the experiment shown in A, a weak basal tyrosine phosphorylation of Vav was detectable in Jurkat cells. UCHT1 and CD28.2 both induced a rapid increase in Vav tyrosine phosphorylation that was maximal at 1 min in the TCRactivated cells and 5 min in the CD28.2-stimulated population. There was no change in Vav tyrosine phosphorylation in cells treated with CD28.5.
The data in B show the effect of the CD28 ligand B7-1 on the tyrosine phosphorylation of Vav. In this experi-ment, the immunoblots were probed with an antiphosphotyrosine antibody and then reprobed with an anti-Vav antibody to control the presence of Vav in all immunoprecipitates. The data show that huB7 + L cells do not express Vav but are able to induce a strong tyrosine phosphorylation of Vav in T cells, that was detectable within 30 s and maintained for 10 min. Untransfected L ceils did not have this effect. The level of tyrosine phosphorylated Vav in cells triggered by B7-1 was markedly higher than the maximal phosphorylation seen in cells activated with UCHT1 or CD28.2 . In addition, the increase in Vav tyrosine phosphorylation was a more prolonged response in the B7-1-stimulated cells.
## The cooh-terminal domain of sos complexes with the p36 tyrosine phosphoprotein injurleat cells stimulated by cd28 mab~ but not by b7-1 ligation. very recently, it was shown
that TCR activation of ras might be controlled, at least in part, by a mechanism involving the formation of a complex of the Ras exchange protein Sos, Grb2, and a membranebound protein of 36 k_D that is tyrosine-phosphorylated in TCR-activated cells [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. The data in compare the ability of CD28 ligation and TCR triggering to induce the tyrosine phosphorylation of the 36-kD Grb2-associated protein.
Grb2 is composed of two SH3 domains that bind to Sos and an SH2 domain that interacts with tyrosine phosphoproteins. To look at the ability of tyrosine-phosphorylated proteins to interact with the Grb2-SH2 domain, we have used a GST fragment of the proline-rich COOH-terminal domain of Sos (amino acids 1135-1336) immobilized on glutathione agarose beads to precipitate Grb2-and Grb2-binding proteins from the lysates of resting and stimulated Jurkat cells. The Grb2 complexes are then analyzed by Western blotting with an antiphosphotyrosine antibody. The GST/Sos fragment used in these experiments binds to the SH3 domains of Grb2 and competes for the binding of endogenous Sos and thereby allows the precipitation of endogenous Grb2 and any proteins complexed to the Grb2-SH2 domain [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. The interaction between Grb2 and GST fusion protein of the COOH-terminal region of Sos is shown from Jurkat cell lysates, using an anti-Grb2 antibody in immunoblotting . W h e n the Grb2 precipitates are probed with an-tiphosphotyrosine antibodies in a Western blot analysis , a 36-kD tyrosine phosphoprotein, p36, is seen in the Sos/Grb2 complexes isolated from TCR-stimulated Jurkat cells, p36 is also present in the Sos/Grb2 complex after stimulation with CD28.2, but was not detectable when the Jurkat cells were unstimulated. The kinetics of p36 tyrosine phosphorylation in TCR-stimulated cells are comparable with the kinetics of Vav tyrosine phosphorylation. The huB7 + L cells . CD28 mAb CD28.2, but not B7-1 ligation induces tyrosine phosphorylation of a 36-kD tyrosine protein that binds to the Grb2 SH2 domain. Jurkat cells (8 x 106) were stimulated with UCHT1 (10/zg/ml), CD28.2 (10/xg/ml), and L cells expressing B7-1 (huB7 + L cells) or not expressing B7-1 (L cells) (at a ratio of 1:2) for the indicated time. Proteins were precipitated from the postnuclear cell lysates with GSTmSosl-COOH terminus fusion protein immobilized on glutathione beads. Bound proteins were analyzed with antiphosphotyrosine 4G10 antibody (top) or anti-Grb2 antibody (bottom). p36 is indicated by an arrowhead. . CD28 mAb CD28.2 or B7-1 ligation induces rapid tyrosine phosphorylation of the ray protooncogene product in Jurkat cells. Vav immunoprecipitates were performed from NP40-soluble cellular proteins. (A) Jurkat cells were stimulated with UCHT1 (10 ~g/ml), CD28.2 (10/~g/ml), and CD28.5 (50/.tg/ml) for the indicated time. (B) Jurkat cells were stimulated with UCHT1 (10/tg/ml), CD28.2 (10/~g/ml), and L cells expressing B7-1 (huB7 + L cells) or not expressing B7-1 (L cells) (at a ratio of 1:2) for the indicated time. In the penultimate lane orB, a Vav immunoprecipitate was performed on a lysate of 2 x 106 huB7 + L cells alone, p95 v~v was detected with antiphosphotyrosine 4G10 antibody (top, A and B) or anti-Vav antibody, [fig_ref] Figure 1: p21 ras is activated in Jurkat cells stimulated by the CD28 mAb... [/fig_ref]. trigger a strong tyrosine phosphorylation of Vav but did not induce the tyrosine phosphorylation of p36. The data in show antiphosphotyrosine immunoblots of Grb2 isolated from Jurkat cells exposed to huB7 + L cells for 10-20 min. Analysis of earlier time points also failed to detect B7-1-induced tyrosine phosphorylation of p36 (data not shown).
# Discussion
The stimulation of the CD28 molecule delivers important costimulatory signals for T cell activation (3). CD28 signaling pathways and the biochemical basis of this costimulatory function are poorly understood. Recent studies have shown that p21TM proteins have an important function in T C R signal transduction [bib_ref] Downward. 1992. p21ras mediates control of IL2 gene promoter function in T..., Rayter [/bib_ref] [bib_ref] p21ras function is important for T cell antigen receptor and protein kinase..., Woodrow [/bib_ref]. The object of the present study was to determine whether p21 ras is involved in CD28 signaling. The activation of p21 ras involves the accumulation of Ras-GTP complexes. Accordingly, the experiments herein compared the ability of CD28 mAbs or the human B7-1 molecule to stimulate p21 TM and p2V ~ dependent functions such as the hyperphosphorylation of Raf-1 and the activation of the MAP kinase ERK2. The data presented show that there is the potential for CD28 to regulate the Ras signaling pathway. Thus CD28.2, a functionally agonistic CD28 mAb, induces a rapid conversion of p 2 P f f G D P to p21~s/GTP. CD28.2 also induces a hyperphosphorylation of a MAPKKK, Raf-1, and stimulates the MAP kinase ERK2. One natural ligand for CD28 is the B7-1 molecule which, when expressed in fibroblasts, can provide costimulatory signals for IL~ secretion in T ceUs (9). The present data show that CD28 interactions with the B7-1 molecule do not induce any increase in the accumulation of p21r~s-GTP. Moreover, huB7 + L cells were not able to induce a hyperphosphorylation of Raf-1 or ERK2 activation in T cells. Collectively the GTP loading data and the Raf-1, ERK2 phosphorylations and ERK2 activation data indicate that unlike the CD28 antibody CD28.2, B7-1 does not initiate a Ras signaling pathway.
TCR stimulation of p21 r~s is dependent on tyrosine kinase function [bib_ref] Role of protein kinase C in T-cell antigen receptor regulation of p21ras:..., Izquierdo [/bib_ref]. Two substrates for TCR. stimulated PTKs have been identified that may mediate the PTK effects on Ras, p95 ray and p36, a protein that complexes to the SH2 domain of the Grb2 adapter molecule, thus recruiting the guanine nucleotide exchange protein Sos [bib_ref] Tyrosine kinase-stimulated guanine nucleotide exchange activity ofvav in T cell activation, Gulbins [/bib_ref] [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. CD28 ligation with both the CD28 mAb CD28.2 and the ligand B7-1 apparently activated cellular PTK pathways as judged by Western blot analysis with an antiphosphotyrosine antibody. We therefore explored whether the differences between CD28.2 and B7-1 with regard to p21 ras activation could be explained by differences in the ability of these CD28 ligands to modulate the tyrosine phosphorylation of Vav or p36.
The data show that Vav is tyrosine phosphorylated in T cells activated via the TCR or CD28 irrespective of whether the CD28 mAb CD28.2 or the ligand B7-1 was used to trigger the CD28 receptor. The overall intensity of tyrosine phosphoproteins in the TCR-activated cells was markedly higher than in the B7-1-activated cells. In particular, in the antiphosphotyrosine blots of total cell lysates a 95-kD protein that migrates at a similar position to Vav is strongly tyrosine phosphorylated in TCR-activated cells and more weakly in T cells activated with huB7 + L cells [fig_ref] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of... [/fig_ref]. This band probably represents multiple 95-kD tyrosine phosphoproteins because when Vav tyrosine phosphorylation was monitored directly, the relative effect of the TCR and B7-1 on Vav phosphorylation was the opposite of that suggested by the analysis of the 95-kD protein in total cell lysates. For direct analysis of Vav tyrosine phosphorylation, Vav immunoprecipitates were probed with an antiphosphotyrosine antibody. These data confirm previous observations that triggering of the TCR stimulation induces a rapid but transient tyrosine phosphorylation of Vav in T cells [bib_ref] Product of the vav proto-oncogene defines a new class of tyrosine protein..., Bustelo [/bib_ref] [bib_ref] Tyrosine phosphorylation of vav proto-oncogene product containing SH2 domain and transcription factor..., Margolis [/bib_ref]. The ligand for CD28, B7-1, also induced Vav tyrosine phosphorylation but with an apparently higher stoichiometry than the TCR stimulus. Moreover, compared with the TCR response, the B7-1induced tyrosine phosphorylation of Vav was a rapid but prolonged event. Previous studies have identified a 100-kD protein (ppl00) as a substrate for CD28-activated PTKs in PMA-treated Jurkat cells [bib_ref] Antibody and B7/BBl-mediated ligation of the CD28 receptor induces tyrosine phosphorylation in..., Vandenberghe [/bib_ref] [bib_ref] CD28-induced T cell activation. Evidence for a protein-tyrosine kinase signal transduction pathway, Lu [/bib_ref]. It is not known whether this ppl00 protein is Vav or another substrate. The 95-kD tyrosine substrate seen in the present study may be the same as ppl00. In the antiphosphotyrosine immunoblots of total lysates, a 65-kD protein appeared selectively in the B7-1activated cells, which supports the hypothesis that CD28 and the TCR may regulate different PTKs. The present study illustrates a second major difference between TCR and CD28/PTK pathways by a difference in the magnitude and kinetics of p95 va~ tyrosine phosphorylation.
Tyrosine phosphorylation of Vav regulates its in vitro guanine nucleotide exchange activity for Ras [bib_ref] Tyrosine kinase-stimulated guanine nucleotide exchange activity ofvav in T cell activation, Gulbins [/bib_ref]. However, one conclusion from the present data is that Vav tyrosine phosphorylation is not sufficient to ensure activation of p21 r~s. In vivo, this does not necessarily exclude that Vav tyrosine phosphorylation plays a role in Ras regulation. First, it is possible that regulation ofVav GTP exchange activity requires tyrosine phosphorylation of particular tyrosine residues and in the present study it was not established that Vav is phosphorylated on identical sites in the TCR-and CD28-activated cells. Second, Vav tyrosine phosphorylation may be important for Ras regulation but there may be a simultaneous requirement for other signals that are triggered by the TCR and not by CD28.
Sos, another guanine exchange factor for Ras, is detected in T cells and interacts with an adaptor protein Grb2 [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref] [bib_ref] Interaction of Shc with the zeta chain of the T cell receptor..., Ravichandran [/bib_ref]. The complex of Sos/Grb2 is present in resting and active T cells (23) and the crucial regulatory step is the recruitment of the Grb2/Sos complex to the plasma membrane. Grb2 is translocated to the membrane by interactions between the Grb2-SH2 domain and tyrosine-phosphorylated proteins. The first example of this phenomenon was in fibroblasts, where EGF promotes the formation of a heterotrimeric complex of EGF receptor, Grb2 adaptor protein and Sos [bib_ref] Association of Sos Ras exchange factor protein with Grb2 is implicated in..., Egan [/bib_ref] [bib_ref] Epidermal growth factor regulates p21ras through the formation of a complex of..., Buday [/bib_ref]. Upon TCR triggering a tyrosine-phosphorylated p36 protein complexes to the Grb2-SH2 domain and it is proposed that this p36 interaction recruits Grb2 and Sos to the membrane. The current data show that T cell activation with the CD28 mAb CD28.2 can also induce tyrosine phosphorylation of a p36 Grb2-associated protein. However, stimulation of T cells with huB7 + L cells does not have this effect. Thus the Grb2-associated p36 protein is not a substrate for B7-1-stimulated PTKs. Furthermore the tyrosine phosphorylation of p36 appears to correlate with activation of the Ras signaling pathway which supports the hypothesis that p36 is crucial for Ras activation.
In fibroblasts another protein that can complex with the Grb2 SH2 domain is Shc [bib_ref] Association of Sos Ras exchange factor protein with Grb2 is implicated in..., Egan [/bib_ref]. Shc can bind to the TCR-~" chain and when it is tyrosine phosphorylated in TCR-activated cells, the formation of a Grb2/Shc complex has been observed [bib_ref] Interaction of Shc with the zeta chain of the T cell receptor..., Ravichandran [/bib_ref]. In CD3-stimulated T cells, it seems difficult to detect tyrosine-phosphorylated Shc in association with Sos (23). This seems to reflect that the stoichiometry of Shc tyrosine phosphorylation in TCR activated cells is low. Thus the only tyro- sine phosphoprotein detectable in Grb2/Sos complexes is a membrane-bound protein of 36 kD [bib_ref] A complex of Grb2 adaptor protein, Sos exchange factor and a 36..., Buday [/bib_ref]. In the present study, we did examine the effect of CD28.2 and B7-1 on Shc tyrosine phosphorylation. CD28 mAb CD28.2 induced a level of Shc tyrosine phosphorylation that was lower than that seen in TCR-activated cells. It proved difficult to look at Shc phosphorylation in B7-1-stimulated T cells since the huB7 + L cells expressed constitutively tyrosine phosphorylated Shc (data not shown). [fig_ref] Table 1: A Comparison of Signal Transduction Pathways by a CD28 mAb CD28 [/fig_ref] shows the signal transduction events triggered by a CD28 mAb and the ligand B7-1. One conclusion is that Ras activation can be initiated when the CD28 molecule is triggered with a CD28 mAb but not the ligand B7-1. One explanation for the discrepancy in the intracellular signals triggered by the CD28 mAb CD28.2 and the ligand B7-1, is that the CD28 mAb induces an artificially high degree of aggregation and cross-linking of the CD28 molecule that triggers a nonphysiological response. There are several precedents for signal transduction artifacts resulting from the inappropriate cross-linking of receptor molecules [bib_ref] Does co-aggregation of the CD45 and CD3 antigens inhibit T cell antigen..., Shivnan [/bib_ref]. An alternative possibility is that the CD28 mAb CD28.2 reveals the full signaling potential of the CD28 receptor, whereas the ligand B7-1 triggers one very important signaling pathway in T cells but it does not necessarily trigger the full stimulatory potential of the CD28 receptor. Recently, another ligand for CD28 has been described, B7-2 (5-8). It has not yet been determined whether the ligands B7-1 and B7-2 trigger common biochemical signals in cells. It is thus possible that the data obtained with the CD28 mAb CD28.2 reflect that B7-2 or indeed another ligand for CD28, as yet undefined, could use Ras-dependent signals to control T cell activation. It should be emphasized, however, that the CD28 costimulatory pathway for IL-2 gene expression is induced when CD28 is triggered with either the CD28 mAb CD28.2 or the ligand B7-1. Hence, the observation that CD28 ligation with antibodies can initiate a p21=s-dependent signaling cascade, whereas the natural ligand B7-1 cannot, suggests that p21 ras activation does not explain the costimulatory effects of CD28 in terms of signal transduction. CD28 ligation with antibodies or B7-1 induces Vav tyrosine phosphorylation and if the relative tyrosine phosphorylation of Vav in B7-1 vs. TClkactivated cells has a functional purpose, then the data suggest that Vav will be more important for CD28 signal transduction than for TCR signals.
We would like to thank J. Downward, L. Buday, S. Egan, C. J. Marshall, and P. Parker for providing antibodies and GST fusion proteins.
This work was supported in part with funds from an Association de la Recherche contre Le Cancer (ARC) fellowship (to J. Nun,s).
[fig] Figure 1: p21 ras is activated in Jurkat cells stimulated by the CD28 mAb CD28.2. (A) TLC of the nucleotides eluted from immunoprecipitates of p21 r~ from [32p]orthophosphate-labded Jurkat cells. The ceils were unstimulated (control) or stimulated for 10 rain with a CD3 mAb UCHT1 (10/zg/ml), CD28 mAbs, (CD28.2 [10 gg/ml] or CD28.5 [50 gg/ml]), and PDBu (100 ng/ml) before lysis and immunoprecipitations of p21 r~s with Y13-259 as described. The position at which GTP and GDP standards ran is indicated. (B) The figure shows the quantitation of guanine nucleotides of p21=s by direct scanning of 3 radiation from the TIC in A. Data are expressed as percentage of GTP to ras with respect to the total amount of guanine nucleotide on p21 r'~. The data given are representative of the results obtained from at least five experiments. (C) Kinetics of p21=~ activation after T cell activation with UCHT1 or CD28.2. Time-course showing the nudeotides bound to Ras upon stimulation of Jurkat cells with UCHT1 (10/~g/ml) or CD28.2 (10 #g/ml). Nucleotides were separated by TLC and quantitated by direct scanning for ~ radiation. [/fig]
[fig] Figure 2: CD28 /ml) for the indicated time. Whole cell lysates were probed with the Raft antibody 66 (top) or the ERK2 antibody 122, which recognizes p42 ~k3 (bottom). [/fig]
[fig] Figure 3: 1I.-2 production in response to CD28 mAb or B7-1-expressed L cells in PHA-stimulated Jurkat cells. Jurkat cells were stimulated by PHA alone or with CD28 mAb CD28.2 (10 ~g/ml) or B7-1-expressed L cells (huB7 + L cells at a ratio of 1:20). Controls were added using untransfected L cells or huB7 + L cells preincubated with anti-B7-1 mAb, 104 (1:100 of ascites) for 5 min. After 18 h, supernatants were assayed for Ib2 as described in Materials and Methods. The experiment shown is representative of three separate experiments. Results are presented as means +_ standard errors of the means of duplicate determinations. [/fig]
[fig] Figure 4: Interaction of CD28 with huB7 + L cells induces tyrosine phosphorylation of cellular proteins, but not Ras activation, Erk2 activation, Raf-1 and Erk2 phosphorylations. L cells were detached from tissue culture plates and washed once in PBS then twice in RPMI medium. Jurkat cells were unstimulated (control) or stimulated with UCHT1 (10/~g/ml), CD28.2 (10/~g/ml), PDBu (100 ng/ml) for 10 min and L cells expressing B7-1 (huB7 + L cells) or not expressing B7-1 (L cells) (at a ratio of 1:2) for the indicated time.(A) Before stimulation, Jurkat cells were labeled with [32p]orthophosphate. After stimulation, an assay of p21 ras activation was performed. The figure shows the quantitation of guanine nucleotides of p21 r~s by direct scanning of/3 radiation. Data are expressed as percentage of GTP to ras with respect to the total amount of guanine nucleotide on p21 ras. (B) Whole cell lysates were probed with the anti-Raft antibody 66 (top) or the anti-Erk antibody 122 (bottom). (C) Regulation of Erk2 in Jurkat cells. The cells were unstimulated (control) or stimulated with UCHT1 (10 #g/ml), CD28.2 (10/zg/ml), PDBu (100 ng/ml), L cells (at a ratio of 1:2) for 10 min and L cells expressing B7-1 (huB7 + L cells) (at a ratio of 1:2) for the indicated time. A MBP phosphorylation assay was performed as indicated in Materials and Methods. Data are expressed as cpm of 32p incorporated into MBP as assessed by Ambis /3 scanning. (D) Immunoblots from B were stripped of antibody and redevetoped with the antiphosphotyrosine 4G10 antibody. The position of molecular mass markers are shown on the left; arrowheads indicate the position of 65-and 95-kD phosphoproteins. [/fig]
[table] Table 1: A Comparison of Signal Transduction Pathways by a CD28 mAb CD28.2 and the CD28 Ligand B7-1Rating scale: -, not detectable; +, positive. * Determined by altered molecular weight on SDS-PAGE gels. * Determined by MBP phosphorylation assay. [/table]
|
Structural similarity-based predictions of protein interactions between HIV-1 and Homo sapiens
Background: In the course of infection, viruses such as HIV-1 must enter a cell, travel to sites where they can hijack host machinery to transcribe their genes and translate their proteins, assemble, and then leave the cell again, all while evading the host immune system. Thus, successful infection depends on the pathogen's ability to manipulate the biological pathways and processes of the organism it infects. Interactions between HIV-encoded and human proteins provide one means by which HIV-1 can connect into cellular pathways to carry out these survival processes.Results:We developed and applied a computational approach to predict interactions between HIV and human proteins based on structural similarity of 9 HIV-1 proteins to human proteins having known interactions. Using functional data from RNAi studies as a filter, we generated over 2000 interaction predictions between HIV proteins and 406 unique human proteins. Additional filtering based on Gene Ontology cellular component annotation reduced the number of predictions to 502 interactions involving 137 human proteins. We find numerous known interactions as well as novel interactions showing significant functional relevance based on supporting Gene Ontology and literature evidence.Conclusions:Understanding the interplay between HIV-1 and its human host will help in understanding the viral lifecycle and the ways in which this virus is able to manipulate its host. The results shown here provide a potential set of interactions that are amenable to further experimental manipulation as well as potential targets for therapeutic intervention.
# Background
Pathogen invasion and survival requires that the pathogen interact with and manipulate its host. Human immunodefficiency virus type 1 (HIV-1) encodes only 15 proteins and must therefore rely on the host cell's machinery to accomplish vital tasks such as the transport of viral components through the cell and the transcription of viral genes [bib_ref] Host factors exploited by retroviruses, Goff [/bib_ref] [bib_ref] HIV-1: fifteen proteins and an RNA, Frankel [/bib_ref]. HIV-1 infects human cells by binding to CD4 and a coreceptor, fusing with the cell membrane and uncoating the virion core in the cytoplasm [bib_ref] HIV-1: fifteen proteins and an RNA, Frankel [/bib_ref]. The genomic RNA is then reverse transcribed and the DNA enters the nucleus as part of a viral preintegration complex (PIC) containing both viral and host proteins. Afterwards, the viral DNA is inserted into the genome by viral integrase (IN) [bib_ref] Host factors exploited by retroviruses, Goff [/bib_ref]. The integrated provirus is transcribed by host RNA polymerase II from a pro-moter located in the provirus long terminal repeat (LTR), and the RNA is exported to the cytoplasm [bib_ref] Host factors exploited by retroviruses, Goff [/bib_ref] [bib_ref] HIV-1: fifteen proteins and an RNA, Frankel [/bib_ref]. Host machinery translates HIV-1 mRNA, and several of the resulting proteins are transported to the cell membrane to be packaged into the virion along with the genomic RNA and multiple host proteins. The virus then buds from the cell and undergoes a maturation process, which enables it to infect other cells [bib_ref] HIV-1: fifteen proteins and an RNA, Frankel [/bib_ref]. Throughout this process, host proteins play an indispensable role.
To understand the interface through which the pathogen connects with and manipulates its host requires knowledge of the molecular points of interaction between them. Specifically, knowledge of the protein interactions between pathogen and host is of particular value. While the prediction of protein interactions within species such as S. cerevisiae and H. sapiens has been pursued for some time, it is only recently that host-pathogen interactions have come under greater scrutiny. Indeed, computational approaches are of significant value in the host-pathogen context as large-scale experimental characterization of these interactions is non-trivial [bib_ref] Epstein-Barr virus and virus human protein interaction maps, Calderwood [/bib_ref] [bib_ref] Hepatitis C virus infection protein network, De Chassey [/bib_ref] [bib_ref] Systems biology and the host response to viral infection, Tan [/bib_ref].
As a result of the need for computational approaches, several recent methods have been developed and applied to host-pathogen interactions, suggesting additional potential interactions in different host-pathogen systems. For instance, interactions between P. falciparum and human using statistics about domains involved in within-species interactions [bib_ref] Computational prediction of hostpathogen protein protein interactions, Dyer [/bib_ref]. Also focusing on malaria, Lee and colleagues generated predictions based on interactions between orthologous proteins from eukaryotes [bib_ref] Orthologbased protein-protein interaction prediction and its application to inter-species interactions, Lee [/bib_ref]. In the context of HIV-human interactions, at least two computational methods have been applied. In the first study, Tastan et al. used a computational approach based on the random forest method to predict protein interactions using features taken from human proteins and the human interactome [bib_ref] Prediction of interactions between HIV-1 and human proteins by information integration, Tastan [/bib_ref]. In the second study, Evans et al. predicted possible interactions using short sequence motifs conserved in both HIV-1 and human proteins [bib_ref] Prediction of HIV-1 virus-host protein interactions using virus and host sequence motifs, Evans [/bib_ref].
While of value, most approaches have not utilized the significant amount of protein structure information that is increasingly available. Specifically, rapid progress in structure determination technologies has led to the establishment and deposition of massive numbers of protein structures into the Protein Data Bank, with over 60,000 protein structures currently deposited [bib_ref] Announcing the worldwide Protein Data Bank, Berman [/bib_ref]. In combination with documented protein-protein interactions, the use of protein structure information provides another means for the prediction of possible protein interactions [bib_ref] InterPreTS: protein interaction prediction through tertiary structure, Aloy [/bib_ref] [bib_ref] MULTIPROSPECTOR: an algorithm for the prediction of protein-protein interactions by multimeric threading, Lu [/bib_ref] [bib_ref] Protein complex compositions predicted by structural similarity, Davis [/bib_ref]. The central premise in such approaches is that, given a set of proteins with defined structures and associated interactions, proteins with similar structures or substructures will tend to share interaction partners. In the context of host-pathogen interactions, Davis et al., used homology modeling to ascertain potential protein interactions for pathogens responsible for several tropical diseases [bib_ref] Host pathogen protein interactions predicted by comparative modeling, Davis [/bib_ref]. Unfortunately, despite their potential value, such computational structure approaches have not been widely applied to the problem of predicting host-pathogen interactions.
Here, we develop a map of interactions between HIV-1 and human proteins based on protein structural similarity. In this approach, we first retrieve structural similarity between host and pathogen proteins identified by an established method which compares known crystal structures. Human proteins identified as having a region of high structural similarity to an HIV protein are referred to as "HIV-similar." Next, we identify known interactions for these HIV-similar proteins, with the one or more human proteins that they interact with referred to as "targets." We then assume that HIV proteins have the same interactions as their human, HIV-similar counterparts, allowing HIV to plug into the host cell protein network at these points [fig_ref] Figure 1: Diagram of approach [/fig_ref]. Using data from recent RNAi screens and cellular co-localization information, we refine this interaction map so as to enrich for those interactions having the greatest potential to be correct based on the available information. Evaluation of these predictions shows a statistically significant enrichment of known interactions as well as numerous novel interactions with potential functional relevance. These predictions provide an additional tool for further investigations into the lifecycle of HIV-1 and identification of potential clinical targets.
# Results and discussion
## Identification of hiv-similar human proteins
To construct a map of interactions between HIV-1 and human proteins, we established a multi-step protocol that begins with the identification of human proteins having significant structural similarity to HIV-1 proteins [fig_ref] Figure 2: Structural prediction workflow [/fig_ref]. We used the Dali Database [bib_ref] Searching protein structure databases with DaliLite v. 3, Holm [/bib_ref] [bib_ref] Protein Structure Comparison by Alignment of Distance Matrices, Holm [/bib_ref] , which contains 3D structure comparisons for all protein structures in the Protein Data Bank (PDB); all publicly available crystal structures for HIV-1 and H. Sapiens are contained within PDB. While the crystal structure for many human proteins is unknown, most HIV-1 proteins have been at least partially resolved. Specifically, crystal structures exist for PR, RT, IN, CA, MA, NC, Gag p2, gp120, gp41, Nef, Tat, Vpr, and Vpu [fig_ref] Table 1: HIV-1 protein structuresThe number of structures representing each HIV-1 protein in Dali [/fig_ref]. The three enzymes encoded by HIV-1, protease (PR), reverse transcriptase (RT), and integrase (IN) are the best characterized structurally, having at least 25 structures each in the PDB, with PR having over 300. CA, gp41, and gp120 are also fairly well studied. We note, however, that many of these structures represent only part of the full-length protein. HIV-1 proteins having regions of high similarity to at least one human protein include: gp41, gp120, CA, MA, Gag p2, PR, IN, RT, and Vpr (Additional File 1). Therefore, predictions were made for nearly every HIV-1 protein that has a published structure. Selected examples of structural similarities between the HIV-1 proteins IN, RT, and gp41 and human proteins determined by Dali are shown in [fig_ref] Figure 3: Selected Structural Similarities [/fig_ref]. The structural similarities frequently involve only part of each protein. However, since in most cases the precise location of protein interaction sites is not known, we used the entire structure in our investigation.
## Protein interaction prediction
Upon obtaining the knowledge of which specific HIV-1 and human proteins have high structural similarity, we extract all known interactions for human proteins from the Human Protein Reference Database, which contains over 37,000 documented protein interactions [bib_ref] Human protein reference database-2006 update, Mishra [/bib_ref]. Again, the central premise is that given a network of protein interactions, proteins with similar structures or substructures will tend to have similar interaction partners. Thus, our hypothesis is that HIV-1 proteins having similar structure to one or more human proteins are also likely to participate in the same set of protein interactions [fig_ref] Figure 1: Diagram of approach [/fig_ref]. Under these assumptions, we directly mapped HIV-1 proteins to their high-similarity matches within this network.
To reduce the number of predictions and provide an additional line of functional evidence for interactions and their possible biological relevance, we filtered these results using two types of datasets on host proteins involved in HIV-1 infection; collectively referred to as "Literature Filters" hereon. The first type represents host proteins that have been shown to impair HIV-1 infection or replication when knocked down by siRNA or shRNA. Three genome-scale siRNA screens have been conducted in HeLa or 293T cells [bib_ref] Identification of Host Proteins Required for HIV Infection Through a Functional Genomic..., Brass [/bib_ref] [bib_ref] Genome-scale RNAi screen for host factors required for HIV replication, Zhou [/bib_ref]. A fourth study with a similar goal was conducted using shRNA in Jurkat T-cells, a more realistic model of HIV-1 infection [bib_ref] A Genome-wide Short Hairpin RNA Screening of Jurkat T-cells for Human Proteins..., Yeung [/bib_ref]. Each of the four screens found over 250 host proteins involved in HIV-1 infection. Remarkably, very little overlap exists between these studies, perhaps due to differences in methods, including the cell lines and stages of the HIV-1 life cycle investigated.
The second type of data used to filter predictions is literature data identifying human proteins present in the HIV-1 virion. During budding, host proteins from both the cell surface and the cytoplasm, including some involved in the cytoskeleton, signal transduction, metabolism, and chaperones, may be incorporated into the virion [bib_ref] Proteomic and Biochemical Analysis of Purified Human Immunodeficiency Virus Type 1 Produced..., Chertova [/bib_ref]. While some of these proteins may be taken up by the budding virus simply by chance, others are known to be specifically incorporated into the virion and may play key roles in viral life cycle or pathogenesis. For example, TSG101 may be incorporated due to its interaction with Gag, and facilitates budding [bib_ref] Proteomic and Biochemical Analysis of Purified Human Immunodeficiency Virus Type 1 Produced..., Chertova [/bib_ref] [bib_ref] Plunder and Stowaways: Incorporation of Cellular Proteins by Enveloped Viruses, Cantin [/bib_ref].
We considered only predicted interactions where the target protein was observed in at least one of the previously described Literature Filters. The resulting predicted HIV-human interaction network consists of 2143 interactions, considering all unique combinations of Uniprot accessions for an HIV-1 protein and a predicted human interactor [fig_ref] Figure 2: Structural prediction workflow [/fig_ref]. Of the predictions that were made, 62 were verified as true interactions based on data from two databases of known host-pathogen interactions, HHPID and PIG (Additional Files 2 and 3). There were 347 human proteins predicted to have structural similarities with an HIV-1 protein and the predictions implicate a total of 406 unique human proteins as potentially interacting with HIV-1 [fig_ref] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction... [/fig_ref].
We visually examined some of the structural similarities that led to predictions that were already known. SMN2 is structurally similar to integrase (IN) [fig_ref] Figure 3: Selected Structural Similarities [/fig_ref] , Additional File 1) and both SMN2 and IN are known to interact with SIP1 (Gemin2) [bib_ref] Human protein reference database-2006 update, Mishra [/bib_ref]. SIP1, part of the large SMN complex involved in the assembly of snRNPs, may also be part of the pre-integration complex during HIV-1 infection and may aid viral reverse transcription [bib_ref] Identification of a novel human immunodeficiency virus type 1 integrase interactor, Gemin2,..., Hamamoto [/bib_ref]. There are also several predicted interactions between IN and host proteins that interact with SMN2 that have not yet been tested (Additional File 1). The structural similarities shown in [fig_ref] Figure 3: Selected Structural Similarities [/fig_ref] -D also led to predictions of known interactions, even though only part of the proteins are structurally similar.
## Protein co-localization
To further narrow the list of likely interactions, we refined these results by requiring both the HIV-1 protein and the target human protein to be present in the same location within the cell, based on GO cellular component (CC) annotation. The refined set of predictions is shown in [fig_ref] Figure 4: Predicted interaction network after cellular component filtering [/fig_ref]. Including this filtering step reduced the number of interaction predictions to 502, involving 189 HIVsimilar proteins having 137 known different binding partners. There are 31 predictions corresponding to already known HIV-human interactions [fig_ref] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction... [/fig_ref] , Additional File 4). Using the criterion that interacting proteins must have some evidence of co-localization not only reduced the size of the predicted interactome, but also increased the percentage of true positive predictions from ~3% true positives before filtering to over 6% after filtering [fig_ref] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction... [/fig_ref].
Taking localization into account, gp41 has many more predicted interactors than any other HIV-1 protein. This is most likely due to the relatively large number of GO cellular component terms that were annotated to gp41 and also relevant to the host cell. Since gp41 is known to be found in more parts of the cell than other HIV-1 proteins, a larger number of human proteins were able to meet the co-localization criterion. The interaction predictions made by this method are specific for structures, and we note that different structures for a single protein may lead to different predictions about its interactions. Therefore, some information is lost if predictions are described at a gene level. Nevertheless, it may be of interest to consider interactions on a gene basis (See Additional File 5 for the mapping of HIV-1 IDs). When counted according to the HIV-1 protein node names and human target Entrez Gene IDs, we made 883 interaction predictions, 56 of which were true positives according to HHPID and PIG. Following CC filtering, we had 22 true positive predictions among 265 total predictions (~10% of known true positives). While these results tend to suggest higher rates of predictive accuracy when using our method, we report our more conservative Uniprot-based accuracy values as our best estimates.
## Properties of human proteins predicted to interact with hiv-1
Using the CC-filtered predictions, we next examined the function of human proteins predicted to interact with HIV-1 during infection. In this instance, we sought biological process and molecular function GO terms that were enriched among these target proteins. Examining the function of human proteins found in our filtered list of interactions, significant enrichment is observed in the processes of protein transport, nucleic acid transport, sig- naling, cell death, and post-translational modifications [fig_ref] Figure 5: Significantly enriched Gene Ontology terms in the Human-HIV-1 interaction network [/fig_ref] ; all of these processes are known to be manipulated or altered by HIV-1 during infection. During the course of the HIV-1 lifecycle, viral protein and nucleic acids must be transported from one part of the cell to another to ensure viral replication. The Pre-Initiation Complex (PIC), consisting of a number of viral and host proteins and the viral genome, must be transported from the site of viral entry to the nucleus for integration of the provirus. In addition, Env and Vpr are known to play both pro-and anti-apoptotic roles by manipulating host signaling. For instance, there is evidence that HIV-1 may inhibit apoptosis in infected cells to prevent the cell from dying before the virus can replicate and assemble. On the other hand, HIV-1 can also promote apoptosis of immune cells using several pathways; indeed, the progressive destruction of CD4 + T cells is a well known indication of AIDS [bib_ref] Apoptosis as an HIV strategy to escape immune attack, Gougeon [/bib_ref]. Interestingly, all of the significantly enriched molecular function GO terms relate to GTP binding or hydrolysis [fig_ref] Figure 5: Significantly enriched Gene Ontology terms in the Human-HIV-1 interaction network [/fig_ref]. GTPases are involved in a number of host processes that HIV-1 may take advantage of, including nuclear transport and cytoskeletal rearrangements that facilitate viral entry and cellular motility. Statins, a class of drugs that lowers cholesterol levels in the blood, have also been shown to inhibit HIV-1 infection by preventing viral fusion with the cell membrane through a mechanism that involves inhibition of Rho GTPases [bib_ref] Statins inhibit HIV-1 infection by down-regulating Rho activity, Del Real [/bib_ref]. In addition, p115-RhoGEF inhibits HIV-1 gene expression through the activation of RhoA [bib_ref] Modulation of HIV-1 replication by a novel RhoA effector activity, Wang [/bib_ref]. Furthermore, both Rho and Rho kinase play a role in the cellular motility that allows HIV-1 infected monocytes to cross the bloodbrain barrier to cause HIV-1 encephalitis [bib_ref] Rho-mediated regulation of tight junctions during monocyte migration across the blood-brain barrier..., Persidsky [/bib_ref].
Actin microfilaments of the cytoskeleton are regulated by actin-binding proteins as well as Rho family small GTPases including Rho, Rac, and Cdc42 [bib_ref] Human immunodeficiency virus (HIV)-1 proteins and cytoskeleton: partners in viral life and..., Matarrese [/bib_ref]. IN, RT, and gp41 were all predicted to interact with RhoA, Rac1, and Cdc42 [fig_ref] Figure 4: Predicted interaction network after cellular component filtering [/fig_ref]. We found that gp41 has regions of structural similarity with many cytoskeleton related proteins, including erythrocytic spectrin alpha (SPTA1), erythrocytic spectrin beta (SPTB), alpha actinin 4 (ACTN4), alpha actinin 2 (ACTN2), moesin (MSN), Rhoassociated coiled-coil containing protein kinase 1 (ROCK1), and arfaptin 2 (ARFIP2). IN resembles NCK adaptor proteins 1 and 2 (NCK1/2), dynactin 1 (DCTN1), and RAS GTPase activating protein 1 (RASA1), among others (Additional File 4). The cytoskeleton has been suggested to be manipulated by HIV-1 during virion fusion, assembly, and budding [bib_ref] Human immunodeficiency virus (HIV)-1 proteins and cytoskeleton: partners in viral life and..., Matarrese [/bib_ref]. HIV-1 movement through the cell can be blocked by drugs that cause depolymerization of microtubules and actin filaments. Actin has also been found within HIV-1 virions, and is incorporated through binding with NC [bib_ref] Actin associates with the nucleocapsid domain of the human immunodeficiency virus Gag..., Wilk [/bib_ref]. Thus, our predictions may aid further investigation into the ways in which HIV-1 manipulates the cytoskeleton.
By integrating a variety of high-quality functional data sets in the Literature Filter, we created a smaller interaction map that has the potential to provide a physical interaction context for a number of experimental findings. As an example, retroviral budding is known to involve members of the endosomal sorting complexes (ESCRTs). The ESCRT complexes normally induce the formation of multivesicular bodies in the endosome, but can be recruited to the plasma membrane by Gag to aid in viral budding. Many members of the ESCRT machinery appear in our results, including VPS4A, STAM2, EEA1, RAB5A, and TSG101 [bib_ref] Host factors exploited by retroviruses, Goff [/bib_ref]. Early endosomal autoantigen 1 (EEA1) is recruited to early endosomes by Rab5 and phosphatidylinositol 3-phosphate [bib_ref] FYVE and coiled-coil domains determine the specific localisation of Hrs to early..., Raiborg [/bib_ref]. Our results show that gp41 and Gag p2 may interact with RAB5A, since they are structurally similar to EEA1 [fig_ref] Figure 4: Predicted interaction network after cellular component filtering [/fig_ref] , Additional Files 1 and 3). EEA1 contains a FYVE domain and colocalizes with human hepatocyte growth factorregulated tyrosine kinase substrate (Hrs) protein [bib_ref] FYVE and coiled-coil domains determine the specific localisation of Hrs to early..., Raiborg [/bib_ref] [bib_ref] Endosomal membrane trafic: convergence point targeted by Mycobacterium tuberculosis and HIV, Deretic [/bib_ref]. Gp41 is also known to interact with AP1G2, an important component of clathrin-coated vesicles. AP1G2 interacts with RAB5A and provides further support for the possibility that gp41 interacts physically with RAB5A, but through a potentially different structural motif [bib_ref] Definition of the consensus motif recognized by gamma-adaptin ear domains, Mattera [/bib_ref]. The Gag p6 protein is a known mimic of Hrs, and like Hrs can recruit TSG101, which is required for the formation multivesicular bodies (MVBs) and viral budding [bib_ref] HIV Gag mimics the Tsg101-recruiting activity of the human Hrs protein, Pornillos [/bib_ref]. Gag p2, as well as a model of gp41, show structural similarity to the human protein CEP55, which recruits TSG101 to the thin membrane that separates the daughter cells, where it is needed for the final separation of two cells [bib_ref] Parallels Between Cytokinesis and Retroviral Budding: A Role for the ESCRT Machinery, Carlton [/bib_ref]. Our results suggest that gp41, IN, and the p2 region of Gag may all be able to interact with TSG101 [fig_ref] Figure 4: Predicted interaction network after cellular component filtering [/fig_ref] , Additional File 4). Overall, interaction predictions are supported by a variety of studies implicating host mechanisms of vesicle formation in HIV-1 infection.
# Additional method assessment
To further assess our predictions, we determined how many known interactions, curated within either HHPID or PIG, could have possibly been predicted using our method and the available data. First, in order for our approach to suggest a possible HIV-human interaction, the HIV protein must be represented among the crystal structures from PDB that are included in the Dali Database. In addition, any host factors predicted to interact with HIV-1 must have at least 1 known interaction with another human protein, and to be considered further, each of these must also have representative structures within Dali. Finally, in this work we included only those proteins that have been implicated in playing a role in HIV-1 infection through RNAi studies or studies of the protein composition of the virion. Since we removed any human target proteins that did not pass the Literature Filter, we did not make predictions for human proteins not mentioned in previous studies. A total of 319 known host-pathogen interactions satisfied these criteria. Sixty-two of these interactions (~19%) were predicted by our methodology, and are the set of predictions considered to be true positives (shown in [fig_ref] Table 3: Method evaluation [/fig_ref]. We also investigated how many of these possible interactions could have been found after using the cellular component filter, and determined that only 166 known interactions met the additional criterion of being annotated to the same cellular component. Within this set, our method found 31 of these (~19%). This result suggests that while the number of interactions considered was decreased by considering cellular localization, the number of true positive predictions did not improve. Obviously, without experimental validation we cannot determine whether the CC filter led to better prediction accuracy within the set of predictions not previously described in the literature or elsewhere. It is clear, however, that GO cellular component annotation is incom-plete and the lack of shared annotation does not completely exclude the possibility that two proteins may interact; inclusion of the CC filter did double the percentage of true positives predicted when considering unknown potential interactions as well as those previously known.
As an additional form of assessment, we investigated how often we could expect to find previously known interactions by chance alone. Starting from proteins in HPRD, we found that ~0.17% of the known interactions could be found at random (see Methods). Cellular Component filtering of these random predictions gave a slight improvement with an average of 0.29% true positives [fig_ref] Table 4: Accuracy of Random Predictions [/fig_ref]. Using only HPRD human target proteins that pass the Literature Filter increased the true positive accuracy of random predictions to 0.57%. This value can be compared to the value of 2.89% indicated in [fig_ref] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction... [/fig_ref]. When these random predictions were also run through the CC Filter, an average of 1.03% true positives were found [fig_ref] Table 4: Accuracy of Random Predictions [/fig_ref] versus a 6.18% when using our method [fig_ref] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction... [/fig_ref]. Thus the Literature Filter and the CC Filter improved the accuracy of the true positive predictions individually, and to an even greater extent when com- Comparison of the number of known interactions predicted to the number of known interactions that could have theoretically been found using the available data. bined. However, even with both filters, at best ~1% of the random predictions were found to be true positives, further indicating that incorporating structural information generates predictions with enhanced accuracy and biological validity.
## Overlap with other studies
We also compared our predictions to those made by two previous computational studies predicting protein-protien interactions between HIV-1 and humans, namely the studies by Evans et al. and Tastan et al. [bib_ref] Prediction of interactions between HIV-1 and human proteins by information integration, Tastan [/bib_ref] [bib_ref] Prediction of HIV-1 virus-host protein interactions using virus and host sequence motifs, Evans [/bib_ref]. Since these investigations reported their results in terms of genes, we compared them to our predictions as counted by gene, to find interactions predicted by multiple methods [fig_ref] Figure 6: Overlap with previous studies [/fig_ref]. We did not find a high degree of overlap between the predictions made by the various studies. There were a few predictions that were shared between all methods. For our results before CC filtering, we found that there were 9 interactions predicted by all three methods [fig_ref] Figure 6: Overlap with previous studies [/fig_ref]. Of these, four were determined to be true positives in our results: RT and MAPK1, gp41 and LCK, gp41 and PTPRC, and IN and PRKCH. The other five interactions (RT and PIN1, p2 and MAPK1, p2 and YWHAZ, gp41 and PLK1, gp41 and MAPK1, gp41 and CLTC, IN and XPO1, and IN and YWHAZ) are not known to occur, and may be good candidates for further investigation since they were predicted by three diverse methods. After we filtered our predictions by shared cellular components, three predictions were still common between all three studies, gp41 and LCK, gp41 and PLK1, IN and XPO1, one of which is a known interaction [fig_ref] Figure 6: Overlap with previous studies [/fig_ref]. In summary, although few predictions were shared by all three studies, a large proportion of them are already known to occur, suggesting that the others may be worthy of high priority in future experimental efforts.
# Conclusions
We have generated a map of potential protein-protein interactions between HIV-1 and its human host. The computational methodology used to create this map is based on the assumption that proteins with similar structures will share similar interaction partners. Thus HIV-1 proteins having a structure similar to one or more human proteins may potentially "plug in" to the host protein interactome at these points; providing the interface through which manipulation of downstream host processes can occur. From previous literature, many human proteins are known to play some role in HIV-1 infection. However, in most cases the nature of this role is unknown. Here, we provide specific predictions of how these human proteins may influence viral infection, namely by interacting with certain HIV-1 proteins.
In principle, our approach is applicable to any hostpathogen system with known protein structures. HIV-1 has a small proteome, with most of its protein structures at least partially determined. In addition, HIV-1 also has a large set of identified interactions that can be used for model validation. While few pathogens currently have such rich data sets, continued progress in protein structure determination will help to remedy such deficiencies.
Identification of points of modulation between a host and pathogen requires multiple lines of evidence. Computational methods can help integrate these data, providing a promising avenue for the discovery of novel hostpathogen interactions mediated by structural similarities as well as enhancing our understanding of functional relationships characterized through modern screening methods such as siRNA. Knowledge of the protein interaction network between the pathogen and human will not only further our basic understanding of pathogen survival mechanisms, but may also provide clinical targets to combat infectious disease.
# Methods
## Data sources
We used the Dali Database for structure comparisons (downloaded in January 2009), and the Human Protein Reference Database (HPRD), HHPID and PIG for protein interactions (downloaded February, July and June 2009, respectively) [bib_ref] Searching protein structure databases with DaliLite v. 3, Holm [/bib_ref] [bib_ref] Protein Structure Comparison by Alignment of Distance Matrices, Holm [/bib_ref] [bib_ref] Human protein reference database-2006 update, Mishra [/bib_ref]. The literature sources and various databases used each have their own system of identifiers. PDB codes obtained from Dali were mapped to their corresponding taxonomy and Uniprot accessions using data from the SIFTS initiative [bib_ref] Announcing the worldwide Protein Data Bank, Berman [/bib_ref] [bib_ref] E-MSD: improving data deposition and structure quality, Tagari [/bib_ref]. Other identifier mappings were carried out using DAVID Gene ID Conversion or Uniprot ID mapping [bib_ref] DAVID: database for annotation, visualization, and integrated discovery, Dennis [/bib_ref] [bib_ref] Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources, Huang [/bib_ref]. Network diagrams were created in Cytoscape [bib_ref] Cytoscape: a software environment for integrated models of biomolecular interaction networks, Shannon [/bib_ref]. Images of protein structures were created in MacPyMol.
## Determination of structural similarity between hiv-1 and host proteins
We used the Dali database to ascertain structural similarity. Dali compares the 3D structural coordinates of two PDB entries by alignment of alpha carbon distance matrices, allowing for differences in domain order, and produces a structural similarity score [bib_ref] Announcing the worldwide Protein Data Bank, Berman [/bib_ref] [bib_ref] Searching protein structure databases with DaliLite v. 3, Holm [/bib_ref] [bib_ref] Protein Structure Comparison by Alignment of Distance Matrices, Holm [/bib_ref]. The Dali Database includes structural comparisons where proteins from PDB90, a subset of the PDB where no two proteins share more than 90% sequence similarity, were used as queries against the full PDB [bib_ref] Removing near-neighbour redundancy from large protein sequence collections, Holm [/bib_ref]. For this study, we took into consideration all human proteins that were listed in the Dali database as being similar to an HIV-1 protein (NCBI Taxonomy ID: 11676) and having a z score above 2.0, with the HIV-1 protein being either the query or the hit. We refer to these human proteins as "HIV-similar" proteins. No proteins of unknown structure were considered.
## Interaction prediction
We found known interactions between HIV-similar proteins and target human proteins, using data from the HPRD database, which contains literature curated interactions between pairs of human proteins [bib_ref] Human protein reference database-2006 update, Mishra [/bib_ref]. For each HIV-similar protein, we predict that the target proteins, which are known to both interact with the HIV-similar protein and pass the Literature Filter, might also interact with the corresponding HIV-1 protein. Therefore, interactions between the HIV-similar and the human target proteins were mapped directly to the corresponding HIV protein.
## Filtering
To reduce the number of predictions as well as add information from functional studies, predictions were filtered based on previous implication of the human protein's involvement in the HIV-1 infection process. One criterion was presence of the host protein in the HIV-1 virion. Host proteins known to be incorporated into or onto HIV-1 during budding were taken from several literature sources [bib_ref] Proteomic and Biochemical Analysis of Purified Human Immunodeficiency Virus Type 1 Produced..., Chertova [/bib_ref] [bib_ref] Plunder and Stowaways: Incorporation of Cellular Proteins by Enveloped Viruses, Cantin [/bib_ref] [bib_ref] Proteomic Analysis of Human Immunodeficiency Virus Using Liquid Chromatography/Tandem Mass Spectrometry Effectively..., Saphire [/bib_ref]. The presence of host proteins in or on HIV-1 may be a result of specific recruitment and serve a functional role, may result from localization of the protein near the site of budding, or may simply occur by chance. Predicted interactions between HIV-1 proteins and human proteins that are incorporated into the HIV-1 virion were retained. In addition, any human protein that is incorporated into the virion and is itself structurally similar to an HIV-1 protein was also included as a possible interaction.
Another filtering criterion was the host protein's essentiality for HIV-1 infection. Recently, several large-scale experiments using siRNA or shRNA knockdowns to identify host proteins involved in the HIV-1 life cycle have been published [bib_ref] Identification of Host Proteins Required for HIV Infection Through a Functional Genomic..., Brass [/bib_ref] [bib_ref] Genome-scale RNAi screen for host factors required for HIV replication, Zhou [/bib_ref] [bib_ref] A Genome-wide Short Hairpin RNA Screening of Jurkat T-cells for Human Proteins..., Yeung [/bib_ref]. The probe ids of the genes implicated by Yeung et al. were mapped to their Entrez Gene IDs using the appropriate Affymetrix annotation file http://www.affymetrix.com/products_services/ arrays/specific/hgu133plus.affx#1_4 [bib_ref] A Genome-wide Short Hairpin RNA Screening of Jurkat T-cells for Human Proteins..., Yeung [/bib_ref]. This filter is referred to as the "Literature Filter." Host proteins that were implicated in at least one of these studies as having a possible role in HIV-1 infection or replication, and which are also known to interact with an HIV-similar protein, were predicted to interact with an HIV-1 protein in the final predicted network.
To create a smaller and potentially more reliable list for further experimental validation, we further filtered the predictions based on shared sub-cellular localization. The Cellular Component (CC) Filtered dataset contains interaction predictions where the two proteins share Gene Ontology (GO) cellular component annotation. Pairs of HIV-1 and human proteins predicted to interact were only included in this dataset if both proteins were annotated by DAVID as being present in the same cellular compartment [bib_ref] DAVID: database for annotation, visualization, and integrated discovery, Dennis [/bib_ref] [bib_ref] Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources, Huang [/bib_ref]. Pairs with only the terms "cell" and "cell part" in common were excluded due to a large num-ber of such pairs and the relative lack of specificity of these high level terms.
## Validation of predictions
Since within Dali there may be multiple PDB structures representing the same protein, there is some redundancy in the interaction predictions. In certain cases, multiple PDB structures for the same HIV-1 protein were found to be similar to multiple PDB structures for an HIV-similar protein, leading to the same interaction predictions. Therefore, the predictions were counted as unique pairs of Uniprot accessions. In addition, for ease of viewing the predicted interactome, each node representing an HIV-1 protein is labeled with the protein name while each human protein is represented by Entrez Gene ID. To determine the correct mapping of PBD codes to HIV-1 proteins, the molecule name associated with each PDB chain was searched for keywords indicating the protein, with ambiguous cases treated on an individual basis. For example, PDB molecule names containing the word "capsid" but not "nucleocapsid" were assigned to the node "capsid." Furthermore, molecule names indicating polyproteins, such as those containing the phrase "gag-pol" were checked individually to determine which specific part of the polyprotein was represented by the entry. Two PDB structures were found to represent more than one mature HIV-1 protein: 1l6nA contains both capsid and matrix, while 1bajA contains capsid and p2 [bib_ref] PDB ID: 1L6N Structure of the Nterminal 283-residue fragment of the immature..., Tang [/bib_ref] [bib_ref] PDB ID: 1BAJ Structures of the HIV-1 capsid protein dimerization domain at..., Worthylake [/bib_ref] ; these structures are represented as "capsid, matrix" and "capsid, p2" respectively. When counting predictions at the gene level, we considered pairs of HIV-1 node names and human target Entrez Gene IDs.
To determine which predictions are true positives, PIG and HHPID entries for the predicted human interactors were examined to see if they contained the HIV-1 protein they were predicted to interact with. These interaction databases consist of PPIs curated from the literature. HHPID uses keywords to characterize the different types of interactions listed in this database. Since this work attempts to predict physical interactions, only entries with keywords representing direct interactions were included [bib_ref] Prediction of interactions between HIV-1 and human proteins by information integration, Tastan [/bib_ref]. The Uniprot accessions associated with the HIV-1 protein PDB entry, in the case of PIG, or the Entrez Gene ID mapped to that Uniprot accesion, in the case of HHPID, was checked to see if it was present as an already known interaction of the human protein.
## Go term enrichment
The Gene Ontology (GO) provides a system of terms to consistently describe and annotate gene products [bib_ref] Gene Ontology: tool for the unification of biology, Ashburner [/bib_ref]. GO term enrichment was performed using the DAVID Functional Annotation Chart tool. The GO is organized as a tree structure, with terms becoming more specific as distance from the root increases. Therefore, to avoid very general and uninformative GO terms, only those that are found at least 5 steps removed from the overall root of GO were considered. The p-values were corrected for multiple testing using the Bonferroni procedure and transformed by taking the -log 10 for easier visualization.
## Computational evaluation
Two forms of computational validation of the method were conducted. As it is not possible to predict all known interactions due to lack of protein structures, as well as other factors, we first determined the largest set of known interactions that it is theoretically possible to predict using our approach. To do this, we first determined the sets of all proteins that could be considered. This includes the set of all HIV-1 proteins in Dali (HIV set), the set of all human proteins that are represented in both Dali and HPRD (possible HIV-similar set), and the set of all human proteins in HPRD that are known to interact with at least one protein in the possible HIV-similar set as well as pass the literature filter (possible target set). Next, every pairwise combination of proteins in the HIV set and the possible target set was checked to see if it represented a known interaction curated in HHPID or PIG. The resulting number of true interactions that could have been found by the method was compared to the number of true positives that were actually found, both before and after filtering by cellular components.
In the second approach, actual prediction results were compared to predictions based on randomly selected HIV-human protein pairs. The HIV-1 proteins were chosen from the 69 Uniprot accessions represented at least once by structures in the Dali Database. For human proteins, two different sets of human Uniprot proteins were created, one containing all the proteins in HPRD, and the other containing the subset of human proteins that also passed the Literature Filter. The set of all human proteins in HPRD consisted of 8582 proteins and was used to see the accuracy of purely random predictions, while the second set of 830 proteins was used to observe the effect of the Literature Filter.
Since the structural similarity step was omitted, the predictions based on a human protein being similar to an HIV-1 protein and incorporated into the virion could not be simulated with the random selection procedure. We found that if we excluded this class of predictions from our real results, the number of unique predictions made was reduced to 2139, but all 62 true positives were still included. Therefore, we randomly selected 2139 pairs of HIV-1 proteins and human proteins from the entire HPRD, and a second set of 2139 pairs of HIV-1 proteins and Literature Filtered human proteins for evaluation. Next, any known interactions between the randomly chosen pairs were found using HHPID and PIG. Additionally, both the unfiltered and Literature Filtered random predictions were then subjected to the CC Filter to gauge the improvement due to this step of the method. The CC Filter reduced the number of predictions to a variable degree, depending on how many of the random predictions were annotated with the same GO cellular component term. The entire procedure was repeated 1000 times. The mean and standard error of the mean for each of the four variously filtered random prediction sets was calculated using R. The distributions of random predictions after Literature Filtering were approximately normal, so one-sided single sample t-tests were performed to determine if the method performed significantly better than random. In addition, we performed Wilcoxon signed-rank tests that do not make assumptions about normality. When comparing our results to random predictions that had undergone the same filtering steps, either the Literature Filter or both the Literature and CC Filters, the p-values were less than 2.2e-16 for all statistical tests. In addition, even when performing the randomization procedure 10000 times, none of the randomly selected interaction sets had a true positive rate higher than that observed in our results, suggesting a p-value of no greater than 0.0001.
To compare our predictions to those made by Evans et al. and Tastan et al., we found the intersection of the prediction sets, counted by HIV-1 protein name and human Entrez Gene ID [bib_ref] Prediction of interactions between HIV-1 and human proteins by information integration, Tastan [/bib_ref] [bib_ref] Prediction of HIV-1 virus-host protein interactions using virus and host sequence motifs, Evans [/bib_ref]. Since each study used different names for the HIV-1 proteins, we had to map the naming schemes to each other to find common predictions. For example, Evans et al.'s "CA" and "GAG" and Tastan et al.'s "gag_capsid" and "gag_pr55" were mapped to our "capsid." Proteins for which we made no predictions, such as Rev, were not mapped to anything in our results, but were converted between Evans et al. and Tastan et al. to find overlap between these two studies.
# Additional material
Additional file 2 Interaction Predictions. Interaction predictions, before the CC Filter. Column descriptions: HIV PDB code-PDB code of an HIV-1 protein in Dali, HIV node name-protein name used to represent the HIV protein in the interaction network, HIV-sim Human PDB code-PDB code of a human protein similar to the HIV protein, HIV-sim Human Gene Symbolthe Entrez Gene Symbol of the human protein that is similar to the HIV protein, HIV-sim Human GeneID-the Entrez GeneID of the human protein that is similar to the HIV protein, HIV-sim Human Uniprot-the Uniprot accession of the human protein that is similar to the HIV protein, Human node Gene Symbol-the Entrez Gene Symbol of a human protein predicted to interact with the HIV protein. The Gene Symbol was used to represent the human protein in the interaction network, Human interactor GeneID-the Entrez GeneID of a human protein predicted to interact with the HIV protein, Human interactor Uniprot -the Uniprot accession of a human protein predicted to interact with the HIV protein, Source Datasets-paper(s) with support for a role of the human protein in HIV infection, True Positive?-whether or not the predicted interaction is already represented in PIG or NIAID, Type-whether the prediction was made because the human protein is known to interact with an HIV-similar protein or because the human protein is HIV-similar itself and known to be incorporated into the HIV virion, NIAID-HIV proteins listed in NIAID as interacting with the human protein, PIG-HIV proteins listed in PIG as interacting with the human protein. This file is provided in tab-separated text format and may be viewed in a text editor or Excel.
[fig] Figure 1: Diagram of approach. HIV-1 proteins showing structural similarity to one or more human proteins are first identified. Interactions for these "HIV-similar" proteins with other human proteins are then identified. Following appropriate filtering, this methodology predicts the existence of a physical interaction between the HIV protein and the human "target" protein(s). [/fig]
[fig] Figure 2: Structural prediction workflow. Structural similarities from Dali and known interactions between human proteins from HPRD are used to predict interactions between HIV-1 and human proteins. These predictions are filtered based on functional information from previous studies to make a first set of predictions. This set is further filtered using GO cellular component terms to yield a final prediction set including fewer predictions with higher confidence. Numbers represent the number of interactions, or structural similarities in the case of Dali, at each stage of the process. [/fig]
[fig] Figure 3: Selected Structural Similarities. Structures of HIV-1 and human proteins aligned using Dali. (A) IN (1ex4A) aligned with SMN2 (1g5vA)[51,52]. (B) NXF1 (1ft8E) aligned with RT (1tl3A)[53,54]. (C) gp41 (2cmrA) aligned with PTK2 (1k04A).[55,56]. (D) RT (1lwcA) aligned with PLEC1 (1mb8A)[57,58]. HIV-1 proteins are in blue, human proteins are in yellow. [/fig]
[fig] Figure 4: Predicted interaction network after cellular component filtering. In addition to the prediction of a physical interaction, the human proteins included in this prediction set are known to have a role in HIV-1 infection or replication as supported by 1) evidence of incorporation into the HIV-1 virion or 2) their reduced expression is known to prevent HIV-1 infection (node line color corresponds to source). Predictions were filtered to contain only those pairs of proteins that share at least one Gene Ontology cellular component term. Red lines represent predicted interactions that are already known to occur. [/fig]
[fig] Figure 5: Significantly enriched Gene Ontology terms in the Human-HIV-1 interaction network. GO Terms removed at least 5 levels from the root for (A) Biological process and (B) Molecular function. Bonferroni corrected p-values (α = 0.01) were -log 10 transformed. [/fig]
[fig] Figure 6: Overlap with previous studies. Venn diagrams of the overlap between between our method and previous computational studies by Evans et al. and Tastan et al. (A) with Literature filter and (B) with Literature and CC filter [9,10]. [/fig]
[table] Table 1: HIV-1 protein structuresThe number of structures representing each HIV-1 protein in Dali. [/table]
[table] Table 2: Summary of Predicted InteractionsThe number of proteins found as well as interaction predictions made by the method are shown. HIV-1 Structure Nodes refers to the number of HIV-1 proteins represented in Dali, while HIV-1 Uniprot refers to the number of HIV-1 Uniprot accessions present in the predictions. Human proteins and predicted interactions are counted by unique Uniprot accessions. [/table]
[table] Table 3: Method evaluation [/table]
[table] Table 4: Accuracy of Random Predictions [/table]
|
Mind the Localized Skeletal Pain: Chronic Recurrent Multifocal Osteomyelitis
Chronic recurrent multifocal osteomyelitis (CRMO) is a rare idiopathic aseptic inflammatory bone disorder affecting primarily children and adolescents characterized by an insidious onset of pain, swelling, and tenderness over the affected bones. The clinical signs and symptoms of CRMO are nonspecific, radiological and histopathological tests are essential for its diagnosis.We present a case of an 18-year-old young man who was diagnosed with CRMO by a combination of clinical data, laboratory results, radiological imaging, and bone biopsy. The patient started anti-inflammatory and immunosuppressant therapy, and his lower extremity pain and swelling improved.This report highlights to investigate promptly in children and adolescents with chronic leg pain, to emphasize the importance of combined clinical, laboratory, and imaging tests for early identification, to have a greater understanding of the imaging appearance and increasing knowledge of this condition, which help shorten time to reach a diagnosis and prevent permanent osseous damage and long-term disabilities.
# Introduction
Chronic recurrent multifocal osteomyelitis (CRMO) also known as chronic nonbacterial osteomyelitis (CNO) is a non-infectious autoinflammatory bone disease affecting primarily children and adolescents. It is characterized by an insidious onset of pain, swelling, and tenderness localized over the affected bones. We report the case who presented with chronic bilateral leg swelling and pain subsequently found to have CRMO after a combination of clinical, laboratory, and radiological data and biopsy.
## Case presentation
An 18-year-old young man was referred to our internal medicine clinic for bilateral leg swelling. During the routine medical check-up, he was found to have lower extremity swelling and directed to a vascular surgeon for possible chronic venous insufficiency. The vascular clinic confirmed his venous system was competent, excluded venous reflux and venous obstruction after bilateral venous duplex ultrasonography.
On visiting our clinic, he reported having bilateral lower limb swelling for ten years. He had minimal swelling of his legs in the morning. By the end of the day, he noted swelling of legs up to the distal third of both legs associated with pain. He experienced lower leg pain that began with running activity. The pain scale was quantified as 4 out of 10. It increased with continued exertion, resolved with rest, or when he stopped engaging in activities. He was asymptomatic at rest and during the activity of daily living.
He denied early morning stiffness, back pain, or positive family history. He tried massage therapy, footwear modifications, and over-the-counter analgesics as needed. Recently he observed more swollen ankles associated with pain on activity. His bowel habit is regular with no diarrhea, abdominal pain or weight loss, or skin rashes. He did not have a cough, fever, or night sweating. There was no significant family history of autoimmune disease and bone disease. On examination, he was able to walk with a normal gait. There was bilateral ankle nonpitting edema with no tenderness, no increase of the local temperature, and peripheral pulses were present. Fine tremors of the hands and sweaty palms were noted. The rest of the examination was unremarkable. At that time, clinical impressions were likely hyperthyroidism and bilateral ankle arthropathy of uncertain etiology.
Laboratory tests disclosed he was biochemically hyperthyroid, free thyroxine 16 pmol/L ( 8.8-14.4 pmol/L ) and Thyroid Stimulating Hormone (TSH) 0.588 MU/L ( 0.65-3.70 MU/L ). Complete blood count, renal panel, liver panel, serum calcium, urine protein creatinine ratio all were in the normal range and Chest X-ray finding was also normal. Radiographic findings of anklesrevealed periosteal reaction seen along with the medial aspects of bilateral distal tibial meta-diaphysis. Soft tissue swelling was noted around the ankles bilaterally. On reviewing him a month later, he reported well. He was not given antibiotics. He had occasional pain in the ankles on long-standing and prolonged walking. The repeated TSH and free thyroxine tests turned out normal. Triiodothyronine and thyroid-stimulating receptor antibody levels were normal. An abnormal X-ray finding of the ankle prompted him to go for further investigations. Inflammatory marker levels were nonsignificant (erythrocyte sedimentation rate [ESR] 10 mm/hour [1-10 mm/hour], C-reactive protein [CRP] 10.3 mg/L [0.2-9.1 mg/L]). Antinuclear antibody, extractable nuclear antigen (ENA) antibodies screening, and antineutrophil cytoplasmic antibody tests were negative. MRI of the anklesrevealed erosions along the medial cortex of the distal tibial metaphysis and epiphysis with periosteal reaction and marked surrounding soft tissue edema. Bony edema was also noted within the talus, calcaneus, navicular, medial cuneiform, and base of the first metatarsal. Findings were fairly symmetrical. On following clinic visit, he experienced ankle pain at night which was worse on the left side. Given the relatively prolonged clinical course over several years, bland biochemistry tests, combined with the multifocal osteolytic bone lesions on imaging, the provisional diagnosis of CRMO was made. He was given indomethacin for pain control. A radioisotope bone scintigraphywas arranged to detect other silent lesions. Increased uptake in bilateral distal tibia, calcaneum, and midfoot bones corresponding to prior MRI findings was observed on bone scan in keeping with inflammatory bone disease. No suspicious focus of increased radiotracer uptake is seen in the rest of the skeletal system. Screening test for tuberculosis (TB) is important to detect latent disease before consideration of starting biologic or disease-modifying anti-rheumatic drugs. TB QuantiFERON test was positive. Sputum for acidfast bacilli (AFB) smear, culture, and TB nucleic acid amplification test all were negative. Subsequently, a bone biopsy was performed to exclude chronic infections (like bacterial osteomyelitis, TB) and malignancy (bone tumors, leukemia, lymphoma, Langerhans cell histiocytosis) of bone. The left medial malleolus core biopsies showed compact bone, minimal intertrabecular areas with the presence of inflammatory cells. There were no granulomas, no malignant cells identified. No AFB were seen on the Ziehl-Neelsen stain and bone culture was sterile.
CRMO can be a difficult diagnosis as there are no validated diagnostic criteria for CRMO. Based on the clinical, radiological, histopathological findings and sterile bone culture, diagnosis of CRMO was subsequently established. In view of the findings on MRI ankles and bone formation on biopsy, methotrexate was initiated to reduce symptoms as well as prevent further osseous expansion. He was also treated for latent TB. He was feeling well with no more ankle pain. He continues to have close follow-up arrangements.
# Discussion
CRMO is a rare disease, likely to be underdiagnosed and underreported due to unawareness in medical practice. There are insufficient data to determine the true prevalence of the CRMO, and over 500 existing patients are reported worldwide based on case series studies. It mostly affects children and adolescents and the peak onset is the age between 7 and 12 years. The pathogenetic mechanism of the sterile bone inflammation remains unknown. Studies suggest that bone inflammation may be the result of an imbalance between pro-and anti-inflammatory cytokine expression from innate immune cells and subsequent osteoclast differentiation and activation, resulting in osteolytic lesions. Environmental and genetic factors, as reported by finding of CRMO in siblings and positive family history of autoimmune diseases, may interplay with the underlying bone inflammation. Infections are not involved in disease etiology. No infectious agents are detectable at the site of the bone lesion.
CRMO is characterized by the insidious onset of pain and swelling localized over the affected bones. Some of the patients complained of nocturnal bone pain which may be misinterpreted as growing pain or nonmalignant bone tumor. The symptoms of CRMO can be variable from asymptomatic inflammation of single bones to chronic, recurrent multifocal osteitis. The bone lesions are typically located in the metaphyses of long bones of lower extremities near knees and ankles. Symmetric involvement is common. The other frequent locations include pelvis, spine, clavicle, and long bone of upper extremity. Patients generally appear well but may experience concomitant systemic symptoms including low-grade fever and generalized malaise.
Physical examinations are mostly unremarkable. There may be localized tenderness and warmth over the affected bones. CRMO may be accompanied by other extra-osseous inflammatory disorders such as pustulosis palmoplantaris, acne, psoriasis, inflammatory bowel disease (IBD), spondyloarthropathy, and infrequently pyoderma gangrenosum, Sweet syndrome, or granulomatosis with polyangiitis.
CRMO remains a diagnosis of exclusion. It is essential to rule out other more common causes of recurrent inflammation such as infection, malignancy, and autoimmune disorders. In most CRMO patients complete blood count is usually normal, and inflammatory markers such as CRP, ESR may be mildly raised. All cultures of blood and tissue specimens are negative. Many patients were initially treated for bacterial osteomyelitis due to similarity in presentation with bacterial acute osteomyelitis. Imaging techniques are crucial for the diagnosis of CRMO. Plain X-ray usually discloses cortical irregularities, osteolytic destruction with surrounding sclerosis in the metaphysis areas of long bones.
Our case had the typical location of chronic symmetric ankle swelling and pain, but no systemic symptoms and no associated extra-osseous disorders. Initial findings of lytic lesion on plain X-ray of ankles directed him to proceed with further workup including MRI, bone scan, and bone biopsy.
MRI can demonstrate marrow edema associated with periostitis, soft-tissue inflammation. It can also locate asymptomatic silent lesions. Whole-body MRI, a gold standard diagnostic modality, allows detection of disease extent of CRMO, facilitates monitoring of treatment efficacy. Although the sensitivity of bone scintigraphy is lower than MRI, bone scintigraphy may be used alternatively to identify all silent lesions in the whole body. There is no specific feature in bone biopsy for CRMO. Histologic findings are consistent with chronic inflammation which include destruction of normal bone structure and presence of nonspecific inflammatory cells. Bone biopsies are usually performed to exclude chronic infections such as bacterial osteomyelitis, TB, and malignancies in particular leukemia, lymphoma, bone tumors.
A combination of clinically well-looking patient, chronic bone pain, normal laboratory results, symmetrical bony lesions on imaging, culture-negative biopsies shortened the length of time to the diagnosis of our patient. Our patient had more than the threshold diagnostic criteria for CRMO by Jansson's by the presence of two major features such as radiologically proven osteolytic lesion, multifocal lesion and three minor requirements including normal blood count and good general health state, unremarkable CRP and ESR, and chronic duration more than six months. He was doing well with non-steroidal anti-inflammatory drugs (NSAID) and methotrexate.
The goal of therapy for CRMO is to control disease activity in order to decrease disease complications and prevent skeletal damage. There is no consensus treatment of CRMO, and most evidence comes from small case series or retrospective cohorts. NSAID is the first-line treatment for all CRMO patients for symptomatic pain relief. In a prospective study, almost half of the cases achieved remission in pain and swelling within the first year of treatment with NSAID. However, relapse is common, and new lesions may develop. The efficacy of NSAID therapy is associated with the initial number of affected bony lesions. The patients with a high number of bony lesions at the time of disease onset had a poor response to NSAIDs. The total duration of NSAID is not clear for patients who responded to NSAID therapy.
In the Childhood Arthritis and Rheumatology Research Alliance (CARRA) meeting, the members decided that the NSAIDs should be tried for at least four weeks. The patients who still have persistent pain with local tenderness or warmth are considered non-responsive to NSAID therapy. Those non-responders require additional second-line treatment which included (1) methotrexate or sulfasalazine, (2) tumor necrosis factor (TNF)-alpha inhibitors, and (3) bisphosphonates.
A multidisciplinary approach involving pediatricians, physical and occupational therapists, rheumatologists, and geneticists is the most efficacious way of managing these patients.
Early diagnosis and effective treatment may prevent or decrease long-term sequelae of the disease. Permanent skeletal damage may include vertebral compression fractures, leg-length discrepancy, and disabling bone deformity like angulation secondary to growth plate damage. Earlier literature suggested that CRMO was a self-limiting disease that eventually resolved after 12-18 months in most cases. However, recent data indicate that long-term outcomes in patients with CRMO may be less favorable, 50 percent of CRMO patients develop disease flares after a median of 29 months. Hence, in patients with CRMO, close follow-up is critical to monitor disease activity, treatment response and capture recurrence of the disease by a combination of clinical, laboratory markers, and radiological imaging.
# Conclusions
A rare disorder, CRMO, should be suspected as one of the differential diagnoses in children and adolescents with chronic bone pain. They should be investigated promptly. A combination of imaging techniques such as plain radiograph, MRI, and bone scintigraphy is useful in the diagnosis of CRMO. Being unfamiliar with CRMO in clinical practice, many physicians commonly misdiagnosed it as bacterial osteomyelitis, delayed in referral and diagnosis. Increasing awareness of this disease and its features may prevent unnecessary diagnostic procedures, and mistreatment with prolonged antibiotic regimens. Timely effective treatment is crucial to prevent permanent skeletal damage.
# 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. |
Anifrolumab, an Anti–Interferon‐α Receptor Monoclonal Antibody, in Moderate‐to‐Severe Systemic Lupus Erythematosus
Objective. To assess the efficacy and safety of anifrolumab, a type I interferon (IFN) receptor antagonist, in a phase IIb, randomized, double-blind, placebocontrolled study of adults with moderate-to-severe systemic lupus erythematosus (SLE).Methods. Patients (n 5 305) were randomized to receive intravenous anifrolumab (300 mg or 1,000 mg) or placebo, in addition to standard therapy, every 4 weeks for 48 weeks. Randomization was stratified by SLE Disease Activity Index 2000 score (<10 or ‡10), oral corticosteroid dosage (<10 or ‡10 mg/day), and type I IFN gene signature test status (high or low) based on a 4-gene expression assay. The primary end point was the percentage of patients achieving an SLE Responder Index (SRI[4]) response at week 24 with sustained reduction of oral corticosteroids (<10 mg/day and less than or equal to the dose at week 1 from week 12 through 24). Other end points (including SRI[4], British Isles Lupus Assessment Group [BILAG]-based Composite Lupus Assessment [BICLA], modified SRI[6], and major clinical response) were assessed at week 52. The primary end point was analyzed in the modified intent-to-treat (ITT) population and type I IFN-high subpopulation. The study result was considered positive if the primary end point was met in either of the 2 study populations. The Type I error rate was controlled at 0.10 (2-sided), within each of the 2 study populations for the primary end point analysis.Results. The primary end point was met by more patients treated with anifrolumab (34.3% of 99 for 300 mg and 28.8% of 104 for 1,000 mg) than placebo (17.6% of 102) (P 5 0.014 for 300 mg and P 5 0.063 for 1,000 mg, versus placebo), with greater effect size in patients with a high IFN signature at baseline (13.2% in placebo-treated patients versus 36.0% [P 5 0.004] and 28.2% [P 5 0.029]) in patients treated with anifrolumab 300 mg and 1,000 mg, respectively. At week 52, patients treated with anifrolumab achieved greater responses in SRI(4)ClinicalTrials.gov identifier: NCT01438489.
(40.2% versus 62.6% [P < 0.001] and 53.8% [P 5 0.043] with placebo, anifrolumab 300 mg, and anifrolumab 1,000 mg, respectively), BICLA (25.7% versus 53.5% [P < 0.001] and 41.2% [P 5 0.018], respectively), modified SRI(6) (28.4% versus 49.5% [P 5 0.002] and 44.7% [P 5 0.015], respectively), major clinical response (BILAG 2004 C or better in all organ domains from week 24 through week 52) (6.9% versus 19.2% [P 5 0.012] and 17.3% [P 5 0.025], respectively), and several other global and organ-specific end points. Herpes zoster was more frequent in the anifrolumab-treated patients (2.0% with placebo treatment versus 5.1% and 9.5% with anifrolumab 300 mg and 1,000 mg, respectively), as were cases reported as influenza (2.0% versus 6.1% and 7.6%, respectively), in the anifrolumab treatment groups. Incidence of serious adverse events was similar between groups (18.8% versus 16.2% and 17.1%, respectively).
Conclusion. Anifrolumab substantially reduced disease activity compared with placebo across multiple clinical end points in the patients with moderate-tosevere SLE.
Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease (1) leading to significant morbidity and shortened lifespan [bib_ref] The economic burden of systemic lupus erythematosus, Panopalis [/bib_ref] [bib_ref] Overall and cause-specific mortality in patients with systemic lupus erythematosus: a meta-analysis..., Yurkovich [/bib_ref]. Treatment of SLE is challenging because of the limited efficacy and poor tolerability of standard therapy [bib_ref] Lupus: improving long-term prognosis, Doria [/bib_ref] [bib_ref] Glucocorticoid use and abuse in SLE, Ruiz-Irastorza [/bib_ref].
Evidence supports activation of the type I interferon (IFN) system as a central pathogenic mediator in SLE [bib_ref] Targeting of type I interferon in systemic autoimmune diseases, Crow [/bib_ref] [bib_ref] Type I interferon in the pathogenesis of lupus, Crow [/bib_ref] [bib_ref] The type I interferon system in the development of lupus, R€ Onnblom [/bib_ref]. Cell signaling by all type I IFNs, including IFNa, IFNb, IFNE, IFNk, and IFNv, is mediated by the type I IFN-a/b/v receptor (IFNAR), resulting in IFNstimulated gene transcription, otherwise known as the IFN gene signature [bib_ref] Regulation of type I interferon responses, Ivashkiv [/bib_ref]. Consequently, blockade of IFNAR may reverse some of the immune dysregulation that occurs in SLE [bib_ref] Emerging therapies for systemic lupus erythematosus: focus on targeting interferon-a, Lichtman [/bib_ref] [bib_ref] Advances in understanding the role of type I interferons in systemic lupus..., Crow [/bib_ref].
Results of recent phase II clinical studies of anti-IFNa antibodies have been mixed. Rontalizumab did not meet the primary or secondary end points in a recent study [bib_ref] A phase II study of the efficacy and safety of rontalizumab (rhuMAb..., Kalunian [/bib_ref]. However, a post hoc analysis suggested potential benefit in a small subset of patients with a low baseline IFN signature. In contrast, sifalimumab met primary and some secondary end points [bib_ref] Sifalimumab, an anti-interferon-a monoclonal antibody, in moderate to severe systemic lupus erythematosus:..., Khamashta [/bib_ref] , but the treatment effects were modest. Both of these molecules have specificity only for IFNa, leaving other type I IFNs unaffected and able to bind IFNAR.
Anifrolumab is a fully human, IgG1k monoclonal antibody that binds to IFNAR and prevents signaling by all type I IFNs [bib_ref] Molecular basis for antagonistic activity of anifrolumab, an anti-interferon-a receptor 1 antibody, Peng [/bib_ref]. Anifrolumab exhibited a favorable safety profile and sustained inhibition of the type I IFN gene signature in a phase I study of patients with scleroderma [bib_ref] Dose-escalation of human anti-interferon-a receptor monoclonal antibody MEDI-546 in subjects with systemic..., Goldberg [/bib_ref]. Given the similarities in type I IFN activity in SLE and scleroderma [bib_ref] Pharmacogenomics and translational simulations to bridge indications for an anti-interferon-a receptor antibody, Wang [/bib_ref] the present phase IIb study was conducted to evaluate the efficacy and safety of 2 fixed intravenous dosages of anifrolumab in adults with moderately to severely active SLE with inadequate responses to standard therapy.
# Patients and methods
Patients. This study included patients ages 18-65 years with SLE. Patients had to weigh $40 kg and fulfill $4 of the 11 American College of Rheumatology 1997 classification criteria for SLE [bib_ref] Updating the American College of Rheumatology revised criteria for the classification of..., Hochberg [/bib_ref] [bib_ref] The 1982 revised criteria for the classification of systemic lupus erythematosus, Tan [/bib_ref]. At screening, all patients were required to have antinuclear antibodies and/or anti-double-stranded DNA (anti-dsDNA) antibodies, and/or anti-Sm antibodies, and to be receiving treatment with at least one of the following: oral prednisone (#40 mg/day or equivalent), azathioprine (#200 mg/day), an antimalarial, mycophenolate mofetil/mycophenolic acid (#2.0 gm/ day), or methotrexate (#25 mg/week).
Treatments for SLE had to be administered for at least 24 weeks prior to study entry and at stable dosages for $2 weeks (for prednisone and nonsteroidal antiinflammatory drugs) or $8 weeks (for other therapies) before screening. Biologic agents and protocol-prohibited immunosuppressants had to be discontinued before the study.
Patients had to meet all of the following disease activity criteria at screening: a score of $6 on the SLE Disease Activity Index 2000 (SLEDAI-2K) [bib_ref] Systemic Lupus Erythematosus Disease Activity Index, Gladman [/bib_ref] , excluding points attributable to lupus headache or organic brain syndrome; a British Isles Lupus Assessment Group (BILAG) 2004 (23) organ domain score of $1A or $2B (24); and a physician's global assessment of disease activity of $1 on a visual analog scale from 0 (none) to 3 (severe disease). In addition, patients were required to have a score of $4 in clinical components of the SLEDAI-2K (clinical SLEDAI-2K; points attributed to laboratory components were excluded) at week 1 (prior to receiving the study drug ). Patients were randomized 1:1:1 to receive intravenous infusions of placebo, anifrolumab 300 mg, or anifrolumab 1,000 mg. Treatment was administered every 4 weeks with the final dose administered at week 48. The primary and secondary efficacy measures were assessed at week 24 and week 52, respectively. All patients were required to complete a 12-week follow-up period after administration of the final dose of the study drug. During the treatment period, there were 13 scheduled monthly visits. The follow-up period consisted of 3 visits, conducted 4, 8, and 12 weeks after the last dose. Randomization was stratified by type I IFN gene signature (IFN high or IFN low), dosage of oral corticosteroids (,10 mg/day or $10 mg/ day of prednisone or equivalent), and SLEDAI-2K score (,10 or $10) at screening. IFN gene signature was determined at a central laboratory using an analytically validated 4-gene (IFI27, IFI44, IFI44L, and RSAD2) quantitative polymerase chain reaction (PCR)-based test from patients' whole blood [bib_ref] Use of type I interferon-inducible mRNAs as pharmacodynamic markers and potential diagnostic..., Yao [/bib_ref]. A predetermined, DC t -based cutoff point, in the trough of the bimodal distribution, was used to segregate patients with a high IFN gene signature from patients with a low IFN gene signature at baseline.
Tapering of oral corticosteroids was encouraged but was at the discretion of the investigators. Tapering was allowed after randomization except within 8 weeks of the primary (week 24) and secondary (week 52) end point assessments. A maximum of 2 oral corticosteroid bursts for increased SLE disease activity were allowed during the study. Patients with increased SLE disease activity could receive an oral corticosteroid burst between week 1 and week 10 (increase of #40 mg/day prednisone or equivalent), which had to be tapered to the week 1 dosage within 2 weeks of initiation of the burst; alternatively, a single intramuscular dose of methylprednisone (80 or 160 mg or equivalent) was permitted. The course of oral corticosteroid burst could not extend beyond week 10. Patients could receive an additional oral corticosteroid burst for increased SLE disease activity between week 24 and week 40 (increase of #20 mg/day prednisone or equivalent), which had to be tapered to the week 24 dosage within 2 weeks of initiation; alternatively, a single intramuscular dose of methylprednisone (80 mg or equivalent) was permitted. The course of oral corticosteroids could not extend beyond week 40.
This study was conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonisation Guidelines for Good Clinical Practice. Independent ethics committee or independent institutional review board approvals were obtained, and all patients provided written informed consent in accordance with local requirements.
An independent data safety and monitoring board was appointed for the study. Screening assessments of SLE organ system involvement and disease activity requirements were confirmed by an independent external adjudication group that also approved patient randomization and monitored study assessments.
Efficacy and safety evaluations. The primary efficacy end point was a composite of the SLE Responder Index (SRI [bib_ref] Overall and cause-specific mortality in patients with systemic lupus erythematosus: a meta-analysis..., Yurkovich [/bib_ref] [bib_ref] Novel evidence-based systemic lupus erythematosus responder index, Furie [/bib_ref] at week 24 with a sustained reduction in oral corticosteroids from week 12 through week 24 (,10 mg/day and less than or equal to the dose received at week 1). SRI [bib_ref] Overall and cause-specific mortality in patients with systemic lupus erythematosus: a meta-analysis..., Yurkovich [/bib_ref] Secondary efficacy measures included the SRI(4) response rates at week 52 with a sustained oral corticosteroid reduction from week 40 through week 52 and reduction of oral corticosteroid dosage at week 52 to #7.5 mg/day in patients who were receiving $10 mg/day at baseline. Other efficacy measures included percentages of patients with $50% improvement in Cutaneous Lupus Erythematosus Disease Area and Severity Index (CLASI) [bib_ref] The CLASI (Cutaneous Lupus Erythematosus Disease Area and Severity Index): an outcome..., Albrecht [/bib_ref] [bib_ref] Development of the CLASI as a tool to measure disease severity and..., Klein [/bib_ref] for patients with at least moderate skin involvement (CLASI $10); $50% improvement in swollen and tender joint count (28 joints assessed) for patients with $8 swollen and $8 tender joints at baseline; response in BILAG-based Composite Lupus Assessment (BICLA) [bib_ref] Efficacy and safety of epratuzumab in patients with moderate/severe active systemic lupus..., Wallace [/bib_ref] ; modified SRIs requiring SLEDAI-2K reductions of 5-8 points to be considered a responder; physician's global assessment; proportion of subjects with a SLEDAI-2K score of #2; proportion of subjects with a SLEDAI-2K score of 0; clinical SLEDAI; major clinical response defined as BILAG 2004 score of C or better in all organ domains at week 24 with maintenance of this response through week 52; .3-point improvement in Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) (30); Short Form 36 (SF-36) health survey; anti-dsDNA; and C3 and C4 complement concentrations.
To ensure that evaluation of SLE disease activity was consistent across study sites, training and certification was provided to investigators and designated site personnel, as appropriate, who were responsible for completing the following disease evaluation assessments: SLEDAI-2K, BILAG 2004, physician's global assessment, CLASI, and swollen and tender joint count evaluation. Where possible, evaluation of each patient was performed by a single assessor. Independent expert adjudication was undertaken throughout the study. Safety end points included adverse events, adverse events of special interest, laboratory assessments, and vital signs.
Statistical analysis. Analysis of the primary end point compared response rates at week 24 between each anifrolumab group and placebo using a logistic regression model adjusted for randomization stratification factors. The secondary end points and other binary end points were analyzed using the same approach as for the primary end point. Continuous end points were analyzed using an analysis of covariance model adjusted for randomization stratification factors, with the relevant baseline value as the covariate. For responder analyses, patients who were withdrawn from treatment, had increased use of oral corticosteroids beyond the protocol-permitted dosage, or had initiation of or an increase in dosage of immunosuppressant treatment any time after baseline were defined as nonresponders.
The primary end point was analyzed in the modified intent-to-treat (ITT) population (all randomized patients who received any dose of study drug and had baseline primary efficacy measurements) and a modified ITT subpopulation of patients with a high IFN gene signature at screening (IFN-high subpopulation). The study result was considered positive if the primary end point was met in either of the 2 study populations.
The Type I error rate was controlled at 0.10 (2-sided), within each of the 2 study populations (modified ITT population and IFN-high subpopulation) for the primary end point analysis, by performing a Cochran-Armitage trend test of all treatment groups prior to performing pairwise comparisons between each anifrolumab group and placebo. No multiplicity adjustment for the 2 study populations or other end points was applied.
The target sample size of 100 patients per group was based on providing 88% power at the 0.10 alpha level to detect at least 20% absolute improvement in SRI(4) response rate at week 24 for anifrolumab relative to placebo, assuming a 40% placebo response rate. All data analyses were conducted using the SAS system (SAS Institute).
# Results
Study population. Between January 2012 and January 2014, 626 patients were screened, with 307 randomized to receive treatment (see Supplementary [fig_ref] Figure 1: Efficacy results at week 52 in patients with systemic lupus erythematosus [/fig_ref] , available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.399 62/abstract). One patient in the anifrolumab 300-mg group had poor peripheral venous access, and 1 patient in the placebo group had an entry criteria violation and did not receive study treatment. The modified ITT population therefore consisted of 305 patients (102 receiving placebo, 99 receiving anifrolumab 300 mg, and 104 receiving anifrolumab 1,000 mg). One patient randomized to the placebo group mistakenly received a single dose of anifrolumab (1,000-mg). This was classified as a protocol deviation. The patient was included in the placebo group of the modified ITT population for the efficacy analyses, but in the anifrolumab 1,000-mg group of the safety population for the safety analyses. Baseline characteristics [fig_ref] Table 1: Demographic and baseline clinical characteristics of the patients with SLE [/fig_ref] were similar between groups, with the following exceptions: disease duration was shorter for the placebo group than for the anifrolumab groups (mean 90.6 months versus 95.9 months and 100.1 months), and fewer patients in the placebo group were receiving methotrexate (15.7% versus 19.2% and 24.0%). The anifrolumab 1,000-mg group had a greater percentage of white patients than the anifrolumab 300-mg and placebo groups (49.0% versus 35.4% and 40.2%, respectively).
Efficacy. Based on the mechanism of action of anifrolumab, the primary end point (SRI [bib_ref] Overall and cause-specific mortality in patients with systemic lupus erythematosus: a meta-analysis..., Yurkovich [/bib_ref] response with sustained reduction of oral corticosteroids at week 24) was analyzed in the overall modified ITT population and the IFN-high subpopulation. The predefined statistical significance level was 0.1; therefore, P values less than 0.1 were considered significant for the analyses of the primary end point. Multiplicity of comparing 2 active dosages with placebo within each study population was adjusted by performing a Cochran-Armitage trend test of all treatment groups prior to performing pairwise comparisons between each anifrolumab group and placebo. No multiplicity adjustment for the 2 study populations was applied. The Cochran-Armitage trend test of all treatment groups showed that the number of patients in whom the primary end point was achieved was greater for anifrolumab versus placebo in the modified ITT population (P 5 0.072) and in the IFN-high subpopulation, which included 75% of the patients (P 5 0.034). In pairwise comparisons, both anifrolumab groups had greater response rates compared with placebo (17.6%). A response was achieved in 34.3% of the patients treated with anifrolumab 300 mg (P 5 0.014 versus placebo) and in 28.8% of the patients treated with anifrolumab 1,000 mg (P 5 0.063 versus placebo). The response rates in the IFN-high subpopulation were 13.2% for the placebo group, 36.0% for the anifrolumab 300-mg group (P 5 0.004 versus placebo), and 28.2% for the anifrolumab 1,000-mg group (P 5 0.029 versus placebo) (see [fig_ref] Figure 2: Efficacy results over time in patients with systemic lupus erythematosus [/fig_ref] , available on the Arthritis & Rheumatology web site at http://onlinelibrary. wiley.com/doi/10.1002/art.39962/abstract). At week 52, anifrolumab treatment groups demonstrated improvements over placebo across a broad range of global disease activity measures. In the whole modified ITT population, 25.5% of the placebo group achieved SRI(4) response with sustained reduction of oral corticosteroids, compared with 51.5% in the anifrolumab 300-mg group (P , 0.001) and 38.5% in the anifrolumab 1,000-mg group (P 5 0.048). Response rates in the IFNhigh population were 19.7% for placebo, 52.0% for the anifrolumab 300-mg group (P , 0.001), and 38.5% for the anifrolumab 1,000-mg group (P 5 0.013). No differences from placebo were seen in the patients with a low IFN gene signature at screening (IFN-low subpopulation) at week 24 or week 52 (Table 2 and [fig_ref] Figure 1: Efficacy results at week 52 in patients with systemic lupus erythematosus [/fig_ref].
Consistent with the primary end point results, a greater percentage of patients receiving anifrolumab met criteria for SRI(4) response without the oral corticosteroid taper requirement [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref] and [fig_ref] Figure 1: Efficacy results at week 52 in patients with systemic lupus erythematosus [/fig_ref] , and the modified SRI responses, requiring 5-, 6-, 7-, or 8-point reductions in SLEDAI-2K scores. Both anifrolumab dosages were associated with higher rates of response in the BICLA compared with placebo (Table 2, [fig_ref] Figure 2: Efficacy results over time in patients with systemic lupus erythematosus [/fig_ref] maintained or increased through week 52 [fig_ref] Figure 2: Efficacy results over time in patients with systemic lupus erythematosus [/fig_ref]. Improvements in physician's global assessment at week 52 were also greater for the anifrolumab-treated groups than for the placebo-treated group [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref] , with a greater effect observed in the 300-mg treatment group compared with the 1,000-mg treatment group. For all outcomes, the treatment effects observed in the overall population were driven by the IFN-high subset [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref] and [fig_ref] Figure 2: Efficacy results over time in patients with systemic lupus erythematosus [/fig_ref].
The study also assessed various measures of low disease activity. A SLEDAI-2K score of #2 at week 52 was achieved in 17.6% in the placebo group, 35.4% in the anifrolumab 300-mg group (P 5 0.004 versus placebo), and 32.7% in the anifrolumab 1,000-mg group (P 5 0.012 versus placebo) (see , available on the Arthritis & Rheumatology web site at http://onlinelibrary. wiley.com/doi/10.1002/art.39962/abstract), and a score of 0 was achieved in 7.8% in the placebo group, 18.2% in the anifrolumab 300-mg group (P 5 0.033 versus placebo), and 19.2% in the anifrolumab 1,000-mg group (P 5 0.018 versus placebo). Major clinical response, defined as a BILAG 2004 score of C or better in all organ domains at week 24 with maintenance of this response through week 52, was achieved in 6.9%, 19.2% (P 5 0.012), and 17.3% (P 5 0.025) in the placebo, anifrolumab 300-mg, and anifrolumab 1,000-mg groups, respectively [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref] and .
The numbers of patients who developed new BILAG 1A or 2B flares at any time during the study were reduced by 28% and 29%, and the numbers who developed new BILAG A flares were reduced by 46% and 35%, in the anifrolumab 300-mg and anifrolumab 1,000-mg groups, respectively [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref].
A reduction in background oral corticosteroid dosage to #7.5 mg/day at week 52 in those taking $10 mg/day at baseline was achieved in 56.4% of the patients in the anifrolumab 300-mg group (P 5 0.001) and 31.7% of the patients in the anifrolumab 1,000-mg group (P 5 0.595), compared with 26.6% of the patients receiving placebo.
Anifrolumab-treated patients demonstrated greater improvements than those receiving placebo across a range of organ-specific disease measures and patient-reported outcomes. The percentage of patients with a baseline CLASI activity score of $10 who had a $50% reduction in this score by week 52 was greater for both anifrolumab dosages (63.0% for 300 mg [P 5 0.013] and 58.3% for 1,000 mg [P 5 0.077]) compared with placebo (30.8%) [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref] and [fig_ref] Figure 3: Organ-specific efficacy results over time in patients with systemic lupus erythematosus [/fig_ref]. Improvement in arthritis was also greater in patients receiving anifrolumab [fig_ref] Figure 3: Organ-specific efficacy results over time in patients with systemic lupus erythematosus [/fig_ref] and . In patients with $8 swollen and $8 tender joints at baseline, a $50% decrease in both the swollen and tender joint count was achieved in a greater percentage of anifrolumab-treated patients [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref]. Assessments of biologic parameters were undertaken to investigate response following administration of anifrolumab. In patients with anti-dsDNA antibodies detectable by the AtheNA Multi-Lyte ANA-II Plus assay at baseline, a numerically greater decrease in anti-dsDNA antibodies was observed with anifrolumab than with placebo (mean 6 SD 224. The median neutralization ratios of a 21gene type I IFN signature in the IFN-high subpopulation at week 24 were 89.7 and 91.7 for the anifrolumab 300-mg and anifrolumab 1,000-mg groups, respectively. This degree of suppression was maintained until week 52. No neutralization in gene expression was observed with placebo. The rates of antidrug antibodies at any time postbaseline were low, with no difference between treatment groups (3% for placebo, 5.1% for anifrolumab 300 mg, and 2.0% for anifrolumab 1,000 mg).
Safety. The safety population consisted of the 305 patients who received at least 1 dose of study drug. One patient randomized to the placebo group mistakenly received a single dose of anifrolumab (1,000 mg). This patient was included in the 1,000-mg anifrolumab group for the safety analyses. The percentages of patients with at least 1 adverse event (77.2% in the placebo group, 84.8% in the anifrolumab 300-mg group, and 85.7% in the anifrolumab 1,000-mg group) and serious adverse events (18.8% in the placebo group, 16.2% in the anifrolumab 300-mg group, and 17.1% in the anifrolumab 1,000-mg group) were similar across all treatment groups. Adverse events led to treatment discontinuation in 7.9% of the patients receiving placebo and 3.0% and 9.5% of the patients receiving anifrolumab 300 mg and anifrolumab 1,000 mg, respectively. The most frequently reported adverse events were headache, upper respiratory tract infection, nasopharyngitis, and urinary tract infection [fig_ref] Table 3: Adverse events [/fig_ref]. There was 1 death (due to sepsis syndrome) prior to randomization, and 1 death during the study in a patient who received 1 dose of anifrolumab 1,000 mg. The patient had acute colitis with a rapidly deteriorating clinical course and had features of macrophage activation syndrome prior to death. No infectious agent was identified. Autopsy confirmed the cause of death to be acute colitis and revealed portal vein thrombosis and increased numbers of phagocytic macrophages in the bone marrow, consistent with macrophage activation syndrome. This was not considered related to the study drug by the investigator.
Herpes zoster infections were reported in 5.1% and 9.5% of the patients in the anifrolumab 300-mg and 1,000-mg groups, respectively, compared with 2.0% in the placebo group. All patients responded promptly to antiviral treatment. One patient in the 300-mg group discontinued treatment due to transverse myelitis with a positive PCR test for herpes zoster in the cerebrospinal fluid; the patient responded to antiviral and high-dose corticosteroid treatment. The percentages of patients who were reported as developing influenza were greater in the anifrolumab 300-mg group (6.1%) and anifrolumab 1,000-mg group (7.6%) than in the placebo group (2.0%). Infusion-related reactions were reported in 6 (5.9%), 2 (2.0%), and 4 (3.8%) of the patients in the placebo, anifrolumab 300-mg, and anifrolumab 1,000-mg groups, respectively. Upper respiratory infections (reported under various terms) occurred more frequently in patients treated with anifrolumab (36.4% in the anifrolumab 300mg group and 41.9% in the anifrolumab 1,000-mg group) compared with placebo (28.7%). No clinically important worsening was observed in hematology or chemistry panels, urinalysis, vital signs, lipid parameters, or electrocardiograms in any of the treatment groups.
# Discussion
The primary and secondary end points of this study were met. Furthermore, compared with placebo, anifrolumab treatment resulted in significantly greater rates of improvement across a broad range of composite and organ-specific disease activity measures as well as in the achievement and maintenance of low disease activity, corticosteroid tapering (in the 300-mg group), and a trend toward flare rate reduction. These results provide compelling evidence that blocking IFNAR is a promising strategy in the treatment of SLE.
The analysis of the primary end point in 2 populations (the modified ITT population and a modified ITT subpopulation of patients with a high IFN gene signature at screening) was based on the mechanism of action of anifrolumab and the assumption that demonstrating statistically significant benefit in the primary end point at a 2-sided alpha level of 0.1 in either of the populations would provide sufficient preliminary evidence of efficacy to continue phase III studies in the appropriate population. A 2-sided alpha level of 0.1 represents a 5% chance of declaring a positive study result when there is no treatment effect (risk of proceeding with an ineffective drug). The present study had a power of 88%, representing a 12% chance of declaring a negative study result when there was a positive treatment effect (risk of discontinuing development of a potentially efficacious drug). This combination of statistical risks was chosen to balance the continuation and discontinuation risks while maintaining a feasible phase IIb study. Further, to keep the sample size within reasonable limits, no adjustment was made for the multiplicity of testing the primary end point in the 2 populations. Therefore, the results of this study are clinically relevant, but are not definitive until prospectively replicated in larger studies with a more stringent alpha level and a more robust control for multiplicity.
The anifrolumab dosages were selected based on pharmacokinetic and pharmacodynamic modeling and simulation from a previous study [bib_ref] Dose-escalation of human anti-interferon-a receptor monoclonal antibody MEDI-546 in subjects with systemic..., Goldberg [/bib_ref] [bib_ref] Pharmacogenomics and translational simulations to bridge indications for an anti-interferon-a receptor antibody, Wang [/bib_ref]. Both the 300mg and 1,000-mg dosages were more efficacious than placebo, with the lower dosage showing greater numerical response rates at week 52 for many end points [fig_ref] Table 2: Summary of efficacy results for patients with SLE treated with placebo or... [/fig_ref]. The fact that the degrees of IFN gene signature inhibition were similar with the low and high anifrolumab dosages supports a plateau effect rather than an inverse dosage response. However, other mechanisms, such as an increased response in inducible immunologic feedback leading to a lower rate of response with the greater anifrolumab dosage, cannot be excluded. Biomarker studies to address this question are currently underway.
The fact that greater efficacy was seen in all end points in patients with high baseline IFN gene signatures compared with those with low baseline IFN gene signatures supports the pathobiology of this treatment strategy. The lack of treatment effect in the IFN-low subgroup was due to a greater response to placebo and standard therapy and not to a lower response to anifrolumab, suggesting that the IFN-high subgroup represents a subpopulation more likely to benefit from the addition of anifrolumab to standard therapy. The interpretation of the results in the IFN-low subgroup is limited by its small size, and thus larger studies are needed to fully evaluate the effect of anifrolumab in this population.
Anifrolumab treatment was well tolerated, and adverse events reported were similar across the 3 treatment groups. Consistent with the mechanism of action of anifrolumab, a dosage-related increase was observed in the occurrence of upper respiratory infections and reactivation of herpes zoster. Importantly, responses to antiviral treatment were prompt in the herpes zoster cases. In addition, there were more cases reported as influenza in patients treated with anifrolumab; however, most of these cases were not confirmed by laboratory testing or the application of strict clinical criteria. To accurately assess the true risk of influenza, future studies should require laboratory confirmation of influenza infections. Ensuring that patients' immunizations are up to date prior to administration of anifrolumab may decrease the risk of influenza and other infections. The dosage-related increase in these infections, together with the fact that increasing the dosage from 300 mg to 1,000 mg did not lead to an increase in efficacy, indicates a more favorable risk-benefit profile for the 300mg dosage. This is the most successful phase II study in SLE to date. The benefits achieved across multiple global and organ-specific disease activity measures compare favorably to those observed in other recent studies of potential lupus therapies [bib_ref] A phase II study of the efficacy and safety of rontalizumab (rhuMAb..., Kalunian [/bib_ref] [bib_ref] Sifalimumab, an anti-interferon-a monoclonal antibody, in moderate to severe systemic lupus erythematosus:..., Khamashta [/bib_ref] [bib_ref] Efficacy and safety of epratuzumab in patients with moderate/severe active systemic lupus..., Wallace [/bib_ref] [bib_ref] A phase 2, randomised, placebo-controlled clinical trial of blisibimod, an inhibitor of..., Furie [/bib_ref]. The phase IIb study evaluating sifalimumab in a similar cohort of patients met its primary and some of its secondary end points, albeit with smaller treatment effects [bib_ref] Sifalimumab, an anti-interferon-a monoclonal antibody, in moderate to severe systemic lupus erythematosus:..., Khamashta [/bib_ref]. The greater efficacy and broader impact demonstrated by anifrolumab are likely the result of achieving greater suppression of the type I IFN pathway. The positive results of the present study suggest that targeting IFNAR is a promising and novel therapeutic approach for patients with SLE whose disease does not respond to currently available therapies.
[fig] Figure 1: Efficacy results at week 52 in patients with systemic lupus erythematosus (SLE) receiving anifrolumab 300 mg, anifrolumab 1,000 mg, or placebo. Anifrolumab treatment led to a greater rate of response across multiple end points. The benefit observed in the overall modified intent-to-treat population was driven by the patients with a high interferon (IFN) gene signature (IFN test high subpopulation), which represents ;75% of the entire cohort. The odds ratios (ORs), 90% confidence intervals (90% CIs), and P values are from a logistic regression model adjusted for stratification factors. SRI(4) 5 SLE Responder Index requiring a $4-point reduction in SLE Disease Activity Index 2000 (SLEDAI-2K) score; OCS 5 oral corticosteroid; BICLA 5 British Isles Lupus Assessment Group 2004-based Combined Lupus Assessment. [/fig]
[fig] Figure 2: Efficacy results over time in patients with systemic lupus erythematosus (SLE) receiving anifrolumab 300 mg, anifrolumab 1,000 mg, or placebo. Treatment was given every 4 weeks from week 1 to week 48. A, Proportion of patients achieving an SLE Responder Index response over time in the modified intent-to-treat (ITT) population and in the interferon (IFN)-high and IFN-low subsets. B, Proportion of patients achieving a British Isles Lupus Assessment Group 2004-based Combined Lupus Assessment response over time in the modified ITT population and in the IFN-high and IFN-low subsets.382 FURIE ET AL than placebo-treated patients (48.6% for placebo, 69.6% for anifrolumab 300 mg [P 5 0.038], and 64.6% for anifrolumab 1,000 mg [P 5 0.156]). Compared with placebo, greater proportions of patients in the anifrolumab 300-mg group achieved a .3-point improvement in the FACIT-F scale, a $3.1-point improvement in the physical component summary score of the SF-36 health survey, and a $3.8-point improvement in the mental component summary score of the SF-36 health survey. However, none of these comparisons achieved statistical significance ( [/fig]
[fig] 4: 6 177.6 for placebo, 270.3 6 166.9 for anifrolumab 300 mg [P 5 0.067], and 243.8 6 86.3 for anifrolumab 1,000 mg [P 5 0.144]). Similarly, in patients with low complement C3 concentrations at baseline, a nonsignificant numerically greater increase in complement C3 concentrations was observed at week 52 in patients treated with anifrolumab compared with those treated with placebo (mean 6 SD 7.7 6 18.9 for placebo, 12.9 6 19.1 for anifrolumab 300 mg [P 5 0.277], and 12.0 6 16.5 for anifrolumab 1,000 mg [P 5 0.242]) (see Supplementary Figure 6, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10. 1002/art.39962/abstract). No effect on complement C4 was observed. [/fig]
[fig] Figure 3: Organ-specific efficacy results over time in patients with systemic lupus erythematosus (SLE) receiving anifrolumab 300 mg, anifrolumab 1,000 mg, or placebo. Treatment was given every 4 weeks from week 1 to week 48. A, Left, Proportion of patients with a Cutaneous Lupus Erythematosus Disease Area and Severity Index (CLASI) activity score of $10 at baseline (n 5 77) who had $50% improvement in the CLASI score. Right, A representative example of skin response following anifrolumab treatment. B, Left, Proportion of patients with $8 swollen and $8 tender joints at baseline (n 5 37 in the placebo group, n 5 46 in the anifrolumab 300-mg group, and n 5 48 in the anifrolumab 1,000-mg group) who had $50% improvement in the swollen and tender joint count. Right, Mean 6 SEM change from baseline in the active joint count in the modified intent-to-treat population. Joint counts are based on the assessment of 28 joints. [/fig]
[table] Table 1: Demographic and baseline clinical characteristics of the patients with SLE (modified ITT population)* * Except where indicated otherwise, values are the mean 6 SD. Treatment was given every 4 weeks from week 1 to week 48. SLE 5 systemic lupus erythematosus; ITT 5 intent-to-treat; IFN 5 interferon; BILAG 2004 5 British Isles Lupus Assessment Group 2004; CLASI 5 Cutaneous Lupus Erythematosus Disease Area and Severity Index; anti-dsDNA 5 anti-double-stranded DNA.† From diagnosis to study entry. ‡ Based on the assessment of 28 joints. § The multiplex assay, an AtheNA Multi-Lyte ANA-II Plus test system, was used for screening and to calculate Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) scores throughout the study. The differences between the multiplex and Farr assays were due to a low sensitivity cutoff point of the multiplex assay (Farr assay $5 IU/ml; multiplex assay $100 IU/ml)(32). ¶ Data were available for 82 patients receiving placebo, 77 patients receiving anifrolumab 300 mg, and 82 patients receiving anifrolumab 1,000 mg. [/table]
[table] Table 2: Summary of efficacy results for patients with SLE treated with placebo or anifrolumab* Values are the number of patients/number assessed (%). Treatment was given every 4 weeks from week 1 to week 48. The observed response rates of each treatment group are shown. SLE 5 systemic lupus erythematosus; SRI 5 SLE Responder Index; IFN 5 interferon; BICLA 5 British Isles Lupus Assessment Group (BILAG)-based Combined Lupus Assessment; SRI(5) 5 SRI requiring SLE Disease Activity Index 2000 (SLEDAI-2K) reductions of 5 points; SF-365 Short Form 36; MCS 5 mental component summary; PCS 5 physical component summary. † Odds ratio (ORs), 90% confidence intervals (90% CIs), and nominal P values are from a logistic regression model for comparisons of each anifrolumab group versus placebo adjusted for randomization stratification factors.‡ Reduction of oral corticosteroid dosage to #7.5 mg/day in patients who were receiving $10 mg/day at baseline. § $50% decrease in the Cutaneous Lupus Erythematosus Disease Area and Severity Index (CLASI) activity score in patients who had a score of $10 at baseline. ¶ $50% decrease in the swollen and tender joint count from baseline in patients with $8 swollen and $8 tender joints at baseline. [/table]
[table] Table 3: Adverse events (safety population)* * Values are the number (%). Treatment was given every 4 weeks from week 1 to week 48. † New or reactivated tuberculosis infection, herpes zoster infection, malignancy, or reactions associated with infusion, hypersensitivity, or anaphylaxis. ‡ Adverse events (preferred term) reported by .5% of patients in the total (both doses) anifrolumab group. § One patient also had transverse myelitis with a quantitatively positive test result for varicella-zoster virus in the cerebrospinal fluid. [/table]
|
Electricity, Neurology, and Noninvasive Brain Stimulation: Looking Back, Looking Ahead
Electricity and neurology evolved synchronously over the past few centuries. is article looks at their origins and their journey into noninvasive brain stimulation technique of transcranial direct current stimulation (tDCS), which is now popular in neuroscience research.
# Introduction
On a long journey, it is useful to look back and figure out how far we have come, to know which way we are heading. Centuries ago, electricity and neurology started a journey together which has taken us a long way. is article is a brief account of that journey into noninvasive brain stimulation.
From prehistoric times, electricity fascinated humans, whether it came as thunderous lightning bolts "from the heavens" or as electric sparks from rubbing stones. From time immemorial, we have attempted to harness the power of electricity. An electric fish applied to the scalp for treating headache was one of the earliest documented therapies. From violet ray to tDCS (transcranial direct current stimulation), there have been waves of interest in noninvasive brain stimulation. Attempts have been plentiful, results neither convinces nor ignorable. Some have enjoyed great success, while others have perished.
## Part 1: looking back
## Electricity: origin.
e word "Electricity" came from the Latin word "electricus" meaning "amber like," amber being a Greek word. Materials like fur when rubbed against amber could accumulate enough static charge and create a spark of "electricity." ales of Miletus (600 BC), also known as the "Father of Science," was the first person to study static electricity. is was in an era where anything that moved was alive, even if it were a lodestone attracted to iron by magnetism. William Gilbert (1544-1603), a physician to Queen Elizabeth I, who wrote volumes on magnetism coined the term "Electricity".
## E electric eel (electrophorus electricus).
If we look further back, it is conceivable that the greatest contributor to electricity and neurology is the electric eel. e characteristic feature of this fish is its ability to deliver an electric shock, either to its prey or to its opponent of such high magnitude as to kill them or numb them in an instant. It was called the " underer of the Nile" for it could even immobilize a horse. ere are several species of electric fish, with the torpedo fish and electric eel being the common ones. e numbing and shocking effects were used to treat headaches, seizures, and pain. e fish made its way into ancient Greek, Roman, and Arabic medicine. e drawings of the fish can be found in Egyptian tombs.
## Electricity: from
Fish to Jar. Early researchers attempted to conduct the electricity generated by the fish into circuits using wires. Interest later evolved from conducting electricity to the generation of electricity. Otto von Guericke developed the first static generator in 1672. It got perfect over the next century, with the Leyden jar in 1742 becoming the most promising device. e first case of electric healing came soon after with Jean Jallabert, Professor of Physics in Geneva, who connected the paralyzed foot of a stroke victim to a Leyden jar and made his paralyzed extremity move. e patient improved weeks later and could use his extremity. In 1752, Benjamin Franklin attempted to treat epileptic fits using electricity from Leyden jar, and the therapy came to be known as "Franklinism".
## Animal electricity.
e focal point setting forth the synchronous evolution of electricity and neurology was a serendipitous experiment in 1780 by Luigi Galvani, the famous Italian physicist and physician. Galvani was dissecting a dead frog on a table used for electrical experiments. e frog's leg kicked as his assistant touched the sciatic nerve with the scalpel. Unaware of the static currents picked up by the scalpel, Galvani thought an "animal electricity" had "manifested." e finding created the notion of "bioelectricity" driving the muscle contraction, which got termed "Galvanism." Giovanni Aldini, nephew and disciple of Galvani, was a strong proponent of "Galvanism." He held public demonstrations of "Galvanism" where he stimulated the face, diaphragm, and extremities of a guillotined criminal to generate facial expression, breathing, and dancing movements. e horror of watching the "dead come alive with electricity" became widespread. Mary Shelley's "Frankenstein" is a product of an era of "Franklinism" and "Galvanism".
## E electric era.
e concept of "Galvanism" got refuted and proven wrong by Alessandro Volta. Based on the anatomic descriptions of the electric fish's nerve plates, he used dissimilar metal plates soaked in brine to invent the first electric "battery".
Faraday's experiment helped create alternating currents. We can find scores of reports during the following two centuries where electricity flourished in the treatment of neurologic and psychiatric conditions ranging from the use of Faradic currents to treat tabes by Joseph Babinski and "melancholy" treatment using voltaic battery by Aldini.
Nicola Tesla, a genius way ahead of his times, developed the alternating current from its origins by Michael Faraday. He met with Paul Oudin and conceived ideas of developing electrotherapeutic instruments. Paul Oudin built the "Violet Ray," which found a variety of therapeutic uses from 1890 to 1910s. An array of electrical devices ranging from violet rays for transcutaneous stimulation to Garceau nerve stimulator for direct cortical stimulation were made. However, these devices were proven nonbeneficial, got discontinued, or wiped out by great depression. Some of them also went into disrepute as "Quack Devices" ending up as exhibits like the ones at the Museum of Questionable Medical Devices at Bakken in Minneapolis.
Postdepression era saw a resurgence in electrostimulation techniques in psychiatry. An inverse relation was seen between schizophrenia and epilepsy. Treatment of schizophrenia involved inducing convulsions by injecting camphor or insulin, which were suboptimal, unsafe, and life-threatening. Ugo Cerletti and Lucio Bini watched butchers using electric pincers to cause convulsions and immobilize pigs in slaughterhouses. ey adapted the technology to the field of psychiatry for electroconvulsive therapy (ECT), and it became the greatest electrical therapy to date.
## Brain polarization.
Masked by the success of ECT, brain polarization experiments with smaller currents remained less prominent during this period. In the 1950s, there was renewed interest in peripheral nerve stimulation. Peripheral nerve polarization was shown to change the potentials recorded from the brain. In addition, direct brain polarization studies in animals showed long-lasting changes in cortical excitability. Influenced by these findings, polarization studies were conducted in humans that found significant benefits in depressed and psychotic patients resistant to other forms of treatment, including ECT. Currents varying from 50 to 500 μA were used, which showed that anodal currents improved alertness, mood, and motor activity, whereas cathodal polarization produced quietness and apathy. However, the interest in noninvasive stimulation faded after inability to replicate the results.
Cranial electrostimulation (CES) device which is a derivative of the earlier studies was used for the treatment of depression, fibromyalgia, and sleep till recently, when it got downgraded to a class III device by the FDA because of inadequate efficacy.
## Part 2: looking ahead
## Revival of brain polarization using noninvasive
Measures. In a 1962 article in Nature, Bindman et al. described short polarizing currents on the cerebral cortex of the rat, which created prolonged changes in cortical excitability. Decades later, Priori et al. studied these effects in humans using weak anodal and cathodal currents of <0.5 mA for 7 seconds with excitability changes determined by transcranial magnetic stimulation (TMS). Anodal followed by cathodal currents resulted in a lowered cortical excitability. Nitsche et al. further refined this technique with stronger currents of 0.2-1.0 mA for a longer duration of 1-5 minutes and showed an increase in motor cortical excitability following anodal stimulation and a decrease in excitability with cathodal stimulation. e residual effects were detectable up to 90 minutes and sensorimotor and cognitive effects up to 30 minutes after stimulation. Long-term potentiation or depression is thought to be responsible for the long-lasting effects. ese findings paved the way for a flurry of new research into the field of transcranial direct current stimulation.
## Polarizing the brain using tdcs.
e materials required for tDCS are a constant current stimulator and scalp electrodes. ese stimulators use currents of 1-4 mA. Electrodes are placed over predetermined sites based on the 10-20 EEG system. Direction of flow determines the polarization of the underlying tissues. When current goes from 2
Neurology Research International anodal pole to the reference electrode, it creates a negative field in the underlying brain tissue that lowers the threshold for depolarization. When the current is reversed, as in cathodal stimulation, an opposite effect of hyperpolarization and inhibition occurs. tDCS modulates the threshold for firing, without generating a neural impulse. Studies on brain models with current density distributions have shown that despite a large percentage of charge being shunted through the scalp, tDCS carries adequate currents to the underlying cortex to impact neuronal excitability. Surface electrodes induced currents of around 1.5 mA are decipherable by intracranial electrodes during presurgical evaluation studies. tDCS has been shown to change the regional blood flow of targeted sites and their connected regions.
## Neuroplastic effects of brain
Polarization. tDCS holds the potential for helping regeneration of injured brain cells and aids neuroplasticity. It may also help regulate the interhemispheric inhibition from the contralesional hemisphere into the ipsilesional hemisphere. We can try to understand this concept with an exercise. Raise your right arm and try drawing the number "6" in the air while trying to draw a "9" with the other arm. e thing that is troubling you is the interhemispheric inhibition. In a healthy individual, there is a synchronous inhibition and facilitation that helps in bimanual tasks. at gains importance when there is an interhemispheric imbalance as in a stroke. e inhibitory contralesional hemisphere worsens ipsilesional hemispheric function. We have shown bi-hemispheric stimulation with cathodal inhibition over the normal hemisphere and anodal facilitation of the lesional hemisphere to improve functional recovery in chronic stroke patients.
## Tdcs as a research
Tool. During the past decade, there has been an explosion in tDCS-based studies. Being a low cost set up, minimal to no side effects, ease of operation, and availability of commercial devices, tDCS has become popular in research settings. ere are over a thousand active trials going on with tDCS. In addition, newer techniques like tACS-transcranial alternating current stimulation, tRNStranscranial random noise stimulation, and tPCS-transcranial pulsed current stimulation are being marketed. Expensive high definition arrays devices to target focal areas limiting the current spread are also available. From cognitive research to clinical applications, there are not many areas left in neuroscience, which tDCS has not touched by now. In addition, commercial devices that work like tDCS devices are being made available as "neuromodulation devices" to treat migraine, depression, pain, fibromyalgia, and other conditions, even though there are no FDA approved indications to use tDCS as a treatment modality.
## Tdcs as a cognitive
Enhancer. tDCS enhances a variety of cognitive tasks and performance, which includes working memory, decision making, motor learning, and speed and vocabulary, just to name a few. is led to the trials of tDCS in autistic disorders, mood disorders, dementia, and neuropsychiatric illnesses, which have all shown benefits. tDCS devices adopted and commercialized for gaming head bands and cognitive enhancers are now available to purchase for the public. tDCS is also being used in military training for improving reaction times.
3.6. tDCS: Caveats. Even though tDCS has found widespread usage, the validity is being questioned. e fundamental issue has been a lack of explanation on how the current passes through the skin and skull to the targeted brain tissue and creates only desired effects. e role of the magnetic fields created by tDCS on an electrical brain is still unclear.
# Conclusion
We have a never-ending fascination with electricity and its application to neuroscience. Every few decades, there seems to be a rise and fall in a different version of the same technology. From the violet rays in the 1910sto the polarization studies in the 1950sand cranial electrotherapy stimulation (CES) devices in the 1980s, there has been much interest in the application of electricity to neurology. tDCS could be the new iteration for this decade, trying to unlock a potential that can achieve success akin to ECT. It remains to be seen if that promise delivers. For all we know-in this long journey-a small possibility remains that we are going around in circles.
## Conflicts of interest
e authors declare that they have no conflicts of interest.
Neurology Research International |
Long noncoding RNA HCP5 participates in premature ovarian insufficiency by transcriptionally regulating MSH5 and DNA damage repair via YB1
The genetic etiology of premature ovarian insufficiency (POI) has been well established to date, however, the role of long noncoding RNAs (lncRNAs) in POI is largely unknown. In this study, we identified a down-expressed lncRNA HCP5 in granulosa cells (GCs) from biochemical POI (bPOI) patients, which impaired DNA damage repair and promoted apoptosis of GCs. Mechanistically, we discovered that HCP5 stabilized the interaction between YB1 and its partner ILF2, which could mediate YB1 transferring into the nucleus of GCs. HCP5 silencing affected the localization of YB1 into nucleus and reduced the binding of YB1 to the promoter of MSH5 gene, thereby diminishing MSH5 expression. Taken together, we identified that the decreased expression of HCP5 in bPOI contributed to dysfunctional GCs by regulating MSH5 transcription and DNA damage repair via the interaction with YB1, providing a novel epigenetic mechanism for POI pathogenesis.
# Introduction
Premature ovarian insufficiency (POI), defined as a cessation of menstruation prior to the age of 40 years, mainly manifests with irregular menstruation, elevated follicle-stimulating hormone (FSH >25 IU/l), and estrogen deficiency. POI is one of the most common reproductive endocrine disorders, which affects ∼1-2% of women of childbearing age. POI encompasses three stages in clinic, i.e. occult, biochemical and overt (formerly called premature ovarian failure) stage. Patients with biochemical POI (bPOI) typically have regular menstruation, but elevated FSH levels and reduced fertility. The etiology of POI is clinically challenging involving genetic, autoimmune, metabolic, and infectious factors. However, the pathogenesis remains to be elucidated in most cases. Of all the causes, genetic defects account for 20-25% of patients, including chromosomal abnormalities and gene mutations. Causative genes of POI have been extensively studied to date in coding regions with alterations to disrupt protein function. However, the protein-coding region only accounts for 1.5% of the whole human genome. The noncoding RNAs, including microRNAs, long noncoding RNAs (lncRNAs) and circRNAs, have recently begun to be explored in ovaries and human diseases.
LncRNAs are a group of noncoding RNAs which are longer than 200 nucleotides and poorly conserved among species. Large-scale transcriptome studies suggest that lncRNAs modulate the expression of protein-coding genes by altering chromatin modification, transcription, mRNA decay, protein subcellular localization and other processes. Also, lncRNAs have been proved to participate in the physiology and pathophysiology of neural, endocrine, and cardiovascular systems. Recently, 20 differentially expressed lncRNAs have been discovered in ovarian cortex from POI patients, which suggested that lncRNAs may be involved in the maintenance of ovarian function. However, the mechanism of lncRNAs contributing to human POI has yet to be determined.
Folliculogenesis is a delicate process regulated by a complex network. The cross-talk between oocyte and somatic cells plays a vital role during follicle development. Granulosa cells (GCs), as one group of ovarian somatic cells, provide essential nutrients, growth factors and secret steroids for oocyte development and maturation. GCs dysfunction would initiate follicle atresia and apoptosis, and eventually lead to oocyte loss. Therefore, exploring the role of lncRNAs in GCs will provide a comprehensive understanding of POI pathogenesis.
In the present study, a down-expressed lncRNA HCP5 was identified in GCs from bPOI patients through microarray analyses. Interestingly, its location is adjacent to the MSH5 gene, a known POI-causing gene. Functional experiments further revealed a regulatory role of HCP5 in MSH5 expression via YB1 to affect the DNA damage repair (DDR) progress of granulosa cells. Our findings highlighted HCP5 as a novel transcriptional activator of MSH5, and provided a new epigenetic mechanism for human POI.
# Materials and methods
## Patients
The study was approved by the Institutional Review Board of Reproductive Medicine of Shandong University, and informed consents were obtained from all participants. Thirty women with bPOI receiving in vitro fertilization or intracytoplasmic sperm injection and embryo transfer (IVF/ICSI-ET) at Reproductive Hospital affiliated to Shandong University were recruited. Inclusion criteria for bPOI consisted of (i) basal serum FSH ≥10 IU/l; (ii) <40 years of age; (iii) with regular menstruation (23-35 days) and (iv) unilateral ovarian antral follicle counts (AFC) <5. Women with known chromosomal abnormalities, history of ovarian surgery, chemotherapy, or radiotherapy were excluded. Thirty-two women with regular menstrual cycles and normal FSH level (<10 IU/l), who sought infertility treatment due to tubal obstruction or male factors were enrolled as controls. Clinical characteristics of all participants are summarized in.
## Granulosa cells isolation
The follicular fluid of large follicles (>14 mm) was collected from each participant on the day of oocyte retrieval. After centrifugation, the supernatant was discarded. The precipitates were incubated with hyaluronidase (80 IU/ml) (Sigma, USA) for 30 min at 37 - C, and then transferred into lymphocyte separation medium (Solarbio, CHN) and subjected to horizontal centrifugation at 1600 rpm for 10 min. GCs were isolated from the interlayer phase, then washed and resuspended in phosphate buffer saline (PBS). Finally, isolated GCs were snap-frozen and stored at −80 - C until used.
# Microarray analysis
Total RNA was extracted from granulosa cells of 10 bPOI and 10 control women using TRIzol Reagent (TaKaRa, China). RNA quantity and quality were measured by NanoDrop ND-1000, and RNA integrity was assessed by standard denaturing agarose gel electrophoresis. LncRNA expression profile was detected by Arraystar Human LncRNA Microarray V3.0 (GSE135697) (Arraystar, USA). Microarray scanning and data analyses were conducted by KangChen Bio-tech (China).
## Rna isolation and rt-qpcr analysis
RNA isolation from GCs or cell lines was performed following standard protocols as described above. All mRNAs and lncRNAs were reverse transcribed using PrimeScript™ RT reagent Kit (TaKaRa, CHN) and followed by quantitative RT-PCR labeled by SYBR Green Master Mix (TaKaRa, CHN). The primers for qRT-PCR are listed in Supplementary.
## Cell culture and transfection
The human granulosa-like tumor cell line KGNwas gifted from RIKEN BioResource Center in Japan. Another human granulosa-like tumor cell line COV434 was obtained from Prof. Ying Xu of Nanjing University in China. The luteinized granulosa cell line SVOG was a gift from Prof. Peter C.K. Leung of University of British Columbia in Canada to The Chinese University of Hong Kong-Shandong University (CUHK-SDU) Joint Laboratory on Reproductive Genetics. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM)/F-12 (KGN and SVOG) (HyClone, USA) or DMEM/High Glucose (COV434) supplemented with 10% fetal bovine serum (Biological Industries, ISV), and 1% penicillin-streptomycin (Invitrogen, USA) at 37 - C in a humidified atmosphere of 5% CO 2 . Transfection of specific siRNAs for MSH5 or ILF2 and negative control (Genepharma, CHN) was performed using Lipofectamine 3000 Reagent (Invitrogen, USA) according to the manufacturer's protocol. After transfection, cells were incubated for 48 hours before further treatment. The sequences of all siRNAs are provided in Supplementary .
## Shrna knockdown
The lentivirus vector carrying HCP5-specific shRNA with puromycin resistance gene (shRNA-HCP5) and negative control (shRNA-NC) lentivirus vector were synthesized and packaged into lentivirus particles by HanBio biotechnology (CHN). For infection, lentivirus was added to KGN, COV434 and SVOG cell culture medium, and positive cells were selected by 2 mg/ml puromycin for 7 days.
## Western blot
Cells were washed twice with ice-cold PBS. After adding sodium dodecyl sulfonate (SDS) lysis buffer (Beyotime, CHN) including proteinase inhibitor cocktail (Cell Signaling Technology, USA), lysates were collected immediately .
## Cytoplasm/nucleus fractionation
The separation of cytoplasmic and nuclear RNA was carried out using the Cytoplasmic & Nuclear RNA Purification Kit (Norgen Biotek, Canada) according to the manufacturer's instructions. RNA was purified and reversetranscribed following standard protocols as described above from total and individual fractions, and then subjected to qPCR analysis. Expressions of HCP5 along with cytoplasmic and nuclear references in individual fractions were normalized to their expression levels in the input RNA, which was set as 100%. The proteins were extracted from each fraction using the PARIS™ Kit (Invitrogen, USA) according to the manufacturer's instructions.
## Fluorescence in situ hybridization
A mix of probes targeting HCP5 was synthesized and labeled with Cy3 (RiboBio, China). The experiment was performed using Fluorescent in situ Hybridization Kit (Ribo-Bio, China) according to the manufacturer's instructions and visualized under a laser confocal microscope (AN-DOR, UK). The Cy3-labeled U6 and 18S probes were hybridized simultaneously as controls.
## Rna pull-down assay
The plasmid pcDNA3.1 (Invitrogen, USA) containing sense or antisense full-length HCP5 cDNA was linearized with restriction enzyme XhoI and purified by phenol/chloroform extraction and ethanol precipitation. Biotin-labeled RNA was transcribed from linearized DNA templates in vitro using the MEGAscript™ T7 Transcription Kit (Invitrogen, USA) and purified using the MEGA-clear™ Transcription Clean-Up Kit (Invitrogen, USA) according to the manufacturer's instructions. After incubating with the KGN cell lysates, proteins co-precipitated by the biotin-labeled transcripts were isolated with streptavidin beads and subjected to SDS-PAGE analysis. Then protein bands were visualized by silver staining followed by mass spectrometry (MS) or western blot.
## Chromatin immunoprecipitation (chip)
The ChIP experiment was performed using the EZ-Magna ChIP™ A/G Kit (Millipore, USA) according to the manufacturer's instructions. For an individual ChIP assay, a total of 5 × 10 6 KGN cells were fixed in 1% formaldehyde at room temperature for 20 min and quenched by 0.125 M glycine for 5 min. Then the isolated chromatin DNA was sonicated and sheared to a length between 200 bp and 1000 bp by Bioruptor Pico sonication device (Diagenode, BEL). The sheared chromatin was immunoprecipitated at 4 - C overnight with anti-YB1 (Abcam, USA) or anti-RNA polymerase II antibody (Abcam, USA). A homologous IgG was used as the isotype control. The co-precipitated DNA was purified by QIAquick Nucleotide Removal Kit (Qiagen, Germany) and then subjected to qPCR. The details of antibodies and specific primers for ChIP-qPCR are listed in, respectively.
## Rna immunoprecipitation (rip)
The RIP experiment was performed using the EZ-Magna RIP kit (Millipore, USA) according to the manufacturer's instructions. For each RIP assay, about 1 × 10 7 KGN, COV434 or SVOG cells were lysed in RIP lysis buffer supplemented with RNase inhibitor and protease inhibitor cocktail. After one freeze-thaw cycle, cell lysates were immunoprecipitated at 4 - C overnight with anti-YB1 (Abcam, USA) or anti-ILF2 antibody (Abcam, USA). A homologous IgG was used as the isotype control. The coprecipitated RNA was extracted using TRIzol™ reagent according to the manufacturer's protocol. Then purified RNA was reverse-transcribed and subjected to qPCR analysis. Proteins isolated from the magnetic beads were detected using immunoblotting analysis. The details of antibodies are listed in .
Nucleic Acids Research, 2020, Vol. 48, No. 8 4483
## Immunofluorescence
The processed KGN cells were washed by PBS and fixed for 20 min in 4% paraformaldehyde at room temperature. After that, cells were permeabilized in 0.3% Triton X-100 and blocked by 5% bovine serum albumin in PBS for 1 h. Then samples were incubated with specific primary antibody at 4 - C overnight and Alexa Fluor 568-conjugated goat antirabbit IgG secondary antibody for 1 h at room temperature. Nuclei were counterstained with DAPI. Images were visualized under a laser confocal microscope (ANDOR, UK). The details of antibodies are listed in .
## Co-immunoprecipitation
Co-immunoprecipitation assay for YB1 and ILF2 was performed by lysing KGN cells with NP-40 lysis buffer supplemented with protease inhibitor cocktails on ice for 30 min. Then lysates were sonicated three cycles (30 s on, 30 s off) to release nuclear proteins completely. A total of 0.5 mg of lysates was incubated overnight at 4 - C with PurePro-teome™ Protein A/G Magnetic Beads (Millipore, USA) as well as anti-YB1 antibody (Abcam, USA). A rabbit IgG was used as negative control. After two times of washing using 0.05% Tween-20 in PBS, the immunoprecipitates were resuspended in loading buffer, degenerated and further analyzed by western blot.
## Statistics
Statistical analysis was implemented using SPSS 13.0 (IBM, USA) and GraphPad Prism 8 (GraphPad Software, USA). Unless otherwise noted, statistical difference was analyzed by the two-tailed Student's t-test. The Mann-Whitney Utest was used when the data were not normally distributed. The Pearson correlation and linear regression analyses were used for the expression correlation between HCP5 and MSH5. Differences were considered statistically significant when P < 0.05.
# Results
## The long noncoding rna hcp5 is down-regulated in granulosa cells from bpoi patients
To explore aberrantly expressed lncRNAs involved in POI, we firstly conducted microarray analyses of granulosa cells from 10 patients with bPOI and 10 age-matched controls. According to the criteria by P < 0.0005 and fold change > 1.75, a total of 159 lncRNAs was discovered to be differentially expressed between patients and controls, B). Among these candidates, we narrowed them down to consider only those lncRNAs which were co-localized and co-expressed with genes associated with POI. As a result, lncRNA HCP5 (RefSeq accession number NR 040662), which is from 276kb upstream of the MSH5 gene on chromosome 6, was the only POI-causing gene colocalized lncRNA identified to be differentially expressed in granulosa cells from bPOI patients compared with controls. Furthermore, quantitative RT-PCR analysis confirmed that HCP5 was differentially expressed in GCs from an independent cohort of 20 bPOI patients and 22 control women. HCP5 is associated with autoimmune diseases, infections, and cancers in human; however, nothing is known about its role in ovary.
Our results showed that HCP5 was expressed ubiquitously in multiple tissues of 21-week human fetus, relatively high in the ovary. Meanwhile, we assessed the coding potential of HCP5 using bioinformatic tools, including Coding Potential Calculator (CPC2)and Coding-Potential Assessment Tool (CPAT) (24), and confirmed the annotation of HCP5 as a long noncoding RNA (Supplementary. Then fluorescence in situ hybridization (FISH) was performed to detect the subcellular localization of HCP5. As shown by the result, HCP5 mainly located in the nucleus, and particularly around the nuclear membrane of KGN, COV434 and SVOG cells, which was further verified by cell fractionation combined with qPCR assay. Together, the decreased expression of lncRNA HCP5 was firstly identified in GCs from bPOI patients.
## Hcp5 regulates msh5 expression in granulosa cells
LncRNAs have been reported to play a cis-regulatory role in their neighboring genes. Therefore, given the adjacent localization between HCP5 and MSH5, we proposed that HCP5 may participate in the regulation of MSH5 expression. RT-qPCR results showed that MSH5 was downregulated in the GCs from bPOI women compared with controls, which was consistent with the expression of HCP5. Moreover, knockdown of HCP5 by shRNA in KGN cells resulted in an obviously decreased expression of MSH5 in both mRNA and protein levels.
Since lncRNA HCP5 could regulate MSH5 expression at transcriptional level, whether HCP5 was co-expressed with MSH5 mRNA in GCs was further examined. We measured the expression of HCP5 and MSH5 in GCs from 20 cases and 22 controls. As shown in, lncRNA HCP5 was significantly correlated with MSH5 mRNA level. Conversely, HCP5 expression was not affected by siRNA-mediated MSH5 knockdown. These results indicated a predominant role of HCP5 in modulating MSH5 expression.
## Msh5 is the target of hcp5 during dna double-strand breaks (dsbs) repair
MSH5 principally participates in mismatch repair, and the MSH4-MSH5 heterodimers play an essential role in homologous recombination (HR) repair for DSBs. To assess whether silencing HCP5 can adversely affect DNA repair process, three independent specific shRNAs were transfected into KGN cells to silence HCP5. A qPCR assay was performed to test the knockdown efficiency (Supplementary. The most efficient shRNA-HCP5#1 was used to decrease the expression of HCP5 in vitro (Supplementary.
KGN cells were stably transfected with HCP5-specific shRNA#1 and negative control, following incubation with 5 g/ml etoposide (ETO) for 1 h to induce DSBs. The ␥ H2AX foci, a sensitive marker of DSBs, were observed by immunofluorescence, and gradually disappeared along with the repair processing. Then protein levels of ␥ H2AX were detected by western blot. Results showed that knockdown of HCP5 distinctly extended the time for DSB repair induced by ETO in KGN, COV434 and SVOG cells. Given the essential role of MSH5 in HR pathway upon DSBs, we therefore examined the effect of HCP5 knockdown on the repair of camptothecin (CPT)induced DNA damage, which could be only repaired by HR but not NHEJ pathway. Briefly, KGN cells were treated with 10 M CPT for 2 h. Consistent with ETO treatment, ␥ H2AX foci were observed and faded away with the recovery as well. Also, silencing HCP5 distinctly prolonged the time for CPT-induced DSB repair in KGN, COV434 and SVOG cells. This indicated the HR repair of DSBs was delayed in HCP5-deficiency cells. Moreover, knockdown of HCP5 markedly enhanced ETOinduced apoptosis as evidenced by cleaved poly (ADPribose) polymerase 1 (PARP1).
Given the role of HCP5 in the regulation of MSH5 expression, we further explored the functional effect of MSH5 downregulation. The siRNA-mediated MSH5 knockdown was validated effective in KGN cells. Similarly, the impaired DNA damage repair and increased apoptosis after MSH5 knockdown were observed, which recapitulated the phenomena of HCP5knockdown (Supplementary. Collectively, these results indicated that HCP5 participated in DNA damage repair and apoptosis of GCs through regulating MSH5 expression.
## Hcp5 directly binds to yb1
To investigate the protein partner of HCP5, RNA pulldown assay combined with MS was conducted with biotinlabeled HCP5 full-length transcript and its antisense transcript. As shown in, a differential HCP5-binding protein band was visualized at ∼50 kDa after silver staining. The specific band was cut off, digested by trypsin, and subjected to MS. As a result, Y-box binding protein 1 (YB1), which is a member of the family of DNAand RNA-binding proteins with an evolutionarily ancient and conserved cold shock domain, was identified as. Western blot confirmed that YB1 was enriched in the HCP5 pull-down proteins. In addition, the potential binding fragments of HCP5 to YB1 were predicted by the catRAPID algorithm (29) (Supplementary.
Since RNA pull-down assay uncovered HCP5-YB1 interaction in vitro, we performed RNA-binding protein immunoprecipitation (RIP) to detect the endogenous association between HCP5 and YB1. RIP assay using anti-YB1 antibody, followed by qRT-PCR, confirmed a direct interaction between HCP5 and YB1 in KGN, COV434 and SVOG cells. Taken together, YB1 protein was a direct binding partner of lncRNA HCP5.
## Hcp5 is essential for yb1 localization to nucleus by acting as an rna scaffold for ilf2 and yb1
We then explored the regulatory role of HCP5 on its binding protein YB1. Knockdown of HCP5 did not alter YB1 expression in either mRNA or protein level, indicating that HCP5 did not affect the overall expression of YB1 in KGN cells. Considering that YB1 activity was mainly determined by its shuttling between nucleus and cytoplasm, whether HCP5 participated in the cellular localization of YB1 was further examined. We detected YB1 expression in the nuclear and cytoplasmic fraction of KGN cells after HCP5 silencing. Western blot showed that YB1 was present in both nuclear and cytoplasmic compartments of KGN cells; however, HCP5 silencing obviously reduced YB1 residing in the nucleus. Intriguingly, immunofluorescence assay confirmed the reduced nuclear staining of YB1 protein and excluded nucleolar localization of YB1 when HCP5 was knocked down, G and Supplementary. These results indicated that HCP5 silencing prevented YB1 from transporting into the nucleus.
Mechanistically, it is still unclear how HCP5 induced YB1 localization into the nucleus. It was previously proposed that activated Akt could bind to and phosphorylate cold shock domain at Ser102 of YB1 and then promoted its nuclear translocation. However, YB1 phosphorylation status was not altered after HCP5 silencing (Supplementary. Given that Interleukin Enhancer Binding Factor 2 (ILF2) could interact with YB1 and modulate its nuclear localization, a co-immunoprecipitation assay was conducted to assess the impact of HCP5 on YB1-ILF2 complex. Notably, knockdown of HCP5 decreased the relative amount of ILF2 that co-precipitated by YB1 antibody. Considering HCP5 localization in the nucleus, especially the nuclear membrane, we hypothesized that HCP5 might serve as a modular scaffold between YB1 and ILF2. Therefore, we performed RIP assay to confirm the interaction between HCP5 and ILF2. As expected, anti-ILF2 antibody co-precipitated HCP5 in KGN, COV434 and SVOG cells, whereas knockdown of HCP5 did not affect subcellular localization of ILF2 (Supplementary. In addition, we observed diminished association between YB1 and ILF2 upon RNase A treatment, which further validated the essential role of an RNA scaffold in YB1-ILF2 interaction. Finally, siRNA-mediated silencing of ILF2 resulted in a marked reduction of nuclear localization of YB1, which was in accordance with the results when HCP5 knockdown-F and Supplementary. Thus, it was concluded that HCP5 served as an RNA scaffold for the interaction between YB1 and ILF2 and could be responsible for the nuclear localization of YB1.
## Knockdown of hcp5 reduced yb1 binding to msh5 promoter and inhibited the transcriptional activation of msh5
Notably, nuclear YB1 could locate to the promoters or other key regulatory regions of target genes, where the transcription is initiated. We further explored whether HCP5 knockdown affected YB1 occupancy to the MSH5 promoter. Chromatin immunoprecipitation (ChIP) assay followed by qPCR demonstrated significant enrichment of YB1 on MSH5 promoter . Furthermore, knockdown of HCP5 significantly reduced the binding of YB1 to the promoter of MSH5 .
It was previously reported that recruitment of YB1 to the specific binding site on promoter is crucial for RNA polymerase II (Pol II) loading, and activating gene transcription. Therefore, we further investigated the influence of HCP5 knockdown on RNA Pol II binding to MSH5 promoter. As shown in , shRNA-mediated HCP5 si- . HCP5 directly binds to YB1 and is essential its nuclear localization. (A) Detection of HCP5-binding proteins by RNA pull-down assays. A specific band of ∼50 kDa (red box) was specific pulled down by HCP5 and subsequently identified as YB1 using mass spectrometry (Upper). Western blot showing the specific association between YB1 and HCP5 in the samples obtained from RNA pull-down (lower). The antisense transcript of HCP5 was used as a negative control. (B) Confirmation of the interaction between YB1 and HCP5 by RNA-binding protein immunoprecipitation (RIP) using YB1 antibody in KGN, COV434 and SVOG cells. Results are expressed as the mean ± SD (n = 3). *P < 0.05 and **P < 0.01. Two-tailed Student's t-test. (C) YB1 mRNA levels were analyzed by qRT-PCR after HCP5 silencing by shRNA in KGN cells. Values of qRT-PCR were obtained from triplicates and expressed as the mean ± SD (n = 3). Two-tailed Student's t-test. (D) YB1 protein levels were analyzed by western blot after HCP5 knockdown in KGN cells. Data shown represent three independent experiments. (E) Subcellular localization of YB1 protein was detected by western blot after HCP5 knockdown in KGN cells. Lamin B1 was used as nuclear control. GAPDH was used as cytoplasmic control. (F) Subcellular localization of YB1 protein was confirmed by immunofluorescence assay after HCP5 silencing. White triangles indicate representative YB1 expression in the nucleus. (G) Quantification of nuclear immunofluorescence intensity (mean gray value) in HCP5-knockdown (n = 10) and negative control (n = 11) KGN cells. Data are presented as the median ± interquartile range. **P < 0.01. Two-tailed Mann-Whitney U-test. lencing led to a reduced binding between RNA Pol II and MSH5 promoter, which was consistent with the inhibition of HCP5 shRNA on MSH5 mRNA level. Collectively, it could conclude that HCP5 regulated the transcription of MSH5 by facilitating the recruitment of YB1 and RNA Pol II to MSH5 promoter.
# Discussion
Before exploring noncoding RNAs responsible for POI, the identification of causative variations for POI is limited within protein-coding genes. However, the causinggene mutations could only explain 10-15% of POI cases. Except for small noncoding RNAs, studies on the roles of lncRNAs played in human POI are lacking. To our knowledge, this is the first finding of a lncRNA involved in GCs DNA damage repair process and responsible for POI.
The ovarian reserve is principally determined by the quantity and quality of oocytes. Either initially inadequate or accelerated exhaustion of primordial follicle pool will lead to ovarian insufficiency. In mammals, the role of GCs determining the fate of follicles has been well documented. Moreover, diminished GCs induced follicle atresia and finally resulted in POI in vivo. Several lncR-NAs expressed in granulosa cells have been discovered to be specifically associated with polycystic ovary syndrome, endometriosis and ovarian hyperstimulation syndrome. Herein, we explored the lncRNAs expression profile in GCs from patients with bPOI and identified downregulated lncRNA HCP5. It has been recognized that increased DNA damage and repair deficiency in granulosa cells facilitate follicle atresia and ovarian aging. Consistently, our results indicated that HCP5-knockdown disrupted the DSBs repair activity of GCs, and thereby led to GCs apoptosis in vitro.
In this study, MSH5, which has been nominated as POF13 by Online Mendelian Inheritance in Man (OMIM) was proved to be the executor of HCP5. MSH5, as a member of the MutS protein family, forms a heterodimeric com- . Knockdown of HCP5 reduced YB1 binding to MSH5 promoter region and inhibited the transcriptional activation of MSH5. (A) ChIP analyses of KGN cells were performed with IgG or YB1 antibody. ChIP products were amplified by PCR with specific primers of the MSH5 promoter region and subjected to electrophoresis analyses. (B) ChIP analyses of KGN cells were performed with IgG or YB1 antibody after HCP5 silencing, followed by qPCR in the MSH5 promoter region. Values of qPCR were obtained from triplicates and expressed as the mean ± SD (n = 3). *P < 0.05. Two-tailed Student's t-test. (C) ChIP analyses of KGN cells were performed with IgG or RNA pol II antibody after HCP5 silencing, followed by qPCR in the MSH5 promoter region. Values of qPCR were obtained from triplicates and expressed as the mean ± SD (n = 3). **P < 0.01. Two-tailed Student's t-test. (D) Schematic summary of the critical role of HCP5 in MSH5 expression and GCs function through direct binding and modulating the subcellular localization of YB1. plex with MSH4, plays an essential role during homologous recombination repair for DSBs. Causative mutations of MSH5 gene have been reported in women with POI. In the present study, the adjacent location of HCP5 to MSH5 gene indicated that it may act as a cis-regulatory factor of MSH5. Furthermore, HCP5 silencing inhibited MSH5 expression. MSH5 silencing also led to a longer time for DSBs repair and promoted ETO-induced apoptosis in GCs. Combined with the similar loss-of-function influence of HCP5 and MSH5 on GCs, we elucidated that MSH5 was the effector of HCP5 during DNA damage repair, which further enriched the understanding of MSH5 on ovarian function, not only during meiotic recombination in oocyte, but also protecting GCs from DNA damage.
It has been proved that lncRNAs exert their biological functions mainly through interacting with RNA-binding proteins. The biotin-labeled RNA pull-down/MS and RIP assay discovered that HCP5 bound with YB1. YB1 is a pleiotropic protein which participates in DNA transcription, mRNA splicing, mRNA stabilization, and mRNA translation by shuttling between the nucleus and the cytoplasm. Although HCP5 did not affect the overall expression of YB1, strikingly, the nuclear enrichment of YB1 was dramatically reduced after HCP5 silencing. In the nucleus, YB1 usually serves as a transcriptional activator. It was further demonstrated that silencing HCP5 decreased the binding of YB1 to the promoter of MSH5. Also, the co-precipitated MSH5 promoter level by RNA Pol II was reduced after silencing HCP5, which validated a protranscription regulation of HCP5 on MSH5 gene. These findings elaborated that HCP5 regulated MSH5 transcription and expression, by controlling the binding between YB1 and MSH5 promoter. Of importance, given that proteins are synthesized in cytoplasm where away from the transcription start site, HCP5 is timely suited to regulate MSH5 in situ once DNA damage occurs in GCs.
Different approaches by which YB1 was transported into the nucleus have been investigated. Interestingly, ILF2 was proved to modulate YB1 nuclear localization by directly interacting with the protein itself, which was further supported by a positive correlation between ILF2 expression and YB1 nuclear enrichment. Our results demonstrated that YB1 and ILF2 could both co-precipitate with HCP5. When HCP5 was silenced, the binding of YB1 with ILF2 showed significantly diminished. More strikingly, ILF2 knockdown resulted in obvious reduction of YB1 nuclear localization. Thus, HCP5 served as an RNA scaffold for the interaction between YB1 and ILF2 and induced YB1 shuttling to the nucleus in GCs. Combined with the interaction between YB1 and ILF2 reported previously (32), we revealed a direct connection between lncRNA HCP5 and YB1-ILF2 heterodimerization during DNA damage repair process in GCs.
In conclusion, we elucidated the mechanism of LncRNA HCP5 responsible for human POI, i.e. HCP5 regulated MSH5 expression and GCs function through direct binding with YB1 and modulating its subcellular localization . This study discovered a novel lncRNA HCP5 contributed to dysfunctional GCs by transcriptionally regulating MSH5 and DNA damage repair via YB1, providing a novel epigenetic mechanism for POI pathogenesis.
## Data availability
GEO accession ID for the microarray data is GSE135697.
## Supplementary data
Supplementary Data are available at NAR Online. |
Efficacy and safety of metformin in the management of type 2 diabetes mellitus in older adults: a systematic review for the development of recommendations to reduce potentially inappropriate prescribing
Background: Metformin is usually prescribed as first line therapy for type 2 diabetes mellitus (DM2). However, the benefits and risks of metformin may be different for older people. This systematic review examined the available evidence on the safety and efficacy of metformin in the management of DM2 in older adults. The findings were used to develop recommendations for the electronic decision support tool of the European project PRIMA-eDS. Methods: The systematic review followed a staged approach, initially searching for systematic reviews and meta-analyses first, and then individual studies when prior searches were inconclusive. The target population was older people (≥65 years old) with DM2. Studies were included if they reported safety or efficacy outcomes with metformin (alone or in combination) for the management of DM2 compared to placebo, usual or no treatment, or other antidiabetics. Using the evidence identified, recommendations were developed using GRADE methodology. Results: Fifteen studies were included (4 intervention and 11 observational studies). In ten studies at least 80% of participants were 65 years or older and 5 studies reported subgroup analyses by age. Comorbidities were reported by 9 studies, cognitive status was reported by 4 studies and functional status by 1 study. In general, metformin showed similar or better safety and efficacy than other specific or non-specific active treatments. However, these findings were mainly based on retrospective observational studies. Four recommendations were developed suggesting to discontinue the use of metformin for the management of DM2 in older adults with risk factors such as age > 80, gastrointestinal complaints during the last year and/or GFR ≤60 ml/min. Conclusions: On the evidence available, the safety and efficacy profiles of metformin appear to be better, and certainly no worse, than other treatments for the management of DM2 in older adults. However, the quality and quantity of the evidence is low, with scarce data on adverse events such as gastrointestinal complaints or renal failure. Further studies are needed to more reliably assess the benefits and risks of metformin in very old (>80), cognitively and functionally impaired older people.
# Background
Type 2 diabetes mellitus (DM2) is a prevalent chronic disease worldwide. Around 9% of adults have DM2, increasing to more than 20% of those aged 65 years or older [bib_ref] National, regional, and global trends in fasting plasma glucose and diabetes prevalence..., Danaei [/bib_ref]. DM2 and its complications are an important cause of morbidity, and people with DM2 have substantially reduced life expectancy [bib_ref] Diabetes mellitus statistics on prevalence and mortality: facts and fallacies, Zimmet [/bib_ref]. Duration of DM2 and degree of metabolic control are important factors determining the prognosis for people with DM2. However, the use of drugs for managing DM2 has been associated with preventable drug-related causes of admission to emergency units in older populations [bib_ref] Emergency hospitalizations for adverse drug events in older Americans, Budnitz [/bib_ref] [bib_ref] Percentage of patients with preventable adverse drug reactions and preventability of adverse..., Hakkarainen [/bib_ref] [bib_ref] Adverse drug reactions as cause of admission to hospital: prospective analysis of..., Pirmohamed [/bib_ref].
Metformin is one of the most widely prescribed first and second line oral glucose-lowering drugs. While it has low risk for hypoglycaemia, the risk for gastrointestinal effects is higher and it is contraindicated in patients with renal insufficiency [bib_ref] Diabetes in older adults, Kirkman [/bib_ref]. Renal function declines with age and, therefore, should be monitored closely in older adults who are prescribed metformin [bib_ref] Chronic kidney disease and automatic reporting of estimated glomerular filtration rate: new..., Johnson [/bib_ref] [bib_ref] Renal function and renal disease in the elderly: part II, Mulder [/bib_ref]. Clinical guidelines (in the United Kingdom, Canada and Australia) have advised that the use of metformin is contraindicated, or that lower doses be used, depending on renal function [bib_ref] Use of metformin in the setting of mild-tomoderate renal insufficiency, Lipska [/bib_ref]. The use of metformin has also been associated with a higher risk of lactic acidosis but this has not been widely reported [bib_ref] Metformin use and risk of lactic acidosis in people with diabetes with..., Li [/bib_ref].
Currently, there is little empirical data about patient safety and effectiveness data on to the use of oral antidiabetics including metformin among older adults. Evidencebased clinical guidelines for the treatment of DM2 have acknowledged the lack of direct evidence in older people. STOPP/START criteria version 2 considered metformin as a potentially inappropriate medication for older people with severe renal failure [bib_ref] STOPP/START criteria for potentially inappropriate prescribing in older people: version 2, O'mahony [/bib_ref]. Inappropriate prescribing may involve the prescription of a wrong dose, the lack of a clear indication or the lack of evidence-base, among others [bib_ref] Appropriate prescribing in elderly people: how well can it be measured and..., Spinewine [/bib_ref].
The objectives of this systematic review (SR) are:
to identify and collect existing literature on the risks and benefits of use of metformin in the treatment of DM2 in older adults, to assess the quality of the evidence identified, and develop recommendations when to discontinue or to adjust the dose of metformin in the treatment of DM2 in older adults.
This evidence was used to develop recommendations on discontinuation or dose adjustment of metformin in older people for the management of DM2 in order to reduce potentially inappropriate prescribing. These recommendations will be used in the electronic decision support tool of the "Polypharmacy in chronic diseases: Reduction of Inappropriate Medication and Adverse drug events in elderly populations by electronic Decision Support" (PRIMA-eDS) project [bib_ref] Appropriate prescribing in elderly people: how well can it be measured and..., Spinewine [/bib_ref].
# Methods
This systematic review was developed following the methods proposed by both the Cochrane Handbook for Systematic Reviews of Interventionsand the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [bib_ref] The PRISMA statement for reporting systematic reviews and meta-analyses of studies that..., Liberati [/bib_ref]. A full description of the methods has been published previously [bib_ref] A set of systematic reviews to help reduce inappropriate prescribing to older..., Martinez [/bib_ref].
## Study inclusion criteria types of studies
We included systematic reviews, meta-analyses, controlled interventional studies and observational studies reporting on risks and benefits of the use of metformin in the treatment of DM2 in older adults. We excluded abstracts, pooled analyses, editorials, opinion papers, case reports, case series, narrative reviews, letters, and qualitative studies.
## Types of participants
The population of interest were older people with DM2. We considered the age of 65 as cut-off point for defining older people, which has been traditionally used because of its association with retirement age in some developed countries [bib_ref] Is 65 the Best Cutoff for Defining "Older Americans?, Moon [/bib_ref]. The criteria for inclusion in this systematic review were:
For existing systematic reviews and meta-analyses:
overall mean or median age ≥ 65 years; or overall mean or median age < 65 with subgroup analysis reporting on participants ≥65 years; or overall mean or median age not reported but 80% or more of the included studies reporting a mean or median age ≥ 65 years.
For individual controlled interventional studies and observational studies:
≥80% of participants ≥65 years; or <80% of participants ≥65 years with subgroup analysis reporting on participants ≥65 years.
## Types of interventions
Studies reporting on the efficacy and/or safety of metformin for the management of DM2 were included irrespective of whether metformin was prescribed as monotherapy or in combination with any other. Included studies compared metformin versus placebo, usual or no treatment, and other drugs to treat DM2 or a non-pharmacological intervention.
## Types of outcomes
The following clinically relevant endpoints were included either as primary or secondary outcomes:
Quality of life Mortality Life expectancy Hospitalisations Cognitive impairment or cognitive status Functional impairment or status Cardiovascular event including stroke Renal failure Composite end points including any of the above (extraction of individual outcomes was undertaken if reported by original studies) Adverse drug event including hypoglycaemia Any of the above evaluated as safety endpoints.
Studies evaluating only glycaemic control or lactate levels. To aid interpretation of findings outcomes were classified into two tiers according to their anticipated impact on longer-term health and quality of life: Tier 1 outcomes have shorter-term impact including hypoglycaemia and adverse events (including serious adverse events); tier 2 outcomes have longer-term impact including, but not limited to, cardiovascular and cerebrovascular events, related admissions, and death.
## Setting
All settings were included.
## Language
Language restrictions were not applied for study searches. However, the inclusion of studies was restricted to languages that could be read by the research team English, German, Finish, Italian, and Spanish.
# Search method
Database searches were conducted by YVM and AW following staged methodology comprising four sequential literature searches. Each search being performed only if the preceding one yielded high quality results or if evidence insufficient to enable any evidence based recommendations to be made. Each search was conducted on 09 December 2015 using the OVID interface for each database. The searches included the following databases and types of studies: References of included studies were checked to identify further articles for inclusion, and we also considered studies identified from manual searches and snowballing.
Protocols for yet-to-be published studies were collected to inform future updates of this systematic review. Studies excluded after full-text check are listed in Additional file 1 together with reasons for exclusion.
The PICOS-framework was used to develop the search terms (population: older people, intervention: metformin, comparison: no limits, outcomes: see list above "Types of outcomes" and study design: systematic reviews, metaanalyses, controlled interventional studies and observational studies). We also developed search filters specific for different study designs, described in detail in the protocol [bib_ref] A set of systematic reviews to help reduce inappropriate prescribing to older..., Martinez [/bib_ref]. Additional file 2 lists the full search terms for each search (i.e. search 1, 2 and 3B).
## Data management
Search results were uploaded to Endnote X7 reference management software where search results were retrieved and de-duplicated.
## Selection of studies
Two reviewers assessed titles and abstracts from each search independently to identify studies to consider for inclusion. Full manuscripts were then obtained for all titles and abstracts meeting the inclusion criteria or where there was any uncertainty about inclusion. YVM, ARG, CA, BF, CS and LS were involved in this task.
Reviewers discussed any disagreement about studies to include. AS was consulted when YVM and ARG could not reach an agreement on whether or not to include a study. YVM and ARG were consulted when CA, CS, BF and LS could not reach agreement.
## Data extraction
YVM, ARG, CA, BF, CS and LS (reviewers) independently conducted data extraction of the included studies using a standardised and piloted data collection form previously published with the protocol [bib_ref] A set of systematic reviews to help reduce inappropriate prescribing to older..., Martinez [/bib_ref]. This extraction form included information related to the study design and aim, characteristics of participants (i.e. age, sex, setting, comorbidity, use of concomitant medications, functional status, and cognitive status), the intervention (i.e. metformin) and comparison, time to follow-up, and reported outcomes. Completeness and accuracy of data extraction was then double-checked by a second reviewer.
## Quality appraisal
For each study design we used separate validated assessment tools to evaluate quality (AMSTAR) [bib_ref] Development of AMSTAR: a measurement tool to assess the methodological quality of..., Shea [/bib_ref] [bib_ref] AMSTAR is a reliable and valid measurement tool to assess the methodological..., Shea [/bib_ref] was used for systematic reviews/meta-analyses, for intervention studies the Cochrane Collaboration's tool for assessing risk of biaswas used, and for observational studies the Critical Appraisal Skills Programme (CASP).
## Dealing with duplicate and companion publications
All relevant data from publications relating to a single primary study were included. The staged approach carries a risk of 'double counting patients whose trials are included in a systematic review. Any such instances have been identified, reported and corrected for in our data synthesis.
## Data synthesis
A narrative synthesis describing all included systematic reviews, meta-analyses, intervention and observational studies, participants and findings was carried out. The included studies were highly heterogeneous in comparison treatments, length of follow-up, type of design, and definition of outcomes; therefore no additional metaanalyses were performed. The quality of the included studies is also reported.
## Identification of "references of interest" for the development of recommendations
During the search process, reviewers identified additional references which did not fulfil the inclusion criteria of the SRs but which they considered of interest for the development of recommendations according to the methodology described by Martinez-Renom Guiteras (2016) [bib_ref] A set of systematic reviews to help reduce inappropriate prescribing to older..., Martinez [/bib_ref].
## Development of recommendations
Included studies and references of interest were summarised in a document used by the research team to develop and discuss recommendations to discontinue the use of metformin for the management of DM2 in older people including: a) study design or type of reference, target population and sample size, metformin dose (if available) and comparison groups, outcomes, main results, subgroup analysis if applicable; b) quality appraisal ratings of included studies; and c) proposed recommendations. Each recommendation was given a strength (weak or strong) and quality (low, moderate or high) rating following the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology [bib_ref] Going from evidence to recommendations, Guyatt [/bib_ref] [bib_ref] What is "quality of evidence" and why is it important to clinicians?, Guyatt [/bib_ref] [bib_ref] GRADE: an emerging consensus on rating quality of evidence and strength of..., Guyatt [/bib_ref]. Recommendations were written following a standardised schema and reflecting the strength and the quality of the evidence. The Finnish team of editors from Duodecim Medical Publications Ltd. participated in the later stages and approved the recommendations.
# Results
Results of the search Searches 1, 2, 3A and 3B were all conducted. Search 1 identified one relevant meta-analysis, by Lamanna et al. [bib_ref] Effect of metformin on cardiovascular events and mortality: a meta-analysis of randomized..., Lamanna [/bib_ref] which did not provide summary results for our SR targeted at old age and was excluded. However eligible individual studies were identified from it and added to search 3A. The date of the search by Lamanna (2011) [bib_ref] Effect of metformin on cardiovascular events and mortality: a meta-analysis of randomized..., Lamanna [/bib_ref] was used as the start date for our search for additional individual studies under search 3B. No relevant meta-analyses were identified from search 2.
In total 2185 records were found through initial database searching (126 from search 1, 175 from search 2, 1884 from search 3B). Additionally, we identified 66 records from search 3A (individual studies from excluded systematic reviews and meta-analyses) and 461 records from reference lists of included studies. After removing duplicates, we screened 2318 records and excluded 1878 checking titles and abstracts. We assessed 440 full-texts for eligibility and excluded 425. Main reasons for exclusion were wrong population, wrong intervention and wrong outcome. We included 15 studies reported by one publication each. The PRISMA flow diagram is presented in [fig_ref] Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses [/fig_ref].
## Included studies
Fifteen studies were included [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref] [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref] [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref] including 426,549 participants of all ages of which 230,229 were 65 years and older. However, analyses of efficacy and risks of antidiabetic drugs did not always include all participants. [fig_ref] Table 1: Summary of study characteristics Summary of study characteristics [/fig_ref] shows the summary of the study characteristics of included studies.
## Study designs
Only four studies were randomised controlled trials (RCTs) [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref] [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. Most of the included studies were observational (11 studies), three of these prospective [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] , seven retrospective [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref] , and one cross-sectional [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref]. Data on outcomes was extracted for the end of the follow-up period of each included study.
## Participants
In ten studies at least 80% of participants were aged 65 years or older . In these studies, the lowest mean age was 69.3 years and the highest 80.5 years. The remaining five studies provided subgroup analyses of older people (≥65 or ≥71 or >80 years) [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref]. Length of follow-up varied from none (cross-sectional study) [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref] to 11 years [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref].
Representation of males ranged from 24.1% [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref] to 97% [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref]. Eight studies reported ethnicity with most participants being white/Caucasian (up to 81%) [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref]. Participants from four different continents were involved in the studies including America (n = 8), Europe (n = 7), Asia (n = 3) and Australia (n = 2). Comorbidities were reported in 11 studies, and congestive heart failure, myocardial infarction, hypertension, chronic kidney disease and hypothyroidism were the most commonly reported. Functional status was reported in one study [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref]. No studies reported on the frailty level of the participants. Cognitive status was reported in four studies [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref]. The percentage of participants with dementia was low in most studies ranging from 2.8% [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] to 67% [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref]. Participant characteristics are summarised in Additional file 3: [fig_ref] Table 1: Summary of study characteristics Summary of study characteristics [/fig_ref].
## Covariates in models
Adjustment for relevant covariates is important in observational studies to reduce confounder bias. The use of covariates indicates that a study's authors have considered this issue, although by no means implies that bias has been eliminated. Many of the studies used long lists of covariates, therefore Additional file 3: [fig_ref] Table 2: Summary of study findings [/fig_ref] is given for online publication.
## Interventions and outcomes
Most of the included studies investigated the effect of metformin as monotherapy . We classified comparison treatments as either "non-specific active treatments" (the comparator was not a single specific drug or treatment e.g. usual care, no insulin sensitizer, not on metformin), or as "specific active treatments" (the comparator was a specified treatment such as insulin, sulfonylureas, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, and other specific drugs and combinations).
One RCT addressed both tier 1 and tier 2 outcomes [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] ; the remaining RCTs investigated tier 1 outcomes only [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref] [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. All the included observational studies addressed tier 2 outcomes only . We did not find any studies reporting on the following The sum of patients in each subgroup is greater than the number of participating patients, as the study design allowed patients to take different drugs at different time periods c SAEs comprised any experience that was fatal, life-threatening, permanently or substantially disabling, resulted in permanent or significant disability or incapacity, required or prolonged hospitalization, an important medical event that jeopardized the patient or required intervention to prevent a serious outcome, a congenital abnormality, a cancer, an overdose of medication, or a condition that resulted in the development of drug dependency or drug abuse relevant endpoints: quality of life, life expectancy, functional impairment or status, and renal failure. [fig_ref] Table 2: Summary of study findings [/fig_ref] summarises the results for each study for both metformin and comparison groups, with estimated risk ratios with 95% confidence intervals, together with statistical comparisons from the study. The results are organised by type of outcome (safety or efficacy), and then by type of comparator within each type of outcome (metformin against non-specific active treatments; metformin against other specific active treatments). A Additional file 3: [fig_ref] Table 2: Summary of study findings [/fig_ref] is available with the covariates that were taken into account in the statistical models of the included studies.
## Main findings
## Tier 1 outcomes (hypoglycaemia and adverse events)
Tier 1 outcomes were investigated by all the included trials but none of the observational studies. In one large trial metformin was not significant different than usual care (non-specific active treatment) for serious adverse events [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref]. Compared against other specific active treatments, a combination of insulin glargine plus glimepiride and metformin demonstrated significantly fewer hypoglycaemic events (both confirmed and unconfirmed hypoglycaemia) compared to premixed insulin [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref]. However, in other trials, participants taking tolbutamide reported fewer side effects than participants on metformin [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] and vildagliptin outperformed metformin on all safety outcomes reported [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref] , but in neither case was any formal statistical comparison reported.
Tier 2 outcomes: Metformin compared to other non-specific active treatments
One trial and five observational studies compared metformin as monotherapy with non-specific active treatments for efficacy-related outcomes [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref]. Three large observational studies reported significantly fewer deaths in participants taking metformin compared to participants taking either no insulin sensitizer [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] , no antidiabetic drugs [bib_ref] Treatment of type 2 diabetes and outcomes in patients with heart failure:..., Macdonald [/bib_ref] , or no metformin [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref]. However, there was no significant difference in mortality for patients aged over 80 years, and so was it for those patients with GFR ≤60 [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref]. [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] in another large study also reported that there was no significant difference in mortality in the metformin group. Admissions for various types of causes were evaluated by three studies [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] without significant differences between metformin and other active treatments, except for re-admission for heart failure which was significantly different favouring metformin [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref].
## Tier 2 outcomes: metformin compared to other specific active treatments
Six observational studies compared metformin with specific active treatments on their effect on efficacy outcomes [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref]. One of these studies [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] reported substantially reduced mortality in the metformin monotherapy group (16% of the 422 participants) compared to groups taking metformin and sulfonylureas in combination (32%) and sulfonylurea monotherapy (51%). The remaining studies compared monotherapy with metformin against a range of other mostly monotherapy drug treatments. Metformin outperformed sulfonylureas with significantly fewer hospital admissions for acute myocardial infarction, stroke or death [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] ; fewer hospitalisations for hypoglycaemia [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] ; fewer events of non-fatal cardiovascular disease [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] ; fewer fractures [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref] and fewer events of myocardial infarction, congestive heart failure and allcause mortality [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref]. Other study results favoured metformin with significant differences in comparisons against thiazolidinediones [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] (mortality and all-cause and HF readmissions). There were no significant differences when metformin was compared to rosiglitazone and pioglitazone [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref] (myocardial infarction, congestive heart failure and all-cause mortality).
## Excluded studies
Additional file 1 has the full list of reasons for exclusion of studies after full text analysis.
## Quality appraisal of included studies randomised trials
Four randomised trials were included and assessed for risk of bias [fig_ref] Table 3: Quality appraisal for intervention studies LR low risk of bias [/fig_ref]. Two of these trials did not provide enough information to assess the risk of bias [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. Another trial was judged to be of high risk of selection, performance and other bias [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref]. One trial was judged to be of high risk of performance and detection bias [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref].
## Observational studies
Quality appraisal was assessed with the CASP tool for the included 11 observational studies [fig_ref] Table 4: Quality appraisal for observational studies according to the Critical Appraisal Skills Programme [/fig_ref]. Most of the included studies reported sufficient detail to assess their quality. All studies addressed a clearly focused issue and all but one [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] used an appropriate method to answer their research question. Selection bias was not a problem in 8 of the 11 included studies as the recruitment method was adequate for the design (either cohort or case-control study) . Potential confounding factors were taken into account in the design or analysis in 10 of the 11 included studies . It was clear that the results of all studies (apart from the cross-sectional study) could be applied to our population of interest (older people). Nearly half the studies did not report accurately how the exposure or outcome were measured, which could lead to high risk of measurement or classification bias [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref]. For example, the comparison group was not clearly defined in one study [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref]. It was unclear whether the follow-up was sufficiently long in 5 studies [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] and it was considered that the follow-up insufficient in 3 studies [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] A cohort study of thiazolidinediones and fractures in older adults with diabetes, Solomon [/bib_ref].
## Additional references of interest for the development of recommendations
One further reference was incorporated as additional reference of interest for the development of the recommendations. A clinical guideline from the American Geriatrics Society which included recommendations about the management of DM2 in older people with renal insufficiencywas included as an additional reference of interest for the development of recommendations.
## Recommendations
Four recommendations about stopping the use of metformin in older people with DM2 [fig_ref] Table 5: Recommendations to stop the use of metformin in older people with type... [/fig_ref] were developed related to halting In order to discuss and agree on these recommendations three meetings took place between YVM (researcher) and ARG (researcher and clinician). IK (senior clinician and researcher) and AS (senior clinician and researcher) participated in one of these meetings. The whole body of evidence identified in the SR was taken into consideration for the development of the recommendations. However, each recommendation was specially supported by the following specific studies included in the SR or considered as additional references of interest: a clinical guideline, an observational study without high quality [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] and two randomised trials with insufficient information to assess their risk of bias [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. All recommendations were considered to be weak and based on evidence of low quality, and the reasons for this are reported in [fig_ref] Table 5: Recommendations to stop the use of metformin in older people with type... [/fig_ref].
The recommendations were included in the Comprehensive Medication Review (CMR) tool developed as part of the PRIMA-eDS project, and they were formulated according to their strength and the quality of their evidence [bib_ref] A set of systematic reviews to help reduce inappropriate prescribing to older..., Martinez [/bib_ref].
# Discussion
Our aim was to systematically review the existing evidence on the risks and benefits of the use of metformin for the management of DM2 in older people. We therefore included only those studies where a high proportion of participants were aged 65 years or older, as specified in our inclusion criteria. No systematic review or metaanalysis fulfilled our inclusion criteria, and we finally included 4 RCTs and 11 observational studies, with most observational studies being retrospective. When comparing metformin with sulfonylureas, results suggest that metformin may be better than sulfonylureas in reducing several outcomes such as cardiovascular outcomes, mortality, hospitalisation for hypoglycaemia, or risk of falls in people aged 65 and older with DM2 [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Oral hypoglycaemic agents and the development of non-fatal cardiovascular events in patients..., Hung [/bib_ref] [bib_ref] Sulfonylureas and risk of falls and fractures among nursing home residents with..., Lapane [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref]. When comparing metformin with no insulin sensitizer antidiabetic drugs, divergent results were found depending on the study population [bib_ref] Insulin-sensitizing antihyperglycemic drugs and mortality after acute myocardial infarction: insights from the..., Inzucchi [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] Generally, these results are in line with clinical guidelines recommending metformin as the first-line drug treatment for adults with DM2. Guidelines also suggest that if initial drug treatment with metformin fails to control levels of glycated haemoglobin, dual therapy should be considered. Only one study was identified which specifically analysed risks and benefits of combining metformin with other antidiabetic drugs, where risk of hypoglycaemia with the combination of metformin, sulfonylurea and insulin glargine compared to premixed insulin was decreased [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref]. Thus, there seems to be a lack of evidence analysing Prospective, observational
[formula] Y U Y N A Y N N Y U U Y Y Y Inzucchi 2005 [34] Retrospective cohort study Y Y Y N A N Y Y Y Y N U Y U Lapane 2015 [37] Retrospective cohort study Y Y Y N A Y N Y Y U N Y Y Y MacDonald 2010 [38] Case-control study Y Y Y Y Y Y Y Y Y Y Y YYregistry Y Y N N A U Y U Y Y U Y Y Y Solomon 2009 [44] Retrospective cohort Y Y Y N A Y Y Y Y Y N U Y Y Tzoulaki 2009 [45] Retrospective cohort study Y Y Y N A U Y Y Y Y Y U Y Y Y yes; [/formula]
N no; U unclear; NA not applicable. Items 11 and 12 are part of the findings in [fig_ref] Table 2: Summary of study findings [/fig_ref] the benefits and risks of combined therapy including metformin in older people with DM2. Few studies provided data on adverse events other than hypoglycaemia and falls with the use of metformin in this population such as bloatedness, nausea, and diarrhoea [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref] [bib_ref] Increased risk of cognitive impairment in patients with diabetes is associated with..., Moore [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. Furthermore, only two studies analysed the adverse event of lactic acidosis, and no increased risk was found for metformin [bib_ref] Comparative outcomes study of metformin intervention versus conventional approach the COSMIC approach..., Cryer [/bib_ref] [bib_ref] Thiazolidinediones, metformin, and outcomes in older patients with diabetes and heart failure:..., Masoudi [/bib_ref]. All these studies had considerable methodological limitations. We did not identify any study reporting on renal failure as adverse event of metformin. Thus, further prospective studies should evaluate the adverse events of the use of metformin in older people.
The benefits of metformin on the mortality of very old people (aged 80 years over) were investigated by only one included study [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref]. Here, mortality was significantly decreased with the use of metformin in people aged 65-80 but the effect was not significant for the population aged 80 and older. The study could been underpowered for this subgroup analysis, but evidence this may also suggest that the benefits of metformin on mortality may be non-existent for very old people, especially those with limited life expectancy, as suggested by other authors [bib_ref] Recommendations to prescribe in complex older adults: results of the CRIteria to..., Onder [/bib_ref].
The included studies rarely reported on the functional level and cognitive status of the participants; the use of concomitant drugs and the presence of other diseases were more frequently reported but focused mostly on cardiovascular drugs and diseases. Thus, the present systematic review demonstrates that not only very old people, but also cognitively and functionally impaired people and old people with multimorbidity are underrepresented or at least underreported in existing studies, which limits the generalisability of already scarce evidence for this heterogeneous group of older people. A growing body of literature presents functional and cognitive status as well as multimorbidity as predictors of mortality among older people independently from their chronological age [bib_ref] Comprehensive geriatric assessment predicts mortality and adverse outcomes in hospitalized older adults, Avelino-Silva [/bib_ref] [bib_ref] Prognostic factors of long-term survival in geriatric inpatients. Should we change the..., Bien [/bib_ref] [bib_ref] The impact of chronic multimorbidity and disability on functional decline and survival..., Marengoni [/bib_ref] , which supports the idea that further studies analysing these aspects are necessary.
## Weak low
Observational study [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] ; RCTs [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref] ; clinical guidelineReason: uncertainty about the magnitude of the benefits and harms.
It was considered to downgrade the quality of the evidence to low quality because there were study limitations (1 observational study with limitations and 2 RCTs with unclear risk of bias), indirectness (observational study with subgroup analysis), inconsistency (different types of comparisons evaluated).
It is suggested to discontinue metformin for the management of type 2 diabetes mellitus in patients 80 years and older taking the life expectancy, physical and functional status of the patient into account. Patients who are concerned about adverse events or appear to experience AE may reasonably choose not to take metformin.
## Weak
Low Observational study [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] Reason: uncertainty about the magnitude of the benefits and harms.
It was considered to keep the quality of the evidence as low quality because this observational study had limitations: data in older people was from subgroup analysis, lack of reporting on recruitment and confounding factors.
It is suggested to discontinue metformin for the management of type 2 diabetes mellitus in patients with gastrointestinal complaints taking the possible benefit and the severity of the patient complaints as possible dverse drug reactions into account.
Weak Low RCTs [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref] Reason: small RCTs with low quality and no significant benefits with metformin; uncertainty about the magnitude of the benefits and harms.
It was considered to downgrade the quality of the evidence to low quality because there were study limitations (2 RCTs with unclear risk of bias) and inconsistency (different types of comparisons evaluated).
It is suggested to discontinue metformin for the management of type 2 diabetes mellitus in patients with renal insufficiency because metformin may increase the risk of lactic acidosis.
## Weak low
Clinical guidelineReason: evidence from a clinical guideline; uncertainty about the magnitude of the benefits and harms.
It was considered to keep the quality of the evidence as low quality because it was from a clinical guideline.
Although the lack of evidence has been previously commented on by several authors [bib_ref] Recommendations to prescribe in complex older adults: results of the CRIteria to..., Onder [/bib_ref] [bib_ref] Anonymous. ASCOT -blood pressure arm confirms efficacy of amlodipine/perindopril (Norvasc)/Coversyl) regimen, Huelgas [/bib_ref] , to the best of our knowledge, ours is the first study to systematically review all available evidence on the risks and benefits of metformin for the management of DM2 in older people.
Current guidelines suggest that for older people with limited life expectancy or functional limitation, intensive glycaemic control is not recommended [bib_ref] Recommendations to prescribe in complex older adults: results of the CRIteria to..., Onder [/bib_ref]. However, older people who are functionally and cognitively intact and have significant life expectancy should be treated with goals similar to those developed for younger people. In the present systematic review, six studies [bib_ref] Effect of Metformin on mortality in patients with heart failure and type..., Evans [/bib_ref] [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref] [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref] [bib_ref] Risk of cardiovascular disease and all cause mortality among patients with type..., Tzoulaki [/bib_ref] reported on mean baseline levels of glycated haemoglobin which ranged between 7.0% [bib_ref] Comparative effectiveness of sulfonylurea and metformin monotherapy on cardiovascular events in type..., Roumie [/bib_ref] and 10.2% [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref]. One of these studies considered a glycated haemoglobin level of 7% or less without experiencing nocturnal hypoglycaemia to represent successful therapy [bib_ref] Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone..., Janka [/bib_ref].
None of the other studies reported their target glycaemic control level clearly. It would be useful for future studies on the management of DM2 to report the target glycaemic control level, especially for frail older people.
Our research team developed four recommendations using both the results of the systematic review and the additional references identified [bib_ref] Comparison of tolbutamide and metformin in elderly diabetic patients, Josephkutty [/bib_ref] [bib_ref] Metformin use and mortality among patients with diabetes and atherothrombosis, Roussel [/bib_ref] [bib_ref] Comparison of vildagliptin and metformin monotherapy in elderly patients with type 2..., Schweizer [/bib_ref]. The recommendations advise clinicians to consider discontinuing metformin in people aged 80 and older, those with gastrointestinal complaints during the last year, and/or those with Glomerular Filtration Rate (GFR) ≤60 ml/min. These recommendations have been incorporated in the trial version of an electronic decision support tool that aims to help general practitioners to reduce inappropriate prescriptions for older people with multimorbidity. Decisions on prescription or de-prescription of metformin should be made taking the symptoms and individual characteristics of each patient into account and clinicians receive instructions on that. Currently, the tool is being tested in a cluster randomised controlled trial in four European countries [bib_ref] Polypharmacy in chronic diseases-reduction of inappropriate medication and adverse drug events in..., Sönnichsen [/bib_ref].
This systematic review has limitations. The search strategy and inclusion criteria were designed to identify studies focusing on older people; studies on the general population that may have contained relevant information for the older population might have been overlooked. However, using independent reviewers for study selection and our peer reviewed process of development of recommendations should have minimised this problem. Our recommendations focus only on the discontinuation of metformin, as it was not the aim of the PRIMA-eDS project to develop recommendations when to use metformin. Nevertheless, this systematic review aims at providing an overview of the existing evidence on both the benefits and risks of the use of metformin in older people.
# Conclusions
This study highlights the lack of good quality evidence on the risks and benefits of metformin for the management of DM2 in older people. The use of metformin seems associated with benefits to lower mortality risk in older people, and may also be associated with a reduced risk of adverse events such as hypoglycaemia and nonfatal cardiovascular events, than other antidiabetic drugs, especially sulfonylureas. However, no prospective studies focussing on very old (80 and older) and functionally and cognitively impaired older people are available. In very old people, those with renal insufficiency (GFR ≤60 ml/min) and those with gastrointestinal complaints during the last year, the discontinuation of metformin should be considered, especially for those with limited life expectancy or functional impairment. There is an urgent need for studies on the risks and benefits of metformin for the management of DM2 in these populations in order to guide clinicians in planning of individualised patient care.
[fig] Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram [/fig]
[table] Table 2: Summary of study findings [/table]
[table] Table 3: Quality appraisal for intervention studies LR low risk of bias; HR high risk of bias; UR unclear risk of bias [/table]
[table] Table 4: Quality appraisal for observational studies according to the Critical Appraisal Skills Programme (CASP) [/table]
[table] Table 5: Recommendations to stop the use of metformin in older people with type 2 Diabetes mellitus [/table]
|
Aerobic Exercise Effects on Ocular Dominance Plasticity with a Phase Combination Task in Human Adults
Supplemental Information Sample size calculations
Supplemental Information Sample size calculations
To be able to compare Lunghi & Sale's (Lunghi and Sale 2015) approach with ours (Binocular rivalry and Binocular combination) we have now provided in each of our results figures an analysis of our results in terms of area (phase x time) so it can be compared with Lunghi & Sale's (eye rivalry x time).
The ratios between the area of the Physical activity and the inactive control in our two experiments are: 1. moderate/rest: 0.981 ± 0.856 (mean ± SD), which was not significantly different with 1: t(9)=-0.07, p = 0.95 (2-tailed one sample t-test); 2. hard/rest: 1.383 ± 1.144 (mean ± SD), which was not significantly different with 1: t(9)=1.06, p = 0.32 (2-tailed one sample t-test).
The performance increase in Lunghi & Sale's cycling paper (Lunghi and Sale 2015) could be calculated from the relative difference between the area under the fitted curves (AUC) in the cycling and resting conditions in their figure 1B (equation (S1), parameters a and b of the fits and the baseline ratio could be found in their supplementary):
[formula] AUC = ∑ t 1+ ( a ( log(t +1)) ) b (S1) [/formula]
Then the relative difference between their cycling and resting state is 2.1784 (between 1 minute and 45 minute).
To calculate the sample size for matched samples we would need to obtain the same margin of error E=2.1784, we use the equation (S2):
[formula] n= ( ( Z β + Z α / 2 ) σ d E ) 2(S2) [/formula]
where Z β = 0.84 corresponds to the selected 80% power, Z α /2 = 1.96 corresponds to the 2-tailed significance level at α =0.05 and σd is the standard deviation of the matched relative differences between our cycling and resting conditions. So, with σd = 0.856 in the moderate condition and σd = 1.144 in the hard condition, we obtain a desired sample size n = 1.21 and n = 2.16 respectively for the moderate and hard cycling condition. |
Diffusion weighted imaging and blood oxygen level-dependent MR imaging of kidneys in patients with lupus nephritis
Background: Lupus nephritis (LN) is one of most common secondary glomerulonephritis. There is no ideal method to simultaneously assess renal structure and function in patients with LN. The aim of this study is to investigate the utility of diffusion weighted imaging (DWI) and blood oxygen level-dependent (BOLD) MR imaging in the assessment of renal involvement and pathological changes in patients with LN.Methods: Sixty-five patients with LN and 16 healthy volunteers underwent coronal echo-planar DWI and BOLD MR imaging of the kidneys. The apparent diffusion coefficient (ADC) and R2* values of the kidneys were calculated with b values of 0 and 500 s/mm 2 . The relationship between the renal injury variables and the ADCs or R2* values were evaluated. And 16 of 65 patients with LN underwent a repeated evaluation after the induction treatment for 9 to 12 months.Results:The mean ADC values of kidneys in patients with LN were 2.40 ± 0.25 × 10 −3 mm 2 / s, the mean R2* values of the renal cortex and medulla were 11.03 ± 1.60/sec and 14.05 ± 3.38/sec respectively, which were all significantly lower than that in volunteers. In patients with LN, the mean ADC values were correlated with eGFR (r = 0.510, p < 0.01). There was a negative correlation between the mean ADC values and renal pathology chronicity indexes (r = −0.249, p < 0.05), the R2* values of the renal medulla and proteinuria (r = −0.244, p < 0.05), and the degree of tubulointerstitial lesions (r = −0.242, p < 0.05). The ADC and R2* values of kidneys were significantly higher than those of pre-treatment in complete remission patients.Conclusions: DWI and BOLD MR imaging of kidneys may be used to noninvasively monitor the disease activity and evaluate therapeutic efficacy in lupus nephritis.
# Background
Systemic lupus erythematosus (SLE) is an autoimmune disease with variable presentations, course and prognosis. Approximately half a million people in Europe and seventy per 100 000 people in China suffer from SLE. Renal involvement is one of the major determinants of the outcome in patients with SLE. Lupus nephritis (LN) is one of the most common secondary glomerulonephritis in China. Renal interstitial infiltrates, tubular necrosis and interstitial fibrosis can be found in 60-70% patients with LN. Interstitial lesions are often associated with glomerular function, and may play a pivotal role on progression and prognosis of lupus nephritis [bib_ref] Tubulointerstitial lesions of patients with lupus nephritis classified by the 2003 International..., Yu [/bib_ref] [bib_ref] Mycophenolate mofetil, tacrolimus and intravenous cyclophosphamide for the induction treatment of active..., Li [/bib_ref] [bib_ref] Lupus nephritis: correlation of interstitial cells with glomerular function, Alexopoulos [/bib_ref]. Nowadays there is no ideal method to simultaneously assess renal structure and function in patients with LN. Moreover, effective means to evaluate tubulointerstitial hypoxia state is lack as well. Therefore, it's difficult to dynamically monitor the disease progression in these patients.
Functional magnetic resonance (MR) imaging techniques such as diffusion-weighted imaging (DWI) and blood oxygen level-dependent (BOLD) imaging have shown considerable value in the evaluation of renal function in health and renal diseases [bib_ref] Renal Functional MRI: are we ready for clinical application?, Chandarana [/bib_ref] [bib_ref] Imaging techniques in the management of chronic kidney disease: current developments and..., Herget-Rosenthal [/bib_ref] [bib_ref] Renal relevant radiology: renal functional magnetic resonance imaging, Ebrahimi [/bib_ref]. The aim of this study is to investigate the feasibility of functional renal MR imaging in the assessment of renal functional and structural changes in patients with lupus nephritis.
# Methods
## Study population
Sixty-five patients with lupus nephritis, diagnosed according to clinical manifestation, laboratory findings and renal histopathological changes, had been followed up since their admission in our department during the recent two years. They all met the diagnosis criteria of systemic lupus erythematosus defined by the American Rheumatism Association in 1997. All the patients were included after signing informed consent. There were 10 male and 55 female patients, with mean age of 34 ± 12 years old. Sixteen healthy volunteers, 2 men and 14 women with mean age of 35 ± 11 years old, had no history of primary or secondary renal diseases.
## Mr imaging
All the LN patients and healthy volunteers underwent coronal echo-planar DW and BOLD MR imaging of the kidneys with a single breath-hold time. And 16 of the 65 patients with LN received functional MR imaging of kidneys again after 9 to 12 months induction treatment with immunosuppression. All of them fasted for 4-8 hours before the examination. MR imaging was performed with a 1.5 T MR imager (GE Twin-Speed, General Electric Medical Systems, Milwaukee, WI). DW MR imaging was performed with a spin echo-echo planar imaging (SE-EPI) sequence, and BOLD MRI was performed with a multiple gradient-recalled-echo (mGRE) sequence. The following parameters were used for DWI: 11 sections (section thickness, 5 mm; intersection gap, 1 mm); repetition time (TR), 2000 ms; echo time (TE), 52.9 ms; NEX, 2; field of view, 380 × 380 mm; matrix, 128 × 128; acquisition time, 16 s. The diffusion gradient b values of 0 and 500 s/mm 2 were used. The following parameters were used for BOLD MRI: TR/TE, 110 ms/2.2-27.5 ms; flip angle, 60°; matrix, 132x128; section thickness, 5 mm; intersection gap, 1 mm. Three coronal sections were obtained for each kidney and eight T2*-weighted images were acquired for each section within one 16-second breath hold.
# Imaging analysis
The apparent diffusion coefficient (ADC) and R2* values of the kidneys were calculated on a GE workstation (Sun Microsystems, ADW4.2) with Functool 2 image analysis software. In the transverse and coronal ADC map, freehand regions of interest (ROIs) were placed on the parenchyma of the kidneys, and fat within renal sinus were excluded. Three such ROIs were placed-one each in the upper pole, inter-polar region, and lower pole-and the mean of these three values were calculated.
R2* values of renal cortex and medulla were measured on T2* maps. The size of ROI was set to reduce the noise and the influence of partial volume effects. Avoiding visible vessels, ROIs were placed manually on the clear corticomedullary differentiation image. Three ROIs were placed on each side of renal cortex and medulla, and the mean of R2* values were calculated.
The relationship between the renal injury variables and the ADCs or R2* values were evaluated. The renal function was estimated using the abbreviated modification of diet in renal disease (MDRD) study equation. Renal tubulointerstitial lesions were scored with semi quantitative method: no lesion was 0, mild lesion (≤25%) was 1, moderate lesion (25%~50%) was 2, and severe lesion (≥50%) was 3. There were 0-9 scores for tubulointerstitial lesions including 0-3 for tubular atrophy, 0-3 for interstitial fibrosis, 0-3 for interstitial inflammatory cells infiltration. The renal pathology activity index (AI) and chronicity index (CI) were evaluated in patients with LN.
# Statistical analysis
The SPSS software was used (version 13.0; SPSS; Chicago, Illinois, USA). All results were expressed as mean ± SD for normally distributed data, median and range for skewed data and frequency (%) for categorical data. The distribution of clinical and laboratory attributes among the groups were evaluated by t test. The Pearson correlation analysis was used to evaluate the relation between the clinical or pathological variables and the ADCs or R2* values. P < 0.05 was considered as statistically significant.
# Results
The mean ADC and R2* values of kidneys in patients with LN were lower than that in healthy volunteers
The mean ADC values of kidneys in healthy volunteers were 2.52 ± 0.17 × 10 −3 mm 2 / s, the mean R2* values of the renal cortex and medulla were 12.63 ± 1.40/sec and 18.14 ± 2.51/sec respectively. The mean ADC values of kidneys in patients with LN were 2.40 ± 0.25 × 10 −3 mm 2 / s, the mean R2* values of the renal cortex and medulla were 11.03 ± 1.60/sec and 14.05 ± 3.38/sec respectively, which were all significantly lower than the ADC and R2* values of kidneys in healthy volunteers (p = 0.048, p = 0.045 and p = 0.008, respectively) [fig_ref] Table 1: The mean ADC and R2* values of kidneys in patients with LN... [/fig_ref]. There was no significant difference of the ADC and R2* values between bilateral kidneys either in volunteers or in patients with LN. Moreover, the ADC and R2* values were much lower in patients with severe lupus nephritis [fig_ref] Figure 1: Renal functional MR imaging of Patient A with LN-IV [/fig_ref].
## The correlation between imaging values and renal function or pathological changes in patients with ln
In the patients with LN, the mean ADC values were correlated with glomerular filtration rate estimated by abbreviated MDRD equation (eGFR) (r = 0.510, p < 0.01) . There was a negative correlation between the mean ADC values and pathology chronicity indexes (r = −0.249, p < 0.05). Moreover, the R2* values of the renal medulla were negatively correlated with 24 hours proteinuria (r = −0.244, p < 0.05), the degree of tubulointerstitial lesions (r = −0.242, p < 0.05).
Comparison of the mean ADC and R2* values of kidneys in two groups of LN [fig_ref] Table 2: Comparison of the mean ADC and R2* values of kidneys in two... [/fig_ref].
The changes in ADC and R2* values of kidneys after the induction treatment in patients with LN Among 16 patients who repeated MR scan after prednisone combined with immunosuppressant treatment in 9 to 12 months, 12 patients got complete remission (CR) and 4 had no response (NR). Complete remission was defined as urinary protein excretion that was less than 0.3 g/24 h, with normal urine sediment, and serum albumin concentration of greater than 35 g/L, and stabilization (±15%) or improvement in serum creatinine. The ADC and R2* values of kidneys were significantly higher than pretreatment in CR patients (all p < 0.05), while the ADC and R2* values being lower than pre-treatment in NR patients with no statistical significance (p > 0.05) .
# Discussion
The current evaluation of renal lesions in patients with lupus nephritis mainly depends on renal biopsy. Moreover, renal biopsy usually has to be repeated due to the possibility of changes in pathological patterns in LN. Although renal biopsy can provide the pathological information directly, it is not ideal for follow-up of treatment effects. Since it is invasive, SLE patients tend to have a great risk of bleeding due to coagulant function abnormality, or some patients are intolerable for renal biopsies if their kidneys are atrophy. Ultrasonography, commonly used in clinical practice, is simple and noninvasive. However, the accuracy of ultrasonic inspection is unstable depending on the skill of inspectors and body fat mass in patients. Scintigraphy is regarded as the gold standard for evaluating glomerular filtration rate (GFR), but the radioactive pollution, special equipment and other factors limit its clinical use scope [bib_ref] Comparison of magnetic resonance imaging with radionuclide methods of evaluating the kidney, Durand [/bib_ref]. Therefore, it seems critical to find a noninvasive and effective modality to dynamically evaluate renal function and pathological changes in patients with LN.
Because of its great promise in assessing physiological and pathophysiological parameters in vivo, functional magnetic resonance imaging has emerged as a useful technique to investigate organ functional characteristics beyond morphologic changes [bib_ref] Functional renal MR imaging, Huang [/bib_ref] [bib_ref] Schoenberg SO: Functional renal MR imaging: nonvascular renal disease, Michaely [/bib_ref]. Kidneys are located at retro peritoneum and less affected by breath, with plenty of blood supply, high water content in tissues, and urine dilution and concentration function, that make kidneys become one of the most ideal organs for functional MR imaging [bib_ref] MRI to assess renal structure and function, Artunc [/bib_ref] [bib_ref] Blood oxygen level-dependent MR in renal disease: moving toward clinical utility, Zhang [/bib_ref]. DWI is a MR modality using strong bipolar gradients to create a sensitivity of the signal to the thermally induced Brownian motion of water molecules in tissue. Although the etiologies of chronic kidney diseases are different, their physiological and pathophysiological changes are usually common, including reduced blood supply of the kidneys, glomerulosclerosis, tubular atrophy and interstitial fibrosis, inducing abnormal diffusion of water molecules within renal tissue, which will contribute to ADC changes in diffusion-weighted imaging [bib_ref] Diffusion-weighted MR imaging of kidneys in healthy volunteers and patients with parenchyma..., Thoeny [/bib_ref]. Diffusion of water molecules and perfusion of tissues can be affected by various renal lesions such as inflammation, necrosis, edema and fibrosis. The ADC (related only with water diffusion capacity) and perfusion scores (related to perfusion in the microvasculature) measured by DWI of multiple b-values can provide information concerning the characteristics of renal lesions and renal filtration function [bib_ref] Diffusion-weighted MRI in the evaluation of renal lesions: preliminary results, Cova [/bib_ref]. Chronic hypoxic injury in tubulointerstitium may be the common pathway of various chronic kidney diseases progressing to end stage renal disease (ESRD). BOLD MR is used for noninvasive assessment of the intrarenal oxygen content, based on paramagnetic properties of deoxyhemoglobin [bib_ref] Blood oxygen level-dependent MR in renal disease: moving toward clinical utility, Zhang [/bib_ref]. The change of R2* values can be found in the early stage of kidney damage in case of renal artery stenosis, unilateral ureteral obstruction, diabetes, hypertension and rejection after kidney transplantation [bib_ref] BOLD MRI of the kidneys, Li [/bib_ref] [bib_ref] Clinical application of blood oxygen level-dependent MR imaging of kidney: a preliminary..., Xu [/bib_ref].
Togao et al. [bib_ref] Assessment of renal fibrosis with diffusion-weighted MR imaging: study with murine model..., Togao [/bib_ref] [bib_ref] Science to practice: can diffusion-weighted MR imaging findings be used as biomarkers..., Thoeny [/bib_ref] reported that DWI can reflect and monitor the progression of renal fibrosis in mice with unilateral ureteral obstruction, and ADC values were well associated with the changes in renal microstructure. It was found in our previous study [bib_ref] Diffusion-weighted MR imaging of kidneys in patients with chronic kidney disease: initial..., Xu [/bib_ref] that DWI can be helpful to detect the early stage renal failure of CKD. We also found a linear correlation between the ADC values and the stages of CKD. Moreover, the ADC values reflected the unilateral renal function, while renal DWI The relationship of mean ADC and eGFR. In the patients with LN, the mean ADC values were significantly correlated with eGFR (r = 0.510, p < 0.01). was closely associated with renal tubulointerstitial lesions, especially tubular atrophy.
Karadeli et al. [bib_ref] Diffusion-weighted MR imaging of kidneys in patients with systemic lupus erythematosus: initial..., Karadeli [/bib_ref] found significant difference between the ADC low values and the urine protein in patients with SLE, but no significant correlation between GFRs and ADCs. The limitation of their study was the small number of patients and mild SLE (only 10 cases with LN). To the best of our knowledge, our investigation is the first pilot study to evaluate and re-evaluate the renal functional and pathological changes before and after induction treatment in patients with LN by MR functional imaging. We found that renal ADC values and R2* values of cortex and medulla in LN were significantly lower than those in normal kidneys, and these values were significantly higher than before in complete remission patients (all p < 0.05). The results indicated that MR functional imaging had emerging as a new noninvasive method in the evaluation of renal prognosis and therapeutic effects.
There was no significant difference of the ADC values and R2* values of cortex or medulla between bilateral kidneys, reflecting the characteristics of diffuse lesions in LN. We analyzed the correlation of the mean ADC values of bilateral kidneys and eGFR which represents the overall filtration function of bilateral kidneys in patients with LN. It was found that the mean ADC values of kidneys were significantly correlated to eGFR in this study. The mechanism might be: Glomerular sclerosis, interstitial fibrosis and tubular atrophy leads to the decline in glomerular filtration rate, and the limitation in the intercellular movement of water molecules, which causing the abnormal diffusion of water molecules within renal tissues, and then the decrease in renal ADC values. It also can explain why there was a negative correlation between the mean ADC values and pathology chronicity indexes in LN.
Moreover, the R2* values of the renal medulla were negatively correlated with 24 hours proteinuria, the degree of tubulointerstitial lesions. In active lupus nephritis, glomerular cells proliferation or crescent formation, interstitial inflammatory cells infiltration, epithelial and endothelial immune complex deposition is common. Inflammatory factors and immune injury may lead to damage of microvasculature in glomeruli and surrounding renal tubules, which contribute to hemodynamic disturbance and hypoxia within kidneys. Chronic hypoxic injury may further induce glomerular sclerosis, interstitial fibrosis and tubular atrophy, which result in lower metabolism and reduction of oxygen consumption within kidneys, and finally cause a decrease in R2* values. Mixed membranous and proliferative LN (class V + III and V + IV) are usually known with more serious clinical presentation and more severe tubulointerstitial lesions than pure membranous LN (class V) or pure proliferative LN (class III and IV) [bib_ref] Submit your next manuscript to BioMed Central and take full advantage of:..., Najafi [/bib_ref]. In this study, R2* values of cortex in class V + III and V + IV were much lower than that in class III, IV or V, suggesting abnormal microcirculation and chronic hypoxia injury within kidneys with LN (V + III)/(V + IV). These findings should be considered preliminary.
# Conclusions
In summary, diffusion-weighted imaging and blood oxygen level-dependent MR imaging are novel functional MRI techniques with repeatable nature that have potential to assess glomerular filtration, parenchymal oxygenation and renal pathological changes. This pilot study indicated that DWI and BOLD MR imaging might be used to noninvasively monitor disease activity and evaluate therapeutic efficacy in lupus nephritis. Further studies involve larger group of patients with lupus nephritis are necessary.
[fig] Figure 1: Renal functional MR imaging of Patient A with LN-IV(S)A. (a) DWI imaging. (b) T2* imaging. The proteinuria was 1.3 g/24 h and serum creatinine was 52 μmol/L, the renal pathology AI was 12 and CI was 0. Renal BOLD images showed a clear demarcation between cortex and medulla. The ADC value of left kidney was 2.81 x10 −3 mm 2 /s and that of right kidney was 2.78 x10 −3 mm 2 /s, the R2* values were similar to normal kidneys. [/fig]
[fig] Figure 2: Renal functional MR imaging of Patient B with LN-IV(G)A/C + V. (a) DWI imaging. (b) T2* imaging. The proteinuria was 6.1 g/24 h and serum creatinine was 422 μmol/L, the renal pathology AI was 19 and CI was 4. The differentiation between cortex and medulla became less apparent in T2* imaging. The ADC value of left kidney was 2.14 x10 −3 mm 2 /s and that of right kidney was 2.20 x10 −3 mm 2 /s, the R2* values of cortex and medulla were 10.79/s and 8.67/s respectively. [/fig]
[table] Table 1: The mean ADC and R2* values of kidneys in patients with LN compared with healthy volunteersx AE sÞ ð ADC: apparent diffusion coefficient; LN: lupus nephritis. [/table]
[table] Table 2: Comparison of the mean ADC and R2* values of kidneys in two groups of LN x AE sÞ ð [/table]
|
Overexpression of SENP3 in oral squamous cell carcinoma and its association with differentiation
Small ubiquitin-like modifier (SUMO) modification is an important post-translational protein modification that can be reversed by SUMO-specific proteases (SENPs); however, the physiological function of SENPs remains largely unexplored, and little is known about the regulation of SENPs themselves. As one of the crucial members of the SUMO system, SENP3 is essential for rRNA processing and cell proliferation. In the present study, we analysed the expression of SENP3 in human oral squamous cell carcinoma (OSCC) and investigated the correlation between its expression and clinicopathological parameters in OSCC patients. The expression of SENP3 was higher in OSCC tissues than that in the normal mucosa adjacent to the tumor, and a modest increase in reactive oxygen species (ROS) regulated SENP3 stability and localization. ROS induced SENP3 redistribution from the nucleoli to the nucleoplasm. Taken together, these results indicated that the expression level of SENP3 may be associated with the differentiation of OSCC and that SENP3 may play an important role in the development of OSCC under oxidative stress.
# Introduction
Oral squamous cell carcinoma (OSCC) is one of the most common malignant tumors in Southeast Asia, afflicting approximately 300,000 patients worldwide each year [bib_ref] The World Oral Health Report 2003: continuous improvement of oral health in..., Petersen [/bib_ref]. The 5-year survival rate of OSCC patients is approximately 50-60%, and the rate is even lower in patients diagnosed at later stages [bib_ref] Global cancer statistics, Parkin [/bib_ref]. However, the mechanisms involved in the tumorigenesis of have not yet been well characterised.
Small ubiquitin-like modifier (SUMO) modification (SUMOylation) is an important post-translational protein modification that has gained much prominence due to the large number of SUMO substrates [bib_ref] SUMO: a history of modification, Hay [/bib_ref]. SUMOylation is catalysed by SUMO-specific E1, E2 and E3 enzymes, and the covalent modification of proteins is reversed by a family of Sentrin/SUMO-specific proteases (SENPs) [bib_ref] Ubiquitin-like proteins: new wines in new bottles, Yeh [/bib_ref]. SUMOylation can regulate a broad spectrum of cellular processes, including DNA replication/repair, cell division, cell signal transduction and nuclear transportation. De-SUMOylation mediated by SENPs has been shown to play an important role in these processes as well [bib_ref] Differential regulation of c-Jun-dependent transcription by SUMO-specific proteases, Cheng [/bib_ref] [bib_ref] SENP1 enhances androgen receptor-dependent transcription through desumoylation of histone deacetylase 1, Cheng [/bib_ref] [bib_ref] SUMO modification of the Ets-related transcription factor ERM inhibits its transcriptional activity, Degerny [/bib_ref] [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] The nucleolar SUMO-specific protease SENP3 reverses SUMO modification of nucleophosmin and is..., Haindl [/bib_ref]. SENPs deconjugate modified proteins and are thus critical for maintaining the level of SUMOylated and un-SUMOylated substrates required for normal physiology. The altered expression of SENPs has been observed in several carcinomas. SENP1 can transform normal prostate epithelia into a dysplastic state and directly modulate several oncogenic pathways in prostate cells, including the androgen receptor, c-Jun and cyclin D1 pathways [bib_ref] Role of desumoylation in the development of prostate cancer, Cheng [/bib_ref]. The assessment of tissues from human prostate cancer patients has revealed elevated mRNA levels of SENP1 and SENP3 [bib_ref] Role of desumoylation in the development of prostate cancer, Cheng [/bib_ref]. As one of the essential members of the SENP family, SENP3 is located in the nucleolus and appears to have a distinct preference for deconjugating SUMO2/3 [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] Characterization of a family of nucleolar SUMO-specific proteases with preference for SUMO-2..., Gong [/bib_ref]. The induction of SENP3 in cancer cells initiates the angiogenic pathway, and SENP3 regulates the transcriptional activity of hypoxia-inducible factor 1α (HIF1α) via de-SUMOylation of the co-regulatory protein, p300 [bib_ref] SENP3 is responsible for HIF-1 transactivation under mild oxidative stress via p300..., Huang [/bib_ref]. SENP3 expression is also increased in prostate cancer and other carcinomas including, ovarian, lung, rectal and colon carcinomas [bib_ref] SENP3-mediated de-conjugation of SUMO2/3 from promytic leukemia is correlated with accelerated cell..., Han [/bib_ref]. However, the mechanisms which trigger de-SUMOylation and regulate SENP3 under various physiological and pathological conditions have not yet been elucidated, and the functions and related mechanisms of SENP3 in OSCC remain unknown.
It has previously been reported that SUMO conjugation, in particular SUMO2/3 conjugation, is a major response to oxidative stress, and the balance between SUMOylation and de-SUMOylation may play a critical role in the cellular adaptive response to reactive oxygen species (ROS) production [bib_ref] Functional heterogeneity of small ubiquitin-related protein modifiers SUMO-1 versus SUMO-2/3, Saitoh [/bib_ref] [bib_ref] Global shifts in protein sumoylation in response to electrophile and oxidative stress, Manza [/bib_ref] [bib_ref] Expression of SUMO-2/3 induced senescence through p53-and pRB-mediated pathways, Li [/bib_ref] [bib_ref] SUMO on the road to neurodegeneration, Dorval [/bib_ref] [bib_ref] Increased protein SUMOylation following focal cerebral ischemia, Cimarosti [/bib_ref] [bib_ref] Transient global cerebral ischemia induces a massive increase in protein sumoylation, Yang [/bib_ref]. A number of studies have found that the development of OSCC correlates with oxidative stress [bib_ref] A review of betel quid chewing, oral cancer and precancer in mainland..., Zhang [/bib_ref] [bib_ref] Blood iron, glutathione, and micronutrient levels and the risk of oral cancer, Richie [/bib_ref]. It is unknown, however, whether the SUMO2/3-specific protease, SENP3, is involved in OSCC development in response to oxidative stress.
To date, research on SUMOylation related to OSCC are limited. Therefore, the aim of this study was to investigate the expression of SENP3 in OSCC cell lines and clinical tissue samples from OSCC patients, as well as the possible correlations between SENP3 expression and the clinicopathological characteristics of OSCC patients.
# Materials and methods
Cell lines and reagents. The HIOEC cell line and the human OSCC cell lines, HB, CAL-27, WSU-HN6, SCC-4, SCC-9, SCC-25 and LEUK-1, were obtained from the Laboratory of Oral Oncology, Ninth People's Hospital, School of Medicine, Shanghai Jiao Tong University. HIOEC cells were maintained in defined keratinocyte-SFM medium (Gibco-BRL, Carlsbad, CA, USA) and the other cells were maintained in DMEM supplemented with 10% heat-inactivated fetal bovine serum (Gibco). All cells were cultured in a humidified atmosphere of 5% CO 2 at 37˚C. To investigate the association with ROS, cells were treated with H 2 O 2 and N-acetyl cysteine (NAC) (Gibco). RFP-SENP3 was constructed by our laboratory. [fig_ref] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent... [/fig_ref]. For immunohistochemical examination, OSCC tissues were fixed with 4% paraformaldehyde and embedded with paraffin. Sections of the samples were blocked with 10% goat serum in PBS and incubated overnight at 4˚C with anti-SENP3 antibody (Abcam). After 3 washes with PBS, the sections were incubated with peroxidase-conjugated goat anti-rabbit antibody for 1 h, followed by incubation with 3,3'-diaminobenzidine (DAB) substrate for 3 min. Counter-staining was performed with hematoxylin, and dehydration was then performed with ethanol and dimethyl benzene. Slides were mounted with Permount (Santa Cruz Biotechnology, Inc.), and visualised under a light microscope (Axio Imager; Carl Zeiss, Jena, Germany). The histological grade of the tumors was determined according to the degree of differentiation in the WHO histological criteria.
## Reverse transcription-pcr (rt-pcr).
To determine SENP3 expression at the mRNA level, RT-PCR was performed. Total RNA was isolated using TRIzol reagent (Invitrogen, San Diego, USA) and the reverse transcription of cDNA was performed using an Omniscript RT kit (Promega) according to the manufacturer's instructions. Primer sequences used were as follows: SENP3, 5'-GGCAGAATAATGACAGTGAC-3' (forward) and 5'-AGTGACACAGCTCCTTGT-3' (reverse); GAPDH, 5'-CTC CTCCACATCCCTTCC-3' (forward) and 5'-CCGCACGT TCAAGAACAGAGA-3' (reverse). Thermocycler conditions comprised of an initial denaturation at 94˚C for 5 min, 35 cycles of PCR program: denaturing at 94˚C for 30 sec, annealing at 55˚C for 30 sec and elongation at 72˚C for 30 sec and a final extension at 72˚C for 10 min. The amplified PCR products were electrophoresed on a 1.5% agarose gel and visualised with ethidium bromide. The electrophoretic bands were documented with the Gene Genius Gel Documentation System (Syngene Inc., Cambridge, UK).
Western blot analysis. CAL-27 cells were stimulated with H 2 O 2 at various concentrations at 37˚C and then lysed with M-PER ® Mammalian Protein Extraction (Pierce). Proteins were quantified by the BCA Protein Assay kit (Pierce) according to the manufacturer's instructions. Samples containing a total of 50 µg protein were incubated at 100˚C for 5 min, separated by SDS-polyacrylamide gel electrophoresis, and subsequently electrotransferred onto a polyvinylidene difluoride membrane. Essential component detection in the cells was performed with anti-SENP3 (Abcam) antibody at overnight incubation at 4˚C, and then HRP-conjugated secondary antibody (1:5,000 dilution; Pierce Chemical) was added for 1 h at room temperature, followed by the development of reactions in a chemiluminescent detection system. β-actin antibody was used as the control.
Immunostaining. CAL-27 cells deposited on glass were transfected with RFP-SENP3 for 24 h with Lipofectamine 2000 (Invitrogen), and stimulated with H 2 O 2 for 1 h. After washing, the cells were fixed with 4% paraformaldehyde in PBS for 20 min. The sections were then mounted in medium containing DAPI for 5 min to visualize cell nuclei. Slides were evaluated with a laser scanning confocal microscope (TCS SP2; Leica, Wetzlar, Germany).
Statistical analysis. Statistical analysis was performed using SPSS 10.0 software (SPSS Inc., Chicago, IL, USA). The statistical difference of the initial data was analysed by non-parametric tests. P<0.05 was considered to indicate a statistically significant difference.
# Results
## Senp3 protein expression is higher in tumors and is associated with clinicopathological characteristics in patients with oscc.
The immunohistochemistry results showed that the expression of SENP3 was mainly found in the cellular cytoplasm and nucleolus [fig_ref] Figure 1: Expression of SENP3 in oral epithelial tissue adjacent to tumor and oral... [/fig_ref] (30) 10 (25) 14 (35) SENP3 protein expression in the cancerous tissues were consistently stronger than those in the paired adjacent non-malignant epithelia (Wilcoxon signed-rank test Z=-2.337, P<0.05, [fig_ref] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent... [/fig_ref]. The difference in the SENP3 protein positivity score in cancerous tissues was analysed by classification of the different clinicopathological characteristics of the OSCC patients. The protein expression level of SENP3 in the cancerous tissues was found to correlate with the pathological differentiation grade of OSCC (P<0.05). Well-differentiated cancerous tissues displayed the strongest staining [fig_ref] Figure 1: Expression of SENP3 in oral epithelial tissue adjacent to tumor and oral... [/fig_ref] , moderately differentiated cancerous tissues stained weaker [fig_ref] Figure 1: Expression of SENP3 in oral epithelial tissue adjacent to tumor and oral... [/fig_ref] and poorly differentiated cancerous tissues displayed the weakest staining [fig_ref] Figure 1: Expression of SENP3 in oral epithelial tissue adjacent to tumor and oral... [/fig_ref]. No statistically significant associations were detected among SENP3 expression and other clinicopathological characteristics [fig_ref] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent... [/fig_ref].
## H 2 o 2 induces a rapid increase of senp3 protein expression in
OSCC cell lines. SENP3 mRNA levels were higher in OSCC cell lines than in the human immortalized oral epithelial cells HIOEC [fig_ref] Figure 2: Expression of SENP3 in OSCC cell lines [/fig_ref]. CAL-27 cells were exposed to various concentrations of H 2 O 2 and the expression of SENP3 was evaluated. The increase of SENP3 protein expression was observed following treatment with 50 µM H 2 O 2 , and this increase occurred in a dose-dependent manner [fig_ref] Figure 2: Expression of SENP3 in OSCC cell lines [/fig_ref]. As ROS generation is common to some stress inducers, we examined whether the increase in the SENP3 protein level could be blocked by anti-oxidants such as NAC. Indeed, the increase in SENP3 protein expression was blocked by NAC [fig_ref] Figure 2: Expression of SENP3 in OSCC cell lines [/fig_ref] , suggesting that the content of SENP3 protein is regulated by changes in the cellular redox state.
SENP3 is redistributed from the nucleolus to the nucleoplasm upon exposure to H 2 O 2 . SENP3 has been reported to be preferentially localized in the nucleolus [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] A novel mammalian Smt3-specific isopeptidase 1 (SMT3IP1) localized in the nucleolus at..., Nishida [/bib_ref]. Therefore, we examined whether the localization of SENP3 in OSCC cells could be altered following exposure of the cells to H 2 O 2 . RFP-SENP3 was overexpressed in the CAL-27 cells and it was found to be mainly localized in the nucleolus in the untreated cells as expected, while it was almost invisible in the nucleoplasm [fig_ref] Figure 3: SENP3 is redistributed between the nucleolus and the nucleoplasm upon exposure to... [/fig_ref]. However, SENP3 appeared in the nucleoplasm when the cells were exposed to H 2 O 2 , observed as a dim dispersion with numerous bright foci [fig_ref] Figure 3: SENP3 is redistributed between the nucleolus and the nucleoplasm upon exposure to... [/fig_ref] , which indicated that the localization of SENP3 was actually regulated by H 2 O 2 .
# Discussion
The conjugation and de-conjugation of SUMOylation to substrates are sophisticatedly controlled and play an impor-tant physiological and pathological role in determining the outcome of numerous cellular processes. Although a number of studies have revealed the correlation between SUMOylation and tumors, there are only a few studies on SENP family expression in tumors. SENP1 is the most potent regulator of androgen receptor-dependent transcription [bib_ref] SENP1 enhances androgen receptor-dependent transcription through desumoylation of histone deacetylase 1, Cheng [/bib_ref]. Certain studies have shown that SENP1 is associated with the development of prostate cancer and that SENP1 overexpression is found in thyroid oncocytic tumors [bib_ref] Characterization of a family of nucleolar SUMO-specific proteases with preference for SUMO-2..., Gong [/bib_ref] [bib_ref] Two-step differential expression analysis reveals a new set of genes involved in..., Jacques [/bib_ref]. SENP5 is required to maintain mitochondrial morphology and function, as well as cell division [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] The SUMO protease SENP5 is required to maintain mitochondrial morphology and function, Zunino [/bib_ref]. SENP5 has been shown to be expressed in OSCC and correlates with the differentiation of OSCC [bib_ref] Overexpression of SENP5 in oral squamous cell carcinoma and its association with..., Ding [/bib_ref]. However, the association between other members of the SENP family and OSCC has not yet been extensively investigated. As one of the essential members of the SENP family, SENP3 is localized in the nucleolus, which suggests that it may be involved in regulating certain aspects of nucleolar function [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] A novel mammalian Smt3-specific isopeptidase 1 (SMT3IP1) localized in the nucleolus at..., Nishida [/bib_ref]. In addition, SENP3 prefers SUMO2/3 as its substrates and is required for rRNA processing and maturation through de-conjugation of SUMO2/3 from nucleophosmin [bib_ref] The SUMO-specific protease SENP5 is required for cell division, Bacco [/bib_ref] [bib_ref] The nucleolar SUMO-specific protease SENP3 reverses SUMO modification of nucleophosmin and is..., Haindl [/bib_ref]. SENP3-mediated de-conjugation of SUMO2/3 from promyelocytic leukemia has been shown to correlate with accelerated cell proliferation under mild oxidative stress [bib_ref] SENP3-mediated de-conjugation of SUMO2/3 from promytic leukemia is correlated with accelerated cell..., Han [/bib_ref]. SENP3 is also responsible for enhancing HIF-1 transactivation under mild oxidative stress via p300 de-SUMOylation [bib_ref] SENP3 is responsible for HIF-1 transactivation under mild oxidative stress via p300..., Huang [/bib_ref]. The stability of SENP3 is regulated by ROS via interactions with CHIP and Hsp90 [bib_ref] Redox regulation of the stability of the SUMO protease SENP3 via interactions..., Yan [/bib_ref]. However, the correlation between the overexpression of SENP3 and its association with OSCC has yet not been elucidated.
In the present study, we found that SENP3 was overexpressed in OSCC cell lines compared to HIOEC, a normal oral epithelial cell line [fig_ref] Figure 2: Expression of SENP3 in OSCC cell lines [/fig_ref]. SENP3 protein expression was higher in OSCC tissues compared to epithelial adjacent to tumor tissues [fig_ref] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent... [/fig_ref] , which indicated that SENP3 was associated with OSCC differentiation. A significant correlation between SENP3 protein expression and the pathological differentiation grade (P<0.05) was observed in our study [fig_ref] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent... [/fig_ref]. This correlation implies that the well-differentiated cases have a higher SENP3 expression. This suggests the possibility that SENP3 may be a marker of differentiation in OSCC and has implications in the induced differentiation therapy of OSCC.
As is well known, smoking, alcohol consumption and chewing betel nut are the main factors in inducing oral squamous cell carcinogenesis, and all can produce ROS [bib_ref] A review of betel quid chewing, oral cancer and precancer in mainland..., Zhang [/bib_ref] [bib_ref] Blood iron, glutathione, and micronutrient levels and the risk of oral cancer, Richie [/bib_ref]. At the same time, mild ROS can induce SENP3 stabilization and activation. SENP3 is a redox sensor, which can regulate transcription factor (TF) activity under oxidative stress through the de-SUMOylation of proteins [bib_ref] Molecular basis of the redox regulation of SUMO proteases: a protective mechanism..., Xu [/bib_ref]. We also found that the accumulation and redistribution of SENP3 in OSCC cell lines in response to oxidative stress, may enable SENP3 to de-SUMOylate its extra-nucleolar substrates and regulate related nuclear events, such as regulating the SUMOylation status of TFs. However, although the functional analysis of the SUMO2/3-specific protease, SENP3, has recently been reported [bib_ref] The nucleolar SUMO-specific protease SENP3 reverses SUMO modification of nucleophosmin and is..., Haindl [/bib_ref] [bib_ref] Overexpression of SENP5 in oral squamous cell carcinoma and its association with..., Ding [/bib_ref] [bib_ref] A novel mammalian Smt3-specific isopeptidase 1 (SMT3IP1) localized in the nucleolus at..., Nishida [/bib_ref] [bib_ref] Redox regulation of the stability of the SUMO protease SENP3 via interactions..., Yan [/bib_ref] , the mechanisms by which the protease functions in response to a variety of cellular stresses in the development of OSCC remain to be elucidated.
Our findings are in agreement with the concept that SENP3 acts as a redox sensor [bib_ref] Molecular basis of the redox regulation of SUMO proteases: a protective mechanism..., Xu [/bib_ref]. We also demonstrate that SENP3 is redistributed in the subnuclear compartments upon H 2 O 2 treatment. It is likely that SENP3, preferentially residing in the nucleolus, is continuously degraded in the nucleoplasm in the basal state. Several TF luciferase reporters have been used to investigate whether SENP3 affects transcriptional activity, such as nuclear factor-κB (NF-κB), promyelocytic leukemia protein (PML) and signal transducer and activator of transcription 3 (STAT3); however, the effects have varied [bib_ref] SENP3 is responsible for HIF-1 transactivation under mild oxidative stress via p300..., Huang [/bib_ref]. This indicates that the de-SUMOylation of TFs by SENP3 is crucial for certain TF activity. In addition, some of these TFs themselves may be indirect substrates of SENP3, which makes the regulation more complex. SENP3 is rapidly stabilised and redistributed from the nucleolus to the nucleoplasm in response to mild oxidative stress, thus serving as a redox sensor. The increased nucleoplasmic SENP3, in turn, acts as an initiator to affect certain TF activity through de-SUMOylation, which may play an important role in the development of OSCC. However, the direct substrates of SENP3 are largely unknown in OSCC. For example, the expression of SUMO-2/3 induces senescence through p53-and pRB-mediated pathways, which can stimulate the transcriptional activity of p53 [bib_ref] Expression of SUMO-2/3 induced senescence through p53-and pRB-mediated pathways, Li [/bib_ref]. It remains unknown, however, whether the transcriptional activity of p53 is affected by SENP3.
In conclusion, the present study demonstrates the expression pattern of SENP3 in OSCC and the correlation between SENP3 expression and the differentiation of OSCC. As a redox sensor, SENP3 may play an important role in the course of OSCC tumorigenesis and progression in response to oxidative stress. Further studies are required to clarify the function of overexpression of SENP3 in OSCC and to identify the direct substrates of SENP3. Understanding the molecular mechanisms of OSCC development may provide valuable information on diagnosis and precaution, and may aid in the development of novel therapeutic strategies.
[fig] Figure 2: Expression of SENP3 in OSCC cell lines. H 2 O 2 induced a rapid stabilization of SENP3 protein. (A) SENP3 mRNA levels in the human immortalised oral epithelial cells and OSCC cell lines were evaluated by RT-PCR. GAPDH was used as the internal control. (B) CAL-27 cells were pre-treated with 10 mM NAC for 4 h prior to H 2 O 2 exposure for another 1 h. SENP3 protein levels were evaluated by immunoblotting. β-actin was used as the loading control. [/fig]
[fig] Figure 1: Expression of SENP3 in oral epithelial tissue adjacent to tumor and oral squamous cell carcinoma (OSCC) tissues. SENP3 expression in (A) oral epithelial tissue adjacent to tumor, (B) well-differentiated, (C) moderately differentiated and (D) poorly differentiated OSCC (x200). [/fig]
[fig] Figure 3: SENP3 is redistributed between the nucleolus and the nucleoplasm upon exposure to H 2 O 2 . CAL-27 cells were transfected with RFP-SENP3 for 24 h, and the cells were then treated with 100 µM H 2 O 2 for 1 h. Cell monolayers were fixed and cell nuclei were stained with DAPI (x200). [/fig]
[table] Table I: Immunohistochemical evaluation of SENP3 in OSCC patients and oral epithelial tissue adjacent to tumor (Wilcoxon signed-rank test). [/table]
|
Reliability of a rapid hematology stain for sputum cytology*
Objective:To determine the reliability of a rapid hematology stain for the cytological analysis of induced sputum samples. Methods: This was a cross-sectional study comparing the standard technique (May-Grünwald-Giemsa stain) with a rapid hematology stain (Diff-Quik). Of the 50 subjects included in the study, 21 had asthma, 19 had COPD, and 10 were healthy (controls). From the induced sputum samples collected, we prepared four slides: two were stained with May-Grünwald-Giemsa, and two were stained with Diff-Quik. The slides were read independently by two trained researchers blinded to the identification of the slides. The reliability for cell counting using the two techniques was evaluated by determining the intraclass correlation coefficients (ICCs) for intraobserver and interobserver agreement. Agreement in the identification of neutrophilic and eosinophilic sputum between the observers and between the stains was evaluated with kappa statistics. Results: In our comparison of the two staining techniques, the ICCs indicated almost perfect interobserver agreement for neutrophil, eosinophil, and macrophage counts (ICC: 0.98-1.00), as well as substantial agreement for lymphocyte counts (ICC: 0.76-0.83). Intraobserver agreement was almost perfect for neutrophil, eosinophil, and macrophage counts (ICC: 0.96-0.99), whereas it was moderate to substantial for lymphocyte counts (ICC = 0.65 and 0.75 for the two observers, respectively). Interobserver agreement for the identification of eosinophilic and neutrophilic sputum using the two techniques ranged from substantial to almost perfect (kappa range: 0.91-1.00). Conclusions: The use of Diff-Quik can be considered a reliable alternative for the processing of sputum samples.ResumoObjetivo: Determinar a confiabilidade da coloração hematológica rápida para a análise do escarro induzido. Métodos: Estudo transversal comparando a técnica padrão (coloração May-Grünwald-Giemsa) com a coloração hematológica rápida (panótico rápido). Participaram do estudo 50 indivíduos (21 asmáticos, 19 portadores de DPOC e 10 controles). Após a coleta do escarro induzido, foram preparadas 4 lâminas, sendo 2 coradas por May-Grünwald-Giemsa e 2 por panótico rápido. As lâminas foram lidas de forma independente por dois pesquisadores capacitados para o exame de escarro induzido e cegados para a identificação das lâminas. A confiabilidade para as contagens celulares dos dois métodos foi avaliada pela determinação dos coeficientes de correlação intraclasse (CCI) para as concordâncias intraobservador e interobservador. As concordâncias na identificação de escarro neutrofílico e eosinofílico entre observadores e entre as duas colorações foram calculadas por estatística kappa. Resultados: Nas duas colorações, os CCI apontaram concordância interobservador quase perfeita para as contagens de neutrófilos, eosinófilos e macrófagos (variação do CCI: 0,98-1,00) e substancial para as contagens de linfócitos (variação do CCI: 0,76-0,83). Na análise intraobservador, a concordância foi quase perfeita para as contagens de neutrófilos, eosinófilos e macrófagos (variação do CCI: 0,96-0,99) e de moderada a substancial para as contagens de linfócitos (CCI = 0,65 e 0,75 para observadores 1 e 2, respectivamente). A concordância interobservador na identificação de escarro eosinofílico e neutrofílico para os dois métodos de coloração variou entre substancial e quase perfeita (variação kappa: 0,91-1,00). Conclusões: O panótico rápido pode ser considerado uma alternativa confiável para o processamento de amostras de escarro.
## 251
http://dx.doi.org/10.1590/S1806-37132014000300008 shorter; whereas the standard stain requires 34 min, the rapid stain is performed within 2 min.
Shortening the total processing time would be extremely important for improving the viability of the technique. In addition, the cost of Diff-Quik is considerably lower than that of the standard stain. Therefore, the validation of this technique is important, especially for developing countries, such as Brazil.
The objective of the present study was to assess the reliability of Diff-Quik for the cytological analysis of induced sputum samples.
# Methods
The study included 50 patients, of whom 21 were adult patients with uncontrolled asthma, characterized by an Asthma Control Questionnaire [bib_ref] Development and validation of a questionnaire to measure asthma control, Juniper [/bib_ref] score greater than 1.7 in the previous week and objectively confirmed (in the 3 previous years) by reversible airflow limitation (a > 12% increase in FEV 1 and a > 200 mL increase in FEV 1 after inhalation of a shortacting bronchodilator) in participants with airflow limitation (an FEV 1 /FVC ratio < 0.7); 19 COPD patients aged > 40 years who had a history of respiratory symptoms associated with moderate or severe airflow obstruction (an FEV 1 < 50% of predicted and an FEV 1 /FVC ratio < 0.7), were receiving any type of medication for COPD, and were (current or former) heavy smokers with a smoking history of > 20 packyears; and 10 healthy nonsmokers who had no respiratory symptoms and whose diagnostic status was objectively confirmed by normal spirometry results. The study excluded patients who had respiratory infection in the four previous weeks, those who had severe diseases of other systems, those who had other known pulmonary diseases, and pregnant women.
# Introduction
The understanding of the mechanisms of diseases and their correct diagnosis has been made possible by analysis of body fluids in several areas of medicine. In the past, sputum analysis was considered not to be reliable or reproducible enough to assist in the understanding of the mechanisms of respiratory diseases.More recently, significant advances related to the processing of sputum samples have allowed the method of sputum examination to become feasible, reproducible, valid, and responsive to interventions. Several researchers have used this method to study the various aspects of airway inflammation in asthma. The use of sputum examination has been further extended to COPD, cystic fibrosis, chronic cough, idiopathic pulmonary fibrosis, and other respiratory diseases.
Inflammation is central in the pathogenesis of airway diseases and is considered responsible for their symptoms, airflow obstruction, exacerbations, and secondary structural changes. [bib_ref] Sputum Examination for indices of airway inflammation: laboratory procedures, Pizzichini [/bib_ref] Therefore, airway inflammation plays an extremely significant role in two major obstructive respiratory tract diseases: asthma and COPD.Both in asthma and in COPD, there is great heterogeneity in clinical and inflammatory characteristics, which result in different clinical phenotypes. [bib_ref] Clinical applications of assessment of airway inflammation using induced sputum, Pavord [/bib_ref] Consequently, there is a need to characterize the phenotype of patients in order to optimize their clinical management, particularly in more severe cases. [bib_ref] Non-eosinophilic corticosteroid unresponsive asthma, Pavord [/bib_ref] Examination of induced sputum is currently considered reliable, reproducible, discriminative of different types of inflammation, and responsive to interventions. Therefore, it has been an important tool in the management of inflammatory diseases of the airways. [bib_ref] Exame físico-valorização do exame de escarro, Polombini [/bib_ref] In addition, sputum induction is a safe minimally invasive technique, [bib_ref] Safety of sputum induction, Pizzichini [/bib_ref] making it possible to identify the type of inflammation and its intensity.However, its use in clinical practice is still very restricted because of the method for sputum induction and for the processing of sputum samples, which is laborious and timeconsuming and requires highly trained personnel.
Induced sputum slides for differential cell counts, both in research and in clinical practice, have been stained with May-Grünwald-Giemsa or with Wright-Giemsa. These stains are part of a group called "Romanowsky stains". [bib_ref] How Romanowsky stains work and why they remain valuable -including a proposed..., Horobin [/bib_ref] Diff-Quik (a rapid hematology stain), which is also based on the Romanowsky technique,uses similar reagents, but the staining time is considerably grade methanol, and then they were immersed in a May-Grünwald stain solution and a dilute May-Grünwald solution (1:1). Immediately afterward, the slides were immersed in a freshly prepared Giemsa stain solution (dilution 1:10) and subsequently dried. The whole procedure lasted exactly 34 min, as recommended by the equipment manufacturer.
Diff-Quik was performed manually. The process included initial immersion of the slides in solution no. 1 (0.1% triarylmethane), moving up and down continuously for 5-10 seconds (5-10 one-second immersions). Subsequently, extensions were immersed in solution no. 2 (0.1% xanthene), repeating the same procedure. After draining, the slides were immersed in solution no. 3 (0.1% thiazine), repeating the same procedure. The slides were rinsed with distilled water and allowed to air-dry.This staining procedure took a maximum of 2 min to complete. Two researchers, trained in reading induced sputum slides, independently counted 400 non-squamous cells on the slides stained either with May-Grünwald-Giemsa or Diff-Quik. Because the slides were coded, the slide readers were prevented from identifying their respective pairs or previous reading results.
The reliability for cell counting using the two techniques (standard stain vs. tested stain) was evaluated by determining the intraclass correlation coefficients (ICCs) for intraobserver and interobserver agreement, and Bland & Altman plots were used. [bib_ref] Comparing methods of measurement: why plotting difference against standard method is misleading, Bland [/bib_ref] The interpretation of ICCs was based on the classification proposed by Landis & Koch. [bib_ref] The measurement of observer agreement for categorical data, Landis [/bib_ref] Agreement in the identification of eosinophilic sputum (eosinophils ≥ 3%)and neutrophilic sputum (neutrophils > 64%) [bib_ref] Induced sputum cell counts in healthy adults, Belda [/bib_ref] between the observers and between the stains was evaluated with kappa statistics. [bib_ref] The measurement of observer agreement for categorical data, Landis [/bib_ref] Differences between the characteristics of the groups studied were examined by ANOVA and the Bonferroni test in the post hoc analysis. The statistical tests were two-tailed, and the level of significance was set at 5%. The Statistical Package for the Social Sciences, version 18.0 (SPSS Inc., Chicago, IL, USA) was used for the analyses.
# Results
Sputum induction was performed in 62 individuals, 50 (80.6%) of whom were able to produce an adequate sample, with cell viability greater than 50%. Twelve induced sputum samples were considered inadequate because American Thoracic Society guidelines The reference values used were those of Crapo et al. [bib_ref] Reference spirometric values using techniques and equipment that meet ATS recommendations, Crapo [/bib_ref] Subsequently, sputum induction was performed in accordance with the method described by Pizzichini et al. [bib_ref] Measurement of inflammatory indices in induced sputum: effects of selection of sputum..., Pizzichini [/bib_ref] The procedure involved inhalation of an isotonic saline aerosol (0.9%) followed by serial inhalation of increasing concentrations of hypertonic saline aerosol (3%, 4%, and 5%) via a Fisoneb ultrasonic nebulizer (Fisons, Pickering, Ontario, Canada). Aerosol inhalation was continued for 1-2 min, according to the level of bronchoconstriction severity before the procedure, and was followed by measurement of FEV 1 . Participants were instructed to rinse their mouths with water, swallow the water, and blow their noses in order to reduce contamination by saliva or postnasal discharge. They were then asked to cough and expectorate the sputum into a clean container. These procedures were repeated consecutively, with the solution concentration being increased every 7 min for 21 min or until there was a ≥ 20% decrease in FEV 1 .
The processing of sputum samples was started within 2 h of collection, which is the longest time reported in the literature. [bib_ref] Sputum Examination for indices of airway inflammation: laboratory procedures, Pizzichini [/bib_ref] The thick portions of the expectorated material were selected with the naked eye or under visualization with an inverted microscope, and the sputum was separated from the saliva. The selected fractions were treated with 0.1% DTT at a ratio of four times the fraction volume. This mixture was homogenized with a Pasteur pipette and agitated on a desktop shaker for 15 min. Dulbecco's PBS was added thereto in an amount that was four times the initial volume of sputum selected, and the resulting suspension was filtered to remove cell debris and undissolved mucus. Subsequently, total leukocyte counts were performed with a modified Neubauer hemocytometer, excluding squamous cells. Cell viability was determined by the trypan blue exclusion test. The sample was adjusted to 1.0 × 10 6 cells/mL, and we prepared four slides, which were coded. In the present study, all of the slides were prepared using the cytocentrifugation method (cytospin). After air-drying, two slides were stained with May-Grünwald-Giemsa, and the other two were stained with Diff-Quik (the technique under study).
May-Grünwald-Giemsa staining was performed with an automated system (Sysmex sp1000i TM ; Sysmex Co., Kobe, Japan). For this technique, the slides were fixed by immersion in analytical 253 http://dx.doi.org/10.1590/S1806-37132014000300008 was substantial (ICC = 0.76). For the slides stained by the Diff-Quik technique, the medians and percentiles were also very close between the two observers, and the ICC indicated almost perfect interobserver agreement for eosinophil, neutrophil, macrophage, and lymphocyte counts (ICC = 1.00, 0.99, 0.99, and 0.83, respectively).
Regarding intraobserver agreement, the ICC values for the two observers for the differential cell counts on the pairs of cytospin slides stained by either of the two studied techniques indicated that it was almost perfect for neutrophils (ICC = 0.97 for both), eosinophils (ICC = 0.99 and 0.98), and macrophages (ICC = 0.96 for both). For lymphocyte counts, intraobserver agreement was substantial for observer 1 (ICC = 0.75) and moderate for observer 2 (ICC = 0.65). These results are shown graphically in [fig_ref] Figure 3 -: Bland & Altman plots [/fig_ref]. [bib_ref] Comparing methods of measurement: why plotting difference against standard method is misleading, Bland [/bib_ref] Interobserver agreement for the identification of eosinophilic and neutrophilic sputum using the two techniques ranged from substantial to almost perfect, as shown in [fig_ref] Table 2 -: Interobserver and intraobserver agreement for the identification of eosinophilic and neutrophilic sputum... [/fig_ref]. However, although intraobserver agreement was substantial, it was lower than was interobserver agreement.
# Discussion
The results of the present study show that cytospin slides stained either by the May-Grünwald-Giemsa technique or with Diff-Quik yield similar cell counts, with high intraobserver and intraobserver agreement. These results demonstrate the reliability of the Diff-Quik technique for use in the processing of induced sputum samples. This fact is relevant because the Diff-Quik technique is simpler, allows a reduction in sample processing of excessive salivary contamination (> 20% of squamous cells), low cell viability (< 50%), or insufficient material to prepare the slides. The demographic, clinical, and functional characteristics of the participants are shown in [fig_ref] Table 1 -: Demographic, clinical, and functional characteristics of the participants [/fig_ref]. The groups were distinct and well characterized, as demonstrated by their demographic, clinical, and functional characteristics.
The cellular characteristics of induced sputum were as expected for the different groups studied. The sputum of asthma patients was characterized by a significantly higher proportion of eosinophils than that found in the sputum of COPD patients and healthy controls. In contrast, the sputum of COPD patients showed a significant increase in total cell counts and in the proportion of neutrophils when compared with that of controls. The control group showed a significantly higher proportion of macrophages than did the other two groups. [fig_ref] Figure 1 -: Induced sputum differential cell counts in the three groups studied [/fig_ref] shows the proportions of neutrophils, eosinophils, and macrophages in the different groups studied. In [fig_ref] Figure 2 -: Median differential cell counts on the cytospin slides stained either with May-Grünwald-Giemsa [/fig_ref] , the median proportions of neutrophils, eosinophils, and macrophages, in the study sample as a whole, are separated by type of stain used. No significant differences were found for the cell counts on the slides stained by either of the two techniques used.
The results for interobserver agreement for induced sputum differential cell counts on the cytospin slides stained by the May-Grünwald-Giemsa technique show that the medians and percentiles were similar between the two observers, and the ICCs indicated almost perfect agreement for eosinophil, neutrophil, and macrophage counts (ICC = 1.00, 0.99, and 0.98, respectively). Interobserver agreement for lymphocyte counts staining techniques was found to be substantial. In addition, interobserver agreement for the identification of eosinophilic and neutrophilic sputum was almost perfect. These results again demonstrate the reliability of Diff-Quik, because they confirm its accuracy for identifying the different inflammatory phenotypes. However, the results also showed that interobserver agreement was higher than intraobserver agreement for the identification of the phenotypes. This difference could be due to variability in cell content on the slides stained by either of the two techniques. Although intraobserver agreement was substantial, this particular result suggests caution and a need for further studies to identify the reason for this variability. Interobserver reproducibility for differential leukocyte counts in induced sputum samples has been previously reported. [bib_ref] The influence of cell viability and squamous epithelial cell contamination on the..., Efthimiadis [/bib_ref] [bib_ref] Induced sputum to assess airway inflammation: a study of reproducibility, Spanevello [/bib_ref] In 1997, one group of authors reported high interobserver reproducibility for all cell types studied. Those authors also found lower agreement for lymphocytes than for the other cell types. The lower agreement for the proportions of lymphocytes was considered to be due to the very small amount of this cell type in the induced sputum samples. In the present study, we also found lower agreement for lymphocyte counts. However, this was not emphasized because it is a known fact that these variations are not clinically relevant.
In one study,there was good interobserver agreement for neutrophil, eosinophil, and macrophage counts, and, again, this agreement was lower for lymphocyte counts; the justification for the lower repeatability rate was related not only to the scarcity of lymphocytes in the samples time of up to 32 min without impairing sample quality, and is considerably cheaper.
To our knowledge, this was the first study to evaluate the reliability of the Diff-Quik technique for use in induced sputum cytology, by comparing it with a standard staining technique, i.e., the May-Grünwald-Giemsa technique. It is important to evaluate the reliability and reproducibility of the results in order to confirm the accuracy of the results obtained by using Diff-Quik. In the present study, the reliability of Diff-Quik was tested by two distinct strategies. The first strategy was to calculate the ICCs for the cell counts performed by two independent observers, blinded to the identification of the slides. Although the stains used in the present study could be identified by the appearance of the slides, differing codes were used to prevent the identification of the respective pairs of slides. Previous studies [bib_ref] The influence of cell viability and squamous epithelial cell contamination on the..., Efthimiadis [/bib_ref] [bib_ref] Induced sputum to assess airway inflammation: a study of reproducibility, Spanevello [/bib_ref] have shown that intraobserver and interobserver agreement for cell counts on cytospin slides stained with Wright-Giemsa and May-Grünwald-Giemsa could be considered perfect, but that it depended on the degree of salivary contamination on the slides.The results obtained with Diff-Quik in the present study are in line with those of the aforementioned studies. The second strategy was to examine the reliability of Diff-Quik for identifying eosinophilic sputum and neutrophilic sputum. This is relevant because, in clinical practice, sputum examination is used to identify phenotypes of severe asthma, predict response to treatment, and decrease the number of asthma exacerbations by control of eosinophilic inflammation.
Intraobserver agreement for the identification of eosinophilic and neutrophilic sputum by the two Methods for refinement of sputum examination have greatly contributed to its accuracy and reproducibility. [bib_ref] Methods of sputum processing for cell counts, immunocytochemistry and in situ hybridisation, Efthimiadis [/bib_ref] However, previous studies have primarily focused on the steps preceding the preparation of cytospin slides and the liquid phase of sputum. [bib_ref] Indices of airway inflammation in induced sputum: reproducibility and validity of cell..., Pizzichini [/bib_ref] [bib_ref] Induced sputum: validity of fluid-phase IL-5 measurement, Kelly [/bib_ref] [bib_ref] Some technical factors influencing the induction of sputum for cell analysis, Popov [/bib_ref] [bib_ref] Induced sputum cell and fluid-phase indices of inflammation: comparison of treatment with..., Efthimiadis [/bib_ref] In 2003, a study comparing the results and costs of three techniques for analysis of induced but also to the difficulty in identifying these cells. One group of authorsconfirmed the results of a previous study, [bib_ref] The influence of cell viability and squamous epithelial cell contamination on the..., Efthimiadis [/bib_ref] demonstrating that the reproducibility of induced sputum differential cell counts is affected by salivary contamination and low cell viability. In the present study, samples with > 20% salivary contamination or < 50% viability were considered inadequate for analysis. In summary, the high degree of agreement for the cell counts on the cytospin slides stained either by the May-Grünwald-Giemsa technique or with Diff-Quik attests to the reliability of the latter stain, which justifies the recommendation that it be used when the objective is to reduce sample processing time and induced sputum costs.
sputum samples was published. [bib_ref] Nunes Mdo P. Comparison of three methods for differential cell count in..., Saraiva-Romanholo [/bib_ref] The techniques consisted of one of the following: preparation of smears from sputum not treated with DTT; preparation of smears from sputum treated with DTT; or preparation of cytospin slides from sputum treated with DTT. Although the first two techniques reduced the time and cost of sputum sample analysis, Spearman's correlation coefficient, which ranged from 0.57 to 0.64 for eosinophil counts and from 0.51 to 0.57 for neutrophil counts ((p < 0.01 for both), can be considered inadequate for these types of techniques. The authors concluded that the technique that uses cytospin slides prepared from samples treated with DTT is more suitable for research purposes and for use in specialized centers. [bib_ref] Nunes Mdo P. Comparison of three methods for differential cell count in..., Saraiva-Romanholo [/bib_ref] Some of the aspects that hinder the widespread use of induced sputum are its complexity and the sample processing time. Two previous studies have attempted to reduce the complexity of sample processing, [bib_ref] Induced sputum (IS) cell counts: a new device to standardize processing in..., Pizzichini [/bib_ref] [bib_ref] A kit to facilitate and standardize the processing of sputum for measurement..., Chaboillez [/bib_ref] both using a sputum filtration device (Accufilter; Cellometrics, Hamilton, Ontario, Canada), which would be an alternative for standardizing the processing of induced sputum samples. The device consists of a kit with a tube used for weighing and treating sputum selected from saliva, connected to a filter and to a reception tube containing DTT, saline solution, and trypan blue. One of those studies [bib_ref] A kit to facilitate and standardize the processing of sputum for measurement..., Chaboillez [/bib_ref] sought to determine the validity of using this device in the analysis of sputum samples, by comparing it with the standard method. The study reported ICCs that indicate good reproducibility for eosinophil and neutrophil counts between the methods; however, the ICCs indicated a reduction in cell viability and total cell counts, as well as an increase in the proportion of squamous epithelial cells, with the use of that device. Reliability of a rapid hematology stain for sputum cytology
[fig] Figure 1 -: Induced sputum differential cell counts in the three groups studied. [/fig]
[fig] Figure 2 -: Median differential cell counts on the cytospin slides stained either with May-Grünwald-Giemsa (dark gray bar) or Diff-Quik (light gray bar). //dx.doi.org/10.1590/S1806-37132014000300008 [/fig]
[fig] Figure 3 -: Bland & Altman plots. Interobserver reproducibility for the proportion of neutrophils (in A), eosinophils (in B), and macrophages (in C) on the cytospin slides prepared from induced sputum samples and stained with May-Grünwald-Giemsa (MGG). Intraobserver reproducibility for the proportion of neutrophils (in D), eosinophils (in E), and macrophages (in F) on the cytospin slides prepared from induced sputum samples and stained by either of the two staining techniques. The plots refer to the differences between the readings by observers 1 and 2 (y axis) in relation to the mean readings by observers 1 and 2 (x axis). The central broken line indicates absence of differences, and the peripheral broken lines indicate two standard deviations of the mean of the differences. ICC: intraclass correlation coefficient; NEU.: neutrophils; EOS.: eosinophils; and MAC.: macrophages. J Bras Pneumol. 2014;40(3):250-258 http://dx.doi.org/10.1590/S1806-37132014000300008 [/fig]
[table] Table 1 -: Demographic, clinical, and functional characteristics of the participants. a [/table]
[table] Table 2 -: Interobserver and intraobserver agreement for the identification of eosinophilic and neutrophilic sputum in the evaluation of slides stained by either of the two studied techniques. [/table]
|
Mens sana in corpore sano: Does the Glycemic Index Have a Role to Play?
Although diet interventions are mostly related to metabolic disorders, nowadays they are used in a wide variety of pathologies. From diabetes and obesity to cardiovascular diseases, to cancer or neurological disorders and stroke, nutritional recommendations are applied to almost all diseases. Among such disorders, metabolic disturbances and brain function and/or diseases have recently been shown to be linked. Indeed, numerous neurological functions are often associated with perturbations of whole-body energy homeostasis. In this regard, specific diets are used in various neurological conditions, such as epilepsy, stroke, or seizure recovery. In addition, Alzheimer's disease and Autism Spectrum Disorders are also considered to be putatively improved by diet interventions. Glycemic index diets are a novel developed indicator expected to anticipate the changes in blood glucose induced by specific foods and how they can affect various physiological functions. Several results have provided indications of the efficiency of low-glycemic index diets in weight management and insulin sensitivity, but also cognitive function, epilepsy treatment, stroke, and neurodegenerative diseases. Overall, studies involving the glycemic index can provide new insights into the relationship between energy homeostasis regulation and brain function or related disorders. Therefore, in this review, we will summarize the main evidence on glycemic index involvement in brain mechanisms of energy homeostasis regulation.
# Introduction
Nutrition has been part of the treatment employed for diabetes and obesity for decades. Therefore, specific food choices to help control weight and glucose homeostasis represent an important step in the establishment of a suitable diet. To assist patients, Jenkins et al. established the glycemic index (GI) concept in 1981 [bib_ref] The Glycaemic Index of Foods Tested in Diabetic Patients: A New Basis..., Jenkins [/bib_ref]. GI measures the impact of an individual food on the blood glucose level over time when compared to the effect of glucose itself (GI = 100). The glycemic response will thus depend on both the quantity and quality of carbohydrates (sugars, starch, or fibers) in the food. Consequently, a low-GI food (GI ≤ 55) contains high-quality carbohydrates and will not raise glycemia as much as a high-GI food (GI ≥ 70) for the same amount of carbohydrate. However, since the GI does not consider the amount of carbohydrate ingested, a glucose load (GL) value was developed. Thereby, GL represents a product of GI (quality of carbohydrate) and the quantity of carbohydrate ingested. Low GL is considered to be below 10, while high GL is above 20 when 10 g of glucose has a GL of 10 . . Example of common foods with their corresponding serving size in g, glycemic index (GI) carbohydrates per serving size in g, and the resulting glucose load (GL). [fig_ref] Table 2: Example of various diets' composition for macronutrients with some examples of common... [/fig_ref] Blue indicates low, orange indicates medium, and green indicates high GI or GL (www.glycemicindex.com) [bib_ref] The Glycaemic Index of Foods Tested in Diabetic Patients: A New Basis..., Jenkins [/bib_ref].
## Food
Nutrients 2020, 12, 2989
## Of 33
In fact, low-GI foods are digested and absorbed slowly compared to high-GI foods. Therefore, low-GI foods induce a limited increase in blood glucose that lasts for an extended period. Conversely, high-GI foods are easily digested and absorbed, and induce a rapid and high increase in blood glucose, followed by insulin secretion that often leads to transient hypoglycemia [fig_ref] Figure 1: Schematic diagram of the influence of GI or GL on blood glucose [/fig_ref] [bib_ref] Glycemic index and glycemic load: Measurement issues and their effect on diet-disease..., Venn [/bib_ref] [bib_ref] Acute effect of meal glycemic index and glycemic load on blood glucose..., Galgani [/bib_ref].
Nutrients 2020, 12, x FOR PEER In fact, low-GI foods are digested and absorbed slowly compared to high-GI foods. Therefore, low-GI foods induce a limited increase in blood glucose that lasts for an extended period. Conversely, high-GI foods are easily digested and absorbed, and induce a rapid and high increase in blood glucose, followed by insulin secretion that often leads to transient hypoglycemia [fig_ref] Figure 1: Schematic diagram of the influence of GI or GL on blood glucose [/fig_ref] [bib_ref] Glycemic index and glycemic load: Measurement issues and their effect on diet-disease..., Venn [/bib_ref] [bib_ref] Acute effect of meal glycemic index and glycemic load on blood glucose..., Galgani [/bib_ref]. Diet interventions are used in physiological conditions, and are also used in sports for weight management, to lose fat mass or gain lean mass. In addition, metabolic disorders are commonly associated with nutritional recommendations to induce weight loss, control glucose levels, or manage dyslipidemia. Most notably, specific diets have also been used in non-metabolic conditions for centuries. For instance, epilepsy treatment has included a ketogenic diet (KD) for almost 100 years [bib_ref] The role for ketogenic diets in epilepsy and status epilepticus in adults, Williams [/bib_ref]. Interestingly, specific diets are used in cancer, cognitive improvement, neurological disorders, and mental diseases [bib_ref] Ketogenic Diet and Other Dietary Intervention Strategies in the Treatment of Cancer, Vergati [/bib_ref] [bib_ref] Ketogenic diets and protective mechanisms in epilepsy, metabolic disorders, cancer, neuronal loss,..., Li [/bib_ref] [bib_ref] Ketogenic diets, mitochondria, and neurological diseases, Gano [/bib_ref] [bib_ref] Brain foods: The effects of nutrients on brain function, Gómez-Pinilla [/bib_ref]. Most of the diets involved are characterized by a low carbohydrate content or low-GI food. Although GI and GL are determined for individual foods, diets based on these values have been developed, as well as methods of evaluation of GI/GL meals [bib_ref] The Use of the Glycemic Index in Predicting the Blood Glucose Response..., Wolever [/bib_ref] [bib_ref] Predicting mixed-meal measured glycaemic index in healthy subjects, Ballance [/bib_ref]. Therefore, the diets used in medical nutrition therapies are expected to be low-GI diets. In fact, the impact of a meal on glycemia is likely to be a key factor in disease development. Indeed, since glucose represents the main source of energy for our body in a healthy state, a disturbed glucose supply will alter the normal function of cells.
The brain represents 2% of the body weight, while it is the main glucose consumer (20%) in humans [bib_ref] Sugar for the brain: The role of glucose in physiological and pathological..., Mergenthaler [/bib_ref]. Therefore, brain control of the glucose supply is finely regulated by the brain itself. Indeed, the brain possesses the ability to sense glucose levels and to trigger an adaptive response when glucose levels are low. Then, the brain will stimulate food intake or glucose production by peripheral organs. This will maintain a constant energy supply for brain activity [bib_ref] The selfish brain: Competition for energy resources, Peters [/bib_ref]. Therefore, brain glucose detection is a key mechanism for both brain activity control and energy homeostasis regulation. The brain plays a central role in the maintenance of energy homeostasis of the whole body. Moreover, nutrient sensing is one of the most regulated mechanisms and involves specific regions such as the hypothalamus [bib_ref] Hypothalamic nutrient sensing in the control of energy homeostasis, Blouet [/bib_ref]. Among the nutrients sensed, glucose is the most studied and characterized [bib_ref] Brain Glucose-Sensing Mechanism and Energy Homeostasis, López-Gambero [/bib_ref]. Since the GI impacts the blood glucose level, the composition of meals based on GI is expected to involve the brain circuits of glucose sensing. Moreover, low-GI diets are also prone Diet interventions are used in physiological conditions, and are also used in sports for weight management, to lose fat mass or gain lean mass. In addition, metabolic disorders are commonly associated with nutritional recommendations to induce weight loss, control glucose levels, or manage dyslipidemia. Most notably, specific diets have also been used in non-metabolic conditions for centuries. For instance, epilepsy treatment has included a ketogenic diet (KD) for almost 100 years [bib_ref] The role for ketogenic diets in epilepsy and status epilepticus in adults, Williams [/bib_ref]. Interestingly, specific diets are used in cancer, cognitive improvement, neurological disorders, and mental diseases [bib_ref] Ketogenic Diet and Other Dietary Intervention Strategies in the Treatment of Cancer, Vergati [/bib_ref] [bib_ref] Ketogenic diets and protective mechanisms in epilepsy, metabolic disorders, cancer, neuronal loss,..., Li [/bib_ref] [bib_ref] Ketogenic diets, mitochondria, and neurological diseases, Gano [/bib_ref] [bib_ref] Brain foods: The effects of nutrients on brain function, Gómez-Pinilla [/bib_ref]. Most of the diets involved are characterized by a low carbohydrate content or low-GI food. Although GI and GL are determined for individual foods, diets based on these values have been developed, as well as methods of evaluation of GI/GL meals [bib_ref] The Use of the Glycemic Index in Predicting the Blood Glucose Response..., Wolever [/bib_ref] [bib_ref] Predicting mixed-meal measured glycaemic index in healthy subjects, Ballance [/bib_ref]. Therefore, the diets used in medical nutrition therapies are expected to be low-GI diets. In fact, the impact of a meal on glycemia is likely to be a key factor in disease development. Indeed, since glucose represents the main source of energy for our body in a healthy state, a disturbed glucose supply will alter the normal function of cells.
The brain represents 2% of the body weight, while it is the main glucose consumer (20%) in humans [bib_ref] Sugar for the brain: The role of glucose in physiological and pathological..., Mergenthaler [/bib_ref]. Therefore, brain control of the glucose supply is finely regulated by the brain itself. Indeed, the brain possesses the ability to sense glucose levels and to trigger an adaptive response when glucose levels are low. Then, the brain will stimulate food intake or glucose production by peripheral organs. This will maintain a constant energy supply for brain activity [bib_ref] The selfish brain: Competition for energy resources, Peters [/bib_ref]. Therefore, brain glucose detection is a key mechanism for both brain activity control and energy homeostasis regulation. The brain plays a central role in the maintenance of energy homeostasis of the whole body. Moreover, nutrient sensing is one of the most regulated mechanisms and involves specific regions such as the hypothalamus [bib_ref] Hypothalamic nutrient sensing in the control of energy homeostasis, Blouet [/bib_ref]. Among the nutrients sensed, glucose is the most studied and characterized [bib_ref] Brain Glucose-Sensing Mechanism and Energy Homeostasis, López-Gambero [/bib_ref]. Since the GI impacts the blood glucose level, the composition of meals based on GI is expected to involve the brain circuits of glucose sensing. Moreover, low-GI diets are also prone to an increased production of ketone bodies since they are low in carbohydrates [bib_ref] Ketogenic diets, mitochondria, and neurological diseases, Gano [/bib_ref]. Finally, the carbohydrate composition of a meal induces changes in gut microbiota and in signals involved in food intake regulation and neurological disorders [bib_ref] The Impact of Dietary Fiber on Gut Microbiota in Host Health and..., Makki [/bib_ref]. Overall, by involving key signals of both metabolic regulations and neuronal functions, GI diets can improve cellular and whole-body metabolism via brain regulation. Indeed, recent reports demonstrate a clear relationship between brain function and energy homeostasis. For instance, Alzheimer's disease (AD) shows metabolic defects, including glucose uptake deficiency in the brain, insulin resistance, and even food intake alterations [bib_ref] Alzheimer's Disease and Diabetes: Insulin Signaling as the Bridge Linking Two Pathologies, Shieh [/bib_ref] [bib_ref] Insulin resistance and impaired lipid metabolism as a potential link between diabetes..., Kulas [/bib_ref]. Furthermore, neurodegeneration is also associated with metabolic impairment and diabetes [bib_ref] Diabetes and neurodegeneration in the brain, Toth [/bib_ref] [bib_ref] Insulin signaling pathway and related molecules: Role in neurodegeneration and Alzheimer's disease, Akhtar [/bib_ref].
Although low-GI diets have been developed to help diabetic people to manage their body weight and glycemia, numerous effects on brain function have also been described. Therefore, the cognitive function (memory, attention, etc.) in healthy people, as well as improvements in the brain function of patients with brain dysfunctions, have been measured in relation to diets. Autism Spectrum Disorders (ASD), epilepsy, neurological disorders, and seizures have all been tested and displayed interesting results.
Therefore, in this review, we aim to present the current understanding of the effect of diets and their GI/GL on brain functions, including cognition and energy homeostasis regulation. Furthermore, the mechanisms involved will be described. Indeed, common mechanisms in both cognition and energy sensing can provide new insights to develop novel therapeutic approaches in diseases associated with these functions. Finally, learning about the mechanisms involved should help us understand the relationship between metabolism and neurological function. The results presented here were obtained from PubMed and Web of Science research using combinations of the following key words: brain, neurodegenerative disease, energy homeostasis, cognition, glucose load, glycemic index, and nutrition. Such an understanding is of great importance for developing novel nutritional approaches for disease treatment, in addition to generating new nutritional recommendations for healthy people.
## Effect of a glycemic index diet on brain function
Several studies have highlighted the role of a low-GI diet in insulin sensitivity, vascular system function, and weight management [bib_ref] Relevance of the glycemic index and glycemic load for body weight, diabetes,..., Vega-López [/bib_ref]. Recommendations in diabetic patients to help control their blood glucose level represent one of the most important applications of GI/GL indexes, despite some caveats in their interpretation. Besides this metabolic role, diets are used in neurodegenerative disorders, cancer, and even seizures. Such diet interventions started to gather interest following the discoveries of the influence of nutrients on brain function and notably on cognition, brain plasticity, and synaptic function, among others [bib_ref] Brain foods: The effects of nutrients on brain function, Gómez-Pinilla [/bib_ref]. More recently, studies on the effects of specific diets on brain function gave rise to new evidence on the importance of nutrition in alterations and thus improvements. Low-GI diets have been used to ameliorate cognitive function, but also improve several pathological symptoms observed in specific neurological disorders, from dementia and depression, to ASD and AD (Table 2) [bib_ref] Ketogenic diets and protective mechanisms in epilepsy, metabolic disorders, cancer, neuronal loss,..., Li [/bib_ref].
Indeed, since glucose represents the main energy source for the brain, glucose level control appears critical for maintaining normal brain activity. Furthermore, neurological disorders are often associated with changes in neuronal activity, which can be targeted by modifying the availability of energetic substrates. Nevertheless, diets can also be used in healthy people in a non-metabolic context. Indeed, different attempts to find out how to improve health through diet have been tested for decades in terms of physical activity, memory, and attention [bib_ref] Macronutrients and Mental Performance, Dye [/bib_ref]. The low-GI diet highlighted in green is taken as a reference for a healthy diet.
## Effect of the glycemic index on cognitive function in healthy people
Normal life requires a balanced diet with adapted macro-and micronutrients to maintain optimal cellular functions. Among other factors, cognition is likely to be altered by diet due to the high energy needs of the brain. Furthermore, the poor feeding habits of modern societies can very much alter normal cognitive function in healthy people. Therefore, a healthy diet can benefit healthy people and unhealthy populations. Aging, for instance, is often accompanied by cognitive decline. Therefore, determining the effects of dietary habits on cognition could be important for delaying aging-related declines. In this regard, the study of cognition in healthy elderly populations has tried to determine the role of GI/GL.
A recent study revealed that a low-GL diet contributes to maintaining a better cognitive function in the elderly. This result, along with others, support the role of GI/GL during the aging process [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref] [bib_ref] Dietary glycaemic load associated with cognitive performance in elderly subjects, Power [/bib_ref] [bib_ref] Dietary glycemic load and risk of cognitive impairment in women: Findings from..., Simeon [/bib_ref] [bib_ref] Blood glucose, diet-based glycemic load and cognitive aging among dementia-free older adults, Seetharaman [/bib_ref]. Furthermore, these studies show either a decreased risk of dementia or AD occurrence. These observations confirm a negative effect of Western diets on cognition, which has been previously documented [bib_ref] The longer-term impacts of Western diet on human cognition and the brain, Francis [/bib_ref] [bib_ref] Western diet consumption and cognitive impairment: Links to hippocampal dysfunction and obesity, Kanoski [/bib_ref] [bib_ref] Dietary patterns are associated with cognition among older people with mild cognitive..., Torres [/bib_ref]. However, the study of indicates that only people with poor glucose regulation display a positive effect of GL [bib_ref] Association between glycemic load and cognitive function in community-dwelling older adults: Results..., Garber [/bib_ref]. High fat associated with high GI has been shown to induce insulin resistance, while the same fat content with a low GI improved insulin sensitivity [bib_ref] Effects of a high-fat, low-versus high-glycemic index diet: Retardation of insulin resistance..., Schothorst [/bib_ref]. Therefore, here, it is possible that the effects observed are due to insulin sensitivity improvement. Indeed, insulin is known to participate in cognitive function [bib_ref] Brain insulin resistance: A treatment target for cognitive impairment and anhedonia in..., Hamer [/bib_ref]. Furthermore, the improvement of glucose homeostasis could improve the energy supply to the brain and thus cognitive function.
Overall, the results presented do not completely address a role in healthy people since the effects observed are those on low glycemic control people. In support of that, previous analyses performed in younger populations over the past years have failed to give a precise answer on the effect of GI/GL on cognition in healthy people. In fact, a meta-analysis and study conducted by Philipou et al. revealed discrepancies in the results obtained on the role of GI/GL in cognition in healthy persons [bib_ref] The influence of glycemic index on cognitive functioning: A systematic review of..., Philippou [/bib_ref] [bib_ref] Dietary glycaemic index and cognitive function: Prospective associations in adults of the..., Philippou [/bib_ref]. Such observations make it difficult to draw a conclusion on the relationship between GI/GL and cognition in healthy people [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
Younger populations of schoolchildren could also be targeted by diet adjustments to improve learning and memory functions. Indeed, different studies on adolescents have described a positive relationship between the GI/GL of breakfast meals and cognition through improving learning, but also attention, stress, and even mood [bib_ref] The glycaemic potency of breakfast and cognitive function in school children, Micha [/bib_ref] [bib_ref] Breakfast is associated with enhanced cognitive function in schoolchildren. An internet based..., Wesnes [/bib_ref] [bib_ref] Glycaemic index and glycaemic load of breakfast predict cognitive function and mood..., Micha [/bib_ref] [bib_ref] Breakfast glycaemic index and cognitive function in adolescent school children, Cooper [/bib_ref] [bib_ref] Breakfast glycaemic index and exercise: Combined effects on adolescents' cognition, Cooper [/bib_ref] [bib_ref] The effect of breakfast composition and energy contribution on cognitive and academic..., Edefonti [/bib_ref]. Here, schoolchildren were divided into lowor high-GI breakfasts or no breakfast at all. Then, cognitive tasks were used to test the ability of the diets to inhibit cognitive interference (Stroop test [bib_ref] The stroop color and word test, Scarpina [/bib_ref] , memory tasks, focus, learning and mood, hunger and thirst, and fatigue in adolescents.
It was previously shown that adolescents who consume breakfast exhibit improved cognitive function compared to those who do not consume breakfast [bib_ref] Breakfast consumption and cognitive function in adolescent schoolchildren, Cooper [/bib_ref]. The higher glucose supply was then expected to help maintain the better performance displayed in the cognitive tests. This result confirms previous results discussed in the elderly. Nonetheless, introducing low-and high-GI breakfast groups to a breakfast omission group provides a more precise picture on the putative mechanism involved. Thereby, both low-and high-GI breakfasts show improved adolescent cognition compared to the group without breakfast. Such a result supports the need of an energy supply for brain function. Moreover, low-GI breakfasts are more effective in these improvements than high GI breakfasts. This greater improvement related to low-GI breakfasts is also associated with lower glycemic and insulinemic responses. In fact, a high-GI breakfast group presented the lowest reaction time during the Stroop test. However, this increase in reaction was to the detriment of accuracy, which was significantly better in the low-GI group. Furthermore, this gain of accuracy was better maintained across the morning [bib_ref] Breakfast glycaemic index and cognitive function in adolescent school children, Cooper [/bib_ref]. Finally, low-GI breakfast children display better results in cognitive tests assessing their working memory, as well as attention [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
All of these studies support the role of glucose in cognition improvement since both low-and high-GI breakfasts are beneficial compared to no breakfast. However, since a low-GI breakfast gives better results than a high-GI breakfast, the role of a high circulating glucose level in such an improvement is disputed [bib_ref] Effects of differences in postprandial glycaemia on cognitive functions in healthy middle-aged..., Nilsson [/bib_ref]. Indeed, since a low-GI meal induces a lower increase in blood glucose compared to a high-GI meal, a high glucose level cannot be the main or only contributor [bib_ref] Breakfast glycaemic index and cognitive function in adolescent school children, Cooper [/bib_ref]. In a previous study, a breakfast with low GI and high GL gave better results in terms of cognitive improvement than a low-GI/low-GL breakfast [bib_ref] Glycaemic index and glycaemic load of breakfast predict cognitive function and mood..., Micha [/bib_ref]. This result indicates that the energy intake is as important as the origin, which in this study, mainly came from carbohydrates. Therefore, high-or low-GL diets differ by the amount of carbohydrate. However, low-GI/low-GL and low-GI/high-GL diets contain the same amount of total energy (twice that of low-GL diets). Furthermore, high GL induces a more important increase in the glucose level than high GI. Interestingly, the cortisol level was higher in the high-GI group, suggesting that low GI could protect against the stress response (cognitive tests in the study). Finally, the authors also reported that adolescents fed the high-GL meal felt more confident, less sluggish, and less hungry or thirsty before the tests. The low-GI fed group, on the other hand, were happier, more alert and less nervous and thirsty before the tests. Previous research supports the findings of Micha et al., showing improved alertness and decreased fatigue following a low-GI/high-GL breakfast [bib_ref] Nutritional influences on cognitive function: Mechanisms of susceptibility, Leigh Gibson [/bib_ref]. However, even so, cognitive test results are not affected by GL, since the observed effects were similar in all groups. This observation confirms the results mentioned above indicating that a high glucose level is not the main vehicle of cognition amelioration. Moreover, the glucose increase measured could suggest stimulation of the hypothalamic-pituitary-adrenal axis. In turn, this activation would result in the increased cortisol level measured. This loop would then serve as an anticipatory response to a stress [bib_ref] Glycaemic index and glycaemic load of breakfast predict cognitive function and mood..., Micha [/bib_ref]. Consequently, this decreased stress will contribute to the improved results obtained during the cognitive tests. Interestingly, a low blood glucose level after a fast alters the hypothalamic-pituitary-adrenal axis [bib_ref] Effects of nutrition on neuro-endocrine stress responses, Rohleder [/bib_ref]. Low GI could then be responsible for lowering the cortisol level by inducing a lower blood glucose level. During a learning and memory task, this effect could represent an advantage for both memorizing and recalling [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. High GI or low GI diet Animal studies, social behavior analysis High GI increase ASD phenotype while low GI improve social behavior [bib_ref] Impact of Clostridium Bacteria in Children with Autism Spectrum Disorder and Their..., Kandeel [/bib_ref] [bib_ref] Autistic Children Exhibit Distinct Plasma Amino Acid Profile, Naushad [/bib_ref] [bib_ref] Oxidative stress-related biomarkers in autism: Systematic review and meta-analyses. Free Radic, Frustaci [/bib_ref] [bib_ref] Re: Biomarkers of Environmental Toxicity and Susceptibility in Autism, Martin [/bib_ref] Depression and Anxiety High GI/GL Rate of disease in a population Increased depression and anxiety rate [bib_ref] The role of inflammation and the gut microbiome in depression and anxiety, Peirce [/bib_ref] According to studies in both adolescent and older populations, the glucose level and the source of carbohydrates are important in cognition, memory, and mood. However, healthy young populations display a strong effect, while the elderly only present a positive effect of a low-GI diet in groups with poor glucose regulation. It is possible that aging alters the sensitivity to GI/GL. Moreover, elderly studies have been conducted via questionnaires. This means that there could have been a large variety of meals, as well as nutritional habits, within the group studied. This could have influenced the results observed. Finally, it is worth noting that the elderly are cognitively healthy, but can display other medical histories and medications that could alter the real impact of diet on cognition. A longitudinal study should help determine how a specific dietary habit followed for a long time period will impact cognition in healthy populations. Additionally, the selection of healthy participants could help determine the effect on the development of age-related diseases, including cognitive deficits. To date, there is no study on the impact of a GI diet on cognition, mood, or memory in an adult population exhibiting healthy conditions. Furthermore, the role of the glucose supply in brain activity is supported by the demonstration of blood glucose level variation, depending on a mental task and emotion in younger adults (≈25 years) [bib_ref] Blood glucose changes and memory: Effects of manipulating emotionality and mental effort, Scholey [/bib_ref]. More recently, younger adults (18-23 years old) and older adults (65 to 85 years old) were tested for memory recognition after a glucose or placebo injection. Only the older group showed an improvement in cognition. However, as previously mentioned, this population displays poor glucose control [bib_ref] Glucose administration and cognitive function: Differential effects of age and effort during..., Macpherson [/bib_ref]. Overall, it is suggested that the glucose supply is still critical for maintaining a normal brain function, as shown by a decreased blood glucose level during high brain activity [bib_ref] Blood glucose changes and memory: Effects of manipulating emotionality and mental effort, Scholey [/bib_ref] [bib_ref] Cognitive Functioning Is Susceptible to the Level of blood glucose, Donohoe [/bib_ref]. In addition, a higher cognitive decline in people with poor blood glucose control or insulin resistance was previously observed [bib_ref] Impairments in glucose tolerance can have a negative impact on cognitive function:..., Lamport [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
It is worth noting that Philippou et al. failed to draw a conclusion on the consensual effect of GI on cognition in their review [bib_ref] The influence of glycemic index on cognitive functioning: A systematic review of..., Philippou [/bib_ref]. Nevertheless, in regard of the key role of the blood glucose level in cognition, GI values can still be important. In fact, in previous studies, the task or parameter analyzed (recognition, learning, memory, mood, accuracy, etc.) and the population studied (different ages, ethnicities, and health conditions) could have affected the results, making comparisons difficult. Finally, the meal composition and time of the experiment (morning vs. noon vs. evening) are also important parameters that change between studies. Despite all of these issues, the review by Philippou et al. helps us make assumptions about possible mechanisms by which GI affects cognition. First, it is suggested that the blood glucose concentration, rather than the amount provided, influences memory enhancement. Therefore, since high GI induces a transient increase in blood glucose, while a low GI leads to a more sustained increase, although lower, a low GI is more likely to induce long-term effects [bib_ref] Effects of differences in postprandial glycaemia on cognitive functions in healthy middle-aged..., Nilsson [/bib_ref] [bib_ref] Effects on cognitive performance of modulating the postprandial blood glucose profile at..., Nilsson [/bib_ref] [bib_ref] The delivery rate of dietary carbohydrates affects cognitive performance in both rats..., Benton [/bib_ref]. In support of that, it has been described that the GI enhancement of cognition appears in the post prandial phase following a meal [bib_ref] Effects of differences in postprandial glycaemia on cognitive functions in healthy middle-aged..., Nilsson [/bib_ref] [bib_ref] Effects on cognitive performance of modulating the postprandial blood glucose profile at..., Nilsson [/bib_ref] [bib_ref] The delivery rate of dietary carbohydrates affects cognitive performance in both rats..., Benton [/bib_ref]. Another putative mechanism involves insulin that is affected by GI and plays a role in cognition. Indeed, insulin resistance alters this role in cognition regulation [bib_ref] Insulin in the brain: There and back again, Banks [/bib_ref] , while low GI improves the insulin sensitivity and should thus improve cognition. Cortisol is another hormone involved in cognitive function modulation via the above-mentioned hypothalamic-pituitary-adrenal axis stimulation induced by a lower blood glucose level due to low-GI food [bib_ref] Glycaemic index and glycaemic load of breakfast predict cognitive function and mood..., Micha [/bib_ref]. Consequently, low GI is associated with a decrease in the stress response, triggering improved results in cognitive tasks [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
Altogether, it is difficult to conclude a clear effect of GI on cognition. Nevertheless, the schoolchildren studies gave the most solid results. Indeed, the results are interesting since GI and nutritional approaches could be important in childhood and during learning processes. Therefore, efforts need to be made to better understand the impact of nutrition on cognition. This knowledge would be useful for setting up new nutritional recommendations for children. Finally, the studies presented so far do not address in depth the mechanisms involved in cognition and the role of GI/GL. However, pathological studies have allowed the effects of diet on brain function to be tested with more attention.
## Low-gi diet and neurological dysfunctions
The previous studies are somewhat confusing and difficult to interpret, in addition to providing very little insight on the mechanisms involved. Interestingly, brain dysfunction research generates more information on the nutritional impact on brain function. Indeed, nutrition therapies used in brain dysfunctions have presented promising results in improvement of the pathology. Therefore, several reports on epilepsy, seizure, ASD, AD, and others have studied brain function after diet interventions. It should be noted that these studies have also permitted hypotheses on the mechanisms involved to be drawn. Therefore, such research could provide greater insights on the mechanisms in play in neurological diseases and metabolic regulation.
## Epilepsy
Epilepsy has been associated with a ketogenic diet (KD) intervention for a long time [bib_ref] The role for ketogenic diets in epilepsy and status epilepticus in adults, Williams [/bib_ref]. KD is a low-carbohydrate and high-fat diet. Additionally, because of its side effects (ketoacidosis, hyperlipidemia, and hypoglycemia), other diets have been tested more recently [fig_ref] Table 2: Example of various diets' composition for macronutrients with some examples of common... [/fig_ref]. Most of these diets are modified KDs with more carbohydrates, including a low-GI diet [bib_ref] Mechanisms and Uses of Dietary Therapy as a Treatment for Epilepsy: A..., Sadeghifar [/bib_ref]. Overall, these diets have been shown to decrease the number of seizures in pediatric patients by at least 50% [bib_ref] Efficacy, safety, and tolerability of the low glycemic index treatment in pediatric..., Muzykewicz [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. Several attempts to elucidate the mechanisms in play have led to different hypotheses involving ketone bodies, mitochondria, or gene regulation. In these diets, the decrease in carbohydrates is compensated for by a higher amount of fat, which induces a shift in nutrient utilization in favor of lipid oxidation. This high rate of lipid oxidation in turn generates ketone bodies [bib_ref] Ketogenic diets, mitochondria, and neurological diseases, Gano [/bib_ref] [bib_ref] Regulation of Fatty Acid Oxidation in Mammalian Liver, Guzmiirp [/bib_ref]. Ketogenesis usually occurs in the liver during fasting periods, but also in type 1 diabetes (due to a defect of glucose utilization because of the absence of insulin) or obesity. Neurological disorders have been associated with decreased glucose utilization. In this condition, the brain becomes dependent on ketone bodies for energy supply [bib_ref] Sugar for the brain: The role of glucose in physiological and pathological..., Mergenthaler [/bib_ref]. Ketone bodies are used as an alternative source of acetyl-coA, which is paralleled by a consumption of oxaloacetate. This stimulates the Krebs cycle and increases α-ketoglutarate. In turn, α-ketoglutarate forms high amounts of glutamate by consuming aspartate, whose level is then lowered. Finally, the glutamate produced is decarboxylated by the glutamic acid decarboxylase to produce the inhibitory neurotransmitter GABA (γ-aminobutyric acid) [bib_ref] The Neuropharmacology of the Ketogenic Diet, Hartman [/bib_ref]. Interestingly, GABA is described as an anti-seizure substance, and drug agonists targeting it are used in epilepsy [bib_ref] Low brain GABA level is associated with poor seizure control, Petroff [/bib_ref] [bib_ref] Isobolographic Analysis of Antiseizure Activity of the GABA Type A Receptor-Modulating Synthetic..., Chuang [/bib_ref] [bib_ref] Mechanisms of action of currently used antiseizure drugs, Sills [/bib_ref]. Among these molecules, benzodiazepines enhance GABA's action [bib_ref] GABAergic mechanisms in epilepsy, Treiman [/bib_ref]. In support of this mechanism, children treated with low-carbohydrate diets (low GI) present high levels of GABA in their cerebrospinal fluid [bib_ref] The ketogenic diet influences the levels of excitatory and inhibitory amino acids..., Dahlin [/bib_ref].
Another hypothesized mechanism of ketone bodies is that they directly enter mitochondria and the tricarboxylic acid cycle (TCA) to be oxidized. In turn, the stimulated oxidative metabolism inhibits phosphofructokinase 1 and glycolysis. This direct metabolization of ketone bodies decreases the ATP produced in glycolysis that will open the ATP sensitive potassium channels (K-ATP) and decrease neuronal activity [bib_ref] Ketogenic diet metabolites reduce firing in central neurons by opening KATP channels, Ma [/bib_ref]. In support of that, a genetic model of drosophila exhibiting seizure-like activity upon mechanical stimulation showed a reduced number of seizures when given ketone bodies. Moreover, blocking the K ATP channels or adding a GABA antagonist has been shown to partially reverse the effect of ketone bodies [bib_ref] The ketogenic diet metabolite beta-hydroxybutyrate (β-HB) reduces incidence of seizure-like activity (SLA)..., Li [/bib_ref]. This partial effect suggests that other mechanisms are also involved. For instance, other studies indicate a blockade of vesicular glutamate transporter (VGLUT) transfer to the synapse by ketone bodies. Such a blockade will decrease the excitatory glutamate neurotransmitters and thus neuronal activity [bib_ref] Vesicular neurotransmitter transporter: Bioenergetics and regulation of glutamate transport, Omote [/bib_ref] [bib_ref] Metabolic Control of Vesicular Glutamate Transport and Release, Juge [/bib_ref].
Mitochondria have also been linked to the anti-seizure effect of KD. Here, ketone bodies increase mitochondrial respiration and NADH oxidation, inhibit reactive oxygen species (ROS) production, and enhance ATP production [bib_ref] Mitochondrial Permeability Transition: New Findings and Persisting Uncertainties, Izzo [/bib_ref]. All of these should prevent the mitochondrial permeability transition (mPT), which ultimately leads to cell death [bib_ref] Ketone bodies mediate antiseizure effects through mitochondrial permeability transition, Kim [/bib_ref]. In support of that, mice with recurrent epileptic seizures show an increased threshold of mPT. This anti-mPT effect depends on cyclophilin D modulation (part of the mPT). Furthermore, learning and memory and long-term potentiation are decreased in these mice, suggesting a role in cognition. However, this theory is contradictory to previous activity decreasing ATP production. Therefore, more research is needed to completely determine the role of ketone bodies in ATP production. Nevertheless, mitochondrial activity has been linked to neuronal function and diseases, while KD is known to improve mitochondrial activity [bib_ref] Reduced mitochondrial reactive oxygen species production in peripheral nerves of mice fed..., Cooper [/bib_ref]. Therefore, KD induces decreased mitochondrial ROS production compared to a normal chow, via changes in the gene expression of the oxidative pathway. Another possibility is that the increased NAD/NADH ratio induced by ketone bodies dampens ROS production [bib_ref] Metabolic dysfunction and oxidative stress in epilepsy, Pearson-Smith [/bib_ref]. Ketone bodies also improve the consumption of O2 within the respiratory chain. By doing so, KD decreases the rate of ROS production by the mitochondria. Since seizure is associated with oxidative stress, KD or modified KD could diminish the seizure occurrence in epileptic patients [bib_ref] Ketogenic diet regulates the antioxidant catalase via the transcription factor PPARγ2, Knowles [/bib_ref]. Moreover, besides having a direct effect on ROS production, KD also stimulates the antioxidant protein catalase, whose expression is stimulated by the activated peroxisome proliferator activated receptor γ2 (PPARγ2) transcription factor [bib_ref] Ketogenic diet regulates the antioxidant catalase via the transcription factor PPARγ2, Knowles [/bib_ref]. Here, PPAR activation occurs after histone hyper-acetylation following the inhibition of histone deacetylases (epigenetic regulation) [bib_ref] Ketone Bodies as Anti-Seizure Agents, Simeone [/bib_ref] [bib_ref] Regulation of brain PPARgamma2 contributes to ketogenic diet anti-seizure efficacy, Simeone [/bib_ref]. In turn, PPAR upregulates antioxidant genes and downregulates pro-inflammatory genes (NFKappaB, cyclooxygenase 2, and iNOS) [bib_ref] Ketogenic diet-induced peroxisome proliferator-activated receptor-γ activation decreases neuroinflammation in the mouse hippocampus..., Jeong [/bib_ref]. Histone deacetylase inhibitors are used as anti-inflammatory and anti-epileptogenic molecules [bib_ref] Histone deacetylase inhibitors: An attractive therapeutic strategy against breast cancer, Damaskos [/bib_ref]. Interestingly, ketone bodies are shown to inhibit histone deacetylase, although the precise mechanism still needs to be determined [bib_ref] Suppression of oxidative stress by β-hydroxybutyrate, an endogenous histone deacetylase inhibitor, Shimazu [/bib_ref].
Finally, Rahman et al. hypothesized that ketone body neuroprotection could rely on the G-protein coupled receptor GPR109 recently identified, which is activated by ketone bodies and found in the brain. Moreover, mice fed a KD or infused with ketone bodies showed a decrease of ischemic infarcts dependent on GPR109. Furthermore, the authors demonstrated that the activation of GPR109 by ketone bodies occurs in infiltrated monocytes and macrophages. In turn, these cells produce prostaglandins and induce an anti-inflammatory response that reduces seizures [bib_ref] Immunity and inflammation in epilepsy. Cold Spring Harb, Vezzani [/bib_ref] [bib_ref] The b-hydroxybutyrate receptor HCA 2 activates a neuroprotective subset of macrophages, Rahman [/bib_ref]. Finally, the anti-inflammatory role of ketone bodies inhibits NLRP3 inflammasome assembly through the blockade of K+ efflux. However, the precise mechanism remains to be elucidated [bib_ref] The ketone metabolite β-hydroxybutyrate blocks NLRP3 inflammasome-mediated inflammatory disease, Youm [/bib_ref].
## Stroke
Vegetarian diets (low GI and GL) have been described to lower the risk of stroke [bib_ref] Lower risk of stroke with a vegetarian diet, Spence [/bib_ref] [bib_ref] Overall glycemic index and glycemic load of vegan diets in relation to..., Waldmann [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. Therefore, a relationship between GI/GL and stroke likely exists. On the other hand, a high-fructose diet that should have a high GI worsened post ischemic brain injury in rats [bib_ref] High fructose diet-induced obesity worsens post-ischemic brain injury in the hippocampus of..., Pérez-Corredor [/bib_ref]. In this study, the authors highlighted an important effect in the hippocampus with increased inflammation while neuronal plasticity was decreased, in parallel with neuronal loss. These changes lowered the neuronal performances during ischemic recovery compared to normal chow fed mice. These results support a role of the diet during the recovery period following a stroke. Here, the glucose level is also at play during the acute stage of a stroke and the recovery period. Therefore, diabetes is a risk factor for stroke [bib_ref] Reducing the risk of stroke in type 2 diabetes: Pathophysiological and therapeutic..., Sander [/bib_ref] [bib_ref] Interactive effect of acute and chronic glycemic indexes for severity in acute..., Lee [/bib_ref]. The oxidative stress associated with hyperglycemia is expected to be the mechanism involved in the poor outcome following a stroke. This suggests that better glucose level control could help decrease the risk of stroke [bib_ref] Opposing effects of glucose on stroke and reperfusion injury: Acidosis, oxidative stress,..., Robbins [/bib_ref]. Diet intervention studies have provided interesting results in this regard. Therefore, a high-GL diet is associated with a poor outcome in patients with acute ischemic stroke [bib_ref] High dietary glycemic load is associated with poor functional outcome in patients..., Song [/bib_ref]. Recently, Song et al. found that, in diabetic patients, this poor outcome does not depend on diabetes, but only on the diet's effect on glycemic variation. Furthermore, the authors suggested that chronic hyperglycemia is the main anticipated cause. Chronic hyperglycemia induced by a high-GL diet was previously shown to induce a poor outcome after a stroke [bib_ref] Chronic hyperglycemia is related to poor functional outcome after acute ischemic stroke, Luitse [/bib_ref] [bib_ref] Prestroke glycemic control is associated with the functional outcome in acute ischemic..., Kamouchi [/bib_ref]. One possible explanation for this is the cerebral hypo-perfusion or edema associated with hyperglycemia, which leads to poor recovery from stroke [bib_ref] Glycemic Variability and Acute Ischemic Stroke: The Missing Link? Transl, González-Moreno [/bib_ref] [bib_ref] Perfusion Deficit Parallels Exacerbation of Cerebral IschemiaIReperfusion Injury in Hyperglycemic Rats, Quast [/bib_ref]. Moreover, stroke is often linked to cytotoxicity, whose association with hyperglycemia dampens the recovery [bib_ref] Hyperglycemia is associated with more severe cytotoxic injury after stroke, Bevers [/bib_ref]. In addition, mitochondrial dysfunction due to lactate production from glucose metabolism has also been suggested. Here, lactate acidifies the intracellular environment and triggers mitochondrial dysfunction [bib_ref] Effects of Glucose and PaO 2 Modulation on Cortical Intracellular Acidosis, NADH..., Anderson [/bib_ref]. A rapid increase in blood glucose also leads to oxidative stress and endothelial dysfunction observed in diabetes [bib_ref] Oscillating glucose is more deleterious to endothelial function and oxidative stress than..., Ceriello [/bib_ref]. Furthermore, high increases followed by periods of low levels are more deleterious than a continuous rise. Interestingly, high GI/GL increases glucose transiently, and thus has a negative effect on stroke outcome. Indeed, a dysfunctional endothelium has been described as a negative factor for acute ischemic stroke outcome [bib_ref] Brachial arterial flow mediated dilation in acute ischemic stroke, Santos-García [/bib_ref]. Moreover, high-GI/GL diets induce insulin resistance [bib_ref] Effects of a high-fat, low-versus high-glycemic index diet: Retardation of insulin resistance..., Schothorst [/bib_ref] that will increase serum levels of fibrinogen and the von Willebrand factor (endothelial function), thus increasing the risk of stroke [bib_ref] Relationships Between Fibrinogen and Insulin Resistance, Raynaud [/bib_ref]. Therefore, high GI/GL should induce hypercoagulability and an increased risk of thrombosis. This phenotype is paralleled by an increase in small vessel diseases that could impair stroke outcome [bib_ref] High dietary glycemic load was associated with the presence and burden of..., Song [/bib_ref]. Finally, in a cohort of Chinese women followed for 10 years, a positive association between GI and the risk of stroke was observed, supporting the previous hypothesis [bib_ref] Dietary glycemic index, glycemic load, and refined carbohydrates are associated with risk..., Yu [/bib_ref]. Furthermore, the low-GI Mediterranean diet is associated with a reduction in the risk of stroke [bib_ref] Nutrition and risk of stroke, Spence [/bib_ref]. Finally, Lim et al. studied the role of diets in stroke recovery and cognition impairments. In their study, the authors described that glycemic variability with hyperglycemic episodes is detrimental to cognitive function recovery. Therefore, a diet avoiding amplitude oscillations for glycemia is more suitable for both vascular and cognitive function recovery after a stroke [bib_ref] Effects of glycemic variability and hyperglycemia in acute ischemic stroke on post-stroke..., Lim [/bib_ref].
Overall, evidence indicates that high glucose increases induced by diet impact both the risk and outcome of stroke. The mechanisms in play are related to mitochondria and inflammation, as well as hypercoagulation. Moreover, chronic and rapid hyperglycemia following a meal is deleterious for both the risk and outcome. Therefore, a low-GI diet that induces a sustained elevation in blood glucose should be beneficial for recovery after a stroke, and decreases the risk of occurrence by improving cardiovascular function.
## Alzheimer's disease (ad)
Diets are widely use in AD to help delay or slow the development of the disease [bib_ref] The Role of Nutrition for the Aging Population: Implications for Cognition and..., Power [/bib_ref] [bib_ref] Alzheimer's disease: Food for thought, Otaegui-Arrazola [/bib_ref]. Additionally, the relationship between AD and metabolic disorders (insulin resistance) means that this disease is considered as type III diabetes [bib_ref] Metabolic Syndrome as a Risk Factor for Alzheimer's Disease: Is Aβ a..., Campos-Peña [/bib_ref]. Therefore, nutritional approaches to studying AD have provided insights into the brain function in the pathology related to diet changes. Recently, it was reported that a high-GI diet increases the accumulation of amyloid β (Aβ) in brains of the elderly, which is a marker of AD, as well as a risk factor for the onset of the disease. In addition, these individuals showed a cognitive decline, and PET scan imaging demonstrated an increased association between Aβ accumulation and a high-GI diet [bib_ref] A high-glycemic diet is associated with cerebral amyloid burden in cognitively normal..., Taylor [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. This interaction being independent of all other factors (age, sex, and education) indicates that high GI is highly involved in Aβ accumulation in aging populations and represents an increased risk factor for AD. In support of a key role of diet in AD onset, high-fat diet (HFD)-induced insulin resistance increases brain Aβ accumulation. However, mice deleted for IRS2 (insulin signaling) become insulin resistant, but do not accumulate Aβ, unless fed an HFD. Others have also shown that Aβ accumulation is accompanied by pTau aggregation (characteristic of AD) in an insulin-resistant AD mice model. Finally, oxidative stress and inflammation have been described in this mice model.
Overall, these results indicate that the Aβ burden is related to the diet, confirming that nutritional adjustments could help prevent AD onset and slow its progression [bib_ref] Differential effects of diet-and genetically-induced brain insulin resistance on amyloid pathology in..., Wakabayashi [/bib_ref]. In support of this, a recent review describes that a healthy diet can decrease the risk of AD by lowering oxidative stress and dampening Aβ accumulation [bib_ref] Dietary pattern in relation to the risk of Alzheimer's disease: A systematic..., Samadi [/bib_ref]. Moreover, this review discusses the anti-inflammatory effect of a healthy diet. In AD, inflammation is caused by Aβ accumulation that stimulates the recruitment of microglia and astrocytes parallel to interferon gamma (IFNγ), interleukin 1β (IL1β), and tumor necrosis factor α (TNF α) secretion [bib_ref] Contribution of inflammatory processes to Alzheimer's disease: Molecular mechanisms, Sastre [/bib_ref] [bib_ref] The NLRP3 inflammasome in alzheimer's disease, Tan [/bib_ref].
AD is characterized by a decreased glucose uptake and utilization by brain cells. However, the brains of AD individuals can still use ketone bodies [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref] [bib_ref] Lower brain 18F-fluorodeoxyglucose uptake but normal 11C-acetoacetate metabolism in mild Alzheimer's disease..., Castellano [/bib_ref]. Therefore, KD could be used in AD to slow progress of the disease or to delay the cognitive deficit. A medium-chain triglyceride diet (MCT diet) is a diet less restrictive in carbohydrates, but still ketogenic, and is used in AD [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref]. The low content of carbohydrates forces the organism to use lipid oxidation as the energy source, which also produces ketone bodies. Therefore, such diets have a low impact on glycemia, making them low-GI diets. Brain energy metabolism and glucose uptake decreases are accompanied by mitochondrial dysfunction in AD [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref]. Brain cells increased the utilization of ketone bodies during brain energetic deficiency, leading to a new "neuroketotherapeutic" strategy to compensate for the lack of glucose as an energy source [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref]. Besides this energetic role, ketone bodies are also involved in neurotransmission and a reduction in oxidative stress and inflammation relevant to AD. Ketone bodies involve the mitochondria and their functional changes. As AD is associated with mitochondrial dysfunctions [bib_ref] The Alzheimer's disease mitochondrial cascade hypothesis: Progress and perspectives, Swerdlow [/bib_ref] , ketone bodies could thus improve the disease phenotype. KD has also been shown to increase the number of mitochondria in the hippocampus, which could contribute to the improvement of AD [bib_ref] Mitochondrial biogenesis in the anticonvulsant mechanism of the ketogenic diet, Bough [/bib_ref]. Furthermore, since ketone bodies produce fewer ROS than glucose, they can participate in a decrease in the oxidative stress in AD [bib_ref] Cerebral metabolic adaptation and ketone metabolism after brain injury, Prins [/bib_ref] [bib_ref] β-Hydroxybutyrate in the Brain: One Molecule, Multiple Mechanisms, Achanta [/bib_ref]. This decrease in ROS production could be induced by stimulating the expression of uncoupling proteins (UCP), as shown previously [bib_ref] The Ketogenic Diet Increases Mitochondrial Uncoupling Protein Levels and Activity, Sullivan [/bib_ref] [bib_ref] Mitochondrial uncoupling and longevity-A role for mitokines?, Klaus [/bib_ref]. Another possible mechanism is that reducing glutamate transport and improving GABA activity decreases the excitability of neurons and thus ROS production (see Section 3.1). In addition, KD was shown to upregulate antioxidant proteins (MnSOD, Glutathione, and Nrf2) [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref]. Finally, KD inhibition of histone deacetylase can allow the expression of proteins improving cellular homeostasis and function (brain-derived neurotrophic factor (BDNF)), and in turn cognitive deficit in AD patients [bib_ref] The role of histone acetylation in memory formation and cognitive impairments, Peixoto [/bib_ref] [bib_ref] HDAC3 negatively regulates spatial memory in a mouse model of Alzheimer's disease, Zhu [/bib_ref] [bib_ref] Brain-Derived Neurotrophic Factor/TrkB Signaling in Memory Processes, Yamada [/bib_ref] [bib_ref] 3-Hydroxybutyrate regulates energy metabolism and induces BDNF expression in cerebral cortical neurons, Marosi [/bib_ref] [bib_ref] Differential regulation of Bdnf expression in cortical neurons by class-selective histone deacetylase..., Koppel [/bib_ref]. In support of the benefit of KD diets, mice models of AD also show decreased Aβ accumulation in the brain when fed a KD [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref] , while in humans, a medium-chain triglyceride (MCT) diet has been shown to result in an encouraging improved memory or at least the stabilization of cognitive function in AD individuals [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref]. The direct effect of a KD in humans has been summarized by Taylor et al., who present an improvement of almost all the memory and verbal communication tests in AD patients under a KD [bib_ref] Dietary neuroketotherapeutics for Alzheimer's disease: An evidence update and the potential role..., Taylor [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
Altogether, these results support an improvement of AD pathology by decreasing the carbohydrate supply and thus lowering the GI of the meals. The benefits depend on ketone bodies that help maintain neuronal activity, decrease oxidative stress, and stimulate gene regulation. However, a KD displays important side effects. Therefore, more research should be conducted to help better understand the mechanisms at play, and eventually develop a therapeutic approach targeting these mechanisms. Moreover, other low-GI diets could have similar beneficial effects to KDs, and thus must be tested. Overall, although there is solid evidence for ketone body involvement, the diets used in AD have several other effects that could be involved. Therefore, conducting more research to understand the role of low GI on the AD phenotype could help develop a diet adapted to the disease.
For instance, the Mediterranean diet (MD) is also protective against cognitive decline in AD [bib_ref] Alzheimer's disease: Food for thought, Otaegui-Arrazola [/bib_ref]. MD is characterized by: a high intake of vegetables, legumes, fruits and cereals, and extra virgin oil (unsaturated fatty acids); low intake of saturated fats and meat; a moderate intake of fish; and a low to moderate intake of dairy products and wine during meals [bib_ref] Mediterranean and MIND diets containing olive biophenols reduces the prevalence of Alzheimer's..., Omar [/bib_ref] [bib_ref] Adherence to a Mediterranean Diet and Survival in a Greek Population, Trichopoulou [/bib_ref]. MD is low in carbohydrates, while high in fibers, and so can be classified as a low-GI diet [fig_ref] Table 2: Example of various diets' composition for macronutrients with some examples of common... [/fig_ref]. MD is known to have numerous health benefits in cardiovascular or metabolic diseases, and also in cognition (lower decline) and in reducing the risk of dementia or AD [bib_ref] Mediterranean diet, cardiovascular disease and mortality in diabetes: A systematic review and..., Becerra-Tomás [/bib_ref] [bib_ref] Mediterranean diet, cognitive function, and dementia: A systematic review of the evidence, Petersson [/bib_ref]. Furthermore, people with mild cognitive impairment fed an MD show a decreased risk of AD onset, and improved memory, delayed recall, and global cognitive function [bib_ref] Mediterranean and MIND diets containing olive biophenols reduces the prevalence of Alzheimer's..., Omar [/bib_ref]. Different studies suggest antioxidant, anti-inflammatory, or cognitive function enhancement, depending on the nutrient components. For instance, olive oil, as a main source of fats enriched in omega-3 and phenolic acid, is considered to be a main factor [bib_ref] Mediterranean and MIND diets containing olive biophenols reduces the prevalence of Alzheimer's..., Omar [/bib_ref]. Olive oil decreases the glycemic response to a high-GI meal [bib_ref] Extra-virgin olive oil reduces glycemic response to a high-glycemic index meal in..., Bozzetto [/bib_ref]. As such, olive oil helps lower the GI of a meal, and thus decreases the glycemic increase induced by a meal that could participate in the effects observed. In addition, since the carbohydrate content of MD is low or non-digestible, ketogenesis is expected and repeats the action described above. Accordingly, a modified Mediterranean-ketogenic diet has been described to be associated with changes in AD biomarkers in cerebro spinal fluid (CSF), suggesting that ketone bodies could be involved. In addition, the observed changes in the brain could result from gut microbiota alterations, suggesting that nutrient digestion and/or absorption are important steps in AD onset and cognition. These results reinforce the relationship between nutrition, metabolism, and AD and cognitive function. Microbiota have been shown to interact with brain function through the metabolism of dietary fibers in certain bacteria that produce propionate or butyrate (short chain fatty acids (SCFA)). In turn, these SCFA exert brain effects via histone deacetylase, transcription factors, or antioxidant regulations. All of these effects will then provide neuroprotection and therefore protect against neurodegenerative disorders [bib_ref] Effect of ultra-processed diet on gut microbiota and thus its role in..., Leo [/bib_ref]. On the other hand, Western diets containing processed food and carbohydrates decrease the number of bacteria producing SCFA. By doing so, these diets have deleterious effects on brain and cognition. Interestingly, highly processed food will most likely have a high GI, while dietary fibers are low-GI carbohydrates, indicating that low-GI foods are protective against neurodegenerative diseases.
The gut-brain axis and nutrition have also been shown to participate in the pathophysiology of Autism Spectrum Disorder (ASD) [bib_ref] Autism and nutrition: The role of the gut-brain axis, Van De Sande [/bib_ref]. Moreover, Parkinson's disease (PD) is improved by a KD. Altogether, nutrition and GI could have an impact on other neurological conditions.
## Others: dementia, depression, mental health, etc.
Metabolism is often associated with pathological brain conditions. Therefore, the AD and PD risk increases with malnutrition and insulin resistance, while diet control is protective. Furthermore, insulin has been shown to increase dopamine transporter mRNA levels in the substantia nigra [bib_ref] Insulin and Neurodegenerative Disease: Shared and Specific Mechanisms, Craft [/bib_ref]. Therefore, it is assumed that a high-GI/GL diet could prevent PD by inducing high insulin secretion [bib_ref] Dietary glycemic index is inversely associated with the risk of Parkinson's disease:..., Murakami [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. However, the study suggesting a role for insulin is controversial, since high GI induces insulin resistance, while low GI improves insulin sensitivity [bib_ref] Effects of a high-fat, low-versus high-glycemic index diet: Retardation of insulin resistance..., Schothorst [/bib_ref]. Nonetheless, whilst the population studied presented a lower rate of PD while consuming high-GI food such as rice, it is likely that other dietary factors could be involved. Indeed, the Japanese diet is considered a healthy diet contributing to the expanded life expectancy observed in Japan. Therefore, cardiovascular-related death and neurodegenerative disease rates are amongst the lowest in the world [bib_ref] Mediterranean and Argentinean diets and their potential roles in neurodegenerative diseases, Dohrmann [/bib_ref]. The Japanese diet is characterized by raw ingredients used in meal preparation. The diet is low in fat and calories because most of the foods used are vegetables, fish, meat, and rice, providing an excellent nutritional balance. This diet is low in red and processed meats, whole milk, refined grains, sweet drinks or alcohol, candy, and sweets, but enriched in fruits, vegetables, stevia sweeteners, and whole grains and fish products [fig_ref] Table 2: Example of various diets' composition for macronutrients with some examples of common... [/fig_ref]. Overall, the Japanese diet is expected to be low GI (stevia has a GI/GL = 0) [bib_ref] Dietary glycemic index is inversely associated with the risk of Parkinson's disease:..., Murakami [/bib_ref]. Therefore, the observed decrease in PD in the study of Murakami et al. could be due to an improved insulin sensitivity, rather than increased insulin levels. It was also described that a high-fat (high-GI) diet is responsible for a decreased number of dopaminergic neurons in the substantia nigra due to a reduced PPAR and inflammation. Therefore, a low-fat diet such as the Japanese diet could prevent these PPAR and dopaminergic neuron decreases and protect against PD [bib_ref] High fat diet suppresses peroxisome proliferator-activated receptors and reduces dopaminergic neurons in..., Kao [/bib_ref]. The MD is also associated with a decreased risk of PD. Here, the microbiota changes induced by the diet are interesting since they are related to dietary fibers (lowering the GI) as a carbohydrate source in the diet. In addition, it strengthens the importance of the gut-brain axis in brain function and disease. Dietary fibers are known to stimulate the production of SCFA that could be part of the mechanism of protection against PD in the MD. Indeed, SCFA can improve insulin sensitivity, reduce inflammation, and stimulate brain-derived neurotrophic factor (BDNF) production, helping to protect against PD. Another possible mechanism occurs through the antioxidant-rich content of the MD. It should be noted that some of the antioxidant products in the diet can also stimulate SCFA production, reinforcing the role of these molecules in brain function and pathology [bib_ref] Diet in Parkinson's Disease: Critical Role for the Microbiome, Jackson [/bib_ref]. In support of that, the SCFA concentration in the feces of PD patients is decreased compared to control individuals [bib_ref] Short chain fatty acids and gut microbiota differ between patients with Parkinson's..., Unger [/bib_ref]. However, Shin et al. also observed an increase of plasma SCFA correlated with the PD severity [bib_ref] Plasma Short-Chain Fatty Acids in Patients with Parkinson's Disease, Shin [/bib_ref]. This result suggests putative SCFA leakage from the intestinal lumen into the bloodstream. Nevertheless, further studies are needed to clearly address the impact of SCFA on PD, and more generally, the impact of diets.
Depression and anxiety are other brain diseases that exhibit a relationship with energy homeostasis [bib_ref] Type 2 diabetes and mental disorders; A plausible link with inflammation, Hajebrahimi [/bib_ref] [bib_ref] Circadian cortisol profiles, anxiety and depressive symptomatology, and body mass index in..., Pervanidou [/bib_ref]. Furthermore, depression has been studied in relation to GI/GL [bib_ref] Glycemic index, glycemic load, and depression: A systematic review and meta-analysis, Salari-Moghaddam [/bib_ref]. Although the results are inconsistent, it seems that high-GI/GL diets increase the risk of depression, as well as aggravate the score of the disease [bib_ref] Glycemic index, glycemic load, and depression: A systematic review and meta-analysis, Salari-Moghaddam [/bib_ref] [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref]. Rodent studies showed that HFD impaired 5-HT neurotransmission, which increases anxiety behavior. The same group also reported decreased anxiety following Connexin 43 downregulation or phosphorylation, or treatment with metformin (activator of AMPK) [bib_ref] Metformin promotes anxiolytic and antidepressant-like responses in insulin-resistant mice by decreasing circulating..., Zemdegs [/bib_ref] [bib_ref] Themed Section: Updating Neuropathology and Neuropharmacology of Monoaminergic Systems High-fat diet-induced metabolic..., Zemdegs [/bib_ref] [bib_ref] Attenuated levels of hippocampal connexin 43 and its phosphorylation correlate with antidepressant-and..., Quesseveur [/bib_ref]. In this context, metformin triggers a decrease in circulating branch chain amino acids (BCAA) that in turn contribute to the anxiolytic and antidepressant effect. BCAA have been previously linked to glutamate transport between intracellular to extracellular medium [bib_ref] Effects of Branched Chain Amino Acids, Pyruvate, or Ketone Bodies on the..., Palaiologos [/bib_ref]. It was also observed that a metformin decrease of BCAA is due to the inhibition of ketone-derived BCAA production, suggesting that ketone bodies could be involved in depression and anxiety. Interestingly, in addition to ketones, anxiety and depression have been associated with oxidative stress, inflammation, and gut-brain communication [bib_ref] The role of inflammation and the gut microbiome in depression and anxiety, Peirce [/bib_ref] [bib_ref] Environmental Noise-Induced Effects on Stress Hormones, Oxidative Stress, and Vascular Dysfunction: Key..., Hahad [/bib_ref]. Moreover, drugs have been developed to target glutamate signaling and its recycling in the synapse [bib_ref] Targeting the glutamatergic system to treat major depressive disorder: Rationale and progress..., Mathews [/bib_ref]. Therefore, it is likely that diet interventions could involve similar processes to those described above.
Autism Spectrum Disorder (ASD) is a neurological condition associated with behavioral and social interaction defects and that shows energy homeostasis dysregulation [bib_ref] Hertz-Picciotto, I. Maternal metabolic conditions and risk for autism and other neurodevelopmental..., Krakowiak [/bib_ref]. In fact, type 2 diabetes and obesity during pregnancy are risk factors for ASD in offspring [bib_ref] Hertz-Picciotto, I. Maternal metabolic conditions and risk for autism and other neurodevelopmental..., Krakowiak [/bib_ref] [bib_ref] Maternal early life factors associated with hormone levels and the risk of..., Lyall [/bib_ref]. Chronic inflammation observed in metabolic disorders and immune system activation that occurs during pregnancy appear to be key in the risk of ASD [bib_ref] Neuroglial activation and neuroinflammation in the brain of patients with autism, Vargas [/bib_ref] [bib_ref] Immune involvement in schizophrenia and autism: Etiology, pathology and animal models, Patterson [/bib_ref] [bib_ref] Immune system gene dysregulation in autism and schizophrenia, Michel [/bib_ref]. It is noteworthy that chronic inflammation can be decreased by a low-GI diet in obese subjects [bib_ref] A low-glycemic load diet reduces serum C-reactive protein and modestly increases adiponectin..., Neuhouser [/bib_ref]. On the other hand, a high-GI diet increases advanced glycation end products (AGEs) that are involved in inflammation during obesity and diabetes. Indeed, the AGE activation of specific receptors results in C Reactive Protein (CRP) production and oxidative stress [bib_ref] Glycation-altered proteolysis as a pathobiologic mechanism that links dietary glycemic index, aging,..., Uchiki [/bib_ref] [bib_ref] Reactive metabolites and AGE/RAGE-mediated cellular dysfunction affect the aging process-A mini-review, Fleming [/bib_ref]. In support of this possible mechanism, Currais et al. demonstrated that mice offspring from high-GI-fed parents displayed increased brain inflammation, reduced neurogenesis, and characteristic ASD behaviors [bib_ref] Dietary glycemic index modulates the behavioral and biochemical abnormalities associated with autism..., Currais [/bib_ref]. Dietary strategies have triggered improvement of the ASD phenotype after birth. For instance, KD given to a mice model of ASD improved the social behavior and decreased repetitive behaviors [bib_ref] Ketogenic Diet Improves Core Symptoms of Autism in BTBR Mice, Ruskin [/bib_ref]. In another attempt, gluten-free foods decreased the inflammatory grade and improved the ASD phenotype. However, only some of the subjects showed improvement, while others failed to observe a benefit [bib_ref] The Role of Gluten in Autism, Sumathi [/bib_ref] [bib_ref] Nutritional interventions for autism spectrum disorder, Karhu [/bib_ref]. These studies, which used different diets, suggest that a specific nutrient or compound could be at play. Furthermore, inflammation is involved in the onset of the disease; however, other brain energetic alterations could also be present, but remain to be determined. Therefore, more research on both the understanding of ASD brain alterations and diet interventions to determine the role of specific nutrients or metabolic products needs to be conducted.
Microbiota are also associated with ASD behaviors [bib_ref] Impact of Clostridium Bacteria in Children with Autism Spectrum Disorder and Their..., Kandeel [/bib_ref] [bib_ref] Autism and nutrition: The role of the gut-brain axis, Van De Sande [/bib_ref] [bib_ref] Dietary Patterns Impact Temporal Dynamics of Fecal Microbiota Composition in Children with, Berding [/bib_ref]. Even if the role of gut microbiota is poorly understood, amino acid metabolism and inflammation are possible mechanisms participating in the phenotype observed. A low-GI diet induces changes in microbiota that are associated with ASD improvement. Therefore, the participation of a product with low-GI carbohydrates could be important. Finally, the ASD mice model shows decreased blood levels of methionine, also described in humans [bib_ref] Autistic Children Exhibit Distinct Plasma Amino Acid Profile, Naushad [/bib_ref] [bib_ref] Oxidative stress-related biomarkers in autism: Systematic review and meta-analyses. Free Radic, Frustaci [/bib_ref] [bib_ref] Re: Biomarkers of Environmental Toxicity and Susceptibility in Autism, Martin [/bib_ref]. Interestingly, a high-GI diet also decreases the methionine levels [bib_ref] Dietary glycemic index modulates the behavioral and biochemical abnormalities associated with autism..., Currais [/bib_ref]. Methionine is a precursor for DNA methylation and gene regulation [bib_ref] Genetics and epigenetics of autism: A Review, Waye [/bib_ref]. In comparison, a low-GI diet maintains higher levels of methionine, suggesting that diet could improve ASD through epigenetic regulations. In line with the role of microbiota and low GI, sulforaphane, produced from low-GI vegetables (cauliflower and broccoli), has been identified as a putative treatment in ASD. Indeed, sulforaphane given to autistic children improved their behavioral phenotype. Moreover, although the mechanism of action is not known, sulforaphane has been described to modulate oxidative stress, methylation, or apoptosis, and to be a potent neuroprotective molecule [bib_ref] Dietary Phytochemicals as Neurotherapeutics for Autism Spectrum Disorder: Plausible Mechanism and Evidence, Bhandari [/bib_ref] [bib_ref] Biomarker Exploration in Human Peripheral Blood Mononuclear Cells for Monitoring Sulforaphane Treatment..., Liu [/bib_ref] [bib_ref] Sulforaphane and iberin are potent epigenetic modulators of histone acetylation and methylation..., Mitsiogianni [/bib_ref] [bib_ref] The neuroprotective mechanisms and effects of sulforaphane, Klomparens [/bib_ref] [bib_ref] Sulforaphane treatment of autism spectrum disorder (ASD), Singh [/bib_ref]. Overall, the identification of this molecule supports a beneficial role of a low-GI diet in ASD. It is noteworthy that, similar to other neurological conditions, ASD improvement through the diet is linked to oxidative stress, gene regulation, and inflammation. Moreover, a sulforaphane extract from vegetables is a promising molecule for the treatment of ASD. Further, other molecules produced or present in low-GI foods could also exist and participate in the beneficial effects, and would thus be useful in other neurological diseases [fig_ref] Table 3: Summary of human studies on diet effects on cognitive function in different... [/fig_ref].
Intriguingly, oxidative stress, inflammation, ketone bodies, the gut-brain axis, microbiota, glutamate metabolism, and neurogenesis are all involved in the neurological conditions described. All of them can also be affected by the diet and especially the carbohydrate content and source. Therefore, low-GI diets are likely to improve neurological disorders, but may also be of interest in physiological conditions, since they help improve cognition and neuronal activity. A more generalized use of a low-GI/GL diet could help prevent or delay the onset of neurological disorders, while it could also help during the growing up period and in learning performances in childhood.
Although most of the mechanisms described appear to be common to the diseases described and are linked to diet intervention changes, no exact mechanisms and molecules involved in these beneficial effects have been precisely identified. Nevertheless, a lot of common pathways indicate a role of ketone bodies and SCFA. Therefore, more studies to determine diet involvement in the production of these molecules should be conducted to determine precise diet recommendations. Considering this, neurological disorders are likely to be successfully targeted by nutritional therapies.
## Glycemic index and brain regulation of energy homeostasis
GI/GL values predict the impact of food on the metabolic response, such as blood glucose and insulin levels. Therefore, a high GI/GL will induce a rapid and high increase in blood glucose and insulin, while a low GI/GL will induce a slow and limited increase of blood glucose that slightly affects insulin levels. Due to these effects, diets based on low-GI foods have been developed to help control weight and prevent cardiometabolic disorders [bib_ref] Relevance of the glycemic index and glycemic load for body weight, diabetes,..., Vega-López [/bib_ref] [bib_ref] A low-glycemic index diet combined with exercise reduces insulin resistance, postprandial hyperinsulinemia,..., Solomon [/bib_ref]. However, studies to address GI/GL's effect on weight loss or glucose management in diabetic people have given rise to contradictory results. Among other reasons, the differences can likely be attributed to a range in GI values for the same food making food choices based on such criteria more difficult. In addition, GI only considers the carbohydrate content of food, while lipids, proteins, and vitamins can also influence the metabolic impact of the food [bib_ref] Metabolic effects of low glycaemic index diets, Radulian [/bib_ref]. The brain consumes a large part of glucose from the body for its activity and needs a constant supply [bib_ref] Sugar for the brain: The role of glucose in physiological and pathological..., Mergenthaler [/bib_ref]. Therefore, according to the "selfish brain" theory, the brain competes with the body for glucose. Because of this, the brain sends a message to increase glucose availability when its activity increases. Furthermore, the brain is overly sensitive to glucose availability, and responsive to changes in the glucose supply [bib_ref] The selfish brain: Competition for energy resources, Peters [/bib_ref]. Such sensitivity makes the brain a key regulator, not only for its needs, but also for maintaining a constant glucose level in the blood [bib_ref] Brain Glucose-Sensing Mechanism and Energy Homeostasis, López-Gambero [/bib_ref]. Blood and brain parenchyma glucose levels are linked through transport across the blood-brain barrier via GLUT1 [bib_ref] Supply and Demand in Cerebral Energy Metabolism, Role [/bib_ref]. While hyperglycemia decreases GLUT1 expression, a drop in blood glucose will stimulate GLUT1 expression [bib_ref] Regulation of expression of glucose transporters by glucose: A review of studies..., Klip [/bib_ref]. However, the brain glucose level does not change in association with the blood concentration [bib_ref] Extracellular Glucose Concentration in Mammalian Brain: Continuous Monitoring of Changes during Increased..., Silver [/bib_ref]. Therefore, micro-dialysis experiments have revealed a brain to blood glucose ratio of about 0.5. Furthermore, this ratio decreases when blood glucose increases [bib_ref] Brain metabolism is significantly impaired at blood glucose below 6 mM and..., Meierhans [/bib_ref]. This observation could be the result of a decreased GLUT1 expression that limits glucose uptake by the brain. Altogether, these results indicate a tightly regulated glucose supply to the brain to prevent both hypo-and hyperglycemia from reaching the brain.
Physiological conditions involve complex inter-organ communication to keep body weight and energy homeostasis in a balanced state. The brain plays a crucial role by integrating information on the body's energy status and adapting food intake and energy expenditure via signals sent to peripheral organs [bib_ref] Neuronal Regulation of Energy Homeostasis: Beyond the Hypothalamus and Feeding, Waterson [/bib_ref]. Furthermore, specific areas of the brain are involved in the sensing of glucose, as well as hormone concentration changes (insulin, ghrelin, leptin, etc.) [bib_ref] Signalling from the periphery to the brain that regulates energy homeostasis, Kim [/bib_ref]. In this regard, GI/GL is expected to act in the brain control of energy homeostasis. However, little is known about the impact of high or low GI/GL on those mechanisms. Nevertheless, due to changes in blood insulin and glucose levels, changes in brain responses are likely to occur. Whether these changes are due to direct (glucose or insulin level) or indirect (hormonal, nutrient, or cytokine level changes due to the meal) effects on the brain control of metabolism should be determined. Moreover, the relationship between GI and metabolic disorders remains unclear, despite several studies having been conducted [bib_ref] Low-glycemic index diets as an intervention for diabetes: A systematic review and..., Zafar [/bib_ref] [bib_ref] Energy-restricted diets based on a distinct food selection affecting the glycemic index..., Abete [/bib_ref] [bib_ref] Impact of fat, protein, and glycemic index on postprandial glucose control in..., Bell [/bib_ref] [bib_ref] The metabolic syndrome in relation with the glycemic index and the glycemic..., Vrolix [/bib_ref] [bib_ref] Carbohydrate-restricted versus low-glycemic-index diets for the treatment of insulin resistance and metabolic..., Wood [/bib_ref]. Intriguingly, the mechanisms involved in cognition or neurological disorders have been described in brain energy homeostasis regulation. Therefore, ROS signaling, mitochondrial activity, the gut-brain axis and SCFA, ketone bodies, astrocyte-neuron communication, glutamate metabolism, and gene regulation and neurogenesis or plasticity have been involved in brain metabolic regulations.
Nevertheless, so far, only a few studies have tried to determine the impact of GI/GL on the brain control of metabolism. However, nutrition can affect the body weight, insulin sensitivity, adiposity, and other metabolic parameters. Therefore, we will discuss results suggesting a GI/GL impact on the brain control of energy homeostasis.
## Gi/gl and brain glucose detection
Within the brain, specific regions are dedicated to the sensing of signals involved in energy homeostasis regulation. The hypothalamus is the first area to sense these peripheral signals on the energy status of the body. Within the different nuclei forming the hypothalamus, various neuronal populations have been well described [bib_ref] Hypothalamic nutrient sensing in the control of energy homeostasis, Blouet [/bib_ref]. Furthermore, glucose sensing and the regulatory response to the periphery have been well characterized in the hypothalamus [bib_ref] Brain Glucose-Sensing Mechanism and Energy Homeostasis, López-Gambero [/bib_ref]. This precise sensing of blood glucose is possible due to the blood-brain barrier permeability being weaker in this area of the brain. Due to this, the impact of foods on the blood glucose level is directly integrated by the hypothalamus to restore euglycemia. In this regard, GI/GL should involve a hypothalamic activation of neuronal circuits of glucose and energy homeostasis regulations. Low-GI food leads to a controlled increase in the blood glucose level that is progressively returned to euglycemia. In this case, insulin release is only limited, and the glucose level is decreased due to tissue utilization. On the other hand, high GI/GL results in a high increase of glucose levels in the first phase. Such a rise in blood glucose stimulates brain hyperglycemia detection. In turn, high glucose excited (HGE) neurons are activated, leading to the release of neurotransmitters (GABA in VMN and NTS or anorexigenic neuropeptides POMC/CART in the ARC) [bib_ref] Systemic glucoregulation by glucose-sensing neurons in the ventromedial hypothalamic nucleus (VMH), Shimazu [/bib_ref]. In a second phase, this should lead to a decreased blood glucose level that involves the same area, but stimulates other neurons called glucose-inhibited neurons. The activation of these neurons occurs when glucose levels decrease. Following their stimulation, the brain will send adaptive signals to stimulate food intake and glucose production [bib_ref] Central Mechanisms of Glucose Sensing and Counterregulation in Defense of Hypoglycemia, Stanley [/bib_ref]. The brain sensing a high glucose level will induce a peripheral response through inducing insulin release by the pancreas [bib_ref] The glycemic index: Physiological mechanisms relating to obesity, diabetes, and cardiovascular disease, Ludwig [/bib_ref]. This response contributes to the rapid decrease of the glucose level and the hypoglycemia induced. Following this second phase, hypoglycemia detection is activated and involves the brain in contributing to stimulating the food intake by activating orexigenic neurons.
Interestingly, the redox balance is involved in both hyper-and hypoglycemia sensing. Therefore, hyperglycemia induces mitochondrial ROS production in the hypothalamus. The inhibition of this ROS production inhibits the insulin secretion normally observed due to a lack of nervous activation [bib_ref] Mitochondrial reactive oxygen species are required for hypothalamic glucose sensing, Leloup [/bib_ref]. In addition, changes in the morphology of mitochondria are required to finely regulate this ROS signaling [bib_ref] Importance of mitochondrial dynamin-related protein 1 in hypothalamic glucose sensitivity in rats, Carneiro [/bib_ref]. During obesity and diabetes, a hypersensitivity to glucose is observed. This hypersensitivity is characterized by insulin secretion by the pancreas after detection of a glucose level lower than the usual level. This hypersensitivity is linked to dysfunctions of mitochondrial respiration activity. Consequently, ROS levels are higher than in the non-obese group. Finally, an antioxidant treatment to decrease the ROS level in the hypothalamus restores the normal glucose sensing function, demonstrating the importance of redox balance [bib_ref] Enhanced hypothalamic glucose sensing in obesity: Alteration of redox signaling, Colombani [/bib_ref] [bib_ref] Balancing Mitochondrial redox signaling: A key point in metabolic regulation, Leloup [/bib_ref]. Dysfunctions of mitochondrial dynamics and ROS signaling have also been shown during metabolic unbalance [bib_ref] Mitochondrial Dynamin-Related Protein 1 (DRP1) translocation in response to cerebral glucose is..., Desmoulins [/bib_ref]. Hypoglycemia detection also depends on ROS signaling. Here, ROS are produced during normal hypoglycemic counter-regulation. However, after recurrent hypoglycemic episodes, there is no ROS production [bib_ref] Hypothalamic S-Nitrosylation Contributes to the Counter-Regulatory Response Impairment following Recurrent Hypoglycemia, Fioramonti [/bib_ref]. These results support a key role of mitochondria and the redox balance in the hypothalamic regulation of energy homeostasis, as also shown in other studies [bib_ref] Fasting-and ghrelin-induced food intake is regulated by NAMPT in the hypothalamus, De Guia [/bib_ref] [bib_ref] Mitochondrial dysfunction in obesity, De Mello [/bib_ref] [bib_ref] Mild Impairment of Mitochondrial OXPHOS Promotes Fatty Acid Utilization in POMC Neurons..., Timper [/bib_ref] [bib_ref] Oxidative Stress in the Hypothalamus: The Importance of Calcium Signaling and Mitochondrial..., Gyengesi [/bib_ref] [bib_ref] Hypothalamic reactive oxygen species are required for insulin-induced food intake inhibition: An..., Jaillard [/bib_ref]. In addition, lipids, insulin, and ghrelin signaling in the hypothalamus also participate in ROS signaling. Therefore, according to the mechanisms putatively involved in a low-vs. high-GI diet, and its impact on ROS production (see Section 3), it is likely that the hypothalamic regulation of metabolism will be affected by the GI of a food or diet. For instance, a high-GI diet could mimic a recurrent increase in ROS and therefore disrupt the mechanism of hypoglycemia counter-regulation. Additionally, recurrent increases in glucose and insulin could participate in the development of insulin resistance in the brain, as has been observed for a high-GI diet in peripheral tissues [bib_ref] Effects of a high-fat, low-versus high-glycemic index diet: Retardation of insulin resistance..., Schothorst [/bib_ref]. On the other hand, a low-GI diet would produce ketone bodies that are inhibitors of ROS production by interfering with the mitochondrial respiration, as described above [bib_ref] Reduced mitochondrial reactive oxygen species production in peripheral nerves of mice fed..., Cooper [/bib_ref] [bib_ref] Metabolic dysfunction and oxidative stress in epilepsy, Pearson-Smith [/bib_ref] [bib_ref] Ketogenic diet regulates the antioxidant catalase via the transcription factor PPARγ2, Knowles [/bib_ref] [bib_ref] Monocarboxylate transporters: New players in body weight regulation, Carneiro [/bib_ref] [bib_ref] Evidence for hypothalamic ketone body sensing: Impact on food intake and peripheral..., Carneiro [/bib_ref] [bib_ref] Hypothalamic Fatty Acids and Ketone Bodies Sensing and Role of FAT/CD36 in..., Foll [/bib_ref] [bib_ref] Fatty acid-induced astrocyte ketone production and the control of food intake, Foll [/bib_ref] [bib_ref] Role of VMH ketone bodies in adjusting caloric intake to increased dietary..., Le Foll [/bib_ref] [bib_ref] Regulation of hypothalamic neuronal sensing and food intake by ketone bodies and..., Le Foll [/bib_ref]. Interestingly, ketone bodies have recently been shown to be involved in metabolic disorders and the brain control of energy homeostasis, notably by affecting the food intake [bib_ref] Monocarboxylate transporters: New players in body weight regulation, Carneiro [/bib_ref] [bib_ref] Evidence for hypothalamic ketone body sensing: Impact on food intake and peripheral..., Carneiro [/bib_ref] [bib_ref] Hypothalamic Fatty Acids and Ketone Bodies Sensing and Role of FAT/CD36 in..., Foll [/bib_ref] [bib_ref] Fatty acid-induced astrocyte ketone production and the control of food intake, Foll [/bib_ref] [bib_ref] Role of VMH ketone bodies in adjusting caloric intake to increased dietary..., Le Foll [/bib_ref] [bib_ref] Regulation of hypothalamic neuronal sensing and food intake by ketone bodies and..., Le Foll [/bib_ref]. Discrepancies in results could be due to various parameters, including the origin of ketone bodies. Indeed, a high-fat diet is associated with hyperketonemia induced by an increase in fat metabolism from the diet, while fasting ketone bodies are produced from lipid stores in the body. Therefore, the results observed can be altered by other signals, such as fatty acids, hyperglycemia, hyperinsulinemia, and leptin, among others, during the consumption of a high-fat diet. A high-GI diet probably falls into this situation since it is high in carbohydrate and high-GI foods. A low-GI diet, however, will induce ketone body production from the fatty acids produced from adipose tissue. Moreover, in addition to ketone bodies produced in the liver, in this condition, astrocyte ketogenesis will also be stimulated [bib_ref] Is There an Astrocyte-Neuron Ketone Body Shuttle, Blázquez [/bib_ref] [bib_ref] Ketone body production and disposal: Effects of fasting, diabetes, and exercise, Balasse [/bib_ref]. Indeed, it is possible that the signal exhibited by ketone bodies produced by astrocytes locally would be different from that displayed by hepatic ketone bodies that would affect the whole body.
Ketone bodies could also be involved in metabolism through the acetylation or methylation of genes, thus modulating their transcription, as described before [bib_ref] Ketone Bodies as Anti-Seizure Agents, Simeone [/bib_ref] [bib_ref] Regulation of brain PPARgamma2 contributes to ketogenic diet anti-seizure efficacy, Simeone [/bib_ref] [bib_ref] Ketogenic diet-induced peroxisome proliferator-activated receptor-γ activation decreases neuroinflammation in the mouse hippocampus..., Jeong [/bib_ref]. Indeed, such gene regulation has previously been linked to the brain control of energy homeostasis [bib_ref] Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood..., Mcgowan [/bib_ref] [bib_ref] Transcriptional and epigenetic regulation of POMC gene expression, Drouin [/bib_ref] [bib_ref] The role of epigenetics in hypothalamic energy balance control: Implications for obesity, Obri [/bib_ref] [bib_ref] Environmental and hormonal regulation of epigenetic enzymes in the hypothalamus, Stevenson [/bib_ref]. Of particular interest, these processes have been shown to be involved in neurogenesis, brain plasticity, and neuronal function. Therefore, low-GI diet-produced ketone bodies should lead to changes in these mechanisms and modify the neural regulation of metabolism. Since ketone bodies are associated with fat metabolism, they would rather signal a lack of energy that would stimulate food intake or increase the glucose supply. However, this assumption remains to be tested, since others have described an inhibition of food intake by ketone bodies. Nevertheless, epigenetic mechanisms stimulated by low GI should help control the energetic balance, while high GI should lower ketone body production and thus block their signal. However, further studies are needed to completely understand the role of ketone bodies in the brain regulation of energy homeostasis.
Interestingly, ketone bodies are involved in changes in ATP production. Such a role has been described in neurological disorders. Either ketone bodies, by inhibiting glycolysis, decrease the ATP production, or by forming substrates for the mitochondrial respiration, maintain ATP production.
Such hypotheses remain to be tested. Nevertheless, ATP levels would be responsible for the regulation of K ATP channel closure [bib_ref] Ketogenic diet metabolites reduce firing in central neurons by opening KATP channels, Ma [/bib_ref] [bib_ref] The ketogenic diet metabolite beta-hydroxybutyrate (β-HB) reduces incidence of seizure-like activity (SLA)..., Li [/bib_ref] [bib_ref] Vesicular neurotransmitter transporter: Bioenergetics and regulation of glutamate transport, Omote [/bib_ref] [bib_ref] Metabolic Control of Vesicular Glutamate Transport and Release, Juge [/bib_ref]. Such a role of ATP production would then regulate the depolarization of glucose excited (GE) neurons that are K ATP -dependent channels.
Astrocyte ketogenesis also supports a role for glial cells in energy homeostasis. First, astrocytes provide energy to neurons to maintain normal activity. Furthermore, astrocytes play a role as a sensor of the metabolic status, and signal this information to the neurons [bib_ref] Brain metabolic sensing and metabolic signaling at the level of an astrocyte, Marina [/bib_ref] [bib_ref] Glucose and hypothalamic astrocytes: More than a fueling role, Leloup [/bib_ref]. Therefore, astrocyte-neuron communication is key in the regulation of energetic signal transport and sensing by neurons. Indeed, studies have described the role of astrocytes in the detection of leptin, glucose, fatty acids, and amino acids and their involvement in the metabolic balance [bib_ref] Involvement of astrocytes in mediating the central effects of ghrelin, Frago [/bib_ref] [bib_ref] Physiological and pathophysiological roles of hypothalamic astrocytes in metabolism, Chowen [/bib_ref] [bib_ref] Glial Fatty Acid-Binding Protein 7 (FABP7) Regulates Neuronal Leptin Sensitivity in the..., Yasumoto [/bib_ref] [bib_ref] Time-Dependent Lactate Production and Amino Acid Utilization in Cultured Astrocytes Under High..., Wang [/bib_ref] [bib_ref] Fatty acid increases cAMP-dependent lactate and MAO-B-dependent GABA production in mouse Astrocytes..., Lee [/bib_ref]. The role of astrocytes in glutamate recycling is also associated with the lactate supply to neurons for neurotransmitter action [bib_ref] Time-Dependent Lactate Production and Amino Acid Utilization in Cultured Astrocytes Under High..., Wang [/bib_ref]. In addition, disrupting the communication between astrocytes by blocking the Connexin 43 protein that participates in the astroglial network in the hypothalamus inhibits the glucose sensing response. This result supports indispensable transport through astrocytes for normal glucose detection [bib_ref] Hypothalamic astroglial connexins are required for brain glucose sensing-induced insulin secretion, Allard [/bib_ref]. The astrocyte-neuron lactate shuttle theory (ANLS) hypothesizes that astrocytes metabolize glucose taken up from the bloodstream, and then convert it into lactate in glycolysis. The lactate produced is then transferred to neurons, where it can be oxidized. Therefore, lactate serves as a signaling molecule for the blood glucose level. Accordingly, lactate infusion to the brain mimics the response obtained with glucose [bib_ref] Alteration of hypothalamic glucose and lactate sensing in 48h hyperglycemic rats, Allard [/bib_ref]. Overall, glucose changes associated with ANLS and glutamate recycling represent targets for a GI diet producing ketone bodies following similar mechanisms described in neuronal disorders. However, such a hypothesis has not been tested in the context of brain nutrient sensing.
Finally, the gut-brain axis is also involved in the brain regulation of energy homeostasis. In addition, the diet composition and GI alter the gut function via microbiota that have an effect on neurological disorders [bib_ref] Dietary fibers as emerging nutritional factors against diabetes: Focus on the involvement..., Gowd [/bib_ref] [bib_ref] Impact of dietary fiber consumption on insulin resistance and the prevention of..., Weickert [/bib_ref] [bib_ref] Pomegranate juice, but not an extract, confers a lower glycemic response on..., Kerimi [/bib_ref]. Here, the same molecules identified in neurological conditions are also known to affect brain circuits of metabolic control. Indeed, dietary fibers are expected to provide resistance to the onset of metabolic disorders, as well as the discussed effect on neurological disorders and cognition. Dietary fibers are found in low-GI food such as vegetables [bib_ref] The Impact of Dietary Fiber on Gut Microbiota in Host Health and..., Makki [/bib_ref]. Therefore, the action of GI on microbiota will affect the brain control of energy homeostasis via SCFA. Here, SCFA have been described to have a satiating effect [bib_ref] Glycaemic index and glycaemic load of breakfast predict cognitive function and mood..., Micha [/bib_ref] [bib_ref] The short-chain fatty acid acetate reduces appetite via a central homeostatic mechanism, Frost [/bib_ref]. More recently, microbiota changes have been shown to activate POMC neurons (anorexigenic), reinforcing the role of the gut-brain axis in food intake control [bib_ref] Gut commensal E. coli proteins activate host satiety pathways following nutrient-induced bacterial..., Breton [/bib_ref]. Other works indicate that leptin sensitivity inhibition and pro-glucagon and BDNF decreases induced by microbiota changes could participate in obesity [bib_ref] The gut microbiota reduces leptin sensitivity and the expression of the obesity-suppressing..., Schéle [/bib_ref]. Interestingly, leptin sensitivity and BDNF are linked to AD, strongly suggesting that microbiota could be key elements in the relationship between brain disorders and metabolism. Therefore, meal composition and the GI of foods and meals represent an interesting target in the elucidation of this relationship, but also in preventing, delaying, and treating the diseases.
# Conclusions and perspectives
GI/GL values have given rise to new diets to help people with metabolic diseases control their body weight and glucose homeostasis. Encouraging results on weight management and insulin sensitivity improvement have been observed. In addition, more historical diets, such as ketogenic or modified ketogenic diets, the Mediterranean diet, and the Japanese diet, which are likely to have a low GI, are associated with a healthy lifestyle. In addition to their metabolic improvement properties, these diets also exhibit positive effects on neuronal diseases. Therefore, people following these types of diets present a decreased risk of developing brain diseases while aging. Furthermore, these diets have also been used for decades to treat epilepsy, for example. However, even though it is known that these nutritional interventions improve wellbeing, little is known on the mechanisms involved. Interestingly, despite the primary role in metabolic control, the most important progress in the understanding of the brain impact of a low-GI diet has come from neurological condition studies. Nevertheless, it is worth noting that the mechanisms identified or hypothesized from these studies have been described in the context of the brain control of energy homeostasis. Among these mechanisms, the redox balance, mitochondrial function, ATP production, insulin sensitivity, gene regulation, astrocyte-neuron lactate shuttle, and neurotransmitter regulation are all involved [fig_ref] Figure 2: Schematic representation of the different mechanisms putatively involved in the beneficial effect... [/fig_ref].
Nutrients 2020, 12, x FOR PEER REVIEW 20 of 32 the relationship between brain disorders and metabolism. Therefore, meal composition and the GI of foods and meals represent an interesting target in the elucidation of this relationship, but also in preventing, delaying, and treating the diseases.
# Conclusions and perspectives
GI/GL values have given rise to new diets to help people with metabolic diseases control their body weight and glucose homeostasis. Encouraging results on weight management and insulin sensitivity improvement have been observed. In addition, more historical diets, such as ketogenic or modified ketogenic diets, the Mediterranean diet, and the Japanese diet, which are likely to have a low GI, are associated with a healthy lifestyle. In addition to their metabolic improvement properties, these diets also exhibit positive effects on neuronal diseases. Therefore, people following these types of diets present a decreased risk of developing brain diseases while aging. Furthermore, these diets have also been used for decades to treat epilepsy, for example. However, even though it is known that these nutritional interventions improve wellbeing, little is known on the mechanisms involved. Interestingly, despite the primary role in metabolic control, the most important progress in the understanding of the brain impact of a low-GI diet has come from neurological condition studies. Nevertheless, it is worth noting that the mechanisms identified or hypothesized from these studies have been described in the context of the brain control of energy homeostasis. Among these mechanisms, the redox balance, mitochondrial function, ATP production, insulin sensitivity, gene regulation, astrocyte-neuron lactate shuttle, and neurotransmitter regulation are all involved [fig_ref] Figure 2: Schematic representation of the different mechanisms putatively involved in the beneficial effect... [/fig_ref]. Therefore, these mechanisms appear to be key in brain function, and an alteration of them could lead to neurological dysfunctions. These dysfunctions would then lead to neurological disorders or defects in brain regulatory mechanisms such as energy homeostasis. Therefore, the inability of the brain to precisely modulate energy needs for its own needs will trigger neurological dysfunctions. Therefore, these mechanisms appear to be key in brain function, and an alteration of them could lead to neurological dysfunctions. These dysfunctions would then lead to neurological disorders or defects in brain regulatory mechanisms such as energy homeostasis. Therefore, the inability of the brain to precisely modulate energy needs for its own needs will trigger neurological dysfunctions.
However, a better understanding of these mechanisms is still needed to completely understand both neurological disorders and metabolic disorders. Moreover, improved knowledge could help to decipher the increasingly evident relationship between neuronal disorders and the metabolic balance.
Finally, a better understanding of the mechanisms involved in diets could help develop and improve nutritional recommendations for improving the health of people with these diseases. In addition, the diet could be a good target for preventing the development of both metabolic and neuronal disorders. Moreover, deciphering how these mechanisms involve diets could result from such studies and thus also improve cognitive function in healthy people.
Author Contributions: Writing-original draft preparation, L.C. and C.L.; writing-review and editing, L.C. and C.L. All authors have read and agreed to the published version of the manuscript.
## Funding: this research received no external funding
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Schematic diagram of the influence of GI or GL on blood glucose (left axis) or insulin (right axis). Low vs. medium vs. high GI or GL and their corresponding value range are indicated. [/fig]
[fig] Figure 2: Schematic representation of the different mechanisms putatively involved in the beneficial effect of a low-GI diet on neurological disorders (left panel), paralleled with known mechanisms involved in the brain control of energy homeostasis (right panel). [/fig]
[table] Table 2: Example of various diets' composition for macronutrients with some examples of common foods associated with them. The low-GI diet highlighted in green is taken as a reference for a healthy diet. [/table]
[table] Table 3: Summary of human studies on diet effects on cognitive function in different neurological conditions. [/table]
|
Association of acute kidney injury with frailty in elderly population: a systematic review and meta-analysis
Aim: The objective of this study was to assess whether an elderly patient's frailty was associated with acute kidney injury (AKI) and to examine whether severe frailty group had an increased risk of AKI than mild-moderate group. Methods: We searched The Cochrane Library, PubMed, and EMBASE for relevant studies without language limitations before 1 March 2019 with a priori defined inclusion and exclusion criteria. Five population-based cohort studies were included for systematic review and meta-analysis. Results: Compared with the control group, the frailty group is significantly associated AKI (Odds Ratio ¼ 2.05; 95% CI: 1.23-3.43). The moderate-severe frailty group has an increased risk of AKI than mild frailty group (Hazard Ratio ¼ 2.87; 95% CI: 1.60-5.17. Conclusion: In conclusion, the available best evidence support an association between frailty and AKI among elder patients, thus relevant interventions should be taken among elderly under potential risk of AKI.ARTICLE HISTORY
# Introduction
Frailty is multidimensional syndrome defined as a 'loss of physiologic reserve and the ability to resist stress', and frailty is common in geriatric populations and leads to a high risk of falls, disability, hospitalization and death. The consequences of frailty are often measured mortality, morbidity and institutionalization. Frailty also has been recognized as a risk factor for diabetes, cardiovascular, kidney dysfunction and depressed mood. Frailty was thought to have physical and cognitive components. Cognitive frailty may lead to impaired executive function, falls, balance disturbances functional decline, urge incontinence, disability. Physical frailty severely limit patients' ability to perform activities in home environment.
From the clinical viewpoint, acute kidney injury (AKI) is a complex clinical syndrome that significantly influences the disease process and worsens the outcome of a large number of patients in hospitalizations. Evidence-based studies defined the syndrome more accurately and elucidate the pathogenesis of AKI. Previous epidemiologic studies also indicated that AKI in the elderly is increasingly common and that there is an age-dependent relationship between AKI and older age. In addition, AKI is obvious that all clinical phenotypes of AKI could not fit into a single pathophysiologic pathway.
Whether frailty is associated AKI is an important clinical question warranting investigation. However, the impacts of frailty on AKI among elderly is still unclear. The objective of this systematic review and metaanalysis was to assess the risk of AKI in elder people with frailty and to examine the respective risk estimates in those with mild and moderate-severe frailty.
'(((elder people OR elderly OR the aged OR aging OR old people OR elder OR old man OR the old OR the elderly)) AND (frail OR frailty OR weak OR fragile OR feeble OR in poor health)) AND (acute renal injury OR acute kidney failure OR acute renal failure OR acute kidney injury OR acute kidney insufficiency OR acute renal insufficiency).' with no limitations on language. We check the reference list of screening studies or other relevant comments to further identify other similar studies. The search strategy without handsearching is shown in Supplementary Material 1. This study is conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The protocol of this systematic review was registered in the PROSPERO under the number CRD136855.
## Study selection
These studies are included if the following inclusion criteria are met: 1. the study design was cohort study; 2. the exposure group was composed of people diagnosed with AKI and the control group was composed non-AKI; and 3. the outcome was either the odds or risk of AKI in subjects with frailty. The outcome of our interest was whether frailty existed before diagnosed with AKI. We did not include frailty after AKI. The titles and/or abstracts of the search results were scanned, and we obtained the full text when a study had appeared to meet the criteria mentioned before. In order to determine if there were potentially related information, we examined the full text. We screened the titles and abstracts of the search papers and obtained the full text when a study revealed to meet the inclusion criteria. We further examined the full text to decide whether information was potentially related. Two authors (Miss Zhu Liduzi Jiesisibieke and Dr. Tao-Hsin Tung) selected relevant studies independently, with disagreements resolved via discussion with a third senior author (Prof. Ching-Wen Chien).
## Data extraction and quality assessment
The following data were extracted from the included studies by using a data extraction form: first author, publication year, country, database used, study duration, study design, study subjects, mean age of study subjects, assigned groups and outcomes. The Newcastle-Ottawa Scale (NOS) were applied to assess the quality of the included studies. The NOS application includes three aspects: selection of study groups, comparability and outcome assessment, and they were used to assess the quality of the cohort studies. For each item in the selection and result fields, a study of up to one star can be awarded, and for comparability, up to two stars can be obtained. If we get seven or more stars, we will consider conducting high quality research.
In addition, to enhance the reproducibility and comparability of this review to future reviews of a similar topic, we also included a risk of bias assessment using Risk of Bias in Nonrandomized studies of Interventions (ROBINS-I), since it is the newest and most robust method of assessing risk of bias in systematic reviews and meta-analyses.
## Inter-rater reliability for the selection and data extraction
To establish a consistent selection and data extraction, the kappa statistic was used to assess interobserver reliability between the two independent reviewers. Our study on interobserver reliability showed a kappa value of 0.798 for the data extraction (95% confidence interval (CI): 0.718-0.878).
# Statistical analysis
In this study, we used the Review Manager 5.3 (The Nordic Cochrane Center, The Cochrane Collaboration, 2014). We presented the risk of AKI as OR with 95% CI and assessed heterogeneity by using the I 2 statistic. The I 2 statistic evaluates the degree of variation across studies due to heterogeneity rather than by chance alone. An I 2 values of 50% or more represents substantial heterogeneity. Since I 2 was below 50% (I 2 ¼ 43%), fixed-effect model were carried in meta-analysis.
# Results
## Characteristics of included studies
As illustrated in, our search identified 1096 articles after removing duplicates. Eventually, 4 publications, reporting a total of four cohort studies with a total of 1052 study subjects met our inclusion criteria. The 2012 article by Aitken et al. reported one cohort study. The characteristics of the included studies are shown in. These studies were published between 2016 and 2018. All four cohort studies were population based cohort studies from the UK, the USA, and Korea. Furthermore, all selected studies were rated nine stars on the NOS scale considered with high quality and low-to-moderate risk using ROBINS-I. As shown in, various sets of diagnostic codes were used for outcome measurement.
In addition, Khaled-Adbel-Kader et al., Seon Ha Baek et al., and Sarah Morton et al.staged AKI based on the criteria of Kidney Disease Improving Global Outcomes (KDIGO) by the difference between baseline and peak serum creatinine in the elder population. Bellal Joseph et al.defined AKI as one of inhospital complications. For the frailty, Khaled Abdel-Kader et al.defined clinical frailty status using the Clinical Frailty Scale at 3 and 12 months after discharge. In Bellal Joseph et al.study, frailty was measured by Trauma Specific Frailty Index (TSFI). Seon Ha Baek et al.defined frailty based on the criteria of comprehensive geriatric assessment (CGA). Sarah Morton et al.used 'The Clinical Frailty Score' (CFS) as a tool to assess elder patients' frailty status.
## Association between frailty and aki among the elderly population
Four studies provided data on this outcome. The study by Joseph et al.included 136 patients with frailty, 232 control subjects and concluded that frailty is a significant predictor of incident in-hospital complications, total mortality, and failure to rescue. The study by Abdel-Kader et al. included 242 patients with frailty and 75 control subjects. The results suggested that the main survivors of critical complications by AKI are clinically frail at three and twelve months after discharge, and AKI is independently related to frailty. The study by Morton et al. included 45 patients with frailty and 119 control subjects. They showed that that the incident AKI in elderly acute medical patients is significantly related to severe frailty and higher mortality. All these study provide odds ratio (OR) or risk ratio (RR), the total OR is 2.05 (95%CI: 1.23-3.43). The study by Atiken et al. include 248 frailty and 1577 control subjects. No meta-analysis for this outcome was undertaken as only this paper provide hazard ratio (HR ¼ 3.4, 95%CI: 2.96-3.84, p < 0.01). In this study,shows that frailty was significantly associated with AKI than nonfrailty group (OR ¼ 2.05, 95%CI: 1.23-3.34).. In this study,indicates that moderate-severe frailty was significantly related to AKI than mild-frailty group (HR ¼ 2.87, 95%CI: 1.60-5.17).
# Discussion
## Clinical implications
To the best of our knowledge, this study is the first systematic review and meta-analysis to examine the impacts of frailty on AKI among the elderly population.
Our study results support the hypothesis that elder patients with frailty had an increased risk of AKI, and as the degree of frailty deepens, the risk of AKI is increased. Thus, the overall evidence supported an association between frailty and AKI in the elder population. AKI is common in hospital and is important for patients' health. Previous study estimated that the incidence of AKI is highest among elderly patients. Frailty occurs most frequently in elderly population, and similar to AKI, carries high risk of adverse results include physical disability, decreased function, more hospitalization and higher mortality, especially in ill elderly patients critically admitted to the ICU. Although the level of serum creatinine may not totally reflect patients' kidney dysfunction in the elderly population, previous study found the J-shaped relationship between serum creatinine and limited ability and suggested that a lower serum creatinine may also be related to functional limitation and kidney function in this subgroup.
Some injury patterns of AKI, for example, inflammation and immune system dysregulation, may predispose to frailty. Other AKI consequences excluded fluid overload, anemia, cardiovascular disease, and metabolic derangements may also affect the frail status of some susceptible critical disease survivors. Although frailty and AKI often occur in critically ill and elderly patients, the interactions between them remain unclear. Despite this, they may tend to cross each other in a vicious circle, thereby worsening the overall prognosis of the patients. To compare with younger patients, elder patients with AKI have decreased kidney function recovery and increased further mortality. After adjustment for other confounding factors, old age is viewed as an independent predictor for AKI and a multifactorial disease process. In addition, previous studies also demonstrated that both AKI and frailty were correlated to longer hospital stay or re-attendance rates. Based on the results showed that no statistical difference in mortality between aged 65-80 years and more than 80 years, subjects' frailty rather than biological age may be as a predictive factor for incident AKI. Frailty is still the most problematic manifestation of population ageing. Further studies are needed to validate the relationship, underpin potential mechanisms of possible bidirectional and develop treatment regimens focused on ameliorating the burden of frailty in elder survivors of critical illness and AKI.
To examine whether frailty had a significant impact on the rates of AKI among the old, our study included 4 cohort studies and excluded cross-sectional or case-control studies. Ordinary meta-analyses on the efficacy of interventions obtain high-quality evidence from randomized controlled trials only. Randomized controlled trials plays an important role in health research because of its potential for control bias; however, randomized trials are often not the best source of evidence on harm as the study duration is often too short to detect long-term or rare adverse outcomes. In addition, it is not possible to randomly classify people into categories that are 'Frail' or 'No Frailty'. All of our studies are cohort analysis is an advantage because these studies can detect the effects of long-term exposure to varying degrees of frailty. Our findings are important for identifying older patients who are more likely to have AKI. As result of that, medical staff could pay attention to nutrition, training and cognitive interventional approaches, which were proved to be effective to reverse frailty among the elderly patients.
# Methodological considerations
There are a few limitations in this study. Firstly, due to the amount of selected studies which could be search were not sufficient, the relative lower statistical power with smaller sample sizes is inevitable. Secondly, surrounding random-effects model is controversial, that is, the assumption of normal distribution for random effects against the principle of randomization in statistical inference. If there were no random effects, the variance of random effects would be only as an encumbrance variable. The statistical application of this nuisance variable to meta-analysis weights would then be to markedly increase variance estimated and uniform the weights through grueling the larger studies. Thirdly, it is difficult to conduct subgroup analyses based on demographic variables such as age, sex, and concurrent health status because the selected studies did not include adequate information. Future studies should be conducted to examine outcomes and confirm whether elderly patients' frailty is an independent risk factor for AKI. Fourthly, I 2 test seeks to determine whether there are real differences according to the results of the selected studies, namely, heterogeneity, or whether the variation in findings is reconcilable with chance alone, that is, homogeneity. I 2 values of 0-24.9%, 25-49.9%, 50-74.9%, and 75-100% were viewed as none, low, moderate, and high heterogeneity, respectively. In this study, we used the fixed-effect model when I 2 statistics was 43% less than 50%. However, we aggregate studies that are different methodologies, heterogeneity in the results is still inevitable. Fifthly, it is very difficult to explore the results of subgroup analysis for AKI and AKI requiring dialysis due to insufficient information for selected studies. Further studies are needed to estimate the pathological mechanisms associating frailty and AKI only or AKI with dialysis in elderly patients. Finally, although the kappa statistics for the agreement of interobserver reliability seemed acceptable, nondifferential misclassification or bias data extraction still may have occurred.
# Conclusion
In conclusion, our study suggests a relationship between frailty and AKI among elderly population. To further examine this finding and establish a stronger result, more large-scale prospective studies are warranted to provide more information about the details of the association between frailty and AKI. An increased rate of AKI occurs in frailty elderly patients and clinicians should be aware of this possibility. |
De novo transcriptome analysis of Lantana camara L. revealed candidate genes involved in phenylpropanoid biosynthesis pathway
Lantana camara L. is an economically important essential oil producing plant belonging to familyVerbenaceae. it is used in medication for treating various diseases like cancer, ulcers, tumor, asthma and fever. The plant is a useful source of essential bioactive compounds such as steroids, flavonoids and phenylpropanoid glycosides etc. nonetheless, very little is known about the genomic or transcriptomic resources of L. camara, and this might be the reason of hindering molecular studies leading to identification of improved lines. Here we used Illumina sequencing platform and performed the L. camara leaf (LcL) and root (LcR) de novo transcriptome analyses. A total of 70,155,594 and 84,263,224 clean reads were obtained and de novo assembly generated 72,877 and 513,985 unigenes from leaf (LcL) and root (LcR) respectively. furthermore, the pathway analysis revealed the presence of 229 and 943 genes involved in the phenylpropanoid biosynthesis in leaf and root tissues respectively. Similarity search was performed against publically available genome databases and best matches were found with Sesamum indicum (67.5%) that were much higher than that of Arabidopsis thaliana (3.9%). To the best of our knowledge, this is the first comprehensive transcriptomic analysis of leaf and root tissues of this non-model plant from family Verbenaceae and may serve as a baseline for further molecular studies.Lantana camara L. belonging to family Verbenaceae is an evergreen shrub, native to the Neotropics and grown worldwide for its medicinal and ornamental value 1 . The plant may grow up to 4 m, forming dense thickets 2 and is extensively used in traditional herbal medicines of many cultures including Saudi Arabia 3 . The medicinal properties of L. camara are attributed to the presence of many bioactive compounds with therapeutic potential such as steroids, flavonoids, triterpenoids, oligosaccharides, iridoide glycosides, naphthoquinones and phenylpropanoid glycosides 4-6 . Furthermore, L. camara is used in medication for treating cancer, ulcers, tumors, tetanus, cuts, eczema, measles, chicken pox, fevers, rheumatism and asthma1,7,8. Several important phytochemicals have been isolated from L. camara including ursolic acid, oleanolic acid, linaroside, lantanoside, verbascoside, camarinic acid, phytol and umuhengerin etc. and their biological activities such as anticancer, antibacterial, antioxidant, antiulcer and nematocidal have been reported 9-11 . Nonetheless, L. camara is also known as a major source of easily available plant essential oil known as Lantana oil 12,13 . Literature survey indicates the substantial diversity in the composition of essential oils isolated from L. camara growing in diverse localities 14-18 .In spite, much is known about the phytochemistry, toxicology and medicinal properties of L. camara 19 very little is known about the genomic architecture of the plant. To date, only 41 sequences including (rps3, atpB, ccsA, rpoC1, rpoC2, FT, GLO1, rpl32 and rbcL) have been deposited to the NCBI Genbank database 20 . To stop the spread of L. camara as an invasive weed, transcriptomes of young ovaries were recently studied and open possible mechanisms of unreduced female gametes formation were elaborated 21 . Complete genome sequences provide invaluable insights into the biological functions of individual genes and proteins and therefore, hold immense promise for crop improvement and breeding of new cultivars. The genome of L. camara has not been sequenced yet and no details are available on the genomics or transcriptomics of the plant. Therefore, a detailed research strategy is needed to elucidate various aspects of genomics and transcriptomics of this important medicinal plant. Recently we have also reported the complete chloroplast genome of L.camara 22 . Nowadays, transcriptomics using next generation sequencing (NGS) technology has emerged to be one of the most powerful and cost-effective approaches generating huge data on transcribed sequences that may be implemented in wide research purposes 23 . The practical applications of transcriptome sequencing have been utilized in various medicinal plants as well as crop plants including rice 24 , maize 25 , chickpea 26 , wheat 27 , Camelina sativa 28 , Quercus pubescens 29 , finger millet 30 , Stellera chamaejasme 31 and Aconitum heterophyllum 32 . As a step forward, here we report the first de novo transcriptome assembly of L. camara leaf and root tissues that may lay the foundation for rapid identification of functional genes discovery and genomics-assisted breeding for developing more efficient L. camara lines with desired traits.Materials and methodsPlant material and RNA quantification. Plants were grown in the laboratory of plant physiology King Abdulaziz University, Jeddah, Saudi Arabia under 28 °C/22 °C day and night temperature in a semi-controlled green house. No specific permits were required for the described study. Fresh leaf (LCL) and root (LCR) tissues were harvested from 02 months old plant, washed thoroughly with sterile water, and immediately frozen in liquid nitrogen. RNA isolation was carried out using RNeasy Plus Mini Kit (Qiagen, Cat. No: 74134). RNA degradation and contamination were monitored on 1% agarose gel. Purity of RNA was checked on NanoPhotometer ® spectrophotometer (IMPLEN, CA, USA). Quantification and integrity of RNA was assessed using the RNA Nano 6000 Assay Kit (Agilent Technologies, CA, USA).
Library preparation for transcriptome sequencing. NEBNext ® Ultra™ RNA Library Prep Kit was used to generate sequencing libraries. Briefly, mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Divalent cations were used to carry out fragmentation under high temperature in NEBNext (FSSR) first strand synthesis reaction Buffer (5×). Random hexamer primers were used to synthesize first strand cDNA using M-MuLV Reverse Transcriptase (RNase H-). Second strand cDNA synthesis was carried out using DNA Polymerase I and RNase H. Exonuclease/polymerase activity was performed to convert remaining overhangs into blunt ends. Ligation of NEBNext adaptor with hairpin loop was performed after adenylation of 3′end of DNA fragments in order to prepare sample for hybridization. The cDNA fragments of 250-300 bp were preferentially selected and purification of the library fragments was carried out with AMPure XP system (Beckman Coulter, Beverly, USA). 3 µl of Enzyme (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37 °C for 15 min, followed by 5 min at 95 °C. PCR was carried out with Phusion High-Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer. PCR products were purified (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 system. Paired-end library of 150 bp (PE150) was prepared following Illumina protocol/instructions. Clustering, quality control and transcriptome assembly. Cluster generation system (PE Cluster Kit cBot-HS Illumina) was used to perform clustering of the index coded samples following manufacturer's instructions. The library preparations were sequenced on an Illumina platform after cluster generation and paired-end reads were generated. Transcriptome assembly was performed using Trinity software [bib_ref] Full-length transcriptome assembly from RNA-Seq data without a reference genome, Grabherr [/bib_ref] with min_kmer_cov set to 2, and all other parameters with default settings.
Gene functional annotation and biological pathways assignment. Libraries of LCL and LCR tissues were annotated using BLASTX against the NCBI database and all unigenes were applied for homology searches. Gene functional annotation was performed using the following 7 databases: NCBI non-redundant protein sequences (NR), NCBI non-redundant nucleotide sequences (Nt), Protein family (Pfam), Clusters of Orthologous Groups of proteins (KOG/COG), manually annotated and reviewed protein sequence database (Swiss-Prot), KEGG Ortholog database (KO) [bib_ref] Automated genome annotation and pathway identification using the KEGG Orthology (KO) as..., Mao [/bib_ref] and Gene Ontology (GO) database. Best aligning results were selected to annotate the unigenes; if aligning results of these databases were not similar, NR database results were preferentially selected. For assignment of function to unigenes, Gene Ontology (GO) enrichment analysis was performed using AgriGO (https ://bioin fo.cau.edu.cn/agriG O/analy sis.php) [bib_ref] agriGO v2.0: A GO analysis toolkit for the agricultural community, 2017 update, Tian [/bib_ref] and annotated sequence/s may have more than one GO term. Similarly, to understand the high-level functions and gain an overview of gene pathway networks KEGG [bib_ref] KEGG as a reference resource for gene and protein annotation, Kanehisa [/bib_ref] [bib_ref] New perspectives on genomes, pathways, diseases and drugs, Kanehisa [/bib_ref] was used. Enzyme commission (EC) numbers were assigned to unique sequences, based on the BLASTx search of protein databases, using a cutoff E value 10 − 5. The statistical enrichment of DEGs in KEGG pathways was carried out using KOBAS Quantification and differential expression analysis. The RNA-seq by Expectation Maximization (RSEM) package was used to estimate gene expression levels of both root and leaf tissues [bib_ref] RSEM: Accurate transcript quantification from RNA-Seq data with or without a reference..., Li [/bib_ref]. At first clean data was mapped back onto the assembled transcriptome and then read count for each gene was obtained from the mapping results. Prior to differential gene expression (DEGs) analysis of the sequenced library of root and leaf tissues, read counts were adjusted through one scaling normalized factor and analysis was performed using DEGseq (2010) R package [bib_ref] DEGseq: An R package for identifying differentially expressed genes from RNA-seq data, Wang [/bib_ref]. P value was adjusted using q value of < 0.005 and |log2(foldchange)| > 1 was set as the threshold for significantly differential expression. To identify DEGs between the LCR and LCL datasets, raw read counts were initially filtered to exclude orphan transcripts. Pairwise comparison was performed using the
# Results and discussion
RNA sequencing and de novo assembly. Advances in genomics have led to the development of NGS based trait mapping approaches that have tremendously increased the efficiency of traits selection and mapping in complex genomes [bib_ref] Statistical significance for genomewide studies, Storey [/bib_ref]. Among the high throughput genotyping approaches, Next Generation RNA Sequencing technology has contributed to a more comprehensive understanding of functional genes and is widely used for characterization of transcriptome profiles of various model and non-model plants [bib_ref] RNA-Seq analysis and transcriptome assembly of raspberry fruit (Rubus idaeus "Heritage") revealed..., Travisany [/bib_ref]. De novo transcriptome analysis is not only providing an excellent platform for finding novel genes, molecular markers development but also cater a base for the construction of networks of gene expressions for various tissues and organs of animals as well as plants. Although, the number of available high-quality reference genomes has been constantly growing still, de novo transcriptome approaches are mainly used for non-model species where whole genomes information are missing. The remarkable advances in RNA sequencing provide a cost-effective way to obtain large amounts of transcriptome data from many organisms and tissue types, especially in the complete absence of a reference genome thereby, allowing us to identify all expressed transcripts [bib_ref] Full-length transcriptome assembly from RNA-Seq data without a reference genome, Grabherr [/bib_ref] [bib_ref] Automated genome annotation and pathway identification using the KEGG Orthology (KO) as..., Mao [/bib_ref]. Here we provide, the first report on transcriptome profiling and DEGs of L. camara leaf and root tissues; RNA-Seq library was constructed and RNA samples with RIN value more than 6 were used (Supplementary file 1). A total of 76,315,644 and 87,218,768 raw reads were obtained for LCL and LCR respectively. For the removal of adapters, poly-A tail, primer sequences and short as well as low quality sequences trimming process was performed that resulted in a total of 70,155,594 clean reads from LCL and 84,263,224 from LCR. The total size of clean bases generated was 10.5 GB for LCL sample with percent error (0.01%), Q20 (97.39%), Q30 (93.47%) and GC content (45.61%), whereas for LCR sample 12.6 GB with a percent error (0.03%), Q20 (96.81%), Q30 (92.01%) and GC content (43.73%) . The sequences (raw data) generated were deposited to NCBI as PRJNA503321 (for leaf sample) and PRJNA605469 (for root sample). For samples lacking reference genomes, clean reads need to be assembled to get a reference sequence. We used Trinity assembler [bib_ref] Full-length transcriptome assembly from RNA-Seq data without a reference genome, Grabherr [/bib_ref] functional annotation of genes. Unigenes of the LCL and LCR tissues were annotated using BLAST search against the NR, NT, PFAM, GO and KOG databases. Gene Ontology assignments based on the protein match and annotation results are based on Götz et al. [bib_ref] High-throughput functional annotation and data mining with the Blast2GO suite, Götz [/bib_ref] Of the 72,877 and 513,985 unique transcript sequences of LCL and LCR annotated, a large proportion of the sequences (73.27% and 70.11%) had hits in databases. Of the LCL tissue, a total of 50,540 (69.34%) unigenes matched with known proteins in the NR database, while 38,975 (53.48%) unigenes were annotated with entries in the Swiss-prot database, 36,209 (49.68%) unigenes matched . Quality of reads obtained after RNA-sequencing of leaf and root transcriptomes.
## Samples
Raw reads Clean reads Clean bases (GB) Error (%) Q20 (%) Q30 (%) GC (%) [fig_ref] Table 3: The ratio of successfully annotated genes [/fig_ref]. Of the assembled unigenes, only 9,849 (13.51%) and 45,051 (8.76%) in LCL and LCR respectively, were successfully annotated in all databases. Reasons for the non-annotated sequences could be, lack of conserved protein domains in short sequences, or in some cases the transcriptome have noncoding genes, UTRs, random transcriptional noise or incomplete spliced introns which are non-homologues to the sequences available in the public databases. This may also be one of the possible reasons that genes have not shown expression at the time of RNA extraction or those genes are expressed at very low levels [bib_ref] Sequencing and de novo assembly of visceral mass transcriptome of the critically..., Kang [/bib_ref].
The assembled unigenes annotation of LCL and LCR tissues is given as Venn diagram [fig_ref] Figure 1: Venn diagram mapping with database annotation [/fig_ref] against PFAM, GO, KOG, NR and NT databases. For leaf tissues, a total of 13,534 unigenes were annotated in 5 databases; 6,060 unigenes sequences showed homology in 3 databases including PFAM, NR and GO. While 1,167 unigenes were annotated in KOG as well as the above three databases and 2,030 unigenes were annotated in PFAM and GO databases [fig_ref] Figure 1: Venn diagram mapping with database annotation [/fig_ref]. Likewise for root tissues, a total of 57,114 unigenes were annotated in five databases where, 28,396 unigenes were homologous in PFAM, NR and GO databases [fig_ref] Figure 1: Venn diagram mapping with database annotation [/fig_ref].
Annotation in nR database. After the high annotation hit score was ascertained in NR database, we considered identifying homologues features of BlastX hits for the annotated unigenes. The E-value distribution in leaf tissue revealed that 35.6% of the annotated unigenes had E-value of 0-1e-100, followed by 19.9% unigenes with 1e−100 to 1e−60 E-value [fig_ref] Figure 2: E-value distribution [/fig_ref]. Similarity distribution showed that 45.2% of the unigenes had 80-90% similarities. Furthermore, 39.1% of the unigenes had 60-80% and 5.2% had 95-100% similarities [fig_ref] Figure 2: E-value distribution [/fig_ref]. Additionally, comparison of the unigenes annotated with homologous sequences of other plant species was also performed 44 . Among the nucleotide sequences, highest similarity was noted for Sesamum indicum (with the highest similarity score of 67.5%). Other species with sequence homology below 13% included Erythranthe guttata (12.3%), Arabidopsis thaliana (3.9%), Coffea canephora (1.4%) and Vitis vinifera (1.0%) [fig_ref] Figure 2: E-value distribution [/fig_ref].
E-value distributions in root tissue revealed that 25.8% of the annotated unigenes had E-value of 1e−15 to 1e−5, followed by 23.9% unigenes with 1e−30 to 1e−15 E-value [fig_ref] Figure 2: E-value distribution [/fig_ref]. Similarity distribution revealed that 2.8% of the unigenes had 95-100% similarities, 26.6% had 80-95% and 42.9% unigenes had 60-80% similarities [fig_ref] Figure 2: E-value distribution [/fig_ref]. Comparison of the annotated unigenes with the homologous sequences revealed highest similarity with Sesamum indicum (26.4%) followed by Erythranthe guttata (5%), Ricinus communis (3.1%) and Guillardia theta (2.8%) [fig_ref] Figure 2: E-value distribution [/fig_ref].
Gene ontology (GO) classification. GO assignments were used to classify the functions of the LCL and LCR unigenes, which classified unigenes under the categories of biological process (BP), molecular function (MF) and cellular component (CC). Sum of 36,209 LCL unigenes were classified into three major categories [fig_ref] Figure 3: X-axis is the GO term under the three main GO domains [/fig_ref]. The biological process category, the unique sequences were classified into 25 groups. The most characterized biological processes were 'cellular process' (57.59%, GO-ID: 0009987) followed by 'metabolic process' (53.72%, GO-ID: 0008152). The cellular components were divided into 21 groups. Interestingly, we found both the represented cellular components 'Cell' (31.80%, GO-ID: 0005623) and 'Cell part'' (31.80%, GO: 0044464) were similar to previous report [bib_ref] High-throughput functional annotation and data mining with the Blast2GO suite, Götz [/bib_ref]. The molecular functions category, the unique sequences clustered into 10 classes. Where, the highest sub-category was 'binding' (58.35%, GO-ID: 0005488) followed by 'catalytic activity' (46.37%, GO: 0003824) [fig_ref] Figure 2: E-value distribution [/fig_ref].
Similarly, 255,964 LCR unigenes were annotated and also likewise classified into three main categories and further divided into 57 sub-groups [fig_ref] Figure 3: X-axis is the GO term under the three main GO domains [/fig_ref]. Among the biological processes classification, the highest was cellular process 142,929 (55.83%), followed by metabolic process 135,458 (52.92%) and single-organism process 107,769 (42.1%). Furthermore, cellular component was classified in 22 groups, among them high amount of unigenes were related to "cell" 75,691 (29.57%) and "cell part" 75,607 (29.53%), followed by "organelle" 50,774 www.nature.com/scientificreports/ (19.83%) and "membrane" 35,704 (13.94%) [fig_ref] Figure 2: E-value distribution [/fig_ref]. Only few unigenes were assigned to extracellular matrix, extracellular region and extra cellular region part. In the molecular function category, the majority of the unigenes were assigned to "binding" 133,482 (52.14%) and "catalytic activity" 114,050 (44.55%). The annotation and sequence information from Gene Ontology results provide important gene sources for future molecular level studies that underline selection and improvements of L. camara.
## Kog classification.
On the basis of conserved domain alignment, the annotated unigenes were searched against the KOG database to find the functionally classified orthologous gene products. A total of 19,818 and 155,979 identified genes were annotated in LCL and LCR tissues respectively and these were divided into 26 protein group families [fig_ref] Figure 4: X-axis [/fig_ref]. Among these protein families predicted in LCL tissues, general function prediction (2,719 unigenes, 15.41%) was identified as the highest annotated group, followed by post translational modification, protein turnover, chaperones (2,335 unigenes, 13.23%), signal transduction mechanisms (1,557 unigenes, 8.82%), translation, ribosomal structure and biogenesis (1,355 unigenes, 7.68%), RNA processing and modification (1,257 unigenes, 7.12%), intracellular trafficking, secretion, and vesicular transport (1,241, 7.03%), whereas the smallest groups were nuclear structure (89), extracellular structures (26) and cell motility (15) (Supplementary file 3). In LCR tissues, "Translation, ribosomal structure and biogenesis" consisted the largest category with 21,097 genes (14.95%), followed by "Posttranslational modification, protein turnover, chaperones" with 20,612 genes (14.60%) and "General function prediction only" 17,296 genes (12.25%). These results are not in alignment to previous report where the authors found "General function prediction only" as the highest category with 1,006 (13.86%) genes [bib_ref] novo transcriptome analysis of Liriodendron chinense petals and leaves by Illumina sequencing, Yang [/bib_ref]. Fewer genes were assigned to "cell motility" (92) and unnamed (4) proteins [fig_ref] Figure 3: X-axis is the GO term under the three main GO domains [/fig_ref].
functional characterization using KeGG. For biological functioning of genes, pathway-based analyses are imperative and this assignment was performed in the KEGG database. In LCL sample, of the 20,506 unigenes, 18,853 were assigned to 5 major groups in KEGG with 26 sub-categories and 131 biochemical pathways [fig_ref] Figure 4: X-axis [/fig_ref]. These major groups were cellular processes (1,004 unigenes), environmental www.nature.com/scientificreports/ information processing (694 unigenes), genetic information processing (4,285 unigenes), metabolism (8,737 unigenes) and organismal systems (878 unigenes). In these 5 groups, the topmost was the metabolism with 8,737 unigenes, and its further evaluation classified it into 10 subgroups [fig_ref] Figure 6: Sub categorization of the metabolism group [/fig_ref]. Of these, 'carbohydrate metabolism' with (1,751) genes involved the highest number of genes, followed by 'amino acid metabolism' (1,064) genes and 'energy metabloism' with (886) genes [fig_ref] Figure 6: Sub categorization of the metabolism group [/fig_ref]. Furthermore, 936 genes were identified that coded for important proteins and have matches with 1,113 enzymes. The function of these enzymes was assigned to 21 secondary metabolite pathways [fig_ref] Figure 7: Assigned enzymes function to secondary metabolite pathways with number of genes [/fig_ref]. Among these pathways 546 genes encoded key enzymes involved in terpeniods biosynthesis including monoterpenoid (28 genes), diterpenoids (30 genes), terpenoid backbone (166 genes), sesquiterpenoid and triterpenoids . Similarly, 41 genes were related to flavonoid biosynthesis pathway including flavone and flavonol (32 gene). Among all the secondary metabolite pathways, the phenylpropanoid biosynthesis pathway involved the highest numbers (229) of genes [fig_ref] Figure 7: Assigned enzymes function to secondary metabolite pathways with number of genes [/fig_ref]. These results provide valuable insight into the metabolic pathways in leaves of L. camara. These results may provide basis for future studies to identify and characterize genes and transcripts that are involved in important metabolic pathways and may provide insights into the understanding of functions of these genes in the biosynthesis of active compounds in L. camara.
Of the 124,076 unigenes in LCR transcriptome, 115,383 were assigned into 5 groups in KEGG with 19 subcategories and 131 biochemical pathways [fig_ref] Figure 4: X-axis [/fig_ref]. The major groups were cellular processes (7,793 unigenes), environmental information processing (3,384 unigenes), genetic information processing (39,630 unigenes), metabolism (50,349 unigenes) and organismal systems (3,111 unigenes). In these 5 groups, www.nature.com/scientificreports/ the topmost was the metabolism with 50,349 unigenes was further analysed, which we took for further consideration. The metabolism group was further categorized into 10 subgroups [fig_ref] Figure 6: Sub categorization of the metabolism group [/fig_ref]. In all the 10 sub-groups of metabolism 'carbohydrate metabolism' was the highest with (11,724) genes, followed by 'amino acid metabolism' with (9,120) genes and 'energy metabolism' with (6,791) genes [fig_ref] Figure 6: Sub categorization of the metabolism group [/fig_ref]. Furthermore, 4,166 genes were identified that coded for proteins and have important matches to 5,368 enzymes. The function of these enzymes was assigned to 21 KEGG secondary metabolite pathways [fig_ref] Figure 7: Assigned enzymes function to secondary metabolite pathways with number of genes [/fig_ref]. Among these pathways 2,796 genes encoded key enzymes involved in terpeniods biosynthesis including monoterpenoid (92 genes), diterpenoids (75 genes), terpenoid backbone (853 genes), sesquiterpenoid and triterpenoids (103 genes). The results showed 119 genes were related to flavonoid biosynthesis pathway including flavone and flavonol biosynthesis (75 gene), whereas www.nature.com/scientificreports/ the highest number of genes (943 genes) among all the secondary metabolite were involved in phenylpropanoid biosynthesis pathway [fig_ref] Figure 7: Assigned enzymes function to secondary metabolite pathways with number of genes [/fig_ref].
phenylpropanoid biosynthesis. The current study identified the maximum number of genes for phenylpropanoid biosynthesis pathway and this was taken into further consideration. Phenylpropanoids are phytobased natural compounds that are usually derived from phenylalanine. Phenylpropanoid plays vital role in plant response to various biotic and abiotic stresses [bib_ref] Phenylpropanoid biosynthesis, Vogt [/bib_ref]. The process of phenylpropanoid biosynthesis starts with the formation of cinnamic acid from phenylalanine. This cinnamic acid is then converted into cinnamoyl-CoA, p-Coumaryl-CoA, p-coumaryl quinic acid, caffeoyl quinic acid, caffeoyl-CoA, feruloyl-CoA, and sinapoyl-CoA. caffeoyl quinic acid also known as chlorogenic acid, which is a highly soluble phenylpropanoid in Solanaceae and it has been mentioned to play a vital role as an antioxidant and defense molecule [bib_ref] Phenylpropanoid biosynthesis, Vogt [/bib_ref]. Further, we have identi- www.nature.com/scientificreports/ fied a total of 229 and 943 genes as well as 17 and 19 enzymes in LCL and LCR transcriptomes respectively and these enzymes are required for phenylpropanoid biosynthesis pathway [fig_ref] Table 4: Important enzymes identified within phenylpropanoid biosynthesis pathway [/fig_ref]. Our results are different from [bib_ref] RNA sequencing and de novo assembly of Solanum trilobatum leaf transcriptome to..., Lateef [/bib_ref] where the authors reported only 11 genes from Solanum trilobatum that were involved in the biosynthesis of Gene expression analysis. De novo transcriptome filtered by Corset was used as a reference [bib_ref] Corset: Enabling differential gene expression analysis for de novo assembled transcriptomes, Davidson [/bib_ref]. RSEM 40 that map reads back to transcriptome and quantify their expression level was applied. Total reads in gene expression analysis of LCL tissues were 70,155,594, followed by total mapped 54,875,118 (78.22%); whereas total reads in gene expression analysis of LCR tissues were 84,263,224 followed by total mapped 47,187,580 (56.00%). To calculate the gene expression level, RSEM analyzed the mapping results of Bowtie and read count for each gene was converted into FPKM value. In RNA-seq, it is the most common method of estimating gene expression levels, which takes into account the effects of both sequencing depth and gene length on counting of fragments. These results are summarized in Supplementary File 6. Venn diagram, heat map and volcano plot highlights the DEGs and shows genes that are unique and common to leaf and root samples. The total numbers of expressed genes in LCL and LCR tissues were 67,714 and 359,010 respectively, while a total of 49,072 unigenes were found as commonly expressed in both tissues. We used DESeq software, to analyze the expression of unigenes in both LCL and LCR transcriptomes by normalizing the values to Fragments Per Kilobase Million (FPKM). The Benjiamini-Hochberg method was used to verify and revise the p-values. Volcano plot demonstrates the fold changes in the expression and statistical comparison. A total of 20,044 differentially expressed genes (DEGs) were identified. Within these DEGs 11,496 genes were up regulated and 8,548 genes were down regulated. Further, the hierarchical clustering analysis was used to screen the DEGs and cluster them according to their expression in LCL and LCR samples [fig_ref] Figure 1: Venn diagram mapping with database annotation [/fig_ref]. On the basis of pathway enrichment, 20 metabolic and/or biosynthetic processes were predominantly involved. Of these spliceosome and protein processing in the endoplasmic reticulum was comprised by approximately 300 genes [fig_ref] Figure 1: Venn diagram mapping with database annotation [/fig_ref]. Other important processes identified included; synaptic vesicle cycle, starch and sucrose metabolism, plant hormone signal transduction, conclusion Molecular studies on L. camara are rare and high-throughput genotyping efforts are almost non-existent. This study investigated the transcriptomes assembly of L. camara leaf and root tissues. A massive data of 70,155,594 and 84,263,224 clean reads were de novo assembled that revealed 72,877 and 513,985 unigenes from leaf and root tissues respectively. Further, the identified unigenes were annotated and functionally characterized in 7 databases. Notably, the numbers of expressed genes in LCL and LCR tissues varied; and 49,072 unigenes were commonly expressed in both tissues. Still, of the 20,044 DEGs 11,496 were up regulated and 8,548 genes were down regulated. Pathway analysis revealed the involvement of 229 and 943 genes (coding for 17 and 19 enzymes) in the biosynthesis of phenylpropanoid pathway in leaf and root tissues respectively. Nonetheless, the genomic resources will not only provide foundation of genomic research in L. camara but it will also provide detailed insight into the expression as well as functional analysis, gene cloning and avenues for genomics-assisted breeding in L. camara. This study will also serve as baseline to understand the regulation and biosynthesis of crucial bioactive compounds and to select superior alleles/haplotypes of L. camara with desired traits in the future.
[fig] Figure 1: Venn diagram mapping with database annotation. Scientific RepoRtS | (2020) 10:13726 | https://doi.org/10.1038/s41598-020-70635-5 [/fig]
[fig] Figure 2: E-value distribution (A), similarity distribution (B) and species distribution (C). Scientific RepoRtS | (2020) 10:13726 | https://doi.org/10.1038/s41598-020-70635-5 [/fig]
[fig] Figure 3: X-axis is the GO term under the three main GO domains; Y-axis is the number and percentage of the annotated genes in the term (sub-term included). [/fig]
[fig] Figure 4: X-axis: names of the 26 KOG group; Y-axis: percentage of annotated genes under this group in the total annotated genes. [/fig]
[fig] Figure 5: Y-axis: names of KEGG pathways; X-axis is the number of the genes annotated in the pathway and the ratio between the number in this pathway and the total number of annotated genes. The KEGG metabolic pathways gene involved in are divided into five branches; (A) Cellular Processes, (B) Environmental Information Processing, (C) Genetic Information Processing, (D) Metabolism and (E) Organismal Systems. Scientific RepoRtS | (2020) 10:13726 | https://doi.org/10.1038/s41598-020-70635-5 [/fig]
[fig] Figure 6: Sub categorization of the metabolism group. [/fig]
[fig] Figure 7: Assigned enzymes function to secondary metabolite pathways with number of genes. Scientific RepoRtS | (2020) 10:13726 | https://doi.org/10.1038/s41598-020-70635-5 [/fig]
[fig] Figure 9: (A) Venn diagram showing differentially expressed genes identified in leaf and root of L.camara, (B) volcano plot showing map of DEGs. carbon fixation in photosynthetic organisms, amino sugar and nucleotide sugar metabolism and galactose metabolism. [/fig]
[table] Table 3: The ratio of successfully annotated genes. [/table]
[table] Table 4: Important enzymes identified within phenylpropanoid biosynthesis pathway. various compounds of this pathway. The important enzymes identified were cinnamyl-alcohol dehydrogenase [EC:1.1.1.195], followed by caffeoyl-CoA O-methyltransferase [EC:2.1.1.104], trans-cinnamate 4-monooxygenase [EC:1.14.13.11], Cinnamoyl-CoA-reductase [EC: 1.2.1.44], phenylalanine ammonia-lyase [EC:4.3.1.24], 4-Coumarate CoA-ligase [EC: 6.2.1.12], and shikimate O-hydroxycinnamoyl-transferase [EC:2.3.1.133]. Interestingly, in root transcriptome two protein coding genes namely cytochrome P450, family 98, subfamily A, polypeptide 8 [EC:1.14.13.-], and Aromatic-L-amino-acid decarboxylase [EC: 4.1.1.28] were identified, that were not detected in leaf transcriptome for phenylpropanoid biosynthesis pathway. The presence of all these enzymes indicates to the phytotherapeutic potential of L. camara (Fig. 8, Table 4, Supplementary File 5). [/table]
|
Psychometric functioning, measurement invariance, and external associations of the Relationship Assessment Scale in a sample of Polish Adults
The current article reports data from three Polish samples to examine the Relationship Assessment Scale (RAS) with respect to its unidimensionality, invariance across countries, gender, formal and informal relationships, degree of precision (or information) across latent levels of relationship satisfaction, and the functioning of individual items. The analyses of the data from the reference sample (n = 733) confirmed a clear 1-factor structure of the RAS-PL and good internal consistency. Configural, metric, and scalar invariance for countries (Poland, Hungary, USA), gender (women and men) and relationship types (formal and informal relationships) were achieved. Item Response Theory Analysis (IRT) suggested that the RAS-PL assesses relationship satisfaction most reliably at low to average levels. Analyses of the data from validation samples (n = 203 and n = 209) confirmed the convergent and divergent validity by weak, medium, and large correlations of the RAS-PL with measures of other theoretically related constructs. Concurrent criterion validity was demonstrated by a strong positive correlation between the RAS-PL and the intent to continue the current relationship.This investigation provides considerable psychometric information about the items and scale of the RAS-PL.Relationship satisfaction is associated with various outcomes for both individuals' and families' well-being 1,2 . Therefore, "[…] the importance of knowing more about how to measure adjustment in intimate partnerships" 3 , p 1029 as well as the psychometric properties of the research instruments used to measure relationship satisfaction is particularly crucial 2,4 . Among multiple research instruments designed to measure relationship satisfaction, the 7-item Relationship Assessment Scale (RAS) 5,6 is one of the most widely and frequently used tools to assess satisfaction with marital and nonmarital relationships 7,8 and has been adapted for the assessment of satisfaction with other types of relationships, such as relationships with parents, friends, and other types of relatives 9 .Many psychometric aspects of the RAS have been investigated in prior research, e.g., unidimensionality, convergent, divergent, and discriminant validity, and test-retest stability 1,5,6 . However, the measurement invariance of the RAS across diverse groups and the item and test properties have not yet been examined within an Item Response Theory (IRT) framework, except in the study by Funk and Rogge 2 . These authors employed a principal-component analysis (PCA) and IRT analysis to examine a pool of 180 items originating from eight tools measuring relationship satisfaction (including the RAS) to assess these items' precision in assessing satisfaction to create a new tool that assesses relationship satisfaction 7 . Although the primary aim of Funk and Rogge's analysis was not to provide comprehensive information on the properties of the RAS, the authors showed that the RAS items are globally worded and relatively homogeneous (e.g., RAS item no. 2) and provide relatively high amounts of information to assess satisfaction 7 . In a recent Hungarian study, Fülöp et al. attempted to develop and validate a single-item version of the original RAS 8 . In two studies utilizing samples of Hungarian individuals, Fülöp et al. employed structural equation modeling (SEM) to explore the linkages between relationship satisfaction measured by the full 7-item RAS and the 1-item RAS and other proximal and distal psychological constructs 8 . Prior research has documented the link between high satisfaction, higher levels of mental and physical health, personal and family well-being among married individuals 2 , and greater emotional and psychological wellbeing 10 . Relationship satisfaction is related to greater support and less conflict 11 , and marital satisfaction is negatively related to impulsivity 12 . High relationship satisfaction is related to satisfaction with relationship status, i.e., satisfaction with having a partner 13 , while lower satisfaction is linked to problems in the mental health domain, including depression and loneliness 10,14-16 . Poorer marital functioning and low marital satisfaction are also related to work stress 17 . For instance, low marital satisfaction is associated with a lack of work enjoyment and high work involvement among female professionals 18 . At the same time, other studies have revealed that workaholism might not be related to adverse relationship outcomes, including lower relationship satisfaction 19 .Finally, high marital satisfaction was found to be related to a lower likelihood of divorce. Individuals in low-satisfaction marriages are approximately twice as likely as those in high-satisfaction marriages to divorce 20 . Notably, relationship satisfaction measured by the employment of the RAS in comparison to the DAS seemed to be a more effective discriminating factor between couples who continued their relationship and couples who broke up, making the RAS a useful tool in detecting couples who may be "at risk" of ending a relationship 6 .Relationship satisfaction across countriesCultural variation affects marital and relationship satisfaction 21 . Tai et al. demonstrated several differences in relationship satisfaction among diverse countries and showed that Australian and French individuals experienced lower satisfaction than German individuals 22 . Wiik et al. showed that cohabitation, in contrast to marriage, was related to lower quality across eight European countries 23 . The cohabitation gap in relationship satisfaction was lower in countries with a higher prevalence of cohabitation 23 . In past research, the measurement invariance of relationship quality assessed in terms of relationship satisfaction, commitment, intimacy, and trust across four diverse national samples (United States, Canada, Indonesia, and China) was tested by Gere and MacDonald 24 . The authors demonstrated that four scales measuring the above-indicated constructs met the criteria for weak measurement equivalence; however, three of them did not meet the criteria for strong measurement equivalence 24 .Relationship satisfaction and genderEarly studies 25 and recent cross-cultural research 26 revealed that men experienced higher marital satisfaction than women; however, in other studies, no gender differences were observed 2,26 . A meta-analysis of 226 studies revealed that for nonclinical community-based samples, the effect size showed a lack of significant gender differences in marital satisfaction 27 . Nevertheless, the role of gender in marital satisfaction may differ across diverse cultures as a function of culturally specific factors such as sex roles or sexual egalitarianism 28 . Although a substantial body of research has focused on sex and gender differences in relationship satisfaction and has found mixed results, we are unaware of prior research that has tested the gender invariance of the RAS.
## Relationship satisfaction across intimate relationships
A more recent comparison between different union types showed that married couples who premaritally cohabited and cohabitating couples who planned to marry did not differ with regard to relationship quality [bib_ref] Relationship quality among cohabiting versus married couples, Brown [/bib_ref]. As a result of changes in the contemporary context of intimate relationships, whether the RAS assesses the same construct of relationship satisfaction within diverse types of relationships has been questioned. This question was proposed approximately 9 years ago by Graham et al., who performed a reliability-generalization meta-analysis of seven measures of relationship satisfaction, including the RAS 1 . However, these authors did not establish the measurement invariance of the relationship satisfaction construct across diverse types of relationships, such as nonmarital, engaged, and marital relationships 1 . Therefore, there is little knowledge of the psychometric properties of the original RAS in terms of its measurement invariance across different romantic relationships.
## The current study
Although a Polish translation of the RAS exists (RAS-PL) [bib_ref] Capitalizing on the success of romantic partners: A laboratory investigation on subjective,..., Monfort [/bib_ref] and has been utilized by Polish researchers, the psychometric properties of the RAS have yet to be subjected to a solid and comprehensive investigation with a Polish sample. Thus, the primary aim of the current study was to assess the psychometric functioning of the existing Polish translation of the RAS30 to provide additional information on the psychometric properties of the RAS by assessing its measurement invariance across countries (Poland, Hungary, the USA), genders (women vs. men) and relationship type groups (marital and nonmarital relationships) and its general functioning based on an IRT analysis.
To establish country invariance, we compared three national samples from Poland (reference sample; n = 733), Hungary (n = 703), and the USA (n = 200). We chose Hungary and the USA since (a) the original RAS was developed, validated and subsequently used in samples originating from the USA, and (b) different marital and relational contexts characterize Poland and Hungary compared to the USA. Moreover, Poland and Hungary differ from each other. For instance, in Hungary, cohabitation was not prevalent until the late 1980s, but cohabitation and living apart together relationships (LAT) have become more common, whereas the marriage rate is declining [bib_ref] Choosing between marriage and cohabitation: Women's first union patterns in Hungary, Brădăţan [/bib_ref] [bib_ref] Marriage and cohabitation in recent Hungarian Society, Ragadics [/bib_ref]. Hungary falls in the mid-range of countries with the highest proportion of married people (Romania and Poland) and the lowest proportion (Norway, Austria, the Netherlands, and Belgium) [bib_ref] Marriage and cohabitation in recent Hungarian Society, Ragadics [/bib_ref]. To test whether the RAS is invariant, or unbiased, in how it assesses relationship satisfaction across countries, we utilized (a) data collected in the scope of the current investigation in Poland, (b) data collected in Hungary by and (c) the subset of data collected in the USA in the scope of another investigation conducted and described in detail by Adamczyk et al. [bib_ref] Relationship status and mental and physical health among Polish and American young..., Adamczyk [/bib_ref] www.nature.com/scientificreports/ Some previous studies provided a theoretical basis to expect gender differences in relationship satisfaction [bib_ref] Global perspective on marital satisfaction, Dobrowolska [/bib_ref] [bib_ref] Marital happiness and psychological well-being across the life course, Kamp Dush [/bib_ref]. However, given mixed findings in this domain, we considered testing gender invariance an essential and current task. Therefore, to establish invariance across gender, we compared women and men in the Polish sample (n = 733) and additional tests within Hungary and the USA. Invariance across gender was tested separately for the Polish, Hungarian, and U.S. samples since prior country invariance might not have been assumed. Analogically, establishing measurement invariance across contemporary relationship types of different levels of commitment (i.e., nonmarital through engaged to marital relationships) is also justified. Previous research has revealed differences in the quality of cohabitating and marital relationships, with more recent studies demonstrating the lack of these differences [bib_ref] Relationship quality among cohabiting versus married couples, Brown [/bib_ref]. Thus, to test invariance across relationship types, for this analysis with the Polish reference sample (n = 733), we combined participants in marital and engaged relationships into one group termed formal relationships (n = 322). In contrast, individuals who did not indicate being in a marital or engaged relationship (n = 401) were classified as a group of informal relationships. Due to the varying size of the formal and informal groups in the Hungarian and U.S. samples, we could not provide a robust invariance test for the relationship type groups in these two countries; therefore, one model is for both nonmarital relationship groups.
Third, to provide detailed information about the psychometric properties of the Polish RAS, the current study drew on Item Response Theory (IRT) methods. The prominent feature of IRT is the assumption that item responses are a function of the characteristics of the properties of the item (item parameters) and the characteristics of an individual (person parameters) [bib_ref] An introduction to using Multidimensional Item Response Theory to assess latent factor..., Osteen [/bib_ref]. Regarding the deficit of the employment of IRT methods for the original RAS, using IRT analysis to identify which Polish RAS items are most closely related to the latent construct of relationship satisfaction may contribute to a better understanding of this construct.
The integration of IRT/MIRT and CFA analyses is recommended in the literature because these two classes of analyses provide different indicators of psychometric performance; therefore, they can jointly provide complementary information pertaining to the dimensionality of instruments and item functioning [bib_ref] An introduction to using Multidimensional Item Response Theory to assess latent factor..., Osteen [/bib_ref] [bib_ref] Psychometric analysis of the adult separation anxiety symptom questionnaire: Item functioning and..., Finsaas [/bib_ref]. The simultaneous use of IRT/MIRT and CFA analyses offers several benefits. Specifically, (1) CFA provides the assessment of item fit in terms of error variances, communalities, and factor loadings, whereas in IRT, item fit is performed by unweighted (outfit) and weighted (infit) mean square errors 37 ; (2) in CFA, the link between the indicator and latent variable is limited to a linear relationship, whereas in IRT, it can be nonlinear 38 ; (3) in CFA, the factor loading is used to indicate the relationship between the indicator and the latent variable across all levels of the latent variable, whereas in IRT, this relationship is provided across the range of possible values for the latent variable [bib_ref] Managerial experience and the measurement equivalence of performance ratings, Greguras [/bib_ref]. In IRT, indices of item information functions (IIF) and test information functions (TIF) can be established; while these indices are available in CFA, they are conditional on the other items on the measure [bib_ref] An introduction to using Multidimensional Item Response Theory to assess latent factor..., Osteen [/bib_ref] ; and (4) a broad spectrum of indices assessing model fit is available in CFA, whereas in IRT, only the χ 2 deviance statistic is available to assess model fit [bib_ref] An introduction to using Multidimensional Item Response Theory to assess latent factor..., Osteen [/bib_ref].
Furthermore, convergent, divergent, and concurrent criterion validity were assessed by testing the associations between the Polish RAS and a series of psychological constructs. Specifically, the convergent and divergent validity of the Polish RAS were examined in validation sample 1 (N = 203) by testing the associations between the Polish RAS and measures of related constructs, i.e., the Couples Satisfaction Index (CSI-4) 7 assessing relationship satisfaction, the Relationship Satisfaction Status Scale (ReSta) [bib_ref] Satisfaction with relationship status: Development of a new scale and the role..., Lehmann [/bib_ref] assessing satisfaction with relationship status, and the Quality of Relationships Inventory (QRI) [bib_ref] General and relationship-based perceptions of social support: Are two constructs better than..., Pierce [/bib_ref] measuring relational depth, support and conflict, impulsive behavior (SUPPS-P) [bib_ref] Examination of a short English version of the UPPS-P Impulsive Behavior Scale, Cyders [/bib_ref] , work enjoyment and work involvement (WorkBAT) [bib_ref] Workaholism: Definition, measurement, and preliminary results, Spence [/bib_ref] and fear of being single (FBSS) [bib_ref] Settling for less out of fear of being single, Spielmann [/bib_ref].
To provide additional information on the Polish RAS's convergent and divergent validity of the Polish RAS, we utilized the subset of Polish data collected in the scope of another project by Adamczyk et al. [bib_ref] Relationship status and mental and physical health among Polish and American young..., Adamczyk [/bib_ref] [bib_ref] Psychometric analysis of the Dating Anxiety Scale for Adolescents in samples of..., Adamczyk [/bib_ref]. This subset of the data (validation sample 2; N = 209) allowed us to assess the associations between the Polish RAS and the Mental Health Continuum-Short Form 43 , which assesses emotional and psychological well-being, the SF-12v2 44 , which measures physical and mental health, the Centre for Epidemiological Studies-Depression Scale (CES-D) [bib_ref] A self-report depression scale for research in the general population, Radloff [/bib_ref] , which assesses depressive symptoms, and the Social and Emotional Loneliness Scale for Adults SELSA (SELSA-S) [bib_ref] Measurement and validity characteristics of the short version of the Social and..., Ditommaso [/bib_ref] , which measures romantic loneliness.
Finally, concurrent criterion validity was investigated by the associations between the Polish RAS and intent to continue the current relationship. Based on prior research, we expected that relationship satisfaction measured by the Polish RAS (a) would be positively associated with relationship satisfaction measured by the CSI-4, satisfaction with relationship status (ReSta), relationship depth and support (QRI), mental and physical health (the SF-12v2), and emotional and psychological well-being (MHC-SF); (b) would be negatively related to relationship conflict (QRI), impulsive behavior (SUPPS-P), depression (CES-D), romantic loneliness (SELSA-S), work enjoyment and work involvement (WorkBAT); and (c) would not be or would be weakly associated with fear of being single (FBSS). We hypothesized that higher relationship satisfaction would be related to higher intent to continue the current relationship with respect to concurrent criterion validity.
# Method
The study protocol was approved by the Ethics Committee for Research with People as Study Participants at the www.nature.com/scientificreports/ through advertisements posted on Facebook that included a description of the study's goals, informed consent, and a link to the survey. The survey took approximately 15 min to complete. The minimal sample size of the validation study was determined based on an a priori power calculation (https:// sample-size. net/ corre lation-sample-size/). Specifically, to detect small-sized correlation coefficients (0.20) with sufficient statistical power (0.80), the study would require at least 194 respondents. Furthermore, since our analyses employed the IRT-based methodology that requires a large sample size (N > 500), we allowed a larger reference sample size to be recruited to meet the requirements of IRT analysis.
Samples. Reference Polish sample. In total, 805 Polish respondents began the study; however, two did not consent to participate, and 30 were single and did not meet the criterion of being in a relationship. Hence, the final sample consisted of 733 participants, including 416 women (56.80%) and 317 men (43.20%), with a mean age of 32.70 years (ranging from 18 to 75 years). For relationship type, 401 participants (54.70%) were in nonmarital relationships, 121 were engaged (16.50%), and 211 were in marital relationships (28.80%). Further characteristics of the sample, including place of residence, highest education level, having a child/children, living with a partner, and duration of the relationship, are provided in in the online supplementary materials. in the online supplementary materials.
The second validation sample came from a study on relationship status and mental health among Polish and U.S. young adults [bib_ref] Relationship status and mental and physical health among Polish and American young..., Adamczyk [/bib_ref] [bib_ref] Psychometric analysis of the Dating Anxiety Scale for Adolescents in samples of..., Adamczyk [/bib_ref] in the online supplementary materials. A description of the study from which sample 2 was derived is provided in the papers by Adamczyk et al. [bib_ref] Relationship status and mental and physical health among Polish and American young..., Adamczyk [/bib_ref] [bib_ref] Psychometric analysis of the Dating Anxiety Scale for Adolescents in samples of..., Adamczyk [/bib_ref].
The Hungarian validation sample came from the study by Fülöp et al. [bib_ref] A two-study validation of a single-item measure of relationship satisfaction: RAS-1, Fülöp [/bib_ref]. The full sample contained 703 participants who were aged 18 to 64 years (M = 25.61, SD = 8.00), including 360 females (51.20%). Among Hungarian participants, 67 individuals (9.50%) had casual relationships, 349 individuals (49.60%) were in a nonmarital relationship but did not live with a partner, 180 individuals (25.60%) were in a nonmarital relationship and lived with a partner, 49 individuals (7.00%) were engaged, and 58 individuals (8.30%) were married. The mean duration of a relationship was 2.93 years (SD = 4.13). Detailed characteristics of the Hungarian participants are provided in the paper by Fülöp et al. [bib_ref] The utility of the relationship assessment scale in multiple types of relationships, Renshaw [/bib_ref].
The fourth U.S. sample consisted of participants enrolled in a study of relationship status and mental health among Polish and U.S. young adults [bib_ref] Psychometric analysis of the adult separation anxiety symptom questionnaire: Item functioning and..., Finsaas [/bib_ref] [bib_ref] Settling for less out of fear of being single, Spielmann [/bib_ref]. The subset of data utilized in the current study involved data from 200 U.S. participants who were in relationships, including 139 women (69.50%) and 61 men (30.50%) aged 19-40 years (M = 21.94, SD = 3.19). Among the U.S. participants, 185 were in nonmarital relationships (92.50%) and 15 were engaged (7.50%). The mean duration of a relationship was 2.18 years (SD = 2.74). Most participants had graduated from high school (n = 72, 36%). The detailed characteristics of the U.S. participants are provided in in the online supplementary materials and the papers by Adamczyk et al. [bib_ref] Relationship status and mental and physical health among Polish and American young..., Adamczyk [/bib_ref] [bib_ref] Psychometric analysis of the Dating Anxiety Scale for Adolescents in samples of..., Adamczyk [/bib_ref].
# Materials
The Relationship Assessment Scale. The Relationship Assessment Scale (RAS) [bib_ref] The Relationship Assessment Scale, Hendrick [/bib_ref] [bib_ref] A generic measure of relationship satisfaction, Hendrick [/bib_ref] in the Polish translation by Monfort et al. [bib_ref] Capitalizing on the success of romantic partners: A laboratory investigation on subjective,..., Monfort [/bib_ref] was completed by participants in the reference sample and validation sample 1. The RAS consists of 7 items assessing global relationship satisfaction (e.g., "How much do you love your partner?"). It was created by utilizing a sample of dating university students. Each item is rated on a 5-point Likert scale from 1 (i.e., not satisfied) to 5 (i.e., very satisfied). Higher scores indicate higher relationship satisfaction. In the current study, we computed a mean score. The internal consistency of the Polish RAS is provided in . The omega coefficient for the RAS was 0.90 and 0.86 in the Hungarian and U.S. samples, respectively.
## Demographic information.
Respondents in the reference sample and validation sample 1 were asked to indicate their age (in years), the gender they identified with most ("male", "female", "other"), their sexual orientation ("heterosexual", "homosexual", "bisexual", "do not know"), their place of residence (from "village" to . Descriptive and reliability statistics for the Polish RAS. RAS Relationship Assessment Scale. a The mean inter-item correlation of the original RAS was determined to be 0.49 .
## M (sd)
Cronbach's alpha Omega total www.nature.com/scientificreports/ "city > 500,000 "), their highest educational level obtained (from "primary education" to "higher education"), relationship status ("single", "partnered"), type of relationship ("informal relationship", "engaged relationship", "marriage"), living with a partner ("yes", "no"), duration of living with a partner (in months), duration of the relationship (in months), having a child/children ("yes", "no"), and intent to continue the current relationship rated on a 4-point Likert scale from 0 (not at all) to 3 (very much).
## Range of inter-item correlations
Validation questionnaires. For validation, in validation sample 1, we used the Polish versions of the Couples Satisfaction Index (CSI-4), the Relationship Satisfaction Status Scale (ReSta) [bib_ref] Development and validation of a Polish-language version of the Satisfaction with Relationship..., Adamczyk [/bib_ref] [bib_ref] Satisfaction with relationship status: Development of a new scale and the role..., Lehmann [/bib_ref] , the Quality of Relationships Inventory (QRI) [bib_ref] General and relationship-based perceptions of social support: Are two constructs better than..., Pierce [/bib_ref] , the Fear of Being Single Scale (FBSS) [bib_ref] Settling for less out of fear of being single, Spielmann [/bib_ref] , the Workaholism Battery 41 , and the Impulsive Behavior Scale Short Version (SUPPS-P) [bib_ref] Examination of a short English version of the UPPS-P Impulsive Behavior Scale, Cyders [/bib_ref]. Participants in validation sample 2, from a study by Adamczyk et al. [bib_ref] Psychometric analysis of the Dating Anxiety Scale for Adolescents in samples of..., Adamczyk [/bib_ref] , completed the Polish versions of the Mental Health Continuum-Short Form (MHC-SF) [bib_ref] The mental health continuum: From languishing to flourishing in life, Keyes [/bib_ref] , the Centre for Epidemiological Studies-Depression Scale (CES-D) [bib_ref] A self-report depression scale for research in the general population, Radloff [/bib_ref] , the Social and Emotional Loneliness Scale for Adults SELSA (SELSA-S) [bib_ref] Measurement and validity characteristics of the short version of the Social and..., Ditommaso [/bib_ref] and the SF-12v2. For details regarding the validation questionnaires in validation samples 1 and 2, see the online supplementary materials.
Data analysis. Data analysis proceeded in five steps.
The first step verified the 1-factor structure of the Polish RAS in Polish reference sample 1 using confirmatory factor analysis (CFA). Given that the RAS items are ordinal, normality was not assumed. The weighted least square mean and variance adjusted estimator was employed because it is recognized to be appropriate when the data are not normally distributed [bib_ref] On the performance of maximum likelihood versus means and variance adjusted weighted..., Beauducel [/bib_ref]. Based on recommendations provided by Whittaker [bib_ref] Structural equation modeling, Whittaker [/bib_ref] , CFI and TLI values ≥ 0.90 showed acceptable model fit, whereas RMSEA and SRMR values < 0.08 demonstrated good model fit. Due to deviations from the multivariate normality of the distribution, the correction proposed by Satorra-Bentler was applied [bib_ref] A scaled difference chi-square test statistic for moment structure analysis, Satorra [/bib_ref]. The CFA was performed with the lavaan package 51 in R version 4.2.1.
The second step assessed the internal consistency of the Polish RAS scores in reference and validation samples 1 and 2 using both Cronbach's alpha (α) and McDonald's omega (ω) coefficients. Cronbach's and McDonald's omega values of ≥ 0.70 and ≥ 0.80 were considered to demonstrate acceptable and good internal consistency, respectively [bib_ref] Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment..., Cicchetti [/bib_ref] [bib_ref] Cronbach's α, Revelle's β, and Mcdonald's ω H: Their relations with each..., Zinbarg [/bib_ref]. Furthermore, the range of interitem correlations and the mean of each interitem correlation of the Polish RAS in the Polish samples 1-3 were also determined. Internal consistencies and descriptive statistics were conducted in SPSS 27.0 (IBM SPSS Statistics).
In the third step, we tested three levels of measurement invariance between the groups distinguished based on (1) the country of origin (Poland, Hungary, and the USA), (2) gender (women and men), and (3) type of relationship (marital and nonmarital relationships). First, we tested the weakest form of measurement invariance, configural invariance, which assumes that the latent constructs (i.e., the factors) exhibit the same dimensionality and that the indicators (i.e., the items) may be identically assigned to the latent constructs in the analyzed groups [bib_ref] Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): Results..., Kliem [/bib_ref]. Establishing configural invariance is a prerequisite to performing unbiased measurement comparisons between groups 55 . To examine this type of invariance, we allowed all factor loadings and item thresholds to vary freely in each group.
Second, we examined the more restrictive form of measurement invariance, metric invariance (weak invariance), to determine whether the factor loadings were equivalent across samples [bib_ref] Assessing relationship quality across cultures: An examination of measurement equivalence, Gere [/bib_ref] [bib_ref] Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): Results..., Kliem [/bib_ref]. To assess this type of invariance, the factor loadings were constrained to be equivalent across groups, whereas the item thresholds were allowed to vary freely [bib_ref] Assessing relationship quality across cultures: An examination of measurement equivalence, Gere [/bib_ref]. The establishment of metric invariance allows for comparisons of structural relationships among latent constructs (e.g., correlation coefficients) between groups [bib_ref] Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): Results..., Kliem [/bib_ref].
Third, we assessed the scalar (strong) invariance to examine whether the item thresholds were equivalent across samples by constraining the item thresholds and factor loadings to be equivalent [bib_ref] Assessing relationship quality across cultures: An examination of measurement equivalence, Gere [/bib_ref] [bib_ref] Marital happiness and psychological well-being across the life course, Kamp Dush [/bib_ref]. Confirming strong invariance allows for the assessment of the between-group differences in the constructs' means [bib_ref] Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): Results..., Kliem [/bib_ref]. We did not test strict invariance (the equivalence of the residual variance of the items) because this level of invariance is considered to be highly restrictive and is rarely established using real data [bib_ref] Assessing relationship quality across cultures: An examination of measurement equivalence, Gere [/bib_ref]. Passing or failing the metric and scalar invariance testing was determined by the size of the change in CFI calculated by subtracting the CFI of the less-constrained model (e.g., the configural invariance model) from the more-constrained model (e.g., the metric invariance model). Invariance was supported if the change in CFI was below 0.010 and an increase in RMSEA was below 0.015 56 . When invariance was not supported, we identified the specific items that created variance across groups.
In the fourth step, the Item Response Theory (IRT) analysis with the generalized partial credit model (GPCM) scaling method was used to analyze the reliability of individual test items of the Polish RAS and for additional testing of the Hungarian RAS. Information function values and information curves were presented. The curves showed the amount of information provided for a given feature level, which was adequate for the precision of the position in the area of a given feature level [bib_ref] Itm: An R package for latent variable modelling and Item Response Theory..., Rizopoulos [/bib_ref]. IRT analyses were conducted using the ltm package 57 in R version 4.0.2.
In the last step, convergent, divergent, and concurrent criterion validity were assessed in validation samples 1 and 2 by computing Pearson correlations (two-tailed) between the Polish RAS scores and the scores on self-report measures of theoretically related constructs (i.e., CSI-4, ReSta, QRI, FBSS, SUPPS-P, WorkBAT, MHC-SF, CES-D, SELSA-S, and the SF-12v2). In turn, to assess the criterion validity of the Polish RAS in the Polish validation sample 1, a hierarchical regression analysis was performed to test whether the RAS score (dependent variable) was predictive of the intent to continue the current relationship. These analyses were conducted in SPSS 27.0 (IBM SPSS Statistics). www.nature.com/scientificreports/
# Results
Factor structure. The results of a CFA analysis in the Polish reference sample (n = 733) showed that the initial one-factor model fit the data well (χ 2 (14) = 109.252, p < 0.001, CFI = 0.998, TLI = 0.998, RMSEA = 0.058, see [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref]. The path diagram of the model is presented in [fig_ref] Figure 1: The path diagram of the one-factor model of the Polish RAS in... [/fig_ref]. All loadings for the seven items were statistically [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref] in the online supplementary materials.
Internal consistency. Country. The invariance of the 1-factor model was first tested across countries. Baseline measurement models were tested for the Polish, Hungarian, and U.S. samples separately and fit the data adequately, Poland: . The configural invariance model was then tested and fit the data adequately, indicating that the factor structure was invariant across countries (see [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref] for configural model fit statistics). Next, the metric (weak) invariance model, which constrained all factor loadings to invariance, was tested and compared with the configural invariance model. This model fit the data adequately according to the Δχ 2 and SRMR because the threshold for Δ RMSEA was slightly exceeded (see [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref]. Finally, the scalar (strong) invariance model was tested and compared with the metric invariance model, which constrained all item thresholds to invariance. This model fit well to the data. A decrease in CFI greater than 0.01 and an increase in RMSEA greater than 0.015 were not observed in this case (see [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref].
Because the U.S. sample was smaller and uneven group sizes might attenuate the sensitivity to detect noninvariant parameters in multiple group CFA analyses, we used Monte Carlo sampling techniques to test invariance across countries [bib_ref] Testing factorial invariance with unbalanced samples, Yoon [/bib_ref]. We randomly chose 200 participants from Poland and Hungary (the size of the U.S. sample) . Model fit for multiple group by gender models and measurement invariance comparisons. CFI comparative fit index, TLI Tucker-Lewis index, RMSEA root mean square error of approximation, CI confidence interval, df degrees of freedom, SRMR squared root mean residuals; S-B Satorra-Bentler correction, Δ comparisons of nested model: configural-metric, metric-scalar. In the U.S. sample, one outlier was excluded due to Cook's distance above 1. *p < 0.05, ***p < 0.001. www.nature.com/scientificreports/ and computed all fit indices. We repeated this procedure 100 times. Generally, the results were the same as in the approach above; only metric (weak) invariance needed freed loadings to accomplish the level of fit. Model modification indices were examined to identify whether particular item loadings contributed significantly to the model misfit, and a partial metric model with freed loadings for item no. 1 ("How well does your partner meet your needs?") was tested. The loading for item 1 was lower in the U.S. sample than in the Polish sample and the Hungarian sample, suggesting that the assessment of the degree to which a current partner meets an individual's needs is more strongly related to relationship satisfaction in the Polish and Hungarian samples than among U.S. participants. The results are displayed in [fig_ref] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries [/fig_ref].
Gender. The invariance of the 1-factor model was tested across women and men in each of the countries. Baseline measurement models were fit for women and men separately and fit the data adequately (see for model fit statistics for each group). For all samples, we achieved scalar (strong) invariance for gender.
Relationship types. Finally, the invariance of the 1-factor model was examined across relationship type groups only within the Polish reference sample, in which we achieved scalar (strong) invariance. The data fit the 1-factor model well for every level of the restrictions (see [fig_ref] Table 4: Model Fit for Multiple Groups by the Relationship Type Models and Measurement... [/fig_ref].
Test and item properties. After meeting the IRT assumptions (unidimensionality: Velicer MAP has a minimum of 0.04 for 1 factor; local independence: pairwise correlation indices do not exceed 0.70), we determined the test and item properties of the Polish RAS (see [fig_ref] χ 2 14: [/fig_ref] and [fig_ref] Figure 1: The path diagram of the one-factor model of the Polish RAS in... [/fig_ref] in the online supplementary materials). In addition, to provide more robust results, we performed an IRT analysis using the data collected in [fig_ref] χ 2 14: [/fig_ref] and in the online supplementary materials). As [fig_ref] χ 2 14: [/fig_ref] shows, the total information curve indicates that the Polish RAS provides the most information or assesses with the slightest error, at approximately 2 SDs below the mean of latent relationship satisfaction. It provides very little information above 2 SDs. In the Hungarian sample, we determined that the RAS also provides the most information or assesses with the slightest error at approximately 2 SDs below the mean of latent relationship satisfaction (see [fig_ref] χ 2 14: [/fig_ref] in the online supplementary materials). [fig_ref] χ 2 14: [/fig_ref] also presents the histogram that depicts the number of cases at all levels of latent relationship satisfaction. The scatterplot of factor scores against sum scores depicts the impact of weighting items by the discrimination parameters. At higher levels of latent relationship satisfaction (approximately 1 SD and above), the sum and factor scores were less strongly related. Analogically, in the Hungarian sample, with higher levels of latent relationship satisfaction (approximately 1 SD and above), the sum and factor scores were less strongly related.
The estimated item parameters (i.e., item discrimination and difficulty) are presented in . The Polish RAS items no 2 ("In general, how satisfied are you with your relationship?") had the highest discrimination, whereas item no. 7 had the lowest discrimination among the seven RAS items (see . The four sets of difficulty parameters were quite variable (1: − 1.920 to − 2.359; 2: − 1.847 to − 2.858; 3: − 0.936 to − 1.670; 4: − 0.841 to 0.561). These parameters were based on item thresholds and demarcate response categories 0 from 1, 1 from 2, 2 from 3, 3 from 4. Very high values are undesirable since they indicate that the measured construct is conceptually narrow. Good values should fit ranges of 0.5-2.5. The data in show that the estimated parameters were of reasonable to very high magnitude except for items no. 1 and no. 2. Additionally, the estimated parameters for item no. 2 in the Hungarian sample were very high, whereas the parameters for item no. 1 were of acceptable magnitude (for the rest of the items, the parameters ranged from 0.0864 to 2.541; see in the online supplement material).
The inspection of the threshold parameters for the Polish RAS showed that most of the items were located to the left of the mean θ, indicating that most of the items were uninformative about individual differences at the range of the θ scale, where a distinction is made between moderately high levels of relationship satisfaction and very high levels. This is reflected in the test information function, which drops sharply above the mean θ in the Polish and Hungarian samples. From the inspection of the parameter estimates and test information functions for the Polish and Hungarian samples, it can be concluded that both scales discriminate best between individuals with low scores and average scores. Convergent, divergent, and criterion validity. The convergent, divergent, and criterion validity of the Polish RAS were assessed using data collected in the current validation sample 1 and validation sample 2 36,42 (see .
As demonstrated in , the results show weak to medium, large and very large correlations of the RAS score with all theoretically related constructs in both samples except for the constructs of workaholism and fear of being single. As expected, the RAS was positively correlated with relationship satisfaction assessed by a different instrument (CSI-4) and relationship quality measured in terms of relationship depth and support as well as positive outcomes such as mental health and emotional and psychological well-being. Furthermore, the RAS showed the expected opposite pattern in which the RAS was negatively correlated with relationship conflict, impulsive behavior, depressive symptoms, and romantic loneliness. Finally, relationship satisfaction was not correlated with two workaholism dimensions, i.e., work enjoyment (enjoying work to such a degree that it is difficult to stop working) and work involvement (the drive to work that involves the feeling of the obligation to work hard and thinking about work even when an individual wants to avoid it) or with the construct of fear of being single.
Concerning the assessment of the concurrent criterion validity of the Polish RAS, we performed a hierarchical regression analysis in which the intent to continue the current relationship was predicted from the demographic and relational variables entered in Step 1 (age, gender, type of a relationship, living vs. not living with a partner, duration of the relationship, duration of living with a partner, having children) and the RAS score in Step 2. The . Item Difficulty Location and Thresholds For Informational Curves for the Polish RAS. n = 733; a = discrimination parameter estimates in the RAS; b-1 = difficulty parameter estimates between response categories 0 and 1; b-2 = difficulty parameter estimates between response categories 1 and 2; b-3 = difficulty parameter estimates between response categories 2 and 3; b-4 = difficulty parameter estimates between response categories 3 and 4. Area under curve with range [− 3, 3] SD raw and percent, compared to total result of information function. www.nature.com/scientificreports/ analysis revealed that in the last step, the only significant, positive predictors of the intent to continue the current relationship were living vs. not living with a partner (β = 0.16, p = 0.008) and the RAS score (β = 0.65, p < 0.001) as the strongest predictors. This model explained 38% of the variance to continue the current relationship. This means that individuals with higher relationship satisfaction and those living with a partner had higher intent to continue their current relationships. For the detailed results of the analysis, see [fig_ref] Table 4: Model Fit for Multiple Groups by the Relationship Type Models and Measurement... [/fig_ref] in the online supplementary materials.
[formula] Items a SE b-1 b-2 b-3 b-4 [/formula]
# Discussion
The primary aim of the current investigation was to examine the Relationship Assessment Scale (RAS) with respect to its unidimensionality, invariance across countries, gender, formal and informal relationships, degree of precision (or information) across latent levels of relationship satisfaction, and the functioning of individual items. Measurement invariance analysis using multiple-group CFA supported configural measurement invariance across countries (Poland, Hungary, USA), genders (women and men), and relationship types (formal and informal relationships). This means that the seven items of the RAS were representative of the relationship satisfaction construct across the analyzed groups. Our analyses also demonstrated that the RAS achieved metric (weak) measurement invariance across countries, genders, and relationship types. This means that each RAS item had equal loadings on the relationship satisfaction factor across the country, gender, and relationship type groups and each item had an equal contribution to the total score on the relationship satisfaction construct, which is the most critical criterion for establishing construct validity [bib_ref] Confirmatory factor analytic procedures for the determination of measurement invariance, French [/bib_ref]. The establishment of metric (weak) measurement invariance also means that the RAS may be utilized in examining the associations between theoretical constructs, such as relationship satisfaction and other constructs 24 , in groups of Polish, Hungarian, and U.S. individuals, women and men, and groups of individuals in formal and informal relationships. We noted that item no. 1, "How well does your partner meet your needs?", did not have equal loadings on the satisfaction factor in the USA compared to Poland and Hungary. This finding implies that item no. 1 may be less relevant to relationship satisfaction in the USA compared to Poland and Hungary 62 .
Finally, we established full scalar (strong) invariance across countries, genders, and relationship type groups; that is, we demonstrated the equivalence of the factor loadings and the thresholds of items across the analyzed groups. The measurement invariance at the level of the thresholds indicates that items measure features of the trait that manifest to similar degrees in each of the countries, genders, and relationship types, contributing to similar probabilities of endorsement of the items [bib_ref] Measurement invariance of three narcissism questionnaires across the US, the UK, and..., Wetzel [/bib_ref]. This finding means that the RAS met the criterion of strong invariance, which allowed us to perform the mean-level differences across the analyzed groups [bib_ref] Assessing relationship quality across cultures: An examination of measurement equivalence, Gere [/bib_ref].
Concerning the characteristics of individual RAS items (specifically, item difficulty and discrimination) and the information (or precision) of the total RAS, the major finding is that the Polish and Hungarian versions of the RAS were found to provide the most accurate assessment, or the most information, at low levels of latent RAS. We demonstrated this in several ways, including the test information and individual item information curves, which peaked near the middle of latent RAS, and the strong positive relationship between the total raw scores and the model-estimated factor scores below the mean of latent RAS, which showed that differences between individuals at low levels are more meaningful than differences between individuals at high levels. For future research, this finding may imply the need to consider modifying the Likert scale employed in the RAS www.nature.com/scientificreports/ to increase the differentiation of options depicting the values at the high end of the scale (the end of the scale depicting higher levels of relationship satisfaction). The low precision at higher levels of relationships suggests that the RAS is most suitable for assessing individuals with low to average relationship satisfaction levels. Although the instrument would be expected to have high precision across the entire trait continuum 63 , we also note a benefit of this precision in the low to average ranges of the relationship satisfaction continuum. Specifically, the RAS appears to be reliable in detecting individuals or couples who are in relationships with low satisfaction, which, as we have already noted, puts them at risk of negative individual and relational outcomes, including the risk of divorce/break up [bib_ref] Patterns of change in marital satisfaction over the newlywed years, Lavner [/bib_ref] , depression [bib_ref] Marital satisfaction and depression: Different causal relationships for men and women?, Fincham [/bib_ref] and loneliness [bib_ref] Development and validation of a Polish-language version of the Satisfaction with Relationship..., Adamczyk [/bib_ref] [bib_ref] Lonely me, lonely you: Loneliness and the longitudinal course of relationship satisfaction, Mund [/bib_ref]. For instance, most clients in couple therapy experience depression and/or problems in their relationships [bib_ref] Couples' depression and relationship satisfaction: Examining the moderating effects of demand/withdraw communication..., Li [/bib_ref].
To explain the precision of the RAS in low ranges of relationship satisfaction, we may refer to an analogy between the relationship satisfaction construct and the term "quasitrait" proposed by Reise and Waller 63 regarding psychopathology constructs. Reise and Waller 63 suggested that the peaked (in the severe trait range) information curve in IRT analyses of clinical scales arises from the unipolar character of the trait and is relevant only in one direction. Although we do not conceptualize relationship satisfaction as a psychopathology construct, we may tentatively consider the unipolar character of relationship satisfaction with reverse relevance involving low but not high ends of the continuum. Specifically, the relevant low end might indicate dissatisfaction with a relationship (negativity in a relationship), while the high end would not indicate relationship satisfaction (positivity in a relationship) but rather a lack of dissatisfaction with a relationship (the absence of negativity). As a result, if relationship satisfaction were considered a unipolar construct, it would be consequential for the employment of the RAS. According to Reise and Waller 63 , in the case of unipolar constructs, researchers often suggest rewriting the items included in the instrument that measure such constructs to elaborate items that provide information spread across the entire trait continuum, which is difficult to achieve. Finally, relationships and relationship satisfaction may be more affected by the absence of negative aspects of relationships than by the presence of diverse positive aspects of relationships [bib_ref] Bad is stronger than good, Baumeister [/bib_ref]. Therefore, the RAS is a reliable measure of relationship satisfaction at low levels of the relationship satisfaction continuum that captures the presence of negativity in a relationship but not the absence of positivity in a relationship.
Regarding the validity of the Polish RAS, the results were satisfactory for the convergent, divergent, and criterion validity tests. These results were consistent with those of previous studies [bib_ref] Development and validation of a Polish-language version of the Satisfaction with Relationship..., Adamczyk [/bib_ref]. Furthermore, in line with past studies, we found that relationship satisfaction was negatively associated with impulsive behavior [bib_ref] Impulsivity, communication, and marital satisfaction in newlywed couples, Tan [/bib_ref] and was not related to fear of being single [bib_ref] Development and validation of a Polish-language version of the Satisfaction with Relationship..., Adamczyk [/bib_ref]. We also identified the lack of a link between workaholism and relationship satisfaction. Our results corroborate prior findings that showed the lack of a link between relationship adjustment and relationship satisfaction and workaholism [bib_ref] Workaholism: Definition, measurement, and preliminary results, Spence [/bib_ref].
Finally, the concurrent validity of the Polish RAS was tested, and the results showed that relationship satisfaction was a strong predictor of the intent to continue the current relationship (in connection with a weak predictor of living vs. not living with a partner). Although various factors determine the stability of a relationship 66 , individuals in the current study who were more satisfied with their relationships were also more willing to continue their relationships, which suggests that they may also be less willing to consider terminating their relationships. This finding is consistent with literature showing the links between low relationship satisfaction and the risk of divorce/break-up [bib_ref] Patterns of change in marital satisfaction over the newlywed years, Lavner [/bib_ref] and between higher relationship satisfaction and greater relationship stability [bib_ref] Relationship effort, satisfaction, and stability: Differences across union type, Shafer [/bib_ref]. Overall, these results verify the convergent, divergent, and criterion validity of the Polish RAS and provide a multilevel assessment of relationship satisfaction that can be applied to evaluate the links between relationship satisfaction and diverse outcomes, including mental and physical health, emotional and psychological well-being, depression and romantic loneliness.
Limitations. The current analyses have both strengths and important limitations. The Polish target samples utilized in the current study were recruited via Facebook, so our sample was not representative of the entire Polish population. It excluded participants who do not use the internet or social media, including Facebook. Furthermore, we could not determine whether individuals who participated in the reference and validation samples 1 and 2 experienced higher or lower relationship satisfaction compared to participants who did not participate in the studies. Furthermore, the Polish reference and validation samples (and the Hungarian sample) included participants who were in nonmarital, engaged and marital relationships. Therefore, our results cannot be extended to individuals whose relational history includes the experience of divorce or widowhood and who have remarried or repartnered. The majority of participants in the Polish samples were childless, limiting the generalizability of our findings to parents. It is also important to note that the ratio of the Hungarian and Polish sample sizes to the U.S. sample size was not high. Since we utilized previously collected data in the USA, we could not increase the U.S. sample size to the required sample size of N > 500 for IRT analyses, which might affect the power of the IRT model [bib_ref] Checking equity: Why differential item functioning analysis should be a routine part..., Martinková [/bib_ref]. In addition, our focal groups of individuals in formal and informal relationships did not reach the size of N > 500. Therefore, we encourage other researchers to obtain the required sample size for the focal groups when employing the IRT methodology. Finally, both the Hungarian and U.S. data and the Polish validation sample 2 utilized in this paper were collected on different occasions than the data collected for the Polish reference and validation sample 1. Notably, in Poland, the data in the reference and validation sample 1 were collected during the period of the COVID-19 pandemic. Thus, the possibility cannot be excluded that COVID-19-related distress might affect the obtained results because the COVID-19 pandemic produced adverse psychological consequences in the domain of mental health [bib_ref] The Polish COVID Stress Scales: Considerations of psychometric functioning, measurement invariance, and..., Adamczyk [/bib_ref]. Furthermore, some recent studies have revealed that both men and women experienced decreased relationship satisfaction before and during the COVID-19 pandemic 69 . www.nature.com/scientificreports/
## Data availability
The raw and processed data collected in the Polish reference and validation samples 1 and 2, the data collected in the United States and the analysis code for this study are available at the Open Science Framework (OSF) repository (https:// osf. io/ 7hv32/? view_ only= 53083 d2e28 69472 490be ece1b 35b72 94). The data collected in Hungary are available from the Hungarian authors by e-mail request [bib_ref] A two-study validation of a single-item measure of relationship satisfaction: RAS-1, Fülöp [/bib_ref]
[fig] Figure 1: The path diagram of the one-factor model of the Polish RAS in the Polish reference sample (N = 733). RAS Relationship Assessment Scale. Scientific Reports | (2022) 12:22157 | https://doi.org/10.1038/s41598-022-26653-6 www.nature.com/scientificreports/ significant (p < 0.001). The loadings of each item on the relationship satisfaction latent construct in the 1-factor model are reported in [/fig]
[fig] χ 2: 14) = 109.252, p < 0.001, CFI = 0.998, TLI = 0.998, RMSEA = 0.058, 90% CI [0.048, 0.069], Hungary: 14) = 119.282, p < 0.001, CFI = 0.998, TLI = 0.996, RMSEA = 0.063, 90% CI [0.053, 0.074], U.S.: 14) = 57.521, p < 0.001, CFI = 0.997, TLI = 0.995, RMSEA = 0.083, 90% CI [0.061, 0.106] [/fig]
[fig] 6: Convergent and divergent validity of the Polish Relationship Assessment Scale. Numbers represent correlations. CSI Couples Satisfaction Index, ReSta Relationship Satisfaction Status Scale, QRI Quality of Relationships Inventory, SF-12v2 SF-12v2 ® Health Survey, CES-D The Center for Epidemiologic Studies Depression Scale, MHC-SF Mental Health Continuum-Short Form, SELSA-S Social and Emotional Loneliness Scale for Adults-Short Form, SUPPS-P The Impulsive Behavior Scale Short Version, WorkBAT Workaholism Battery, FBSS Fear of Being Single Scale. [/fig]
[table] Table 2: Model fit for multiple group models and measurement invariance comparisons across countries. 010. * based on SRMR < 0.06, because CFI and RMSEA indices are inconsistent. CFI comparative fit index, TLI Tucker-Lewis index, RMSEA root mean square error of approximation, CI confidence interval, df degrees of freedom, SRMR squared root mean residuals, S-B Satorra-Bentler correction, Δ comparisons of nested model: configural-metric, metric-scalar. [/table]
[table] Table 4: Model Fit for Multiple Groups by the Relationship Type Models and Measurement Invariance Comparisons. CFI comparative fit index, TLI Tucker-Lewis index, RMSEA root mean square error of approximation,CI confidence interval, df degrees of freedom, SRMR squared root mean residuals, S-B Satorra-Figure 2. Test information curve for the Polish RAS, histogram of relationship satisfaction factor scores, and a scatterplot of raw v. factor scores for the Polish sample. n = 733. RAS Relationship Assessment Scale. [/table]
|
Frailty syndrome: implications and challenges for health care policy
Older adults are a highly heterogeneous group with variable health and functional life courses. Frailty has received increasing scientific attention as a potential explanation of the health diversity of older adults. The frailty phenotype and the Frailty Index are the most frequently used frailty definitions, but recently new frailty definitions that are more practical have been advocated. Prevalence of frailty among the community-dwelling population aged 65 years and older is ~10% but varies depending on which frailty definitions are used. The mean prevalence of frailty gradually increases with age, but the individual's frailty level can be improved. Older adults, especially frail older adults, form the main users of medical and social care services. However, current health care systems are not well prepared to deal with the chronic and complex medical needs of frail older patients. In this context, frailty is potentially a perfect fit as a risk stratification paradigm. The evidence from frailty studies has not yet been fully translated into clinical practice and health care policy making. Successful implementation would improve quality of care and promote healthy aging as well as diminish the impact of aging on health care systems and strengthen their sustainability. At present, however, there is no effective treatment for frailty and the most effective intervention is not yet known. Based on currently available evidence, multi-domain intervention trials, including exercise component, especially multicomponent exercise, which includes resistance training, seem to be promising. The current challenges in frailty research include the lack of an international standard definition of frailty, further understanding of interventions to reverse frailty, the best timing for intervention, and education/training of health care professionals. The hazards of stigmatization should also be considered. If these concerns are properly addressed, widespread application of public health approaches will be possible, including screening, identification, and treatment of frailty, resulting in better care and healthier aging for older people.
# Introduction
Life expectancy has markedly increased worldwide during the past 100 years, mainly due to public health improvements.This demographic transformation of the population has resulted in growing numbers of older adults in both developing and developed countries.Between 2000 and 2050, the proportion of people aged 60 years or older in the world is projected to double from about 11% to 22%, an increase from 605 million to 2 billion adults aged ≥60 years.The number of older adults aged over 80 years is expected to quadruple to 395 million during the same period.In general, we tend to develop more health problems and become frailer as we age. The increase in submit your manuscript | www.dovepress.com Dovepress Dovepress 24 life expectancy allows chronic diseases to develop while physical and cognitive functions decline, which predisposes older people to disability or dependency.
Older adults are a highly heterogeneous group. Their life courses of health and functional status vary substantially, depending on their genetic, biological, and environmental backgrounds as well as other physical, psychological, and social factors. Therefore, individuals with the same chronological age can have different biological ages. [bib_ref] Frailty, fitness and late-life mortality in relation to chronological and biological age, Mitnitski [/bib_ref] Frailty has received increasing scientific attention as a way of understanding health diversity among older adults. [bib_ref] Frailty in elderly people, Clegg [/bib_ref] In the past, the term frailty was used almost interchangeably with aging, disability, or comorbidity, partly because of the similarity and high coexistence rate of these descriptive states. [bib_ref] Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting..., Fried [/bib_ref] However, there are clear differences between frailty, aging, disability, and comorbidity. First, advanced age on its own does not necessarily mean vulnerability to negative health outcomes so typical of frailty. [bib_ref] Old or frail: what tells us more?, Schuurmans [/bib_ref] Frailty is at least partly programmed in early life and is also associated with lower socio-economic status in adulthood. [bib_ref] Early life determinants of frailty in old age: the Helsinki Birth Cohort..., Haapanen [/bib_ref] Second, frailty is conceptualized as a state of decreased physiological reserve and compromised capacity to maintain homeostasis as a consequence of agerelated, multiple, accumulated deficits. [bib_ref] Frailty in elderly people, Clegg [/bib_ref] Frail older people are highly vulnerable to adverse health outcomes when exposed to an internal or external stressor. [bib_ref] Frailty in elderly people, Clegg [/bib_ref] Third, whereas frailty refers to instability and risk of loss of function, disability indicates loss of function and often assessed based on difficulty or dependency in performing activities necessary to live independently, such as activities of daily living (ADL), eg, bathing, dressing, eating, toileting, continence, and transferring 8 and instrumental activities of daily living (IADL), eg, shopping, telephone use, meal preparation, housekeeping, laundry, transportation, medication, and finances. [bib_ref] Assessment of older people: self-maintaining and instrumental activities of daily living, Lawton [/bib_ref] Finally, comorbidity is defined as having two or more medically diagnosed diseases. [bib_ref] Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting..., Fried [/bib_ref] Thus, frailty is clearly different and distinguishable from advanced age, disability, and comorbidity.
## Definition of frailty
A number of definitions have been proposed to conceptualize and operationalize frailty. [bib_ref] Toward a conceptual definition of frail community dwelling older people, Gobbens [/bib_ref] [bib_ref] The identification of frailty: a systematic literature review, Sternberg [/bib_ref] Despite the long-lasting and extensive debates on how best to define frailty, international consensus has yet to be reached and a gold standard definition of frailty is still lacking. [bib_ref] Frailty in elderly people, Clegg [/bib_ref] Nonetheless, most conceptual frailty definitions have some factors in common, such as decreased reserves/capacity to tolerate minor stressors, increased vulnerability to adverse health outcomes, and impairment in multiple physiological systems. [bib_ref] Toward a conceptual definition of frail community dwelling older people, Gobbens [/bib_ref] After the inclusion of these factors, frailty is conceptually defined as "a clinically recognizable state in which the ability of older people to cope with everyday or acute stressors is compromised by an increased vulnerability brought by age-associated declines in physiological reserve and function across multiple organ systems".This definition by the WHO has been widely accepted 4,13 and adopted in the Joint Action ADVANTAGE, a recently launched European Union (EU) initiative.Among various frailty definitions, the most commonly used is the frailty phenotype, [bib_ref] Frailty assessment instruments: systematic characterization of the uses and contexts of highly-cited..., Buta [/bib_ref] developed by Fried et al using the Cardiovascular Health Study cohort in 2001. [bib_ref] Frailty in older adults: evidence for a phenotype, Fried [/bib_ref] The Fried frailty phenotype consists of five physical components to define frailty: unintentional weight loss, self-reported exhaustion, weakness, slow walking speed, and low physical activity. [bib_ref] Frailty in older adults: evidence for a phenotype, Fried [/bib_ref] Individuals are considered to be frail when they meet three or more criteria, and they are considered to be robust when they have none. [bib_ref] Frailty in older adults: evidence for a phenotype, Fried [/bib_ref] Individuals who have one or two criteria are defined as prefrail, a state between robust and frail. [bib_ref] Frailty in older adults: evidence for a phenotype, Fried [/bib_ref] The Frailty Index is another popular approach, [bib_ref] Frailty assessment instruments: systematic characterization of the uses and contexts of highly-cited..., Buta [/bib_ref] based on a cumulative deficit model advocated by Mitnitski et al using the Canadian Study on Health and Aging. [bib_ref] Accumulation of deficits as a proxy measure of aging, Mitnitski [/bib_ref] In contrast to the frailty phenotype, this cumulative deficit approach describes frailty as a state caused by the accumulation of health deficits during the life course, and the more deficits individuals have, the more likely they are to be frail. [bib_ref] Frailty assessment instruments: systematic characterization of the uses and contexts of highly-cited..., Buta [/bib_ref] The Frailty Index, a continuous score ranging from 0 (no deficit) to 1 (all deficits present), is calculated as the ratio of the number of deficits present in the individual to the number of total deficits considered. [bib_ref] A standard procedure for creating a frailty index, Searle [/bib_ref] For the Frailty Index, at least 30-40 deficits, associated with age and adverse health outcomes that are prevalent in 1% or more of the population, should be considered, with fewer than 5% of missing values, and it can include symptoms, signs, diseases, disabilities, abnormalities of laboratory, radiographic, and electrocardiographic findings, and social characteristics. [bib_ref] A standard procedure for creating a frailty index, Searle [/bib_ref] There has been some debate regarding the practical application of these two most commonly used frailty definitions. [bib_ref] Frailty measurement in research and clinical practice: a review, Dent [/bib_ref] The frailty phenotype requires special equipment to measure the handgrip strength and space for measuring gait speed. The population-based lowest 20% of the handgrip strength, gait speed, and physical activity need to be calculated based on the population distributions. Calculation of the Frailty Index requires the recording of various types of deficits (typically more than [bib_ref] Prevalence of frailty in end-stage renal disease: a systematic review and meta-analysis, Kojima [/bib_ref] [bib_ref] The prevalence of frailty in heart failure: a systematic review and meta-analysis, Denfeld [/bib_ref] [bib_ref] Prevalence of frailty in mild to moderate Alzheimer's disease: a systematic review..., Kojima [/bib_ref] [bib_ref] Prevalence of frailty in nursing homes: a systematic review and meta-analysis, Kojima [/bib_ref] [bib_ref] Prevalence of frailty and associated factors in the community-dwelling population of China, Ma [/bib_ref] [bib_ref] Prevalence of frailty in Canadians 18-79 years old in the Canadian Health..., Kehler [/bib_ref] [bib_ref] Prevalence and correlates of frailty in an older rural African population: findings..., Payne [/bib_ref] [bib_ref] The prevalence and characteristics of frailty by frailty phenotype in rural Tanzania, Lewis [/bib_ref] [bib_ref] Transitions between frailty states among community-dwelling older people: a systematic review and..., Kojima [/bib_ref] [bib_ref] Frailty transitions in the San Antonio longitudinal study of aging, Espinoza [/bib_ref] [bib_ref] Transition patterns of frailty syndrome in community-dwelling elderly individuals: a longitudinal study, Azevedo Da Silva [/bib_ref] and dividing the number of deficits present by the number of deficits considered, which may take ~20-30 minutes, [bib_ref] Frailty measurement in research and clinical practice: a review, Dent [/bib_ref]
## Feasible frailty instruments in a clinical setting
According to the International Association of Nutrition and Aging (IANA) Task Force, a frailty tool should be quick, inexpensive, reliable, and easy to use in clinical settings because the identification of frail older people at risk is an important initial step potentially leading to appropriate preventive and/ or treatment interventions and ultimately to higher quality care for this vulnerable population. [bib_ref] The I.A.N.A. Task Force on frailty assessment of older people in clinical..., Abellan Van Kan [/bib_ref] From this perspective, the frailty phenotype and the Frailty Index may be rather impractical and unfeasible especially in a busy clinical setting. Based on a systematic review of the literature as well as input from a panel of geriatric experts, the IANA's working group advocated a new frailty tool. [bib_ref] The I.A.N.A. Task Force on frailty assessment of older people in clinical..., Abellan Van Kan [/bib_ref] The FRAIL scale is a simple tool consisting of five yes/no questions: Fatigue, Resistance (inability to climb stairs), Ambulation (inability to walk a certain distance), Illnesses (more than five of comorbidities), and Loss of weight (more than 5%), [bib_ref] The I.A.N.A. Task Force on frailty assessment of older people in clinical..., Abellan Van Kan [/bib_ref] and has been shown to be able to predict mortality and incident ADL and IADL disabilities among community-dwelling older people in recent meta-analysis studies. [bib_ref] Frailty defined by FRAIL scale as a predictor of mortality: a systematic..., Kojima [/bib_ref] [bib_ref] Quick and simple FRAIL scale predicts incident activities of daily living (ADL)..., Kojima [/bib_ref] The FRAIL scale is also recommended as one of the tools to detect frailty by the Joint Action ADVANTAGE, a EU co-funded initiative launched in 2017 involving 22 member states and over 40 organizations.The main goal of ADVANTAGE is to establish a common European framework for addressing the problems of frailty, including: 1) improvements in screening, diagnosis, prevention, and treatment for frailty, 2) health care system reforms adapted to population aging, and 3) facilitation of research and education on frailty.ADVANTAGE has proposed that tools used for frailty screening should be quick to administrate (no more than 10 minutes to complete); require no special equipment; validated; and be meant for screening. Existing frailty instruments meeting these four criteria are Clinical Frailty Scale, Edmonton Frail Scale, FRAIL scale, INTER-FRAIL, Prisma-7, Sherbrooke Postal Questionnaire, Short Physical Performance Battery, and Study of Osteoporotic Fractures Index. 14 The Kihon Checklist, a self-reported comprehensive questionnaire consisting of 25 simple yes/no questions covering multiple domains, is another relatively new frailty tool. [bib_ref] Systematic review of the Kihon Checklist: is it a reliable assessment of..., Sampaio [/bib_ref] [bib_ref] Prevalence of frailty in Japan: a systematic review and meta-analysis, Kojima [/bib_ref] This tool was originally developed by the Japanese Ministry of Health, Labour and Welfare in 2005-2006 as a screening tool to identify vulnerable older adults who are at high risk of dependency and more recently has been recognized as a useful frailty assessment tool. [bib_ref] Validity of the Kihon Checklist for assessing frailty status, Satake [/bib_ref] [bib_ref] Are the Kihon Checklist and the Kaigo-Yobo checklist compatible with the frailty..., Kojima [/bib_ref] This is another brief, simple, quick, and cost-effective instrument which does not need special equipment and takes <10 minutes to complete, [bib_ref] Frailty measurement in research and clinical practice: a review, Dent [/bib_ref] and therefore may be appropriate for screening. These frailty instruments can easily be incorporated into comprehensive geriatric assessment or primary care in a clinical setting for screening frail older adults. Some of them consist of short lists of simple questions and can be administered by not only physicians or other health care professionals but also care givers and non-professionals, in person as well as by phone, mail, or email.
## Prevalence and natural course of frailty
The mean prevalence of frailty among the communitydwelling population aged 65 years and older is ~10% but can range widely from 4.0% to 59.1% depending on the frailty criteria used. [bib_ref] Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic..., Collard [/bib_ref] Advanced age is a significant risk factor for frailty and a quarter of those aged 80 years or older are frail. [bib_ref] Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic..., Collard [/bib_ref] A higher prevalence of frailty is also observed in selected populations with specific diseases or conditions, such as patients with cancer (42%), 29 end-stage renal disease (37%), [bib_ref] Prevalence of frailty in end-stage renal disease: a systematic review and meta-analysis, Kojima [/bib_ref] heart failure (45%), [bib_ref] The prevalence of frailty in heart failure: a systematic review and meta-analysis, Denfeld [/bib_ref] Alzheimer disease (32%), [bib_ref] Prevalence of frailty in mild to moderate Alzheimer's disease: a systematic review..., Kojima [/bib_ref] and nursing home residents (52%). [bib_ref] Prevalence of frailty in nursing homes: a systematic review and meta-analysis, Kojima [/bib_ref] Although the mean prevalence of frailty gradually increases with age, 28,34-37 the individual course of frailty varies and the level of frailty can be reduced even in old age. [bib_ref] Frailty in elderly people, Clegg [/bib_ref] [bib_ref] Transitions between frailty states among community-dwelling older people: a systematic review and..., Kojima [/bib_ref] Several longitudinal population-based studies have showed that 8.3%-17.9% of older adults actually improved their frailty status [bib_ref] Frailty transitions in the San Antonio longitudinal study of aging, Espinoza [/bib_ref] [bib_ref] Transition patterns of frailty syndrome in community-dwelling elderly individuals: a longitudinal study, Azevedo Da Silva [/bib_ref] [bib_ref] Transitions in frailty states among community-living older adults and their associated factors, Lee [/bib_ref] [bib_ref] Frailty transitions and types of death in Chinese older adults: a population-based..., Liu [/bib_ref] [bib_ref] Patterns and predictors of frailty transitions in older men: the osteoporotic fractures..., Pollack [/bib_ref] [bib_ref] Factors influencing transitions between frailty states in elderly adults: the Progetto Veneto..., Trevisan [/bib_ref] and that some of them made frequent and dynamic transitions over time. [bib_ref] Transitions between frailty states among community-living older persons, Gill [/bib_ref]
## Impacts of frailty on health care systems
Frail older adults are at increased risk of premature death [bib_ref] Evaluating frailty scores to predict mortality in older adults using data from..., Stow [/bib_ref] [bib_ref] Frailty defined by FRAIL scale as a predictor of mortality: a systematic..., Kojima [/bib_ref] [bib_ref] Are the Kihon Checklist and the Kaigo-Yobo checklist compatible with the frailty..., Kojima [/bib_ref] [bib_ref] Frailty index as a predictor of mortality: a systematic review and meta-analysis, Kojima [/bib_ref] and various negative health outcomes, including falls, [bib_ref] Frailty as a predictor of future falls among communitydwelling older people: a..., Kojima [/bib_ref] fractures, 48 disability, [bib_ref] Quick and simple FRAIL scale predicts incident activities of daily living (ADL)..., Kojima [/bib_ref] [bib_ref] Frailty as a predictor of disabilities among communitydwelling older people: a systematic..., Kojima [/bib_ref] and dementia, [bib_ref] Frailty as a predictor of Alzheimer disease, vascular dementia, and all dementia..., Kojima [/bib_ref] all of which could result in poor quality of life 51 and increased cost 52 and use of health care resources, 53 such as emergency department visits, [bib_ref] Frailty as a predictor of emergency department utilization among community-dwelling older people:..., Kojima [/bib_ref] hospitalization, [bib_ref] Frailty as a predictor of hospitalisation among community-dwelling older people: a systematic..., Kojima [/bib_ref] and institutionalization. [bib_ref] Frailty as a predictor of nursing home placement among community-dwelling older adults:..., Kojima [/bib_ref] Multiple studies using cohorts of community-dwelling older adults have showed that the health care costs of frail individuals are sometimes several-fold higher than those of non-frail counterparts. [bib_ref] Associations of frailty with health care costs -results of the ESTHER cohort..., Bock [/bib_ref] [bib_ref] Frailty phenotype and healthcare costs and utilization in older women, Ensrud [/bib_ref] [bib_ref] Effect of frailty on resource use and cost for Medicare patients, Simpson [/bib_ref] [bib_ref] Frailty and healthcare costslongitudinal results of a prospective cohort study, Hajek [/bib_ref] [bib_ref] Healthcare costs of frailty: implications for long-term care, Salinas-Rodriguez [/bib_ref] Older adults form the main users of medical and social care services, 62 and the majority of health care costs are incurred by them. In the context of ongoing population aging, with an unprecedented growing number and proportion of older adults, this epidemiological and demographic population shift is starting to have a major impact on health care systems. Current health care systems are mostly designed to address organ-specific and disease-specific problems one at
## 26
Kojima et al a time and are not well prepared to deal with the chronic and complex medical needs of frail older patients and to provide seamless care for them in the long term. [bib_ref] Frailty: an emerging concept for general practice, De Lepeleire [/bib_ref] [bib_ref] The end of the disease era, Tinetti [/bib_ref] Therefore, older patients often receive suboptimal care due to the fragmented delivery of appropriate treatments and services. 65
## Interventions for frailty
There is no standard treatment of choice specifically for frailty, but there is a need for high quality cost-effective health care strategies to counter frailty. [bib_ref] Frailty research moves beyond risk assessment, Walston [/bib_ref] Although various types of frailty intervention models have been developed and investigated, there is a considerable degree of heterogeneity in terms of optimal intervention type, sample size, population characteristic, setting, baseline frailty status, frailty definition, and outcomes, and most findings are inconclusive. [bib_ref] The home-based older people's exercise (HOPE) trial: a pilot randomised controlled trial..., Clegg [/bib_ref] [bib_ref] Specifying the content of homebased health behaviour change interventions for older people..., Gardner [/bib_ref] [bib_ref] Health promotion interventions for community-dwelling older people with mild or pre-frailty: a..., Frost [/bib_ref] [bib_ref] Identifying acceptable components for home-based health promotion services for older people with..., Frost [/bib_ref] [bib_ref] Home-based health promotion for older people with mild frailty: the HomeHealth intervention..., Walters [/bib_ref] [bib_ref] van den Heuvel W. Interventions to prevent disability in frail community-dwelling elderly:..., Daniels [/bib_ref] [bib_ref] Effects of multi-domain interventions in (pre)frail elderly on frailty, functional, and cognitive..., Dedeyne [/bib_ref] [bib_ref] The (cost-)effectiveness of preventive, integrated care for community-dwelling frail older people: a..., Looman [/bib_ref] At present, it is therefore not possible to conclude what intervention is the most effective and appropriate. Overall, multi-domain intervention trials, which have been frequently conducted, 75,76 and many of which included an exercise component, seemed to have some favorable effects (although not in all trials) compared with mono-domain interventions or the control. [bib_ref] Effects of multi-domain interventions in (pre)frail elderly on frailty, functional, and cognitive..., Dedeyne [/bib_ref] [bib_ref] Outcomes of coordinated and integrated interventions targeting frail elderly people: a systematic..., Eklund [/bib_ref] Among the various outcome measures, functional ability, disability, and falls have been commonly examined while only a limited number of trials investigated changes in frailty status as an outcome. [bib_ref] Interventions to prevent or reduce the level of frailty in community-dwelling older..., Puts [/bib_ref] Although it is still not clear which frailty intervention is the most effective, exercise programs, especially multicomponent exercise including resistance training, have been consistently successful and seem likely to play a pivotal role in frailty interventions. [bib_ref] Effects of different exercise interventions on risk of falls, gait ability, and..., Cadore [/bib_ref] [bib_ref] The functional effects of physical exercise training in frail older people: a..., Chin A Paw [/bib_ref] [bib_ref] Effects of physical exercise interventions in frail older adults: a systematic review..., De Labra [/bib_ref] [bib_ref] Frailty and exercise interventions: evidence and barriers for exercise programs, Freiberger [/bib_ref] [bib_ref] The effectiveness of exercise interventions for the management of frailty: a systematic..., Theou [/bib_ref] [bib_ref] The Asia-Pacific clinical practice guidelines for the management of frailty, Dent [/bib_ref]
## Implications and challenges for health care policy
One of the plausible implementations of frailty into clinical practice is to identify frail patients using electronic health record data. [bib_ref] Defining and assessing geriatric risk factors and associated health care utilization among..., Kan [/bib_ref] [bib_ref] The value of unstructured electronic health record data in geriatric syndrome case..., Kharrazi [/bib_ref] [bib_ref] Comparing clinician descriptions of frailty and geriatric syndromes using electronic health records:..., Anzaldi [/bib_ref] In a UK study, Clegg et al developed the electronic Frailty Index (eFI) from 36 deficits, [bib_ref] Development and validation of an electronic frailty index using routine primary care..., Clegg [/bib_ref] based on the Frailty Index of cumulative deficit model. [bib_ref] Accumulation of deficits as a proxy measure of aging, Mitnitski [/bib_ref] The eFI was automatically populated from routinely collected data stored in the existing primary care electronic health record where general practitioners (GPs) list all patient diagnoses. [bib_ref] Development and validation of an electronic frailty index using routine primary care..., Clegg [/bib_ref] The authors showed eFI was able to stratify patients according to the degree of frailty and had robust predictive validity for mortality, hospitalization, and nursing home admission. [bib_ref] Development and validation of an electronic frailty index using routine primary care..., Clegg [/bib_ref] In 2017, NHS England started to require GPs to identify patients aged 65 years or older with moderate and severe frailty using validated frailty instruments including eFI, which is now freely available at most of GP practices.Following clinical assessment, patients with severe frailty are monitored using an annual medication review and other clinically relevant interventions if appropriate.This is probably the first attempt of nation-wide population-based frailty risk stratification and health utilization predictions in health care systems. [bib_ref] The challenge of ageing populations and patient frailty: can primary care adapt?, Reeves [/bib_ref] Population-based screening for frailty could be expensive and resource intensive, and currently there is no clear evidence for potential benefit, cost-effectiveness, or improved outcomes.Nonetheless, at the Frailty Consensus Conference in 2012, it was concluded that screening for frailty should be recommended for people with specific conditions or in certain settings. [bib_ref] Frailty consensus: a call to action, Morley [/bib_ref] One of the four consensus points was that those aged 70 years or older and those with significant weight loss (≥5%) due to chronic disease should be screened for frailty. [bib_ref] Frailty consensus: a call to action, Morley [/bib_ref] This recommendation is supported by the ADVANTAGE initiative, which advocates opportunistic frailty screening of people aged over 70 years receiving health care at any level of the system. 14 The French Society of Geriatrics and Gerontology suggests that people aged over 75 years who do not have difficulty with simple ADL but with early IADL would be good candidates for screening.The UK practice guidelines for frailty published from the British Geriatrics Society, Age UK, and Royal College of General Practitioners in 2014 recommend conducting a frailty assessment using all the encounters between health and social care staff and older people in community and outpatient settings. [bib_ref] British Geriatrics Society; Age UK; Royal College of General Practioners. Best practice..., Turner [/bib_ref] Research and development efforts aimed at establishing and disseminating best practice in frailty should not lack policy attention to older people with early (pre-) frailty that misses an opportunity to address some demands on health and social care services. [bib_ref] Moving upstream in health promoting policies for older people with early frailty..., Drennan [/bib_ref] Among the current challenges in the field of frailty research, one of the most important issues is the lack of an international standard definition of frailty. [bib_ref] Frailty consensus: a call to action, Morley [/bib_ref] [bib_ref] Research on frailty: continued progress, continued challenges, Sloane [/bib_ref] Despite accumulated research evidence on frailty, the variability in frailty definitions used in existing studies influences interpretation of the evidence, comparison with other studies, generalization of findings, and its implementation in the health care policy. In order to further advance and improve the health care services for frail older adults, it is imperative to come to an agreement in terms of frailty definition. [bib_ref] Frailty scales -their potential in interprofessional working with older people: a discussion..., Poltawski [/bib_ref] Ideally, the definition should be not only reliable but also feasible and easy to apply. [bib_ref] Family physicians need easy instruments for frailty, De Lepeleire [/bib_ref] Based on the currently available evidence on frailty intervention, there is strong evidence that exercise is central and possibly the optimal treatment of frailty. This needs to be explored further through multi-domain interventions that include exercise. The best timing for frailty intervention is not known but could range from intervening proactively to decrease risk of developing frailty or targeting those who are found to be prefrail or frail at the time of screening. If these concerns are properly addressed, widespread application of public health approaches will be possible, including screening, identification, and treatment of frailty, resulting in better care and healthier aging for older people. Involving frail older people in exploration of responses to frailty is likely to be fruitful; a Swedish study showed that financial situation, self-rated health, and social networks were determinants of life satisfaction. Actions that benefit life satisfaction -social and financial support -should be promoted. [bib_ref] The impact of socioeconomic conditions, social networks, and health on frail older..., Berglund [/bib_ref] Another important area of frailty research is education and training. In order to deliver high quality care and services effectively and efficiently to frail older adults, health care professionals including physicians, nurses, and other medical workers need to understand basic principles of care for frail older adults and to be able to detect frailty and provide treatment/interventions. Different understandings of frailty may impede communication between disciplines and need to be addressed. However, the evidence on education or training for frailty management is lacking. A recent systematic review that involved searches until May 2017 found no relevant article on education and/or training interventions for health care professionals in the field of frailty. [bib_ref] Educational strategies to train health care professionals across the education continuum on..., Windhaber [/bib_ref] There are currently ongoing frailty projects including educational components targeting health care professionals, patients, and caregivers, and new findings from these projects are expected to contribute to the field of frailty. [bib_ref] Educational strategies to train health care professionals across the education continuum on..., Windhaber [/bib_ref] There is an increasing interest in frailty in other medical fields than the geriatrics. [bib_ref] Frailty and cancer: implications for oncology surgery, medical oncology, and radiation oncology, Ethun [/bib_ref] [bib_ref] Integrating frailty research into the medical specialties -report from a U13 conference, Walston [/bib_ref] One example is that frailty has recently been pursued as a potential risk assessment measure for older surgical patients and has shown to be promising in predicting post-operative complications, such as mortality or length of hospital stay. [bib_ref] Frailty and post-operative outcomes in older surgical patients: a systematic review, Lin [/bib_ref] However, the lack of knowledge about frailty is a major barrier to the preoperative frailty assessment by surgeons, which may be addressed by education and training. [bib_ref] Surgical frailty assessment: a missed opportunity, Eamer [/bib_ref] A US educational interventional study involving cardiothoracic surgery residents showed that online short courses on frailty significantly improved residents' knowledge of frailty and influenced surgical risk estimates. [bib_ref] The impact of a frailty education module on surgical resident estimates of..., Ferguson [/bib_ref] Finally, frailty research may have evolved without taking into account the patient perspective. Frailty can be considered as a highly negative term and being labeled as frail may affect negatively the most vulnerable individuals in various ways. Those who were labeled as "old and frail" by others were more likely to be associated with a loss of interest in social and physical activities, poor physical health, and increased stigmatization. [bib_ref] Thinking you're old and frail': a qualitative study of frailty in older..., Warmoth [/bib_ref] The future research and development efforts need to acknowledge the risks of labeling older people in stigmatizing ways, and avoid frailty from becoming the new cloak of ageism and a tool for discrimination. 107
# Conclusion
There is an urgent need to identify and implement effective long-term care schemes to meet the complex demands of older adults. Frailty is potentially a perfect fit as a risk stratification paradigm and has therefore been recognized as an emerging public health priority. [bib_ref] Frailty: an emerging public health priority, Cesari [/bib_ref] Although a growing number of frailty studies have been conducted over the last two decades, their findings have not yet been fully translated into clinical practice 108 and the implementation of evidence on frailty in health care policy-making is further underrepresented. [bib_ref] Moving upstream in health promoting policies for older people with early frailty..., Drennan [/bib_ref] Successful implementation has the potential to improve quality of care for frail older adults and promote healthy aging as well as diminish the impact of frailty on health care systems and strengthen their sustainability. Such actions further demonstrate the substantial public health importance of frailty. Given the multidimensional and heterogeneous nature of frailty and the complex care needs of frail older adults, a multidisciplinary collaborative approach is needed between researchers, clinicians, policy makers, and older people themselves to improve the health and well-being of this subgroup of older adults. [bib_ref] Giving voice to older adults living with frailty and their family caregivers:..., Holroyd-Leduc [/bib_ref] [bib_ref] Designing randomized, controlled trials aimed at preventing or delaying functional decline and..., Ferrucci [/bib_ref] The field of frailty is still evolving and expanding and will need much more time and effort for further progress to occur. [bib_ref] Research on frailty: continued progress, continued challenges, Sloane [/bib_ref] Better outcomes for older people are likely to come with a time lag, and addressing frailty may require a massive cultural (perhaps generational) shift in the organization of health and care systems. 90
# Disclosure
The authors report no conflicts of interest in this work. submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## Risk management and healthcare policy
## Publish your work in this journal
Submit your manuscript here: https://www.dovepress.com/risk-management-and-healthcare-policy-journal Risk Management and Healthcare Policy is an international, peer-reviewed, open access journal focusing on all aspects of public health, policy, and preventative measures to promote good health and improve morbidity and mortality in the population. The journal welcomes submitted papers covering original research, basic science, clinical and epidemiological studies, reviews and evaluations, guidelines, expert opinion and commentary, case reports and extended reports. The manuscript management system is completely online and includes a very quick and fair peerreview system, which is all easy to use. Visit http://www.dovepress.com/ testimonials.php to read real quotes from published authors.
## Dovepress
## 30
|
Oncogenic effects of RAB27B through exosome independent function in renal cell carcinoma including sunitinib-resistant
Exosomes are 40-100 nm nano-sized extracellular vesicles. They are released from many cell types and move into the extracellular space, thereby transferring their components to recipient cells. Exosomes are receiving increasing attention as novel structures participating in intracellular communication. RAB27B is one of the leading proteins involved in exosome secretion, and oncogenic effects have been reported in several cancers. In recent years, molecularly targeted agents typified by sunitinib are widely used for the treatment of metastatic or recurrent renal cell carcinoma (RCC). However, intrinsic or acquired resistance to sunitinib has become a major issue. The present study aimed to elucidate the role of RAB27B in RCC including sunitinib-resistant and its role in exosomes. Bioinformatic analyses revealed that high expression of RAB27B correlates with progression of RCC. The expression of RAB27B protein in RCC cell lines was significantly enhanced compared with that in normal kidney cell lines. Furthermore, RAB27B protein expression was enhanced in all of the tested sunitinib-resistant RCC cell lines compared to parental cells. Although no specific effect of RAB27B on exosomes was identified in RCC cells, loss-of-function studies demonstrated that knockdown of RAB27B suppressed cell proliferation, migration and invasive activities. Moreover, anti-tumor effects of RAB27B downregulation were also observed in sunitinib-resistant RCC cells. RNA sequence and pathway analysis suggested that the oncogenic effects of RAB27B might be associated with MAPK and VEGF signaling pathways. These results showed that RAB27B is a prognostic marker and a novel therapeutic target in sunitinib-sensitive and -resistant RCCs. Further analyses should improve our understanding of sunitinib resistance in RCC.
# Introduction
Renal cell carcinoma (RCC) constitutes approximately 90-95% of all kidney neoplasms, making it the seventh most common site for tumours in 2013. The incidence of RCC has in RCC cells and sunitinib-resistant RCC cells were evaluated. The alteration of exosome secretion by knockdown of RAB27B and cell proliferative effects of exosomes derived from RAB27B down-regulated RCC cells were examined. Loss of function assays were performed in sunitinib-sensitive and -resistant RCC cells by examining cell proliferation, migration and invasion. The mechanisms of the effects of RAB27B were investigated by RNA sequencing and pathway analysis.
# Materials and methods
## Analysis of the correlation between rab27b and rcc
Kaplan-Meier and log-rank methods were used to analyze overall survival (OS) time using data in the OncoLnc dataset (http://www.oncolnc.org/), which contains survival data for 8,647 patients from 21 cancer studies included in TCGA. Also, OncoLnc is a useful tool for exploring survival correlations, and for downloading clinical data coupled to expression data for mRNAs, miRNAs, or long noncoding RNAs as previously described . In order to evaluate the clinical relevance, a TCGA cohort database of 534 patients with ccRCC was used. Full sequencing information and clinical information were acquired using UCSC Xena (http:// xena.ucsc.edu/), cBioPortal (http://www.cbioportal.org/publicportal/), and TCGA (https:// tcga-data.nci.nih.gov/tcga/). The present study met the criteria for the publication guidelines provided by TCGA (http://cancergenome.nih.gov/publications/publicationguidelines).
## Human rcc cell lines and cell culture
Human . SU-R-A498, SU-R-ACHN and SU-R-Caki1 were established by the same method. These cell lines were validated sunitinib resistance in xenograft assays. Parental and SU-R cells were subcutaneously injected into flanks of female nude mice (BALB/c nu/nu, 6-to 8-weeks-old, n = 4 for each group). After tumor formation was confirmed, we started gavage feeding of sunitinib (25mg/kg, five times a week). The tumors were harvested 20 days after injection. Comparison of the tumor volume of parental and SU-R cells were shown in S1 Fig. Human RCC cell lines were grown in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Equitech-Bio, Inc., Kerrville, TX, USA), 50 μg/mL streptomycin, and 50 U/mL penicillin. For the experiments involving exosomes, exosome-depleted FBS (System Bioscience, LLC, Palo Alto, CA, USA) was used as a supplement in place of standard FBS. The HK2 cell line was grown in Keratinocyte Serum-Free Medium (Invitrogen) supplemented with 0.05 mg/mL bovine pituitary extract (BPE) and 5 ng/mL epidermal growth factor (EGF). These cell lines were maintained in a humidified atmosphere of 95% air/5% CO 2 at 37˚C.
## Rna extraction and reverse transcription-quantitative polymerase chain reaction (rt-qpcr)
Total RNA was isolated using Isogen (NIPPON GENE CO., LTD., Tokyo, Japan) according to the manufacturer's protocol, using SYBR-Green qPCR for RT-qPCR. First, 500 ng of total RNA was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific, Inc.) under the following incubation condition: 25˚C for 10 min, 37˚C for 120 min and 85˚C for 5 min. cDNA was used for q-PCR performed with the Power SYBR Green Master Mix (cat. no. 4367659; Applied Biosystems, Foster City, CA, USA) on a 7300 Real-Time PCR System (Applied Biosystems). The specificity of amplification was monitored using the dissociation curve of the amplified product. All data values were normalized with respect to glucuronidase β (GUSB), and the ΔΔCq method was used to calculate the fold-change. The following primers were used:
RAB27B, forward primer, 5 0 -TAGACTTTCGGGAAAAACGTGTG-3 0 and reverse primer, 5 0 -AGAAGCTCTGTTGACTGGTGA-3 0 ; and GUSB, forward primer, 5 0 -CGTCCCACCTAG AATCTGCT-3 0 and reverse primer, 5 0 -TTGCTCACAAAGGTCACAGG -3 0 .
## Western blotting
## Transfection with small interfering rna (sirna)
As described previously , human RCC cells were transfected using Lipofectamine RNAi-MAX transfection reagent and Opti-MEM (both Thermo Fisher Scientific, Inc.) containing 10 nM siRNA. RAB27B siRNA (product ID, HSS184177 and HSS143561) or negative control siRNA (product ID, D-001810-10) (all Thermo Fisher Scientific, Inc.) to achieve loss-of-function. The knockdown efficiency of RAB27B siRNA was validated by confirming downregulation of RAB27B mRNA using RT-qPCR and RAB27B protein using western blotting.
## Isolation and quantification of exosomes
Exosomes were purified by differential centrifugation procedures, as described previously . Supernatants were collected from cells that had been cultured for 48 h in medium containing exosome-depleted FBS, and they were subsequently subjected to sequential centrifugation steps at 300g for 10 min, 2,000g for 10 min and 10,000g for 30 min to remove cell debris, dead cells and EVs other than exosomes. Supernatants were then centrifuged at 100,000g for 70 min at 4˚C (himac CP80WX, Hitachi, Ltd., Tokyo, Japan). The pelleted exosomes were suspended in PBS and collected by ultracentrifugation at 100,000g for 70 min. The purified exosomes were resuspended in PBS and used in subsequent experiments.
Exosome abundance was estimated with the ExoELISA-ULTRA CD63 kit (System Bioscience, LLC.) according to the manufacturer's protocol. This assay is a sensitive, direct Enzyme-Linked ImmunoSorbent Assay (ELISA) to quantitate exosome abundance in a given sample. The amount of exosomes is estimated by detecting CD63 on the exosome surface by specific antibody.
## Cell proliferation, migration, and invasion assays
Human RCC cells were seeded in 96-well plates with 3x10 3 cells/well for XTT assays. After 72 h, cell proliferation was determined using a Cell Proliferation Kit II (Roche Diagnostics GmbH, Mannheim, Germany) as described previously .
Cell migration activity was evaluated with wound healing assays. Cells were plated in 6-well plates at 2x10 5 cells per well, and after 48 h of transfection the cell monolayer was scraped using a P-20 micropipette tip. The initial (0 h) and residual gap lengths 24 h after wounding were calculated from photomicrographs.
Cell invasion assays were performed using modified Boyden chambers consisting of Matrigel-coated Transwell membrane filter inserts with 8-μM pores in 24-well tissue culture plates (BD Biosciences, San Jose, CA, USA). 48 h following transfection, the cells were seeded in the upper chamber of 24-well plates at 1x10 5 /well with serum-free RPMI 1640 medium. RPMI containing 10% exosome-depleted FBS in the lower chamber served as the chemoattractant, as previously described . 24 h after seeding, the cells that had passed through the pores and attached to the surface of the chamber were stained by Diff-Quick (a modified Giemsa stain) (Richard Allan Scientific, San Diego, CA, USA) and counted from photomicrographs.
## Rna sequencing analysis and pathway analysis
Total RNAs from 786-o and A498 cells transfected with control siRNA or RAB27B siRNA were subjected to RNA sequencing, which was performed by RIKEN GENESIS CO., LTD., Tokyo, Japan. mRNA profiles were generated by NovaSeq 6000 (Illumina, Inc., San Diego, CA, USA).
Genes with significantly downregulated expression after transfection with RAB27B siRNA were compared with control siRNA (fold change < -1.0) and were then categorized with the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways through GeneCodis analysis (genecodis.cnb.csic.es) [43-45].
# Statistical analysis
Data are presented as means ± standard deviation of at least 3 independent experiments. The relationships between two groups were analyzed using Mann-Whitney U tests. The relationships between three or more variables and numerical values were analyzed using Bonferroniadjusted Mann-Whitney U tests. All analyses were performed with Expert StatView software version 5.0 (SAS Institute, Inc., Cary, NC, USA). When P < 0.05, the data were accepted as showing a statistically significant difference.
# Results
## Clinical significance of rab27b expression in rcc
We first characterized the correlation between RAB27B expression and OS by performing a Kaplan-Meier analysis using the OncoLnc dataset. This analysis demonstrated that the group of patients with high expression of RAB27B (Z-score > 0) exhibited significantly lower OS rates compared with those in the low expression group (Z-score < 0) in both the ccRCC (P = 0.00179,left panel) and pRCC (P = 0.00407,cohorts. Next, we evaluated the correlations between RAB27B expression levels and patient clinicopathological parameters. Among the ccRCC cohort of TCGA, we found that the expression levels of RAB27B were significantly increased in pathological T4 categoryand pathological high grade G4 cases. There was no significant difference of RAB27B expression between normal tissues and ccRCCs/pRCCs (S2 .
Then, RAB27B protein expression levels in RCC cell lines, sunitinib-resistant RCC cell lines and HK2 cells were evaluated by Western blot analyses. The expression levels of RAB27B protein in RCC cell lines were elevated in comparison with the levels in HK2. Furthermore, it was revealed that the expression of RAB27B protein in all of the sunitinib-resistant RCC cell lines was enhanced compared to their parent cell lines.
## Functional investigation of rab27b for exosomes in rcc
To analyze the function of RAB27B for exosomes, we employed A498 cells because of their high expression level of RAB27B. First, we investigated whether knockdown of RAB27B reduced the secretion of exosomes. RT-qPCR analyses indicated effective downregulation of RAB27B mRNA in the si-RAB27B-tansfected RCC cells. Western blot analyses revealed that RAB27B protein levels were also downregulated in the cells transfected with si-RAB27B.
## Plos one
Oncogenic effects of RAB27B in sunitinib-sensitive and -resistant RCCs A498 cells transfected with si-control or si-RAB27B were cultured in RPMI 1640 medium supplemented with exosome-depleted FBS. Cell culture conditioned media were collected 48 h after transfection, and exosomes in the medium were isolated by ultracentrifugation and resuspended in PBS. We measured the amount of exosomes isolated from 40 mL of conditioned medium with the ExoELISA-ULTRA CD63 kit. Although the quantity of exosomes in the medium of si-RAB27B-transfected cells tended to be less than that of si-control-transfected cells, there was no statistically significant difference between them (S3A .
We evaluated the effect of exosomes on cell proliferation with the XTT assay. We initially checked the effect of A498 exosomes on their own proliferation. A498 exosomes were added into the medium of A498 cells seeded in 96-well plates, and cell proliferation was determined after 72 h. XTT assays revealed that exosomes derived from si-RAB27B-transfected cells had no significant effect on cell proliferation compared to those from mock and si-control transfectants (S3B . Subsequently, we examined the effect on the proliferation of cells other than the cells from which the exosomes were derived. We obtained the same results in the experiment when we added A498 exosomes to 786-o cells (S3B and SU-R-786-o cells (S3B . Thus, no specific effect of RAB27B on exosomal function was revealed in RCC cells in regard to both quantity and function.
## Oncogenic effects of rab27b in rcc cells
Loss-of-function studies using si-RAB27B were conducted to investigate the functional role of RAB27B in RCC cells. XTT assays with 786-o, A498, ACHN and Caki1 cells demonstrated that cell proliferation was inhibited in si-RAB27B-transfectants in comparison with that in the
## Plos one
Oncogenic effects of RAB27B in sunitinib-sensitive and -resistant RCCs mock-or si-control-transfectants. Wound healing assays revealed that cell migration activity was also inhibited in si-RAB27B-transfected 786-o and A498 cells. Similarly, Matrigel invasion assays revealed that the number of invading cells was significantly decreased in si-RAB27B-transfected 786-o and A498 cells. From the above, the oncogenic effects of RAB27B in RCC cells were demonstrated.
## Oncogenic effects of rab27b in sunitinib-resistant rcc cells
Subsequently, we investigated the functional role of RAB27B in sunitinib-resistant RCC cells. Loss-of-function assays by transfection with si-RAB27B were performed in sunitinib-resistant RCC cells as well as parental cells. Cell proliferation assessed by XTT assays was reduced in si-RAB27B-transfected SU-R-786-o, SU-R-A498, SU-R-ACHN and SU-R-Caki1 cells. Cell migration activity was suppressed in si-RAB27B-transfected SU-R-786-o and SU-R-A498 cells. Cell invasion activity was also suppressed by si-RNA27B transfection in SU-R-
## Plos one
Oncogenic effects of RAB27B in sunitinib-sensitive and -resistant RCCs 786-o and SU-R-A498 cells. These results indicated that high expression of RAB27B was associated with oncogenic effects even in sunitinib-resistant RCC cells.
## Analysis of the mechanism of oncogenic effects of rab27b
In order to investigate the mechanism of oncogenic effects of RAB27B, RNA sequence and pathway analyses were performed (DRA009693: https://ddbj.nig.ac.jp/DRASearch/ submission?acc=DRA009693). A heatmap of RNA sequencing data was shown in S4 Fig. Based on the RNA sequencing data, we found 1069 genes downregulated in 786-o cells transfected with si-RAB27B, and 436 genes downregulated in A498 cells transfected with si-RAB27B compared with each cell type transfected with si-control. Among the 185 genes common to both, the most downregulated 20 genes are displayed in.
Next, we used GeneCodis analysis to categorize the si-RAB27B-downregulated genes into KEGG pathways. The results revealed that these genes were contained in 24 pathways, including the Vascular Endothelial Growth Factor (VEGF) signaling pathway, which is the main PLOS ONE extracellular environment as exosomes, where they function in a multitude of intercellular signaling processes . RAB27A and RAB27B are thought to function in MVEs' docking to the plasma membrane . Furthermore, RAB27B has been suggested to be involved in the transfer of MVEs to the plasma membrane . Contrary to expectations, suppression of exosome release was not observed in cells transfected with si-RAB27B in the present study. Since exosome secretion has been reported to increase in environments that are not suitable for cell survival, including chemotherapeutic treatment, hypoxia, heat stress, etc.
[49], it is possible that the tumor-suppressive effects of si-RAB27B acted to promote exosome secretion. Further studies are necessary to elucidate the mechanism underlying RAB27B and exosome secretion in RCC.
Exosomes have a range of different roles in cancer and they can be used as diagnostic markers and predict therapeutic responses. They may also constitute targets in therapeutic applications. Chen et al. reported that PD-ligand 1 (PD-L1) on exosome surfaces released from metastatic melanoma cells suppressed tumor immunity, and circulating exosomal PD-L1 can be a response predictor of anti-PD-1 therapy [50]. Kamerkar et al. demonstrated that exosomes derived from mesenchymal cells artificially incorporated siRNA or short hairpin (sh) RNA specific for oncogenic KRAS suppressed cancer in a plurality of mouse models of pancreatic cancer [51]. In renal cancer, differential protein profiling in urinary exosomes [52] and miRNAs contained in serum exosomes [53, 54] represent potential diagnostic markers. In addition, exosomes released from renal cancer stem cells contribute to triggering angiogenesis at premetastatic niches in the lung . Further on, exosomes containing carbonic anhydrase 9 (CA9), a cellular response to hypoxia, were released from hypoxic RCC cells, and they are suggested to enhance angiogenesis in the microenvironment, thereby contributing to cancer
## Plos one
Oncogenic effects of RAB27B in sunitinib-sensitive and -resistant RCCs In the present study, the relevance of RAB27B to exosome secretion was not identified. Nevertheless, knockdown of RAB27B showed inhibitory effects of cell proliferation as well as cell migration and invasion. Therefore, we hypothesized that another functional role of RAB27B might be involved in epithelial-mesenchymal transition (EMT). Even though the western blot analyses of several representative EMT markers showed no common alterations in the RCC cells, the protein expression levels of Vimentin and N-cadherin were faintly decreased in 786-O or SU-R-A498 cells after si-RAB27B transfection (S6 suggesting that RAB27B might partially contributed to the EMT process in RCC as was previously reported in breast cancer.
As well as our demonstration, the oncogenic effects of RAB27B have been reported in several types of cancer. RAB27B has been shown to regulate invasive growth in vitro and in vivo in estrogen receptor (ER)-positive breast cancer cell lines and to be involved in osteosarcoma cell migration and invasion. Furthermore, it was suggested that disposal of tumorsuppressive miRNA via exosome release through the function of RAB27B is associated with acquisition of metastatic properties in bladder cancer . According to our analysis, the expression level of RAB27B was associated with poor prognosis in RCCs. It has been demonstrated that high expression of RAB27B is correlated with poor prognosis in hepatocellular carcinoma, colorectal cancerand ovarian cancer, consistent with our results. Although hypomethylated RAB27B is reported to be a progression-associated prognostic biomarker in glioma [32], the regulatory mechanisms of RAB27B expression remain to be defined. Thus, further research on the mechanisms is important to better understand these processes.
In terms of acquisition of drug resistance, a number of studies have demonstrated the contribution of exosomes. For instance, there are several reports indicating a relationship between chemoresistance and exosomes in prostatic cancer. In neuroblastomaand ovarian cancer,exosomal transfer of miR21 between cancer cells and stromal cells was indicated to contributes to development of chemoresistance. Recently, it was also demonstrated that ALK in the exosomes secreted by BRAF inhibitor-resistant melanoma cells transferred drug resistance through activation of the MAPK signaling pathway in recipient cells. Furthermore, in renal cancer, Qu et al. showed that long noncoding RNA transmitted by exosomes promoted acquisition of sunitinib resistance. With respect to RAB27B, metabolic reprogramming mediated by RAB27B was verified to induce doxorubicin resistance in breast cancer cells, and it was found that RAB27B is involved in chemoresistance to cisplatin in pancreatic cancer. Moreover, exosome transfer from stroma to cancer cells regulated by stromal RAB27B is involved in therapeutic resistance in breast cancer. In our study, RAB27B showed oncogenic effects in RCC cell lines with sunitinib resistance. Since the expression of RAB27B protein in all of the sunitinib-resistant RCC cell lines was enhanced compared to the parental cells, RAB27B may have some involvement in sunitinib resistance acquisition.
In order to investigate the mechanism underlying the oncogenic effects of RAB27B and its relevance to sunitinib resistance, pathway analysis was performed. GeneCodis analysis demonstrated that the VEGF signaling pathway and the MAPK signaling pathway were downregulated in RAB27B-knockdown cells. VEGF plays an important role in pathological angiogenesis associated with tumor growth, and also act as an autocrine growth factor. VEGF have long been regarded as a promising therapeutic target in RCC, and is a major target molecule of sunitinib. Ishibashi et al. reported that tyrosine kinase inhibitors (TKIs) treatment to 786-o cells enhanced the expression of IL-6 and VEGF, and suggested that combination therapy of IL-6 inhibitor and TKIs may overcome TKI resistance. MAPK pathway, which play essential roles in cell differentiation, proliferation and survival, also has been reported to be possible therapeutic target in RCC. In addition, Gao et al. demonstrated that treatment targeting the ERK/MAPK pathway suppressed sunitinib resistance. In ER-negative breast cancer, RAB27B-mediated modulation of β-catenin and VEGF was reported. Further, in pancreatic cancer, EVs released by upregulated RAB27B activated p38 MAPK. From the above, RAB27B may be associated with activated VEGF and MAPK signaling. Although knockdown of RAB27B showed no common alterations of the expression pattern of VEGF-or MAPK-related proteins in the western blot analyses, Erk1/2 or p38 MAPK protein was faintly decreased in 786-O, A498, or SU-R-786-O after si-RAB27B transfection (S5 suggesting that RAB27B might partially contributed to accelerating MAPK pathway in RCC. Interestingly, pathway analysis also showed an association between RAB27B and neurological diseases such as Huntington's disease, Parkinson's disease and Alzheimer's disease (AD). In fact, there are several reports showing the involvement of exosomes in these diseases. Additionally, the upregulation of some synaptic GTPases, including RAB27B, was detected in tissues from patients with higher degrees of AD, and aberrant synaptic trafficking was suggested to modulate the progression of AD.
In conclusion, the present study investigated the association between RAB27B, an exosome secretory protein, and RCC. Although specific effects of RAB27B on exosomes were not identified, the oncogenic effects of RAB27B in RCC cell lines were demonstrated. Furthermore, the oncogenic effects of RAB27B were also demonstrated in sunitinib-resistant RCC cell lines. RAB27B may be a novel therapeutic target for sunitinib-sensitive and -resistant RCC. In addition, further studies may provide new insights into improving our understanding of sunitinib resistance in RCC.
## Supporting information
|
Cardiovascular risk assessment in children and adolescents with congenital solitary kidneys
[bib_ref] Born with a solitary kidney: at risk of hypertension, Scola [/bib_ref] [bib_ref] Arterial stiffness in children and adolescents with masked and sustained hypertension, Kollios [/bib_ref] [bib_ref] Changes in central aortic pressure levels, wave components and determinants associated with..., García-Espinosa [/bib_ref] [bib_ref] Central blood pressure and pulse wave amplification across the spectrum of peripheral..., Lurbe [/bib_ref] [bib_ref] ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for..., Mancia [/bib_ref] [bib_ref] Changes in central aortic pressure levels, wave components and determinants associated with..., García-Espinosa [/bib_ref] [bib_ref] Central blood pressure and pulse wave amplification across the spectrum of peripheral..., Lurbe [/bib_ref] [bib_ref] Subclinical ventricular repolarization abnormality in uncontrolled compared with controlled treated hypertension, Al-Nimer [/bib_ref] [bib_ref] The meaning of the Tp-Te interval and its diagnostic value, Kors [/bib_ref] [bib_ref] QT ratio as an index of arrhythmogenesis, Gupta [/bib_ref]
## | patients and me thods
The study protocol was approved by the local ethics committee The control group consisted of age-and gender-matched otherwise healthy children who were evaluated in the pediatric cardiology clinic for cardiac murmur and found to have physiological murmur without any underlying cardiac pathology.
Bodyweight, height, and office systolic BP (SBP) and diastolic BP (DBP) were measured in all participants. Body mass index (BMI) was calculated as a ratio of body weight (kg) per square body height (m 2 ). Standard deviation scores (SDSs) of BMI, and office SBP and DBP were determined via the Child Metrics program according to published normal values. [bib_ref] New features for child metrics: further growth references and blood pressure calculations, Demir [/bib_ref] [bib_ref] Clinical practice guideline for screening and management of high blood pressure in..., Flynn [/bib_ref] [bib_ref] Reference values for weight, height, head circumference, and body mass index in..., Neyzi [/bib_ref] Laboratory blood and urine tests were estimated in all patients with standard methods. Estimated glomerular filtration rates were calculated due to the original Schwartz formula. [bib_ref] A simple estimate of glomerular filtration rate in adolescent boys, Schwartz [/bib_ref] In addition, PWA, ambulatory blood pressure monitoring (ABPM), electrocardiographic, and echocardiographic evaluations were performed in all cases. Patients were evaluated with an automated oscillometric PWA-ABPM device (Mobil-O-Graph; IEM, Stolberg, Germany) validated for office and ambulatory pBP and cBP measurements according to the British Hypertension Society and the European Society of Hypertension recommendations. [bib_ref] Evaluation of the Mobil-O-Graph new generation ABPM device using the ESH criteria, Franssen [/bib_ref] The device has shown good accuracy for cBP when compared with invasive measurements in children. [bib_ref] The accuracy of central blood pressure obtained by oscillometric noninvasive method using..., Shiraishi [/bib_ref] By using this device, cSBP and cDBP, aortic pulse wave velocity (PWV), validated by referring to tonometric devices and/or invasive methods, and augmentation index normalized for an HR of 75 bpm (AIx@75) can be assessed. [bib_ref] Recommendations for improving and standardizing vascular research on arterial stiffness: a scientific..., Townsend [/bib_ref] The cuff of the device is placed in the non-dominant arm. Brachial SBP, DBP, pulse pressure, and waveforms are obtained using an algorithm (ARC Solver algorithm). [bib_ref] Wave reflection quantification based on pressure waveforms alone-methods, comparison, and clinical covariates, Hametner [/bib_ref] [bib_ref] New features for child metrics: further growth references and blood pressure calculations, Demir [/bib_ref] [bib_ref] Distribution of 24-h ambulatory blood pressure in children: normalized reference values and..., Wühl [/bib_ref] With PWA and wave separation analysis, PWV and AIx@75 were calculated. [bib_ref] Wave reflection quantification based on pressure waveforms alone-methods, comparison, and clinical covariates, Hametner [/bib_ref] AIx@75 represents increased wave reflection and/or early return of the reflected wave due to increased arterial stiffness presented by PWV. [bib_ref] Validation of a brachial cuff-based method for estimating central systolic blood pressure, Weber [/bib_ref] As a result, PWV is considered as a direct indicator of vascular stiffness, while AIx@75 is an indirect measure. [bib_ref] Utility of arterial stiffness assessment in children, Panchangam [/bib_ref] With the combination of office and ABPM measurements, study patients were designated into four classical phenotypes: "Normotension (NT)" was defined as having normotensive BP measurements on both office and ABPM measurements; "masked HT (MHT)" was defined as having normotensive office BP measurements with hypertensive ABPM measures; "white coat HT (WCHT)" was defined as having hypertensive office BP measurements with normotensive ABPM measures; and "ambulatory (sustained) HT (AHT)" was defined as having hypertensive BP measurements on both office and ABPM measurements. [bib_ref] Changes in central aortic pressure levels, wave components and determinants associated with..., García-Espinosa [/bib_ref] In addition, we grouped patients with any abnormalities on ABPM irrespective of the office BP measurements, as HT according to ABPM (HT ABPM ) including hypertensive ABPM measures and/or high BP loads (≥25%). [bib_ref] Cardiovascular disease in children with chronic kidney disease, Mitsnefes [/bib_ref] QTcd was calculated as the difference between the maximum and the minimum QTc intervals. [bib_ref] QTc dispersion predicts cardiac mortality in the elderly: the Rotterdam Study, De Bruyne [/bib_ref] Tp-e was measured from the highest point to the endpoint of T wave. If reverse T waves were present, the measurement was taken from the lowest point to the endpoint of the T wave. Tp-ed was the difference between the maximum and minimum Tp-e values. [bib_ref] Tpeak-tend interval in 12-lead electrocardiogram of healthy children and adolescents in childhood, Bieganowska [/bib_ref] [bib_ref] Echocardiographic determination of left ventricular mass in man, Devereux [/bib_ref] [bib_ref] Age-specific reference intervals for indexed left ventricular mass in children, Khoury [/bib_ref] An LVMI exceeding the 95th percentile for sex and age in normal children and adolescents was used to define LVH. [bib_ref] Age-specific reference intervals for indexed left ventricular mass in children, Khoury [/bib_ref] Abdominal aortic stiffness was evaluated with "strain" that showed the elasticity or distensibility and "pressure strain elastic modulus (Ep)" that represents the stiffness of the aortic wall. To calculate these parameters, minimum diastolic and maximum systolic aortic diameters were measured in the subxiphoid long axis. The systolic aortic diameter was recorded during the maximum anterior movement of the aorta, and the diastolic diameter was recorded as the aortic diameter measured during the QRS peak, according to simultaneous ECG. The values measured from 10 consecutive heartbeats were averaged. Accordingly, strain, Ep, and normalized Ep were calculated according to the following formulas [bib_ref] Utility of arterial stiffness assessment in children, Panchangam [/bib_ref]
## | electrocardiographic assessment
## | echocardiographic assessment
# | statistical analysis
The SPSS 24.0 (SPSS Inc.) package program was used for statistical analysis. The Kolmogorov-test was used to evaluate the normal distribution of continuous variables between groups. Normally distributed continuous variables were presented as mean ± standard deviation and compared by Student's t test, whereas parameters not distributed normally were presented as median (interquartile range, IQR) and compared by the Mann-Whitney U test. The chi-square test was used to compare categorical variables between groups, which were expressed as frequency. Depending on the distribution type of the variables, Pearson's or Spearman's correlation analysis was performed. Factors affecting specific parameters adjusted for age, gender, and BMI were assessed by linear regression analysis.
Logistic regression analysis was used to determine the odds ratio for probability to have HT ABPM . A p-value <.05 was considered statistically significant for all statistical evaluations.
## | re sults
A total of 36 children with CSK and 36 healthy age-and gendermatched controls were enrolled in the study. We had excluded one patient with severe hydroureteronephrosis and vesicoureteral reflux, and two cases with heterogeneous Tc99m DMSA uptake in renal scan advised to be followed up for scarring although they had no reflux. Of the patients with CSK, 25 (69%) had MCDK, and 11 (31%) had URA. Single functioning kidneys were on the right side in 20 (56%) and on the left side in 16 (44%) of the patients with CSK.
Eight (22%) of the patients had a prenatal diagnosis. The rest of the patients were found to have URA or MCDK by chance upon abdominal ultrasounds performed for abdominal complaints. One of the cases with URA had a history of maternal URA. None of the patients were active smokers.
## | laboratory findings
Serum creatinine, uric acid, and total cholesterol levels were significantly higher, while eGFR levels were significantly lower in patients with CSK . In patients with CSK, only one patient had an eGFR < 90 ml/min/1.73 m 2 , which was 82.5 ml/min/1.73 m 2 and all other eGFR values were ≥ 90 ml/min/1.73 m 2 .
## | bp measurements
Although office SBP and DBP levels were similar between the groups, most of the ABPM parameters, as well as cBP parameters, were significantly higher in patients with CSK (Table 2, .
When we compared the BP phenotypes between the groups, the rate of patients with NT, MHT, WCHT, or AHT was similar between the groups (p > .05), while patients with HT ABPM were significantly higher in the CSK group (p = .001, . The probability of HT ABPM significantly increased with having CSK (OR: 6.00, 95% confidence interval: 2.15-16.71, p = .001).
## | ventricular repolarization indices
Electrocardiographic ventricular repolarization parameters, namely QTd, QTcd, Tp-e, Tp-ed, Tp-e/QT, Tp-e/QTc, and Tp-e/QTcd, were all similar between the groups (p > .05).
## | arterial stiffness parameters
Abdominal arterial stiffness was assessed with PWV and AIx@75 by the oscillometric method and with strain, elastic modulus, and normalized elastic modulus (Ep and Ep*, respectively) calculated based on echocardiographic measurements. Although PWVs were significantly higher in patients with CSK , AIx@75 values were similar between patients with CSK and healthy controls .
Strain, Ep, and Ep* were similar between the groups . Since strain represents the "elasticity," and Ep and Ep* represent the "stiffness" of the arteries, we expected a negative correlation between PWV and strain and positive correlations between PWV and Ep and Ep*. However, a correlation in the reverse direction than expected was observed in the control group between 24-hour PWV and strain and between 24-hour PWV and Ep. When adjusted for age, gender, and BMI SDS, the significance only remained between 24-hour PWV and strain. In patients with CSK, 24-hour PWV tended to correlate with strain negatively and it was positively and significantly correlated with Ep and Ep*, which remained significant after adjustment for age, gender, and BMI SDS . and/or microalbuminuria may be present in approximately half of the cases with CSK, especially in those with subtle dysplastic changes. [bib_ref] Ambulatory blood pressure monitoring is recommended in the clinical management of children..., Westland [/bib_ref] [bib_ref] Hypertension and microalbuminuria in children with congenital solitary kidneys, Schreuder [/bib_ref] In addition, patients with SKs, including both CSK and ASK, have also been found to have lower eGFR levels, higher risk of HT, and MHT when evaluated with ABPM compared to healthy children. [bib_ref] Evaluation of hypertension by ambulatory blood pressure monitoring in children with solitary..., Tabel [/bib_ref] [bib_ref] Ambulatory blood pressure monitoring and renal functions in children with a solitary..., Dursun [/bib_ref] Therefore, ABPM has been advised to be applied to all cases with SKs. [bib_ref] Ambulatory blood pressure monitoring is recommended in the clinical management of children..., Westland [/bib_ref] However, other studies stated that the risk of HT and levels of microalbuminuria and eGFR might be similar to those of healthy children in patients with unilateral SKs, in the event the SK was completely healthy or in cases who had a prenatal diagnosis and early postnatal follow-up. [bib_ref] Blood pressure, renal function, and proteinuria in children with unilateral renal agenesis, Seeman [/bib_ref] [bib_ref] Is microalbuminuria a risk factor for hypertension in children with solitary kidney?, Shirzai [/bib_ref] [bib_ref] Outcomes of a cohort of prenatally diagnosed and early enrolled patients with..., Marzuillo [/bib_ref] Since subtle dysplastic anomalies may accompany CSK, we preferred to include only patients without any uptake defect or confirmed renal scarring in Tc 99m -DMSA.
## | discuss ion
In studies comparing cases with CSK and ASK, the results were much more conflicting. In one study, GFR was better in patients with CSK, which was thought to be associated with functional adaptation, while HT and microalbuminuria were similar. [bib_ref] Congenital solitary kidney size at birth could predict reduced eGFR levels later..., Marzuillo [/bib_ref] In other studies,
HT and/or ABPM parameters were significantly increased in patients with CSK, [bib_ref] Ambulatory blood pressure monitoring is recommended in the clinical management of children..., Westland [/bib_ref] Central BP, which is considered to represent BP in the aortic root, is calculated from the pulse wave measured from the peripheral arteries via the transfer function and is generally lower than the value measured from the brachial artery. [bib_ref] Central blood pressure and pulse wave amplification across the spectrum of peripheral..., Lurbe [/bib_ref] In recent years, cBP measurement is superior in determining cardiovascular risk especially in young adults. [bib_ref] Isolated systolic hypertension in young and middle-aged adults, Yano [/bib_ref] Since the middle-sized arteries such as the brachial artery are more elastic in children than in adults, the accumulation of pulse waves by the brachial artery results in higher brachial BP measurements in children although cBPs are within normal ranges. [bib_ref] Central systolic blood pressure and central pulse pressure predict left ventricular hypertrophy..., Litwin [/bib_ref] Thus, measuring cBPs may prevent the overestimation of hypertension determined by peripheral ABPM measurements in children and young adults. The number of studies in childhood is scarce. In a recent study, children with primary hypertension had normal cBP values even in those with severe AHT, and cBP measures had at least the same or higher power as ABPM in predicting end-organ damage. [bib_ref] Central systolic blood pressure and central pulse pressure predict left ventricular hypertrophy..., Litwin [/bib_ref] No studies have encountered cBP in children with CSK. In our study, cBP values were higher in patients with CSK and peripheral ABPM measurements confirming a real increase in BP. Also, we have demonstrated that having CSK increased the risk of HT ABPM sixfold.
Thus far, we may conclude that ABPM should be performed in cases with CSK.
Arterial stiffness can be evaluated by pulse wave analysis (PWA) or echocardiographic methods. The gold standard for PWA measurement is tonometry; however, this technique is operator-dependent and may be challenging in childhood and obesity. [bib_ref] The accuracy of central blood pressure obtained by oscillometric noninvasive method using..., Shiraishi [/bib_ref] [bib_ref] Utility of arterial stiffness assessment in children, Panchangam [/bib_ref] Thus, the oscillometric method has been increasingly preferred as an easier way to assess PWA in recent years. [bib_ref] Utility of arterial stiffness assessment in children, Panchangam [/bib_ref] [bib_ref] Non-invasive 24-hour ambulatory monitoring of aortic wave reflection and arterial stiffness by..., Papaioannou [/bib_ref] A pulse wave is generated when the left ventricle contracts. When a vessel wall hardens, its elasticity decreases, leading to early wave reflection and increased PWV. PWV is considered as a direct measure of vascular stiffness, whereas AIx@75 is considered as an indirect measure. [bib_ref] Utility of arterial stiffness assessment in children, Panchangam [/bib_ref] Increased PWV was reported to be closely associated with the risk of cardiovascular mortality and morbidity. [bib_ref] Tpeak-tend interval in 12-lead electrocardiogram of healthy children and adolescents in childhood, Bieganowska [/bib_ref] In our study, cBP, PWV, and AIx@75 were measured using an oscillometric device for PWA along with ABPM. [bib_ref] The accuracy of central blood pressure obtained by oscillometric noninvasive method using..., Shiraishi [/bib_ref] Further studies seem to be needed to define the most efficient parameter of the PWA-ABPM device for defining and predicting cardiovascular problems in children with CSK.
## Co n fli c t o f i nte r e s t
The authors declare that they have no conflict of interest.
## Auth o r co ntr i b uti o n s
Belde Kasap-Demir served as the principle investigator and involved in all aspects of the study, including study concept, study design, recruitment of patients, data analysis, and manuscript preparation.
Eren Soyaltın and Seçil Arslansoyu-Çamlar involved in the recruitment of patients, data analysis, and manuscript preparation. Caner
## Alparslan involved in data acquisition. demet alaygut and önder
Yavaşcan involved in study design. Tülay Demircan performed echocardiographic evaluation. Fatma Mutlubaş involved in data interpretation. Cem Karadeniz involved in electrocardiographic data evaluation and manuscript preparation.
## I n fo r m e d co n s e nt
Written informed consent was obtained from all individual participants included in the study and their parents or legal guardians.
## O rci d
Belde Kasap-Demir https://orcid.org/0000-0002-5456-3509
Seçil Arslansoyu-Çamlar https://orcid. org/0000-0002-2402-0722
## R e fe r e n c e s
[fig] 14: 12.2017/128). Written informed consent was obtained from all guardians, and participants were informed about the purpose of the study, based on their age. Patients with CSK being followed up in our clinic between January 2018 and June 2019 were recruited for this prospective cohort study. Those with unilateral renal agenesis (URA) or multicystic dysplastic kidneys (MCDK) were included in the study. Patients with any uptake defect or confirmed renal scarring in Tc 99 m dimercaptosuccinic acid scan (Tc 99 m -DMSA); estimated glomerular filtration rates (eGFRs) ≤60 ml/min/1.73 m 2 ; vesicoureteral reflux, ureteropelvic junction obstruction, duplex system, megaureter, posterior urethral valve, or neurogenic bladder; those with type 1 or type 2 diabetes mellitus, or those under anti-hypertensive medication were excluded. [/fig]
[fig] A 12 -: lead digital ECG was performed in all patients and healthy volunteers. The standard 12-lead ECG (Cardiofax GEM, Model 9022 K; Nihon Kohden) was recorded at a speed of 25 mm/sec and an amplitude of 1 mV/cm. QT dispersion (QTd), corrected QT dispersion (QTcd), Tpeak to Tend (Tp-e) interval, Tp-e dispersion (Tp-ed), and Tp-e/QT and Tp-e/QTc ratios were calculated to evaluate ventricular and repolarization periods. For QTd, we measured the QT interval from the beginning of the QRS complex to the end of the T wave. The QTd was calculated as the difference between the maximum and minimum of QT interval. QTc interval was calculated according to the Bazett formula [QTc = QT/ √RR (ms)]. [/fig]
|
A Misleading Trismus
We report the case of a 73-year-old women admitted to ICU with a diagnosis of tetanum supported by a mild trismus and treated with potentially unsafe drugs until the correct diagnosis was made.
# Introduction
Acute intermittent porphyria is a disease common enough to be encountered by many physicians but rare enough to lead to a wrong diagnosis.
## Report of the case
A 73-year-old male patient came to our attention after admission to ICU from infective disease ward with the diagnosis of tetanus. His past clinical history was significant for recent use of a nonreported antibiotic for a sinusitis and a recent injury on his left hand with a garden tool. At the admission the patient was conscious with intense abdominal pain, blood pressure was 210/110 mmHg, 100 bpm, Sp02 = 89% in air, red urine output due to difficult bladder catheterization, and mild trismus. Abdomen was soft, with mild tenderness. The patient began specific antibiotic treatment, immunoglobulins, and benzodiazepines (diazepam) to treat muscle rigidity, diclofenac, and metoclopramide for abdominal pain and vomiting with no significant clinical improvement. Tetanus antibodies resulted negative. After admission, abdominal pain worsened with no positive radiological examinations (abdominal TC and echography) and bleeding from catheter also continued in spite of frequent bladder washing with continuous cold saline. On the base of abdominal pain and red urine output, acute intermittent porphyria (AIP) was suspected and the qualitative test for urinary porphyrins was positive. The patient was given dextrose at rate of 200 gr/24 h, and the day after abdominal pain resolved and the urine colour became clearer. The patient never required intubation as hypoxemia responded to oxygen supplementation through facemask, and neurological status never deteriorated, so the trismus did not interfere with airway patency.
The patient was then readmitted to infective disease ward for genetic diagnosis.
# Discussion
Many drugs used during recovery were useful in the hypothesis of tetanus but not safe (metoclopramide) or contraindicated (diazepam and diclofenac) in AIP and may have worsened the attack, which was probably induced by the unspecified antibiotic used to treat sinusitis. Red urine output should have risen before the suspicion of AIP, but very difficult catheterization was misleading, until associated with nonimproving abdominal pain, but the most misleading factor was the previous diagnosis of tetanus on the basis of anamnesis and mild trismus only, though abdomen was soft, with no muscular spasm. The presence of trismus contrasts the usual weakness or paralysis typical of AIP, and in this case remains unexplained. Differential diagnosis of trismus includes malignant neuroleptic syndrome, but in this case no neuroleptic was administered before or after admission to intensive care unit. We enhance then the importance of performing early urinary test for porphyrins in the presence of abdominal pain and red urine output. |
Inference of Gene Flow between Species under Misspecified Models
Genomic sequence data provide a rich source of information about the history of species divergence and interspecific hybridization or introgression. Despite recent advances in genomics and statistical methods, it remains challenging to infer gene flow, and as a result, one may have to estimate introgression rates and times under misspecified models. Here we use mathematical analysis and computer simulation to examine estimation bias and issues of interpretation when the model of gene flow is misspecified in analysis of genomic datasets, for example, if introgression is assigned to the wrong lineages. In the case of two species, we establish a correspondence between the migration rate in the continuous migration model and the introgression probability in the introgression model. When gene flow occurs continuously through time but in the analysis is assumed to occur at a fixed time point, common evolutionary parameters such as species divergence times are surprisingly well estimated. However, the time of introgression tends to be estimated towards the recent end of the period of continuous gene flow. When introgression events are assigned incorrectly to the parental or daughter lineages, introgression times tend to collapse onto species divergence times, with introgression probabilities underestimated. Overall, our analyses suggest that the simple introgression model is useful for extracting information concerning between-specific gene flow and divergence even when the model may be misspecified. However, for reliable inference of gene flow it is important to include multiple samples per species, in particular, from hybridizing species.
# Introduction
Hybridization can enhance variation in recipient species, and has long been recognized as an important process in plants that can stimulate the origin of new species . Analyses of genomic data in the past decade have highlighted the prevalence of hybridization or introgression in animals as well, including bears [bib_ref] Population genomics reveal recent speciation and rapid evolutionary adaptation in polar bears, Liu [/bib_ref] [bib_ref] The evolutionary history of bears is characterized by gene flow across species, Kumar [/bib_ref] , birds [bib_ref] The genomic landscape of species divergence in Ficedula flycatchers, Ellegren [/bib_ref] , and butterflies [bib_ref] Genome-wide evidence for speciation with gene flow in Heliconius butterflies, Martin [/bib_ref]. Between-species gene flow may involve either sister or non-sister species and may play an important role in ecological adaptation [bib_ref] How reticulated are species?, Mallet [/bib_ref] [bib_ref] Interpreting the genomic landscape of introgression, Martin [/bib_ref]. Gene flow can be a major contributor of genealogical variation across the genome and gene tree-species tree discordance, in addition to ancestral polymorphism or delayed coalescence [bib_ref] Gene trees in species trees, Maddison [/bib_ref] [bib_ref] Gene trees and species trees are not the same, Nichols [/bib_ref].
There is a long history of studies in population genetics of models of population subdivision and migration [bib_ref] Isolation by distance, Wright [/bib_ref] [bib_ref] Les mathematiques de I'heredite. Paris: Masson. Mallet J. 2007. Hybrid speciation, Malecot [/bib_ref] [bib_ref] Gene flow and the geographic structure of natural populations, Slatkin [/bib_ref] , and a number of methods have been developed to estimate the migration rate between populations [bib_ref] Effect of unsampled populations on the estimation of population sizes and migration..., Beerli [/bib_ref] [bib_ref] Inference from gene trees in a subdivided population, Bahlo [/bib_ref]. An important limitation of models of population subdivision, when applied to data from different species or subspecies, is that they do not account for the divergence history of the populations or species. Introducing a population/ species phylogeny into models of population subdivision not only improves the realism of the model but also opens up opportunities for addressing a number of interesting questions in evolutionary biology, such as estimating species divergence times and ancestral population sizes, delineating species boundaries, and inferring the direction, rate, and timing of gene flow .
Two classes of models of gene flow have been developed that accommodate the phylogeny of the species, both of which are extensions of the multispecies coalescent (MSC) model [bib_ref] Bayes estimation of species divergence times and ancestral population sizes using DNA..., Rannala [/bib_ref]. The first is the MSC-with-migration model (MSC-M, or isolation-with-migration or IM model, [bib_ref] Multilocus methods for estimating population sizes, migration rates and divergence time, with..., Hey [/bib_ref] [bib_ref] Isolation with migration models for more than two populations, Hey [/bib_ref] [bib_ref] Maximum likelihood implementation of an isolation-with-migration model with three species for testing..., Zhu [/bib_ref] [bib_ref] Maximum likelihood implementation of an isolation-with-migration model for three species, Dalquen [/bib_ref] [bib_ref] Phylogeny estimation by integration over isolation with migration models, Hey [/bib_ref] , which assumes that two MBE species exchange migrants at a certain rate over an extended time period. The rate of gene flow from species A to B is measured by the proportion of migrants (m AB ) in the receiving population B or by the population migration rate, M AB = N B m AB , the expected number of immigrants from A to B per generation, where N B is the (effective) population size of species B. We note that the isolation-with-initial-migration (IIM) model of [bib_ref] Inference of gene flow in the process of speciation: an efficient maximum-likelihood..., Costa [/bib_ref] , which assumes that gene flow occurs initially after species divergence but stops after a period of time when reproductive isolation has been fully established, is an instance of the MSC-M or IM model (see below). The second class of models of gene flow is the MSC-with-introgression (MSci) model , also known as multispecies network coalescent model (MSNC; [bib_ref] Coestimating reticulate phylogenies and gene trees from multilocus sequence data, Wen [/bib_ref] [bib_ref] Bayesian inference of species networks from multilocus sequence data, Zhang [/bib_ref] , which assumes that gene flow occurs at fixed time points in the past. The rate of gene flow is measured by the introgression probability (φ or γ), which is the proportion of successful immigrants in the population at the time of introgression.
In the real world, introgressed alleles may be removed by natural selection because they are involved in hybrid incompatibility and are deleterious in the genetic background of the recipient population [bib_ref] Isolating mechanisms, evolution, and temperature, Muller [/bib_ref] or because they are linked to such loci [bib_ref] The effect on neutral gene flow of selection at a linked locus, Petry [/bib_ref] ; [bib_ref] The barrier to genetic exchange between hybridising populations, Barton [/bib_ref] [bib_ref] Adaptive gene introgression after secondary contact, Uecker [/bib_ref]. Thus the rate of gene flow (M in MSC-M or φ in MSci), when those models are used to analyze genomic sequence data, reflect the long-term effects of selection and drift as well as hybridization or introgression [bib_ref] Interpreting the genomic landscape of introgression, Martin [/bib_ref]. Such an effective rate of gene flow may be expected to vary across the genome, influenced by the presence of loci in the genomic region important in ecological adaptation and by the local recombination rate [bib_ref] The effects of linkage and gene flow on local adaptation: a two-locus..., Bürger [/bib_ref] [bib_ref] The effect of linkage on establishment and survival of locally beneficial mutations, Aeschbacher [/bib_ref] [bib_ref] The consequences of gene flow for local adaptation and differentiation: a two-locus..., Akerman [/bib_ref] [bib_ref] Natural selection interacts with recombination to shape the evolution of hybrid genomes, Schumer [/bib_ref] [bib_ref] Genomic architecture and introgression shape a butterfly radiation, Edelman [/bib_ref] [bib_ref] Recombination rate variation shapes barriers to introgression across butterfly genomes, Martin [/bib_ref]. The rate may also vary over time, depending on geological or ecological events that cause changes in the ecology and distribution of the species and in the chance for two species to exchange genes. One can envisage models of gene flow in which the rate varies over time and across genomic regions. For the present, such extended models are not yet implemented in the MSC framework, and the feasibility of fitting such parameter-rich models to genomic datasets is unexplored. MSC-M and MSci models implemented to date [bib_ref] Maximum likelihood implementation of an isolation-with-migration model for three species, Dalquen [/bib_ref] [bib_ref] Phylogeny estimation by integration over isolation with migration models, Hey [/bib_ref] [bib_ref] Coestimating reticulate phylogenies and gene trees from multilocus sequence data, Wen [/bib_ref] [bib_ref] Bayesian inference of species networks from multilocus sequence data, Zhang [/bib_ref] [bib_ref] A Bayesian implementation of the multispecies coalescent model with introgression for phylogenomic..., Flouri [/bib_ref] assume constant rates, and should be considered first approximations when applied to genomic sequence data.
In this paper, we use mathematical analysis and computer simulation to examine the impact of model misspecification on estimation of parameters under the MSci model, such as species divergence and introgression times, population sizes, and introgression probabilities. We use the Bayesian program Bayesian phylogenetics and phylogeography (BPP) [bib_ref] Species tree inference with BPP using genomic sequences and the multispecies coalescent, Flouri [/bib_ref] to analyze multilocus sequence data simulated under various MSci and MSC-M models. Although BPP is our own implementation of the MSci model, our results should apply to similar exact or likelihood methods [bib_ref] Coestimating reticulate phylogenies and gene trees from multilocus sequence data, Wen [/bib_ref] [bib_ref] Bayesian inference of species networks from multilocus sequence data, Zhang [/bib_ref]. Our results may also apply to approximate methods, which use summaries of the data such as the genome-wide sitepattern counts (as in the D-statistic, [bib_ref] A draft sequence of the Neandertal genome, Green [/bib_ref] [bib_ref] Detecting hybrid speciation in the presence of incomplete lineage sorting using gene..., Meng [/bib_ref] [bib_ref] HyDe: a Python package for genome-scale hybridization detection, Blischak [/bib_ref] , reconstructed gene trees (as in SNAQ, Solis-Lemus and Ane 2016), or other summary statistics used in Approximate Bayesian Computation [bib_ref] Approximate Bayesian computation untangles signatures of contemporary and historical hybridization between two..., Dittberner [/bib_ref]. However, approximate methods do not make a full use of information in the data and may not identify all parameters in the model. For example, the D-statistic is agnostic of the mode of gene flow (migration versus introgression) and cannot be applied to data sampled from only two species or populations. The computational strengths and statistical weaknesses of approximate methods have been discussed by a number of authors [bib_ref] Modeling hybridization under the network multispecies coalescent, Degnan [/bib_ref] [bib_ref] Advances in computational methods for phylogenetic networks in the presence of hybridization, Elworth [/bib_ref] [bib_ref] Complexity of the simplest species tree problem, Zhu [/bib_ref] [bib_ref] Phylogenomic approaches to detecting and characterizing introgression, Hibbins [/bib_ref] [bib_ref] Significant cross-species gene flow detected in the Tamias quadrivittatus group of North..., Ji [/bib_ref]. In contrast, likelihood methods integrate over all possible gene trees underlying the sequence alignments, making use of all information about the model and parameters in the sequence data. They typically involve a heavy computational load. However, recent algorithmic improvements have made it possible to apply the MSci model to genome-scale datasets with more than 10,000 loci ). Inferring introgression events or constructing an introgression model using genomic sequence data, however, remains a challenging task, even when a binary species tree is specified, onto which introgression events can be added [bib_ref] Significant cross-species gene flow detected in the Tamias quadrivittatus group of North..., Ji [/bib_ref] [bib_ref] Full-likelihood genomic analysis clarifies a complex history of species divergence and introgression:..., Thawornwattana [/bib_ref] ; see Discussion for an overview of currently available methods for inferring gene flow on a species phylogeny. For these and many other reasons, the model of gene flow assumed in our data analysis may often be incorrect. An important question is to what extent inference of gene flow and estimation of the timing and rate of gene flow can still be achieved when the model of gene flow is misspecified. The impact of model misspecification on estimation of other evolutionary parameters such as species divergence times is also of major concern.
Although there are many ways in which the assumed model is wrong, we are particularly interested in a few types that are likely in real data analyses [bib_ref] Genome-scale data reveal deep lineage divergence and a complex demographic history in..., Finger [/bib_ref] [bib_ref] Full-likelihood genomic analysis clarifies a complex history of species divergence and introgression:..., Thawornwattana [/bib_ref]. First, gene flow may be occurring continuously during a time period but an MSci model is fitted to the genomic data, which assumes that gene flow occurred at a particular time point (e.g., [bib_ref] Coestimating reticulate phylogenies and gene trees from multilocus sequence data, Wen [/bib_ref] [bib_ref] The impact of cross-species gene flow on species tree estimation, Jiao [/bib_ref]. We are here interested in whether species divergence times and ancestral population sizes are affected by the misspecification, and how the migration rate in the migration model (M) corresponds to the introgression probability in the MSci model (φ). The case of two species is analytically tractable. We study the limit of the maximum-likelihood estimates (MLEs) of introgression probability and introgression time when the data size (the number of loci) approaches infinity when the data are generated under the MSC-M MBE model. We use computer simulation to verify and extend the analytical calculation.
Second, the introgression event may be assigned to a wrong branch on the species tree, for example, to a parental or daughter branch of the genuine introgression lineage. Alternatively, introgression may involve species that have since gone extinct or are not included in the data sample. The presence of such ghost species is known to mislead inference of the history of gene flow for the sampled species [bib_ref] Effect of unsampled populations on the estimation of population sizes and migration..., Beerli [/bib_ref] [bib_ref] Ghost lineages highly influence the interpretation of introgression tests, Tricou [/bib_ref]. Thus we conducted simulation to examine the impact of unsampled species on the inference of gene flow. In general, our results demonstrate the usefulness of the simple introgression model in inferring gene flow using genomic sequence data. , we study the asymptotic behavior of Bayesian parameter estimation under the introgression (MSci) model when the data are generated under the migration (IM) model in the case of two species, with one sequence per species per locus. Here we focus on this simple case because it is analytically tractable. Note that our Bayesian implementation in BPP accommodates an arbitrary number of species and an arbitrary number of sequences per species per locus, and the likelihood calculation averages over the gene genealogy for the sequences at each locus. We assume an infinite number of loci, and the data at each locus consist of a pair of sequences (a, b) from the two species, with x differences at n sites. The coalescent time t for the locus is unknown and underlies the observed difference. [bib_ref] The impact of cross-species gene flow on species tree estimation, Jiao [/bib_ref] analyzed the MSC-M modelassuming infinite sequence length (n = ∞) so that the true coalescent time between the two sequences (t) is known. Here we accommodate random fluctuations in the number of mutations due to finite sequence length and consider three variants of the migration model.
# Results
## Correspondence between the msc-m and msci models in the case of two species
## Notation and definition of parameters
In the basic IM model, species A and B diverged at time τ R and there has since been gene flow from A to B at the rate of M AB migrants per generation. The IIM modelassumes that migration occurred initially after species divergence but stopped at time τ T > 0 [bib_ref] Inference of gene flow in the process of speciation: an efficient maximum-likelihood..., Costa [/bib_ref] , and is represented by an MSC-M model for three species including a ghost species. Here the ghost does not necessarily represent a real species but is a mathematical device for specifying the IIM model. The IIM model becomes identical to the IM model when τ T = 0. We also consider a secondary contact (SC) model, in which two species initially had complete isolation but came into contact at a certain time point (τ T ) with ongoing gene flow at the rate of M AB ever since [bib_ref] Inference of gene flow in the process of speciation: efficient maximum-likelihood implementation..., Costa [/bib_ref]. This is similarly specified using a ghost species at time point τ T. The migration model involves three types of parameters: species divergence times (τ R , τ T ), population sizes for extant, and extinct species (θ A , θ B , θ T , θ R ), and the (population) migration rate M AB . The population size parameter for any species with (effective) population size N is defined as θ = 4Nμ, where μ is the mutation rate per site per generation. We refer to a branch on the species tree by its daughter node so that branch RA is also branch A, with population size parameter θ A . Both divergence times (τ) and population sizes (θ) are measured by the expected number of mutations per site.
## Asymptotic theory
We first consider the IIM model of figure 1b, of which the IM model of figure 1a is a special case with τ T = 0. The backwards-in-time process of coalescent and migration in time interval (τ T , τ R ) is described by a Markov chain with
## Mbe
three states: AB, AA, and A [bib_ref] The coalescent and the genealogical process in geographically structured population, Notohara [/bib_ref]. Here AB is the initial state, with two sequences in the sample, one in A and another in B; AA means both sequences are in A (in other words, sequence b is traced back into A); and A means one sequence in A (in other words, sequence b is traced back into A and has coalesced with sequence a). Note that in the Markov chain, time runs backwards, so the transition from AB to AA means migration of a sequence from A to B in the real world. The generator matrix for the Markov chain is (see, e.g., [bib_ref] The coalescent and the genealogical process in geographically structured population, Notohara [/bib_ref] [bib_ref] The impact of cross-species gene flow on species tree estimation, Jiao [/bib_ref]
[formula] Q = AB AA A AB −w w 0 AA 0 − 2 θ A 2 θ A A 0 0 0(1) [/formula]
where w = m AB /μ = 4M AB /θ B is the mutation-scaled migration rate, and 2/θ A is the coalescent rate in population A, with one time unit being the expected time taken to accumulate one mutation per site. Q has eigenvalues λ 1 = 0, λ 2 = −2/θ A , and λ 3 = −w.
Let the transition probability matrix over time t be P(t) = {p ij (t)} = e Qt , where p ij (t) is the probability that the Markov chain will be in state j time t later given that it is in state i at time 0. This is
[formula] P(t) = e −wt θ A w 2−θ A w (e −wt − e − 2 θ A t ) 1 − 2e −wt −θ A we − 2 θ A t 2−θ A w 0 e − 2 θ A t 1 − e − 2 θ A t 0 0 1 ⎡ ⎢ ⎢ ⎣ ⎤ ⎥ ⎥ ⎦ .(2) [/formula]
The probability density of coalescent time t is thus
[formula] f iim (t) = P AB,AA (t − τ T ) 2 θ A , if τ T < t < τ R , [1 − P AB,A (τ R − τ T )] 2 θ R e − 2 θ R (t−τ R ) , if t > τ R = 2w 2−θ A w [e −w(t−τ T ) − e − 2 θ A (t−τ T ) ], if τ T < t < τ R , 2 2−θ A w e −w(τ R −τ T ) − θ A w 2−θ A w e − 2 θ A (τ R −τ T ) 2 θ R e − 2 θ R (t−τ R ) , if t > τ R . ⎧ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎩(3) [/formula]
This is a function of w = 4M AB /θ B but not of M AB and θ B individually. The parameters specifying the density are thus Θ iim = (w, θ A , θ R , τ R , τ T ). Note that the density under the IM model f im (t) is given by f iim (t) with τ T = 0.
Similarly under the secondary-contact (SC) model, the coalescent-with-migration process over the time interval (0, τ T ) is described by the Markov chain of equation (1). Given the parameters Θ m , the probability density of coalescent time t is
[formula] f sc (t) = P AB,AA (t) 2 θ A , if 0 < t < τ T , P AB,AA (τ T ) 2 θ A e − 2 θ A (t−τ T ) , if τ T < t < τ R , P AB,AA (τ T ) 2 θ A e − 2 θ A (τ R −τ T ) + P AB,AB (τ T ) × 2 θ R e − 2 θ R (t−τ R ) , if t > τ R ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎪ ⎩ = wθ A 2−wθ A [e −wt − e − 2 θ A t ] 2 θ A , if 0 < t < τ T , wθ A 2−wθ A [e −wτ T − e − 2 θ A τ T ] 2 θ A e − 2 θ A (t−τ T ) , if τ T < t < τ R , wθ A 2−wθ A [e −wτ T − e − 2 θ A τT ]e − 2 θ A (τR−τT) + e −wτ T × 2 θ R e − 2 θ R (t−τ R ) , if t > τ R . ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎩(4) [/formula]
Under the MSci model, we have (e.g., [bib_ref] The impact of cross-species gene flow on species tree estimation, Jiao [/bib_ref]
[formula] f i (t) = φ 2 θ S e − 2 θ S (t−τ S ) , if τ S < t < τ R , [φe − 2 θ S (τ R −τ S ) + (1 − φ)] 2 θ R e − 2 θ R (t−τ R ) , if t > τ R . ⎧ ⎨ ⎩ (5) [/formula]
This is a function of parameters
[formula] Θ i = (φ, θ R , θ S , τ R , τ S ). [/formula]
Given the coalescent time t for a locus, the probability of observing x differences at n sites under the JC mutation modelis given by the binomial probability
[formula] f (x | t) = ( 3 4 − 3 4 e − 8 3 t ) x · ( 1 4 + 3 4 e − 8 3 t ) n−x .(6) [/formula]
The marginal probability of observing x differences at n sites, under both the migration (IM, IIM, SC) and introgression (MSci) models, is
[formula] f(x | Θ)= ∫ ∞ 0 f(x | t)f (t | Θ) dt,(7) [/formula]
where f(t|Θ) is given by equations (3), (4), or (5). For analytical tractability of the likelihood (eq. 7), we assume the infinite-sites mutation model instead of JC, and replace the binomial likelihood by a Poisson approximation
[formula] f (x | t) = 1 x! (2nt) x e −2nt .(8)D(Θ m ∥ Θ i ) = n x=0 f m (x | Θ m ) log f m (x | Θ m ) f i (x | Θ i ) ,(9) [/formula]
where the subscript "m" stands for any of the three MSC-M models ("im" for IM, "iim" for IIM, or "sc" for SC,. The KL divergence is a measure of distance from the fitting introgression model to the true migration model: here Θ m are fixed, whereas Θ i are being estimated.
The limiting values Θ * i as L → ∞ are also known as the pseudo-true parameter values for the misspecified MSci model. The BFGS optimization routine in PAML [bib_ref] PAML 4: phylogenetic analysis by maximum likelihood, Yang [/bib_ref] is used to minimize equation (9) to obtain the MLEs.
We are in particular interested in the introgression probability φ and the introgression time τ S . Note that under the migration model, the probability that any lineage from species B traces back to A is
[formula] φ 0 = 1 − e −4M AB Δτ/θ B ,(10) [/formula]
where Δτ is the time period of gene flow. Equation (10) gives the expected proportion of migrants under the true migration model. When M AB is small,
[formula] φ 0 ≈ (4M AB /θ B ) [/formula]
Δτ, which is also given by equating the expected total number of migrants under the two models:
[formula] N B φ 0 ≈ m AB N B Δτ/μ. Note that m AB N B [/formula]
is the expected number of migrants per generation and Δτ/μ is the number of generations with gene flow. It may be noted that the theory of equation (9) can be used to study the limiting parameter estimates (when L → ∞) in the migration model when the true model is the introgression model. One has only to flip the roles of f m (x | Θ m ) and f i (x | Θ i ) in equation (9). This is not pursued in this paper.
## Asymptotic results under the im model
We used the asymptotic theory (eq. 9) to obtain the MLEs (Θ We use five methods (a-e) to fit the MSci model, with method d estimating all five parameters, whereas the others have some parameters fixed [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref]. We examined the effects of the sequence length (n) and the migration rate (M AB ). Note that five parameters are identifiable under the MSci model:, and θ A , θ B , θ H are unidentifiable as no coalescent events can occur in those populations given one sequence per species per locus. Population size θ S is identifiable as it is possible for both sequences a and b to be traced back to population S. Nevertheless, one expects the information concerning θ S to be weak in datasets of two sequences per locus. In methods c and d, θ S and φ are estimated as free parameters. Application of the misspecified MSci model (to data generated under the IM model) led to unreasonably large estimates of θ S (as large as 0.5 mutations per site), and the poor estimates of θ S caused φ to be poorly estimated as well. This is due partly to our use of one sequence per species per locus and partly to the confounding effects between θ S and φ. We discuss both effects later when we describe the simulation results. Here we focus on methods a, b, and e, in which θ S is fixed at the true value θ 0 (in methods a and b) or constrained to be equal to θ R (in method e).
[formula] Θ i = (τ R , τ S , θ R , θ S , φ) (fig [/formula]
In the IM model, migration occurs throughout the time interval (0, τ R ), at the rate of M AB migrants per generation. When such data are analyzed under the introgression model, a simple expectation might be that the introgression time τ S should be the average τ R /2, whereas the introgression probability might be given by the expected proportion φ 0 of equation (10). However, as we show below, this expectation is too simplistic.
First, we discuss the introgression time τ S , assuming the true coalescent time (or n = ∞). Given the datagenerating IM model, there is a strictly positive probability for 0 < t < ϵ for any small constant ϵ > 0 (supplementary, Supplementary Material online). In other words, there must exist loci at which t is arbitrarily close to 0. In the MSci model, sequences a and b cannot coalesce until they are in the same population S, so that τ S < t. When the MSci model is fitted to data generated under the IM model, τ S is dominated by the minimum rather than the average coalescent time, and τ S → τ * S = 0 when the number of loci L → ∞ (and when the true coalescent time t is known). Even though migration occurs throughout the time interval (0, τ R ), the MSci model has to lump all migration events to one time point, τ * S = 0 [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref] and e). With finite sequences (n < ∞), t is not observed and is reflected in the number of mutations (x). Whatever the true t, there is a positive probability of observing no mutations between the two sequences, so that an absence of mutations (x = 0) is not strong evidence for t = 0. The MLE τ * S reflects not only the minimum coalescent time, but also the whole distribution (supplementary, Supplementary Material online). Thus τ * S > 0, different from the case where the coalescent time is known without error (n = ∞). Nevertheless, one expects τ * S to be closer to 0 than to τ R , especially if the number of sites is large. Indeed in our calculations, τ * S ≪ 1 2 τ R [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref] and e).
Next we consider the introgression probability φ and again focus on methods a, b, and e [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref]. The estimate φ * increases nearly linearly when M AB is small (< 1 4 , say) but tails off at large M AB . All estimates are smaller than φ 0 of equation (10) . We defer to a later section a detailed discussion of the estimation of φ, contrasting the IM, IIM, and SC models.
Finally the estimated divergence time between the two species (τ R ) matched the true values at low migration rates but was underestimated at high migration rates, with the ancestral population size θ R overestimated [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref]. It may be tempting to interpret the underestimation of τ R (and overestimation of θ R ) by the MSci model as being due to the difficulty of distinguishing complete isolation with recent species divergence from introgression or of distinguishing migration and coalescent events close to species divergence from ancestral polymorphism. However, this does not appear to be a correct interpretation.
We examined the true and fitted distributions of the coalescent time (supplementary fig. S1, Supplementary Material online). If there is no migration (M AB = 0), the MSci model (with φ = 0) will be correct, and the parameter estimates will converge to the true values, with a perfect fit to the density f m (t). At low migration rates (M AB ≤ 0.1, say), the MSci model fits the density f m (t) very well, with the discontinuity point in the true and fitting distributions coinciding: τ * R = τ m R . At the intermediate rate of M AB = 1, the species divergence time τ R is still correctly estimated even though the fit to the density is poor (supplementary, Supplementary Material online). At high rates (with M AB ≥ 1.4, say), the true density has a mode in the interval (0, τ m R ), dropping off at τ m R . The best fitting density starts from 0, with an exponential decay, and has a discontinuity point at τ R with again an exponential decay. This best-fitting density is a poor fit, and the discontinuity point τ * R is moved to smaller values as an attempt to accommodate the migration and coalescent events in the middle of the interval (0, τ m R ) to improve the fit (judged by the KL divergence). Thus τ R is underestimated (τ * R < τ m R ). As a result, the population size parameter θ R is overestimated, as those two parameters tend to be strongly negatively correlated (e.g., [bib_ref] Estimation of hominoid ancestral population sizes under Bayesian coalescent models incorporating mutation..., Burgess [/bib_ref]. In other words, the intermediate coalescent times in the interval (0, τ R ), which occur at a large proportion of loci, are accommodated or misinterpreted by the MSci model using a smaller τ R and larger θ R . Coalescent times in the range τ * R < t < τ m R , which represent true migration events, are misinterpreted as coalescent events in species R, and φ * is much less than φ 0 (eq. 10). show similar patterns to those under the IM model discussed above. Similarly, θ S is difficult to estimate using two sequences per locus in methods c and d, and the poor estimates of θ S affects the estimation of φ. Thus we focus on methods a, b, and e, in which θ S is fixed or constrained, and on the introgression time and introgression probability. In the IIM model, migration events occur throughout the time interval (τ T , τ R ), but the estimate of the introgression time is dominated or influenced by the minimum coalescent time, so that τ * S = τ T when n = ∞, and τ * S > τ T when n is finite. In the latter case, τ * S is much closer to τ T than to τ R (supplementary [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref] , Supplementary Material online).
The introgression probability φ * grew almost linearly with M AB when M AB was small (with M AB ≤ 0.2, say), and this estimate was close to the expectation φ 0 of equation (10) The results show patterns similar to those under the IM and IIM models discussed above. The species divergence time under the MSci model τ * R = τ (sc) R when the migration rate M AB is small but drops at very high rates (with M AB > 2). The introgression time is dominated by the minimum coalescent time, so that τ * S = 0 when n = ∞, and τ * S is much closer to 0 than to τ T when n is finite (supplementary, Supplementary Material online). Note that in the true model migration occurs throughout the time interval (0, τ T ).
The introgression probability φ * grew almost linearly with the migration rate M AB when M AB was small (with M AB ≤ 1 4 , say), and was close to the expectation φ 0 (eq. 10) when M AB < 2 (supplementary, b and e, Supplementary Material online). At very high rates (M AB > 2), φ * was much smaller than φ 0 , and this 'bias' was accompanied by an underestimation of τ R and overestimation of θ R . Similarly to the IM and IIM models discussed above, this is due to the attempt of the MSci model to accommodate the coalescent times in the middle of the interval (0, τ T ) (supplementary [fig_ref] FIG. 5: Posterior means of speciation/introgression times [/fig_ref] , Supplementary Material online).
## The amount of gene flow under the im, iim, and sc models
Although the expected total amount of gene flow measured by φ 0 (eq. 10) is the same under the IM, IIM, and SC models of figure 1a-c, the estimates under the MSci model differ, as summarized in figure 1e.
At low migration rates, τ R , θ S , and θ R in the MSci model are nearly accurately estimated to match those in the true model (figs. 2, supplementary figures S2 and S4, Supplementary Material online). Consider the case of infinitely long sequences with known coalescent time. Let
[formula] τ * R = τ m R , θ * R = θ m R , [/formula]
and let the introgression time be τ * S = 0 for the IM and SC model, and τ * S = τ T for the IIM model. We also match the probability density of coalescent time t, with f i (t) = f m (t), for t > τ R . With those simplifying assumptions, φ * that minimizes the KL divergence (eq. 9) can be derived as
[formula] φ * (IM) ≈ φ 0 − wθ A 2 (1−e − 2 θ A τ R ) (1− wθ A 2 )(1−e − 2 θ A τ R ) , φ * (IIM) ≈ φ 0 − wθ A 2 (1−e − 2 θ A (τ R −τ T ) ) (1− wθ A 2 )(1−e − 2 θ A (τ R −τ T ) ) , φ * (SC) ≈ φ 0 − wθ A 2−wθ A (e −wτ T −e − 2 θ A τ T )e − 2 θ A (τ R −τ T ) 1−e − 2 θ A τ R .(11) [/formula]
At low migration rates, equation (11) (11), we have
[formula] φ 0 > φ * (SC) > φ * (IIM) = φ * (IM) .(12) [/formula]
In other words, recent gene flow (as in SC) is easier to recover by the MSci model than ancient gene flow (as in IM or IIM). Note that φ * (IIM) = φ * (IM) holds only when one sequence is sampled per species; as there is no coalescent over (0, τ T ), IIM is essentially the same model as IM with a time shift. This will not be the case when multiple sequences per species are sampled or when the sequence length is finite.
# Simulation results
As our asymptotic theory was limited to a single sequence per species per locus, we used simulation to verify and augment our analytical calculations above. We simulated data under the IM, IIM, or SC models of figure 1a-c, using the same parameter values as above, and analyzed them using BPP under the MSci model. The JC mutation modelwas assumed. In the basic setting, we used S = 4 sequences per species per locus, n = 1, 000 sites per sequence, and L = 4, 000 loci in each dataset, with the migration rate M AB = 0.2. We varied n, S, L, M AB MBE to examine their effects. With multiple sequences per species (S > 1), all eight parameters of the MSci model,are identifiable . The results are summarized in figure 3. We note a few common features first. In nearly all cases, population sizes for extant species (θ A , θ B ) were very well estimated, with posterior means close to the true values and with very narrow highest-probability-density (HPD) credibility intervals (CIs). The exception was parameter θ B under the IM model (note that B is the species receiving immigrants), which was less well estimated when the dataset was small and had either short sequences (n = 250) or few loci (L ≤ 500), or when the migration rate was very high. The poorer estimation of θ B appeared to be related to the underestimation of φ and τ R ; see below. The population size for the common ancestor θ R was mostly well estimated, although overestimated at very high migration rates. Population sizes for the ancestral species (θ S , θ H ) are harder to estimate; indeed they had larger CIs and were influenced by misspecification of the model of gene Inference of gene flow · https://doi.org/10.1093/molbev/msac237 MBE flow. As expected from the asymptotic results, τ R was very well estimated, except at very high migration rates, in which case τ R was underestimated (and θ R overestimated).
[formula] Θ i = (φ, θ A , θ B , θ R , θ S , θ H , τ R , τ S ) (fig. [/formula]
Next we examine the effects of n, S, L, M AB in turn. First, the number of sites (n) had a relatively small impact on MSci parameters, when other factors were fixed (at the basic setting of S = 4, L = 4,000, and M AB = 0.2). When n = 250, CIs for parameters such as the introgression time and probability (τ S and φ) were wide. When n ≥ 1, 000, the CIs were much smaller for all parameters. The introgression time τ S decreased slightly as the sequence became longer. This is consistent with the asymptotic analysis, which suggests that τ * S is dominated by the smallest coalescent time or sequence divergence and should be 0 for the IM and SC models or τ T for the IIM model (when n → ∞) (figs. 2, supplementary figures S2 and S4, Supplementary Material online). Similarly, for the IM and SC models, φ increased with the increase of n when M AB was low, as observed in the asymptotic analysis. Under the IIM model, small datasets with short sequences (n = 250) produced very uncertain estimates of φ and θ H (and, to a lesser extent, τ S and θ S ). The two parameters are nearly confounded; this is discussed below when we examine the impact of the migration rate (M AB ).
Second, we varied the number of sequences per species (S). When one sequence per species is in the data (S = 1), only five parameters in the MSci modelare identifiable: θ R , θ S , τ R , τ S , φ). When multiple sequences were sampled per species, all eight parameters are identifiable. They were well estimated when the dataset was large (say, with S ≥ 2 for IM and SC or S ≥ 4 for IIM). Even with S = 4 sequences per species, estimates of φ from data generated under the IIM model involved wide CIs, with τ S being close to τ R , and θ S and θ H being very imprecise as well [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref]. This is due to the semi-unidentifiability or the confounding effects of the parameters, and will be discussed below. Here we note that the problem disappeared and all parameter estimates were well-behaved in large datasets when many sequences were sampled (S ≥ 2 for IM and SC or S ≥ 4 for IIM; [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref].
Third, we examined the impact of the number of loci (L). The IIM model was hard to fit in small datasets with a small number of loci (L ≤ 1, 000), generating large CIs for parameters φ and θ H . This is the same pattern as in the case of short loci (n = 250) or few sequences (S ≤ 2), discussed above. In large datasets, the parameters were well estimated. Note that the number of loci L is the sample size in the statistical model as data at different loci are independently and identically distributed. Theory predicts that in large datasets the variance should be proportional to 1/L (see O'Hagan and Forster 2004 for the case of correctly specified models and Yang and Zhu 2018 for the case of misspecified models), and thus the CI should decrease at the rate of L −1/2 . This prediction held for parameters that were well estimated [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref]. As discussed earlier, the introgression time τ S is dominated by the smallest coalescent time or smallest sequence divergence. Thus increasing the number of loci led to a decrease in the estimated introgression time, and the trend was in particular apparent for the IIM model (under which τ S → τ T when L → ∞ if n = ∞). In all cases, the estimated introgression time (τ S ) was closer to the more recent end of the time interval for gene flow than to the midpoint, that is, τ S < τ R /2 for IM, τ S < (τ R + τ T )/2 for IIM, and τ S < τ T /2 for SC (see.
Finally, we evaluated the impact of the migration rate (M AB ) [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref]. Under the IM model, there is a near linear relationship between the introgression probability φ and M AB at low rates. The amount of gene flow estimated under the MSci model is less than the true amount expected under the IM model (φ 0 of eq. 10) but the two were close at low rates (with M AB < 0.1, say). At very high rates (with M AB > 1.0, say), divergence time τ R was increasingly underestimated and the population size θ R was overestimated. These patterns are the same as observed in the asymptotic analysis of infinite data (L = ∞), and are due to the attempt of the Under the IIM model, φ involved very large uncertainties at low rates (M AB < 0.04, say), with θ S , τ S , and θ H affected as well. Given the small M AB , why did φ not converge to ∼ 0 with narrow CIs? Note that if M AB = 0 in the IIM model, the MSC model with no gene flow or MSci with φ = 0 will be the correct model. Similarly in [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] where M AB = 0.2 was fixed, wide CIs for those parameters were observed in small datasets with short loci (n ≤ 250), few sequences (S ≤ 2), or few loci (L ≤ 1, 000), as noted above. Also in the asymptotic analysis (with L = ∞), we noted that θ * S and φ * were grossly wrong but had no sampling errors because the data size was L = ∞ [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref] ; supplementary figures S2 and S4, Supplementary Material online, methods c, d). We suggest that all those results are due to the near unidentifiability of parameters in the MSci model (in particular, θ S and φ); in other words, the parameters are confounded.
If M AB = 0 in the MSC-M model, MSci with φ = 0 will be the correct model, but φ > 0 with a large appropriately adjusted θ S may provide a very good fit to the data (of two sequences per locus). When M AB is small but nonzero, the MSci model will never achieve a perfect fit, and a large φ with appropriately adjusted θ S may provide a better fit than a small φ. Thus in infinite data (L = ∞), we may get grossly wrong estimates with no uncertainty (supplementary figure S2, Supplementary Material online, methods c, d). In finite datasets (L < ∞), there will be a ridge in the posterior surface involving φ and θ S , leading to wide CIs for those parameters, influenced by both model misspecification and the prior [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref]. Including multiple samples from the same species (S > 1) is useful for improving the information content in the data, but strong correlation between φ and θ S may be expected nevertheless. In this regard, the large uncertainties in posterior estimates of parameters may be useful as they help MBE the investigator avoid incorrect inferences of a large φ when gene flow is minimal.
## Introgression events assigned to wrong branches
We conducted simulations to examine the bias in parameter estimates when the introgression event is assigned on either the parental or daughter branch of the lineage genuinely involved in introgression. The data were simulated under model trees A or B and analyzed under models A or B of figure 4a and b.
In the A-A and B-B settings, the correct MSci model was assumed, and the performance of the method serves as a reference for comparison. Most parameters,
[formula] τ R = 4θ 0 , τ S = 3θ 0 , τ T = 2θ 0 , and τ X = τ Y = 1.5θ 0 . In MSci model B, τ R = 4θ 0 , τ S = 3θ 0 , τ T = θ 0 , and τ X = τ Y = 1.5θ 0 . [/formula]
Introgression probability is φ = 0.2. In the IM model C, τ R = 4θ 0 , τ S = 3θ 0 , and τ T = 2θ 0 , with migration occurring from species A to B over time period (0, τ T ) at the rate M = 0.1 migrants per generation. In the IM model D, τ R = 4θ 0 , τ S = 3θ 0 , and τ T = θ 0 , with migration from species A to ST over time period (τ T , τ S ) at the rate M = 0.1. (e) The 95% HPD CIs for parameters in 100 replicate datasets of L = 250, 1,000, and 4,000 loci. Column labels refer to the simulation model followed by the analysis model; for example, "B-A" means that data were simulated under model B and analyzed under model A. The values of θ and τ parameters are multiplied by 10 3 . Black solid line indicates the true value.
## Mbe
including the species divergence times (τ R , τ S , τ T , and τ X = τ Y ) and population sizes for extant species (θ A , θ B , θ C , θ D ), were well estimated. For well-estimated parameters, the CI width reduced by a half as the number of loci (L) quadrupled, as predicted by theory. Population sizes for ancestral species (θ R , θ S , θ T , θ X , and θ Y ) were less well estimated, although performance improved with sample size: with L = 4, 000 loci, these parameters were well estimated. Introgression probability (φ) was well estimated, but thousands of loci were necessary to obtain precise estimates with narrow CIs under the standard settings used here (four sequences per species per locus and 500 sites per sequence).
In the other settings, there was mismatch between the models used to simulate and to analyze data. We note that population sizes for extant species (θ A , θ B , θ C , θ D ) were well estimated, as was the age of the root (τ R ). Performance for estimation of those parameters was very similar whether or not there was model misspecification (e.g., the A-B setting versus the B-B setting and C-A versus A-A). Below we focus on estimation of the other parameters.
In the A-B setting, data were simulated under model A with A → B introgressionbut analyzed under model B with introgression incorrectly assigned to the parental branch ST. Ancestral population sizes θ R and θ S were well estimated, similar to the B-B setting. Divergence times τ R and τ S were well estimated, but τ T and τ X were stuck together. We expect τ T to be mostly determined by the smallest sequence divergence (t bc ) between B and C, which should be close to τ (A) T = 2θ 0 = 0.004. Here, we use the superscript to indicate the model in which the parameter is defined. In the fitting model B, the introgression time τ (B) X (which is >τ (B) T ) should reflect the smallest sequence divergence t ab , whereas in the true model A, t ab is mostly determined by τ X (which is <τ T ). Thus misidentification of the introgression lineage caused τ (B) X to be stuck at τ (B) T [fig_ref] FIG. 5: Posterior means of speciation/introgression times [/fig_ref]. There was virtually no information for θ T as the population was estimated to have near-zero time duration with no chance for coalescence. The introgression probability was seriously underestimated, converging to φ * A−B ≈ 0.12 when the number of loci L increases (table 1), whereas the true value was 0.2. This smaller estimate of introgression probability is explained by the distribution of coalescent times between species in the true and fitting models (supplementary. Under the true model A, sequences from A and B are more similar than those between A and C due to the A → B introgression, with an excess of small coalescence time t ab . Under the analysis model B, t ab and t ac have the same distribution. Thus the true model predicts an excess of small t ab , whereas the fitting model predicts an excess of small t ac , and having a smaller φ in the fitting model helps to reduce the discrepancy.
In the B-A setting, the simulation model (MSci model B ofassumes introgression involving the ancestral branch ST but the analysis model (model A) assigned introgression to the daughter branch TB. Posterior means and CIs for divergence times τ R and τ T were similar to those in the A-A setting. Note that τ (A) T should be mostly determined by the smallest sequence divergence (t bc ) between B and C, and given that this is τ (B) T = 2θ 0 = 0.002, τ (A) T was well estimated, unaffected by mis-assigned introgression event. Although the true introgression time τ X was 0.003, it was forced to be less than τ T by the analysis model A. As the number of loci increases, τ (A) X became stuck at τ (A) T [fig_ref] FIG. 5: Posterior means of speciation/introgression times [/fig_ref]. However, τ (A) S was seriously underestimated. This may be explained as follows. In the analysis model A, τ (A) S was mostly determined by the shortest sequence distance between A and C. In the true model B, this should be close to τ (B) X = 1.5θ 0 = 0.003, due to introgression. With mutational fluctuations in the sequences, one can expect τ (A) S to lie between (τ (B) X , τ (B) S ) = (1.5θ 0 , 3θ 0 ), but closer to τ (B) X in large datasets with many sites and/or many loci. Population sizes θ (A) S and θ (A) Y were affected by the mis-assigned introgression events as well, as those populations are close to the introgression branches. In particular, θ (A) Y was very imprecise as branch YT was very short, and θ (A) S was overestimated because τ (A) S was seriously underestimated (as those two parameters are negatively correlated). Finally, the introgression probability (φ) was underestimated, apparently converging to φ * B−A ≈ 0.02 when the number of loci L increased (table 1), whereas the true value was 0.2. This greatly reduced introgression probability appeared to reflect the very poor fit of the misspecified model A to data generated under model B (see the large differences between the true and fitting distributions of coalescent times in, Supplementary Material online, second row). As τ (A) X and τ (A) S are seriously underestimated, an excess of small coalescent times (t ab , t ac ) is expected in the fitting model A but does not appear in the data, so that having a smaller φ improves the fit.
In summary, assigning introgression events to a wrong parental or daughter branch led to biased estimates of introgression times (causing the introgression events to collapse onto speciation events) and to seriously underestimated introgression probabilities.
## Continuous migration versus episodic introgression
In this set of simulations, we generated data under the IM models C and D of figure 4c and d and analyzed them under the MSci models A and B, with the mode of gene flow misspecified and with gene flow assigned to either the correct branch or a wrong branch on the species tree.
In the C-A and D-B settings, gene flow occurred continuously but the data were analyzed under the MSci model assuming introgression at a time point. The mode of gene flow was misspecified, but the lineages involved were correctly identified. In the C-A setting, gene flow was between non-sister species, whereas in the D-B setting it was between sister species. Speciation times (τ R , τ S , τ T ) and population sizes (θ) were well estimated, similar to the A-A setting. Surprisingly ancestral population sizes MBE θ T , θ X , θ Y appeared to be even better estimated, with narrower CIs, in the C-A setting than in A-A. Speciation times and population sizes were extremely similar between settings D-B and B-B. Those results were consistent with the results for the case of two species [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] , which showed that at low migration rates, species divergence times and population sizes were well estimated under the MSci model when the data were generated under the IM model.
In the C-A setting, the estimated introgression time τ (A) X appeared to converge (when L increased) to 0.0011, much more recent than the average time of gene flow (τ (C) T /2 = 0.002), and the introgression probability φ C−A appeared to converge to φ * C−A = 0.12 [fig_ref] Table 1: Average [/fig_ref] , smaller than the expected proportion of total migrants:
[formula] φ 0 = 1 − e −4M AB τ (C) [/formula]
T /θ (C) B = 0.148. As discussed earlier for the case of two species, the limiting value for τ (A) X was nonzero, as the sequence length is finite, and the MLE φ C−A slightly underestimated the true amount of gene flow. In the D-B setting, the introgression time τ (B) X appeared to converge to 0.0027, larger than τ (D) T = 0.002 but much smaller than the average time of gene flow, 1 2 (τ (D) S + τ (D) T ) = 0.004, and the introgression probability φ D−B appeared to converge to φ * D−B = 0.08 [fig_ref] Table 1: Average [/fig_ref] , much smaller than φ 0 = 0.148 from equation (10). In both the C-A and D-B settings, the estimated introgression time was within the time interval of gene flow, but closer to the time when gene flow stopped, whereas the amount of gene flow was underestimated
[formula] (φ * C−A < φ 0 , φ * D−B < φ 0 ). Moreover, we have φ * D−B < φ * C−A . [/formula]
These patterns are consistent with our analysis of the two-species case at low migration rates (eq. 12, [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] , which suggested that gene flow after a period of isolation (the SC model) is easier to recover than gene flow that starts at speciation but stops some time afterwards (the IIM model).
In the C-B and D-A settings, the mode of gene flow was misspecified and furthermore gene flow was assigned onto the wrong branch of the species tree. In the C-B setting, divergence time τ T was underestimated slightly, due to gene flow assigned to the wrong branch, as observed in the A-B setting. Ancestral population sizes θ T and θ Y were affected by gene flow, similar to the A-B setting. Model B forces τ X > τ T . Thus we expect τ (B) X and τ (B) T to get stuck together, with both being smaller than Inference of gene flow · https://doi.org/10.1093/molbev/msac237 MBE τ (C) T = 2θ 0 = 0.004; as the number of loci L increased, τ (B) X appeared to converge to 0.0029, and φ C−B to φ * C−B = 0.10 [fig_ref] Table 1: Average [/fig_ref].
In the D-A setting, the divergence time τ S was underestimated, due to gene flow assigned to the wrong branch, similarly to the B-A setting. The ancestral population sizes θ R and θ X were well estimated as in the A-A setting, but θ T had a slight positive bias. The ancestral population sizes θ S and θ Y were affected by the gene flow, similar to the B-A setting. The introgression time and probability (τ X and φ) do not exist in the simulation model D. Model A forces τ X < τ T , so we expect τ (A) X to be close to τ (D) T = θ 0 = 0.002; when the number of loci L increased, τ (A) X appeared to converge to 0.00186, and
[formula] φ D−A to φ * D−A = 0.02 (table 1). Note that φ 0 >φ C−B >φ D−A with φ C−B <φ C−A and φ D−A <φ D−B . [/formula]
Those results are consistent with our early results for fitting the MSci model to data generated under the migration model in the two-species case (eq. 12, [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] , and with the results for the A-B and B-A settings that assignment of gene flow to a wrong branch reduces the estimate of φ.
In summary, the estimated introgression probabilities, at 0.12, 0.08, 0.10, and 0.02 for the C-A, D-B, C-B, and D-A settings, respectively, even though the total amount of gene flow was the same in models C and D [fig_ref] Table 1: Average [/fig_ref] , suggest the following general patterns. First, the MSci model underestimates the total amount of gene flow if gene flow occurs continuously in every generation (i.e., φ C−A < φ 0 ,φ D−B < φ 0 ), as discussed in our analysis of the two-species case. Second, assigning gene-flow events to wrong lineages led to serious underestimation of the amount of gene flow (i.e., φ C−B <φ C−A ,φ D−A <φ D−B ). Third, recent gene flow in the data is more easily recovered (i.e., φ C−A >φ D−B ,φ C−B >φ D−A ).
## Isolation with initial migration (iim) model
Next, we assessed the effects of taxon sampling when the mode of gene flow is misspecified. We used the IIM model for three species of figure 6a to simulate data and analyzed them under the MSci model of figure 6b. Species divergence times (τ R , τ S ) and population sizes (θ A , θ B , θ C , θ R , θ S , and even θ X and θ Y ) were well estimated. We expect the estimated introgression time τ X to converge to τ T = θ 0 = 0.002 if the sequence length is infinite and to a higher limit for finite sequence length. In our simulation τ X ≈ 0.00283 at L = 4, 000 (table 1). The estimated introgression probability (φ) converged to a nonzero limit, ∼ 0.08 [fig_ref] Table 1: Average [/fig_ref] , compared with φ 0 = 0.148 by equation (10).
The IIM model ofis very similar to the two-species model of figure 1b except that here the tree is larger with more species, and serves to highlight the fact that the impact of the misspecification of the model of gene flow is local. The case is also similar to the D-B setting of, with the only difference that here the hybridizing species T had only one descendent species sampled in the data, whereas in figure 4 (D-B) it had two descendent species sampled. Thus estimates of parameters such as the introgression probability and introgression time were similar to those in the D-B setting of figure 4 but with wider CIs [fig_ref] Table 1: Average [/fig_ref]. Unlike approximate methods designed to work with species triplets or quartets only, the Bayesian approach accommodates an arbitrary number of species in the data (with arbitrary data configurations at each locus), so that the difference in taxon sampling has only the effect of affecting the information content in the data.
## Ghost species
We considered two scenarios in which a species that contributed migrants to extant species has gone extinct or is otherwise unsampled in the data. Note that existence of extinct or unsampled species that received genetic materials from ancestors of extant species in the sample is not relevant to the analysis of the sampled data and does not constitute a model misspecification. In the first scenario, model A ′ of figure 7a ′ is used to simulate data, which assumes that species XUV contributed migrants to species B but is not included in the sample. Note that this model is equivalent to model A of figure 7a. When we fit model B, the only incorrect assumption is the constraint that τ X = τ Y . This is a minor misspecification. Indeed all parameters shared between the simulation model and the analysis model were well estimated. The estimates of introgression time, τ X =τ Y ≈ 0.0028 [fig_ref] Table 1: Average [/fig_ref] , were close to the average of the two parameters in the true model (0.0025). Introgression probability φ ≈ 0.21 (table 1) was also close to the true value (0.2). The existence of the ghost species (XUV) had very little effect on the inference.
In the second scenario , the true model assumes continuous migration involving intermediate ancestral species that have gone extinct, and the MSci model was fitted to data sampled from extant species. Divergence times τ R and τ T were very well estimated, as were the population sizes shared between the simulation and analysis models (θ A , θ B , θ C , θ R ). We expect τ T in model B to be dominated by the minimum coalescent time t ab between sequences from A and B, and this is given by
[formula] τ (A) [/formula]
T . Gene flow from branches RC to SU over the time interval (τ U , τ S ) and then from SU to TB during (τ U , τ T ) was interpreted as introgression in the MSci model. The effective rate for this migration may be close to M CU M UB = 0.04,. The introgression time τ X should be between τ U = θ 0 = 0.002 and τ T = 2θ 0 = 0.004 and the estimate was ≈ 0.0030. Note that both θ T and θ Y were overestimated . Branch T of figure 8b corresponds to branches RS and ST of figure 8a, with population size θ 0 = 0.002. Branch Y corresponds to a segment of branch TB over the time interval (τ U , τ T ), with θ 1 = 0.01. Overestimation of θ Y (and θ T ) may be because there is a deficit of t bb over the interval (τ U , τ S ) due to gene flow, and the fitting MSci model, with the amount of gene flow underestimated (φ < φ 0 ), used large θ Y and θ T to compensate.
[formula] giving φ 0 = 1 − e −4×M CU M UB ×(τ T −τ U )/θ B = 0.031. The estimate was φ X ≈ 0.02 ( [/formula]
## Mbe
# Discussion
The Mode of Gene Flow and the Utility of Misspecified Introgression Models
The asymptotic theory, even though based on only two species with one sequence sampled per species per locus, has been very useful. It generated a number of insights that were confirmed and extended in our simulation. Together the theory and simulation suggest the following correspondence between the MSC-M and MSci models. When gene flow occurs continuously over an extended time period after divergence of two species and we fit the introgression model, the estimated introgression time tends to be closer to the more recent end of the time period of gene flow, because the introgression time is dominated by the most recent coalescent time or the minimum sequence divergence between species. If the true coalescent time is known and used as data, the introgression time will converge to the time when gene flow stopped. At low migration rates (M < 1 4 , say), the species divergence time is correctly estimated by the MSci model, and the introgression probability φ is lower than but close to the expected proportion of migrants (φ * < φ 0 ). The estimate is particularly close under the secondary-contact model (supplementary, Supplementary Material online). At very high migration rates, the estimated introgression probability φ * may be much less than φ 0 , and furthermore the species divergence time is underestimated to account for intermediate coalescent times generated under the MSC-M model. Recent gene flow (as in the SC model) is easier to recover (with φ * closer to φ 0 ) than ancient gene flow (as in the IIM model).
The accurate estimation of species divergence times under the MSci model despite the misspecification, at least at lower migration rates (e.g., τ R for M ≤ 0.3 in [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] , may be worth emphasizing. It is well known that ignoring gene flow between two species may lead to serious underestimation of the species divergence time. Here our results suggest that if gene flow is continuous, the MSci model assuming introgression at a fixed time point still gives reliable estimates of the species divergence time. The estimated introgression probability (φ) may also serve as a useful guide even though it reflects both the migration rate per generation (m or M) and the time duration of the period of gene flow (eq. 10). Even if gene flow occurs continuously over time (so that the migration model is a more realistic model), the MSci model is effective in extracting historical information about species divergence times and
## Mbe
population sizes. Note that on the evolutionary time scale, a few hundred or thousand generations may count as a fixed time point, in which case the MSci model may provide an adequate approximation. Both the asymptotic theory and simulation have highlighted the semi-unidentifiability or confounding effects between the introgression probability (φ) and the population size of the donor species (θ S in(e.g., [fig_ref] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data... [/fig_ref] , methods c and d). The problem is particularly acute under the IIM model applied to small datasets (with short loci, few sequences per species, or few loci), where high estimates of φ with wide CIs are produced even though migration occurs at very low rates [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref]. One such case has been observed in a recent analysis of genomic data from the erato group of Heliconius butterflies [bib_ref] Full-likelihood genomic analysis clarifies a complex history of species divergence and introgression:..., Thawornwattana [/bib_ref]. The estimated H. sara → H. demeter introgression probability was high with wide CIs for some chromosomal regions with a small number of loci (e.g., chromosome 21 with 4350 noncoding and 3628 coding loci, and an inversion on chromosome 15 with 149 noncoding and 167 coding loci), with the introgression time close to the species divergence time, whereas for the other large chromosomes, the estimates were nearly zero (φ < 0.01). The true rate in this case appeared to be φ ≈ 0, but the limited data from small chromosomal segments led to poorly supported large introgression rates, as in our simulations [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref].
We demonstrated that including multiple samples from the same species (in particular, from recipient species) is important to resolving unidentifiability issues or confounding effects, as well as boosting up the information content concerning the rate of gene flow in the data. In this regard, it may be noted that many approximate methods are designed to use only one sample per species, and it has been claimed that "adding more samples provides little new information with respect to introgression" [bib_ref] Phylogenomic approaches to detecting and characterizing introgression, Hibbins [/bib_ref]. We suggest that this may not be a generally correct statement.
Overall, our simulations using larger species trees with more than two species suggest that misspecification of the mode of gene flow (continuous migration versus episodic hybridization/introgression) has relatively small and localized effects, restricted to divergence times and population sizes around the lineages involved in gene flow, while species divergence times, population sizes for extant species and for ancestral species not involved in gene flow are largely unaffected. If gene flow occurs between species A and B but more distantly related species are included in the data sample, parameters outside the AB clade are largely unaffected (e.g., compare results for the IIM model for two species of [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref] with those for three species of. Similarly, if A represents a clade rather than one species, divergence times and population sizes inside the A clade are not affected by gene flow involving the branch ancestral to the A clade (e.g., compare the D-B setting ofwith the IIM model of [fig_ref] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci... [/fig_ref].
Assigning gene flow to parental or daughter branches causes the introgression probability to be underestimated, and the introgression time to collapse onto the species divergence time. This result may be used to diagnose the mis-assignment of introgression lineages in real data analysis [bib_ref] Significant cross-species gene flow detected in the Tamias quadrivittatus group of North..., Ji [/bib_ref]. A number of authors have discussed the impact of ghost species on detection of betweenspecies gene flow [bib_ref] Effect of unsampled populations on the estimation of population sizes and migration..., Beerli [/bib_ref] [bib_ref] Ghost introgression: spooky gene flow in the distant past, Ottenburghs [/bib_ref]. [bib_ref] Ghost lineages highly influence the interpretation of introgression tests, Tricou [/bib_ref] used simulations to demonstrate that D-statistics can be misled to detect false signals of introgression when the model involved an unsampled (ghost) species. In our simulations, the impact of ghost species on Bayesian estimation of introgression rate and time was minor provided we considered the rate of gene flow in the migration and introgression models to reflect both indirect gene flow via intermediate species and direct gene flow.
## Testing models of gene flow
In this study, we fixed the model of introgression in our analyses, with all introgression events pre-identified, to examine the effects of model misspecification. One may ask what happens if different introgression models (which for example assign introgression events onto different branches of the species tree) are compared using genomic data. Currently, both *BEAST and PHYLONET have implemented cross-model MCMC algorithms under the MSci model, which insert and delete introgression events on the species tree, allowing the Markov chain to move between models. Those algorithms are computationally expensive and currently the two programs can handle only very small datasets (with <100 loci, say). In the BPP program, one may use the Bayes factor to compare two MSci models, using thermodynamic integration [bib_ref] Simulating normalizing constants: from importance sampling to bridge sampling to path sampling, Gelman [/bib_ref] [bib_ref] Computing bayes factors using thermodynamic integration, Lartillot [/bib_ref] combined with Gaussian quadrature to calculate the marginal likelihood values [bib_ref] Efficient Bayesian species tree inference under the multispecies coalescent, Rannala [/bib_ref]. In the case where the compared models are nested (e.g., one with introgression and another without), the Bayes factor may also be calculated through the Savage-Dickey density ratio [bib_ref] The weighted likelihood ratio, linear hypotheses on normal location parameters, Dickey [/bib_ref] , which uses only a within-model MCMC run under the more general model [bib_ref] Significant cross-species gene flow detected in the Tamias quadrivittatus group of North..., Ji [/bib_ref]. This has a computational advantage over reversible jump MCMC [bib_ref] Reversible jump Markov chain monte carlo computation and Bayesian model determination, Green [/bib_ref] , and has recently been applied to formulate and compare introgression models in an analysis of genomic data from the Tamias quadrivittatus group of North American chipmunks [bib_ref] Significant cross-species gene flow detected in the Tamias quadrivittatus group of North..., Ji [/bib_ref]. Calculation of marginal likelihood values or Bayes factors may be feasible if we have only a small number of well-specified models but may not be feasible for searching in the space of MSci models for a given set of species.
Approximate methods have also been developed to infer introgression events or the so-called phylogenetic networks using summaries of the multilocus sequence data. For example, estimated gene tree topologies may be MBE treated as data, as in PHYLONET/GT [bib_ref] Reticulate evolutionary history and extensive introgression in mosquito species revealed by phylogenetic..., Wen [/bib_ref]. Some methods are designed to detect gene flow in a small tree with three or four species, including summary methods based on genome-wide site-pattern counts (such as D and HYDE discussed earlier) or on estimated gene trees (e.g., SNAQ) and maximum likelihood applied to multilocus sequence alignments (e.g., 3S, [bib_ref] Maximum likelihood implementation of an isolation-with-migration model with three species for testing..., Zhu [/bib_ref] [bib_ref] Maximum likelihood implementation of an isolation-with-migration model for three species, Dalquen [/bib_ref]. Results for species subsets may then be combined to formulate an introgression model on the large tree for all species, which is a challenging task [bib_ref] Genomic architecture and introgression shape a butterfly radiation, Edelman [/bib_ref] [bib_ref] Full-likelihood genomic analysis clarifies a complex history of species divergence and introgression:..., Thawornwattana [/bib_ref]. In summary, there is currently an acute need for improving the computational efficiency of Bayesian MCMC algorithms for inference under the MSC model with gene flow and the statistical efficiency of approximate methods.
It will also be interesting to use the same genomic data to compare the MSC-M and MSci models. The two classes of models often predict very different distributions of gene trees and coalescent times (e.g., supplementary figs. S1, S3, S5, Supplementary Material online; see also . Thus, genomic data may be informative to distinguish them. A stochastic search in the combined space of MSC-M and MSci models may be infeasible, as the two types of models are very different. However, they can be compared using Bayes factors.
# Materials and methods
## Simulation to establish a correspondence between the migration and introgression models in the case of two species
We analyzed the relationships between parameters when data are generated under the continuous migration model (IM, IIM, and SC;and analyzed under the episodic introgression (MSci) model. Our theory assumed an infinite number of loci (L = ∞), a finite number of sites per sequence (n), with only one sequence per species per locus. We conducted computer simulations to augment the theoretical analysis. Data of multilocus sequence alignments were simulated under the IM, IIM, and SC models of figure 1a-c, and analyzed under the MSci model. Population sizes on the species treewere θ 0 = 0.002 for the thin branches and θ 1 = 0.01 for the thick branches. Migration occurred from species A to B after their divergence at τ R = θ 0 in the IM model, between τ R = 2θ 0 and τ T = θ 0 in the IIM model, and between τ T = θ 0 and thein [bib_ref] A Bayesian implementation of the multispecies coalescent model with introgression for phylogenomic..., Flouri [/bib_ref] assumes that τ X > τ Y and τ T > τ Y and can represent scenario (a ′ ) in which species X split into two species (A and U), and species XUV contributed migrants into species TYB at time τ Y but has since become extinct. This model was used to simulate data, with θ 0 = 0.002 for the thin branches and θ 1 = 0.01 for the thick branches, τ R = 4θ 0 , τ S = 3θ 0 , τ T = 2θ 0 , τ X = 1.5θ 0 , and τ Y = θ 0 . The introgression probability is φ = 0.2. The number of sequences is S = 4, and the sequence length is n = 500. (b) MSci model Bin [bib_ref] A Bayesian implementation of the multispecies coalescent model with introgression for phylogenomic..., Flouri [/bib_ref] used to analyze the data, which incorrectly assumes τ X = τ Y . (c) The 95% HPD CIs for parameters, with θs and τs multiplied by 10 3 and black solid line indicating the true value.
## Mbe
continuous in the true model but the MSci model assumes episodic introgression at a particular time point, so that the mode of gene flow is misspecified. In settings C-B and D-A, the mode of gene flow was similarly misspecified but we had in addition mis-assignment of gene flow to wrong branches on the species tree. Other parameter settings were the same as above. With two trees, three numbers of loci (L), a total of 600 datasets were generated, each analyzed twice (under models A and B).
## Isolation with initial migration (iim) model
Data were simulated under the IIM model A of, with A → B migration over the time period (τ T , τ S ), and analyzed under the MSci model of, assuming introgression at time τ X = τ Y . The IIM model was specified using a ghost species (U) from which no sequences were available. We generated 100 replicate datasets, each of L = 250, 1,000, or 4,000 loci, with a total of 300 datasets simulated. MCMC settings were the same as above.
## Ghost species
To assess the effects of unsampled ghost population, we simulated data under MSci model A ′ (seein [bib_ref] A Bayesian implementation of the multispecies coalescent model with introgression for phylogenomic..., Flouri [/bib_ref] of figure 7a ′ and analyzed them under the MSci model B of figure 7b, with τ X = τ Y incorrectly assumed. Here introgression involved a ghost species XUV which went extinct or was otherwise unsampled in the data. This scenario is equivalent to model A of figure 7a. With the three values for L (250, 1,000, 4,000), 300 datasets were generated, all analyzed under the MSci model.
We also used the IIM model of to generate data, with migration from species RC to SU and from SU to TB, and with V and W to be unsampled ghost species. Data (i.e., sequences from A, B and C) were analyzed under the MSci model of figure 8b. We used three values for L (250, 1,000, 4,000) and 100 replicates, with 300 datasets simulated in total. Other settings were the same as above.
# Supplementary material
Supplementary data are available at Molecular Biology and Evolution online.
[fig] FIG. 1: (a-c) Three MSC-M models for two species A and B used to generate data: IM (isolation with migration), IIM (isolation with initial migration), and SC (secondary contact). The IIM model is an instance of the MSC-M model with a ghost species at node T and with migration from species A to T (b). Similarly, the SC model (c) is a case of the MSC-M model with τ T > 0. Note that τ T is the time when migration stopped in the IIM model and the time when migration started in the SC model. In the numerical calculations and in the simulations, we assumed the population size θ 0 = 0.002 for the thin branches and θ 1 = 0.01 for the thick branches, and the migration rate was M AB = 0.2 migrants from A to B per generation. Note that in our setup, the time period of gene flow is Δτ = θ 0 in all three models.(d ) The introgression (MSci) model used to analyze the data. (e) A schematic summary of the estimate of the introgression probability (φ) in the MSci model (d ) when the data are generated under the MSC-M models of a-c. The sudden drop in φ as M AB increases coincides with an underestimation of τ R and overestimation of θ R . [/fig]
[fig] FIG. 2: Best-fitting parameter values under the MSci model of figure 1d when data of two sequences per locus (one per species), each of n sites, are generated under the IM model offigure 1a. Five methods (a-e) are used to fit the MSci model, estimating 2, 3, 4, 5, and 4 parameters, respectively, whereas the other parameters are fixed. In (a), τ R and φ are estimated, but θ R and θ S are fixed at their true values in the IM model, and the introgression time τ S = τ H is fixed at τ T = 0. In (b), τ S is estimated as well. In (c), τ S = 0 is fixed, whereas the other four parameters are estimated. In (d ), all five parameters are estimated. In (e), the constraint θ R = θ S is enforced so that four free parameters are estimated. The dotted lines for φ indicate the true total amount of introgression of equation(10). The dashed lines indicate φ * of equation(11). The true and best-fitting distributions of the coalescent time (t) are shown in supplementary figure S1, Supplementary Material online.MBEAsymptotic Results under the IIM Model When data are generated under the IIM model (fig. 1b)and analyzed under the MSci model (fig. 1d), the results (supplementary figs. S2 and S3, Supplementary Material online) [/fig]
[fig] FIG. 3: Average posterior means and 95% HPD CIs for parameters in the MSci model of figure 1d over 30 replicate datasets simulated under the migration (IM, IIM, and SC) models of figure 1a-c, plotted against the number of sites per sequence (n), the number of sequences per species (S), the number of loci (L), and the migration rate (M AB ). Parameters in the migration model are given in the legend to figure 1. In the standard setting, each dataset consists of L = 4,000 loci, with S = 4 sequences per species at each locus and n = 1,000 sites per sequence, and the migration rate was M AB = 0.2 individuals per generation. In the four sets of simulations, one of the factors (n, S, L, M) varies whereas the others are fixed. When S = 1, population sizes θ A , θ B , and θ H are unidentifiable. Estimates of τ and θ parameters are multiplied by 10 3 . Dotted lines indicate true values of identifiable parameters, except in the plot of φ against M AB , where it represents φ 0 of equation(10), (which is identical for the IM, IIM, and SC models offig. 1). Note that the n, S, L, and M AB axes are all on the logarithmic scale. [/fig]
[fig] FIG. 4: (a,b) Two introgression (MSci) models and (c,d) two migration (IM) models used in simulation. The thin branches have the population size θ 0 = 0.002 and the thick branches have θ 1 = 0.01. In MSci model A, the species divergence/ introgression times are [/fig]
[fig] FIG. 5: Posterior means of speciation/introgression times (×10 −3 ) when the introgression event is assigned to a wrong branch. In (a) tree A-tree B, data were simulated using species tree A (fig. 4a), with introgression from species A to B, but are analyzed assuming tree B, with introgression assigned incorrectly to the parental branch ST (so that τ X > τ T ). In (b) tree B-tree A, data were simulated under tree B (fig. 4b) and analyzed assuming tree A, with introgression assigned to the daughter branch B (with τ X < τ T ). For each number of loci (L = 250, 1,000, 4,000), 100 replicate datasets were generated and analyzed. These correspond to the A-B and B-A settings of figure 4e, where estimates of other parameters are shown. [/fig]
[fig] L 4000, Figure 4, Figure 4 C: = 250 L = 1000 L = 4000 L = 250 L = 1000 L = -A 1.12 (0.83, 1.40) 1.11 (0.97, 1.25) 1.11 (1.04, 1.18) 0.12 (0.09, 0.15) 0.12 (0.10, 0.13) 0.12 (0.11, 0.12) Figure 4 D-A 1.69 (1.18, 2.07) 1.80 (1.58, 1.97) 1.86 (1.76, 1.94) 0.02 (0.01, 0.04) 0.02 (0.01, 0.03) 0.02 (0.02, 0.03) Figure 4 A-B 3.82 (3.53, 4.11) 3.75 (3.61, 3.90) 3.73 (3.66, 3.80) 0.18 (0.09, 0.28) 0.13 (0.11, 0.16) 0.12 (0.11, 0.14) Figure 4 B-B 2.98 (2.61, 3.35) 2.99 (2.80, 3.18) 3.00 (2.91, 3.10) 0.23 (0.14, 0.34) 0.20 (0.17, 0.24) 0.20 (0.18, 0.22) Figure 4 C-B 2.98 (2.72, 3.24) 2.93 (2.80, 3.06) 2.91 (2.85, 2.98) 0.11 (0.08, 0.14) 0.10 (0.09, 0.12) 0.10 (0.10, 0.11) Figure 4 D-B 2.83 (2.28, 3.38) 2.71 (2.42, 3.00) 2.73 (2.59, 2.87) 0.11 (0.04, 0.20) 0.08 (0.05, 0.10) 0.08 (0.07, 0.09) Figure 6 IIM 3.40 (2.38, 4.36) 2.93 (2.42, 3.43) 2.83 (2.58, 3.08) 0.24 (0.04, 0.53) 0.10 (0.05, 0.16) 0.08 (0.06, 0.10) [/fig]
[fig] FIG. 6: (a) An isolation-with-initial-migration (IIM) model used to simulate data. The parameter values used are θ 0 = 0.002 for population sizes for the thin branches and θ 1 = 0.01 for the thick branches, τ R = 4θ 0 , τ S = 3θ 0 , τ T = θ 0 for species divergence times. The number of sequences is S = 4, with the sequence length n = 500. The migration rate is M = 0.1. (b) The MSci model used to analyze the data. (c) The 95% HPD CIs for parameters, with black lines indicating the true values. Estimates of θ and τ are multiplied by 10 3 . [/fig]
[fig] FIG. 7: (a) MSci model A (fig. [/fig]
[table] Table 1: Average [/table]
|
Cost-utility analysis of an implant treatment in dentistry
Background: When dealing with the replacement of one missing tooth, the patient has the option of choosing between different types of treatment interventions. Several important factors play a role in his decision-making process, including his limited financial means and his efforts to solve the problem of missing teeth as effectively as possible. The main goal of the study is the economic-clinical evaluation of implant treatment, as a surgical-prosthetic method in dentistry, in case of replacement of one missing tooth of the molar area.Methods:Cost-utility analysis from the patient's perspective is used for evaluation. The selected comparator is a purely prosthetic solution with the help of a three-unit fixed dental prosthesis. Cost-utility analysis is modelled using Markov models, which consider a 30-year time horizon.Results: Based on the results of modelling, the intervention evaluated by the patient, i.e. treatment with the help of implant-supported single crown, brings exactly 15.31 quality-adjusted prosthesis years (QAPY) after 30 years. The value of incremental cost-utility ratio amounted to USD − 1434.Conclusion:The results of the cost-utility analysis suggest that implant treatment with an implant-supported single crown is more cost-effective than treatment with the three-unit fixed dental prosthesis.
# Background
The current modernization and innovation of treatment procedures is leading to a constant increase in health care costs. Increasing quality of provided care goes hand in hand with this trend. The development of technologies and new methods of treatment does not avoid the field of dentistry either. Implantology is one of the examples of a dynamically developing field of dentistry. However, patients are generally very often faced with the decision on choosing the type of treatment, and the financial side of things play an important role in their decision, sometimes the most important. Other factors that influence their final decisions include aesthetics, durability of the treatment, the time required for treatment and also the willingness to undergo surgery.
The dental implants themselves are intended to replace a missing tooth or an entire section of teeth. Various treatment strategies are possible to use in dealing with the loss of one tooth of the molar region. Treatment with a three-unit fixed dental prosthesis (FDP), is considered a conventional treatment, i.e. standard. In the Czech Republic, and also in the world, however, the most common type of dental implants is an implant-supported single crown (ISC).
The three-unit fixed dental prosthesis is a prosthetic replacement that is able to bridge the gap after one or more lost teeth. It is therefore a partial replacement of teeth. The standard FDP has one major drawback, and that is the need for grinding, deterioration and irreversible damage to two teeth adjacent to the gap (so-called abutment teeth). Vogelin his literature review highlights this fact as a major benefit of the use of the implant. The prosthetic solution of a missing tooth in the form of the FDP is today considered inadequate by many experts due to the irreversible deterioration of abutment teeth. In this context, we often talk about harm to the patient. Also, higher incidence of complications is referred in connection with the FDP.
The use of implant treatment in the case of single tooth replacement is still a controversial topic in terms of costs and benefits. There is no cost study in the Czech Republic that would accurately assess the costs and benefits of this treatment. Most published foreign studiescomparing implant treatment and FDP as part of the solution to the replacement of one missing tooth agree that implant treatment entails higher initial costs. However, success, longevity and patient satisfaction are considered more favourable. Foreign studiesfocused on economic-clinical evaluation of dental implants very often use models with long time horizon which proves that after a longer period of time, implant treatment becomes more cost-effective compared to FDP.
Studiesalso agree that implant therapy requires more visits, a longer duration of the initial phase of treatment, while the total time spent in a chair is comparable in most studies. The aim of the study is the economicclinical evaluation of implant treatment in dentistry in comparison with FDP. Based on the aim of the study we set null hypothesis that implantology therapy is more cost-effective compared to standard conventional therapy in the case of treatment of one tooth in the molar area of the teeth.
# Methods
The CUA is used to fulfil the goal with the help of Markov models. The input data of the model were consulted with members of an expert team consisting of four dentists operating in Prague. A necessary prerequisite for the selection of dentists for an expert team was their practical experience in the field of implantology and prosthetics at least 5 years. Another requirement was a complete knowledge of the clinical issues and their operation in Prague. All of them are working in a different private practice. The expert team also included a standard patient (male, 50 years old), who was treated by another private dentist and was chosen randomly in the moment of research. The reason for including the patient in the expert group was primarily the introduction of the voice of the patient perspective. He had to be presented with a decision on the selection of a treatment variant for the replacement of one missing molar area tooth as part of a dentist's treatment. All members of the expert group were male between the ages 30 and 50 (see in. No experimental research was performed within the study, an expert team was created for the purpose of consulting the input data of the model. All addressed experts and one patient were familiar with the purpose of the research and agreed to participate in expert interviews.
# Cost-utility analysis parameters
The selected comparator is a three-unit fixed dental prosthesis. The preferred perspective is the health care payer's perspective. However, in the case of an implant solution for the replacement of one tooth, this is not the most advantageous, as the participation of health insurance companies in payments for this treatment is minimal. The patient's perspective was chosen in this study. The main reason is the role of the patient in the dental care system. The patient is the one who bears the highest part of the costs in the implant-surgical process.
Both treatment strategies offer three treatment variants, which differ in their complexity, length and cost. The choice of treatment variant is always made by the dentist after evaluation of the clinical condition of the tissues and after agreement with the patient. Variant A of ISC treatment involves inserting an implant without the need for bone augmentation and providing the implant with a crown, provided there is open healing. Variant B of ISC is a solution where closed healing is assumed. Variant C of ISC requires moderate bone augmentation prior to implant placement. Variant A of FDP represents tooth The target population in the presented model is represented by 50 years old adult patients, who primarily address the need to replace one missing tooth in the molar area of the tooth. Patients who have their own teeth adjacent to the gap are considered. Based on the nature of the problem, the methodology used in published foreign studies of cost-effectiveness and cost-utility analysis, and the value of life expectancy in the Czech Republic, the maximum time horizon of the model of 30 years was chosen. From the data of the Czech Statistical Office (CSO)for 2018, it is clear that the average life expectancy of individuals living in Prague is 78.3 years (men) and 83 years (women).
The costs and utilities in the CUA model are discounted at a corresponding discount rate beyond one year. A discount rate of 3% is considered, which is considered standard in the Czech Republic.
## The structure of the markov model
As Briggs and Sculpherdescribe the Markov models as decision-analytical models that divide disease (process) into distinct states (see in. Transition probabilities describe the transition between the states over a discrete time period (cycle). Transition probabilities for our model are inand each line represent transition probability from one state to another and the source of the information. We can assign estimates of costs and health outcomes to each the states and then calculate long-term cost and outcomes or CUA for particular healthcare intervention. An adapted Markov model with one-year cycles was designed to calculate the CUA. The licensed TreeAge Prosoftware, specifically its Healthcare module, was used to create the model itself. Markov's model essentially consists of two basic decision trees, each of which has its initial state as a solution to the replacement of one missing tooth, which is one of two alternatives for the treatment of a given health problem-the state "ISC" and "FDP". In the case of both treatment strategies, four health states are described in the model. State diagrams of the Markov model for an implant-supported single crown and a three-unit fixed dental prosthesis are in.
## Probabilities used in the model
Probabilities of mutual transitions are described between the individual health states of the model. Missing tooth (MT) refers to a medical condition associated with a gap after a missing tooth that is not filled with a replacement. The double implant-supported bridge (DIB) is the state that is selected when a second FDP fails. DIB represents three-unit implant-supported bridge. Its incorporation into the model was inspired by a Swiss studyand consulted with members of the expert group.
The mentioned Markov models assume that the value of the survival rate of individual dental prostheses, as health conditions of the model, decreases with increasing number of years of action in the oral cavity. For this purpose, cumulative Kaplan-Meier survival curves of individual replacements were extracted from available foreign clinical studies, from which the probabilities of survival of replacements for individual years of the considered time horizon were derived. The most suitable probability distributions describing the survival rate were found for the relevant reconstructed Kaplan-Meier survival curves. Based on the log-likelihood ratio, the Weibull probability distribution was chosen for all curves. The interpolation of the probability of survival of ISC, FDP and DIB treatments is shown in the following figure.
## Cost data collection
The costs are based on financial treatment plans and price lists of the 13 contacted private dental offices, where 4 of them represent workplaces of members of the expert team. The costs of the initial phase of treatment, both direct and indirect, as well as the costs of possible complications and the maintenance of the replacement, which represent annual costs, were considered. According to Bassi et al., the indirect costs of the initial phase of treatment, which relate to the time spent by the physician and can be referred to as lost patient gain, are calculate according to formula 1. . The unemployment rate must also be included in the calculation of the so-called lost profit. As of January 2020, the CSO states this at 3.1%.
The patient's transport costs to the doctor (formula 2) are processed based on the average number of visits. Data from thirteen Prague dental clinics are included in the cost analyses, therefore the amount of CZK 32 (USD 1.53) was used as a cost unit. The amount corresponds to the rate of the basic fare of Prague Public Transit Company and allows an adult to travel around Prague for 90 min.
## Collection of utility data
For the economic-clinical evaluation of ISC and FDP, it was necessary to obtain clinical results regarding both interventions. For this purpose, utility values corresponding to the health conditions presented in the model were extracted from the available studies. As well as costs, the utilities in the model were discounted at a 3% discount rate.
# Sensitivity analysis
A deterministic one-way sensitivity analysis was performed with the presentation of the results using graphs, tables and a tornado diagram. Selected factors that were gradually varied include initial (input) costs, utility values and the length of the time horizon (or the number of model cycles). When considering 3 cost variants (A, B, C), the standard deviation for FDP is up to 25% of the mean value, and therefore the initial (input) costs and utility values were varied by ± 30%. In the case of the length of the time horizon, these were changes in values in the interval from 5 to 55 years after five years. Subsequently, an analysis of scenarios was presented, in which there were changes in discounting. It was a scenario that does not consider the discount rate (0% discount rate) and a scenario with a 5% discount rate.
# Results
## Costs of health states
Each health state of the model was assigned the appropriate monetary units that the patient spends after finding himself in such a health state.lists the entry costs that the patient will pay after entering the relevant health state, as well as the annual costs, which represent the amount paid by the patient each year for maintenance of the reimbursement and any complications with reimbursement if the patient remains in the relevant health state. According to study, the cost of DIB reimbursement is calculated as one third of the cost of FDP (considered variant A) together with twice the cost of ISC (considered variant B). The cost of entering the MT health state represents the average amount that the patient pays at the clinics for the explanation or extraction of both abutment teeth.
In order to clarify and simplify the input costs of ISC and FDP health conditions, their mean value was used in the model. This value was calculated as the arithmetic mean of all three values corresponding to the respective treatment variants. In the case of ISC treatment, the mean value was set at CZK 40,286 (USD 1,923.24) and in the case of FDP treatment at CZK 40,161 .
## Utility values of health states
The utility values associated with the ISC, FDP and MT health states in the model were drawn from a Canadian study, published in 2007 at BMC Oral Health. The health utility of DIB is based on a studypublished in the United Kingdom. The resulting utility values are shown in.
## Outcomes of markov decision model
The considered cohort in the model in individual cycles, depending on the transition probabilities and probabilities of survival of the given replacement, gradually spreads to individual health states in the entire Markov tree. The proportions of the cohort appearing in both Markov tree models within the 30-year time horizon are shown in the following charts .
The distribution of the cohort is connected also with the division of costs and utility. The cumulative costs of the Markov model ISC correspond to CZK 134,514 (USD 6,421.64) after 30 years. The cumulative costs of the and the cumulative utilities chart is presented in . The evaluated modelled survival curves of individual treatment strategies over the entire considered time horizon are as follows .
## Results of modelling cost-utility analysis
Cost-utility analysis modelling has shown that treatment with ISC appears to be the dominant intervention after 30 cycles of the model. In contrast, treatment with FDP in the model appears to be an absolutely dominated intervention. The final results are presented inlower right quadrant and is therefore defined as clearly cost-effective.
# Sensitivity analysis
Sensitivity analysis showed that the relationship of dominance of interventions by most changes was not disturbed. Changes in the perception of the dominance of interventions occur only if the values of utilities vary in the interval of ± 30%. The situation is presented in the graphs of
# Discussion
The study focuses on the replacement of one lost tooth in the molar area, which in terms of the frequency of occurrence of missing teeth is among the most common. The study by Mack et al.proves that the first molar is the most frequently missing tooth in the population between 20 and 74 years of age. Most published foreign cost-effectiveness studiesused FDP for comparison with implant treatment. Endodontic treatmentor treatment with a removable prosthesiswas also often chosen as a comparator. Partial removable prosthesis was not considered in this study as it appears to be the solution with the lowest quality-adjusted tooth years (QATY) based on Oral Health-related Quality of Life (OHRQoL) measurements. Indirect costs were included in the total cost of the initial treatment phase. In contrast, the study by Bouchard et al.took into account only direct costs. The impact of the inclusion of indirect costs and direct non-medical costs on the total costs was assessed through a sensitivity analysis. The factor that mainly affects the amount of average direct medical costs for the initial phase of treatment is the cooperation of the workplace with the health insurance company. The results of this study show that the differences in costs based on the factor of cooperation with the health insurance company are more pronounced under the assumption of the choice of FDP. We can therefore say that if a patient chooses FDP and a clinic that does not have a contract with health insurance company, he will pay on average CZK 9,403.22 (USD 448.89) more than at a clinic that cooperate with the patient's health insurance company. However, it is important to mention that this result is based on an assumption that does not take into account the selected treatment variant, it is a simple arithmetic average for a given treatment alternative. This difference is only CZK 2,128.77 (USD 101.62) for ISC. It is clear from the results that the costs of the initial phase of implant treatment are not significantly affected by the factor of the clinic's cooperation with the health insurance company. The result confirms the presumption of non-participation of the health insurance company in reimbursement of implant treatment and is in accordance with Alison's study. Bouchard et al.states that even in France, the participation of health insurance company in the payment for ISC is negligible, in contrast to the payment for FDP. The situation is similar in Japan where implant treatment is reimbursed only if the patient loses or damages his jaw due to illness or accident.
A limitation of this study is the use of the QAPY parameter instead of the QALY parameter. However, in the field of dentistry, the presentation of benefits through QALY or Life Year Gained is relatively difficult.
To strengthen the level of evidence, it would be appropriate to obtain data from more than 13 Prague clinics and at best scenario to compile a set of clinics from non-Prague workplaces where the cost of dental care may be lower. Furthermore, it would be appropriate to consider a treatment that does not involve the replacement of only one tooth, but several teeth, or, for example, a dental material other than ceramics (although ceramics are now considered a material with the required level of quality and aesthetics).
The objectives of the study were based on the assumption that treatment of single tooth loss with an ISC is generally more expensive than treatment with FDP. This primary assumption led to the use of a method of modelling using Markov models which should identify a value representing the equality of the cumulative costs of both treatments, the so-called turning point. However, this primary assumption was not confirmed during the data analysis. It seems that in the CUA model thisThe tornado diagram shows the relationship of intervention dominance by changes in inputs in the interval of ± 30% (CZK 1 = USD 0.048 as of 17 th May 2021). Initial costs of three-unit fixed dental prosthesis (c_FDP_initital), utility of double implant-supported bridge (u_DIB), initial costs of implant-supported single crown (c_ISC_initial), initial costs of double implant-supported bridge (c_DIB_initial), utility of implant-supported single crown (u_ISC), utility of three-unit fixed dental prosthesis (u_FDP), incremental cost-utility ratio could reflect the use of the mean value of the input costs of both treatments. However, additional analysis revealed that even under the assumption of using precisely calculated input costs of individual treatment variants, there is ultimately no change in the CUA result. The reason of non-fulfilment of the primary assumption of work (higher initial costs for implant treatment) can also be the consideration of ceramics as a material for the production of FDP and also ISC. However, the advantages of ceramics have already been discussed above. Another reason may be the need for endodontic treatment of abutment teeth which, according to information from the panel of experts, is very often a condition for the deployment of a three-unit fixed dental prosthesis. In joint interviews with the doctors of the expert team, it was even found that many doctors consider this treatment to be automatic before using FDP. However, it must be said that this is a treatment which, based on the data obtained from dental clinics in Prague, will cost an average of CZK 15,731.31 (USD 750,97) with a standard deviation of CZK 4,412.32 (USD 210,63) (endodontics of both abutment teeth is considered).
The structure of the model takes into account not only the input costs of treatment, but also the annual costs that the patient will pay for solving possible complications and for inspections and preventive visits to the dentist and dental hygienist. The probabilities of the occurrence of these events were obtained from a systematic search in literature. The total costs associated with FDP in the model are generally higher than the cost of implant treatment. The same conclusion was reached by a study, which evaluates ISC even with inclusion of indirect costs (due to the higher number of visits per implant treatment) as cheaper treatment. Another studyeven provides evidence of higher total costs of initial FDP compared to ISC because of laboratory costs. Chun et al.demonstrate that although the direct input costs of ISC are initially higher compared to FDP, after a 10-year time horizon, ISC becomes less expensive in terms of direct treatment costs.
Within the model, it was necessary to solve the issue of data transferability. Data on transition probabilities were obtained from individual studies found in systematic search; more accurate data could be obtained from a metanalysis, but this type of study is not suitable for fitting the Kaplan-Meier survival curve. Data on utilities of individual health conditions were transferred to the model. The source of utilities for the condition of implant treatment, fixed bridge and the condition of the missing tooth was a Canadian study. The primary requirement in selecting the study was its complexity to provide utility values for all health conditions considered in the model, it's appropriate year of publication (not older than 15 years), and finally the appropriate age of the target population (50 years). A Canadian studymeets all primary requirements except for the condition of gaining values for all health conditions considered in the model. No other such study was found. In order to obtain data on the utility of DIB treatment, which was not addressed in a Canadian study, another studyhad to be used. This was a study directly from the United Kingdom, which considered the same target population and publication was not older than 15 years. To express utility value for DIB state and to compare it with the utility values from the Canadian study, this studywas chosen for several other reasons. It is a studysimilar to the Canadian study: including comparable options in different treatment options; study that uses one of the recommended valuation methods (Visual Analogue Scale) to express the utility; study with a sufficiently large number of patients.
As the time horizon of this clinical-economic model was longer than one year, the costs and benefits had to be discounted at the recommended discount rate, which according to Czech Pharmacoeconomic Societyis 3%. The recommendation of the State Institute for Drug Controlis the use of 0% and 5% discount rates in the sensitivity analysis.
Modelling of CUA confirmed that after a 30-year time horizon for the use of a prosthesis in the oral cavity, FDP appears to be not clearly cost-effective. The result of CUA modelling coincides with the results of several foreign studies, which also used modelling using Markov models with a longer time frame for more accurate and flexible results. The incremental cost-utility ratio in this study is -30 072 CZK (USD -1,435.56). The evaluated intervention thus became clearly cost-effective. A French studycame to the same conclusion and the authors argue that the unambiguity of the result should lead to the first place of this treatment in all European countries. However, it should be noted that only a 20-year time horizon was considered in the study. Another study, which considers a 10-year time horizon, has the opposite results and ISC is dominated intervention. However, sensitivity analysis states that for ISC to become the dominant intervention, it is sufficient that the cost of ISC treatment is reduced by 20%.
The model is transferable between other countries, it is possible to make specific adjustments typical for a given country and to use collected inputs. In contrast to the clinical outcomes and the probabilities used, it is not possible to consider the cost side of the study. Although proportionally we can assume the same cost results and thus the same results for more developed countries with a similar health care system as in the Czech Republic.
# Conclusion
Clinical-economic evaluations provide physicians as well as patients with important information on the cost-effectiveness of medical technologies. However, such evaluations are still rather rare in dentistry. In addition, there is a significant need for use of appropriate methodology for a correct estimate of cost-effectiveness. This work brings new knowledge about the cost-effectiveness of ISC in the conditions of the Czech healthcare system. Although there is speculation in dentistry that ISC is more expensive than FDP, and this fact was the original premise of this study, the results speak differently. Based on the modelling, it was found that over the entire time horizon of 30 years, ISC shows significantly lower cumulative costs than FDP treatment and brings a higher effect. In addition, the conventional solution of replacing one missing tooth with FDP exposes abutment teeth to many biological and technical problems. ISC is therefore a suitable first-choice solution for adult patients who has one missing tooth in the molar area and are willing to undergo surgery. For them, the treatment becomes cost-effective. |
The burden of oral cancer in China, 1990–2017: an analysis for the Global Burden of Disease, Injuries, and Risk Factors Study 2017
Background: Oral cancer is among the most common malignant tumors worldwide, and it has become an increasingly important public health problem in China. This study systematically assesses the current state of oral cancer in China from 1990 to 2017, providing new information and perspectives for oral health researchers and public health policy makers.Methods: Based on the Global Burden of Disease, Injuries, and Risk Factors Study 2017 (GBD 2017), we evaluated the incidence rates, mortality and disability-adjusted life year (DALY) rates for oral cancer in China and their changing trends between 1990 and 2017, making comparisons by gender and age. We also assessed the DALY rates associated with oral cancer at the provincial level for 33 provinces and their trends over time.Results: From 1990 to 2017, the number of new cases and the age-standardized incidence rate for oral cancer in China increased by 280.0% and 79.7%, respectively; the number of deaths and the age-standardized mortality rose by 196.8% and 29.0%, respectively; and the number of DALYs and the age-standardized DALY rate increased by 149.1% and 21.0%, respectively. The incidence rates for oral cancer rose after 30 years of age and peaked at 65-69 years; the mortality for oral cancer rose after 50 years of age and peaked at 65-69 years; and the DALY rates for oral cancer rose after 45 years of age and peaked at 65-69 years. The incidence rates, mortality and DALY rates for oral cancer in males were significantly higher than those in females and showed an upward trend, while there was a decrease or no significant change in females. The DALY rates increased in 21 provinces and decreased in 12 provinces, with the largest growth in Henan Province and the largest decline in Hong Kong Province.Conclusions:The burden of oral cancer in China continues to increase continuously. More prevention, control and intervention measures should be taken and increased attention paid to common risk factors is essential for the prevention of oral cancer.
floor of the mouth, lip and alveolar mucosa in non-Asian regions. The site of oral cancer depends on the main risk factors for the particular geographical location. The major frequently found pathological classification of oral cancer is oral squamous cell carcinoma, which accounts for up to 90% of all oral malignant tumours. Despite improvements in early diagnosis and early treatment techniques, the 5-year survival rate is still very low for oral cancer, ranging from 40 to 50%.
Global Cancer Statistics 2018, compiled by the International Agency for Research on Cancer (IARC), provides a comprehensive report on the current status of cancer around the world. This recent report estimates that oral cancer accounted for almost 2.0% of all cancer cases and 1.9% of all cancer deaths globally.
According to GLOBOCAN 2018 estimates, the incidence rates for oral cancer varied across different geographic areas. The highest incidence rates for oral cancer were in Melanesia, followed by South Central Asia, with age-standardized incidence rates of more than 12 per 100,000. In contrast, the lowest incidence rates for oral cancer were in Central America and West Africa, which were less than 3 per 100,000 population.
Oral cancer tended to be concentrated in certain highrisk areas. It is very prevalent in South Asia (especially India and Sri Lanka) and the Pacific Islands, with Papua New Guinea having high incidence rates worldwide in both males and females. Oral cancer was also a major cause of death among males in India and Sri Lanka. It is worth noting that the incidence and mortality of oral cancer are more serious in developing countries, as they are usually two to three times higher than those in developed territories and countries. Because of the heavy socioeconomic burden, oral cancer should receive greater attention in developing countries.
The cancer statistics for the incidence and mortality in China usually combine oral cancer and oropharyngeal cancer. The Cancer Registry Report, released by the National Central Cancer Registry of China (NCCR), summarizes the national cancer registration information. The latest report published in 2020 estimates that in 2015, there were about 4.29 million new cancer cases and 2.81 million cancer deaths in China, while there were about 48,100 new cases and 22,100 deaths from oral and oropharyngeal cancer, accounting for 1.12% and 0.79% of all new cancer cases and deaths, respectively. Although the absolute numbers are not high, the burden of oral and oropharyngeal cancer has been on the rise in China in recent years.
The main risk factors for oral cancer in China are smoking and alcohol use. In particular, it is important to note that betel quid chewing is also an important risk factor in China. In 2018, the number of smokers in China was 308 million, 50.5% of whom were male and only 2.1% of whom were female. The incidence of oral cancer in Chinese males is also much higher than that in females. The drinking rate among adults in China is 30.5%, including 53.8% for men and 12.2% for women. In China, betel quid chewing habits are mainly distributed in Hunan, Hainan, and Yunnan provinces, especially Hunan Province. It has been reported that betel quid chewing is common in Hunan Province, and the proportion of betel nut chewers is as high as 64.5-82.7%.
At present, comprehensive data on the burden of oral cancer are lacking at the national and provincial levels. Therefore, an analysis of the change trend in oral cancer in China over the last three decades fills a significant gap and can help to inform oral health researchers and public health policy makers in China to focus more attention on oral cancer. In this study, we showed the incidence rates, mortality, and DALY rates for oral cancer in China from 1990 to 2017.
# Methods
## Data sources
The data for this study were obtained from the 2017 Global Burden of Disease, Injuries and Risk Factor study. The GBD 2017 includes cause-specific mortality from 282 causes of deaths in 195 regions and countries around the world between 1990 and 2017, as well as health losses from 359 diseases and injuries associated with DALYs and 84 risk factors, addressing a total of 23 age groups and two genders.
The provincial administrative units of this study include twenty-two provinces, five autonomous regions, four municipalities directly under the Central Government in the Chinese Mainland, and the Hong Kong and Macao Special Administrative regions. The above 33 locations are collectively termed provinces throughout the whole article.
These data mainly come from surveys and surveillance. The primary data sources for Chinese mainland included the China Disease Surveillance Point (DSP) System, the China Cancer Registry data, and the vital registration gathered by the Chinese Center for Disease Control and Prevention (CDC), supplement by hospital inpatient data, published literature, and some national-level surveys. The National Census and the Maternal and Child Surveillance System were also important complements to the data sources. In addition, the cause of death data for Hong Kong and Macao come from medical data records certified by the World Health Organization (WHO). This study is not included in the data of Taiwan Province because it is not available.
## Eligibility criteria
The aetiological code used in GBD 2017 was based on the International Classification of Disease (ICD) formulated by the WHO, which is the current standard in the world and the most exhaustive cause list. The definition of oral cancer in GBD 2017 consisted of codes C00-C08 (lip and oral cavity cancer) in ICD10. Codes C11 (nasopharyngeal cancer), C09-C10 and C12-C13 (cancer of other parts of the pharynx), and C14 (malignant neoplasm of other and ill-defined sites in the lip, oral cavity, and pharynx) were excluded by GBD 2017. Specific coding by oral cancer type can be found in full detail in the GBD 2017 literature.
## Incidence, mortality and dalys
Details of the methods used in GBD 2017 have been fully described elsewhere. In this study, the disease burden of oral cancer is described by incidence, mortality and DALYs in China. To comprehensively measure and compare the heath loss from fatal and non-fatal disease burden, a summary indicator is created called DALY to estimate trends in population health over the past 27 years, and it is composed of years lived with disability (YLDs) and years of life lost (YLLs). YLDs is an indicator measuring the sequelae of the disease, and it is computed by multiplying the prevalence by the disability weight for the health associated with the sequela of the disease. YLLs is an indicator of health loss used to assess premature death, and it is the result of multiplying the number of deaths from various causes and the standard life expectancy per year.
# Possible methods
The GBD Cause of Death Ensemble modelling tool, CODEm, developed by GBD, was used to estimate causespecific mortality for each location, year, age, and gender. The estimates of nonfatal disease burden, such as incidence, were commonly used with the Bayesian metaregression tool DisMod-MR 2.1. The age-standardized population of GBD 2017 was calculated using the age of the world population as the standard. Our research followed the principles in the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) to obtain transparent as well as replicable data and results. All data and results can be obtained from the Global Health Data exchange (GHDx) on the GBD website.A comprehensive list of core data was provided by the GHDx source tools (http://ghdx.healt hdata .org/data-type/disea se-regis try), such as data on deaths, YLLs, YLDs, DALYs, prevalence, incidence, causes, risks, cause-risk attribution, aetiologies, and impairment.
## Statistical analyses
Analyses were done using Python versions 2.7.12 and 2.7.3, Stata version 13.1, and R version 3.2.2. The incidence of cancer was estimated directly from cancer mortality using the mortality-to-incidence ratios (MIRs). MIR estimation and estimation of cancer mortality have been explained in detail elsewhere. The incidence and mortality data from the cancer registrations were matched according to cancer, age, sex, year and location to produce MIRs. MIRs were estimated using a fixed effect logistic regression model, with covariates for sex, categorical age, and the Healthcare Access and Quality index.
We generated 1000 draws from each step of the evaluation process. Then, 95% uncertainty intervals (UIs) were computed using the 2.5th and 97.5th percentile of the ordered 1000 estimate values, and point estimates were calculated by the mean values across the 1000 draws. Unlike confidence intervals, which derived uncertainty only from sampling errors, UIs acquire uncertainty from multiple modelling steps, model estimation and model specification.
# Results
## Overall trends
From 1990 to 2017, the total number of cases of oral cancer in China increased by 289.2% from 49.3 thousand (95% UI: 47.0 thousand-51.7 thousand) to 192.1 thousand (95% UI: 180.9 thousand-201.8 thousand), and the number of new cases of oral cancer in China increased by 280.0%, from 12.4 thousand (95% UI: 11.8 thousand-13.0 thousand) to 47.2 thousand (95% UI: 44.5 thousand-49.7 thousand). It was estimated that the age-standardized incidence rate per 100,000 rose by 79.7% from 1.36 (95% UI: 1.29-1.43) in 1990 to 2.44 (95% UI: 2.30-2.57) in 2017. Compared with 1990 values, the total number of deaths from oral cancer in 2017 increased by 196.8%, from 6.9 thousand (95% UI: 6.6 thousand-7.4 thousand) to 20.7 thousand (95% UI: 19.6 thousand-21.8 thousand), respectively, and the age-standardized mortality per 100,000 rose by 29.0% from 0.84 (95% UI: 0.80-0.88) in 1990 to 1.09 (95% UI: 1.03-1.14) in 2017 (
## Age-specific trends
We estimated the incidence rates, mortality and DALY rates per 100,000 population for all ages. Between 1990 and 2017, the incidence rates for oral cancer declined among those aged 29 years or younger and rose for those aged 30 years or older, peaking among those aged 65-69 years . During the same period, the mortality for oral cancer declined among those aged 49 years or younger and rose for those aged 50 years or older, peaking among those aged 65-69 years . During the same period, the DALY rates for oral cancer declined for those aged 44 years and younger and rose among those aged 45 years and older, peaking among those aged 65-69 years .
## Gender-specific trends
In 1990, the age-standardized incidence rate, mortality and DALY rate per 100,000 for males were 1.58, 1.03 and 25.2, respectively, while those for females were 1.17, 0.67 and . The values in males were higher than those in females in 1990. In 2017, the age-standardized incidence rate, mortality and DALY rate per 100,000 in males were 3.79, 1.73 and 40.2, respectively, and those in females were 1.19, 0.49 and 11.0. The values were higher for males than for females in 2017 .
Comparing the changes between 1990 and 2017, it was found that the age-standardized incidence rate, mortality and DALY rate in males increased by 140.0%, 67.8% and Changes of incidence, mortality, and DALY rates for oral cancer by age in China, 1990-2017. a Changes of incidence rates from 1990 to 2017 by age.b Changes of mortality from 1990 to 2017 by age.c Changes of DALY rates from 1990 to 2017 by age 59.7%, respectively, while for females, the incidence rate increased by only 1.1%, and the mortality and DALY rate decreased by 27.5% and 35.0%, respectively.
## Location-specific trends of dalys
We estimated the age-standardized DALY rates and their changes for 33 provinces.shows the provincial age-standardized DALY rates per 100,000 and the changing trend in DALY rates between 1990 and 2017. In 1990, the age-standardized DALY rates of oral cancer were lowest in Shaanxi Province (15.2 per 100,000) and highest in Hong Kong Province (56.2 per 100,000). In 2017, the age-standardized DALY rates for oral cancer varied from a low of 13.0 per 100,000 in Ningxia Province to a high of 40.8 per 100,000 in Hunan Province. Comparing the changes in the age-standardized DALY rate of oral cancer between 1990 and 2017, we can see that the age-standardized DALY rates decreased in 12 provinces, the largest decrease of which was 60.4% in Hong Kong Province, and increased in the other 21 provinces, the largest increase of which was 68.4% in Henan Province.
# Discussion
This study comprehensively and systematically assessed the burden of oral cancer in China between 1990 and 2017. The age-standardized DALY rate for oral cancer in China in 2017 was 25.4 per 100,000, corresponding to an increase of 21.0% over the 1990 rate. Over the same period, the age-standardized DALY rates for oral cancer worldwide were 60.6 and 64.2 per 100,000, respectively, representing an increase of 0.6%. A Sub-Saharan African region research showed that the age-standardized DALY rates for oral cancer in this region were 47.8 and 45.0 per 100,000 between 1990 and 2017, respectively, with a decrease of 4.6%. Although the agestandardized DALY rates for oral cancer in China were lower than those in the Sub-Saharan African region and below the worldwide average, the changing trend in the DALY rates for oral cancer in China between 1990 and 2017 was significantly higher than the global level. Thus, the DALY rate for oral cancer in China was characterized by rapid growth.
There are significant differences in the oral cancer data in terms of gender and age. The incidence of oral cancer increases with age. Regarding sex, the incidence rate among males is much higher than that among females. The differences in oral cancer by age and sex are consistent with findings from previous studies. The reason for these differences may be that oral cancer is the result of long-term stimulation of certain risk factors, especially smoking, alcohol use and betel quid chewing, and the proportion of men exposed to these risk factors is much higher than that of women.
The age-standardized DALY rates of oral cancer varied across provinces in China, with the greatest difference being a multiple of three between the highest and lowest provinces. The different distribution of risk factors in different provinces may be part of the reason for the variation in the burden of oral cancer among provinces. High DALY rates are particularly observed in provinces with cultural practices of tobacco use, alcohol use and betel quid chewing. Surveys showed that the smoking prevalence in Yunnan, Hunan, and Guizhou provinces was relatively high compared with the national average, and there was much more economic reliance on the production of tobacco in these compared with other provinces. Another report showed that Sichuan is the province with the highest levels of alcohol consumption. In addition, Hunan Province is recognized as one of the provinces in Chinese mainland where betel quid chewing is most prevalent. This result for China is not surprising. Over the past three decades, Chinese society has experienced tremendous and rapid development. The population ageing and changes in lifestyle have had a significant impact on public health. With the popularity of vaccines and the improvement of living conditions in China, the burden of infectious diseases has declined significantly, but the burden of some cancers, including oral cancer, is steadily increasing.
Smoking is the most crucial risk factor for oral cancer. The study by Yao et al. compared smokers with nonsmokers and showed that Chinese smokers have a prominently increased risk of developing oral cancer, and the risk of oral cancer due to smoking could increase significantly with the increase in years or frequency of smoking. The prevalence of oral cancer in daily smokers was three times higher than that in non-smokers. According to the 2015 tobacco report, the tobacco epidemic in China continues to be serious. Smokers in China accounted for 27.7% of the total number of adults, including 52.1% of men and 2.7% of women. These different smoking habits may be the main reason for the gender and provincial disparity in the incidence of oral cancer.
Alcohol use is also an important risk factor for oral cancer. Epidemiological studies have shown that drinking alcohol increases the risk of oral cancer by three-fold, which is an independent risk factor, with risk increasing with increased alcohol consumption. Compared with smoking alone, smoking combined drinking increased the risk of oral cancer by about eight-fold. The 2014 statistics revealed that approximately 40.9% of oral and pharyngeal cancers worldwide were related to alcohol consumption. Between 1990 and 2017, alcohol use increased significantly in some developing countries, such as Vietnam, India and China, reaching even higher levels than those found in some European regions. The 2016 report showed that the per capita consumption of alcohol in China increased by 76% from the 1990 level, reaching 4.1 L in 2005, 7.1 L in 2010 and 7.2 L in 2016. Reducing alcohol consumption is very important for the prevention and control of oral cancer.
Betel quid chewing is the most common risk factor for oral cancer in China, and betel quid has been recognized as a first-class carcinogen by the IARC. Hundreds of millions of people around the world use different types of betel quid, mainly distributed in South-eastern and Southern Asia, particularly India, Pakistan, and China. Hunan Province is among the regions with the highest use of betel quid in China. The raw betel nuts used in Hunan Province are imported from Hainan and Thailand, but the vast majority of Chinese betel quid products are produced and processed in Hunan Province. Surveillance, clinical services and policies for betel quid and oral cancer should be provided to control and reverse the rapid growth of oral cancer related to betel quid products.
Lifestyle interventions are an effective way to prevent and control oral cancer. To encourage a healthier lifestyle and promote risk reduction of cancers, some targeted policies and regulations have been formulated and promulgated by the Chinese government, including the Healthy China 2030 Plan, the "National Medium to Long Term Plan for Prevention and Treatment of Chronic Disease (2017-2025)", national-level demonstration zones for the comprehensive prevention and control of chronic noncommunicable diseases and the Healthy Lifestyle Action (2019-2025) proposal. These policies can decrease the risk of various cancers, including oral cancers, and have played a major role in some aspects of health promotion. In addition, in view of the risk factor for betel quid chewing, the Chinese government has strengthened the screening, prevention and control of this activity in key areas. In 2019, the central government announced the launch of the country's first clinical screening programme for oral cancer in Hunan Province. The Healthy Lifestyle Action (2019-2025) also proposed strengthening oral education and oral health examinations in regions where betel quid chewing is prevalent.
There are certain limitations of this article that we need to consider. First, this study was based on the GBD study, so the limitations described in detail from the GBD 2017 study are shared by this study. Second, currently, the DSP system in China has been improved by the method of death distribution, which will likely create high deviations in disease death estimation. Third, despite the continuous improvement in cancer registration data, there were still a large number of poorly defined cancer cases that needed to be reassigned to all types of cancers, which may lead to a certain degree of deviation. In addition, revisions and changes in the coding systems over time may introduce artificial differences in disease estimates. Fourth, changes in diagnostic techniques may affect the diagnostic criteria of some diseases. Fifth, few areas in remote and poor counties were covered by disease surveillance systems, which led to a lack of reliable data on cause-specific mortality in some provinces, affecting the accuracy of our estimates. Because of the general lack of cancer registries in rural areas, as these are usually set up in urban areas, the representation of the resulting population may also be problematic. Sixth, in our study, YLLs were independent of YLDs, so the uncertainty in DALY may be undervalued. Seventh, in the GBD study, an accurate MIR was required to estimate the mortality of cancer. We further revised our estimation method of MIR for GBD 2017, but this method was still biased due to missed diagnosis cases or uncertainty deaths. Eighth, this study still lacked detailed explanations for the relationship between the burden of oral cancer and risk factors at the provincial level, thus, more in-depth studies are needed in the future.
# Conclusions
Between 1990 and 2017, the burden of oral cancer in China rapidly increased, accompanied by significant gender and age differences, which made the prevention and control of oral cancer more difficult in China. The disease burden of oral cancer and the trends in its occurrence varied greatly from province to province. These conditions deserve attention. Early screening, lifestyle intervention, oral health care, and regular oral check-ups can improve and reduce the burden of oral cancer. |
Brain Metastases from Lung Cancer Show Increased Expression of DVL1, DVL3 and Beta-Catenin and Down-Regulation of E-Cadherin
The susceptibility of brain to secondary formation from lung cancer primaries is a well-known phenomenon. In contrast, the molecular basis for invasion and metastasis to the brain is largely unknown. In the present study, 31 brain metastases that originated from primary lung carcinomas were analyzed regarding over expression of Dishevelled-1 (DVL1), Dishevelled-3 (DVL3), E-cadherin (CDH1) and beta-catenin (CTNNB1). Protein expressions and localizations were analyzed by immunohistochemistry. Genetic alterations of E-cadherin were tested by polymerase chain reaction (PCR)/loss of heterozygosity (LOH). Heteroduplex was used to investigate mutations in beta-catenin. DVL1 and DVL3 showed over expression in brain metastasis in 87.1% and 90.3% of OPEN ACCESS samples respectively. Nuclear staining was observed in 54.8% of cases for DVL1 and 53.3% for DVL3. The main effector of the Wnt signaling, beta-catenin, was up-regulated in 56%, and transferred to the nucleus in 36% of metastases. When DVL1 and DVL3 were up-regulated the number of cases with nuclear beta-catenin significantly increased (p = 0.0001). Down-regulation of E-cadherin was observed in 80% of samples. Genetic analysis showed 36% of samples with LOH of the CDH1. In comparison to other lung cancer pathologies, the diagnoses adenocarcinoma and small cell lung cancer (SCLC) were significantly associated to CDH1 LOH (p = 0.001). Microsatellite instability was detected in one metastasis from adenocarcinoma. Exon 3 of beta-catenin was not targeted. Altered expression of Dishevelled-1, Dishevelled-3, E-cadherin and beta-catenin were present in brain metastases which indicates that Wnt signaling is important and may contribute to better understanding of genetic profile conditioning lung cancer metastasis to the brain.
# Introduction
Metastasis is a highly selective process conditioned by the genetic profile of the original tumor as well as by organ environment. The susceptibility of the brain to secondary formation from lung cancer primaries is a well-known phenomenon. In contrast, the molecular basis for invasion and metastasis to the brain is largely unknown. Reports indicate that specific set of genes mediate metastasis to the brain according to discrete changes [bib_ref] Identifying site-specific metastasis genes and functions, Gupta [/bib_ref] [bib_ref] Metastasis signatures: Genes regulating tumor-microenvironment interactions predict metastatic behavior, Albini [/bib_ref].
The work presented in this paper focused on the analysis of different expression levels of Dishevelled-1 (DVL1), Dishevelled-3 (DVL3), beta-catenin (CTNNB1) and E-cadherin (CDH1) in brain metastases that originated from primary lung carcinomas. All the molecules studied are key players in the classical Wnt signaling pathway which has today been established as one of the basic cellular pathways and whose misregulation plays an important role in tumorigenesis. Initiation of the Wnt signaling cascade depends on the presence of Wnts, glycoproteins that bind to receptors called frizzled. Thus, activated, the receptors recruit Dishevelleds at cytoplasmic membrane and this interaction also pulls axin from the beta-catenin destruction complex to the plasma membrane. In this fashion the complex is inactivated [bib_ref] A role of Dishevelled in relocating AXIN to the plasma membrane during..., Cliffe [/bib_ref] leading to the stabilization of cytosolic beta-catenin. Stabilized beta-catenin enters the cell nucleus and associates with T-cell factor (TCF)/lymphoid enhancer factor (LEF) transcription factors, which leads to the transcription of Wnt-target genes such as c-myc, matrix metalloproteinase (MMP)7, cyclin D1, etc. [bib_ref] Signaling through β-catenin and Lef/Tcf, Novak [/bib_ref] [bib_ref] TCF: Lady justice casting the final verdict on the outcome of Wnt..., Brantjes [/bib_ref] [bib_ref] Wnt signalling and its impact on development and cancer, Klaus [/bib_ref] [bib_ref] Wnt signaling inside the nucleus, Shitashige [/bib_ref] [bib_ref] Wnt signal transduction pathway and apoptosis: A review, Pecina-Slaus [/bib_ref] [bib_ref] TCF/LEFs and Wnt signaling in the nucleus, Cadigan [/bib_ref] [bib_ref] SOX2 promotes tumor metastasis by stimulating epithelial-to-mesenchymal transition via regulation of Wnt/β-catenin..., Li [/bib_ref]. In the absence of factors that activate Wnt signaling, the beta-catenin destruction complex that consists of adenomatous polyposis coli (APC)-Axin and glycogen synthase kinase 3 (GSK3) binds to beta-catenin with subsequent beta-catenin phosphorylation, ubiquitination, and degradation by proteasomes. In this scenario beta-catenin/TCF/LEF target genes are repressed [bib_ref] Signaling through β-catenin and Lef/Tcf, Novak [/bib_ref] [bib_ref] TCF: Lady justice casting the final verdict on the outcome of Wnt..., Brantjes [/bib_ref] [bib_ref] The β-catenin destruction complex, Stamos [/bib_ref]. However, mutations of the complex APC-Axin and GSK3 can also results in the translocation of beta-catenin to the nucleus ultimately leading to oncogenic transformation and progression.
The cellular function of human Dishevelled genes has not been completely elucidated yet. Nevertheless, Dishevelleds (DSH or DVL) are considered to be the central hub of Wnt signaling and are all multifunctional phosphoproteins that have been shown to shuttle between the cytoplasm and the nucleus [bib_ref] Differential mediation of the wnt canonical pathway by mammalian Dishevelleds-1, Lee [/bib_ref] [bib_ref] Dishevelled: The hub of wnt signaling, Gao [/bib_ref] [bib_ref] Dishevelled 2 signaling promotes self-renewal and tumorigenicity in human gliomas, Pulvirenti [/bib_ref]. The reported data on Dishevelleds' role in tumorigenesis is controversial, although it has been demonstrated that they are over expressed in several types of cancers including lung cancer [bib_ref] Amplification, up-regulation and over expression of DVL1, the human counterpart of the..., Nagahata [/bib_ref] [bib_ref] Dishevelled-1 and Dishevelled-3 affect cell invasion mainly through canonical and noncanonical Wnt..., Zhao [/bib_ref] [bib_ref] The expression of Dishevelled-3 and glutamine metabolism in malignant pleural mesothelioma, Li [/bib_ref] [bib_ref] Activation of the Wnt pathway in non small cell lung cancer: Evidence..., Uematsu [/bib_ref] [bib_ref] Wnt pathway activation in Mesothelioma: Evidence of Dishevelled overexpression and transcriptional activity..., Uematsu [/bib_ref].
Of particular interest is a process very much involved in invasion and metastasis of tumors-the so called Epithelial-to-Mesenchymal Transition (EMT), where noninvasive tumor cells acquire motility and ultimately disseminate to places distant from the primary site. It has been shown that the Wnt signaling has a particularly tight link with EMT. For example, the nuclear translocation of beta-catenin can induce EMT [bib_ref] Wnt signalling and its impact on development and cancer, Klaus [/bib_ref] [bib_ref] Direct evidence for a role of beta-catenin/LEF-1 signaling pathway in induction of..., Kim [/bib_ref] [bib_ref] Correlation of epithelial-mesenchymal transition markers with clinicopathologic parameters in adenocarcinomas and squamous..., Kim [/bib_ref] by activating the transcriptional repressors Snail and Slug that suppress E-cadherin expression [bib_ref] Wnt signalling and its impact on development and cancer, Klaus [/bib_ref] [bib_ref] Differential matrix rigidity response in breast cancer cell lines correlates with the..., Kostic [/bib_ref] [bib_ref] Molecular alterations of E-cadherin and beta-catenin in brain metastases, Zeljko [/bib_ref] [bib_ref] Epithelial-to-mesenchymal transition: Possible role in meningiomas, Pecina-Slaus [/bib_ref]. Another important finding is that lymphoid enhancer factor 1 (LEF1) when over expressed leads to enhanced tumor invasiveness and induces EMT [bib_ref] Metastasis signatures: Genes regulating tumor-microenvironment interactions predict metastatic behavior, Albini [/bib_ref] [bib_ref] Direct evidence for a role of beta-catenin/LEF-1 signaling pathway in induction of..., Kim [/bib_ref] [bib_ref] Wnt pathway component LEF1 mediates tumor cell invasion and is expressed in..., Nguyen [/bib_ref] [bib_ref] LEF-1 and TCF-4 expression correlate inversely with survival in colorectal cancer, Kriegl [/bib_ref].
Besides being the main signaling effector molecule of the pathway, beta-catenin is also involved in the cellular architecture. It is bound to E-cadherin and is an essential component of adherens junctions. In this respect, it is important to remember that the most prominent feature of EMT is the destruction of adherens junctions and the loss of expression of the cell-cell adhesion molecule E-cadherin.
Our hypothesis is that molecular components of the Wnt pathway, Dishevelleds, E-cadherin and beta-catenin, play important roles in metastasis of lung cancer to the brain. We believe that their altered levels of expression contribute to the outline of molecular profile of distant brain metastases.
# Results
DVL1 and DVL3 showed over expression in brain metastasis tissues in 87.1% and 90.3% of samples respectively. The staining was observed mainly in the cytoplasm and was diffusely or granularly distributed. Nevertheless, we also observed nuclear staining of DVL1 in 54.8% of cases and DVL3 in 53.3% of cases.
At the level of the complete brain metastasis sample, our analyses showed that there were 12.9% of samples demonstrating weak expression of DVL1; 45.2% moderate expression and 41.9% strong expression. Considering the expression levels of DVL3 in our total sample we observed 9.7% of samples with weak expression; 51.6% with moderate expression and 38.7% with strong expression. The normal levels of expression in healthy tissues were very weak. The expression levels of the two proteins in metastases, according to the starting points of the specific types of primary lung cancers are shown in [fig_ref] Figure 1: The distribution of the expression levels of the two proteins in metastases... [/fig_ref]. For DVL1 the expression levels were not statistically different between the various pathological diagnoses (χ 2 = 9.375; df = 8; p = 0.312). In contrast to this, the expression levels of DVL3 were statistically significantly different between the same pathological diagnoses (χ 2 = 16.275; df = 8; p = 0.039).
Normal levels of E-cadherin staining were scored as +++, and the protein was located along the cell membrane or inside the cytoplasm. The investigation on E-cadherin showed that overall 80% of metastases had down-regulation of E-cadherin. Intense down-regulation was noticed in 52% of cases. The majority of samples (88.9%) with E-cadherin's gross deletions (LOH) were accompanied with down-regulation of E-cadherin protein. All metastases originated from small cell lung cancer (SCLC), adenocarcinoma and carcinosarcoma showed E-cadherin down-regulation, while in large cell carcinoma and squamous cell carcinomas E-cadherin was down-regulated in 75% and 60% cases respectively. The expression levels of Dishevelleds were not associated to E-cadherin down-regulation (p = 0.856 for DVL1; p = 0.310 for DVL3). Our next step was to detect whether the expression and cellular localization of beta-catenin, correlated with the expression levels of Dishevelleds and E-cadherin. We noticed that beta-catenin was up-regulated in 56%, and transferred to the nucleus in 36% of metastases. It is interesting to note that 40.9% of cases with up-regulated DVL1 expression had beta-catenin transferred to the nucleus. The percentage of metastases with up-regulated DVL3 that also exhibited nuclear staining of beta-catenin was 32%. When DVL1 was up-regulated the number of cases with nuclear beta-catenin significantly increased (McNemar test, p = 0.0001). The same trend was noticed when DVL3 was up-regulated (McNamar, test p = 0.0001).
Approximately 25% of metastases showed simultaneous nuclear localization of Dishevelleds and beta-catenin. Statistical analysis showed a significant correlation between the nuclear localization of beta-catenin and DVL1 (p = 0.05), but no correlation between nuclear DVL3 and beta-catenin (p = 0.261).
Only 4 primary lung cancers autologous to the brain metastases were available for the analysis. Three out of 4 primary lung cancers showed higher E-cadherin's and lower beta-catenin's expression levels than in the autologous brain metastases. All four primaries showed lower DVL1 expression levels when compared to the corresponding metastases, and 2 out of 4 primaries showed lower DVL3 levels.
Detailed histopathological diagnosis, the origin of the lesions of metastasis sample and findings on the protein levels are presented in . Immunostaining of Dishevelleds 1 and 3 is demonstrated in [fig_ref] Figure 2: Brain metastasis samples immunohistochemically stained for the expression of DVL1, DVL3, E-cadherin... [/fig_ref] -D and of beta-catenin protein in [fig_ref] Figure 2: Brain metastasis samples immunohistochemically stained for the expression of DVL1, DVL3, E-cadherin... [/fig_ref]. Loss of E-cadherin's expression is demonstrated in [fig_ref] Figure 2: Brain metastasis samples immunohistochemically stained for the expression of DVL1, DVL3, E-cadherin... [/fig_ref]. The results of our genetic analysis indicate a total of 36% of the sample with LOH of the CDH1 (E-cadherin) gene when assessed with 3 markers. The highest frequency of genetic changes was observed in metastases originating from primary sites with the diagnosis of lung adenocarcinoma and SCLC with 83.3% and 75% of changes, respectively, while squamous cell carcinoma (SCC) exhibited 25%, and large cell carcinoma (LCC) 12.5% of LOHs. In comparison to other lung cancer pathologies, the diagnoses adenocarcinoma and SCLC were significantly associated to E-cadherin genetic changes with χ 2 = 10.364; df = 1; p = 0.001. A type of genomic instability-microsatellite instability (MSI) was detected in one metastasis originated from adenocarcinoma. The results of heteroduplex analysis of exon 3 did not show samples with additional bands which indicated that mutational hot-spot of beta-catenin, was not targeted in our metastasis sample.
# Discussion
Our results on DVL1, DVL3, beta-catenin and E-cadherin bring novel insights on the Wnt signaling role in brain metastases that originated from primary lung carcinomas. We investigated the combination of Wnt key molecules in our collected tumor sample because Wnt signaling is involved in EMT and metastasis formation [bib_ref] Wnt/Tcf signaling through LEF1 and HOXB9 mediates lung adenocarcinoma metastasis, Nguyen [/bib_ref]. The results of investigations of DVL1 and DVL3 protein levels showed over expression in brain metastasis tissues in 87.1% and 90.3% of samples respectively. The involvement of DVL members in lung cancer is still very much unknown, particularly the association between changed expression and the pathohistological type of lung cancer from which the metastasis originated. Moreover, the correlation with tumor prognosis is still not established. Only several papers report the over expression of Dishevelleds in primary lung cancers. Uematsu et al. [bib_ref] Activation of the Wnt pathway in non small cell lung cancer: Evidence..., Uematsu [/bib_ref] report on DVL3 over expression in non-small cell lung cancer. They showed that 75% NSCLC tumors (¾ squamous and ¾ adeno) had DVL3 overexpression. They imply that these events, upstream of beta-catenin, are critical for activation of Wnt signaling. Surprisingly these authors did not detect expression of DVL1 or DVL2 which is different from our findings on the presence of DVL1 over expression.
Li et al. [bib_ref] Transcription expression and clinical significance of Dishevelled-3 mRNA and δ-catenin mRNA in..., Li [/bib_ref] discovered that DVL3 mRNA levels are significantly higher in pleural effusions from patients with adenocarcinoma and offered this finding as a marker for pleural micrometastasis.
The findings that our investigation builds on are brought by Zhao et al. [bib_ref] Dishevelled-1 and Dishevelled-3 affect cell invasion mainly through canonical and noncanonical Wnt..., Zhao [/bib_ref] and Wei et al. [bib_ref] Dishevelled family proteins are expressed in non-small cell lung cancer and function..., Wei [/bib_ref] who found that Dishevelled family proteins are over expressed in primary lung cancers. Zhao et al. [bib_ref] Dishevelled-1 and Dishevelled-3 affect cell invasion mainly through canonical and noncanonical Wnt..., Zhao [/bib_ref] found positive expression of DVL1 and DVL3 in 45.1% and 48% of non small cell lung cancer (NSCLC) (squamous cell carcinoma (SCC) and adenocarcinoma). The expression levels of Dishevelleds were significantly higher in adenocarcinomas than squamous cell carcinomas.
Wei et al. [bib_ref] Dishevelled family proteins are expressed in non-small cell lung cancer and function..., Wei [/bib_ref] showed that the positive expression rate in primary NSCLC tumors was 53.1% for total DVL, 36.3% for DVL1, 36.3% for DVL2 and 41.6% for DVL3. They also demonstrated that the positive expression rates of Dishevelleds were higher in tumor stages III and IV. Nodal metastases included in their study also showed higher expression levels of DVL1 and DVL3 than primary growths. Our findings on very elevated levels of Dishevelleds in brain metastases represent a logical continuation of the findings reported by these authors in primary lung carcinomas. Moreover, Wei et al.'s [bib_ref] Dishevelled family proteins are expressed in non-small cell lung cancer and function..., Wei [/bib_ref] data on the rise of Dishevelleds' expression from primary (53.1%) to nodal metastases (87.2%) is additionally confirmed by our analysis that shows that 85.2% of Dishevelleds' over expression is present when brain metastases are the final destination.
So far it has been reported that DVLs are over expressed in various tumors types. The results from Nagahata et al. [bib_ref] Amplification, up-regulation and over expression of DVL1, the human counterpart of the..., Nagahata [/bib_ref] suggest that amplification and up-regulation of the DVL1 gene are involved in breast carcinogenesis, especially in the acceleration of tumor growth. The involvement of DVLs in invasive ductal carcinoma of the breast was also reported by Prasad et al. [bib_ref] Wnt signaling pathway in invasive ductal carcinoma of the breast: Relationship between..., Prasad [/bib_ref]. Mizutani et al. [bib_ref] Up-regulation and over expression of DVL1, the human counterpart of the Drosophila..., Mizutani [/bib_ref] got similar results in prostate cancer. Their sample of 20 primary prostate cancer showed significant over expression of DVL1 (65%) and correlation to beta-catenin's expression. Pulvirenti et al. [bib_ref] Dishevelled 2 signaling promotes self-renewal and tumorigenicity in human gliomas, Pulvirenti [/bib_ref] showed that DVL2 is over expressed in human high-grade gliomas. Results from Okino et al. [bib_ref] Up-regulation and overproduction of DVL-1, the human counterpart of the Drosophila dishevelled..., Okino [/bib_ref] demonstrate increased expression of DVL1 in over two thirds of primary cervical squamous cell cancers when compared to corresponding non-cancerous uterine squamous cell tissues. All these data collectively indicate that amplification and increased expression of the DVL genes may play important role in the development of a portion of human cancers through derangement of the Wnt signaling pathway.
Although reports indicate DVLs location in the nucleus is only sporadic [bib_ref] Dishevelled family proteins are expressed in non-small cell lung cancer and function..., Wei [/bib_ref] , our study on brain metastases did not show this feature. We found nuclear staining of DVL1 proteins in 54.8% of cases and of DVL3 in 53.3%. It is known that DVLs interact with the wide range of protein partners in both the cytoplasm and the nucleus [bib_ref] Dishevelled: The hub of wnt signaling, Gao [/bib_ref]. Recently collected data [bib_ref] Dishevelled: The hub of wnt signaling, Gao [/bib_ref] [bib_ref] Activation of Wnt/β-catenin/Tcf signaling pathway in human astrocytomas, Sareddy [/bib_ref] [bib_ref] Wnt/Frizzled signaling: Receptor-ligand selectivity with focus on FZD-G protein signaling and its..., Dijksterhuis [/bib_ref] [bib_ref] Wnt signaling pathway in non-small cell lung cancer, Stewart [/bib_ref] suggest that there are two cellular pools of DVLs: one that translocates to the nucleus, while the other remains in the cytoplasm. While it is not clear how it is regulated, when DVL is located in the nucleus it is known to interact with beta-catenin [bib_ref] Dishevelled: The hub of wnt signaling, Gao [/bib_ref] [bib_ref] Nuclear Dvl, c-Jun, β-catenin, and Tcf form a complex leading to stabilization..., Gan [/bib_ref] [bib_ref] Dishevelled and Wnt signaling: Is the nucleus the final frontier?, Habas [/bib_ref] [bib_ref] Dvl3 translocates IPMK to the cell membrane in response to Wnt, Wang [/bib_ref]. Our data support these findings by demonstrating that 25% of metastases simultaneously exhibit nuclear localization of Dishevelleds and beta-catenin.
The majority of metastases had down-regulation of E-cadherin expression with intense down-regulation in 52% of cases. Similar results were reported by McDonald et al. [bib_ref] Elevated phospho-S6 expression is associated with metastasis in adenocarcinoma of the lung, Mcdonald [/bib_ref] who also detected lower E-cadherin expression in metastatic lesions. Additionally, Saad et al. [bib_ref] Immunohistochemical markers associated with brain metastases in patients with nonsmall cell lung..., Saad [/bib_ref] demonstrated that loss of E-cadherin in patients with adenocarcinoma and squamous cell carcinoma (SCC) of the lung is significantly associated with increased risk of developing brain metastases. The results of other authors investigating E-cadherin in brain metastasis [bib_ref] Expression of p53, bcl-2, E-cadherin, matrix metalloproteinase-9, and tissue inhibitor of metalloproteinases-1..., Arnold [/bib_ref] [bib_ref] Epithelial-to-mesenchymal transition in the development and progression of adenocarcinoma and squamous cell..., Prudkin [/bib_ref] [bib_ref] Immunohistochemical expression of E-cadherin in metastatic brain tumors, Shabani [/bib_ref] collectively demonstrate that E-cadherin is constantly expressed in metastatic deposits. E-cadherin was expressed in the majority of our sample too, but contrary to the above studies, we observed different levels of its expression. E-cadherin's expression in primary lung cancers is down-regulated in invasive component of 62% adenocarcinoma [bib_ref] Expression of transforming growth factor β1 and E-cadherin proteins in pulmonary adenocarcinoma:..., Kim [/bib_ref] , and in another study in 72% of SCC and adenocarcinoma [bib_ref] Impact of p120-catenin isoforms 1A and 3A on epithelial mesenchymal transition of..., Zhang [/bib_ref]. In 2 investigated cases of carcinosarcoma 70% of cells showed E-cadherin's down-regulation.
Recent studies [bib_ref] SOX2 promotes tumor metastasis by stimulating epithelial-to-mesenchymal transition via regulation of Wnt/β-catenin..., Li [/bib_ref] also show that DVLs are involved in tumor metastasis. Results from Shi et al. [bib_ref] Regulator of G-protein signaling isoform 1 (PDZ-RGS3) enhances canonical Wnt signaling and..., Shi [/bib_ref] indicate that specific protein domains shared by Dishevelleds and axin 1 (PDZ-RGS3 domains) enhance signals generated by Wnt canonical pathway which plays important role in EMT and tumor migration.
Nuclear location of beta-catenin is an indicator of its acquisition of oncogenic activity. In our study, beta-catenin was up-regulated in 56%, and transferred to the nucleus in 36% of metastases. It is interesting to note that when DVL1 was up-regulated the number of cases with nuclear beta-catenin significantly increased (p = 0.0001). Additionally, when DVL3 was up-regulated, there was a significant number of cases with beta-catenin nuclear staining (p = 0.0001). A study by Guldur et al. [bib_ref] Comparison of osteopontin, beta-catenin and hnRNP B1 expression in lung carcinomas, Guldur [/bib_ref] showed 66.7% of LCC samples, 53.4% of SCC and 42% of adenocarcinoma with positive beta-catenin staining, while all of their SCLC samples showed reduced beta-catenin staining. A study by Rodriguez-Salas et al. [bib_ref] Beta-catenin expression pattern in small cell lung cancer: Correlation with clinical and..., Rodriguez-Salas [/bib_ref] showed cytoplasmic hyperexpression in 28% of SCLC and no nuclear beta-catenin staining. Only two samples were reported for carcinosarcoma staining of beta-cateninand 70% of cells expressed beta-catenin with the strong intensity.
In heteroduplex analysis, the conformational properties of the double stranded molecules are used to distinguish different base pairing (i.e., mutations). Annealing of mutant DNA to wild type probe gives duplexes with one or more mismatched bases (heteroduplexes). Mismatching causes the double helix to take on a conformation which retards its mobility during electrophoresis [bib_ref] Genetic changes of CDH1, APC and CTNNB1 found in human brain tumors, Nikuseva Martic [/bib_ref]. The results of our heteroduplex analysis did not demonstrate mutations of beta-catenin suggesting that mutations in exon 3 of the CTNNB1 gene are not associated to brain metastasis process. This finding is not unusual since many investigations [bib_ref] Wnt signaling in lung cancer, Mazieres [/bib_ref] [bib_ref] Regulation of AXIN2 expression at the levels of transcription, translation and protein..., Hughes [/bib_ref] [bib_ref] Epigenetic silencing of AXIN2/βTrCP and deregulation of p53-mediated control lead to wild-type..., Tseng [/bib_ref] collectively demonstrated that mutation of APC, AXIN2 or beta-catenin are rarely seen in lung cancer [bib_ref] Epigenetic silencing of AXIN2/βTrCP and deregulation of p53-mediated control lead to wild-type..., Tseng [/bib_ref] [bib_ref] Loss of heterozygosity at 5q21 in non-small cell lung cancer: A frequent..., Cooper [/bib_ref] or are mutually exclusive, which is supported by our results. Uematsu et al. [bib_ref] Wnt pathway activation in Mesothelioma: Evidence of Dishevelled overexpression and transcriptional activity..., Uematsu [/bib_ref] also failed to detect mutations in exon 3 of beta-catenin in mesothelioma. Therefore nuclear beta-catenin accumulation may be a result of mechanisms other than mutation.
Since there are no known mutations of beta-catenin in lung cancer it seems that alterations of the upstream effectors of beta-catenin's degradation are responsible for the Wnt activation. The over expression of Dishevelleds act upstream of the destruction complex stealing axin from it and thus disrupting it or preventing its proper formation. This event enables cytoplasmic accumulation of beta-catenin and its translocation to the nucleus which is also observed in our study. The underlying molecular mechanisms suggested by our study are that Dishevelleds over expression enhance transcriptional activity of beta-catenin.
We are aware that our study was carried out on a relatively small number of brain metastases and, therefore, all conclusions made on this data set need to be tempered with caution. It is, however, important to note that despite the small number of cases analyzed, we recorded significant associations between DVL3 expression levels between the metastases originating from various primary tumors, upregulation of DVL1 and DVL3 and nuclear beta-catenin staining and also association of CDH1 genetic changes and diagnoses adenocarcinoma and SCLC. What needs to be taken into account is the fact that the overall number of investigated cases reported in medical literature so far still remains shockingly small-after a comprehensive search of the available literature we found just 270 cases of lung cancer investigated for DVL1 and DVL3, and of these there was no data on potential brain metastasis.
Although we are aware that it would be very interesting to draw comparisons between the primary lung cancer and the metastatic deposits, unfortunately it was not possible to attain the samples from the primary lung cancer except in 4 cases. Nevertheless, many authors performed analyses on primary lung cancer and report on the expression levels. Obviously, what needs to be done in the future is to analyze primary tumors and metastases in the same patient.
Our previous investigations on brain metastases showed the involvement of yet another important component of Wnt signaling, the APC gene [bib_ref] Brain metastases exhibit gross deletions of the APC gene, Pecina-Slaus [/bib_ref]. We found that 55% of metastases from lung carcinoma had LOH of this tumor suppressor gene. Three adenocarcinomas, 1 large cell carcinoma, 1 small cell carcinoma, and 1 squamous cell carcinoma harbored gross deletions of APC gene. Although LOH of the APC gene is a common finding in lung cancer, point mutations of this tumor suppressor have not been frequent [bib_ref] Loss of heterozygosity at 5q21 in non-small cell lung cancer: A frequent..., Cooper [/bib_ref]. It is therefore interesting to report our finding of somatic mutation in exon 15 of the APC gene detected in one case of brain metastase derived from SCLC by direct DNA sequencing of the metastasis and autologous lymphocyte samples. The substitution was at position 5883 G-A in the metastasis tissue [bib_ref] Report on mutation in exon 15 of the APC gene in a..., Pecina-Slaus [/bib_ref].
The biologic spectrum of metastases is unfortunately wide, heterogenic and difficult to predict, resulting in poor prognosis. Their proliferative activity and invasion are important characteristics that should be considered in diagnostics and prognosis. Why is the metastasis process from the lung so fast, particularly in comparison with the breast and the other primary locations? The velocity of metastasis establishment differs depending on the primary site and may therefore differ according to the genes involved. However, relevant information on the genes that are crucial in the metastatic process has not yet been utilized in diagnostics, prognostics, or therapeutic purposes. Novel findings point out the importance of chemokines and their receptors in brain metastasis formation [bib_ref] Pollo, B. CXCL12, CXCR4 and CXCR7 expression in brain metastases, Salmaggi [/bib_ref]. Lung cancer is one of the most common malignancies and is the leading cause of cancer death in industrialized countries. Novel therapies aimed at inhibition of metastasis process based on better understanding of the biology of brain metastases will offer improvement in diagnosis and treatment of this disease.
## Experimental section
## Tumor specimen
Samples of 31 brain metastases from lung cancer, together with autologous blood samples were collected from the Department of Pathology and Department of Neurosurgery University Hospital "Sisters of Mercy", Zagreb, Croatia. Using the magnetic resonance imaging (MRI) metastasis lesions were found in different cerebral regions, most frequently in parietal and frontal regions and cerebellum, with the surrounding zone of perifocal oedema. During the operative procedure the metastasis was maximally reduced using a microneurosurgical technique. The patients had no family history of brain tumors and did not receive radio-or chemotherapy before metastasis excision. All metastases were studied by pathologists and classified according to the WHO criteria [bib_ref] WHO Classification of Tumours of the Central Nervous System, Louis [/bib_ref]. There were 8 large cell carcinoma (LCC), 7 small cell lung cancer (SCLC), 9 adenocarcinoma (AC), 5 squamous cell carcinoma (SCC) and 2 carcinosarcoma (CS) cases. The metastasis tissues for DNA extraction were frozen in liquid nitrogen and transported to the laboratory, where they were immediately transferred at −75 C. The peripheral blood samples were collected in EDTA and processed immediately.
Twenty-two patients were male, and 9 were female. The age of patients varied from 46 to 81 (mean age 60.48, median 60 years). The mean age at diagnosis for males was 59.95, and for females 61.4 years.
The local Ethical Committee approved our study, grant 1.2.1.19., and patients gave their informed consent.
## Immunohistochemistry
Immunohistochemistry was performed in order to establish the levels of expression and cellular localization of Dishevelled-1, Dishevelled-3, beta-catenin and E-cadherin proteins. The samples were fixed in formalin, embedded in paraffin, sliced into 4-μm thick sections, and then fixed onto capillary gap microscope slides (DakoCytomation, Glostrup, Denmark). Sections were immunostained using streptavidin horseradish peroxidase/DAB (3,3'-diaminobenzidine) (Dako REAL™, EnVision™, Glostrup, Denmark). Briefly, sections were deparaffinized and rehydrated and then microwaved twice for 3 min at 700 W in citrate buffer and once for 4 min at 350 W to unmask epitopes. To block endogenous peroxidase activity, cells were fixed in methanol with 3% H 2 O 2 . Non-specific binding was blocked by incubating samples with goat serum for 30 min at 4 °C. Next, the primary antibodies were applied for 30 min at room temperature. The antibodies used for Dishevelled protein detection were: polyclonal rabbit anti-human DVL1 (diluted 1:50), Abcam, Cambridge, UK and monoclonal mouse anti-human DVL3 (diluted 1:50), Santa Cruz Biotehnology, Dallas, TX, USA. For E-cadherin protein detection monoclonal mouse anti-human E-cadherin NCH-38 (diluted 1:100), and for beta-catenin monoclonal mouse anti-human antibody (diluted 1:200), both Dako Corporation, Carpinteria, CA, USA, were used.
Slides were then washed three times in phosphate-buffered saline (PBS)/goat serum, and secondary LINK antibody was applied for 1 h at 4 °C. Slides were again washed three times in PBS/goat serum and were incubated with substrate chromogen solution (EnVision™, Dako REAL™) for 30 s.
Negative controls underwent the same staining procedure but without incubating samples with the primary antibodies. The frontal cortex of a normal brain, normal human placenta and normal bronchial epithelia served as positive controls. Antibody labeling was analyzed by three independent and blinded observers using an Olympus BH-2 microscope. No expression or very weak expression was labeled as 0/+, moderate expression as ++, and strong expression as +++. Two hundred cells of each sample were analyzed.
## Genetic analyses of e-cadherin and beta-catenin
For the genomic DNA extraction approximately 0.5 g of metastasis tissue was homogenized with 1 mL extraction buffer (10 mM Tris-HCl, pH 8.0; 0.1 M EDTA, pH 8.0; 0.5% sodium dodecyl sulfate) and incubated with proteinase K (100 µg/mL; Sigma-Aldrich, St. Louis, MO, USA; overnight at 37 C). Phenol chloroform extraction and ethanol precipitation followed. Blood was used to extract leukocyte DNA. Five milliliter of blood was lysed with 7 mL distilled water and centrifuged (15 min/5000× g). The pellet was then processed as for DNA extraction from the tissue samples.
## Polymerase chain reaction (pcr)
PCR amplification of microsatellite markers for the CDH1 (E-cadherin) gene was used to test loss of heterozygosity (LOH) in brain metastases. Allelic loss at the CDH1 locus was assessed using 3 highly polymorphic microsatellite markers-D16S265, D16S398 and D16S752, which map to 16q21-22.1 [bib_ref] Functional characterization of E-and P-cadherin in invasive breast cancer cells, Sarrio [/bib_ref]. D16S265 and D16S398 are CA dinucleotide repeat polymorphisms and the D16S752 polymorphic region is a tetranucleotide GATA polymorphism (GATA51G03). All markers were amplified in a total volume of [bib_ref] Wnt pathway component LEF1 mediates tumor cell invasion and is expressed in..., Nguyen [/bib_ref] Primers for D16S752: 5'-AATTGACGGTATATCTATCTGTCTG-3'; and 5'-GATTGGAGGAG GGTGATTCT-3'; for D16S265: 5'-CCAGACATGGCAGTCTCTA-3' and 5'-AGTCCTCTGTGCAC TTTGT-3'; and for D16S398: 5'-CTTGCTCTTTCTAAACTCCA-3' and 5'-GAAACCAAGTGGGT TAGGGTC-3'.
PCR products were analyzed on 2% agarose gels, length of the D16S752 repeat was 102-126 bp, length of the D16S265 repeat was 89-117 bp, and length of the D16S398 repeat was 180-196 bp.
## Loss of heterozygosity, microsatellite instability (msi)
Absence or significant decrease in the intensity of one polymorphic allele in metastasis compared to the heterozygous autologous blood sample was considered as LOH of CDH1 gene. The samples were electrophoresed on Spreadex gels EL 300 in the SEA 2000 submarine electrophoresis apparatus (Elchrom Scientific, Cham, Switzerland) at 120 V. Temperature of the running buffer was kept constant at 55 °C. The samples were stained with SyberGold (Molecular Probes, Leiden, The Netherlands). All the PCR experiments were repeated twice and the LOHs were confirmed. The samples with LOHs were additionally electrophoresed on 15% polyacrylamide gels stained with silver.
MSI is a type of genomic instability indicating impaired cellular mismatch repair. Samples demonstrating MSI have bands on different positions in comparison to bands of autologous blood tissue due to a defect in the replication/repair machinery in tumor cells.
# Heteroduplex analysis
Exon 3 of the CTNNB1 (beta-catenin) gene was screened for mutations. Heteroduplexes were formed by heating 3 µL of PCR products (tumor mixed with normal DNA) at 95 °C for 3 min, followed by incubation on ice for 20 min. About 3 µL of each sample was mixed with 7 µL of mixture of formamide and 10 mM NaOH (1:100) prior to loading to a gel. The electrophoresis was performed on the GMA (Gene Mutation Analysis) gels in the SEA 2000 submarine electrophoresis apparatus (Elchrom scientific, Cham, Switzerland). The temperature of the running buffer was kept constant at 9 °C.
# Statistical analysis
All individuals were analyzed for the following features: pathohistological diagnosis (PHD) status, sex, age, CDH1 genetic changes, E-cadherin, Dishevelled 1, 3 and beta-catenin protein expression levels and localizations. Differences in the frequencies of the analyzed features were tested with the Pearson χ 2 test employing Yates correction when appropriate and McNamar's test. All statistical evaluations were performed using the SPSS statistical package, version 14.0 (SPSS Inc., Chicago, IL, USA).
# Conclusions
Our findings suggest that altered expressions of Dishevelled-1, Dishevelled-3, E-cadherin and beta-catenin were present in brain metastases from lung cancer which indicates that Wnt signaling may be essential for the progression of lung cancer.
[fig] Figure 1: The distribution of the expression levels of the two proteins in metastases according to the starting point of specific type of the primary lung cancer. (A) Dishevelled-1 (DVL1); and (B) Dishevelled-3 (DVL3). [/fig]
[fig] Figure 2: Brain metastasis samples immunohistochemically stained for the expression of DVL1, DVL3, E-cadherin and beta-catenin proteins. Strong expression of DVL1, (A) in cytoplasm; (B) in cytoplasm and nucleus; Strong expression of DVL3, (C) in cytoplasm; (D) in cytoplasm and nucleus; Nuclear expression of beta-catenin (E); and Loss of expression of E-cadherin protein (F). Original magnification, 400×; scale bar, 50 μm. [/fig]
[table] Table 1 Table 1: The localization of the metastases, the polymorphic status for microsatellite markers used and genetic changes of the CDH1 gene, expression levels of E-cadherin, beta-catenin, DVL1 and DVL3 proteins, and pathohistological diagnosis of the primary site. Cont. LOH = loss of heterozygosity; MSI = microsatellite instability; HETERO = heterozygous samples; HOMO = homozygous samples; ND = not determined; 0 = no expression; + = weak expression; ++ = moderate expression; +++ = strong expression; C = cytoplasmic; N = nuclear localization; All = all three MS markers D16S752&D16S265&D16S398 showing LOH; SCLC = small cell lung cancer; PR = denotes expression in the autologous primary lung cancer. [/table]
|
Sacral and Rectal Necrosis after Carbon Ion Radiotherapy Reconstructed with Transpelvic Rectus Abdominis Flap
[bib_ref] Working Group for Bone and Soft Tissue Sarcomas. Carbon ion radiation therapy..., Imai [/bib_ref] [bib_ref] Rectotumoral fistula formation occurring more than 5 years after carbon ion radiotherapy..., Ukon [/bib_ref]
## Case presentation
The patient was a 75-year-old woman who had received CIRT (6720 cGy in 16 fractions) for sacral chordoma 3 years earlier. Two years and 6 months after the radiotherapy, she visited Gunma University Hospital because of sacral pain, tissue necrosis, and rectum-to-sacrum fistula. She received a proctostomy and underwent palliative debridement of necrotic tissue by the gastroenterologic and orthopedic surgeons in our hospital. After the operation, the sacral wound was left open, and the patient consulted plastic surgery for reconstruction of the wound.
At the first visit, we observed a sacral skin ulcer. Central necrosis of subcutaneous tissue and bone had led to a fullthickness defect, and the rectum was exposed through the defect . Computed tomography showed the sacral bone destruction and cystic changes, and a colonoscopy revealed that the rectal mucosa was necrotic . The patient did not complain of urination disorder or any other myelopathy.
Daisuke Atomura, MD* † Takaya Makiguchi, MD, PhD* Hideharu Nakamura, MD* Yukie Yamatsu, MD* Katsuya Osone, MD ‡ Yoko Motegi, MD ‡ Hiroomi Ogawa, MD ‡ Ken Shirabe, MD, PhD ‡ Satoshi Yokoo, DMD, DMSc* Summary: Carbon ion radiotherapy (CIRT) has been used for malignant tumors that are difficult to excise surgically, such as sacral chordoma, and the success of its outcomes is attributable to the high dose concentration and biological effects. CIRT has produced successful clinical outcomes, and it is considered to have fewer adverse effects on surrounding normal tissues; moreover, complications have been rarely reported. We describe a 75-year-old woman with a full-thickness sacral defect, who had received CIRT for sacral chordoma 3 years earlier. Computed tomography showed sacral bone destruction, and a colonoscopy revealed rectal necrosis. Rectectomy in addition to sacral bone resection was necessary, which resulted in a huge sacral defect of slightly anxious viability. We performed reconstruction of the sacral defect by using pedicled vertical rectus abdominis myocutaneous (VRAM) flap, obliterating sacral defects and intrapelvic dead space that occurred after rectectomy. Six months after surgery, the wound had healed well, and no complication was observed. Sacral complications after CIRT may affect surrounding normal tissues such as the rectum, and it would be difficult to reconstruct the resulting complications. The vertical rectus abdominis myocutaneous flap is considered useful for the simultaneous obliteration of sacral defects and intrapelvic dead space after CIRT. Orthopedic and gastroenterologic surgeons performed radical debridement and sacrectomy with the patient in the prone position until viable tissue was exposed. Then, after repositioning in the supine position, rectectomy with Miles procedure was performed. Subsequently, a VRAM flap was created and elevated and then transferred to the sacral region through the intrapelvic space, and the donor site was closed in a layered manner. The insertion of the rectus abdominis was preserved to avoid excess tension on the pedicle. After the patient was moved into the prone position, the flap was carefully brought out from intrapelvic space and set into the sacral defect .
Postoperatively, the wound healed completely, and computed tomography revealed that the intrapelvic dead space and sacral region were filled with the rectus abdominis flap . No complication of sacral or donor site was observed at the postoperative 6-month follow-up visit.
# Discussion
Sacral full-thickness tissue defects with rectum necrosis are rarely reported as complications of CIRT; we demonstrated that it could be successfully reconstructed with a transpelvic VRAM flap. Sacral full-thickness tissue necrosis containing rectum after CIRT is extremely rare and dangerous; only 2 cases have been reported, and both patients died. [bib_ref] Rectotumoral fistula formation occurring more than 5 years after carbon ion radiotherapy..., Ukon [/bib_ref] To the best of our knowledge, our case is the first in which reconstruction for sacral complications of CIRT was successful.
CIRT has been used as a treatment for unresectable malignant tumors, such as sacral chordoma, and the success of its outcomes is attributable to the high dose concentration and biological effects. Imai et al 1 reported that the 5-year rate of local control after CIRT for chordoma was 77.2%, whereas those rates after surgery have ranged from 35% to 50%. The mechanism of CIRT is the release of low radiation dose along the travel path, except for the maximum energy at the end of the path (Bragg peak); thus, CIRT allows selective irradiation of the target tumor, . Clinical and endoscopic findings. a, tissue necrosis from the skin to the bone was observed in the sacrum, and the rectum was exposed through the sacral defect. B, Colonoscopy revealed rectal intestinal necrosis. . the rectus abdominis myocutaneous flap was elevated and brought into the sacral defect through the intrapelvic space. . sacral appearance after the reconstruction.
sparing the surrounding normal tissues. [bib_ref] Surgical spacer placement prior carbon ion radiotherapy (CIRT): an effective feasible strategy..., Lorenzo [/bib_ref] Although CIRT causes relatively little damage to surrounding normal tissues, once the damage occurs, the normal tissues become sublethally compromised as a result of extensive tissue destruction, and recovery cannot be expected. In our case, the damage affected not only the surrounding tissue but also the rectum. Because sacral chordoma is close to the rectum, it is important for surgeons to anticipate rectal complications after CIRT, and if the rectum is damaged, resection should be considered. [bib_ref] Rectotumoral fistula formation occurring more than 5 years after carbon ion radiotherapy..., Ukon [/bib_ref] Reconstruction of a sacral full-thickness defect like that of our patient is challenging. [bib_ref] Surgical outcomes after abdominoperineal resection with sacrectomy and soft tissue reconstruction: lessons..., Manrique [/bib_ref] [bib_ref] Outcome and complications following vertical rectus abdominis myocutaneous flap surgery to reconstruct..., Houdek [/bib_ref] Three representative flaps have been reported for sacral reconstruction: gluteus maximus flap, transpelvic VRAM flaps, and free flaps; no significant difference in complications between these methods has been reported. [bib_ref] Flap reconstruction for sacrectomy defects: a systematic review and meta-analysis, Asaad [/bib_ref] However, after CIRT, adjacent blood vessels such as the superior and inferior gluteal vessels may be damaged and may not be reliable as pedicles for pedicled flaps or as recipients for free flaps. Miles et al 7 advocated the use of bilateral gluteal advancement flaps in patients with no preoperative radiation therapy and intact gluteal vessels; in contrast, in patients with a history of radiation to the sacral area, the use of the transpelvic VRAM flap should be considered. In addition, rectectomy may be required because rectal complications after CIRT may result in intrapelvic dead space. Placement of a pedicled VRAM flap after rectal resection could obliterate sacral defects and intrapelvic dead space. We consider such a flap suitable for reconstruction of a sacral defect after CIRT. |
Deciphering the Counterplay of Aspergillus fumigatus Infection and Host Inflammation by Evolutionary Games on Graphs
[formula] e S = E S /E R e S ≥ 1 0.3679 0.0434 e H = E H /E S e H ≥ 1 0.3463 0.0338 α AM = m R /E R 1.1439 - µ R = m R /m H µ R ≥ 1 0.3220 - µ S = m S /m R µ S ≥ 1 0.3324 - α PMN = g R /E R - 0.4847 ρ PMN ρ PMN ∈ [0, 1] - 0.5708 γ S = g S /g R γ S ≥ 1 - 0.0632 γ H = g H /g S γ H ≥ 1 - 0.2161 [/formula]
[table] Table S2: Mutual information between infection score and parameters of Game II and Game III. [/table]
|
320-row CT renal perfusion imaging in patients with aortic dissection: A preliminary study
ObjectiveTo investigate the clinical value of renal perfusion imaging in patients with aortic dissection (AD) using 320-row computed tomography (CT), and to determine the relationship between renal CT perfusion imaging and various factors of aortic dissection.MethodsForty-three patients with AD who underwent 320-row CT renal perfusion before operation were prospectively enrolled in this study. Diagnosis of AD was confirmed by transthoracic echocardiography. Blood flow (BF) of bilateral renal perfusion was measured and analyzed. CT perfusion imaging signs of AD in relation to the type of AD, number of entry tears and the false lumen thrombus were observed and compared. OPEN ACCESS Citation: Liu D, Liu J, Wen Z, Li Y, Sun Z, Xu Q, et al. (2017) 320-row CT renal perfusion imaging in patients with aortic dissection: A preliminary study. PLoS ONE 12(2): e0171235.ConclusionsThis study shows a direct correlation between renal CT perfusion changes and AD, with the size, number of intimal tears, different types of AD, different renal artery origins and false lumen thrombosis, significantly affecting the perfusion values.CT renal perfusion imaging in patients with aortic dissection PLOS ONE | length product; ED, effective dose; FOV, field of view; HU, Hounsfield unit; MIP, maximum intensity projection; MPR, multi planar reconstruction; ROI, region of interest; SD, standard deviation; TDC, time-density curve; TTP, time to peak; VR, volume rendering. Fig 2. Time-density curves (TDCs) of renal cortex from the single input maximum slope. (A) The ROIs were placed in the abdominal aorta and renal cortex. (B) TDC indicates the enhancement characteristics of the aorta and renal cortex during the first pass. Blood flow (BF) can be determined from the maximum gradient of renal cortical TDC by the peak enhancement of the aorta.* N indicates the number of patients, and data in parentheses are percentages. # Represents mean ± standard deviation.Number of entries and Size of entries, 1, one intimal tear; 2, one entry tear and one secondary tear; 3, one entry tear and two secondary tears; 4, one entry tear and three secondary tears; 5, one entry tear and four secondary tears; 6, one entry tear and five secondary tears. Abbreviation: AD, aorta dissection.
# Results
The BF values of patients with type A AD were significantly lower than those of patients with type B AD (P = 0.004). No significant difference was found in the BF between different numbers of intimal tears (P = 0.288), but BF values were significantly higher in cases with a false lumen without thrombus and renal arteries arising from the true lumen than in those with thrombus (P = 0.036). The BF values measured between the true lumen, false lumen and overriding groups were different (P = 0.02), with the true lumen group having the highest. Also, the difference in BF values between true lumen and false lumen groups was statistically significant (P = 0.016), while no statistical significance was found in the other two groups (P > 0.05). The larger the size of intimal entry tears, the greater the BF values (P = 0.044). Introduction Aortic dissection (AD) represents a common cardiovascular disease and it is responsible for the most commonly encountered pathologies in aortic emergency [bib_ref] Acute aortic dissection: clinician update, Braverman [/bib_ref]. AD is characterized by separating the aortic wall layers into true and false lumens through an intimal entrance tear [bib_ref] Multi-modality image-based computational analysis of haemodynamics in aortic dissection, Dillon-Murphy [/bib_ref]. Renal dysfunction is a common complication associated with aortic dissection, with resultant high mortality rate of about 27.27% [bib_ref] Branch vessel complications are increased in aortic dissection patients with renal insufficiency, Beckman [/bib_ref]. The antegrade propagation of the dissection from the proximal aorta to the level of renal arteries and the intervention during surgery may both increase the risk of renal malperfusion [bib_ref] Surgical delay for acute type A dissection with malperfusion, Deeb [/bib_ref] [bib_ref] Management strategies for type A dissection complicated by peripheral vascular malperfusion, Girardi [/bib_ref] ; thus, it is important to assess renal function for pre-and post-operative evaluation and guidance for treatment. This can be achieved through perfusion imaging. Computed tomography (CT) perfusion imaging can not only provide anatomical details, but also can describe vascular physiology [bib_ref] Perfusion CT: a worthwhile enhancement?, Miles [/bib_ref] [bib_ref] Perfusion computed tomography could be a new tool for single-session imaging of..., Yilmaz [/bib_ref]. A time-density curve can be calculated through dynamic CT imaging of tissue perfusion [bib_ref] Unilateral renal artery stenosis: perfusion patterns with electron-beam dynamic CT-preliminary experience, Paul [/bib_ref]. Presently, cerebral CT perfusion is well established, in particular, in the diagnostic assessment of acute stroke and tumors [bib_ref] Perfusion CT of the Brain and Liver and of Lung Tumors: Use..., Cros [/bib_ref] ; and the application of renal CT perfusion imaging is also widely used. There is increasing evidence to show that CT perfusion imaging is not only to measure changes in glomerular filtration [bib_ref] Determination of glomerular filtration rate using dynamic CT-angiography: simultaneous acquisition of morphological..., Helck [/bib_ref] and evaluate angiogenesis [bib_ref] Angiogenesis of renal cell carcinoma: perfusion CT findings, Chen [/bib_ref] , but also to monitor the response of tumors to treatment [bib_ref] Monitoring response to antiangiogenic treatment and predicting outcomes in advanced hepatocellular carcinoma..., Jiang [/bib_ref] [bib_ref] Computed tomography perfusion imaging for therapeutic assessment: has it come of age..., Goh [/bib_ref]. Several studies have demonstrated that multi-slice CT perfusion is a useful method for the evaluation of renal functions [bib_ref] Perfusion computed tomography could be a new tool for single-session imaging of..., Yilmaz [/bib_ref] [bib_ref] Assessment of glomerular filtration rate with dynamic computed tomography in normal Beagle..., Chang [/bib_ref] [bib_ref] Functional renal perfusion imaging with colour mapping: is it a useful adjunct..., Blomley [/bib_ref]. 320-row CT is a recent technological development in CT imaging. Extended longitudinal coverage of 16 cm enables imaging of the entire kidney perfusion. To the best of our knowledge, no study has assessed renal CT perfusion using 320-row dynamic volume CT in patients with acute AD, so the aim of this study is to assess the clinical potential of using 320-row CT in evaluating renal function in patients with AD. We hypothesize that different variables of AD are closely correlated with changes in renal CT perfusion.
# Materials and methods
## Patient population
This prospective study was approved by the Institutional Review Board of Beijing Anzhen Hospital Ethics Committee. During the period from May 2014 to January 2016, 46 patients diagnosed with AD by transthoracic echocardiography in our department, none of whom had any history of impaired functions of the heart, liver, or kidneys, or a history of allergy to the iodine contrast medium, were invited to take part in this study. Further, patients who had accessory renal arteries were excluded from this study. Three cases were excluded from data analysis after CT imaging due to severe motion artifacts. The remaining 43 (35 males and 8 females) with a mean age of 50.47 ± 9.82 years (range 28-73 years) were included in the perfusion analysis. Written informed consent was obtained from all participants.
## Ct contrast protocols
CT scanning was performed using a 320-row dynamic volume CT scanner (Aquilion One, Toshiba Medical Systems, Ottawara, Japan). The following contrast injection protocol was used for perfusion imaging: a bolus of iodinated contrast material 40 ml (Ultravist 370, 370 mg/mL; Bracco S.P.A., Italy) was administered intravenously at a flow rate of 5 mL/s, followed by a 20 mL saline chaser at the same flow rate. The total injection length was 8 s. Image postprocessing was carried out on a Vitrea workstation with special CT body perfusion software.
## Procedure design and scanning techniques
In order to minimize the radiation dose, we established a dynamic CT protocol with the following parameters: tube voltage of 100 kV; tube current of 60 mA; 0.5 s gantry rotation time; 1 mm pixel spacing; 512 X 512 pixel (spatial resolution); and 0.5 mm reconstructed slice thickness.
Instructions were given to patients prior to CT scans to ensure natural breathing. The first and second steps were similar to a previous study [bib_ref] Study of 320-slice dynamic volume CT perfusion in different pathologic types of..., Chen [/bib_ref] , then they were followed by a single-section tracking series (section thickness, 5 mm; tube current, 60 mA; tube voltage, 100 kV) at the renal hilum level. In the third step, volumetric cine scans were automatically triggered at a preset threshold of 90 HU in the lumen of the abdominal aorta above the renal artery. In total, 17 CT volumes of the kidneys were acquired without table movement. It took 0.5 s to complete each of these 17 scans (0.5 s per gantry rotation of 360 degree for one volume data acquisition on 320-slice Toshiba scanner, with different temporal sampling intervals being 1.5 to 4.5 s), thus the whole process took 58 s in total. The data were processed with adaptive iterative dose reduction and were reconstructed automatically with a section thickness of 5 mm and 5 mm spacing.
The volumetric CT dose index (CTDI vol ) and dose-length product (DLP) of CT examination was automatically recorded and displayed on CT console. The effective dose (ED) was estimated using an organ conversion coefficient k-factor of 0.015 mSv/mGy.cm [bib_ref] Assessment of a theoretical formalism for dose estimation in CT: an anthropomorphic..., Brix [/bib_ref].
## Perfusion image post-processing
Post-processing and measurement of CT perfusion images were conducted by two experienced radiologists with 12 and 21 years of diagnostic experience in CT imaging, who reached consensus on the image analysis. We used the same approach as reported in a previous study to deal with motion and breathing artifacts [bib_ref] Study of 320-slice dynamic volume CT perfusion in different pathologic types of..., Chen [/bib_ref]. The new registered datasets were post-processed using Body Perfusion software (Vitrea fx ves 6.0, Licensed software, Toshiba Medical Systems) with details as follows: The region of interest (ROI) was placed in the abdominal aorta, renal cortex separately at the level of the renal hilum to obtain a time-density curve (TDC). A previously reported model was used [bib_ref] Comparison between the deconvolution and maximum slope 64-MDCT perfusion analysis of the..., Djuric-Stefanovic [/bib_ref] to acquire a blood flow (BF) map (BF mL/ min /100 mL) based on TDC .
The perfusion maps [fig_ref] Fig 3: Examples of regions of interest [/fig_ref] were generated by the Body Perfusion software. The ROIs of the renal cortex were selected from four points on the axial section of kidney in an X shape; three points of upper pole, hilum and lower pole of kidney at coronal section separately [fig_ref] Fig 3: Examples of regions of interest [/fig_ref] and whole kidney [fig_ref] Fig 3: Examples of regions of interest [/fig_ref] were defined manually on the axial and coronal planes. The BF values were recorded separately and the BF values of the different types of aortic dissection were compared.
For all patients, dynamic CT examinations were executed on thoracic and abdominal great vessels, and multiplanar reconstruction (MPR), maximum-intensity projection (MIP) and volume rendering (VR) images were generated to observe these various CT imaging signs associated with AD, which included different AD types, origin of renal arteries, the false lumen thrombosis, and the size, number and position of entry tears. A comparison was then performed between the perfusion imaging and these dissection features. We measured the size, number and position of entry tears on MPR reconstruction images [fig_ref] Fig 4: Example of size measurement and numbers of the entry tears [/fig_ref]. Based on different MPR images, we choose the maximum diameter as the size of the entry tears.
# Statistical analysis
Statistical analysis was performed using SPSS (SPSS, V 20.0A IBM, Armonk, NY, USA). Continuous variables of perfusion parameters were presented as mean and standard deviation (mean ± SD). Independent-Samples T test was used to analyze the relationship of renal perfusion to the number of intimal entries and different AD types, and to assess the relationship of false lumen thrombosis and BF. Analysis of variance (ANOVA) was used to evaluate the differences of the perfusion values among the three groups of renal arteries: originating from the true lumen, from the false lumen, or overriding. A simple regression analysis was performed to examine the relationship between the BF and the size of intimal entries. Multiple linear regression was used to assess the relationships between various kinds of variable and the BF. The axial BF and coronal BF values of the same patient were assessed using Bland-Altman plots: a p value less than 0.05 was considered to indicate a significant difference.
# Results
The perfusion imaging examinations were successfully performed in all 43 patients with no technical problems or adverse reactions to the contrast medium. The estimated mean ED was 11.09 ± 0.29 mSv (range, 10.69-11.72 mSv). Type A AD was found in seven patients with the intimal tear located in the ascending aorta, while type B AD was found in 36 patients with the tear located distal to the left subclavian artery orifice or descending aorta. Twenty-two patients had one entry, and the remaining 21 had two or more (3 to 6) entries. Of the 86 renal arteries analyzed, 58 originated from the true lumen, 20 from the false lumen; 8 were overriding [fig_ref] Table 1: Characteristics of study population [/fig_ref]. Patients with accessory renal arteries were not included in the analysis due to potential bias of renal perfusion analysis. (S1 File)
## Ct color maps
The TDC curves of the renal tissue in AD demonstrate slowly upgrading slopes, lower peak values and delayed time to peak (TTP) [fig_ref] Fig 5: Time-density curves [/fig_ref]. The color maps of the kidney reveal that the renal cortex has become thin and BF has decreased. In three cases the BF color maps are inhomogeneous, indicating the possibility of a partially decreasing BF. One patient in this group shows infarction at the inferior pole of the left kidney [fig_ref] Fig 6: Blood flow [/fig_ref].
## Mean perfusion parameters of bilateral kidneys in ad
The average BF values of two types of AD, the different number of intimal tears and the false lumen with or without thrombosis, are summarized in [fig_ref] Table 2: Independent-Samples T test [/fig_ref]. The BF values of patients with type A AD were significantly lower than those of patients with type B AD (P = 0.004). No significant difference was found between BF values and the number of intimal tears (P = 0.288). The BF value of AD with no thrombus in the false lumen was significantly higher than that of the false lumen with thrombus (P = 0.036). Different renal artery origins caused different BF values. The BF values among the true lumen, false lumen and overriding groups were different (P = 0.02): with the true lumen group having highest, and the overriding group lowest. The BF values between the true lumen and false lumen groups were significantly different (P = 0.016), but there was no statistical significance between the false lumen and overriding groups, and the true lumen and overriding groups (P > 0.05), with findings summarized in [fig_ref] Table 4: Multiple comparisons [/fig_ref]. The size of the intimal entry tears has a direct relationship with BF. The regression coefficient is 2.503 (95% CI: 0.063-4.942). A coefficient greater than 0 indicates the size of intimal tears has a positive correlative relationship with BF. With the increasing size of intimal entry tears, the BF value of the kidney increased significantly (P = 0.044, , result of regression equation was shown in [fig_ref] Fig 7: Result of simple regression analysis [/fig_ref] The results of multiple linear regression analysis of the 86 renal arteries from 43 patients including age, gender, type and other factors showed that the higher the numbers of intimal entry tears, false lumen without thrombosis, and renal artery originating from the true lumen, the higher the BF values. The BF value of renal perfusion was significantly related to the number of intimal tears, the false lumen with or without thrombosis, and origin of the renal artery (P < 0.0001). Bland-Altman analysis for differences in BF value from two different observation positions [fig_ref] Fig 9: Bland-Altman analysis on differences between the average blood flow [/fig_ref] reveals that the BF value of the axial section is similar to that of the coronal section in each patient. The analysis was performed on the left and right sides of 86 renal arteries in 43 patients, with good consistency.
# Discussion
In this study we investigate the clinical application of renal perfusion in patients with aortic dissection. Our findings show that blood flow values to the kidneys are directly related to the type of aortic dissection, renal artery origins, the number and size of the intimal tears, and to false lumen thrombosis. CT perfusion, a non-invasive method for hemodynamic and functional evaluation of tissues and organs, can provide organ anatomy and quantitative perfusion parameters in one scan. The theoretical base of CT perfusion is radioactive tracing (from nuclear medicine) and central volume discipline; the application base is the rapid volume scanning technique of a multidetector spiral CT [bib_ref] Measurement of tissue perfusion by dynamic computed tomography, Miles [/bib_ref]. CT perfusion is performed by scanning the same slice or volume at different times after an intravenous injection of a contrast medium, to get the TDC curve of each pixel in the selected slice or volume; perfusion parameters are then calculated according to different algorithms. Renal CT perfusion study without 320-row CT has several limitations in clinical use: first, the longitudinal scan range is limited, with the ROI full definition set at the level of the renal hilum and is unable to reflect the whole kidney perfusion (the upper and lower poles of the kidney). Second, it requires strict breath holding; but for elderly, severely ill, or pediatric patients, holding breath for more than 60 seconds presents challenges and is impossible in some cases. This presents a problem as the target region of the kidney may exceed the image plane in such cases. It is also difficult to obtain satisfactory images and data because of motion artifacts and image mis-registration [bib_ref] Study of 320-slice dynamic volume CT perfusion in different pathologic types of..., Chen [/bib_ref]. The 320-row CT can avoid motion artifacts because of its high temporal resolution and use of image registration post-processing software [bib_ref] Whole-organ perfusion of the pancreas using dynamic volume CT in patients with..., Kandel [/bib_ref].
Hemodynamic changes in tissues or local lesions have been evaluated by these perfusion parameters and color maps [bib_ref] Colour perfusion imaging: a new application of computed tomography, Miles [/bib_ref] [bib_ref] Functional images of hepatic perfusion obtained with dynamic CT, Miles [/bib_ref] [bib_ref] Measurement of human pancreatic perfusion using dynamic computed tomography with perfusion imaging, Miles [/bib_ref]. Daghini et al.studied renal hemodynamic and function changes in stenotic renal arteries in animal models by analyzing renal perfusion compared with that of full-definition electron beam computed tomography(EBCT). They concluded that the multi-slice CT renal perfusion can accurately assess quantitative renal hemodynamic and functional changes. Blood flow was defined as the flow rate through the vasculature in a given tissue region, an important parameter in CT perfusion study equal to blood perfusion volume [bib_ref] A CT method to measure hemodynamics in brain tumors: validation and application..., Cenic [/bib_ref] [bib_ref] Assessment of the reproducibility of postprocessing dynamic CT perfusion data, Fiorella [/bib_ref]. These studies show the feasibility of renal CT perfusion in renal artery disease; however, the current study utilizes the technique in aortic dissection, which has not been investigated before.
Assessment of organ or tumor perfusion has been of interest for years. In the brain, CT perfusion imaging plays an important role from the clinical perspective, detecting ischemic tissue in patients with acute stroke [bib_ref] Dynamics of perfusion CT parameters in the early acute phase of ischemic..., Shamalov [/bib_ref]. Also, it demonstrates hemodynamic changes in brain tissue after acute supratentorial spontaneous intracerebral hemorrhage [bib_ref] CT perfusion imaging evaluation on hemodynamic changes of acute spontaneous intracerebral hemorrhage..., Kuang [/bib_ref]. Perfusion CT can also be applied to the diagnosis and monitoring of renal tumor and liver carcinoma [bib_ref] Perfusion-CT monitoring of cryo-ablated renal cells tumors, Squillaci [/bib_ref] [bib_ref] Liver computed tomographic perfusion in the assessment of microvascular invasion in patients..., Wu [/bib_ref]. In our study, 320-row dynamic volume CT successfully assessed bilateral renal imaging. In AD patients, The TTPs of aorta and kidney are all delayed, and TDC curves of the false lumen have lower peak values and lower upgrade slopes. Our results demonstrate a decrease in the perfusion parameters of the AD kidneys, which represents an abnormal low perfusion in the cortex and medulla in patients to different degrees. In our study, average BF values in the true lumen, the false lumen and the overriding groups were 281.46, 223.90 and 219.24mL/100mL/min respectively, which are lower than the values of normal kidneys reported in previous studies using the same CT system [bib_ref] Study of 320-slice dynamic volume CT perfusion in different pathologic types of..., Chen [/bib_ref] [bib_ref] The clinical application of one-stop examination with 640-slice volume CT for Nutcracker..., Zhong [/bib_ref]. Zhong et al. [bib_ref] The clinical application of one-stop examination with 640-slice volume CT for Nutcracker..., Zhong [/bib_ref] reported that the BF values of the left and right renal cortex were 323.8 and 322.9 mL/100mL/min in control group. Chen et al. [bib_ref] Study of 320-slice dynamic volume CT perfusion in different pathologic types of..., Chen [/bib_ref] reported that the BF value of the normal renal cortex was 305.4 mL/100mL/min. The BF values of the coronal section are similar to those of the axial section in our study.
A study by Kandel et al. [bib_ref] Determination of split renal function using dynamic CT-angiography: preliminary results, Helck [/bib_ref] reported that a relatively slow injection rate of 11.5 seconds in a maximum slope perfusion CT protocol could lead to an 8% underestimation of true perfusion. Yuan et al. [bib_ref] A Simplified Whole-Organ CT Perfusion Technique with Biphasic Acquisition: Preliminary Investigation of..., Yuan [/bib_ref] showed that using the maximum slope method as a reference standard was likely to cause a slight underestimation of the true perfusion value because of the long injection duration of the protocol. In our study, the use of 40 mL contrast agent within eight seconds of the injection guaranteed the accuracy of the perfusion value. This is clinically important, as reducing the contrast medium contributes to the reduction of contrast-induced acute kidney injury, the main concern associated with contrast-enhanced CT imaging.
Studies have found that patients with AD generally present more than two tears [bib_ref] Distribution of intimomedial tears in patients with type B aortic dissection, Khoynezhad [/bib_ref]. These secondary tears are typically found in the vicinity of visceral branches, and especially near the opening of the renal artery. It remains an important question whether secondary tearing has a significant impact on the dissection hemodynamic; thus in our study, including gender, age, and the size, number and position of intimal entries and a variety of other factors were considered in the multiple linear regression analysis. It was found in our study that the number of intimal tears, the false lumen thrombosis and different renal artery origin have important influence on the BF value. We believe these findings could assist in the assessment of renal function for pre-and post-operative evaluation and treatment guiding.
Acute AD is one of the most lethal surgical emergencies of the aorta, resulting from a tear in the aortic wall intima and extending into the aortic wall media to create a false lumen and a dissection flap [bib_ref] Acute Complex Type A Dissection associated with peripheral malperfusion syndrome treated with..., Suliman [/bib_ref]. Determination of renal cortex blood perfusion is a direct index by which to judge renal blood flow. In many pathological conditions such as acute and chronic renal failure, nephritis, urinary system obstruction, renal transplant rejection, renal artery stenosis and other diseases can affect the renal cortex and medulla perfusion [bib_ref] Unilateral renal artery stenosis: perfusion patterns with electron-beam dynamic CT-preliminary experience, Paul [/bib_ref] [bib_ref] Contrast enhanced spiral computerized tomography in live kidney donors: a single session..., El-Diasty [/bib_ref]. Many studies have considered perfusion in this field, but renal blood perfusion changes in patients with AD have rarely been reported. Nearly 30% of acute AD associated with branch artery ischemia and renal artery involvement lead to renal blood perfusion changes. It has been found that the AD postoperative mortality rate is significantly associated with the duration of preoperative ischemic symptoms. Organ blood flow changes should be observed closely for AD patients. If this is delayed for a long time for an emergency operation induced by organ ischemia, multi-organ failure cannot be reversed even if blood flow is recovered. A study has shown that half of the aneurysmal patients with type B dissection who died after operations experienced partial irreversible damage before interventional therapies [bib_ref] Clinical profiles and outcomes of acute type B aortic dissection in the..., Suzuki [/bib_ref]. As CT perfusion imaging can reveal the hemodynamic changes of the kidney in AD patients, the findings of this study will assist clinicians to predict renal blood flow changes and so minimize mortality. Renal perfusion imaging can be used to guide clinicians choose surgery time, for example, patients with lower BF should be operated as soon as possible while higher BF would allow more time for thorough pre-operative planning. Both of surgical and interventional procedures may further affect the renal function. For patients with type A AD, if renal artery originates from the false lumen with lower BF, surgical procedure is recommended. Surgeons need to pay more attention to the involved renal artery during operation because acute kidney injury is one of the common complications that may occur after operation. CT renal perfusion imaging is a useful tool to assess renal functional status pre-operation, reveal the situation of renal ischemia post-operation, and early positive therapy can prevent the development of renal failure. We believe that CT-derived BF measurement could have a prognostic value for AD patients, although further studies with inclusion of large cohort of patients are needed.
Some limitations in this study should be addressed. First, the sample size is small, and is also based on a single-center experience. Further studies based on the analysis of more cases are required. Second, we did not study renal function changes, although no impairment was found in the recruited patients. Longitudinal studies with inclusion of follow-up data would correlate CT perfusion imaging with clinical outcomes, in particular with changes in renal function. Third, since patients with accessory renal artery were not included, their results may not apply to those patient presenting an accessory renal artery. Finally, although CT perfusion imaging in this study afforded statistical power to detect significant changes in renal perfusion in patients with different types of AD, correlation of CT perfusion imaging with CT-derived computational fluid dynamics (CFD) could represent further research investigations. CFD is increasingly used in cardiovascular disease due to its ability of quantifying hemodynamic changes to the aortic dissection and coronary artery disease with promising results reported [bib_ref] Impact of plaques in the left coronary artery on wall shear stress..., Chaichana [/bib_ref] [bib_ref] A systematic review of computational fluid dynamics in type B aortic dissection, Sun [/bib_ref]. In particular, CFD simulations of aortic dissection pre-and post-endovascular stent grafting have been shown to play an important role in the prediction of disease progression and patient outcome [bib_ref] A systematic review of computational fluid dynamics in type B aortic dissection, Sun [/bib_ref].
# Conclusion
This study shows that 320-row dynamic volume CT brings an unprecedented improvement in renal perfusion and provides important information about whole kidney blood perfusion in aortic dissection patients. It is concluded that CT perfusion imaging is an effective and noninvasive technique to evaluate renal perfusion in aortic dissection patients.
Supporting information S1 File.
[fig] Fig 1: Transverse contrast-enhanced CT scan showed the effect of motion correction. (A) Motion artifacts were seen as blurring of the kidney and aorta contours (white arrowheads). (B) Most motion artifacts were not seen after correction. doi:10.1371/journal.pone.0171235.g001 [/fig]
[fig] Fig 3: Examples of regions of interest (ROIs) measurements. (A) ROIs of renal cortex at axial view. (B) ROIs of renal cortex at coronal view. (C) ROIs of whole kidney were defined manually on the axial plane. (D) ROIs of whole kidney on the coronal plane. doi:10.1371/journal.pone.0171235.g003 [/fig]
[fig] Fig 4: Example of size measurement and numbers of the entry tears. (A) Primary entry tear located in the aortic arch. (B) MPR image displaying five small re-entry tears in descending and abdominal aorta, with the primary entry tear locate in the ascending aorta. doi:10.1371/journal.pone.0171235.g004 [/fig]
[fig] Fig 5: Time-density curves (TDCs) from the single input maximum slope. TDC indicates the enhancement characteristics of the aorta and renal cortex during the first pass. BF can be determined from the maximum gradient of renal cortex. doi:10.1371/journal.pone.0171235.g005 [/fig]
[fig] Fig 6: Blood flow (BF) color maps of bilateral kidneys. White arrow indicate the partially decreasing BF. doi:10.1371/journal.pone.0171235.g006 [/fig]
[fig] Fig 7: Result of simple regression analysis. The result indicates the positive correlative relationship between the perfusion values and the size of intimal entries. doi:10.1371/journal.pone.0171235.g007 [/fig]
[fig] Fig 8: Residual diagram of model fitting of multi factor regression analysis of BF value. The result displays all the points of distribution on both sides of central line, the residual plots obey normal distribution. doi:10.1371/journal.pone.0171235.g008 [/fig]
[fig] Fig 9: Bland-Altman analysis on differences between the average blood flow (BF) value of the axial plane and that of the coronal plane. (A) BF value differences between the two planes for right kidney.(B) BF value differences between the two planes for left kidney. Mean difference (blue line), mean difference ± 1.96 SD (red line). doi:10.1371/journal.pone.0171235.g009 [/fig]
[table] Table 1: Characteristics of study population. [/table]
[table] Table 2: Independent-Samples T test. [/table]
[table] Table 4: Multiple comparisons (LSD-t test). [/table]
|
Detection of allelic variations of human gene expression by polymerase colonies
Background: Quantification of variations of human gene expression is complicated by the small differences between different alleles. Recent work has shown that variations do exist in the relative allelic expression levels in certain genes of heterozygous individuals. Herein, we describe the application of an immobilized polymerase chain reaction technique as an alternative approach to measure relative allelic differential expression.Results:Herein, we report a novel assay, based on immobilized polymerase colonies, that accurately quantifies the relative expression levels of two alleles in a given sample. Mechanistically, this was accomplished by PCR amplifying a gene in a cDNA library in a thin polyacrylamide gel. By immobilizing the PCR, it is ensured that each transcript gives rise to only a single immobilized PCR colony, or "polony". Once polony amplified, the two alleles of the gene were differentially labeled by performing in situ sequencing with fluorescently labeled nucleotides. For these sets of experiments, silent single nucleotide polymorphisms (SNPs) were used to discriminate the two alleles. Finally, a simple count was then performed on the differentially labeled polonies in order to determine the relative expression levels of the two alleles. To validate this technique, the relative expression levels of PKD2 in a family of heterozygous patients bearing the 4208G/A SNP were examined and compared to the literature.Conclusions:We were able to reproduce the results of allelic variation in gene expression using an accurate technology known as polymerase colonies. Therefore, we have demonstrated the utility of this method in human gene expression analysis.
# Background
The advent of the PCR played an important role in revolutionizing research in the field of molecular biology. A recent adaptation of this technology, known as polymerase colonies, or "polonies", holds tremendous promise as well [bib_ref] In situ localized amplification and contact replication of many individual DNA molecules, Mitra [/bib_ref] [bib_ref] Functional characterization of mutant yeast PGK1 within the context of the whole..., Joshua [/bib_ref] [bib_ref] Digital genotyping and haplotyping with polymerase colonies, Mitra [/bib_ref] [bib_ref] Molecular colony diagnostics: detection and quantitation of viral nucleic acids by in-gel..., Chetverina [/bib_ref] [bib_ref] Characterization of Mutations and LOH of p53 and K-ras2 in Pancreatic Cancer..., Butz [/bib_ref]. Polony technology is a form of PCR in which the reaction is immobilized in a thin polyacrylamide gel attached to a microscope slide. As the PCR proceeds, the PCR products diffuse radially within the gel from its immobilized template (e.g., cDNA or genomic DNA), giving rise to a circular PCR or polomerase colony [bib_ref] In situ localized amplification and contact replication of many individual DNA molecules, Mitra [/bib_ref]. When the gel is stained and scanned with a microarray scanner, the polymerase colony resembles a bacterial colony on an agar plate, hence its name. In this paper polony technology is adapted to quantify the variations in allelic expression levels for a heterozygous gene.
Recent work has shown that variations exist in the relative allelic expression levels in certain genes of heterozygous individuals [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref]. Herein, we describe the application of polony technology as an alternative approach to measure relative allelic differential expression. Initially, the two alleles for a particular gene were polony amplified with primers which flank the SNP , giving rise to one polony per transcript. The polonies were observed after the gel was stained with SyBr Green I and scanned with a microarray scanner . An additional step was required to discriminate the two alleles which involved: 1) making the polony single stranded by formamide denaturation and electrophoresis ; 2) hybridization of a sequencing primer with the 3' end preceding the SNP; 3) a single-base extension with fluorescent nucleotides to differentially label the two alleles ; and 4) scanning the gel to determine allelic expression levels . This approach was able to successfully identify variations in the 4208A/G alleles of PKD2 in heterozygous patients as previous described [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref].
# Results
In the work by Yan et al. [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref] the allelic expression levels of PKD2 in heterozygous individuals (4208A/G) was investigated in a CEPH family of eight individuals. Of the eight individuals, it was determined that three individuals exhibited a greater than 20% difference in the relative expression levels of 4208A and 4208G PKD2 [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref] [bib_ref] Genetics of gene expression surveyed in maize, mouse and man, Schadt [/bib_ref].
To determine if polony technology could be used to accurately measure the relative allelic expression levels, the relative expression levels of 4208A and 4208G PKD in the same eight patients was examined using polonies in a blind assay and the subsequent results were compared to the original study [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref]. Initially, cDNA for each of the eight patients was obtained as previously described [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref] prior to polony amplification in separate gels. The cDNA for each patient was blindly analyzed in quadruplicate as described in the materials and methods. For each patient the respective 4208A and 4208G polony count for all four of the quadruplicate gels was determined as well as the relative expression levels of the 4208A and 4208G alleles [fig_ref] Table 1: The relative expression of 4208A, 4208G PKD2 alleles using polony technology [/fig_ref].
Polony quantification of relative expression levels was determined to be sensitive enough to detect variation in relative levels of 4208A and 4208G across patients. For example, a polony gel for patient 50 clearly shows elevated expression of PKD2 transcripts bearing the 4208G SNP , red polonies) compared to the 4208A SNP , green polonies). Compared to the previously published data of Yan et al., polony quantification was able to detect the elevated 4208G to 4208A expression levels in the same three patients, namely patients 50, 50-3 and 50-5. Furthermore, similar relative allelic expression levels were detected in all patients using both methods, and the rank order of relative expression levels across patients also remained largely unaffected by the choice of method. It should be noted that analysis of 4208A/G levels in genomic DNA from four of the eight patients was close to the expected 1:1 4208A to 4208G levels [fig_ref] Table 2: The relative levels of 4208A, 4208G PKD2 alleles in genomic DNA using... [/fig_ref].
# Discussion
There are at least two major advantages to the polony approach for measuring allelic variation in gene Determining relative allelic expression using polony technology Determining relative allelic expression using polony technology. (A) Initially, the gene of interest is polony amplified from cDNA. The primers are designed to amplify the region of the gene bearing a silent SNP, which is used to discriminate the two alleles. In this illustration, one allele bears an A-T pairing at a particular position whereas the other allele has a G-C pairing instead. Note: one of primers has a 5' acrydite modification (star), which is eventually required for distinguishing these two alleles. (B) Following polony amplification, the gel is stained with SybrGreenI and scanned with a microarray scanner in order to ensure that the polony amplification worked properly. Each black dot represents the polony amplification of one copy of the desired first-strand cDNA template. (C) Prior to hybridization of sequencing primer, the polony is made single stranded by denaturing the double stranded in formamide followed by the removal of non-acrydited strand using electrophoresis. (D) A sequencing primer is hybridized to the single stranded polony, a single base extension is performed with Cy5-dATP and then the gel is scanned. To label the other allele, the process of denaturation, hybridation, extension, and scanning is repeated, but the extension is performed with Cy5-dGTP instead. (E) To determine the relative expression levels of the two alleles, a composite image from the Cy5-dATP and Cy5-dGTP extensions is first generated and then the number of polonies (transcripts) for each of the two alleles is counted. In this particular example, the "G-C" SNP represents 62.5% of the population (i.e., 5 out of 8 polonies). Figure adapted from Yan et al [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref].
[formula] T A T T A C G C C G A B C D E [/formula]
expression. First, the measurements are digital and hence more accurate than the 'analog' approaches for quantifying allelic variation (see . In addition, the polony approach for measuring allelic variation provides an absolute measurement of the relative expression of the two alleles, whereas other approaches are normalized to the population average which is assumed to be 1:1.
# Conclusions
We have shown that immobilized PCR, or polony technology, can be successfully used to detect variations in relative allelic expression of heterozygous genes. We believe that this is the next evolutionary step in analyzing allelic variation, and trying to relate allelic variation to disease, due to the "digital" nature of the data generated using this technique.
# Methods
## Polony gels
Polony gels were cast on Teflon coated microscope slides for detection of allelic variation. To determine the relative expression levels of the two PDK2 alleles from numerous heterozygous patients, 20 µl of master mix was combined with 1 µl of cDNA (The cDNA for allelic variation analysis was prepared as described [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref] diluted to the appropriate level For all patients, the polony count for both alleles represents the total number of polonies for the respective species observed in quadruplicate gels following polony analysis of cDNA. The P-value indicates the probability that the alleles are expressed at an equal ratio given the data. The P-value indicates that significance of the result.
Representative polony gel for patient 50 following extensions with fluorescently labeled dATP (green) and dGTP (red) Representative polony gel for patient 50 following extensions with fluorescently labeled dATP (green) and dGTP (red). This is a composite image generated from the independent extensions with either Cy5-dATP or Cy5-dGTP as described in .
The colors used in the image are artificial.
to obtain distinguishable non-overlapping polonies as well as 0.23 µl of 50 µM forward (5'-CAGCACTTT-GGGAGGCCGAAC-3') and reverse primer (5'-CCCCAT-AAATTTGGGTGAGACCCTC-3') designed to polony amplify the region of the PKD2 gene bearing the silent SNP, 4208A/G [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref].
The reverse primer in the gel was modified with a 5' acrydite to covalently link one strand to the gel. Immediately prior to casting the polony gel, 1.38 µl of JumpStart Taq Polymerase (Sigma), 0.34 µl of 5% APS, and 0.34 µl of TEMED were added to the master mix/primer/DNA solution. The sample was mixed prior to pipeting 20 µl of the mixture into the void space created by placing a coverslip onto a Teflon-masked, bind-silane treated slide. The polyacrylamide gel polymerized for 10 minutes, and then a hybrid well cover (Grace Bio-labs) was placed on top of the gel and light mineral was pipeted into the hybrid well chamber. The slide was placed in a hybridization tower and PCR was performed (94 C for 2 minutes, 40 cycles of 94 C for 15 seconds, 63 C for 30 seconds and 72 C for 30 seconds, and a final extension step at 72 C for 2 minutes).
Upon completion of the PCR reaction, the hybrid well cover was removed and the slide was placed in hexane for 5 minutes to remove the mineral oil. The coverslip was carefully removed. The slide was then dipped in clean hexane to remove residual oil prior to incubation in a 2X Sybr Green solution (20 µl Sybr GreenI (Molecular Probes) in 100 ml 1x TBE) for 15 minutes. Finally, the slide was washed in 1x TBE for 15 minutes and scanned using a ScanArray 5000 microarray scanner (PerkinElmer) with the FITC laser and filter set.
## Denaturation and electophoresis of polony gels
The polony DNA was denatured and a sequencing primer was hybridized to the 5' side of the 4208A/G SNP prior to in situ sequencing. The double stranded polonies were made single stranded by stripping away the non-acrydited strand in a two step procedure. The first step was to denature the polonies by incubating the gels in a formamide buffer (1x SSC, 70% formamide, 25% dd-water) at 70 C for 15 minutes. Immediately following denaturation, the gels were electrophoresed to remove the non-acrydited strand.
## Hybridization and in situ sequencing
After electrophoresis, the polony slides were washed 4 × in wash 1E (0.1 M Tris-Cl, pH 7.5, 20 mM EDTA, 0.5 M KCl) to prepare for hybridization of the sequencing primer. 200 µl of annealing buffer (6x SSPE, 0.01% Triton X-100) containing 0.5 µM sequencing primer (5'-TGATACT-GTCTAATGTATGAACTTTTAATA-3', [bib_ref] Allelic variation in human gene expression, Yan [/bib_ref] was then pipetted onto the gel and covered with a hybrid well chamber.
The sample was then placed in a hybridization tower and heated for 2 minutes at 94 C followed by 20 minutes at 60 C to facilitate hybridization.
The SNP was sequenced using a single base extension of the sequencing primer with fluorescently labeled dNTPs. Following hybridization, the gels were washed 2 × in For all patients, the polony count for both alleles represents the total number of polonies for the respective species observed in triplicate gels following polony analysis of genomic DNA. All ratios are not significantly different than 1.00.
Sample size required to differentiate various proportions of allelic expression Sample size required to differentiate various proportions of allelic expression. The sample size is shown that is required to differentiate various levels of allelic variation of gene expression. We have required a power of 0.90 for these data.
## Proportion to differentiate
## Sample size
Wash1E and then equilibrated in Klenow extension buffer (50 mM Tris-Cl, pH 7.5, 5 mM MgCl 2 , 0.01% Triton X-100) for 1 minute. For each sample, 50 µl solution containing approximately 5 units of Klenow large fragment (New England Biolabs), 3 µg of single stranded binding protein (US Biochemicals), and either 0.5 µM Cy5-labeled dATP or dGTP (PerkinElmer) was pipetted onto the gel. The single base extension was allowed to proceed for 2 minutes. The gels were then washed in Wash1E to reduce background fluorescence and scanned on the ScanArray5000 with the appropriate lasers and filters. The process of formamide denaturation, hybridization, in situ sequencing, and scanning was repeated an additional time. This was necessary to completely discriminate nucleotides 4208A and 4208G in heterozygous PKD2 transcripts.
# Authors' contributions
JAB performed most of the experiments in this manuscript in addition to writing the preliminary manuscript. HY helped design experiments and also performed all the experiments necessary to procure the cDNA used in this project. VM performed valuable experiments needed to complete this work. JSE conceived and funded this project. All authors read and approved the final manuscript.
[table] Table 1: The relative expression of 4208A, 4208G PKD2 alleles using polony technology. [/table]
[table] Table 2: The relative levels of 4208A, 4208G PKD2 alleles in genomic DNA using polony technology. [/table]
|
Spatial control of the GEN1 Holliday junction resolvase ensures genome stability
Holliday junction (HJ) resolvases are necessary for the processing of persistent recombination intermediates before cell division. Their actions, however, need to be restricted to the late stages of the cell cycle to avoid the inappropriate cleavage of replication intermediates. Control of the yeast HJ resolvase, Yen1, involves phosphorylation changes that modulate its catalytic activity and nuclear import. Here, we show that GEN1, the human ortholog of Yen1, is regulated by a different mechanism that is independent of phosphorylation. GEN1 is controlled exclusively by nuclear exclusion, driven by a nuclear export signal (NES) that restricts GEN1 actions to mitosis when the nuclear membrane breaks down. Construction of a nuclear-localized version of GEN1 revealed that its premature actions partially suppress phenotypes associated with loss of BLM and MUS81, but cause elevated crossover formation. The spatial control of GEN1 therefore contributes to genome stability, by avoiding competition with non-crossover promoting repair pathways.
H omologous recombination promotes the repair of DNA double-strand breaks and facilitates the recovery of stalled or collapsed replication folks [bib_ref] Regulation of homologous recombination in eukaryotes, Heyer [/bib_ref] [bib_ref] Pathways of mammalian replication fork restart, Petermann [/bib_ref]. Repair involves a reciprocal exchange of DNA strands between sister chromatids or homologous chromosomes, leading to the formation of intermediates, such as Holliday junctions (HJs), that provide a covalent linkage between sister chromatids or homologous chromosomes [bib_ref] A mechanism for gene conversion in fungi, Holliday [/bib_ref] [bib_ref] Double Holliday junctions are intermediates of DNA break repair, Bzymek [/bib_ref] [bib_ref] Molecular views of recombination proteins and their control, West [/bib_ref]. The efficient processing of these joint molecules, which is essential for chromosome segregation at cell division, also plays an important role in determining the outcome of recombination: for example, crossovers (COs) between homologous chromosomes are required for meiotic division, whereas non-COs are favored in mitotic cells to avoid loss of heterozygosity [bib_ref] Branching out: meiotic recombination and its regulation, Cromie [/bib_ref] [bib_ref] Holliday junction resolution: regulation in space and time, Matos [/bib_ref].
In mitotic cells, two major mechanisms ensure the timely elimination of joint molecules. First, the human BTR complex (BLM-TOPIIIa-RMI1-RMI2), or yeast STR complex (Sgs1-Top3-Rmi1), promote the dissolution of double HJs to generate NCO products [bib_ref] The Bloom's syndrome helicase suppresses crossing over during homologous recombination, Wu [/bib_ref] [bib_ref] Rmi1 stimulates decatenation of double Holliday junctions during dissolution by Sgs1-Top3, Cejka [/bib_ref]. Second, structure-selective endonucleases resolve joint molecules to form both COs and non-COs. These include MUS81-EME1 (refs 10,11), SLX1-SLX4 and GEN1 . MUS81-EME1 and SLX1-SLX4 act in the same pathway, which is distinct from that mediated by GEN1 .
It is thought that BTR/STR-mediated dissolution provides the primary pathway for the processing of HJs in mitotic cells since cells derived from individuals with Bloom's syndrome, that lack BLM activity, exhibit an elevated frequency of sister chromatid exchanges (SCEs) resulting from MUS81-EME1, SLX1-SLX4 and GEN1 mediated HJ resolution [bib_ref] Coordinated actions of SLX1-SLX4 and MUS81-EME1 for Holliday junction resolution in human..., Wyatt [/bib_ref] [bib_ref] Cooperative control of Holliday junction resolution and DNA repair by the SLX1..., Castor [/bib_ref] [bib_ref] Aberrant chromosome morphology in human cells defective for Holliday junction resolution, Wechsler [/bib_ref] [bib_ref] Roles of SLX1-SLX4, MUS81-EME1 and GEN1 in avoiding genome instability and mitotic..., Sarbajna [/bib_ref]. Depletion of the resolvases from either wild type or BLM-deficient cells results in DNA segregation defects [bib_ref] Coordinated actions of SLX1-SLX4 and MUS81-EME1 for Holliday junction resolution in human..., Wyatt [/bib_ref] [bib_ref] Cooperative control of Holliday junction resolution and DNA repair by the SLX1..., Castor [/bib_ref] [bib_ref] Human GEN1 and the SLX4-associated nucleases MUS81 and SLX1 are essential for..., Garner [/bib_ref] [bib_ref] Aberrant chromosome morphology in human cells defective for Holliday junction resolution, Wechsler [/bib_ref] [bib_ref] Roles of SLX1-SLX4, MUS81-EME1 and GEN1 in avoiding genome instability and mitotic..., Sarbajna [/bib_ref] , suggesting that the resolution pathways serve as a safeguard mechanism for the removal of persistent HJs that elude dissolution. The regulation of HJ dissolution/resolution is therefore important for the maintenance of genome stability, both by limiting CO formation and for ensuring proper chromosome segregation.
Recent studies have shown that the enzymatic activities and subcellular localization of the HJ resolvases are tightly controlled and modulated throughout the cell cycle to avoid competition with the dissolution pathway. For example, Saccharomyces cerevisiae Mms4 (the ortholog of EME1) is phosphorylated by Cdk and Cdc5 at the G2/M transition, leading to a boost of Mus81-Mms4 nuclease activity [bib_ref] Cell cycle-dependent regulation of the nuclease activity of Mus81-Eme1/Mms4, Gallo-Fernandez [/bib_ref] [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref] [bib_ref] Cell cycle kinases coordinate the resolution of recombination intermediates with chromosome segregation, Matos [/bib_ref] [bib_ref] Premature Cdk1/Cdc5/Mus81 pathway activation induces aberrant replication and deleterious crossover, Szakal [/bib_ref]. Similarly, in human cells, EME1 is phosphorylated by CDK/PLK1 at prometaphase, but in this case, the modification promotes interactions between MUS81-EME1 and SLX1-SLX4 leading to the formation of an SLX1-SLX4-MUS81-EME1 complex (SLX-MUS) that promotes HJ resolution [bib_ref] Coordinated actions of SLX1-SLX4 and MUS81-EME1 for Holliday junction resolution in human..., Wyatt [/bib_ref]. The yeast ortholog of GEN1, Yen1, is also regulated by phosphorylation: in S phase, the protein is heavily modified by Cdk-mediated phosphorylation, leading to decreased substrate binding, and additionally, the phosphorylation of S679, which overlaps the nuclear localization signal (NLS), leads to inhibition of its nuclear import [bib_ref] Systematic identification of cell cycle-dependent yeast nucleocytoplasmic shuttling proteins by prediction of..., Kosugi [/bib_ref] [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. Later in the cell cycle, these inhibitory phosphorylation events are reversed by the release/ activation of Cdc14 phosphatase, leading to Yen1's nuclear entry and enzymatic activation. This second boost in resolvase activity processes any remaining HJs and ensures DNA segregation [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. Importantly, the presence of nuclear and constitutively active Yen1 (Yen1 ON ) led to increased CO formation, indicating that prematurely activated Yen1 competes with the NCO pathway [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. The temporal and spatial restriction of Yen1 activity to the late stages of the cell cycle also appears to be important to protect against the cleavage of replication intermediates that arise during S phase, as premature activation of Yen1 increases the sensitivity of the cells to DNA-damaging agents [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref]. HJ resolution catalyzed by GEN1 is also thought to be restricted to the late stages of the cell cycle [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref] , but the underlying mechanism of regulation is unknown. Here, we show that GEN1 is regulated in a manner that is different from that of Yen1. We find that GEN1 contains a powerful nuclear export signale (NES) that mediates its nuclear exclusion such that its actions are restricted until mitosis when the nuclear membrane breaks down. Moreover, when GEN1 is artificially mislocalized to the nucleus an elevated frequency of COs is observed (as indicated by an increased incidence of SCEs), hence increasing the risk of loss of heterozygosity and genomic rearrangements.
# Results
CRM-mediated nuclear export regulates GEN1 localization. Previous studies have shown that GEN1 is cytoplasmic throughout interphase [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref] [bib_ref] A novel role of human Holliday junction resolvase GEN1 in the maintenance..., Gao [/bib_ref]. However, because the subcellular localization of many DNA repair-related proteins (e.g. p53, p38MAPK) is affected by DNA damage 30,31 , we used FLAP (GFP and FLAG)-tagged GEN1, expressed at endogenous levels from a bacterial artificial chromosome [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref] , to determine whether the localization of GEN1 is altered by treatments that either block replication fork progression (hydroxyurea) or introduce DNA lesions (camptothecin, cisplatin). In all cases, we found that GEN1 localized mainly to the cytoplasm [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref]. These results were confirmed by cellular fractionation, which led us to estimate that B80% of the GEN1 was present in the cytoplasmic fraction of both untreated and treated cells, with the remaining 20% in the nuclear fraction [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref]. These results indicate that the bulk of the GEN1 protein gains access to DNA only after the breakdown of the nuclear envelope at mitotic entry. In contrast, MUS81 localized mainly to the nucleus [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref] , indicating that the actions of the two resolvases are spatially distinct in interphase cells.
To determine whether GEN1 localization might be controlled by the nuclear export machinery, we used leptomycin B (LMB) to inhibit CRM1, a major nuclear export receptor [bib_ref] Nucleo-cytoplasmic transport of proteins as a target for therapeutic drugs, Yashiroda [/bib_ref] [bib_ref] CRM1-mediated nuclear export: to the pore and beyond, Hutten [/bib_ref]. After treatment with LMB, a proportion of the interphase cells (23.7 ± 1.5%, n4350 cells) exhibited a clear increase in the nuclear localization of GEN1 [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref]. As not all cells showed this increase in response to LMB treatment, we determined whether the effect was specific to the stage of the cell cycle by staining for cyclin A, which serves as a marker for S/G2 cells. We found that B47% of the G1 cells (that is cyclin A negative) showed nuclear accumulation of GEN1 on LMB treatment (n ¼ 236 cells). In contrast, none of the S/G2 cells responded to LMB treatment (n ¼ 169 cells). These results led us to conclude that the cells responding to LMB treatment and exhibiting nuclear GEN1 after LMB treatment were in G1 [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref]. As the nuclear envelope reforms at the end of mitosis, we speculated that nuclear export of GEN1 must occur at telophase, and that LMB might specifically block the transport of GEN1 to the cytoplasm at this stage of the cell cycle. This was indeed shown to be the case because, in the absence of LMB, GEN1 was excluded from the nucleus in all telophase cells examined (n ¼ 102 cells) as detected by a-tubulin staining of the mid-body, whereas almost all telophase cells exhibited nuclear staining for GEN1 following LMB treatment (109 out of 112 telophase cells) [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref].
Effect of phosphorylation on GEN1 activity and function. The cellular localization and substrate binding activities of yeast Yen1 are directly controlled by changes to its phosphorylation status [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. To determine whether GEN1 is also regulated by phosphorylation during the mitotic cell cycle, cells were synchronized at S phase (using a thymidine block) and at prometaphase (by nocodazole shake-off), respectively. Mitotic arrest was indicated by the accumulation of PLK1 and the phosphorylation of histone H3 at S10 . A phosphorylated form of GEN1 was detected at M phase, as determined by its reduced electrophoretic mobility on phos-tag gels , and by western blotting using an antibody specific for phospho-(Ser) CDK substrates . However, when GEN1 was affinity purified from the S-and M-phase extracts, and assayed for its ability to resolve HJs, we did not observe any great difference in the levels of activity, although slightly more activity was found in the M-phase samples (38% resolution versus 32%). Moreover, the activities were largely unaffected by treatment with l-phosphatase that promotes GEN1's dephosphorylation . The phosphorylation of GEN1 was reduced at mitotic exit, as indicated by its increased electrophoretic mobility and the disappearance of the phosphoserine signal , but its HJ resolving activity remained relatively constant during mitotic exit . Similarly, when cells were synchronized by a double thymidine block followed by release, we again found that GEN1 activity was relatively constant during cell cycle progression from G1/S phase to mitosis [fig_ref] Figure 1 |: Cellular localization of GEN1 [/fig_ref].
In yeast, Yen1 is phosphorylated by Cdk 27 . To determine whether CDK phosphorylation affects GEN1 activity, all putative CDK sites (S/TP) in GEN1 were mutated to alanine (S249A, S269A, S512A, T535A, S598A, S655A, S842A and S893A).
Previously, the phosphorylation of one of these CDK sites (S893) was identified in proteomic studies [bib_ref] Global detection of protein kinase D-dependent phosphorylation events in nocodazole-treated human cells, Franz-Wachtel [/bib_ref] [bib_ref] Quantitative phosphoproteomics identifies substrates and functional modules of Aurora and Polo-like kinase..., Kettenbach [/bib_ref]. The actions of the resulting phosphomutant, GEN1 8A , together with GEN1 as a control, were then analysed in a GEN1 knockout cell line (GEN1 À / À ) generated using the CRISPR-Cas9 system 36,37 that targeted exon 2 of GEN1 locus . The knockout was verified by both sequencing and western blotting . We found that GEN1 8A did not show a reduced electrophoretic mobility on nocodazole-induced mitotic block [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref] , indicating that the phosphorylation of GEN1 is mediated primarily by mitotic CDK. Furthermore, when GEN1 WT and GEN1 8A were affinity purified from the S-and M-phase extracts and assayed for activity [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref] , they exhibited similar levels of HJ resolution activity.
To determine whether CDK-mediated phosphorylation of GEN1 modulates its chromatin binding, cells were synchronized by overnight treatment of nocodazole followed by biochemical fractionation into separate soluble and chromatin fractions [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref]. We found that the same amounts of GEN1 WT and GEN1 8A were found in the chromatin fraction, arguing against the notion that phosphorylation regulates GEN1-chromatin interaction. Finally, we compared the SCE levels of cells expressing GEN1 WT and GEN1 8A . As expected, GEN1 À / À cells
[formula] Untreated Untreated GEN1-GFP MUS81 α-Tubulin Histone H3 C y t o N u c C y t o N u c C y t o N u c C y t o N u c GEN1-GFP GEN1-GFP GEN1-GFP α-tubulinT h y N o c GEN1-GFP (phos-tag) GEN1-GFP (phos-tag) λ PPase (95°C) λ PPase GEN1-GFP (phos-tag) p-Ser (CDK) (GEN1-GFP) p-Ser (CDK) (GEN1-GFP) PLK1 PLK1 [/formula]
H3 (p-S10) H3 (p-S10)
α-Tubulin α-Tubulin | GEN1 activity is independent of its phosphorylation status. (a,b) Extracts were prepared from HeLa cells expressing GEN1-FLAP after synchronization with thymidine (Thy) or nocodazole (Noc), as indicated. Cell extracts were analysed by western blotting for the indicated proteins (a). GEN1-FLAP was affinity purified from each sample by GFP pull down, treated with active or heat-inactivated l-phosphatase (95°C), and assayed for HJ resolution activity (b). Phosphorylated GEN1 was detected using an anti-phosphoserine antibody (p-Ser) that is specific for CDK substrates.
(c,d) Extracts were prepared from HeLa cells expressing GEN1-FLAP after release from nocodazole arrest at the indicated time points. Cell extracts were analysed by western blotting for the indicated proteins (c). GEN1-FLAP was affinity purified from each sample by GFP pull down and assayed for HJ resolution activity (d).
--- showed a significant reduction of cisplatin-induced SCEs when compared with the parental wild-type cells [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref] , but this reduction was fully rescued by both GEN1 WT and GEN1 8A [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref]. Taken together, these results show that the temporal restriction of GEN1 function is achieved primarily through CRM1-mediated nuclear export, rather than by the modulation of GEN1 activity by changes to its phosphorylation status. These results are in stark contrast to the mechanisms by which Yen1 is regulated in yeast [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref].
[formula] - - - + + + + Thy 293 GEN1 -/- +WT GEN1 -/- +8A GEN1 -/-+ GEN1 WT GEN1 -/-GEN1 -/- +WT GEN1 -/- +8A GEN1 -/-+ GEN1 [/formula]
Identification of a nuclear export signal in GEN1. CRM1 mediates nuclear export by interacting with the NES of its cargo proteins [bib_ref] CRM1-mediated nuclear export: to the pore and beyond, Hutten [/bib_ref] , leading us to attempt to map the location of a NES in GEN1. To do this, a series of GEN1 truncations, fused in-frame with GFP, were generated [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref] and the cellular localization of each expressed protein was determined [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. The small size of each truncation protein was expected to allow nuclear entry by passive diffusion through nuclear pores [bib_ref] Protein import into the cell nucleus, Dingwall [/bib_ref] , such that the NEScontaining fragment of GEN1 could be identified by its nuclear exclusion. As observed with GFP alone, most of the truncations localized both in the cytoplasm and the nucleus, except for GEN1 526-683 -GFP that displayed a clear cytoplasmic localization [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. Examination of the sequence of this construct revealed a leucine-rich sequence (660-LLSGITDLCL-669) [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref] , conforming to the NES consensus sequence (L- [bib_ref] NESbase version 1.0: a database of nuclear export signals, La Cour [/bib_ref].
[formula] x(2-3)-[LIVFM]- x(2-3)-L-x-[LI]) [/formula]
When cells expressing GEN1 526-683 were treated with LMB we found that nuclear exclusion of the polypeptide was inhibited [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. Similarly, mutation of four leucine/isoleucine residues within the presumed NES consensus (4A: L660A, L661A, I664A and L667A) abolished the nuclear exclusion of GEN1 526-683 , confirming the presence of a functional NES [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. Interestingly, attempts to overcome the NES by generating a GEN1 fusion protein containing three nuclear localization signals (3xNLS) were insufficient to induce its relocalization to the nucleus [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. However, we were successful in driving the nuclear localization of GEN1 by adding three NLS sequences and by mutating the four leucine/isoleucine residues in the NES [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. In this case, GEN1 was found to be constitutively nuclear, and in the following work, we will refer to this as GEN1 nuc .
Nuclear localization of GEN1 promotes crossover formation. In contrast to yeast Yen1, GEN1 nuclease activity does not appear to be regulated by phosphorylation during the cell cycle, leading us to reason that forced nuclear import of GEN1 would lead to genome instability due to its premature action. To study the effect of GEN1 deregulation, we generated stable HEK293 cells expressing wild-type GEN1 (GEN1 WT ) or GEN1 nuc under the control of a tetracycline-inducible promoter. Both proteins carried carboxy-terminal 3xFLAG tags that allowed us to visualize GEN1 expression by immunostaining. As expected, GEN1 nuc was enriched in the nucleus, whereas GEN1 WT was mainly cytoplasmic [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref]. To determine whether nuclear GEN1 was chromatin bound, cells were arrested in S phase using hydroxyurea followed by biochemical fractionation. We found that a significant amount of the GEN1 nuc was associated with the chromatin fraction in these S-phase cells [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref]. Importantly, affinity purified GEN1 nuc from these extracts displayed a similar HJ resolvase activity as GEN1 WT , showing that inclusion of the 3xNLS sequence and mutation of the NES did not affect protein activity [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref].
To determine whether expression of GEN1 nuc leads to an increased sensitivity to genotoxic stress, cells expressing GEN1 WT or GEN1 nuc were treated with various concentrations of hydroxyurea, camptothecin or cisplatin and analysed for colony formation. We found that GEN1 nuc did not affect cell survival [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref]. Moreover, expression of GEN1 nuc failed to induce an increase in the levels of phosphorylated H2AX (gH2AX) nor affect the DNA damage checkpoint response, measured by the phosphorylation of CHK1 at S345 and KAP-1 at S842, either in untreated or damaged cells [fig_ref] Figure 3 |: Functional activity of a GEN1 phosphomutant [/fig_ref]. These results indicate that premature nuclear import of GEN1 is not particularly harmful to the cells.
One predicted effect of GEN1 nuc is that it would be expected to compete with BTR-mediated HJ dissolution and thereby promote mitotic CO formation. We therefore examined SCE levels in both untreated cells [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref] and those treated briefly with cisplatin to elevate the level of recombination intermediates [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref]. In both cases, cells expressing GEN1 nuc exhibited significantly increased SCE formation, compared with control cells or cells expressing GEN1 WT [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref] ; [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref] shows the expression levels of GEN1). As controls, cells that were not treated with tetracycline (thus no GEN1 induction) showed no significant difference in SCE levels [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref]. Taken together, these results show that restriction of GEN1 action (by nuclear exclusion) to the late stages of the cell cycle is critical for preventing mitotic CO formation.
GEN1 nuc suppresses BLM and MUS81 depletion phenotypes. Persistent unresolved HJs induce various cellular defects including chromosome breakage and aberrant chromosome segregation [bib_ref] Coordinated actions of SLX1-SLX4 and MUS81-EME1 for Holliday junction resolution in human..., Wyatt [/bib_ref] [bib_ref] Human GEN1 and the SLX4-associated nucleases MUS81 and SLX1 are essential for..., Garner [/bib_ref] [bib_ref] Aberrant chromosome morphology in human cells defective for Holliday junction resolution, Wechsler [/bib_ref] [bib_ref] Roles of SLX1-SLX4, MUS81-EME1 and GEN1 in avoiding genome instability and mitotic..., Sarbajna [/bib_ref]. We therefore determined whether expression of nuclear GEN1 would suppress defects in other HJ processing pathways. To do this, BLM, MUS81 or BLM and MUS81 were small interfering RNA (siRNA)-depleted from cells expressing GEN1 WT or GEN1 nuc from a tetracycline-inducible promoter. All proteins were depleted with a high efficiency [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref] , and GEN1 WT or GEN1 nuc were induced at comparable levels in the siRNAs-treated cells [fig_ref] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal [/fig_ref].
First, we determined the effect of GEN1 WT or GEN1 nuc expression on cell viability using clonogenic survival assays. As expected, cells depleted of both MUS81 and BLM exhibited a significant loss of viability, and this was partially rescued by expression of GEN1 nuc but not GEN1 WT [fig_ref] Figure 6 |: Rescue of BLM/MUS81 defects by expression of GEN1 nuc [/fig_ref]. Second, we determined whether GEN1 nuc could suppress the high levels of chromosome breakage seen in metaphase spreads from BLM and MUS81 co-depleted cells [bib_ref] Human GEN1 and the SLX4-associated nucleases MUS81 and SLX1 are essential for..., Garner [/bib_ref] [bib_ref] Roles of SLX1-SLX4, MUS81-EME1 and GEN1 in avoiding genome instability and mitotic..., Sarbajna [/bib_ref]. In these experiments, we scored several types of chromosome aberrations, including chromatid breaks (only one arm of the chromosome is broken), isochromatid breaks (both arms are broken) and radial chromosomes [fig_ref] Figure 6 |: Rescue of BLM/MUS81 defects by expression of GEN1 nuc [/fig_ref]. Importantly, we found that expression of GEN1 nuc , but not GEN1 WT , rescued the increased frequencies of chromatid and isochromatid breaks [fig_ref] Figure 6 |: Rescue of BLM/MUS81 defects by expression of GEN1 nuc [/fig_ref] , upper and center panels). In contrast, there was little effect on the frequency of radial chromosomes [fig_ref] Figure 6 |: Rescue of BLM/MUS81 defects by expression of GEN1 nuc [/fig_ref] , indicating that these do not arise from a loss of HJ processing activity. Together, these results show that the premature actions of nuclear GEN1 lead to the removal of recombinant intermediates that are normally processed earlier in the cell cycle, either by the BTR complex or by MUS81-EME1.
# Discussion
In this work, we have shown that the activity of the GEN1 Holliday junction resolvase is restricted to mitosis by nuclear exclusion. The mechanism of GEN1 regulation contrasts with that observed in S. cerevisiae, where the GEN1 ortholog Yen1 undergoes a dual mechanism of regulation [bib_ref] Systematic identification of cell cycle-dependent yeast nucleocytoplasmic shuttling proteins by prediction of..., Kosugi [/bib_ref] [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. In yeast, Cdk-mediated phosphorylation of Yen1, which takes place during S phase, promotes its nuclear exclusion and inhibits its catalytic activity. Nuclear exclusion of Yen1 is achieved by: (i) impairing its nuclear import by phosphorylation of S679, which overlaps the NLS, and (ii) Msn5-mediated nuclear export.
S-phase phosphorylation also directly inhibits the nuclease activity of Yen1 by reducing its DNA binding affinity [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref]. At the onset of anaphase, however, Cdc14 phosphatase interacts with Yen1 and promotes its dephosphorylation, triggering both its nuclear import and an increase of its DNA binding affinity. In human cells, the action of GEN1 is also restricted in mitosis to avoid competition with the BTR-mediated NCO pathway, although the underlining control mechanism is distinct from that of Yen1. We find that GEN1 was phosphorylated at M phase primarily by CDK. However, the phosphorylation status of GEN1 did not appear to have a significant effect on its nuclease activity or ability to bind chromatin. Moreover, we found that the nonphosphorylatable mutant of GEN1 (GEN1 8A ) was equally active as the wild type in complementing GEN1 knockout cells in terms of SCE formation. The nuclear exclusion of GEN1 appears to be mainly responsible for regulating the activity of GEN1, and is mediated by a NES located within the C-terminal region of the protein. On entry into mitosis, breakdown of the nuclear envelope allows the binding of GEN1 to chromatin. Finally, at mitosis, when the nuclear envelope reforms, GEN1 again becomes excluded from nucleus. A comparison of the mechanisms of regulation of Yen1 and GEN1 is shown in [fig_ref] Figure 7 |: Distinct mechanisms of regulatory control for S [/fig_ref]. Why have lower and higher eukaryotes evolved different mechanisms of Yen1/GEN1 regulation? One explanation is that yeast undergoes a 'closed' mitosis in which the nuclear envelope remains intact. As a consequence, active import machineries have to be employed to shuttle Yen1 into the nucleus. In contrast, in human cells, the physical barrier is automatically disassembled at the G2/M transition, so that cells only need to maintain GEN1 export to prevent its accumulation in the nucleus. Also, again in contrast to the situation in yeast, the human CDC14 orthologs ARTICLE CDC14A and CDC14B do not appear to play essential roles in the reversal of mitotic CDK-mediated phosphorylation to control mitotic exit [bib_ref] Vertebrate cells genetically deficient for CDC14A or CDC14B retain DNA damage checkpoint..., Mocciaro [/bib_ref] [bib_ref] The nucleolar phosphatase CDC14B is dispensable for chromosome segregation and mitotic exit..., Berdougo [/bib_ref]. The consequences of the nuclear presence of deregulated Yen1/GEN1 have been revealed in studies with both Yen1 ON and GEN1 nuc , as summarized in [fig_ref] Figure 7 |: Distinct mechanisms of regulatory control for S [/fig_ref]. Expression of Yen1 ON (in which all nine Cdk consensus sites were mutated to alanine) and GEN1 nuc led to an increased CO frequency, demonstrating that deregulated Yen1/GEN1 competes with the non-crossover repair pathway [bib_ref] Dual control of Yen1 nuclease activity and cellular localization by Cdk and..., Blanco [/bib_ref] [bib_ref] The Cdk/Cdc14 module controls activation of the Yen1 Holliday junction resolvase to..., Eissler [/bib_ref]. However, while expression of Yen1 ON caused a cellular hypersensitivity to genotoxic stress 27 , we did not observe increased DNA damage sensitivity or a checkpoint response in cells expressing GEN1 nuc . One possibility that may account for this difference is the broader substrate specificity exhibited by Yen1 in comparison with GEN1 (M. G. Blanco, Y. W. C. and S. C. W., unpublished data). Indeed, the actions of deregulated/nuclear Yen1 may damage DNA intermediates formed during replication fork regression, leading to a hypersensitivity to DNA-damaging agents that promote replication fork stalling. Further studies of the biochemical properties of Yen1 and GEN1 will be required before we gain a more complete understanding of the actions of these flap/fork endonucleases on recombination/replication intermediates.
Our work shows that mislocalization of GEN1 to the nucleus increases CO formation that may in turn promote loss of heterozygosity. This raises the possibility that uncontrolled GEN1 activity could be linked to tumorigenesis. Interestingly, using plasmid-based systems, it was recently shown that GEN1 can act on palindromic AT-rich repeats identified on human chromosomes 11 and 22, that extrude to form cruciform-like structures, giving rise to translocation-like rearrangements [bib_ref] Chromosomal instability mediated by non-B DNA: cruciform conformation and not DNA sequence..., Inagaki [/bib_ref] [bib_ref] Two sequential cleavage reactions on cruciform DNA structures cause palindrome-mediated chromosomal translocations, Inagaki [/bib_ref]. Carriers of chromosomal translocations t(11,22)(q23;q11), caused by those palindromic AT-rich repeats, may give birth to offspring manifesting Emanuel syndrome [bib_ref] Regions of genomic instability on 22q11 and 11q23 as the etiology for..., Kurahashi [/bib_ref]. However, at the present time, no cancer/disease-associated mutations have been identified in the NES region of GEN1, so it is not presently clear whether GEN1 misregulation is a driver for disease development.
Taken together, our work provides the first insights into the spatial regulation of GEN1, which serves at least two purposes: First, to ensure the removal of persistent HJs that have escaped the dissolution pathway before cell division and thereby safeguard chromosome segregation. This may be particularly important when cells are exposed to DNA-damaging agents, or replication stress, which leads to an increase in recombination events that could 'overload' the BTR complex. Second, to prevent competition with the dissolution pathway mediated by the BTR complex, therefore avoiding mitotic CO formation that can lead to loss of heterozygosity and genomic rearrangements.
# Methods
Plasmids. The GEN1 truncation constructs (full-length GEN1, GEN1 1-220 , GEN1 221-444 , GEN1 444-527 , GEN1 526-662 , GEN1 526-683 and GEN1 662-908 ) were generated by PCR and cloned in-frame into a modified pDONR221 vector (Life Technologies) encoding a C-terminal GFP tag using an In-Fusion HD cloning kit (Clontech). The coding sequences were then shuttled by Gateway recombination (Life Technologies) into a modified pcDNA5/FRT vector (Life Technologies) with a Gateway cassette RfB cloned into the EcoRV site. To generate GEN1 nuc , a 3xNLS sequence (DPKKKRKVDPKKKRKVDPKKKRKV) and a 3xFLAG tag sequence (AGDYKDHDGDYKDHIDYKDDDDK) were added to the C terminus of GEN1 by PCR. Cell culture and transfection. HeLa Kyoto cells, HeLa Kyoto cells expressing GEN1-FLAP from a BAC [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref] and Flp-In T-REx 293 cells were cultured at 37°C in 5% CO 2 in DMEM (Life Technologies) supplemented with 10% fetal bovine serum (Life Technologies) and penicillin/streptomycin. Hygromycin (200 mg ml À 1 ), zeocin (50 mg ml À 1 ) and blasticidin (4 mg ml À 1 ) were obtained from Life Technologies. Nocodazole (100 ng ml À 1 ), thymidine (2 mM), cisplatin, hydroxyurea, camptothecin and BrdU (100 mM) were obtained from Sigma-Aldrich. Colcemid (0.2 mg ml À 1 ) was obtained from Roche. Leptomycin B (25 nM) was obtained from Santa Cruz. HeLa cells were transfected with plasmids encoding various GEN1 truncations for 30 h using TransIT-2020 (Mirus Bio LLC).
To generate GEN1-expressing stable cell lines, Flp-In T-REx 293 cells were co-transfected with plasmids encoding the GEN1 proteins and the Flp recombinase encoding plasmid pOG44 (in 1:9 ratio) using FuGENE (Promega). Hygromycinresistant colonies were picked and expanded. The expression of GEN1 WT and GEN1 8A were induced with 0.1 mg ml À 1 tetracycline (Sigma-Aldrich). The expression of GEN1 nuc was induced with 1 mg ml À 1 tetracycline.
To generate the GEN1 À / À cell line, Flp-In T-REx 293 cells were transfected with pX330 cloned with the GEN1 targeting sequence, along with pSuper.puro (Oligoengine) at a 9:1 ratio. After 24-48 h, cells were selected with 5 ug ml À 1 puromycin, and seeded as single colonies. Clones were picked and first selected on the basis of a negative signal when western blotted for GEN1. Selected clones were then verified by sequencing. To do this, genomic DNA was extracted from cells with DNeasy Blood & Tissue kit (Qiagen), and PCR was carried out with KOD Hot Start DNA Polymerase (Novagen) to amplify the targeted locus using a forward (5 0 -TGGCTTATAATATATTGTTTG-3 0 ) and a reverse (5 0 -GCTTTTAGTATCTG AAGCATC-3 0 ) primer. The PCR product was then purified by agarose gel electrophoresis followed by gel purification with QIAquick Gel Extraction kit (Qiagen) and finally sequenced.
Lysate preparation and HJ resolution assay. Cell lysates were prepared by suspending cells in lysis buffer (50 mM Hepes pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 10% glycerol, 1 mM DTT) supplemented with protease and phosphatase inhibitors. The lysates were passed through a 0.8 Â 40 mm needle (20 Â ) followed by incubation on ice for 30 min. The lysates were then cleared by centrifugation (14,000 r.p.m., 30 min). For HJ resolution assays, tagged GEN1 was affinity purified from the cleared lysates using anti-FLAG M2 agarose beads (Sigma-Aldrich) or GFP-Trap beads (Chromotek). The beads were washed extensively with lysis buffer and then mixed with 10 ml of cleavage buffer (50 mM Tris-HCl pH 7.5, 1 mM MgCl 2 , 1 mM DTT) and B1 nM 5 0 -32 P-end-labelled HJ X0 DNA [bib_ref] Identification of Holliday junction resolvases from humans and yeast, Ip [/bib_ref]. After 10 min of incubation at 37°C, DNA products were deproteinized for 45 min at 37°C by addition of 2.5 ml stop buffer (100 mM Tris-HCl pH 7.5, 50 mM EDTA, 2.5% SDS and 10 mg ml À 1 proteinase K). Loading buffer was added and radiolabelled products were separated by 10% neutral polyacrylamide gel electrophoresis, and analysed by autoradiography or by phosphorimaging using a Typhoon scanner and ImageQuant software (GE Healthcare).
Dephosphorylation assay. GEN1-FLAP was affinity purified with GFP-Trap beads. The beads were then incubated with l-phosphatase (NEB) or l-phosphatase that had been inactivated by heating for 10 min at 95°C. The reaction mixtures were then incubated at 30°C for 20 min before washing extensively with lysis buffer and analysed by western blotting and for HJ resolution.
Western blotting. Western blotting was carried out as described [bib_ref] Regulatory control of the resolution of DNA recombination intermediates during meiosis and..., Matos [/bib_ref]. Phos-tag SDS-polyacrylamide gel electrophoresis gels were made from 7% acrylamide (37.5:1 crosslinking ratio), 70 mM Phos-tag reagent (Wako Pure Chemical Industries) and 140 mM MnCl 2 , and were processed according to the manufacturer's manual. Proteins were detected using the following primary antibodies: mouse anti-GFP (1:1,000, Roche 11814460001), mouse anti-PLK1 (1:2,000, Santa Cruz sc-17783), mouse anti-histone H3 pSer10 (1:5000, Abcam 14955), mouse anti-a-tubulin (1:5000, Sigma-Aldrich T9026), rabbit antiphospho-(Ser) CDK substrate (1:1000, Cell Signaling 2324), rabbit anti-histone H3 (1:2000, Abcam ab1791), mouse anti-FLAG HRP (1:1,000, Sigma-Aldrich A8592), mouse anti-MUS81 (1:1,000, Santa Cruz sc-47692), mouse anti-BLM (1:500, Santa Cruz sc-70426), rabbit anti-KAP-1 pSer842 (1:1000, Abcam ab70369), rabbit anti-CHK1 pSer345 (1:1000, Cell Signaling 2341), mouse anti-histone H2A.X pSer139 (1:1000, Millipore 05-636-1), rabbit anti-GEN1 (1:100, raised against GEN1 890-908 ) [bib_ref] Mechanism of Holliday junction resolution by the human GEN1 protein, Rass [/bib_ref]. Uncropped blots are shown in [fig_ref] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation [/fig_ref].
Immunofluorescence. HeLa Kyoto cells or Flp-In T-REx 293 cells were fixed and processed for immunofluorescence as described [bib_ref] FANCM connects the genome instability disorders Bloom's syndrome and Fanconi anemia, Deans [/bib_ref]. In brief, cells were fixed in 4% paraformaldehyde for 10 min at the room temperature followed by ice-cold methanol for 1 min. The cells were then wash with 1  phosphate-buffered saline (PBS) and permeabilized with 0.5% Triton X-100 for 5 min before antibody staining. The primary antibodies used were as follows: mouse anti-histone H2AX pSer139 (1:1,000, Millipore 05-636-1), mouse anti-a-tubulin (1:5,000, Sigma-Aldrich T9026), rabbit anti-GFP (1:5,000, Abcam ab290), mouse anti-cyclin A (1:200, Santa Cruz sc-56299), mouse anti-FLAG (1:1,000, Sigma-Aldrich A8592). Secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 555 (1:1,000, Molecular Probes) were used for detection. DNA was stained with 4,6-diamidino-2-phenylindole. Images were acquired using Zeiss AXIO Imager M1 with a  40 EC-Plan-Neofluor lens and Hamamatsu photonics camera under the control of Volocity software. Images were processed using Adobe Photoshop. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5844 ARTICLE fraction and nuclear pellet were collected by centrifugation at 1,300 g for 4 min. The cytoplasmic fraction was spun at 20,000 g to remove cell debris and insoluble aggregates. Isolated nuclei were then washed in hypotonic buffer, lysed in solution B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT and protease and phosphatase inhibitors), and incubated on ice for 10 min. The soluble nuclear fraction and chromatin pellet were collected by centrifugation at 1,700 g for 4 min. The chromatin pellet was then washed in hypotonic buffer, spun down at 10,000 g and resuspended in nuclear extraction buffer (50 mM Tris-HCl pH 7.5, 420 mM NaCl, 10% glycerol, 1 mM EDTA, 1 mM DTT, 0.01% NP40, and protease and phosphatase inhibitors) and placed on a rotator at 4°C for 45 min. 1 mM CaCl 2 and 2 ml (0.4 U) of micrococcal nuclease (Sigma-Aldrich) were added and the mixture was incubated at 37°C for 5 min with gentle shaking. EGTA (1 mM) was then added to inhibit the nuclease and the solubilized chromatin was collected by centrifugation at 10,000 g for 10 min.
siRNA treatment. The control (5 0 -UAAUGUAUUGGAACGCAUA-3 0 ) and BLM (5 0 -GGAGCACAUCUGUAAAUUA-3 0 ) siRNAs [bib_ref] PICH and BLM limit histone association with anaphase centromeric threads and promote..., Ke [/bib_ref] were purchased from Eurofins, and the MUS81 siRNAs 1-2 (5 0 -CAGCCCUGGUGGAUCGAUA-3 0 and 5 0 -CAU UAAGUGUGGGCGUCUA-3 0 ) (ref. [bib_ref] Aberrant chromosome morphology in human cells defective for Holliday junction resolution, Wechsler [/bib_ref] were purchased from Dharmacon. One day before transfection, 6 Â 10 5 cells were seeded in 6-cm cell culture plates. In all experiments, cells were transfected with 50 nM of siRNA twice (24 h apart) using Lipofectamine RNAiMAX (Life Technologies). Proteins were analysed by western blotting 48 h after the second transfection.
Clonogenic survival assay. For drug sensitivity tests, B600 cells were seeded in six-well plates. After 2 days, cells were untreated or treated with hydroxyurea, camptothecin or cisplatin for 24 h (three wells per condition), washed with PBS and maintained in fresh media for 12-14 days to allow colony formation. For survival assays after siRNA treatment, cells were transfected with siRNA twice in 6-cm plates, and 24 h after the second transfection, the cells were collected and re-plated in 6-well plates (B600 cells per well, three wells per condition) and incubated for 12-14 days to allow colony formation. Colonies were stained for B5 min with 40 mg ml À 1 crystal violet solution (Sigma-Aldrich) containing 20% ethanol. The percent survival was calculated against untreated cells or the control siRNA sample. SCE analysis. Cells expressing GEN1 (tetracycline induction, 48 h) were treated with or without cisplatin (2 mM) for 1 h. The culture media was then changed and BrdU (100 mM) was added. Cells were grown for further 48 h and treated with Colcemid (0.2 mg ml À 1 ) for 1 h before being collected. Metaphase chromosomes were prepared and assayed for SCEs as previously described [bib_ref] Aberrant chromosome morphology in human cells defective for Holliday junction resolution, Wechsler [/bib_ref].
Analysis of chromosome aberrations. To analyse chromosome aberrations, GEN1 was expressed from a tetracycline-inducible promoter, and the indicated proteins were knocked down by siRNA treatment (2 Â , 24 h apart). At a time (8 h) after the second siRNA transfection, cells were treated with cisplatin (2 mM) for 1 h. The culture media were changed and cells were grown for further 48 h, before being treated with Colcemid (0.2 mg ml À 1 ) for 1 h and then collected. Metaphase chromosomes were prepared as above.
[fig] Figure 1 |: Cellular localization of GEN1. (a) HeLa cells expressing GEN1-FLAP were treated with or without hydroxyurea (HU, 2 mM), camptothecin (CPT, 20 nM) or cisplatin (Cis-Pt, 2 mM) for 16 h. GEN1, gH2AX and DNA were visualized using an anti-GFP antibody (green), an anti-gH2AX antibody (red) and DAPI staining (blue), respectively. (b) Cell fractionation to determine the localization of GEN1. Fractions were analysed by western blotting for the indicated proteins. (c) HeLa cells expressing GEN1-FLAP were treated with or without 25 nM LMB for 4 h. GEN1 and DNA were visualized as in a.(d) As c, except the LMB-treated cells were stained for Cyclin A (red) to identify S-and G2-phase cells. (e) As a, with a-tubulin staining decorating the mid-body and marking late mitotic cells. Scale bars, 10 mm. [/fig]
[fig] Figure 3 |: Functional activity of a GEN1 phosphomutant. (a) Extracts were prepared from GEN1 À / À Flp-In T-REx 293 cells expressing GEN1 WT or GEN1 8A after synchronization with Thymidine (Thy) or Nocodazole (Noc) as indicated. Cell extracts were analysed by western blotting for the indicated proteins. GEN1-FLAG was affinity purified from each sample using anti-FLAG M2 agarose beads and assayed for HJ resolution activity. Parental Flp-In T-Rex 293 cells (293) were used as a control. (b) Biochemical fractionation of GEN1 À / À cells expressing GEN1 WT or GEN1 8A after overnight synchronization with Nocodazole. The whole-cell extract (total), soluble and chromatin (Chr) fractions were analysed by western blotting for the indicated proteins. (c) Flp-In T-Rex 293 cells, GEN1 À / À cells and GEN1 À / À cells expressing GEN1 WT or GEN1 8A were treated with cisplatin (2 mM) for 1 h. SCE formation was quantified as described. Forty metaphase cells (41,600 chromosomes) were counted per condition. Black bars represent the mean number of SCEs per 100 chromosomes per spread. P values were determined using a two-tailed t-test. NS: not significant. [/fig]
[fig] Figure 4 |: Cytoplasmic localization of GEN1 is directed by a nuclear export signal. (a) Schematic representation of a series of GEN1 truncations. The green hexagon, the orange square and the blue ellipse represent the XPG-N, XPG-I and the helix-hairpin-helix domains, respectively. GFP signifies the C-terminal tag. (b) HeLa cells, transfected with the GEN1-GFP tagged constructs, or GFP alone, were visualized by immunofluorescence. GEN1-GFP and DNA were visualized as inFig. 1. (c)Schematic representation of the putative nuclear export signal (NES) of GEN1, which is conserved between various mammalian species. (d) HeLa cells, transfected with GEN1 526-683 -GFP or GEN1 526-683(4A) -GFP, were treated with or without LMB. GEN1 and DNA were visualized as above. (e) Schematic representation of GEN1 WT , GEN1 with 3xNLS sequences (GEN1 3xNLS ), GEN1 with mutated NES (GEN1 NES(4A) ) and GEN1 with a mutated NES and 3xNLS sequences (GEN1 nuc ). (f) HeLa cells, transfected with the GEN1 constructs shown in e, were analysed for GEN1-FLAG and DNA using an anti-FLAG antibody and DAPI staining, respectively. Scale bars, 10 mm. [/fig]
[fig] Figure 5 |: Expression of nuclear GEN1 leads to elevated SCE formation. (a) Flp-In T-REx 293 cells stably expressing GEN1 WT or GEN1 nuc were treated with or without tetracycline (Tet) for 48 h. GEN1 and DNA were visualized using an anti-FLAG antibody and DAPI staining, respectively. Scale bar, 10 mm.(b) Biochemical fractionation of cells expressing GEN1 WT or GEN1 nuc after overnight synchronization with hydroxyurea (2 mM). The fractions were analysed by western blotting for the indicated proteins. (c) Representative images of metaphase spreads from parental Flp-In T-REx 293 cells (control), and cells expressing GEN1 WT or GEN1 nuc . (d) Quantification of SCE formation, as determined in c. Each data point represents a single metaphase. 60 metaphase cells (42,600 chromosomes) were counted per condition. Black bars represent the mean number of SCEs per 100 chromosomes per spread. P values were determined using a two-tailed t-test. (e) As c, except the cells were treated with cisplatin (2 mM). (f) Quantification of SCE formation, as in e. [/fig]
[fig] Figure 6 |: Rescue of BLM/MUS81 defects by expression of GEN1 nuc . (a) Clonogenic cell survival assays were carried out on Flp-In T-REx 293 cells expressing GEN1 WT or GEN1 nuc (±Tet) and depleted for the indicated proteins. Cell survival after treatment with the control siRNA is defined as 100%. The data represent the mean ± s.d. of at least three independent experiments. P values were determined using a two-tailed t-test. (b) Representative images of metaphase spreads of Flp-In T-REx 293 cells treated with the indicated siRNAs. Three types of chromosome aberrations (chromatid breaks, isochromatid breaks and radial chromosomes) were identified and quantified. (c) Flp-In T-REx 293 cells expressing GEN1 WT or GEN1 nuc ( ± Tet) were treated with the indicated siRNAs, and metaphase spreads were analysed and quantified for the indicated chromosome aberrations. One-hundred twenty metaphase cells (45,500 chromosomes) were analysed. The data represent the mean number of aberrations per 100 chromosomes per spread ± s.e.m. P values were determined using a two-tailed t-test. [/fig]
[fig] Figure 7 |: Distinct mechanisms of regulatory control for S. cerevisiae Yen1 and human GEN1. (a) Regulatory control mechanisms (see text for details). (b) Phenotypic effects of expressing Yen1 ON and GEN1 nuc (see text for details). [/fig]
|
Cancer awareness among community pharmacist: a systematic review
Background: The WHO recognises that community pharmacists are the most accessible healthcare professionals to the general public. Most patients regularly visit community pharmacies for health information and also seek advice from pharmacists with respect to signs and symptoms of cancer. As readily accessible health care professionals, community pharmacists are also in the best position to include cancer-screening initiatives into their practice. Pharmacists are therefore in a good position to raise awareness when they counsel people who buy over-thecounter medication for the control of possible cancer-related symptoms. The aim of this review was to critically appraise evidence gathered from studies that; (1) explore or assess knowledge of community pharmacist on signs and symptoms of cancer, (2) explore or assess knowledge of community pharmacist on cancer screening. Methods: EMBASE (ovid), CINAHL (EBSCOhost) and MEDLINE (EBSCOhost) were systematically searched for studies conducted between 2005 to July 2017. Studies that focused on knowledge of community pharmacist in cancer screening, signs and symptoms were included. Results: A total of 1538 articles were identified from the search, of which 4 out of the 28 potentially relevant abstracts were included in the review. Findings of the selected studies revealed lack of sufficient knowledge on breast cancer screening, signs and symptoms. Both studies attributed knowledge limitation as the cause of reason for the key findings of their studies. Conclusion: The selected studies focused largely on breast cancer, which hinder the generalizability and transferability of the findings. Hence there is a need for more studies to be conducted in this area to draw a better conclusion.
# Background
Cancer now causes more deaths than all coronary heart disease or strokes, according to World Health Organisation (WHO) estimates for 2011. This is likely as a result of late presentation of the disease [bib_ref] Asian Cancer registry forum 2014-regional cooperation for cancer registration: priorities and challenges, Moore [/bib_ref] which have been attributed to a number of factors such as poor awareness of the signs and symptoms of cancer, cancer risk factors, poor availability of tests or screening programs [bib_ref] Do not let to be late: overview of reasons for melanoma delayed..., Gajda [/bib_ref] [bib_ref] Delays in diagnosis and treatment of breast cancer: a multinational analysis, Jassem [/bib_ref] [bib_ref] Global cancer statistics, Jemal [/bib_ref].
The continuous global demographic and epidemiological evolution shows an increasing cancer burden over the next decades, especially in low and middle income countries (LMIC), with over 20 million new cancer cases expected annually as early as 2025 [bib_ref] Transitions in human development and the global cancer burden, Bray [/bib_ref].
Contemporary pharmacy practice reflects an emerging paradigm from one in which the pharmacist primarily supervises medication distribution and counsels patients, to a more expanded role providing patient-centered medication therapy management, health improvement, health education, health promotion activities and disease prevention services [bib_ref] Prescribing medications: changing the paradigm for a changing health care system, Webb [/bib_ref]. The role of the pharmacist in cancer care is now growing with community pharmacists advocating, promoting, supporting and providing cancer related health promotion.
The WHO recognises that community pharmacists are the most accessible healthcare professionals to the general public. Studies have shown that community pharmacies provide easy and equitable access to healthcare [bib_ref] Access all areas? An area-level analysis of accessibility to general practice and..., Todd [/bib_ref]. Most patients regularly visit community pharmacies for health information and also seek advice from pharmacists with respect to signs and symptoms of cancer [bib_ref] Cancer detection and the community pharmacist, Lum [/bib_ref]. Pharmacists are therefore in a good position to raise awareness when they counsel people who buy over-the-counter medication for the control of possible cancer-related symptoms. To be able to achieve this, as healthcare providers in the community, pharmacist must be able to differentiate between conditions that require selfmedication and those that need the attention of a physician. They must be able to identify the common signs and symptoms of cancer. As readily accessible health care professionals, community pharmacists are in the best position to include cancer-screening initiatives into their practice. A number of organizations including the US Preventive Services Task Force, American Cancer Society [bib_ref] Cancer screening in the United States, 2010: a review of current American..., Smith [/bib_ref] , and National Comprehensive Cancer Network, have developed cancer screening recommendations. Because clinicians may use different guidelines, pharmacists need a working knowledge of basic recommendations.
Studies that have assessed knowledge on screening, signs and symptoms of cancer among community pharmacist have been conducted, however, no systematic review have been conducted to pool findings from these studies to inform practice. The aim of this review was to critically appraise evidence gathered from studies that; (1) explore or assess knowledge of community pharmacist on signs and symptoms of cancer, (2) explore or assess knowledge of community pharmacist on cancer screening.
# Methods
## Sources and search strategy
Search of EMBASE (ovid), CINAHL (EBSCOhost) and MEDLINE (EBSCOhost) were done to identify evidence. The search period was from 2005 to July, 2017. The MEDLINE search strategy (Appendix 1) used key words such as cancer, community pharmacist, knowledge, awareness, signs and symptoms, screening. This search strategy was adopted for other databases search. Additional search from reference lists of articles selected for full text review yielded no results. The review was designed and carried out following established guidelines on good conduct and reporting of systematic reviews [bib_ref] The PRISMA statement for reporting systematic reviews and meta-analyses of studies that..., Liberati [/bib_ref]. The protocol was registered with PROSPERO, registration number 2017:CRD42017071390.
## Eligibility criteria and study selection
Two investigators (KM and FO) independently read the titles and abstracts of all records retrieved and assessed them against the set criteria [fig_ref] Table 1: Inclusion and exclusion criteria Studies with full text not available in English... [/fig_ref]. Data from the included studies were extracted by the primary reviewer (KM) using a standardized research matrix [bib_ref] Cancer patients' and professional caregivers' needs, preferences and factors associated with receiving..., Goossens [/bib_ref] , and later checked by another reviewer (AB). Author's name, year of publication, country and setting, study design, type of cancer, sample size, findings, where the data collected (Appendix 2). The search results were independently reviewed by two authors (KM and FO). The database search identified 1538 records. A total of 349 duplicate records were deleted. One thousand one hundred and eight nine (1189) articles were independently screened on title and abstract by two authors (KM and FO) and irrelevant articles were excluded. The authors evaluated 32 full-text articles for eligibility. After exclusion of 28 articles, 4 studies met the criteria for inclusion in the review. A flow chart summarising the selection procedure is shown in [fig_ref] Figure 1: Systematic selection process [/fig_ref].
# Results
The database search found 1538 publications between 2005 and July 2017. A total of 349 duplicate records were removed. A further 1173 records were excluded based on their abstracts and titles. Following the exclusion criteria, another 28 records were also excluded. The remaining 4 articles which met the inclusion criteria were read in full. A flowchart summarising the selection process is shown in [fig_ref] Figure 1: Systematic selection process [/fig_ref].
## Study characteristics
The characteristics of the four studies are shown in [fig_ref] Table 2: Characteristic of studies included in the review [/fig_ref]. The studies were published from 2010 to 2016. The studies were conducted in Malaysia, Qatar, UAE and Jordan. Community pharmacists were recruited from commercial community pharmacies. The studies included a total of 1678 pharmacists. Breast cancer was the type of cancer discussed in the selected studies. The smallest sample size in the studies was 35 [bib_ref] Knowledge, perception, practice and barriers of breast cancer health promotion activities among..., Beshir [/bib_ref] and largest sample size was 1113 [bib_ref] Abu Alkishik S. Knowledge, attitudes and barriers towards breast cancer health education..., Ayoub [/bib_ref].
## Quality assessment of selected studies
The quality of the selected studies was assessed using a quality assessment toolScore from 0% -33.9% is regarded as weak, 34% -66.9% is regarded as moderate, and 67% -100% is regarded as strong (n = 4) based on [bib_ref] Males' awareness of benign testicular disorders: an integrative review, Saab [/bib_ref] classification of quality level [fig_ref] Table 3: Quality Assessment of Selected Studies [/fig_ref].
# Discussion
The data from the selected studies were heterogeneous; hence it was not possible to combine it for metaanalysis. Hence the outcomes of the studies were reported as a narrative synthesis. Findings of the four selected studies revealed lack of sufficient knowledge on breast cancer and screening recommendations. Scores of participants on items about knowledge on cancer signs and symptoms were moderate [bib_ref] Abu Alkishik S. Knowledge, attitudes and barriers towards breast cancer health education..., Ayoub [/bib_ref] [bib_ref] Community PHARMACISTS'INVOLVEMENT in breast cancer health promotion in UNITED Arab EMIRATES (UAE), Ibrahim [/bib_ref]. The other two studies [bib_ref] Knowledge, perception, practice and barriers of breast cancer health promotion activities among..., Beshir [/bib_ref] [bib_ref] Breast cancer health promotion in Qatar: a survey of community pharmacists' interests..., El Hajj [/bib_ref] had only one item on cancer signs and symptoms which does not give a proper reflection about participants knowledge on signs and symptoms about breast cancer. With aging population in the world, the global burden of cancer is set to increase. One of the approaches adopted by the World Health Organisation (WHO) is to raise awareness through education regarding warning signs of cancer. Therefore much has to be done to improve the knowledge of community pharmacist on these warning signs. Scores were noticed to be low for items about knowledge on cancer screening recommendations for one of the studies [bib_ref] Community PHARMACISTS'INVOLVEMENT in breast cancer health promotion in UNITED Arab EMIRATES (UAE), Ibrahim [/bib_ref]. All studies attributed knowledge limitation as the cause of reason for the key findings of their studies. Lack of continuous pharmacy education, nonattendance of continuous pharmacy education and different undergraduate pharmacy curricula contribute to knowledge limitations.
Through this systematic review it can be seen that there has not been many studies done to analyse the knowledge of community pharmacist on screening recommendations, signs and symptoms of cancer for the past 12 years. The selected studies focused on breast cancer only, which hinder the generalizability and transferability of the findings. Hence there is a need for more studies to be conducted in this area to draw a better conclusion which will inform policy.
# Conclusion
In conclusion, community Pharmacists possess moderate knowledge on breast cancer signs, symptoms and screening recommendations. However, the findings of this systematic review were highly limited by the fact that only four studies met the review criteria, samples of studies were taken from only one geographic area, Middle East Region and sample size was relatively small. Hence findings may not be applicable to all community pharmacists in general. Further studies should be conducted in other sub regions of the World to generate results for future policy implementation.
# Limitations
The search was limited to three databases and did not include data from grey literature. Also the search was restricted to studies conducted from 2005 to July, 2017 and studies published in English. These create opportunity for study selection bias.
Researcher-designed questionnaires were used in the selected studies, which led to heterogeneous results that could not be combined for meta-analysis or metasynthesis.
The studies were done in only breast cancer hence cannot be generalised for the other cancers. The review was limited to four studies only, and so worldwide survey is required to address certain perception aspects of breast cancer screening, signs and symptoms.
# Appendix 1
An example of search strategy used in the Medline Ovid platform. - Breast cancer knowledge was evaluated using 7 questions to be answered as yes, no or uncertain. More than half (50%) of the participants answered general questions related to breast cancer and its risk factors correctly. The mean percent score of correct answers was 72.6% ± 11. - Pharmacists' perception regarding breast cancer health promotion was assessed using 8 lickert scale type questions. The mean percent score for the pharmacists who strongly agree or agree with all the given statements was 90% ± 8.07. - 91.4% strongly agree or agree that there is a need to integrate breast cancer health promotion activities i to their daily practice and about 71.4% strongly agree or agree that there is a demand from the community to get advice on breast cancer screening and early detection. - The community pharmacists cite lack of time (80%), lack of breast cancer education materials (77.1%), and training (62.9%) as major barriers that limit their involvement in breast cancer promotion activities.
- The cross-sectional survey was confined to community pharmacists in two districts (Malaysia), hence the results could not be generalized to all community pharmacists in Malaysia.
## Community pharmacists' involvement in breast cancer
Health Promotion in United Arab Emirate (UAE).
## Uae community pharmacies 335
- About 47% of the pharmacists reported that they never provided patients with advice or
- The study variables were assessed by self-report, which may be biased by an inclination to provide socially desirable
[fig] Figure 1: Systematic selection process [/fig]
[table] Table 1: Inclusion and exclusion criteria Studies with full text not available in English language. ➢ Studies published before 2005 [/table]
[table] Table 2: Characteristic of studies included in the review [/table]
[table] 1: 'cancer* OR 'tumour * OR 'tumor* 'OR 'neoplasm* OR 'malignant* OR 'neoplasm malignant* OR 'malignant neoplasm* OR 'tumoral disease* OR malignant tumor* OR 'malignant tumour* OR 'cancer morphology' OR 'malignant tumor morphology' OR 'primary malignant neoplasm* 2. 'community pharmacist* OR community pharma* OR 'community druggist* OR 'pharmacist* 3. 'know* OR 'knowledge' OR 'knowledges' 4. 'awareness OR 'awarenesses' OR aware* 5. 'signs and symptoms' OR 'sign* OR 'symptom* 6. screening* OR 'early detection of cancer' [/table]
[table] Table 3: Quality Assessment of Selected Studies [/table]
[table] Table 4 A: summary of the findings of studies that were reviewed [/table]
|
Cardiogenic programming of human pluripotent stem cells by dose-controlled activation of EOMES
Reviewers' Comments:Reviewer #1: Remarks to the Author: GENERAL COMMENTS Pfieffer and colleagues present a manuscript examining the role of eomesodermin (EOMES) in regulating cardiac mesoderm induction using a human ESC model. This transcription factor is known to be essential in germ layer development particularly of endoderm and mesoderm based on prior studies. In this manuscript, the authors generate a knockout of EOMES in human ESCs and show that this clearly blocks cardiac differentiation using a standardized protocol that can be rescued with a genetically engineered EOMES dox-inducible construct. Most importantly, the authors implicate the role of EOMES in inducing Wnt3 expression as the key downstream effector that promotes the formation of cardiac mesoderm using additional chemical inhibition and knockout lines. The authors conclude with a model of a feedback cardiac induction module of EOMES-Wnt. Overall, the manuscript is clearly written and presents novel findings regarding the impact of EOMES on Wnt3 expression. However, there are some significant limitations as currently presented as follow:
1. A single human ESC lines was apparently used in this study, HuES6, and it is important to confirm that the findings are generalizable and not unique to this stem cell line. For example, given the known variations in paracrine signaling in different stem cell lines, different results could be observed in different cell lines (e.g. :228-40 for individual cell line titration of growth factors in cardiac differentiation). Repeating a key experiment or confirmatory experiment in another cell line would strengthen the manuscript.
2. The authors' feedback model of EOMES and Wnt3 signaling is oversimplified. While the authors provide clear evidence that EOMES can stimulate Wnt3 expression and signaling, the converse is not clear from the presented data -that Wnt3 can stimulate EOMES expression or signaling. The effect of Wnt signaling is context-dependent, so Wnt effects on cardiac progenitors are likely quite different than on induction of cardiac mesoderm. Either the authors need to revise their model or provide such evidence. MINOR COMMENTS 1.Should 'eomesodermin' be used in title rather than EOMES for clarity? 2. On page 3, the authors state that "To verify this conclusion in a more functional way, pCMs were treated with E-4031, an inhibitor of the hERG potassium channel that plays a prominent role in ventricular repolarization." Although it is true that hERG channels contribute to ventricular repolarization, they also contribute to atrial repolarization prominently and so do not discriminate between cardiomyocytes from these chambers.
3. On page 3, "…and slowdown after addition of propranolol which is a beta blocker ." Why should a beta blocker slow the endogenous rate in the absence of adrenergic stimulation or innervation? 4. On page 4, "…easily be overdosed." Overdose is the wrong word here as doses are given to patients or animals. 5. Y axis label of QTc interval needs to be more clearly labeled as it is not true QTc which comes from surface electrocardiograms. Some indication that this is not standard QTc such as QTc-like or Pseudo-QTc, etc should be used.
6. Reference 8 is incomplete.
Reviewer #2: Remarks to the Author: This paper investigates EOMES during cardiac differentiation of human embryonic stem cells . The finding that EOMES is essential for cardiac differentiation has been published previously by this group .
Eomes is a key driver of mesoderm formation, including the driving epithelial-to-mesenchymal transitions necessary for nascent mesoderm formation. Thus, it is highly likely that a failure to form mesoderm underlies the phenotype rather than specifically impaired cardiomyocyte differentiation. In this context, the authors should examine the capacity of the EOMES knockout to form other mesodermal lineages. The data establish the endoderm differentiation is compromised in EOMES knockouts. Furthermore, Tet-ON EOMES patterned day 2 differentiated cells should be examined for the capacity to produce other mesodermal lineages (eg. blood) and endodermal derivatives.
Similarly, overexpression of EOMES using PiggyBac vectors promotes mesodermal markers . The cardiogenic potential of this mesoderm can be exploited by Wnt inhibition, as has been established. However, whether mesoderm derived by the TET-ON EOMES over expression is specific to the cardiac lineage has not been evaluated. The authors should test if EOMES induced mesoderm retains the capacity to form other mesodermal derivatives.
The finding that EOMES can substitute for cytokine/small molecules for mesoderm induction is an interesting one and may be of utility for biotechnology approaches. However, at present the expression is based on random integration of the PiggyBac vectors. Have the authors used a single integration site such as AAVS1 to control for possible integration defects (eg. silencing, mutagenesis)?
The major finding is establishing a link between EOMES and WNT3 signalling during early mesoderm specification. Inhibiting WNT signalling or deleting WNT3 prevents the development of cardiomyocytes, presumably because mesoderm does not form. Indeed, Wnt3 knockout mice do not form a primitive streak. No data is shown when WNT3 is deleted in a wild type background, the hypothesis would be that WNT3 is necessary for mesoderm formation. This control is necessary to further establish that EOMES drives mesoderm development via a WNT3-dependant mechanism. Furthermore, WNT and BMP signalling act synergistically during mesoderm formation and BMP4 is unregulated by EOMES . While a BMP inhibition experiment is performed it is done in the context of WNT inhibition. Have the authors tested BMP inhibition alone?
Another concern is the use of only a single hESC line. The key findings need to be confirmed in another genetic background to ensure that these findings are not restricted to the HuES6 cell line.
## Minor points
Regarding "the issue of whether there is a bona fide master regulatory factor specifically promoting the induction of cardiac cells", multiple lines of evidence suggest that there is no single key master cardiogenic factor. This is due to the nature of the highly conserved gene regulatory network that controls cardiogeneis. While removal of a single member of this network often results in lethality in all cases contractile cardiomyocytes are formed. Furthermore, attempts to make induced cardiomyocytes have shown that at least 3 transcription factors (GATA4, TBX5, MEF2c) are required to reprogram fibroblasts. Thus, a "master regulatory" factor for cardiomyocytes does not exist. This should be clarified in the introduction.
Please separate the genes in the gene expression graphs (eg d, e). Combining multiple genes in one bar makes it difficult to interpret the data. This is used throughout the manuscript. Also, when relative gene expression is used please indicate in the figure legend what the expression is relative to (eg. relative to wildtype etc).
Have the authors examined the onset of mesodermal markers (CD13, ROR2, PDGFRa etc) in differentiating EOMES knockout cell lines? The microarray data (Table S1) shows that transcripts for PDGFRa are up-regulated in the TET-On EOMES samples whereas ROR2 levels do not vary. These data may be useful in quantifying mesodermal cell production in the various cell lines used.
How many PiggyBac lines are used in the experiments? Is all the data from a single clone or are multiple lines used to confirm the findings? legend: "~1.5 weeks" and "~3 weeks" are somewhat vague. Please provide exact days of differentiation.
The modified hESC lines were not karyotyped to ensure that no abnormalities arose during the genetic manipulation and no evidence is shown that they maintain pluripotency markers. please provide a schematic of the WNT3 and WNT3a alleles showing the region deleted and the sequence data demonstrating that the allele. Also, a second PCR should be performed using a primer internal to the CRISPR mediated deletion to confirm the knockout. The smaller PCR product may preferentially amplify meaning that a heterozygous genotype may be misclassified as a homozygote. An internal/external PCR combination should only produce a product if a wild type allele is still present. This could be presented in supplementary data. please indicate on the figure that the WNT inhib. was included with the BMP inhibitor. This experiment should also be repeated using the BMP inhibitor alone.
The microarray data presented in could be presented more effectively. A summary of the data for each transcript (rather than each spot on the microarray), including a column showing fold differences between the samples, should be included as a separate table within the excel spreadsheet. This would make it easier for the reader to identify which gene networks are dysregulated.
The model in suggests the EOMES-WNT form a cardiac mesoderm induction module, however, further data are required to confirm that this indeed cardiac specific mesoderm. Further differentation of these cells along blood, endothelial and other mesodermal lineages is required to formally establish that EOMES-WNT specific pre-cardiac mesoderm.
Reviewer #3: Remarks to the Author: The role of Eomes in cell fate specification and in regulating WNT3 expression is novel and important. I find this elegant study very interesting and overall convincing and highly recommend publication after addressing the following points:
1. The treatment of WNT3 KO cells with CHIR compound in is a nice specificity control, but a more quantitative readout of the rescue of cardiac differentiation would strengthen the conclusion. In addition, Methods should state how many independent CRISPR clones were analyzed per genotype.
2. Please indicate the statistical significance of the GO term enrichments shown in 3. The legend of Suppl. states that n was 2 to 5. To estimate the significance of the cumulative effect on cardiac fate, the authors should state which differences did or did not reach statistical significance. 4. In Suppl. , 'time ----50 ms' probably denotes the scale of the x-axis, but this should be clarified (ideally by drawing an x-axis with ticks). seems to be missing in the text.
## A description of
We wish to thank all reviewers for expressing interest in our work and evaluating it in a fair and constructive manner. We do believe that the various changes to the manuscript (highlighted in blue font color) and the additional data provided in this revised version will further strengthen our conclusions and increase the overall value of this work.
Reviewer #1 (Remarks to the Author):
## General comments
Pfieffer and colleagues present a manuscript examining the role of eomesodermin (EOMES) in regulating cardiac mesoderm induction using a human ESC model. This transcription factor is known to be essential in germ layer development particularly of endoderm and mesoderm based on prior studies. In this manuscript, the authors generate a knockout of EOMES in human ESCs and show that this clearly blocks cardiac differentiation using a standardized protocol that can be rescued with a genetically engineered EOMES dox-inducible construct. Most importantly, the authors implicate the role of EOMES in inducing Wnt3 expression as the key downstream effector that promotes the formation of cardiac mesoderm using additional chemical inhibition and knockout lines. The authors conclude with a model of a feedback cardiac induction module of EOMES-Wnt. Overall, the manuscript is clearly written and presents novel findings regarding the impact of EOMES on Wnt3 expression. However, there are some significant limitations as currently presented as follow:
1. A single human ESC lines was apparently used in this study, HuES6, and it is important to confirm that the findings are generalizable and not unique to this stem cell line. Altough we had already demonstrated pCM formation in two clonal TET-ON cell lines in the first version of the manuscript, these admittedly share the same genetic background. We have hence prepared an additional cell line, on induced pluripotent stem cell background, which involved the screening of ~50 candidate clones to identify at least one that would homogeneously express EOMES in a DOX concentration-dependent manner, as required. Validation data of one selected line are now shown in r e,f, and highly efficient cardiomoycyte programming is demonstrated in r c. Hence, these data suggest that EOMES universally induces a cardiac mesoderm fate both in hES and hiPS cells. We thank this reviewer for encouraging these significant but useful efforts.
2. The authors' feedback model of EOMES and Wnt3 signaling is oversimplified. While the authors provide clear evidence that EOMES can stimulate Wnt3 expression and signaling, the converse is not clear from the presented datathat Wnt3 can stimulate EOMES expression or signaling. The effect of Wnt signaling is context-dependent, so Wnt effects on cardiac progenitors are likely quite different than on induction of cardiac mesoderm. Either the authors need to revise their model or provide such evidence.
We do not deny that the model presents a simplified view on the complex process of cardiac induction, as there are certainly many more genes involved downstream of the module. We do think, though, that it brings the key findings of our work across in a crisp and intuitive manner. One aspect that was admittedly not highlighted in the previous version of the model is the dose-dependence. Through additional experimentation, we have now investigated differentiation fates promoted outside the DOX concentration range that is most optimal for cardiac induction. These additional and revealing findings are shown in revised ,d and they have also been incorporated into the model of r h -together with indicating the overall dose-dependence of the system, in agreement with developmental principles.
As to the link between WNT and EOMES: In the legend to r h, we are now referring to our previous work that the present study is in part based on . In that paper, we show that EOMES activation by canonical WNT signaling is a first key event of cardiac mesoderm induction. Indeed, some contemporary cardiac differentiation protocols are exclusively based on canonical WNT signaling activation. We appologize for not having stressed out the link to our previous study in a sufficiently clear manner.
## Minor comments
1.Should 'eomesodermin' be used in title rather than EOMES for clarity?
We intentionally use "EOMES" instead of "eomesodermin" throughout, in part to better highlight the genetic nature of our new approach. Since "eomesodermin" is another commonly used designation, though, we acknowledge the recommendation and now mention this name in the revised abstract.
2. On page 3, the authors state that "To verify this conclusion in a more functional way, pCMs were treated with E-4031, an inhibitor of the hERG potassium channel that plays a prominent role in ventricular repolarization." Although it is true that hERG channels contribute to ventricular repolarization, they also contribute to atrial repolarization prominently and so do not discriminate between cardiomyocytes from these chambers.
The claim that hERG would not significantly contribute to atrial repolarization was based on some of our recent observations in the context of cardiac subtype-specific disease modeling in the hPSC system (Marczenke et al., PMID 28729840). Admittedly, however, this may not present an accepted view and it might overall be confined to the hPSC-CM system. And in general, it is totally true that there are no strict ventricular-specific potassium channels. We have therefore decided to merely mention the result as a selected demonstration of pan-cardiac ion channel functionality (page 3 of revised manuscript). Instead, in r d, we are now showing new data generated using 4-AP as an atrial-specific ion channel inhibitor. We hope R1 agrees that these physiological data, together with the previous gene expression analysis, now support our conclusion -that pCM are ventricular-like by default -in a more convincing way.
3. On page 3, "…and slowdown after addition of propranolol which is a beta blocker ." Why should a beta blocker slow the endogenous rate in the absence of adrenergic stimulation or innervation?
We consistently observe a decrease in beating rates using beta blocker treatment of hPSC-CMs, which is thus not specific to pCMs (see e.g. in . This would simply suggest that there is a certain baseline cAMP pathway activity in the cells.
4. On page 4, "…easily be overdosed." Overdose is the wrong word here as doses are given to patients or animals.
We have followed the recommendation and replaced the phrase by "under or overstimulated" (page 4 of revised ms).
## Suppl fig 2c -y axis label of
QTc interval needs to be more clearly labeled as it is not true QTc which comes from surface electrocardiograms. Some indication that this is not standard QTc such as QTc-like or Pseudo-QTc, etc should be used.
We agree. In new r d, as well as in the text throughout, we now speak of field potential durations or "QT max -like" intervals.
## Reference 8 is incomplete.
This has been corrected (thanks a lot for taking such a close look).
## Reviewer #2 (remarks to the author):
This paper investigates EOMES during cardiac differentiation of human embryonic stem cells (hESCs). The finding that EOMES is essential for cardiac differentiation
## Has been published previously by this group (rao et al cell stem cell 2016).
Eomes is a key driver of mesoderm formation, including the driving epithelial-to-mesenchymal transitions necessary for nascent mesoderm formation. Thus, it is highly likely that a failure to form mesoderm underlies the phenotype rather than specifically impaired cardiomyocyte differentiation. In this context, the authors should examine the capacity of the EOMES knockout to form other mesodermal lineages. The data establish the endoderm differentiation is compromised in EOMES knockouts. Furthermore, Tet-ON EOMES patterned day 2 differentiated cells should be examined for the capacity to produce other mesodermal lineages (eg. blood) and endodermal derivatives. Similarly, overexpression of EOMES using PiggyBac vectors promotes mesodermal markers . The cardiogenic potential of this mesoderm can be exploited by Wnt inhibition, as has been established. However, whether mesoderm derived by the TET-ON EOMES over expression is specific to the cardiac lineage has not been evaluated. The authors should test if EOMES induced mesoderm retains the capacity to form other mesodermal derivatives.
These are interesting and insightful points brought up by R2. Although we felt that delving into alternative non-cardiac differentiation protocols would go beyond the scope of our study, we found it useful to investigate alternative fates based on our present differentiation platform using moderately varied induction conditions. Indeed, we had previsously demonstrated that terminal differentiation outcomes on our platform are highly dependent on the concentrations of the inductive signaling cues, as well as on subsequent WNT inhibition . We have therefore approached the questions from two angles:
In a first set of additional experiments, we investigated non-cardiac mesoderm differentiation competence in EOMES KO cells. This is described in results and methods (page 2/3 and 7 of the revised manuscript, respectively) and a corresponding unbiased expression analysis is shown in r . Although we did not specifically monitor blood differentiation markers, these data interestingly reveal that the EOMES knockout preserves differentiation competence into several mesodermal cell types except cardiomyocytes.
In a second experimental series, we followed up on the point that cardiac induction was strictly dose-dependent, a fact that is now also reflected in the revised model of We strongly hope these efforts adequately address the points by R2 in the context of this overall cardiac-centred work. Certainly, non-cardiac mesoderm fates could be further investigated in dedicated follow-up studies and we will be happy to share our cell lines with the community for this purpose.
The finding that EOMES can substitute for cytokine/small molecules for mesoderm induction is an interesting one and may be of utility for biotechnology approaches. However, at present the expression is based on random integration of the PiggyBac vectors. Have the authors used a single integration site such as AAVS1 to control for possible integration defects (eg. silencing, mutagenesis)?
We have not used safe-harbour integration of the expression cassette. It is true that with random integration, silencing of the transgenic cassette can in principle occur. For this reason, established candidate lines need to be carefully validated. For example, upon establishing our new TET-ON hiPSC line in course of the present revision, we have screened approximately 50 candidate clones for an efficient interplay between the TET-ON cassette and the transactivator, as well as for homogeneous transgene induction over at least several days of course. Once established, though, a good cell line tends to efficiently induce the transgene even after many passages, so it is a reliable and powerful system. Our findings are highly unlikely to result from unspecific effects. Firstly, we are investigating a clearcut gain-of-function approach by which the previously established EOMES knockout could be faithfully rescued. Secondly, we show that this pro-cardiac phenotype is strictly dependent on dose-dependent transgene induction, while confirming in r and S3 r c that the cells are also capable of forming alternative cell fates. Thirdly, we delinate a specific molecular mechanism of EOMES action that is dependent of WNT3 activation and that could be abolished using WNT3 disruption and in turn be rescued using exogenous WNT activation. And finally, we demonstrate highly similar pCM formation outcomes in three independent cells on two different genetic backgrounds. We think these facts collectively exclude the possbility of unspecific mutagenic effects.
The major finding is establishing a link between EOMES and WNT3 signalling during early mesoderm specification. Inhibiting WNT signalling or deleting WNT3 prevents the development of cardiomyocytes, presumably because mesoderm does not form. Indeed, Wnt3 knockout mice do not form a primitive streak. No data is shown when WNT3 is deleted in a wild type background, the hypothesis would be that WNT3 is necessary for mesoderm formation. This control is necessary to further establish that EOMES drives mesoderm development via a WNT3-dependant mechanism. Furthermore, WNT and BMP signalling act synergistically during mesoderm formation and BMP4 is unregulated by EOMES . While a BMP inhibition experiment is performed it is done in the context of WNT inhibition. Have the authors tested BMP inhibition alone?
In our mechanistic analysis, we were particularly interested in elucidating a mechanism by which EOMES drives the process in our TET-ON system. We are not saying this is the only mechanism but we show that the activation of autocrine WNT signaling downstream of EOMES is essential. In the context of mouse in vivo studies alluded to by R2, we think the close interplay between EOMES and WNT revealed here and in our prior study makes it more understandable that both EOMES and WNT3 are such important players, as discussed in the manuscript and indicated in the model.
We are not actually sure whether WNT3 disruption would render signaling factor-driven cardiac differentiation impossible -probably not, because in the standard protocol, WNT is activated by the extrinsic supplementation of CHIR. Indeed, we show that CHIR addition can rescue WNT3 deficiency in the context of the TET-ON approach . Hence, it is likely that e.g. elevating the initial CHIR dose applied to the cells could readily overcome a genetic defect in WNT3 using the signaling factor-based differentiation protocol. In any case, the WNT3 KO phenotype in such a setting would not question our findings about EOMES' mode of action and therefore, we did not pursue knocking out the WNT3 gene in wild-type cells.
As to the role of BMP signaling, we fully acknowledge its contribution in the standard protocol as we have ourselves demonstrated a close collaboration with WNT signaling in the aforementioned Rao et al. study. We therefore considered both pathways as candidates for being involved downstream of EOMES, but our unbiased analysis of WNT and BMP target gene activation together with the pathway inhibition experiments in r a,b just did not reveal any substantial contribution by BMP. In course of this revision, we have carried out the suggested BMP inhibition experiment in the absence of the WNT inhibitor. As the below data shows, however, the WNT inhibition step is indeed crucial for enabling overall cardiomyocyte differentiation as a baseline:
For clarification, we are now stating this fact in the legend to r b while referring to existing data in : "C-59 had to be administered to all samples of the bottom experiments to provide overall CM-permissive differentiation conditions as a baseline (see ." In sum, BMP signaling is a weaker activator of EOMES than WNT and Rao et al.) and conversely, the EOMES TET-ON approach involves an immediate activation of autocrine WNT rather than BMP signaling.
Another concern is the use of only a single hESC line. The key findings need to be confirmed in another genetic background to ensure that these findings are not restricted to the HuES6 cell line.
We agree that showing EOMES-driven cardiac induction in fully independent cell lines is key to demonstrating a universal principle. As already mentioned, therefore, we have prepared such a cell line on an induced pluripotent stem cell background and utilize it to demonstrate DOX dose-dependent cardiac induction. The corresponding data is in r c and S3 r e,f (also see results text on page 4 and added information in methods).
## Minor points
Regarding "the issue of whether there is a bona fide master regulatory factor specifically promoting the induction of cardiac cells", multiple lines of evidence suggest that there is no single key master cardiogenic factor. This is due to the nature of the highly conserved gene regulatory network that controls cardiogeneis. While removal of a single member of this network often results in lethality in all cases contractile cardiomyocytes are formed. Furthermore, attempts to make induced cardiomyocytes have shown that at least 3 transcription factors (GATA4, TBX5, MEF2c) are required to reprogram fibroblasts. Thus, a "master regulatory" factor for cardiomyocytes does not exist. This should be clarified in the introduction.
We wish to mention that MESP1 had previously -and perhaps somewhat prematurely -been termed a cardiac master regulator based on ES cell studies which we refer to in the introduction. Furthermore, we do not make any claims about trans-reprogramming of e.g. fibroblasts into cardiac cells. It is in fact likely that this will not work using EOMES induction alone. Instead, we term it a "context-dependent master regulator", which does not present an overstatement we think. As explained in the discussion section, this activity requires the pluripotent stem cell context, the ability to tightly control EOMES expression levels, and to inhibit endogenous WNT signaling in a subsequent step. Under these conditions, however, EOMES drives near-homogeneous cardiac induction, which is pretty remarkable we think and this should also be reflected in a strong metaphoric designation in our opinion.
Please separate the genes in the gene expression graphs (eg . Combining multiple genes in one bar makes it difficult to interpret the data. This is used throughout the manuscript. Also, when relative gene expression is used please indicate in the figure legend what the expression is relative to (eg. relative to wildtype etc).
It is overall useful to monitor more than just on marker gene in a given context, as this tends to generate more robust data. On the other hand, we think it is important to present the data in a manner as intuitive and simple as possible. At least this is what we attempt to do. Piling up genes in bar charts is one very useful measure to achieve a good compromise in these regards we think. We therefore wish to hold on to this way of presenting our data, also because even detailed information can be extracted from it. For example, the CM characterization data of r c would otherwise look unnecessarily complicated and consume far too much space. qPCR data can be expressed in different ways and it is legitimate to be somewhat flexible with this in order to bring the essence of the data across in the most intuitive manner. When plotting time-course expression data of several genes into one chart, for instance, it can be most useful to normalise the data to the highest expressing sample -to best convey the respective temporal exression patterns (rather than highlighting more "absolute" expression levels for instance). Likewise, normalising the expression of differentiation markers against undifferentiated cells -which essentially do not express those genes -can be problematic due to rather strong baseline fluctuations in the ground state. For clarification, therefore, we have now stated more clearly in the methods section which ways of expressing qPCR data have overall been used in this study (page 8). Moreover, the specific normalization procedure employed is also indicated on the y axis of every chart shown in this work.
Have the authors examined the onset of mesodermal markers (CD13, ROR2, PDGFRa etc) in differentiating EOMES knockout cell lines? The microarray data (Table S1) shows that transcripts for PDGFRa are up-regulated in the TET-On EOMES samples whereas ROR2 levels do not vary. These data may be useful in quantifying mesodermal cell production in the various cell lines used.
In general, even primitive streak markers and early mesoderm genes are severely compromised in EOMES KO cells, as highlighted in the heat map representation of r e. Nonetheless, as correctly pointed out by R2, some cardiac-specific genes seem to be rather unaffected -ISL1 being one striking example r e, Table S1 r ). Clearly, there is much room here for follow-up investigation and it is for this reason that we decided to make the entire time-course expression table available as easily accessible supplementary information.
How many PiggyBac lines are used in the experiments? Is all the data from a single clone or are multiple lines used to confirm the findings?
In the revised manuscript, we have now clarified on several occasions which cell lines have been used. The study is based on three clonal cell lines -an EOMES KO/E.TET-ON hESC line, a WT E.TET-ON hESC line (both on HuES6 background), and a WT E.TET-ON hiPSC line. legend: "~1.5 weeks" and "~3 weeks" are somewhat vague. Please provide exact days of differentiation.
The symbols were to indicate that similar outcomes could independently be obtained "around" the indicated time-points. But we agree to the suggestion and have accordingly modified information in the figure and legend ("d 10" and "21 days").
The modified hESC lines were not karyotyped to ensure that no abnormalities arose during the genetic manipulation and no evidence is shown that they maintain pluripotency markers.
As suggested, we are now providing karyotype analyses for all three cell lines (in r a and S3 r a,e; n = 10 each). In addition, r b,f show robust SOX2 expression in the undifferentiated state of 2 of these lines. Otherwise, we think that the numerous cardiac and non-cardiac differentiation experiments throughout this study inherently provide compelling evidence of the plurpotency of our cell lines. please provide a schematic of the WNT3 and WNT3a alleles showing the region deleted and the sequence data demonstrating that the allele. Also, a second PCR should be performed using a primer internal to the CRISPR mediated deletion to confirm the knockout. The smaller PCR product may preferentially amplify meaning that a heterozygous genotype may be misclassified as a homozygote. An internal/external PCR combination should only produce a product if a wild type allele is still present. This could be presented in supplementary data.
Another useful suggestion. New contains a corresponding schematic of the knockout approach as well as the additional diagnostic PCR data. The corresponding primer and CRISPR sequences are listed in r a and can readily be employed to locate the corresponding sites in the genome using e.g. BLAT searches. In addition, we are now providing the genomic sequences in the WNT3 and WNT3A loci that have been deleted from the genome (in Figure 3d please indicate on the figure that the WNT inhib. was included with the BMP inhibitor. This experiment should also be repeated using the BMP inhibitor alone.
We have modfied r b accordingly. The suggested experiment has been carried out (see above) but we do not wish to include it in the manuscript. Instead, in the legend to , we refer to a related experiment in
[formula] r b. [/formula]
The microarray data presented in could be presented more effectively. A summary of the data for each transcript (rather than each spot on the microarray), including a column showing fold differences between the samples, should be included as a separate table within the excel spreadsheet. This would make it easier for the reader to identify which gene networks are dysregulated.
As suggested, we have now included an additional spreadsheet with expression ratios against the standard WT differentiation time-course (Table S1 r b). Furthermore, it is true that many genes on the array are represented by more than one probe. In many cases, though, these are isoform-specific detection probes, which prohibits averaging the corresponding signal intensities. Instead, we have now included all detection probe sequences as an additional column in the supplementary table and this will enable the readers to selectively look up where they bind to.
The model in suggests the EOMES-WNT form a cardiac mesoderm induction module, however, further data are required to confirm that this indeed cardiac specific mesoderm. Further differentation of these cells along blood, endothelial and other mesodermal lineages is required to formally establish that EOMES-WNT specific pre-cardiac mesoderm.
As strongly confirmed by the additional suggested experiments of r , S3 r c,d, as well as by the new WT E.TET-ON hiPSC-based data and S3 r f), efficient cardiac induction is confined to a cell clone-dependent DOX -i.e. EOMES expression level -range. Outside this range, EOMES drives differentiation into alternative fates. These results and observations are now incorporated into an extended model presented in In essence, we do not claim that EOMES overexpression as such is cardiac-specific, which resembles the effects of BMP/WNT-activating molecules that are likewise not cardiac-specific.
The role of Eomes in cell fate specification and in regulating WNT3 expression is novel and important. I find this elegant study very interesting and overall convincing and highly recommend publication after addressing the following points: is a nice specificity control, but a more quantitative readout of the rescue of cardiac differentiation would strengthen the conclusion. In addition, Methods should state how many independent CRISPR clones were analyzed per genotype.
## The treatment of wnt3 ko cells with chir compound in
We were able to isolate one CRISPR clone per genotype, while given the dispensability of WNT3A in this context, we would argue that the WNT3 KO and the DKO could serve as specificity controls for one another. The CHIR supplementation experiments were performed separately from the primary experiments analyzed in r f, and by contrast, they were only analyzed in a semiquantitative manner, namely by estimating the percentage of beating cell surface area through a stereo microscope. Due to the semiquantitative nature of these data, we did not wish to incorporate them in the manuscript but I would like to emphasise that all cardiac immunostains shown in this work were overall representative ones (as also pointed out in the legend of . Given the high overall workload imposed in this revision, we hope R3 does not insist on repeating these demanding CHIR supplementation experiments, which requires a perfect control of culture conditions and synchronisation of four different cell lines in parallel:
The CHIR dose required for achieving a full rescue here was rather high (8 µM -indicated in legend to . Interestingly, though, this roughly corresponds to the concentration range used in standard "CHIR-only" cardiac differentiation protocols (e.g. Lian et al., PMID 22645348) and suggests that there is rather strong autocrine WNT3 signaling in WT E.TET-ON cells.
2. Please indicate the statistical significance of the GO term enrichments shown in .
The statistical significance of the data in r d is implied in the bar values themselves which denote "-log 10 of the p value" (a bar value of 10 would denote a p value of 10 -10 , for instance). For clarification, we have accordingly expanded the corresponding figure legend in the revised manuscript.
3. The legend of Suppl. states that n was 2 to 5. To estimate the significance of the cumulative effect on cardiac fate, the authors should state which differences did or did not reach statistical significance.
We agree and have accordingly performed 1-sided t-tests against the best condition (with plausible outcomes considering the inter-experimental variation and number of replicates). This is additionally provided for several pairwise comparisons in 4. In Suppl. , 'time ----50 ms' probably denotes the scale of the x-axis, but this should be clarified (ideally by drawing an x-axis with ticks).
In the revised version of the manuscript, we have replaced those data by a new experiment but have followed R3's recommendation when drawing the axis. seems to be missing in the text.
## A description of
This is now included (new r c / page 5). Thanks a lot for pointing this out.
## Reviewers' comments:
Reviewer #1: Remarks to the Author: The revised manuscript by Pfeiffer and colleagues addresses my initial concerns. Using an additional hiPSC line provides some assurance that the findings are more generalizable. The dox concentration-dependent induction of EOMES with the demonstration of alternative lineages also strengthen the findings. Overall, many of these findings are novel and add meaningfully to the literature by defining the critical role of EOMES in cardiogenesis.
Reviewer #2: Remarks to the Author: To address the reviewers concerns the authors present a large collection of new data and have altered the text where appropriate. The majority of the reviewers concerns have been adequately addressed.
However, the key claim that EOMES is a context-dependant cardiac master regulator remains contentious. This group had previously shown that EOMES is essential for the formation of cardiac competent mesoderm, in part via the repression of SOX2 . Further, EOMES has been shown in animal models to be essential for a range of mesendodermal lineages, and in the context of heart development a direct regulator of MESP1 (van den Ameele, EMBO Rep. 2012). In the revised manuscript the authors present evidence that in EOMES is required for endoderm differentiation and EOMES knockout cells readily differentiate down neruoectodermal lineages. In this respect, the title implying that EOMES is a context dependant cardiac master regulator neglects a fundamental requirement for EOMES in endoderm development. The overexpression data using the TET-ON system demonstrate that EOMES is capable of patterning mesoderm in a dose dependent manner. The title should reflect this finding rather than the cardiac using the term cardiac master regulator. Cardiac differentiation in this system still relies up a subsequent suppression of WNT signalling, if this suppression is replace by other signalling (eg VEGF) a different cell population would be generated. Historically, a master regulator is a transcription factor that single handedly converts cell fate, eg. ANTP or MYOD. This is not the case for the data presented in this manuscript. This reviewer feels that this data support the hypothesis that EOMES is critical for the specification of cardiac mesoderm but the term "cardiac master regulator" could be misleading.
## Other issues
As raised previously, the cumulative bar graphs are extremely difficult to interpret and relative expression is used without clearly stating what the values are relative too. For example, in figure 1 c expression of NKX2-5 is given as 3. Firstly, the Y-axis says "rel. expr. on day 8 [max=1]". So if max=1, how can NKX2-5 expression equal 3? Secondly, does this mean that NKX2-5 expression is 3 fold higher than CTNT and 1.5 higher than MYH6? This data is completely inconsistent with the known difference in mRNA levels of transcription factors and sarcomeric proteins. By contrast, in MYH6 is now 3 fold higher than NKX2-5 only 1 day earlier in differentiation. Is this due to line specific differences or is this difference due to the widespread over expression of TET-EOMES or is it due to the timing of Q-PCR? Please present this data in a more appropriate manner for the reader understand.
The authors show "unbiased expression analysis of meso-permissive differentiation" in , however, gene expression data underlying this is not shown. This data should be included. Only select GO annotations are shown and the data for this GO annotation is not presented. Until this data is shown it remains possible that these genes are present in multiple GO terms and that those selected are not mesodermally restricted. The data in clearly show that mesoderm development is dramatically impaired in EOMES knockouts with dramatic reductions in MESP1, MIXL1, T and LHX1. Supporting the hypothesis that this EOMES is required for mesoderm formation rather than cardiac mesoderm specifically.
Reviewer #3: Remarks to the Author: The authors have satisfactorily addressed my comments, except for the first: I could still not find information in the Methods that only one clone per genotype was analyzed. I think this info is important for the readers. I also recommend to include (line 244?) how many correctly targeted clones were obtained: This is useful for anyone seeking to reproduce this targeting experiment.
We wish to thank the reviewers once more for reviewing the changes made to the manuscript. In this second round, we have now addressed the remaining open points as detailed below. Changes to the manuscript text are highlighted in blue font colour.
## Reviewers' comments:
Reviewer #1 (Remarks to the Author):
The revised manuscript by Pfeiffer and colleagues addresses my initial concerns. Using an additional hiPSC line provides some assurance that the findings are more generalizable. The dox concentration-dependent induction of EOMES with the demonstration of alternative lineages also strengthen the findings. Overall, many of these findings are novel and add meaningfully to the literature by defining the critical role of EOMES in cardiogenesis.
## Reviewer #2 (remarks to the author):
To address the reviewers concerns the authors present a large collection of new data and have altered the text where appropriate. The majority of the reviewers concerns have been adequately addressed.
However, the key claim that EOMES is a context-dependant cardiac master regulator remains contentious. This group had previously shown that EOMES is essential for the formation of cardiac competent mesoderm, in part via the repression of SOX2 (Rao Cell Stem Cell 2015). Further, EOMES has been shown in animal models to be essential for a range of mesendodermal lineages, and in the context of heart development a direct regulator of MESP1 (van den Ameele, EMBO Rep. 2012). In the revised manuscript the authors present evidence that in EOMES is required for endoderm differentiation and EOMES knockout cells readily differentiate down neruoectodermal lineages. In this respect, the title implying that EOMES is a context dependant cardiac master regulator neglects a fundamental requirement for EOMES in endoderm development. The over-expression data using the TET-ON system demonstrate that EOMES is capable of patterning mesoderm in a dose dependent manner. The title should reflect this finding rather than the cardiac using the term cardiac master regulator. Cardiac differentiation in this system still relies up a subsequent suppression of WNT signalling, if this suppression is replace by other signalling (eg VEGF) a different cell population would be generated. Historically, a master regulator is a transcription factor that single handedly converts cell fate, eg. ANTP or MYOD. This is not the case for the data presented in this manuscript. This reviewer feels that this data support the hypothesis that EOMES is critical for the specification of cardiac mesoderm but the term "cardiac master regulator" could be misleading.
In a strict sense, R2 is right. The function of EOMES to program hPSCs into a cardiogenic fate may not apply to other starting cell types such as fibroblasts. It appears to us, though, that nowadays the term "master regulator" is also being used in a more general way, to include powerful transcription factors capable of inducing entire developmental programmes as in the present case. As an example, Mesp1 as well has been proposed to be a master regulator in the cardiac context, albeit perhaps somewhat prematurely . So, it seems that the definition of the term has loosened over time.
Nonetheless, to acknowledge the criticism (and to hopefully prevent further delays), we have now replaced the term "master regulator" by other expressions throughout -except for one occurence in the discussion where we, hovewer, toned down the statement ("may be considered"). In addition, we have completely revised the title. This latter modification was also to avoid the impression that EOMES activity would exclusively produce cardiac outcomes. In fact, in abstract, introduction, as well as in results , we fully recognise that EOMES is essential for endoderm differentiation as well. Otherwise, we think the data added in the first revision strongly support our model according to which diverse differentiation outcomes form in an EOMES expression strength-dependent manner.
## Other issues
As raised previously, the cumulative bar graphs are extremely difficult to interpret and relative expression is used without clearly stating what the values are relative too. For example, in figure 1 c expression of NKX2-5 is given as 3. Firstly, the Y-axis says "rel. expr. on day 8 [max=1]". So if max=1, how can NKX2-5 expression equal 3? Secondly, does this mean that NKX2-5 expression is 3 fold higher than CTNT and 1.5 higher than MYH6? This data is completely inconsistent with the known difference in mRNA levels of transcription factors and sarcomeric proteins. By contrast, in MYH6 is now 3 fold higher than NKX2-5 only 1 day earlier in differentiation. Is this due to line specific differences or is this difference due to the widespread over expression of TET-EOMES or is it due to the timing of Q-PCR? Please present this data in a more appropriate manner for the reader understand. Now we understand where the misunderstanding lies -thanks very much for explaining. qPCR data may be normalized to any sample in a given experiment. In many cases, we choose to normalize the expression levels -gene-by-gene -to the sample which shows the highest expression level for a given gene in the experiment ("max"). In , therefore, all three genes of the WT sample are set to a value of 1. Because we assay 3 genes, piling up the data (1+1+1) yields a value of 3 on the y axis for the WT cells. The KO values are simply relative to that. Like with expressing pPCR data as fold changes against a certain sample, the values of different genes do not reflect differences at the absolute expression scale.
As said before, we wish to stick to this way of presenting the data because it significantly reduces complexity while preserving a high information content and robustness. However, we also acknowledge that additional explanation may be useful. Hence, we have modified the axis labelling and chart of to better indicate how the data was normalised. In addition, we have now added an explanatory statement to the figure legend. We strongly hope R2 considers these measures sufficient.
The authors show "unbiased expression analysis of meso-permissive differentiation" in , however, gene expression data underlying this is not shown. This data should be included. Only select GO annotations are shown and the data for this GO annotation is not presented. Until this data is shown it remains possible that these genes are present in multiple GO terms and that those selected are not mesodermally restricted. The data in clearly show that mesoderm development is dramatically impaired in EOMES knockouts with dramatic reductions in MESP1, MIXL1, T and LHX1. Supporting the hypothesis that this EOMES is required for mesoderm formation rather than cardiac mesoderm specifically.
The underlying microarray experiment was a rather rough one -based on analyzing end point samples differentiated based on arbitrarily under or overdosing otherwise cardiac-promoting signaling factors. This was mainly to identify mesodermal marker genes in an unbiased manner -to then confirm these at the protein level, which is shown in previously revised . We think these data provide compelling evidence that EOMES disruption does not disable mesodermal differentiation capability in general. Admittedly, in light of the in vivo Eomes KO phenotype this may seem a bit surprising. We think the explanation may reside in the fact that signaling activity may be added exogenously in a cell culture setting, i.e. the cultured cells do not fully depend on producing these cues in an autocrine fashion -in contrast to Eomes KO embryos where our model would predict a severe lack of autocrine Wnt(3) activity.
We concur with the reviewer that showing the actual expression data in addition to the immunostainings of Suppl. may further increase the credibility of our data. Hence, the filtered expression data set underlying the GO analysis of is now shown in new Supplementary data file 1. |
Congenital Generalized Lipoatrophy (Berardinelli-Seip Syndrome) Type 1: Description of Novel AGPAT2 Homozygous Variants Showing the Highly Heterogeneous Presentation of the Disease
# Introduction
Congenital generalized lipodystrophy was first described by Waldemar and later on further outlined by Martin Seip.
Berardinelli-Seip congenital lipoatrophy (BSCL) is characterized by near total fat atrophy since birth, associated with the progressive development of metabolic complications. The most common are type 2 diabetes, severe hypertriglyceridemia, acute pancreatitis, hepatic steatosis, and hepatomegaly which are usually detected during infancy and adolescence; other features include, but are not restricted to, muscle pseudo-hypertrophy and acromegaloid appearance, umbilical hernia, polycystic ovary syndrome, cysts in the appendicular bones, cardiopathies, and cardiac rhythm disorders. Depending on the underlying molecular cause, mechanical adipose tissue depots of palms, soles, orbits, and under the scalp is preserved or notand other comorbidities can be observed. BSCL is classified in four different subtypes; BSCL type 1 is an autosomal recessive disease due to variants (9) of the gene coding for the enzyme 1-acylglycerol-3phosphate-Oacyltransferase β (lysophosphatidic acid acyltransferase beta or AGPAT2 OMIM# 608594). BSCL type 2 is caused by biallelic mutations in BSCL2 encoding seipin, a transmembrane protein involved in the functional relationships between endoplasmic reticulum and lipid droplets [OMIM# 269700)]. Rarer forms of BSCL are due to mutations of the CAV1 (OMIM# 612526) and PTRF (OMIM# 613327) genes, respectively, encoding caveolin-1 and cavin-1, belonging to the signaling platforms caveolae at the plasma membrane.
AGPAT2 is a lysophosphatidic acid acyltransferase isoform of 278 amino acids localized in the endoplasmic reticulum and highly expressed in white adipocytes. It catalyzes the transformation of lysophosphatidic acid (1,2-diacylglycerol-3phosphate) into phosphatidic acid (glycerol-3-phosphate) which belongs to the glycerophospholipid and triglyceride biosynthetic pathways. AGPAT2 deficiency leads to a depletion of triglycerides inside the adipose organ and to a defective signaling of key elements such as PI3K/AKT and PPARγ involved in proper adipogenesis. AGPAT2 KO mice develop a phenotype resembling that of BSCL type 1 (BSCL1) in humans characterized by near total loss of white and brown adipose tissue with severe insulin resistance, diabetes, and hepatic steatosis.
Abbreviations: BSCL, Berardinelli-Seip congenital lipoatrophy; AGPAT2, 1acylglycerol-3phosphate-O-acyltransferase β; PI3K, phosphoinositide 3-kinase; AKT, serine/threonine-protein kinases; OMIM, online mendelian inheritance in man; PPARγ, peroxisome proliferator-activated receptor gamma; CAV1, caveolin-1; PTRF, polymerase I and transcript release factor; KO, knockout mouse; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ExAC, exome aggregation consortium; HDL, high density lipoprotein-cholesterol; ECG, electrocardiogram; HbA1c, glycated hemoglobin; LH, luteinizing hormone; FSH, follicle-stimulating hormone; E2, estradiol; IGF-1, Insulin-like growth factor-1; DEXA, dual-energy x-ray absorptiometry; FT3, free tri-iodothyronine; CLIA, chemiluminescence immunoassay; FT4, free thyroxine; TSH, thyroid stimulating hormone; ICMA, chemiluminescent immunochemiluminometric assay; PRL, prolactin; GGT, gamma-glutamyl transferase; PCR, polymerase chain reaction; dbSNP, Single Nucleotide Polymorphism Database.
A very relevant aspect of the disease is the increased food intake caused by reduced circulating levels of the adipose-derived hormone leptin, which contributes to the metabolic alterations.
BSCL is a very rare disorder, its prevalence was estimated to be 1-10 case every 10 million, although it could reach 3 per 100,000 people in some areas. Some of the mutations recur in specific geographic areas, probably due to a "founder effect", and their presence is unveiled by consanguineous and endogamic marriages or in small and isolated communities.
Close to 95% of patients affected by BSCL have identified mutations. BSCL1 is an autosomal recessive disease and all patients exhibit either homozygous or compound heterozygous AGPAT2 mutations. Most reported mutations are nonsense or cause altered splicing or frame shifts, which disrupts the protein activity. Only few missense mutations have been so far described.
## Case report
Clinical Case N - 1 A 7 years-old child was referred to our Lipodystrophy Center. The boy was born at term after an uneventful pregnancy. His birth weight was 3.250 Kg. He was the third child of non-consanguineous healthy parents living in a small village in the Northern part of Italy. At the age of 3 months, he was admitted to hospital for vomiting and diarrhea. Physical examination showed a generalized lack of subcutaneous fat, abdominal distension, muscle hypertrophy, low anterior hairline, prominent orbital ridges, large ears, and umbilical hernia. Laboratory tests showed severe hypertriglyceridemia (1565 mg/dl) and increased liver enzymes (AST 543 U/L, ALT 667 U/L), breast feeding was halted and low fat medium chain triglycerides enriched milk was introduced with progressive normalization of the parameters. Liver ultrasound showed diffuse hyperechogenicity consistent with steatosis. BSCL1 syndrome was diagnosed when genetic analysis revealed a novel nonsense homozygous AGPAT2 pathogenic variant. In this patient a nucleotide change (c.430 C > T) was observed in exon 3 of the AGPAT2 gene predicting the substitution of the Glutamine residue at position 144 by a stop codon (p.Gln144 * ). This variant was not found in the ExAC nor in the 1,000 genome project databases and has never been described in patients affected by BSCL1. In silico analysis confirmed the high pathogenic score of the mutation. The allelic variant in the heterozygous form was found in both asymptomatic parents and in three additional family members of the maternal side.
A low-fat hypocaloric diet was then prescribed. At the time we first evaluated the patient (at 7 years of age) his development and school performance were appropriate for age. On physical examination his height and weight were between the 25th and 50th percentile of the growth charts. He was prepubertal. Lipoatrophy affected the entire body with the exception of the palm of the hands and sole of the feet. He presented muscle hypertrophy, particularly of the calf and thigh muscles. On low fat diet, biochemical analysis showed normal levels of fasting blood glucose, triglycerides, total cholesterol, and liver enzymes. He had low level of high density lipoprotein-cholesterol (HDL-cholesterol 28 mg/dl). As expected from the generalized lack of body fat, serum leptin levels were very low (0.1 ng/ml). Fasting insulin (4.4 mU/L) and glycated hemoglobin levels (33 mmol/mol) were in the normal range. Endocrinological evaluation showed normal thyroid, adrenal and pituitary functions. No pathological findings were demonstrated at cardiologic evaluation (ECG and echocardiography). Abdominal ultrasonography revealed mild steatosis. The hepatic left lobe volume, a standardized surrogate marker of liver volume, was enlarged (234 ml). Liver stiffness value assessed by fibroscan was normal (3.7 kPa).
## Clinical case n - 2
The second patient was a 53 years old woman born from consanguineous parents (second-degree cousins,living in a 200 residents village located in the Center of Italy. At 10 years, after development of polyuria and polydipsia she was diagnosed with lipoatrophic diabetes and eruptive xanthomatosis (triglycerides = 718 mg/dl). The patient was advised initially to restrict dietary fats and carbohydrate intake, but over the years she developed uncontrolled hyperglycemia and hyperphagia briefly treated with fenfluramine that exacerbated a bipolar disorder. Later on anti-hypertensive therapy (at age 40) and insulin therapy (at age 43), where administered after she had already developed bilateral proliferative retinopathy and blindness of the left eye because of retinal detachment. At 44 years, she was diagnosed with diabetic and hypertensive nephropathy. She had irregular periods (oligo-menorrhea) throughout her life and no pregnancies. She went into menopause at age of 49.
At the time we first evaluated the patient (53 years of age), physical examinationrevealed diffuse reduction of subcutaneous fat, acromegaloid features (marked prognathism and enlarged hands and feet), acanthosis nigricans on the neck and axillae. In addition, she displayed muscle hypertrophy, phlebomegaly and umbilical hernia. No signs of hirsutism were present. The height and the body mass index of the patient were 1.55 m and 22.1 kg/m 2 , respectively.
Biochemical blood tests, under insulin treatment (51 IU/day) and omega three fatty acids, showed high levels of fasting plasma glucose with normal glycated hemoglobin (178 mg/dl; HbA1c 37 mmol/mol, respectively) and frequent hypoglycemic episodes, hypertriglyceridemia (423 mg/dl), and low HDL-cholesterol levels (22 mg/dl). Serum creatinine and proteinuria were 2.14 mg/dl and 1.9 g/24 h, respectively. Endocrinological evaluation showed normal thyroid, parathyroid, and adrenal functions. Plasma leptin and adiponectin were severely reduced (0.4 ng/ml and 0.1 mcg/ml, respectively). Gonadal function showed low levels of gonadotropins with undetectable estradiol (LH 1.1 UI/l; FSH 4.6 UI/l; E2 <20 ng/L) and slightly increased levels of IGF-1 (231 mcg/L) and prolactin (71.4 mcg/l). Pituitary imaging excluded the presence of adenomas or suprasellar masses and hyperprolactinemia was attributed to the chronic treatment with risperidone.
Abdominal ultrasonography revealed mild liver steatosis with an enlarged hepatic left lobe volume (223 ml) and splenomegaly. Body composition assessed by DEXA showed reduced total fat mass (9.2%). Total body X-rays confirmed the presence of multiple bone cysts, confluent, with osteosclerotic appearance, mainly localized in the femoral heads and humerus diaphysis.
Her cardiovascular evaluation showed hypertension, dilated left atrial chamber, left ventricular hypertrophy but normal ventricular volume. Taking into account the high cardiovascular risk of this patient, a stress myocardial perfusion Single-Photon Emission Computed Tomography (SPECT) SPECT was conducted with negative results. The patient had been asymptomatic during the stress test.
Approximately 3 months after discharge from our institution, she experienced chest pain and was admitted to the emergency room and diagnosed with massive myocardial infarction and acute heart failure. Despite transcutaneous revascularization for diffuse coronary stenosis the patient developed progressive cardiac insufficiency and kidney failure. One month after her hospitalization the patient died in the intensive care unit.
In this patient, genetic screening of the AGPAT2 gene revealed a homozygous gene variant predicting the substitution of the Arginine residue at position 159 with a Cysteine (p.Arg159Cys, nucleotide change c.475C > G).
The missense variant p.Arg159Cysin exon 3 of the gene was found, among others, in the ExAC database with an allelic frequency of 0.00553 and in the 1,000 Genome Project with an allelic frequency of 0.003 in the general population. In silico analysis showed a high pathogenic score (180/215). This aminoacid is significantly conservedamong different species and resides in a highly conserved protein region. Protein stability of the missense variant p.Arg159Cys may be affected as predicted by dedicated software. This variant was found in the heterozygous state in her brotherwho had normal triglycerides, glucose tolerance, glycated hemoglobin, hepatic ultrasound, and total body fat assessed by DEXA. Genetic Testing, Mutation Screening AGPAT2 variants were identified by next generation sequencing (Patient 1) and/or Sanger sequencing (Patient 2), and all were confirmed by Sanger sequencing in an independent DNA sample. Specific primers were designed to amplify AGPAT2 exons and splice junctions from genomic DNA, isolated from whole blood. PCR was performed using PCR Master Mix (Promega Corporation, WI 53711-5399) with an annealing temperature of 60 - C. After purification with ExoProStar (GE Healthcare UK Limited, UK), the PCR products were directly sequenced using Applied Biosystems 3130 xl sequencer (Thermo Fisher Scientific, MA, USA).
## Biochemistry and hormones
## In silico modeling of the pathogenicity of agpat2 variants
In silico analysis was conducted using Mutationtaster software [http://mutationtaster.org/ (Last time accessed December 23, 2019)]: each variant is given a score up to the value of 215; a high score is indicative for pathogenic mutations. ExAC and 1,000 Genome Project, dbSNP, Ensembl, Exome Variant server, and in the Clinvar database databases were also searched to determine allelic frequencies in the general population. PROVEAN (Protein Variation Effect Analyzer) is a software tool which predicts whether an amino acid substitution or indel has an impact on the biological function of a protein (http://provean.jcvi.org/index. php, last time accessed December 23, 2019). DUET a server for predicting effects of mutations on protein stability via an integrated computational approach (http://biosig.unimelb.edu. au/duet/stabilityn, last time accessed December 23, 2019) was also used. Prediction from SIFT, Mutation Assessor, CAD and Revel were obtained from the Ensembl genome database.
# Discussion
BSCL1 pathogenic mutations affect triacylglycerol and glycerophospholipid synthesis in adipose tissue and cause lipodystrophy by impairing adipogenesis and depleting the adipocytes of triglycerides.
Patients with BSCL1 are either homozygous or compound heterozygous for AGPAT2 gene mutations that co-segregate with the phenotype of disease in accordance with an autosomal recessive pattern of inheritance. Heterozygotes carriers are asymptomatic although an increased incidence of diabetes is suggested.
We have identified 150 cases of patients with a genetically confirmed BSCL1 syndrome reported and 42 different mutations. Most of the mutations cause frame-shifts or alter mRNA splicing leading to the synthesis of a non-functional enzyme while rarer pathogenic variants cause amino acid substitutions. Mutations are spread throughout all six exons and intronic junctions of the gene. Approximately 1/4 of the mutations are nonsense mutations, more frequently present in exon 2 and 3. Insertions and deletions localize usually in exons 3 and 4 while over 80% of mutations altering mRNA splicing localize in intron 4 (p.Gln196fsX228 or p.Phe109fsX452 variants). Fewer patients (15%) present a compound heterozygosity for AGPAT2 pathogenic variants and each compound mutation, concentrating predominantly in introns or exons 3 or 4, is almost unique. No relationships between AGPAT2 pathogenic genotypes and the severity of lipodystrophy or specific complications has ever been described. Several are the metabolic complications related to AGPAT2 functional loss, insulin resistance is a hallmark of the disease and usually progresses during childhood (4) leading to diabetes mellitus after adolescence.
In the functional absence of AGPAT2, liver activity of the isoform 1 of the enzyme (AGPAT1) is significantly upregulated leading to overproduction of diacylglycerol, used for triacylglycerol synthesis. This could be an additional mechanism responsible for the development of severe liver steatosis and hepatomegaly besides the ectopic accumulation of triglycerides of alimentary origin.
Patients with BSCL1 display specific skeletal abnormalities, in particular lytic lesions at the extremities of appendicular bones associated with a diffuse serous transformation of the bone marrow, which can help clinicians with an early diagnosis in pauci-symptomatic patients.
Many patients with generalized lipodystrophy show evidence of cardiac hypertrophy, as measured by echocardiographic parameters. Lupsa et al.have documented that out of 19 patients affected by BSCL1, 10 had some evidence, in some cases severe, of left ventricular hypertrophy. ECG abnormalities were seen in 53% of patients. At variance with the numerous cardiovascular risk factors displayed by patients with congenital generalized lipodystrophy, myocardial infarction and sudden death were not reported to be the leading cause of decease.
The two cases presented illustrate well the clinical heterogeneity of BSCL1. In the younger patient, born from non-consanguineous parents, the metabolic complications, at the age of 8 years, were mild and controlled by hypocaloric diet and physical activity. The older woman, who was diagnosed during puberty, showed the multiple and severe metabolic derangements typically occurring in the natural history of the disease. Regarding the cardiac health, ECGs in both our patients were normal but echocardiography revealed dilated left atrial chamber and left ventricular hypertrophy in the adult patient, who also had hypertension. Taking into account the numerous and long standing cardiovascular risk factors of this 53 years old woman, we performed one of the most accurate test available for diagnosing coronary artery disease, which, however, failed to foresee the underlining ischemic disease. The etiology of the frequent hypertrophic cardiomyopathy in lipodystrophy remains unclear; in part, it could be related to the underlying hypertension, but also to the increased growth stimulatory action of excess insulin. A third potential mechanism of cardiac disease is related to the increased fatty acid influx or ectopic deposition of fat in the myocardium. "Lipotoxic cardiomyopathy" in lipodystrophy has not proved yet but is a very plausible explanation since a significant number of patients develop cardiac hypertrophy and left ventricular dysfunction in the absence of hypertension. Lipotoxic cardiomyopathy, together with the vascular complications of diabetes, hypertension and dyslipidemia, favor cardiovascular morbidity and mortality. Retrospectively, the adoption of an insulin treatment favoring hypoglycemic episodes rather than insulin sensitizers and drugs with a protective cardiovascular profile may have been further detrimental in this case. Furthermore, the role for implementation of a protective cardiac treatment (ace-inhibitors, anti-aldosterone drugs, anti-aggregants, PCSK9 inhibitors) should be assessed in adult patients with BSCL.
Another consequence of chronically decreased leptin levels is the reduction of GnRH pulsatile activity which causes oligomenorrhea. Indeed, our patient presented with a clinical history of oligo-menorrhea and showed almost undetectable gonadotropin levels, probably determined by the combination of reduced leptin levels and iatrogenic hyperprolactinemia. Notably, ∼10% of BSCL1 patients, display cognitive impairment, but our patients did not.
Described individuals with mutations in the AGPAT2 gene mainly originated from America or Northern Europe and occasionally from sub-Saharan Africa or Middle Eastern Countries. As for Caucasian Italian population the two herein described subjects add up to one other single case previously reported.
The first mutation (p.Gln144 * ) is not present in databases collecting variants identified in the general population.
In this case it is straightforward to predict that the mutation affects the protein activity via the synthesis of a truncated enzyme approximately half (143 amino acids) of the proper size. Missense mutations located close (p.Gly136Arg)or nonsense mutations more distal to the C-terminal part of the protein have been previously reported as pathogenic (e.g., p.Phe189X, p.Tyr190X, p.Lys216X, p.Gln226X).
The substitution of the Arginine residue at position 159 with a Cysteine (AGPAT2 p.Arg159Cys) was present in the homozygous state in Proband 2 and found in the heterozygous state in the unaffected brother. This variation alters a conserved amino acid site in a highly conserved region among mammalian species and its in silico score was predictive of disruption of protein activity. Furthermore, the homozygous variant segregated with the disease in the family making its pathogenic role possible.
We speculate that the substitution of hydrophilic charged Arginine with a polar, non-charged Cysteine, capable of forming disulfide bonds may disrupt the folding or activity of the enzyme. Results from dedicated software tools are not fully concordant in this regard: SIFT, Mutation Assessor, CAD, and PROVEAN predict in fact the aminoacid change to be neutral whereas Revel consider this substitution pathogenic, in line with the reduced stability of the R159C variant predicted by DUET, an integrated computational approach that takes into account the three dimensional structure of the protein.
Databases collecting variants existent in the general population report the heterozygous p.Arg159Cys to be present at a frequency around 0.3% which is higher than expected from the disease prevalence. The presence of few homozygous carriers (three subjects in ExaC and nine subjects in gnomAD) is reported by these databases but no information regarding the clinical history of the homozygous subjects can be retrieved. This evidence is difficult to reconcile with our findings. We therefore cannot rule out different pathogenic mechanisms. Pathogenic AGPAT2 missense mutations are thought to reduce protein expressionand possibly enzymatic activity. Testing in vitro the effects of the described mutation may give us further clues, but the assessment of the pathogenic potential of AGPAT2 variants has been proven challenging and has been rarely performed. Seip syndrome is frequently unveiled by consanguineous marriages or in small and relatively isolated communities. This was indeed the case for both patients. The first one was born from apparently unrelated parents living in a small village of few thousands inhabitants. The occurrence of a unique variant of AGPAT2 was found in members on both sides of the family, suggesting a founder effect from a common ancestor. The second patient was born from second-degree cousins.
In conclusion we describe two patients with BSCL type 1 lipodystrophy harboring novel homozygous variants. The first one presented with a null variant (p.Gln144 * ) occurring in exon 3 of AGPAT2 gene. The second patient presented with a homozygous AGPAT2 missense variant (p.Arg159Cys), never described previously in BSCL1 patients, raising the question on its pathogenicity. The segregation of the disease with the genotype in the family and the in silico analysis are compatible with a causative role of the p.Arg159Cys variant.
We remark that BSCL1 can be clinically very heterogeneous at presentation and that the associated complications, occurring in the natural history of the disease, may reduce lifeexpectancy. We point to the necessity for medical treatments capable of reducing the risk of cardiovascular death. In BSCL1 patients, the assessment of cardiovascular disease with conventional diagnostic means maybe particularly challenging. Future studies will be essential to investigate if specific cardiovascular diagnostic approach is recommended in adult BSCL patients.
# Data availability statement
All datasets generated for this study are included in the article.
# Ethics statement
Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. Written informed consent was obtained from the individual(s), and minor(s)' legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.
# Author contributions
GC, SM, FF, and FS carried out the characterization of the patients. CP, MS, GS, MM, and OL performed the genetic studies of AGPAT2 gene. GC, SM, and FS wrote the manuscript. IJ and CV reviewed and edited the manuscript. All authors read and approved the final version of the manuscript.
# Funding
This research was funded by the Italian Ministry of the University, Project code 2015JSWLTN: Metabolic flexibility and ectopic fat. Adiposity phenotype, mitochondrial dysfunction, hepatic inflammation, gut microbiota, cardiac failure, and genetics for a comprehensive understanding of the cross-talk among adipose tissue, liver, and musculo-skeletal system. IJ, OL, and CV were supported by Assistance Publique-Hôpitaux de Paris, Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (Inserm), Institut Hospitalo-Universitaire de Cardiométabolisme et Nutrition (ICAN), and Fondation pour la Recherche Médicale, Paris, France. |
Different Dimensions of Cognitive Style in Typical and Atypical Cognition: New Evidence and a New Measurement Tool
We developed the Sussex Cognitive Styles Questionnaire (SCSQ) to investigate visual and verbal processing preferences and incorporate global/local processing orientations and systemising into a single, comprehensive measure. In Study 1 (N = 1542), factor analysis revealed six reliable subscales to the final 60 item questionnaire: Imagery Ability (relating to the use of visual mental imagery in everyday life); Technical/Spatial (relating to spatial mental imagery, and numerical and technical cognition); Language & Word Forms; Need for Organisation; Global Bias; and Systemising Tendency. Thus, we replicate previous findings that visual and verbal styles are separable, and that types of imagery can be subdivided. We extend previous research by showing that spatial imagery clusters with other abstract cognitive skills, and demonstrate that global/local bias can be separated from systemising. Study 2 validated the Technical/Spatial and Language & Word Forms factors by showing that they affect performance on memory tasks. In Study 3, we validated Imagery Ability, Technical/Spatial, Language & Word Forms, Global Bias, and Systemising Tendency by issuing the SCSQ to a sample of synaesthetes (N = 121) who report atypical cognitive profiles on these subscales. Thus, the SCSQ consolidates research from traditionally disparate areas of cognitive science into a comprehensive cognitive style measure, which can be used in the general population, and special populations.
# Introduction
An individual's cognitive style refers to his or her preferred method for acquiring and processing information and this is regarded as relatively stable over time. Cognitive styles can encompass attitudes, preferences or strategies used by individuals that influence functions such as perceiving, remembering, thinking, and problem solving [bib_ref] The nature of cognitive styles: Problems and promise in educational practice, Messick [/bib_ref] (see [bib_ref] Cognitive styles in the context of modern psychology: Toward an integrated framework..., Kozhevnikov [/bib_ref] , for review). We aimed to develop a new questionnaire-the Sussex Cognitive Styles Questionnaire (SCSQ)-to investigate the interrelationships between visual and verbal cognitive styles and to incorporate recent findings on global/local processing biases and systemising. To preface our study, we describe below the different facets of cognitive styles that one might wish to capture in a comprehensive cognitive styles questionnaire. We then describe the design of our own questionnaire and the way we have validated it in different populations of typical and atypical individuals.
One traditional view of visual and verbal cognitive styles places those who prefer visual processing strategies and those who prefer verbal processing strategies as on opposite ends of a visual-verbal continuum [bib_ref] Verbalizer-Visualizer: A cognitive style dimension, Richardson [/bib_ref]. That is, visualisers who prefer to use visual mental imagery to process information are placed at one end of the continuum and verbalisers, who prefer verbal techniques, are on the other. The self-report Verbaliser-Visualiser Questionnaire (VVQ; [bib_ref] Verbalizer-Visualizer: A cognitive style dimension, Richardson [/bib_ref] was originally designed to classify people along this dimension. Recent neuroimaging results support the distinction between visualisers and verbalisers. For example Kraemer, Rosenberg and Thompson-Schill [bib_ref] The neural correlates of visual and verbal cognitive styles, Kraemer [/bib_ref] , demonstrated that participants who score highly on verbal preference exhibit greater activation in phonological regions when presented with pictorial stimuli, and those who score highly on visual preference exhibit greater activation in visual cortex when presented with written descriptions of visual stimuli (see also [bib_ref] Colour, context, and cognitive style: Variations in colour knowledge retrieval as a..., Hsu [/bib_ref] [bib_ref] Individual differences in cognitive style and strategy predict similarities in the patterns..., Miller [/bib_ref]. However, research has also demonstrated independence between visual and verbal dimensions on self-report measures [bib_ref] Verbal and visual learning, Kirby [/bib_ref] [bib_ref] Factor analysis of a questionnaire on imagery and verbal habits and skills, Paivio [/bib_ref] , and it is therefore theoretically possible for an individual to score high (or low) on both visualising and verbalising [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref].
Recent research has shown visual style can be further subdivided into two relatively independent imagery factors-object imagery and spatial imagery [bib_ref] Object-spatial imagery: A new self-report imagery questionnaire, Blajenkova [/bib_ref] [bib_ref] Spatial versus object visualisers: A new characterisation of visual cognitive style, Kozhevnikov [/bib_ref]. Object imagery refers to the subjective quality of cognitive representations in terms of resemblance to an object's form (encompassing, for example, size, shape, colour, brightness), whereas spatial imagery refers to the quality of representations in terms of spatial relationships between objects (encompassing, for example, location, movement, spatial transformation). Blajenkova et al. [bib_ref] Object-spatial imagery: A new self-report imagery questionnaire, Blajenkova [/bib_ref] developed the two-factor self-report Object-Spatial-Imagery Questionnaire (OSIQ) to capture dissociations between object and spatial imagery, with items such as "My images are very colourful and bright" loading on to the former factor and items such as "I was very good in 3D geometry as a student" loading on to the latter. Object imagery scores correlated with performance on a degraded pictures task (identifying a picture from partial information) and spatial imagery correlated with a performance on paper folding and mental rotation tasks, thereby demonstrating predictive validity of the measure. Blazhenkova and Kozhevnikov [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] subsequently developed the Object-Spatial Imagery and Verbal Questionnaire (OSIVQ) to include a third factor representing verbal style. This factor predicted performance on a task of generating sentences using a set of four nouns.
Spatial mental imagery, in particular, has been shown to correlate with another cognitive style, namely systemising [bib_ref] The relationship between systemising and mental rotation and the implications for the..., Brosnan [/bib_ref] [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] [bib_ref] Are systemising and autistic traits related to talent and interest in mathematics..., Morsanyi [/bib_ref]. Systemising refers to the drive to understand rules, regularities or variables that govern how a system works, or the tendency to construct systems to understand the world [bib_ref] Autism, hypersystemising, and truth, Baron-Cohen [/bib_ref] [bib_ref] Talent in autism: Hyper-systemising, hyper-attention to detail and sensory hypersensitivity, Baron-Cohen [/bib_ref] [bib_ref] The systemising quotient: An investigation of adults with Asperger syndrome or high-functioning..., Baron-Cohen [/bib_ref]. Systemising can be gauged via the self-report questionnaire called the Systemising Quotient (SQ; [bib_ref] The systemising quotient: An investigation of adults with Asperger syndrome or high-functioning..., Baron-Cohen [/bib_ref]. Behavioural studies have shown that systemising is correlated with the ability to mentally rotate 3D shapes [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] , and other aspects of mental rotation such as stimulus encoding and same-different comparison processes [bib_ref] The relationship between systemising and mental rotation and the implications for the..., Brosnan [/bib_ref]. Furthermore, scores on the short-form SQ [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] correlate well with the spatial imagery subscale of the OSIVQ (r = .61; [bib_ref] Are systemising and autistic traits related to talent and interest in mathematics..., Morsanyi [/bib_ref]. Spatial abilities such as mental rotation, or everyday activities such as map reading or solving the Rubik's cube, are thought to involve systemising as input features must be transformed via the application of rules in order to arrive at the correct outcome [bib_ref] Autism, hypersystemising, and truth, Baron-Cohen [/bib_ref]. The present research considers whether systemising and spatial imagery reflect a unitary cognitive style or two separate, but related, processes.
Systemising also presupposes excellent attention to detail [bib_ref] Autism: The empathising-systemising (E-S) theory, Baron-Cohen [/bib_ref] and as such bears at least a superficial similarity to a local processing bias. When attending to a complex stimulus, such as a scene, one can focus on it holistically or focus on its constituent parts. These can be thought of as global (holistic) or local (featural or analytic) processing biases respectively. Systemising is presumably dependent on sufficient local processing for relevant task details to be attended to and then acted upon. Indeed, Billington, Baron-Cohen and Bor [bib_ref] Systemising influences attentional processes during the Navon task: An fMRI study, Billington [/bib_ref] demonstrated that scores on the Revised Systemising Quotient (SQ-R; [bib_ref] Predicting Autism Spectrum Quotient (AQ) from the Systemising Quotient-Revised (SQ-R) and Empathy..., Wheelwright [/bib_ref] correlate with local bias in the Navon [bib_ref] Forest before trees: The precedence of global features in visual perception, Navon [/bib_ref] letters task (r = .45).
Global/local biases can influence visual processing more generally. For example, a global bias is positively correlated with performance on judgments of line orientation [bib_ref] Global-local visual biases correspond with visual-spatial orientation, Basso [/bib_ref] and inappropriate local processing orientations can disrupt recognition of complex visual stimuli such as faces and cars [bib_ref] Do I know you? Processing orientation and face recognition, Macrae [/bib_ref] [bib_ref] Verbal overshadowing in a multiple face presentation paradigm: Effects of description instruction, Brown [/bib_ref] [bib_ref] Verbal overshadowing of multiple face and car recognition: Effects of within-versus across-category..., Brown [/bib_ref]. Additionally, Bouvet, Rousset, Valdois and Donnadieu [bib_ref] Global precedence effect in audition and vision: Evidence for similar cognitive styles..., Bouvet [/bib_ref] found correlations between local and global processing across different modalities (e.g., a participant exhibiting a faster response time in identifying global patterns in visual stimuli was likely to exhibit a similar response time advantage for global audio stimuli, r = .55), and they suggest that global/local biases may be indicative of a particular cognitive style (see also [bib_ref] Investigating the stability of and relationships among global/local processing measures, Dale [/bib_ref] for a similar suggestion). Thus there is some support for the notion that global/local biases represent somewhat stable traits. The key difference between local processing and systemising is that systemising requires understanding the operations involved in a task and noting the consequences of actions [bib_ref] Talent in autism: Hyper-systemising, hyper-attention to detail and sensory hypersensitivity, Baron-Cohen [/bib_ref]. A local processing bias does not in and of itself presuppose a drive to understand the relations between elements in a problem, but does suggest a predisposition towards attention to detail. Thus it is not clear whether systemising tendencies and global/local bias would be driven by the same latent factor, or whether they are largely independent. Despite this, we are aware of no published self-report measure which gauges whether a person thinks of themselves as naturally inclined towards local or global processing, which we address here. Furthermore, the relationship between systemising and local processing, on the one hand, and visual-verbal measures of cognitive style, on the other, is essentially unknown.
We sought to design a comprehensive new questionnaire measure-the SCSQ-with items designed to capture self-reported imagery, systemising and global/local bias. We aimed to replicate the finding that visual and verbal styles can be divided, and that certain types of selfreported mental imagery can be further subdivided. We also aimed to extend the scope of such measures by additionally considering systemising and global/local bias. Specifically, we aimed to determine whether systemising and global/local bias items would or would not cluster together (cf [bib_ref] Systemising influences attentional processes during the Navon task: An fMRI study, Billington [/bib_ref] , and how these variables would relate to different forms of visual imagery (cf [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] , as it was unclear a priori whether (some) systemising and global/local bias items would cluster with aspects of visual imagery. Study 1 describes the new SCSQ, and presents factor and reliability analyses as well as the correlations between factor scores. Gender differences are also considered. Following the factor analysis and subscale establishment in Study 1, we aimed to validate the SCSQ in two further studies. Study 2 sought to validate relevant subscales of the SCSQ via correlation with performance on visual and verbal memory tasks in a subset of participants from Study 1.
Study 3 aimed to validate the SCSQ via testing on a group of unusual individuals. We asked whether the presence of certain forms of synaesthesia-which have been linked to particularly vivid imagery (e.g., [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref] [bib_ref] Synaesthesia is associated with enhanced, self-rated visual imagery, Barnett [/bib_ref] -would predict scores on SCSQ subscales. Individuals with synaesthesia have additional perceptual-like experiences ('concurrents') in the presence of certain materials ('inducers'), such as experiencing colours when reading or listening to music [bib_ref] Mechanisms of synaesthesia: Cognitive and physiological constraints, Grossenbacher [/bib_ref] , and we might therefore expect the presence of certain forms of synaesthesia to be predictive of visual, verbal, systemising and global/local processing preferences in particular ways compared to non-synaesthetes. We focus on two variants of synaesthesia. Firstly grapheme-colour synaesthesia, where letters and/or numbers result in additional colour experiences (e.g., the letter A might be crimson red; for review see [bib_ref] Beyond perception: Synaesthesia as a psycholinguistic phenomenon, Simner [/bib_ref] , and secondly sequence-space synaesthesia, where particular sequences (e.g., numbers, months, years) are consciously visualised as having spatial arrays (e.g., the months of the year might be experienced in an ellipse in front of the body, with January, say, to the right; for review see [bib_ref] Synaesthetic visuo-spatial forms: Viewing sequences in space, Simner [/bib_ref]. Following the results of the factor analysis in Study 1, we were able to make predictions about how synaesthetes' cognitive styles may differ from those of non-synaesthetes, and thereby validate the SCSQ subscales. We address this in Study 3.
## Study 1
The aim of Study 1 was to design a new cognitive styles questionnaire (the SCSQ). We then sought to assess its factor structure and reliability, explore the correlations between the factors, and investigate gender differences on the obtained factors.
# Method
Participants. One thousand five hundred and forty-two participants took part in our study and completed the questionnaire (956 female, 586 male). Ages ranged from 16 to 90 (M = 27, SD = 13). As the questionnaire only asks for birth year, not date of birth, the ages are estimated. Participants were recruited from the student population at the University of Sussex and from volunteers at the University of Edinburgh. Sussex participants (n = 370) were recruited in exchange for course credit. Edinburgh participants (n = 1172) were recruited as part of a large-scale, centrally co-ordinated undergraduate research project. Every student registered on the 2 nd year of the Psychology undergraduate course at the University of Edinburgh acted as a research assistant, and each recruited 8 participants (4 male and 4 female) over 16 years of age.
Ethics statement. The study procedures and methods of providing consent for all three studies in the current paper were approved by the local ethics committees at the Universities of Sussex and Edinburgh. On the first page of the questionnaire, participants were informed they would be asked to consent to taking part in the study. If they wished to participate, they continued to the second page and clicked a button, agreeing to complete the measure and for their data to be used. The Sussex cohort were undergraduates aged 18 or over. The Edinburgh cohort were recruited from student and non-student populations. The ethics committee who approved the study confirmed that, as per British Psychological Society guidelines, no specific consent procedures were required for prospective participants aged 16 and over, and these participants could therefore provide consent for themselves. Thus, [bib_ref] Are systemising and autistic traits related to talent and interest in mathematics..., Morsanyi [/bib_ref] [bib_ref] Autism, hypersystemising, and truth, Baron-Cohen [/bib_ref] year olds taking part in this study used the standard consent procedure provided to all participants.
Design of the questionnaire. Eighty-four items were initially included on the questionnaire (see S1 Table for the list of items). Previous questionnaires such as the OSIVQ [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] were developed on the basis that items should reflect preferences for the use of object, spatial and verbal processing in everyday life. Therefore, we aimed to select and generate items which could reasonably be said to reflect everyday events, activities or affairs for the SCSQ.
Twenty-two items were taken from the OSIVQ, 11 of which loaded on to their Object Imagery factor and 11 on to their Spatial Imagery factor. Four items were additionally taken from the 'Habitual Use of Imagery' subscale of the Individual Differences Questionnaire (IDQ; [bib_ref] Factor analysis of a questionnaire on imagery and verbal habits and skills, Paivio [/bib_ref]. The IDQ measures imagery and verbal thinking habits. The Habitual Use of Imagery subscale was most relevant to our research question (i.e. processing preferences in everyday life). Twenty-four items were taken from the SQ to assess systemising [bib_ref] The systemising quotient: An investigation of adults with Asperger syndrome or high-functioning..., Baron-Cohen [/bib_ref]. Seven items reasonably related to systemising or global/local biases were additionally taken from the 'Attention to Detail' subscale of the Autism Quotient (AQ; [bib_ref] The autism-spectrum quotient (AQ): Evidence from Asperger syndrome/high-functioning autism, males and females,..., Baron-Cohen [/bib_ref].
Twenty-seven further items were generated by us to measure how global/local processing bias, systemising, imagery, and verbal habits may be used in everyday life. In particular, we generated five items concerned with the visual appearance of facets of language (e.g., I tend to notice if a word has the same letter repeated in its spelling). We did this to determine whether verbal items that focus on the visual appearance of written language would form a separate factor from visual imagery items more generally. All items were measured on a five point Likert scale (from 'strongly disagree' to 'strongly agree', with the midpoint being 'neither agree nor disagree') in line with the measurement scale of the OSIVQ (note the original AQ and SQ are measured on four point scales). The IDQ was originally measured on a true/false response scale, but performs well when a five-point scale is used instead [bib_ref] Factor analysis of a questionnaire on imagery and verbal habits and skills, Paivio [/bib_ref] [bib_ref] Imagery assessment through self-report: What do imagery questionnaires measure?, Hiscock [/bib_ref]. The presentation order of items was randomised, but fixed between participants (see S1 .
Procedure. Questionnaire data were collected via the online portal Bristol Online Surveys. All participants completed the SCSQ in a location of their choosing, aside from 44 Sussex participants who completed it under laboratory conditions (as part of Study 2). After providing consent to participate on the first and second web page, participants were asked to enter their demographic information (gender, birth year, level of achieved formal education, nationality). All questionnaire items were presented on the next webpage. The final page gave a brief description of grapheme-colour and sequence-space synaesthetic experiences and asked the question-"do you suspect you have synaesthesia?". This final question was for the purposes of screening out potential synaesthetes at this stage from the initial factor structure analysis as this forms the basis of a separate study.
# Results and discussion
Of the 1542 participants who completed the measure, 87 (5.6%) reported they thought they may experience some form of synaesthesia and these participants were excluded from further consideration. The analyses were conducted on the remaining 1455 participants (907 females, 548 males). Age data were available for 1437 of these participants (range 16-90 years; M = 27; SD = 13). Data are included in S1 File.
Exploratory factor analysis. Exploratory factor analysis (EFA) with weighted least squares estimation was conducted on all 84 items to determine the model that best described the data. Spearman correlations were used in the analysis due to the ordinal nature of the items. Initially, we determined the number of factors to extract using Velicer's [bib_ref] Determining the number of components from the matrix of partial correlations, Velicer [/bib_ref] Minimum Average Partial (MAP) method. This suggested a nine factor solution at this stage. Thus an EFA was specified with a nine factor solution, with an oblique rotation (Oblimin-Quartimin) to allow for correlations between factors. Oblimin rotations parametrise the obliqueness of the rotated factors. Quartimin rotations are a special case, where the solution is most oblique by minimising the product of the squared loadings on each factor. Items with factor loadings > .30 were considered worthy of retention. Seventeen items did not load at .30 at this stage and were removed from further analysis (Q03, Q12, Q13, Q14, Q34, Q35, Q37, Q39, Q49, Q50, Q68, Q70, Q74, Q75, Q76, Q79, Q83). Five further items were removed for cross-loading on more than one factor (Q26, Q54, Q63, Q66, Q82).
A second MAP analysis was conducted, suggesting a six factor solution to the remaining 62 items. This resulted in a clear and interpretable factor structure. Regarding this second EFA, the Kaiser-Meyer-Olkin measure of sampling adequacy suggested a factorable solution (KMO = .87). At this stage, two remaining items did not load at the .30 level on to the obtained factors and were removed from further consideration (Q30, Q80). [fig_ref] Table 1: Rotated factor matrix of factor loadings for the final sixty items of... [/fig_ref] presents the results of the EFA. The reduced version of the questionnaire contained sixty items, which explained 32% of the total variance. See [fig_ref] Table 2: Correlation coefficients [/fig_ref] the inter-item correlation matrices for the six obtained factors.
Factor 1 contained items mostly related to the use of visual imagery in everyday life (e.g., "I often use mental images or pictures to help me remember things" from the IDQ), with some items which gauge the subjective quality of visual imagery (e.g., "My mental images are very vivid and photographic" from the OSIVQ), hence this factor was termed 'Imagery Ability'. Factor 2 contained items relating to spatial mental imagery (e.g., "I can easily imagine and mentally rotate three-dimensional geometric figures" from the OSIVQ), but also items relating to mathematics abilities (e.g., "I am fascinated by numbers" from the AQ) and items relating to technical interests (e.g., "If I were buying a computer, I would want to know exact details about its hard drive capacity and processor speed" from the SQ), thus this factor was termed 'Technical/Spatial'. All items on Factor 3 related to aspects of language. In particular most of these items indicate an interest in the visual appearance of written language as opposed to oral verbal ability (e.g., "When I hear a new word, I am curious to know how it is spelled", a new item; "When I read something, I always notice whether it is grammatically correct", from the SQ). These items clearly formed their own factor separate from visual imagery. Therefore this factor was termed 'Language & Word Forms'.
Factor 4 contained items relating to organisation (e.g., "If I had a collection (e.g., CDs, coins, stamps), it would be highly organised" from the SQ) and order (e.g., "Order is important to me", a new item) and was termed 'Need for Organisation'. Negatively loaded items on Factor 5 indicate attention to detail or a local processing preference (e.g., "I tend to focus on details in a scene rather than the whole picture", a new item), whereas positively loaded items indicate a more holistic preference (e.g., "I usually concentrate on the whole picture, rather than the small details" from the AQ), therefore this factor was termed 'Global Bias'. Finally, items on Factor 6 refer to the drive to categorise (e.g., "When I look at an animal, I like to know the precise species it belongs to" from the SQ), or interest in systems (e.g., "I am fascinated by dates" from the AQ) and was termed 'Systemising Tendency'. One item ("When I look at a tree I focus on its features such as branches and leaves rather than the whole") cross-loaded onto two factors: Factors 5 and 6. This item better captures attention to detail or local bias and was therefore included on Global bias (Factor 5). Interestingly, as both the globally and locally phrased items loaded on to the same factor as opposed to two different factors, it seems that these preferences are at opposite ends of a spectrum in terms of self-report.
Reliability analysis. Next, we assessed the reliability of the obtained factors via ordinal α (see [fig_ref] Table 1: Rotated factor matrix of factor loadings for the final sixty items of... [/fig_ref]. All factors exhibited good levels of reliability (all α .73). Next we consider improvements to α with the removal of items. The reliability of Imagery Ability would not improve with the removal of any item. The reliability of Technical/Spatial would increase from .885 to .887 with the removal of Q48. The reliability of Language & Word Forms would increase from .798 to .804 with the removal of Q1. The reliability of Need for Organisation would increase from .770 to .776 with the removal of Q38. The reliability of Global Bias would increase from .736 to .746 with the removal of Q27. The reliability of Systemising Tendency would not benefit from the removal of any item. As the removal of any item would only result in minor improvement in ordinal α ( .01) all items were deemed worthy of retention.
Next we calculated the mean scores for each factor (negatively loading items were reverse coded prior to calculation; see [fig_ref] Table 1: Rotated factor matrix of factor loadings for the final sixty items of... [/fig_ref] for mean factor scores). [fig_ref] Table 2: Correlation coefficients [/fig_ref] displays the correlation coefficients between the six obtained factors. Although most factors tended to correlate the effect sizes were small (rs < .30) in all but one case (based on the categorisation of effect sizes by Cohen [bib_ref] A power primer, Cohen [/bib_ref]. Global Bias and Systemising Tendency correlated negatively, and one might expect sufficient local processing to systemise effectively. Systemising Tendency correlated positively with Technical/Spatial, showing that spatial skills and systemising are related. Technical/ Spatial also correlated negatively with Global Bias. It is not surprising that Systemising Tendency and Technical/Spatial were correlated; one might expect that sufficient local detail is needed for spatial transformation and attention to detail in systemising domains. Imagery Ability also correlated negatively with Global Bias.
Gender differences. Next we assessed gender differences on SCSQ factor scores. Mean scores for females and males were calculated and compared via independent samples t-tests .
Consistent with Blazhenkova and Kozhevnikov [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] , females scored higher on Imagery Ability (similar to their object imagery factor) and males scored higher on Technical/Spatial (similar to their spatial factor, with additional mathematics and technical interest items). Similar results for imagery according to gender are also presented by Campos [bib_ref] Gender differences in imagery, Campos [/bib_ref]. Males also scored higher than females on Systemising Tendency, consistent with Baron-Cohen et al. [bib_ref] The systemising quotient: An investigation of adults with Asperger syndrome or high-functioning..., Baron-Cohen [/bib_ref] , Ling et al. [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] and Wheelwright et al [bib_ref] Predicting Autism Spectrum Quotient (AQ) from the Systemising Quotient-Revised (SQ-R) and Empathy..., Wheelwright [/bib_ref].
## Study 2
As the SCSQ used items pertaining to the everyday use of dimensions on the measure, the goal of Study 2 was to validate the visual and verbal subscales of the new measure by correlating mean factor scores with performance on modified versions of tasks taken from a standard memory battery, the Wechsler Memory Scale Revised (WMS-R;. Specifically, we adapted the visual paired associates and verbal paired associates tasks. Both tasks included additional trials to prevent ceiling effects in the current student sample (the original versions of these tasks are used in clinical samples), and both tasks were computerised using E-Prime 2 software. We also sought to validate the measure with an additional visual recognition task using fractal images as stimuli. The fractal image recognition task has previously been used by Ward, Hovard, Jones and Rothen [bib_ref] Enhanced recognition memory in grapheme-colour synaesthesia for different categories of visual stimuli, Ward [/bib_ref]. The task was chosen here because the visual stimuli are hard to verbalise and performance on the task tends to be overall lower and more variable than those used in the WMS-R (which was designed to measure memory impairment).
# Method
Participants. Forty-four participants recruited from the University of Sussex (37 female, 7 male) who had provided data for Study 1 completed these additional tasks. Ages ranged from 18 to 42 years (M = 21; SD = 4). All had normal or corrected to normal vision and had a good level of English. None indicated they experienced any form of synaesthesia. Prior to the task, participants signed a consent form, thereby providing written consent to take part in the memory tasks. Consent to complete the questionnaire was acquired in the same way as Study 1.
Materials and procedure. The visual paired associates task requires participants to remember specific colours associated with monochrome line drawings. The original task utilises six of these colour-shape pairs. We included four additional pairs, meaning ten pairs were used in total. Each pair consisted of two boxes, with the shape on the left hand side and a square of colour on the right hand side. The screen background colour was white. The task alternated between three learning blocks and three memory blocks with a final delayed memory block which was completed after a retention interval of approximately thirty minutes. During the learning blocks, each colour-shape pair was presented for 3 seconds, in the centre of the screen (with a 3 second inter-trial interval). Here, participants were instructed to memorise the pairings. During the memory blocks, a shape prompt appeared in the top third of the screen. All ten colours appeared in a row at the bottom third of the screen, which corresponded to keys 1 to 0 on the top number pad of the keyboard. Participants were required to pick the colour associated with the shape via the corresponding key input (this response was unspeeded), and the next prompt appeared. This continued for all ten pairs. Corrective feedback was provided onscreen (duration of 1.5 seconds). Participants completed the delayed block without an additional learning block, and no feedback was provided during this block. Presentation order of pairings was pseudorandomised between participants. The verbal paired associates task required the participants to associate two words in memory. The original version of the task includes eight pairs of concrete nouns. We included four additional pairs. The task alternated between four learning blocks and four memory blocks. During the learning blocks, participants heard the twelve word pairs spoken by a male voice through headphones (with a 3 second inter-trial interval). Here, participants were instructed to memorise the pairings. During the memory blocks, they were prompted with one of the word pairs and were required to type its associate from the learning blocks (prompts and associates were the same throughout). Corrective feedback was provided onscreen (duration of 1.5 seconds). Participants completed the delayed block without an additional learning block, and no feedback was provided during this block. Presentation order of pairings was pseudorandomised between participants.
The fractal recognition stimuli were similar to those used in a previous study [bib_ref] Enhanced recognition memory in grapheme-colour synaesthesia for different categories of visual stimuli, Ward [/bib_ref]. Sixty greyscale fractal images were used in total. Stimulus size was 400 x 400 pixels. The images were split into two sets of thirty. One set was used as target stimuli and the other set as distractors (counterbalanced so each list was used as targets or distractors for 50% of the participants). Participants were shown the target images during an encoding phase. Each image was presented for 1 second. Participants were instructed to press N on the keyboard as soon as the stimulus disappeared, to maintain attention. A new stimulus was presented upon key press. During the test phase, target and distractor images were randomly intermixed between participants. Participants had to indicate whether each image was 'old' or 'new' via a button press on the keyboard (Z and M, respectively). At test, images were presented for one second, and the next image was presented after the participant's response.
Participants completed the associative learning tasks first (half completed visual first, half completed verbal first) before the fractal recognition task. After the retention interval (during which they performed visual psychophysics tasks) they completed the delayed blocks of the associative learning tasks in the same order as the learning blocks. Finally they completed the full, original 84 item SCSQ (note they only completed the questionnaire once).
# Results and discussion
Data are included in S2 File. Firstly, we consider performance on the memory tasks before investigating the relationships between memory task performance and scores on the SCSQ subscales (note that we do not consider gender in the following analyses due to the small number of males in the sample). The proportions of correct responses for the verbal and visual paired associates tasks were calculated (see [fig_ref] Table 4: Mean proportion of correct responses over blocks for the visual and verbal... [/fig_ref] for descriptive statistics). Considering performance on the visual paired associates task, there was a significant linear trend over the learning blocks, F(1, 43) = 66.78, p < .001, η 2 p = .61, indicating performance improved over the blocks, and performance in each block differed significantly from performance in every other block, ps .001. Considering performance on the verbal paired associates task, there was a significant linear trend over the learning blocks, F(1, 43) = 80.15, p < .001, η 2 p = .65, indicating performance improved over the blocks, and performance in each block differed significantly from performance in every other block, ps .002. Thus we can be confident that learning occurred during both the visual and verbal paired associates tasks.
Performance in the fractal recognition task was measured via the signal detection measure, d'. This statistic represents the difference between standardised hit rate (correctly endorsing test stimuli present during learning) and false alarm rate (erroneously endorsing test stimuli not present during learning). Above chance recognition is shown if d' is greater than zero. Mean d' was 0.92 (SE = .09), which was significantly greater than zero, t(43) = 10.33, p < .001. Thus sufficient discrimination between targets and lures was demonstrated.
We then correlated task performance, averaging across the initial learning blocks in the paired associates tasks, with scores on the questionnaire subscales (see . Technical/Spatial scores correlated significantly with average visual associates performance, visual associates delayed performance, and fractal recognition d'. Language & Word Forms correlated significantly with average verbal associates performance and verbal associates delayed performance. Thus these factors demonstrated convergent validity. Perhaps surprisingly, Imagery Ability did not correlate significantly with performance on these tasks. In the visual associates task, participants were required to associate abstract line drawings with colours and in the fractal recognition task were required to recognise abstract images, thus performance on these tasks may be plausibly related to more abstract representational styles (captured by Technical/Spatial scores), rather than concrete representational style (captured by Imagery Ability scores). Convergent validity of the Imagery Ability factor could potentially be demonstrated using more naturalistic stimuli (such as natural scenes). Global Bias correlated negatively with performance in the learning blocks of the Visual Paired Associates task, indicating better performance was related to local bias. However, this factor did not correlate with performance in the delayed block. This could indicate that local bias confers a benefit to memory encoding when associating complex, abstract shapes with other visual materials. It may also reflect a Type I error (given the number of analyses performed we would predict 1.5 significant results by chance). Although most results would not survive a Bonferroni correction for multiple comparisons, the significant results represent medium effect sizes [bib_ref] A power primer, Cohen [/bib_ref] , generally form a stable pattern across measures, and are in line with theory.
## Study 3
In Study 3 we sought to validate the SCSQ subscales identified in Study 1 by issuing it to a sample of participants with atypical experiences, namely grapheme-colour and sequence-space synaesthetes. It has been suggested that synaesthetes' cognitive preferences may reflect affinities for materials associated with their synaesthesia, either in the domain that induces the experience, or the domain of the experience itself [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref]. Thus, in the case of grapheme-colour synaesthesia we would expect a higher rating than that of non-synaesthetes on verbal and visual styles (here, Language & Word Forms and Imagery Ability respectively) because the inducer is verbal and the concurrent is visual.
Although the bulk of research on imagery, cognitive styles and individual differences in synaesthesia has focused on grapheme-colour variants (e.g., [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref] [bib_ref] Mechanisms of synaesthesia: Cognitive and physiological constraints, Grossenbacher [/bib_ref] [bib_ref] Mental imagery and synaesthesia: Is synaesthesia from internally generated stimuli possible, Spiller [/bib_ref] , recent work has begun to address imagery in sequence-space synaesthesia. For example, since sequence-space synaesthetes appear to be superior in tasks of mental rotation (e.g., [bib_ref] A foundation for savantism? Visuo-spatial synaesthetes present with cognitive benefits, Simner [/bib_ref] [bib_ref] Enhanced mental rotation in time-space synaesthesia, Brang [/bib_ref] ; but see [bib_ref] Do sequence-space synaesthetes have better spatial imagery skills? Maybe not, Rizza [/bib_ref] we might predict they would score highly in our questionnaire on spatial imagery. Interestingly, these synaesthetes had not scored highly on previous questionnaires of spatial imagery [bib_ref] Spatial forms and mental imagery, Price [/bib_ref] , which conflicts with their performance in the equivalent behavioural task (e.g., rotation; [bib_ref] A foundation for savantism? Visuo-spatial synaesthetes present with cognitive benefits, Simner [/bib_ref]. It may therefore be possible that previous self-report questionnaires were failing to detect spatial skills in this group. This may be because spatial imagery forms part of constellation of abilities, as shown by the Technical/Spatial factor, and only by taking into account related abilities would we observe the posited advantage. Additionally, as spatial abilities and systemising are correlated [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] , we sought to determine whether sequence-space synaesthetes in particular would also score highly on Systemising Tendency. We therefore look with interest at how sequence-space synaesthetes might perform in our own questionnaire, particularly in relation to spatial imagery and systemising. Such results would help verify these subscales of the SCSQ.
# Method
Participants. We tested 121 synaesthete participants, 78 with sequence-space synaesthesia, 22 with grapheme-colour synaesthesia and 21 with both. Participants were recruited by email from a database of synaesthete volunteers at the University of Sussex (these were not the same self-declared synaesthetes excluded from the analysis in Study 1). The consent procedure was the same as used as in Study 1.
All individuals completed a questionnaire detailing their synaesthetic experiences, and participants were classified as having sequence-space experiences on the basis of answering affirmatively to the question "Some people always experience sequences in a particular spatial arrangement. Do you think this applies to you?". Example diagrams of spatial forms were given when answering this question. All indicated having sequence-space synaesthesia for at least one of the following sequences: Numbers, days, months, years, letters of the alphabet, temperature, height, weight. Grapheme-colour synaesthetes were first identified from a question asking whether they experience coloured letters or digit, but were further validated using the most widely available online testing tool hosted at www.synesthete.org (for detailed methods see [bib_ref] A standardised test battery for the study of synaesthesia, Eagleman [/bib_ref]. In this test, participants are shown each grapheme (letters A-Z, digits 0-9) three times in a random order and must select their synaesthetic colours from an online palette of colours. Synaesthetes typically show very high internal consistency in this test (e.g., repeatedly giving the same shade of red for the letter A) and this metric is used to generate a standardised colour-distance score. A score below 1.43 indicates the consistency typical of grapheme-colour synaesthetes (e.g., [bib_ref] Diagnosing synaesthesia with online colour pickers: Maximising sensitivity and specificity, Rothen [/bib_ref] [bib_ref] Validating a standardised test battery for synaesthesia: Does the Synaesthesia Battery reliably..., Carmichael [/bib_ref]. Our 43 grapheme-colour synaesthetes all scored 1.43 (M = 0.76, SD = 0.25).
The mean age of our synaesthete participants was 35 years (SD = 15) and 80% were female. To match the sample in terms of gender, we took a stratified random sample from the control group (Study 1 participants) to yield a gender ratio of 80:20 female:male participants. To maximise use of the data, nine gender-matched controls were randomly selected for each synaesthete (874 females, 216 males).
Materials and procedure. Synaesthetes completed the SCSQ via Bristol Online Surveys in the same manner reported for Study 1. Mean factor scores were calculated for each participant for each subscale of the SCSQ.
# Results and discussion
Data are included in S3 File. The results are summarised in [fig_ref] Table 6: Mean factor scores as a function of presence/absence of grapheme-colour and sequence-space... [/fig_ref]. The three groups of synaesthete are defined by the presence of either or both types of synaesthesia and the controls are defined by the absence of both.
We conducted separate multiple regression analyses to determine whether the presence of different types of synaesthesia predicted scores on each of the SCSQ subscales. It is important to note that individuals with synaesthesia often have multiple forms (e.g., [bib_ref] Synaesthesia: The prevalence of atypical cross-modal experiences, Simner [/bib_ref]. By specifically controlling for this in the regression analyses, we can address which forms of synaesthesia contribute most to different forms of cognitive styles. The presence/absence of grapheme-colour and sequence-space synaesthesia were coded as 0 (absent) and 1 (present) and entered into regression models as binary predictors for each SCSQ subscale. Importantly, Study 1 revealed gender differences on the Imagery Ability, Technical/Spatial, and Systemising Tendency subscales. Therefore we entered gender as a third predictor into each regression to assess whether effects of synaesthesia were detectible when accounting for gender differences. Females were coded as 1 and males as 0. [fig_ref] Table 7: Regression weights [/fig_ref] shows the results of the regression analyses. Both grapheme-colour synaesthesia and sequence-space were linked to increased Imagery Ability and increased Language and Word forms scores. Sequence-space synaesthesia, but not grapheme-colour synaesthesia, was linked to increased Technical/Spatial, increased Systemising Tendency and reduced Global Bias (i.e. a greater local bias) scores. The differences observed between grapheme-colour and sequence-space synaesthetes on SCSQ scales shows that different forms of synaesthesia may predict different aspects of cognition, and should be accounted for in future research. The relationship between gender and SCSQ scores reflects that observed in Study 1, even when accounting for synaesthesia. Females scored higher on Imagery Ability, and males scored higher on both Technical/Spatial, and Systemising Tendency. Thus gender and the presence of different forms of synaesthesia make independent contributions to SCSQ scores.
# General discussion
The current research aimed to produce a new cognitive styles questionnaire-the SCSQ-that links previously unconnected domains: the verbaliser-visualiser tradition alongside global/local processing and systemising. To achieve this, we issued the SCSQ to a large number of participants and assessed its factor structure and reliability. We found six factors: Imagery Ability, Technical/Spatial, Language & Word Forms, Need for Organisation, Global Bias, and Systemising Tendency which all exhibited good levels of internal consistency. We also found generally low to modest, but statistically significant, correlations between almost all of the factors.
We replicated the finding that visual and verbal styles are separable [bib_ref] Verbal and visual learning, Kirby [/bib_ref] [bib_ref] Factor analysis of a questionnaire on imagery and verbal habits and skills, Paivio [/bib_ref] and that visual imagery is not a unitary construct, but can be subdivided further [bib_ref] Object-spatial imagery: A new self-report imagery questionnaire, Blajenkova [/bib_ref] [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref]. Here we found Imagery Ability and Technical/Spatial factors, when new items were included in the measure. We also found that global/local bias items formed their own factor, as did systemising items. Notably, each factor contained items from more than one source (i.e. the factor analysis did not simply cluster items according to their original measure). Furthermore, all factors (with the exception of Need for Organisation) demonstrated construct validity. Firstly Study 1 revealed gender differences on SCSQ subscales. Secondly SCSQ scores predicted performance on visual and verbal memory tasks in Study 2, in-line with previous findings that these cognitive styles predict objective task performance [bib_ref] The neural correlates of visual and verbal cognitive styles, Kraemer [/bib_ref] [bib_ref] Colour, context, and cognitive style: Variations in colour knowledge retrieval as a..., Hsu [/bib_ref] [bib_ref] Individual differences in cognitive style and strategy predict similarities in the patterns..., Miller [/bib_ref]. Thirdly, SCSQ scores differed according to the presence and absence of synaesthesia in Study 3.
The Imagery Ability factor is conceptually similar to the object visual factor from the OSIVQ containing a number of items related to detail in object imagery. Indeed, this factor contained six items from this subscale (and one negatively loading item from the spatial subscale of the OSIVQ). The Imagery Ability factor also contained items from a number of other sources-four items from the IDQ and six new items designed to capture the use of imagery in everyday life (e.g., "When I can't find something I'm looking for, I automatically visualise the last place I saw it"). Thus it may be the case that self-rated object imagery is tightly linked with its use in everyday life.
Our Technical/Spatial factor focuses on spatial imagery and other activities based on abstract cognitive skills such as mathematical abilities and technical interests. This subscale contained seven items from the spatial subscale of the OSIVQ, but also contained six items from the SQ and two from the AQ, thus this latent factor captures more than spatial mental imagery. That mathematical items clustered with spatial imagery items perhaps is not surprising; mathematical problem solving has previously been shown to correlate positively with spatial imagery and negatively with pictorial imagery (e.g., [bib_ref] Types of visual-spatial representations and mathematical problem solving, Hegarty [/bib_ref] ; see [bib_ref] Visualising numbers in the mind's eye: The role of visuo-spatial processes in..., De Hevia [/bib_ref] , for review) and latent factors driving performance on spatial imagery tasks could covary with mathematical ability (see [bib_ref] Visualising numbers in the mind's eye: The role of visuo-spatial processes in..., De Hevia [/bib_ref] for the neuro-anatomical basis of numerical and spatial processes). Interestingly, previous factor analyses of the SQ show that items relating to interest in technology form their own component of systemising [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref]. That items relating to an interest or concern with technical details also loaded on the Technical/Spatial factor also indicates a more general capacity for dealing with abstract information.
In terms of scores for our imagery factors, females rated themselves higher on Imagery Ability than males, and males rated themselves higher on Technical/Spatial than females. The findings with respect to gender and different forms of imagery are in good agreement with other studies [bib_ref] Object-spatial imagery: A new self-report imagery questionnaire, Blajenkova [/bib_ref] [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] [bib_ref] Gender differences in imagery, Campos [/bib_ref] , and thus validate the imagery subscales of the SCSQ.
Both the presence of grapheme-colour and sequence-space synaesthesia significantly predicted scores on the Imagery Ability factor. This supports previous research showing these groups rate themselves highly on object imagery (e.g., [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref] [bib_ref] A foundation for savantism? Visuo-spatial synaesthetes present with cognitive benefits, Simner [/bib_ref] [bib_ref] Enhanced mental rotation in time-space synaesthesia, Brang [/bib_ref]. Additionally, sequencespace synaesthesia predicted scores on the Technical/Spatial factor where previous research failed to find an advantage for this group in terms of self-report spatial imagery [bib_ref] Do sequence-space synaesthetes have better spatial imagery skills? Maybe not, Rizza [/bib_ref] [bib_ref] Spatial forms and mental imagery, Price [/bib_ref] , despite others' sometimes showing greater objective spatial imagery skills (e.g., [bib_ref] A foundation for savantism? Visuo-spatial synaesthetes present with cognitive benefits, Simner [/bib_ref]. Our finding of superior scores in the Technical/Spatial factor is more in-line with most behavioural findings in this domain. Additionally, this may reflect the factor's capture of abilities correlated with spatial processing, such as mathematics, on which synaesthetes may also differ. For example, 77% of a sample of sequence-space synaesthetes reported consciously using their number forms in calculations [bib_ref] The impact of visuo-spatial number forms on simple arithmetic, Ward [/bib_ref].
Language & Word Forms contained one item (pertaining to language) from the SQ and five new items. Most of the items on this subscale relate to the appearance of words and sentences (such as spelling and grammar structure), thus this factor emphasises orthographic qualities. That all language items formed their own factor supports the tripartite model of visual imagery and verbal skills (e.g., [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref] , even though most of the verbal items used in the SCSQ we used were novel and de-emphasised oral aspects of verbal skills. In addition to Imagery Ability scores, the presence of grapheme-colour synaesthesia predicted scores on Language & Word Forms. This supports the hypothesis that synaesthetes' cognitive styles relate to inducing-or concurrent-related information, i.e. verbal and visual information in the case of grapheme-colour synaesthesia, replicating findings by Meier and Rothen [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref] who issued the VVQ to grapheme-colour synaesthetes. That the SCSQ performed similarly to the VVQ offers further validation. Interestingly, the presence of sequence-space synaesthesia also predicted scores on Language & Word Forms. Notably, letters of the alphabet are often cited as inducing spatial forms in sequence-space synaesthetes. This also fits well with Meier and Rothen's [bib_ref] Grapheme-colour synaesthesia is associated with a distinct cognitive style, Meier [/bib_ref] suggestion that synaesthetes' cognitive styles reflect a particular affinity for their inducing materials.
Need for Organisation emerged as a factor, which contained one item from the SQ and five new items, and was not predicted a priori. Neither gender, nor the presence of either form of considered synaesthesia predicted scores on this factor. Future validation work could explore this factor's relationship to orderliness factors on other questionnaire measures (e.g. Comrey Personality Scales, [bib_ref] Factor analysis of the NEO-PI and The Comrey Personality Scales, Hahn [/bib_ref].
Global bias contained three items from the AQ, four new items and one item from the OSIVQ object factor (which focused on attention to specific details). As globally oriented items loaded positively on this factor and locally oriented items negatively, it appears that selfreported global/local bias can be conceptualised as opposite ends of a single dimension. Systemising Tendency contained four items from the SQ, one item from the AQ and one new item. All items related to systemising, and in particular precision in categorisation (of dates, animals, plants etc.).
Although Systemising Tendency and Global Bias emerged as separate factors, there was a modest negative correlation between them which indicates that those who exhibit higher systemising are also likely to exhibit a local bias (cf. [bib_ref] Systemising influences attentional processes during the Navon task: An fMRI study, Billington [/bib_ref]. The correlation between Global Bias and Systemising Tendency supported the notion that these traits are related, but not identical, presumably as one needs to pay sufficient attention to detail to systemise effectively [bib_ref] Autism, hypersystemising, and truth, Baron-Cohen [/bib_ref]. Males also scored more highly than females on Systemising Tendency, replicating work by others [bib_ref] The systemising quotient: An investigation of adults with Asperger syndrome or high-functioning..., Baron-Cohen [/bib_ref] [bib_ref] Psychometric analysis of the systemising quotient (SQ) scale, Ling [/bib_ref] [bib_ref] Predicting Autism Spectrum Quotient (AQ) from the Systemising Quotient-Revised (SQ-R) and Empathy..., Wheelwright [/bib_ref].
Bouvet et al. [bib_ref] Global precedence effect in audition and vision: Evidence for similar cognitive styles..., Bouvet [/bib_ref] and Dale and Arnell [bib_ref] Investigating the stability of and relationships among global/local processing measures, Dale [/bib_ref] suggest there may be a cognitive style that underpins performance on global/local tasks and this is the first attempt to create a self-report subscale that attempts to gauge this style. Future work should aim to further validate the Global Bias scale by assessing its relationship with objective tests of local and global processing, such as the Navon [bib_ref] Forest before trees: The precedence of global features in visual perception, Navon [/bib_ref] task. It could be the case that scores on this subscale predict performance on this task, or predict the extent to which individuals are affected by inappropriate processing orientations in a more direct manner than the SQ-R. This subscale may also provide a useful proxy for global/local orientations when laboratory-based testing is impractical (e.g. online data collection).
In terms of the relationship between processing orientation and imagery, we observed that increased systemising and local processing were associated with increased Imagery Ability and Technical/Spatial scores. Kozhevnikov et al. [bib_ref] Spatial versus object visualisers: A new characterisation of visual cognitive style, Kozhevnikov [/bib_ref] suggest that spatial imagers encode and process images in an analytic manner to arrange and analyse components. Additionally, Kozhevnikov, Hegarty and Mayer [bib_ref] Revising the visualiser-verbaliser dimension: Evidence for two types of visualisers, Kozhevnikov [/bib_ref] show that, although object visualisers tend to encode images as a single unit (global processing), they can also apply sequential, local reasoning in the verbal domains. This suggests that they may be adept at switching between processing orientations but self-report a local bias. The only non-significant correlation was between Technical/Spatial and Language & Word Forms.
The presence of sequence-space synaesthesia, but not grapheme-colour synaesthesia was associated with a lower Global Bias score (i.e. a local bias) and a higher Systemising Tendency score. Synaesthetic experiences are by their nature systematic (e.g., consistency between inducer and concurrent or application of rules to the spatial form to arrive at some outcome), and a proclivity to understand or construct systems is a hallmark of systemising. Being able to locally inspect, manipulate and analyse forms may well depend on local processing. For instance, Simner et al. [bib_ref] Determining the number of components from the matrix of partial correlations, Velicer [/bib_ref] (also [bib_ref] Time-space synaesthesia-A cognitive advantage?, Mann [/bib_ref] showed that synaesthetes with time-space forms have superior recall of dates and events in time and that most reported that target events were retrieved from their spatial forms. (See also [bib_ref] An extended case study on the phenomenology of sequence-space synaesthesia, Gould [/bib_ref] , for a detailed phenomenological report of sequence-space synaesthesia.). Additionally, sequence-space synaesthesia is often linked to, amongst others, calendar and numerical systems, that themselves are often subject to systemising [bib_ref] Autism: The empathising-systemising (E-S) theory, Baron-Cohen [/bib_ref].
Finally, past research into cognitive styles has had trouble connecting theory with measurement [bib_ref] Cognitive styles in the context of modern psychology: Toward an integrated framework..., Kozhevnikov [/bib_ref] [bib_ref] The new object-spatial-verbal cognitive style model: Theory and measurement, Blazhenvoka [/bib_ref]. We consolidated research from seemingly disparate areas (object, spatial and verbal processing, systemising, global/local bias, and synaesthesia) into a novel cognitive style measure that can be applied to many domains of research (e.g. in typical and atypical populations). In Study 3 we predicted how the subscales of the SCSQ would behave with respect to a special population of synaesthetes, and these predictions were largely borne out through validation with external measures.
Supporting Information S1 File. Study 1 Data-Item scores on each SCSQ item. Item numbers are given, and can be crossed-referenced with S1 Table for the item wordings. Gender is coded as 0 = female; 1 = male. (XLSX) S2 File. Study 2 Data-Factor scores and memory task performance. Mean factor scores for every participant have been calculated, and performance for each task given as a proportion of correct responses. Full data for each block in the associates tasks are included, along with d' scores and response criterion for the fractal recognition task. Gender is coded as 0 = female; 1 = male. (XLSX) S3 File. Study 3 Data-Factor scores and presence of synaesthesia. Mean factor scores for participants included in Study 3. Grapheme-colour synaesthesia is coded as 0 = absent; 1 = present. Sequence-space synaesthesia is coded as 0 = absent; 1 = present. Gender is coded as 0 = female; 1 = male. GCS score refers to the consistency scores for the Eagleman et al. [bib_ref] Diagnosing synaesthesia with online colour pickers: Maximising sensitivity and specificity, Rothen [/bib_ref] consistency test previously taken by grapheme-colour synaesthetes in Study 3. (XLSX) S1 Items used in the SCSQ, their original source (and subscale, if appropriate, in parentheses), and item number in the current measure. (DOCX) S2
[table] Table 1: Rotated factor matrix of factor loadings for the final sixty items of the SCSQ.When thinking about an abstract concept (or building), I imagine an abstract schematic building in my mind or its blueprint rather than a specific concrete building. d When I hear a new word, I am curious to know how it is spelled. b I keep my workspace highly organised (e.g., all files/folders on the same subject are in the same colour). bIf I had a collection (e.g., CDs, coins, stamps), it would be highly organised. e [/table]
[table] Table 2: Correlation coefficients (Pearson's r) between scores on the scales of the SCSQ. [/table]
[table] Table 4: Mean proportion of correct responses over blocks for the visual and verbal paired associates tasks.Table 5. Correlation coefficients (Spearman's rho) between SCSQ factor scores and memory task performance. [/table]
[table] Table 6: Mean factor scores as a function of presence/absence of grapheme-colour and sequence-space synaesthesia. [/table]
[table] Table 7: Regression weights (unstandardised b, standard error, and standardised β) for each subscale of the SCSQ, as a function of the presence/ absence of grapheme-colour synaesthesia, sequence-space synaesthesia, and gender. [/table]
|
Rationale for a Pediatric-Inspired Approach in the Adolescent and Young Adult Population with Acute Lymphoblastic Leukemia, with a Focus on Asparaginase Treatment
In the last two decades great improvements have been made in the treatment of childhood acute lymphoblastic leukemia, with 5-year overall survival rates currently approaching almost 90%. In comparison, results reported in adolescents and young adults (AYAs) are relatively poor. In adults, results have improved, but are still lagging behind those obtained in children. Possible reasons for this different pattern of results include an increased incidence of unfavorable and a decreased incidence of favorable cytogenetic abnormalities in AYAs compared with children. Furthermore, in AYAs less intensive treatments (especially lower cumulative doses of drugs such as asparaginase, corticosteroids and methotrexate) and longer gaps between courses of chemotherapy are planned compared to those in children. However, although favorable results obtained in AYAs receiving pediatric protocols have been consistently reported in several international collaborative trials, physicians must also be aware of the specific toxicity pattern associated with increased success in AYAs, since an excess of toxicity may compromise overall treatment schedule intensity. Cooperative efforts between pediatric and adult hematologists in designing specific protocols for AYAs are warranted.
# Introduction
Acute lymphoblastic leukemia (ALL) is the most common cancer of childhood. [bib_ref] Treatment of acute lymphoblastic leukemia in adolescents and young adults, Mohan [/bib_ref] [bib_ref] Childhood and adolescent lymphoid and myeloid leukemia, Pui [/bib_ref] Improvements in overall survival (OS) in children with ALL are among the major successes in the history of cancer treatment. 1,2 OS rates obtained in countries adopting modern intensive chemotherapy schedules are in fact in the range of 85-90%. 1,2 ALL represents almost 30% of all childhood cancers, but only 6% of cancers in adolescents and young adults (AYAs aged between 20 and 34 years). [bib_ref] Treatment of acute lymphoblastic leukemia in adolescents and young adults, Mohan [/bib_ref] Survival rates in AYA patients are lower than in children, for example, 5-year OS rates were 89% for children aged under 15 years versus 50% OS for those aged 15-19 years in the early 2000s. [bib_ref] Treatment of acute lymphoblastic leukemia in adolescents and young adults, Mohan [/bib_ref] It has therefore been suggested that treatment of AYA patients should be closer to the strategies included in pediatric ALL trials, i.e. intensified post-remission strategies including high-dose chemotherapy agents (i.e. steroids, methotrexate) and intensive use of asparaginase (ASP). [bib_ref] Treatment of acute lymphoblastic leukemia in adolescents and young adults, Mohan [/bib_ref] [bib_ref] Childhood and adolescent lymphoid and myeloid leukemia, Pui [/bib_ref] [bib_ref] Optimal therapy for acute lymphoblastic leukemia in adolescents and young adults, Schafer [/bib_ref] [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] [bib_ref] What determines the outcomes for adolescents and young adults with acute lymphoblastic..., Stock [/bib_ref] [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] In the main, adolescents tend to start receiving adult protocols at around age 18 years. [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] Asparaginase are valuable agents widely used in the treatment of childhood ALL. Three forms are currently available: two are derived from E. coli (one native and its pegylated form, PEG-ASP) and one from Erwinia chrysanthemi (asparaginase Erwinia chrysanthemi; crisantaspase). [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] These ASP products are not interchangeable due to their different pharmacological and antigenic properties; in addition, their use is associated with considerable variations in efficacy and toxicity depending on several factors such as the individual patient, the dosage/schedule adopted and also the ongoing line of treatment. [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] The biological mechanism underlying ASP-related therapeutic effects is the same for all three forms, i.e. a deep and prolonged asparagine (ASN) depletion induced in plasma immediately after its administration induces apoptosis in leukemic blasts. [bib_ref] Use of L-asparaginase in childhood ALL, Müller [/bib_ref] Response to ASP varies from patient to patient; it has been suggested that the microenvironment of bone marrow-derived mesenchymal cells where leukemic cells grow has high levels of ASN-synthetase, up to 20times higher than the leukemic blast, and that ASN produced within the microenvironment may provide protection against ASP. [bib_ref] Mesenchymal cells regulate the response of acute lymphoblastic leukemia cells to asparaginase, Iwamoto [/bib_ref] Downregulation of ASN-synthetase could reduce the capacity of the microenvironment to protect against ASP, whilst upregulation of ASN-synthetase could conversely confer enhanced protection against ASP.
Allergic reactions or silent inactivation may develop, both of which may potentially reduce the therapeutic benefit of ASP. [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] For this specific reason modern treatment protocols often include guidelines for timely identification of allergic reactions (and switch to another ASP product) and therapeutic drug monitoring (TDM) programs. The latter programs allow the early identification of patients with silent inactivation who do not benefit from current ASP treatment and facilitate a switch to a different ASP product. This switch ensures continued depletion of ASN, completion of the treatment schedule and maintenance of outcomes. [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] This report summarizes the rationale for a pediatric-inspired approach in AYAs with ALL as presented and discussed during a symposium held in the framework of the 2013 European ALL Working Group (EWALL) International Meeting. A special effort to focus on how ASP treatment might contribute to achieve better results in AYAs was one of the aims of the symposium.
## Current guidelines in acute lymphoblastic leukemia: focus on adolescents and young adults
Outcomes in patients with ALL vary by age and phenotype. [bib_ref] Childhood and adolescent lymphoid and myeloid leukemia, Pui [/bib_ref] Patients with B-cell ALL have better outcomes than those with T-cell ALL. Indeed, optimal outcomes are seen in children aged 1-5 years with B-cell ALL, with 10-year event free survival (EFS) of around 80%. EFS falls to around 70% in children with B-cell ALL aged 10 and over, in contrast EFS rates are somewhat less favorable in children with T-cell ALL but remain fairly static when older ages are concerned. [bib_ref] Childhood and adolescent lymphoid and myeloid leukemia, Pui [/bib_ref] Survival rates in AYAs are poor compared with those in younger children. Data from Surveillance Epidemiology and End Results (SEER) 2000-2004 reported 10-year OS of around 80% in children aged under 15 years, falling to 60% in adolescents aged 15-20 years and 30% in young adults aged 20-30 years; rates have improved by a further 10-15% over the past decade in the AYA group. The steepest decline in survival is seen in mid-adolescence, the sudden decrement at 18 years coincides with newly diagnosed patients receiving adult rather than pediatric regimens. [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] Acute lymphoblastic leukemia can be challenging to treat in AYA. There is an increased incidence of unfavorable and decreased incidence of favorable cytogenetic abnormalities in adolescents compared with children [fig_ref] Table 1: Cytogenetic and immunophenotypic features of acute lymphoblastic leukemia in adolescents and young... [/fig_ref]. [bib_ref] Treatment of acute lymphoblastic leukemia in adolescents and young adults, Mohan [/bib_ref] As we will discuss later in this paper, data from adult cooperative groups demonstrates improved outcomes in AYAs treated with intensified post-remission strategies as per pediatric regimens. [bib_ref] What determines the outcomes for adolescents and young adults with acute lymphoblastic..., Stock [/bib_ref] However, there is a lack of European guidance for the treatment of AYA patients, although the US-based National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology (NCCN Guidelines) do provide guidance and consider AYA separately from the adult population. 6
## National comprehensive cancer network guidelines
National Comprehensive Cancer Network is an alliance of 25 US-based cancer centers, which work together to develop treatment guidelines and carry out research into cancer. The NCCN Guidelines provide recommendations based on the best evidence available at the time they are derived. The guidelines are continuously updated and revised to reflect new data and clinical information; however, they are not necessarily directly reflective of established practice. The guidelines define AYAs as aged 15-39 years and are further subdivided by the presence of the Philadelphia (Ph) chromosome into Ph-positive ALL and Phnegative ALL.The NCCN guidelines recommend pediatricinspired chemotherapy regimens for Ph-positive and Ph-negative ALL in AYAs for induction as outlined below. Maintenance therapy consisting of weekly methotrexate plus daily 6mercaptopurine (6-MP) plus monthly vincristine/prednisone pulses (for 2-3 years) is recommended, with the addition of tyrosine kinase inhibitors (TKI) (imatinib or dasatinib) in Ph-positive patients.Ph-positive disease Patients should be treated in a clinical trial whenever possible. In the absence of an appropriate clinical trial, induction therapy should be a pediatric-inspired multi-agent chemotherapy combined with a TKI. Treatment regimens should include adequate central nervous system prophylaxis for all patients. In those patients achieving a complete response (CR) following initial induction therapy, consolidation with allogeneic hematopoietic stem cell transplantation (HSCT) should be considered if a matched donor is available.Emerging data suggests that in younger AYA patients (aged ≤21 years), allogeneic HSCT may not confer an advantage over chemotherapy plus TKIs, [bib_ref] Improved early event-free survival with imatinib in Philadelphia chromosome-positive acute lymphoblastic leukemia:..., Schultz [/bib_ref] and long-term data is eagerly awaited to determine whether younger patients can be successfully treated without allogeneic HSCT. After HSCT, TKI should be considered. For patients without a donor, consolidation therapy following a CR is a continuation of multi-agent chemotherapy plus a TKI. Such patients should continue to receive post-consolidation maintenance therapy with a regimen including a TKI.Adolescents and young adults patients with Ph-positive relapsed/refractory ALL should participate in a clinical trial. In the absence of an appropriate trial, the patient may be consid-ered for second-line therapy with multi-agent chemotherapy combined with an alternative TKI, allogeneic HSCT (if a second CR is achieved) or donor lymphocyte infusion, if the patient relapses after allogeneic HSCT.Ph-negative disease Patients should be treated in a clinical trial whenever possible. In the absence of an appropriate trial, induction therapy should be based on pediatric-inspired protocols. Treatment regimens should include adequate central nervous system prophylaxis for all patients. Testing for TPMT gene polymorphism should be considered for patients receiving 6-MP as part of maintenance therapy, especially in patients who experience severe bone marrow toxicities. Monitoring for minimal residual disease (MRD) should be considered in patients achieving CR after initial induction therapy.In patients achieving CR, multi-agent based chemotherapy in consolidation, re-induction and maintenance phases must be given. If a matched donor is available, consolidation with allogeneic HSCT may be considered, particularly for patients with residual disease as assessed by MRD, or with high-risk features. In patients achieving less than CR after initial induction therapy, the treatment approach is similar to patients with relapsed/refractory ALL.For patients with relapsed/refractory disease following an initial CR, the approach to second-line treatment depends on the duration of the initial response. In patients with a late relapse (i.e., relapse occurring ≥36 months from initial diagnosis) re-treatment with the same induction regimen may be reasonable. Participation in a clinical trial is preferred; in the absence of an appropriate trial consider second-line therapy with previously unused induction regimens, salvage chemotherapy, allogeneic HSCT (if a second CR is achieved). 6
## Gimema all 1308
The Gruppo Italiano Malattie EMatologiche dell'Adulto (GIMEMA) protocol provides an example of current treatment in AYAs in Europe. Patients are included if they are aged between 18 and 35 years, with a diagnosis of non-B-mature, Ph-negative ALL.All patients receive steroids and methotrexate prior to induction therapy. Patients receive induction Ia (vincristine, daunorubicin, ASP and prednisone) followed by induction Ib (cyclophosphamide, 6-MP and cytarabine). Patients who achieve hematological remission proceed to consolidation therapy. Patients receive consolidation therapy according to their risk group. Standard-risk patients receive high-dose methotrexate and 6-MP. High-risk patients receive consolidation therapy in three steps; step 1 (dexamethasone, vincristine, methotrexate, cytarabine and ASP), step 2 (dexamethasone, vindesine, methotrexate, ifosfamide, ASP and daunorubicin hydrochloride) and step 3 (dexamethasone, cytarabine and ASP). Two step re-induction follows consolidation: IIa (vincristine, doxorubicin hydrochloride, ASP and dexamethasone) and IIb (cyclophosphamide, thioguanine and cytarabine).Early data from GIMEMA ALL 1308 presented at the European Hematology Association meeting in 2014 suggest that the intensified protocol is effective and well tolerated in AYA.Sixty-six patients have been enrolled into GIMEMA ALL 1308, of whom 61 were eligible for treatment. Complete response rate was 98%, which is higher than that found in previous studies GIMEMA ALL 2000 (84%) and GIMEMA ALL 0904 (83%). At 24 months OS in GIMEMA ALL 1308 was 72.3% compared with 61% and 72% in the earlier GIMEMA studies. Severe adverse events have been reported in 11 patients (18%), ASP-related adverse events accounted for six serious adverse events and infection for the other five.
## Review
## Asparaginase in children, adolescents and young adults
The ASPs are a universal component of ALL therapy and are used for remission induction and intensification treatment in every pediatric regimen for ALL. [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] Leukemic cells are unable to synthesize asparagine (ASN) and rely on extracellular sources. In the presence of ASP, ASN is rapidly de-aminated in serum depleting extracellular sources and reducing the supply of ASN to leukemic cells. Leukemic cells are unable to undertake protein biosynthesis leading to cell death. [bib_ref] Use of L-asparaginase in childhood ALL, Müller [/bib_ref] Studies using intensive ASP have revealed significant benefit in terms of EFS, 13,14 disease free survival (DFS) [bib_ref] Long-term results of a randomized trial on extended use of high dose..., Pession [/bib_ref] and continuous complete remission rate, [bib_ref] Intensive high-dose asparaginase consolidation improves survival for pediatric patients with T cell..., Amylon [/bib_ref] when compared with less intensive ASP treatment. Also the completion of the treatment schedule is essential to ensure the expected full clinical benefit. In a study carried out by the Dana-Farber Clinical Institute (DFCI), children were treated with an extended 30 weeks of highdose ASP during intensification (n=352). At 5year follow-up, EFS in children who received less than 25 weeks of planned ASP therapy was significantly poorer than in those who received 26 weeks or more of therapy: 73% versus 90%, P<0.01. [bib_ref] Improved outcome for children with acute lymphoblastic leukemia: results of Dana-Farber Consortium..., Silverman [/bib_ref] A significant improvement in EFS with continued ASP therapy was also seen in a retrospective analysis by the Tokyo Children's Cancer Study Group, wherein children who received more than 50% of the scheduled dose had a significantly improved 5-year EFS versus those who received less than 50% of the scheduled dose: 92.9% versus 74.1%, P<0.025. 14
## Use of pediatric protocols in adolescents and young adults
There is considerable evidence from retrospective analyses that treating AYAs with a pediatric protocol may improve clinical outcomes compared with treatment adopted in adult protocols. [bib_ref] Optimal therapy for acute lymphoblastic leukemia in adolescents and young adults, Schafer [/bib_ref] [bib_ref] What determines the outcomes for adolescents and young adults with acute lymphoblastic..., Stock [/bib_ref] [bib_ref] Intensified chemotherapy inspired by a pediatric regimen combined with allogeneic transplantation in..., Rijneveld [/bib_ref] [bib_ref] Should adolescents with acute lymphoblastic leukemia be treated as old children or..., Boissel [/bib_ref] [bib_ref] Difference in outcome of adolescents with acute lymphoblastic leukemia (ALL) enrolled in..., Testi [/bib_ref] [bib_ref] Significant difference in outcome for adolescents with acute lymphoblastic leukemia treated on..., De Bont [/bib_ref] Pediatric protocols have higher cumulative dosing of drugs (ASP, corticosteroids, methotrexate, vinca-alkaloids) and shorter gaps between courses of chemotherapy compared with adult protocols. [bib_ref] Intensified chemotherapy inspired by a pediatric regimen combined with allogeneic transplantation in..., Rijneveld [/bib_ref] A systematic review and meta-analysis of comparative trials of AYA patients receiving induction therapy with either adult or pediatric-inspired chemotherapy identified 11 trials (n=2489). The AYA patients receiving a pediatric-inspired regimen had a significantly lower all cause mortality at 3 years compared to those receiving an adult regimen: relative risk 0.58, 95%CI 0.51-0.67, P<0.05. [bib_ref] Adolescents and young adults with acute lymphoblastic leukemia have a better outcome..., Ram [/bib_ref] The absolute risk reduction for all cause mortality at 3 years was 0.2 and the number needed to treat to prevent one death with pediatricinspired regimens was 5 . Secondary end-points included all cause mortality at the end of the trial, complete remission, 3-year EFS and relapse rate. Significant benefit was seen in the patients receiving the pediatric-inspired regimen (P<0.05 for all secondary end-points). Non-relapse mortality was similar in both groups.
The German multicenter ALL (GMALL) protocols were originally based on pediatric Berlin-Frankfurt-Munster (BFM) protocols and have been optimized for AYAs since 1981. A retrospective analysis compared outcomes from GMALL 05/93 (an earlier study) and GMALL 07/03 (a later study). The main innovations in GMALL 07/03 were intensified shortened induction with dexamethasone rather than with prednisone, PEG-ASP rather than native ASP, intensified first consolidation, six doses of high dose methotrexate and ASP during consolidation, matched unrelated SCT for high risk and very high risk patients without sibling donor and SCT indication in patients with persistent minimal residual disease. AYA patients receiving the later protocol (GMALL 07/03) had significant improvements in 5-year OS compared with GMALL 05/93 (65% in GMALL 07/03 versus 46% in GMALL 05/93). This data represents the largest cohort of AYA patients treated to date with pediatric-inspired protocols (642 in GMALL 05/93 and 887 in GMALL 07/03). [bib_ref] Significant improvement of outcome in adolescents and young adults (AYAs) aged 15-35..., Gokbuget [/bib_ref] A number of other studies have been carried out using retrospective data to compare outcomes in AYAs receiving pediatric and adult inspired protocols. The results are shown in and demonstrate that outcomes are significantly improved in AYA patients receiving a pediatric-inspired protocol compared with an adult-inspired protocol.
A retrospective study compared outcomes in 177 AYAs aged 15-20 years entering either a pediatric [French Acute Lymphoblastic Leukemia Group (FRALLE)-93] or an adult protocol [Leucémie Aiguë Lymphoblastique de l'Adulte (LALA)-94]. The cumulative doses of treatment (vincristine/vindesine, prednisone, dexamethasone, ASP, daunorubicin/ doxorubicin/mitoxantrone, vepeside/cyclophosphamide) were higher in the pediatric protocol than in the adult protocol. The overall dose of ASP was 20-times higher in the pediatric protocol: 180,000 IU/m 2 in the pediatric regimen versus 9000 IU/m 2 in the adult regimen. [bib_ref] Should adolescents with acute lymphoblastic leukemia be treated as old children or..., Boissel [/bib_ref] A retrospective study compared outcomes in adolescents aged 14-18 years treated on the pediatric Associazione Italiana Ematologia Oncologia Pediatrica (AIEOP) ALL 95 and 2000 protocols with those treated on adult GIMEMA ALL 0496 and 2000 protocols. [bib_ref] Difference in outcome of adolescents with acute lymphoblastic leukemia (ALL) enrolled in..., Testi [/bib_ref] The pediatric protocols had seven-drug induction followed by risk-modulated post-remission therapy. SCT was recommended for very high-risk patients. Another retrospective study compared the Dutch Children's Oncology Group (DCOG) pediatric regimen with the Hemato-Oncologie voor Volwassenen Nederland (HOVON) adult protocols in AYAs. The main differences between the regimens were shorter intervals between courses (≤1 week versus ≤4 weeks) and more ASP (mean cumulative dosage: 101,000 IU/m 2 versus 70,000 IU/m 2 ) in the pediatric regimen. [bib_ref] Significant difference in outcome for adolescents with acute lymphoblastic leukemia treated on..., De Bont [/bib_ref] A similar study compared the Children cytarabine. [bib_ref] What determines the outcomes for adolescents and young adults with acute lymphoblastic..., Stock [/bib_ref] Although ASP is a pivotal drug in the treatment of ALL, protocols based on strategies not including ASP have also demonstrated benefit in AYAs. A recent study compared outcomes in 85 patients aged 12-40 years with Ph-ALL treated with the pediatric augmented BFM regimen with outcomes in 71 historical controls who received hyper-CVAD (cyclophosphamide, vincristine, doxorubicin and dexamethasone). Outcomes year complete remission duration (CRD) and OS] were comparable between the two groups: CRD was 70% in the BFM arm versus 66% in the hyper-CVAD arm and OS 74% versus 71%.Toxicity is a key issue in the use of pediatric-inspired regimens in older patients, and may limit the potential benefit of high intensity pediatric-inspired regimens. Toxicity may has the potential to lead to increased adverse events, potentially lethal toxicities and a reduction in the total dose due to dose interruption, dose reduction or early cessation of therapy. Some recent protocols include monitoring for MRD to inform clinical decisions and ensure treatment intensity is appropriate to each individual's needs, whilst maintaining efficacy and minimizing adverse events.
Recent evidence presented at the American Society of Hematology (ASH) meeting in December 2013 suggests that pediatricinspired protocols are feasible and well tolerated in AYA patients. [bib_ref] Frontline treatment of acute lymphoblastic leukemia (ALL) in older adolescents and young..., Advani [/bib_ref] [bib_ref] Impact of age on toxicity associated with chemotherapy for acute lymphoblastic leukemia..., Hough [/bib_ref] Data from the prospective US intergroup trial C10403 in 318 AYA patients aged 16-39 years receiving a pediatric-inspired regimen revealed grade 3-5 adverse events as follows: hyperglycemia (29.3%), febrile neutropenia (19.2%), hyperbilirubinemia (15.9%), thrombosis (2.9%), pancreatitis (1.1%) and allergic reaction (0.7%) in induction. [bib_ref] Frontline treatment of acute lymphoblastic leukemia (ALL) in older adolescents and young..., Advani [/bib_ref] Data was also presented from UK Acute Lymphoblastic Leukemia (UKALL) 2003, a large study of 3129 patients, which included 229 patients aged 16-24 years, 56 of whom were aged 20 years or above. Most observed toxicities were more common in patients aged over 10 years than in younger patients, however, there was no difference in incidence between patients aged 16-24 years and 10-15 years. Age did not appear to impact on some adverse events (infection, hypersensitivity to ASP, neurotoxicity to vincristine), however, the incidence of thrombotic events due to ASP increased with increasing patient age and avascular necrosis was most commonly observed in patients aged 10-19 years. [bib_ref] Impact of age on toxicity associated with chemotherapy for acute lymphoblastic leukemia..., Hough [/bib_ref] Further data on toxicity comes from the Erwinaze Master Treatment Protocol (EMTP) was a compassionate use program which allowed patients with a grade 2 or higher hypersensitivity reaction to PEG-ASP or native ASP to switch to crisantaspase (trade named Erwinaze in the US). Adverse event reports or case report forms were completed for 940 patients aged 0-76 years (mean 9.7 years), of whom 16% (n=147) were aged 16-39 years. Post hoc analyses revealed that the adverse event profile in the AYA population was consistent with the profile in the full population in this trial. [bib_ref] Safety profile of asparaginase Erwinia chrysanthemi in a large compassionate-use trial, Plourde [/bib_ref] However, despite the potential benefits of a pediatric regimen as shown in , many AYA patients do not currently receive a pediatric regimen. This may be due to a number of factors including fear of increased incidence of side-effects and the potential for increased mortality due to toxicity, referral patterns (AYAs are referred by oncologists preferably to adult rather than pediatric centers), cost of treatment, lack of insurance (where applicable), lack of parental vigilance and poor compliance, lack of information, guidance and patient involvement in the decision about where to be treated. [bib_ref] Optimal therapy for acute lymphoblastic leukemia in adolescents and young adults, Schafer [/bib_ref] [bib_ref] Intergroup Trial C10403: a pediatric treatment approach to improve outcomes in adolescents..., Breitenbach [/bib_ref] [bib_ref] Disparity in outcome of young adults with ALL (response), Stock [/bib_ref]
## Management of toxicities with asparaginase
Asparginase treatment is associated with a number of adverse events, [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] which can lead to discontinuation of ASP or delay in treatment and a reduction in clinical benefit.
A retrospective analysis reviewed the records of 214 patients aged 15-59 years with ALL or lymphoblastic lymphoma and considered the reasons for early discontinuation of ASP. [bib_ref] Changes in antithrombin and fibrinogen levels during induction chemotherapy with L-asparaginase in..., Hunault-Berger [/bib_ref] All patients received six doses of native E. coli ASP during induction. ASP was delayed in 22% of patients and the number of doses was reduced in 41%. The most common reasons for delay were coagulation abnormalities (47%) and logistical reasons (34%). The most common reasons for dose reduction were coagulation abnormalities (35%), liver toxicity (17%), logistical reasons (16%) and pancreatic reaction (12%). [bib_ref] Changes in antithrombin and fibrinogen levels during induction chemotherapy with L-asparaginase in..., Hunault-Berger [/bib_ref] The majority of data on adverse events with ASP is from studies using E. coli-derived ASP. Different definitions of adverse events make it very difficult to compare data across studies; however, the pegylated formulation has reduced immunogenicity and consequently lower rates of hypersensitivity. [bib_ref] Tolerability and efficacy of L-asparaginase therapy in pediatric patients with acute lymphoblastic..., Raetz [/bib_ref] provides key information on adverse events and management strategies. [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] [bib_ref] Tolerability and efficacy of L-asparaginase therapy in pediatric patients with acute lymphoblastic..., Raetz [/bib_ref] [bib_ref] Prevalence of transient hyperglycemia during induction chemotherapy for pediatric acute lymphoblastic leukemia, Lowas [/bib_ref] [bib_ref] Hyperglycemia during induction therapy is associated with increased infectious complications in childhood..., Sonabend [/bib_ref] [bib_ref] Toxicities of intravenous (IV) pegasparaginase (ONCASPAR) in adults with acute lymphoblastic leukemia..., Advani [/bib_ref] [bib_ref] Hyperglycemia during induction therapy is associated with poorer survival in children with..., Sonabend [/bib_ref] [bib_ref] Asparaginase-associated pancreatitis in children, Raja [/bib_ref] [bib_ref] Predicting asparaginase-associated pancreatitis, Knoderer [/bib_ref] [bib_ref] Successful management of extreme hypertriglyceridemia from pegaspargase with omega-3, Bostrom [/bib_ref] [bib_ref] Successful management of extreme hypertriglyceridemia in a child with acute lymphoblastic leukemia..., Tong [/bib_ref] Two adverse events, hypersensitivity and coagulation disorders, are discussed in greater depth below.
## Hypersensitivity
Asparginase use may lead to the development of anti-ASP antibodies, which may result in a clinical hypersensitivity reaction or be symptom-free (known as silent inactivation). [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] Hypersensitivity is the most commonly reported adverse reaction with all ASP; incidence varies according to a number of factors including type of ASP, dosing schedule, route of administration, concomitant medication and duration of treatment. [bib_ref] Lasparaginase treatment in acute lymphoblastic leukemia, Pieters [/bib_ref] [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] [bib_ref] Optimizing asparaginase therapy for acute lymphoblastic leukemia, Rizzari [/bib_ref] Rates can be as high as 36% with native E. coli ASP and tend to be lower with PEG-ASP and crisantaspase. [bib_ref] The anti-asparagines antibodies correlate with L-asparagines activity and may affect clinical outcome..., Zalewska-Szewczyk [/bib_ref] Silent inactivation occurs in around 30% of patients receiving native E. coli ASP and rates are lower with PEG-ASP and crisantaspase.
There is cross-reactivity between E. coliderived ASP (native ASP and PEG-ASP) but not between E. coli-derived ASPs and crisantaspase, which is derived from Erwinia chrysanthemi. [bib_ref] The cross-reactivity of anti-asparaginase antibodies against different Lasparaginase preparations, Zalewska-Szewczyk [/bib_ref] Therefore, it has been suggested that a change to crisantaspase in cases of allergy to native or pegylated E. coli ASP might ensure advantages in continuation of treatment and clinical benefit. [bib_ref] Optimizing asparaginase therapy for acute lymphoblastic leukemia, Rizzari [/bib_ref] Two recent studies demonstrate the adverse event profile seen with crisantaspase, given as second-line treatment to patients with a hyper- . Outcomes in adolescents and young adults patients receiving a pediatric or an adult regimen. The comparison of complete response, event free survival and overall survival rates is reported as pediatric vs adult protocols, respectively. sensitivity reaction to E. coli-derived ASP. In EMTP, the most frequently reported adverse events included hypersensitivity (13.6%), local hypersensitivity reaction (3.3%) and anaphylaxis (0.9%), infection/sepsis (3.9%), pancreatitis (3.9%), fever (3.8%), hyperglycemia (3.7%) and increased transaminases (3.5%). [bib_ref] Safety profile of asparaginase Erwinia chrysanthemi in a large compassionate-use trial, Plourde [/bib_ref] In AALL07P2, which enrolled 55 patients, hypersensitivity was seen in 10.9% of patients, hyperglycemia in 10.9% and pancreatitis in 1.8%. [bib_ref] Erwinia asparaginase achieves therapeutic activity after pegaspargase allergy: a report from the..., Salzer [/bib_ref] The safety profile of second-line crisantaspase seems to compare favorably with that observed in patients treated with first-line native E. coli ASP; however, the limitations of these two studies (small numbers in one and the compassionate-use design in the other) should be taken into consideration and confirmed in larger cohorts and prospective studies.
## Review
## Country (years of recruitment) age range pediatric protocols adult protocols cr (%) efs (%) os (%)
## Coagulation disorders
Reduced protein synthesis with ASP leads to falls in the serum levels of key proteins. Reduced serum albumin levels impact on the clearance and metabolism of some agents, including steroids, with a potential reduction in efficacy. [bib_ref] The right dose for the right patient, Asselin [/bib_ref] [bib_ref] Dexamethasone exposure and asparaginase antibodies affect relapse risk in acute lymphoblastic leukemia, Kawedia [/bib_ref] Reduced serum levels of immunoglobulins and lectins may also increase the risk of infection.
Coagulation disorders result from the effect of ASP on protein synthesis, which leads to reductions in plasminogen, fibrinogen, antithrombin, protein C and S, factors IX and X. Reductions in anti-coagulant proteins can impair thrombin inhibition or result in elevated thrombin levels which may increase the risk of bleeding or thrombosis. Therefore, ASP treatment has been associated with an increased risk of thrombo-hemorrhagic disorders. Thrombosis, mostly at venous sites, is considered the main risk. [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] Coagulation disorders may occur in up to one-third of patients receiving ASP and generally occur early in treatment. [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] It is difficult to compare rates across the ASPs, however, data from EMTP and AALL07P2 showed low rates of thrombosis/hemorrhage with crisantaspase (0% in AALL07P2, [bib_ref] Erwinia asparaginase achieves therapeutic activity after pegaspargase allergy: a report from the..., Salzer [/bib_ref] thrombosis rates of 2.1% and hemorrhage rates of 1% in EMTP). [bib_ref] Safety profile of asparaginase Erwinia chrysanthemi in a large compassionate-use trial, Plourde [/bib_ref] In adults, work has shown that a single dose of PEG-ASP leads to a reduction in plasma antithrombin III to <50% of normal in twothirds of patients , with an overall median nadir level of 45% of normal. The reduction in plasma antithrombin III lasted for approximately 21 days, suggesting that there may be greater potential for thrombotic adverse events with PEG-ASP due to an extended duration of asparagine depletion. [bib_ref] Pharmacodynamics and safety of intravenous pegaspargase during remission induction in adults aged..., Douer [/bib_ref] As with all thrombotic conditions, rates are higher in adults than in children. [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] A review of 548 ALL patients treated at the DFCI between 1991 and 2008 revealed that venous thrombotic events (VTE) occurred in 8% of patients, including 5% of pediatric patients and 34% of adult patients . [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] Median time to VTE in this study was 3.5 months (0.5-10.1 months) with no difference between adult and pediatric patients.
Risk factors for VTE include older age at diagnosis, T-cell phenotype, high-risk ALL, use of a central line, steroid treatment, presence of mediastinal mass and inherited thrombophilia traits. [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] [bib_ref] Validation of a predictive model for identifying an increased risk for thromboembolism..., Mitchell [/bib_ref] [bib_ref] Thrombotic complications in children with haematologic malignancies, Putti [/bib_ref] Older children and adolescents (15-20 years) have a increased risk of thrombosis compared with younger children (6-10 years), odds ratio: 4.0 versus 11,7, P<0.01. [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] However, although changes in coagulation proteins are commonly observed during ASP treatment, one recent study failed to show a clear association between coagulation derangement and the occurrence of VTE in a pediatric population. [bib_ref] Screening for coagulopathy and identification of children with acute lymphoblastic leukemia at..., Santoro [/bib_ref] VTE are severe events with significant morbidity and mortality. Inpatients receiving ASP at DFCI between 1998 and 2008 the most common complication was VTE recurrence, which occurred in 44% of adults and 15% of children. [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] Thrombosis has an adverse impact on outcome. In a retrospective analysis of 214 patients aged 15-59 years with ALL or lymphoblastic leukemia, treated with six doses of native E. coli ASP during induction, 9.8% of patients experienced a thrombotic event during induction. [bib_ref] Changes in antithrombin and fibrinogen levels during induction chemotherapy with L-asparaginase in..., Hunault-Berger [/bib_ref] Treatment with ASP was stopped in 10 of the patients who experienced a thrombosis. Patients who experienced thrombosis had significantly poorer outcomes compared with those without thrombosis; 7year OS of 19 months versus 53 months, P=0.06. Therefore, strategies to prevent VTE and allow administration of adequate doses of ASP are warranted but there is still no consensus on which diagnostic and prophylactic strategies should be used to prevent them. Prophylactic administration of antithrombin concentrates and the use of heparin have been attempted but results and indications are still far from consistent. [bib_ref] Validation of a predictive model for identifying an increased risk for thromboembolism..., Mitchell [/bib_ref] [bib_ref] Trend to efficacy and safety using antithrombin concentrate in prevention of thrombosis..., Mitchell [/bib_ref] Recommendations suggest that in general, ASP should be withheld i) in case of severe VTE (Grade 3-4) and discontinued if symptoms do not resolve or ii) in cerebral thrombosis. In resolved non-cerebral VTE, resumption of ASP at lower doses may be tried under anticoagulation treatment. [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] Re-exposure to ASP has been indeed safely performed. [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref]. Main toxicities related to asparaginase treatment.
## Incidence impact management options
Hyperglycemia 20-35% children [bib_ref] Prevalence of transient hyperglycemia during induction chemotherapy for pediatric acute lymphoblastic leukemia, Lowas [/bib_ref] [bib_ref] Hyperglycemia during induction therapy is associated with increased infectious complications in childhood..., Sonabend [/bib_ref] Increased infection rates Resolves within 2-4 weeks; 31 treat with insulin therapy; 29 glycemic control should be 25% adults [bib_ref] Toxicities of intravenous (IV) pegasparaginase (ONCASPAR) in adults with acute lymphoblastic leukemia..., Advani [/bib_ref] and poor survival outcomes [bib_ref] Hyperglycemia during induction therapy is associated with increased infectious complications in childhood..., Sonabend [/bib_ref] [bib_ref] Hyperglycemia during induction therapy is associated with poorer survival in children with..., Sonabend [/bib_ref] improved using diet and exercise; [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] monitoring of blood glucose essential to ensure that patients are identified 29 Pancreatitis 5-10% of patients, [bib_ref] Tolerability and efficacy of L-asparaginase therapy in pediatric patients with acute lymphoblastic..., Raetz [/bib_ref] [bib_ref] Asparaginase-associated pancreatitis in children, Raja [/bib_ref] Generally mild, but can present as a severe Bowel rest (tube feeding), correction of electrolytes and glucose disturbances and rates similar in adults complication ,37 long-term sequelae include prophylactic antibiotic treatment. Use of octreotide and protease inhibitors has also been and children the formation of pancreatic cysts and suggested; 29 ASP treatment must be discontinued in patients with symptomatic persistent diabetes mellitus [bib_ref] Predicting asparaginase-associated pancreatitis, Knoderer [/bib_ref] pancreatitis 29 Liver toxicity 20% children [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] Commonly presents as elevation of liver enzymes Test liver function prior to each ASP dose and when drugs metabolized by the liver e.g. 33% adults [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] (aspartate transaminase and alanine transaminase), anthracycline and vinca alkaloids, are given after ASP; 4 ASP treatment should be but may also present as hyperbilirubinemia [bib_ref] Prevention and management of asparaginase/pegasparaginase-associated toxicities in adults and older adolescents: recommendations..., Stock [/bib_ref] interrupted in patients with Grade 3 or 4 liver toxicity and resumed with careful and reduction of hepatic protein synthesis, [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] monitoring if toxicity resolves to Grade 1 4 generally mild and transient [bib_ref] Incidence and management of asparaginase-associated adverse events in patients with acute lymphoblastic..., Earl [/bib_ref] Serum amylase --Monitor during treatment;4 withhold treatment if levels increase and lipase to >2-3 times the upper limit of normal; 4 discontinue if levels continue to be changes >3 times the upper normal limit for more than 2-3 days; 4 .rechallenge may be possible, but only for very mild cases (e.g. asymptomatic cases only and resolving within 48 hours) Elevation of -May mask initial signs of pancreatic distress; Treatment is poorly defined and may include a wide range of measures, plasma however clinical symptoms do not normally i.e. from none to concomitant treatment and dietary modifications, triglycerides accompany laboratory changes [bib_ref] Successful management of extreme hypertriglyceridemia from pegaspargase with omega-3, Bostrom [/bib_ref] [bib_ref] Successful management of extreme hypertriglyceridemia in a child with acute lymphoblastic leukemia..., Tong [/bib_ref] hydration, use of lipid-lowering agents or even plasmapheresis [bib_ref] Successful management of extreme hypertriglyceridemia from pegaspargase with omega-3, Bostrom [/bib_ref] [bib_ref] Successful management of extreme hypertriglyceridemia in a child with acute lymphoblastic leukemia..., Tong [/bib_ref] years revealed a thrombotic event rate of 3.2%. [bib_ref] Asparaginase-related venous thrombosis in UKALL 2003-re-exposure to asparaginase is feasible and safe, Qureshi [/bib_ref] Fifty of the 59 patients with thrombotic events required ongoing ASP and 38 (73%) were re-exposed to PEG-ASP, including 10 patients with cerebral venous sinus thrombosis. There was no recurrence of thrombosis during re-exposure and no excess bleeding due to heparin. Low molecular weight heparin was used during re-exposure in three-quarters of patients. In the DFCI study, [bib_ref] The frequency and management of asparaginase-related thrombosis in paediatric and adult patients..., Grace [/bib_ref] ASP was withheld after diagnosis of VTE for a median of 9 weeks in children and 4 weeks in adults. ASP was restarted in 77% of patients and most (70%) received at least 85% of the scheduled dose of ASP. Recurrence of VTE occurred in 33% of patients restarted on ASP. There was no significant difference in clinical outcomes in the patients with VTE compared to those without VTE; 2-year OS of 86±7% versus 95±1%, P=0.12. There were no deaths directly related to VTE in either group. Hematologists should be aware of possible treatment complications with ASP; careful vigilance can lead to necessary modulation and safe completion of treatment.
# Conclusions
In conclusion, the design of modern chemotherapy protocols for AYA should be the result of cooperative efforts between pediatric and adult hematologists. It is important to consider the specific biological and response patterns of ALL subtypes affecting AYA and also their well known propensity to develop severe side-effects. In this context, ASP may represent a great opportunity, given its specific mechanism of action, the possibility of effective TDM and the established pattern of toxicity. Toxicity with ASP is easily preventable with careful ASP treatment dosage modulation and manageable with advanced supportive treatment currently available to hematologists.
[table] Table 1: Cytogenetic and immunophenotypic features of acute lymphoblastic leukemia in adolescents and young adults and in children. [/table]
|
Primary prevention of venous thromboembolism in elderly medical patients
Primary prophylaxis with the use of an effective and safe intervention appears the best approach of venous thromboembolism (VTE) management in medical elderly patients, the most affected by VTE. With increasing life expectancy, prevention of VTE, particularly in elderly patients, will arise as a major public health problem. Few well designed clinical trials evaluating thromboprophylaxis in medical settings were conducted in the specifi c population of geriatric patients. However, among the several pharmacological treatments evaluated, low molecular weight heparins enoxaparin 40 mg daily or dalteparin 5000 IU daily appeared effective and safe in the prevention of VTE in elderly patients. Despite available data, and recommendations for VTE prevention in medical patients, thromboprophylaxis is underused or misused in practice. Heterogeneity of clinical studies, selected populations, concern about bleeding, and lack of a clear clinical benefi t are some of the reasons that could explain the gap between theory and practice. In this review, after a brief report of epidemiologic data and specifi cities of VTE in elderly patients, the authors discuss the available results of VTE primary prevention trials for elderly medical patients, the limitations of these data, and the challenges to improve the practice and to reduce the incidence of this frequent but preventable disease.
# Introduction
Venous thromboembolism (VTE) is a major cause of morbidity and mortality in hospitalized patients [bib_ref] A population-based perspective of the hospital incidence and case-fatality rates of deep..., Anderson [/bib_ref]. Risk of VTE associated with surgical procedures is well established, and the benefi t of thromboprophylaxis has been clearly demonstrated in surgical settings. Use of thromboprophylaxis is now a routine practice in patients undergoing surgery. However, recent epidemiological data suggested that the majority of VTE events acquired in hospitals occurred in medical, rather than in surgical patients [bib_ref] Incidence of venous thromboembolism verifi ed by necropsy over 30 years, Lindblad [/bib_ref]. Despite this, prevention of VTE in medical settings has been less extensively studied than in surgical settings. Nevertheless, some good-quality clinical trials have been conducted in the last decade in hospitalized medical patients, and guidelines now exist for these patients. Consensus guidelines published by the American College of Chest Physicians (ACCP) recommend prophylaxis with unfractionated heparin (UFH) or low molecular weight heparin (LMWH) in patients admitted to hospital with congestive heart failure or severe respiratory disease, or who are confi ned to bed and have one or more additional risk factors, including active cancer, previous VTE, sepsis, acute neurological disease, or infl ammatory bowel disease.
Among medical patients, elderly patients deserve probably more attention than others as regards the risk and the prevention of VTE for many reasons. They are more affected by VTE than younger patients [bib_ref] Incidence of venous thromboembolism: a community-based study in Western France, Oger [/bib_ref] ; the proportion of severe and fatal VTE events is higher in those patients [bib_ref] The epidemiology of venous thromboembolism in the community, Heit [/bib_ref] ; the association with other comorbidities is frequent [bib_ref] The epidemiology of diagnosed pulmonary embolism and deep venous thrombosis in the..., Kniffi N Wd [/bib_ref] ; diagnosis of pulmonary embolism (PE) is more diffi cult (Le [bib_ref] The challenge of diagnosing pulmonary embolism in elderly patients: infl uence of..., Righini [/bib_ref] ; and fi nally the risk of bleeding related to anticoagulant treatments is high [bib_ref] Venous thromboembolism in very elderly patients: fi ndings from a prospective registry..., Lopez-Jimenez [/bib_ref] , even more due to the high frequency of polymedication and interaction between used drugs. Therefore, prevention of VTE in this specifi c population is particularly challenging.
Despite relative high proportion of elderly patients included in the largest recent clinical trials of VTE prevention in medical settings [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref] , none were specifi cally conducted in the elderly population, and only a few other large, well designed, randomized controlled trials have been performed addressing this specifi c issue [bib_ref] Prevention of deep vein thrombosis in elderly medical in-patients by a low..., Dahan [/bib_ref] [bib_ref] Heparin prophylaxis in bedridden patients, Bergmann [/bib_ref]. Nevertheless some data from these studies are available for elderly patients and are discussed in this review. However, specifi c concerns, such as bleeding risk or lack of clinical benefi t remain, and could explain why the systematic use of prophylaxis in elderly is not well implemented in medical wards [bib_ref] A multinational observational cohort study in hospitalized medical patients of practices in..., Anderson [/bib_ref].
After a report of VTE epidemiology and specifi cities in elderly medical patients, this review gives some practical advice, based on available data, to manage VTE prevention in older patients in medical wards. We also discuss the limits of current data and the challenges to improve the practice and reduce the incidence of VTE in elderly patients.
## Venous thromboembolism in elderly medical patients vte: a disease of elderly patients
The incidence of VTE increases with age. A populationbased study of patients discharged from hospital showed that the incidence of deep venous thrombosis (DVT) increased from 119 per 100,000 patients between 60-69 years old, to 291 per 100,000 patients Ն80 years old [bib_ref] A population-based perspective of the hospital incidence and case-fatality rates of deep..., Anderson [/bib_ref]. In France, the global annual incidence of symptomatic VTE was 1.83 per 1000 patients, but reached 10 per 1000 patients aged 75 years or older [bib_ref] Incidence of venous thromboembolism: a community-based study in Western France, Oger [/bib_ref]. The increased incidence was noted for DVT, but also for PE. Moreover, because of the silent nature of the disease, the incidence of VTE was probably underestimated.
In a post-hoc analysis of MEDENOX study, age Ͼ75 years was found to be an independent risk factor of VTE [bib_ref] Risk factors for venous thromboembolism in hospitalized patients with acute medical illness:..., Alikhan [/bib_ref] , and prevalence of VTE reached 18% in patients Ն75 years old from the placebo group [bib_ref] Prevention of venous thromboembolism in medical patients with enoxaparin: a subgroup analysis..., Alikhan [/bib_ref]. Furthermore, elderly patients have often other acquired conditions recognized as risk factors for VTE that contribute to the high incidence of the disease (Kniffi n et al 1994).
## High mortality rate in elderly patients
The mortality of acute VTE in elderly is increased: 16% of patients over 80 years died early after the diagnosis of VTE, in comparison with 2% before 40 years, a difference that is confi rmed and increased over 3.5 years of follow up [bib_ref] A population-based perspective of the hospital incidence and case-fatality rates of deep..., Anderson [/bib_ref]. In the analysis of the RIETE registry, at three months after the venous thromboembolic event, 3.7% of patients aged Ն80 years died of PE compared to 1.1% of patients Ͻ80 years (OR 3.6, 95% CI 2.7-4.7) [bib_ref] Venous thromboembolism in very elderly patients: fi ndings from a prospective registry..., Lopez-Jimenez [/bib_ref].
## More severe initial presentation in elderly patients
In the study by , there is a trend of a higher proportion of symptomatic PE in elderly: in patients over 70 years, PE is the manifestation of VTE in 70% of cases, in comparison with 55% of cases under 70 years. This fi nding is confi rmed in the RIETE registry, where patients Ն80 years old were more often enrolled with symptomatic PE [bib_ref] Venous thromboembolism in very elderly patients: fi ndings from a prospective registry..., Lopez-Jimenez [/bib_ref]. In this study, patients with VTE aged 80 or over more often weighed Ͻ65 kg, and had a higher rate of chronic lung disease, heart failure and creatinine clearance Ͻ60 mL/ min.
## Diffi culties of diagnosis of pe in elderly patients
Clinical symptoms and sign of PE are not specifi c, and PE can be a silent disease as well. Sometimes sudden death is the fi rst clinical manifestation of PE. In autopsy studies, diagnosis of PE was suspected before death in only 10% of patients who died from proven PE [bib_ref] Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep..., Sandler [/bib_ref]. Diagnosis of PE is particularly diffi cult in elderly patients because the clinical presentation is less discriminative. The frequency of many cardiopulmonary underlying conditions that may mimic clinical presentation of PE is increased with age, and on the other hand, some classical signs of PE such as hemoptysis, tachycardia, hemidiaphragmatic elevation and pleural effusion were not associated with PE in patients aged 75 years or over (Le . The characteristics of diagnostic tests for PE are also unfavorably infl uenced by age . In a study of 1029 consecutive patients presenting to the emergency department with clinically suspected PE, this diagnosis could be ruled out by a negative D-dimer test in 67% of patients 40 years or under, but in only 10% of patients aged 80 years or older [bib_ref] Effects of age on the performance of common diagnostic tests for pulmonary..., Righini [/bib_ref]. Lung scan for PE diagnosis demonstrates a lower performance in elderly due to frequent underlying pulmonary diseases [bib_ref] Effects of age on the performance of common diagnostic tests for pulmonary..., Righini [/bib_ref] , and fi nally, the frequency of renal impairment in elderly patients limits the use of CT-scan.
## High bleeding risk in elderly patients
The main risk of treatment with anticoagulants, at therapeutic or prophylactic doses, is bleeding. Major data on the risk of bleeding is provided by trials on long term oral anticoagulation. These studies have shown that elderly patients with VTE have a higher risk of bleeding complications [bib_ref] Anticoagulant related bleeding: clinical epidemiology, prediction and prevention, Landefeld [/bib_ref] [bib_ref] Major bleeding after hospitalization for deep-venous thrombosis, White [/bib_ref] [bib_ref] Hemorrhagic complications of anticoagulant treatment, Levine [/bib_ref]. In a prospective evaluation of an index for predicting the risk of major bleeding in outpatients treated with warfarin, age 65 years or greater was found to be an independent risk factor of major bleeding [bib_ref] Prospective evaluation of an index for predicting the risk of major bleeding..., Beyth [/bib_ref]. In an analysis from the RIETE registry (Lopez-Jimenez et al 2006), 3.4% of patients Ն80 years old had major bleeding events compared to 2.1% of patients Ͻ80 years. Fatal bleeding occurred in 0.8% in the oldest and 0.4% among patients Ͻ80 years. Recent bleeding, abnormal renal function (creatinine clearance Ͻ60 mL/min), use of corticosteroids or long term therapy with LMWH were independently associated with an increased risk of major bleeding. Other results from the same registry have also shown that major and fatal bleeding events occurred more frequently in acutely ill medical patients than in surgical patients [bib_ref] The outcome after treatment of venous thromboembolism is different in surgical and..., Monreal [/bib_ref].
In the case of severe renal insuffi ciency, UFH can be used, while LMWH are usually contraindicated. However, UFH was also associated with a higher risk of bleeding in the elderly [bib_ref] Aging and heparin-related bleeding, Campbell [/bib_ref].
During VTE prophylaxis, results from a meta-analysis of clinical trials conducted in medical settings showed that the incidence of major bleeding was 0.4% in heparin groups, compared to 0.2% in control groups [bib_ref] Prevention of venous thromboembolism in internal medicine with unfractionated or low-molecular-weight heparins:..., Mismetti [/bib_ref]. In the recent trials, this incidence varied from 0.2% to 1.7% in the treatment groups [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref].
## Vte: a complication of medical illness
VTE is a major cause of morbidity and mortality in hospitalized patients [bib_ref] A population-based perspective of the hospital incidence and case-fatality rates of deep..., Anderson [/bib_ref]. It has been reported that 10% of deaths occurring during hospitalization were related to PE [bib_ref] Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep..., Sandler [/bib_ref]. Autopsy studies revealed that 70% to 80% of all in-hospital deaths related to PE were not associated with surgical procedures, but occurred in medical patients [bib_ref] Risk factors for pulmonary embolism: the Farmingham Study, Goldhaber [/bib_ref] [bib_ref] Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep..., Sandler [/bib_ref] [bib_ref] Incidence of venous thromboembolism verifi ed by necropsy over 30 years, Lindblad [/bib_ref] [bib_ref] Fatal pulmonary embolism in hospitalised medical patients, Baglin [/bib_ref]. Hospitalization for an acute medical illness was also recognized as an independent factor of VTE associated with about an 8-fold increased relative risk for VTE [bib_ref] Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control..., Heit [/bib_ref]. Moreover, the initial presentation and outcomes of VTE were more severe and frequent in patients who developed VTE after an acute medical illness than in patients who developed VTE after a surgical intervention [bib_ref] The outcome after treatment of venous thromboembolism is different in surgical and..., Monreal [/bib_ref].
Before the publication of the more recent large clinical trials, the incidences of DVT (including asymptomatic DVT) and clinical PE in medical patients not receiving prophylaxis were respectively estimated at about 19% and 1% [bib_ref] Prevention of venous thromboembolism in internal medicine with unfractionated or low-molecular-weight heparins:..., Mismetti [/bib_ref]. Results from recent large clinical trials for VTE incidence rates in placebo groups were: 15% in the MEDENOX study [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] , 5% in PREVENT study [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] , and 11% in ARTEMIS study [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref]. Differences in asymptomatic outcomes defi nition (proximal alone or distal and proximal DVT), and methods to measure these outcomes (venography or compression ultrasonography) could explain this relatively wide interval for incidence rates.
Among hospitalized patients, elderly medical patients are probably the most affected by VTE. As a result, the proportion of preventable events by an effective thromboprophylaxis is expected to be high in this population. Moreover, because of VTE specifi cities in elderly (high mortality rate, severe initial presentation, diffi cult diagnosis of PE and high risk of major bleeding with anticoagulant drugs at therapeutic doses), primary prophylaxis with the use of an effective and safe intervention appears to be the best approach to VTE management.
## Management of thromboprophylaxis in elderly medical patients do all hospitalized elderly medical patients need a thromboprophylaxis?
The fi rst question in front of a newly admitted elderly patient is whether they will require prevention against VTE; in other words, we have to systematically assess the individual risk of venous thrombosis, and select the patients for whom thromboprohylaxis would be helpful.
This assessment, and therefore the decision of prescribing a prophylaxis, should be based on the presence or absence of specifi c risk factors [fig_ref] Table 1: Risk factors of venous thromboembolism in medical patients [/fig_ref]. Previous studies have identifi ed a number of risk factors for VTE in medical patients [bib_ref] A population-based perspective of the hospital incidence and case-fatality rates of deep..., Anderson [/bib_ref] [bib_ref] An epidemiologic study of risk factors for deep vein thrombosis in medical..., Samama [/bib_ref] [bib_ref] The epidemiology of venous thromboembolism in the community, Heit [/bib_ref] [bib_ref] Relative impact of risk factors for deep vein thrombosis and pulmonary embolism...., Heit [/bib_ref] [bib_ref] The management and outcome of acute venous thromboembolism: a prospective registry including..., Arcelus [/bib_ref]. Many of these factors are frequent in elderly hospitalized patients, alone or combined and generally cumulative (Kniffi n et al 1994; [bib_ref] Venous thrombosis: a multicausal disease, Rosendaal [/bib_ref]. However, individual risk assessment in medical settings can be problematic in practice.
In clinical trials on thromboprophylaxis in medical settings, eligibility criteria included most of the various risk factors (Table 2) but they differ between studies and cannot be easily repeated [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref]. Thromboprophylaxis has been studied specifi cally in some medical conditions such as stroke [bib_ref] What is the optimal pharmacological prophylaxis for the prevention of deep-vein thrombosis..., Kamphuisen [/bib_ref] [bib_ref] The effi cacy and safety of enoxaparin versus unfractionated heparin for the..., Sherman [/bib_ref] , myocardial infarction [bib_ref] Low dose heparin in the prevention of deep-vein thromboses in patients with..., Pitt [/bib_ref] or cancer [bib_ref] Double-blind randomised trial of a very-low-dose warfarin for prevention of thromboembolism in..., Levine [/bib_ref] [bib_ref] Low molecular weight heparin, therapy with dalteparin, and survival in advanced cancer:..., Kakkar [/bib_ref]. Instead of providing clarifi cation, the decision of whom and how to treat may become more diffi cult as a result.
Nevertheless, assessment of the risk of VTE is probably easier for elderly than for other medical patients. Because age over 75 years is an independent risk factor of VTE in medically ill patients [bib_ref] Risk factors for venous thromboembolism in hospitalized patients with acute medical illness:..., Alikhan [/bib_ref] , the presence of only one additional acute medical condition is required to consider thromboprophylaxis for these patients, using for example the risk assessment model developed by . A thromboprophylaxis should be proposed if the hospitalization is motivated by one of these following conditions: acute myocardial infarction, acute heart failure, active cancer requiring therapy, acute infectious disease, respiratory disease, acute rheumatologic disease, ischemic stroke, paraplegia, and infl ammatory disorder with immobility or infl ammatory bowel disease. In conclusion, elderly medical patients should receive a thromboprophylaxis if an acute medical illness is the leading cause of hospitalization.
## Which prophylaxis should be used?
Among thromboprophylactic methods, anticoagulant treatments with UFH or LMWH are the regimens of choice, but they share main contraindications: untreated hemophilia and other hemorrhagic disorder, active bleeding or perceived bleeding risk ( peptic ulcer, recent cerebral hemorrhage, severe hypertension, and severe liver disease), thrombocytopenia with platelets Ͻ60 × 10 9 /L, a history of heparin-induced thrombocytopenia, and hypersensitivity to heparin.
Because of its principally renal route of elimination, LMWH, even at prophylactic doses, must be used with caution in patients with renal failure. When creatinine clearance is less than 30 mL/min, UFH may be preferred.
Before prescribing an anticoagulant regimen, we need to check systematically the potential contraindications: baseline risk of bleeding, platelets count and renal function. In the event that there is a contraindication to pharmacological thromboprophylaxis, consider nonpharmacological treatments. Nonpharmacological strategies include graduated compression stockings, intermittent pneumatic compression, leg elevation, and early mobilization. No randomized clinical trials have evaluated mechanical methods of prophylaxis in general medical patients. These measures should be considered as alternative methods when contraindications to anticoagulants exist.If there is no contraindication to pharmacological treatments, use LMWH or UFH.
Prophylactic treatment versus placebo [fig_ref] Table 3: Double-blind randomized clinical trials evaluating thromboprophylaxis versus placebo in various medical patients [/fig_ref] Seven trials comparing a prophylactic heparin treatment to a control (15,095 patients) in medical patients were selected in the meta-analysis by . Four studies evaluated UFH at 10000 IU or 15000 IU daily [bib_ref] Prevention of deep vein thrombosis in medical patients by low-dose heparin, Belch [/bib_ref] [bib_ref] Comparison of sodium and calcium heparin in prevention of venous thromboembolism, Cade [/bib_ref] [bib_ref] Randomised, controlled trial of low-dose heparin for prevention of fatal pulmonary embolism..., Gardlund [/bib_ref] and three studies evaluated a LMWH (enoxaparin or nadroparin) [bib_ref] Prevention of deep vein thrombosis in elderly medical in-patients by a low..., Dahan [/bib_ref] [bib_ref] Heparin prophylaxis in bedridden patients, Bergmann [/bib_ref] [bib_ref] Nadroparin in the prevention of deep vein thrombosis in acute decompensated COPD...., Fraisse [/bib_ref]. A signifi cant decrease in DVT and in clinical PE were observed with heparins as compared to control (risk reductions of 56% and 58% respectively, p Ͻ 0.001 in both cases), without signifi cant difference in the incidence of major bleedings or deaths.
In the MEDENOX study, patients of 75 years or older accounted for more than half of acute ill medical patients included in the trial [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref]. A post-hoc analysis was performed in the subgroup of older patients [bib_ref] Prevention of venous thromboembolism in medical patients with enoxaparin: a subgroup analysis..., Alikhan [/bib_ref]. Those who received 40 mg enoxaparin had a 78% reduction in VTE when compared with placebo (absolute risk reduction 14.4%; relative risk [RR] 0.22; 95% CI, 0.09-0.51). The incidence of adverse effects, including hemorrhages, did not differ signifi cantly between the placebo group and either enoxaparin group in the primary analysis, but no data were reported for the subgroup of elderly patients.
The PREVENT trial included medical patients randomized to receive dalteparin, 5000 IU once daily, or placebo [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref]. In the subgroup of patients 75 years or older, who represented one third of the population included in the trial, the primary end point was reported in 4.2% with dalteparin and 8.0% with placebo (RR 0.52; 95% CI, 0.31-0.87) [bib_ref] Effi cacy and safety of fi xed low-dose dalteparin in preventing venous..., Kucher [/bib_ref]. Dalteparin was not associated with a signifi cant increase in major hemorrhage (1.1% vs 0.7%; p = 0.12).
The ARTEMIS study was designed to evaluate the effi cacy and safety of the anticoagulant fondaparinux in older acute medical inpatients at moderate to high risk of VTE [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref]. To our knowledge, no specifi c data concerning the subgroup of elderly patients were already published.
Two double-blind randomized studies evaluated thromboprophylaxis without systematic assessment of asymptomatic DVT at the end of the treatment period [bib_ref] Heparin prophylaxis in bedridden patients, Bergmann [/bib_ref] [bib_ref] The prophylaxis of medical patients for thromboembolism pilot study, Lederle [/bib_ref]. In the fi rst study, nadroparin did not have a signifi cant effect on mortality and the study provided no data suggesting that nadroparin may reduce the incidence of thromboembolic events in patients hospitalized for an acute medical disease [bib_ref] Heparin prophylaxis in bedridden patients, Bergmann [/bib_ref] [bib_ref] Lack of effect of a lowmolecular-weight heparin (nadroparin) on mortality in bedridden..., Mahe [/bib_ref]. In the second study, the primary outcome was the total mortality at 90 days after randomization, but authors reported also the number of readmission due to objectively proven symptomatic venous thromboembolic events occurring within 90 days of randomization [bib_ref] The prophylaxis of medical patients for thromboembolism pilot study, Lederle [/bib_ref]. Overall mortality and occurrence of clinical thromboembolic events did not differ between enoxaparin and placebo.
## Comparison of prophylactic anticoagulant regimens
A meta-analysis of eight trials comparing UFH and LMWHs in medical patients found no signifi cant differences between the two treatment groups on the incidence of DVT, clinical PE or mortality [bib_ref] Prevention of venous thromboembolism in internal medicine with unfractionated or low-molecular-weight heparins:..., Mismetti [/bib_ref]. However, LMWH seemed to be safer than UFH by reducing the risk of major hemorrhage (RR 52%, p = 0.049).
In the recent meta-analysis by [bib_ref] Pharmacological venous thromboembolism prophylaxis in hospitalized medical patients. A meta-analysis of randomized..., Wein [/bib_ref] that included studies of patients with acute ischemic stroke, LMWH was associated with a reduced risk of DVT (RR, 0.68; 95% CI, 0.52-0.88) and injection site hematoma (RR, 0.47; 95% CI, 0.36-0.62) compared with UFH. No statistically signifi cant differences were observed between the two treatments with respect to PE, mortality, total bleeding, and major or minor bleeding.
Enoxaparin 20 mg once daily and UFH 5000 IU twice daily were considered equivalent in terms of effi cacy in patients aged 65 years or more (mean age 83 years) and no signifi cant difference was found in terms of safety . However, MEDENOX study suggested that enoxaparin at 20 mg was ineffective for preventing VTE in acutely ill medical patients compared to placebo [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref].
Enoxaparin 40 mg subcutaneously once daily was shown to be as effective as UFH 5000 IU three times daily in the PRIME and PRINCE studies [bib_ref] The venous thrombotic risk in nonsurgical patients: epidemiologal data and effi cacy/safety..., Lechler [/bib_ref] [bib_ref] Study to compare the effi cacy and safety of the LMWH Enoxaparin..., Kleber [/bib_ref]. The PRIME study included various medical patients (mean age 74 years), and the PRINCE study included patients (mean age 70 years) with severe respiratory disease or heart failure.
The recent PREVAIL study was designed to assess the effi cacy and safety of enoxaparin versus UFH for the prevention of VTE in patients with acute ischemic stroke [bib_ref] The effi cacy and safety of enoxaparin versus unfractionated heparin for the..., Sherman [/bib_ref]. The mean age of included patients was 66 years, and 25% of patients were older than 75 years. More than No studies have directly compared different modalities of UFH administration. However, a recent meta-analysis comparing UFH given twice daily or 3 times daily from 12 studies found no difference in the overall rate (per 1000 patient-days) of VTE (twice daily: 5.4 vs 3 times daily: 3.5; p = 0.87), but an increased risk of major bleedings with three times daily heparin (twice daily: 0.35 vs 3 times daily: 0.96; p Ͻ 0.001) [bib_ref] Twice vs three times daily heparin dosing for thromboembolism prophylaxis in the..., King [/bib_ref].
In regards to the risk of heparin-induced thrombocytopenia (HIT) that should be considered when providing patients with thromboprophylaxis, LWMHs appeared safer than UFH in surgical patients [bib_ref] Heparin-induced thrombocytopenia in patients treated with low-molecular-weight heparin or unfractionated heparin, Warkentin [/bib_ref]. However, in the recent meta-analysis by , no statistically signifi cant differences were observed between UFH and LMWH with respect to thrombocytopenia (RR, 0.25; 95% CI, 0.05-1.16).
## Cost effectiveness analyses
Several studies conducted in different countries have evaluated the cost effectiveness of LMWH relative to UFH or to no pharmacological prophylaxis for DVT prevention in medical inpatients [bib_ref] Economic evaluation of the MEDENOX trial: a Canadian perspective, Lamy [/bib_ref] [bib_ref] Cost-effectiveness of enoxaparin as thromboprophylaxis in acutelly ill medical patients from the..., Nuijten [/bib_ref] [bib_ref] Cost effectiveness of enoxaparin as prophylaxis against venous thromboembolic complications in acutely..., Schadlich [/bib_ref] [bib_ref] Low-molecular weight heparin for deep vein thrombosis prophylaxis in hospitalized medical patients:..., Shorr [/bib_ref]. Regardless of some methodological differences, they conclude that, despite higher acquisition costs, LMWHs for thromboprophylaxis in medical patients resulted in savings compared to no pharmacological prophylaxis and were associated with lower costs or at least neutral cost compared with UFH.
To conclude, based on the results of the clinical trials presented above, elderly patients hospitalized for an acute medical illness should receive a prophylactic anticoagulant treatment unless there is a contraindication. LMWHs, enoxaparin at the dose of 40 mg and dalteparin at the dose of 5000 IU, are effective and safe to prevent VTE in elderly patients. Compared to UFH, they are more convenient, requiring only one subcutaneous injection daily. They are associated with a lower risk of heparin-induced thrombocytopenia, and fi nally are equal or superior in cost effectiveness analyses. In case of severe alteration of renal function however, the safety of LMWHs is uncertain and UFH should be considered. More data are needed to recommend the use of fondaparinux as a routine treatment in elderly patients. In case of contraindication to heparins, graduated stockings or intermittent pneumatic compression should be used.
## For how long should the treatment be given?
In the main randomized trials, prophylactic treatment was given for 1 to 2 weeks. While no specifi c recommendations exist regarding the duration of anticoagulant treatment, treatment would therefore not exceed this length. However, some elements could be discussed in favor of duration extension in medical patients, and especially in elderly medical patients.
First, it is likely that the thrombotic process has already started in some patients before their admission to hospital. In a study evaluating the prevalence of VTE in medical patients hospitalized for another reason than VTE, asymptomatic VTE at admission was diagnosed in 17.8% of patients over 80 years [bib_ref] High prevalence of asymptomatic deep vein thrombosis on admission in a medical..., Oger [/bib_ref].
Second, in the MEDENOX, PREVENT and ARTEMIS studies, additional symptomatic VTE events and fatal PE occurred after discontinuation of prophylaxis during the follow up period in the treatment groups [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref].
Third, extension of prophylaxis duration was shown to be effective in other settings. In major orthopedic surgery, a meta-analysis of randomized trials confi rmed that prolonged prophylaxis with LWMH for one month postoperatively, compared with the standard practice of prescribing prophylaxis for 7-10 days, reduced the relative risk of all DVT events by 59% (p Ͻ 0.001), and reduced the risk of symptomatic VTE by 64% (p Ͻ 0.001), without increasing bleeding complications [bib_ref] Extended out-of-hospital lowmolecular-weight heparin prophylaxis against deep venous thrombosis in patients after..., Hull [/bib_ref]. Benefi t of prolonged prophylaxis to 25-31 days was also demonstrated in patients with active cancer who were undergoing surgery .
The Extended Clinical Prophylaxis in Acutely III Medical Patients (EXCLAIM) trial was designed to demonstrate the superiority of extended-duration prophylaxis with enoxaparin (28 ± 4 days) compared with the standard regimen of enoxaparin (10 ± 4 days) for prevention of VTE in acutely ill medical patients with recent reduced mobility [bib_ref] Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recent reduced..., Hull [/bib_ref]. Mean age of included patients was 70 years, with more than 40% of patients being 75 years or older. Extendedduration prophylaxis with enoxaparin was more effective than standard regimen of enoxaparin on the primary effi cacy endpoint (asymptomatic DVT detected by ultrasonography, symptomatic DVT, symptomatic PE, fatal PE) (2.8% vs 4.9%, RR 0.56, 95% CI 0.39-0.80). However, major hemorrhagic complications occurred signifi cantly more frequently with the extended-duration prophylaxis (0.60% vs 0.15%, RR 4.03; 95% CI 1.14-14.3) [bib_ref] Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recent reduced..., Hull [/bib_ref].
One or two weeks is the common duration of prophylactic treatment with LMWH or UFH. Extension of prophylaxis duration to one month appeared to be more effective than standard duration but was also associated with an excess of major bleedings. Awaiting further analyses of the EXCLAIM study, extended-duration prophylaxis in medical patients should be discussed on a case-by-case basis.
## Is a monitoring of pharmacological prophylaxis required?
The highest risk of anticoagulant prophylaxis is bleeding. Even if major bleedings are less frequent with prophylactic than with therapeutic doses, physicians should be aware of this risk, and should observe their patients regularly by clinical examination for all potential bleedings. Moreover, underlying conditions or associated treatments could change during hospitalization, and this could increase the bleeding risk associated with thromboprophylaxis. A regular reassessment of the benefi t and risk of thromboprophylaxis decided on admission should be performed taking into account the potential modifi cations since admission, for example, the modifi cations of renal function or the presence of a newly diagnosed contraindication. However, no specifi c biological monitoring is needed to evaluate the anticoagulant activity of UFH or LMWH at prophylactic doses.
Less frequent, but often fatal, is the potential development of HIT. Even if HIT appeared to be less frequently observed with LMWH than with UFH, a monitoring of platelets count is required at least twice weekly for all heparins during the treatment period [bib_ref] Heparin-induced thrombocytopenia: recognition, treatment, and prevention, Warkentin [/bib_ref].
Regular assessment of bleeding risk associated with thromboprophylaxis should be performed, taking into account the potential changes in underlying conditions and associated treatments. A monitoring of platelets count is required for UFH as well as for LMWH.
A proposition of management of thromboprophylaxis in elderly medical patients is resumed in [fig_ref] FourthFigure 1: step [/fig_ref]. Check the creatinine level and estimate the creatinine clearance
## First step: assessment of vte risk does the patient have one of the after acute medical illnesses?
Acute myocardial infarction Acute heart failure Active cancer requiring therapy Acute infectious disease Respiratory disease Acute rheumatologic disease Ischemic stroke Paraplegia Inflammatory disorder with immobility Inflammatory bowel disease Second step: Is there a contraindication to pharmacological thromboprophylaxis? Assessment of baseline bleeding risk, history of heparin induced thrombocytopenia or heparin hypersensitivity, check platelets count
## Vte prophylaxis in elderly patients
## Why is thromboprophylaxis in the elderly not better adopted in clinical practice?
The gap between theory and practice Despite available data, and recommendations for VTE prevention in medical patients, underuse or misuse of thromboprophylaxis remains a problem. A case series study showed that among 208 patients who developed acute VTE during hospitalization in medical wards, only 48% had received VTE prophylaxis [bib_ref] New onset of venous thromboembolism among hospitalized patients at Brigham and Women's..., Goldhaber [/bib_ref]. Recently, the Canadian CURVE study aimed to determine the proportion of hospitalized, acutely ill medical patients who were eligible to receive thromboprophylaxis, and to evaluate the frequency, determinants and appropriateness of its use [bib_ref] Multicenter evaluation of the use of venous thromboembolism prophylaxis in acutely ill..., Kahn [/bib_ref]. Following ACCP guidelines, authors found that of the 1894 medical hospitalized patients who were eligible for study inclusion, 90% should receive a thromboprophylaxis. However, only 23% received a prophylaxis, and only 16% received an appropriate thromboprophylaxis.
Use of prophylaxis in elderly medical patients is unknown, but thought to be as low as in general medical patients. Nevertheless, in CURVE study, age was not found as a determinant of thromboprophylaxis use [bib_ref] Multicenter evaluation of the use of venous thromboembolism prophylaxis in acutely ill..., Kahn [/bib_ref] , and in IMPROVE study, prophylaxis with LMWH or UFH was more likely to be used if patients were aged Ն85 years . In the study by , patients treated by thromboprophylaxis were older than those without prophylaxis. The rate of patients needing thromboprophylaxis and not receiving it was lower for older patients (Ͼ75 years) than for younger patients. However, the rate of patients not needing thromboprophylaxis and receiving it was higher for older patients than for younger patients.
## Explaining the gap
The gap between theory and practice for the use of thromboprophylaxis could be explained by many reasons. Heterogeneity of clinical studies, selected populations, concern about bleedings, and the lack of a clear clinical benefi t are some of these reasons.
## Heterogeneity of clinical trials
Clinical trials highlighted the benefi t of pharmacological treatment for VTE prevention in medical patients [bib_ref] A comparison of enoxaparin with placebo for the prevention of venous thromboembolism..., Samama [/bib_ref] [bib_ref] Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in..., Leizorovicz [/bib_ref] [bib_ref] Effi cacy and safety of fondaparinux for the prevention of venous thromboembolism..., Cohen [/bib_ref]. However, the different rates of VTE incidence from 4.8% to 14.9% in the placebo groups suggested disparities between studies. This may refl ect the heterogeneous medical conditions of the different populations studied, as well as differences in the endpoints and outcomes measure. In fact, any study could be directly compared to another. When two studies shared the same inclusion criteria, treatments and/or primary end points and/or modalities to measure the outcomes were different.
As presented in [fig_ref] Table 2: Eligibility criteria in MEDENOX, PREVENT, and ARTEMIS clinical trials Additional risk factors... [/fig_ref] , eligibility criteria of clinical trials, and the resulting validated indications of thromboprophylaxis evaluated in these trials are various and different between studies. Compared to surgical settings, there is a lack of legibility of eligibility criteria for the use of thromboprophylaxis in medical settings. This complexity is discouraging and could explain its underuse.
Homogenization of outcomes definition, outcomes measure and simplifi cation of eligibility criteria are probably needed for further trials. This may help to clarify the recommendations and to improve the practice.
## Selected medical populations, and probably high selected elderly populations
Inclusion and exclusion criteria of clinical trials lead to selected populations, and results should be interpreted with caution before being extrapolated to a larger population. Taking into account the high proportion of patients hospitalized in medical wards, the number of patients included in the clinical trials was low. In IMPROVE, only 13%-19% of all patients would have been eligible for inclusion in MEDENOX, PREVENT, or ARTEMIS studies [bib_ref] A multinational observational cohort study in hospitalized medical patients of practices in..., Anderson [/bib_ref]. This problem can be also illustrated by the results of the CURVE study [bib_ref] Multicenter evaluation of the use of venous thromboembolism prophylaxis in acutely ill..., Kahn [/bib_ref]. Over the 3-week study period, in 29 hospitals across Canada, 1702 medical patients had an indication of thromboprophylaxis. This number is higher than the number of included patients in MEDENOX study over 18 months in 60 centers from nine countries.
Furthermore, many patients who were usually hospitalized in geriatric units, were unable to sign a written informed consent, presented alteration of renal function or low weight were excluded from these trials. Renal insuffi ciency and low weight, two frequent conditions in elderly patients and associated with higher risk of bleeding [bib_ref] Venous thromboembolism in very elderly patients: fi ndings from a prospective registry..., Lopez-Jimenez [/bib_ref] , were often exclusion criteria. Therefore, selection of a population with low risk of bleeding in clinical trials was likely and concern about bleeding with thromboprophylaxis could be a reason of its underuse in elderly patients. Moreover, as an informed consent is required, patients need to understand and approve the enrolment in the study. Unfortunately, only few patients among the oldest are able to give such consent, because of impairment of cognitive functions, or too severe weakness to be able to participate.
Despite fi ndings of clinical trials suggesting that thromboprophylaxis is effective and safe in the elderly patients, few trials enrolled specifi cally geriatric patients. Additionally, elderly patients were probably highly selected and therefore not representative of overall elderly medical patients.
## Lack of evident clinical benefi t?
Only few trials were conducted using symptomatic VTE and/or mortality as the primary end point. In the majority of clinical trials, the primary end point was a combined outcome including symptomatic VTE (DVT and/or fatal and non fatal PE) and asymptomatic DVT systematically assessed at the end of the treatment period. Asymptomatic DVT, however, were the great majority of events detected during the studies. In medical patients, there is a lack of evidence that prophylaxis prevents clinically important outcomes, such as PE, which has been shown in surgical patients [bib_ref] Reduction in fatal pulmonary embolism and venous thrombosis by perioperative administration of..., Collins [/bib_ref]. Therefore, some could argue about the clinical relevance of the results from trials conducted in medical settings.
A recent meta-analysis aimed to evaluate the effects of anticoagulant prophylaxis in reducing clinically important VTE outcomes in hospitalized medical patients [bib_ref] Meta-analysis: anticoagulant prophylaxis to prevent symptomatic venous thromboembolism in hospitalized medical patients, Dentali [/bib_ref]. Prophylaxis (with UFH, LMWH or fondaparinux) was associated with a signifi cant reduction of any PE (RR: 0.43, CI, 0.26-0.71) and fatal PE (RR: 0.38, CI, 0.21-0.69), a nonsignificant reduction in symptomatic DVT and a nonsignifi cant increase in major bleeding. Anticoagulant prophylaxis had no effect on all-cause mortality. However, the main limitation of this analysis is that the results concerned only the treatment period. How are they maintained after the treatment has been stopped (which is really the relevant question) is unknown.
The benefi t of prophylaxis in medical patients is uncertain for clinical nonfatal VTE events, and no benefi t on overall mortality was demonstrated. How these results are relevant for elderly population remains to be determined.
## Challenges to improve the practice
As life expectancy is increasing in developed countries, incidence of VTE is also expected to increase. How to prevent this potential fatal disease will arise as a major public health problem for the next decades. According to the Agency for Healthcare Research and Quality in the United States, prevention of VTE in adult patients in hospital was the main challenge to patient safety in 2001 (Agency for Healthcare Research and Quality 2001). Therefore, many efforts have to be done to improve thromboprophylaxis particularly in the elderly. Determining the optimal strategy of prevention is essential and further trials conducted specifi cally in the elderly nonselected patients and evaluating the benefi t-risk ratio of thromboprophylaxis are needed. However, available results suggested that some prophylactic treatments were effective and safe even in the older patients. These treatments should be more used in practice following the available guidelines. But the use of such treatments in the "real life" and for patients that would have not been included in the clinical trials must be evaluated. Reducing the gap between theory and practice shown by many studies should reduce the number of patients exposed to high risk of thrombosis and not receiving effective prevention.
To improve practice, some suggestions should be proposed in all hospitals:
- Results of epidemiologic studies highlighted the importance of VTE in in-hospital morbi-mortality should be largely taught. This would reduce the lack of awareness that VTE is a real public health problem among all practitioners. - Models of assessment of the VTE risk should be validated and more used. Assessment of the risk of VTE for all medical patients is the fi rst recommendation of current guidelines [bib_ref] Prevention of venous thromboembolism. International Consensus Statement. Guidelines compiled in accordance with..., Nicolaides [/bib_ref]. However, assessing the individual risk of VTE in medical settings is diffi cult. Models to assess this risk have been developed to assist the physicians to determine if thromboprophylaxis is warranted or not [bib_ref] Quantification of risk factors for venous thromboembolism: a preliminary study for the..., Samama [/bib_ref] [bib_ref] An electronic tool for venous thromboembolism prevention in medical and surgical patients, Samama [/bib_ref] [bib_ref] Assessment of venous thromboembolism risk and the benefi ts of thromboprophylaxis in..., Cohen [/bib_ref]. But these models have not been validated in specifi cally designed prospective clinical trials, they are less structured than in surgical settings and are not conceived for elderly patients. Nevertheless, systematic assessment of VTE risk at admission has been shown to improve the prophylaxis practice in an original initiative conducted in Brigham and Women's Hospital in Boston [bib_ref] Electronic alerts to prevent venous thromboembolism among hospitalized patients, Kucher [/bib_ref]. A computer program linked to the patient database was developed to identify consecutive hospitalized patients at risk for DVT in the absence of prophylaxis. Eligible patients were randomly assigned to an intervention group, in which the responsible physician was alerted to a patient's risk of DVT, and to a control group, in which no alert was issued. The computer alert reduced the risk of VTE at 90 days by 41% (HR, 0.59; 95% CI, 0.43-0.81). These encouraging fi ndings suggested that alerting physicians to the individual risk of VTE could contribute to improve the rate of clinical outcomes.
- Regular updates on prevention of VTE in medical settings should also be proposed in Continuing Medical Education programs. On these occasions, last data and last recommendations of experts could be presented and discussed.
Finally, further studies are needed to determine the place of the new anticoagulant drugs, such as direct Factor Xa inhibitors or antithrombin inhibitors, in the prevention of VTE in medical settings. Besides their effectiveness and safety demonstrated in surgical settings, their pharmacokinetic properties (that allow one oral daily dose) provide other advantages that should be useful in daily practice.
# Conclusions
Elderly patients are the most concerned by VTE, and development of optimal thromboprophylaxis strategies should be a priority in this population. Further clinical trials evaluating effi cacy and risk of thromboprophylactic methods, including new anticoagulant drugs, are needed for elderly patients. Awaiting such trials, and despite some limitations, available data from existing clinical trials conducted in hospitalized medical patients supported a good benefi t-risk ratio of different thromboprophylaxis modalities in geriatric patients. Nevertheless, in practice, the balance between expected benefi t and risk of prophylactic anticoagulation should be carefully assessed for each individual elderly patient taking into account the specifi c risk of VTE in older patients and the risk of anticoagulant related bleeding probably underestimated in clinical trials. Risk assessment models developed for overall medical patients are useful tools to improve the use of thromboprophylaxis even in elderly patients, but remain to be evaluated in this specifi c population.
[fig] FourthFigure 1: step: LMWH ca n be used: Enoxaparin 40 mg s ubcutaneous once daily or dalteparin 5000 IU s ubcutaneous once daily for 1Management of thromboprophylaxis in elderly (Ͼ75 years) medical patients. [/fig]
[table] Table 1: Risk factors of venous thromboembolism in medical patients [/table]
[table] Table 2: Eligibility criteria in MEDENOX, PREVENT, and ARTEMIS clinical trials Additional risk factors of VTE were: age Ն75 years, cancer, previous VTE, obesity, varicose veins, hormone therapy (antiandrogen or estrogen, except for postmenopausal hormone-replacement therapy), and chronic heart or respiratory failure Additional risk factors of VTE were: age Ն75 years, cancer, previous VTE, obesity, varicose veins and/or chronic venous insuffi ciency, hormone replacement therapy, history of chronic heart failure, chronic respiratory failure, or myeloproliferative syndrome. [/table]
[table] Table 3: Double-blind randomized clinical trials evaluating thromboprophylaxis versus placebo in various medical patients [/table]
|
Revealing Differences in Metabolic Flux Distributions between a Mutant Strain and Its Parent Strain Gluconacetobacter xylinus CGMCC 2955
A better understanding of metabolic fluxes is important for manipulating microbial metabolism toward desired end products, or away from undesirable by-products. A mutant strain, Gluconacetobacter xylinus AX2-16, was obtained by combined chemical mutation of the parent strain (G. xylinus CGMCC 2955) using DEC (diethyl sulfate) and LiCl. The highest bacterial cellulose production for this mutant was obtained at about 11.75 g/L, which was an increase of 62% compared with that by the parent strain. In contrast, gluconic acid (the main byproduct) concentration was only 5.71 g/L for mutant strain, which was 55.7% lower than that of parent strain. Metabolic flux analysis indicated that 40.1% of the carbon source was transformed to bacterial cellulose in mutant strain, compared with 24.2% for parent strain. Only 32.7% and 4.0% of the carbon source were converted into gluconic acid and acetic acid in mutant strain, compared with 58.5% and 9.5% of that in parent strain. In addition, a higher flux of tricarboxylic acid (TCA) cycle was obtained in mutant strain (57.0%) compared with parent strain (17.0%). It was also indicated from the flux analysis that more ATP was produced in mutant strain from pentose phosphate pathway (PPP) and TCA cycle. The enzymatic activity of succinate dehydrogenase (SDH), which is one of the key enzymes in TCA cycle, was 1.65-fold higher in mutant strain than that in parent strain at the end of culture. It was further validated by the measurement of ATPase that 3.53-6.41 fold higher enzymatic activity was obtained from mutant strain compared with parent strain.
# Introduction
Plant-based cellulose, due to its abundance and low-cost, has attracted increasing attention in recent years [bib_ref] Biofuels in China: past, present and future, Zhong [/bib_ref] [bib_ref] Optimization of enzymatic hydrolysis and ethanol fermentation from AFEX-treated rice straw, Zhong [/bib_ref]. However, the presence of lignin, hemicelluloses and other molecules in lignocellulosic biomass makes it complicated processes for medical use [bib_ref] Biofuels in China: past, present and future, Zhong [/bib_ref] [bib_ref] Bacterial cellulose-based materials and medical devices: current state and perspectives, Petersen [/bib_ref]. Bacterial cellulose (BC) is an insoluble, extracellular polysaccharide that is produced by certain types of microorganisms, such as Acetobacter species [bib_ref] Cellulose: fascinating biopolymer and sustainable raw material, Klemm [/bib_ref]. It is a highly pure form of cellulose with a fine nano-scale structure and has been widely used in food and biomedical fields for various applications [bib_ref] Preparation and application of bacterial cellulose sphere: a novel biomaterial, Zhu [/bib_ref] [bib_ref] The influence of fermentation conditions and post-treatment methods on porosity of bacterial..., Tang [/bib_ref] [bib_ref] Water holding and release properties of bacterial cellulose obtained by in situ..., Ul-Islam [/bib_ref]. BC has the same chemical structure as plant cellulose, but exhibits superior physical and chemical properties, including high mechanical tensile strength, purity, biodegradability and waterholding capacity [bib_ref] Bacterial cellulose-based materials and medical devices: current state and perspectives, Petersen [/bib_ref] [bib_ref] Preparation and application of bacterial cellulose sphere: a novel biomaterial, Zhu [/bib_ref] [bib_ref] Water holding and release properties of bacterial cellulose obtained by in situ..., Ul-Islam [/bib_ref].
Recently, investigations on developing cost-effective culture processes have been broadly carried out using Gluconacetobacter, Acetobacter etc., to obtain maximum productivity of BC [bib_ref] Statistical optimization of culture conditions for bacterial cellulose production by Acetobacter xylinum..., Zeng [/bib_ref] [bib_ref] Improvement of bacterial cellulose production in Acetobacter xylinum using byproduct produced by..., Ha [/bib_ref]. Till now, increasing BC yield is still a big challenge which limits its further application in various fields [bib_ref] Improvement of bacterial cellulose production by manipulating the metabolic pathways in which..., Li [/bib_ref]. Bae and Shoda [bib_ref] Production of bacterial cellulose by Acetobacter xylinum BPR2001 using molasses medium in..., Bae [/bib_ref] reported that adding molasses at a lower concentration of 20 g/L into the cultivation medium improved the efficiency and the econometrics. Jung et al. [bib_ref] Influence of glycerol on production and structural-physical properties of cellulose from Acetobacter..., Jung [/bib_ref] used glycerol as the sole carbon source as opposed to glucose, after seven days of cultivation, BC yield and crystallinity index were 380% and 9% higher, respectively. It has also been reported that cellulose production by G. xylinus on glucose medium was enhanced in batch culture when ethanol was present in the media [bib_ref] Factors affecting the yield and properties of bacterial cellulose, Krystynowicz [/bib_ref]. However, there is limited knowledge regarding the metabolism of existing strains of Gluconacetobacter, which makes it difficult to understand the metabolic network and relate it to the production of interesting products (such as acetic acid, cellulose, etc.) [bib_ref] The unified metabolism of Gluconacetobacter entanii in continuous and batch processes, Velasco-Bedrán [/bib_ref] [bib_ref] Metabolic flux analysis of Gluconacetobacter xylinus for bacterial cellulose production, Zhong [/bib_ref].
Metabolic flux analysis provides theoretical foundations for analyzing and understanding BC biosynthesis pathway and extraand intracellular metabolism at specific conditions [bib_ref] The benefits of being transient: isotope-based metabolic flux analysis at the short..., Nöh [/bib_ref]. In an effort to obtain highly productive strains, researchers have applied genetic engineering methods combined with metabolic flux analysis to adjust or shift the desired metabolic pathway toward desirable products [bib_ref] Genome-derived minimal metabolic models for Escherichia coli MG1655 with estimated in vivo..., Taymaz-Nikerel [/bib_ref]. Velasco-Bedrán and Lopezsunza-Isunza [bib_ref] The unified metabolism of Gluconacetobacter entanii in continuous and batch processes, Velasco-Bedrán [/bib_ref] analyzed the catabolic and anabolic pathway for Gluconacetobacter central metabolism, but the metabolic fluxes towards biomass accumulation were not taken into consideration. In one of our recent researches, the metabolic network in Gluconacetobacter xylinus CGMCC 2955 was constructed, and the metabolic flux distributions in this strain cultured on three different carbon sources (glucose, fructose and glycerol) were compared [bib_ref] Metabolic flux analysis of Gluconacetobacter xylinus for bacterial cellulose production, Zhong [/bib_ref]. It has been revealed that byproducts accumulation in bypasses would limit the BC productivity. Meanwhile, the nucleic acids, proteins, ATP, NADPH, and other molecules weregenerated as well, which could be used as precursors for cell growth and BC production [bib_ref] Improvement of bacterial cellulose production by manipulating the metabolic pathways in which..., Li [/bib_ref]. In our previous study [bib_ref] Metabolic flux analysis of Gluconacetobacter xylinus for bacterial cellulose production, Zhong [/bib_ref] , metabolic flux analysis for the central carbon metabolism revealed that about 47.96% of glycerol was transformed into BC, while only 19.05% of glucose and 24.78% of fructose were transformed into BC. Instead, when glucose was used as the sole carbon source, 40.03% of the carbon source was turned into the by-product gluconic acid.
In this study, a mutant strain was obtained by combined chemical mutation, with a high productivity of BC. Using this metabolic network and knowledge of the precursor molecules, metabolic flux distributions were calculated and compared between the BC high-yield mutant and the parent strain G. xylinus CGMCC 2955. This information provides a theoretical foundation of metabolism in G. xylinus CGMCC 2955 that could inform future genetic manipulations.
# Materials and methods
Microorganism G. xylinus CGMCC2955 was screened by the Key Laboratory of Industrial Microbiology, Ministry of Education, Tianjin University of Science and Technology, stored in China General Microbiological Culture Collection Center with the registered number, NO. 2955 [bib_ref] Preparation and application of bacterial cellulose sphere: a novel biomaterial, Zhu [/bib_ref] [bib_ref] Improvement of bacterial cellulose production by manipulating the metabolic pathways in which..., Li [/bib_ref].
## Culture medium and growth conditions
Pre-cultures were grown in 500 mL baffled shake flasks with 100 mL medium containing 25 g/L glucose, 7.5 g/L yeast extract, 10 g/L peptone, and 10 g/L Na 2 HPO 4 ?2H 2 O. The medium was acidified to pH 5.5-6.0 by hydrochloric acid and sterilized at 121uC for 20 min. After culture at 30uC for 24 h, cell suspension was inoculated into other 500 mL baffled shake flasks with chemically defined medium at a ratio of 8% (v/v), static culture at 30uC for 8 days. Chemically defined medium contained glucose (25 g/L), Na 2 HPO 4 (3 g/L), KH 2 PO 4 (1 g/L), (NH 4 ) 2 SO 4 (5 g/L), MgCl 2 (0.02 g/L), CaCl 2 (0.02 g/L) and para-aminobenzoic acid (PABA) (0.0015 g/L), was used for metabolic flux analysis [bib_ref] Improvement of bacterial cellulose production by manipulating the metabolic pathways in which..., Li [/bib_ref]. The glucose was separately sterilized at 115uC for 15 min.
## Mutagenesis procedure
Combined chemical mutation was used to mutate the parent strain by DES (diethyl sulfate) with LiCl, intending to obtain a mutant strain that generates a lower quantity of byproducts, specifically gluconic acid and acetic acid. Biomass suspensions that had been treated by DES for different intervals were spread on a plate containing 0.3% (w/v) LiCl. The dose-dependent mutation lethality curve was illustrated by counting of the colonies on the above plate. The lethal dose and screening reagent were 30 min and bromophenol blue (0.4 g/L), respectively. When a colony grows in the petridish after being cultured for 4-5 d, the pH of the medium around the colony will decrease, which would lead to the color change of bromophenol blue (from yellow to blue), the ones that get lager colony and smaller color ring, are supposed to be high yield mutant.
## Analytical techniques
Acetic acid and gluconic acid were analyzed by a Waters highpressure liquid chromatography (HPLC) at 230 nm with UV detector, equipped with a Shodex NH2P-50 column (150 mm * 4.6 mm, Showa Denko K.K., Japan). The mobile phase consisted of 0.05 M KH 2 PO 4 at a flow rate of 1 mL/min.
To separate the cells from the cellulose, 2% (v/v) cellulase (5000 U/mL) was added into medium at the end of the culture, processed 2 h to degrade cellulose and release cells. Cell dry weight was determined using 10 mL cell suspensions in five replicates. Samples were centrifuged at 6,000 rpm, washed twice with 30 mL distilled water, and then dried at 80uC for 10 h to a constant weight. Cellulose production was measured based on at least five replicates, which was collected by soaking and washing the cellulose pellicle with NaOH (0.1 M) to lyse cells, washed by water until pH-neutral and dried at 80uC for 10 h to a constant weight.
## Assay of enzymatic activities
The enzymatic activity of phosphoenolpyruvate carboxylase (PEPC) in mutant and parent strains was assayed spectrophotometrically by monitoring the disappearance of NADH at 340 nm (PEPC kit, Nanjing Jiancheng Institute of Biotechnology, Nanjing, China) [bib_ref] Screening of certain mangroves for photosynthetic carbon metabolic pathway, Venkatesalu [/bib_ref]. The activity of succinate dehydrogenase (SDH) was detected using succinate/DCPIP oxidation-reduction assay kit (Nanjing Jiancheng Institute of Biotechnology, Nanjing, China) [bib_ref] Antimicrobial activity of nobiletin and tangeretin against Pseudomonas, Yao [/bib_ref]. PEPC and SDH activities were both represented as units/ OD 600 : The ATPase activity in parent and mutant strain was determined by measuring the formation of phosphoric acid from ATP [bib_ref] Scutellaria flavonoid reduced memory dysfunction and neuronal injury caused by permanent global..., Shang [/bib_ref] [bib_ref] Inhibition of pyrethroid insecticides on nerve na-K-ATPase in house files (Musca domestica), Yunzhuan [/bib_ref]. The measurement procedure was referred to the reagent kit instruction (Nanjing Jiancheng Institute of Biological Engineering, Nanjing, China) as follows:
[formula] PEPC [/formula]
ATPase activity intracellular = (A 1 2A 3 ) 4 A 2 6 C 6 D 4 T mmol/OD 600 , where A 1 , A 2 and A 3 are the absorbance of sample, standard and control, respectively. C is the phosphoric content of standard. D is the dilute times of sample and T is the optical density of cells.
Cells concentration was measured each day as optical density (OD 600 ). Samples were prepared as follows: cultured cells were harvested by centrifugation at 5000 rpm for 5 min at 4uC, 400 mL of extraction solution was added into 2,000,000 cells. Suspension was exposed to ultrasonic cell disruption system (power 20%, ultrasonic 3 seconds, interval 10 seconds, repeated for 30 times). The obtained supernatant was centrifuged at 8000 6 g for 10 min at 4uC, and then put in the ice bath before measurement.
## Flux balance model
A stoichiometric model combined with extracellular metabolite measurement was applied to the estimation of intracellular fluxes [bib_ref] Metabolic flux analysis for biosynthesis of poly (b-hydroxybutyric acid) in Alcaligenes eutrophus..., Shi [/bib_ref] [bib_ref] Metabolic flux distributions in Corynebacterium glutamicum during growth and lysine overproduction, Vallino [/bib_ref] [bib_ref] Metabolic flux analysis of lactic acid fermentation: effects of pH and lactate..., Venkatesh [/bib_ref]. A bioreaction network with branch point-associated metabolites was assembled to calculate intracellular fluxes. Network reactions in the central metabolism were determined, and the biosynthetic pathway to BC was constructed as described by Ross et al. [bib_ref] Cellulose biosynthesis and function in bacteria, Ross [/bib_ref] and Tonouchi et al. [bib_ref] Coenzyme specificity of enzymes in the oxidative pentose phosphate pathway of Gluconobacter..., Tonouchi [/bib_ref]. The reaction network consisted of 23 metabolites and 21 reactions with some unknown fluxes. A (pseudo)-steady-state approximation, in which the sum of the fluxes to and from any particular intermediary metabolite equals to zero, was used to generate the following linear mass balance equation for intracellular metabolite pools, and then calculate the flux distributions:
[formula] S Ã v~bð1Þ [/formula]
In this equation, S is the matrix of the stoichiometric bioreactions (23*21) based on the metabolic map, v is the vector of 21 unknown metabolic fluxes to be determined, and b is the vector of 23 known fluxes from measurable products and substrates for each of the 23 metabolite balances. The solution to equation (1) was determined by a constrained least-squares approach with the objective of minimizing the sum of the squares of residuals from the metabolite mass balances. The only constraint in the least-squares problem was that the fluxes be non-negative for irreversible reactions. Here we used LINGO [bib_ref] Reaction path synthesis methodology for waste minimization, Shanying [/bib_ref] [bib_ref] Modelling and optimisation of gasification for palm kernel shell (PKS) in a..., Ng [/bib_ref] to solve the equation and obtain the values of metabolic fluxes.
# Results
## Obtaining mutant strain with high bc productivity
To reduce the amount of byproducts and improve BC yield, different groups have carried out various methods that are summarized as follows. A mutant strain with high BC yield was obtained by MNNG (N-methyl-N'-nitro-N-nitrosoguanidine) or EMS (ethyl methane sulfonate) mutation. Following this, a glucose dehydrogenase (GDH)-deficient mutant of strain BPR2001, GD-I, was generated via gene disruption using the cloned gene fragment. Strain GD-I did not produce any gluconic acid, but produced 4.1 g/L of BC aerobically using glucose as the carbon source [bib_ref] Cellulose production from glucose using a glucose dehydrogenase gene (gdh)-deficient mutant of..., Shigematsu [/bib_ref]. In another study, a mutant lacking the genes for acetan production was found to have no improvement in cellulose synthesis [bib_ref] Role of water-soluble polysaccharides in bacterial cellulose production, Ishida [/bib_ref].
In this study, chemical-compound mutation was used to mutate the parent strain by DES (diethyl sulfate) with LiCl. The mutation lethality increased rapidly from 0 to 30 min, but after 30 min the increase of lethality leveled off. Thus, the DES treatment time was set at 30 minutes for which the lethality was 85.9% .
After 2 rounds of compound mutation, a mutant strain with a high BC yield and relatively high final pH (indicating low production of acid byproducts) was obtained from the parent strain G. xylinus CGMCC 2955. The BC production reached 11.75 g/L in the mutant strain (named G. xylinus AX2-16); final pH value in culture broth was 4.86, which was higher than that of the parent strain; and gluconic acid concentration was 5.71 g/L, which was about 56% lower than that of parent strain.
In order to determine the genetic stability of the mutant strain, G. xylinus AX2-16, the strain was transferred to Erlenmeyer flasks for observation of BC and byproduct generation (estimated by pH) during serial passages [fig_ref] Figure 2: Transgenerational variation in the production of bacterial cellulose and the pH value... [/fig_ref]. From passage 1 to 3, the BC production and pH value decreased to 11.75 g/L and 4.86, however the two parameters were stable through all passages following passage 3.
## Metabolic network construction of g. xylinus (cgmcc no. 2955)
In our recent work [bib_ref] Metabolic flux analysis of Gluconacetobacter xylinus for bacterial cellulose production, Zhong [/bib_ref] , we have constructed the reaction network of central carbon metabolism in G. xylinus (CGMCC No.2955). It has to be stated that the inability to metabolize glucose under anaerobic condition in G. xylinus results from the lack of phosphofructokinase that is required for glycolysis [bib_ref] The unified metabolism of Gluconacetobacter entanii in continuous and batch processes, Velasco-Bedrán [/bib_ref] [bib_ref] Fermentation and metabolic characteristics of Gluconacetobacter oboediens for different carbon sources, Sarkar [/bib_ref]. Hence, flux from F6P to T3P does not exist. Gluconeogenesis occurs in G. xylinus CGMCC 2955 from oxalacetate via pyruvate (r 18 ) due to the unusual regulation of the oxaloacetate decarboxylase and pyruvate phosphate dikinase [bib_ref] Purification and regulatory properties of the oxaloacetate decarboxylase of Acetobacter xylinum, Benziman [/bib_ref] [bib_ref] The PEP-pyruvate-oxaloacetate node as the switch point for carbon flux distribution in..., Sauer [/bib_ref]. As shown in [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref] , there were fluxes from phosphoenolpyruvate to oxaloacetate (r 25 ) and oxaloacetate to pyruvate (r 26 ) for glycolysis and gluconeogenesis, respectively. Thus, cellulose is produced by G. xylinus CGMCC 2955 from a metabolic pool of hexose phosphate that is converted directly by the phosphorylation from exogenous hexoses, as well as indirectly via the pentose cycle and the gluconeogenic pathway. The determined fluxes (r 1 , r 2 , r 3 , r 4 ) from glucose to cellulose are glucose R glucose 6-phosphate R glucose 1-phosphate R UDP-glucose R cellulose. Other studies have shown that lipid-and protein-linked cellodextrins may function as intermediates between UDP-glucose and cellulose in G. xylinus [bib_ref] Intermediatry steps in Acetobacter xylinum cellulose synthesis: studies with whole cells and..., Swissa [/bib_ref].
It was assumed that NADPH was produced only to fulfill biosynthetic requirements, and the NADH flux was assumed to be proportional to the oxygen uptake. The biosynthetic requirements for ribose-5-phosphate, 3-phosphoglycerate, and NAD(P)H were specified as follows. Although the NAD-and NADP-linked dehydrogenases are known to catalyze the same reaction in G. xylinus, the NAD-related glucose 6-phosphate dehydrogenase of G. xylinus CGMCC 2955 is involved mainly when the pentosephosphate pathway is directed toward oxidation and energy generation, while the NADP-related enzyme functions in an anabolic capacity. Furthermore, the NAD-specific enzyme rather than NADP-dependent dehydrogenase is sensitive to inhibition by ATP. Therefore, the reactions from glucose 6-phosphate dehydrogenase to ribulose 5-phosphate (r 6 ) involving NAD or NADP, depend on the different growth conditions. It has also been reported that the malate dehydrogenase of G. xylinus is an FADenzyme containing an iron-binding site that is essential for its activity [bib_ref] Flavine adenine dinucleotide-linked malic dehydrogenase from Acetobacter xylinum, Benziman [/bib_ref] , hence we specify that this flux (r 24 ) in our strain is involving FAD.
In G. xylinus, since phosphofructokinase, an essential enzyme in the glycolytic pathway, could not be synthesized, the fructose-6-P phosphoketolase pathway may represent a very desirable meta-bolic feature. Meanwhile, conversion from fructose-6-P to acetate (r 13 ) produces 3 moles of ATP per mole of fructose-6-P, or 2 moles of ATP per mole of glucose, the yields of which are identical with that of glycolysis. Therefore, a unique pathway (r 13 ) exists from fructose-6-P to acetate in G. xylinus CGMCC 2955.
## Enzymatic activities
In G. xylinus, PEPC catalyzes the addition of bicarbonate to phosphoenolpyruvate (PEP) to form the four-carbon compound oxaloacetate (OAA) and inorganic phosphate, corresponding to the bioreaction of r 25 in [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref] [bib_ref] The effect of NADP-dependent malic enzyme expression and anaerobic C4 metabolism in..., Kwon [/bib_ref]. As shown in , the enzymatic activity of PEPC in mutant strain AX2-16 was 1.97-fold of that in parent strain on the 1 st day. As culture time extended, PEPC in G. xylinus AX2-16 showed lower activity than that in parent strain. For example, on the 4 th -8 th day, its activity in mutant strain was only 71.60%, 22.91%, 32.20%, 27.65% and 12.19% of that in parent strain, respectively. This result was consistent with the value of r 25 from flux analysis. As shown in [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref] , for r 25 , the flux in G. xylinus AX2-16 was 0.73, while it was 10.54 in the parent strain G. xylinus CGMCC 2955.
SDH is one of the key enzymes in TCA cycle, with its activity tightly correlated with the cellular energy metabolism. It catalyzes the oxidation of succinate to fumarate with the reduction of ubiquinone to ubiquinol [bib_ref] Structure and function of succinate dehydrogenase and fumarate reductase, Ackrell [/bib_ref]. The changes of SDH activity in both strains showed similar trends as a function of culture time. The activity of SDH in G. xylinus AX2-16 was shown at high level since the 6 th day. It has to be noted that the SDH activity in G. xylinus AX2-16 was obtained at 8.50 U/OD 600 on the 7 th day, compared with 5.35 U/OD 600 in G. xylinus CGMCC 2955 at the same time. At the end of culture (8 d), the SDH activity decreased in both the two strains, but the SDH activity in mutant strain still showed 1.65-fold of that in parent strain.
The ATPase activity in G. xylinus AX2-16 exhibited a much higher value than that in G. xylinus CGMCC 2955 since the 2 nd day. Higher ATPase activity in G. xylinus AX2-16 was obtained at 152.08 mmol/OD 600 on the 5 th day, which was almost 3.56-fold of that in parent strain. As culture proceeded, the ATPase activities in both strains decreased. However, at the end of culture, the enzymatic activity of ATPase in G. xylinus AX2-16 was still 6.41fold of that in G. xylinus CGMCC 2955.
# Discussion
## Comparison of cultivation process between g. xylinus (cgmcc no. 2955) and ax2-16 in batch culture
Batch cultures of the parent strain G. xylinus CGMCC 2955 were conducted to determine the growth parameters using glucose as the sole carbon source. BC, gluconic acid, and biomass production in G. xylinus CGMCC 2955 were measured. shows that for the parent strain, the maximum production of BC (7.26 g/L) and dry biomass accumulation (1.72 g/L) were observed on the 7 th day. As shown in [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref] , the biomass concentration of G. xylinus AX2-16 was always higher than that of G. xylinus CGMCC 2955 during the cultivation process. Particularly, at the exponential phase from 3 rd to 5 th day, the biomass accumulation of G. xylinus AX2-16 was about 1.5 times higher than that of G. xylinus CGMCC 2955 . It was interesting to find that the gluconic acid concentration of AX2-16 was only half of the parent strain's . Final BC production from G. xylinus AX2-16 was also 62% greater than G. xylinus CGMCC 2955 . Correspondingly, the BC productivity by G. xylinus AX2-16 was obtained at 2.75 mmol/g?h, which was 1.34 fold of that in G. xylinus CGMCC 2955 .
## Comparison of metabolic flux distributions in the mutant and parent strain
Chemical mutagenesis is a random method to alter microorganisms. In other words, genetic or metabolic changes of microorganisms in this process were unpredictable. However, by estimating metabolic fluxes based on stoichiometric models, it provides a valuable and convenient tool to reveal the metabolic distributions in the parent and mutant strains. A stoichiometric mass balance analysis was used to give a quantitative description of the flux distribution within the defined bioreaction network, as shown in [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref]. In this network, the quantity of glucose that entered the cell was normalized to 100. For the parent strain, almost 59% and 9.5% of carbon was directed to gluconic acid and acetic acid, respectively, which were regarded as byproducts in the generation of BC, and a waste of the carbon source. Only 24% of the carbon source converted to the final product, BC. Therefore, the key objective of the induced mutation was to increase the efficiency of converting glucose to BC.
It was shown that 1.7% and 2.4% of the source carbon (derived from glucose) was used to generate cell biomass for G. xylinus CGMCC 2955 and G. xylinus AX2-16, respectively. The flux of carbon source to the desired end-product, BC in parent strain increased to 40%, compared to 24.2% of the parent strain. In addition, the carbon flux to the byproduct, gluconic acid, decreased for G. xylinus AX2-16 at only 32.7%, compared with 58.5% of that in the parent strain G. xylinus CGMCC 2955. The fraction of source carbon that ended in the secondary byproduct, acetic acid, was fairly similar for parent and mutant, with 9.5% and 4.0%, respectively. Within the metabolic pathway of G. xylinus CGMCC 2955, 33.0% of the carbon source (glucose) entered into the pentose phosphate pathway (PPP) (r 6 ) and approximately 17% entered into TCA cycle (r 22 , r 23 and r 24 ) [fig_ref] Figure 3: Metabolic network of mutant and parent strains [/fig_ref]. However, for strain G. xylinus AX2-16, a dramatically higher TCA cycle flux (57.0%) was obtained. Bass et al. reported that the enzyme activities within each metabolic pathway are closely related [bib_ref] Metabolic differentiation of distinct muscle types at the level of enzymatic organization, Bass [/bib_ref]. As shown in , the enzymatic activity of SDH from G. xylinus AX2-16 was 1.90-fold, 1.59-fold and 1.65-fold higher than that from parent strain on the 6 th , 7 th and 8 th day, respectively. This confirmed the flux analysis result that higher TCA cycle activity was obtained in mutant strain than that in parent strain. Furthermore, the PPP was the primary metabolic pathway through which glucose was able to enter TCA cycle, which also requires additional ATP. TCA cycle is a series of chemical reactions to generate energy (in the form of ATP) through the oxidation of acetate. The relatively high level of TCA cycle flux may also be explained by higher ATP requirements for biosynthesis of cellulose. As illustrated by the ATPase activity in G. xylinus AX2-16 was much higher than that in the parent strain after the 2 nd day. It has to be noted that the enzymatic activity of ATPase in G. xylinus AX2-16 was still 6.41fold of that in G. xylinus CGMCC 2955 at the end of culture, indicating that more ATP was demanded in mutant strain AX2-16 compared with parent strain. For strain G. xylinus CGMCC 2955, 15.8% glucose returned to glucose-6P, while the reversible reactions between glucose-6P and fructose-6P remained balanced in the G. xylinus AX2-16, thereby, reducing the generation of gluconic acid. Moreover, for strain G. xylinus AX2-16, the majority of the flux entered into F6P and then into Ac-P and TCA cycle in order to produce energy. A number of fluxes for G. xylinus AX2-16 were equal to zero (r 7 , r 17 and r 19 ). For r 7 , this flux may be zero if strain AX2-16 does not have the G6P recycle to produce F6P. For r 17 , the parent strain probably requires more carbon source for gluconeogenesis (PYR to PEP, the major bioreaction of gluconeogenesis) and from there to gluconic acid and BC. For r 19 , some Ac-P was spontaneously hydrolyzed to phosphorus and acetate. Finally, for r 13 , the flux in strain AX2-16 was about 3.5 times of that in G. xylinus CGMCC 2955, indicating that there was more flux into the Ac-P, and then into the TCA cycle to produce more ATP for biosynthesis in the mutant strain.
Generally, NADPH formation is strictly coupled to the biosynthetic NADPH requirements, which causes the PPP to operate only to satisfy the cellular requirements for NADPH and pentose precursors. However, in G. xylinus, the operation of the pentose phosphate pathway is not exclusively governed by the demand for NADPH and pentose precursors. In phosphoglucoisomerase-negative mutants of E. coli, glucose cannot be metabolized via glycolysis, but only through the PPP [bib_ref] Pathways of NADPH in Escherichia coli, Csonka [/bib_ref]. Instead, for these mutants, G6P is metabolized by pentose-phosphate (PP) and phospho-ketolase (PK) pathways. These pathways are linked to the intracellular acyl-phosphate pool [bib_ref] The acetate switch, Wolfe [/bib_ref]. In addition to anabolic utilization of NADPH, G. xylinus could also reoxidize to NADH. PPP flux is regarded to be in the range of 20% to 30% of the total glucose uptake in bacteria, which exceeds the requirements for NADPH and pentose formation under most conditions. These bacteria also could use NADH produced in the TCA cycle via G6P dehydrogenation, and NADPH produced in the PP pathway [bib_ref] Characterization of enzymes involved in the central metabolism of Gluconobacter oxydans, Rauch [/bib_ref]. Other bacteria also have been shown to produce NADPH in excess of their biosynthetic requirements for NADPH. Using labeling studies combined with metabolite balancing, the actual NADPH formation in lysine-producing Corynebacterium glutamicum was found to exceed the biosynthetic requirements by 21.1% [bib_ref] Determination of the fluxes in the central metabolism of Corynebacterium glutamicum by..., Marx [/bib_ref]. Similar to the mutant E. coli above, G. xylinus CGMCC 2955 lacked phosphofructose kinase required for glycolysis, and the glucose was only metabolized by PPP flux by 27-33% of the glucose. Hence, as the pentose phosphate pathway is necessary for conversion of G6P to F6P in G. xylinus CGMCC 2955, constraining the flux through this pathway to the amount required solely for NADPH requirements, may lead to a biased solution, in which all of the substrate carbon not in biomass and products will be converted to CO 2 in the TCA cycle. In order to obtain an unbiased estimation of the flux pattern, it was assumed that interchangeability of reducing equivalents, NADH and NADPH, could be explained biologically by the presence of a transhydrogenase. Based on this assumption, the estimated oxidative PPP flux was adjusted and reached relative fluxes higher than unity at the expense of the TCA cycle. If no further assumptions regarding the biological function of the PPP are made under these conditions, the estimated PPP flux is very sensitive to the respiratory quotient. The best estimate was obtained by allowing a realistic 20% exchange of reducing equivalents from NADPH to NADH in the optimized calculations, based on the experimental results of Marx et al. [bib_ref] Determination of the fluxes in the central metabolism of Corynebacterium glutamicum by..., Marx [/bib_ref]. The flux estimates obtained with this assumption for the PPP and TCA cycles indicated the robustness of the solution for a transfer of reducing equivalents from NADPH to NADH in the physiological range. Therefore, the NADPH and NADH are considered as equivalents.
In another words, all of the NADPH overproduced by the PPP will be converted into NADH and completely oxidized in the TCA cycle.
Using chemical mutation by DES and LiCl, a mutant strain, G. xylinus AX2-16, was obtained with highest BC productivity of 11.75 g/L. However, gluconic acid, the main byproduct, was only produced at 5.71 g/L by mutant strain, which was 55.7% lower than that of parent strain. Metabolic flux analysis demonstrated that 40.1% of the source carbon was diverted into the desired product BC in mutant strain, compared with 24.2% for parent strain. Additionally, only 32.7% of carbon source was fluxed into gluconic acid in mutant strain, compared with 58.5% of that for parent strain. A higher flux of TCA cycle was obtained in mutant strain (57.0%) compared with parent strain (17.0%), which matched well with the results from enzymatic analysis. It indicated that the increased TCA cycle flux and ATP content in mutant strain would be attributed to the acceleration of BC biosynthesis.
[fig] Figure 2: Transgenerational variation in the production of bacterial cellulose and the pH value for mutant strain AX2-16. doi:10.1371/journal.pone.0098772.g002 [/fig]
[fig] Figure 3: Metabolic network of mutant and parent strains (Upper data-G. xylinus AX2-16, lower data-G. xylinus CGMCC 2955 [/fig]
[fig] Figure 4, Figure 5: The changes of enzymatic activities of PEPC, SDH and ATPase in mutant and parent strains. The activities of PEPC, SDH and ATPase in mutant and parent strains were detected as described in Materials and methods. Cells were cultured statically at 30uC for 8 days. doi:10.1371/journal.pone.0098772.g004 Changes in biomass, gluconic acid, and bacterial cellulose production during batch culture using glucose as the sole carbon source. (a) biomass, (b) bacterial cellulose, (c) gluconic acid. doi:10.1371/journal.pone.0098772.g005 [/fig]
|
Stability of the Mitigating Effect of Students’ Perceived Teacher Enthusiasm on Class-related Boredom: Moderating Role of Boredom Proneness and Perceived Task Difficulty
The aim of the current study was to explore the stability of the mitigating effect of students' perceived teacher enthusiasm on class-related boredom and the moderating role of boredom proneness and perceived task difficulty in such effect. A total of 984 students from five universities in China participated in the study. Questionnaires on class-related boredom, perceived teacher enthusiasm, boredom proneness, and perceived task difficulty were used to measure the respective variables. Results showed that boredom proneness and perceived task difficulty significantly moderated the relationship between perceived teacher enthusiasm and class-related boredom. Moreover, when considering perceived task difficulty, boredom proneness became silent in the moderating path between perceived teacher enthusiasm and class-related boredom. Even so, the mitigating effect of students' perceived teacher enthusiasm on class-related boredom was stable in students with different levels of boredom proneness and perceived task difficulty. The implications for learning and teaching are discussed. stable mitigating effect of perceived teacher enthusiasm on students' class-related boredom among students with different levels of boredom proneness and perceived task difficulty[4,5]. Exploring the relationships among these variables could provide an important research basis for future interventional studies. Therefore, it is necessary to further investigate the stability of the effect of students' perceived teacher enthusiasm on their class-related boredom and the moderating role of boredom proneness and perceived task difficulty in this effect.Class-Related BoredomBoredom is one of the most commonly emotions which can be experienced in many settings[3]. Class-related boredom refers to a negative and low physiological arousal emotion experienced by students in the process of learning in the classroom[3,8]. Different from the general academic boredom, class-related boredom is the experienced boredom in the classroom setting. Although general boredom is perceived as a high arousal state by some researchers [9], the control-value theory of achievement emotions and more research in recent decades suggested class-related boredom to be a low arousal state[3]. Class-related boredom is a negative emotion which involves dissatisfaction and low arousal, and causes irrelevant thoughts (i.e., daydreaming), temporal extension, and motivation to leave the boring situation[10][11][12]. The behavioral expression of this negative emotion mainly consists of sleepiness, yawning, flabby body posture, cold hands, and vacant eyes[3].Class-related boredom is widely experienced by undergraduate students in the classroom. Mann and Robinson found that 59% of undergraduates reported experiencing boredom during half of the time during their lectures, and 30% reported that most or all of the time in their lectures was boring[13]. Pekrun et al. found that in the process of class-related learning, 42.2% undergraduates experienced boredom, which is higher than the rate of undergraduates experiencing emotions such as anxiety (28.0%), anger (19.3%), and hopelessness (13.6%)[3]. Tze and colleagues found that Chinese university students experienced more boredom in their classrooms than did their Canadian counterparts[14].The influence of class-related boredom on students' learning manifests itself, mainly, in the following aspects. Firstly, it was found to have effects on students' class-related emotions, achievement goals, task value, self-efficacy in learning, learning engagement, use of learning strategies, and behavior problems[15][16][17]. Secondly, research found that students' class-related boredom affected their achievement and performance[14,18,19]. Lastly, students' class-related boredom was confirmed to have an influence on learning burnout [20], career aspirations[21], and lifelong learning[22]. Therefore, the effects of students' class-related boredom on their learning and vocational development should not be underestimated; thus, investigating the antecedents of class-related boredom and identifying an efficient method to mitigate it is urgently required.In the classroom environment, beside students' individual antecedents, characteristics of teachers and teaching are important environmental antecedents of students' class-related emotions[23]. In the view of control-value theory of achievement emotions, Pekrun suggested that important classroom environmental variables included instruction, value induction, autonomy support, goal structures/expectations, and achievement[24]. Goetz and Hall classified the antecedents of boredom into three categories: Individual (e.g., boredom proneness), environmental (e.g., monotony, isolation, or repetitive task), and related to the fit of individuals and environment (e.g., too hard or too easy task)[11]. Although some researchers suggested the importance of latent cognitive antecedents such as attention and hyperactivity disorder and impulsivity[25], current research has mainly focused on antecedents related to individual and environmental aspects.
# Introduction
Classrooms are not neutral spaces, but full of emotions; such class-related emotions are at the core of teaching and learning. Class-related boredom is a negative emotion widely experienced by students, which may affect their learning process and outputs as well as health. The latest research has shown that students' perceived teacher enthusiasm negatively predicted their class-related boredom significantly, which suggests that increasing teacher enthusiasm may be an efficient way to mitigate students' class-related boredom. In the theories of boredom antecedents and related empirical research, individuals' boredom proneness and perceived task difficulty were considered to be important factors influencing boredom in specific situations. Cui et al. confirmed the predicting effect of students' perceived teacher enthusiasm on class-related boredom after controlling for the effects of boredom proneness and perceived task difficulty; however, neither of the studies confirmed a proposing the definition of dispositional teacher enthusiasm, which included positive affect and positive emotional expressivity, and verifying the construct.
Teacher enthusiasm reflects a positive emotion and has been found to improve a range of students' learning outcomes, such as school achievement, recall performance, and learning motivation. However, only a few empirical studies have explored the relationship between teacher enthusiasm and students' class-related boredom. For instance, in the Latin class setting, teacher enthusiasm was found to be one of the teaching variables highly related to students' boredom. Goetz and colleagues suggested teacher enthusiasm as one of teachers' core teaching characteristics, and found that supportive presentation style, which included teacher enthusiasm, could significantly predict students' class-related boredom. Moreover, a group of researchers recently reported the direct and indirect relationships between teacher enthusiasm and students' class-related boredom. In a similar study, an outcome variable model of teacher enthusiasm was constructed, and an indirect predictive role of teacher enthusiasm on class-related boredom was supported. In addition, students' interest was considered a major antecedent of their boredom in theoretical research, and empirical research also found that interest could significantly predict class-related boredom.
Furthermore, some researchers suggested teacher enthusiasm as a core index of students' perceived classroom environmentsand instruction quality. In fact, teacher enthusiasm has been known as a core index of instruction quality and teaching effectiveness for a long time, due to earlier educational studies. In recent years, with the increasing research on students' academic emotions as well as teacher emotions, some researchers have integrated related concepts of teacher enthusiasm and examined their effect on teachers' teaching and students' learning.
Although teacher enthusiasm may play an important role in mitigating students' class-related boredom, few studies have investigated this relationship. In terms of a theoretical basis, theories explaining the mitigating role of teacher enthusiasm on class-related boredom have been limited. Previous theories on the antecedents of boredom focused on boredom in the work environment, and the repetition of tasks and monotony of the environment were recognized as major antecedents of boredom. Few theories have explained the antecedents of boredom in the classroom environment, especially focusing on the roles played by teachers. Pekrun suggested that teacher enthusiasm may affect students' achievement emotions through the mechanism of emotional contagion and observational learning in the control-value theory of achievement emotions; however, no definite relationship between teacher enthusiasm and students' class-related boredom is explained by this theory.
To the best of our knowledge, few studies have explored the relationship between teacher enthusiasm and class-related boredom in the domain of teacher emotions and academic emotions of students. Based on the control-value theory of achievement emotions and related results of empirical studies, we inferred that teacher enthusiasm may predict students' class-related boredom.
## Boredom proneness
Although some researchers have defined boredom proneness, there is no widely accepted definition to date. Zuckerman, Eysenck, and Eysenck provided a definition for a similar concept, that is, boredom susceptibility as "an aversion to repetition, routine, and dull people, and restlessness when things are unchanging". Famer and Sundberg defined boredom proneness as "one's connectedness with one's environment on many situational dimensions, as well as the ability to access adaptive resources and realize competencies". Boredom proneness was significantly related to depression, anxiety, stress, life satisfaction, and autonomy orientation.
In the classroom setting, students' boredom proneness may be one of the major predictors of their class-related boredom. Empirical and theoretical evidence exists to support such association. For instance, students' boredom proneness significantly predicted their scores on cognitive failure. Based on the control-value theory of achievement emotions, students' cognitive failure may affect their evaluation of control and value, and then induce class-related boredom. Therefore, it could be concluded that students' boredom proneness is associated with class-related boredom. Furthermore, boredom proneness was found to be predictive of the level of flow and mood monitoring, which may lead to class-related boredom. Students with high boredom proneness were more likely to have internet addiction, but had a low tendency to engage in online learning. More direct evidence was provided by a study on boredom in lecture, that students with a high level of boredom proneness reported higher levels of class-related boredom, displayed corresponding behaviors, used more boredom coping strategies, and were more likely to miss a lecture. To sum up, boredom proneness is highly likely to be a core predictor of class-related boredom.
## Perceived task difficulty
Perceived task difficulty has been found to affect academic boredom. Specifically, Daschmann and colleagues found that students being over-or under-challenged in class-related learning were major predictors of their boredom in learning, and both variables have been confirmed as important precursors of boredom. Tanaka and Murayama found that students' perception of difficulty was associated with their boredom. Asseburg and Frey found that the level of ability-difficulty fit of ninth-graders was significantly related to their boredom in a test, with students with a higher level of ability-difficulty fit having lower levels of boredom. An explanation as to why high task difficulty and overload are associated with students' boredom may be that high-difficulty tasks lead to excessive cognitive load and lack of fluency, which may decline students' perceived value of the tasks and increase their experience of boredom in learning. In contrast, a study found a decline in boredom with increasing difficulty, which seems to be inconsistent with the literature. To be noted, this study was conducted in the condition of easy tasks and students had a low level of anxiety. In such condition, an increase in difficulty indicated an increase in perceived challenge, and hence students perceive more control and value.
The framework of the control-value theory of achievement emotions suggests that students' perceived control and value are proximate causes of achievement emotions, and that environmental variables affect emotions through the mediating role of students' perceived control and value. In accord with this, Fisher's theory identified qualitative underload and overload as major causes of boredom. According to the aforementioned studies and theories, we inferred that in learning tasks with medium difficulty, students' increased perceived task difficulty will lead to a decline in their perceived control and value; therefore, negative emotions including boredom will be generated.
To sum up, students' perceived task difficulty may be a major antecedent of boredom in various specific settings (e.g., boredom in learning and boredom in a test). However, there has been scarce research exploring the effect of task difficulty on students' class-related boredom, and whether task difficulty can moderate the effects of antecedent variables on students' class-related boredom.
## Research questions and hypothesis
In previous theoretical and empirical studies on antecedents of boredom, boredom proneness and perceived task difficulty were conceived as important antecedent variables of specific and situational boredom. For example, in the control-value theory of achievement emotions and the study by Mann and Robinson, students' boredom proneness and perceived task difficulty were important antecedent variables of class-related boredom. Cui et al. found that after controlling for the effects of boredom proneness and perceived task difficulty, students' perceived teacher enthusiasm predicted their class-related boredom negatively and significantly. Based on the previous research by Cui and colleagues, the current study aimed to further explore whether boredom proneness and perceived task difficulty may moderate the relationship between perceived teacher enthusiasm and class-related boredom, and whether students' perceived teacher enthusiasm can predict their class-related boredom significantly and stably among students with different levels of boredom proneness and perceived task difficulty. The present study aimed to explore the following two questions: Question 1: Can students' boredom proneness moderate the relationship between their perceived teacher enthusiasm and class-related boredom, and can perceived teacher enthusiasm predict class-related boredom stably and significantly among students with different levels of boredom proneness? Question 2: Can students' perceived task difficulty moderate the relationship between their perceived teacher enthusiasm and class-related boredom, and can perceived teacher enthusiasm predict class-related boredom stably and significantly among students with different levels of perceived task difficulty?
Research population in the present study was college students. Although a number of research studies on teacher enthusiasm and boredom involved primary, middle, or high school settings, the present study with college students could investigate these variables and their associations in the setting of higher education. Results might provide significant practical implications to the teaching and learning of college students. Corresponding to the research questions, two research hypotheses were proposed. Hypothesis 1: Boredom proneness can moderate the relationship between perceived teacher enthusiasm and class-related boredom significantly, and students' perceived teacher enthusiasm can predict their class-related boredom stably, negatively, and significantly among students with different levels of boredom proneness. Hypothesis 2: Students' perceived task difficulty can moderate the relationship between perceived teacher enthusiasm and class-related boredom significantly, and students' perceived teacher enthusiasm can predict their class-related boredom stably, negatively, and significantly among students with different levels of perceived task difficulty.shows the relationship indicated in the hypotheses. Question 1: Can students' boredom proneness moderate the relationship between their perceived teacher enthusiasm and class-related boredom, and can perceived teacher enthusiasm predict class-related boredom stably and significantly among students with different levels of boredom proneness?
Question 2: Can students' perceived task difficulty moderate the relationship between their perceived teacher enthusiasm and class-related boredom, and can perceived teacher enthusiasm predict class-related boredom stably and significantly among students with different levels of perceived task difficulty?
Research population in the present study was college students. Although a number of research studies on teacher enthusiasm and boredom involved primary, middle, or high school settings, the present study with college students could investigate these variables and their associations in the setting of higher education. Results might provide significant practical implications to the teaching and learning of college students. Corresponding to the research questions, two research hypotheses were proposed. Hypothesis 1: Boredom proneness can moderate the relationship between perceived teacher enthusiasm and class-related boredom significantly, and students' perceived teacher enthusiasm can predict their class-related boredom stably, negatively, and significantly among students with different levels of boredom proneness. Hypothesis 2: Students' perceived task difficulty can moderate the relationship between perceived teacher enthusiasm and class-related boredom significantly, and students' perceived teacher enthusiasm can predict their class-related boredom stably, negatively, and significantly among students with different levels of perceived task difficulty.shows the relationship indicated in the hypotheses.
# Materials and methods
## Participants and procedures
To test the research hypotheses, 984 students (73.0% female) at five colleges in Henan and Shanxi provinces aging from 17 to 24 years (mean = 20.08, standard deviation [SD] = 1.24) participated in this research. These participants were recruited from 21 class subjects across 19 majors in 15 departments, and they were in the first, second, or third school year (n = 567, 310, and 107, respectively). The present study was part of a large longitudinal survey; the questionnaire on boredom proneness was administered at two weeks after the beginning of a new term, and the remaining questionnaires at four weeks after the previous questionnaire. Each student in the study completed the survey only once, in relation to a single class. Teachers and students provided informed consent before the survey, which was conducted at the end of the class, and the questionnaires were collected immediately.
## Measures
# Class-related boredom
Eleven-item class-related boredom subscale in the Achievement Emotions Questionnaire (AEQ)was used to assess students' class-related boredom in this study. Two example questions were
# Materials and methods
## Participants and procedures
To test the research hypotheses, 984 students (73.0% female) at five colleges in Henan and Shanxi provinces aging from 17 to 24 years (mean = 20.08, standard deviation [SD] = 1.24) participated in this research. These participants were recruited from 21 class subjects across 19 majors in 15 departments, and they were in the first, second, or third school year (n = 567, 310, and 107, respectively). The present study was part of a large longitudinal survey; the questionnaire on boredom proneness was administered at two weeks after the beginning of a new term, and the remaining questionnaires at four weeks after the previous questionnaire. Each student in the study completed the survey only once, in relation to a single class. Teachers and students provided informed consent before the survey, which was conducted at the end of the class, and the questionnaires were collected immediately.
## Measures
# Class-related boredom
Eleven-item class-related boredom subscale in the Achievement Emotions Questionnaire (AEQ)was used to assess students' class-related boredom in this study. Two example questions were "The lecture bores me" and "I think about what else I might be doing rather than sitting in this boring class". Responses were indicated on a five-point Likert scale ranging from one (strongly disagree) to five (strongly agree). A higher aggregated score indicated a higher level of class-related boredom. Cronbach's alpha was 0.94 in the current study.
## Perceived teacher enthusiasm
Three items (i.e., "Our teacher in this class teaches with enthusiasm," "Our teacher in this subject enjoys teaching compared to other courses", and "Our teacher in this class tries to inspire students about the subject") from the study by Keller et al., were used to assess perceived teacher enthusiasm. Responses were indicated on a five-point Likert scale ranging from one (not at all true of me) to five (very true of me). A higher aggregate score indicated a higher level of perceived teacher enthusiasm. Cronbach's alpha was 0.87 for this construct in this study.
## Boredom proneness
The 12-item Boredom Proneness Scale Short Form (BPS-SF) was adapted to measure boredom proneness. Consistent with previous research, two items were deleted to adapt to the Chinese culture (i.e., "I find it easy to entertain myself" and "It seems that the same old things are on television or the movies all the time; it's getting old") Ten items were maintained. An example question: "In any situation I can usually find something to do or see to keep me interested". Responses were indicated on a seven-point Likert scale ranging from one (strongly disagree) to seven (strongly agree). A higher aggregate score indicated a higher level of boredom proneness. Cronbach's alpha was 0.69 for this scale in the present study.
## Perceived task difficulty
Two questions were included in the study to assess students' perceived task difficulty, which were commonly used in previous research. The two questions are "Today's class was hard for me" and "Compared to other courses, today's class was hard for me". Responses to both questions are indicated on a five-point Likert scale ranging from one (not at all true of me) to five (very true of me). A higher aggregate score indicates a higher level of perceived task difficulty. Cronbach's alpha was 0.84 for this measure in the present study.
## Statistical analyses
All analyses were conducted using PASW statistics for Windows (Version 18, IBM Corp., Armonk, NY, USA). Firstly, we examined descriptive statistics (mean and SD) and intercorrelations of the variables. Subsequently, we tested the moderating effects of boredom proneness and perceived task difficulty on the relationship between perceived teacher enthusiasm and class-related boredom using a hierarchical regression analysis.
The data of perceived teacher enthusiasm, boredom proneness, and perceived task difficulty had been standardized. Tests on data in the present study showed that the assumptions of hierarchical regressions (e.g., normality, homoscedasticity, independence of errors of prediction, and linearity.) were met. Multicollinearity was tested by calculating tolerance for each independent variable and values of tolerance indicated that multicollinearity did not occur among the data.
# Results
## Means, sds, and intercorrelations of all measures
Means, standard deviations, and variable intercorrelations are presented in. The significant intercorrelations provided a foundation for the further analysis of the moderating effects of boredom proneness and perceived task difficulty.
## Moderating effects of boredom proneness
To test Hypothesis 1, a hierarchical regression analysis was used to test the moderating role of boredom proneness on the relationship between perceived teacher enthusiasm and class-related boredom. After controlling for the effect of demographic variables on class-related boredom, the predictive effects of perceived teacher enthusiasm, boredom proneness, and the interaction between perceived teacher enthusiasm and boredom proneness on class-related boredom were tested individually. The results are presented in. As shown in, in the first step, major, subject, and age (among demographic variables) were significant predictors of class-related boredom (p < 0.01). In the second step, after controlling for the effects of demographic variables on class-related boredom, perceived teacher enthusiasm (B = −0.204, p < 0.001) and boredom proneness (B = 0.124, p < 0.001) were significant predictors of class-related boredom. In the third step, after controlling for the effects of demographic variables on class-related boredom, perceived teacher enthusiasm (B = −0.211, p < 0.001), boredom proneness (B = 0.118, p < 0.001), and the interaction between perceived teacher enthusiasm and boredom proneness (B = 0.058, p < 0.01) were significant predictors of class-related boredom. Thus, Hypothesis 1 was confirmed, as the results manifested that boredom proneness moderated the relationship between perceived teacher enthusiasm and class-related boredom. To be noted, perceived teacher enthusiasm had a significantly negative predicting effect on students' class-related boredom. When perceived teacher enthusiasm increased, students' class-related boredom could decrease. Furthermore, simple slope tests were used to test differences in the predictive effect of perceived teacher enthusiasm on class-related boredom between college students with high or low levels of boredom proneness (see. class-related boredom between college students with high or low levels of boredom proneness (see. The results of simple slope tests are shown in. High or low boredom proneness refers to the level of boredom proneness above or below one SD of the mean, respectively. Similarly, high or low teacher enthusiasm refers to the level of teacher enthusiasm above or below one SD of the mean, respectively. For the students with high levels of boredom proneness, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively weak (simple slope = −0.153, t = −5.141, p < 0.001). In comparison, for the students with low levels of boredom proneness, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively strong (simple slope = −0.269, t = −8.061, p < 0.001). In summary, although boredom proneness significantly moderated the relationship between perceived teacher enthusiasm and class-related boredom, the mitigating effects of perceived teacher enthusiasm on class-related boredom were stable and significant.
## Moderating effects of perceived task difficulty
To test Hypothesis 2, a hierarchical regression analysis was used to test the moderating role of perceived task difficulty between perceived teacher enthusiasm and class-related boredom. After controlling for the effects of demographic variables and boredom proneness on class-related boredom, the predictive effect of perceived teacher enthusiasm, perceived task difficulty, and interaction between perceived teacher enthusiasm and perceived task difficulty on class-related boredom were tested individually. The results are presented in.. Moderating effect of perceived task difficulty between perceived teacher enthusiasm and class-related boredom (N = 984). The results of simple slope tests are shown in. High or low boredom proneness refers to the level of boredom proneness above or below one SD of the mean, respectively. Similarly, high or low teacher enthusiasm refers to the level of teacher enthusiasm above or below one SD of the mean, respectively. For the students with high levels of boredom proneness, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively weak (simple slope = −0.153, t = −5.141, p < 0.001). In comparison, for the students with low levels of boredom proneness, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively strong (simple slope = −0.269, t = −8.061, p < 0.001). In summary, although boredom proneness significantly moderated the relationship between perceived teacher enthusiasm and class-related boredom, the mitigating effects of perceived teacher enthusiasm on class-related boredom were stable and significant.
## First step
## Moderating effects of perceived task difficulty
To test Hypothesis 2, a hierarchical regression analysis was used to test the moderating role of perceived task difficulty between perceived teacher enthusiasm and class-related boredom. After controlling for the effects of demographic variables and boredom proneness on class-related boredom, the predictive effect of perceived teacher enthusiasm, perceived task difficulty, and interaction between perceived teacher enthusiasm and perceived task difficulty on class-related boredom were tested individually. The results are presented in.
The results of simple slope tests are shown in. Consistent with the previous simple slope tests, perceived task difficulty and teacher enthusiasm were divided into high, medium, and low levels based on the scores above or below one SD of the mean. In the condition of high task difficulty, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively strong (simple slope = −0.269, t = −8.574, p < 0.001). In comparison, the mitigating effect was relatively weak in the condition of low task difficulty (simple slope = −0.148, t = −5.068, p < 0.001). In summary, although perceived task difficulty significantly moderated the relationship between perceived teacher enthusiasm and class-related boredom, the mitigating effects of perceived teacher enthusiasm on class-related boredom were stable and significant. levels based on the scores above or below one SD of the mean. In the condition of high task difficulty, the mitigating effect of perceived teacher enthusiasm on class-related boredom was relatively strong (simple slope = −0.269, t = −8.574, p < 0.001). In comparison, the mitigating effect was relatively weak in the condition of low task difficulty (simple slope = −0.148, t = −5.068, p < 0.001). In summary, although perceived task difficulty significantly moderated the relationship between perceived teacher enthusiasm and class-related boredom, the mitigating effects of perceived teacher enthusiasm on class-related boredom were stable and significant.
## Comparison of the moderating effects of boredom proneness and perceived task difficulty
To further test the moderating effects, an additional hierarchical regression was conducted, which considered both boredom proneness and perceived task difficulty. The results are presented in. After controlling for the effect of demographic variables in Step one and the predictive effects of perceived teacher enthusiasm on class-related boredom in Step two, the moderating effects of boredom proneness and perceived task difficulty were tested in Step three and four, respectively. The significant moderating effect of boredom proneness showed in Step three turned to be nonsignificant after including perceived task difficulty in Step four (p = 0.102). Meanwhile, the moderating effect of perceived task difficulty is still significant even after controlling for the boredom proneness (p = 0.005). Comparison between the models tested in Step three and four showed that perceived task difficulty was a more direct and important moderator than boredom proneness between perceived teacher enthusiasm and class-related boredom. 0.212 *** R 2 change 0.135 *** Step 3 PTE × PTD −0.061 ** 0.020 R 2 0.219 *** R 2 change 0.007 ** Note. PTD = perceived task difficulty. * p < 0.05, ** p < 0.01, and *** p < 0.001.
## Comparison of the moderating effects of boredom proneness and perceived task difficulty
To further test the moderating effects, an additional hierarchical regression was conducted, which considered both boredom proneness and perceived task difficulty. The results are presented in. After controlling for the effect of demographic variables in Step one and the predictive effects of perceived teacher enthusiasm on class-related boredom in Step two, the moderating effects of boredom proneness and perceived task difficulty were tested in Step three and four, respectively. The significant moderating effect of boredom proneness showed in Step three turned to be non-significant after including perceived task difficulty in Step four (p = 0.102). Meanwhile, the moderating effect of perceived task difficulty is still significant even after controlling for the boredom proneness (p = 0.005). Comparison between the models tested in Step three and four showed that perceived task difficulty was a more direct and important moderator than boredom proneness between perceived teacher enthusiasm and class-related boredom. 0.221 *** R 2 change 0.071 ** Note. N = 984. BP = Boredom proneness. * p < 0.05, ** p < 0.01, and *** p < 0.001.
# Discussion
## Moderating effect of boredom proneness
According to the results of the present study, after controlling for the effects of demographic variables, boredom proneness played a significant moderating role in the relationship between students' perceived teacher enthusiasm and class-related boredom. Perceived teacher enthusiasm was a relatively weaker predictor on class-related boredom among students with a high level of boredom proneness compared to those with a low level of boredom proneness.
The present results showing the moderating effect of boredom proneness on the relationship between perceived teacher enthusiasm and class-related boredom are consistent with other studies. For instance, Farmer and Sundberg suggested that boredom proneness as a trait could moderate individuals' boredom experience in specific settings. They found that, compared to students with low boredom proneness, students with high boredom proneness had a higher level of boredom emotions and related behaviors. Mann and Robinson found that boredom proneness was an important antecedent of college students' class-related boredom, and students with a high level of boredom proneness reported spent less time in lectures and missing more lecture time. In addition, they found that compared to students with low boredom proneness, students with high boredom proneness used more coping strategies for boredom, and spent more time playing mobile phone games, sending text messages, and making shopping lists, and tended to "switch off," write notes to others, daydream, and decide not to attend the next lecture. Finally, Liu and colleagues found that students' total scores and scores on all sub-scales of the multi-dimensional state boredom scale were positively and significantly correlated with their scores on boredom proneness. According to these results, students' boredom proneness may have important effects on their class-related boredom, especially for those with high boredom proneness levels, who easy experience boredom and find it hard to mitigate it. Therefore, boredom proneness may weaken the mitigating effect of perceived teacher enthusiasm on students' class-related boredom.
## Moderating effect of perceived task difficulty
Perceived task difficulty is an important task or environmental variable that may affect class behaviors and emotions of students. In the theories on antecedents of boredom, earlier research suggested perceived repetitiveness, monotony, and low difficulty of tasks as major causes of boredom. Subsequently, it was suggested that students' perceived low or high task difficulty were causes of their class-related boredom, and those with perceived middle task difficulty may experience minimal boredom. Furthermore, it was easier to elicit class-related boredom in settings with high task difficulty than in those with low difficulty.
Based on the previous contradictory findings, the present study explored the predictive role of students' perceived task difficulty on their class-related boredom, and found that perceived task difficulty positively and significantly predicted class-related boredom, which is consistent with previous theoriesand empirical results. According to the control-value theory, the cause of achievement emotion is probably students' perceived control and value, and environmental variables influent boredom through affecting students' perceived control and value. Higher perceived task difficulty results in a decline in students' perceived control and value, and hence, boredom will be increased. Although previous theories and researchers considered that both low and high task difficulty were important antecedents of class-related boredom, empirical findings showed that high task difficulty had more serious effects on class-related boredom than did low task difficulty. In factual educational settings, most classes in higher education are above the middle level of difficulty. The present study was conducted among college students and the actual learning tasks for them are usually medium or high.
The results of the present study showed that students' perceived task difficulty moderated the relationship between perceived teacher enthusiasm and class-related boredom. Specifically, the perceived teacher enthusiasm of students who perceived high task difficulty had a stronger mitigating effect on their class-related boredom compared to students who perceived low task difficulty. This effect may due to the learning settings in college where most learning tasks involve middle or high difficulty. Compared to the students in primary or middle schools who perceived low task difficulty and experienced low level of boredom, college students who perceived high task difficulty experienced a higher level of boredom, thus providing enough space for mitigating class-related boredom. Therefore, perceived teacher enthusiasm had a stronger negative predictive effect on class-related boredom among students who perceived high task difficulty than among those who perceived low task difficulty.
In summary, the current study showed that students' perceived teacher enthusiasm had a negative and significant mitigating effect on their class-related boredom in most classes with middle task difficulty. It was difficult to find classes with a very high or very low task difficulty, so that perceived teacher enthusiasm would have no mitigating effects on class-related boredom in these extreme situations.
Interestingly, the comparison between the models with the two moderating variables suggested that when considering perceived task difficulty, boredom proneness became silent in the moderating path between perceived teacher enthusiasm and class-related boredom. Statistically, the moderating effect by boredom proneness could be mainly explained by perceived task difficulty. This means that the moderating effect of perceived task difficulty on the association is more direct than boredom proneness. This is a new finding and has not been explored in previous literature. However, the control-value theory could provide a possible explanation to this novel finding and indirect empirical evidence has existed in previous literature. According to the control-value theory, individual characters, such as boredom proneness, work on class-related boredom by affecting students' perceived control and value, while perceived task difficulty can determine students' perceived control and value. Consistently, basic need satisfaction was found to matter more than personality in affecting students' engagement and boredom. To sum up, the effect of boredom proneness on boredom in a specific setting (e.g., class-related boredom) is less than setting-related variables (e.g., perceived task difficulty). Boredom proneness is a personality trait which is not easily changed across settings, whereas perceived task difficulty, perceived teacher enthusiasm, and class-related boredom are variables easily affected by learning settings. Therefore, perceived task difficulty contributes more, compared with boredom proneness, to moderating the association between teacher enthusiasm and class-related boredom.
## Negative and stable predictive effect of perceived teacher enthusiasm on class-related boredom
The present study showed that although college students' boredom proneness and perceived task difficulty moderated the relationship between perceived teacher enthusiasm and class-related boredom, perceived teacher enthusiasm still had a stable, negative, and significant predictive effect on class-related boredom among students with various levels of boredom proneness and perceived task difficulty. That is to say, although both boredom proneness as a trait antecedent variable and perceived task difficulty as an environmental antecedent variable had important effects on college students' class-related boredom, their perceived teacher enthusiasm had a widespread, stable, and significant mitigating effect on class-related boredom. When increasing perceived teacher enthusiasm, students' class-related boredom could decrease accordingly across various learning subjects.
## Practical implications
Considering learning and teaching practice, the present study confirmed the stable and widespread effect of teacher enthusiasm on students' class-related boredom. Based on these results, educational researchers, administrators, and college teachers should pay more attention to the dampening effects of teacher enthusiasm on students' class-related boredom. Therefore, more programs and measures for the cultivation of teacher enthusiasm should be implemented for teachers and pre-teachers, such as improving wages and treatment, and providing a better working environment and more humanistic care for teachers. In addition, it is important to train teachers to naturally and truly express their enthusiasm for their subjects and teaching before their students.
The present study showed that the dampening effect of teacher enthusiasm on class-related boredom was weaker among students with high boredom proneness, compared to those with low boredom proneness. Therefore, before intervening to decrease students' boredom, for best results, it is necessary to measure their level of boredom proneness, and offer different intervention programs based on grouping according to their level of boredom proneness. For the students with low level of boredom proneness, increasing teacher enthusiasm could make a great difference. However, for the students with high level of boredom proneness, additional teaching strategies need to be considered to improve their learning effectiveness.
Compared to boredom proneness, which is difficult to change, there is a larger space for controlling for course difficulty. Results of the present study showed that the dampening effect of teacher enthusiasm on class-related boredom was stronger among students who had a high level of course difficulty, compared to those with a low level of course difficulty. Moreover, the moderating effect of perceived task difficulty between perceived teacher enthusiasm and class-related boredom is more direct than boredom proneness. Usually, college students face a medium or high level of learning challenge leading to relatively high levels of perceived course difficulty, which is one of the important antecedents of their class-related boredom. For the challenging courses, promoting teacher enthusiasm is an efficient and easy-to-implement intervention route to decrease students' class-related boredom.
## Limitations and future research directions
The present study had some limitations. Firstly, the study was based on cross-sectional data and hence, it was difficult to draw causal inferences. Future research should further examine the relationships among these variables using longitudinal data. Secondly, the study focused on the college students. Findings might not apply for younger students in primary, middle, or high schools who typically face simpler learning tasks. Thirdly, the data were based on students' perceptions and self-report measures, and other sources of data such as observation or video analysis were not included. Therefore, to some extent, the stringency of the results of the present study is relatively low. Future research should, as far as possible, use multi-source data to generate stronger evidence. However, some researchers consider students' perceptions and self-report measures of class, teachers, teaching, and learning to be reliable.
# Conclusions
To sum up, the current study suggested that boredom proneness and perceived task difficulty moderated the relationship between perceived teacher enthusiasm and class-related boredom; however, in general, perceived teacher enthusiasm could predict their class-related boredom stably and significantly among students with different levels of boredom proneness and perceived task difficulty. |
Beech and silver fir’s response along the Balkan’s latitudinal gradient
At the 1000 km geographical distance in Dinaric montane forests of silver fir (Abies alba Mill.) and European beech (Fagus sylvatica L.), the tree response from the north-western sites towards southern, warmer and dryer sites was performed during three consecutive growing seasons (2011, 2012 and 2013). On eleven permanent plots, positioned in uneven-aged beech and fir forests above 800 m along the geographical gradient, the physiological and morphological response to light intensity were measured in predefined light categories based on the analysis of hemispherical photos. Radial growth was analysed on all plots and compared to precipitation, temperature and two drought indexes. Analysis showed a decrease in the cumulative precipitation and no change in temperature between plots. Beech was most efficient in the open area light conditions, while fir proved most efficient under shelter. Physiological response for beech increased towards SE and reached its maximal values in the middle of transect, while fir's response decreased from the NW towards SE. Tendency to plagiotropic growth decreased from NW to SE in both species. Growth response to climatic parameters is weak, stronger in fir than in beech and decreasing towards SE.Montane forests of silver fir (Abies alba Mill.) and European beech (Fagus sylvatica L.) in the Dinaric region as the largest contiguous forest area in Central Europe 1 harbour several protected areas (e.g. National parks, Natura 2000) and habitats for many endemic and endangered species. As a long-lived tree, fir is considered a significant ecological and functional species 2 , which stabilizes soils and retains water, and is less susceptible to windthrow and snow or ice breakage than Norway spruce 3 . Fir is also considered a fundamental species for maintaining high biodiversity in forest ecosystems because of its shade tolerance, ability to survive long periods in the understory and to respond when light conditions become more favourable, plasticity to environmental conditions and ability to coexist with many tree species 3,4 . Both fir and beech, as main coexisting species in montane, mixed-species forests, are shade tolerant, and could thrive under conditions of deep shade for longer time periods 5 . Fir is late successional tree, more sensible to water deficits than beech 6 on drier sites 7 .Despite the high degree of forest naturalness, fir's regression is one of the major concerns in the whole region. It was observed already in the 1930s and 1950s, and was attributed mostly to climatic extremes coupled with bark beetle calamities e.g.8,9. Later, polluted air contributed to fir's decline 10 , while its regeneration was exposed to overbrowsing in parts of the region 11 . Intensity of fir regression varied due to different combinations of causes of decline across the region 12 . Current size structure and regeneration characteristics indicate further regression of fir in the next decades. In the Mediterranean region, fir is more sensitive to drought and changes in the seasonal distribution of precipitation 13 compared to non-native Norway spruce. It is a large tree important for site productivity, which forms many special habitats as a veteran or slow decomposing dead tree. Fir is economically much appreciated and the most important conifer tree species within the Dinaric region.European forests are facing enormous threats from rapid global climate change (GCC) with increasing frequency and intensity of summer droughts; considerable uncertainties exist about plants potential to respond to future warming and declining moisture availability 14,15 . In the Mediterranean Basin, drought is the main limiting factor for tree growth 16 , where extreme events are expected to increase 17 and lead to even higher soil moisture deficits during growing periods. Changes in the forest productivity 18 and species distribution in many regions are likely to be expected 19 .Current predictions of climate change impacts on plant demography rely on the association between species' current geographical distribution and corresponding climate characteristics, invoking hypothetical constraints imposed by temperature and/or moisture availability extremes on one or more stages of a species' life cycle 20 . The exact nature of such constraints is, however, unknown and lacking a sufficiently mechanistic basis21,22. To predict species responses to climate changes, physiological limits should be evaluated to obtain a complete representation
of the fundamental niche of a species and then constrain it with biotic interaction and dispersal limitation effects. In northern and western Europe, the increasing atmospheric CO 2 content and higher temperatures are expected to result in favourable effects on forest growth and wood production, while increasing drought and disturbance risks are likely to outweigh positive trends in southern and eastern Europe.
Studies of tree response along the latitudinal gradient have been performed in boreal zones, in eastern part of the Iberian Peninsula on black pine, and across southern distribution limits in Spain, Italy and Romania on fir. The dependency of tree growth on precipitation has increased during the last century and drought has experienced an upward trend after the 1950's. The latitudinal progression of the radial growth decline and proportion of positive trends strongly support the rapid northward advance of Mediterranean climate caused by GCC and its effect on tree ecology and growth.
Recent studies suggest a different response of fir along its distribution range. Its disappearance from warmer and drier areas has been observed in Slovenia in fragmented forests, at the limit of its distributionand in southwestern Europe, particularly in the Mediterranean region where fir's decline is strongly related to increase in aridity. Extreme weather events such as storms, droughts, frosts, lack of precipitation and an increase in average temperature will influence fir demography. A shift in fir distribution toward higher elevations and northwards is expected. On the Balkan Peninsula, where different and well-expressed ecological factors intertwine at relatively short geographical distance (approx.1000 km), the studied tree response from the southern, warmer and dryer sites may serve as a most probable future prediction for the same species-response on currently less extreme sites.
Our aim wasto compare physiological and morphological responses of beech and fir along the defined 1000 km geographical gradient,to evaluate differences in same light categories of both species between managed and old growth forest, andto verify the connection between radial growth of adult stand with ecophysiological and morphological traits of both species along the gradient.
# Material and methods
Location and plot description. In one of the largest mountain regions in Europe, the Dinaric mountain chain extends from the southern edge of the Eastern Alps in Slovenia to the mountain massif in North Macedonia; it is bounded by the Adriatic Sea along its western border and the Pannonian Basin toward the east. The main part of the range is formed of Mesozoic rocks, predominately of limestone and dolomite. The depth of the limestone and dolomite is unique, typically 1-3 km, with considerable local variation.
Westerly winds blowing over the Adriatic Sea bring large amounts of humidity to higher elevations along the western side of the range. Precipitation throughout the upper elevation zone is relatively evenly distributed during the year, with snowpack often lasting up to six months. Forest structure and composition in the region is strongly influenced by the interaction of the mountain relief, karst terrain, soils, and climatic gradient. Mountain forests above 800 m include mainly beech dominated forests and mixed uneven-aged forests dominated by varying amounts of beech, fir, and occasionally spruce. Large forested regions in the interior range have been left intact until the present days and have been managed with low intensity silvicultural systems for more than a century, with several protected old-growth remnants scattered throughout the area.
Eleven permanent plots were established in the optimally developed managed beech and fir adult forest stands, distributed from Slovenia from its furthest NW part over Croatia, Bosnia & Herzegovina (BiH) and Montenegro along the mountainous region of the Balkan Peninsula to North Macedonia on the SE part of the range. All selected plots were positioned at elevations above 800 m. Among selected locations three (plots No 3, 7 and 8) belong to the old growth reserves, .
On every plot, three categories of different light intensities were defined based on the analysis of hemispherical photos: under closed canopy with Indirect Site Factor (ISF) < 15%, at the forest edge (15% < ISF < 25%) and in the open ISF > 25% 37 . Assimilation response was measured in saplings of fir and beech in June and July during three consecutive growing seasons . Age of the trees varied between 5-12 years. In every light category and site, at least 12 trees were measured.
Weather and climate. Data of mean monthly temperatures (°C) and total monthly precipitation were interpolated for the 0.5° grids including each sampled stand and corresponding to the CRU TS 4.01 dataset. Climatic data were obtained from the Royal Netherlands Meteorological Institute 'Climate Explorer' web page (http:// climexp.knmi.nl). For the comparison between A max , Φ, morphological response and climate (temperature and precipitation) data from the last 30-year average period (1985-2015) was used. For the long-term comparison between climate and tree growth, we defined four regions (A, B, C, D) where A is the northernmost and D the southernmost one. For each region we extracted gridded climate data for the mean monthly temperature and sum of monthly precipitation using CRU TS 4.01 dataset with 0.5 × 0.5-degree resolution from the KNMI web site.
Nitrogen content (N tot ) and leaf mass per area (LMA). Leaves and needles were sampled in the upper crown position of minimal 12 trees per light category and location, then cool-stored in airtight conditions. Same trees were used for the assimilation response measurements. Nitrogen concentration (N tot ) [mg/g] was determined to compare macronutrient status (Leco CNS-2000 analyser)for open-, forest edge-and closed canopy-category below mature trees. Fresh leaves were weighed and scanned for the leaf area. Leaves were dried at 105° for 24 hours until constant weight and weighed for the dry mass in the lab to provide leaf mass per area
[formula] (LMA) [g/m 2 ]. [/formula]
Assimilation light response. The assimilation of beech and fir was measured on randomly distributed saplings along same light categories along three relative diffuse light (ISF) categories: of <15%, 15 to 25% and >25% defined by the hemispherical photos. The hemispherical photos were taken at each sampling group of saplings separately, prior to further assimilation measurements. At least 12 young trees of the same height, unobstructed by their neighbours, were randomly chosen for light saturation measurements, performed during three sequential growing seasons (2011, 2012 and 2013) (sensu. The light-response was measured with an LI-6400 portable system on at least four leaves/locations per tree, located in the upper third of the tree-crown.
Light saturation curves were established to compare the net assimilation (A max ) in young beech and fir trees in the same light conditions. All assimilation measurements were performed in field at a constant temperature of the measurement block (20 °C), a CO 2 concentration of 420 µmol/l, airflow 500 µmols −1 and different light intensities: 0, 50, 250, 600 and 1500 µmolm −2 s −1 . Maximum assimilation (A max ) rates for the light saturation curves were used for comparisons of responses between different light categories and plots.
The characteristic points of maximal quantum yield (Φ), defined as the maximal amount of fixed CO 2 per amount of absorbed light quanta, were established for each light category, species and plot, as described in Čater et al.. www.nature.com/scientificreports www.nature.com/scientificreports/ Morphological response. Changes in crown morphology play an important role in the acclimation capacity of species to reduced light intensity under a mature canopy and in younger development stages. Crown morphological plasticity has been found to be especially important in shade-tolerant species. Plagiotropism, result of deviation from the vertical axis, is the most unwanted effect with respect to future timber that increases with the degree of shade.
To evaluate the morphological response of beech, the quotient between the length (l) and height of trees (h) was used, which increased in the case of plagiotropic growth. As young trees show slight deviation in growth from the vertical axis, which is not necessarily a function of plagiotropic growth, a wider threshold value of l/h ≥ 1.1 (110%) was chosen to separate plagiotropic from orthotropic growth. For fir maximal distance from the stem to the furthest branch, tip (d) was measured and compared with tree height (h) in all light conditions (ISF%); a threshold 2d/h ≥ 1.15 (115%) was used accordingly. The limiting value of light was defined after measurements of the ratio after three consequential growing seasons . In the analysis, the values under same light-intensity conditions were compared between plots (how closed/open the mature stand was). Plagiotropic behaviour between different plots was assessed by comparison of data into an exponential-decay-3 parameter curve (1),
[formula] = + * − Y A B e (1) [/formula]
Cx where x is the measured light (ISF%), Y is the quotient between the length (l) and the height of saplings (h) and A, B and C are the curve parameters.
Growth response. Radial growth analysis for the last 100 years was conducted for adult beech and fir trees;
two cores per tree from 15 adult dominant beech and firs on each location were taken, 660 in total. Each core was mounted and sanded to a high polish using sanding paper of progressively finer grit. The cores were then scanned using ATRICS systemand annual radial increments measured to the nearest 0.01 mm using CooRecorder and CDendro software (www.cybis.se). We used the same software for quality control of measured tree-ring width sequences. Tree ring series were then visually and statistically cross-dated using PAST-5.
## Statistical analysis. differences between same year (2011, 2012 and 2013) for the lma, n tot , a max and φ
were tested with the two-way ANOVA with species (beech and fir) and light (open, edge, canopy) as a dependent variable. Analyses of variance (ANOVA) and HSD Tuckey post hoc test were used after testing data to meet conditions of normality. Probability values of p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) were considered significant. Data analysis and correlation between measured variables was performed with Statistica data analysis software system (2011). Tree-ring widths were standardised, using R software and "dpl" package 46 with double detrending approach. First, a negative exponential curve was fitted to tree-ring width sequence to remove age trend, in the second step flexible, a 30-year cubic spline was fitted to the tree-ring width indices from the first step to remove any other disturbances and retain only climatic signal. Residual chronologies were calculated as the quotient between measured and fitted value and used in all climate-growth relationships, using R software and package "treeclim". Statistical relationship between meteorological months and residual chronologies for the 1902-2010 period from September prior to year of ring formation till September of current ring formation was analysed. We also ran a 31-year moving window correlation using only statistically significant combinations of meteorological data and tree-ring widths. Pointer years (PY) defined by Schweingruber et al.were calculated for every region.
# Results
Climate. On the studied range, climatic conditions in the SE differ from those in the NE: generally, the temperature in the southern part is higher, while the precipitation is lower. Sub-mediterranean mountainous climate prevails, characterized by high winter precipitation and markedly low summer precipitation. Climatic condition for the last 30-year period indicate homogenous average temperature range on all studied plots within 12°-14 °C range and a significant decrease of annual amount of precipitation towards SE -from plot No.1 to plot No.11.
When comparing precipitation and temperature values of two different consecutive referential periods 1995-2004 and 2005-2015 within April-September, the drop of total precipitation and rise of average annual air temperature was confirmed. On plots 1-9, evident decrease of precipitation between periods amounted to between 10-20 mm with the exception of plots 10 and 11, while the temperature rise ranged between 0.33°-0.65 °C on all studied plots.
## Nutrient status and leaf mass per area (lma). the foliar nitrogen amount (n tot ) for beech and fir was
highest in the open and lowest under canopy conditions on all plots, without confirmed differences between same light categories on different plots and observed years. The content was found within optimal threshold values reported by Grassi and Bagnaresi 50 or Mellert and Göttlein 51 . LMA increased from the shelter towards the open light category, with significant differences between shelter and open categories for both beech and fir, .
## Assimilation response (a max ) and quantum yield (φ).
For both beech and fir no difference was confirmed in A max and Φ between observed years 2011-2012, 2012-2013 or 2011-2013, respectively. Evident increase in A max in all light categories from NW to SE was confirmed for both species. Φ for beech increased from NW, reached peak in the middle of the studied transect and decreased towards SE, while maximal Φ for fir were in NW, followed by the evident decrease in all light categories towards SE. www.nature.com/scientificreports www.nature.com/scientificreports/ Differences in A max and Φ between same light categories were significantly different for both beech and fir.
[formula] Scientific RepoRtS |( [/formula]
Different response in beech and fir is confirming our former research, showing increasing Φ in beech with increasing light intensity and the opposite, decreasing in fir, respectively . Differences are even more pronounced along the geographical gradient, showing increase in Φ for beech towards the SE in all light categories and decrease in Φ for fir.
Post-hoc analysis confirmed statistically significant differences between all categories of light for both A max and Φ, except on locations with old growth reserves, where no significance has been confirmed between forest edge and open light categories for both beech and fir.
Maximal Φ for beech was observed in open light, for fir under maximal shelter. When comparing light responses between old growth reserves and the neighbouring managed forests, A max and Φ were in all cases significantly higher than in managed forests. We believe the reason is the microclimate, as the former study within same forest complex highlighted higher relative humidity (RH), higher water use efficiency (WUE) and also photosynthetic nitrogen use efficiency (PNUE) in old growth reserve compared to managed forests in all light categories.
Evident shift of Φ in the edge category towards the open light in all old growth reserves was observed for both species compared to managed forests, while in managed forests the distribution of light categories was more even. www.nature.com/scientificreports www.nature.com/scientificreports/ A max and Φ for both fir and beech confirmed stronger relation between average annual precipitation than between average annual temperature. Relation between Φ and temperature was more pronounced for fir than beech.
## Morphological response.
Morphological reaction between light intensity and all studied variables (e.g., total openness or direct light) was highest in the case of ISF (data not shown). On all plots, plagiotropic growth was triggered below 17% ISF for beech and fir. Breaking or deflection points (DP), where orthotropic growth changed to plagiotropic growth due to lower light intensity were lowest according to ISF (%) values in old growth reserves -13.5% for beech, 13.7% for fir, respectively, and increased towards SE. In terms of dispersion of l/h (beech) and 2d/h (fir), the ratio in old growth reserves were in all cases smaller compared to the Relation between DP and precipitation along the studied range was more pronounced for the growing season (R 2 beech = 0.82 and R 2 fir = 0.75) than for the whole year-period (R 2 beech = 0.64 and R 2 fir = 0.67). Stronger relation between the morphological response (DP) and temperature on the whole gradient was confirmed more for the growing period (R 2 beech = 0.36; R 2 fir = 0.26) than for the entire year (R 2 beech = 0.25; R 2 fir = 0.12), respectively.
Climate-growth response. Radial growth of firis more sensitive to climate than beech. In regions A and B we confirmed clear summer-precipitation signal, where above average precipitation (p) positively influenced radial growth of fir. This was also supported by the positive correlation with 3-month standardised precipitation index (SPI3) and Palmer Drought Severity Index (PDSI), both showing the same direction of correlation as p. Temperature (T) influenced the growth of fir in regions A and B, but not to such an extent as p. In region A, the correlation values for p, SPI3 and PDSI are similar; in region B, correlation between p and radial growth decreases, while correlation between SPI3 and PDSI are the same as in region A. We believe these results indicate tree growth as more sensitive to long-term rather than short term water deficit. In more southern regions, C and D, we confirmed only weak correlation between climate variables and fir radial growth. In region C, only PDSI shows correlation between February PDSI (till September of the preceding year) and radial growth. Weak negative correlations between T and radial growth in August and September indicate that water required for the growth is accumulated during the winter in a form of snow and slowly released during the growing period. In southernmost region D, no significant correlations between any of the studied climatic variables and radial growth were confirmed. Only two weak correlations between T-a positive in December of the preceding year and a negative in September of the year of ring formation were evident. Both were weak and on the edge of significance.
Growth response of beechto climate is different than in fir, although trees from the same locations and forest stands were sampled. In the northernmost region A, beech shows positive response to p in March and June as well as significant correlation between drought indices, SPI3 and PDSI and radial growth. PDSI shows more significant correlations, highest in June and July. Response of radial growth to T is weak and significant only for www.nature.com/scientificreports www.nature.com/scientificreports/ November and December of the year prior to ring formation. In region B, significant positive correlation between radial growth and p in June was evident. Both drought indices and T in the growing period were without significance. In southernmost regions (C and D), growth is characterised by positive response to T in May and negative response to p in September, which can be associated with abrupt stop of growth. In region D, beech responded positively to April p and negatively to above average T in April. Relatively high values of correlation with SPI3 in September and October of the year prior to ring formation indicate long-term demand for water, also visible through positive correlation with p.
Comparison of tree growth response to climate between managed and old-growth reserves did not confirm any differences; management of studied forests is evidently close-to-natural processes in old-growth reserves, without significant disturbances caused by forest management.
Within the 1901-2010 period, we identified altogether 53 pointer years (PY) for fir and 50 for beech. For fir, 26 PY were positive and 27 were negative; for beech, 27 years were positive and 23 were negative. Number of PY varies between the regions, with B and C having significantly more than regions A and D. We couldn't find only a single PY year that would be common to all four regions and both tree species. Only one positive PY (1958) was.
# Discussion
Many studies have predicted substantial changes in forest dynamics during the next century because of GCC. As some spatial distribution models have projected reduction of the distribution areas of fir and beech forests by 2100 owing to climate change to the benefit of more drought-tolerant species, the question of the prevailing impact of ecological as opposed to macroecological and phytogeographical gradients on vegetation is highlighted in many studies.
All selected study sites were located above 800 m above sea level; study sites in the SE were located at higher elevations than in the NW. That is understandable, because there is a gradient in T and humidity in the area under study. The mean annual T differed very little among the study sites, showing that despite differences in elevation the study plots represent similar climatic conditions.
According to selection criteria of our studied fir-beech sites, no significant relation was confirmed between measured physiological and morphological parameters and the T, while more pronounced relation was confirmed with p, particularly for fir. Change of responses in studied species is in accordance with Marinšek et al., who confirmed increase in proportion of chamaephytes, hemicryptophytes and therotypes towards SE in mesophylous beech forests of SE Europe. We relate comparable amount of nitrogen content in leaves on all plots with similar temperature and site conditions.
Elevations of studied plots above 800 m show constant average and relatively narrow T belt without expressed change in T along the entire geographical gradient. We believe this is the reason for statistically insignificant relation between studied parameters and T or surface solar radiation (both from KNMI database) either during the entire year or during the April-September growing period, respectively. High correlation was confirmed only between locations and different cloud coverage towards SE direction (r 2 = 0.86), but only during growing period, which could be related with increasing A max values in both species towards the SE and increasingly longer growing period in the SE. Differences in A max between light categories were significant on all locations and became more pronounced for beech towards SE and in smaller degree for fir. Φ for beech was highest in the central range (Bosnia and Herzegovina), and for fir in the NW part of the transect (Slovenia, Croatia).
We may relate higher Φ in fir with sites and micro locations with predominating diffuse light and lower Φ in SE with limiting edge of its natural distribution 54 , while Φ for beech culminates on sites with predominately direct light component. Φ on each location for beech was maximal in microsites with maximal light intensity, while for fir it was maximal in shelter, with predominating diffuse light, which is in accordance with our previous work. Φ for beech increases with light intensity, so its adaptation ability to light increase is better than in fir. The lower Φ of fir in exploiting high-intensity solar radiation compared to beech may be a competitive disadvantage in large canopy gaps, which could limit species recruitment to the forest understory or small gaps, especially in admixture with beech. It is not clear what caused the shift of Φ in the edge light category in the old growth reserves towards the open category uniformly for both beech and fir, as leaf nitrogen values were comparable between sites and in optimal range on all plots. Beech can tolerate a broad range of understory light levels and manages to recruit in a variety of light conditions in young stages of growth, as it is capable of decurrent and polycyclic growth. Fir is more shade-tolerant than beech, but is also considered a "late-successional species. Its competitive strength is, compared to beech, in low light conditions greater, but in an intermediate and ample light condition it is consequently smaller; in gap-openings, beech adapts better and much faster to rapid changes in light intensity, while fir's adjustment of growth rate to light environment occurs gradually over several years. Competition for light may be more pronounced in old growth reserves compared to managed forests, where larger proportion of forest edge category is artificially created. In old growth reserves, gaps are created after disturbances and present an opportunity for overgrowing of the present tree species.
Morphological adaptation of juvenile trees to various light levels represents an important species-specific characteristic. Relation reflects the control exerted by the apical shoot over the outgrowth of the lateral buds. Plagiotropic growth, usually evident under conditions with lower light intensities and under dense canopies with predominating diffuse light component, may be also influenced by the forest management approach. Beech and firs are better adapted to higher light intensities on sites with applied irregular shelterwood system than on single-tree selection sites, while maximum shade tolerance was confirmed in old growth 37 . Wagner and Müller-Using 59 quote a limiting value of 10%, while Zang and Biondi et al.state 15% of relative light intensity, below which plagiotropic growth is evident, which accords better with our results.
We believe the selected criteria encompasses the whole plant size and thus responds to the cumulative light environment of the plant during its lifespan and not only the last growing year, as in case of apical dominance ratio (ADR) used by Ripullone et al., who used the ratio between apical shoot length and length of first whorl lateral twigs. There is, however, no clear evidence that shade-tolerant species are morphologically more plastic than less tolerant ones.
Decrease of shade tolerance along the geographical gradient towards SE in both beech and fir is in accordance with Marinšek et al., who confirmed significant increase in ecological indicator values (EIV), especially for light and decrease in EIV for moisture and nutrients for beech forests along the geographical gradient towards SE. The plagiotropic growth responses of trees to different light intensities in our study were non-linear ; the ratio between canopy density (ISF%) and plagiotropic shape increased exponentially after the light dropped below 13.9% on the greater part of NE plot and below 15.8% on the greater part of SE plot. www.nature.com/scientificreports www.nature.com/scientificreports/ In old growth reserves, plagiotropic growth in beech and fir was triggered by smaller light intensities (DP) than in managed forests. Assimilation rates (A max ) and efficiency (Φ) were also higher than on neighbouring managed forest sites.
Fir growth response to climate was slightly stronger than in beech. Both variables -T and p have stronger influence on growth of fir than on beech. Climate signal in fir diminished from NW to SE, where only drought indices remain significant, while beech response to climate was weaker on all plots and diminished, similar as in fir, from NW to SE.
We confirmed a clear spatial effect in climate response for both species. In northern regions A and B, the response to climate is better than in southern C and D regions, which contradicts our expectations. In southern regions one would expect increasing water deficit, pronounced climate sensitivity and higher correlations with climatic parameters related to precipitation or drought. In two southernmost regions, low climate sensitivity with significant response to above-average temperature and negative response to precipitation (visible in 3-month SPI) in March was confirmed for beech. Similar high values for Slovenia and northern parts of Croatia (e.g. region A) were found by Čufar et al., while similarly low values for the climate-growth relationship for the B, C and D regions were found by Tegel et al.for the sites in Albania and North Macedonia and by Hacket-Pain 65 for sites in Greece, who confirmed diminishing climate signal in beech after 1990 and increasing growth despite continued dry and hot summer conditions. Low values for the climate growth relationship were also found for old-growth forestsand managed forests 67 in Bosnia and Herzegovina and for managed forest in Serbia 68,69 . Stjepanović et al., who studied beech growth and climate response in a relatively dense network of beech sites in Serbia and Bosnia & Herzegovina, which correspond to our regions B and C, confirmed that beech responded to climate only at a lower, warmer and drier elevations, and in some cases also at the upper timberline. Beech in the optimal altitudinal distribution range shows weak response to climate and mainly responses to other, non-climatic factors.
Phenomenon of weak to non-existing responses to different climate variables found as we move southwards along the transect were found by many studiesand is suggested to be connected to either genetic adaptation, phenotypic acclimation of the species or combination of both. If the results are genetically based, trees within a given location could be much more sensitive to climate change than indicated by the very broad geographic distributions of these temperate tree species, but if the results are phenotypic, this would represent local acclimation that could help buffer species in the face of climate change.
Fir, with higher demand for water than beech, was never regarded as a drought-tolerant species. After a pronounced period of its dieback e.g., species recovered and shows signs of growth increase. The majority of studies agree that below-average summer T and above-average p positively affect its growth e.g.. Our study confirms this significant relation in regions A, B and C, but not D. Fir response to climate has a clear spatial element: in NW July T and p (together with drought indices) were the most influential factors, in region B a wider time frame of the response ranging from May till August for p, T, PDSI and SPI-3 was confirmed, while in region C the main driving factors are p in May and PDSI, indicating importance of water availability at the beginning of the growth. In region D, only PDSI and SPI-3 correlated with growth, confirming again water availability as a major factor influencing the radial growth.
In the middle of the transect, in regions B and C, trees are particularly responsive to extreme years, showing higher number of positive and negative PY compared to regions A and D. We found such response hard to interpret; some studies show that beech might be more vulnerable to drought stress than initially assumed or visible from the measured parameters.
Well preserved condition of studied beech-fir forests is in tight relation with their low management intensity in the past. Their uneven-aged structure above 700 m is similar to the structure of old growth reserves 57 , where elevation represents the key factor controlling the microclimate in temperate mountain forest stands. Liberal selection of felling regimes applied in uneven-aged beech-fir forests, also known as close-to-nature silviculture 77 , employ relatively low intensity and small-scale felling regimes to mimic natural forest composition, structures, and natural disturbances on the lower end of the disturbance severity gradient at stand scales. Forest stands managed in such a way generally create stands with small-scale heterogeneous structure and are thought to be both resistant and resilient to disturbance, as was also confirmed by our results. Disadvantages of uneven-aged forestry include the reliance on shade tolerant species, which can be hampered by climatic conditions of open areas created by disturbances.
Further decline of fir may be expected because of its higher proportion in the stand volume than in near-natural forests and by recruitment failure due to overbrowsing and strong loss of vitality. Limited silvicultural options for preserving fir in an adequate proportion in mixed mountain forests, where both climate change and browsing pressure are present, may lead to a non-compensatory effect, where increase in browsing pressure may enhance the shift in dominance of tree species.
Different response in beech and fir is confirming our former research, showing increasing efficiency in beech with increasing light intensity and the opposite, decreasing in fir, respectively. Physiological and morphological differences are even more pronounced along the geographical gradient, showing efficiency increase in beech towards the SE in all light categories and decrease in fir. Morphologic changes indicate reduced shade tolerance towards SE for both species. Reasons for such response might be in the natural range of the species distribution or better plasticity in beech compared to fir.
Fir in the SE part reaches its southernmost range of natural distribution, where droughts and precipitation deficit are more pronounced compared to the NW region of the study. |
The TRIMendous Role of TRIMs in Virus–Host Interactions
The innate antiviral response is integral in protecting the host against virus infection. Many proteins regulate these signaling pathways including ubiquitin enzymes. The ubiquitin-activating (E1), -conjugating (E2), and -ligating (E3) enzymes work together to link ubiquitin, a small protein, onto other ubiquitin molecules or target proteins to mediate various effector functions. The tripartite motif (TRIM) protein family is a group of E3 ligases implicated in the regulation of a variety of cellular functions including cell cycle progression, autophagy, and innate immunity. Many antiviral signaling pathways, including type-I interferon and NF-κB, are TRIM-regulated, thus influencing the course of infection. Additionally, several TRIMs directly restrict viral replication either through proteasome-mediated degradation of viral proteins or by interfering with different steps of the viral replication cycle. In addition, new studies suggest that TRIMs can exert their effector functions via the synthesis of unconventional polyubiquitin chains, including unanchored (non-covalently attached) polyubiquitin chains. TRIM-conferred viral inhibition has selected for viruses that encode direct and indirect TRIM antagonists. Furthermore, new evidence suggests that the same antagonists encoded by viruses may hijack TRIM proteins to directly promote virus replication. Here, we describe numerous virus-TRIM interactions and novel roles of TRIMs during virus infections.Vaccines 2017, 5, 23 2 of 38UbiquitinUbiquitin is a key protein in orchestrating immune responses. Ubiquitin is a 76-amino-acid protein that associates either covalently or non-covalently with proteins[1][2][3]. Interaction of ubiquitin with its targets may influence subcellular localization, signaling, trafficking, protein stability, and cell cycle progression[1][2][3]. The addition of ubiquitin to a protein target is catalyzed by three classes of ubiquitin enzymes. The ubiquitin activating enzyme (E1) binds the C-terminal glycine of the ubiquitin molecule to its active-site cysteine in an ATP-dependent manner to form a high-energy thioester bond[1,2]. Interaction of the E1 with ubiquitin exposes a site to enable the recruitment of the next enzyme[4]. Via a trans-thiolation reaction, the ubiquitin is transferred from the E1 to the E2 active-site cysteine[1,2]. Once the ubiquitin is bound, an E3 ubiquitin ligase can interact with an E2-ubiquitin conjugase and assist in the transfer of the ubiquitin from the E2 onto a target protein[1,2]. When covalently attaching ubiquitin to a protein, most often the ubiquitin is covalently linked at the lysine (K) ε-amino group[1,2].Since ubiquitin has seven lysine residues itself, K6, 11, 27, 29, 33, 48, and 63, the ubiquitin enzymes can coordinate the formation of covalently linked polyubiquitin (poly-Ub) chains[1,2]. In addition, ubiquitin molecules can be linked in a head-to-tail orientation in which the donor ubiquitin's C-terminal glycine is linked to the acceptor ubiquitin's methionine-1 amino group [1,2]. The E2 primarily determines the ubiquitin chain topology[4,5]while the E3 is more important in identification and recruitment of the appropriate target[4]. Ubiquitin can be covalently linked to a protein as a single ubiquitin at one site (monoubiquitination), single ubiquitin molecules at multiple lysine residues of the same target protein (multi-monoubiquitination), or a covalently linked chain of ubiquitin to a single lysine (poly-Ub) [1]. Additionally, ubiquitination enzymes can synthesize unanchored poly-Ub chains, which are not covalently attached to any other protein, and act as a ligand for proteins containing a ubiquitin-binding domain (UBD)[1,4,6,7]. In addition to the complexity of chain length and topology, phosphorylation and acetylation of ubiquitin has also been identified[8,9]. The unique combination of poly-Ub chain topologies, covalent or non-covalent modifications, and chain length variation allows for precise signaling regulation.The host encodes a variety of mechanisms to enact, regulate, and interpret this complex 'ubiquitin code.' Ubiquitin modification can directly influence the target molecule through induction of a conformational change. Such changes may recruit other molecules, expose a subcellular localization sequence, or alter protein stability[1,3]. Recognition of ubiquitin by UBD-containing proteins is analogous to a receptor binding its ligand. The UBD-containing proteins may recognize a specific poly-Ub chain topology (e.g., UBD specifically recognizing K48-linked poly-Ub chains) and/or chain length (e.g., a UBD that interacts with monoubiquitin) and then recruit specific molecules to create a signaling complex[1,3,10]. Although some exceptions exist, different poly-Ub chain topologies are associated with particular effector responses. The best-characterized examples of chain topologies include K48-linked poly-Ub chains, which stereotypically target proteins for proteasome-mediated degradation, or K63-linked poly-Ub chains, which generally coordinate cell signaling complexes and subcellular localization[1,3]. Despite the canonical roles of these chain topologies, K48-linked poly-Ub chains have been noted to be involved in promoting signaling complex assembly [11] and K63-linked poly-Ub chains can mark protein targets for degradation[12]. To regulate the ubiquitin response, the host also encodes deubiquitinating enzymes (DUB) or ubiquitin specific proteases (USP), which can disassemble poly-Ub chains or coordinate the removal of ubiquitinated enzymes[1,3]. Overall, the host encodes the enzymes necessary to generate, respond to, and eliminate ubiquitin modifications.Tripartite Motif (TRIM) ProteinsTripartite motif proteins (TRIMs) are an E3 ligase family critical in many cellular functions, including the regulation and coordination of innate immunity and antiviral responses[6,7,[13][14][15][16]. TRIMs are known to coordinate the ubiquitination of target proteins to assemble signaling complexes, mediate proteolytic degradation, alter subcellular localization, or modulate the host's transcript or Vaccines 2017, 5, 23 3 of 38 protein composition (reviewed in: [2,6,7,17]). TRIM proteins are such named due to their conserved RBCC domain. The RBCC domain includes a really interesting new gene (RING) E3 ligase domain (R), one or two B-box domains (B), and a coiled-coil domain (CC) [2,18]. Like other RING motif-containing E3 ligases, the RING domain of TRIMs usually mediates the interaction with ubiquitin-bound E2 via the zinc finger motifs[2,4,17]. These motifs are comprised of key cysteine and histidine residues that coordinate binding to two zinc ions and facilitate protein-protein interactions[2,4,17]. The function of the B-box is less well characterized, however studies suggest that this domain is important in coordinating TRIM self-association and protein-protein interactions, as well as promoting formation of higher-order TRIM oligomerization (reviewed in[17]). Since the B-box domain also contains a zinc finger motif, it has been suspected to confer E3 ubiquitin ligase activity to some TRIMs[19,20], although it does not appear to be a general feature of the TRIM family[21]. Some TRIMs include only a B-box 2 domain, while others encode a B-box 1 followed by a B-box 2[17,18].The coiled-coil domain is a hyper-helical structure that allows for dimerization and self-association of TRIMs[17,18,21,22]. Crystal structures and biochemical analyses of the coiled-coil regions of TRIM5α, TRIM20, TRIM25 and TRIM69 indicate that the TRIM dimers are formed by coiled-coil antiparallel helical structures[17,[22][23][24][25], and bioinformatics analysis suggest that this could be a general characteristic of the TRIM family[22]. Often, dimerization of TRIMs via the coiled-coil domain and formation of higher order oligomerization through the B-box domain is critical for TRIM function and E3 ligase activity[17,21,23,[26][27][28][29]. Furthermore, in some cases, the antiparallel nature of the coiled-coil dimers and the formation of higher order oligomers allows for dimerization of RING domains, which are also important for enzymatic activity[17,23]. However, the presence of one or two B-box domains, or the presence of different C-terminal regions, may affect dimerization and/or higher order oligomerization of TRIMs. In fact, the oligomerization mode of TRIM32 and TRIM25, which contain one and two B-box domains respectively, differs greatly and is critical for their function[21]. Prediction of TRIM structure and function is further complicated by differences in TRIM domain composition. For example, although TRIM members are characterized by the presence of an RBCC, some TRIM-like members, including TRIM14 and TRIM16, lack a RING domain within the RBCC[19,30].The variable C-terminal region of TRIMs is responsible primarily for interaction with target proteins and subcellular localization[2,18,31]. To date, 11 classes of TRIM C-terminal domains have been characterized[7,17,18]. The most prevalent TRIM C-terminal domain is the B30.2, or PRY-SPRY, domain with approximately 40 members identified in humans[6,17,32]. TRIMs post-translational modifications[11,[33][34][35][36][37], alternatively spliced isoforms[16,38,39], and heterodimerization[18,19]expand the diversity and functionality of this protein family. The expanding interest in TRIMs relates to the realization that these proteins play crucial roles in regulating responses to pathogens, autoimmune pathologies, and cancer.The evolutionary history of TRIMs in vertebrates correlates with the role of this E3 ligase family in regulating responses to viral infection. Though present in invertebrates, the TRIM family is greatly diversified in vertebrates coinciding with the evolution of the interferon (IFN) pathway and adaptive immune system[32,38]. Further, several TRIMs are clustered within chromosomes[38][39][40]. The clustering of closely related TRIMs suggests they evolved due to gene duplication[38][39][40][41]. The close proximity of several TRIMs, particularly those with a B30.2 domain, near known immune genes such as HLA (Human Leukocyte Antigen)[38][39][40]supports immune regulation as a TRIM-mediated function. Additionally, the regulatory regions of several TRIMs include target sequences recognized by transcription factors involved in coordinating immune responses[42]. The species-specific expansion of B30.2-containing TRIMs[32,43]suggests that this group is specialized. In humans, the 10 TRIMs without murine orthologs all contain a B30.2 domain[32]. Additionally, the region encoding the C-terminus of many TRIMs is under positive selection[32,39,41,43,44]. The non-coding, transcriptional regulatory regions of more recently evolved TRIM genes are also
# Introduction
Eukaryotes are constantly exposed to a variety of pathogens, including viruses. As with other environmental signals, viral invasion triggers tightly regulated intracellular signaling cascades to optimally respond to infection. Mammals enact both an innate and an adaptive immune response to identify an infecting pathogen, to clear the foreign agent, and to protect against subsequent invasion. A primary mechanism for fine-tuning molecular pathways is utilization of post-translational modifications. Altering the functional proteome influences protein interactions, transcriptional programs, translation, secretion, and cytoskeletal arrangement. A variety of molecules, including phosphates, sugars, lipids, or proteins, can be attached or removed enzymatically to modulate protein function. Post-translational modifications thus enable rapid and reversible regulation. under positive selection [bib_ref] Positive selection of the TRIM family regulatory region in primate genomes, He [/bib_ref]. This species-specific pattern of positive selection of closely related TRIMs suggests that individual TRIMs play specific antiviral roles.
In addition to differential mRNA expression upon viral infection, several TRIM family members are intimately involved in the antiviral response. Type-I IFNs and other cytokines, such as pro-inflammatory cytokines induced via the NF-κB pathway, have been noted to differentially regulate the expression of a significant population of TRIMs [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Human trim gene expression in response to interferons, Carthagena [/bib_ref] [bib_ref] Induction of TRIM22 by IFN-gamma involves jak and pc-plc/pkc, but not mapks..., Gao [/bib_ref] [bib_ref] Expression of the immune regulator tripartite-motif 21 is controlled by IFN regulatory..., Sjostrand [/bib_ref] [bib_ref] Type I interferon-dependent and -independent expression of tripartite motif proteins in immune..., Rajsbaum [/bib_ref] [bib_ref] TRIM38 negatively regulates TLR3/4-mediated innate immune and inflammatory responses by two sequential..., Hu [/bib_ref] [bib_ref] TRIM13 is a negative regulator of MDA5-mediated type I interferon production, Narayan [/bib_ref]. Likewise, TRIM overexpression influences the transcription of type-I IFN, pro-inflammatory cytokines, and IFN-stimulated genes (ISGs) [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] The E3-ligase trim family of proteins regulates signaling pathways triggered by innate..., Versteeg [/bib_ref]. The roles of TRIMs in viral infection include intrinsic restriction of viral pathogens, positive regulation of immune pathways that promote viral clearance, and negative regulation of antiviral pathways to limit immunopathology [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref] [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Tripartite-motif proteins and innate immune regulation, Mcnab [/bib_ref]. The incorporation of TRIM antagonists into viral genomes exemplifies the importance of TRIMs in antiviral responses [bib_ref] Regulatory role of TRIM21 in the type-I interferon pathway in japanese encephalitis..., Manocha [/bib_ref] [bib_ref] The interferon signaling antagonist function of yellow fever virus NS5 protein is..., Laurent-Rolle [/bib_ref] [bib_ref] Dengue subgenomic RNA binds TRIM25 to inhibit interferon expression for epidemiological fitness, Manokaran [/bib_ref] [bib_ref] The matrix protein of nipah virus targets the E3-ubiquitin ligase TRIM6 to..., Bharaj [/bib_ref] [bib_ref] Influenza a virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition..., Gack [/bib_ref] [bib_ref] Viral evasion mechanisms of early antiviral responses involving regulation of ubiquitin pathways, Rajsbaum [/bib_ref]. Here, we will focus on the role of TRIMs in the direct and indirect inhibition of viruses and novel mechanisms of viral-mediated antagonism and hijacking of TRIMs. Excellent reviews on the roles of TRIMs in autophagy, cancer, and other diseases have been recently published [bib_ref] TRIM family proteins: Roles in autophagy, immunity, and carcinogenesis, Hatakeyama [/bib_ref] [bib_ref] Precision autophagy directed by receptor regulators--Emerging examples within the TRIM family, Kimura [/bib_ref] [bib_ref] TRIM proteins and diseases, Watanabe [/bib_ref].
## Trim-mediate regulation of antiviral signaling
## Introduction to innate antiviral responses
Cells identify pathogen invasion due to the presence of pathogen-associated molecular patterns (PAMPs) contained in viral components, which are recognized by host pattern recognition receptors (PRRs) [bib_ref] Pattern recognition receptors and inflammation, Takeuchi [/bib_ref]. Examples of viral PAMPs include some envelope or capsid proteins, viral nucleic acid, or intermediates of genome replication [bib_ref] Pattern recognition receptors and inflammation, Takeuchi [/bib_ref]. Upon PAMP engagement of a PRR, a signaling cascade is initiated that relies on post-translational modifications for proper coordination. These modifications include ubiquitination and phosphorylation, which facilitate the assembly of adaptor and enzymatic molecules needed to activate and inactivate transcription factors and other effector molecules [bib_ref] Ubiquitin enzymes in the regulation of immune responses, Ebner [/bib_ref] [bib_ref] Pattern recognition receptors and inflammation, Takeuchi [/bib_ref] [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. These transitions in the transcriptional profile and functional proteome enable the cell to respond optimally to the pathogen and to communicate (e.g., cytokine secretion) with neighboring cells to limit viral replication and promote clearance. Examples of pathways critical in response to viral infection include IFN induction and signaling and NF-κB activation [bib_ref] Convergence of the NF-κb and IRF pathways in the regulation of the..., Hiscott [/bib_ref]. TRIM E3 ligases regulate IFN production and signaling as well as NF-κB induction at multiple levels, from PRR-mediated PAMP recognition to regulation of transcription factors and from promotion of signaling complex assembly to degradation of inhibitors [bib_ref] Ubiquitin enzymes in the regulation of immune responses, Ebner [/bib_ref] [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref]. In this section, we discuss TRIM regulation of antiviral pathways (summarized in [fig_ref] Figure 1: Regulation of pattern recognition receptor [/fig_ref].
## Trims and the retinoic acid-inducible gene i (rig-i)-like receptor pathway
Viral double-stranded RNA (dsRNA) or single-stranded RNA (ssRNA) containing 5'-triphophates produced during virus replication, in the cytoplasm of a host cell, act as a Retinoic Acid-Inducible Gene I RIG-I-like receptor (RLR) agonist [bib_ref] In vivo ligands of MDA5 and RIG-I in measles virus-infected cells, Runge [/bib_ref] [bib_ref] The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate..., Yoneyama [/bib_ref] [bib_ref] 5'-triphosphate rna is the ligand for RIG-I, Hornung [/bib_ref] [bib_ref] Reis e Sousa, C. Rig-i-mediated antiviral responses to single-stranded RNA bearing 5'-phosphates, Pichlmair [/bib_ref]. RIG-I and melanoma differentiation-associated protein (MDA5), encoded by Ddx58 and Ifih1, respectively, bind distinct viral RNA agonists yet they induce similar downstream antiviral pathways [bib_ref] Differential roles of MDA5 and RIG-I helicases in the recognition of RNA..., Kato [/bib_ref]. RLRs are ATP-dependent RNA helicases that have two N-terminal caspase-activated recruitment domains (CARDs), a central DEAD box, and an auto-inhibitory C-terminal domain [bib_ref] RIG-I-like receptor regulation in virus infection and immunity, Chan [/bib_ref]. The unique PAMPs recognized by these receptors enable the host to respond to a broader range of pathogens [bib_ref] Differential roles of MDA5 and RIG-I helicases in the recognition of RNA..., Kato [/bib_ref]. Upon engagement of the PAMP with the RLR, a conformational shift exposes the CARDs, which allows homo-oligomerization and recruitment of the RLRs to their adaptor mitochondrial antiviral signaling protein (MAVS) at the mitochondrial outer membrane (MOM) [bib_ref] The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate..., Yoneyama [/bib_ref] [bib_ref] Ubiquitin-induced oligomerization of the RNA sensors RIG-I and MDA5 activates antiviral innate..., Jiang [/bib_ref] [bib_ref] Structural basis for dsRNA recognition, filament formation, and antiviral signal activation by..., Wu [/bib_ref] [bib_ref] Identification and characterization of mavs, a mitochondrial antiviral signaling protein that activates..., Seth [/bib_ref]. A variety of factors influence the activation of RLRs downstream of PAMP recognition including ATP hydrolysis [bib_ref] The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate..., Yoneyama [/bib_ref] [bib_ref] RIG-I forms signaling-competent filaments in an ATP-dependent, ubiquitin-independent manner, Peisley [/bib_ref] , RLR oligomerization [bib_ref] Ubiquitin-induced oligomerization of the RNA sensors RIG-I and MDA5 activates antiviral innate..., Jiang [/bib_ref] [bib_ref] Structural basis for dsRNA recognition, filament formation, and antiviral signal activation by..., Wu [/bib_ref] [bib_ref] RIG-I forms signaling-competent filaments in an ATP-dependent, ubiquitin-independent manner, Peisley [/bib_ref] , and post-translational modifications [bib_ref] Phosphorylation-mediated negative regulation of RIG-I antiviral activity, Gack [/bib_ref] [bib_ref] The ubiquitin ligase riplet is essential for RIG-I-dependent innate immune responses to..., Oshiumi [/bib_ref] [bib_ref] Reconstitution of the RIG-I pathway reveals a signaling role of unanchored polyubiquitin..., Zeng [/bib_ref] [bib_ref] Negative regulation of the RIG-I signaling by the ubiquitin ligase RNF125, Arimoto [/bib_ref] [bib_ref] Ubch8 regulates ubiquitin and ISG15 conjugation to RIG-I, Arimoto [/bib_ref]. Interaction of the N-terminal CARDs of both the RLRs and MAVS induces the adaptor to form prion-like aggregates and exposes domains to recruit critical RING E3 ligases including tumor necrosis factor (TNF) receptor-associated factors (TRAFs) 3 and 6 [bib_ref] Mavs forms functional prion-like aggregates to activate and propagate antiviral innate immune..., Hou [/bib_ref].
Downstream of TRAF6, the UBD-containing adaptor TGF-β-activated kinase 1(TAK1)/Mitogen activating protein 3K7 (MAP3K7)-binding protein (TAB) 2/3 recruits the critical kinase TAK1 [bib_ref] Traf molecules in cell signaling and in human diseases, Xie [/bib_ref] [bib_ref] Activation of the iκb kinase complex by traf6 requires a dimeric ubiquitin-conjugating..., Deng [/bib_ref] [bib_ref] TAB2 and TAB3 activate the NF-κB pathway through binding to polyubiquitin chains, Kanayama [/bib_ref]. TAK1 auto-phosphorylates to enable the phosphorylation of NF-κB essential modulator (NEMO), the regulatory domain of inhibitor of NF-κB (IκB) kinase (IKK) complex, to activate the enzymatic domains of IKKα and β [bib_ref] Activation of the iκb kinase complex by traf6 requires a dimeric ubiquitin-conjugating..., Deng [/bib_ref] [bib_ref] Direct activation of protein kinases by unanchored polyubiquitin chains, Xia [/bib_ref]. IKKα and β then phosphorylate IκB, resulting in the recruitment of another E3 ligase, β-TrcP (β-Transducin Repeat Containing E3 Ubiquitin Protein Ligase). β-TrcP ligates K48-linked ubiquitin to IκB, inducing the proteasome-mediated degradation of the NF-κB inhibitor [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. Once the inhibitor is destroyed, NF-κB is phosphorylated and its nuclear localization sequence is exposed allowing nuclear translocation [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. Inside the nucleus, NF-κB regulates the transcription of a variety of genes including pro-inflammatory cytokines and chemokines, such as pro-IL-1β, TNF-α, and IL-6, and negative regulators of the pathway to limit an exacerbated inflammatory response.
Vaccines 2017, 5,5 of 36 another E3 ligase, β-TrcP (β-Transducin Repeat Containing E3 Ubiquitin Protein Ligase). β-TrcP ligates K48-linked ubiquitin to IκB, inducing the proteasome-mediated degradation of the NF-κB inhibitor [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. Once the inhibitor is destroyed, NF-κB is phosphorylated and its nuclear localization sequence is exposed allowing nuclear translocation [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. Inside the nucleus, NF-κB regulates the transcription of a variety of genes including pro-inflammatory cytokines and chemokines, such as pro-IL-1β, TNF-α, and IL-6, and negative regulators of the pathway to limit an exacerbated inflammatory response. TRIMs play an integral role in the positive and negative regulation of antiviral pathways. TRIMs can act as pathogen PRRs, as is the case for TRIM21 in the recognition of non-enveloped viruses bound by immunoglobulin (Ig). Additionally, these TRIMs can regulate the activation of other PRRs that recognize viral pathogen-associated molecular patterns (PAMPs) in the cytosol (DDX41 (DEAD-box helicase 41), cyclic GMP-AMP synthase (cGAS), DEAH-box helicase 33 (DHX33), nucleotide-binding oligomerization domain-containing protein 2 (NOD2), retinoic acid-inducible gene I (RIG-I), and melanoma differentiation-associated protein (MDA5)) and at membrane surfaces (toll-like receptors, TLRs). Downstream of the initial pattern recognition, TRIMs also influence the recruitment and interaction of adaptor molecules (stimulator of IFN genes (STING), mitochondrial antiviral signaling protein (MAVS), TGF-β-activated kinase 1(TAK1)/MAP3K7-binding protein (TAB) 2, Myeloid differentiation primary response gene 88 (MyD88), TIR-domain-containing adapterinducing interferon-β (TRIF), NF-κB essential modulator (NEMO), nucleosome assembly protein (NAP-1), and tumor necrosis factor (TNF) receptor-associated factors (TRAF) family memberassociated NF-κB activator (TANK)) and enzymes (TRAF3, TRAF6, TAK1, inhibitor of NF-κB (IκB) kinase (IKK) α,β,ε, TANK binding kinase 1 (TBK1)) to signaling complexes in order to activate transcription factors. This includes IFN regulatory factor (IRF)3 and IRF7, important in type-I interferon (IFN) signaling, and NF-κB, important in expression of pro-inflammatory genes, which regulate the expression of antiviral effectors. Type-I IFN production is critical for an effective antiviral response. TRIMs play an integral role in the positive and negative regulation of antiviral pathways. TRIMs can act as pathogen PRRs, as is the case for TRIM21 in the recognition of non-enveloped viruses bound by immunoglobulin (Ig). Additionally, these TRIMs can regulate the activation of other PRRs that recognize viral pathogen-associated molecular patterns (PAMPs) in the cytosol (DDX41 (DEAD-box helicase 41), cyclic GMP-AMP synthase (cGAS), DEAH-box helicase 33 (DHX33), nucleotide-binding oligomerization domain-containing protein 2 (NOD2), retinoic acid-inducible gene I (RIG-I), and melanoma differentiation-associated protein (MDA5)) and at membrane surfaces (toll-like receptors, TLRs). Downstream of the initial pattern recognition, TRIMs also influence the recruitment and interaction of adaptor molecules (stimulator of IFN genes (STING), mitochondrial antiviral signaling protein (MAVS), TGF-β-activated kinase 1(TAK1)/MAP3K7-binding protein (TAB) 2, Myeloid differentiation primary response gene 88 (MyD88), TIR-domain-containing adapter-inducing interferon-β (TRIF), NF-κB essential modulator (NEMO), nucleosome assembly protein (NAP-1), and tumor necrosis factor (TNF) receptor-associated factors (TRAF) family member-associated NF-κB activator (TANK)) and enzymes (TRAF3, TRAF6, TAK1, inhibitor of NF-κB (IκB) kinase (IKK) α,β,ε, TANK binding kinase 1 (TBK1)) to signaling complexes in order to activate transcription factors. This includes IFN regulatory factor (IRF)3 and IRF7, important in type-I interferon (IFN) signaling, and NF-κB, important in expression of pro-inflammatory genes, which regulate the expression of antiviral effectors. Type-I IFN production is critical for an effective antiviral response.
## Figure 2.
TRIMs in cytokine signaling. Downstream of the initial pathogen recognition and induction of pro-inflammatory cytokines, TRIMs can regulate their cytokine signaling pathways through interactions with cytokine receptor adaptors (TAB2/3) and enzymatic proteins (IKKα, IKKβ and IKKε) within the signaling complexes, the activity and stability of pathway negative regulators (Protein Inhibitor of Activated STAT 3 (PIAS3), suppressor of cytokine signaling (SOCS), and influence the transcription of various cytokine-effector genes (NF-κB-induced pro-inflammatory cytokines, signal transducer and activator of transcription (STAT)-induced genes, interferon stimulated genes (ISGs)) or cytokine signaling regulators (tumor necrosis factor (TNF) receptors (TNFR1/2, and STAT-induced genes)).
In addition to activation of NF-κB, RLR signaling induces type-I IFN [bib_ref] Identification and characterization of mavs, a mitochondrial antiviral signaling protein that activates..., Seth [/bib_ref]. TRAF3, in cooperation with NEMO, recruits and stabilizes TRAF family member-associated NF-κB activator (TANK) or nucleosome assembly protein (NAP1) which are critical in linking TANK binding kinase 1 (TBK1), and in some cases inhibitor of kappa light polypeptide gene enhancer in B cells (IKKε), to the MAVS signalosome [bib_ref] Traf molecules in cell signaling and in human diseases, Xie [/bib_ref]. Once activated, TBK1 and/or IKKε phosphorylate the IFN regulatory factor (IRF) 3 and IRF7 [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref] [bib_ref] Are the IKKs and IKK-related kinases TBK1 and IKK-epsilon similarly activated?, Chau [/bib_ref]. Upon phosphorylation, the IRFs homodimerize and translocate to the nucleus where they bind to DNA regulatory regions [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref] [bib_ref] Primary activation of interferon A and interferon B gene transcription by interferon..., Juang [/bib_ref]. To induce optimal IFN-β transcription, activated IRF3, NF-κB, and AP-1 (activator protein 1) must translocate to the nucleus and bind to their respective regulatory regions of the Ifnb1 promoter [bib_ref] Convergence of the NF-κb and IRF pathways in the regulation of the..., Hiscott [/bib_ref]. The resulting IFN-β is then secreted and signals in a paracrine and autocrine manner. Binding of IFN-β to its heterodimeric receptor results in the activation of tyrosine kinases, Janus kinase 1 (JAK1) and tyrosine kinase 2 (Tyk2), which phosphorylate signal transducer and activator of transcription (STAT) 1 and STAT2. Following phosphorylation, STAT1 and STAT2 heterodimerize and associate with IRF9 to form IFN-stimulated gene factor 3 (ISGF3) and translocate to the nucleus [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. Within the nucleus, ISGF3 binds to genes with an IFN stimulated response element (ISRE) in their promoter to activate transcription [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. The resulting proteins expressed from these ISGs, such as PKR (protein kinase R), MxA (myxovirus resistance gene A), ISG15, and TRIMs, are involved in creating a cellular environment prohibitive to viral entry and replication [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. As with other immune pathways, ISGF3 also promotes the transcription of type-I IFN negative regulators to mitigate deleterious effects [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref].
TRIMs play a critical role in both the positive and negative regulation of the RLR pathway to ensure optimal virus restriction while minimizing self-inflicted damage [fig_ref] Figure 1: Regulation of pattern recognition receptor [/fig_ref]. Several TRIMs have been shown to positively regulate the receptors RIG-I and MDA5 [bib_ref] TRIM65-catalized ubiquitination is essential for MDA5-mediated antiviral innate immunity, Lang [/bib_ref] [bib_ref] TRIM4 modulates type I interferon induction and cellular antiviral response by targeting..., Yan [/bib_ref] [bib_ref] TRIM25 ring-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity, Gack [/bib_ref]. The best characterized example of TRIM-mediated RIG-I activation involves TRIM25. TRIM25 ligates K63-linked poly-Ub chains onto the N-terminal CARD at K172, which induces downstream signaling [bib_ref] TRIM25 ring-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity, Gack [/bib_ref]. Additionally, TRIM25 catalyzes the synthesis of unanchored K63-linked poly-Ub chains, which facilitate RIG-I oligomerization and stabilization [bib_ref] Reconstitution of the RIG-I pathway reveals a signaling role of unanchored polyubiquitin..., Zeng [/bib_ref]. Both oligomerization and stabilization of RIG-I promotes the interaction of its CARDs with MAVS [bib_ref] Ubiquitin-mediated modulation of the cytoplasmic viral RNA sensor RIG-I, Oshiumi [/bib_ref]. Adding complexity to this interaction, TRIM25 K48-linked . TRIMs in cytokine signaling. Downstream of the initial pathogen recognition and induction of pro-inflammatory cytokines, TRIMs can regulate their cytokine signaling pathways through interactions with cytokine receptor adaptors (TAB2/3) and enzymatic proteins (IKKα, IKKβ and IKKε) within the signaling complexes, the activity and stability of pathway negative regulators (Protein Inhibitor of Activated STAT 3 (PIAS3), suppressor of cytokine signaling (SOCS), and influence the transcription of various cytokine-effector genes (NF-κB-induced pro-inflammatory cytokines, signal transducer and activator of transcription (STAT)-induced genes, interferon stimulated genes (ISGs)) or cytokine signaling regulators (tumor necrosis factor (TNF) receptors (TNFR1/2, and STAT-induced genes)).
In addition to activation of NF-κB, RLR signaling induces type-I IFN [bib_ref] Identification and characterization of mavs, a mitochondrial antiviral signaling protein that activates..., Seth [/bib_ref]. TRAF3, in cooperation with NEMO, recruits and stabilizes TRAF family member-associated NF-κB activator (TANK) or nucleosome assembly protein (NAP1) which are critical in linking TANK binding kinase 1 (TBK1), and in some cases inhibitor of kappa light polypeptide gene enhancer in B cells (IKKε), to the MAVS signalosome [bib_ref] Traf molecules in cell signaling and in human diseases, Xie [/bib_ref]. Once activated, TBK1 and/or IKKε phosphorylate the IFN regulatory factor (IRF) 3 and IRF7 [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref] [bib_ref] Are the IKKs and IKK-related kinases TBK1 and IKK-epsilon similarly activated?, Chau [/bib_ref]. Upon phosphorylation, the IRFs homodimerize and translocate to the nucleus where they bind to DNA regulatory regions [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref] [bib_ref] Primary activation of interferon A and interferon B gene transcription by interferon..., Juang [/bib_ref]. To induce optimal IFN-β transcription, activated IRF3, NF-κB, and AP-1 (activator protein 1) must translocate to the nucleus and bind to their respective regulatory regions of the Ifnb1 promoter [bib_ref] Convergence of the NF-κb and IRF pathways in the regulation of the..., Hiscott [/bib_ref]. The resulting IFN-β is then secreted and signals in a paracrine and autocrine manner. Binding of IFN-β to its heterodimeric receptor results in the activation of tyrosine kinases, Janus kinase 1 (JAK1) and tyrosine kinase 2 (Tyk2), which phosphorylate signal transducer and activator of transcription (STAT) 1 and STAT2. Following phosphorylation, STAT1 and STAT2 heterodimerize and associate with IRF9 to form IFN-stimulated gene factor 3 (ISGF3) and translocate to the nucleus [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. Within the nucleus, ISGF3 binds to genes with an IFN stimulated response element (ISRE) in their promoter to activate transcription [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. The resulting proteins expressed from these ISGs, such as PKR (protein kinase R), MxA (myxovirus resistance gene A), ISG15, and TRIMs, are involved in creating a cellular environment prohibitive to viral entry and replication [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref]. As with other immune pathways, ISGF3 also promotes the transcription of type-I IFN negative regulators to mitigate deleterious effects [bib_ref] Type I interferons in infectious disease, Mcnab [/bib_ref].
TRIMs play a critical role in both the positive and negative regulation of the RLR pathway to ensure optimal virus restriction while minimizing self-inflicted damage [fig_ref] Figure 1: Regulation of pattern recognition receptor [/fig_ref]. Several TRIMs have been shown to positively regulate the receptors RIG-I and MDA5 [bib_ref] TRIM65-catalized ubiquitination is essential for MDA5-mediated antiviral innate immunity, Lang [/bib_ref] [bib_ref] TRIM4 modulates type I interferon induction and cellular antiviral response by targeting..., Yan [/bib_ref] [bib_ref] TRIM25 ring-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity, Gack [/bib_ref]. The best characterized example of TRIM-mediated RIG-I activation involves TRIM25. TRIM25 ligates K63-linked poly-Ub chains onto the N-terminal CARD at K172, which induces downstream signaling [bib_ref] TRIM25 ring-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity, Gack [/bib_ref]. Additionally, TRIM25 catalyzes the synthesis of unanchored K63-linked poly-Ub chains, which facilitate RIG-I oligomerization and stabilization [bib_ref] Reconstitution of the RIG-I pathway reveals a signaling role of unanchored polyubiquitin..., Zeng [/bib_ref]. Both oligomerization and stabilization of RIG-I promotes the interaction of its CARDs with MAVS [bib_ref] Ubiquitin-mediated modulation of the cytoplasmic viral RNA sensor RIG-I, Oshiumi [/bib_ref]. Adding complexity to this interaction, TRIM25 K48-linked polyubiquitination negatively regulates RLR activation, but the ubiquitin specific protease 15 (USP15) can specifically disassemble these poly-Ub chains to stabilize TRIM25 [bib_ref] The ubiquitin-specific protease usp15 promotes RIG-I-mediated antiviral signaling by deubiquitylating TRIM25, Pauli [/bib_ref]. TRIMs 4, 13, and 38 have also been implicated in positive regulation of the RIG-I pathway. Similar to TRIM25, TRIM4 also catalyzes the ligation of K63-linked poly-Ub chains onto RIG-I CARD [bib_ref] TRIM4 modulates type I interferon induction and cellular antiviral response by targeting..., Yan [/bib_ref]. Additionally, TRIM38 functions as an E3 SUMO (small ubiquitin-like modifier) ligase and SUMOylates both RIG-I and MDA5 to prevent the ligation of K48-linked poly-Ub chains thus stabilizing these PRRs [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Multifaceted roles of TRIM38 in innate immune and inflammatory responses, Hu [/bib_ref] [bib_ref] Innate immunity to RNA virus is regulated by temporal and reversible sumoylation..., Hu [/bib_ref]. The capacity of multiple TRIMs to activate RIG-I suggests that ubiquitination is crucial in RIG-I signaling, but the relative contribution of each TRIM is not well understood. Perhaps multiple TRIMs allow for redundancy in the instance that one TRIM is inhibited or if TRIMs play cell-type specific roles in RLR signaling.
Recently TRIM65 was identified as an E3 ligase of MDA5. Unlike TRIM25, TRIM65 ubiquitinates MDA5 at the RNA helicase domain [bib_ref] TRIM65-catalized ubiquitination is essential for MDA5-mediated antiviral innate immunity, Lang [/bib_ref]. The covalent linkage of K63-linked poly-Ub onto K743 promotes MDA5 oligomerization and downstream activation of IRF3 [bib_ref] TRIM65-catalized ubiquitination is essential for MDA5-mediated antiviral innate immunity, Lang [/bib_ref]. Demonstrating the specificity of MDA5 activation, TRIM65 only promotes the restriction of encephalomyocarditis virus (EMCV), a picornavirus, and not vesicular stomatitis virus (VSV), a rhabdovirus [bib_ref] TRIM65-catalized ubiquitination is essential for MDA5-mediated antiviral innate immunity, Lang [/bib_ref]. In mouse cells, TRIM13 was shown to impair MDA5-mediated activation of the IFN pathway through an unclarified mechanism [bib_ref] TRIM13 is a negative regulator of MDA5-mediated type I interferon production, Narayan [/bib_ref]. Another TRIM inhibitor of the MAVS pathway is TRIM59, which interacts with evolutionarily conserved signaling intermediate in Toll pathways (ECSIT) and MAVS and subsequently inhibits the transcription of IRF3 and NF-κB target genes [bib_ref] TRIM59 interacts with ecsit and negatively regulates NF-κB and IRF-3/7-mediated signal pathways, Kondo [/bib_ref]. Although TRIMs have not been identified as RIG-I negative regulators, their role in MDA5 inhibition suggests there may be unidentified TRIM-mediated RIG-I inhibition.
The role of TRIM25 in the regulation of RLR pathways and/or type-I IFN induction has been shown to be conserved among different species. In fact a diverse range of vertebrates encode RIG-regulating TRIMs. In salmonids, TRIM25, MAVS, MDA5, and RIG-I were induced following infection with an alphavirus although the signaling pathways were not addressed directly [bib_ref] De novo transcriptome analysis shows that sav-3 infection upregulates pattern recognition receptors..., Xu [/bib_ref]. Duck TRIM25 catalyzes the synthesis of unanchored poly-Ub chains to activate RIG-I [bib_ref] Activation of duck rig-i by TRIM25 is independent of anchored ubiquitin, Miranzo-Navarro [/bib_ref]. Despite lacking lysine 172 in duck RIG-I, duck TRIM25 ubiquitinates RIG-I's CARD domains and promotes RLR signaling [bib_ref] Activation of duck rig-i by TRIM25 is independent of anchored ubiquitin, Miranzo-Navarro [/bib_ref]. In chicken cells, despite lacking a functional RIG-I gene, knockdown of chicken TRIM25 results in reduced IFN-β upon infection with specific strains of the influenza A virus (IAV) [bib_ref] Species-specific inhibition of RIG-I ubiquitination and IFN induction by the influenza a..., Rajsbaum [/bib_ref] , suggesting that TRIM25 is involved in activation of IFN signaling through a RIG-I-independent mechanism, perhaps activation of MDA5 or MAVS. Expression of chicken TRIM25 is induced after Newcastle disease virus (NDV), poly(I:C) treatment, or poly(dA:dT) treatment [bib_ref] Molecular characterization, tissue distribution and expression analysis of TRIM25 in gallus gallus..., Feng [/bib_ref] , probably via a type-I IFN signaling-dependent pathway. Similar to human TRIM25, TRIM27-L stimulates IFN-β production in response to IAV infection in ducks [bib_ref] Duck TRIM27-l enhances mavs signaling and is absent in chickens and turkeys, Blaine [/bib_ref]. This anti-viral benefit is absent in chickens and turkeys, as they do not carry the TRIM27-L gene in their TRIM cluster [bib_ref] Duck TRIM27-l enhances mavs signaling and is absent in chickens and turkeys, Blaine [/bib_ref]. However, expression of duck TRIM27-L and d2CARD in chicken DF1 cell lines was shown to facilitate IFN-β and MX1 expression [bib_ref] Duck TRIM27-l enhances mavs signaling and is absent in chickens and turkeys, Blaine [/bib_ref]. This difference may account for the different pathologies in avian species as waterfowl are typically more resistant to some strains of IAV as compared to chickens.
Downstream of RLR activation, a variety of TRIMs promote MAVS signaling. TRIM25 has been implicated in the K48-linked polyubiquitination of MAVS, which results in its proteasome-mediated degradation and release of downstream signaling molecules (TBK1, NEMO, and possibly TRAF3) to induce type-I IFN production [bib_ref] Mavs ubiquitination by the E3 ligase TRIM25 and degradation by the proteasome..., Castanier [/bib_ref]. Recently, TRIM31 has been described to mediate the K63-linked ubiquitination of MAVS at lysines 10, 311, and 461 [bib_ref] The ubiquitin E3 ligase TRIM31 promotes aggregation and activation of the signaling..., Liu [/bib_ref]. This ubiquitination promotes the prion-like aggregation of MAVS needed for optimal signaling, exemplified by the decrease in TBK1 and IKKε phosphorylation [bib_ref] The ubiquitin E3 ligase TRIM31 promotes aggregation and activation of the signaling..., Liu [/bib_ref]. In a TRIM31-deficient murine model, TRIM31 knock-out mice demonstrated an increased susceptibility to VSV infection and an upregulation in IFN-β production. Although RIG-I was demonstrated to be required for activation of TRIM31 function, its role in MDA5-mediated signaling was not investigated in-depth [bib_ref] The ubiquitin E3 ligase TRIM31 promotes aggregation and activation of the signaling..., Liu [/bib_ref]. TRIM14 has been demonstrated to play a crucial role in linking the NF-κB and IRF3 branches of RLR signaling [bib_ref] TRIM14 is a mitochondrial adaptor that facilitates retinoic acid-inducible gene-i-like receptor-mediated innate..., Zhou [/bib_ref]. Despite lacking a RING domain, TRIM14 interacts with NEMO via its PRY-SPRY domain and promotes K63-linked ubiquitination of NEMO, which is critical in recruiting NEMO to MAVS [bib_ref] TRIM14 is a mitochondrial adaptor that facilitates retinoic acid-inducible gene-i-like receptor-mediated innate..., Zhou [/bib_ref]. The role of TRIM44 stabilization of MAVS has also been characterized [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Novel function of TRIM44 promotes an antiviral response by stabilizing visa, Yang [/bib_ref]. Although no TRIM inhibitors of MAVS have been described, screens of TRIMs that inhibit MAVS-mediated type-I IFN production may reveal such TRIMs.
In addition to regulating RLRs and MAVS, TRIMs modulate downstream adaptors and enzymes. Several of these TRIMs can likewise mitigate signaling downstream of other PRRs that converge on shared molecules such as NEMO and TAK1, but TRIMs identified to be involved at the level of signaling using RLR induction are described below. Recently, the short isoform of TRIM9 (TRIM9s) was shown to bridge GSK3β ( Glycogen synthase kinase 3 beta) to phosphorylated TBK1 to promote TBK1 oligomerization and activation of IRF3 [bib_ref] TRIM9 short isoform preferentially promotes DNA and RNA virus-induced production of type..., Qin [/bib_ref]. To facilitate the interaction between GSKβ and pTBK1, TRIM9s must be auto-ubiquitinated [bib_ref] TRIM9 short isoform preferentially promotes DNA and RNA virus-induced production of type..., Qin [/bib_ref]. This activation of TBK1-signaling occurs downstream of RLR and STING signaling [bib_ref] Crosstalk between cytoplasmic RIG-I and sting sensing pathways, Zevini [/bib_ref] and biases the immune response toward the type-I IFN pathway while limiting NF-κB-induced transcription [bib_ref] TRIM9 short isoform preferentially promotes DNA and RNA virus-induced production of type..., Qin [/bib_ref]. TRIM11's coiled-coil domain interacts with the coiled-coil domain 2 of TBK1 to prohibit the kinase's interaction with adaptors NAP1 or TANK [bib_ref] TRIM11 negatively regulates ifnβ production and antiviral activity by targeting TBK1, Lee [/bib_ref]. Impairing this interaction results in lack of IFN-β production [bib_ref] TRIM11 negatively regulates ifnβ production and antiviral activity by targeting TBK1, Lee [/bib_ref]. NAP1 is targeted for degradation following TRIM38-mediated K48-linked poly-Ub, which likewise decreases activation of IFN-β [bib_ref] Tripartite motif-containing protein 38 negatively regulates TLR3/4-and RIG-I-mediated IFN-β production and antiviral..., Zhao [/bib_ref]. Interaction between the ARF (ADP-ribosylation factor domain) (C-terminus) domain of TRIM23, and both the coiled-coil 1 and LZ domains of NEMO, allow TRIM23 to facilitate K27-linked ubiquitination of NEMO [bib_ref] Polyubiquitin conjugation to nemo by triparite motif protein 23 (TRIM23) is critical..., Arimoto [/bib_ref]. This ubiquitination of NEMO facilitates the activation of IRF3 and NF-κB signaling downstream of pathogen recognition, but not TNF-α signaling [bib_ref] Polyubiquitin conjugation to nemo by triparite motif protein 23 (TRIM23) is critical..., Arimoto [/bib_ref]. Similar to the described TRIM9s mechanism, TRIM26 is able to interact with TBK1 and auto-phosphorylates K27-linked poly-Ub chains to bridge TBK1 and NEMO [bib_ref] Autoubiquitination of TRIM26 links TBK1 to nemo in RLR-mediated innate antiviral immune..., Ran [/bib_ref]. This interaction was demonstrated downstream of MAVS signaling and promoted IRF3 activation [bib_ref] Autoubiquitination of TRIM26 links TBK1 to nemo in RLR-mediated innate antiviral immune..., Ran [/bib_ref]. Indirectly, TRIM68 antagonizes IFN-β transcription [bib_ref] TRIM68 negatively regulates IFN-β production by degrading TRK fused gene, a novel..., Wynne [/bib_ref]. This TRIM is able to inhibit both toll-like receptor (TLR) and RLR-driven activation of the type-I IFN pathway [bib_ref] TRIM68 negatively regulates IFN-β production by degrading TRK fused gene, a novel..., Wynne [/bib_ref].
Several TRIMs are likewise implicated in the inhibition of the NF-κB activation branch of PRR signaling. Murine specific TRIM30α negatively regulates the TAB2/TAB3 complex, which impedes the recruitment and activation of TAK1, thus preventing the phosphorylation of NEMO and subsequent NF-κB activation [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Trim30 α negatively regulates TLR-mediated NF-kappa B activation by targeting TAB2 and..., Shi [/bib_ref]. Through a different mechanism in the brain, the long isoform of TRIM9 (TRIM9L) inhibits β-TrcP [bib_ref] Negative regulation of NF-κB activity by brain-specific tripartite motif protein 9, Shi [/bib_ref]. Perturbation of β-TrcP inhibits both canonical and non-canonical NF-κB activation [bib_ref] Negative regulation of NF-κB activity by brain-specific tripartite motif protein 9, Shi [/bib_ref]. The TRIM9L protein sequence includes a degron motif that, when phosphorylated at serine residues 76 and 80, recruits β-TrcP [bib_ref] Negative regulation of NF-κB activity by brain-specific tripartite motif protein 9, Shi [/bib_ref]. Titrating β-TrcP from the NF-κB inhibitors prevents their degradation and subsequent pro-inflammatory signaling [bib_ref] Negative regulation of NF-κB activity by brain-specific tripartite motif protein 9, Shi [/bib_ref]. Several other TRIMs also inhibit NF-κB activation including TRIM11 [bib_ref] The human antiviral factor TRIM11 is under the regulation of HIV-1 VPR, Yuan [/bib_ref] , TRIM29 [bib_ref] Identification of a role for TRIM29 in the control of innate immunity..., Xing [/bib_ref] , and TRIM39 [bib_ref] TRIM39 negatively regulates the NFκB-mediated signaling pathway through stabilization of cactin, Suzuki [/bib_ref]. The mechanisms for TRIM11 regulation have not been clearly elucidated, but TRIM29 targets NEMO for degradation in alveolar macrophages [bib_ref] Identification of a role for TRIM29 in the control of innate immunity..., Xing [/bib_ref] and TRIM39 stabilizes Cactin [bib_ref] TRIM39 negatively regulates the NFκB-mediated signaling pathway through stabilization of cactin, Suzuki [/bib_ref] , a nuclear, negative regulator of NF-κB. The bias of TRIMs in the negative regulation of the NF-κB branch of RLR signaling suggests that TRIMs may play a role in promoting the type-I IFN pathway at the cost of NF-κB activation, and further supports the hypothesis that groups of TRIMs may have evolved as part of the antiviral type-I IFN system [bib_ref] Type I interferon-dependent and -independent expression of tripartite motif proteins in immune..., Rajsbaum [/bib_ref]. However, it is important to note that some studies showing negative regulatory roles of TRIMs have only used overexpression assays with large concentrations of TRIM expressing vectors, which could lead to artifacts. However, it is now clear that overall TRIMs act at several levels to regulate RLR signaling to balance viral clearance and cell survival.
## Trims and sting signaling
In addition to regulating cytosolic RNA-stimulated responses, TRIMs also regulate the pathways following cytosolic DNA recognition. In the cytoplasm, the host expresses multiple DNA and RNA receptors aside from RLRs, including IFI16 (Interferon Gamma Inducible Protein 16), cyclic GMP-AMP synthase (cGAS), and DDX41. The listed double-strand (ds) DNA receptors activate the adaptor molecule stimulator of IFN genes (STING) at the endoplasmic reticulum (ER) [bib_ref] Sting is an endoplasmic reticulum adaptor that facilitates innate immune signalling, Ishikawa [/bib_ref] [bib_ref] Sting regulates intracellular DNA-mediated, type I interferon-dependent innate immunity, Ishikawa [/bib_ref] [bib_ref] Sting is a direct innate immune sensor of cyclic di-gmp, Burdette [/bib_ref] [bib_ref] Structure of sting bound to cyclic di-gmp reveals the mechanism of cyclic..., Shu [/bib_ref] [bib_ref] Structural and functional analysis of ddx41: A bispecific immune receptor for DNA..., Omura [/bib_ref]. Upon recognition of dsDNA, which can result from infection with DNA viruses, cGAS oligomerizes and catalyzes cyclic dinucleotide (c-GMP-AMP) synthesis [bib_ref] Multifaceted roles of TRIM38 in innate immune and inflammatory responses, Hu [/bib_ref]. c-GMP-AMP and other dinucleotides activate STING and induce STING dimerization [bib_ref] Multifaceted roles of TRIM38 in innate immune and inflammatory responses, Hu [/bib_ref] [bib_ref] Sting is a direct innate immune sensor of cyclic di-gmp, Burdette [/bib_ref] [bib_ref] Structure of sting bound to cyclic di-gmp reveals the mechanism of cyclic..., Shu [/bib_ref]. Consequently, STING recruits TBK1, which phosphorylates IRF3 for type-I IFN induction [bib_ref] Sting is an endoplasmic reticulum adaptor that facilitates innate immune signalling, Ishikawa [/bib_ref]. The DNA helicase DDX41 recognizes both cytoplasmic dsDNA and cyclic dinucleotides, both of which promote DDX41 activation of STING [bib_ref] Structural and functional analysis of ddx41: A bispecific immune receptor for DNA..., Omura [/bib_ref] [bib_ref] The helicase ddx41 recognizes the bacterial secondary messengers cyclic di-gmp and cyclic..., Parvatiyar [/bib_ref].
Five TRIMs are known to influence STING-mediated signaling [fig_ref] Figure 1: Regulation of pattern recognition receptor [/fig_ref]. The PRY-SPRY domain of TRIM21 interacts with DDX41's helicase domain and catalyzes ubiquitination at lysine residues 9 and 115, targeting the PRR for degradation [bib_ref] The E3 ubiquitin ligase TRIM21 negatively regulates the innate immune response to..., Zhang [/bib_ref]. This degradation pathway occurs in myeloid dendritic cells and restricts type-I IFN induction [bib_ref] The E3 ubiquitin ligase TRIM21 negatively regulates the innate immune response to..., Zhang [/bib_ref]. The murine-specific TRIM, TRIM30α, ubiquitinates STING following herpes simplex virus 1 (HSV-1) and targets the adaptor protein for degradation [bib_ref] TRIM30α is a negative-feedback regulator of the intracellular DNA and DNA virus-triggered..., Wang [/bib_ref]. Both TRIM32 and TRIM56 facilitate K63-linked polyubiquitination of STING to promote dimerization and activation of STING-mediated antiviral responses [bib_ref] The ubiquitin ligase TRIM56 regulates innate immune responses to intracellular double-stranded DNA, Tsuchida [/bib_ref] [bib_ref] TRIM32 protein modulates type I interferon induction and cellular antiviral response by..., Zhang [/bib_ref]. Finally, TRIM38 SUMOylates cGAS and STING similar to the SUMOylation of RIG-I and MDA5 [bib_ref] Multifaceted roles of TRIM38 in innate immune and inflammatory responses, Hu [/bib_ref]. SUMOylation inhibits the ligation of K48-linked ubiquitin and results in their stabilization [bib_ref] Multifaceted roles of TRIM38 in innate immune and inflammatory responses, Hu [/bib_ref].
## Trims and tlr signaling
In addition to regulating cytosolic PRR signaling, TRIMs also modulate membrane-bound PRRs including toll-like receptors (TLRs) [fig_ref] Figure 1: Regulation of pattern recognition receptor [/fig_ref]. Several TLR family members recognize viral PAMPs. The main TLRs involved in virus recognition include endosome-localized TLRs 3, 7, 8, and 9 [bib_ref] The critical role of toll-like receptors-from microbial recognition to autoimmunity: A comprehensive..., Jimenez-Dalmaroni [/bib_ref]. TLR3 recognizes both double-stranded RNA and the viral RNA mimic poly(I:C), TLR7 and TLR8 recognize single-stranded RNA, and TLR9 recognizes CpG [bib_ref] The critical role of toll-like receptors-from microbial recognition to autoimmunity: A comprehensive..., Jimenez-Dalmaroni [/bib_ref]. TLR4 is associated mainly with the plasma membrane, although it can also be internalized in endosomes, and can recognize some viral surface antigens. TLRs 4, 7, 8, and 9 signal through IRAKs (Interleukin-1 receptor-associated kinase) 1 and 4, which are recruited via the adaptor molecule MyD88 (Myeloid differentiation primary response gene 88) [bib_ref] The critical role of toll-like receptors-from microbial recognition to autoimmunity: A comprehensive..., Jimenez-Dalmaroni [/bib_ref]. TRAF6 then re-localizes to the TLR signaling complex to activate TAK1, inducing NF-κB [bib_ref] Traf molecules in cell signaling and in human diseases, Xie [/bib_ref] [bib_ref] The critical role of toll-like receptors-from microbial recognition to autoimmunity: A comprehensive..., Jimenez-Dalmaroni [/bib_ref] similar to the pathway described in the aforementioned RLR section. TLR3 and 4 also signal through IRAKs 1 and 4, and interact with the adaptor TRIF (TIR-domain-containing adapter-inducing interferon-β), resulting in the activation of TRAF3. Ubiquitination downstream of TRAF3 induces TBK1-and IKKε-mediated phosphorylation of IRF3 and IRF7 to induce type-I IFN [bib_ref] The critical role of toll-like receptors-from microbial recognition to autoimmunity: A comprehensive..., Jimenez-Dalmaroni [/bib_ref]. TRIM21's PRY-SPRY domain interacts with IRFs, including IRF3, 5, and 7, to induce their degradation [bib_ref] The E3 ubiquitin ligase ro52 negatively regulates IFN-β production post-pathogen recognition by..., Higgs [/bib_ref] [bib_ref] Self protection from anti-viral responses-ro52 promotes degradation of the transcription factor IRF7..., Higgs [/bib_ref] [bib_ref] Tripartite motif 21 (TRIM21) differentially regulates the stability of interferon regulatory factor..., Lazzari [/bib_ref]. Although TRIM21 may promote degradation of IRFs downstream of other PRRs, the interaction between the two proteins may depend on the induction of a specific TLR pathway. TRIM38 targets TLR signaling at multiple points. Downstream of TLR2, 3, 4, or 7, TRIM38 targets TRAF6 for proteasome-mediated degradation following K48-linked polyubiquitination [bib_ref] E3 ubiquitin ligase tripartite motif 38 negatively regulates TLR-mediated immune responses by..., Zhao [/bib_ref]. Downstream of TLR3 and TLR4 signaling, TRIM38 also targets TRIF and NAP1 for degradation [bib_ref] Tripartite motif-containing protein 38 negatively regulates TLR3/4-and RIG-I-mediated IFN-β production and antiviral..., Zhao [/bib_ref] [bib_ref] Hung, T. TRIM38 negatively regulates TLR3-mediated IFN-β signaling by targeting trif for..., Xue [/bib_ref]. Finally, TRIM56 has been shown to promote TLR3 activation via interaction with TRIF in a RING ligase-independent manner [bib_ref] TRIM56 is an essential component of the TLR3 antiviral signaling pathway, Shen [/bib_ref]. Perhaps this interaction promotes the stability of TRIF to facilitate downstream signaling. Disruption of adaptor protein availability thus bottlenecks antiviral signaling.
## Trims and the nucleotide-binding domain and leucine-rich repeat-containing receptors (nlr) pathway
Nucleotide-binding domain and leucine-rich repeat-containing receptors (NLRs) are another class of cytosolic receptors that recognize both PAMPs and damage-associated molecular patterns (DAMPs). DAMPs are host-derived molecules that are expressed only after a cell experiences stress and/or damage commonly due to inflammation [bib_ref] Mechanisms and functions of inflammasomes, Lamkanfi [/bib_ref] [bib_ref] Activation and regulation of the inflammasomes, Latz [/bib_ref]. Upon activation of a NLR, the receptor assembles an inflammasome in cooperation with the adaptor ASC to recruit pro-caspases [bib_ref] Mechanisms and functions of inflammasomes, Lamkanfi [/bib_ref] [bib_ref] Activation and regulation of the inflammasomes, Latz [/bib_ref]. The NLRP3 (NLR Family Pyrin Domain Containing 3) -induced inflammasome promotes the cleavage of pro-caspase 1 to caspase 1, which then promotes a pro-apoptotic response involving the caspase-1-mediated cleavage of pro-IL-1β and pro-IL-18 to their active forms IL-1β and IL-18 [bib_ref] Mechanisms and functions of inflammasomes, Lamkanfi [/bib_ref] [bib_ref] Activation and regulation of the inflammasomes, Latz [/bib_ref]. The secreted pro-inflammatory cytokines then promote further inflammatory responses, such as pyroptosis, which may be damaging to the host when uncontrolled [bib_ref] Mechanisms and functions of inflammasomes, Lamkanfi [/bib_ref] [bib_ref] Activation and regulation of the inflammasomes, Latz [/bib_ref]. In some instances another NLR, NOD2, may function as a cytosolic dsRNA receptor and converge with the RLR pathway at the level of MAVS [bib_ref] Activation of nucleotide oligomerization domain 2 (NOD2) by human cytomegalovirus initiates innate..., Kapoor [/bib_ref] [bib_ref] NOD1 and NOD2 signaling in infection and inflammation, Moreira [/bib_ref].
At this point only a few TRIMs have been described to interact with NLRs in the control of viral infections. TRIM33 binds to and ubiquitinates DHX33, a cytosolic dsRNA receptor that acts upstream of NLRP3, at lysine K218 to facilitate inflammasome activation [bib_ref] The E3 ubiquitin ligase tripartite motif 33 is essential for cytosolic RNA-induced..., Weng [/bib_ref]. The knockdown of TRIM33 diminishes the activation of caspase 1 and likewise decreases the release of IL-1β and IL-18 [bib_ref] The E3 ubiquitin ligase tripartite motif 33 is essential for cytosolic RNA-induced..., Weng [/bib_ref]. The interaction of TRIM33 with DHX33 relies on the B-box and coiled-coil domains [bib_ref] The E3 ubiquitin ligase tripartite motif 33 is essential for cytosolic RNA-induced..., Weng [/bib_ref]. In contrast, the murine-specific TRIM30α impairs the NLRP3 inflammasome through an unknown mechanism [bib_ref] Tripartite-motif protein 30 negatively regulates NLRP3 inflammasome activation by modulating reactive oxygen..., Hu [/bib_ref]. Another NLRP3 inhibitor identified using a dextran sodium sulfate-induced colitis model is TRIM31 [bib_ref] The E3 ubiquitin ligase TRIM31 attenuates NLRP3 inflammasome activation by promoting proteasomal..., Song [/bib_ref]. The coiled-coil domain of TRIM31 interacts with the leucine rich and NACHT domains of NLRP3 and ligates K48-linked poly-Ub chains to target NLRP3 for proteasome-mediated degradation [bib_ref] The E3 ubiquitin ligase TRIM31 attenuates NLRP3 inflammasome activation by promoting proteasomal..., Song [/bib_ref]. Although the authors did not evaluate virus infection directly, Trim31-deficient cells stimulated with poly(I:C) express higher levels of NLRP3 compared to wild-type cells suggesting TRIM31 may regulate NLRP3 downstream of virus recognition [bib_ref] The E3 ubiquitin ligase TRIM31 attenuates NLRP3 inflammasome activation by promoting proteasomal..., Song [/bib_ref]. TRIM27 is able to promote degradation of NOD2 [bib_ref] TRIM27 negatively regulates NOD2 by ubiquitination and proteasomal degradation, Zurek [/bib_ref] , which may prohibit this receptor from recognizing DNA and RNA virus infection [bib_ref] Activation of nucleotide oligomerization domain 2 (NOD2) by human cytomegalovirus initiates innate..., Kapoor [/bib_ref] [bib_ref] NOD1 and NOD2 signaling in infection and inflammation, Moreira [/bib_ref]. The further investigation of TRIM interactions with NLRs is important for understanding the immune response during bacteria-virus co-infections.
## Trims and cytokine signaling
Several TRIMs are also involved in regulating the cytokine signaling following initial pathogen recognition. Downstream of TNF-α engagement with its receptor TNFR1, the TAK1 complex is recruited to activate NF-κB [bib_ref] Expanding role of ubiquitination in NF-κb signaling, Liu [/bib_ref]. TRIM8 specifically promotes the activation of TNF-α-induced NF-κB signaling through inhibition of the NF-κB nuclear repressor protein inhibitor of activated STAT (PIAS) 3 [bib_ref] Nucleo-cytoplasmic trafficking of TRIM8, a novel oncogene, is involved in positive regulation..., Tomar [/bib_ref]. The TRIM8-mediated repression of PIAS likewise promotes IL-6-dependent activation of STAT3 [bib_ref] TRIM8 modulates STAT3 activity through negative regulation of PIAS3, Okumura [/bib_ref]. Specifically downstream of IL-1β and TNF-α signaling in IFN-β-primed cells, TRIM38 targets TAB2 for lysosomal degradation [bib_ref] TRIM38 negatively regulates TLR3/4-mediated innate immune and inflammatory responses by two sequential..., Hu [/bib_ref] [bib_ref] TRIM38 inhibits TNFα-and IL-1β-triggered NF-κB activation by mediating lysosome-dependent degradation of TAB2/3, Hu [/bib_ref]. As described above, the degradation of TAB2 inhibits the recruitment and activation of TAK1 which blocks pro-inflammatory signaling. Zheng and colleagues showed that TRIM27 catalyzes K48-linked poly-Ub at residues K251 and K372 of TBK1 to promote proteasome-mediated degradation downstream of IFN signaling [bib_ref] Siglec1 suppresses antiviral innate immune response by inducing TBK1 degradation via the..., Zheng [/bib_ref]. TRIM27-TBK1 interactions require the coiled-coil and B-box TRIM domains [bib_ref] Siglec1 suppresses antiviral innate immune response by inducing TBK1 degradation via the..., Zheng [/bib_ref]. In endothelial cells, TRIM28 sustains the expression of TNFR1 and TNFR2 to promote the activation of pro-inflammatory pathways. The role of TRIM28 in the activation of the endothelium may play an important, unexplored role in immune cell trafficking in response to infection.
In response to cytokines (i.e., IL-12) secreted from activated dendritic cells (DCs), IFN-γ (type-II IFNs) is released from T cells and natural killer cells [bib_ref] Interferons and viral infections, Fensterl [/bib_ref]. After this, type-II IFN binds its receptor, and the downstream kinases JAK1 and JAK2 promote STAT1 phosphorylation and homodimerization to form gamma activating factor, which translocates to the nucleus to bind gamma-stimulated elements in gene promoters [bib_ref] Interferons and viral infections, Fensterl [/bib_ref]. TRIM24 inhibits STAT1 transcription via binding to the STAT1 promoter [bib_ref] Tripartite motif 24 (TRIM24/TIF1α) tumor suppressor protein is a novel negative regulator..., Tisserand [/bib_ref]. In contrast, TRIM8 is able to destabilize SOCS-1 (Suppressor Of Cytokine Signaling 1), a negative regulator of the IFN-γ signaling pathway, resulting in increased type-II IFN signaling [bib_ref] TRIM8/gerp ring finger protein interacts with socs-1, Toniato [/bib_ref]. The regulation downstream of IFN signaling suggests that this is a negative regulatory mechanism to prevent inflammatory response overactivation. In response to type-I IFNs, TRIM6 catalyzes the formation of K48-linked unanchored poly-Ub chains to facilitate the activation of IKKε, which favors ISGF3 formation due to STAT1 phosphorylation at S708 [bib_ref] Unanchored k48-linked polyubiquitin synthesized by the e3-ubiquitin ligase TRIM6 stimulates the interferon-ikkepsilon..., Rajsbaum [/bib_ref]. This modification increases STAT1-STAT2 dimerization [bib_ref] Multiple functions of the IKK-related kinase ikkepsilon in interferon-mediated antiviral immunity, Tenoever [/bib_ref] [bib_ref] Iκb kinase ε (IKKε) regulates the balance between type I and type..., Ng [/bib_ref]. This TRIM6-enahnced activation of IKKε may also play a role in activating IFN-β transcription and translation downstream of RLR activation [bib_ref] Unanchored k48-linked polyubiquitin synthesized by the e3-ubiquitin ligase TRIM6 stimulates the interferon-ikkepsilon..., Rajsbaum [/bib_ref]. TRIM22 induces the lysosome-mediated destruction of FOXO4 and consequently impairs the transcription of IFN-β downstream of TLR3 and RLR signaling [bib_ref] Control of foxo4 activity and cell survival by TRIM22 directs TLR3-stimulated cells..., Oteiza [/bib_ref].
## Trims and adaptive immunity
After the innate response to pathogen invasion, the host evolves an adaptive immune response. Establishment of an adaptive response requires the presentation of pathogen antigens to T and B lymphocytes. Although TRIM proteins have been mostly studied as regulators of innate immune responses and the role of TRIMs in the regulation of antigen presentation has not been investigated intensively, several studies indicate that TRIMs play an important role in regulating T cell activation. Expression of TRIM22, for example, has been shown to be down-regulated upon CD28/CD2-mediated activation despite being expressed at high levels in resting T cells [bib_ref] The interferon-inducible STAF50 gene is downregulated during T cell costimulation by CD2..., Gongora [/bib_ref]. In contrast, TRIM22 expression in T cells is increased following IL-2 and IL-15 cytokine signaling [bib_ref] Regulation of the interferon-inducible p53 target gene TRIM22 (STAF50) in human T..., Obad [/bib_ref] , both of which act as pro-survival signals. Recently, TRIM24 was shown to influence the expression of Th2-type cytokines but not other CD4 + T cell sub-types [bib_ref] T-cell-intrinsic TIF1α/TRIM24 regulates IL-1r expression on Th2 cells and Th2 cell-mediated airway..., Perez-Lloret [/bib_ref]. As Th2 cells predominantly act in allergy-and parasite-induced responses while Th1 cells are more important for viral clearance, deregulation of TRIM24 expression may impact the CD4 + population composition and the capacity of the host to efficiently clear the virus. In the past, TRIM27 was described as impairing the activation of CD4 + T cells via K48-linked poly-Ub of PI3KC2B (Phosphatidylinositol 4-phosphate 3-kinase C2 domain-containing subunit beta) [bib_ref] Tripartite motif containing protein 27 negatively regulates CD4 T cells by ubiquitinating..., Cai [/bib_ref]. This impairment was observed specifically in CD4 + T cells downstream of T cell receptor (TCR) engagement and not CD8 + T cells [bib_ref] Tripartite motif containing protein 27 negatively regulates CD4 T cells by ubiquitinating..., Cai [/bib_ref]. Additionally, TRIM28 depletion in vivo promotes expansion of the Th17 population resulting in an autoimmune phenotype [bib_ref] TRIM28 prevents autoinflammatory T cell development in vivo, Chikuma [/bib_ref]. Following TCR-mediated activation, TRIM28 was phosphorylated suggesting it plays a role T cell activation [bib_ref] TRIM28 prevents autoinflammatory T cell development in vivo, Chikuma [/bib_ref]. Throughout the lifespan of Trim30 knockout mice, the ratio of CD4 + to CD8 + progressively increased and the CD4 + T cells proliferated abnormally [bib_ref] Tripartite motif-containing protein 30 modulates TCR-activated proliferation and effector functions in CD4+..., Choi [/bib_ref]. The role of TRIM30 is intrinsic to CD4 + T cells, because the same defect was observed upon adoptive transfer of knockout CD4 + T cells to wild-type mice [bib_ref] Tripartite motif-containing protein 30 modulates TCR-activated proliferation and effector functions in CD4+..., Choi [/bib_ref]. It remains to be seen whether other TRIMs play important roles in adaptive immunity. It is likely that TRIMs have unexplored functions in recognition of antigen presentation by the T cell receptor, and or in differentiation of CD4 + /CD8 + T cells.
## Trim-mediated virus inhibition
Despite the numerous host evasion mechanisms pathogens employ, a variety of host encoded molecules, such as TRIMs, are able to restrict viruses [bib_ref] Ubiquitin enzymes in the regulation of immune responses, Ebner [/bib_ref]. In addition to conferring an antiviral state indirectly by regulating cytokine production downstream of PRR signaling, TRIMs are capable of restricting the effectiveness of pathogens through direct interactions with viral proteins crucial to their entry, dissemination, or life cycle [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref]. The categories of viral restriction include: inhibition of viral transcription, replication or translation, degradation or interference of viral proteins, and impairment of virus entry or exit. We next outline the various means by which hosts deploy TRIMs to counter and clear pathogens.
## Trims and retroviruses
## Trim5α and retroviruses
TRIM5α is an example of a TRIM functioning as both a direct virus restriction factor as well as a pathogen-recognition receptor. In old-world monkeys, such as African green monkeys and rhesus macaques, TRIM5α enables natural resistance to HIV-1 while new-world monkeys, like owl monkeys, are protected from HIV-1 by fusion of TRIM5 with cyclophilin A (TCypA) [bib_ref] The cytoplasmic body component TRIM5α restricts HIV-1 infection in old world monkeys, Stremlau [/bib_ref] [bib_ref] TRIM5α protein restricts both HIV-1 and murine leukemia virus, Yap [/bib_ref] [bib_ref] A TRIM5-cyclophilin a fusion protein found in owl monkey kidney cells can..., Nisole [/bib_ref] [bib_ref] Cyclophilin a retrotransposition into TRIM5 explains owl monkey resistance to HIV-1, Sayah [/bib_ref]. Progress on the topic ranges from revealing the α isoform of TRIM5 as the restriction factor to discerning the role in restriction of each structural motif [bib_ref] The cytoplasmic body component TRIM5α restricts HIV-1 infection in old world monkeys, Stremlau [/bib_ref] [bib_ref] TRIM5α protein restricts both HIV-1 and murine leukemia virus, Yap [/bib_ref]. . The role of TRIMs in retrovirus replication. TRIM5α oligomerization into a hexagonal lattice associates directly with HIV-1 capsids to promote premature uncoating. Recognition of HIV-1 capsid by TRIM5α also triggers NF-κB/AP-1-mediated innate immune signaling via synthesis of unanchored K63-linked poly-Ub chains that activate the TAK1 kinase. One potential mechanism of TRIM5αmediated restriction could be involved ubiquitination of TRIM5α and proteasomal degradation of TRIM5α-capsid complexes. TRIM11 mobilizes cellular microtubule formation to prematurely uncoat HIV-1 and facilitate rapid release of the vRNA from the viral core, resulting in inhibition of virus replication. TRIM19 translocates to the cytoplasm and binds Daxx to prevent its degradation by the proteasome, allowing for Daxx-mediated disruption of HIV-1 reverse transcription (RT). TRIM22 inhibits Sp1, preventing LTR-mediated transcription. The connection between TRIM5α-mediated restriction and proteasomal-mediated degradation of HIV-1 requires further characterization. However, one potential model may involve improved proliferation of HIV-1-specific CD8 + T cells due to enhanced production of viral peptides from proteasome-mediated degradation of HIV-1 capsids.
Several components inherent to TRIM5α play a role in HIV-1 capsid restriction, including dimerization, oligomerization, and ubiquitination [bib_ref] TRIM5α requires ube2w to anchor Lys63-linked ubiquitin chains and restrict reverse transcription, Fletcher [/bib_ref] [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref]. There have been different models proposed for the mechanism of restriction and all have been attributed to the interaction of the SPRY domain of TRIM5α with the retrovirus capsid. Binding of TRIM5α to viral cores serves to prematurely uncoat the viral particle and induces early release of the viral genome. Elucidation of the TRIM5α higher order structure has revealed the significant contributions made by each component of the TRIM5α protein [bib_ref] A b-box 2 surface patch important for TRIM5α self-association, capsid binding avidity,..., Diaz-Griffero [/bib_ref] [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref] [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. Monomeric TRIM5α can dimerize in an antiparallel fashion through the coiled-coil domains, thereby generating the most fundamental component needed to bind the viral capsid [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref] [bib_ref] Recent insights into the mechanism and consequences of TRIM5α retroviral restriction, Sastri [/bib_ref]. Once bound to its target, TRIM5α can form higher-order structures resembling a hexagonal net [bib_ref] Hexagonal assembly of a restricting TRIM5αprotein, Ganser-Pornillos [/bib_ref] [bib_ref] TRIM5 is an innate immune sensor for the retrovirus capsid lattice, Pertel [/bib_ref]. TRIM5α and TRIM5 variants from several primate species, including rhesus macaques, African green monkeys, and owl monkeys, are capable of forming flexible, hexagonal frameworks encompassing HIV-1 capsid surfaces. The hexagonal nets are formed from TRIM5α trimers and are dependent on the TRIM5α's B-box domain [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. These trimers have been shown to be flexible and may allow for ideal binding of the SPRY domains on the highly variable HIV-1 capsid [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. The ability of rhesus macaque TRIM5α (rhTRIM5α) to form higher-order structures upon binding to the capsid also allows the formation of TRIM5α RING dimers, which enhances its E3-ubiquitin ligase activity and innate anti-HIV-1 activity [bib_ref] Ring dimerization links higher-order assembly of TRIM5alpha to synthesis of k63-linked polyubiquitin, Yudina [/bib_ref]. As one potential mechanism of rhTRIM5α-mediated restriction it was proposed that as the linker 2 regions of the dimer change their conformation, the SPRY domains that are bound to the HIV-1 capsid in their hexagonal lattice formation disturb the viral structure [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref]. The induction of this conformational change disrupts the integrity of the viral core and may be responsible for pre-mature viral uncoating. capsid by TRIM5α also triggers NF-κB/AP-1-mediated innate immune signaling via synthesis of unanchored K63-linked poly-Ub chains that activate the TAK1 kinase. One potential mechanism of TRIM5α-mediated restriction could be involved ubiquitination of TRIM5α and proteasomal degradation of TRIM5α-capsid complexes. TRIM11 mobilizes cellular microtubule formation to prematurely uncoat HIV-1 and facilitate rapid release of the vRNA from the viral core, resulting in inhibition of virus replication. TRIM19 translocates to the cytoplasm and binds Daxx to prevent its degradation by the proteasome, allowing for Daxx-mediated disruption of HIV-1 reverse transcription (RT). TRIM22 inhibits Sp1, preventing LTR-mediated transcription. The connection between TRIM5α-mediated restriction and proteasomal-mediated degradation of HIV-1 requires further characterization. However, one potential model may involve improved proliferation of HIV-1-specific CD8 + T cells due to enhanced production of viral peptides from proteasome-mediated degradation of HIV-1 capsids.
Several components inherent to TRIM5α play a role in HIV-1 capsid restriction, including dimerization, oligomerization, and ubiquitination [bib_ref] TRIM5α requires ube2w to anchor Lys63-linked ubiquitin chains and restrict reverse transcription, Fletcher [/bib_ref] [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref]. There have been different models proposed for the mechanism of restriction and all have been attributed to the interaction of the SPRY domain of TRIM5α with the retrovirus capsid. Binding of TRIM5α to viral cores serves to prematurely uncoat the viral particle and induces early release of the viral genome. Elucidation of the TRIM5α higher order structure has revealed the significant contributions made by each component of the TRIM5α protein [bib_ref] A b-box 2 surface patch important for TRIM5α self-association, capsid binding avidity,..., Diaz-Griffero [/bib_ref] [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref] [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. Monomeric TRIM5α can dimerize in an antiparallel fashion through the coiled-coil domains, thereby generating the most fundamental component needed to bind the viral capsid [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref] [bib_ref] Recent insights into the mechanism and consequences of TRIM5α retroviral restriction, Sastri [/bib_ref]. Once bound to its target, TRIM5α can form higher-order structures resembling a hexagonal net [bib_ref] Hexagonal assembly of a restricting TRIM5αprotein, Ganser-Pornillos [/bib_ref] [bib_ref] TRIM5 is an innate immune sensor for the retrovirus capsid lattice, Pertel [/bib_ref]. TRIM5α and TRIM5 variants from several primate species, including rhesus macaques, African green monkeys, and owl monkeys, are capable of forming flexible, hexagonal frameworks encompassing HIV-1 capsid surfaces. The hexagonal nets are formed from TRIM5α trimers and are dependent on the TRIM5α's B-box domain [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. These trimers have been shown to be flexible and may allow for ideal binding of the SPRY domains on the highly variable HIV-1 capsid [bib_ref] Crystal structure of the TRIM5α bbox2 domain from rhesus macaques describes a..., Keown [/bib_ref]. The ability of rhesus macaque TRIM5α (rhTRIM5α) to form higher-order structures upon binding to the capsid also allows the formation of TRIM5α RING dimers, which enhances its E3-ubiquitin ligase activity and innate anti-HIV-1 activity [bib_ref] Ring dimerization links higher-order assembly of TRIM5alpha to synthesis of k63-linked polyubiquitin, Yudina [/bib_ref]. As one potential mechanism of rhTRIM5α-mediated restriction it was proposed that as the linker 2 regions of the dimer change their conformation, the SPRY domains that are bound to the HIV-1 capsid in their hexagonal lattice formation disturb the viral structure [bib_ref] Dynamic conformational changes in the rhesus TRIM5α dimer dictate the potency of..., Lamichhane [/bib_ref]. The induction of this conformational change disrupts the integrity of the viral core and may be responsible for pre-mature viral uncoating.
Proteasomal-mediated degradation of viral components is a host restriction strategy characteristic of many TRIM family proteins involved in host innate immunity. However, the functional role of the proteasome in TRIM5α-mediated HIV-1 restriction has been difficult to discern and has been subject of debate. Early work with proteasomal inhibitors, like MG132, suggested a two phase restriction of HIV-1 by TRIM5α. In the proteasome-independent phase, TRIM5α inhibits nuclear entry of the reverse transcription (RT) products [bib_ref] Proteasome inhibition reveals that a functional preintegration complex intermediate can be generated..., Anderson [/bib_ref] [bib_ref] Proteasome inhibitors uncouple rhesus TRIM5α restriction of HIV-1 reverse transcription and infection, Wu [/bib_ref]. In the second phase, TRIM5α can inhibit late RT products in the presence of a functional proteasome, suggesting TRIM5α may utilize the proteasome to disrupt the viral RT complex [bib_ref] Proteasome inhibition reveals that a functional preintegration complex intermediate can be generated..., Anderson [/bib_ref] [bib_ref] Proteasome inhibitors uncouple rhesus TRIM5α restriction of HIV-1 reverse transcription and infection, Wu [/bib_ref]. While proteasome inhibition can block disassembly and/or degradation of viral core components, it does not appear to rescue infectivity, which has led some investigators to conclude that degradation of capsid by the proteasome is not the mechanism of TRIM5α-mediated restriction [bib_ref] Fates of retroviral core components during unrestricted and TRIM5-restricted infection, Kutluay [/bib_ref]. In addition, thus far no study has described ubiquitination sites on the HIV-1 capsids, which could potentially link degradation of the capsid with proteasomal function. However, additional studies have shown proteasomal involvement in TRIM5α restriction of HIV-1 by degrading TRIM5α itself [bib_ref] Proteasomal degradation of TRIM5α during retrovirus restriction, Rold [/bib_ref]. TRIM5α's E3 ligase activity can facilitate auto-ubiquitination, thereby signaling for its own destruction by the host proteasomal machinery [bib_ref] Rapid turnover and polyubiquitylation of the retroviral restriction factor TRIM5, Diaz-Griffero [/bib_ref] [bib_ref] Proteasomal degradation of TRIM5α during retrovirus restriction, Rold [/bib_ref]. These occurrences have led some to speculate on a potential model where TRIM5α interacts with the HIV-1 capsid leading to auto-ubiquitination, capsid uncoating, and delivery to the proteasome [bib_ref] Proteasomal degradation of TRIM5α during retrovirus restriction, Rold [/bib_ref] [bib_ref] Nonhuman TRIM5 variants enhance recognition of HIV-1-infected cells by CD8+ T cells, Jimenez-Moyano [/bib_ref] [bib_ref] Capsid-binding retrovirus restriction factors: Discovery, restriction specificity and implications for the development..., Sanz-Ramos [/bib_ref] [bib_ref] HIV-1 capsid recognition, and innate immune signaling, Grütter [/bib_ref]. Despite numerous investigations centered around proteasomal involvement, controversy within the literature surrounding the exact model connecting TRIM5α, HIV-1 cores, and proteasomes remains, and proteasome-dependent and independent mechanism have been proposed [bib_ref] Capsid-binding retrovirus restriction factors: Discovery, restriction specificity and implications for the development..., Sanz-Ramos [/bib_ref] [bib_ref] HIV-1 capsid recognition, and innate immune signaling, Grütter [/bib_ref].
Polyubiquitination through the E3 ligase activity of TRIM5α's RING domain may also be a factor involved in inhibition of retroviral replication [bib_ref] TRIM5α-mediated ubiquitin chain conjugation is required for inhibition of HIV-1 reverse transcription..., Campbell [/bib_ref]. In addition to its restriction activity, TRIM5α can induce antiviral type-I IFNs. Multiple TRIM5 orthologs induce AP-1-mediated innate immune signaling [bib_ref] TRIM5 is an innate immune sensor for the retrovirus capsid lattice, Pertel [/bib_ref] [bib_ref] TRIM5 retroviral restriction activity correlates with the ability to induce innate immune..., Lascano [/bib_ref]. In the presence of the HIV-1 capsid, the E3 ligase activity of TRIM5 facilitates the generation of unanchored K63-linked poly-Ub chains. These K63-linked chains promote TAK1 autophosphorylation and activation, resulting in the induction of AP-1-and NFκB-mediated transcription [bib_ref] TRIM5 is an innate immune sensor for the retrovirus capsid lattice, Pertel [/bib_ref]. Collectively, these recent findings serve to better characterize the mechanisms through which TRIM5α achieves inhibition of HIV-1 replication.
The restriction benefits conferred by TRIM5α can be observed in the progression to disease upon infection with simian immunodeficiency virus (SIV) in rhesus macaques. Course of infection studies with either a TRIM5α-sensitive or -resistant strain of SIV revealed that inoculation with a TRIM5α susceptible strain of SIV provided higher survival rates for rhesus macaques. Also, delayed development of the pathology associated with TRIM5α-sensitive SIV compared to subjects treated with a TRIM5α-resistant strain was observed [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. In addition to these benefits, the restriction of the TRIM5α sensitive SIV strain correlated with prolonged maintenance of CD4 + central memory T cells [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. The preservation of CD4 + central memory T cells has been extensively characterized as important in resisting viremia and generating CTL (cytotoxic T lymphocytes) subsets [bib_ref] Preserved CD4+ central memory T cells and survival in vaccinated SIV-challenged monkeys, Letvin [/bib_ref] [bib_ref] Defective CD8 T cell memory following acute infection without CD4 T cell..., Sun [/bib_ref] [bib_ref] Requirement for CD4 T cell help in generating functional CD8 T cell..., Shedlock [/bib_ref]. The prolonged presence of these CD4 + T cells in the TRIM5α-susceptible SIV-infected treatment may be responsible for the higher survival rate [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. In spite of this, TRIM5α escape mutants were present [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. Half of the subjects that received the TRIM5α-sensitive strain of SIV still developed AIDS, and a third succumbed to infection at a similar time as the TRIM5α-resistant group [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. Sequencing of the SIV capsids from these macaques revealed two mutations in the region encoding the gag protein, which resulted in nonsynonymous substitutions that mimicked the alterations made by the authors to generate their TRIM5α resistant strain [bib_ref] TRIM5α restriction affects clinical outcome and disease progression in simian immunodeficiency virus-infected..., Wu [/bib_ref]. Subversion of TRIMs is a central evolutionary strategy for viral innate immune evasion, as demonstrated by the high mutational rate of retroviruses.
Although first described as a viral restriction factor against HIV-1 in old-world monkeys, humans and other species are capable of utilizing TRIM5α to inhibit other retroviruses. Human TRIM5α is incapable of restricting HIV-1, yet a single amino acid mutation in its PRY-SPRY domain can confer resistance to the pathogen [bib_ref] A single amino acid change in the spry domain of human TRIM5αleads..., Yap [/bib_ref]. Instead, human TRIM5α can bind to and restrict N tropic mouse leukemia virus (N-MLV) [bib_ref] A single amino acid change in the spry domain of human TRIM5αleads..., Yap [/bib_ref]. Interestingly, the 13-amino-acid stretch in the PRY-SPRY domain of primate TRIM5α is under strong positive selection [bib_ref] Positive selection of primate TRIM5alpha identifies a critical species-specific retroviral restriction domain, Sawyer [/bib_ref]. TRIM5α paralogues have been identified in bovine [bib_ref] Evolution of a cytoplasmic tripartite motif (TRIM) protein in cows that restricts..., Si [/bib_ref] [bib_ref] Isolation of an active lv1 gene from cattle indicates that tripartite motif..., Ylinen [/bib_ref] , ovine [bib_ref] Ovine TRIM5α can restrict visna/maedi virus, Jauregui [/bib_ref] , and piscine [bib_ref] Origin and evolution of TRIM proteins: New insights from the complete trim..., Boudinot [/bib_ref] species, and have also been shown to restrict retroviruses. Overall, TRIM5α plays a convergent role in the recognition and restriction of retroviruses.
Aside from its role in non-human primates, TRIM5α has been described functioning in a cell-type specific manner [bib_ref] Receptor usage dictates HIV-1 restriction by human TRIM5α in dendritic cell subsets, Ribeiro [/bib_ref] [bib_ref] Lack of endogenous TRIM5α-mediated restriction in rhesus macaque dendritic cells, Arhel [/bib_ref] [bib_ref] Endogenous TRIM5α function is regulated by sumoylation and nuclear sequestration for efficient..., Portilho [/bib_ref]. In contrast to conventional DC-SIGN + (Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin) DCs, Langerhans cells (LCs) benefit from the anti-retroviral capabilities of TRIM5α through an increased activity of the LC autophagocytic components [bib_ref] Receptor usage dictates HIV-1 restriction by human TRIM5α in dendritic cell subsets, Ribeiro [/bib_ref]. Upon Langerin-mediated HIV-1 uptake, the presence of cellular autophagosomes increases as a result of TRIM5α-directed assembly, leading to targeting of the viral capsid for destruction [bib_ref] Receptor usage dictates HIV-1 restriction by human TRIM5α in dendritic cell subsets, Ribeiro [/bib_ref]. Further evidence of cell-type specific HIV-1 restriction by TRIM5α was found in rhesus macaque DCs, which in contrast to macrophages appear to be permissive to HIV-1 infection [bib_ref] Lack of endogenous TRIM5α-mediated restriction in rhesus macaque dendritic cells, Arhel [/bib_ref]. The lack of TRIM5α-mediated restriction in both human and macaque conventional DCs may be due to SUMOylation of TRIM5α, which promotes its sequestration in nuclear bodies [bib_ref] Endogenous TRIM5α function is regulated by sumoylation and nuclear sequestration for efficient..., Portilho [/bib_ref]. However, this lack of TRIM5α-mediated restriction in DCs provides an innate immune sensing advantage by allowing recognition of HIV-1 reversed transcribed DNA by cGAS, resulting in type-I IFN production [bib_ref] Endogenous TRIM5α function is regulated by sumoylation and nuclear sequestration for efficient..., Portilho [/bib_ref]. It remains to be seen whether HIV-1 can actively promote or enhance SUMOylation of TRIM5α in DCs as mechanism to antagonize type-I IFN production.
The ability of TRIM5α to promote restriction of HIV-1 may go beyond direct intervention of the viral lifecycle. TRIM5α has also been linked to the activity of other components of cellular innate immunity including antigen presentation to cytotoxic T lymphocytes (CTL). TCypA and rhTRIM5α enhanced the ability of CD8 + T cells to identify HIV-1 infected cells, and promoted a HIV-1-specific immune response [bib_ref] Nonhuman TRIM5 variants enhance recognition of HIV-1-infected cells by CD8+ T cells, Jimenez-Moyano [/bib_ref]. The presence of these TRIM5 orthologs was associated with increased associations between HIV-1 particles and the host proteasome [bib_ref] Nonhuman TRIM5 variants enhance recognition of HIV-1-infected cells by CD8+ T cells, Jimenez-Moyano [/bib_ref]. It is speculated that this may facilitate improved HIV-1-specific CTL development, as amplified peptide concentrations could support enhanced antigen presentation to CD8 + T cells. The mechanism linking direct restriction of viral capsids with heightened CD8 + T cell activation warrants further investigation.
## Other trims and retroviruses
A plethora of TRIM-mediated restriction mechanisms targeting HIV-1 and other retroviruses have been proposed [bib_ref] TRIM E3 ligases interfere with early and late stages of the retroviral..., Uchil [/bib_ref] , and have been reviewed previously [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref]. More recent reports acknowledged TRIM11 as a potent host restriction factor of HIV-1 [bib_ref] The human antiviral factor TRIM11 is under the regulation of HIV-1 VPR, Yuan [/bib_ref] [bib_ref] An HIV-1 capsid binding protein TRIM11 accelerates viral uncoating, Yuan [/bib_ref]. The mechanisms of TRIM11-mediated restriction include curbing the amount of viral reverse transcription products allowed to accumulate in the host. Through an interaction with the viral capsid-nucleocapsid protein (CA-NC) complexes, TRIM11 promotes premature uncoating and release of the viral genetic material, reducing transduction efficiency. Neither proteasomal nor lysosomal inhibitor treatments recovered viral p24 protein in the pellets of TRIM11 overexpressing cells, suggesting that ubiquitin-mediated degradation by the proteasome or lysosomal acidification is not required for TRIM11-mediated uncoating [bib_ref] An HIV-1 capsid binding protein TRIM11 accelerates viral uncoating, Yuan [/bib_ref]. Furthermore, while rhTRIM5α-mediated inhibition of HIV-1 is rescued by proteasome inhibitors [bib_ref] Proteasome inhibition reveals that a functional preintegration complex intermediate can be generated..., Anderson [/bib_ref] [bib_ref] Proteasome inhibitors uncouple rhesus TRIM5α restriction of HIV-1 reverse transcription and infection, Wu [/bib_ref] , TRIM11-mediated inhibition was not, indicating that TRIM11 and TRIM5α restrict HIV-1 by different mechanisms [bib_ref] The human antiviral factor TRIM11 is under the regulation of HIV-1 VPR, Yuan [/bib_ref] [bib_ref] An HIV-1 capsid binding protein TRIM11 accelerates viral uncoating, Yuan [/bib_ref]. In contrast, using the microtubule dynamics inhibitors nocodazole and taxol, the authors were able to demonstrate a restoration of HIV-1 capsid levels in cells supplemented with exogenous TRIM11, suggesting that microtubules may contribute to TRIM11-mediated HIV-1 restriction [bib_ref] An HIV-1 capsid binding protein TRIM11 accelerates viral uncoating, Yuan [/bib_ref]. Earlier work with the use of nocodazole and taxol in a study examining HIV-1 and TRIM5α also demonstrated recovered HIV-1 infectivity in the absence of functional microtubules [bib_ref] Functional evidence for the involvement of microtubules and dynein motor complexes in..., Pawlica [/bib_ref].
Although the exact mechanism of TRIM11-mediated restriction of HIV-1 in the aforementioned study has yet to be determined, the authors demonstrated that purified TRIM11 associated with in vitro assembled HIV-1 capsids [bib_ref] An HIV-1 capsid binding protein TRIM11 accelerates viral uncoating, Yuan [/bib_ref] , indicating that TRIM11 interacts directly with the capsid and it probably does not require TRIM5α or other cellular proteins for promoting untimely capsid uncoating. Microtubule involvement in viral uncoating has been implicated as both a host restriction and a viral propagation mechanism [bib_ref] Influenza a virus uses the aggresome processing machinery for host cell entry, Banerjee [/bib_ref] [bib_ref] HIV-1 uncoating is facilitated by dynein and kinesin 1, Lukic [/bib_ref] [bib_ref] Cytoplasmic dynein promotes HIV-1 uncoating, Pawlica [/bib_ref] [bib_ref] Unanchored ubiquitin in virus uncoating, Rajsbaum [/bib_ref]. Functional microtubules and their associated motor proteins, like dynein, have been suggested as key components in genome release for both IAV and HIV-1 [bib_ref] Influenza a virus uses the aggresome processing machinery for host cell entry, Banerjee [/bib_ref] [bib_ref] HIV-1 uncoating is facilitated by dynein and kinesin 1, Lukic [/bib_ref] [bib_ref] Cytoplasmic dynein promotes HIV-1 uncoating, Pawlica [/bib_ref] [bib_ref] Unanchored ubiquitin in virus uncoating, Rajsbaum [/bib_ref]. Interestingly, in the case of IAV, unanchored poly-Ub chains contained in the virion are recognized by the host histone deacetylase 6 (HDAC6), a component of the aggresome-autophagy pathway, which interacts with dynein and microtubules to promote viral uncoating [bib_ref] Influenza a virus uses the aggresome processing machinery for host cell entry, Banerjee [/bib_ref] [bib_ref] Unanchored ubiquitin in virus uncoating, Rajsbaum [/bib_ref]. It will be interesting to examine whether HIV-1 utilizes a similar mechanism and whether this is mediated by TRIM11. Additionally, several members of the TRIM family (TRIMs 1, 9, 18, 36, 46, and 67) have been shown to possesses a C-terminal domain motif allowing for association with cytoskeletal elements like microtubules [bib_ref] Subclassification of the RBCC/TRIM superfamily reveals a novel motif necessary for microtubule..., Short [/bib_ref]. So far only TRIM1 (also called MID2) and TRIM18 (also called MID1) have been shown to be directly involved in microtubule stabilization [bib_ref] MID1 and MID2 are required for xenopus neural tube closure through the..., Suzuki [/bib_ref] , and TRIM1 has been implicated in restriction of N-MLV [bib_ref] TRIM5α protein restricts both HIV-1 and murine leukemia virus, Yap [/bib_ref]. Whether TRIM1-mediated restriction is dependent on microtubules, or whether other microtubule-interacting TRIMs may have viral restriction activity by microtubule-dependent mechanisms, remains to be seen.
TRIM-mediated suppression of HIV-1 replication has revealed additional avenues through which TRIMs subdue pathogens. TRIM22 prevents normal viral transcription events by regulating the effectiveness of the transcription factor Sp1 to bind the HIV-1 long terminal repeat (LTR) promoter region [bib_ref] HIV-1 transcriptional silencing caused by TRIM22 inhibition of sp1 binding to the..., Turrini [/bib_ref]. This restriction was independent of TRIM22's E3 ligase activity and did not involve direct interaction between TRIM22 and Sp1, implying that the observed reduction in HIV-1 LTR-mediated transcription requires additional unspecified factors [bib_ref] HIV-1 transcriptional silencing caused by TRIM22 inhibition of sp1 binding to the..., Turrini [/bib_ref]. TRIM37 exhibits anti-retroviral functions through interference of HIV-1 infection, replication, and transcription, possibly interfering with DNA synthesis [bib_ref] Anti-HIV-1 activity of TRIM 37, Tabah [/bib_ref]. However further investigation into TRIM-mediated transcriptional inhibition will be required to reveal additional pathways in which TRIMs play a significant role.
Another example of indirect inhibition of virus replication by TRIMs is illustrated by the promyelocytic leukemia protein (PML)/TRIM19. PML/TRIM19 interferes with HIV-1 infection in human and murine fibroblasts through the reduction of reverse transcriptase products [bib_ref] The interferon-induced antiviral protein pml (TRIM19) promotes the restriction and transcriptional silencing..., Masroori [/bib_ref] [bib_ref] TRIM19/pml restricts HIV infection in a cell type-dependent manner, Kahle [/bib_ref]. This effect was dependent on two events; the translocation of PML/TRIM19 from the nucleus to the cytoplasm in the presence of HIV-1 infection, as well as association of the PML/TRIM19 cytoplasmic bodies (CB) with Daxx [bib_ref] Pml/TRIM19-dependent inhibition of retroviral reverse-transcription by daxx, Dutrieux [/bib_ref]. PML/TRIM19 interacts with Daxx in a protective fashion in order to prevent its degradation by the proteasome, thereby making PML/TRIM19 a necessary component for Daxx-mediated inhibition of HIV-1 RT [bib_ref] Pml/TRIM19-dependent inhibition of retroviral reverse-transcription by daxx, Dutrieux [/bib_ref]. The inhibition of HIV-1 by PML/TRIM19 and Daxx may be cell-type dependent, since studies using different cell types have shown different results [bib_ref] TRIM19/pml restricts HIV infection in a cell type-dependent manner, Kahle [/bib_ref] [bib_ref] Pml/TRIM19-dependent inhibition of retroviral reverse-transcription by daxx, Dutrieux [/bib_ref].
## Trim21 and antibody-dependent intracellular neutralization of viruses
Another well characterized TRIM involved in pathogen recognition is TRIM21, which has been shown to detect intracellular antibody-opsonized viruses (reviewed previously in [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref] [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref]. Some immunoglobulin-coated non-enveloped viruses are internalized into the host cell. Within most cells, a high affinity antibody receptor, TRIM21, binds to the highly conserved Fc region of virion-bound immunoglobulin (Ig)G, IgM, or IgA [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Translocalized IGA mediates neutralization and stimulates innate immunity inside infected cells, Bidgood [/bib_ref] , and targets the virus for degradation by the proteasome before the virus can transcribe its genes [bib_ref] Antibodies mediate intracellular immunity through tripartite motif-containing 21 (TRIM21), Mallery [/bib_ref]. TRIM21's PRY-SPRY domain interacts with Fc residues conserved across mammalian species [bib_ref] TRIM21 is an igg receptor that is structurally, thermodynamically, and kinetically conserved, Keeble [/bib_ref]. TRIM21 binding with an immunoglobulin-virion (Ig-V) complex triggers tightly regulated intracellular antibody neutralization and pro-inflammatory pathways [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref] [bib_ref] Translocalized IGA mediates neutralization and stimulates innate immunity inside infected cells, Bidgood [/bib_ref]. Upon engagement of TRIM21 with the Ig-V complex within the cytoplasm, the E2 Ube2W (Ubiquitin Conjugating Enzyme E2 W) monoubiquitinates TRIM21, which promotes Ube2N/Ube2V2 E2 complex recruitment to TRIM21 for K63-linked poly-Ub [bib_ref] Sequential ubiquitination and deubiquitination enzymes synchronize the dual sensor and effector functions..., Fletcher [/bib_ref]. Following poly-Ub, TRIM21 is recruited to the proteasome to initiate degradation of the antibody-bound virus. Concurrently, with proteasome-mediated degradation, Poh1 de-ubiquitinates TRIM21 [bib_ref] Sequential ubiquitination and deubiquitination enzymes synchronize the dual sensor and effector functions..., Fletcher [/bib_ref]. This antibody-dependent intracellular neutralization appears to be limited to non-enveloped DNA and RNA viruses that can enter the host cytosol with immunoglobulin attached to the virion surface [bib_ref] Intracellular antibody-bound pathogens stimulate immune signaling via the fc receptor TRIM21, Mcewan [/bib_ref]. The de-ubiquitinated TRIM21 is then able to promote both IFN induction and NF-κB activation [bib_ref] Sequential ubiquitination and deubiquitination enzymes synchronize the dual sensor and effector functions..., Fletcher [/bib_ref] [bib_ref] Intracellular antibody-bound pathogens stimulate immune signaling via the fc receptor TRIM21, Mcewan [/bib_ref] [bib_ref] Simultaneous neutralization and innate immune detection of a replicating virus by TRIM21, Watkinson [/bib_ref] , hypothesized to result from the release of the viral genome into the cytoplasm for PRR-mediated recognition [bib_ref] TRIM21 promotes cgas and RIG-I sensing of viral genomes during infection by..., Watkinson [/bib_ref]. However, some antiviral cytokines are induced independent of PRRs [bib_ref] TRIM21 promotes cgas and RIG-I sensing of viral genomes during infection by..., Watkinson [/bib_ref] , but the pathway is not well characterized. In addition, it has also been proposed that recognition of the Ig-V complex in the cytoplasm by TRIM21 triggers the synthesis of unanchored K63-linked poly-Ub chains that activate NF-κB, AP-1 and IRF3 pathways [bib_ref] Intracellular antibody-bound pathogens stimulate immune signaling via the fc receptor TRIM21, Mcewan [/bib_ref]. Notably, when the affinity of TRIM21 for the Fc portion of an antibody is decreased, the viral neutralization efficiency is maintained while the activation of cytokine signaling is diminished [bib_ref] TRIM21 immune signaling is more sensitive to antibody affinity than its neutralization..., Foss [/bib_ref]. This suggests that the association of TRIM21 with Ig-V complexes is important for triggering antiviral responses. TRIM21-deficient mice infected with mouse adenovirus 1 experienced lethal disease while wild-type mice were protected, exemplifying the importance of TRIM21 in viral neutralization in vivo [bib_ref] Intracellular antibody receptor TRIM21 prevents fatal viral infection, Vaysburd [/bib_ref]. Most likely, these mechanisms are shared in other species due to the highly conserved Fc and TRIM21 interacting residues, however the activity of TRIM21 as an intracellular antibody receptor in non-human, non-murine models has only been demonstrated in pig cells infected with foot-and-mouth disease virus [bib_ref] Swine TRIM21 restricts fmdv infection via an intracellular neutralization mechanism, Fan [/bib_ref].
## Trims and negative sense rna viruses
TRIM family proteins can target several components of pathogens ranging from structural elements to factors essential for transcription and replication. A variety of influenza A virus (IAV) and influenza B virus (IBV) proteins are targets of TRIM-mediated inhibition of virus replication [fig_ref] Figure 4: Effects of TRIMs on influenza virus components and IFN antagonism [/fig_ref]. Despite being a target of IAV-NS1 (nonstructural protein 1)-dependent IFN antagonism, TRIM25 interacts with the N-terminus of IBV-NS1 preventing the viral protein's C-terminal component from binding viral RNA [bib_ref] Robust lys63-linked ubiquitination of RIG-I promotes cytokine eruption in early influenza b..., Jiang [/bib_ref]. Preventing IBV-NS1 from interacting with RIG-I allows signaling through this RLR pathway to proceed [bib_ref] Robust lys63-linked ubiquitination of RIG-I promotes cytokine eruption in early influenza b..., Jiang [/bib_ref]. Another TRIM-mediated anti-IAV mechanism is polyubiquitination of IAV nucleoprotein (NP) by TRIM22, which leads to NP proteasome-mediated degradation and reduced virus replication [bib_ref] TRIM22 inhibits influenza a virus infection by targeting the viral nucleoprotein for..., Pietro [/bib_ref]. The polymerase basic protein 1 (PB1) of IAV is also a target of TRIM-mediated inhibition. PB1 is one of the three components that make up the RNA-dependent RNA-polymerase responsible for transcribing the eight segments of the IAV genome [bib_ref] Structure of influenza a polymerase bound to the viral RNA promoter, Pflug [/bib_ref]. TRIM32 facilitates K48-linked polyubiquitination of PB1, resulting in enhanced turnover of this viral subunit, and a reduction in viral titers [bib_ref] TRIM32 senses and restricts influenza a virus by ubiquitination of pb1 polymerase, Fu [/bib_ref].
Aside from targeting pathogen components for proteasomal degradation via ubiquitination, some members of the TRIM family have been shown to enact their restriction factor capabilities through other means. The TRIM56 C-terminal domain, rather than its E3 ligase activity, is required to reduce IAV and IBV replication. The mechanism TRIM56 employs to diminish vRNA levels of both influenza viruses is currently unknown, although inhibition of translation through direct interaction between TRIM56 and the vRNAs has been proposed [bib_ref] The C-terminal tail of TRIM56 dictates antiviral restriction of influenza A and..., Liu [/bib_ref].
## Trims and positive sense rna viruses
TRIMs have also been reported to mediate restriction against Flaviviruses [fig_ref] Figure 5: The roles of TRIMs in flavivirus infection and innate immune antagonism [/fig_ref]. The Flavivirus genus comprises more than 70 viruses including a number of important human pathogens such as Dengue virus (DENV), Zika virus (ZIKV), West Nile virus (WNV), tick-borne encephalitis virus (TBEV), Japanese encephalitis virus (JEV), hepatitis C virus (HCV), and yellow fever virus (YFV). Flaviviruses are small enveloped viruses hosting a positive-sense single-stranded RNA genome. Several flaviviral proteins are associated with viral persistence, immune system evasion, or viral replication [bib_ref] Pathogenesis of flavivirus infections: Using and abusing the host cell, Fernandez-Garcia [/bib_ref]. influenza viruses is currently unknown, although inhibition of translation through direct interaction between TRIM56 and the vRNAs has been proposed [bib_ref] The C-terminal tail of TRIM56 dictates antiviral restriction of influenza A and..., Liu [/bib_ref].
## Trims and positive sense rna viruses
TRIMs have also been reported to mediate restriction against Flaviviruses [fig_ref] Figure 5: The roles of TRIMs in flavivirus infection and innate immune antagonism [/fig_ref]. The Flavivirus genus comprises more than 70 viruses including a number of important human pathogens such as Dengue virus (DENV), Zika virus (ZIKV), West Nile virus (WNV), tick-borne encephalitis virus (TBEV), Japanese encephalitis virus (JEV), hepatitis C virus (HCV), and yellow fever virus (YFV). Flaviviruses are small enveloped viruses hosting a positive-sense single-stranded RNA genome. Several flaviviral proteins are associated with viral persistence, immune system evasion, or [bib_ref] Ubiquitin enzymes in the regulation of immune responses, Ebner [/bib_ref]. RNA genome is released into the cytoplasm (3). The positive-sense genomic ssRNA is translated into a polyprotein, which is cleaved into all structural and non-structural proteins (4). Replication takes place at the surface of endoplasmic reticulum in cytoplasmic viral factories [bib_ref] Ube2d3 and ube2n are essential for RIG-I-mediated mavs aggregation in antiviral innate..., Shi [/bib_ref]. In this step, the TRIMs restrict virus replication, degrading viral proteins such as NS2A in Japanese encephalitis virus (JEV) by TRIM52 and viral RNA inhibition in JEV and Dengue virus (DENV) by TRIM56. TRIM22 and TRIM79 degrade NS5 protein in hepatitis C virus (HCV) and tick-borne encephalitis virus (TBEV), respectively. Virus assembly occurs at the endoplasmic reticulum. The virion buds at the endoplasmic reticulum and is transported to the Golgi apparatus [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref]. The prM protein is cleaved in the Golgi, thereby maturing the virion, which is fusion competent [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref]. Release of new virions by exocytosis [bib_ref] Ubiquitin acetylation inhibits polyubiquitin chain elongation, Ohtake [/bib_ref]. TRIMs and antagonism function also are used by flaviviruses. TRIM21 inhibits IFN-β production during JEV infection. TRIM23 promotes yellow fever virus replication. DENV short noncoding sfRNAs bind TRIM25 to inhibit IFN expression.
HCV encodes a nonstructural protein, NS5A, which inhibits the phosphorylation and nuclear translocation of STAT1 in the IFN-α2-induced JAK/STAT pathway via their IFN sensitivitydetermining region [bib_ref] Understanding the molecular mechanism(s) of hepatitis C virus (HCV) induced interferon resistance, Qashqari [/bib_ref] [bib_ref] Mutations in the nonstructural protein 5a gene and response to interferon in..., Enomoto [/bib_ref]. TRIM14's SPRY domain specifically interacts with NS5 of HCV and induces NS5A degradation [bib_ref] TRIM14 inhibits hepatitis C virus infection by spry domain-dependent targeted degradation of..., Wang [/bib_ref] , which is an example of TRIM-mediated IFN-independent inhibition. TRIM22 specifically binds the NS5A-D1 protein (Domain 1) via its SPRY domain and The positive-sense genomic ssRNA is translated into a polyprotein, which is cleaved into all structural and non-structural proteins (4). Replication takes place at the surface of endoplasmic reticulum in cytoplasmic viral factories [bib_ref] Ube2d3 and ube2n are essential for RIG-I-mediated mavs aggregation in antiviral innate..., Shi [/bib_ref]. In this step, the TRIMs restrict virus replication, degrading viral proteins such as NS2A in Japanese encephalitis virus (JEV) by TRIM52 and viral RNA inhibition in JEV and Dengue virus (DENV) by TRIM56. TRIM22 and TRIM79 degrade NS5 protein in hepatitis C virus (HCV) and tick-borne encephalitis virus (TBEV), respectively. Virus assembly occurs at the endoplasmic reticulum. The virion buds at the endoplasmic reticulum and is transported to the Golgi apparatus [bib_ref] The roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity, Rajsbaum [/bib_ref]. The prM protein is cleaved in the Golgi, thereby maturing the virion, which is fusion competent [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref]. Release of new virions by exocytosis [bib_ref] Ubiquitin acetylation inhibits polyubiquitin chain elongation, Ohtake [/bib_ref]. TRIMs and antagonism function also are used by flaviviruses. TRIM21 inhibits IFN-β production during JEV infection. TRIM23 promotes yellow fever virus replication. DENV short noncoding sfRNAs bind TRIM25 to inhibit IFN expression.
HCV encodes a nonstructural protein, NS5A, which inhibits the phosphorylation and nuclear translocation of STAT1 in the IFN-α2-induced JAK/STAT pathway via their IFN sensitivity-determining region [bib_ref] Understanding the molecular mechanism(s) of hepatitis C virus (HCV) induced interferon resistance, Qashqari [/bib_ref] [bib_ref] Mutations in the nonstructural protein 5a gene and response to interferon in..., Enomoto [/bib_ref]. TRIM14's SPRY domain specifically interacts with NS5 of HCV and induces NS5A degradation [bib_ref] TRIM14 inhibits hepatitis C virus infection by spry domain-dependent targeted degradation of..., Wang [/bib_ref] , which is an example of TRIM-mediated IFN-independent inhibition. TRIM22 specifically binds the NS5A-D1 protein (Domain 1) via its SPRY domain and utilizes its E3 ubiquitin ligase activity to target NS5A for removal [bib_ref] Interferon α(ifnα)-induced TRIM22 interrupts HCV replication by ubiquitinating NS5a, Yang [/bib_ref].
Wenchun and colleagues have shown that TRIM52 interacts with the NS2A protein of JEV and targets the protein for proteasome-mediated destruction [bib_ref] TRIM52 inhibits japanese encephalitis virus replication by degrading the viral NS2A, Fan [/bib_ref]. NS2A is a small, hydrophobic transmembrane protein involved in the virus life cycle and subversion of host antiviral responses [bib_ref] Dengue virus inhibits α interferon signaling by reducing stat2 expression, Jones [/bib_ref] [bib_ref] Inhibition of interferon signaling by dengue virus, Munoz-Jordan [/bib_ref] , including inhibition of the double-stranded RNA-activated protein kinase PKR during JEV infection [bib_ref] Blocking double-stranded RNA-activated protein kinase PKR by japanese encephalitis virus nonstructural protein..., Tu [/bib_ref]. TRIM52-dependent inhibition of JEV NS2A protein occurs in BHK-21 and 293T cells and is therefore important for restricting JEV replication [bib_ref] TRIM52 inhibits japanese encephalitis virus replication by degrading the viral NS2A, Fan [/bib_ref].
Recent evidence also suggests that the RING and C-terminal domains of TRIM56 may be important in the restriction of other Flaviviruses, including YFV and DENV [bib_ref] Overlapping and distinct molecular determinants dictating the antiviral activities of TRIM56 against..., Liu [/bib_ref] , but the mechanism of action remains elusive. TRIM56 might modulate post-translational modification of one or more viral proteins and/or host factors to suppress viral replication [bib_ref] Identification of three interferon-inducible cellular enzymes that inhibit the replication of hepatitis..., Jiang [/bib_ref]. Although TRIM56 fails to restrict HCV replication when overexpressed in human hepatoma Huh7 cells [bib_ref] TRIM56 is a virus-and interferon-inducible E3 ubiquitin ligase that restricts pestivirus infection, Wang [/bib_ref] , TRIM56 overexpression in HEK293 cells support some selectable HCV RNA replicons at very low efficiencies [bib_ref] Identification of three interferon-inducible cellular enzymes that inhibit the replication of hepatitis..., Jiang [/bib_ref]. Perhaps TRIM56 facilitates degradation of viral proteins similar to mechanisms observed with TRIMs 14 and 22 against HCV and TRIM52 against JEV. Further studies are needed to elucidate the underlying molecular mechanisms of TRIM56-mediated restriction of DENV and YFV. Interestingly, TRIM56 is capable of binding to the protease N pro of bovine diarrheal virus (BVDV), a Pestivirus of the Flaviviridae family [bib_ref] TRIM56 is a virus-and interferon-inducible E3 ubiquitin ligase that restricts pestivirus infection, Wang [/bib_ref]. The restriction of this viral protein is critical considering that N pro is capable of mediating IRF3 degradation, thus impairing production of IFN-β [bib_ref] TRIM56 is a virus-and interferon-inducible E3 ubiquitin ligase that restricts pestivirus infection, Wang [/bib_ref]. TRIM56-mediated restriction of BVDV is specific, as closely related viruses are not likewise impaired in their replication [bib_ref] TRIM56 is a virus-and interferon-inducible E3 ubiquitin ligase that restricts pestivirus infection, Wang [/bib_ref].
TRIM79α, also known as TRIM30-3 or TRIM30D, is present only in rodents. TRIM79α is highly expressed in the spleen, lymph node, and bone marrow in a type-I IFN-dependent manner, and is required for effective restriction of TBEV replication [bib_ref] Tick-borne flaviviruses antagonize both IRF-1 and type I IFN signaling to inhibit..., Robertson [/bib_ref]. TRIM79α is an important mediator of the innate cellular response to restrict Langat virus (a member of the TBEV serogroup) infection by targeting the viral RNA polymerase and major IFN antagonist, NS5 [bib_ref] Tick-borne flaviviruses antagonize both IRF-1 and type I IFN signaling to inhibit..., Robertson [/bib_ref]. NS5 has a methyltransferase and RNA-dependent RNA polymerase activity that associates with NS3 and NS2B to form the viral replication complex. NS5 inhibits IFN-α/β-dependent responses by preventing JAK-STAT signaling and thus suppresses IFN-stimulated gene (ISG) expression [bib_ref] Tick-borne flaviviruses antagonize both IRF-1 and type I IFN signaling to inhibit..., Robertson [/bib_ref] [bib_ref] Mechanisms of evasion of the type I interferon antiviral response by flaviviruses, Diamond [/bib_ref] [bib_ref] Inhibition of interferon-stimulated jak-stat signaling by a tick-borne flavivirus and identification of..., Best [/bib_ref] [bib_ref] TRIM79α, an interferon-stimulated gene product, restricts tick-borne encephalitis virus replication by degrading..., Taylor [/bib_ref]. Taylor and colleagues demonstrated that TRIM79α interacts with NS5 from LGTV and TBEV and blocks the replication of these viruses via a lysosomal-targeting mechanism. Despite NS5 being the most conserved of the flaviviral proteins, TRIM79α did not target NS5 from WNV, nor could it inhibit WNV replication [bib_ref] TRIM79α, an interferon-stimulated gene product, restricts tick-borne encephalitis virus replication by degrading..., Taylor [/bib_ref].
Besides Flaviviruses, other positive-sense RNA viruses have been reported to be inhibited by TRIM-mediated mechanisms and can occur through both direct and indirect means. For example, TRIM25 acts as a cofactor for the Zinc-finger antiviral protein (ZAP), a member of the poly(ADP-ribose) polymerase family that is known to bind and promote viral RNA degradation [bib_ref] TRIM25 enhances the antiviral action of zinc-finger antiviral protein (zap), Li [/bib_ref]. TRIM25 enhances ZAPs' ability to inhibit Sindbis virus translation [bib_ref] TRIM25 enhances the antiviral action of zinc-finger antiviral protein (zap), Li [/bib_ref]. However, although TRIM25 is capable of promoting K48-and K63-linked polyubiquitination of ZAP isoforms, ubiquitination does not appear to affect ZAP antiviral activity [bib_ref] TRIM25 enhances the antiviral action of zinc-finger antiviral protein (zap), Li [/bib_ref]. It will be of interest to elucidate what other factors may be ubiquitinated by TRIM25 that could affect ZAP antiviral activity. TRIM19-IV has been shown to confer resistance to encephalomyocarditis virus (EMCV) hampering viral replication and protein synthesis [bib_ref] Promyelocytic leukemia isoform iv confers resistance to encephalomyocarditis virus via the sequestration..., Maroui [/bib_ref]. This ability was shown to originate from TRIM19-IV's C-terminus, specifically through an interaction with the viral 3D polymerase (3Dpol), leading to nuclear body sequestration of the viral polymerase [bib_ref] Promyelocytic leukemia isoform iv confers resistance to encephalomyocarditis virus via the sequestration..., Maroui [/bib_ref]. Additionally, SUMOylation of TRIM19-IV was required to facilitate restriction of EMCV [bib_ref] Promyelocytic leukemia isoform iv confers resistance to encephalomyocarditis virus via the sequestration..., Maroui [/bib_ref].
## Trims and dna viruses
TRIMs play an additional role in restricting DNA virus replication. Eight different TRIM proteins (TRIM5, were identified to inhibit hepatitis B virus (HBV) [bib_ref] Identification and characterization of multiple trim proteins that inhibit hepatitis B virus..., Zhang [/bib_ref]. In particular, TRIM41 was the only TRIM from this group of eight that specifically reduced both the enhancers I and II components of HBV. This inhibition of viral transcription required TRIM41's RING and PRY/SPRY domains, implicating its E3 ligase activity [bib_ref] Identification and characterization of multiple trim proteins that inhibit hepatitis B virus..., Zhang [/bib_ref]. TRIM22 is associated with HBV clearance in acutely infected chimpanzees, and possesses anti-HBV activity under physiological conditions [bib_ref] TRIM22: A diverse and dynamic antiviral protein, Hattlmann [/bib_ref]. Gao and colleagues also reported that TRIM22 was one of the most strongly induced TRIM family molecules in human hepatoma HepG2 cells after treatment with IFNs [bib_ref] Tripartite motif-containing 22 inhibits the activity of hepatitis B virus core promoter,..., Gao [/bib_ref].
Epstein-Barr virus (EBV) replication and transcription activator (Rta) protein activates EBV lytic genes for proliferation [bib_ref] TRIM5α promotes ubiquitination of RTA from epstein-barr virus to attenuate lytic progression, Huang [/bib_ref]. TRIM5α promotes the ubiquitination of Rta upon interaction, thereby blocking the EBV lytic cycle [bib_ref] TRIM5α promotes ubiquitination of RTA from epstein-barr virus to attenuate lytic progression, Huang [/bib_ref]. The involvement of TRIM5α in a non-retroviral innate immune response implies TRIMs possess diverse, situation-specific functions that have yet to be characterized. TRIMs may also operate as recognition receptors for other components of host immunity. TRIM19 isoforms are able to capture varicella-zoster virus nucleoproteins to impede nuclear egress and thus block the release of new virions [bib_ref] Emerging role of PML nuclear bodies in innate immune signaling, Scherer [/bib_ref]. Comprehension of the multiple roles TRIM-family proteins play may become critical in discerning the impact that they have on the innate immune response.
## Viral antagonism of trims
As we have described, TRIM family proteins play an important role in innate immunity and counter pathogens [bib_ref] Intrimsic immunity: Positive and negative regulation of immune signaling by tripartite motif..., Versteeg [/bib_ref]. In retaliation to this evolutionary pressure, pathogens have adapted to antagonize TRIMs. Mechanisms of viral-mediated TRIMs restriction range from impairment of TRIM-promoted innate immune signaling complex assembly to direct hindrance of the TRIM proteins. Interfering with TRIMs, either directly or indirectly, undermines their intended function, and enables viruses to gain an early advantage over the host [bib_ref] Viral evasion mechanisms of early antiviral responses involving regulation of ubiquitin pathways, Rajsbaum [/bib_ref]. The following section highlights several recent examples that elucidate viral manipulation of TRIMs.
## Antagonism of trims by rna viruses
A broad range of RNA viruses have evolved effective evasion strategies to manipulate host antiviral immunity. Influenza viruses employ well-studied examples of TRIM antagonism [fig_ref] Figure 4: Effects of TRIMs on influenza virus components and IFN antagonism [/fig_ref]. The nonstructural protein 1 (NS1) of influenza A and B virus (IAV/IBV) has been characterized as a viral antagonist of host innate immunity through interactions with TRIM25 [bib_ref] Influenza a virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition..., Gack [/bib_ref] [bib_ref] Species-specific inhibition of RIG-I ubiquitination and IFN induction by the influenza a..., Rajsbaum [/bib_ref] [bib_ref] Influenza A virus TRIMs the type I interferon response, Ludwig [/bib_ref] [bib_ref] The role of TRIM25 in development, disease and RNA metabolism, Heikel [/bib_ref]. As described above, TRIM25 plays a critical role in the activation of the RLR pathways [bib_ref] Influenza a virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition..., Gack [/bib_ref] [bib_ref] Ubiquitin in influenza virus entry and innate immunity, Rudnicka [/bib_ref] [bib_ref] TRIM25 identification in the chinese goose: Gene structure, tissue expression profiles, and..., Wei [/bib_ref]. The NS1 protein from IAV directly interacts with the coiled-coil domain of TRIM25 to impede its multimerization [bib_ref] Influenza a virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition..., Gack [/bib_ref] [bib_ref] Subcellular localizations of RIG-I, TRIM25, and MAVS complexes, Sanchez-Aparicio [/bib_ref]. Since dimerization is required for TRIM25 E3 ligase activity, NS1 binding to TRIM25 leads to impaired ubiquitination of the RIG-I and downstream signaling, resulting in a reduced antiviral response [bib_ref] Influenza a virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition..., Gack [/bib_ref] [bib_ref] Subcellular localizations of RIG-I, TRIM25, and MAVS complexes, Sanchez-Aparicio [/bib_ref].
The TRIM25 interaction with IAV NS1 is species-specific. Human TRIM25 interacts with IAV strains isolated from many species while chicken TRIM25 binds only NS1 from avian strains and murine TRIM25 did not bind any NS1 [bib_ref] Species-specific inhibition of RIG-I ubiquitination and IFN induction by the influenza a..., Rajsbaum [/bib_ref]. Riplet, a close relative of TRIM25, lacks a B-box domain but shares homologous RING and SPRY domains. Its predicted coiled-coil structure also binds NS1 to inhibit RIG-I signaling in mice and humans [bib_ref] Species-specific inhibition of RIG-I ubiquitination and IFN induction by the influenza a..., Rajsbaum [/bib_ref]. Aside from interacting with TRIM25, IAV NS1 also associates with host RIG-I directly [bib_ref] Subcellular localizations of RIG-I, TRIM25, and MAVS complexes, Sanchez-Aparicio [/bib_ref]. As seen with IAV, the N-terminal domain of IBV NS1 is also able to block the Lys63-linked ubiquitination of RIG-I and subsequent antiviral signaling downstream of the RLR pathway [bib_ref] Robust lys63-linked ubiquitination of RIG-I promotes cytokine eruption in early influenza b..., Jiang [/bib_ref].
Members of the Coronaviridae, Flaviviridae, and Bunyaviridae families also encode viral proteins that inhibit TRIM-mediated regulation of RLR signaling. Severe acute respiratory syndrome and Middle East respiratory syndrome coronavirus (SARS/MERS-CoV) are large, positive-sense single-stranded RNA viruses. Akin to the influenza virus NS1 protein, the nucleoprotein of SARS-CoV was demonstrated to interact with the SPRY domain of TRIM25, preventing the necessary interaction and subsequent ubiquitination of the RIG-I CARD domains [bib_ref] The severe acute respiratory syndrome coronavirus nucleocapsid inhibits type I interferon production..., Hu [/bib_ref]. A similar loss of RIG-I-induced IFN-β is achieved when MERS-CoV nucleoprotein associates with TRIM25 [bib_ref] The severe acute respiratory syndrome coronavirus nucleocapsid inhibits type I interferon production..., Hu [/bib_ref]. For DENV, Manokaran et al. compared two viral sequences (PR1 and PR-2B) and identified mutations that resulted in the increased production of subgenomic flavivirus non-coding RNAs (sfRNAs) by the PR-2B strain [bib_ref] Dengue subgenomic RNA binds TRIM25 to inhibit interferon expression for epidemiological fitness, Manokaran [/bib_ref] [fig_ref] Figure 5: The roles of TRIMs in flavivirus infection and innate immune antagonism [/fig_ref]. The PR-2B sfRNAs were capable of binding to host TRIM25 and prevented USP15-mediated deubiquitination [bib_ref] Dengue subgenomic RNA binds TRIM25 to inhibit interferon expression for epidemiological fitness, Manokaran [/bib_ref] , which is crucial for activation of RIG-I [bib_ref] The ubiquitin-specific protease usp15 promotes RIG-I-mediated antiviral signaling by deubiquitylating TRIM25, Pauli [/bib_ref]. This data provides unique molecular insight into the epidemiological fitness of DENV, suggesting that DENV sfRNAs can bind to host proteins to promote viral evasion of innate immunity [bib_ref] Dengue subgenomic RNA binds TRIM25 to inhibit interferon expression for epidemiological fitness, Manokaran [/bib_ref]. The NSs protein of severe fever with thrombocytopenia syndrome virus (SFTSV), a negative-sense RNA virus in the family Bunyaviridae, interacts directly with TRIM25, and indirectly with RIG-I and TBK1 to isolate these signaling molecules from associating with MAVS [bib_ref] Hijacking of RIG-I signaling proteins into virus-induced cytoplasmic structures correlates with the..., Santiago [/bib_ref]. As with the other viral protein-TRIM25 interactions described above, downstream activation of IRF3 and subsequent IFN-β production are impaired [bib_ref] Hijacking of RIG-I signaling proteins into virus-induced cytoplasmic structures correlates with the..., Santiago [/bib_ref]. TRIM25 is a common target of a diverse group of RNA viruses, suggesting that other pathogens may also impair TRIM25-mediated stimulation of the RLR pathway.
Similar to IAV, HIV-1 encodes multiple proteins that restrict TRIM function. For example, HIV-1 Vpr protein has been suggested to play a role in TRIM manipulation [bib_ref] The human antiviral factor TRIM11 is under the regulation of HIV-1 VPR, Yuan [/bib_ref]. Interestingly, protein expression levels of TRIM11 have shown to be under the control of Vpr in a dose-dependent manner, where the presence of TRIM11 was decreased when the concentration of Vpr was low [bib_ref] The human antiviral factor TRIM11 is under the regulation of HIV-1 VPR, Yuan [/bib_ref]. Currently, the mechanism HIV-1 Vpr employs to antagonize TRIM11 is unknown. Following infection of human neural precursor cells (hNPCs) with HIV-1, TRIM32 becomes upregulated, primarily due to the viral trans-activator of transcription protein (Tat) [bib_ref] Tripartite containing motif 32 modulates proliferation of human neural precursor cells in..., Fatima [/bib_ref]. This Tat-mediated upregulation of TRIM32 induces proliferation arrest in hNPCs, eventually leading to neurodegeneration [bib_ref] Tripartite containing motif 32 modulates proliferation of human neural precursor cells in..., Fatima [/bib_ref].
Flaviviruses also encode viral proteins that antagonize TRIM-mediated innate immunity [fig_ref] Figure 5: The roles of TRIMs in flavivirus infection and innate immune antagonism [/fig_ref]. YFV and DENV NS5 protein have 10 amino acid residues on the N-terminus, which are essential for antagonism of type-I IFN signaling [bib_ref] The interferon signaling antagonist function of yellow fever virus NS5 protein is..., Laurent-Rolle [/bib_ref] [bib_ref] Ns5 of dengue virus mediates STAT2 binding and degradation, Ashour [/bib_ref]. YFV NS5 binds STAT2 only after IFN treatment, and appears to inactivate ISGF3 within the nucleus [bib_ref] The interferon signaling antagonist function of yellow fever virus NS5 protein is..., Laurent-Rolle [/bib_ref]. found in DENV NS5 a glycine and a threonine residue within the N-terminus that are required for binding with UBR4 (Ubiquitin Protein Ligase E3 Component N-Recognin 4) to mediate STAT2 degradation [bib_ref] Dengue virus co-opts ubr4 to degrade Stat2 and antagonize type I interferon..., Morrison [/bib_ref]. UBR4, a member of the N-recognin family, is a potential E3 ligase that recognizes and degrades proteins containing destabilizing N termini.. UBR4 interacts preferentially with proteolytically-processed DENV NS5, but not with YFV NS5 or WNV NS5 [bib_ref] Dengue virus co-opts ubr4 to degrade Stat2 and antagonize type I interferon..., Morrison [/bib_ref]. Although UBR4 does not belong to the TRIM family, it is possible that TRIM members may also be involved in STAT2 degradation by NS5. For example, TRIM23 was identified as an essential factor in YFV replication due to its interaction and poly-Ub of residue K6 on YFV-NS5, promoting binding with STAT2 and inhibition of type-I IFN signaling [bib_ref] The interferon signaling antagonist function of yellow fever virus NS5 protein is..., Laurent-Rolle [/bib_ref]. Another flavivirus, Japanese encephalitis virus (JEV), induces expression of TRIM21 in human microglial cells, which attenuates JEV-induced antiviral signaling [bib_ref] Regulatory role of TRIM21 in the type-I interferon pathway in japanese encephalitis..., Manocha [/bib_ref]. The study by Manocha et al. demonstrated that TRIM21 overexpression suppressed phosphorylation of IRF3 and activation of IFN-β, while silencing TRIM21 permitted efficient type-I IFN responses in JEV-infected human microglial cells [bib_ref] Regulatory role of TRIM21 in the type-I interferon pathway in japanese encephalitis..., Manocha [/bib_ref]. This study provides evidence that JEV suppress the IFN-I response due to induction of TRIM21.
Finally, we recently showed that Nipah virus (NiV), a single-stranded negative-sense highly pathogenic RNA virus (Paramyxoviridae family, genus Henipavirus) that causes fatal diseases in humans [bib_ref] Hendra and nipah viruses: Different and dangerous, Eaton [/bib_ref] , can inhibit TRIM6-mediated type-I IFN responses [bib_ref] The matrix protein of nipah virus targets the E3-ubiquitin ligase TRIM6 to..., Bharaj [/bib_ref] [fig_ref] Figure 6: TRIM6 is targeted by Nipah and Ebola viruses to enhance virus replication [/fig_ref]. Mechanistically, the NiV matrix (NiV-M) structural protein, which is required for virus assembly and budding [bib_ref] Evidence for ubiquitin-regulated nuclear and subnuclear trafficking among paramyxovirinae matrix proteins, Pentecost [/bib_ref] [bib_ref] Ubiquitin-regulated nuclear-cytoplasmic trafficking of the nipah virus matrix protein is important for..., Wang [/bib_ref] , targets TRIM6 for degradation. Interestingly, NiV budding requires trafficking of NiV-M from the cytoplasm to the nucleus before reaching the cell membrane for virus assembly [bib_ref] Evidence for ubiquitin-regulated nuclear and subnuclear trafficking among paramyxovirinae matrix proteins, Pentecost [/bib_ref] [bib_ref] Ubiquitin-regulated nuclear-cytoplasmic trafficking of the nipah virus matrix protein is important for..., Wang [/bib_ref]. The reason for NiV-M trafficking to the nucleus is still unclear, however, a lysine residue (K258) in the NiV-M bipartite nuclear localization signal that is conserved in divergent henipaviruses and is required for trafficking, is critical for the IFN antagonist function [bib_ref] The matrix protein of nipah virus targets the E3-ubiquitin ligase TRIM6 to..., Bharaj [/bib_ref]. Consistent with this, the matrix proteins of Ghana, Hendra and Cedar viruses were also able to inhibit IFN-β induction [bib_ref] The matrix protein of nipah virus targets the E3-ubiquitin ligase TRIM6 to..., Bharaj [/bib_ref]. It is currently unknown whether TRIM6 E3-ligase activity affects NiV-M trafficking or whether it directly interferes with NiV-replication. NiV-M-induced TRIM6 degradation did not appear to require the proteasome, however. Although the precise mechanism of TRIM6 degradation remains to be elucidated, inhibitors that recover TRIM6 protein levels could potentially be used as therapeutic drugs against NiV infections. These findings highlight the importance of TRIM6 as an antiviral factor of the type-I IFN system.
## Antagonism of trims by dna viruses
A variety of DNA viruses also encode viral proteins that interfere with TRIMs. Alternative methods for subverting TRIM-induced innate immune responses can be noted through alterations made to host transcription by hepatitis B virus (HBV). The HBV-encoded protein X (HBx) methylates a single CpG in TRIM22's promoter region. This viral-mediated epigenetic modification blocks IFN-α/γ-induced IRF1, a transcriptional activator, from binding the TRIM22 promoter and consequently facilitates viral proliferation and escape. In Epstein-Barr virus (EBV), KAP1 (KRAB [Kruppel-Associated Box Domain]-Associated Protein 1)/TRIM28 regulates activation of the viral lytic cycle [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref]. In cells undergoing lytic stage activation, KAP1/TRIM28 becomes phosphorylated at serine residue 824 by ataxia telangiectasia mutated (ATM). This post-translational modification impairs KAP1/TRIM28's restriction factor function and allows EBV to transition from latency to lytic stage [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref]. Additionally, the anti-malarial drug chloroquine acts as an activator for ATM by phosphorylating KAP1/TRIM28, which ultimately leads to promotion of the EBV lytic cycle and escape of the virus particles [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref].
Mutations in gene expression serve as an additional focal point highlighting the role that TRIMs play in a dysfunctional antiviral response. When a population of 765 HBV-infected individuals was screened, chronic HBV infection was found to correlate to a single T to C silent mutation single nucleotide polymorphism (SNP) in the TRIM22 RING domain [bib_ref] Tripartite motif-containing 22 gene-364t/c polymorphism associated with hepatitis B virus infection in..., Zhao [/bib_ref]. Although the implications to patient outcomes are considerable, additional investigations into the basic biological mechanisms are warranted to discern the importance of these SNPs in TRIM-mediated innate immunity. [fig_ref] Figure 6: TRIM6 is targeted by Nipah and Ebola viruses to enhance virus replication [/fig_ref]. TRIM6 is targeted by Nipah and Ebola viruses to enhance virus replication. Ebola virus (EBOV) inhibits type-I IFN production by multiple mechanisms. EBOV VP35 binds and inhibits RIG-I, IKKε, and TBK-1 to inhibit IFN production. TRIM6 ubiquitinates VP35 on K309 and promotes VP35 activity as the cofactor of the viral polymerase and enhances virus replication. Additional unidentified ubiquitination sites of VP35 exist. Whether TRIM6 enhances EBOV replication by promoting viral genome replication or viral gene transcription is not known. In Nipah virus infection, the viral matrix protein (NiV-M) promotes TRIM6 degradation, resulting in reduced synthesis of K48-linked unanchored polyubiquitin chains, IKKε oligomerization, IKKε-T501 autophosphorylation, IRF3 phosphorylation, and reduced IFN induction. These combined losses confer an impaired host-antiviral response.
## Antagonism of trims by dna viruses
A variety of DNA viruses also encode viral proteins that interfere with TRIMs. Alternative methods for subverting TRIM-induced innate immune responses can be noted through alterations made to host transcription by hepatitis B virus (HBV). The HBV-encoded protein X (HBx) methylates a single CpG in TRIM22's promoter region. This viral-mediated epigenetic modification blocks IFN-α/γ-induced IRF1, a transcriptional activator, from binding the TRIM22 promoter and consequently facilitates viral proliferation and escape. In Epstein-Barr virus (EBV), KAP1 (KRAB [Kruppel-Associated Box Domain]-Associated Protein 1)/TRIM28 regulates activation of the viral lytic cycle [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref]. In cells undergoing lytic stage activation, KAP1/TRIM28 becomes phosphorylated at serine residue 824 by ataxia telangiectasia mutated (ATM). This post-translational modification impairs KAP1/TRIM28's restriction factor function and allows EBV to transition from latency to lytic stage [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref]. Additionally, the anti-malarial drug chloroquine acts as an activator for ATM by phosphorylating KAP1/TRIM28, which ultimately leads to promotion of the EBV lytic cycle and escape of the virus particles [bib_ref] Chloroquine triggers epstein-barr virus replication through phosphorylation of kap1/TRIM28 in burkitt lymphoma..., Li [/bib_ref].
Mutations in gene expression serve as an additional focal point highlighting the role that TRIMs play in a dysfunctional antiviral response. When a population of 765 HBV-infected individuals was screened, chronic HBV infection was found to correlate to a single T to C silent mutation single nucleotide polymorphism (SNP) in the TRIM22 RING domain [bib_ref] Tripartite motif-containing 22 gene-364t/c polymorphism associated with hepatitis B virus infection in..., Zhao [/bib_ref]. Although the implications to patient outcomes are considerable, additional investigations into the basic biological mechanisms are warranted to discern the importance of these SNPs in TRIM-mediated innate immunity.
In addition to impairing TRIM gene expression, DNA viruses can mimic host proteins to prohibit TRIM function. The immediate early protein (ICP0) of herpes simplex virus 1 (HSV-1) possesses a RING domain and facilitates the degradation of TRIM27 through ubiquitination [bib_ref] Identification of TRIM27 as a novel degradation target of herpes simplex virus..., Conwell [/bib_ref]. This destruction of TRIM27 is facilitated through the host's proteasome [bib_ref] Identification of TRIM27 as a novel degradation target of herpes simplex virus..., Conwell [/bib_ref] , mimicking the numerous instances of TRIM-mediated restriction. Gammaherpesvirus MHV-68 similarly targets TRIM19 for proteasome-mediated degradation [bib_ref] Emerging role of PML nuclear bodies in innate immune signaling, Scherer [/bib_ref]. The IE1 proteins of human cytomegalovirus (HCMV) mimics the coiled-coil domain of TRIMs to recruit and sequester TRIM19 from nuclear bodies thus impairing the activation of IFN responses [bib_ref] Emerging role of PML nuclear bodies in innate immune signaling, Scherer [/bib_ref].
## Hijacking of antiviral trims as a novel mechanism to directly enhance virus replication
So far, most studies on TRIMs have focused on their roles as antiviral factors by directly restricting virus replication or indirectly by inducing antiviral cytokines, as described above. The fact that TRIMs are targeted by viruses for immune evasion further highlights their important roles in protecting the host against infections. However, whether TRIMs may directly act as host factors required for virus replication, or "pro-viral" factors, has not been addressed. Some studies have shown that certain viral antagonists can hijack TRIMs to activate their IFN antagonist activity (e.g., TRIM23 ubiquitinates YFV-NS5 for antagonism of STAT2 function [bib_ref] The interferon signaling antagonist function of yellow fever virus NS5 protein is..., Laurent-Rolle [/bib_ref] , but these are indirect effects that provide an advantage to the virus by reducing host antiviral responses. Since ubiquitination of viral proteins may positively influence specific steps of the replication cycle, it would not be surprising if TRIMs are involved in directly promoting virus replication by non-degradative ubiquitination of viral proteins.
Indeed, we recently reported the first example of such a role for TRIM6 [bib_ref] The host E3-ubiquitin ligase TRIM6 ubiquitinates the ebola virus VP35 protein and..., Bharaj [/bib_ref]. As described above in Section 2.6, TRIM6 is involved in antiviral type-I IFN responses by catalyzing the synthesis of unanchored K48-linked polyubiquitin chains that activate IKKε. Further evidence of TRIM6 as an antiviral factor is highlighted by our findings that the NiV can inhibit IFN-I responses by targeting TRIM6 [bib_ref] The matrix protein of nipah virus targets the E3-ubiquitin ligase TRIM6 to..., Bharaj [/bib_ref]. Furthermore, knockdown of TRIM6 in lung A549 cell lines and primary human monocyte derived dendritic cells has shown increased replication of multiple viruses, including IAV, EMCV, and Sendai virus (SeV), most probably due to reduced type-I IFN responses [bib_ref] Unanchored k48-linked polyubiquitin synthesized by the e3-ubiquitin ligase TRIM6 stimulates the interferon-ikkepsilon..., Rajsbaum [/bib_ref]. In addition, TRIM6 A549 knockout cells showed reduced IFN responses [bib_ref] The host E3-ubiquitin ligase TRIM6 ubiquitinates the ebola virus VP35 protein and..., Bharaj [/bib_ref]. Unexpectedly, despite a defect in the IFN response, infectious Ebola virus (EBOV: Filoviridae family) replicated less efficiently in TRIM6 knockout cells as compared to parental wild type (WT) cells. This observation raised the question whether TRIM6 may be acting as a pro-viral factor or an enhancer of virus replication. In support of this hypothesis, we found that TRIM6 interacts with EBOV-VP35, which is a major viral IFN antagonist by targeting RIG-I [bib_ref] Ebola virus VP35 protein binds double-stranded RNA and inhibits α/β interferon production..., Cardenas [/bib_ref] [bib_ref] Mutual antagonism between the ebola virus VP35 protein and the RIG-I activator..., Luthra [/bib_ref] and the kinases IKKε and TBK-1 [bib_ref] Ebola virus protein VP35 impairs the function of interferon regulatory factor-activating kinases..., Prins [/bib_ref]. However, since VP35 also plays a critical role as the cofactor of the virus polymerase, TRIM6 could directly affect polymerase function. Mass spectrometry analysis and co-immunoprecipitation assays demonstrated that TRIM6 ubiquitinates VP35 on K309 [bib_ref] The host E3-ubiquitin ligase TRIM6 ubiquitinates the ebola virus VP35 protein and..., Bharaj [/bib_ref] , a lysine residue located on its IFN antagonist domain. Moreover, minigenome reporter experiments showed that TRIM6 can enhance VP35-mediated polymerase activity, and this effect requires the E3-ubiquitin ligase activity of TRIM6. Although VP35 is ubiquitinated in multiple (currently unidentified) residues in addition to K309, the TRIM6-dependent effects on VP35 minigenome activity required an intact K309 residue. VP35 was also able to inhibit RIG-I induced TRIM6-enhanced IFNβ in reporter assays, however the precise mechanism was not elucidated [bib_ref] The host E3-ubiquitin ligase TRIM6 ubiquitinates the ebola virus VP35 protein and..., Bharaj [/bib_ref]. Collectively, these findings suggest that TRIM6 is a host factor hijacked by EBOV-VP35 for both immune evasion and for promoting virus replication via ubiquitination of VP35 [fig_ref] Figure 6: TRIM6 is targeted by Nipah and Ebola viruses to enhance virus replication [/fig_ref]. Future studies will address whether ubiquitination of VP35 regulates virus RNA replication or transcription. It remains to be seen if other TRIMs that are targeted by viruses may also directly enhancing virus replication via ubiquitination of viral proteins.
## Potential roles of trim-mediated autophagy during virus infections
Autophagy (self-eating) is a highly conserved catabolic mechanism through which eukaryotic cells deliver dispensable, or potentially dangerous, cytoplasmic material to lysosomes for degradation [bib_ref] Eaten alive: A history of macroautophagy, Yang [/bib_ref]. The process is characterized by the formation of autophagosomes, which sequester the cytoplasmic structures targeted for destruction. Autophagy has been linked to a wide range of physiological processes, including cell differentiation and development, the degradation of aberrant structures and turnover of damaged organelles, as well as innate and adaptive immunity [bib_ref] Autophagy, immunity, and microbial adaptations, Deretic [/bib_ref] [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref]. A growing number of studies indicate that several TRIMs are linked to autophagy and recent excellent reviews are available on the emerging roles of TRIMs in autophagy [bib_ref] TRIM family proteins: Roles in autophagy, immunity, and carcinogenesis, Hatakeyama [/bib_ref] [bib_ref] Precision autophagy directed by receptor regulators--Emerging examples within the TRIM family, Kimura [/bib_ref] [bib_ref] Autophagy in leukocytes and other cells: Mechanisms, subsystem organization, selectivity, and links..., Deretic [/bib_ref] [bib_ref] TRIM-directed selective autophagy regulates immune activation, Kimura [/bib_ref].
A good example of the potential role of TRIMs in autophagy and virus infections is rhesus TRIM5α, which acts both as a regulator of autophagy by providing a platform for the assembly of activated ULK1 and Beclin 1 (key components of the autophagy regulatory complexes) and as a receptor for selective autophagy [bib_ref] TRIM proteins regulate autophagy and can target autophagic substrates by direct recognition, Mandell [/bib_ref] [bib_ref] TRIM proteins regulate autophagy: TRIM5 is a selective autophagy receptor mediating HIV-1..., Mandell [/bib_ref]. In its role as an autophagic cargo receptor, TRIM5α directly recognizes viral capsid sequences via its SPRY domain [bib_ref] The cytoplasmic body component TRIM5α restricts HIV-1 infection in old world monkeys, Stremlau [/bib_ref] [bib_ref] TRIM proteins regulate autophagy and can target autophagic substrates by direct recognition, Mandell [/bib_ref] [bib_ref] TRIM proteins regulate autophagy: TRIM5 is a selective autophagy receptor mediating HIV-1..., Mandell [/bib_ref] [bib_ref] Specific recognition and accelerated uncoating of retroviral capsids by the trim5α restriction..., Stremlau [/bib_ref]. This is an example of selective autophagy in mammalian cells, which could occur via direct substrate recognition by TRIMs and connects autophagy with a role in defense against viral pathogens. The rhesus TRIM5 can execute precision autophagy of the HIV-1 capsid. In contrast, the weak affinity of human TRIM5 for the HIV-1 capsid precludes effective precision autophagy. As a consequence, rhesus TRIM5, but not human TRIM5, could contribute to defense against HIV-1 through precision autophagy. Since this recognition depends on the C-terminal region of TRIMs (e.g., SPRY domain of TRIM5α), other types of C-terminal domains on TRIMs could selectively recognize diverse protein targets. Thus, TRIM proteins, as a group, could comprise a class of broad-repertoire, high-fidelity, selective autophagic receptors. Given the breadth of the role of TRIMs in various diseases, it will be important to explore precision autophagy-in addition to bulk autophagy-as a therapeutic target against viral infections.
The proposed role of TRIMs in autophagy raise questions. For example, how many TRIMs act as autophagic receptors and what are their specific targets? Do TRIM proteins function as hubs connecting different signaling pathways or different systems? How is the autophagic role of TRIMs integrated with the other functions of TRIMs, including regulation of gene expression and pro-inflammatory signaling? What is the interplay between the E3 ligase activity of TRIMs and precision autophagy? It is important to determine inhibitory compounds of TRIM proteins for their use as therapeutic tools in infectious diseases. However, because some TRIM proteins have simultaneous dual functions in carcinogenesis and the immune response, it should be considered that putative drugs (inhibitors of some TRIM proteins) for cancer therapy may affect immunological reactions as a side effect. Further detailed analysis of TRIM proteins is needed for their use as novel therapeutics with minimal side effects.
## Conclusions and future perspectives
This review highlighted the roles of TRIMs in virus-host interactions. Extensive reports detail TRIM involvement in immune signaling and direct virus restriction. In addition, viral antagonism of TRIMs exemplifies the importance of this protein family in antiviral responses. Despite these advancements, many TRIMs have yet to be characterized. Additionally, the molecular mechanisms underlying TRIM-mediated virus restriction or viral protein-mediated TRIM inhibition are not fully elucidated. The role of TRIMs in regulating poly-Ub chain topology is also of interest. In several examples, including TRIMs 5 [bib_ref] TRIM5 is an innate immune sensor for the retrovirus capsid lattice, Pertel [/bib_ref] , 6 [bib_ref] Unanchored k48-linked polyubiquitin synthesized by the e3-ubiquitin ligase TRIM6 stimulates the interferon-ikkepsilon..., Rajsbaum [/bib_ref] , 21 [bib_ref] Sequential ubiquitination and deubiquitination enzymes synchronize the dual sensor and effector functions..., Fletcher [/bib_ref] , and 25 [bib_ref] Reconstitution of the RIG-I pathway reveals a signaling role of unanchored polyubiquitin..., Zeng [/bib_ref] , TRIMs facilitate the synthesis of unanchored poly-Ub chains. The relative contribution of TRIMs in synthesizing specifically unanchored poly-Ub chains, versus covalently linked poly-Ub chains, is unclear. Classically, the E2 is considered more important in determining poly-Ub chain characteristics [bib_ref] Building ubiquitin chains: E2 enzymes at work, Ye [/bib_ref] [bib_ref] Ube2d3 and ube2n are essential for RIG-I-mediated mavs aggregation in antiviral innate..., Shi [/bib_ref] , but post-translational modifications may influence the decision between covalent and non-covalent linkage [bib_ref] Sequential ubiquitination and deubiquitination enzymes synchronize the dual sensor and effector functions..., Fletcher [/bib_ref] as may the choice of the partnering E3. Perhaps TRIMs play a unique role in the synthesis of unanchored chains. In the future, evaluating the factors that regulate E2-TRIM pairing may add an additional layer of complexity to TRIM-mediated regulation and the ubiquitin code. Although one study addressed this question by testing interactions between TRIMs and the E2-conjugases and a few TRIM-E2 pairs were identified [bib_ref] Functional interactions between ubiquitin E2 enzymes and TRIM proteins, Napolitano [/bib_ref] , the complexity of potential transient interactions, and the possibility of cell-type specific expression for the combination of TRIMs and E2-conjugases makes this task a huge challenge.
Importantly, another question is how poly-Ub chains of unconventional linkages may affect virus replication. For example, unanchored poly-Ub chains can be incorporated into the IAV virion to enable hijacking of the host's aggresomal pathway to facilitate viral replication [bib_ref] Influenza a virus uses the aggresome processing machinery for host cell entry, Banerjee [/bib_ref]. The E2 and E3 enzymes responsible for synthesizing these poly-Ub chains have not been identified, but perhaps IAV hijacks a TRIM to make these poly-Ub chains critical in efficient infection. Therefore, unanchored poly-Ub chains may have both pro-viral and antiviral functions [bib_ref] Unanchored ubiquitin in virus uncoating, Rajsbaum [/bib_ref] and elucidating how to tip the balance towards antiviral responses may help develop novel antiviral therapies. Additionally, IAV replication via host ubiquitin and aggresome systems relies on the host's cytoskeletal network [bib_ref] Influenza a virus uses the aggresome processing machinery for host cell entry, Banerjee [/bib_ref]. This may further implicate the role of TRIMs in similar pathways as several TRIMs associate with microtubules [bib_ref] The tripartite motif family identifies cell compartments, Reymond [/bib_ref] [bib_ref] Subclassification of the RBCC/TRIM superfamily reveals a novel motif necessary for microtubule..., Short [/bib_ref]. Many TRIMs assemble into cytoplasmic bodies, which can be dynamic structures as described with TRIM32 [bib_ref] 14-3-3 proteins sequester a pool of soluble TRIM32 ubiquitin ligase to repress..., Ichimura [/bib_ref]. The exact role of cytoplasmic bodies is not well characterized. Possible TRIM-cytoplasmic body functions may include facilitation of signaling complex assembly [bib_ref] Unanchored k48-linked polyubiquitin synthesized by the e3-ubiquitin ligase TRIM6 stimulates the interferon-ikkepsilon..., Rajsbaum [/bib_ref] , regulation of active TRIM solubility, and/or generation of autophagosome-like complexes to target proteins for degradation. Viruses can also disrupt or reorganize TRIM-cytoplasmic bodies, further supporting a role in antiviral responses. Understanding how TRIM sub-cellular localization influences activity will also benefit the field.
The applicability of the information garnered from studying TRIM-virus interactions may facilitate the identification of novel antiviral targets. Further studies will advance our understanding of the complexities involved with TRIM signaling such as isoform-specific roles, cell-type specific activity, cytoplasmic body assembly and function, and post-transcriptional modification-induced TRIM function. Finally, future studies will need to address whether other TRIMs, in addition to TRIM6, may be hijacked by viruses to directly enhance replication, acting as pro-viral factors.
[fig] Figure 1: Regulation of pattern recognition receptor (PRR) signaling by tripartite motifs (TRIMs). [/fig]
[fig] Figure 2: TRIMs in cytokine signaling. Downstream of the initial pathogen recognition and induction of pro-inflammatory cytokines, TRIMs can regulate their cytokine signaling pathways through interactions with cytokine receptor adaptors (TAB2/3) and enzymatic proteins (IKKα, IKKβ and IKKε) within the signaling complexes, the activity and stability of pathway negative regulators (Protein Inhibitor of Activated STAT 3 (PIAS3), suppressor of cytokine signaling (SOCS), and influence the transcription of various cytokine-effector genes (NF-κB-induced pro-inflammatory cytokines, signal transducer and activator of transcription (STAT)-induced genes, interferon stimulated genes (ISGs)) or cytokine signaling regulators (tumor necrosis factor (TNF) receptors (TNFR1/2, and STAT-induced genes)). [/fig]
[fig] Figure 3: The role of TRIMs in retrovirus replication. TRIM5α oligomerization into a hexagonal lattice associates directly with HIV-1 capsids to promote premature uncoating. Recognition of HIV-1 capsid by TRIM5α also triggers NF-κB/AP-1-mediated innate immune signaling via synthesis of unanchored K63-linked poly-Ub chains that activate the TAK1 kinase. One potential mechanism of TRIM5αmediated restriction could be involved ubiquitination of TRIM5α and proteasomal degradation of TRIM5α-capsid complexes. TRIM11 mobilizes cellular microtubule formation to prematurely uncoat HIV-1 and facilitate rapid release of the vRNA from the viral core, resulting in inhibition of virus replication. TRIM19 translocates to the cytoplasm and binds Daxx to prevent its degradation by the proteasome, allowing for Daxx-mediated disruption of HIV-1 reverse transcription (RT). TRIM22 inhibits Sp1, preventing LTR-mediated transcription. The connection between TRIM5α-mediated restriction and proteasomal-mediated degradation of HIV-1 requires further characterization. However, one potential model may involve improved proliferation of HIV-1-specific CD8 + T cells due to enhanced production of viral peptides from proteasome-mediated degradation of HIV-1 capsids. [/fig]
[fig] Figure 4: Effects of TRIMs on influenza virus components and IFN antagonism. The TRIM25-binding domain of the viral NS1 protein prevents TRIM25-mediated activation of RIG-I. TRIM56 has been proposed to target viral 5' triphosphate RNA and inhibit vRNA synthesis. TRIM22 targets the viral nucleoprotein, while TRIM32 prevents association of PB1 to the viral polymerase complex through K48-linked polyubiquitin chains that target the viral components for proteasome-mediated degradation. [/fig]
[fig] Figure 5: The roles of TRIMs in flavivirus infection and innate immune antagonism. Flaviviruses are internalized by receptor-mediated endocytosis and by clathrin-mediated endocytosis (1). Fusion of virus membrane with host endosomal membrane [/fig]
[fig] Figure 6: TRIM6 is targeted by Nipah and Ebola viruses to enhance virus replication. Ebola virus (EBOV) inhibits type-I IFN production by multiple mechanisms. EBOV VP35 binds and inhibits RIG-I, IKKε, and TBK-1 to inhibit IFN production. TRIM6 ubiquitinates VP35 on K309 and promotes VP35 activity as the cofactor of the viral polymerase and enhances virus replication. Additional unidentified ubiquitination sites of VP35 exist. Whether TRIM6 enhances EBOV replication by promoting viral genome replication or viral gene transcription is not known. In Nipah virus infection, the viral matrix protein (NiV-M) promotes TRIM6 degradation, resulting in reduced synthesis of K48-linked unanchored polyubiquitin chains, IKKε oligomerization, IKKε-T501 autophosphorylation, IRF3 phosphorylation, and reduced IFN induction. These combined losses confer an impaired hostantiviral response. [/fig]
|
Prognostic value of NT-proBNP in patients with severe COVID-19
Background: The outbreak of coronavirus disease 2019 caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in China has been declared a public health emergency of international concern. The cardiac injury is a common condition among the hospitalized patients with COVID-19. However, whether N terminal pro B type natriuretic peptide (NT-proBNP) predicted outcome of severe COVID-19 patients was unknown. Methods: The study initially enrolled 102 patients with severe COVID-19 from a continuous sample. After screening out the ineligible cases, 54 patients were analyzed in this study. The primary outcome was in-hospital death defined as the case fatality rate. Research information and following-up data were obtained from their medical records.Results: The best cut-off value of NT-proBNP for predicting in-hospital death was 88.64 pg/mL with the sensitivity for 100% and the specificity for 66.67%. Patients with high NT-proBNP values (> 88.64 pg/mL) had a significantly increased risk of death during the days of following-up compared with those with low values (≤88.64 pg/mL). After adjustment for potential risk factors, NT-proBNP was independently correlated with in-hospital death. Conclusion: NT-proBNP might be an independent risk factor for in-hospital death in patients with severe COVID-19.
# Background
The outbreak of coronavirus disease 2019 (COVID-caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in China has been declared a public health emergency of international concern on 30 January 2020. Despite lower case fatality rate, SARS-CoV-2 has killed more people than SARS and MERS and the number keeps growing. Epidemic studies have described that patients with severe COVID-19 were more likely to develop adverse clinical outcomes with more complications including acute respiratory distress syndrome, acute cardiac injury, acute kidney injury and shock. Investigating prognostic markers for severe patients provides insights for early therapeutic strategies.
Cardiac injury is a common condition among the hospitalized patients with COVID-19. It was recently reported that 19.7% patients from a total of 416 cases with COVID-19 had cardiac injury with more adverse clinical outcomes compared to those without cardiac injury. reported that COVID-19 patients with elevated TnT levels had higher mortality. A retrospective, single-center case series of the 138 COVID-19 patients study reported that 7.2 and 16.7% patients had complications of acute cardiac injury and arrhythmia, respectively. The fraction of acute cardiac injury and arrhythmia was even higher in severe patients with the percentage of 22.2 and 44.4%, respectively. The patients with severe COVID-19 also showed higher creatine kinase-MB (CK-MB) and hypersensitive troponin I (hs-TnI) levels than others.
A recent study demonstrated that the heart failure marker, N terminal pro B type natriuretic peptide (NT-proBNP), increased significantly during the course of hospitalization in those who ultimately died. However, there is no research concerning whether NT-proBNP predicted the outcome of severe COVID-19 patients.
# Methods
## Subjects
The study initially enrolled 102 patients with severe COVID-19 from a continuous sample in Hubei General Hospital during the management by national medical team. The study is a retrospective, observational registry with clinicaltrials.gov identifier NCT04292964. The study was also registered on Chinese medical research registration information system. All procedures were followed the instructions of local ethic committee (approval NO. 20200701). Criteria for severe conditions included respiratory rate ≥ 30/min or rest oxyhemoglobin saturation (SPO 2 ) ≤93% or oxygenation index (arterial oxygen tension/ inspired oxygen fraction, PaO2/ FiO2) ≤300 mmHg. All the data was collected using a same protocol by well-trained researchers with a doubleblind method. Patients lacking NT-proBNP results (n = 45) were excluded. Patients who had stroke (n = 2) and acute myocardial infarction (n = 1) were excluded. Other exclusion criteria including patients with malignant tumor (n = 0) and pregnancy (n = 0) were also taken account. Finally, 54 patients with COVID-19 were studied in this research.
## Baseline data and follow-up
Demographic data, clinical features and medical history were available and collected according to the patient record system. Data collection of laboratory results were defined using the first-time examination at admission (within 24 h after admission). All the laboratory data was tested in a same laboratory with the same standard. To observe the risk of in-hospital death, patients were followed up from admission to discharge (1 to 15 days). The primary outcome was in-hospital death defined as the case fatality rate. The follow-up data were collected from reviewing medical records by trained researchers using double-blind method.
# Statistical analysis
Data is presented as mean ± standard deviation, frequency (%) or median (interquartile ranges). Intergroup comparisons between NT-proBNP high group and low group were made by the independent-samples T-test (normally distributed continuous variables), Mann-Whitney U test (nonnormally distributed continuous variables) and chi-square test (categorical variables). The best NT-proBNP cut-off was that of the highest product of sensitivity and specificity for in-hospital death prediction. Cumulative survival curves of in-hospital death were estimated using the Kaplan-Meier product-limit estimation method with the log-rank test. Spearman correlation analysis was used to investigate the coefficients of NT-proBNP with selected covariates. Cox proportional hazards models were used to screening out the potential risk factors and analyzing the independent effect of NT-proBNP for in-hospital death. Statistical analyses were performed by SPSS 22.0 (SPSS, Chicago, IL, USA) and a two-sided P < 0.05 was considered statistically significant.
# Results
## Baseline characteristics
Baseline characteristics of participants were divided into two groups by low and high NT-proBNP levels (NT-proBNP≤88.64 pg/mL and NT-proBNP> 88.64 pg/ mL,according to the cut-off value determined in the ROC curve . Patients in NT-proBNP high group were significantly older with more comorbidities of hypertension (HP) and coronary heart disease (CHD), higher levels of diastolic blood pressure (DBP), myohemoglobin (MYO), CK-MB, hs-TnI, blood urea, creatinine, white blood cell (WBC), C-Reactive Protein (CRP) and procalcitonin (PCT) and lower level of lymphocyte (LYM) than the participants in NT-proBNP low group. Other characteristics like sex, temperature, pulse rate, respiratory rate, systolic blood pressure and the history of chronic obstructive pulmonary disease (COPD) and diabetes showed no significance between two groups with the different levels of NT-proBNP.
## Receiver operator characteristic (roc) curve for prediction in-hospital death
Receiver operation characteristic (ROC) curves were shown in to analyze the prognostic value and the best cut-off of NT-proBNP for prediction in-hospital death. The area under the curve (AUC) for in-hospital death was 0.909 (95%CI 0.799-0.970, P < 0.001). The best cut-off of NT-proBNP for predicting in-hospital death was 88.64 pg/mL with the sensitivity for 100% and the specificity for 66.67% .
## Cumulative survival curves of in-hospital death
Cumulative survival rate curves between two groups categorized by NT-proBNP cut-off value were shown in
## Spearman correlation coefficients of nt-probnp with selected covariates
In present study, plasma NT-proBNP was positively correlated with age, urea, cardiac injury markers of MYO, CK-MB and hs-TnI and systematic inflammation makers of WBC, CRP, Hs-CRP and PCT (Supplemental .
## Results of cox proportional hazards analyses of inhospital death
Cox proportional hazards regression analysis was used to evaluate potential associations between NT-proBNP and in-hospital death. Results of univariate analyses showed that the hazard ratio (HR) of NT-proBNP associated to in-hospital death was 1.369 (95% CI 1.217-1.541, P < 0.001) for an increase of 100 pg/mL. Meanwhile, age, male, history of hypertension (HP), myoglobin (MYO), creatine kinase-MB (CK-MB), high-sensitivity troponin-I (Hs-TnI), urea, creatinine, white blood cell (WBC), lymphocytes (LYM), c-reactive protein (CRP) and procalcitonin (PCT) were correlated with the risk of inhospital death. Multivariate Cox proportional hazards regression analyses were used to evaluate the independent prognostic effect of NT-proBNP level. After adjusting for sex and age (Mode 1), the HR of NT-proBNP for in-hospital death was 1.323 (95% CI 1.119-1.563, P = 0.001) for an increase of 100 pg/mL. After adjusting for HP and CHD history (Mode 2), the HR was 1.342 (95% CI 1.185-1.520, P < 0.001). After adjusting for MYO, CK-MB and hs-TNI (Mode 3), the HR was 1.360 (95% CI 1.177-1.572, P < 0.001). After adjusting for urea and creatinine .
# Discussion
The present study for the first time showed the relationship between plasma NT-proBNP level and the risk of in-hospital death in severe COVID-19 patients. Severe COVID-19 patients with high NT-proBNP levels tended to be older with increased cardiac injury markers and higher levels of systematic inflammation markers. Patients with high NT-proBNP (> 88.64 pg/mL) level had lower cumulative survival rate. After adjusting for potential cofounders in separate modes, NT-proBNP presented as an independent risk factor of in-hospital death in patients with severe COVID-19. Previous studies have found that NT-proBNP is a powerful and independent predictor of mortality in community-acquired pneumonia (CAP). In these studies, the best cut-off values of NT-proBNP for prediction 30-day mortality were 1434.5 pg/mL and 1795.5 pg/ mL, respectively. The elevated NT-proBNP in these cases was believed owing to the cardiac complications resulted from complex interactions among preexisting conditions, relative ischemia, up-regulation of the sympathetic system, systemic inflammation and direct pathogenmediated damage to the cardiovascular system.
However, the cutoff value of NT-proBNP to predict the adverse outcome of severe COVID-19 patients was far lower than the threshold to diagnose heart failure (450 pg/mL for < 50 years, 900 pg/mL for 50-75 years and 1800 pg/mL for > 75 years)in present study. It was suggested that the prognostic effect of plasma NT-proBNP in severe COVID-19 patients could not fully ascribe to heart failure induced by the virus or hypoxia. Further understanding of physiological and pathological significance of plasma NT-proBNP elevation in severe COVID-19 patients might help clinicians make corresponding decisions to reduce the risks of adverse outcome.
NT-proBNP is secreted in response to increased myocardial wall stress. It is also controlled by acute renal injury and proinflammatory molecules such as lipopolysaccharide, interleukin 1, C-reactive protein, and cardiotrophin I, which are independent of ventricular function. It was consisted with the study finding that NT-proBNP level was positively correlated to the makers of cardiac injury, renal injury and systematic inflammation. And these makers also constituted the risks of in-hospital death according to the univariate Cox proportional-hazards regression analysis. However, NT-proBNP was an independent risk factor after accounting these factors in multivariate Cox. The prognostic effect of NT-proBNP might be a specific index of reflecting the overall state of SARS-CoV-2 infection. The mechanism of SARS-CoV-2-induced cardiac injury was still unclear. From the result of autopsy by Xu and colleagues, a few interstitial mononuclear inflammatory infiltrates were observed in heart biopsy, indicating an inflammation induced cardiac injury. Other factors including the SARS-CoV-2 infection and invasion cardiomyocytes via the binding site of angiotensin-converting enzyme-related carboxypeptidase (ACE2), the pulmonary infection induced inadequate oxygen supply to the myocardium and the influences of cytokine storm syndromemight also contribute to the cardiac injury. All these contributes to the elevation of NT-proBNP and risks of poor prognosis in patients with COVID-19.
The virus itself may also elevate the NT-proBNP level in COVID-19 patients. SARS-CoV-2 binds with ACE2, resulting the uncontrolled releasing of angiotensin 2 (ANGII) and restricted synthesis of ANG1-7. The latter exerts anti-inflammation effect to protect tissue while ANGII plays in an opposite role and facilitates the secretion of NT-proBNP. It indicated that NT-proBNP level might associated with the severity of infection thus leading an adverse outcome, which needs further verification.
By investigating the prognostic effect of NT-proBNP level of severe COVID-19 patients at admission, it might be helpful to early identifying patients with poor prognoses. However, this study was limited by sample size and a single test of NT-proBNP at admission. Larger studies with continuous monitoring of NT-proBNP are necessary to further confirm the prognostic effect of NT-proBNP in patients with severe COVID-19.
# Conclusion
In conclusion, NT-proBNP might be an independent risk factor for in-hospital death in patients with severe COVID-19.
## Supplementary information
Supplementary information accompanies this paper at https://doi.org/10. 1186/s12931-020-01352-w.
Additional file 1:. Spearman correlation coefficients of NT-proBNP with selected covariates. |
Comparative metabolomics analysis of amphotericin B high-yield mechanism for metabolic engineering
Background: The polyene macrocyclic compound amphotericin B (AmB) is an important antifungal antibiotic for the clinical treatment of invasive fungal infections. To rationally guide the improvement of AmB production in the main producing strain Streptomyces nodosus, comparative metabolomics analysis was performed to investigate the intracellular metabolic changes in wild-type S. nodosus ZJB20140315 with low-yield AmB production and mutant S. nodosus ZJB2016050 with high-yield AmB production, the latter of which reached industrial criteria on a pilot scale.Results:To investigate the relationship of intracellular metabolites, 7758 metabolites were identified in mutant S. nodosus and wildtype S. nodosus via LC-MS. Through analysis of metabolism, the level of 26 key metabolites that involved in carbon metabolism, fatty acids metabolism, amino acids metabolism, purine metabolism, folate biosynthesis and one carbon pool by folate were much higher in mutant S. nodosus. The enrichment of relevant metabolic pathways by gene overexpression strategy confirmed that one carbon pool by folate was the key metabolic pathway. Meanwhile, a recombinant strain with gene metH (methionine synthase) overexpressed showed 5.03 g/L AmB production within 120 h fermentation, which is 26.4% higher than that of the mutant strain.Conclusions:These results demonstrated that comparative metabolomics analysis was an effective approach for the improvement of AmB production and could be applied for other industrially or clinically important compounds as well.
## Highlight
## Comparative metabolomics analysis was applied on
wild-type Streptomyces nodosus and its mutant to investigate the mechanism of amphotericin B overproduction for the first time.
2. One carbon pool by folate was confirmed as the key metabolic pathway related to amphotericin B biosynthesis. 3. The overexpression of genes MetH, GlyA and PurN in high AmB-yield strain all exerted positive impacts on AmB accumulation in comparison with the control.
# Background
Amphotericin B (AmB) is an important macrocyclic polyketide mainly produced by Streptomyces nodosus, it has various biological and pharmacological activities,
## Open access
Microbial Cell Factories such as broad-spectrum antifungal, antiparasitic, antitumor and anti-HIV activities. The fungal infections, especially invasive fungal infections, are expanding in recent years and becoming more stubborn to be cured. AmB plays a crucial role in inhibiting the growth of fungi by affecting the permeability of cell membranes and thus was used to effectively treat invasive fungal infections. In order to reduce the toxic and side effects of AmB, some strategies have been developed, including water-soluble dextrin-AmB (Dex-AmB) formulations and AmB loaded pluronic F127 (PF 127) micelles which was then coated with chitosan (Cs-PF-AmB-M). As AmB is a complex compound, the synthesis of AmB by chemical way is extremely difficult. In contrast, the biosynthesis of AmB by microorganism is relatively convenient and environmental-friendly. The gene cluster for AmB biosynthesis belongs to the type I polyketide synthase (PKS), the process of which including three steps: (i) the biosynthesis of precursors, i.e. three propionate and 16 acetate precursors, (ii) the extension of the PKS chain and (iii) the modification of post-PKS. The precursors are cyclized under the action of PKS to form the macrolactone core. On this basis, the target product AmB is then obtained by the oxidation of a methyl branch at C 41 , mycosaminylation at Cand hydroxylation at C 8 sequentially. The biosynthesis of these precursors involved in multiple primary metabolic pathways, such as central carbon metabolism, fatty acid metabolism and amino acid metabolism, etc. The variations of intracellular metabolite in organism related to the above pathways are essential to study the potential key factors for AmB biosynthesis. Recently, metabolomics-based approach has been applied as a powerful tool to measure and evaluate the dynamic multiparametric metabolic responses of living systems to the inside and outside disturbance in a comprehensive manner. For example, comparative metabolomics approach was employed to analyze metabolite concentration differences of Streptomyces tsukubaensis cultivated in two media. Metabolomics results illuminated that oxygen, S-adenosine methionine (SAM) and the transport protein were vital limiting factors for AmB accumulation in S. nodosus. Meanwhile, through optimization of fermentation process, changes of α-ketoglutarate, pyruvate and citric acid concentration were identified as the most critical metabolite nodes for AmB biosynthesis. Based on these results, further comparative metabolomics analysis can be useful for metabolic engineering of strain for AmB overproduction. Metabolic engineering has been applied based on analysis of metabolic network in Escherichia coli, Streptomyces and Corynebacterium glutamicum for the industrial production of valuable amino acids and chemicals, such as L-2-aminobutyric acid, shikimic acid, ascomycin and so on.
In this work, comparative metabolomics analysis was employed to analyze the intracellular metabolite variations between mutant and the wild-type S. nodosus. The potential key metabolites that might be responsible for AmB overproduction were identified for the first time. Afterwards, the variations of these metabolite were evaluated and analyzed to reveal the possible factors that influencing the AmB biosynthesis. Because the changes of metabolites reflected the changes of metabolic pathway fluxes, gene overexpression strategy at key metabolic nodes was carried out to further improve AmB production based on the above analysis.
# Results
## Differences of fermentation kinetics between the wild-type and mutant s. nodosus
To find out the internal changes of mutant S. nodosus, fermentation kinetics for AmB production was studied. Wild-type and mutant strains under the same cultivation condition showed the same fermentation cycle which could be divided into lag phase (phase I, 0-24 h), logarithmic phase (phase II, 24-108 h), stationary phase (phase III, 108-132 h) and decline phase (phase IV, 132-168 h). In the lag phase, pH values in the medium showed no significant differences, but AmB could be detected in mutant S. nodosus at 24 h while no AmB was detected in wild-type S. nodosus. In addition, the biomass increased rate and total sugar consumption rate of wildtype S. nodosus were faster than that of mutant S. nodosus. The average sugar consumption rate between two strains showed significant discrepancy between 24 and 48 h, which indicate the metabolism distinction between two strains.
In the logarithmic phase, the biomasses of two strains increased rapidly and the pH values declined. Meanwhile, the average sugar consumption rate between 60 and 108 h remained stable. Additionally, the production of AmB at 72 h reached about 0.5 g/L in wild-type S. nodosus and 2.8 g/L in mutant S. nodosus, respectively, and the difference of AmB production in two strains began to be noticeable (> 2 g/L) at this time.
In the stationary phase, the biomasses and the pH values between two strains were obviously different, especially for the average sugar consumption rate, which increased rapidly in mutant S. nodosus. To be noticed, a large percent of AmB was accumulated in this phase, indicating that this phase was the principal phase for AmB production in wild-type and mutant strains. In the decline phase, the yield of AmB began to decrease and the pH increased significantly in wild-type and mutant S. nodosus.
Taken together, the fermentation advantages of mutant S. nodosus for AmB production were obvious, with a 26.4% increase for AmB production. However, the potential factors that affect AmB production were obscure and could not be known through the analysis of fermentation kinetics. It was necessary to fully study the changes of intracellular metabolites as well as the strain metabolism discriminations between the wild-type and the mutant strain.
# Comparative metabolomics analysis
To investigate the discrepancies of intracellular metabolites between wild-type and mutant strain, the raw data of LC-MS were imported into Progenesis QI software, results indicated that 7758 compounds were identified. Here, PCA analysis examined the stability of the process and the overall distribution between samples. (O)PLS-DA distinguished more clearly for the sample groups and could improve the model effectiveness. Concretely, in order to assess the accuracy and stability of the equipment state during detection and collection process, the points of the quality control (QC) samples were closely clustered together in the PCA score plot. Most samples within 95% confidence intervals were close to each other and the sample repeatability qualities were reliable. The differences were small when the sample marks gathered, whereas the differences were great when the sample marks dispersed.
In the (O)PLS-DA score plot, both groups of samples marks were distributed on different sides of the Y axis in each fermentation stage, and the distinction was significant. The results showed that (O)PLS-DA method analysis could reduce the difference between the same group of samples and make the difference between different groups of samples more significant. The significant discrimination on the (O)PLS-DA score plot profilingof 0.164 based on the metabolic analysis and AmB production. In order to assess the accuracy and stability of the equipment state during detection and collection process, the QC samples were prepared in advance and tested every 10 samples revealed the metabolic discrimination of the wild-type strain and its mutant strain. The variable importance in the projection plot demonstrated the contribution of each identified metabolite to sample class separation and AmB accumulation. Generally, a metabolite with a VIP value greater than 1 indicated a significant contribution to the separation of sample groups within (O)PLS-DA model. The higher the VIP value of a metabolite, the higher contribution of this metabolite was shown for the AmB accumulation. The identified metabolites with VIP values greater than 1 were listed in Additional file 1: , and most of them were positively correlated to the AmB yield. These metabolites (VIP > 1) were mainly involved in carbon metabolism, fatty acid metabolism, amino acid metabolism, purine metabolism, folate biosynthesis and one carbon pool by folate. To explore the key factors involving in the AmB production, the key metabolites with VIP values greater than 1 were comparatively analyzed.
Significant discrepant metabolites were identified among four groups with VIP > 1 and p value < 0.05. Results indicated that 674 differential metabolites between mutant S. nodosus and wild-type S. nodosus were identified at 24 h, 613 differential metabolites at 72 h, 721 differential metabolites at 120 h, and 386 differential metabolites at 156 h. In this work, 60 differential metabolites were determined to display a marked difference between two strains according to the statistical analysis of differential metabolites at 24 h, 72 h, 120 h and 156 h of fermentation process. Metabolite data were analyzed to study the differences during the whole fermentation process of S. nodosus, including carbon metabolism, fatty acid metabolism, amino acid metabolism, purine metabolism, folate biosynthesis and one carbon pool by folate.
## Comparative metabolites in central carbon metabolism analysis
The results of (O)PLS-DA analysis indicated that glucose, lactate, tagatose-6-phosphate, sorbitol, sedoheptulose-7-phosphate and succinate had greater contribution (VIP > 1) to the AmB accumulation. The changes of intracellular metabolites involved in carbon metabolism were shown in. Compared to wild-type S. nodosus, the levels of glucose, tagatose-6-phosphate, sorbitol and intermediate metabolites sedoheptulose-7-phosphate were higher in mutant S. nodosus, and they all decreased during the main period of amphotericin biosynthesis (24-120 h). This indicated that mutant S. nodosus could utilize sugar better than the wild-type strain. However, the levels of lactate were very little in mutant S. nodosus, and it showed a decreasing trend in wild-type S. nodosus during the logarithmic period (24-72 h). One possible explanation for above observations was that lactate was a competitive branch of pyruvate metabolism flowing to propionic acid and acetyl-CoA. Reducing the accumulation of lactate would result to increased acetyl-CoA pool which was then entered into the TCA cycle.
## Comparative metabolites in fatty acid metabolism analysis
Compared to wild-type S. nodosus, the levels of major fatty acids and phospholipids were higher in mutant S. nodosus, and they all decreased during the main period of amphotericin biosynthesis (24-120 h). It has been reported that under aerobic conditions, metabolism of fatty acids can decompose and release a large amount of energy and acetyl-CoA, which can be carboxylated to form malonyl-CoA. To be noticed, acetyl-CoA, malonyl-CoA and methylmalonyl-CoA were important precursors for the synthesis of AmB.
Hexadecanoic acid, heptadecanoic acid, myristic acid, pentadecanoic acid, phosphatidic acid and triglycerides (TGs)were the key metabolites that affected metabolism and the production of AmB according to (O) PLS-DA analysis. Phosphatidic acid is a component of cell membrane as a prerequisite for the synthesis of phosphoglycerides, and its corresponding phosphate could be combined with some molecules to form a variety of phospholipids.
TGs are lipids formed by glycerol and fatty acids, and it is reported that triacylglycerol (TAGs) plays a key role in linking primary metabolism to polyketide synthesis in streptomyces. Intracellular TAGs, which were accumulated in primary metabolism, could be degraded during stationary phase in mutant S. nodosus. This process could channel carbon flux from both intracellular TAGs and extracellular substrates into polyketide biosynthesis pathway. Compared with mutant S. nodosus, concentrations of TGs in wild-type S. nodosus were almost unchanged. It showed that the synthesis of polyketide was weak in wild-type S. nodosus.
Therefore, it could be speculated that the fatty acids could regulate cell metabolism from primary to secondary metabolism and then affected the synthesis of AmB polyketide.
## Comparative metabolites in amino acid metabolism analysis
The results of (O)PLS-DA analysis indicated that methionine, cysteine, SAM, arginine, tetrahydrodipicolinate and glutathione were the key metabolites that affected cell metabolism and the AmB biosynthesis. SAM is an important molecule to maintain normal cell function and survival, involved in three important metabolic pathways: transmethylation, transmercaptylation and The VIP values represented the contribution of metabolites, and the higher the VIP value of a metabolite, the more contribution that this metabolite showed. The variables with VIP value greater than 1 are separated from the others by a red line polyamine synthesis. A number of literatures have been reported that gene metK (coding SAM synthetase) can improve the synthesis of SAM and increase the production of secondary metabolites in Streptomyces. Compared to wild-type S. nodosus, the levels of methionine supply were especially insufficient in mutant S. nodosus. Methionine is a precursor to SAM, and gene metH (methionine synthase) has been reported to improve methionine synthesis.
Similarly, compared to wild-type S. nodosus, the levels of glutathione (GSH) were higher in mutant S. nodosus, and they all decreased during the main period of amphotericin biosynthesis (24-120 h). It is reported that GSH maintains bacteria homeostasis through enzymatic reactions and participates in metabolic redox reactions. In addition, GSH plays an important role in cell resistance to high oxygen, high osmotic pressure, metal ions and environmental changes during the fermentation process.
Results also showed that the level of arginine decreased a lot in the fermentation early stage of mutant S. nodosus, and arginine could be metabolized to spermine containing two amino groups and two imino groups. Spermine can make DNA molecules more flexible and stable in bacteria and is deemed as an important substance to promote cell proliferation. Moreover, a high level of tetrahydrodipicolinate was measured in mutant S. nodosus, while tetrahydrodipicolinate is involved the formation of racemic 2,6-diaminyl heptadiic acid, which can be used in the biosynthesis of lysine and also be used as one of the bacterial peptidoglycan components to form the cell wall structure. These results indicated that changes of amino acid metabolism in mutant S. nodosus would help the strain for growth and thus promote the biosynthesis of AmB.
## Comparative metabolites in purine metabolism analysis
Nucleotides are mainly involved in the formation of nucleic acids, and many single nucleotides also involved in important biological functions. Compared to wildtype S. nodosus, the levels of purine metabolic related chemicals showed more obvious decreasing trend in mutant S. nodosus, such as glutamine, 5′-phosphoribosyl-N-formylglycinamide (FGAR) and 5-aminoimidazole ribonucleotide (AIR). Additionally, there was little difference for the 5-aminoimidazole-4-carboxamide (AICA) between the wild-type S. nodosus and mutant S. nodosus.
## Comparative metabolites in folate biosynthesis and one carbon pool by folate analysis
The results of (O)PLS-DA analysis indicated that tetrahydrofolic acid, methyl-THF and methylene-THF showed significant differences between wild-type and mutant strains, and they might be the key metabolites that affected cell metabolism and the AmB production. Compared to wild-type S. nodosus, the levels of tetrahydrofolic acid and methyl-THF were higher in mutant S. nodosus, and the level of methyl-THF decreased during 24 to 72 h, while the level of tetrahydrofolic acid decreased during 72 to 120 h. Almost all the synthesis of nucleic acids, amino acids and pantothenic acids in bacteria required tetrahydrofolic acid as a carrier of carbon units. Combined with the results of methionine metabolism and purine metabolism, it might be concluded that tetrahydrofolic acid was the key metabolite in the mutant S. nodosus. The role of tetrahydrofolate in the one carbon pool by folate is based on the conversion between methyl-THF and methylene-THF, and tetrahydrofolate can be obtained by the dihydrofolate reduction catalyzed by dihydrofolate reductase.
## The relationships between intracellular metabolites and amphotericin accumulation
To find out the important metabolic pathways for AmB biosynthesis, pathway enrichment analysis and metabolite addition strategies were further employed. The differential metabolites were then analyzed through pathway enrichment based on KEGG database (see Additional file 1:and the effects of different addition strategies on the production of AmA and AmB were studied, finally, 26 intracellular metabolites were chosen (as shown in Additional file 1: . Compared to wild-type S. nodosus, the levels of these 26 key metabolites were higher in mutant S. nodosus, and almost all of them showed decreasing trends during the main fermentation period for amphotericin biosynthesis (24-120 h). These metabolites involved in carbon metabolism, fatty acid metabolism, amino acid metabolism, purine metabolism, folate biosynthesis and one carbon pool by folate, etc. The relationships between these metabolism pathways and amphotericin biosynthesis were revealed and summarized in .
## Gene overexpression strategies for amb improvement
In this study, mutant strain was obtained by UV-NTG mutagenesis of wild-type S. nodosus. The mutations in the genome caused by UV-NTG treatment resulted to the metabolite change in mutant strain. In brief, the changes of metabolite were due to the changes of metabolic flux, which were controlled by gene expression. Hence, we firstly employed the comparative metabolism analysis between mutant and wild-type S. nodosus to find out the potential key metabolites. To further improve the AmB production, strategies including addition of the critical metabolites or overexpression of the key genes in the pathway were employed, in which the latter would be more efficient.
We had studied the obvious different intracellular tetrahydrofolate levels between wild-type and mutant strains, inadequate supply of methionine and continuously decreased S-adenosylmethionine level in mutant S. nodosus, which helped to identify that one carbon pool by folate metabolic pathway might be the key metabolic pathway. To verify these results, 50 mg/L methionine, serine, glycine, SAM and THF were selected to study their effect on AmB production in mutant S. nodosus. As shown in Additional file 1:, methionine, serine, glycine, SAM and THF showed positive affections for AmB biosynthesis which had been reported by our group. Besides this, several gene expression vectors had been constructed (gcvT, thyX, fmt, metF, purH, metH, purN, DHFR and glyA), these genes were overexpressed under the control of promoter ermE*p from Saccharopolyspora erythraea. The expression vectors were introduced into the mutant S. nodosus by optimized conjugative transferred method and the engineered strains were successfully obtained. Strain that overexpressed metH showed a 5.03 g/L AmB yield at 120 h in flask shake fermentation, about 26.4% higher than that of the mutant strain S. nodosus. RT-qPCR analysis showed the increased transcriptional levels for purN, metH, glyA, metF, purH, and fmt genes in the engineered strains (Additional file 1:.
# Discussion
In previous work, we obtained a mutant strain named S. nodosus ZJB2016050 with high AmB yield by environmental/chemical mutagen (UV+ N-methyl-N-nitroso-N-nitroguanidine) treatment of wild-type S. nodosus ZJB20140315. Then, new strategies with overexpressing some key genes involved in oxygen-taking or precursor-acquiring were carried out in our lab. Based on comprehensive analysis of key metabolites and pathways during different AmB fermentation time in mutant strain, we have identified 28 metabolites as key factors and 6 pathways were closely associated with AmB production and studied to optimize the biosynthesis of AmB. However, the results of genes overexpression did not support metabolome analysis results strongly, which indicated the complexity of the metabolic network. Meanwhile, the metabolome analysis results of a single strain were not enough as they only provided the metabolite changes at different fermentation times. What's more, the differences of the two strains at the same fermentation time were obtained more intuitively, The proposed relationships of amphotericin biosynthesis with relevant metabolic pathways in S. nodosus. The key metabolic pathways were marked by red, including central carbon metabolism, fatty acids metabolism, purine metabolism and one carbon pool by folate. EMP Embden-Meyerhof-Parnas pathway, TCA tricarboxylic acid cycle, Glu glucose, G6P glucose 6-phosphate, F6P fructose 6-phosphate, G3P glucose 3-phosphate, G2P glucose 2-phosphate, PEP phosphoenolpyruvate, Pyr pyruvate, CIT citrate, ISO-CIT iso-citrate, AKG α-ketoglutarate, SUC-CoA succiny-CoA, SUC succinate, FUM fumaric acid, MAL malic acid, OAA oxalacetic acid, Lac lactate, Ser serine, Gly glycine, Cys cysteine, Glu glutamic, Gsh glutathione, Asp aspartic, Dap diaminopimelate, Lys lysine, His histidine, THF tetrahydrofolic, Met methionine, SAM S-adenosine methionine, SAH S-adenosyl-l-homocysteine, PRPP 5-Phospho-alpha-d-ribose-1-diphosphate, PR-AMP phosphoribosyl-adenosine monophosphate, GAR glycinamide ribonucleotide, FGAR formyl-glycinamide ribonucleotide, FGAM formyl glycinamidine ribonucleotide, AIR 5-aminoimidazole ribonucleotide, AICAR 5-aminoimidazole-4-carboxamide ribonucleotide for example, compared with wild-type S. nodosus, the level of substances related to fatty acid metabolism especially the TG metabolism decreased obviously in mutant S. nodosus. This result indicated that oxygen supply was a crucial factor for providing sufficient precursors for AmB biosynthesis. In addition, some results showed that SAM precursor-methionine was inadequately supplied and we found obvious different tetrahydrofolate levels between the mutant and wild-type strain, which could not be obtained from previous metabolome analysis. Hence, the comparative metabolome analysis between mutant S. nodosus and wild-type S. nodosus was necessary to obtain the potential clues for the high-production of AmB by mutant strain.
In this study, in order to rationally guide the improvement of AmB production, the discrepancy between intracellular metabolism and AmB accumulation in mutant and wild-type S. nodosus were studied by comparative metabolomics analysis. The results of (O)PLS-DA analysis indicated that there were strong correlations between intracellular metabolites and AmB accumulation.
In carbon metabolism, the decrease of lactate levels could improve acetyl-CoA pool which was then entered to the TCA cycle. As an important intermediate metabolite of primary metabolites, sedoheptulose-7-phosphate was also considered as a node for AmB production. In fatty acid metabolism, most of the fatty acids could be catabolized to release large amount of energy and acetyl-CoA, phosphatidic acid with high level in mutant strain is an important component of the cell membrane and participates in the recognition and signal transduction of membrane proteins. In addition, TGs can regulate cell primary metabolism to secondary metabolism, thereby affecting the synthesis of secondary metabolite AmB. In the study of amino acid metabolism, GSH maintained a relative high level at the fermentation early stage in mutant strain, as GSH could keep the cell intracellular redox balance through enzymatic reactions which helped strain adapt to environmental changes. As a methyl donor of the secondary metabolite, SAM was a key metabolite and was found to be lack at the fermentation early stage for AmB biosynthesis. In the study of purine metabolism, purine and pyrimidine were growth factors, but significant down-regulations of FGAM, AIR and other precursors in the biosynthesis of AICAR were showed in the mutant strain. In the one carbon pool by folate, the levels of tetrahydrofolic acid and methyl-THF were higher in mutant S. nodosus than in the wild-type strain. All these results showed the advantages in mutant S. nodosus for high AmB production. As most of nucleic acids, amino acids and pantothenic acids in bacteria required tetrahydrofolic acid as a carrier of carbon units, tetrahydrofolic acid was deemed as the key metabolite in the mutant S. nodosus.
Moreover, 26 key metabolites that contributed to AmB productivity were identified, most of which were related to carbon metabolism, fatty acid metabolism, amino acid metabolism, purine metabolism and folate biosynthesis. The relationships between these metabolism and amphotericin biosynthesis are summarized in . According to the analyses and results of comparative metabolomics, one carbon pool by folate metabolic pathway was identified as the key metabolic pathway. Based on this, further study of the effects of precursors and key metabolites including serine, glycine and SAM addition on AmB production were carried out in our previous studiesand the addition of methionine and THF showed improved AmB yield (Additional file 1:. Later, gene overexpression strategies were conducted in mutant S. nodosus to improve AmB production according to the above results. We screened the genes that were involved in one carbon pool by folate metabolic pathway. Through overexpression of single gene purN, metH, glyA, metF, purH or fmt, a strain with gene metH overexpression was obtained with its optimal AmB yield at 5.03 g/L in 120 h fermentation, which increased by 26.4% compared with that of the mutant strain S. nodosus in flask shake fermentation. However, the yield of by-product AmA in those recombinant bacteria was unstable, as AmA decreased in the strain with purN overexpressed and obviously increased in the strain with other genes overexpressed. Recently, Huang et al.speculated that the decrease synthesis of AmA was related to the insufficient NADPH or the weak binding force between ER5 domain and NADPH. Based on critical role of ER5 domain of amphC in the biosynthesis of AmA and AmB reported by Cafrey, reducing the by-product AmA would become the focus of our subsequent studies.
# Conclusions
Overall, based on the comparative metabolome analysis between mutant and wild-type S. nodosus, one carbon pool by folate metabolic pathway was identified as the key metabolic pathway in the mutant strain. By choosing the genes related to the one carbon pool by folate metabolic pathway for overexpression, three recombinant strains with the overexpression of key metabolic genes (purN, glyA and metH) in S. nodosus were obtained and they showed positive impacts on AmB yield in flask fermentation, which is 22.8% to 26.4% higher than that of the original mutant S. nodosus. The production of AmB in metH overexpression strain reached 5.03 g/L within 120 h fermentation time, about 26.4% higher than that of mutant S. nodosus, which needed 144 h fermentation time for almost the same AmB production titer. Our results indicated the application of gene overexpression strategy based on the comparative metabolomics analysis would be very effective for the rational design of other antibiotics production improvement.
# Methods
## Bacterial strain, medium, and culture conditions
S. nodosus ZJB2016050 (CCTCC M2017426, China Center for Type Culture Collection, Wuhan, China) used throughout this study was isolated after N-methyl-Nnitroso-N-nitroguanidine (NTG) and ultraviolet (UV) mutagenesis of wild-type S. nodosus. The detailed ultraviolet mutagenesis method was reported by Zhang and coworkers. All strains and recombinant plasmids with characteristics and resources used in this article were listed in Additional file 1: .
The composition of the fermentation medium contained 69 g/L glucose, 25 g/L beef extract, 9 g/L CaCO 3 and 0.1 g/L KH 2 PO 4 . The composition of the seed medium contained 15 g/L tryptone, 10 g/L yeast extract, 10 g/L glucose, 5 g/L NaCl and 1 g/L CaCO 3 . Luria-Bertani (LB) medium and GYM medium were used for E. coli and S. nodosus cultivation, respectively. The composition of the LB medium contained 10 g/L tryptone, 5 g/L NaCl, 5 g/L yeast extract (additional 20 g/L agar for solid medium). The composition of the GYM medium contained 4 g/L glucose, 4 g/L yeast extract, 10 g/L malt extract, 2 g/L CaCO 3 and 20 g/L agar. The composition of the MS medium contained 20 g/L soybean meal, 20 g/L mannitol and 20 g/L agar. Before autoclaving, the pH value of the fermentation, seed and LB medium was adjusted to 7.0, and the pH value of the GYM and MS medium was adjusted to 7.2. The agar medium was prepared by adding soybean meal into a certain volume of distilled water and boiling for 1 h, then the volume was filled up to 1 L by distilled water. Finally, mannitol and agar were quantitatively added into the medium and autoclaved. All used primers are listed in Additional file 1: .
The strain was cultivated on MS agar plate at 26 °C for 4-6 days to isolate a single colony. The isolated pure culture was then grown on the GYM medium for 7 days for the harvest of spores. The 1 mL resulting spore suspension was then transferred into a 50 mL seed medium in a 250-mL shake flask and incubated for 48 h at 26 °C. Seed cultures (2 mL) were then inoculated into 50 mL fermentation medium in a 500-mL shake flask and incubated at 26 °C for 168 h.
# Analytical methods
AmB concentration was measured by high-performance liquid chromatography (HPLC) (3200; LDC ANALYTI-CAL INC., New York, USA) equipped with a Agilent C18 reversed-phase column (5 μm, 4.6 × 150 mm, Agilent Technologies Inc., USA) and a UV-vis detector. Culture samples (0.15 mL) mixed with the nine times volume DMSO was shaken intermittently in a rotary shaker at 50 °C for 30 min. After centrifugation at 12,000g for 2 min, the supernatant filtered through a 0.22 μm organic filter was subjected to HPLC analysis. The mobile phase was comprised of water/acetonitrile/methanol (9:7:4, v/v/v) with a flow rate of 1.0 mL/min. the pH value of the mobile phase was adjusted to 5.0 with glacial acetic acid. The column temperature was 25 °C and UV detector was set at 304 nm (0-12 min) and 405 nm . The AmB standard sample was purchased from Sigma-Aldrich (CAS: 1397-89-3).
The dry cell weight (DCW) of strains were measured as follows. In detail, 1 mL of culture was sampled and centrifuged at 12,000g for 10 min. To remove calcium carbonate, appropriate amount of acetic acid was added and reacted for 10 min. The mycelium was washed twice with sterilized water, centrifuged at 12,000g for 10 min, and dried to constant weight at 80 °C. The residual sugar in the fermentation broth was quantified using a commercial glucose assay kit. The pH value was detected using a commercial pH meter.
## Sample preparation for metabolic analysis
According to the AmB production curve, samples for metabolic analysis were harvested at 24 h when the AmB accumulation began to be divergent (> 0.5 g/L) between mutant S. nodosus and wild-type S. nodosus. Samples were taken at 24, 72, 120 and 156 h, respectively.
For metabolites extraction, 2 mL cultures were sampled and centrifuged below 4 °C at 12,000g for 5 min. the supernatant was collected and washed three times with 1.5 mL cold saline, centrifuged again below 4 °C at 12,000g for 5 min. The supernatant was collected and froze rapidly in liquid nitrogen, then stored at − 80 °C. Before testing, the samples were placed at − 20 °C for 30 min, and then thawed in an ice bath for 4 min. Another 0.8 mL cold methanol/water (1:1, v/v) and two steel beads were added into the cell pellet.
Cell samples were placed in the TissueLyser at 35 Hz and grind for 4 min. After grinding, the samples were precipitated at − 20 °C for two hours, and centrifuged at 12,000g for 20 min at 4 °C. The supernatant was then taken out for liquid chromatography-mass spectrometry (LC-MS) analysis.
# Lc-ms analysis
LC-MS was performed to analyze the metabolite samples using a Waters 2777C UPLC system (Waters, Milford, USA) equipped with an ACQUITY UPLC BEH C18 spectrum column (100 mm × 2.1 mm, 1.7 μm, Waters, Milford, USA). The spectrum column temperature was 50 °C and flow rate was set at 0.4 mL/min. The metabolites were eluted with the following gradient: 0-2 min, 100% mobile phase A; 2-11 min, 0-100% mobile phase B; 11-13 min, 100% mobile phase B; 13-15 min 0-100% mobile phase A. mobile phase A is deionized water with 0.1% formic acid, and the mobile phase B is methanol with 0.1% formic acid. The mass spectrometer model was high-resolution tandem mass spectrometry Xevo G2-XS QTOF (Waters, Milford, USA). Small molecules washed from liquid chromatography columns were collected using the positive ion mode and negative ion mode modes. The capillary voltage and conical hole voltage in two positive and negative ion modes are 3 kV, 40 V and 1 kV, 40 V., respectively centroid data were collected using full-information tandem mode, the molecular size of primary mass spectrometry scanning was 50-1200 Da, scanning time was 0.2 s. Then break all the parent ions with 20-40 eV set energy, and the LE signal correction is performed every 3 s for all the collected fragments scanning time 0.2. Normally, to assess the accuracy and stability of the equipment during detection and collection process, QC samples (mixture of all samples) were prepared in advance and tested every 10 samples. Perform data alignment and normalization for the complete data set, composed of multiple analytical blocks.
The LC-MS raw data were analyzed by Progenesis QI (Waters Corporation, Milford, USA) software using the following parameters. The parameters used were retention time range 0.5-14.0 min, mass range 50-1000 Da, mass tolerance 0.01 Da. Isotopic peaks were excluded for analysis, noise elimination level was set at 10.00, minimum intensity was set to 15% of base peak intensity, and finally, RT tolerance was set at 0.01 min. The Excel file was obtained with three-dimension data sets including m/z, peak RT and peak intensities, and RT-m/z pairs were used as the identifier for each ion. The resulting matrix was further reduced by removing any peaks with missing value in more than 60% samples. The internal standard was used for data QC. The positive and negative data were combined to get a combine data set which was imported into SIMCA-P+ 14.0 software package (Umetrics, Umeå, Sweden). And then the data was processed by two different standardized methods such as centralization (mean-centering) and automatic specification (unit variance scaling, UV).
LC/MS raw data uses Progenesis QI software (Waters Corporation, Milford, USA) for baseline filtering, peak identification, integration, peak alignment, retention time correction, standardization and normalization, and finally a retention time, mass-to-charge ratio and peak are obtained Intensity data matrix. 7758 metabolites were identified under the scoring standard of Progenesis QI software. Use the p value obtained by univariate analysis and the VIP value obtained by multivariate analysis to screen for differential metabolites. The screening conditions are p value < 0.05 (T-test) and VIP > 1 (Simca-p 14.1), which satisfy both the metabolites of the condition are the differential metabolites. The selected differential metabolites and their corresponding genes are integrated into the data and used as the raw data of the Metabo-Analyst software (https ://www.metab oanal yst.ca). Enter the MetaboAnalyst software homepage, select the "Joint pathway Analysis" module, upload the data, enter the above processed gene list and the corresponding metabolite list data, perform ID mapping on the designated gene and metabolite identifier, and then perform the gene with the KEGG database map with metabolites to get the result table, click submit to enter the parameter analysis, and get the pathway enrichment analysis result. It is similar to the metabolic pathway analysis result figure (Additional file 1:. We plotted the top 10 metabolic pathways with the lowest p value (Additional file 1:.
## Chemical reagent
Unless otherwise noted, all the reagents and solvents used in this study were purchased from Sigma Chemical Company (St. Louis, MO, USA) and Sangon Biotech (Shanghai, China) at the analytical grade.
# Data analysis
All the experiments were carried out for at least five replicates, and the values were shown as mean ± standard deviations. The p value was performed with t tests and fold change analysis to analyze the data obtained from LC-MS. The levels of metabolites in mutant S. nodosus and wild-type S. nodosus were considered significantly different at p value < 0.05.
Principle component analysis and (orthogonal) partial least-squares-discriminant analysis were carried out to visualize the metabolic alterations among experimental groups, after mean centering and unit variance scaling. Variable importance in the projection ranks the overall contribution of each variable to the (O)PLS-DA model, and those variables with VIP > 1 are considered relevant for group discrimination. In this study, in order to guard against overfitting, the default 7-round cross-validation was applied with a seventh of the samples being excluded from the mathematical model in each round. |
The Incidence and Health Economic Burden of Ischemic Amputation in Minnesota, 2005-2008
# Introduction
Lower extremity peripheral artery disease (PAD) is defined as blockages in the major arteries that supply the leg. PAD is 1 of the 3 major systemic atherosclerotic diseases and has a prevalence similar to that of coronary heart disease and ischemic stroke. People with PAD may be asymptomatic but often experience claudication as the primary ischemic symptom that causes exertional leg muscle pain. When atherosclerotic artery blockages in the leg progress and become severe or when a clot travels to completely block a leg artery (arterial thromboembolism), unremitting ischemia at rest can occur. This presentation of PAD is defined as critical limb ischemia (CLI) and signifies an imminent risk of amputation if revascularization is not achieved. The etiology of more than 80% of amputations that occur in all developed nations, including those that occur in people with diabetes, are due to PAD and CLI [bib_ref] Lower limb amputations in southern Finland 1984-1985, Pohjolainen [/bib_ref]. The incidence of CLI will likely increase as the population ages because of the increasing prevalence of diabetes. If so, these preventable, morbid, and costly limb ischemic events will demand greater medical and revascularization resources [bib_ref] The influence of arterial reconstructive surgery on the outcome of critical leg..., Luther [/bib_ref]. This burden has an effect at the family, city, and state levels.
Major lower extremity amputation occurs at an estimated rate of 120 to 500 per million per year in Western countries [bib_ref] Major amputations: clinical patterns and predictors, Dormandy [/bib_ref] , and this rate increases with age [bib_ref] Impact of arterial surgery and balloon angioplasty on amputation: a populationbased study..., Hallett [/bib_ref]. The rate is estimated at 25% in patients with CLI. Ischemic amputation is still a commonly performed operation, even in hospitals that emphasize aggressive revascularization for limb salvage, and occurs more often in the United States among blacks than whites [bib_ref] Do differences in hospital and surgeon quality explain racial disparities in lower-extremity..., Regenbogen [/bib_ref]. The age-adjusted annual rate of lower extremity amputation in the United States was reported to remain stable at approximately 30 per 100,000 per year from 1979 to 1989, despite an increase in the annual rate of percutaneous transluminal angioplasty and peripheral bypass surgery [bib_ref] The use of angioplasty, bypass surgery, and amputation in the management of..., Tunis [/bib_ref]. Hallett et al reported a decrease in the rate of major (but not minor) lower extremity amputation (from 36.7 to 19.0 per 100,000 annually from 1973 to 1992), while the rate of percutaneous and operative revascularization increased during the same period [bib_ref] Impact of arterial surgery and balloon angioplasty on amputation: a populationbased study..., Hallett [/bib_ref]. More recent data from Goodney et al suggest a decrease in major lower extremity amputation rates and lower extremity bypass surgery, with an associated increase in rates of endovascular procedures in older adults [bib_ref] National trends in lower extremity bypass surgery, endovascular interventions, and major amputations, Goodney [/bib_ref]. These conflicting estimates use different case ascertainment strategies, and we are not aware of any effort in North America or Europe to conduct surveillance of ischemic amputation rates.
The aim of this investigation was to create a surveillance methodology for the incidence of ischemic amputation in Minnesota and to identify potential geographic and demographic disparities in incidence. This method was also designed to permit creation of a conservative estimate of the direct costs and health economic burden associated with ischemic amputation at the state level. We hypothesized that there would be heterogeneity in ischemic amputation rates on the basis of geographic regions (eg, urban vs rural); that ischemic amputation would represent a significant cardiovascular disease burden compared with both coronary heart disease (CHD) and stroke mortality; that hospital-related charges and costs would be comparable to those for CHD and stroke; and that trends in the incidence of amputation may be evident during a 4-year surveillance period. We intended to close a knowledge gap on ischemic amputation and describe a surveillance strategy that could be modeled by other states.
# Methods
We used data from a claims-based administrative data set compiled annually by the Minnesota Hospital Association, using the Uniform Billing Code of 1992 and the Uniform Billing Code of 2004 established by the National Uniform Billing Committee. This claims dataset contains information on inpatient hospitalizations at participating hospitals in Minnesota, including approximately 97% of the staffed inpatient beds at acute care hospitals. Information on Minnesota residents hospitalized outside the state was limited to the bordering states of North Dakota, South Dakota, and Iowa; information on hospitalizations in neighboring Wisconsin was not available. We used discharges that took place from January 1, 2005, through December 31, 2008. The dataset includes disease and procedure codes from the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM), and patient sex, age, residential zip code, county of residence, date of admission, date of discharge, payer type, total dollar charges, and discharge disposition.
We identified lower limb amputations through ICD-9-CM codes. For each hospitalization record, up to 24 disease codes and 8 procedure codes are present. We first identified hospitalizations with any diabetes or cardiovascular disease code in the primary or secondary positions (ICD-9-CM 250.xx or 390.xx-459.xx), excluding patients with an injury code in the primary position (ICD-9-CM 800. xx-999.xx). We then identified lower limb ischemic amputations by ICD-9-CM procedure codes for lower extremity amputation (ICD-9-CM 84. , excluding cases with an ICD-9-CM disease code for cancer (including malignant neoplasm of the bone, malignant neoplasm of skin, melanoma, and Kaposi's sarcoma). We further classified amputations as major (at or above the ankle) or minor (below the ankle). For comparison, we also identified CHD, stroke, and transient ischemic attack (TIA) hospitalizations by the presence of ICD-9-CM disease codes 410.xx-414.xx (CHD) and 430.xx-438.xx (stroke and TIA) in the primary position.
To estimate the total costs for each hospitalization, we adjusted total dollar charges by the hospital-specific cost-to-charge ratio at the time of patient discharge. For hospitalizations at facilities with no cost-to-charge ratio, depending on the hospital's location, we used the average rural hospital or urban hospital ratio for the state. Hospitalizations in South Dakota (n = 69) were not identified by facility, so we substituted the urban hospital cost-
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
to-charge ratio, on the basis of known patterns of care for Minnesota residents hospitalized in that state. We applied this adjustment to estimate the total resource cost associated with inpatient care only, as it does not reflect costs associated with out-of-hospital imaging, procedures, officebased care, outpatient rehabilitation, or use of limb prostheses or chronic medications [bib_ref] 42 CFR Parts 411, 412, 413, 422, and 489 Medicare Program; changes..., Us Department [/bib_ref] Continuous data are presented as medians (interquartile range); categorical data are presented as frequencies. Both inpatient hospitalization charges and costs are presented using medians, means, and sums. All analyses were conducted using SAS software version 9.2 (SAS Institute, Inc, Cary, North Carolina).
# Results
There were 4,302 hospitalizations for lower limb amputations among Minnesotans from 2005 through 2008 . Of these, 1,831 (43%) were major, 2,470 (57%) were minor, and 1 was unspecified. The median patient age was 67, and major amputations were more likely to occur in older people. Most amputations took place among men and people who lived in an urban county. Of total amputations, 72% of cases were performed in people with diabetes, most of whom had a minor amputation. The number of CHD hospitalizations declined during the study period, while the number of ischemic amputations and stroke hospitalizations were steady ( for all ischemic amputation hospitalizations approached 20% of the total charges for stroke and TIA.
The inflation-adjusted estimated median cost for ischemic amputation remained stable from 2005 through 2008, but the mean cost per hospitalization increased by 9.2% [fig_ref] Table 3: Inflation-Adjusted Inpatient Hospital Costs for Ischemic Amputation in Minnesota, 2005-2008 a Costs... [/fig_ref]. Compared with 2005, the number of annual events in 2008 was 11.5% lower, and total annual inflation-adjusted inpatient hospital costs were 3.8% lower. The pattern in the intervening years was not monotonic.
# Discussion
We used a representative hospital discharge data set from 2005 through 2008 to evaluate the incidence of ischemic amputation in Minnesota. These data demonstrated that most patients who underwent ischemic amputation were male and older. The age-adjusted annual incidence of ischemic amputation in Minnesota was low compared with the incidence reported elsewhere [bib_ref] The use of angioplasty, bypass surgery, and amputation in the management of..., Tunis [/bib_ref]. Diabetes was prevalent in people who underwent ischemic amputation; 72% of the total ischemic amputation population had diabetes, and 63% of patients who had a major amputation had diabetes. Other studies have reported a prevalence of diabetes in ischemic amputees that ranges from 45% to 61% [bib_ref] The fate of bilateral lower limb amputees in end-stage vas-The findings and..., Inderbitzi [/bib_ref] [bib_ref] Age, diabetes and smoking in lower limb amputation for arterial occlusive disease, Liedberg [/bib_ref]. Furthermore, we found that the resource allocation by hospitals and economic burden was high; people who underwent ischemic amputation had long hospital stays. Median inpatient hospitalization charges were similar to those for CHD patients and double those for stroke and TIA patients.
Though the systemic atherosclerosis disease process provides a common etiology for stroke, myocardial infarction, and PAD, and though systemic risk is elevated for people with each of these syndromes, the incidence of amputation, CHD mortality, and stroke mortality were not uniform across Minnesota. This finding likely reflects differing care standards for each condition, because reporting for each condition is likely to be similar (each outcome represents a "hard ischemic endpoint"). These data do not provide insight into distinctions in the process of care between coronary and stroke or CLI patients that may clarify these varying rates. Our results underscore -on a state-based and county-based level -the health disparity in care that affects the CLI population.
Visser et al estimated a cost of $45,200 for a major amputation in the first year, followed by an annual cost of $11,000 beyond the initial hospitalization [bib_ref] Cost-effectiveness of diagnostic imaging work-up and treatment for patients with intermittent claudication..., Visser [/bib_ref]. In a more selected population of patients treated for CLI, the immediate cost directly attributed to amputation was $40,000 (1). The addition of subsequent rehabilitation is usually associated with a doubling of these costs (1). Our cost findings in a more general population provide additional information about immediate hospitalization charges associated with ischemic amputation. The Healthcare Cost and Utilization Project Nationwide Inpatient Sample of 39.5 million discharges in the United States in 2007 estimated median charges for amputation due to "circulatory system disorders" (exclusive of the upper limb and toe) of $45,900 and median hospital costs of $15,900. This value compares well with our estimates.
Minnesota is home to less than 2% of the total US population, and Minnesotans experience lower rates of atherosclerotic disease and lower medical costs than the national average [bib_ref] Slowing the growth of health care costs -lessons from regional variation, Fisher [/bib_ref]. Nevertheless, our results delineate the health cost implications in the context of nationally limited health care resources. Ziegler-Graham et al estimated the prevalence of limb loss in the United States, projected these values through the year 2030, and suggested that ischemic limb loss is likely to rise with the aging population and an increase in diabetes and "dysvascular" disease [bib_ref] Estimating the prevalence of limb loss in the United States: 2005 to..., Ziegler-Graham [/bib_ref].
Differences in ischemic amputation rates in the literature may be due to the unique populations included in those distinct datasets, differing case identification criteria for ischemic amputation, inclusion or exclusion of minor amputations or people with diabetes, inclusion of first amputation events versus cumulative per-patient event rates, or multiple amputations in the ipsilateral limb. Only through use of a consistent methodology will patients, payers, departments of health, and other agencies be able to make state-to-state comparisons. Consistent application of a common cross-state case ascertainment strategy would foster creation of estimates of the incidence of lower limb ischemic amputation and costs, permitting future comparison of population-based treatment strategies. Although racial and ethnic disparities in amputation rates have been elucidated, disparities based on geography, sex, and health care coverage may become visible from regional data. Health disparities, once revealed, provide an opportunity to implement preventive health strategies that can reduce high ischemic amputation rates across and within states.
Despite the use of a representative hospital discharge data set, this study had several limitations. The limited avail-
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
ability of clinical covariates in public data sets is acknowledged in the literature [bib_ref] Disparities in level of amputation among minorities: implications for improved preventative care, Lefebvre [/bib_ref]. Multiple hospitalizations for amputation (of the same or different limb) cannot be identified because of the absence of unique patient identifiers. Another limitation is that inpatient hospitalization charges, and the subsequent adjustment with cost-to-charge ratios, do not optimally reflect the resources expended in managing this condition. These data do not distinguish between use of health care resources that were expended to treat the episode of CLI versus any associated illnesses that existed during the index hospitalization. Additionally, these charges only reflect resources expended during the inpatient hospitalization and do not reflect costs after discharge. Our cost estimate is conservative, as it does not reflect any ongoing disability costs associated with ischemic amputation. Furthermore, data for counties with few cases of ischemic amputation (ie, [bib_ref] 42 CFR Parts 411, 412, 413, 422, and 489 Medicare Program; changes..., Us Department [/bib_ref] were reported, but the estimates are potentially unstable and should be interpreted with caution. Finally, our results were derived from a state that is known to be characterized by lower racial and ethnic diversity than the nation. Therefore, these data are not presumed to be generalizable to other states or regions. Despite these limitations, these data provide insight into the incidence and cost of ischemic amputation in Minnesota.
In conclusion, these data provide the first contemporary population-based estimate of ischemic amputation and hospitalization charges in a nonselected state-based population. Although the incidence of ischemic amputation in Minnesota is lower than has been reported from other studies, the direct inpatient hospital charges and estimated costs are high. The incidence of ischemic amputation was higher in rural counties. Ischemic amputation represents a significant statewide cardiovascular disease burden compared with both CHD and stroke mortality. This methodology could underpin an ischemic amputation surveillance strategy that could be modeled by other states nationally. Centers for Disease Control and Prevention - www.cdc.gov/pcd/issues/2011/nov/11_0023.htm
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
[fig] Figure: Lower limb ischemic amputations, coronary heart disease (CHD) mortality, and stroke mortality per 100,000 by Minnesota county, 2005-2008. Data for lower limb ischemic amputations obtained from Minnesota Hospital Uniform Billing Claims Data; Health Economics Program, Minnesota Department of Health; and the Minnesota Hospital Association. Data for CHD mortality obtained from Minnesota Department of Health Center for Health Statistics and International Classification of Diseases, 10th revision (ICD-10) codes I00-I09, I11, I13,and I20-I51. Data for stroke mortality obtained from Minnesota Department of Health Center for Health Statistics and ICD-10 codes I60-I69. [/fig]
[table] Table 2: The median charge for ischemic amputation hospitalizations was similar to that for CHD and about twice as high as for stroke and TIA hospitalizations. The total direct inpatient hospital charges Centers for Disease Control and Prevention • www.cdc.gov/pcd/issues/2011/nov/11_0023.htmThe findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. [/table]
[table] Table 3: Inflation-Adjusted Inpatient Hospital Costs for Ischemic Amputation in Minnesota, 2005-2008 a Costs calculated from hospital-specific cost-to-charge ratios in effect at time of patient discharge. All cost values are inflated to 2008 $US using the Consumer Price Index for inpatient hospital care. [/table]
|
Healthcare Professionals’ and Users’ Experiences of Intersectoral Care between Hospital and Community Mental Healthcare
This paper explores healthcare professionals' and users' experience of coherent intersectoral care between hospital mental healthcare and community mental healthcare. A total of 20 healthcare professionals, primarily nurses, and 14 users with a range of mental illnesses participated in nine focus group interviews (FGIs). Participants were encouraged in the FGIs to reflect upon their experience of coherency in intersectoral care. The analysis of FGIs was informed by a phenomenological-hermeneutic approach in a research group from 2016-2019. The Consolidated Criteria for Reporting Qualitative Research checklist was used as a guideline to ensure complete and accurate reporting of the study. The analysis led to the generation of several themes from a professional perspective and from a user perspective, addressed barriers to coherent intersectoral care. The healthcare professionals experienced barriers such as a lack of common language and knowledge of partners. The users did not feel involved and lacked coherence in their recovery processes and, as such, intersectoral care was often experienced as being lost in a maze.
# Introduction
In Western countries, coherent intersectoral care is one of the cornerstones of the healthcare debate. It requires a focus on long-term care instead of concentrating only on the here and now in one sector. According to healthcare policy, coherent intersectoral care is presented as a stated objective, with requirements for healthcare professionals to ensure that users systematically experience coherent efforts across the long-term process through hospital mental healthcare and community mental healthcareAt the political level, a good coherent intersectoral care service through improving collaboration is seen as key to improving and developing the mental healthcare system. However, despite such political aims, great awareness, and many initiatives, there are still challenges within and criticisms of the mental healthcare system in Scandinaviaand other Western countries. This can, for instance, be explained as a lack of coherence in the users' experience of the cross-sectoral process.
The Ministry of Health and the Elderly defines the measure of coherent intersectoral care as the sum of the activities, contacts, and events in the health service that a user or a defined group of users experiences in relation to the healthcare service. In some studies, it is assumed that 'intersectoral collaboration' is an important tool in achieving coherence. Additionally, coherent intersectoral care
# Materials and methods
The method of analysis is inspired by French philosopher Paul Ricoeur, in which the structure in the analysis is the narrative theory about mimesis divided into three; naive reading, structural analyses, and critical interpretation. The study's epistemology was framed by a qualitative approach using Ricoeur's phenomenological-hermeneutic theory of interpretation for healthcare professionals' and users' experience of coherent intersectoral care. The unique contribution from this investigation is part of a study where the data were used to perform iterative analyses with different theoretical framings in order to improve methods and depth by triangulation.
## Recruitment and sampling
We used purposive samplingto establish contacts and recruit participants for this study. Once we had received the names of the participants from the managements of hospital mental healthcare, Recovery College, community mental healthcare, and user associations, the first and second authors contacted the participants by email or telephone to request an appointment. The time period for the study was established as 2016 to 2019, where all of the authors were colleagues in the nurse education program. The first author holds a Ph.D. degree and has several years of theoretical and clinical experience in mental health; the second author is a graduate with no previous experience in psychiatry; and the third author holds a Ph.D. degree and has project experience, but not mental health experience.
As the research group, we reflected upon and discussed our positions in the research field, as well as our experiences.
As participants, the mental healthcare professionals (nurses, social workers, educators, and physiotherapists) were employed in either hospital mental healthcare or community mental healthcare. In total, 34 informants gave their informed consent to participate. Of these, 14 represented a user perspective, while 20 were healthcare professionals, of which 12 were from hospital mental healthcare and eight from the community mental healthcare.
## Focus groups
Context Participants
## 1.
Hospital mental healthcare N = 6 Healthcare professionals
## 2.
Hospital mental healthcare N = 3 Healthcare professionals
## 3.
Hospital mental healthcare N = 3 Healthcare professionals
## 4.
Community mental healthcare N = 4 Healthcare professionals
## 5.
Community mental healthcare N = 4 Healthcare professionals 6.
Recovery college N = 3 Users 7.
Community mental healthcare N = 4 Users 8.
Hospital mental healthcare N = 4 Users 9.
User organization N = 3 Users Number of participants 34
The healthcare professionals had 2-25 years of experience in mental healthcare, most having lengthy experience as specialists within the field of mental healthcare and employed for more than 5 years.
The users had different types of challenges and diagnoses (although mainly schizophrenia) and had experience using the services in both sectors. Most lacked an occupation, but several were active in volunteering.
Users with experience of coherent intersectoral care in both sectors (hospital mental healthcare and community mental healthcare) were included, and the healthcare professionals included had previous experience in the delivery of intersectoral care.
Users with distinctly psychotic and paranoid episodes lacking the energy or desire to participate were excluded. We have chosen not to involve the relatives in the study.
## Data collection-focus group interviews
Focus group interviews (FGIs) were conducted, which aimed to discuss and gain a deeper understanding of the findings. Data were collected through nine FGIs. A key strength of a FGI is that it encourages interaction in the group, which creates new perspectives and a deepened discussion. The participants in the focus groups were unfamiliar with each other and were selected because they had certain characteristics in common that related to the topic of the focus group. Within FGIs, the moderator and interviewer create a permissive and nurturing environment that encourages different perceptions and points of view, without pressuring participants to vote, plan, or reach a consensus.
Focus groups with 4-8 participants provide an opportunity for dynamic and mutual inspiration, which can make the empirical data more comprehensive. Some users joined the group for a cup of coffee, but left shortly afterwards as they were not due to participate. The participants in the focus group were very talkative.
All FGIs were audio-recorded and transcribed verbatim. Inspired by existing knowledge in the field, we formulated a list of topics to explore our participants' experience of coherent intersectoral care within mental healthcare; these themes guided the narrative. Based on the topics, we asked a few broad questionsand focused on follow-up questions to obtain richness and depth. The data have previously been used in a published document analysis, where we explored the discourses that emerge when healthcare professionals and users relate to intersectoral care. Two interviewers were present at each FGI interview in order to ensure a dialogic conversation where narrative phenomena were disclosed and developed. The first author was present in most FGIs, which reinforced the role of interviewer. There was no prior relationship between researchers and participants. The Consolidated Criteria for Reporting Qualitative Research checklist was used as a guideline to ensure accurate and complete reporting of the study. This guide contains recommendations for qualitative research, which we have tried to achieve throughout.
The topics served as a guiding framework for the knowledge we wanted to extract, but in the FGIs, we let ourselves flow with the thoughts voiced by the participants and inquired more deeply into these to achieve a nuanced understanding of their opinions.
# Data analysis
The aim of a hermeneutic interpretation of a text is to understand the participants' experiences as revealed from the text itself. We interpreted what the text says and speaks about. Aiming to find a deeper understanding and for creating new knowledge, FGIs were explored through descriptions in FGIs, where the interpretation already begun.
The analysis was carried out by all authors inspired by the phenomenological-hermeneutic approach described by Ricoeur's theory of text interpretation, by following three levels of interpretation: naive reading, structural analysis, and critical interpretation.
The FGIs were analyzed individually. Once the analysis of the narrative interviews was completed, we then analyzed across all of the FGIs. The naive reading gave an overall impression of what had been said. In the naive reading, the text was read repeatedly and with an open mind to achieve an initial understanding of how users experienced coherent intersectoral care. The units of opinion can be part of a sentence or several sentences, and it is important to emphasize that, in all cases, they can express only one opinion.
A structural analysis is a step between the naïve surface interpretation and comprehensive understanding. In structural analysis, units from the naive reading were used to find out what had been discussed and helped to formulate units of significance and corresponding themes. In the structural analysis of the FGIs, the focus was on deepening our understanding of the structures and patterns in the texts. To explain the structures and patterns in the FGIs, we examined whole, multiple, or parts of sentences from the text. Structural analyses were performed several times until we came to the point where no further subthemes and themes could be found in the data. The structural analysis aims to develop a possible explanation about the text, thus opening up the text and enabling the researcher to reach a deeper and more critical interpretation. This is situated at the end of the hermeneutic arc, thus creating a comprehensive interpretation of the text. The analysis is based on the total material and consists of the 'data', 'text', 'condensation', 'subtheme', and 'theme' phases. The community mental healthcare draws up a plan for the users that defines the focus of the intersectoral care; that is to say, the aim to be achieved by the users. Some users do not agree with the plans and they do not attend the meetings.
The intersectoral care is decided by the healthcare professional in hospital mental healthcare and community mental healthcare.
## Municipal
## Requirement.
Barriers to good intersectoral care.
## Fgi: users
The day after I arrived at Gentofte hospital's mental healthcare department, I was admitted to a closed ward. I did not really get any information about why I should be in the closed ward. I was quite confused about this. And then the day after I was moved to the open ward.
I did not feel consulted or involved.
Lack of information about transitions to the other wards. Discharge and intersectoral care is planned by the healthcare professionals, but not necessarily for the users.
Healthcare professionals are steering intersectoral care.
The comprehensive understanding emerged from our reading and rereading, checking and rechecking, until we had reached a common understanding of the phenomena. The themes from the structural analysis became the basis of the comprehensive understanding and were presented in the discussion, where theoretical perspectives were brought forward to achieve new insight and gain new knowledge regarding coherent intersectoral care.
# Ethical considerations
This study was conducted in adherence to the ethics of scientific work. Informed consent was obtained from all participants after they had received verbal and written information regarding the purpose of the study. The participants were informed that they were able to withdraw from the study with no consequences at any time, and that all data would be treated in such a way that no unauthorized person could have access to the material. Data were treated confidentially and anonymously, and all data material will be destroyed following publication.
We adhered closely to the ethics of scientific work. The study was accepted by the Danish Data Protection Agency According to the Helsinki Declarationand Danish law, no formal permit from a biomedical ethics committee was required, as the purpose of the research was not to influence the informants, either physically or psychologically.
## Findings
Initially, we first analyzed the healthcare professionals' experience, followed by the users'. The analysis of themes is supported by citations from the FGIs' data material.
## Common language and culture: experience of healthcare professionals
The healthcare professionals' experience of coherence in intersectoral care was coded into one main theme: 'common language and culture'. The main theme was based on three themes and nine subthemes. The first theme was 'coherence in intersectoral care involves good internal and external collaboration' with its three subthemes 'both sectors are located in close proximity', 'treatment and rehabilitation are based on a recovery-oriented approach', and 'contact person in the coherent intersectoral care'. The next theme was the 'healthcare professionals' focus on the methods to create a recovery-oriented approach to the intersectoral care' with its two subthemes 'cognitive behavioral therapy' and 'the users often do not fit the treatment methods'. The third theme was 'barriers to successful intersectoral care' with its four subthemes 'lack of knowledge of the other employees in the cooperating intersectoral units', 'healthcare professionals do not understand medicine and are afraid of patients', 'users must be in a life-threatening condition before the mental health center will activate a treatment judgment', and 'frequent change of employees hurts 'know-how".
## Coherence in intersectoral care involves good internal and external collaboration
Under the above theme, three subthemes were unfolded: 'both sectors are located in close proximity', 'treatment and rehabilitation are based on a recovery-oriented approach', 'contact person in the intersectoral care', as presented below.
## Both sectors are located in close proximity
As experienced by healthcare professionals, intersectoral care was most coherent when the units were located in close proximity. We are an outpatient clinic and operate close to an inpatient ward. So, we visit them and have a slightly easier communication path' (Professional FG 1).
## Treatment and rehabilitation are based on a recovery-oriented approach
The intersectoral collaboration was experienced as recovery-oriented, and the focus was always on the users' needs and hopes.
'The intersectoral care is very recovery-oriented. When users are discharged, the healthcare professionals in community mental healthcare support in the home' (Professional FG 1). We have the opportunity to reach out to users on the first visit, if appropriate and necessary' (Professional FG 3).
Various methods were described by the healthcare professionals to promote recovery-oriented intersectoral care. Tools that could be used to coordinate and create transparency in healthcare professionals acting across sectors were emphasized as part of the solution. Network meetings, weekly and treatment plans were among these tools. We experience that the approach to coordinate coherent intersectoral care was through network meetings, weekly and treatment plans as methods to keep the focus on the user's recovery' (Professional FG 2). 'There are many parts of the sector transition. In our work, it is very much about coordination' (Professional FG 5).
## Contact person in the coherent intersectoral care
The contact person in this coherent intersectoral care was experienced as being important to ensure continuity, and, as an organizational structure, as one of the best ways to build bridges between the sectors and to ensure cohesion. 'The contacts we have are the ones who coordinate. They are the ones who follow up appointments and also follow-up to ensure that we get to hold those meetings and medical appointments' (Professional FG 3).
The healthcare professionals focus on the methods to create a recovery-oriented approach to the coherent intersectoral care.
In the theme below, the healthcare professionals reflected on different structural methods as recovery-oriented. From this perspective, their methods were reflected as approaches to promoting coherent intersectoral care. The healthcare professionals' focus on coherent intersectoral care was formulated into two subthemes; 'cognitive behavioral therapy' and 'the users often do not fit the treatment methods', as presented below.
## Cognitive behavioral therapy
The healthcare professionals experience cognitive behavioral therapy as a framework for coherent intersectoral care in hospital mental healthcare. The healthcare professionals in hospital mental healthcare were convinced that the use of cognitive behavioral therapy was the best basis for a recovery-oriented practice. By using these methods, the users learned about their illnesses and symptoms and were empowered to manage their conditions. 'The cognitive model is the way we have a conversation, the way we structure agreements. We have an agenda, what has happened lately, what topics should we be dealing with, what home tasks, and what medicine' (Professional FG 2).
## The users often do not fit the treatment methods
According to the healthcare professionals in the hospital mental healthcare, the users did not always fit the treatment methods. Many users with severe mental illnesses found it difficult to reflect on their situation and to answer all the questions within the cognitive behavioral therapy model. 'Users do not have a completely realistic understanding of what their share is in working with cognitive behavioral therapy, where I think that, in psychiatry. I think it is unbearable and too difficult for everyone' (Professional FG 2).
## Barriers to successful coherent intersectoral care
Several barriers were found to creating a successful and coherent intersectoral care process between hospital mental healthcare and community mental healthcare. These barriers were categorized into four subthemes: 'lack of knowledge of the other employees in the cooperating intersectoral units', 'healthcare professionals do not understand medicine and are afraid of patients', 'the users must be in a life-threatening condition before the hospital mental healthcare activates a treatment judgment' and 'frequent change of employees hurts know-how,' as presented below.
## Lack of knowledge of the other employees in coherent intersectoral care
The healthcare professionals experience a lack of knowledge about the employees in the collaborating intersectoral service units. The healthcare professionals did not know the employees of the other service units, and thus communication and collaboration between intersectoral employees becomes difficult. We have a lack of knowledge about each other's workflows and structures. They will probably not be happy when they hear that they will receive the patient tomorrow because they need days to prepare' (Professional FG 1).
'Physically, it is an advantage that we as an ambulatory section live right on top of the other wards, as we can visit inpatient users below. That gives an easier communication path. It creates consistency and confidence for the user' (Professional FG 1).
## Healthcare professionals do not understand medicine and are afraid of patients
Healthcare professionals in the hospital mental healthcare claimed that healthcare professionals in community mental healthcare did not understand medicine and were afraid of the users. In community mental healthcare, healthcare professionals were predominantly focused on pedagogical activities. The following comment of a healthcare professional in mental healthcare illustrates the lack of mutual understanding of the users' problems and who should help solve them: 'At one time, we held a dialogue with the healthcare professionals in community mental healthcare who were sending users to the hospital's mental healthcare. We could not see the purpose when the user was the same as he was back when he was discharged from the hospital's mental healthcare department. The healthcare professionals at community mental healthcare claimed he could not stay in the institution because he was too ill. Sometimes healthcare professionals arrive at the mental healthcare with a user, but the user will absolutely not be admitted' (Professional FG 1).
## The users must be in a life-threatening condition before the hospital mental healthcare activates a treatment judgment
The healthcare professionals from community mental healthcare experienced that healthcare professionals in hospital mental healthcare did not take them seriously. When they informed the hospital mental healthcare that a user needs to be admitted, the user was often rejected or quickly discharged. Coherence in intersectoral care is challenged by a lack of collaboration due to the different cultures; healthcare professionals in community mental healthcare could assess the user as, for example, violent and psychotic and in need of admission, while healthcare professionals in the hospital mental healthcare do not believe there is any need for hospitalization. The healthcare professionals in community mental healthcare did not feel listened to and felt that they were powerless when they view a user as needing acute treatment but healthcare professionals in the hospital mental healthcare would not accept the user. Users should be in a life-threatening condition before the hospital mental healthcare activate a treatment judgment (this is a requirement for legislated treatment):
'Our collaboration with hospital mental healthcare is challenged when a user does not take his medicine, and a treatment judgment is made. When we call the ward to activate the judgment, it will often be rejected on the grounds that we have overreacted' (Professional FG 5).
## Frequent change of employees diminishes understanding
Frequent change of healthcare professionals is detrimental to the establishment of 'know-how' and was experienced as a serious problem because the healthcare professionals and users did not understand the work routines and schedules of their colleagues in intersectoral care, and therefore feel that they have to constantly restart the collaboration. The coordination during the intersectoral transition between the two sectors was challenged by frequent changes of physicians and other healthcare professionals:
'Frequent change of healthcare professionals diminishes understanding. When I call, they are no longer there. The expression of the condition is the same all over in healthcare service' (Professional FG 4).
## Not experiencing being seen as an individual person: experience of the users
In the FGIs, the users' experiences of intersectoral care were reflected in one main theme: 'users did not experience being seen as an individual person'. The users did not experience that the healthcare professionals focus on individual needs, but rather that the users have to adapt to the professionally preferred mode of treatment and rehabilitation. The main theme was based on two themes and five subthemes found in the analysis. The first theme was 'limited opportunity to participate in their treatment and rehabilitation' with its two subthemes 'lack of information about transitions to other units' and 'hospitalization starts with a conversation about the discharge'. The next theme was 'the users did not participate much in their treatment and rehabilitation' with its three subthemes 'not feeling respect for lived experience', 'not invited to meetings where the treatment and rehabilitation is planned', and 'not listening to the user's individual wishes to the treatment and rehabilitation' as presented below . . Intersectoral care as perceived by users: subthemes, themes, and main theme.
## Subthemes themes main theme
Lack of information about transitions to other units. Hospitalization starts with a conversation about the discharge.
Limited opportunity to participate in their treatment and rehabilitation.
Not experiencing being seen as an individual person: experience of the users. Not feeling respect for lived experience Not invited to meetings where the treatment and rehabilitation is planned. Not listening to the users' individual wishes regarding the treatment and rehabilitation.
The users did not participate much in their treatment and rehabilitation.
## Limited opportunity to participate in their treatment and rehabilitation
This theme reflects how the creation of coherence in intersectoral care was largely dependent on the user's will and ability to take responsibility for information moving from one unit to another. Coherence in intersectoral care requires that the user has resources to participate and collaborate actively 'I am surprised at the lack of influence on the process. Also compared to when I felt ready. The two treatment sites did not know what each other was doing. I think that is very strange and unsafe. The context is left to me' (User FG 9).
## Lack of information about transitions to other units
Lack of information about transitions to other units means that the user's treatment plans did not follow him, and he must start all over again every time he gets to a new unit in mental healthcare: 'I have been in the hospital's district psychiatry and community mental healthcare for a while, and my experience is that through every transition, it's like starting over anew. The ideal would be someone who could follow me. And how should information follow me? There is probably a journal somewhere' (User FG 8).
## Hospitalization starts with a conversation about the discharge
As experienced by the users, admission starts with a conversation about the discharge. The users found admission to be very stressful when the healthcare professionals start the conversation with a discussion about when he must be discharged. 'One thing some healthcare professionals say when you are admitted is that in fourteen days or three weeks, you will come out again. So, I say, how do they know? So, I tell them that it's of no use to me, when they give me a date for my discharge. Because I don't know how I got there' (User FG 6).
The above citation could relate to a bigger problem: Many users did not feel ready to be discharged shortly after admission. The focus was on efficiency and rapid treatment so that users could move forward as quickly as possible.
## The users did not participate much in their treatment and rehabilitation
The users did not participate much in their treatment and rehabilitation, while they did not feel any respect for lived experience, were not invited to meetings where the treatment and rehabilitation is planned, and did not feel their individual wishes for the treatment and rehabilitation were listened to.
Many users did not have the resources to be active and participate when they were admitted with serious illnesses. This problem was a general one across the sectors where healthcare professionals expected the users to follow the plan and participate actively. The healthcare professionals steered the plans for the treatment and rehabilitation. For many users, the experience was that participation was not possible to the extent that they had hoped for. The users did not experience that they could participate much in decision-making. 'I would like to be involved a little more in the treatment. Now, I didn't know that I should be in here for two-and-a-half months, and there was probably no one here who knew anything about it. But I would like to think that once a month had passed, you can say: "Okay, now a month has passed, it has happened and that and that. We have these problems and these thoughts going forward with the medicine. For example, we have put your medicine down here, or we want to put your medicine up". And if they, for example, wanted to put up my medicine, then I also want to be involved in it. And I wasn't' (User FG 9).
## Not feeling respect for lived experience
Users felt that their life experience was not being respected, which we interpreted as the users attempting to share their experiences of suffering and needing help but not having those needs met: 'For me, a good patient process is when you arrive at, for example, the district psychiatrist's outpatient treatment facilities, that they actually know who you are when you arrive. Not necessarily your entire life story, but at least they know your name. They know what your diagnosis is, they are aware of what problems you are currently battling with' (User FG 6).
## Not being invited to meetings where the treatment and rehabilitation is planned
Not being invited to meetings where treatment plans are conducted was a significant problem for users who had wishes and hopes for the future, and who hoped to discuss these details with the healthcare professionals. Moreover, the users did not receive any information about why they should not attend the network meetings. 'I have not attended network meetings and my experience is that the meetings are mostly held for the sake of the healthcare professionals rather than listening to us users' (User FG 5).
## Not listening to the users' individual wishes to the treatment and rehabilitation
Not listening to the users' individual wishes for the future was an experience that users problematized. We interpreted this as the users wishing that the intersectoral collaboration was more focused on support their recovery process. 'When it comes to purely medical treatment, you have good contact with each other. But when it comes to physiotherapy, occupational therapy and that path, it is as if there is no communication at all' (User FG 7).
# Discussion
In the analysis, we interpreted that healthcare professionals and users show different perceptions and experiences of coherent intersectoral care between hospital mental healthcare and community mental healthcare. In the analysis, we interpreted that healthcare professionals and users show different perceptions and experiences of coherent intersectoral care between hospital mental healthcare and community mental healthcare. There are differences between the professionals' and users' perception of 'coherence'. While the professionals relate closely to their own practice and context, users perceive coherence related to not just continuity between sectors and easy handovers. Coherence also seems to involve the integration of the psychiatric services/help in user's daily life. The overriding theme from a healthcare professional perspective 'common language and culture' (the interpretation of which differs between the sectors) and the theme from the users' perspective 'not experiencing being seen as an individual person', is seen as opening up the discussion for the different ways in which healthcare professionals understand and solve users' problems.
Overall, we interpreted that the healthcare professionals primarily focus on the treatment of symptoms and of fostering self-care and rely less on involving the user. The healthcare professionals experience the treatment and rehabilitation in intersectoral care as being recovery-oriented, but their focus was framed by clinical recovery where the goal is remission and maintenance of the functional level, as we also see among other researchers. The healthcare professionals in community mental healthcare thought that relational work is the most effective way to help the users towards recovery.
In addition, other studies show that healthcare professionals in hospital mental healthcare are focused on disorders and symptoms and do not prioritize the involvement of users in the treatment and rehabilitation. The lack of a common language and culture becomes a barrier to improving coherence in intersectoral care, as is also shown in other studies. In addition, frequent replacement of healthcare professionals, meagre economic resources and low numbers of healthcare professionals, all make the transition between hospital mental healthcare and community mental healthcare difficult; all of these issues are detrimental to the creation and maintenance of a shared understanding.
Healthcare professionals in community mental healthcare have voiced that plans prepared for continued mental healthcare after discharge from hospital mental healthcare do not necessarily make sense for the users.
As Bonfils et al. emphasize, differences in language and culture are visible through the healthcare professionals' different views about when a person is ill and in need of help .
Coherent intersectoral care requires the communication of essential information to be given in a way that other units understand and accept. We believe, in accordance with, that there is need for a cultural change, placing users at the center of their own recovery process. Therefore, it is essential that coherent intersectoral care supports the users' need for knowledge of care rather than following predefined treatment procedures that do not necessarily match the users' real needs. and Borg believe that we should rethink the healthcare system so that it emphasizes the differences between individual users, rather than seeing them as a collective mass.
In this and other studies, healthcare professionals agreed that the users should participate in their treatment and rehabilitation, but, because of their severe symptoms, it often seems difficult to involve them. In this and other studies, users with serious symptoms were not invited to participate in planning the process of help, and the healthcare professionals thus decided the users' treatment without involving or informing them, as they expected this group of users to be unable to fully participate. The consequence of such expectations is that the users do not gain ownership of a treatment that should address their individual needs.
In hospital mental healthcare, there is confidence that methods such as psychoeducation and cognitive behavioral therapy provide a coherent fostering basis for intersectoral care and collaboration. According to healthcare professionals, when the users gain more knowledge and ownership of their treatment, the recovery-oriented process is experienced as being more coherent. The municipality contributes on many fronts and has the expectation that all users will, as far as possible, recover in order to be able to care for themselves at some point. Users are thus subject to the respective plans formulated by the hospital mental healthcare and community mental healthcare, but their participation in the preparation of these plans (often) occurs only in the most peripheral sense.
# Methodological considerations
Ricoeur's hermeneutics is a challenging and time-consuming method that has led to in-depth discussions in the research team in order to arrive at relevant themes. As consequence of the phenomenological-hermeneutic methodology, we deliberately chose a small number of participants for this study. We found the data collected to be adequate with respect to our opportunities to perform the thorough analysis that is required by the phenomenological-hermeneutic tradition, as well as sufficient to formulate well-founded themes and main themes from the interpretation of the texts from FGIs with users and healthcare professionals.
The findings from this study have implications for practice and research. The healthcare professionals can learn from the healthcare professionals' and users' experience of intersectoral care and collaboration. Our findings may provide greater insight into the obstacles that may be associated with creating a link between intersectoral care between hospital and community mental healthcare.
The healthcare professionals clearly have a central role in the promotion of coherent intersectoral care. Reflection on their own practice may make healthcare professionals aware of their current approach towards engaging users, which can lead to nurses acknowledging the patients' role in nursing care and a more patient-centered manner.
We advocate for the relevance of new research seeking to improve continuity and coherence for users' recovery process by developing a user-participated intervention drawing on users' hopes, goals and possibilities. This will require a discussion on how successful transitions between hospital mental healthcare and community mental healthcare should be understood and implemented in the nursing practice. Anchoring the transition between hospital mental healthcare and community mental healthcare requires changes in organizational structures and procedures.
# Conclusions
The aim of this study was to explore coherency in intersectoral care between hospital mental healthcare and community mental healthcare, as experienced by healthcare professionals and users. The study shows that healthcare professionals from both sectors experience that coherent intersectoral care was defined by the time in which users received treatment and/or rehabilitation. The healthcare professionals experience that intersectoral care was dependent on the intersectoral collaboration and entails a recovery-oriented focus, where users were involved in decisions, and that communication was optimized when the healthcare professionals across the sectors knew each other in advance and the units were physically close. Neither healthcare professionals nor users experience the intersectoral care and collaboration as a coherent whole because information was often lost in the transitions between sectors, as well as in the structural conditions such as network meetings where user-involvement in decision-making was not complied with. This leads us to understand the experience of intersectoral care as an ideal often lost between sectors.
The healthcare professionals in hospital mental healthcare and community mental healthcare are divided on the question of how the users' problems should be understood and resolved. Different paradigmatic approaches (e.g., biomedicine and holismseem to challenge what should be the focus of the users' recovery process between hospital mental healthcare and community mental healthcare. The users had no doubt that they want to be seen more as whole persons and were demanding a more holistic paradigm across the sectors. The users did not experience being sufficiently involved in the intersectoral care and being seen as a whole person with a life outside hospital mental healthcare. The users did not feel the professionals listened to their perceptions of mental challenges and their narratives about their lives were not passed on between treatment sites.
Author Contributions: K.J. was responsible for the coordination, data generation, data analysis, and drafting the manuscript. J.F. and M.B.D. participated in the iterative analyses, interpretations of the data, drafting the manuscript and revising and approving the final manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by the University College Copenhagen. |
The pleiotropic roles of adipocyte secretome in remodeling breast cancer
Background: Breast cancer is the leading female cancer type and the cause of cancer-related mortality worldwide. Adipocytes possess important functions of energy supply, metabolic regulation, and cytokine release, and are also the matrix cell that supports mammary gland tissue. In breast cancer tumor microenvironment (TME), adipocytes are the prominent stromal cells and are implicated in inflammation, metastatic formation, metabolic remodeling, and cancer susceptibility.Main body:It is well-established that adipocyte secretome is a reservoir engaged in the regulation of tumor cell behavior by secreting a large number of cytokines (IL-6, IL-8, and chemokines), adipokines (leptin, adiponectin, autotaxin, and resistin), lipid metabolites (free fatty acids and β-hydroxybutyrate), and other exosome-encapsulated substances. These released factors influence the evolution and clinical outcome of breast cancer through complex mechanisms. The progression of breast cancer tumors revolves around the tumor-adipose stromal network, which may contribute to breast cancer aggressiveness by increasing the pro-malignant potential of TME and tumor cells themselves. Most importantly, the secretome alterations of adipocytes are regarded as distinctly important targets for breast cancer diagnosis, treatment, and drug resistance.Conclusion: Therefore, this review will provide a comprehensive description of the specific adipocyte secretome characteristics and interactions within TME cell populations, which will enable us to better tailor strategies for tumor stratification management and treatment.
# Background
Breast cancer is the most common female cancer type and the leading cause of cancer-related mortality worldwide . Breast cancer is a highly heterogeneous entity with multiple tumor subtypes, accompanied by frequent alterations in genetics and malignant cellular behaviors [bib_ref] Tissue-resident FOLR2+ macrophages associate with CD8+ T cell infiltration in human breast..., Ramos [/bib_ref]. Adipocytes are the major component of breast tissue, approximately constituting 90% of the stromal tissue. According to the anatomical distribution, the fully differentiated breast primarily consists of two cellular compartments: the epithelial cell zone with glandular and lobular alveolar differentiated units containing branching ducts, and the connective tissue zone, commonly referred to as the breast fat pad, which is composed primarily of adipose tissue (AT) [bib_ref] Breast biomechanics: what do we really know?, Mcghee [/bib_ref]. In general, normal mature adipocytes are separated from epithelial cells by the basement membrane, thus limiting the possibility of heterotypic crosstalk between the two cell types. Breast cancer usually begins in the epithelial cells surrounding the ducts and lobular tissue of the breast. Extracellular matrix (ECM) and epithelial cell degeneration may lead to the eventual destruction of the basement membrane that separates epithelial and mesenchymal tissues for increased tumor aggressiveness [bib_ref] Chemotherapy as a regulator of extracellular matrix-cell communication: implications in therapy resistance, Gonzalez-Molina [/bib_ref].
The tumor microenvironment (TME) refers to the highly complex, heterogeneous, and spatiotemporally changing microenvironment in tumor entities, which profoundly and ubiquitously affects the biological behaviors and malignant evolutions of tumors [bib_ref] Accessories to the crime: functions of cells recruited to the tumor microenvironment, Hanahan [/bib_ref]. The TME contains plentiful soluble secreted factors, coding and non-coding RNAs, and a wide variety of cell types, thus establishing an intricate niche for cellular crosstalk. Specifically, the types of cell components in the TME mainly include fibroblasts, immune cells, adipocytes, and vascular endothelial cells (ECs). Breast cancer cells can reprogram the TME through paracrine and autocrine, and altered TME in turn further promotes biological events, including breast cancer growth, invasion, angiogenesis, and metastasis [bib_ref] The breast tumor microenvironment: role in cancer development, progression and response to..., Mittal [/bib_ref] [bib_ref] Heparanasedriven sequential released nanoparticles for ferroptosis and tumor microenvironment modulations synergism in..., Zhang [/bib_ref] [bib_ref] Multi-omics profiling reveals distinct microenvironment characterization and suggests immune escape mechanisms of..., Xiao [/bib_ref]. Therefore, these series of related events lead TME to play a key role in supporting breast cancer development and tumor diagnosis and treatment. And very importantly, adipocytes are the most prominent cell type in the breast cancer TME . Adipocytes have been traditionally considered as energy storage cells, AT is broadly characterized by three storage types consisting of white, beige, and brown [bib_ref] Exosomes from the tumour-adipocyte interplay stimulate beige/brown differentiation and reprogram metabolism in..., Wu [/bib_ref]. The endocrine, metabolic and inflammatory distribution of adipose depends on and is influenced by the estrogenic and androgenic status of AT [bib_ref] Impact of estrogens on the regulation of white, beige, and brown adipose..., Bernasochi [/bib_ref]. The abundance of intratumoral adipocytes in breast cancer is implicated in The adipocyte-containing tumor microenvironment (TME) in breast cancer (BC) tissue. A According to the distribution characteristics, adipose tissue (AT) can be cataloged into breast AT, subcutaneous AT, visceral AT, and intramuscular AT. B The adipocytes in breast AT (nBAs) represent a unique cell population to sustain the breast morphology. The main structure of the mammary gland consists of a system of acinus and lobules. C The TME in breast cancer consists of multiple adipocyte types, including ADSCs, pre-adipocytes, TSAs, and CAAs. Process of carcinoma in situ (CIS) and invasive carcinoma (IC) of the breast: the ducts consist of luminal cells (inner layer) and myoepithelial cells (outer layer), and are surrounded by the basement membrane, adipocytes, and other mesenchymal cells. CIS is characterized by tumor tissue distribution confined to epithelial cells, not breaking through the basement membrane, and not invading the interlobular space. Once the tumor cells have penetrated the basement membrane, invasive breast cancer occurs. At this point, breast cancer cells can interact with adipocytes at the invasion front, and educate them into CAAs to support breast cancer progression inflammation, metastatic formation, metabolic changes, cancer stemness, and a supportive tumor immune microenvironment [bib_ref] Adipocytes induce distinct gene expression profiles in mammary tumor cells and enhance..., Nickel [/bib_ref].
It is well-established that adipocytes can secrete tremendous factors engaged in the regulation of inflammation, metastatic formation, metabolic remodeling, and cancer susceptibility [bib_ref] Obesity and breast cancer: the role of crown-like structures in breast adipose..., Faria [/bib_ref]. Specifically, adipocytes regulate tumor cell behavior either through cell-to-cell contact or through the release of proteins, lipids, metabolites, or exosomes, and vice versa. There are distribution regional, structural and functional differences among different types of adipocytes, including tumor-stromal adipocytes (TSAs), cancer-associated adipocytes (CAAs), normal breast adipocytes (nBAs), adipose-derived stem cells (ADSCs) and adipocytes in other parts of the body . TSA-breast cancer cell crosstalk has been recognized as a crucial event in facilitating breast cancer growth via paracrine release [bib_ref] Interaction with adipocyte stromal cells induces breast cancer malignancy via S100A7 upregulation..., Sakurai [/bib_ref]. The breast cancerneighbored aberrant adipocytes within invasive frontiers are denoted as CAAs. CAAs are characterized by smaller cell size, reduced lipid droplets, and a fibroblast-like phenotype, as well as reduced biomarkers of adipogenic differentiation, implying that these microscopic features of CAAs undergo dedifferentiation under the influence of neighboring cancer cells [bib_ref] Cancer-associated adipocytes: emerging supporters in breast cancer, Zhao [/bib_ref]. ADSCs possess multi-lineage differentiation potential and self-renewal capability, which is highly significant for developing regenerative strategies [bib_ref] Hydrogel scaffolds for differentiation of adipose-derived stem cells, Huang [/bib_ref]. CAA-originated adipokines synergistically orchestrate the interaction between adipocytes, tumor cells, and other stromal cells [bib_ref] Cancer-associated adipocytes: key players in breast cancer progression, Wu [/bib_ref]. Metabolically, adipocytes and ADSCs can fuel tumor growth by acting as energy reservoirs for adjacent breast cancer cells or by secreting key signaling molecules. A variety of physiological and pathological factors can trigger the changes in the morphology, quantity, and functional status of adipocytes. In the physiological status, pre-and postmenopausal, pre-and post-lactation, epinephrine level fluctuation, lymphedema status, and even hot and cold stimulation, are important factors affecting adipocytes [bib_ref] The effect of adipocyte-macrophage crosstalk in obesity-related breast cancer, Engin [/bib_ref] [bib_ref] Epinephrine infiltration of adipose tissue impacts MCF7 breast Cancer cells and Total..., Avril [/bib_ref] [bib_ref] Lymphedema alters lipolytic, lipogenic, immune and angiogenic properties of adipose tissue: a..., Koc [/bib_ref]. Moreover, some specific disease states, such as mastitis, benign and malignant tumors, and endocrine diseases, can arouse persistent inflammation and hormone level disorders, which can cause AT inflammation, endoplasmic reticulum stress, ECM changes, and other mechanisms to cause adipocyte alteration and AT disorders [bib_ref] Obesity and Cancer mechanisms: tumor microenvironment and inflammation, Iyengar [/bib_ref]. In particular, epidemiological evidence highlights a strong association between obesity and higher cancer risk and poor prognosis in breast cancer patients [bib_ref] Metabolic pathways in obesity-related breast cancer, Brown [/bib_ref]. Obesity makes the risks of breast cancer occurrence much more imminent. Adipose cachexia caused by obesity triggers malignant crosstalk between adiposity and tumors. The obesity-related adipocyte metabolism synthetically facilitates this condition by producing local and circulating pro-inflammatory secretome and initiating the malignant behavior of breast cancer cells [bib_ref] Obesity and adverse breast cancer risk and outcome: mechanistic insights and strategies..., Picon-Ruiz [/bib_ref] [fig_ref] Table 1: The morphology and functon comparison of nBAs, TSAs, and CAAsAbbreviations [/fig_ref].
Therefore, the progression of breast cancer tumors revolves around the tumor-adipose network, which may promote breast cancer aggressiveness by increasing the pro-malignant potential of TME and tumor cells themselves. Here, this review emphasizes the pleiotropic roles of adipocyte secretome in remodeling breast cancer progression, including representative cytokines (IL-6, IL-8, and chemokines), adipokines (leptin, adiponectin, autotaxin, and resistin), lipid metabolites (free fatty acids and β-hydroxybutyrate), and other exosome-encapsulated substances, as well as their performance in vicious tumor-adipose interaction, immune regulation, and breast cancer diagnosis breast cancer therapy. The thorough understanding will provide novel perspectives on adipocyte-based therapeutic strategies for combating breast cancer.
## Adipocyte secretome in breast cancer progression
## Cytokines
## Il-6
Interleukin-6 (IL-6) is a widely studied, multifunctional, and complex cytokine and is involved in pathophysiological processes such as immune response, inflammation, hematopoiesis, bone metabolism, and tissue repair [bib_ref] IL-6 in inflammation, autoimmunity and cancer, Hirano [/bib_ref]. Adipocytes are reverted to an immature and proliferative phenotype in the presence of breast cancer cells. Meanwhile, the cancerous CAAs more significantly promote the proliferation and migration in comparison to nBAs, by secreting the adipokines represented by IL-6 and monocyte chemoattractant protein-1 (MCP-1). Pre-adipocytes showed increased IL-6 secretion and expression of cancer-associated fibroblast (CAF) markers FSP1 and α-SMC and their pathological IL-6 overproduction led to the proliferation and migration of breast cancer cells and xenografts [bib_ref] IL-6-mediated cross-talk between human preadipocytes and ductal carcinoma in situ in breast..., Kim [/bib_ref]. IL-6-mediated interactions between pro-adipocytes and breast cancer cells contribute to the progression of breast cancer.
Adipose stromal cells with plasticity could secrete IL-6 to promote migration and invasion of breast cancer cells, which could be mediated by cofilin-1 dependent pathway [bib_ref] Interleukin 6 secreted from adipose stromal cells promotes migration and invasion of..., Walter [/bib_ref]. Lee et al. identified the pivotal role of mmu-miR-5112-Cpeb1-Il6 axis in promoting the inflammatory nature of CAAs in the breast cancer microenvironment [bib_ref] Transition into inflammatory cancer-associated adipocytes in breast cancer microenvironment requires microRNA regulatory..., Lee [/bib_ref]. Adipocyte-derived IL-6 and leptin enhanced the PLOD2 expression by activating the janus tyrosine kinase (JAK)/signal transducer and activator of transcription 3 (STAT3) and phosphatidylinositide 3-kinases (PI3K)/ protein kinase B (AKT) signaling pathways, thereby promoting breast cancer metastasis in vivo [bib_ref] Adipocyte-derived IL-6 and leptin promote breast Cancer metastasis via upregulation of Lysyl..., He [/bib_ref]. Oncostatin M (OSM) is a pluripotential cytokine of IL-6 family members, and can inhibit interferon-β (IFN-β) expression and signaling in triple-negative breast cancer (TNBC) [bib_ref] The opposing effects of interferon-beta and oncostatin-M as regulators of cancer stem..., Doherty [/bib_ref]. Lapeire et al. confirmed that cancer-associated AT boosted breast cancer progression by paracrine OSM and JAK/STAT3 axis activation [bib_ref] Cancer-associated adipose tissue promotes breast cancer progression by paracrine oncostatin M and..., Lapeire [/bib_ref]. reported that paracrine IL-6 from matured adipocytes enhanced the aggressive epithelial-mesenchymal transition (EMT) behavior of breast cancer cells via inducing IL-6/STAT3 signaling [bib_ref] Interleukin-6/STAT3 signalling regulates adipocyte induced epithelial-mesenchymal transition in breast cancer cells, Gyamfi [/bib_ref]. Adipose-induced EMT in breast cancer cells can be reversed by inhibition of the IL-6/STAT3 signaling axis, by using U.S. food and drug administration (FDA)-approved anthelmintic niclosamide [bib_ref] Niclosamide reverses adipocyte induced epithelial-mesenchymal transition in breast cancer cells via suppression..., Gyamfi [/bib_ref].
## Il-8
Interleukin-8 (IL-8), also known as C-X-C motif ligand 8 (CXCL8), is a granulocytic chemokine with high expression levels in breast cancer [bib_ref] CCL8 mediates crosstalk between endothelial colony forming cells and triple-negative breast cancer..., Kim [/bib_ref]. IL-8 is responsible for immunosuppressive cell recruitment, inflammation, angiogenesis, and metastasis. Adipose stromal cellderived IL-8 predominantly enhanced the breast cancer cell invasiveness and non-adherent growth ability in a paracrine manner [bib_ref] Interleukin-8 derived from local tissue-resident stromal cells promotes tumor cell invasion, Welte [/bib_ref]. CAA secreted a significantly higher IL-8 level and could enhance the EMT effect in normal breast luminal cells in an IL-8-dependent way [bib_ref] Interleukin-8 activates breast Cancer-associated adipocytes and promotes their angiogenesis-and tumorigenesis-promoting effects, Al-Khalaf [/bib_ref]. Moreover, the ectopic expression of IL-8 in adipocytes in histologically noncancerous region activated them and enhanced their oncogenesis effects in a STAT3dependent manner. This study also highlights multiple patterns of IL-8 inhibition that could effectively inhibit CAA features and their tumor-promoting effects. In addition, compared with single cultured breast cancer cells, the co-culture of adipocytes significantly enhanced the IL-8 secretion, which could increase the cell-adhesion molecules and influence the metastatic capacity of breast cancer cells [bib_ref] Adipocytes promote early steps of breast cancer cell dissemination via Interleukin-8, Vazquez Rodriguez [/bib_ref]. Anti-IL-8 intervention altered the cytokine secretion diversion of breast adipocytes with reduced vascular endothelial growth factor (VEGF) secretion and affected the propagation of breast cancer.
## Il-1β
Interleukin-1 (IL-1) is composed of two main isoforms, IL-1α and IL-1β, each of which is encoded by two separate genes [bib_ref] Function and regulation of IL-1α in inflammatory diseases and cancer, Malik [/bib_ref]. IL-1β is a pro-inflammatory cytokine, and its abnormally high expression is a predictive biomarker of tumor malignancy and increased risk of bone metastasis from breast cancer [bib_ref] The role of IL-1B in breast cancer bone metastasis, Tulotta [/bib_ref]. IL-1β can be triggered by inflammatory signaling in multiple immune types and is an important factor and driver of tumor malignant evolution, therefore IL-1β blockade may be an important strategy to inhibit tumor growth and metastasis. Under obesity, IL-1β from primary adipocytes could induce the angiopoietin-like 4 (ANGPTL4) expression through activation of NF-κB and mitogen-activated protein (MAP) kinases, therefore promoting angiogenesis and tumor suppression [bib_ref] Obesity-associated inflammation promotes angiogenesis and breast cancer via angiopoietin-like 4, Kolb [/bib_ref]. Adipocyte-derived ANGPTL4 was beneficial to breast cancer angiogenesis and progression in obesity and was a potential therapeutic target for obese breast cancer patients.
## Ccl2
The MCP-1/C-C motif chemokine ligand 2 (CCL2) and its receptor interaction, including CC-chemokine receptor 2 (CCR2) and atypical chemokine receptor 2 (ACKR2) are renowned for their capability to drive the chemotaxis of myeloid cell and lymphocyte recruitment [bib_ref] CCL2/ACKR2 interaction participate in breast cancer metastasis especially in patients with altered..., Zhong [/bib_ref]. CCL2-CCR2 signaling axis function has various impacts on breast cancer progression, such as invasiveness and metastasis, recruitment of immunosuppressive cells, and stem cell self-renewal [bib_ref] Roles of CCL2-CCR2 Axis in the tumor microenvironment, Kadomoto [/bib_ref]. CAA-originated CCL2 participated in mediating the migration of ER+ and TNBC cells [bib_ref] Cancer-mediated adipose reversion promotes cancer cell migration via IL-6 and MCP-1, Fujisaki [/bib_ref]. CCL2 produced by 3 T3-L1 cells was an important mediator in the crosstalk of adipocytes and 4 T1 cells [bib_ref] Aspirin breaks the crosstalk between 3T3-L1 Adipocytes and 4T1 breast cancer cells..., Hsieh [/bib_ref] [bib_ref] Lunasin attenuates obesity-associated metastasis of 4T1 breast Cancer cell through anti-inflammatory property, Hsieh [/bib_ref]. Anti-inflammatory drugs, such as aspirin and lunasin, can inhibit this crosstalk effect and thus prevent breast cancer progression. Liu et al. established a protein trap with CCL2 targeting ability to locally inhibit the CCL2 expression of tumor-associated adipocytes and thus lightened the immunosuppressive TME, which manifested as more abundant T cells and fewer M2 and myeloid-derived suppressor cells (MDSCs) [bib_ref] Tumor-targeted gene therapy with lipid nanoparticles inhibits tumor-associated adipocytes and remodels the..., Liu [/bib_ref]. Arendt et al. showed that in obesity, the enhanced production of CCL2 and IL-1β by the breast AT induced macrophage recruitment and crown-like structure (CLS) formation, which further induced angiogenesis and promoted the expanding tissue. This result demonstrated that the CCL2/IL-1b/CXCL12 was involved in the link of adipocytes and macrophages for carcinogenesis.
## Ccl5
CC-motif chemokine ligand 5 (CCL5) belongs to an important CC subfamily of chemotactic cytokines that can be secreted either by tumor cells or by stromal cells. CCL5/CCR5 combination is known for initiating and facilitating inflammatory responses, and acts as a tumorpromoting factor that is tightly associated with the invasive and metastatic stages of breast cancer [bib_ref] The chemokine CCL5 as a potential prognostic factor predicting disease progression in..., Yaal-Hahoshen [/bib_ref]. The human adipocytes enhanced the invasiveness of MDA-MB-231 cells in the co-culture system. This phenomenon was partially mediated by CCL5, and could be antagonized by using CCL5-blocking peptides and antibodies. Song et al. showed that the increased secretion of CCL5 in adipocytes enhanced the EMT effect of co-cultured TNBC cell lines, thereby promoting tumor growth and lung/liver metastasis in a mouse model [bib_ref] Emodin inhibits epithelial-mesenchymal transition and metastasis of triple negative breast cancer via..., Song [/bib_ref]. Hence, selectively CCL5 targeted inhibition, such as emodin in adipose microenvironment might arouse positive therapeutic effects for breast cancer.
## Igf-1 and igfbp
Tumor stroma-derived insulin-like growth factor 1(IGF-1) and insulin-like growth factor binding protein (IGFBP) signaling have the potential to be novel targets for cancer therapy or adjuvant treatment, but these therapies have so far failed to yield significant efficacy [bib_ref] Insulin/IGF-driven cancer cell-stroma crosstalk as a novel therapeutic target in pancreatic cancer, Mutgan [/bib_ref]. IGFBPs show a tendency to be overexpressed and phosphorylated in several breast cancer subtypes, resulting in cancer cell proliferation, metastasis, and drug resistance-related properties [bib_ref] Role of the IGF-1 Axis in overcoming resistance in breast cancer, Ianza [/bib_ref]. The expression of IGF-1 content, which was regulated by glucose and fatty acids, was observably up-regulated in undifferentiated cells and differentiated adipocytes from obese individuals than those from lean individuals [bib_ref] Obesity-associated systemic interleukin-6 promotes pre-adipocyte aromatase expression via increased breast cancer cell..., Bowers [/bib_ref]. Correspondingly, the IGF-1 inhibition in adipocytes almost suppressed the growth-promoting effect of adipocytes on MCF-7 cells in vitro. The IGFBP-2 protein level in adipocytes was higher in adjacent peritumoral metastatic breast cancer tissues, compared to that in metastatic tissues [bib_ref] Human Adipocytes stimulate invasion of breast cancer MCF-7 Cells by Secreting IGFBP-2, Wang [/bib_ref]. Meanwhile, IGFBP-2 secreted by mature adipocytes greatly affected the metastatic capability of breast cancer cells.
## Vegf
VEGF is an important angiogenic factor that initiates downstream signaling cascades by binding to VEGF receptors with different affinities and specificities, thereby stimulating tumor cell proliferation, growth, and neovascularization [bib_ref] Anti-vascular endothelial growth factor therapy in breast cancer: molecular pathway, potential targets,..., Zhang [/bib_ref]. VEGF signaling also activates matrix metalloproteinase 2 (MMP2), matrix metalloproteinase 9 (MMP9), and urokinase plasminogen activator to degrade the ECM and basement membrane system [bib_ref] The role of VEGF in triple-negative breast cancer: where do we go..., Dent [/bib_ref]. Overall, VEGF affects the survival, proliferation, and migration of ECs, thus triggering angiogenic formation, liberating malignancy from the dormant phase, and inducing tumor vascularization. Several studies have demonstrated the contribution of adipocyte-derived VEGF to the malignant progression of breast cancer. Lauriane et al. reported that the mature adipocytes, obtained after differentiation of human adipose-derived stem cells (hADSCs) from normal weight (MA20) and obese (MA30) women, were co-cultured with breast cancer cells to mimic the adipocyte microenvironment surrounding mammary tumors [bib_ref] Supernatants of adipocytes from obese versus normal weight women and breast Cancer..., Bougaret [/bib_ref]. Compared with MA20, MA30 exhibited increased expression of leptin and cytochrome P450 family 19 subfamily a member 1(CYP19A1), especially VEGF. Supernatants from co-cultures of mature adipocytes and MCF-7 cells increased the proliferation, migration, and germination of human umbilical vein endothelial cells (HUVECs) independent of female body mass index (BMI). This promotive effect could be counteracted by neutralizing antibodies of VEGF and leptin. In addition, Francesca et al. found that the tumors co-injected with hADSCs grew faster and larger than those without hAD-SCs, accompanied by high vascularized biomarker VEGF and CD31 [bib_ref] Human adipose stem cell differentiation is highly affected by cancer cells both..., Paino [/bib_ref]. The evidence suggests that the pro-angiogenic effect of adipocytes in breast cancer may be partially mediated by VEGF.
## G-csf
Granulocyte-colony stimulating factor (G-CSF) is an important member of the hematopoietic growth factor family, mainly affecting neutrophil activation and regulation. G-CSF becomes aberrantly overexpressed in different types of cancer cells, such as lung cancer, gastric cancer, invasive bladder cancer, glioma, and a significant proportion of breast cancer [bib_ref] Opposing effects of granulocyte Colonystimulating factor on the initiation and progression of..., Hiraga [/bib_ref]. G-CSF and neutrophils may play an essential role in synergistically promoting tumor metastasis. G-CSF is identified as one of the most highly up-regulated genes in CAAs and was increased in breast cancer-associated AT. Mechanically speaking, CAA-derived G-CSF was capable of promoting breast cancer migration and invasion via STAT3 signaling and other corporated CAA-derived factors IL-6 and granulocyte-macrophage-colony stimulating factor (GM-CSF) in vitro, thus forming an amplification effect to mediate breast cancer progression [bib_ref] Cancer-associated adipocyte-derived G-CSF promotes breast cancer malignancy via Stat3 signaling, Liu [/bib_ref]. It posed a feasible strategy for interventing breast cancer invasivity by targeting the CAA-derived G-CSF axis.
## Adipokines leptin
Leptin is a key adipokine that is produced mainly by mature adipocytes. Leptin appears to be a key driver in energy homeostasis and breast cancer tumorigenesis, and its expression is associated with AT mass, BMI, and leptin receptor (OB-R) [bib_ref] The leptin Axis and its association with the adaptive immune system in..., García-Estevez [/bib_ref]. During the breast cancer progression, TSAs stimulated by external conditions can oversecrete leptin, which could bind to receptors on tumor cells and exert biological effects, triggering the effects on proliferation, migration, and tumor invasion [bib_ref] Association between obesity and breast cancer: molecular bases and the effect of..., Martínez-Rodríguez [/bib_ref].
Leptin was enriched in mammary adipocytes/AT, and notably suppressed CD8+ T cell function by activating the STAT3-fatty acid oxidation (FAO) axis and reducing glycolysis [bib_ref] STAT3 activation-induced fatty acid oxidation in CD8+ T effector cells is critical..., Zhang [/bib_ref]. Targeting the leptin-STAT3-FAO axis might potentiate robust anti-tumor CD8 + T immune killing effect, in addition to inhibiting cancer stem cells (CSCs) and chemo-resistant tumor cells in breast cancer. Plasminogen activator inhibitor-1 (PAI-1) was a key effector of the metastasis in breast cancer cells co-cultured with adipocytes [bib_ref] Leptin promotes migration and invasion of breast cancer cells by stimulating IL-8..., Cao [/bib_ref]. Then, adipocytesecreted leptin promoted PAI-1-mediated breast cancer metastasis via activating the STAT3/miR-34a pathway in vivo. In addition, leptin promoted breast cancer cell migration and invasion via FAK-Src-dependent manner in vitro [bib_ref] Leptin induces cell migration and invasion in a FAK-Src-dependent manner in breast..., Juárez-Cruz [/bib_ref]. verified that leptin could enhance the cAMP response element binding protein (CREB)-dependent aromatase, which was related to the up-regulated cyclooxygenase-2 (COX-2) expression in breast cancer cells [bib_ref] Leptin induces CREB-dependent aromatase activation through COX-2 expression in breast cancer cells, Kim [/bib_ref]. Wei et al. reported that caused the EMT effect of breast cancer via pyruvate kinase M2 (PKM2) up-regulation and PI3K/AKT signaling pathway activation [bib_ref] Leptin promotes epithelial-mesenchymal transition of breast cancer via the upregulation of pyruvate..., Wei [/bib_ref]. Leptin derived from obese adipose stromal and ADSCs were highly expressed and promoted the proliferation and metastasis behaviors in the ER+ MCF-7 cells [bib_ref] Leptin produced by obese adipose stromal/stem cells enhances proliferation and metastasis of..., Strong [/bib_ref]. These studies suggest that functional crosstalk between leptin and estrogen signaling contributes to breast cancer development and progression.
## Adiponectin
ADIPOQ/adiponectin is an abundant adipocytokine secreted by AT and adipocytes of breast cancer tissue and possesses insulin sensitizing, anti-inflammatory function, and antiatherogenic performances [bib_ref] Cross-talk between adiponectin and IGF-IR in breast Cancer, Mauro [/bib_ref]. Epidemiological investigations have confirmed that breast cancer tumors in women with low circulating adiponectin levels exhibit more aggressive characteristics, such as higher histological grade and enhanced angiogenesis and metastasis [bib_ref] Novel insights into adiponectin action in breast cancer: evidence of its mechanistic..., Andò [/bib_ref]. Low circulating adiponectin level was a risk indicator for breast cancer patients regardless of postmenopausal status [bib_ref] Circulating adiponectin and breast cancer risk: a systematic review and meta-analysis, Macis [/bib_ref]. Compared with distant and node-negative sites, peritumoral adipokines were significantly altered, especially in the obesity setting and advanced breast cancer stage [bib_ref] Peritumoral expression of Adipokines and fatty acids in breast Cancer, Gnerlich [/bib_ref]. analyzed that in addition to enhancing the chemotherapy efficacy of breast cancer xenograft, the high expression of adiponectin was intensively related to the better overall survival (OS) in breast cancer patients with combined chemotherapy [bib_ref] ADIPOQ/adiponectin induces cytotoxic autophagy in breast cancer cells through STK11/LKB1-mediated activation of..., Chung [/bib_ref].
Many breast cancer cell lines, such as MCF-7, T47D, and SK-BR-3, could express adiponectin receptors Adi-poR1 and AdipoR2 [bib_ref] Effects of adiponectin on breast cancer cell growth and signaling, Grossmann [/bib_ref]. By binding to its receptor, adiponectin negatively the malignant behavior of tumors. Adiponectin could inhibit the leptin-promoted cell proliferation thus manifesting as a potent leptin antagonist [bib_ref] Integral role of PTP1B in adiponectin-mediated inhibition of oncogenic actions of leptin..., Taliaferro-Smith [/bib_ref]. Mechanically speaking, adiponectin inhibited the activation of leptin-mediated extracellular signal-regulated kinase (ERK) and AKT, and increased the leptin-inhibited protein tyrosine phosphatase 1B (PTP1B) expression and activation in vivo. Chung et al. deciphered that adiponectin promoted the autophagy-related cytotoxic death in breast cancer cells both in vitro and in vivo, which was mediated by STK11/LKB1-AMPK-ULK1 axis [bib_ref] ADIPOQ/adiponectin induces cytotoxic autophagy in breast cancer cells through STK11/LKB1-mediated activation of..., Chung [/bib_ref].
## Atx
Autotaxin (ATX) is a secreted form of lysophospholipase D that yields extracellular lysophosphatidic acid (LPA) by hydrolyzing the choline of lysophosphatidylcholine (LPC) [bib_ref] The expression regulation and biological function of, Zhang [/bib_ref]. ADSCs and adipocytes, especially those adjacent to breast cancer tumor, were the major contributors to enhanced ATX level and plasma LPA in breast cancer patients [bib_ref] ADSCs and adipocytes are the main producers in the autotaxin-lysophosphatidic acid axis..., Schmid [/bib_ref] [bib_ref] Tumor-induced inflammation in mammary adipose tissue stimulates a vicious cycle of autotaxin..., Benesch [/bib_ref]. ATX-LPA signaling triggers key cellular events that lead to enhanced aggressiveness and motility of breast cancer cells.
The conditioned medium (CM) from mammary AT could enhance the ER-breast cancer cell proliferation in vitro, by promoting ATX secretion [bib_ref] Mammary adipose tissue-derived lysophospholipids promote estrogen receptor-negative mammary epithelial cell proliferation, Volden [/bib_ref]. The bidirectional action between breast cancer cells and breast adipocytes modified the local ATX/LPA axis during the ER-breast cancer progression and increased ATX expression in the breast adipose pad. Cha et al. investigated that ATX-LPA signaling-related proteins, including LPA1, LPA2, and LPA3, were overexpressed in breast cancer adipose stroma, accompanied by abundantly infiltrating macrophages [bib_ref] Expression of Autotaxin-Lysophosphatidate signalingrelated proteins in breast Cancer with adipose stroma, Cha [/bib_ref]. It indicated that ATX-LPA axis was a key coordinator in the inflammatory of adipose stroma. Benesch et al. verified that ONO-8430506 (an ATX inhibitor) efficiently reduced tumor growth by down-regulating inflammatory mediators in cancerinflamed AT [bib_ref] Tumor-induced inflammation in mammary adipose tissue stimulates a vicious cycle of autotaxin..., Benesch [/bib_ref]. Therefore, the malignant inflammation cycle played an important role in breast cancer progression and could be interrupted by the inhibition of ATX.
## Resistin
Resistin is acknowledged as a novel secretory factor from adipocytes, and may regulate inflammatory responses, affect tumor angiogenesis and modulate insulin sensitivity. Resistin is one of the highest regulatory adipokines secreted by 3 T3-L1 adipocytes under obesity-related metabolic settings compared to normal physiology. Lee et al. confirmed that breast cancer tissues, rather than adjacent normal breast tissues, possess an abundantly high expression of resistin, which level was positively correlated with tumor size, stage, lymph node (LN) metastasis, and ER status [bib_ref] Resistin expression in breast cancer tissue as a marker of prognosis and..., Lee [/bib_ref] [bib_ref] Adipocytes and obesity-related conditions jointly promote breast cancer cell growth and motility:..., Rosendahl [/bib_ref]. High expression of resistin receptor CAP1 in tumors implied shorter OS, and relapse free survival (RFS) among ER+ tumors or LN-positive tumors [bib_ref] Adipocytes and obesity-related conditions jointly promote breast cancer cell growth and motility:..., Rosendahl [/bib_ref]. Furthermore, resistin was proved to accelerate EMT and breast cancer cell stemness by toll-like receptor 4 (TLR4)-induced NF-κB activation [bib_ref] Resistin facilitates breast cancer progression via TLR4-mediated induction of mesenchymal phenotypes and..., Wang [/bib_ref].
Resistin could stimulate both ER+ breast cancer and TNBC cell proliferation and motility in vitro [bib_ref] Adipocytes and obesity-related conditions jointly promote breast cancer cell growth and motility:..., Rosendahl [/bib_ref]. Lee et al. verified that resistin enhanced the phosphorylation of ezrin, radixin, and moesin via protein kinase C α (PKCα) [bib_ref] Resistin, a fatderived secretory factor, promotes metastasis of MDA-MB-231 human breast cancer..., Lee [/bib_ref]. This effect mediated the increased vimentin expression and facilitated breast cancer invasion. Gao et al. identified that resistin acted as the TAZ (a transcription cofactor) target to promote oncogenesis and was correlated with the adipocytic TAZ expression in TNBC tissue [bib_ref] Adipocytes promote breast tumorigenesis through TAZ-dependent secretion of Resistin, Gao [/bib_ref]. They suggested that the adipocyte-related TAZ/resistin signal was conducive to the development of breast cancer tumors, and the resistin neutralization represented a promising therapeutic strategy.
## Visfatin
Visfatin/nicotinamide phosphoribosyltransferase (NAMPT) is an adipokine and an enzyme that is related to metabolic and immune disorders [bib_ref] Adipocytokines visfatin and resistin in breast cancer: clinical relevance, biological mechanisms, and..., Wang [/bib_ref]. Human visfatin is predominantly produced by adipocytes and macrophages residing in AT. In addition, circulating visfatin concentrations are significantly higher in the obese state, and visfatin is negatively feedback regulated by insulin and leptin [bib_ref] Involvement of novel Adipokines, Chemerin, Visfatin, Resistin and Apelin in reproductive functions..., Estienne [/bib_ref].
Hung et al. demonstrated that extracellular visfatinpromoted growth and lung metastasis of breast cancer were mediated by activating c-Abl and STAT3 [bib_ref] Extracellular Visfatin-promoted malignant behavior in breast cancer is mediated through c-Abl and..., Hung [/bib_ref]. Visfatin promoted the MMP2, MMP9, and VEGF gene expression and proliferation of MCF-7 cells [bib_ref] Visfatin stimulates proliferation of MCF-7 human breast cancer cells, Kim [/bib_ref]. Extracellular visfatin induced nicotinamide adenine dinucleotide (NAD), subsequently leading to the sirtuin 1 (SIRT1) activation and p53 deacetylation in vitro [bib_ref] Extracellular NAMPT/visfatin causes p53 deacetylation via NAD production and SIRT1 activation in..., Behrouzfar [/bib_ref]. showed that pERK/CXCL1 activation was involved in visfatin-induced breast cancer progression in lung metastasis, which could be blocked by CXCL1 antibody. In ADSCs and breast cancer cell co-culture, visfatinpreconditioned ADSCs (vADSCs) produced more activated tumor behaviors and tumorsphere formation [bib_ref] Visfatin mediates malignant behaviors through adipose-derived stem cells intermediary in breast cancer, Huang [/bib_ref]. In vivo, vADSCs promoted the tumor to form a larger entity. This result was mediated by visfatin-induced growth differentiation factor 15 (GDF15)-AKT activation in vADSCs.
## Mmp
Matrix metalloproteinases (MMPs) are regarded to be a category of proteins that mediate the degradation of the ECM of tumors and are primarily involved in tissue remodeling, inflammatory cell recruitment, and tumor invasion [bib_ref] Therapeutic potential of matrix metalloproteinase inhibition in breast cancer, Radisky [/bib_ref]. Thus, MMPs are important remodeling enablers of the TME. MMPs drive the malignant phenotype of breast cancer cells, including the acquisition of CSC characteristics and the induction of EMT. The overexpression of MMPs, especially MMP-9, MMP-2, and MMPs in serum, may indicate a higher risk of poor prognosis in breast cancer [bib_ref] Overexpression of MMP family members functions as prognostic biomarker for breast cancer..., Ren [/bib_ref]. Leitner et al. demonstrated that osteopontin (OPN) and MMP were upregulated in obese AT [bib_ref] Osteopontin promotes aromatase expression and estradiol production in human adipocytes, Leitner [/bib_ref]. OPN and MMP-cleaved OPN could both contribute to inducing aromatase activity and estradiol production in human adipocytes, and then possibly facilitated breast cancer cell proliferation.
## Adipsin
Adipsin, known as complement factor D, can maintain AT homeostasis, and is primarily expressed in monocytes, macrophages AT, and ADSC within tumor tissue [bib_ref] Adipsin serum concentrations and adipose tissue expression in people with obesity and..., Milek [/bib_ref]. showed that adipsin originated from ADSCs distinctly promoted the proliferation and CSClike features of human breast cancer patient-derived xenograft (PDX) cells [bib_ref] Adipose-derived stem cells enhance human breast cancer growth and cancer stem cell-like..., Goto [/bib_ref]. Notably, hepatocyte growth factor (HGF) was secreted by mammary ADSCs and acted as a downstream molecular of adipsin in mammary ADSCs to improve the CSC properties in breast cancer [bib_ref] Adipsin-dependent secretion of hepatocyte growth factor regulates the adipocyte-cancer stem cell interaction, Mizuno [/bib_ref].
## Apelin
Apelin, also denoted as angiotensin-like-receptor 1 (APJ), is an endogenous modulatory peptide that is a ligand for the APJ receptor belonging to the G protein-coupled receptor family [bib_ref] The apelin/APJ system in the regulation of vascular tone: friend or foe?, Rikitake [/bib_ref]. Besides, apelin is expressed in a large range of organs including brain, lung, kidney, pancreas, testis, prostate and AT [bib_ref] Role of Apelin/APJ axis in cancer development and progression, Masoumi [/bib_ref]. Apelin expression is enhanced in various cancers, and the apelin/APJ axis has a crucial impact in tumor progress by reinforcing angiogenesis, metastasis, cell proliferation, and CSC plasticity and drug resistance. Gourgue et al. showed that apelin was an elevated adipokine in the subcutaneous adipose and tumor tissues of obese mice, while the apelin antagonist could inhibit the TNBC growth and brain metastasis formation [bib_ref] Obesity and triple-negative-breast-cancer: is apelin a new key target?, Gourgue [/bib_ref]. In addition to obesity, the high tumor apelin expressions were related to a blunted response to N-acetylcysteine in their breast cancer cohort [bib_ref] Tumor apelin and obesity are associated with reduced neoadjuvant chemotherapy response in..., Gourgue [/bib_ref].
## Sfrp5
Secreted frizzled-related protein 5 (SFRP5) is considered to be a tumor suppressor gene and its circulating level is a feature in obesity-related metabolic disorders [bib_ref] Circulating Sfrp5 is a signature of obesity-related metabolic disorders and is regulated..., Hu [/bib_ref]. SFRP5 derived from 3 T3-L1 cells could suppress the migration and invasion of breast cancer cells by inhibiting Wnt and downstream EMT processes [bib_ref] Adipocyte-derived SFRP5 inhibits breast cancer cells migration and invasion through Wnt and..., Zhou [/bib_ref].
## Lipid metabolites
Adipocyte lipolysis could prepare metabolic substrates to fuel breast cancer cells, thus favoriting breast cancer progression [bib_ref] Adipocyte lipolysis links obesity to breast cancer growth: adipocyte-derived fatty acids drive..., Balaban [/bib_ref].
## Ffa
Fatty acid (FA) is enriched in adipocytes, and could be transported to breast cancer cells. There is a complex metabolic symbiosis between adipose stromal cells and breast cancer cells that stimulate their malignant activity. After tumor-induced lipolysis, free fatty acid (FFA) released by adipocytes is transferred and stored in the tumor cells as triglycerides. FFA can be chronically released from lipid droplets via adipose triglyceride lipase (ATGL)-dependent lipolytic pathway [bib_ref] Mammary adipocytes stimulate breast cancer invasion through metabolic remodeling of tumor cells, Wang [/bib_ref]. In vivo, ATGL is expressed in human tumors, and its expression correlates with tumor aggressiveness and is upregulated by contact with adipocytes. Inhibiting the ATGL-dependent lipolytic/FAO pathway impeded the invasiveness of breast cancer cells induced by co-culture. Balaban et al. elucidated that breast cancer cells in an FA-rich environment were able to accumulate extracellular FA in the form of intracellular triacylglycerols (TAG), and the increased intracellular TAG levels, which could protect breast cancer cells from palmitate-induced apoptosis [bib_ref] Heterogeneity of fatty acid metabolism in breast cancer cells underlies differential sensitivity..., Balaban [/bib_ref]. Zaoui et al. reported that the breast adipocyte secretome could promote malignant proliferation of breast cancer cells by enhancing CD36-mediated FA uptake. This function of breast adipocyte donors was unrelated to the state of BMI, menopausal status, and mammary density [bib_ref] Breast-associated adipocytes Secretome induce fatty acid uptake and invasiveness in breast cancer..., Zaoui [/bib_ref].
Byon et al. proved that FFAs could enhance PAI-1 expression in MDA-MB-231 cells and facilitate breast cancer progression in vitro, by partially inducing SMAD4. breast cancer cells affect lipolysis in adipocytes, and adipocytes affect β-oxidation in breast cancer cells [bib_ref] Energy transfer from adipocytes to cancer cells in breast cancer, Kim [/bib_ref]. Concretely, Adipocytes promoted breast cancer growth through mitochondrial β-oxidation by using fatty acid from adipocytes, of which process fatty-acid-binding protein 4 (FABP4) was a key activated target for metabolic interaction. Bellanger et al. adipocytes activated breast cancer cell autophagy by lysosome acidification, thus leading to tumorous survival and migration [bib_ref] Adipocytes promote breast cancer cell survival and migration through autophagy activation, Bellanger [/bib_ref]. Adipocytes increased autophagic flux, autophagosome maturation, lysosomal acidification, and intracellular degradation in breast cancer cells, and that disruption of membrane phospholipid composition with reduced arachidonic acid content was responsible for adipocyteinduced activation of autophagy in breast cancer cells.
## Β-hb
Ketone body β-hydroxybutyrate (β-HB), is extracted from FAO and acts as a nutrient supply component for peripheral tissues [bib_ref] Ketone bodies as signaling metabolites, Newman [/bib_ref]. Huang et al. identified that primary mammary gland-derived adipocytes could induce histone H3K9 acetylation and tumor-promoting gene upregulation by secreting β-HB, thus increasing the malignancy of breast cancer cells [bib_ref] Adipocytes promote malignant growth of breast tumours with monocarboxylate transporter 2 expression..., Huang [/bib_ref].
## Etp
Endotrophin (ETP) is a signaling molecule derived from collagen type VI and is richly expressed in AT [bib_ref] Human endotrophin as a driver of malignant tumor growth, Bu [/bib_ref]. Park et al. showed that ETP could promote collagen VI-stimulated tumor growth and TGF-β-mediated tissue fibrosis and EMT in tumor invasion [bib_ref] Adipocyte-derived endotrophin promotes malignant tumor progression, Park [/bib_ref]. ETP enhances fibrosis, angiogenesis, and inflammation by recruiting macrophages and ECs.
## Lcn2
Lipocalin-2 (Lcn2), also denoted as neutrophil-gelatinase-associated lipocalin (NGAL), is a transporter protein related to energy homeostasis, inflammatory response, tumorigenesis, and tumor growth [bib_ref] The effect of lipocalin-2 (LCN2) on apoptosis: a proteomics analysis study in..., Wu [/bib_ref]. Drew et al. verified that reduced ERα expression in AT was associated with breast cancer proliferation and migration by increasing adipocyte-specific Lcn2 production and enhancing Lcn2 sensitivity of breast cancer cells [bib_ref] Estrogen receptor (ER)α-regulated Lipocalin 2 expression in adipose tissue links obesity with..., Drew [/bib_ref].
## Adipocyte secretome in regulating immune cells in breast cancer
Complex interactions of environmental factors TME can have profound effects on the metabolic activities of immune cells, stromal cells, and tumor cell types. Macrophages are the most studied characteristic cells in adipose-tumor-immune cell interaction network. Crosstalk between adipocytes and macrophages profoundly affects breast cancer progression. In particular, the adipocytemacrophage-breast cancer cell interaction in obesity involves multiple sequential mechanisms, including the release of inflammatory factors, ER stress, insulin resistance (IR), aromatase, and other hormone levels [bib_ref] The effect of adipocyte-macrophage crosstalk in obesity-related breast cancer, Engin [/bib_ref]. Focusing on tumor-associated macrophages (TAMs) and tumor-associated adipocytes and their crosstalk will provide new clues for breast cancer treatment.
Tumor-associated AT is typically featured by thin/ fragile adipose membranes, necrosis, a high abundance of pro-inflammatory high-mobility group box 1 protein (HMGB1), as well as the loss of the lipid storage medium perilipin-1 [bib_ref] Loss of adipocyte specification and necrosis augment tumor-associated inflammation, Wagner [/bib_ref]. The loss of adipocyte specification and necrosis directly activated macrophages and thus enhanced the inflammation state in TME. The combination of soluble breast adipocyte-derived leptin, insulin, IL-6, and TNF-α were inducers of the release of angiogenic mediators in macrophages, representing a potential mechanism for the enhanced risk of breast cancer progression in obese individuals [bib_ref] Breast adipocyte co-culture increases the expression of pro-angiogenic factors in macrophages, Yadav [/bib_ref]. Yadav et al. indicated that some adipocyte-derived mediators, such as leptin, insulin, IL-6, and TNF-α, could facilitate VEGF content in THP-1 macrophages in co-culture model. It confirmed the angiogenesis-promoting effect of adipocytes by enhancing the paracrine function of angiogenic factors of macrophage. Adipocyte-secreted IL-8 polarized the tumor-promoting neutrophils and increased the cell-adhesion molecules in breast cancer cells, thus facilitating the pro-inflammatory state in breast cancer [bib_ref] Adipocytes promote early steps of breast cancer cell dissemination via Interleukin-8, Vazquez Rodriguez [/bib_ref]. Reciprocally, tumor-cell-derived factors and macrophage-related factors can also influence the inflammatory properties of adipocytes [bib_ref] Lunasin attenuates obesity-associated metastasis of 4T1 breast Cancer cell through anti-inflammatory property, Hsieh [/bib_ref]. Targeting and reducing the inflammatory state associated with adipocytes is conducive to suppressing the progression of breast cancer with obesity.
Breast AT could intensify the breast cancer invasiveness via breast AT macrophages, and had a stronger relationship with breast cancer survival than breast tumor stroma macrophages [bib_ref] Breast adipose tissue macrophages (BATMs) have a stronger correlation with breast cancer..., Lin [/bib_ref]. There is an obesity-inflammation-aromatase axis in mouse models and female patients, which is featured by CLS [bib_ref] Molecular pathways: adipose inflammation as a mediator of obesity-associated Cancer, Howe [/bib_ref] [bib_ref] Obesity is associated with inflammation and elevated aromatase expression in the mouse..., Subbaramaiah [/bib_ref]. CLS is formed by macrophages surrounding dead or dying adipocytes and is denoted as a biomarker of the proinflammatory status of AT [bib_ref] The relationship between white adipose tissue inflammation and overweight/obesity in Chinese female..., Zhao [/bib_ref]. In the human cohort, CLSs of the breast are positively associated with older age, obesity, dyslipidemia, and higher levels of glucose, insulin, C-reactive protein, and IL-6 [bib_ref] Crown-like structures in breast adipose tissue of breast cancer patients: associations with..., Chang [/bib_ref]. Macrophage-derived pro-inflammatory mediators induce aromatase and estrogen-dependent gene expression in adipocytes. Santander et al. verified that the AT of breast cancer tissue in obese mice presented a larger number of infiltrated macrophages, CLS, and hypertrophic adipocytes, compared with lean breast cancer tumor [bib_ref] Paracrine interactions between adipocytes and tumor cells recruit and modify macrophages to..., Santander [/bib_ref].
Leptin mediates the bidirectional action between breast cancer cells and TAMs. Leptin promotes migration and invasion of breast cancer cells by enhancing IL-8 expression in M2 macrophages [bib_ref] Leptin promotes migration and invasion of breast cancer cells by stimulating IL-8..., Cao [/bib_ref]. Li et al. showed that leptin promoted the IL-18 expression and secretion in TAMs via NF-κB/NF-κB1 and breast cancer cells via PI3K-AKT/ATF-2 pathway, ultimately leading to breast cancer cell migration and invasion [bib_ref] Leptin promotes breast cancer cell migration and invasion via IL-18 expression and..., Li [/bib_ref]. The adipocytereleased leptin could increase macrophage recruitment and activation by pleiotropic function regulation, such as various pro-inflammatory transcription factors and CD295 receptors in macrophages [bib_ref] Paracrine interactions between adipocytes and tumor cells recruit and modify macrophages to..., Santander [/bib_ref]. In Cha et al. study, CD68+/CD163+ macrophages were infiltrating and CLSs existed in AT surrounding breast cancer tissue. This specific macrophage and structures had a higher histologic grade of breast cancer [bib_ref] Tumor-associated macrophages and crown-like structures in adipose tissue in breast cancer, Cha [/bib_ref].
In general, macrophages are a highly heterogeneous cell population with strong plasticity that communicates closely with adipocytes. Factors released by adipocytes, represented by leptin, insulin, and IL-6, can domesticate macrophages toward the function of promoting tumor malignancy, manifested by the changes in the secretion profile of macrophages, including proangiogenic VEGF and pro-inflammatory IL-6 [bib_ref] Cancer-associated adipocytes in breast Cancer: causes and consequences, Rybinska [/bib_ref]. On the other hand, the adipose-secreting profile can induce an increase in the number and activity of macrophages residing in ATT, which leads to the activation and amplification of inflammatory signals for favoring breast cancer progression. This functional alteration is closely relevant to the endocrine, paracrine, and autocrine pathways of adipocyte secretome [bib_ref] The effect of adipocyte-macrophage crosstalk in obesity-related breast cancer, Engin [/bib_ref]. Therefore, from the perspective of treatment, specifically targeting the function of macrophages in TME or reducing the number of immunosuppressive macrophages can block the progression of breast cancer by interrupting the communication between adipocytes and macrophages, promoting immune escape, and inhibiting angiogenesis. Of course, the role of the function and activity of other immune cells in the adipose-tumor interaction is also worthy of further investigation.
## Exosome-mediated horizontal transfer of adipocyte secretome in breast cancer
Exosomes are naturally occurring membranous extracellular vesicles (EVs) with 30-150 nm in diameter, characterized by conserved tetraspanins and key proteins associated with endocytosis and cargo sorting. Exosomes are a reservoir that contains genetic and functional molecules, such as nucleic acids (DNA, mRNA, ncRNAs), proteins, enzymes, and lipids. Functionally, exosomes are mediators of the interplay between adipocytes and tumor cells in breast cancer, and can promote the formation of the inflammatory microenvironment and systematically pre-conditioning metastasis niches for remote transmission and transfer [bib_ref] 20160485. research data, including large and complex data types • gold Open..., Robado De Lope [/bib_ref]. Specific exosomal cargoes transported from AT might serve as endocrine or paracrine modes of intracellular communication between AT and other tissue types [bib_ref] The clinical potential of circulating microRNAs in obesity, Ji [/bib_ref]. Exosomal mediators-dissociated from visceral adipocyte is capable of accommodating the key organ inflammatory and fibrosis signaling pathways [bib_ref] Adipocyte-derived exosomal miRNAs: a novel mechanism for obesityrelated disease, Ferrante [/bib_ref].
EVs produced by metabolically abnormal adipocytes potentially fuel the progression of breast cancer. For instance, exosome proteins from IR adipocytes were associated with characteristics of EMT and CSCs in a cohort of breast cancer patients [bib_ref] Adipocyte-derived exosomes may promote breast cancer progression in type 2 diabetes, Jafari [/bib_ref]. These adipocyteexosomes carrying thrombospondin 5 (TSP5) dramatically enhanced the tumor EMT behaviors, and promote breast cancer malignancies in type 2 diabetes. hADSCderived exosomes could induce MCF-7 migration by activating Wnt signaling in vitro [bib_ref] Exosomes from human adipose-derived mesenchymal stem cells promote migration through Wnt signaling..., Lin [/bib_ref]. Exosomes released by ADSC-differentiated adipocytes altered the transcriptome and improved MCF-7 growth via activation of Hippo signaling pathway [bib_ref] Exosomes secreted by mesenchymal stromal/stem cell-derived adipocytes promote breast cancer cell growth..., Wang [/bib_ref]. And, the consumption of adipocyte exosomes weakened the tumor-promoting effect of adipocytes. Gernapudi et al. confirmed that exosomes secreted from preadipocytes modulated earlystage breast cancer via influencing stem cell renewal, breast cancer cell migration, and neoplasia through miR-140/SOX2/SOX9 axis [bib_ref] Targeting exosomes from preadipocytes inhibits preadipocyte to cancer stem cell signaling in..., Gernapudi [/bib_ref]. Therefore, exosome-mediated horizontal transfer of adipocyte secretome induces breast cancer cells to exhibit metabolically reprogram and aggressive phenotypes both in vitro and in vivo [bib_ref] Exosomes from the tumour-adipocyte interplay stimulate beige/brown differentiation and reprogram metabolism in..., Wu [/bib_ref]. Targeted exosome therapy is an essential approach to cut off the interaction between adipocytes and tumor cells and has potential for clinical therapy [fig_ref] Table 2: The role of adipocytes-derived exosomes in regulating BC progression Abbreviations [/fig_ref].
## Adipocyte secretome in breast cancer diagnosis
Some adipokines expression levels in serum, like resistin, leptin, adiponectin, and visfatin, can serve as risk factors for clinicopathological features of post-menopausal breast cancer, including tumor-node-metastasis (TNM) stage, tumor size, LN metastasis, and histological grade [bib_ref] Resistin, Visfatin, adiponectin, and leptin: risk of breast Cancer in pre-and postmenopausal..., Assiri [/bib_ref]. These adipokines are excellent candidates for surveillance and stratified management for postmenopausal breast cancer. Serum expression levels of leptin, resistin, and visfatin are reliable diagnostic markers of breast cancer and are independent predictive factors of LN invasion and ER-negative breast cancer patients [bib_ref] Evaluation of diagnostic and predictive value of serum adipokines: leptin, resistin and..., Assiri [/bib_ref].
Leptin and ObR are positively correlated, and both are highly expressed in primary and metastatic breast cancer tissue [bib_ref] Increased expression of leptin and the leptin receptor as a marker of..., Garofalo [/bib_ref]. These two biomarkers are confirmed to be associated with poor prognosis, especially in G3 tumors. The high ratio of serum leptin/adiponectin might indicate the presence of aggressive breast cancer [bib_ref] Serum adiponectin and leptin levels in Taiwanese breast cancer patients, Chen [/bib_ref]. Lee et al. emphasized that resistin was a potential independent prognostic factor for breast cancer patients and for hormone therapy stratification and tumor therapy [bib_ref] Resistin expression in breast cancer tissue as a marker of prognosis and..., Lee [/bib_ref]. The epigenetic inactivation of secreted SFRP5 gene, especially SFRP5 promoter hypermethylation was associated with poor prognosis in breast cancer [bib_ref] Epigenetic inactivation of the secreted frizzled-related protein-5 ( SFRP5 ) gene in..., Veeck [/bib_ref]. Afterward, the high level of SFRP5 in plasma and tumor tissue was related to better clinical factors in breast cancer patients, such as BMI and LN metastasis, TNM stage, as well as low high Ki67 expression [bib_ref] Adipocyte-derived SFRP5 inhibits breast cancer cells migration and invasion through Wnt and..., Zhou [/bib_ref]. Similarly, high serum endolipin and visfatin levels were linked to advanced tumor stage, greater tumor size, and poor outcomes in breast cancer patients, thus serving as a prognostic biomarker in breast cancer survival [bib_ref] Extracellular Visfatin-promoted malignant behavior in breast cancer is mediated through c-Abl and..., Hung [/bib_ref].
Histologically related adipokine staining is also a potential diagnostic method. There is a close connection between adipocyte hypertrophy and the staging and grading of breast cancer [bib_ref] Associations between markers of mammary adipose tissue dysfunction and breast cancer prognostic..., Laforest [/bib_ref]. The staining assay showed that CCL5 expression located in peri-tumor AT was related to LN and distant metastases, and OS in TNBC patients. Yamaguchi et al. underlined that AT invasion (ATI) at the tumor margin was very concerned about the lymph node involvement in patients with invasive breast ductal carcinoma [bib_ref] Prognostic impact of marginal adipose tissue invasion in ductal carcinoma of the..., Yamaguchi [/bib_ref]. Besides, ATI was related to patient age and invasive tumor size and acted as an excellent prognosis for this breast cancer subtype. Calligaris et al. found it was available to recognize the margin of breast cancer by utilizing desorption electrospray ionization mass spectrometry imaging (DESI-MSI) based on lipid profiles, since several fatty acids, including oleic acid, expressed higher levels in breast cancer tissue than in normal region [bib_ref] Application of desorption electrospray ionization mass spectrometry imaging in breast cancer margin..., Calligaris [/bib_ref]. This study supposed that the method based on lipid profiles was feasible for rapid intraoperative detection of residual breast cancer tissue during breast-conserving surgery.
## Adipocyte secretome in breast cancer therapy
Targeting these tumor-associated adipocytes may intervene or reverse the progression of breast tumors. Inhibition of these adipocytes can significantly impair breast cancer metastasis [bib_ref] Lunasin attenuates obesity-associated metastasis of 4T1 breast Cancer cell through anti-inflammatory property, Hsieh [/bib_ref]. For instance, Alvarez-García et al. confirmed that melatonin could induce the decrease of the TNF-a, IL-6, and IL-11 gene expression in MCF-7 and 3 T3-L1 in their co-culture system [bib_ref] Melatonin interferes in the desmoplastic reaction in breast cancer by regulating cytokine..., Alvarez-García [/bib_ref]. Indeed, the appropriate intervention of specific adipocyte secretory factors described above may result in inhibition of tumor growth or metastasis.
The signatures of adipokines expression can affect the effect of hormone therapy in breast cancer patients. Tamoxifen efficacy could be antagonized by mature adipocytes derived from ADSCs of patients with obesity but normal-weight in vitro. This was related to the elevated level of IL-6, leptin, and TNFα in adipose TME that counteracted on anti-proliferative efficacy of tamoxifen. The up-regulated IL-6 and FGF-2 abrogated or inhibited the anti-VEGF therapy in obese breast cancer patients and mouse model [bib_ref] Obesity promotes resistance to anti-VEGF therapy in breast cancer by up-regulating IL-6..., Incio [/bib_ref]. This effect could be reversed by IL-6 blockade through normalizing tumor vasculature and immunosuppression. Leptin sustained the resistance of aromatase inhibitor anastrozole in MCF-7 cells,
## Exosomes sources exosomes contents target cells function and mechanisms
Ref. [fig_ref] Table 1: The morphology and functon comparison of nBAs, TSAs, and CAAsAbbreviations [/fig_ref] Induced EMT and CSC-related gene expression in BC cells, promoted BC malignancy in type 2 diabetes [bib_ref] Adipocyte-derived exosomes may promote breast cancer progression in type 2 diabetes, Jafari [/bib_ref] hADSCs N/A MCF-7 Promoted MCF-7 migration by activating Wnt signaling in vitro [bib_ref] Exosomes from human adipose-derived mesenchymal stem cells promote migration through Wnt signaling..., Lin [/bib_ref] hADSCs N/A MCF-7 Promoted MCF-7 migration, reduced BC cell apoptosis and altered the transcriptome, promoted MCF-7 growth via activating Hippo signaling pathway [bib_ref] Exosomes secreted by mesenchymal stromal/stem cell-derived adipocytes promote breast cancer cell growth..., Wang [/bib_ref] Pre-adipocytes miR-140 DCIS cells Influenced CSC renewal, BC cell migration, and neoplasia through miR-140/SOX2/ SOX9 axis [bib_ref] Targeting exosomes from preadipocytes inhibits preadipocyte to cancer stem cell signaling in..., Gernapudi [/bib_ref] thus leptin disturbance would be beneficial for patients with hormone-resistant breast cancer [bib_ref] Leptin signaling contributes to aromatase inhibitor resistant breast cancer cell growth and..., Gelsomino [/bib_ref]. Adipocytesecreted leptin also enhanced the HER2 protein levels through a STAT3-mediated up-regulation of Hsp90 in breast cancer cells in vitro [bib_ref] Leptin increases HER2 protein levels through a STAT3-mediated upregulation of Hsp90 in..., Giordano [/bib_ref]. Long-term exposure to leptin could markedly inhibit the sensitivity of breast cancer cells to the estrogen tamoxifen [bib_ref] Hormonal therapy resistance and breast cancer: involvement of adipocytes and leptin, Delort [/bib_ref]. verified that in radiotherapy, AT developed an inflammatory wound-healing response and increased the secretion of ATX as well as other inflammatory cytokines, which could help breast cancer resistance to radiotherapy [bib_ref] Implications for breast cancer treatment from increased autotaxin production in adipose tissue..., Meng [/bib_ref].
## Ir adipocytes
CSCs play an important role in tumor evolution and treatment resistance. Adipocyte-derived leptin was certified to propel the breast CSC self-renewal and paclitaxel chemoresistance via activating JAK/STAT3-regulated FAO in mouse breast cancer tumors in vivo [bib_ref] JAK/STAT3-regulated fatty acid β-oxidation is critical for breast Cancer stem cell self-renewal..., Wang [/bib_ref]. Therefore, the secretion of adipocyte factors could endow drug resistance in breast cancer cells, by enhancing the release of inflammatory mediators and reducing FFA [bib_ref] The paracrine effects of adipocytes on lipid metabolism in doxorubicin-treated triple negative..., Mentoor [/bib_ref]. This process causes inflammation and lipid metabolic remodeling, and ultimately confers to the chemotherapeutic drug resistance.
## Future perspectives and challenges
Peritumoral tissue, stromal tissue, AT, and secreted adipokines have important roles in breast cancer biology. Adipocyte secretion profiles, especially TSAs and CAAs, are becoming critical contributors to breast carcinogenesis, diagnosis, and treatment. Currently, there are still many unresolved questions in terms of this field. Firstly, there is a varied, dynamic, and complicated interplay between adipocytes and tumor cells. Breast cancer stimulation can induce adipocytes to revert to an immature and proliferative phenotype, thus in turn promoting breast cancer cell malignancy [bib_ref] Cancer-mediated adipose reversion promotes cancer cell migration via IL-6 and MCP-1, Fujisaki [/bib_ref]. Adipocytes can also integrate the substances from the metabolic environment and promote the growth of breast cancer cells. Additionally, fatty acids released by adipocytes are taken up by breast cancer cells and they are utilized in the biosynthesis of new lipid compounds in breast cancer cells [bib_ref] MCF-7 drug resistant cell lines switch their lipid metabolism to triple negative..., Adriá-Cebrián [/bib_ref]. Adipocytes and breast cancer cells are competent to respond to endocrine and paracrine signals derived from both sides. Currently, most of the related studies have focused on the role of the adipocyte secretion profile in promoting breast cancer. However, some adipocyte factors have certain oncogenic effects or dual functions, thus playing multifaceted roles in different breast cancer processes and types. For example, oleic acid is deemed as an abundant monounsaturated fatty acid in the human body, which can release from adipocytes and affects different energy metabolism reactions [bib_ref] Role of oleic acid in the gut-liver Axis: from diet to the..., Piccinin [/bib_ref]. The highly metastatic breast cancer cells could utilize oleic acid to sustain malignancy by AMPK activationmediated β-oxidation. Nevertheless, oleic acid inhibited growth and survival in low metastatic MCF-7 cells in vitro [bib_ref] High metastaticgastric and breast cancer cells consume oleic acid in an AMPK..., Li [/bib_ref]. Leptin influences the secretion of inflammatory cytokines in breast cancer cells, such as IL-1, IL-6, IL-17, and TNF-α, as well as the secretion of various growth factors and cytokines by the tumor stroma [bib_ref] Adipokines as therapeutic targets in breast cancer treatment, Cha [/bib_ref]. Obesity status leads to changes in the traits and functions of adipocytes, which in turn affects more aggressive changes in tumors. For example, Bowers et al. found that obesity-induced systemic IL-6 significantly increased the aromatase expression of pre-adipocytes through up-regulating breast cancer-derived prostaglandin E2 (PGE2) in postmenopausal patients [bib_ref] Obesity-associated systemic interleukin-6 promotes pre-adipocyte aromatase expression via increased breast cancer cell..., Bowers [/bib_ref]. In different metabolic environments, such as obesity, the precise mechanisms controlling these interactions are more complex. Targeting factors of adipocyte origin have the potential to be a useful approach to improving the prognosis of breast cancer patients in the setting of obesity.
Secondly, there is also a network of interactions between adipocyte factors that promote or constrain respective physiopathological functions. There were synergistic or antagonistic expressions among different factors. The expression of ADSC-derived IGF-1, HGF, VEGF, and IL-8 was higher in breast cancer tissue compared with the normal group [bib_ref] Expression profile of IL-8 and growth factors in breast cancer cells and..., Razmkhah [/bib_ref]. Therefore, the diversiform ADSCs distributed in breast cancer entity release large amounts of tumor-promoting factors to support breast cancer development and evolution. The co-culture of breast cancer cells and primary human adipocytes enhanced VEGF and leptin secretion in adipocytes, and inhibition of VEGF in vitro also resulted in a dosedependent decrease in leptin levels [bib_ref] Adipokines and vascular endothelial growth factor in Normal human breast tissue in..., Morad [/bib_ref]. Therefore, it is a thorny and pressing problem to determine which secreted factor plays the main predominant function at a particular stage of tumor progression.
The development of novel sequencing technologies, including metabolomics, proteomics, single-cell omics, and spatial transcriptomics, can uncover potential adipocyte secretion profiles and key molecules that are (See figure on next page.) The pleiotropic roles of adipocyte secretome in remodeling breast cancer (BC). Adipocytes can affect themselves by autocrine and also affect or fuel tumors by paracrine. Reciprocally, breast cancer cells are also able to change the dedifferentiation, and remodel the function, secretome, and metabolism of adipocytes. The adipocyte-released secretome components, including cytokines (IL-6, IL-8, IL-1β, CCL2, CCL5, IGF-1, IGFBP, VEGF, and G-CSF), adipokines (leptin, adiponectin, ATX, resistin, visfatin, adipsin, MMP, apelin, and SFRP5), and lipid metabolites (FFA, β-HB, ETP, and Lcn2), profoundly impact on remodeling breast cancer progression and drug-resistance via multiple complicated mechanisms. Particularly, adipocyte exosomes are cellular vesicles to transport TSP5, miR-140, and other messages, leading to the altered behaviors of breast cancer cells expected to serve as potential diagnostic and therapeutic targets. Using scRNA-seq and snRNA-seq acting on adipose-specific knockout mice, will shed light on novel cell types or paracrine effects involved in the secreted signature of adipocytes [bib_ref] Single-cell analysis of human adipose tissue identifies depot-and disease-specific cell types, Vijay [/bib_ref]. The application of singlecell profiling in human AT can identify particular cell types for lipid storage and disease, including AT aging, adaptive thermogenesis, and tumor remodeling [bib_ref] The impact of Single-cell genomics on adipose tissue research, Deutsch [/bib_ref]. However, currently isolating intact and single adipocytes is complex due to their fragile characteristics. Similar sequencing strategies have not been thoroughly applied in human AT. But most certainly, these techniques will be instrumental in elucidating the broad range of molecular functions and adipocyte heterogeneity. Given the complex role of the aforementioned adipocytes and factors in TME, targeting the pro-tumor pathways of these cells and factors, such as paracrine effects, metabolite reprogramming, and immunomodulatory effects, would be an effective way to combat breast cancer. However, targeting CAA or related components alone may not be able to completely ablate malignant cancer cells and seeds. More promisingly, a combination of cytostatic agents, specific factor inhibitors, and other conventional therapies will help to treat breast cancer.
# Conclusion
Adipocytes are phenotypically and functionally heterogeneous. In setting of breast cancer, adipocytes act as active facilitators, but not silent bystanders. There are pleiotropic roles of adipocyte secretome in remodeling breast cancer progression. The released cytokines, adipokines, and lipid metabolites from multiple adipocytes, are crucial orchestrators in shaping breast cancer growth, metastasis, immune regulation, exosome-mediated transfer, and breast cancer diagnosis and therapy . The comprehensive understanding of specific adipose secretome characteristics and interactions within TME cell populations will enable us to better tailor strategies for tumor stratification management and treatment.
[table] Table 1: The morphology and functon comparison of nBAs, TSAs, and CAAsAbbreviations: BC Breast cancer, CAAs Cancer-associated adipocytes, nBAs Normal breast adipocytes, TSAs Tumor-stromal adipocytes Adipocytes located at the stroma of tumor Abnormal adipocytes located at the front of the tumor inva-Morphology Normal size, round shape, riched in lipid droplets that occupy 90% of the cell volume Smaller size, spindle or ellipsoidal, reduced lipid contends Dedifferentiated state, smaller size, fibroblast-like phenotype, smaller and reduced lipid contends [/table]
[table] Table 2: The role of adipocytes-derived exosomes in regulating BC progression Abbreviations: BC Breast cancer, CSC Cancer stem cell, DCIS Ductal carcinoma in situ, EMT Epithelial-mesenchymal transition, hADSCs Human adipose-derived stem cells, IR Insulin resistance, TSP5 Thrombospondin 5 [/table]
|
Placental Pathology in Pregnancies with Maternally Perceived Decreased Fetal Movement - A Population-Based Nested Case-Cohort Study
Background: Decreased fetal movements (DFM) are associated with fetal growth restriction and stillbirth, presumably linked through an underlying placental dysfunction. Yet, the role of placental pathology has received limited attention in DFM studies. Our main objective was to explore whether maternal perceptions of DFM were associated with placental pathology in pregnancies recruited from a low-risk total population. Methods/Principal Findings: Placentas from 129 DFM and 191 non-DFM pregnancies were examined according to standardized macro-and microscopic protocols. DFM was defined as any maternal complaint of DFM leading to a hospital examination. Morphological findings were timed and graded according to their estimated onset and clinical importance, and classified in line with a newly constructed Norwegian classification system for reporting placental pathology. With our population-based approach we were unable to link DFM to an overall measure of all forms of placental pathology (OR 1.3, 95% CI 0.8-2.2, p = 0.249). However, placental pathology leading to imminent delivery could be a competing risk for DFM, making separate subgroup analyses more appropriate. Our study suggests a link between DFM and macroscopic placental pathology related to maternal, uteroplacental vessels, i.e. infarctions, placental lesions (intraplacental hematomas) and abruptions. Although not statistically significant separately, a compound measure showed a significant association with DFM (OR 2.4, 95%CI 1.1-5.0, p = 0.023). This association was strengthened when we accounted for relevant temporal aspects. More subtle microscopic materno-placental ischemic changes outside the areas of localized pathology showed no association with DFM (OR 0.5, 95%CI 0.2-1.4, p = 0.203). There was a strong association between placental pathology and neonatal complications (OR 2.9, 95% CI 1.6-5.1, p,0.001).Conclusions: In our population-based study we were generally unable to link maternally perceived DFM to placental pathology. Some associations were seen for subgroups. Citation: Winje BA, Roald B, Kristensen NP, Frøen JF (2012) Placental Pathology in Pregnancies with Maternally Perceived Decreased Fetal Movement -A Population-Based Nested Case-Cohort Study. PLoS ONE 7(6): e39259.
# Introduction
Fetal growth restriction (FGR) is associated with significant risk for severe disabilities and death [bib_ref] Perinatal outcome in SGA births defined by customised versus population-based birthweight standards, Clausson [/bib_ref] [bib_ref] Restricted fetal growth in sudden intrauterine unexplained death, Frøen [/bib_ref] [bib_ref] Pregnancy outcomes associated with intrauterine growth restriction, Gilbert [/bib_ref]. Risk can be reduced by appropriate assessment, but the ability to detect FGR in antenatal care remains weak [bib_ref] Does antenatal identification of small-forgestational age fetuses significantly improve their outcome?, Lindqvist [/bib_ref]. The mother's perception of fetal movement (FM) is still the simplest source of information about the baby's well-being and should not be underrated. A maternal perception of decreased fetal movement (DFM) is widely reported to be associated with FGR [bib_ref] What investigation is appropriate following maternal perception of reduced fetal movements?, Heazell [/bib_ref] [bib_ref] Correction: Reduction of late stillbirth with the introduction of fetal movement information..., Tveit [/bib_ref] [bib_ref] Reduction of late stillbirth with the introduction of fetal movement information and..., Tveit [/bib_ref] [bib_ref] Subjective recording of fetal movements. III. Screening of a pregnant population; the..., Valentin [/bib_ref] [bib_ref] Maternal perception of decreased fetal movement as an indication for antepartum testing..., Whitty [/bib_ref]. DFM is also reported the days preceding an unexplained stillbirth [bib_ref] Risk factors for sudden intrauterine unexplained death: Epidemiologic characteristics of singleton cases..., Frøen [/bib_ref] [bib_ref] Fetal activity and fetal wellbeing: an evaluation, Pearson [/bib_ref] [bib_ref] Daily fetal movement recording and fetal prognosis, Sadovsky [/bib_ref] , suggesting that interventions could have prevented morbidity and mortality [bib_ref] Stillbirths: the way forward in high-income countries, Flenady [/bib_ref] [bib_ref] A kick from within -fetal movement counting and the cancelled progress in..., Frøen [/bib_ref]. The majority of women examined for perceived DFM in third trimester, however, continues with uncomplicated pregnancies [bib_ref] Maternal characteristics and pregnancy outcomes in women presenting with decreased fetal movements..., Tveit [/bib_ref]. So even if a maternal perception of DFM is rightly recognized as a good indicator of fetal compromise, its predictive value is low.
The well-documented association between DFM, FGR and stillbirth [bib_ref] Perinatal outcome in SGA births defined by customised versus population-based birthweight standards, Clausson [/bib_ref] [bib_ref] Restricted fetal growth in sudden intrauterine unexplained death, Frøen [/bib_ref] [bib_ref] Pregnancy outcomes associated with intrauterine growth restriction, Gilbert [/bib_ref] [bib_ref] Does antenatal identification of small-forgestational age fetuses significantly improve their outcome?, Lindqvist [/bib_ref] is presumably linked to an underlying placental dysfunction [bib_ref] Identifying placental dysfunction in women with reduced fetal movements can be used..., Warrander [/bib_ref]. Pathological processes in the placenta may lead to fetal hypoxia [bib_ref] Review of some causes of stillbirth, Bendon [/bib_ref] , either following profound acute circulatory insults such as abruptions and hemorrhages, or longstanding processes resulting in prolonged chronic hypoxia. When exposed to nutrient and oxygen restriction, it is hypothesized that the fetus will redistribute blood to vital organs [bib_ref] Middle cerebral artery flow velocity waveforms in normal and small-for-gestational-age fetuses, Mari [/bib_ref] and will reduce nonvital activities such as gross fetal movements, [bib_ref] Effects of hypoxemic events on breathing, body movements, and heart rate variation:..., Bekedam [/bib_ref] [bib_ref] Assessment of fetal heart rate and fetal movements in detecting oxygen deprivation..., Bocking [/bib_ref] [bib_ref] Prediction of fetal acidaemia in intrauterine growth retardation: comparison of quantified fetal..., Ribbert [/bib_ref]. Studies have reported that growth restricted fetuses have reduced fetal movement compared to controls [bib_ref] Motor behaviour in the growth retarded fetus, Bekedam [/bib_ref] and that they demonstrate an almost dose-dependent reduction in FM during hypoxia [bib_ref] Prediction of fetal acidaemia in intrauterine growth retardation: comparison of quantified fetal..., Ribbert [/bib_ref] [bib_ref] The Effect of Intrauterine Growth-Retardation on the Quality of General Movements in..., Sival [/bib_ref] [bib_ref] Computerized analysis of behavior in fetuses with congenital abnormalities, Vindla [/bib_ref] [bib_ref] Comparison of unstimulated and stimulated behaviour in human fetuses with congenital abnormalities, Vindla [/bib_ref]. DFM has been found to be associated with abnormal placental morphology paralleling those seen in placentas in FGR pregnancies [bib_ref] Identifying placental dysfunction in women with reduced fetal movements can be used..., Warrander [/bib_ref]. Although it is generally assumed and clinically plausible that DFM reflects fetal adjustment to a negative energy balance induced by reduced placental function, evidence to support this is limited. The first study on placental morphology was published just recently and reported altered placental structure and function with DFM [bib_ref] Maternal perception of reduced fetal movements is associated with altered placental structure..., Warrander [/bib_ref].
A perception of DFM often causes anxiety [bib_ref] Low maternal awareness of fetal movement is associated with small for gestational..., Saastad [/bib_ref] [bib_ref] Implementation of uniform information on fetal movement in a Norwegian population reduces..., Saastad [/bib_ref] and results in frequent unscheduled third trimester antenatal visits [bib_ref] Correction: Reduction of late stillbirth with the introduction of fetal movement information..., Tveit [/bib_ref] [bib_ref] Reduction of late stillbirth with the introduction of fetal movement information and..., Tveit [/bib_ref] [bib_ref] A kick from within -fetal movement counting and the cancelled progress in..., Frøen [/bib_ref]. So far, however, placental pathology in DFM studies have been inadequately pursued [bib_ref] Identifying placental dysfunction in women with reduced fetal movements can be used..., Warrander [/bib_ref]. A prospective FM counting study with a subsequent blinded study of the placenta was initiated to reveal information that may help to identify the DFM pregnancies at greatest risk.
The placenta substudy forms the basis for this report. Our main objective was to explore whether maternal perceptions of DFM were associated with placental pathology in pregnancies recruited from a low-risk total population. We hypothesized that DFM placentas would show morphological changes consistent with reduced placental function.
# Methods
# Ethics statement
Written informed consent was obtained from all participants, both for the FM counting study and the morphological examination of the placenta following delivery. The study was approved by the Regional Committee for Medical Research Ethics, S-08694d, 2008/18353, 06. . There were no minors or legally incompetent participants in the study.
## The fm counting study
The placenta study is a case-cohort nested within a broader prospective FM counting study initiated to explore FM counting patterns and their relation to adverse pregnancy outcome. Thus all pregnancies included in this placenta study were selected among women who were already included in the population-based FM counting study.
From July 2009 to July 2011, all women with singleton pregnancies attending Østfold Hospital Trust for routine ultrasound screening in pregnancy week 17-19 were invited to the study. After written informed consent, a total of 2468 women were enrolled in the FM counting study, representing 42% of the eligible population. Among them, 1445 (59%) later submitted their FM chart and thus form the study group. Compared to the total population of pregnant women at Østfold Hospital Trust (data from Medical Birth Registry of Norway, year 2009 used as a reference, the study group included more primiparous women (RR 1.2, 95%CI 1.2-1.3, p,0.001), fewer smoking mothers (RR 0.5, 95%CI 0.4-0.5, p,0.001), fewer cesarean sections (RR 0.8, 95%CI 0.7-0.8, p,0.001), fewer preterm (RR 0.8, 95%CI 0.6-0.9, p = 0.028), and low birth weight babies (RR 0.6, 95%CI 0.4-0.9, p = 0.006) (data not shown).
Participating women systematically recorded FM daily with a modified ''count-to-ten'' approach, i.e. the time needed to perceive ten movements. The counting protocol was according to guidelines from the international collaboration Fetal Movement Intervention Assessment (FEMINA) [bib_ref] Correction: Reduction of late stillbirth with the introduction of fetal movement information..., Tveit [/bib_ref] [bib_ref] Reduction of late stillbirth with the introduction of fetal movement information and..., Tveit [/bib_ref] [bib_ref] Analysis of 'count-to-ten' fetal movement charts: a prospective cohort study, Winje [/bib_ref]. The information provided to women about DFM and when to seek medical attention is presented in full in Textbox S1 (Other 1). Women were not provided with any fixed limits for DFM, but advised to report significant and sustained decreases in the baby's normal activity. In the current report DFM is defined as any maternal concern leading to a hospital examination.
## The placenta substudy
From this prospective FM counting cohort there were two different criteria for eligibility to the placenta study: (i) if the mother had been examined in hospital care for a concern for DFM after 24 0 pregnancy week, or (ii) if the mother was among pregnancies preselected to the placenta study at time of enrollment in the FM counting study, independent of pregnancy outcome (a population-based sample as controls). Some of the women preselected to the population sample also experienced DFM and were included as DFM pregnancies in the analyses, [fig_ref] Figure 1: Flowchart for data collection [/fig_ref]. Only babies without malformations were included in the analyses.
Since the placenta study was complementary to the FM counting study, a preceding power calculation for a case-control design was not performed. However, with our placenta sample a power calculation shows that we would have been able to detect an odds ratio for placental pathology of 2.0 in DFM compared to non-DFM pregnancies.
## Placental examination
Information on gestational age, birth weight, and Apgar scores was available for pathologists at time of examination. All placentas were examined macroscopically by four designated pathologists according to a standardized protocol. All DFM placentas and a selection of non-DFM placentas were examined microscopically by a single, experienced pathologist with special interest in placental pathology (BR). The non-DFM placentas as controls were selected independently from their birth outcome. For every two DFM placentas we selected three non-DFM placentas (casecontrol ratio1:1.5). The pathologist performing the microscopic examinations was blinded for macroscopic findings and DFM information.
Placentas were weighed (without cord and membranes), measured and inspected for focal lesions. Focal lesions were reported as estimated % of total placental volume, location central or peripheral and arbitrarily timed as acute (hemorrhagic changes) (,48 hours), subacute (hemorrhagic and fibrous changes) (2-20 days), or longstanding (fibrous changes) (. = 21 days). The gross macroscopic pathology was graded according to assumed clinical impact as:
(1) no pathology: placentas without abnormalities (2) minor impact: abnormalities in placental shape, bilobate placenta, circumvallate placenta without bleeding, meconium stained membranes, peripheral infarctions (,10%) (3) potential impact: velamentous or marginal cord insertion, true knots (4) moderate impact: infarctions (central infarctions 5-9% or peripheral infarctions $10%) (5) significant impact: focal lesions like central infarctions and hemorrhages $10%, abruptions.
Standard tissue sections were taken from (i) membranes and umbilical cord, (ii) cord insertion site and placental near cord, (iii) full thickness macroscopically normal placenta, and (iv) two sections from the maternal plate. Additional sections were taken from centrally located focal parenchymal lesions. The sections were routinely formalin fixed, processed and embedded in paraffin. For the microscopic review only sections stained with Hematoxylin and Eosin (HE) were used.
Placental pathology from the microscopic examinations was categorized into nine groups according to a new Norwegian classification scheme [bib_ref] A systematic and standardized classification of placental pathology: A Norwegian enterprise, Roald [/bib_ref] , [fig_ref] Table 1: A systematic and standardized classification of placental pathology[34] [/fig_ref]. The assumed clinical impact of the various processes was graded similarly to the macroscopic examination: (0) no pathology, (1) minor-, (2) potential-(3) moderate-and (4) significant impact, and timed accordingly. Only pathologies with clinical impact grade 3 and 4 are included as pathology in the analyses. Separate analyses are presented for pathologies with clinical impact grade 2 (potential impact). Analyses are based on the last DFM consultation if several.
Linking placental pathology present at term to early third trimester DFM consultations may be dubious. To get a more valid estimate of the association between DFM and placental pathology, we performed two separate subanalyses. First, we delimited the subset to DFM consultations occurring within the last seven days before birth and compared placental pathology between DFM and non-DFM pregnancies. With this approach the placental pathology would most likely precede the DFM consultation in time. Second, we delimited the subset to DFM consultations occurring within the last 21 days before birth and included only acute and subacute placental pathology, i.e. with estimated onset within the last 21 days, and compared these pathologies between DFM and non-DFM pregnancies. We were then able to assess whether DFM and placental pathology likely coincided in time.
Demographic indicators and information on birth outcome were collected from antenatal pregnancy charts and hospital records. Birth weight was adjusted for gestational age and sex. Baby weight below the 10 th percentile was classified as small for gestational age (SGA) [bib_ref] A direct method for ultrasound prediction of day of delivery: a new,..., Gjessing [/bib_ref]. We have defined neonatal complications as preterm birth, SGA, infections, Apgar scores ,7 5min , or transfer to neonatal care unit for conditions relevant to fetal growth restriction or fetal distress, including respiratory syndrome and cerebral irritation. Classifications comply with definitions from Medical Birth Registry of Norway. Respiratory distress is defined as typical signs or X ray findings, and cerebral irritation is defined as unrest, trembling, stiffness, and other signs of cerebral excitation.
## Statistical analyses
We used SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). To compare the likelihood of events between groups, we calculated odds ratios (OR) with 95% confidence intervals (CI) or relative risks (RR) when appropriate. Two samples t test was used to explore relationships between continuous variables. The level of statistical significance was set at p,0.05.
# Results
The final sample of 320 placentas with complete macroscopic and microscopic examinations consisted of 129 DFM and 191 non-DFM placentas. All babies included in the analyses were live born. For data collection, see flow chart [fig_ref] Figure 1: Flowchart for data collection [/fig_ref]. There were no statistically significant differences in maternal characteristics between DFM and non-DFM pregnancies, .
Generally, we were unable to link DFM to an overall measure of all forms of placental pathology with statistical significance (OR 1.3, 95% CI 0.8-2.2, p = 0.249), . However, DFM seems more closely associated with macroscopic placental pathology related to maternal, uteroplacental vessels, i.e. infarctions, placental lesions (intraplacental hematomas) and abruptions. All odds ratios for these subcategories were higher than unity, although not statistically significant separately. Yet, when these subcategories were merged, the compound measure showed a significant association with DFM (OR 2.4, 95%CI 1.1-5.0, p = 0.023), . This association was strengthened when we restricted the analysis to DFM consultations within the last seven days before birth (OR 3.0, 95%CI 1.1-7.6, p = 0.025). The same applied to the subsample with DFM consultations and estimated onset of placental pathology within the last 21 days preceding birth (OR 3.5. 95%CI 1.1-11.3, p = 0.038).
There was no association with DFM for more subtle microscopic materno-placental ischemic changes outside the areas of localized pathology (OR 0.5, 95%CI 0.2-1.4, p = 0.203), . In cases of acute chorioamnionitis we found no association, as the DFM consultations preceded the pathology onset by large margins and thus were unrelated. Placentas from DFM and non-DFM pregnancies were similar in mean trimmed weight and mean fetal placental weight ratio across samples. A quintile distribution of placental weight showed no differences between DFM and non-DFM pregnancies. Fetal vessels in the membranes are vulnerable to injury and thrombosis, and are more susceptible to compression by fetal parts resulting in obstruction of blood flow. We found cord anomalies, including true knots and velamentous and marginal cord insertions, to be similar between the groups, . Since abnormal cord insertion site has been linked to SGA [bib_ref] Gross morphological changes of placentas associated with intrauterine growth restriction of fetuses:..., Biswas [/bib_ref] as well as DFM [bib_ref] Maternal perception of reduced fetal movements is associated with altered placental structure..., Warrander [/bib_ref] , we restricted the analysis to include only velamentous and marginal cord insertions. However, results were the same (data not shown). Ten women had more than one DFM consultation. This is presented in [fig_ref] Table 4: Characteristics of consultations for DFM from 129 pregnancies [/fig_ref] with information on gestational age at time of DFM and days between DFM and birth. We found no association between having recurrent DFM consultations and overall placental pathology (OR 1.5, 95% CI 0.4-5.8, p = 0.525) or between having recurrent DFM consultations and neonatal complications (OR 1.9, 95% CI 0.8-4.6, p = 0.167). For DFM infants later diagnosed as SGA, the median time between diagnosis of intrauterine growth restriction (fetal weight estimate ,-10% by ultrasound measurement) and delivery was 20 days, range 2-63.
On birth outcome a strong association between placental pathology and neonatal complications was found, [fig_ref] Table 5: Placental pathology by birth outcome [/fig_ref]. The strongest associations with birth outcome were seen for placental pathology in category four according to the Norwegian classification system, i.e. materno-placental circulatory disorders. This category relates to maternal vascular pathology. It includes both longstanding, chronic placental processes such as old infarctions and diffuse ischemic changes, and acute episodes that have occurred closer to birth like abruptions. These associations were present both for the more abrupt circulatory insults and the subtle ischemic changes.
Neither placental ascending infections (placental pathology category two) nor cord anomalies were associated with the birth outcomes, [fig_ref] Table 5: Placental pathology by birth outcome [/fig_ref]. The remaining categories were small, which limited subgroup analyses.
We found no statistically significant differences in birth outcome between DFM and non-DFM pregnancies, .
## Representativeness of sample
Placentas from approximately two thirds of the DFM pregnancies were eventually included: 85% from DFM pregnancies that were among pregnancies initially preselected to the population sample and 46% from DFM pregnancies outside the population sample, [fig_ref] Figure 1: Flowchart for data collection [/fig_ref]. In 78 (38%) of the DFM pregnancies the placenta was lost to the study. A sensitivity analysis showed that DFM pregnancies with (n = 129) or without (n = 78) the placenta included were similar in terms of mean infant birth weight (3555 grams (SD 609) versus 3634 grams (SD 523), p = 0.350) and mean gestational age at birth (40 0 weeks (SD 2) versus 40 3 weeks (SD 1.6), p = 0.156). They were also similar in terms of neonatal complications (OR 1.1, 95% CI 0.5-2.1, p = 0.872), SGA (OR 0.9, 95% CI 0.4-2.3, p = 0.854) and preterm infants (OR 1.2, 95% CI 0.4-4.2, p = 0.749).
# Discussion
To our knowledge, this is the first population-based study that compares placental morphology in DFM and non-DFM pregnancies. We were unable to link DFM to placental pathology with statistical significance, although our data suggest higher odds for a subgroup of placental pathology in DFM-pregnancies, primarily related to abrupt circulatory insults. The maternal ability to detect SGA and neonatal complications was limited. With our population-based approach we faced the well-known challenge of low power when studying rare events in prospective cohort designs. However, the placenta study is part of a broader FM counting study where this design was most suitable.
Generally, placental pathology contributed little to explain third trimester maternally perceived DFM. Two factors need to be mentioned. First, some forms of placental pathology are known to trigger imminent delivery, among them acute chorioamnionitis, and may serve as competing risk for DFM. An overall measure of placental pathology could therefore be misleading, since placental pathology relevant for DFM may be underestimated. So subgroup analyses may be more appropriate. Second, linking placental pathology present at term to early third trimester DFM consultations may be dubious. Placental pathology must precede the DFM consultation in time to be relevant for DFM. By necessity, a retrospective estimate of pathology onset is broad, especially for pathology with estimated onset .21 days prior to birth.
To get a more valid estimate of the associations between DFM and placental pathology, we therefore restricted the analyses to include only DFM consultations occurring within the last seven days before birth. The result from this subgroup analysis confirmed and even strengthened the result from the total sample; DFM seemed more associated with abrupt, major circulatory insults resulting from obstruction of maternal uteroplacental vessels. Discrepancies between the two analyses were primarily linked to cases of acute chorioamnionitis, which were unrelated in time with the DFM consultation, similar to what was seen for the more diffuse ischemic changes.
The same result emerged when we restricted the analysis to include DFM pregnancies occurring within the last 21 days before birth and compared only placental pathology with acute or subacute onset (,21 days) between DFM and non-DFM pregnancies, implying that events were more likely to have coincided in time. Again the associations between DFM and macroscopic maternal vascular pathology were strengthened. Thus associations between DFM and more abrupt circulatory events remained also when temporal associations were accounted for. The clinical implications are, however, not clear. Important macroscopic indicators of placental function, such as placental trimmed weight and fetal placental weight ratio, were not different between DFM and non-DFM pregnancies in our study, indicating overall healthy placentas with substantial reserve capacity in both groups. Placental weight has previously been found to be predictive of maternal disease, obstetric outcome and perinatal morbidity and mortality [bib_ref] Placenta weight percentile curves for singleton and twins deliveries, Almog [/bib_ref].
There are few studies to support or refute our findings, as research linking placental dysfunction to DFM is scarce. The first study actually investigating placenta morphology in DFM pregnancies was presented just recently [bib_ref] Maternal perception of reduced fetal movements is associated with altered placental structure..., Warrander [/bib_ref]. In comparing placentas from 36 DFM pregnancies with 36 healthy controls, striking differences were reported. Placentas from DFM pregnancies were smaller (lighter, with smaller surface area), had more macroscopic infarctions, and were more likely to have abnormal shape and eccentric cord insertion than those from healthy pregnancies. Microscopic examination revealed ischemic changes indicating maternal vascular pathology with increased number of syncytial knots, fewer blood vessels, and reduced area of trophoblast per villus [bib_ref] Maternal perception of reduced fetal movements is associated with altered placental structure..., Warrander [/bib_ref]. While these results apparently differ from our findings, direct comparison may be deceiving. The previous study included only pregnancies where perceived DFM lasted more than 12 hours and where the baby was delivered within seven days of presentation, representing 12% of the DFM consultations in the study (36/305). These selected high-risk DFM pregnancies were compared with selected healthy controls, i.e. sick versus healthy. In our population-based approach, we compared women with and without a maternal complaint for DFM without further selection, a measure known to have low predictive value, but important in clinical practice. These differences in design are clearly reflected in the study samples. Their DFM sample included a substantial number of smaller placentas with lower fetal placental weight ratio, whereas DFM and non-DFM placentas in our populationbased sample were comparable in size. The differences in placental ischemic changes in the two studies may mainly reflect the differences in the study cohorts, partially also differences in criteria and classification. The pathology examination procedure differed between the studies. Again our study has a focus on every-day approaches. We have thus used standard, routine examination protocols, both in the macroscopic and microscopic examinations, assessing HE sections only.
While acknowledging that differences in our study were expected to be smaller, our design deliberately aimed at being relevant for the everyday situations facing obstetricians and midwives. With the similarities in aims between our and the previous study, we have purposely amended our analysis where appropriate to facilitate comparison. The studies are thus complementary and each provides building blocks to fill in the knowledge gaps.
The association between placental pathology and FGR and stillbirth has previously been documented [bib_ref] Placental findings contributing to fetal death, a study of 120 stillbirths between..., Kidron [/bib_ref] [bib_ref] Placental pathology: a systematic approach with clinical correlations, Redline [/bib_ref] [bib_ref] Placental pathology of fetal growth restriction, Salafia [/bib_ref]. Consistent with these studies we found strong associations between SGA and placental pathology, both for the non-macroscopic, microscopically identified ischemic changes and for the macroscopically more abrupt, circulatory insults. However, we only found an association with DFM for the latter category. Birth outcomes appear similar between DFM and non-DFM pregnancies. This is in line with the overall result of minor differences in placental pathology between DFM and non-DFM placentas in our study. In addition, the effect of focal obstruction of maternal uteroplacental vessels is potentially less severe in normally sized placentas, as in our sample, with capacity for compensatory mechanisms. Thus major differences in birth outcome should not be expected.
Other factors may also have improved birth outcomes for DFM pregnancies. The effect of being included in a study often ii Small for gestational age: birth weight for gestational below 10 th percentile adjusted for maternal height and pre pregnancy weight and infant sex. influences participant behavior, usually in a beneficial direction, i.e. maternal care seeking behavior. Improved clinical care such as appropriate management, timing of delivery and delivery interventions may be more likely in DFM pregnancies. Our study was, however, neither designed for nor powered to measure such effects of maternal monitoring of FM on birth outcome. The strong association between placental pathology and neonatal complications may have been reinforced as pathologists were informed about gestational age, birth weight and Apgar scores. This information is, however, vital in placental examination since placenta is a dynamic organ, constantly developing and maturing throughout pregnancy. In terms of DFM, a number of macro examinations were conducted during parts of the study period when DFM placentas were the only placentas collected, implying that pathologists were inevitably aware of DFM status during part of the study period. The strict protocol for macroscopic registration and standardized sectioning makes this awareness less prone to bias. Importantly, when micro examinations were conducted, the pathologist was unaware of DFM status. The well defined microscopic criteria of the newly constructed Norwegian classification scheme were strictly applied.
Consistent with earlier studies [bib_ref] Analysis of 'count-to-ten' fetal movement charts: a prospective cohort study, Winje [/bib_ref] [bib_ref] Fetal movement assessment, Frøen [/bib_ref] [bib_ref] Methods of fetal movement counting and the detection of fetal compromise, Heazell [/bib_ref] , we found that the predictive value of maternally perceived FM for identification of SGA and neonatal complications was low. Women's perception of FM is known to be affected by pathological and non-pathological entities [bib_ref] Methods of fetal movement counting and the detection of fetal compromise, Heazell [/bib_ref]. A valuable contribution from the previous study was that it managed to link DFM to placental pathology [bib_ref] Maternal perception of reduced fetal movements is associated with altered placental structure..., Warrander [/bib_ref]. However, this was based on a highly selected risk group of DFM pregnancies representing less than one percent of its obstetric source population. With our population-based approach we were unable to replicate these results, illustrating how difficult it can be to interpret DFM both for the mothers and health care professionals. Since women will continue to report concerns for DFM [bib_ref] Correction: Reduction of late stillbirth with the introduction of fetal movement information..., Tveit [/bib_ref] [bib_ref] Reduction of late stillbirth with the introduction of fetal movement information and..., Tveit [/bib_ref] [bib_ref] A kick from within -fetal movement counting and the cancelled progress in..., Frøen [/bib_ref] , simple tools to help mothers maintain a safe pregnancy is needed.
In a recent Lancet series on stillbirth prevention, screening for placental insufficiency and better management of DFM pregnancies were rated among top ten research priorities [bib_ref] Stillbirths: the way forward in high-income countries, Flenady [/bib_ref]. A natural first step would be to improve women's ability to recognize the important changes in FM so as to ensure appropriate care-seeking behavior. In the broader FM counting study, placenta data will be included as an objective measure to explore whether FM counting patterns contain information that can support maternal common sense. Given appropriate care seeking behavior, the potential role of placental biomarkers may provide a promising supplement to identify those DFM pregnancies at highest risk of poor outcome. Preliminary results from a DFM study have reported that DFM pregnancies with poor birth outcome showed reduced plasma concentrations of hCG and hPL compared to DFM pregnancies with normal outcome [bib_ref] Effects of hypoxemic events on breathing, body movements, and heart rate variation:..., Bekedam [/bib_ref] [bib_ref] Placental pathology of fetal growth restriction, Salafia [/bib_ref]. Dysregulation of placental function was suggested as a clue to the underlying pathology.
# Conclusion
In our population-based study we were generally unable to link maternally perceived DFM to underlying placental pathology, although some associations were seen for subgroups. Maternal ability to identify FGR was low. In order to enhance the role of FM counting, further research must focus on ways to help women to identify fetal compromise from chronic placental pathology.
## Supporting information
Textbox S1 Information about decreased fetal movement (DFM) provided to the mothers. (DOC)
[fig] Figure 1: Flowchart for data collection. doi:10.1371/journal.pone.0039259.g001 [/fig]
[fig] 1: Includes the cord anomalies; true umbilical cord knots (n = 7), velamentous (n = 5) and marginal cord (n = 9) insertion. doi:0.37/journal.pone.0039259.t005 [/fig]
[table] Table 2 Table 3: Maternal and fetal characteristics and birth outcome for DFM versus non-DFM pregnancies, from pregnancy week 24 0 (n = 320). Placental pathology in DFM pregnancies versus non-DFM pregnancies (n = 320). [/table]
[table] Table 4: Characteristics of consultations for DFM from 129 pregnancies. [/table]
[table] Table 5: Placental pathology by birth outcome (n = 320). MATERNO-PLACENTAL CIRCULATORY DISORDERS, abrupt circulatory insults CORD ANOMALIES WITH POTENTIAL IMPACT 1 "p-values refer to odds ratio for categorical data and t-test for continuous variables for comparisons between pregnancies with or without placental pathology.Includes all cases with pathology with assumed moderate to important clinical impact from macroscopic or microscopic examination.Neonatal complications: preterm birth, SGA, infections, Apgar scores ,7 5min or transfer to NCU for conditions relevant to fetal growth restriction or fetal distress (respiratory syndrome or cerebral irritation). [/table]
|
Trace Minerals Supplementation with Great Impact on Beef Cattle Immunity and Health
# Introduction
Beef cattle operations from cow/calf farms to feedlot systems achieve their production goals by ensuring the health of the animals under their care. Of the health issues that can interfere with reaching these goals, bovine respiratory disease (BRD) is the most prevalent. Multiple factors, such as suboptimal nutrition, stress related to weaning, transportation, commingling, environmental extremes, and high stocking density contribute to the occurrence of BRD [bib_ref] Bovine respiratory disease in feedlot cattle: Environmental, genetic, and economic factors, Snowder [/bib_ref]. Therefore, an effective beef cattle health program should integrate approaches designed to offer adequate nutrient balance (quality and quantity), minimize stress, reduce the level of exposure to pathogens (through adequate biosecurity and biocontainment), generate protective immunity through colostrum and vaccination, identify and isolate sick animals, and identify the associated pathogens.
Balanced nutrition that includes adequate supply of protein, energy, vitamins and minerals is one of the pillars of achieving optimal health in beef cattle operations. Trace minerals (TM) play critical roles in the immune system's response to potential pathogens, which will be the focus of this review. As a result, TM are important for the cattle health and performance [bib_ref] Interaction of cattle health/immunity and nutrition, Galyean [/bib_ref]. Although these elements are needed in small quantities, they are fundamental for many biological processes. Examples of these processes include energy production, signal transduction, nucleic acid replication, transcription and translation, and antioxidant protection against cellular damage. Because TM have specific functions in the animal's innate and acquired immune responses [bib_ref] Trace Mineral Deficiency in Cattle: A Review, Smart [/bib_ref] , adequate TM supplementation is crucial for the development of protective immunity, especially in stressed animals. Four TM, namely Se, Cu, Zn, and Mn, exert specific positive effects on the immune response to different pathogens. In addition, Cr is known to be beneficial for the health of cattle during periods of stress by reducing cortisol levels and modulating insulin and glucose concentrations. As a result, these elements are currently included in commercial TM formulations.
Numerous factors influence the TM status of cattle, including the mineral content of soils, forages, and feedstuffs, cattle age and production stage, TM requirements and dry matter intake (DMI), presence of antagonists in feed, water, and forages and the type of supplement formulations. Supplementation with TM can be achieved using either oral (e.g., powder formulations, salt blocks, drench and rumen boluses) or injectable routes of administration. It is important to recognize the differences in the bioavailability of the TM between these administration routes. Clinical and scientific evidence supports TM supplementation in beef cattle herds and growing steers. Numerous studies have shown the effects of TM on the cattle immune response as well as their health and performance . However, many of those studies differ substantially in their experimental designs, protocols, TM formulations, baseline diets, and outcomes. Consequently, there is considerable variation in their conclusions. In addition, the majority of the studies did not objectively assess the effects of TM on the cell-mediated immune response to specific pathogens affecting cattle (e.g., antigen recall leukocyte proliferation or proportion of lymphocyte subsets after challenge). The present manuscript reviews the research-based evidence of the effects of supplementation with Se, Cu, Zn, Mn and Cr on the immune response and health of beef cattle. This review highlights the effects of strategic supplementation based on injectable TM during critical times of cattle management (e.g., vaccination) in conjunction with oral free-choice supplements, on the humoral and cell-mediated immune responses and protection against respiratory disease pathogens in beef cattle, which is a relatively new area of investigation and the author's main research focus.
## Beef cattle health program: basic principles
The overall health of cattle depends on an appropriately functioning immune system, its interaction with the environment (e.g., climate, nutrition, management, cattle density), exposure to potential pathogens, and their virulence. In most instances, the level of exposure to pathogens is determined by hygienic conditions and biosecurity. Health is achieved when the animal's level of immunity exceeds the disease challenge. Consequently, the goal of all beef cattle health programs is to simultaneously increase the level of immunity while reducing pathogen exposure [bib_ref] Herd health management, Chenoweth [/bib_ref]. Several factors can increase a herd's susceptibility to disease, including poor hygiene and biosecurity, low immuno-competence due to lack or ineffective vaccination, stress, inadequate management, nutritional deficiencies, high population density, and failure in passive transfer, among others [bib_ref] Herd health management, Chenoweth [/bib_ref]. Therefore, the health of the cattle can be improved through preventative management strategies focused on optimizing nutrition, enhancing colostrum quality, minimizing stress, improving biosecurity, and safe and efficacious vaccination procedures. Appropriate nutrient supply, including supplementation with key minerals (e.g., Cu, Se, Mn, and Zn) is a key component in cattle health programs as it provides the foundation for other health management strategies (e.g., vaccination).
Adequate maternal antibody transfer [e.g., Immunoglobulin G (IgG) serum concentration >1600 mg/dL] is the first line of defense for the newborn calf [bib_ref] Neonatal immune development in the calf and its impact on vaccine response, Chase [/bib_ref]. Inadequate passive transfer occurs in 10-25% of newborn beef calves [bib_ref] A guide to colostrum management in beef cows and calves, Perino [/bib_ref]. Satisfactory antibody passive transfer is achieved when newborn calves absorb enough immunoglobulins (200-400 g) after suckling high quality colostrum (>50 g IgG/L) from their dams within the first six hours of life [bib_ref] A guide to colostrum management in beef cows and calves, Perino [/bib_ref]. Thus, it is vital to increase the colostrum immunoglobulin concentration through effective vaccination (e.g., BRD, Clostridium complex, and pinkeye) of female cattle, especially in the last third of pregnancy. Not only does vaccination increase the concentration of antibodies, it also increases the number of immunologically relevant cells (lymphocytes) and concentration of interferons in colostrum. It also is important for cows to maintain an adequate body condition score (6-7, on a scale from 1 to 9, with 1 being severely emaciated and 9 extremely obese) and mineral status before calving to enhance IgG concentration. In beef cattle, marginal TM deficiencies negatively affect antibody production after vaccination, as well as colostrum quality and quantity, and subsequent passive antibody transfer to the calves [bib_ref] Injection of Zn, Se, Cu and Mn enhances antibody response to Bovine..., Palomares [/bib_ref].
Vaccination of calves before or after weaning is a major tool to prevent the many diseases affecting beef cattle, such as BRD. However, vaccines are not 100% efficacious, as the subsequent level of protection depends on many factors. Low response to BRD vaccination may occur as the result of several factors, including nutritional deficiencies, poor management, high level of stress, concurrent diseases, young age at vaccination, high levels of maternal antibodies, and genetics [bib_ref] Keys to Effective Vaccination of Calves, Woolums [/bib_ref]. Mineral deficiencies have been identified as an important risk factor contributing to the development of BRD in nursing calves. Additionally, many bovine practitioners are concerned that marginal deficiencies in Se, Cu, and Zn are responsible for the failure of some animals to respond appropriately to vaccination against BRD.
It is critical for cattle to develop sufficient immunity before encountering stressful situations or being exposed to pathogens. For example, management practices (e.g., castration, weaning, transportation, and commingling) can affect their ability to produce an adequate immune response after vaccination [bib_ref] Stress, immunity, and the management of calves, Hulbert [/bib_ref]. During stressful periods, animals undergo an increased metabolic demand that results in oxidative stress. This involves an imbalance between the production and neutralization of pro-oxidant substances known as reactive oxygen species (ROS). These free radicals cause significant cell damage by peroxidation of lipid cell membranes [bib_ref] Role of antioxidants and trace elements in health and immunity of transition..., Spears [/bib_ref] to which leukocytes are particularly sensitive. This impairs the immune function and increases the animal's susceptibility to disease. For this reason, stressed cattle are more prone to BRD shortly after arrival at the feed yard than preconditioned cattle, which experience lower levels of stress [bib_ref] Board-invited review: Recent advances in management of highly stressed, newly received feedlot..., Duff [/bib_ref].
Importantly, Cu, Zn, Se and Mn are structural components of the enzymes that neutralize ROS, and strategic supplementation with TM can minimize the negative effects of stress on the immune system. Supplementation with TM has been suggested as a husbandry strategy to improve the immune response to vaccination (especially against respiratory viruses) and the overall health status of cattle. Therefore, it should be considered as an essential component of beef cattle health programs.
## Cattle immunity
The immune system has two main components: innate and adaptive immunity. The innate immunity is the first line of protection against pathogens. It includes physical barriers (dermal and mucosal epithelia), antimicrobial substances produced by epithelial cells, the complement proteins, and phagocytic cells (e.g., neutrophils, macrophages and dendritic cells) and their cytokines and chemokines. Furthermore, the innate immunity includes the type I interferon (IFN)-induced antiviral state, gamma delta T cells (γδ T), mucosa-associated invariant T cells, non-phagocytic granulocytes (e.g., eosinophils and basophils) and natural killer cells [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref] [fig_ref] Figure 1: The bovine immune system [/fig_ref]. Neutrophils identify infectious agents by recognition of pathogen-associated molecular patterns using Toll-like receptors [bib_ref] Innate Immunology of Bovine Respiratory Disease, Ackermann [/bib_ref] [bib_ref] Functional Polymorphisms in Toll-Like Receptor Genes for Innate Immunity in Farm Animals, Novak [/bib_ref]. After pathogen recognition, phagocytic cells activate and initiate phagocytosis and killing of the engulfed microorganisms. Importantly, the innate immune system is not pathogenspecific and responds to different antigens in a similar manner. Moreover, this component of the immune system does not produce a memory or anamnestic response.
## Cattle immunity
The immune system has two main components: innate and adaptive immunity. The innate immunity is the first line of protection against pathogens. It includes physical barriers (dermal and mucosal epithelia), antimicrobial substances produced by epithelial cells, the complement proteins, and phagocytic cells (e.g., neutrophils, macrophages and dendritic cells) and their cytokines and chemokines. Furthermore, the innate immunity includes the type I interferon (IFN)-induced antiviral state, gamma delta T cells (γδ T), mucosa-associated invariant T cells, non-phagocytic granulocytes (e.g., eosinophils and basophils) and natural killer cells [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref] [fig_ref] Figure 1: The bovine immune system [/fig_ref]. Neutrophils identify infectious agents by recognition of pathogen-associated molecular patterns using Toll-like receptors [bib_ref] Innate Immunology of Bovine Respiratory Disease, Ackermann [/bib_ref] [bib_ref] Functional Polymorphisms in Toll-Like Receptor Genes for Innate Immunity in Farm Animals, Novak [/bib_ref]. After pathogen recognition, phagocytic cells activate and initiate phagocytosis and killing of the engulfed microorganisms. Importantly, the innate immune system is not pathogenspecific and responds to different antigens in a similar manner. Moreover, this component of the immune system does not produce a memory or anamnestic response. The adaptive immune response is initiated after vaccination or natural exposure to foreign agents. This response consists of humoral (antibody production by B cells) and cell-mediated immunity (handled by T lymphocytes and their cytokines) [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref]. This part of the immune system is pathogen-specific, requiring recognition of the pathogen, activation of T and B lymphocytes, differentiation into effector and memory cells, and repair of damaged tissues [fig_ref] Figure 1: The bovine immune system [/fig_ref]. This branch of the immune system creates memory that subsequently responds to repeated exposure to the same pathogen.
B lymphocytes mature in the bone marrow and are released into blood where they circulate and populate lymphoid tissues. Plasma cells, which are differentiated effector B cells, function as the main component of the humoral immune response, producing specific antibodies after vaccination or antigen exposure [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref]. Antibodies are Y-shaped proteins having different isotopes (e.g., IgG, IgM, IgD, IgA and IgE) and specific roles, including neutralization, complement activation, opsonization and phagocytosis, and antibody- The adaptive immune response is initiated after vaccination or natural exposure to foreign agents. This response consists of humoral (antibody production by B cells) and cellmediated immunity (handled by T lymphocytes and their cytokines) [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref]. This part of the immune system is pathogen-specific, requiring recognition of the pathogen, activation of T and B lymphocytes, differentiation into effector and memory cells, and repair of damaged tissues [fig_ref] Figure 1: The bovine immune system [/fig_ref]. This branch of the immune system creates memory that subsequently responds to repeated exposure to the same pathogen.
B lymphocytes mature in the bone marrow and are released into blood where they circulate and populate lymphoid tissues. Plasma cells, which are differentiated effector B cells, function as the main component of the humoral immune response, producing specific antibodies after vaccination or antigen exposure [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref]. Antibodies are Y-shaped proteins having different isotopes (e.g., IgG, IgM, IgD, IgA and IgE) and specific roles, including neutralization, complement activation, opsonization and phagocytosis, and antibody-dependent cellular cytotoxicity [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref] [bib_ref] The Unusual Genetics and Biochemistry of Bovine Immunoglobulins, Stanfield [/bib_ref]. B cells also recognize foreign antigens and present them to T lymphocytes, which further stimulate greater antibody production by plasma cells. In addition, B cells located in the mucosa secrete IgA antibodies as part of the humoral response. The IgA antibodies help neutralize antigens and prevent attachment of pathogens to the epithelial cells, thus preventing the pathogens from colonizing the main routes of entrance and underlying tissues.
T lymphocytes are produced in the bone marrow and mature in the thymus. They are categorized as helper, cytotoxic, and gamma-delta T cells. T helper cells are characterized by the CD4 protein that collects information from antigen presenting cells expressing the major histocompatibility complex class II (MHC-II) [bib_ref] Dissection of the Clonal Composition of Bovine Alphabeta T Cell Responses Using..., Connelley [/bib_ref] [bib_ref] Towards a Systems Understanding of MHC Class I and MHC Class II..., Neefjes [/bib_ref]. One of the important roles of T helper cells is to release cytokines (e.g., IFN-γ). These proteins stimulate a variety of immune cells, including neutrophils, macrophages, additional T helper cells, cytotoxic T cells and B lymphocytes as part of the development of the adaptive immune response [bib_ref] Cellular responses to interferon-gamma, Boehm [/bib_ref]. For instance, production of antibodies by plasma cells involves the secretion of interleukins-2 (IL-2) IL-4 and IFN-γ by CD4 + T cells.
In contrast, cytotoxic T cells express CD8 through which they recognize and destroy virus-infected cells via MHC-I [bib_ref] Contribution of in vivo and ex vivo studies to understanding the role..., Howard [/bib_ref]. Gamma-delta T cells, which express the surface protein WC1, comprise the most abundant subset of T cells in cattle and have the ability to activate independently of the MHC proteins. They are present in high numbers in epithelial sites, haemal lymph nodes, and the thymus [bib_ref] Bovine γδ T cells are a major regulatory T cell subset, Guzman [/bib_ref]. These cells play wide-ranging roles in both innate and adaptive immune responses, including immune regulation, cytokine production, and local defenses at the mucosal epithelium, [bib_ref] Diverse immunological roles of γδ T cells, Chen [/bib_ref].
## Role of trace minerals in bovine immunity
The main roles of Zn, Cu, Se, Mn and Cr in the bovine immune system are summarized in [fig_ref] Figure 2: The main roles of trace minerals [/fig_ref].
Animals 2022, 12, x FOR PEER REVIEW 5 of 21 dependent cellular cytotoxicity [bib_ref] Bovine Immunology: Implications for Dairy Cattle, Vlasova [/bib_ref] [bib_ref] The Unusual Genetics and Biochemistry of Bovine Immunoglobulins, Stanfield [/bib_ref]. B cells also recognize foreign antigens and present them to T lymphocytes, which further stimulate greater antibody production by plasma cells. In addition, B cells located in the mucosa secrete IgA antibodies as part of the humoral response. The IgA antibodies help neutralize antigens and prevent attachment of pathogens to the epithelial cells, thus preventing the pathogens from colonizing the main routes of entrance and underlying tissues. T lymphocytes are produced in the bone marrow and mature in the thymus. They are categorized as helper, cytotoxic, and gamma-delta T cells. T helper cells are characterized by the CD4 protein that collects information from antigen presenting cells expressing the major histocompatibility complex class II (MHC-II) [bib_ref] Dissection of the Clonal Composition of Bovine Alphabeta T Cell Responses Using..., Connelley [/bib_ref] [bib_ref] Towards a Systems Understanding of MHC Class I and MHC Class II..., Neefjes [/bib_ref]. One of the important roles of T helper cells is to release cytokines (e.g., IFN-γ). These proteins stimulate a variety of immune cells, including neutrophils, macrophages, additional T helper cells, cytotoxic T cells and B lymphocytes as part of the development of the adaptive immune response [bib_ref] Cellular responses to interferon-gamma, Boehm [/bib_ref]. For instance, production of antibodies by plasma cells involves the secretion of interleukins-2 (IL-2) IL-4 and IFN-γ by CD4 + T cells.
In contrast, cytotoxic T cells express CD8 through which they recognize and destroy virus-infected cells via MHC-I [bib_ref] Contribution of in vivo and ex vivo studies to understanding the role..., Howard [/bib_ref]. Gamma-delta T cells, which express the surface protein WC1, comprise the most abundant subset of T cells in cattle and have the ability to activate independently of the MHC proteins. They are present in high numbers in epithelial sites, haemal lymph nodes, and the thymus [bib_ref] Bovine γδ T cells are a major regulatory T cell subset, Guzman [/bib_ref]. These cells play wide-ranging roles in both innate and adaptive immune responses, including immune regulation, cytokine production, and local defenses at the mucosal epithelium, [bib_ref] Diverse immunological roles of γδ T cells, Chen [/bib_ref].
## Role of trace minerals in bovine immunity
The main roles of Zn, Cu, Se, Mn and Cr in the bovine immune system are summarized in [fig_ref] Figure 2: The main roles of trace minerals [/fig_ref]. and Se is a component of the enzyme glutathione peroxidase (GPx). These enzymes protect leukocytes against membrane damage by free radicals. These TM play important functions mainly in epithelial integrity, leukocyte migration, phagocytosis and bacteria killing, cytokine production, antibody secretion and cell-mediated immunity (CMI). Se is a component of the enzyme glutathione peroxidase (GPx). These enzymes protect leukocytes against membrane damage by free radicals. These TM play important functions mainly in epithelial integrity, leukocyte migration, phagocytosis and bacteria killing, cytokine production, antibody secretion and cell-mediated immunity (CMI).
## Zinc
Zinc is an important structural and functional element of several enzymes and proteins required in different metabolic pathways [bib_ref] Zinc Signals and Immunity, Maywald [/bib_ref]. Zinc plays an essential role in tissues that undergo mitosis (e.g., lymphoid tissue, skin, testicular seminiferous tubules, and mucous membranes). In doing so, Zn contributes to the strength of the mucosal lining of the respiratory and digestive systems. It also is involved in mucosal tissue repair and epithelial integrity. Its ability to affect proliferation of epithelial cells is due to its role in nucleic acid replication (e.g., DP1-DP2 zinc fingers, ribonucleotide reductases), transcription (e.g., RNA polymerases, aminoacyl-tRNA synthetases and transcription factors), and protein synthesis via mechanistic target of rapamycin kinase, mTOR [bib_ref] Zinc Signals and Immunity, Maywald [/bib_ref]. Zinc is essential for clonal expansion during lymphocyte proliferation and differentiation before effector functions are executed [bib_ref] Micronutrients and immune function in cattle, Spears [/bib_ref] [bib_ref] Trace elements and host defense: Recent advances and continuing challenges, Failla [/bib_ref] [bib_ref] Dietary antioxidants: Immunity and host defense, Puertollano [/bib_ref]. In addition, its role in mRNA expression and subsequent translation into proteins makes it critical for cytokine production by macrophages and lymphocytes during both innate and acquired immunity [bib_ref] Zinc finger protein 64 promotes Toll-like receptor-triggered proinflammatory and type I interferon..., Wang [/bib_ref].
Zinc is a powerful antioxidant as a structural component of superoxide dismutase (Zn-SOD), a cytosolic and mitochondrial intermembrane enzyme involved in conversion of superoxide radicals (O 2 − ) into hydrogen peroxide (H 2 O 2 ) [bib_ref] Zinc and its role in immunity and inflammation, Bonaventura [/bib_ref]. Zinc also plays an important role in the synthesis of nitric oxide by macrophages [bib_ref] Cellular and Molecular Immunology, Abbas [/bib_ref] , which is an essential molecule for bacteria killing.
As might be anticipated, Zn deficiency impairs both branches of the immune response. In the innate immunity, Zn deficiency negatively affects neutrophil and macrophage migration, phagocytosis and killing of microbes, as well as dermal and mucosal integrity. In the adaptive immunity, Zn deficiency affects thymic function as well as activation, proliferation, differentiation and cytokine secretion by lymphocytes [bib_ref] Zinc and the immune system, Rink [/bib_ref]. Zinc status has been associated with the risk of neonatal diarrhea in calves [bib_ref] Effects of copper, zinc and selenium status on performance and health in..., Enjalbert [/bib_ref]. Further, serum Zn concentrations have been negatively correlated with abortion rates [bib_ref] Serum zinc and copper concentrations in relation to spontaneous abortion in cows:..., Graham [/bib_ref].
## Copper
Copper is required for various enzymes and proteins involved in mitochondrial energy production and other metabolic pathways [bib_ref] Trace elements and host defense: Recent advances and continuing challenges, Failla [/bib_ref]. Its primary role in immunity is through its involvement in the enzyme Cu-SOD, a metalloprotein responsible for production of hydrogen peroxide radicals by neutrophils [bib_ref] Selected vitamins and trace elements support immune function by strengthening epithelial barriers..., Maggini [/bib_ref]. These free radicals are used to generate powerful oxidizing agents such as hydroxyl radicals by the myeloperoxidase system [bib_ref] Zinc Signals and Immunity, Maywald [/bib_ref] [bib_ref] Dietary antioxidants: Immunity and host defense, Puertollano [/bib_ref]. Cu-SOD is responsible for maintaining a balance in the production of free radicals needed to kill bacteria via the oxidative burst activity. Furthermore, this enzyme is important in mitigating oxidative stress as it scavenges superoxide radicals. In doing so, Cu-SOD prevents damage to leukocyte membranes by avoiding excessive accumulation of superoxide free radicals. Copper also is a component of ceruloplasmin, an acute phase protein with a prominent role in bovine immunity [bib_ref] Copper and immunity, Percival [/bib_ref] , especially in stressed animals. In this regard, concentrations of ceruloplasmin have been used as a biomarker to evaluate Cu status and the effects of TM supplementation in cattle [bib_ref] The effect of molybdenum-induced copper deficiency on acute-phase protein concentrations, superoxide dismutase..., Arthington [/bib_ref].
Copper deficiency may occur as a consequence of low Cu dietary levels or consumption of molybdenum-rich pastures [bib_ref] The effect of molybdenum-induced copper deficiency on acute-phase protein concentrations, superoxide dismutase..., Arthington [/bib_ref]. Similarly, excessive sulfur in cattle diets (e.g., high sulfate water or feeding ethanol coproducts such as distillers grains) negatively affects Cu absorption. In the rumen, S can interact with molybdate to form thiomolybdate complexes which have very high affinity for Cu, negatively affecting its absorption and bioavailability, resulting in Cu deficiency [bib_ref] The Interactions between Copper, Molybdenum, and Sulphur in ruminant nutrition, Suttle [/bib_ref]. Copper deficiency is associated with impaired neutrophil phagocytic function [bib_ref] Copper and immunity, Percival [/bib_ref] and affects the acute phase protein response. However, Arthington et al. [bib_ref] Effect of copper depletion and repletion on lymphocyte blastogenesis and neutrophil bactericidal..., Arthington [/bib_ref] showed that Cu deficiency in crossbred beef heifers did not affect in vitro or in vivo neutrophil chemotaxis after injection of an inflammation-inducing compound. In another study, Arthington et al. [bib_ref] The effect of molybdenum-induced copper deficiency on acute-phase protein concentrations, superoxide dismutase..., Arthington [/bib_ref] observed that calves with molybdenuminduced Cu deficiency did not elevate plasma ceruloplasmin concentrations after BHV1 challenge, but had a significant increase of haptoglobin and fibrinogen concentrations compared with calves with normal Cu status. Overall, Cu deficiency has been associated with poor calf health and performance (mainly diarrhea, growth retardation and mortality) [bib_ref] Effects of copper, zinc and selenium status on performance and health in..., Enjalbert [/bib_ref] , likely through reduced innate immunity, failure of passive transfer, decreased antibody production and impaired cell-mediated immunity [bib_ref] The immune system as a physiological indicator of marginal copper status?, Bonham [/bib_ref]. Severe Cu deficiency in cattle may result in sudden death due to heart failure [bib_ref] Biochemical and physiological indicators of mineral status in animals: Copper, cobalt and..., Mills [/bib_ref].
## Selenium
Selenium is a structural component of the enzyme glutathione peroxidase (GPx) which catalyzes the reduction of hydrogen peroxide to water and oxygen. This enzyme is important to providing nicotinamide adenine dinucleotide phosphate (NADPH) for the respiratory burst along with glutathione reductase, and to maintaining the bactericidal capacity of neutrophils [bib_ref] Increased neutrophil adherence and adhesion molecule mRNA expression in endothelial cells during..., Maddox [/bib_ref] [bib_ref] Recent developments in trace element metabolism and function: Trace elements, disease resistance..., Suttle [/bib_ref]. In addition, this enzyme inactivates ROS production, preventing subsequent leukocyte damage [bib_ref] Increased neutrophil adherence and adhesion molecule mRNA expression in endothelial cells during..., Maddox [/bib_ref].
Selenium plays a major role in neutrophil migration into tissues and subsequent inflammation [bib_ref] Increased neutrophil adherence and adhesion molecule mRNA expression in endothelial cells during..., Maddox [/bib_ref]. Selenium deficiencies have been associated with decreased neutrophil migration and killing activity [bib_ref] Effects of selenium and copper deficiency on neutrophil function in cattle, Boyne [/bib_ref] , reduced GPx antioxidant functions, and impaired B-cell function and antibody production. Se-deficient calves had reduced antibody production 14 d after BHV1 challenge [bib_ref] Effect of dietary selenium on the primary and secondary immune response in..., Reffett [/bib_ref]. However, in that study, Se deficiency did not affect DMI or rectal temperatures [bib_ref] Effect of dietary selenium on the primary and secondary immune response in..., Reffett [/bib_ref].
Selenium deficiency causes multiple health problems. Se-deficient calves may have muscular degeneration known as White muscle disease. This condition may result in immediate death due to myocardial degeneration. Marginally Se-deficient calves may display weakness and low vigor. A retrospective study in France indicated that deficient and marginal Se status of cows was associated with increased risk of infectious abortion, high calf morbidity and mortality due to poor immune response [bib_ref] Effects of copper, zinc and selenium status on performance and health in..., Enjalbert [/bib_ref] , which may be related to altered iodine metabolism [bib_ref] Thyroid hormone concentrations in selenium deficient and selenium sufficient cattle, Arthur [/bib_ref]. Further, Se and vitamin E deficiency in cows have been associated with increased incidence of retained fetal membranes [bib_ref] Effects of copper, zinc and selenium status on performance and health in..., Enjalbert [/bib_ref].
## Manganese
Manganese is a component of several enzymes (e.g., arginase, pyruvate carboxylase, and SOD). It also functions as an enzyme activator for various hydrolases, kinases, and transferases. Evidence of the roles of Mn in the immune response is limited. Manganese has an essential function in scavenging ROS produced by phagocytic cells since it is a structural component of the Mn-SOD enzyme [bib_ref] Role of Trace Minerals in the Immune System, Tomlinson [/bib_ref]. Dietary manganese deficiency has relatively mild effects on immunity, with the exception of impaired humoral immune response [bib_ref] The effect of microelements-Copper, manganese and cobalt-On the antibody forming function of..., Bazhora [/bib_ref]. Manganese deficiency (plasma Mn < 76 nmol/L) has been associated with abortion, long bone congenital deformities (including enlarged joints, stiffness, twisted legs, shorter bones, as well as general physical weakness) in newborn calves [bib_ref] Manganese deficiency in the etiology of deformed calves, Dyer [/bib_ref] [bib_ref] Assessment of the role of manganese in congenital joint laxity and dwarfism..., Hidiroglou [/bib_ref] and tongue neuromuscular disorder in cows [bib_ref] Tongue play and manganese deficiency in dairy cattle, Karatzias [/bib_ref].
## Chromium
Chromium is essential for the immune response during stress, which is particularly important in feedlot cattle. Chromium is required to promote the action of insulin and insulin-like growth factor I in tissues such as muscle and liver. Thus, it is implicated in different metabolic pathways (e.g., glucose tolerance and clearance rate) related to growth and homeostasis [bib_ref] Supplemental chromium: Its benefits to the bovine immune system, Burton [/bib_ref]. Moreover, Cr appears to decrease circulating cortisol concentrations during highly stressful situations [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref]. Chromium affects several pathways of the immune system. The effects of Cr on lymphocytes show a biphasic pattern with a stimulatory effect at low concentrations but an inhibitory effect at high concentrations [bib_ref] Cadmium, nickel, chromium and lead accumulate in human lymphocytes and interfere with..., Borella [/bib_ref]. Chromium reduced the levels of the immunomodulatory cytokines IL-2, IFN-γ, and TNF-α following in vitro mitogen stimulation in lactating Holstein cows [bib_ref] Effects of supplemental chromium on production of cytokines by mitogen-stimulated bovine peripheral..., Burton [/bib_ref]. In addition, Cr stimulated lymphocyte proliferation after mitogen stimulation [bib_ref] Effects of supplemental chromium on immune responses of periparturient and early lactation..., Burton [/bib_ref] , cell-mediated immunity after injection of phyto-hemagglutinin in young calves [bib_ref] Immune response and disease resistance of calves fed chromium nicotinic acid complex..., Kegley [/bib_ref] , as well as antibody production following immunization with infectious bovine rhinotracheitis virus [bib_ref] Effects of supplemental chromium on immune responses of periparturient and early lactation..., Burton [/bib_ref] and tetanus toxoid in cattle [bib_ref] Chromium supplementation enhances antibody response to vaccination with tetanus toxoid in cattle, Faldyna [/bib_ref].
## Trace mineral status in beef cattle
In general, the TM status results from a balance between dietary TM intake and requirements. Moreover, the TM status is affected by the intestinal absorption, chemical interactions, and mineral antagonisms in the gastrointestinal tract. Liver samples to determine hepatic TM concentrations represent the gold standard and the best choice for determining the TM status in cattle. Liver is the organ where most TM of interest in cattle are stored, incorporated into enzymes and then released when needed. Blood TM concentrations may be variable, inconsistent and have low predictive value of the real TM status. Assessment of the TM status of beef herds is used to determine whether a nutrient deficiency exists (marginal or clinically evident), to assess the prevalence of a deficiency, or to estimate the endogenous reserves of TM. Copper levels in liver samples should be above 25-50 ppm to be considered normal (range: 50-600 µg/g; on a dry matter (DM) basis, [fig_ref] Table 1: Reference values of trace mineral requirements, maximum tolerable concentrations and hepatic concentrations... [/fig_ref] [bib_ref] The use of blood analysis to evaluate trace mineral status in ruminant..., Herdt [/bib_ref]. Liver levels of Se of 0.8-1.0 ppm (range: 0.7-2.5 µg/g; [bib_ref] The use of blood analysis to evaluate trace mineral status in ruminant..., Herdt [/bib_ref] are considered to be adequate, and levels below 0.2 ppm are considered to be critically deficient. Liver Mn concentrations below 9 ppm (range: 5-15 µg/g [bib_ref] The use of blood analysis to evaluate trace mineral status in ruminant..., Herdt [/bib_ref] can be considered an indication of marginal deficiency. Hepatic Zn concentrations under 80 ppm can be considered marginal or deficient(range: 90-400 µg/g [bib_ref] The use of blood analysis to evaluate trace mineral status in ruminant..., Herdt [/bib_ref] , [fig_ref] Table 1: Reference values of trace mineral requirements, maximum tolerable concentrations and hepatic concentrations... [/fig_ref]. Nevertheless, liver tissue is not an ideal sample for accurate determination of Zn status, probably due to its particular bioavailability, dynamic chemical interactions and extensive use throughout the body. Muscle and bone tissues represent the main Zn pools. However, sampling these tissues is not a practical procedure to be performed in the field. Several factors affect the TM status of cattle, including mineral content of soils, forages and feedstuffs, high variability of mineral intake, presence of mineral antagonists in water, forages, and feedstuffs, and variable mineral requirements. The requirements are the guidelines to determine the animal's needs and formulate TM supplements. Trace minerals requirements derive from refereed research summarized by the National Academies of Sciences, Engineering, and Medicine (NASEM), Board on Agriculture Subcommittee on Beef Cattle Nutrition [fig_ref] Table 1: Reference values of trace mineral requirements, maximum tolerable concentrations and hepatic concentrations... [/fig_ref]. Trace mineral requirements vary and may be affected by several factors including cattle biological type, breed, physiological state, health and nutritional status, management practice, and environmental conditions, among others.
Concentration of TM in the soil is variable across the USA, several regions containing deficient levels of Se, Zn, Co, and Cu in forages. Therefore, TM supplementation is highly recommended in those areas where forages do not supply the mineral requirements [bib_ref] Trace mineral feeding and assessment, Swecker [/bib_ref]. A study performed by Kansas State University and USDA on 352 forage samples from cow/calf operations in 18 USA states revealed deficiency of important TM in a high proportion of samples. Only 2.5% of the forage samples contained adequate zinc levels. Copper and cobalt levels were adequate in 36% and 34.1% of the samples, respectively. In addition, 49.7% of the samples were considered marginal for Cu, and 14.2% of the samples were classified as critically deficient. Another report of 23 USA states showed similar results where only 4%, 15% and 26% of fescue samples from Midwest USA had adequate levels of Se, Zn and Cu, respectively. Moreover, only 18, 24 and 24% of native forage samples were adequate in Cu, Se and Zn, respectively. In both surveys, high percentage of the samples had very high (8-13%) or marginally high (17-48%) levels of iron, molybdenum and sulfur, which results in Cu deficiency due to antagonistic effects limiting Cu absorption through the gastrointestinal tract. The TM status of cattle is highly dependent on mineral intake by the animal, which is affected by stress, the animal's physiologic status, the feeding protocol and the type of diet. For instance, DMI decreases after arrival at the feedyard, where a high proportion of calves do not eat sufficiently during the first 48-72 h. In general, stressed calves reduce their feed intake to an average of 1.5% of body weight during the first two weeks after arrival [bib_ref] Effects of season, health, and management on feed intake by beef cattle, Galyean [/bib_ref]. This deprives the immune system of the nutrients needed to generate an effective response. Short-term reduced intake and transient deficiency of Mn, Cu, Zn, and Se will probably not induce an immediate clinical deficiency response in cases where prior intake was adequate. However, a chronic insufficient TM intake for several weeks may lead to depletion of the body stores, resulting in altered metabolism, impaired immunity, and poor health status and production performance. The rate of tissue (e.g., liver) depletion for different TM in cattle (and the switch to clinical deficiency) is not completely understood. Deficiency of TM progressively affects different physiologic and metabolic functions. A marginal deficiency in TM is not visually evident, and borderline levels may continue to decrease over time, affecting the immune response, which makes the animals more susceptible to common respiratory pathogens. A subclinical TM deficiency eventually may become evident at the herd level, with increased morbidity and reduced growth and reproductive performance [bib_ref] Differential diagnosis of micronutrient-responsive disorders in beef suckler herds, Suttle [/bib_ref]. A clinical TM deficiency occurs when TM levels are so low that they result in specific clinical signs of disease (e.g., stiffness and muscle weakness due to muscle degeneration known as White muscle disease, caused by severe Se deficiency). It is highly recommended to always evaluate the herd's mineral profile, the amount of TM that the animals may be consuming, and therefore determine the TM formulation the animals may need.
## Oral trace minerals supplementation, effects on beef cattle health
Oral TM supplementation is a basic management practice in beef cattle operations. There are different sources of trace minerals to supplement cattle depending on their organic or inorganic composition. Organic sources (e.g., chelates, amino acid or polysaccharide complexes, propionates, proteinates and yeast derivatives) are more biologically available to the animal. Thus, organic sources may be beneficial at periods of high stress and low feed intake [bib_ref] Effects of feeding organic trace minerals on milk production and reproductive performance..., Rabiee [/bib_ref] and have been reported to offer benefits on health and performance compared with the inorganic forms [bib_ref] Effects of organic or inorganic cobalt, copper, manganese, and zinc supplementation to..., Marques [/bib_ref] [bib_ref] Supplemental trace minerals (zinc, copper, manganese, and cobalt) as Availa-4 or inorganic..., Kegley [/bib_ref]. Inorganic forms of TM (e.g., sulfates, oxides, carbonates, and chlorides) are structured with molecules of sulfur, oxygen, and chloride, among others. The most bioavailable inorganic source of TM is generally the sulfate forms, followed by the chloride forms of Zn, Cu, and Mn. In contrast, the oxide inorganic form is the least bioavailable and should be avoided for beef cattle feeding programs. Oral mineral supplements can be provided as free-choice mixture of mineralized salt, blocks or in combination with energy and/or protein supplements in the total mixed ration.
The effects of oral TM supplementation on health and disease resistance have been assessed by providing single or combined TM compounds. Studies using TM combinations provide limited information about the specific effects of TM on the immune response and health. Thus, this section of the manuscript focuses on the effects of oral supplementation with individual TM on cattle health and immunity. During the last decades, numerous studies have demonstrated the benefits of oral TM supplementation on beef cattle immunity, health and performance . However, other studies have shown marginal or insignificant effects of oral TM supplementation on beef cattle metabolism, growth, immunity and health [bib_ref] Effect of parenteral selenium and vitamin E on performance, health, and humoral..., Droke [/bib_ref] [bib_ref] Effect of vitamin E and selenium supplementation on some immune parameters following..., Nemec [/bib_ref] [bib_ref] Effects of zinc source and level and added copper lysine in the..., Galyean [/bib_ref] [bib_ref] Zinc oxide and amino acids as sources of dietary zinc for calves:..., Kincaid [/bib_ref] [bib_ref] The effects of pre-weaning vitamin E and selenium supplementation on the performance,..., Wright [/bib_ref] [bib_ref] Effects of supplemental zinc concentration and source on performance, carcass characteristics, and..., Malcolm-Callis [/bib_ref] [bib_ref] Effects of inorganic and organic copper supplemented to first-calf cows on cow..., Muehlenbein [/bib_ref] [bib_ref] Performance of steers supplemented with copper before and during receiving at the..., Beck [/bib_ref] [bib_ref] Effects of copper supplementation on the copper status of peripartum beef cows..., Enjalbert [/bib_ref] [bib_ref] Effect of Se on selenoprotein activity and thyroid hormone metabolism in beef..., Rowntree [/bib_ref] [bib_ref] Effects of copper and zinc source on performance and humoral immune response..., Salyer [/bib_ref] [bib_ref] Growth, reproductive performance, and manganese status of heifers fed varying concentrations of..., Hansen [/bib_ref] [bib_ref] Effects of source of supplemental zinc on performance and humoral immunity in..., Nunnery [/bib_ref] [bib_ref] Parenteral selenium and vitamin E supplementation of weaned beef calves, Swecker [/bib_ref] [bib_ref] Effect of moderate Cu supplementation on serum metabolites, enzymes and redox state..., Castillo [/bib_ref] [bib_ref] Supplemental trace minerals (zinc, copper, and manganese) as sulfates, organic amino acid..., Ryan [/bib_ref] [bib_ref] Supplementation of organic and inorganic selenium to late gestation and early lactation..., Muegge [/bib_ref] [bib_ref] Effects of feeding pregnant beef cows selenium-enriched alfalfa hay on selenium status..., Wallace [/bib_ref]. Many of these studies have provided inconsistent results due to differences in mineral forms supplemented (organic versus inorganic), high variability in TM concentration in control diets, the animals' TM requirements and initial TM status, variation in animal intake and absorption, chemical mineral interactions and antagonisms, level of stress, and differences in experimental models and endpoints. These factors have made interpretation and comparison of data from different studies extremely difficult. Further, in some of these studies, the reported effects of TM on the immune function and protection may not be clinically relevant or influence the cattle health status. Thus, this review will mainly discuss studies assessing the effects of oral supplementation with individual TM on important immunological or clinical outcomes for beef cattle production systems.
## Oral zinc supplementation
According to the NASEM, the recommended Zn dietary level is 30 mg/kg DM [fig_ref] Table 1: Reference values of trace mineral requirements, maximum tolerable concentrations and hepatic concentrations... [/fig_ref]. Multiple studies have demonstrated the effects of Zn supplementation on the humoral and cell-mediated immune responses and health [bib_ref] The effect of zinc methionine in a mineral mixture on gain and..., Brazle [/bib_ref] [bib_ref] Serum IgG and IgM responses to sheep red blood cells (SRBC) in..., Prasad [/bib_ref] [bib_ref] Effects of zinc propionate supplementation on growth performance, skeletal muscle fiber, and..., Wellmann [/bib_ref] [bib_ref] Immune response and performance of heifers supplemented with zinc from an organic..., Kegley [/bib_ref] [bib_ref] Dietary zinc and manganese sources administered from the fetal stage onwards affect..., Chirase [/bib_ref] [bib_ref] Effects of zinc source and level and added copper lysine in the..., Galyean [/bib_ref]. Supplementation of steers with Zn-methionine resulted in greater DMI and improved health status after inoculation of BHV1 compared with non-supplemented steers or animals receiving supplementation with ZnO [bib_ref] Dietary zinc and manganese sources administered from the fetal stage onwards affect..., Chirase [/bib_ref] [bib_ref] Feed intake, rectal temperature, and serum mineral concentrations of feedlot cattle fed..., Chirase [/bib_ref] , Zinc proteinate (71 mg/kg DM) or Zn-SO 4 (67 mg/kg DM, [bib_ref] Effect of supplemental trace mineral complexes on rectal temperature, feed intake and..., Blezinger [/bib_ref]. In addition, steers supplemented with Zn-methionine or ZnO (25 mg of supplemental Zn/kg DM) had greater BHV1-specific antibody titers compared with steers receiving a control diet (26 mg of Zn/kg DM) [bib_ref] Effects of zinc methionine and zinc oxide on performance, blood characteristics, and..., Spears [/bib_ref]. Moreover, increasing the amount of supplemented Zn-methionine (from 35 to 70 mg/kg DM) in beef calves receiving a 65% concentrate basal diet after weaning and transportation significantly reduced the incidence of BRD (23.6 and 9.7% for low and high Zn-methionine, respectively) during a 28-day receiving period and a 21-day step-up period [bib_ref] Effects of zinc source and level and added copper lysine in the..., Galyean [/bib_ref].
In a subsequent report, supplementation of grazing Angus crossbred heifers with ZnSO 4 (360 mg of Zn/d), or Zn amino acid complex (360 mg of Zn/d) after transportation tended to enhance cell-mediated immune response to phyto-hemagglutinin injection compared with non-supplemented heifers [bib_ref] Immune response and performance of heifers supplemented with zinc from an organic..., Kegley [/bib_ref]. Studies showing the effects of feeding Zn-deficient-diets (e.g., 17 mg/kg DM) or abrupt Zn depletion in beef cattle revealed a significant reduction in skin cell-mediated immune response to phyto-hemagglutinin [bib_ref] The effect of feeding organic and inorganic zinc on biochemical parameters in..., Engle [/bib_ref] [bib_ref] Effects of breed, dietary phosphorus, and trace mineral source on immune function,..., Engle [/bib_ref]. Moreover, Zn supplementation has been demonstrated to enhance hoof quality scores [bib_ref] Effect of organic zinc sources on performance, zinc status, and carcass, meat,..., Kessler [/bib_ref] , while Zn deficiency has been previously associated with lameness and hoof deformation [bib_ref] Effects of supplementation of slightly deficient dairy cows with zinc oxide or..., Corbellini [/bib_ref]. Specifically, supplementation of grazing calves and finishing steers with Znmethionine decreased the incidence of foot rot compared with non-supplemented animals [bib_ref] The effect of zinc methionine in a mineral mixture on gain and..., Brazle [/bib_ref] , reflecting the importance of Zn for maintenance of the healthy tegumentary tissues. In contrast, other studies have not shown positive effects of different sources of oral Zn supplementation (methionine, sulfate, and propionate) on beef cattle immunity [bib_ref] Zinc oxide and amino acids as sources of dietary zinc for calves:..., Kincaid [/bib_ref] [bib_ref] Effects of source of supplemental zinc on performance and humoral immunity in..., Nunnery [/bib_ref] , health and performance [bib_ref] Effects of copper and zinc source on performance and humoral immune response..., Salyer [/bib_ref] [bib_ref] Effects of source of supplemental zinc on performance and humoral immunity in..., Nunnery [/bib_ref].
## Oral copper supplementation
According to NASEM, finishing steers and beef cows require 10 mg Cu/kg DM. When forage samples contain less than 8 ppm of copper, they are considered deficient. The effects of oral Cu supplementation on animal health and immunity depends on the animal's physiologic conditions, immune competence and mineral status. Copper supplementation of beef cows resulted in improved health status and increased growth performance of their calves [bib_ref] Diagnosis of copper deficiency and effects of supplementation in beef cows, Naylor [/bib_ref]. In addition, Cu supplementation of weaned calves enhanced their oxidative status [bib_ref] Effects of supplementation with two sources and two levels of copper on..., Correa [/bib_ref] , immune response, protection and overall health [bib_ref] Effects of zinc source and level and added copper lysine in the..., Galyean [/bib_ref] [bib_ref] Effects of copper deficiency and copper deficiency coupled with high dietary iron..., Gengelbach [/bib_ref].
Supplementation of weaned beef calves with Cu-lysine (16 mg of Cu/kg DM) resulted in increased plasma Cu concentration, enhanced phagocytic and oxidative burst activities of monocytes compared with control calves fed a basal diet of corn silage/soybean meal with no supplemental Cu (7 mg of Cu/kg DM) [bib_ref] Effect of copper supplementation and vaccination on cellular immune response in growing..., Saker [/bib_ref]. In addition, Galyean et al. [bib_ref] Effects of zinc source and level and added copper lysine in the..., Galyean [/bib_ref] reported that Cu-lysine supplementation (5 mg of added Cu/kg) of beef steers was associated with a reduction (not statistically significant) in BRD morbidity (13.9%) compared with the control animals (20.1% morbidity) receiving a baseline diet that provided 3.25 mg of Cu/kg from CuO. However, Cu-lysine-supplemented animals had lower average daily gain (ADG) and DMI. Dorton et al. [bib_ref] Effects of copper source and concentration on copper status and immune function..., Dorton [/bib_ref] observed that Cu supplementation of Angus steers (10 or 20 mg Cu/kg DM) tended to increase the antibody response compared with animals receiving a control diet (7 mg Cu/kg DM).
Molybdenum, sulfur and iron are Cu antagonists. High levels of Mo (>1-3 ppm) in forages and feedstuffs (or when the Cu:Mo ratio is <3:1) may negatively impact Cu status, resulting in Cu deficiency-like effects. Calves fed high levels of Mo (5 mg of Mo/kg DM of diet) showed reduced Cu-SOD activity, ceruloplasmin concentration and TNFα production by monocytes after BHV1 inoculation compared with Cu-supplemented calves (10 mg Cu/kg DM), which had adequate Cu status [bib_ref] Effects of copper deficiency and copper deficiency coupled with high dietary iron..., Gengelbach [/bib_ref]. In the same study, the calves supplemented with Cu and inoculated with BHV1 and Pasteurella two days after weaning had greater body temperature and TNF-α concentration after challenge, which was attributed to improved immune response and greater cytokine production [bib_ref] Effects of copper deficiency and copper deficiency coupled with high dietary iron..., Gengelbach [/bib_ref]. Poor neutrophil phagocytic activity was observed in Cu-deficient cattle (induced by feeding Fe or Mo) compared with cattle supplemented with adequate amounts of Cu in the diet [bib_ref] Effects of molybdenum or iron induced copper deficiency on the viability and..., Boyne [/bib_ref].
Other studies failed to demonstrate positive effects of Cu supplementation on beef cattle health, or had inconsistent results [bib_ref] Micronutrients and immune function in cattle, Spears [/bib_ref] [bib_ref] Effects of inorganic and organic copper supplemented to first-calf cows on cow..., Muehlenbein [/bib_ref] [bib_ref] Performance of steers supplemented with copper before and during receiving at the..., Beck [/bib_ref] [bib_ref] Effects of copper and zinc source on performance and humoral immune response..., Salyer [/bib_ref] [bib_ref] Effect of moderate Cu supplementation on serum metabolites, enzymes and redox state..., Castillo [/bib_ref] [bib_ref] Diagnosis of copper deficiency and effects of supplementation in beef cows, Naylor [/bib_ref]. Stressed calves supplemented with Cu (5 mg/kg DM for 133 d) had increased antibody production (to pig red blood cells) and cell-mediated immune response (to dinitrochlorobenzene stimulation) [bib_ref] The effects of low-copper diets with or without supplemental molybdenum on specific..., Ward [/bib_ref]. However, these effects were not observed in unstressed calves [bib_ref] The effects of low-copper diets with or without supplemental molybdenum on specific..., Ward [/bib_ref]. Copper source (Polysaccharide complex of copper SQM ® Cu or CuSO 4 ) did not affect growth performance and BRD morbidity in newly received heifers, but heifers receiving CuSO 4 had greater antigen-specific antibody titers than did heifers on the SQM ® Cu treatment on days 14 and 21 [bib_ref] Effects of copper and zinc source on performance and humoral immune response..., Salyer [/bib_ref]. In steers with adequate Cu status, supplementation with Cu (5 mg/kg DM as Cu sulfate or Cu lysine) did not affect antibody production, cell-mediated immunity or growth performance [bib_ref] Effect of copper level and source (copper lysine vs. copper sulfate) on..., Ward [/bib_ref]. Similarly, Cu supplementation before and after calving (either inorganic, organic, or organic with extra Zn) of 2-year-old crossbred beef cows having Cu hepatic borderline levels (~50 mg/kg) did not improve pregnancy rates, maternal antibody transfer or calf health when compared with non-supplemented cows [bib_ref] Effects of inorganic and organic copper supplemented to first-calf cows on cow..., Muehlenbein [/bib_ref].
## Oral selenium supplementation
According to NASEM, finishing steers and beef cows require 0.10 mg Se/kg DM. In general, oral Se supplementation has been documented to enhance both humoral and cell-mediated immune responses [bib_ref] Selected vitamins and trace elements support immune function by strengthening epithelial barriers..., Maggini [/bib_ref]. However, similar to other reported mineral assays, assessing the effects of oral Se supplementation on cattle health has been challenging, and provided inconsistent conclusions.
Beef cows supplemented with 120 mg of Se/kg of salt-mineral mix had greater colostrum IgG concentration than Se-deficient cows. In addition, calves born to Sesupplemented cows had higher post-suckle serum concentrations of IgG than calves born to control cows [bib_ref] Effect of selenium supplementation on colostral IgG concentration in cows grazing selenium-deficient..., Swecker [/bib_ref]. In another report, feeding Se-biofortified alfalfa hay to pregnant cows resulted in increased colostrum Se and IgG1 concentrations but did not affect short-term serum antibody concentration in their calves [bib_ref] Effects of feeding pregnant beef cows selenium-enriched alfalfa hay on selenium status..., Wallace [/bib_ref]. Hall et al. [bib_ref] Effect of feeding selenium-fertilized alfalfa hay on performance of weaned beef calves, Hall [/bib_ref] [bib_ref] Effects of feeding selenium-enriched alfalfa hay on immunity and health of weaned..., Hall [/bib_ref] reported that weaned calves that had been fed Se-enriched alfalfa hay (0.95, 1.55, or 3.26 mg Se/kg DM) that was harvested from sodium-selenite-fertilized lands .0, or 89.9 g Se/ha, respectively) during the backgrounding phase had considerably lower BRD mortality and greater growth performance in the feedlot stage in a dose-dependent manner compared with control calves receiving unfertilized alfalfa hay (0.07 mg Se/kg DM). Moreover, calves supplemented with high Se-yeast before weaning had enhanced in vitro macrophage phagocytosis on day 22 post-weaning compared with unsupplemented control calves [bib_ref] Case study: Effects of feeding supplemental organic or inorganic selenium to cow-calf..., Beck [/bib_ref]. Similarly, another report showed higher natural killer cell cytotoxicity in bulls that received selenium yeast and vitamin E compared with control bulls [bib_ref] Selenium and vitamin E enriched diet increases NK cell cytotoxicity in cattle, Latorre [/bib_ref]. The positive effects of Se supplementation may be associated with an enhancement of the antioxidant capacity and cholesterol metabolism [bib_ref] Effect of different selenium sources and concentrations on glutathione peroxidase activity and..., Da Silva [/bib_ref] , upregulation of the mitochondrial gene expression capacity, and downregulation of oxidative stress-related genes, as previously demonstrated [bib_ref] Dietary supplementation of selenium in inorganic and organic forms differentially and commonly..., Liao [/bib_ref]. More recent studies showed that feeding Se-biofortified alfalfa hay to weaned beef calves prior to enter-ing the feedlot enhanced nasopharyngeal microbiome diversity, which may explain in part the benefits of high-Se diets for prevention of BRD in beef calves [bib_ref] Weaned beef calves fed selenium-biofortified alfalfa hay have an enriched nasal microbiota..., Hall [/bib_ref] [bib_ref] Feeding seleniumbiofortified alfalfa hay during the preconditioning period improves growth, carcass weight,..., Hall [/bib_ref].
Nevertheless, other trials have shown no effects of Se supplementation on beef cattle immune and health parameters [bib_ref] Effect of parenteral selenium and vitamin E on performance, health, and humoral..., Droke [/bib_ref] [bib_ref] The effects of pre-weaning vitamin E and selenium supplementation on the performance,..., Wright [/bib_ref] [bib_ref] Selenium effects on glutathione peroxidase and the immune response of stressed calves..., Reffett-Stabel [/bib_ref]. Supplementation with Se and vitamin E in preweaning Hereford/Angus calves with normal mineral and vitamin status did not affect the pre-or post-weaning weight gain or the immune response to BVDV and BHV1 following vaccination [bib_ref] The effects of pre-weaning vitamin E and selenium supplementation on the performance,..., Wright [/bib_ref]. Studies aimed to evaluate the effects of Se and Vitamin E supplementation on the immune response, health and performance of feedlot steers showed an increase in antibody production against P. haemolytica, but did not demonstrate positive effects improving health status [bib_ref] Effect of parenteral selenium and vitamin E on performance, health, and humoral..., Droke [/bib_ref]. Intriguingly, Reffett-Stabel et al. [bib_ref] Selenium effects on glutathione peroxidase and the immune response of stressed calves..., Reffett-Stabel [/bib_ref] reported that weaned calves that had been nursed by cows receiving Se-adequate diet had lower antibody titers after challenge with P. haemolytica than calves nursed by cows receiving Se-deficient diet.
Despite Se being the most deficient TM in forage-based dietsand Se supplementation providing multiple known benefits, caution should be taken when supplementing beef cattle with Se to ensure the provision of adequate amounts without reaching toxic levels [bib_ref] Management of mineral supplementation programs for cow-calf operations, Olson [/bib_ref].
## Oral manganese supplementation
According to NASEM, finishing steers and beef cows require 20 and 40 mg Mn /kg DM, respectively. There is limited evidence of the effects of individual Mn supplementation on the immunity and health of beef cattle. A study was performed to evaluate the effects of different dietary Mn levels (10, 30, or 50 mg of supplemental Mn/kg DM from MnSO) on growth and reproduction of Angus and Simmental heifers. Heifers were fed a diet containing cottonseed hulls, corn gluten feed, citrus pulp, and ground corn, and the control diet contained 15.8 mg of Mn/kg DM. Mn supplementation did not affect reproductive performance, indicating that 15.8 mg of Mn/kg DM is adequate for growth and fertility of heifers [bib_ref] Growth, reproductive performance, and manganese status of heifers fed varying concentrations of..., Hansen [/bib_ref].
## Oral chromium supplementation
The Cr requirements for beef cattle have not been established. The positive effects of oral Cr supplementation on health and performance of beef steers have been previously demonstrated [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref] [bib_ref] Chromium supplementation alters the performance and health of feedlot cattle during the..., Bernhard [/bib_ref]. Improvement in growth performance and lower BRD morbidity have been reported in stressed newly-received cattle supplemented with Cr [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref] [bib_ref] Chromium supplementation alters the performance and health of feedlot cattle during the..., Bernhard [/bib_ref]. In the initial study using calves supplemented with either 0 (controls) or 0.4 mg/kg/day of Cr (from high-Cr yeast) for 28 days revealed that Cr supplementation significantly increased ADG (30%) and feed efficiency (27%), suggesting a possible Cr deficiency in the experimental calves [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref]. The antibody levels were greater for the Cr-supplemented calves, with no effect on morbidity [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref]. Another experiment during a 70-day growth period using the control calves from the previous study did not show significant effects of Cr supplementation on ADG or feed efficiency, but it was associated with reduction of blood glucose and cortisol, indicating that Cr might positively influence the immune system of stressed animals. These two studies suggested that timing of Cr supplementation in relation to stress and pathogen exposure should be considered [bib_ref] Supplemental chromium for stressed and growing feeder calves, Chang [/bib_ref].
A subsequent study showed that lower level of Cr supplementation (0.2 mg/kg) was able to improve ADG and reduced morbidity in newly-received stressed beef calves compared with calves receiving higher Cr supplementation (0.5 and 1 mg/kg) and no supplementation [bib_ref] Effect of level of supplemental chromium on performance, serum constituents, and immune..., Moonsie-Shageer [/bib_ref]. This study confirmed the results of the previous reports on the effects of Cr supplementation on growth performance, immuno-competence and reduction of cortisol levels [bib_ref] Effect of level of supplemental chromium on performance, serum constituents, and immune..., Moonsie-Shageer [/bib_ref]. A successive trial reported significant reduction of morbidity when supplementing calves with an amino acid chelated Cr (15%) compared to high-Cr yeast (33%) and non-supplemented control (55%) [bib_ref] Chelated chromium for stressed feeder calves. Can, Mowat [/bib_ref]. Further research testing the effects of Cr supplementation on the immune response elicited by a MLV vaccine containing BHV1 and PI 3 V in stressed calves showed that Cr supplementation had significant effects on BHV1 specific antibody response. The results showed that 75% of the BHV1 seropositive calves were from the Cr-supplemented group, whereas only 25% were from the control group [bib_ref] Chromium supplementation enhances antibody response to vaccination with tetanus toxoid in cattle, Faldyna [/bib_ref]. Further, Cr supplementation has been documented to also affect the cellmediated immunity of beef cattle. A greater local cell-mediated immune response after intradermal injection of phyto-hemagglutinin was observed in Angus and Angus crossbred steers receiving ad libitum supplementation with high-Cr yeast than control steers or those supplemented with CrCl3 or Cr nicotinic acid. In addition, animals supplemented with Cr-nicotinic acid complex in corn silage-based diets had a greater blastogenic response to the same antigen than steers supplemented with CrCl3 [bib_ref] Immune response, glucose metabolism, and performance of stressed feeder calves fed inorganic..., Kegley [/bib_ref].
## Single-use, pulse-dose administration of trace minerals
Oral free-choice TM supplementation may lead to mineral deficiencies due to high variability in mineral composition and intake, mineral binding to undigested feed particles, reduced absorption in the digestive tract, and antagonisms. Strategic extra supplementation using single-administration, pulse-dose trace mineral products (oral or injected) is becoming very popular on beef farms to prevent TM deficiencies or improve the TM status. This provides the required TM amounts during critical points of the production cycle, when the animals have greater TM demands (e.g., before transportation or at vaccination). However, this strategic supplementation is not intended to substitute the regular oral mineral supply.
Oral single pulse TM supplementation (e.g., drench, paste or bolus) is being used in some beef cattle farms. However, there is not enough evidence of the effects of these products on the animals' TM status [bib_ref] Comparison of multiple single-use, pulse-dose trace mineral products provided as injectable, oral..., Jackson [/bib_ref]. A study performed by Dr. Hansen's team at Iowa State University demonstrated that oral single, pulse-dose trace mineral supplements (paste or drench) did not efficiently increase blood or liver trace mineral concentrations in Angus-cross steers with adequate TM status [bib_ref] Comparison of multiple single-use, pulse-dose trace mineral products provided as injectable, oral..., Jackson [/bib_ref]. Administration of injectable trace minerals (ITM) resulted in faster and greater increase in Zn, Se and Mn concentrations (through 24 h) compared with other single-dose supplementation treatments (drench, paste and boluses) [bib_ref] Comparison of multiple single-use, pulse-dose trace mineral products provided as injectable, oral..., Jackson [/bib_ref]. In addition, hepatic Se levels were also higher (through day 29) in the steers treated with ITM compared to the other treatment groups. However, ruminal boluses (Reloader 250™ bolus) generated greater long-term hepatic Se concentration than ITM (on day 91) and the other products (on day 120). The choice of the single-dose supplementation method depends on whether fast effects (e.g., ITM concurrent with vaccination) or a longterm, sustained-release TM supplementation (e.g., Se rumen bolus) are needed to support a particular management strategy [bib_ref] Comparison of multiple single-use, pulse-dose trace mineral products provided as injectable, oral..., Jackson [/bib_ref].
## Injectable trace minerals on beef cattle immunity and health
The use of ITM formulations (containing Se, Cu, Zn and Mn) has several advantages for beef production systems. It provides a known and well-controlled quantity of TM that is rapidly and efficiently absorbed and stored after injection [bib_ref] Mineral concentrations of plasma and liver after injection with a trace mineral..., Pogge [/bib_ref]. This is important for cattle having reduced DMI and mineral consumption (e.g., at the time of weaning and vaccination or during periods of transportation and receiving) or limited access to free-choice mineral supplements (e.g., extensive rangeland systems, flooded pastures, etc). Therefore, administration of ITM decreases the variability in TM status observed in animals with free choice mineral intake [bib_ref] Effects of trace mineral source and feeding method on the productivity of..., Arthington [/bib_ref]. A commercially available ITM product containing Zn (60 mg/mL), Cu (15 mg/mL), Mn (10 mg/mL) and Se (5 mg/mL; Multimin ® -90. Multimin ® , Axiota ® Animal Health) has been proven to be absorbed quickly, increasing blood TM levels within 8 to 10 h after injection and followed by storage in liver at approximately 24 h post injection [bib_ref] Mineral concentrations of plasma and liver after injection with a trace mineral..., Pogge [/bib_ref] [bib_ref] A multielement trace mineral injection improves liver copper and selenium concentrations and..., Genther [/bib_ref]. Administration of this compound results in an increase in enzymatic (e.g., GPx and SOD) activity by 24 h following administration, with a maximum enzymatic function by 15 days post injection. These effects may persist for approximately 90 days [bib_ref] Mineral concentrations of plasma and liver after injection with a trace mineral..., Pogge [/bib_ref] [bib_ref] A multielement trace mineral injection improves liver copper and selenium concentrations and..., Genther [/bib_ref] [bib_ref] The effect of injectable trace minerals (selenium, copper, zinc, and manganese) on..., Machado [/bib_ref].
Studies performed by the author's research team (GRAVID™) at the University of Georgia [bib_ref] Injection of Zn, Se, Cu and Mn enhances antibody response to Bovine..., Palomares [/bib_ref] [bib_ref] Effects of injectable trace minerals on humoral and cell-mediated immune responses to..., Palomares [/bib_ref] [bib_ref] Effects of injectable trace minerals administered concurrently with a modified live virus..., Bittar [/bib_ref] [bib_ref] Effects of injectable trace minerals on the immune response to Mannheimia haemolytica..., Bittar [/bib_ref] [bib_ref] Immune response and onset of protection from Bovine viral diarrhea virus 2..., Bittar [/bib_ref] [bib_ref] T cell populations in calves infected with BVDV + IBR after intranasal..., Hoyos-Jaramillo [/bib_ref] and others [bib_ref] Effect of injectable trace minerals on the humoral immune response to multivalent..., Arthington [/bib_ref] [bib_ref] Effect of supplemental trace minerals from injection on health and performance of..., Richeson [/bib_ref] [bib_ref] Effect of injectable trace mineral administration on health, performance, and vaccine response..., Roberts [/bib_ref] have demonstrated the effects of the administration of ITM (containing Cu, Se, Zn, and Mn) on the protective immunity elicited by vaccination against BRD and neonatal calf diarrhea. Arthington and Havenga [bib_ref] Effect of supplemental trace minerals from injection on health and performance of..., Richeson [/bib_ref] reported that administration of ITM concurrently with BRD MLV vaccination increased and/or mitigated the decrease of trace mineral levels (Cu, Se, and Zn) after vaccination of beef calves. In that study, treatment with ITM enhanced the production of BHV1-specific serum neutralizing antibodies on days 14, 30, and 60 post-vaccination. Similarly, administration of ITM concurrently with a BRD MLV vaccine + M. haemolytica and P. multocida bacterin in weaned Holstein bull calves with normal TM status resulted in increased hepatic Se, Cu and Mn concentrations. In addition, ITM calves had greater antibody titers (26.7% more calves having seroconversion) to BVDV1 on day 28 after primary vaccination compared with the control group [bib_ref] Effects of injectable trace minerals on humoral and cell-mediated immune responses to..., Palomares [/bib_ref]. In the same study, calves treated with ITM showed an earlier and stronger mononuclear leukocyte proliferation to BVDV1 and BRSV following vaccination compared with the control group. Moreover, ITM administration also enhanced antibody production and leukocyte proliferation against M. haemolytica and P. multocida, respectively [bib_ref] Effects of injectable trace minerals on the immune response to Mannheimia haemolytica..., Bittar [/bib_ref]. The observed decrease in TM concentration in plasma [bib_ref] Effect of supplemental trace minerals from injection on health and performance of..., Richeson [/bib_ref] and liver [bib_ref] Effects of injectable trace minerals on humoral and cell-mediated immune responses to..., Palomares [/bib_ref] after vaccination in the untreated control calves is believed to be associated with TM utilization during the development of the immune response. Therefore, strategic supplementation with ITM at the time of vaccination may be a critical control point in regular beef cattle vaccination programs. Studies performed by Bittar et al. [bib_ref] Immune response and onset of protection from Bovine viral diarrhea virus 2..., Bittar [/bib_ref] and Roberts et al. [bib_ref] Effect of injectable trace mineral administration on health, performance, and vaccine response..., Roberts [/bib_ref] using newly-received beef calves demonstrated that treatment with ITM concomitantly with MLV parenteral vaccination induced faster and stronger BVDV-specific antibody response [bib_ref] Immune response and onset of protection from Bovine viral diarrhea virus 2..., Bittar [/bib_ref] [bib_ref] Effect of injectable trace mineral administration on health, performance, and vaccine response..., Roberts [/bib_ref]. A rapid and robust antibody response against BRD pathogens might be of significant value by conferring earlier protection following vaccination in newly-received, highly stressed cattle, which are at high risk of suffering from BRD. Our study indicated that ITM administration mitigated the decrease in CD4 + and CD8 + T cells associated with BVDV2 acute infection in calves that were BVDV2-challenged five days after arrival and immunization [bib_ref] Immune response and onset of protection from Bovine viral diarrhea virus 2..., Bittar [/bib_ref]. These results support our previous findings of greater leukocyte proliferation in ITM-treated calves [bib_ref] Effects of injectable trace minerals on humoral and cell-mediated immune responses to..., Palomares [/bib_ref]. Further studies in our lab also demonstrated that ITM supplementation at the time of intranasal (IN) vaccination of dairy bull calves controlled the CD8 + T cell reduction after BVDV2 + BHV1 challenge [bib_ref] T cell populations in calves infected with BVDV + IBR after intranasal..., Hoyos-Jaramillo [/bib_ref]. In addition, calves treated with ITM at the time of IN vaccination had reduced inflammation and damage of the upper respiratory tract mucosa after BVDV2 + BHV1 challenge compared to vaccinated, saline-treated calves.
In another challenge trial, the calves treated with ITM at the time of vaccination had increased ADG (during the first 14 days) following a BVDV2 challenge given five months after vaccination. Further, ITM-treated calves had greater platelet counts compared with the control calves [bib_ref] Effects of injectable trace minerals administered concurrently with a modified live virus..., Bittar [/bib_ref]. A more recent study examined the effects of a different injectable solution containing Cu and Zn on weight gain, hematological parameters, and immune response of pre-weaning beef calves [bib_ref] Effect of Injectable Copper and Zinc Supplementation on Weight, Hematological Parameters, and..., Mattioli [/bib_ref]. Supplementation with ITM improved ADG and the antibody response to inactivated BHV1 vaccination [bib_ref] Effect of Injectable Copper and Zinc Supplementation on Weight, Hematological Parameters, and..., Mattioli [/bib_ref].
The effects of ITM on growth performance and feed efficiency of cattle have been variable. Arthington et al. [bib_ref] Effects of trace mineral injections on measures of performance and trace mineral..., Arthington [/bib_ref] performed three experiments to assess the effects of ITM (Multimin ® -90) on health and performance of pre-and post-weaned Brangus-crossbred beef calves. Administration of ITM improved Cu, Se, and Zn status and acute phase proteins, but did not improve growth performance of pre-weaned and post-weaned calves. However, at the subsequent developmental stage (peri-pubertal period), the heifers treated with ITM had enhanced ADG, antibody titers to porcine red blood cells, and liver Se concentrations compared with control heifers [bib_ref] Effects of trace mineral injections on measures of performance and trace mineral..., Arthington [/bib_ref]. Genther and Hansen [bib_ref] A multielement trace mineral injection improves liver copper and selenium concentrations and..., Genther [/bib_ref] evaluated the growth performance and meat quality of beef cattle treated with ITM. Supplementation with ITM improved growth of mildly TM-deficient steers and improved rib eye area and marbling score regardless of initial TM status.
The use of another ITM formulation containing 10 mg/mL Cu, 10 mg/mL Mn, 5 mg/mL Se, 40 mg/mL Zn (Adaptador Min, Biogénesis Bago, Buenos Aires, Argentina) administered in conjunction with vitamins A, and E decreased body weight loss in the steers submitted to 26-day preconditioning from weaning to transport [bib_ref] Zn) and vitamins (A, E) on plasma metabolite and hormone concentrations and..., Galarza [/bib_ref]. In contrast, other studies to evaluate the efficacy of ITM at the time of weaning in Nellore calves showed no effect of mineral supplementation on growth performance, even though it caused an increase in plasma concentrations of antioxidant enzymes (SOD and GPx) [bib_ref] Effect of a trace mineral injection at weaning on growth, antioxidant enzymes..., Vedovatto [/bib_ref].
The effects of ITM supplementation on Bovine coronavirus antibody response and passive transfer after vaccination against neonatal calf diarrhea in beef cattle have been recently evaluated [bib_ref] Injection of Zn, Se, Cu and Mn enhances antibody response to Bovine..., Palomares [/bib_ref]. Administration of ITM improved Se and Cu status and enhanced colostrum Bovine coronavirus antibody response to vaccination in beef cows with marginal Se, Cu and Zn deficiencies. In that study, calves that had nursed ITM-treated cows had enhanced Bovine coronavirus antibody passive transfer [bib_ref] Injection of Zn, Se, Cu and Mn enhances antibody response to Bovine..., Palomares [/bib_ref].
The research-based evidence indicates that strategic administration of ITM containing Se, Cu, Zn and Mn in calves receiving free-choice oral TM supplementation improves the oxidative status and the immune response and protection elicited by IN or parenteral (attenuated or inactivated) vaccination against bovine respiratory pathogens. In addition, there is limited evidence that under certain conditions (depending on TM status, developmental state, weaning management, among others) ITM supplementation may positively affect cattle growth performance.
# Conclusions
Trace mineral supplementation is an essential component of beef cattle health programs. Micro-minerals such as Se, Cu, Zn, Mn and Cr play significant roles in numerous metabolic pathways, homeostatic functions, oxidative balance and immune response. Therefore, cattle TM status has a direct impact on health and performance. Adequate TM supplementation is crucial to maintaining an optimal TM status, especially during episodes of stress and oxidative imbalance (e.g., during transportation) or higher metabolic and immune demands (e.g., at vaccination). Oral TM supplementation is a common and highly recommended practice for beef cattle operations, as numerous reports demonstrate its benefits on immunity, health and performance. However, due to the high variation in research experimental conditions and TM formulations, interpretation and comparison of data on the effects of oral TM supplementation on animal health is difficult. In addition, clinical and experimental evidence indicates that there is an enormous source of variation in TM bioavailability in animals receiving oral TM. It is highly recommended to evaluate the herd's mineral profile and the amount of TM that the animals may be consuming to determine the TM supplementation regime. Research studies and the author's field experience support the practice of strategic ITM supplementation (without replacing traditional oral mineral supplementation) to enhance TM and oxidative status, overall cattle health and the immune response prior to challenging episodes of stress or following vaccination. This may result in decreased cattle morbidity and costs of treating sick animals, which would positively impact the productivity of beef cattle systems.
Funding: This review article did not receive external funding. However, some of the research studies performed by the author and described in this review were financially supported by The University of Georgia and Multimin ® USA, Inc. (Grant # 056545-01).
Data Availability Statement: Not applicable.
[fig] Figure 1: The bovine immune system. The innate immunity mainly involves epithelial barriers, phagocytic cells and their chemokines and cytokines, Natural killer (NK) cells, Mucosa-associated invariant T cells (MAIT), Type I interferon and Complement proteins. The adaptive immunity consists of antibody production by Plasma cells (differentiated B lymphocytes) and cell-mediated immunity handled by T lymphocytes. Gamma delta T cells are the most abundant T lymphocytes in cattle and play important roles in both innate and adaptive immune responses. [/fig]
[fig] Figure 2: The main roles of trace minerals (TM) in the bovine immune system. Copper, Mn and Zn are structural and functional components of the antioxidant enzyme superoxide dismutase (SOD) [/fig]
[table] Table 1: Reference values of trace mineral requirements, maximum tolerable concentrations and hepatic concentrations in adults and growing calves. † Requirements Range mg/kg DM † Maximum Tolerable Nutrient Requirements of Beef Cattle, 8th ed. National Academy Press: Washington, DC, USA, 2016[82]. * Values are expressed on a dry tissue basis. Reference ranges of hepatic TM concentrations as reported by Herdt and Hoff[81]. DM: dry matter. [/table]
|
Hypothalamic structural integrity and temporal complexity of cortical information processing at rest in migraine without aura patients between attacks
The hypothalamus has been attributed an important role during the premonitory phase of a migraine attack. Less is known about the role played by the hypothalamus in the interictal period and its relationship with the putative neurocognitive networks previously identified in the pathophysiology of migraine. Our aim was to test whether the hypothalamic microstructure would be altered during the interictal period and whether this co-existed with aberrant connectivity at cortical level. We collected multimodal MRI data from 20 untreated patients with migraine without aura between attacks (MO) and 20 healthy controls (HC) and studied fractional anisotropy, mean (MD), radial (RD), and axial diffusivity of the hypothalamus ROI as a whole from diffusion tensor imaging (DTI). Moreover, we performed an exploratory analysis of the same DTI metrics separately for the anterior and posterior hypothalamic ROIs bilaterally. From resting-state functional MRI, we estimated the Higuchi's fractal dimension (FD), an index of temporal complexity sensible to describe non-periodic patterns characterizing BOLD signature. Finally, we correlated neuroimaging findings with migraine clinical features. In comparison to HC, MO had significantly higher MD, AD, and RD values within the hypothalamus. These findings were confirmed also in the exploratory analysis on the sub-regions of the hypothalamus bilaterally, with the addition of lower FA values on the posterior ROIs. Patients showed higher FD values within the salience network (SN) and the cerebellum, and lower FD values within the primary visual (PV) network compared to HC. We found a positive correlation between cerebellar and SN FD values and severity of migraine. Our findings of hypothalamic abnormalities between migraine attacks may form part of the neuroanatomical substrate that predisposes the onset of the prodromal phase and, therefore, the initiation of an attack. The peculiar fractal dimensionality we found in PV, SN, and cerebellum may be interpreted as an expression of abnormal efficiency demand of brain networks devoted to the integration of sensory, emotional, and cognitive information related to the severity of migraine.OPENIn recent years, studies of the pathophysiology of migraine have seen a renewed interest in the role played by the hypothalamus in the recurrence of migraine attacks. This brain structure seems to play a predominant role during the phase immediately preceding an attack when some patients experience premonitory symptoms 1-3 . During the attack, the hypothalamus is metabolically hyperactive 4 , poorly connected to the spinal trigeminal nucleus 2 , and more strongly coupled with the dorsal pons 5 . The hypothalamus also appears to be involved in the state of a never-ending migraine attack that is chronic migraine 6,7 .Furthermore, the interictal phase is characterized by abnormal connectivity of cortical networks and this has been shown to correlate with clinical variables such as frequency of attacks and severity of migraine pain 8 . We recently reported that abnormal functional connectivity between some of these networks and the hypothalamus is peculiar to patients suffering from chronic migraine 7 . Moulton and colleagues 9 using an ROI-to-ROI approach, found stronger functional connectivity between the hypothalamus and several brain regions involved in the regulation of autonomic functions in interictal migraine patients but little is known at whole-brain level.Here, we sought whether there is abnormal integrity of the hypothalamic structures and of functional connectivity of cortical networks in a group of migraine patients without aura between attacks. Therefore, we measure the hypothalamic microstructure through diffusion tensor imaging (DTI), a useful sensitive method to detect white matter tracts in grey matter nuclei-as is the case for the hypothalamus 10 and the independent cortical networking by acquiring resting-state functional MRI taking the advantages of the non-linearity of the Higuchi's fractal dimension (FD) analysis 11-14 . This non-linear approach is more suitable to describe the irregular and non-periodic patterns characterizing the BOLD signature of discrete cortical areas belonging to a resting-state network (RSN) recorded by neuroimaging 15,16 as well as electrophysiological techniques17,18.Considering that in previous studies the hypothalamus has been involved in the pre-ictal period of migraine and that the activity of cortical networks is dysfunctional even outside attacks, we hypothesized that the microstructure of the hypothalamus could be altered during the pain-free period, as a favoring anatomical substrate to the recurrence of migraine, and that would be independent from aberrant RSN connectivity and migraine clinical features.ResultsDemographic characteristics of MO and HC and clinical features of MO are summarized inTable 1. No significant difference emerged between MO and HC in gender ( χ 2 1 = 0.102, p = 0.749) and age (t 38 = − 1.628, p = 0.112). In patients with MO, we did not detect white matter lesions.
Characterization of hypothalamic DTI. No multivariate outliers were present in the MANOVA model (highest Mahalanobis Distance value: 14.843). Multivariate test revealed a significant GROUPs effect (Wilks' λ = 0.736, F 4,35 = 3.131, p = 0.027). Univariate ANOVA analyses showed that, compared to HC, patients with MO showed significantly higher MD, AD, and RD hypothalamic DTI metrics, with large effect sizes. FA DTI metric did not statistically differ between groups, even though the effect size was moderate to large. In [fig_ref] Table 2: Descriptive and univariate statistics for the hypothalamus fractional anisotropy [/fig_ref] , descriptive and univariate statistics of DTI metrics for hypothalamic ROI are reported.
Regarding the DTI metrics in the four regions of interest of hypothalamus, in the rm-MANOVA model no multivariate outliers were present (highest Mahalanobis Distance value: 32.307). Multivariate test revealed a PARTs × SIDEs × GROUPs effect just above the level of significance (Wilks' λ = 0.779, F 4,35 = 2.487, p = 0.061). Despite Wilks' Lambda multivariate significance criterion not being reached, we proceeded to univariate analysis for exploratory purposes. The rm-ANOVA model of FA did not have a significant PARTs × SIDEs × GROUPs effect (F 1,38 = 0.099, p = 0.755) whereas the rm-ANOVA models for MD, AD, and RD revealed significant PARTs × SIDEs × GROUPs effects (respectively: MD: F 1,38 = 7.674, p = 0.009; AD: F 1,38 = 8.670, p = 0.005; RD: F 1,38 = 6.076, p = 0.018). Univariate ANOVA analyses revealed that FA in the anterior part of the left and right hypothalamus did not differ between HC and MO whereas FA in the posterior part of the left and right hypothalamus was significantly lower in patients with MO than HC, with large effect sizes (respectively: left posterior, d = 1.36; right posterior, d = 1.81). Consistently, MD, AD, and RD in the anterior and posterior part of the left and right hypothalamus were significantly higher in MO compared to HV, with large to huge effect sizes [ds ranging from 0.85 (AD in left anterior hypothalamus) to 2.89 (AD in left posterior hypothalamus)]. In , descriptive . Clinical and demographic of healthy controls (HC) and of patients with migraine without aura (MO). Data are expressed as frequency and mean ± SD.
## Hc (n = 20) mo (n = 20)
Women (number) Age (years) 29 ± 4 32 ± 7
Attacks frequency/month (number) 3 ± 2.0
Disease duration (year) 14 ± 6.5
Severity of headache (0-10) 7.3 ± 0.9
Headache-related disability (number) 2 ± 0.4
Tablet intake/month (number) [bib_ref] Longitudinal neuroimaging over 30 days: Temporal characteristics of migraine, Schulte [/bib_ref] No other significant correlation was found between DTI metrics of the hypothalamus, considered as a whole or in its single ROIs, and FD metrics of the networks, as well as with clinical variables.
# Discussion
Confirming our initial hypothesis, we detected an interictal alteration of hypothalamic diffusion-weighted MRI signal in the migraine group. Compared to HC, DTI MRI showed higher values of MD, AD, and RD within the hypothalamus when measured as a single ROI. In migraine patients we also detected significantly higher FD then in the HCs at the level of the salience network and cerebellum, while FD was lower in the primary visual network. Nonetheless, the complexity of cortical networks correlated with attacks frequency and severity of migraine pain. In the exploratory analysis on the contribution of anterior and posterior bilateral hypothalamic ROIs, we detected higher values of MD, AD, and RD within the all ROIs, with lower values of FA in the posterior hypothalamus bilaterally.
Microstructure of the hypothalamus. The hypothalamus, through its orexinergic and non-orexinergic neuroendocrine system, is a fundamental brain hub devoted to coordinating appropriate physiological and behavioral responses to threatening or potentially threatening internal and external factors [bib_ref] Role of the orexin/hypocretin system in stress-related psychiatric disorders, James [/bib_ref] [bib_ref] Orexin 2 receptor regulation of the hypothalamic-pituitary-adrenal (HPA) response to acute and..., Grafe [/bib_ref]. For this func- www.nature.com/scientificreports/ tion, the hypothalamus, particularly in its posterior part, is anatomically connected to the two most important systems of the brain programmed to control pain, the one located in the frontal lobes 21 and the one located in the midbrain [bib_ref] The migraine generator revisited: continuous scanning of the migraine cycle over 30..., Schulte [/bib_ref] [bib_ref] Trigeminohypothalamic and reticulohypothalamic tract neurons in the upper cervical spinal cord and..., Malick [/bib_ref]. Through these systems, the hypothalamus has an important antinociceptive function [bib_ref] A review of the role of orexin system in pain modulation, Razavi [/bib_ref].
With the help of modern neuroimaging, some recent studies have supported the historical view [bib_ref] Migraine: A cerebral disorder, Pearce [/bib_ref] of the hypothalamus as a cerebral structure playing a pivotal role in the recurrence of migraine attacks, its anterior part showing increased neural activity up to 48 h before an attack, in a time period when premonitory symptoms may occur [bib_ref] Brain activations in the premonitory phase of nitroglycerintriggered migraine attacks, Maniyar [/bib_ref] [bib_ref] The migraine generator revisited: continuous scanning of the migraine cycle over 30..., Schulte [/bib_ref] [bib_ref] Longitudinal neuroimaging over 30 days: Temporal characteristics of migraine, Schulte [/bib_ref]. Hypothalamic activation has also been described during the chronic daily or almost daily presentation of migraine attacks, especially in its anterior part when migraine is chronic, and its posterior region during headaches [bib_ref] Hypothalamus as a mediator of chronic migraine, Schulte [/bib_ref] [bib_ref] Increased neural connectivity between the hypothalamus and cortical resting-state functional networks in..., Coppola [/bib_ref] [bib_ref] Functional connectivity of hypothalamus in chronic migraine with medication overuse, Lerebours [/bib_ref]. Volume of the posterior hypothalamic regions in voxel-based morphometry processing was found lower in episodic migraine patients when compared with controls [bib_ref] Volume of hypothalamus as a diagnostic biomarker of chronic migraine, Chen [/bib_ref].
Mean, axial, and radial diffusivity were significantly higher in patients than in HC, both considering the hypothalamus as a whole and assessing its anterior and posterior bilateral ROIs separately. The MD metric typically is comprised of RD and AD and quantifies the overall magnitude of water diffusion by indicating both cellular swelling and cellular density [bib_ref] The basis of anisotropic water diffusion in the nervous system: A technical..., Beaulieu [/bib_ref]. In particular, AD and RD diffusivity is considered to be in vivo surrogate markers of myelin and axonal damage, respectively. We also observed that despite FA DTI metric did not statistically differ between groups when we considered the hypothalamus as a whole, in the exploratory analysis on the hypothalamic ROIs significantly lower values were detected in the posterior parts bilaterally.
The peculiar diffusivity pattern of higher MD, AD, and RD, with lower FA, we found in patients with MO may reflect slightly decreased cellularity (neuronal and glial cells) and/or loss in the directional organization of highly anisotropic myelinated fibers interconnecting individual hypothalamic nuclei in combination with an increased cell density [bib_ref] Functional diffusion tensor imaging: Measuring task-related fractional anisotropy changes in the human..., Mandl [/bib_ref]. Indeed, anatomically, the hypothalamus is formed by a conglomerate of grey matter nuclei and by interconnecting myelinated fibers; animal models have indicated that increased cell swelling may coincide with increased in neural electric response [bib_ref] Rapid structural changes in nerve fibers evoked by electric current pulses, Tasaki [/bib_ref].
Within network fractal dimensionality. The involvement of salience, visual, and cerebellar networks in migraine pathophysiology is not a novel finding. The salience network includes dorsal anterior cingulate and anterior insular cortices. Among its many functions, the SN is involved in self-awareness through the integration of sensory, emotional, and cognitive information [bib_ref] Saliency, switching, attention and control: A network model of insula function, Menon [/bib_ref]. In previous studies performed in MO patients, the SN showed less intrinsic functional connectivity [bib_ref] Altered brain functional connectome in migraine with and without restless legs syndrome:..., Yang [/bib_ref] and lack of BOLD response habituation to pain [bib_ref] Neuronal correlates of impaired habituation in response to repeated trigemino-nociceptive but not..., Stankewitz [/bib_ref] [bib_ref] Reduced insula habituation associated with amplification of trigeminal brainstem input in migraine, Lee [/bib_ref] in comparison to HC.
The visual areas probably play a major role in the pathophysiology of migraine. In fact, neurophysiological 36 and psychophysiological 37,38 responses, fMRI BOLD evoked activity and spontaneous connectivity [bib_ref] Structural and functional reorganization of the brain in migraine without aura, Soheili-Nezhad [/bib_ref] [bib_ref] Advanced visual network and cerebellar hyperresponsiveness to trigeminal nociception in migraine with..., Russo [/bib_ref] , metabolic activity [bib_ref] Photophobia in migraine: an interictal PET study of cortical hyperexcitability and its..., Boulloche [/bib_ref] [bib_ref] Evidence of an increased neuronal activation-to-resting glucose uptake ratio in the visual..., Lisicki [/bib_ref] and structural abnormalities 43 of the visual areas were all several times found to be altered in patients with episodic migraine when between attacks.
Even though the cerebellum was found to be activated in response to painful stimuli in several studies, it received less attention by researchers who attempt to shed light on migraine pathophysiology [bib_ref] Advanced visual network and cerebellar hyperresponsiveness to trigeminal nociception in migraine with..., Russo [/bib_ref] [bib_ref] Painful heat reveals hyperexcitability of the temporal pole in interictal and ictal..., Moulton [/bib_ref] [bib_ref] Trigeminal nociceptive transmission in migraineurs predicts migraine attacks, Stankewitz [/bib_ref] [bib_ref] Pain processing in patients with migraine: An event-related fMRI study during trigeminal..., Russo [/bib_ref] [bib_ref] Enhanced pain-induced activity of pain-processing regions in a case-control study of episodic..., Schwedt [/bib_ref] [bib_ref] Activity and connectivity of the cerebellum in trigeminal nociception, Mehnert [/bib_ref]. Its involvement in migraineur pain comes not at odd since it is well known that the deep cerebellar nuclei process noxious stimuli [bib_ref] Alteration of nociception by stimulation of cerebellar structures in the monkey, Siegel [/bib_ref] [bib_ref] The cerebellum and pain: Passive integrator or active participator?, Moulton [/bib_ref] [bib_ref] Aversion-related circuitry in the cerebellum: Responses to noxious heat and unpleasant images, Moulton [/bib_ref] and take part in the perception of pain and its inhibition through their connections with the brainstem and the thalamus [bib_ref] Activity and connectivity of the cerebellum in trigeminal nociception, Mehnert [/bib_ref] [bib_ref] Cerebellar stimulation modulates the intensity of a visceral nociceptive reflex in the..., Saab [/bib_ref].
Our findings add a further dimension to the understanding of the pathophysiology of MO; increased fractal dimensionality within the salience and cerebellar networks and decreased FD within the primary visual network could be explained postulating that loss or gain in complexity in brain activity reflects more flexible and/ or efficient information processing, as a result of short and long-range interactions between neuronal structures operating at multiple dimensional levels such as space and time [bib_ref] What is physiologic complexity and how does it change with aging and..., Goldberger [/bib_ref]. Therefore, the complex pattern found in SN and cerebellum may reflect an increased efficiency demand of brain networks devoted to the integration of sensory, emotional, and cognitive information related to the severity of migraine presentation. In support of this interpretation, we detected that both SN and cerebellum FD correlated positively with the frequency of the recurrence of migraine attacks and with the subjective perception of ictal pain intensity, a datum already described by others [bib_ref] Neuronal correlates of impaired habituation in response to repeated trigemino-nociceptive but not..., Stankewitz [/bib_ref] [bib_ref] Activity and connectivity of the cerebellum in trigeminal nociception, Mehnert [/bib_ref]. The reduced efficiency in information processing we detected within the PV network may be the morpho-functional counterpart of the neurophysiological finding of an initially slower interictal visual cortical reactivity 54 as well as of the cortical hypometabolism found with the FDG-PET 42 in between attacks of episodic migraine.
Relevance for migraine pathophysiology. In recent years, the attention of researchers has shifted from the brain stem to the hypothalamus as a possible generator of migraine. The neuroimaging results obtained longitudinally over a period of 1 month led researchers to believe that the hypothalamus, especially the anterior part, plays a primary role in the beginning of the premonitory phase of the attack 1-3 . Our study shows that the hypothalamus presents structural abnormalities, both in its anterior and posterior regions, even when it is not in a premonitory phase or during migraine attacks. We speculate that these microstructural abnormalities may reflect rising in hypothalamic excitability, and that could be considered as a neuroanatomical substrate favoring the beginning of the prodromal phase and, therefore, the ignition of an attack. The involvement of cortical networks detected by us in the same group of patients could be another favorable factor on which external modifiable factors could further lower the threshold for the activation of the trigeminovascular system which, together with the hypothalamic-pituitary neuroendocrine system, tries to maintain correct homeostasis of the body, i.e. to prevent brain dysexcitability. In this regard, it is interesting to note that most of the areas forming part of the networks analyzed by us were active, together with the hypothalamus, even during the pre-monitory phase of migraine 1 , and that, in migraine families, a gene module in the visual cortex can determine a complex genetic www.nature.com/scientificreports/ trait setting peculiar gene-gene interactions, transcriptomic, and gene-networking both at the cortical level and at the level of the neuroendocrine system 56 . Interestingly, our results show no correlation between microstructural metrics of the hypothalamus and cortical complexity metrics in interictal migraine. This could mean both that the hypothalamus is not crucial in the determinism of cortical dysfunction among migraine attacks or that it may become so only when a migraine attack is about to initiate, i.e. at a time when hypothalamus should coordinate appropriate behavioral responses to threatening or potentially threatening internal and external factors [bib_ref] Role of the orexin/hypocretin system in stress-related psychiatric disorders, James [/bib_ref] [bib_ref] Orexin 2 receptor regulation of the hypothalamic-pituitary-adrenal (HPA) response to acute and..., Grafe [/bib_ref]. Overall, the mechanistic underpinnings of these complex multilevel changes, from structure to function, are still far from being understood as they fit into the complex puzzle of migraine pathophysiology. In the context of this complexity, we note that the area belonging to the networks in which we found abnormal fractal dimensionalities are indeed similar to brain regions in which abnormal cortical hyperresponsiveness to sensory stimuli was previously detected, and therefore considered another factor favoring the repetition of migraine attacks [bib_ref] Altered processing of sensory stimuli in patients with migraine, De Tommaso [/bib_ref].
A limitation of our study relies in the nature of migraine syndrome, i.e. its high genotypic and phenotypic variability that implies a less immediate generalizability of the results. A larger cohort of patients might allow a better characterization of phenotypic subgroups of migraine patients. Another possible limitation is the lack of correlation between morpho-functional and psychometric variables, such as pain rumination or anxiety.
Future work must investigate the same diffusivity metrics also in the premonitory phase and during migraine attacks. It remains to be seen whether these interictal microstructural anomalies are permanent or can be normalized by pharmacological and non-pharmacological therapies, commonly used for migraine prophylaxis.
# Materials and methods
Participants. We prospectively enrolled 20 patients affected by episodic migraine without aura (MO) from the headache consultation of Sapienza University of Rome Polo Pontino in Latina (Italy). Patients were initially diagnosed according to the International Classification of Headache Disorders (ICHD) IIIbeta and confirmed according to the ICHD-III [bib_ref] Headache Classification Committee of the International Headache Society (IHS) The International Classification..., Ichd [/bib_ref]. We included only patients who did not have a migraine attack during the scan or in a period between 3 days before or after the scan, as well as those who were not overusing medication or did not have prophylaxis in progress or do it in the 3 months before the scan. We recruited 20 healthy subjects with no personal or family history of migraine or other primary headaches as controls (HC). For all participants in the study, the exclusion criteria were the presence of other concomitant neurological disorders, or obvious psychiatric, endocrinological, autoimmune, connective tissue disorders, and ophthalmological disorders as assessed with a complete neuro-ophthalmological evaluation including best-corrected visual acuity, slit-lamp biomicroscopy, intraocular pressure measurement, and indirect ophthalmoscopy. This study is part of a larger one performed between 2013 and 2019, in which the same participant underwent multiple neuroimaging procedures during the same experimental session. All the scanning sessions took place during the afternoon. Women who participated in the study were scanned outside the days of menstruation (11.4 ± 4.0 days from menstruation onset in HCs, 12.0 ± 3.9 in MO patients). All participants were first informed about the aims of the study and the procedures, then signed informed consent. The study was approved by the ethical committee of the Sapienza University of Rome (RIF.CE 4839). All experiments were performed in accordance with the Declaration of Helsinki. fMRI data acquisition and preprocessing. MRI data were obtained on a Siemens 3 T Verio scanner using a 12-channel head coil. Structural anatomic scans were performed using a T1-weighted sagittal magnetization-prepared rapid gradient echo (MPRAGE) series (TR: 1900 ms, TE: 2.93 ms, 176 sagittal slices, 0.508 × 0.508 × 1 mm 3 voxels) [bib_ref] Haemodynamic activity characterization of resting state networks by fractal analysis and thalamocortical..., Porcaro [/bib_ref]. We acquired an interleaved double-echo Turbo Spin Echo sequence proton density and T2-weighted images (repetition time: 3320 ms, echo time: 10/103 ms, matrix: 384 × 384, field of view: 220 mm, slice thickness: 4 mm, gap: 1.2 mm, 50 axial slices). Diffusion tensor imaging (DTI) was acquired by using single-shot echo-planar imaging, with a 128-channel head coil (TR 12,200 ms, TE 94 ms, 72 axial slices, 2 mm thickness, isotropic voxels). Images from the same participants and during the same session were obtained with diffusion gradients applied along 30 non-collinear directions, effective b values of 0 and 1000 s/mm 2 were used. Functional MRI data were obtained using T2*-weighted, echo-planar imaging (TR: 3000 ms, TE: 25 ms, 40 axial slices, 3.906 × 3.906 × 3 mm, 150 volumes). Functional resting scans lasted seven minutes and 30 s, during which participants were instructed to relax, avoid motion and keep their eyes closed, but not to fall asleep.
Data pre-processing was carried out using SPM12 software (http:// www. fil. ion. ucl. ac. uk/ spm/) implemented in MATLAB (version R2016b, MathWorks, Inc., Natick, MA, USA). Data were realigned to the first volume to correct for head motion using a 6-parameter rigid body process and resliced by cubic spline interpolation. The structural (T1-MPRAGE) and functional data were co-registered for each participant dataset. Normalization procedure transformed structural and realigned EPI images into a common stereotactic space based on Talairach and Tournoux 59 , resampled by 3 mm in each direction. Finally, the spatially normalized functional images were smoothed isotropically at 8 × 8 × 8 mm.
Diffusion tensor imaging (DTI) analysis. FSL 6.0 software package (FMRIB Image Analysis Group, Oxford, England; https:// fsl. fmrib. ox. ac. uk/ fsl/ fslwi ki) was used to process image data.
The FSL Diffusion Toolbox (FDT) was used to correct susceptibility induced distortions 60 , eddy currents 61 , and motion artifacts 62 , while the brain extraction tool (BET) was used to create brain masks from the b0 image of each participant [bib_ref] Fast robust automated brain extraction, Smith [/bib_ref]. An automated quality control framework was used to assess diffusion MRI data [bib_ref] Automated quality control for within and between studies diffusion MRI data using..., Bastiani [/bib_ref].
The FSL toolbox DTIFIT fits the pre-processed image based on a diffusion tensor model to yield AD (axial diffusivity), FA (fractional anisotropy), MD (mean diffusivity), and RD (radial diffusivity).
For each subject, a region of interest (ROI) was defined, which covers the whole of the hypothalamus. In addition, for each subject we defined 4 further regions of interest covering the left and right anterior www.nature.com/scientificreports/ hypothalamus-mostly with neuroendocrine function-and the left and right posterior hypothalamus, including the wake promoting nuclei. For this purpose, we used the coordinates provided by Boes et al. [bib_ref] Connectivity of sleep-and wake-promoting regions of the human hypothalamus observed during resting..., Boes [/bib_ref]. The size of the hypothalamic ROI in the 2 mm space was 6 voxels, 3 per hemisphere, equally for the anterior and posterior ROIs. Mean AD, FA, MD, and RD values in the hypothalamus for every subject were calculated by averaging those voxels in the ROI.
fMRI data analysis. After data preprocessing, resting-state data of all participants were concatenated (HC and MO) and analyzed using spatial independent component analysis (ICA) based on the infomax algorithm, as implemented in the Group ICA of fMRI Toolbox (GIFT-http:// trend scent er. org/ softw are/ gift/) to decompose the data into functional networks that exhibited a unique time course profile. Two data reduction steps were carried out using principal component analysis, subject-specific and group-level steps. Firstly Subject-specific data were reduced to 50 components and subsequently reduced data were concatenated over time. Secondly, at group level, data were reduced into 36 group independent components (ICs) using the expectation-maximization algorithm included in GIFT [bib_ref] Tracking whole-brain connectivity dynamics in the resting state, Allen [/bib_ref]. Spatial ICAs was also performed separately for HC and MO patients to ensure that the resulting components had similar resting-state fluctuations in the two groups as in the resulting components obtained from all 40 participants combined.
The number of ICs was estimated using the minimum description length (MDL) criterion [bib_ref] A method for making group inferences from functional MRI data using independent..., Calhoun [/bib_ref] [bib_ref] Choice of multivariate autoregressive model order affecting real network functional connectivity estimate, Porcaro [/bib_ref]. In our specific case, 33 independent components (ICs) were indicated to be estimated. Subject-specific spatial maps and time courses were obtained using the back-reconstruction approach (GICA) [bib_ref] Spatial and temporal independent component analysis of functional MRI data containing a..., Calhoun [/bib_ref].
From the 33 ICs, we identified the relevant RSNs by applying a previously described procedure [bib_ref] Tracking whole-brain connectivity dynamics in the resting state, Allen [/bib_ref]. Two experienced neuroradiologists (E.T. & F.C.) blindly reviewed the components discarding those showing spatial overlap with vascular, ventricular, edge regions corresponding to artifacts [bib_ref] Hand classification of fMRI ICA noise components, Griffanti [/bib_ref]. This process resulted in 20 meaningful ICs that we sorted into 10 functional networks, based on the spatial correlation between independent components and the template provided by GIFT Toolbox [bib_ref] Tracking whole-brain connectivity dynamics in the resting state, Allen [/bib_ref] An array of mean values L(k) was obtained, and the FD was estimated as follow:
In practice, the original curve or signal can be divided into smaller parts with or without overlap, called "windows". Then, the method for computing FD should be applied to each window where N should be seen as the length of the window. In that case, FD values are calculated for each window, with or without overlap. Individual FD values can be averaged across all windows for the entire curve, and the mean FD value can be used as a measure of curve complexity.
Here, using the single-subject IC time courses for each RSN, we calculated FD in non-overlapped time windows of 150 s (corresponding to 50 of our fMRI volumes). The choice of the free parameter k has a crucial role in FD estimation. For each window we estimated twenty-four values of FD for k = 2, …, 25. The value 25 was equal to half of the samples within our 50 volumes window (i.e. 150 s). kmax is equal to half of the window length the maximum length that can be chosen. There were three windows within our 150 volume scans, therefore we estimated three measures of FD at each value of k (e.g. FD2, FD3, FD4, …., FD24). These three measures were averaged to give one mean value of FD for each k, for each subject 12,17,71 . The process was then repeated In order to control the type I error rate due to the multiple comparisons, we carried out a model of multivariate analysis of variance (MANOVA), a model of repeated-measures of multivariate analysis of variance (rm-MANOVA), and a model of repeated-measures analysis of variance (rm-ANOVA) on the DTI metrics of the hypothalamus, the DTI metrics the four regions of interest of hypothalamus, and the FD values, respectively. MANOVA, followed by univariate ANOVAs, was employed to investigate the GROUPs effect (between-subject factor: HC vs. MO) on AD, FA, MD, and RD (dependent variables). As the presence of outliers can increase type I error rate in MANOVA, Mahalanobis Distance was used to identify potential multivariate outliers. Mahalanobis Distance critical value of chi-square distribution, for degrees of freedom = 4 and p < 0.001, was 18.47. Univariate ANOVA results were examined only if Wilks' Lambda multivariate significance criterion was satisfied. rm-MANOVA, followed by univariate rm-ANOVAs, was employed to investigate the PARTs × SIDEs × GROUPs interaction effect (PARTs and SIDEs are the two within-subject factors: respectively, anterior vs. posterior part of hypothalamus and left vs. right hypothalamus; GROUPs is the between-subject factor: HC vs. MO) on FA, MD, AD, and RD (dependent variables). For the rm-MANOVA, Mahalanobis Distance critical value of chi-square distribution, for degrees of freedom = 16 and p < 0.001, was 39.25. As for MANOVA, univariate ANOVA results were analyzed only if the Wilks' Lambda multivariate significance criterion was achieved. rm-ANOVA was performed on the FD values to investigate the interaction effect GROUPs × ICs (the two GROUPs as a betweensubject factor: HC vs. [fig_ref] Figure 1: Resting State Networks [/fig_ref]. As for MANOVA and rm-MANOVA, univariate ANOVA results were analyzed only if the Wilks' Lambda multivariate significance criterion was achieved. The sphericity of the covariance matrix was verified with the Mauchly sphericity test. In the case of violation of the sphericity assumption, the Greenhouse-Geisser epsilon adjustment was used. Cohen's d was used as a measure of effect size in univariate ANOVAs of MANOVA, rm-MANOVA, and rm-ANOVA models.
[formula] L m (k) = 1 k � i=1,int � N−m k � |X(m + ik) − X(m + (i − 1)k)| · N − 1 int � N−m k � L(k) = k m=1 L m (k) k FD = [/formula]
Analysis of Pearson correlation coefficient was performed respectively between FD values for each IC and DTI metrics values for hypothalamic ROIs and clinical variables (e.g.: the severity of headache attacks, ranging 0 to 10; the duration of migraine history, in years; the number of monthly migraine attacks; the number of days from the last migraine attack; the monthly number of acute medications).
The significance threshold was set at a p-value < 0.05.
## Data availability
Clinical, imaging and statistical data will be available upon request from any qualified investigator. www.nature.com/scientificreports/
[fig] Figure 1: Resting State Networks (RSNs) identified by GIFT. Twenty spatial maps divided into ten functional networks were found: Cerebellum (Cb-IC1); Auditory (AN-IC2); Fronto-Parietal (FPN: IC3-rFPN and IC23-lFPN); Dorsal Attention System (DAS: IC6-lDAS and IC17-rDAS); Sensory Motor (SMN-IC7, IC8, IC33); Salience (SN-IC 10 and IC19); Visual (VN: IC11-Primary Visual and IC24-Lateral Visual); Default Mode (DMN-IC15, IC20, IC28 and IC31); Precuneous (IC25 and IC27) and Language (LN-IC 26) networks based on their anatomical view. Montreal Neurological Institute (MNI) coordinates are shown as well.Correlation analysis. Significant correlation was found between severity of migraine pain, as assessed by 0-10 VAS, and FD values of IC1 (r = 0.448, p = 0.049) and IC10 (r = 0.469, p = 0.037), while attacks frequency correlated with FD values of IC10 (r = − 0.486, p = 0.030)(Fig. 3). [/fig]
[fig] Figure 2: RSNs BOLD characterization by Higuchi's fractal dimension (FD). For each panel (Left, Middle and Right)-Spatial maps of the IC obtained by GIFT toolbox. Grand average and standard error for the FD values (k = 12) are shown for both groups HC (blue) and MO (green). Left panel-Shows the results obtained for IC10 representing the aSN. Middle panel-Shows the results for IC 11 representing the PV. Right panel-Shows the results for IC 1 representing the Cerebellum. All images have been co-registered into the Montreal Neurological Institute (MNI) space. The numbers above each image refers to the X, Y and Z coordinates in MNI space. aSN anterior part of the salience network, PV primary visual network. [/fig]
[fig] Figure 3: Scatterplots between BOLD IC FD values and clinical variables. Correlation analysis between FD values of IC1 (Cb) and IC10 (aSN) BOLD activity and clinical variables (Frequency and VAS). Scientific Reports | (2021) 11:18701 | https://doi.org/10.1038/s41598-021-98213-3 [/fig]
[table] Table 2: Descriptive and univariate statistics for the hypothalamus fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) in HC and MO. [/table]
|
Clinical Implication of Hypoxic Liver Injury for Predicting Hypoxic Hepatitis and In-Hospital Mortality in ST Elevation Myocardial Infarction Patients
www.eymj.orgCoronary intervention has evolved drastically in recent years, and consequentially, the survival of ST elevation myocardial infarction (STEMI) patients has improved. 1 However, despite an abundance of studies and reports on treatment strategies for STEMI, mortality rates remain relatively high, 1 and methods for reaching an accurate prognosis for STEMI patients remain elusive, despite several proposed parameters and risk factors validated for use in risk-scoring systems. 2 While left ventricular ejection fraction (LVEF), Killip class, and presence of multivessel disease are also known risk factors contributing to the survival of STEMI patients, 3,4 convincing biochemical markers Purpose: In this study, we aimed to determine the value of hypoxic liver injury (HLI) in the emergency room (ER) for predicting hypoxic hepatitis (HH) and in-hospital mortality in ST elevation myocardial infarction (STEMI) patients. Materials and Methods: 1537 consecutive STEMI patients were enrolled. HLI in the ER was defined as a ≥2-fold increase in serum aspartate transaminase (AST). HH was defined as a ≥20-fold increase in peak serum transaminase. Patients were divided into four groups according to HLI and HH status (group 1, no HLI or HH; group 2, HLI, but no HH; group 3, no HLI, but HH; group 4, both HLI and HH). Results: The incidences of HLI and HH in the ER were 22% and 2%, respectively. In-hospital mortality rates were 3.1%, 11.8%, 28.6%, and 47.1% for groups 1, 2, 3, and 4, respectively. Patients with HLI and/or HH had worse Killip class, higher cardiac biomarker elevations, and lower left ventricular ejection fraction. Multivariate logistic regression analysis showed that HLI in the ER was an independent predictor of HH [odds ratio 2.572, 95% confidence interval (CI) 1. 166-5.675, p=0.019]. The predictive value of HLI in the ER for the development of HH during hospitalization was favorable [area under the curve (AUC) 0.737, 95% CI 0.643-0.830, sensitivity 0.548, specificity 0.805, for cut-off value AST >80]. Furthermore, in terms of in-hospital mortality, predictive values of HLI in the ER and HH during hospitalization were comparable (AUC 0.701 for HLI at ER and AUC 0.674 for HH). Conclusion: Among STEMI patients, HLI in the ER is a significant predictor for the development of HH and mortality during hospitalization (INTERSTELLAR ClinicalTrials.gov number, NCT02800421).
# Introduction
predictive of prognoses in STEMI patients have not been developed.
Since the pathophysiological nature of STEMI is directly associated with LVEF,one of the organs vulnerable to injury is the liver. Hypoxic liver injury (HLI) is caused by an acute cardiovascular event resulting in decreased hepatic blood flow or hepatic congestion due to increased central venous pressure.Hypoxic hepatitis (HH) is another term used to describe liver damage, and the mechanism of insult is equivalent to that of HLI. While the predictive value of HLI in STEMI patients and the development of HH as a prognostic factor in critically ill patients have been reported, the effects that they have upon each other and their prognostic value in STEMI patients has not been addressed.Therefore, we sought to determine the value of HLI in the emergency room (ER) for predicting HH and inhospital mortality in STEMI patients undergoing primary percutaneous coronary intervention (PCI).
# Materials and methods
## Study design and patient selection
The study protocol was approved by the Institutional Review Board of Inha University Hospital, Inha University College of Medicine (IRB No. 2016-05-015), and written consent was obtained from each patient. We collected data from four different hospitals (Inha University Hospital, Gachon University Gil Hospital, Sejong General Hospital, Soon Chun Hyang University Bucheon Hospital) located in the Incheon-Bucheon province. The four hospitals previously formed a registry on STEMI patients, called INcheon-Bucheon cohorT of patients undERwent primary PCI for acute ST-Elevation myocardial infARction (INTERSTELLAR). A total of 1537 consecutive STEMI patients (79.2% male, mean age 60.5±13.2 years) who had undergone primary PCI between 2007 and 2014 were enrolled. Patients with a prior history of chronic hepatitis, viral hepatitis, alcoholic liver disease, or toxic hepatitis were excluded. Coronary intervention was performed in accordance with current guidelines for myocardial revascularization.Pharmacological treatment and mechanical support related to primary PCI were performed at the operator's discretion.
Patients were divided into four groups according to their HLI and HH states: group 1 had no HLI or HH; group 2 had HLI, but no HH; group 3 had no HLI, but HH; and group 4 had both HLI and HH. Baseline characteristics, risk factors, echocardiographic and coronary angiographic findings and clinical features, such as Killip class, were recorded and analyzed.
## Definitions and measurements
Patients with systolic blood pressure above 140 mm Hg, diastolic blood pressure above 90 mm Hg, or prior use of antihypertensive medication were defined as having hypertension. Diabetes mellitus was defined as 1) prior use of hypoglycemic agents or insulin, 2) fasting plasma glucose above 126 mg/dL or glycosylated hemoglobin above 6.5%, or 3) previously diagnosed, but untreated hyperglycemia. The definition of dyslipidemia was total cholesterol ≥240 mg/dL, low-density lipoprotein cholesterol ≥130 mg/dL, high-density lipoprotein cholesterol <40 mg/dL, triglycerides ≥200 mg/dL, and prior use of lipidlowering agents. A patient who was currently smoking or had smoked until 1 month prior to primary PCI was considered as a smoker. STEMI was diagnosed upon an electrocardiogram showing a ST elevation of >1 mm in at least two consecutive leads or new-onset left bundle branch block, two-fold elevation of serum levels of troponin-I or creatine kinase-MB (CK-MB) above the upper normal limit, and typical anginal chest pain lasting for more than 30 min. Coronary artery disease (CAD) was defined as luminal narrowing of more than 50% in any coronary artery.HLI was defined as ≥2-fold increase in serum aspartate transaminase (AST) and/or alanine transaminase (ALT) above the upper normal limit at admission.HH was defined as a ≥20-fold increase in peak AST and/or ALT above the upper normal limit. 11
## Endpoint determination and follow-up data acquisition
Our primary endpoint was all-cause in-hospital mortality with respect to the presence of HLI at ER admission and HH development during hospitalization. Patient follow-up data were collected through either electronic medical record review or standardized telephone interviews.
## Data analysis and statistical methods
Continuous data are presented as means±standard deviations. Categorical data are presented as a percentage or absolute number. Analyses of continuous data were performed using analysis of variance, and analyses of categorical data were performed using the chi-square test to assess differences among the four study groups. Receiver operating characteristic (ROC) analysis was applied to evaluate the predictive value of HLI in the ER and HH during hospitalization on in-hospital mortality and the predictive value of HLI upon developing HH during hospitalization.Binary logistic regression analyses were performed to identify risk factors associated with developing HH during hospitalization; potential factors included clinical characteristics, laboratory findings, Killip classification, LVEF, and HLI. Hazard ratios (HR) were calculated as an estimate of the risk associated with a particular variable with 95% confidence intervals (CI). All analyses were performed using SPSS version 19.0 (IBM Corp., Armonk, NY, USA) and SAS version 9.3 (SAS Institute, Cary, NC, USA). A p value of less than 0.05 was considered statistically significant.
# Results
## Baseline characteristics of the study population and comparison among patients according to their hli state in the er and development of hh
In total, 1537 patients were enrolled. 1185 patients (77.1%) were allocated to group 1, 321 patients (20.9%) to group 2; 14 patients (0.9%) to group 3, and 17 patients (1.1%) to group 4. A summary of the baseline, laboratory, and angiographic characteristics according to HLI at ER and HH state is provided in . Groups 2, 3, and 4 had lower systolic and diastolic blood pressure (p=0.024 and p=0.048, respectively), higher heart rate (p< . Baseline, Laboratory, and Angiographic Characteristics according to HLI at ER (ER AST >80) and HH (Peak AST and/or ALT >800) 0.001), worse Killip class (p<0.001), higher cardiac biomarker elevation (p<0.001), lower LVEF (p<0.001), less frequent right coronary artery infarct (p=0.008), and more intra-aortic balloon pump usage (p=0.002), compared to group 1. The prevalences of commonly known CAD risk factors, such as diabetes mellitus, hypertension, and dyslipidemia, were not significantly different among the four groups, and the extent of CAD did not differ among these groups. Liver enzyme (AST/ALT) elevations were significantly higher in groups 2, 3, and 4 (p<0.001, respectively). Peak CK-MB levels were significantly higher in groups 2, 3, and 4 (p<0.001). The left anterior descending artery was the most common infarct-related artery in all four groups (p=0.008). Univariate logistic regression analysis showed that presence of Killip class 4, elevated heart rate, LVEF ≤35%, and HLI were predictive factors of HH. Multivariate analysis adjusted for the predictive risk factors mentioned above showed that HLI was an independent predictive factor for developing HH (odds ratio= 2.572, 95% CI 1.166-5.675, p=0.019).
## Predictive value of hli in the er and developing hh during hospitalization
In ROC analysis, the predictive value of the presence of HLI utilizing AST at the ER for the development of HH during hospitalization was more favorable than that using ALT [cut-off value for AST>80: area under the curve (AUC) 0.737, 95% CI 0.643-0.830, sensitivity 0.548, specificity 0.805] (cut-off value for ALT >80: AUC 0.704, 95% CI 0.594-0.813). Therefore, a cut-off value of AST >80 was designated as the definition of HLI. In terms of in-hospital mortality, the predictive value of HLI at the ER and HH during hospitalization in STEMI patients who had undergone primary PCI were acceptable (AUC 0.701, 95% CI 0.635-0.767, sensitivity 0.517, specificity 0.817 for a cut-off value .
## In-hospital mortality according to development of hh and hli state
# Discussion
In our study, more than one-fifth of the total study population had HLI, whereas only 2% developed HH. In-hospital mortality was significantly higher for patients who had either HLI at ER or developed HH during hospitalization. Patients who had HLI in the ER were more susceptible to developing HH during hospitalization, and both HLI at ER and HH development during hospitalization were predictive factors for in-hospital mortality. While it is evident that HH is a poor prognosticator in STEMI patients, it seems that HLI is a better prognostic factor, since the overall incidence of HH development is quite low. Recent developments in interventional cardiology have provided STEMI patients who undergo PCI with higher survival rates.The prognosis of each individual, however, can vary . Flow chart of HLI at ER and HH incidence and outcomes in STEMI patients undergoing primary PCI. HLI, hypoxic liver injury; ER, emergency room; HH, hypoxic hepatitis; AST, aspartate transaminase; STEMI, ST elevation myocardial infarction; PCI, percutaneous coronary intervention.
from patient to patient, and there have been many studies reporting on prognostic factors related to coronary intervention. Tsai, et al.reported that acute kidney injury in patients who have undergone PCI was an independent predictor of in-hospital mortality. Others showed that dysglycemia was another indicator of poor outcomes: dysregulation of hormones in stressful situation precipitates insulin resistance causing elevated serum glucose that leads to organ damage.Many endeavors from recent years have discovered that serum aminotransferase, a biomarker broadly used to assess liver function, is a novel indicator of myocardial injury.The notion of serum aminotransferase as indicative marker for prognosis in STEMI has been reiterated several times in recent studies. One study showed that liver function enzymes (AST/ALT) were independent predictors of in-hospital mortality in STEMI patients, even after adjusting for CK-MB.Meanwhile, another indicated that HLI (even mild to modest serum aminotransferase elevation) results in poor prognosis in STEMI patients.Moreover, HLI in conjunction with acute kidney injury and dysglycemia maintained clinical significance as a prognosticator.Up until recently, the term HLI was used synonymously with HH. Hepatic function impairment is usually derived from either hypotensive state or passive congestion due to heart failure that causes centrilobular necrosis and ultimately elevates liver enzymes.Despite its clinical significance, the terminology used describing hepatic hypoxic insult is ambiguous. We acknowledged HLI and HH as independent syndromes and sought to verify the effect of each condition on the prognosis of STEMI patients. To minimize controversy regarding the type of liver enzyme and the cut-off value selected to define HLI, ROC analysis with both AST and ALT was performed. The AUC for AST with respect to a cut-off value above 80 was not only larger, but also had higher sensitivity than ALT in predicting the development of HH and in-hospital mortality, supporting our selection of AST as an HLI defining biomarker. Our study showed that patients presenting with HLI in the ER were at higher risk of mortality than patients without hepatic dysfunction. This result is consistent with a previous study suggesting that HLI could be a useful predictor of prognosis in STEMI patients.Patients developing HH during hospitalization were also related with higher death rates. In support thereof, a previous study highlighting the clinical implication of HH in ICU patients showed that HH has clinical significance in predicting mortality in critically ill patients.In our study, patients who had HLI presenting in the ER were more prone to developing HH during hospitalization than patients without HLI (5% vs. 1%). Univariate and multivariate logistic analyses also indicated that HLI was a predictive factor for developing HH. Indeed, patients who were absent of HLI were likely not to develop HH or experience in-hospital mortality. However, for patients with HLI, the rate of HH development was five-fold higher than that for patients without HLI, and for individuals who did develop HH, in-hospital mortality was highest. Even for patients who did not develop HH, those with HLI in the ER were at higher risk of death, which suggests that both the presence of HLI and HH development are associated with poor prognosis. Interestingly, conventional risk factors for CAD were not associated with the development of HH, whereas clinical factors (Killip class, heart rate) and hemodynamic status (LVEF ≤35%) were factors indicative of developing HH, supporting the hypothesis that hepatic hypoxic injury is usually due to a hypotensive condition and dysregulation of hormones.Numerous studies have reported the clinical manifestation of hepatic dysfunction in STEMI patients. However, the interpretation of the terms HLI and HH have been arbitrary, causing misconception of the two distinctly different syndromes. Our study differentiated HLI from HH according to degrees of hepatic dysfunction and managed to investigate their prognostic relevance and clinical relationships between the two syndromes.
Our study has several limitations. First, the design of our current study was observational, and the study population only comprised Koreans. Second, histopathological confirmation of centrilobular necrosis of hepatocytes, consistent with HH was not undertaken. Third, elevated AST levels could have been attributed to the sizes of myocardial infarctions, along with symptom to balloon time, in these patients. Patients with large myocardial infarctions and long symptom to balloon times tend to have higher incidences of HLI and HH and could have obscured the results of our study.
In conclusion, our study was able to show that STEMI patients presenting with HLI are at higher risk of developing HH during hospitalization. Both the presence of HLI in the ER and development of HH during hospitalization were independent predictors of in-hospital mortality. Therefore, HLI in the ER and development of HH during hospitalization could emerge as significant predictors for mortality during hospitalization in STEMI patients undergoing primary PCI. In-hospital mortality (%) . In-hospital mortality of STEMI patients according to HLI at ER and HH. STEMI, ST elevation myocardial infarction; HLI, hypoxic liver injury; ER, emergency room; HH, hypoxic hepatitis. |
Presence–Absence Variation in A. thaliana Is Primarily Associated with Genomic Signatures Consistent with Relaxed Selective Constraints
one other accession. Only genes with at least one, but not all, exons missing were considered. For the regression of histogram density estimates against bin midpoints, adj. r 2 = 0.47, p = 5e -4 .
Where there is a significant enrichment (p-value <= 0.05) between the amount of observed and expected E-PAV genes for a particular category an asterisk is shown over the bars.
[fig] 3 Supplementary, Figure 3: n=81)those with their entire coding region missing in at least one accessionby GOslim categories for molecular function (a), biological process (b) and cellular component (c). Both expected and observed number of CDS-PAV affected genes per category represented on each bar. Where there is a significant enrichment (p-value <= 0.05) between the amount of observed and expected CDS-PAV genes for a particular category an asterisk is shown over the bars. Distribution of 'exon presence/absence' (E-PAV) genesthose with at least one, but not all, exons missing in at least one accessionby GO category (n=330). Only GO terms with 200 or greater genes were considered. Both expected and observed number of E-PAV genes per category represented on each bar. [/fig]
|
The S-Layer Glycoprotein of the Crenarchaeote Sulfolobus acidocaldarius Is Glycosylated at Multiple Sites with Chitobiose-Linked N-Glycans
Glycosylation of the S-layer of the crenarchaea Sulfolobus acidocaldarius has been investigated using glycoproteomic methodologies. The mature protein is predicted to contain 31 N-glycosylation consensus sites with approximately one third being found in the C-terminal domain spanning residues L 1004 -Q 1395 . Since this domain is rich in Lys and Arg and therefore relatively tractable to glycoproteomic analysis, this study has focused on mapping its N-glycosylation. Our analysis identified nine of the 11 consensus sequence sites, and all were found to be glycosylated. This constitutes a remarkably high glycosylation density in the C-terminal domain averaging one site for each stretch of 30-40 residues. Each of the glycosylation sites observed was shown to be modified with a heterogeneous family of glycans, with the largest having a composition Glc 1 Man 2 GlcNAc 2 plus 6-sulfoquinovose (QuiS), consistent with the tribranched hexasaccharide previously reported in the cytochrome b 558/566 of S. acidocaldarius. S. acidocaldarius is the only archaeal species whose N-glycans are known to be linked via the chitobiose core disaccharide that characterises the Nlinked glycans of Eukarya.
# Introduction
In many Archaea the surface layer (S-layer) proteins are the sole cell wall component [bib_ref] Archaeobacterial cell envelopes, Koenig [/bib_ref]. These S-layer proteins assemble into a natural 2-D crystal structure with very strong self interactions. In Archaea, which do not possess other cell wall components, the S-layer has to maintain the cell integrity and stabilize as well as to protect the cell against mechanical and osmotic stresses or extreme pH conditions. It is also predicted that the S-layer has to maintain or even determine the cell shape [bib_ref] S-layer proteins, Sára [/bib_ref] [bib_ref] Structural research on surface layers: a focus on stability, surface layer homology..., Engelhardt [/bib_ref] [bib_ref] Are S-layers exoskeletons? The basic function of protein surface layers revisited, Engelhardt [/bib_ref] [bib_ref] Primary structure of selected archaeal mesophilic and extremely thermophilic outer surface layer..., Claus [/bib_ref] [bib_ref] Functions of S-layers, Beveridge [/bib_ref].
In Sulfolobus spp. the S-layer is composed of two proteins: a small protein of approximately 45 kD, SlaB, and a large protein, SlaA, of approximately 120 kDa. SlaB is an integral membrane protein and its strong interaction with SlaA, which covers the whole cell surface, tethers the S-layer to the membrane [bib_ref] Organization and interactions of cell envelope proteins of the extreme thermoacidophile Sulfolobus..., Grogan [/bib_ref] [bib_ref] Acidianus, Sulfolobus and Metallosphaera surface layers: structure, composition and gene expression, Veith [/bib_ref]. Taking into account the harsh growth condition of the thermoacidophilic Sulfolobus spp. (pH 2-3 and 75-80 - C), the S-layer proteins will play an important role in maintaining cell integrity and must be adapted to be functional under these conditions. One possible posttranslational modification proteins can undergo is glycosylation, which has a major effect on stability and half-life [bib_ref] Biological roles of oligosaccharides: all of the theories are correct, Varki [/bib_ref]. Indeed, all archaeal S-layer proteins which have been structurally studied to date, have been found to carry N-glycans [bib_ref] Exchange of Ser-4 for Val, Leu or Asn in the sequon Asn-Ala-Ser..., Zeitler [/bib_ref] [bib_ref] Identification and characterization of the unique N-linked glycan common to the flagellins..., Voisin [/bib_ref] [bib_ref] Primary structure and glycosylation of the S-layer protein of Haloferax volcanii, Sumper [/bib_ref] [bib_ref] Asparaginyl-N-acetylgalactosamine. Linkage unit of halobacterial glycosaminoglycan, Paul [/bib_ref] [bib_ref] Purification and characterization of a prokaryotic glycoprotein from the cell envelope of..., Mescher [/bib_ref] [bib_ref] Three-dimensional structure of the regular surface glycoprotein layer of Halobacterium volcanii from..., Kessel [/bib_ref] [bib_ref] S-layer, surface-accessible, and concanavalin a binding proteins of Methanosarcina acetivorans and methanosarcina..., Francoleon [/bib_ref].
Although Eukarya, Bacteria, and Archaea all share certain characteristics of the N-glycosylation pathway, the resulting glycan structures in Bacteria and Archaea are more diverse than in Eukarya [bib_ref] Protein glycosylation in bacterial mucosal pathogens, Szymanski [/bib_ref] [bib_ref] Not just for Eukarya anymore: protein glycosylation in Bacteria and Archaea, Abu-Qarn [/bib_ref]. Notably a far greater variety of monosaccharides is used, many of which carry functionalities such as sulfate and methyl groups, or even amino acids such as threonine [bib_ref] Identification and characterization of the unique N-linked glycan common to the flagellins..., Voisin [/bib_ref]. In the last two years, substantial progress in describing the enzymes involved in [bib_ref] Cytochrome b 558/566 from the archaeon sulfolobus racidocaldarius has a unique Asn-linked..., Zähringer [/bib_ref]. (b) This shows a symbolic representation of the glycan which is used in the annotations of archaeal N-glycosylation pathways has been made [bib_ref] Not just for Eukarya anymore: protein glycosylation in Bacteria and Archaea, Abu-Qarn [/bib_ref] [bib_ref] Asparagine-linked protein glycosylation: from eukaryotic to prokaryotic systems, Weerapana [/bib_ref] [bib_ref] N-glycosylation in Archaea: on the coordinated actions of Haloferax volcanii AglF and..., Yurist-Doutsch [/bib_ref] [bib_ref] Identification of genes involved in the assembly and attachment of a novel..., Vandyke [/bib_ref] [bib_ref] Identification of genes involved in the biosynthesis and attachment of Methanococcus voltae..., Chaban [/bib_ref].
The archaeal N-glycosylation machinery combines aspects of both the eukaryal and bacterial pathways. For instance Archaea and Eukarya use dolichol as the lipid carrier whereas Bacteria use undecaprenyl. On the other hand, in Archaea and Bacteria the oligosaccharyltransferase is comprised of a single subunit whereas in Eukarya it is a multimeric complex. So far proteins from about 25 species of Archaea have been reported to be glycosylated and about ten have had their N-glycans partially or fully characterized [bib_ref] Not just for Eukarya anymore: protein glycosylation in Bacteria and Archaea, Abu-Qarn [/bib_ref] [bib_ref] Protein glycosylation in Archaea: sweet and extreme, Calo [/bib_ref]. The best understood are S-layers of the halophiles Halobacterium salinarum and Haloferax volcanii, and S-layers and flagellins of the methanogens Methanothermus fervidus, Methanococcus voltae, and M. maripaludis. In contrast to the N-glycans of Eukarya, which are almost always branched and usually considerably greater than six residues in size, these archaeal glycoproteins contain unbranched glycans most of which have fewer than six sugar residues. The exception is Hbt. salinarum which, in addition to bearing a trisaccharide composed of monosulfated glucuronic acid linked via glucose at about ten consensus sites, has one N-linked site occupied by a GalNAc-linked polysaccharide comprised of multiple repeats of a sulfated pentasaccharide (composed of GlcNAc, GalNAc, Gal, GalA, and 3-O-methyl-GalA) [bib_ref] Sequence of the halobacterial glycosaminoglycan, Paul [/bib_ref]. In H. volcanii the S-layer glycoprotein is modified by the attachment of a pentasaccharide, composed of two hexoses, two hexuronic acids, and a methylester of hexuronic acid [bib_ref] Identification of AglE, a second glycosyltransferase involved in N glycosylation of the..., Abu-Qarn [/bib_ref] [bib_ref] Haloferax volcanii AglB and AglD are involved in Nglycosylation of the S-layer..., Abu-Qarn [/bib_ref]. The N-glycans attached to the S-layer of Methanothermus fervidus are hexasaccharides containing 3-O-methylmannose, mannose, and GalNAc [bib_ref] Primary structure of the heterosaccharide of the surface glycoprotein of Methanothermus fervidus, Karcher [/bib_ref]. In M. voltae the flagellins and S-layer proteins are glycosylated with a complex trisaccharide composed of GlcNAc, GlcdiNAcA, and a threonine-substituted ManNAcA [bib_ref] Identification and characterization of the unique N-linked glycan common to the flagellins..., Voisin [/bib_ref]. A second strain of M. voltae has been recently found to carry tetrasaccharides which share this trisaccharide sequence capped by an uncharacterized sugar [bib_ref] AglC and AglK are involved in biosynthesis and attachment of diacetylated glucuronic..., Chaban [/bib_ref]. An even more unusual tetrasaccharide has been found on the pilins and flagellins of M. maripaludis.
This has the sequence Sug-4-β-ManNAc3NAmA6Thr-4-β-GlcNAc3NAcA-3-β-GalNAc where Sug is a highly unusual aldolase diglycoside [bib_ref] A novel N-linked flagellar glycan from Methanococcus maripaludis, Kelly [/bib_ref].
Glycosylation of extreme thermophile members of the Archaea domain is quite poorly understood despite the fact that one member of this class, Thermoplasma acidophilum, from the Euryarchaeota kingdom of Archaea, was amongst the first of the archaea to have its glycoproteins studied by biophysical methods [bib_ref] Purification and partial characterization of a procaryotic glycoprotein from the plasma membrane..., Yang [/bib_ref]. This early study, which reported the presence of branched mannose-rich glycans linked via GlcNAc to Asn, has not, however, been followed up with more rigorous structure analysis. A second extremophile member of the Euryarchaeota kingdom to have its glycosylation studied is Pyrococcus furiosus, which, interestingly, has also been shown to biosynthesis branched glycans [bib_ref] Structure-guided identification of a new catalytic motif of oligosaccharyltransferase, Igura [/bib_ref]. The oligosaccharyltransferase from this species has been purified, and its ability to glycosylate a fluorescently labeled peptide containing a consensus sequence has been assayed in the presence and absence of lipid-linked oligosaccharide (LLO) prepared from Pyrococcus furiosus cells. In the presence of the LLO a glycopeptide was produced which was shown by mass spectrometry to be a branched heptasaccharide having a pentose sugar attached to each of the second and third residues of a pentasaccharide of sequence HexNAc-HexA-Hex-Hex-HexNAc [bib_ref] Structure-guided identification of a new catalytic motif of oligosaccharyltransferase, Igura [/bib_ref].
Branching is also a feature of the only glycan so far determined from a member of the Crenarchaeota kingdom. Thus cytochrome b 558/566 of Sulfolobus acidocaldarius, which grows optimally at 75-80 - C and pH 2-3, was shown to be modified with a tribranched hexasaccharide of composition Glc 1 Man 2 GlcNAc 2 plus 6-sulfoquinovose, an unusual sugar which is characteristic of chloroplasts and photosynthetic bacteria (seefor structure) [bib_ref] Cytochrome b 558/566 from the archaeon sulfolobus racidocaldarius has a unique Asn-linked..., Zähringer [/bib_ref]. As, to date, this is the only characterised glycan structure from a crenarchaeal species our objective is to determine the glycan composition of other extracellular proteins of the Sulfolobales. It is known that nearly all extracellular proteins found in these organisms Archaea 3 are glycosylated [bib_ref] Organization and interactions of cell envelope proteins of the extreme thermoacidophile Sulfolobus..., Grogan [/bib_ref] [bib_ref] Glucose transport in the extremely thermoacidophilic Sulfolobus solfataricus involves a high-affinity membrane-integrated..., Albers [/bib_ref] [bib_ref] Sugar transport in Sulfolobus solfataricus is mediated by two families of binding..., Elferink [/bib_ref]. As a first model protein, the Slayer protein of S. acidocaldarius was isolated and its glycosylation investigated using glycoproteomic methodologies.
# Materials and methods
## Strains and growth
Conditions. S. acidocaldarius (DSM639) was grown in Brock medium at pH 3 and 76 - C [bib_ref] Sulfolobus: a new genus of sulfur-oxidizing bacteria living at low pH and..., Brock [/bib_ref] and the medium was supplemented with 0.1(w/v) % of tryptone as sole carbon and energy source. Growth of cells was monitored by measuring the optical density at 600 nm.
## S-layer isolation.
Fresh cells or frozen cell pellets from a 50 ml culture were resuspended in 40 ml buffer A (10 mM NaCl, 1 mM PMSF, 0.5% Na-Lauroylsarcosine) with the addition of a little bit of DNAse. The samples were shaken for 45 minutes at 37 - C and centrifuged for 20 min in an Optima Max-XU Ultracentrifuge (Beckman Coulter) at 16.000 rcf, yielding a brownish tan pellet. The pellet was resuspended in 1,5 ml buffer A and incubated for 30 min at 37 - C. After centrifugation in a tabletop centrifuge at 14.000 rpm the pellet was purified by repeatable washes in buffer B (10 mM NaCl, 0,5 mM MgSO 4 , 0.5% SDS), incubation for 20 min at 37 - C and subsequent centrifugation, until a translucent tan pellet was obtained. Once the pellet was translucent the Slayer proteins were once washed with water ad then stored in water at 4 - C.
## Proteolytic digestion for glycoproteomic analysis.
Purified S-layer samples of S. acidocaldarius were run on a 2-8% precast gel (Invitrogen, Paisley, UK) and stained with Novex Colloidal blue stain (Invitrogen). The S-layer was observed as a broadband between 116 and 160 kDa. The band was then excised and cut into pieces, destained using 400 μl of 50% (v/v) acetonitrile in 0.1 M ammonium bicarbonate (pH 8.4) and dried in a SpeedVac. Reduction/carboxymethylation was carried out by swelling and incubating the dried gel pieced in Dithiothreitol (10 mM) (200 μl) in ammonium bicarbonate (AMBIC) (50 mM, pH 8.4) (Roche, West Sussex, UK) at 56 - C for 30 min. The DTT solution was then removed and the gel pieces were washed with acetonitrile (200 μl) and dried. The gel pieces were incubated in dark at room temperature in 50 mM iodoacetic acid (IAA) (200 μl) (Sigma-Aldrich Dorset, UK) which was dissolved in ammonium bicarbonate (50 mM, pH 8.4). The IAA was then removed and gel pieces were washed in AMBIC buffer (500 μl) for 15 min. The AMBIC was removed and gel pieces were shrunk in acetonitrile (200 μl) for 5 min. The gel pieces were dried in SpeedVac and then subjected to digestion with trypsin or chymotrypsin. For the tryptic digest the dried gel pieces were reswelled in AMBIC solution and incubated at 37 - C with 25ng/μl trypsin (20 μl) (Promega cat V5111) overnight. The supernatant was removed and placed in a clean eppendorf. The gel pieces were then incubated in 0.1% TFA (50 μl) at 37 - C for 10 min. Acetonitrile (100 μl) was added to the mixture which was incubated at 37 - C for 15 min. The supernatant was then pooled with the previous supernatant and the process was repeated twice. The supernatant volume was then reduced (to about 35 μl) in preparation for LC-MS. For the chymotryptic digest re-swelled gel pieces were incubated at 37 - C in 25 ng/μl chymotrypsin (Sigma C-3142) dissolved in Tris-HCl (100 mM, pH 7.8), overnight. The remainder of the experiment was carried out as for tryptic digestion.
# Lc-ms analysis.
The extracted peptides/glycopeptides from the gel pieces were analyzed a nano-LC-ES-MS/MS employing a quadrupole TOF mass spectrometer (Q-STAR Pulsar I, MDS Sciex). Separation of the peptides/glycopeptides was carried out by using a nano-LC gradient method generated by an Ultimate pump fitted with a Famos autosampler and a Switchos microcolumn switching module (LC Packings, Amsterdam, The Netherlands). The system was coupled to an analytical C 18 nanocapillary (75 m inside diameter × 15 cm, PepMap) and a microprecolumn C 18 cartridge for online peptide/glycopeptide separation. The digest was first loaded onto the precolumn and eluted with 0.1% formic acid (Sigma) in water (HPLC grade, Purite) for 4 min. The eluant was then transferred onto the column and eluted at a flow rate of 150 nL/min using the following gradient of solvent A [0.05% (v/v) formic acid in a 95:5 (v/v) water/acetonitrile mixture] and solvent B [0.04% formic acid in a 95:5 (v/v) acetonitrile/water mixture]: 99% A from 0 to 5 min, 99 to 90% A from 5 to 10 min, 90 to 60% A from 10 to 70 min, 60 to 50% A from 70 to 71 min, 50 to 5% A from 71 to 75 min, 5% A from 75 to 85 min, 5 to 95% A from 85 to 86 min, and 95% A from 86 to 90 min. Data acquisition was performed using Analyst QS software with an automatic information-dependent-acquisition (IDA) function.
# Sugar composition analysis.
Samples were hydrolysed in 1 M methanolic hydrogen chloride at 80 - C for 16 h and the reagent was removed under a stream of nitrogen. Hexosamines were re-N-acetylated in 500 μl of methanol/pyridine/acetic anhydride (500:1:5, v/v/v) for 15 min at room temperature, then dried under nitrogen. Trimethylsilyl derivatisation was performed in 200 μL of Tri-Sil "Z" (Pierce) at room temperature for 30 min, after which the reagent was removed under nitrogen. Derivatized monosaccharides were resuspended in 1 ml of hexanes, centrifuged at 3000 rpm for 10 min, and the supernatant transferred and dried under nitrogen for analysis by gas chromatography-mass spectrometry (GC-MS).
# Gc-ms analysis.
This was carried out using a Perkin Elmer Clarus 500 instrument, fitted with a RTX-5 (30 m × 0.25 mm internal diameter, Restek Corp.). Temperature program: the oven was held at 65 - C for 1 min before being increased to 140 - C at a rate of 25 - C/min, then to 200 - C at a rate of 5 - C/min and finally to a temperature of 300 - C at a rate of 10 o C/min. [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] T13 T11 : Polypeptide sequence of the S. acidocaldarius S-layer. The N-terminal signal sequence has been omitted. Consensus sequences for N-glycosylation are shaded and the predicted products of tryptic digestion are shown by underlining.
[formula] Archaea E T K A N I Q I N S V V T S I V N S S S V Q I N V V F L V L G S A P A G S G T L L Y A Q I G G A S V P P T L L G P L Y K S L Y S L P I N V S T T T Y P Y L F S V V V F V V K D V S D Y Y A N Y G S T P F R V T V S G T T A G T Q A N S V V G E Q M N T N A Q T T Y S N Y S T L V P N T G L I Q Q G T S T D I L S G F P T S S V S T V Q A G T E D Y P Q A T L I V P H Y Y A E S S L S G L N G N V Q M P N Y L G A Q Y P S N S F I F P A Y L I V P V T S G S F S Q G G L T F S T L T G F K Y S L Y T F S T P L N N I S A I W A L Q A M N I T L P G P V L Q F T F P V S G S A Y T D I L I L A R G W V L K T V Y E V P T T S G V V L L Y V Y F K K G N I G S P N I T L A G A T L Q S I I G L F S E Y G N S T P T S G V T I Q V T Y P T S Y W N G T L T I T T G A P G G I E Y V P L T Y K N S S T T V T V G Q P L F N T S S P Y E V S A S T A T L K G Q A T I T S S G T I S I E S P P T S L I A T S Y Y P S Q T T I G E Y A V A K L L I Q S V Y N T L A F S L V Y D S Y L A G I V Y A M A T I A A Y G P T V I I G N F S Y Q T L P I N E Y T L T L L N V T A G F D V L A I Y S T S P V T F F E L Q L P S T G T E T A P V S P A L L T L G F P A L T T V E N N M V Q M P E Q S G P D Y I T V T V Y N N P C I F N Q Y N Y P A T S I S P P I V I G Y F Y G L N V S L K N N V T D Y A S V T V S G W L Y T F T T V A A G K N F N S Y G F V G S P N Y V S S G P F I V A G M N G Y K L P L S S T T F S G S P Y P D L V A D S P L S A S V P S A V I S L G G M L V P M G N V T Q I F Y Y N D T P N N D A Y Q S L Y N T N G T L A V P S F P G S Q A T L T N Q S A V Y L G N I M N N G V T A G D I K Y T I Y I N L N Y N P T A M G P L P S P V Q A A S P L L Y M F G P S S T S Y L P T T S S P F Y L N A P V S V T K F V K Y Q G T S Y F Y H L V A Y D T T T Y T L K I N S G V F P V V Y I S L T V S I S K I F G L Y Q Q V A S V G F L V Y N P S N W L S T G L G I Y T P L P S G Y Y P E S L Y L P V S L Q T V S L I L Q Q V N L N L L K E L G T I S I G F Q F C A Q L E Q G V L Q P F L T Q M N S T T L P I V V T Q K V V N E A I F W P P A F G T Y P Y A T L L F E V V G A G K A V L S L T S L A V L K T T A T L T E G T V Y D H Y S L T A V T Q G Y I N T I V T Q G S Y Y V S A K G G L G V V V T S L S L I A S P N T K A F Q L D S Y T M F P Y A K G N L V I A L T Y V A F S T K T L S S P I Q V V P V I T F Q L S G T F G A A L E T L T P L V N G A P V Q F P A L R I P E A T I P V Q T T S T G V G D S V T I I A A G I A S G T Y T V F P L L F L G A P G L S P S G T H V N F Q S A E Y V F A Q I Q F E Q P Y P L I S T V Y I K L V Q Q Q P T I V I Y E Y K L K P A N P L P V Q S L V P G G A F P V A S L V I V Y K G A P L N S I G Q L Q V Y V L L N P V S T G I G L A P Y A R V A S Y S F Y V P A A T Q T L T F I L Q A P Y E T V N S V N S Y N T S N T F N T S N S V N T K N T G N T A [/formula]
# Results
## Strategy for glycoproteomic analysis of the s. acidocaldarius s-layer.
The polypeptide sequence of the S. acidocaldarius S-layer SlaA protein (Saci 2355) is shown in [bib_ref] The Universal Protein Resource (UniProt) in 2010, The Uniprot Consortium [/bib_ref]. The mature protein is predicted to comprise 1,395 amino acids and to contain 31 consensus sites for N-glycosylation. These are scattered throughout the S-layer with the greatest density being in the C-terminal domain where about 25% of the polypeptide contains one third of the consensus sequences. This domain is rich in Lys and Arg, in contrast to the remainder of the S-layer where these residues are quite sparse. As shown in , the majority of the predicted tryptic peptides from Leu 1003 onwards are smaller than about 40 residues, and are thus well suited to electrospray tandem mass spectrometry (ES-MS/MS) whereas many of those in Archaea 5 the N-terminal domain are much larger and are therefore expected to be far less tractable to proteomic analysis. We decided, therefore, to focus our efforts on defining glycosylation in the C-terminal domain by performing nano-LC-MS/MS analyses of in-gel tryptic digests [fig_ref] Figure 3: SDS PAGE gel of the S [/fig_ref] of the S-layer and manually searching the resulting MS/MS data for potential C-terminal glycopeptides. First, we identified spectra containing fragment ions suggesting the presence of sugars. Then, promising MS/MS data were examined for the presence of peptide sequence ions that could be attributed to predicted tryptic peptides in the C-terminal domain. Finally, glycopeptide structures were deduced taking into account likely peptide and glycan compositions, the latter assignment being assisted by knowledge of the glycan on the cytochrome b 558/566 of S. acidocaldarius. Additionally, all the mass spectra acquired in, and adjacent to, the elution windows of the identified glycopeptides were examined for evidence of molecular ions consistent with glycoforms whose abundance and/or m/z values had precluded their selection for MS/MS analysis by the automatic software. The results of these analyses are presented in the following sections.
## Evidence that t-26 and t-24 are glycosylated. following on-line nano-lc-es-ms analysis of the s-layer tryptic digest
and manual interpretation of the resulting data, a number of multiply charged molecular ions were observed whose product ion spectra were indicative of peptide glycosylation. Once recognised, related glycoforms were identified by summation across the appropriate nano-LC elution time.
The summed mass spectrum of components eluting between 46 and 52 min is shown in [fig_ref] Figure 4: Summed mass spectra recorded between 46 and 52 min from an on-line... [/fig_ref].. The signal at m/z 1111.59 was not selected for MS/MS but its m/z value is consistent with this component having one fewer hexose than m/z 1151.09 [fig_ref] Figure 4: Summed mass spectra recorded between 46 and 52 min from an on-line... [/fig_ref]. Without MS/MS data, we are not able to determine whether Glc or Man are absent from the glycan. It is interesting that we do not observe a molecular ion corresponding to Hex 2 HexNAc 2 .
We observed similar patterns for all the other tryptic glycopeptides, although there was some variation in relative abundances, particularly of minor components. Some sites had significant amounts of glycans that were truncated to a single GlcNAc. An example is shown in [fig_ref] Figure 6: Summed mass spectra of T24 recorded between 48 [/fig_ref] , which is the tryptic glycopeptide T-24 having the sequence LLNL-NVQQLNNSILSVTYHDYVTGETLTATTK. Triply charged [M + 3H] 3+ molecular ions are observed at m/z 1256.38, 1378.09, 1507.44 and 1561.45 corresponding to glycans of composition HexNAc, Hex 1 HexNAc 2 , Hex 2 HexNAc 2 QuiS, and Hex 3 HexNAc 2 QuiS, respectively. The MS/MS data confirmed the identity of the peptide T-24 (data not shown). For this glycopeptide a relatively abundant ion for the glycopeptides truncated to a single HexNAc was observed (m/z 1256.4). This was not a significant glycoform in the case of T-26.
## All observed consensus sequences in the c-terminal
Domain Are Glycosylated. Using the same logic as applied to T-26 and T-24 (see above) we unambiguously identified all the remaining tryptic peptides in the C-terminal domain with the exception of T-29. This is a 66 residue peptide having three consensus sequences (A 1115 SVYY. . .SSLTK 1180 ) and is likely to be too large for the ES-MS experiment. The identified glycopeptides are shown in [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] which summarises the m/z values and compositions for glycoforms observed by ES-MS. In an attempt to identify the consensus sites falling within the T-29 tryptic glycopeptide we carried out a chymotryptic digestion. This is not an ideal enzyme for glycoproteomics because it cleaves relatively nonspecifically at large hydrophobic residues as well as at aromatic residues. Hence it yields a very complex digest and consequently the MS and MS/MS data are an enormous challenge for manual interpretation. Fortunately, however, two useful sets of glycopeptide data from the C-terminal domain of the S-layer were revealed by manual inspection. Firstly high quality spectra were found for the C-terminus itself (AGGPVLSEYPAQLIFTNVTLTQ; designated C-1 in [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] , which includes the consensus sequence at Asn 1390 . MS/MS results (not shown) confirmed the major glycoform was the same as that assigned in the tryptic digest [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] where our evidence had been confined to MS data only. More usefully the chymotryptic digest gave data corresponding to a glycopeptide of sequence TIVPNNTVVQIPSSL which spans the consensus site at Asn 1168 within T-29 (designated C-2). Once again the glycan profile was similar to other observed sites with the hexa-and trisaccharides being the most abundant components [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref].
## Sugar analysis confirms man, glc, and glcnac content.
The S. acidocaldarius S-layer was analysed for its sugar composition by GC-MS of trimethylsilyl (TMS) methyl glycoside derivatives and the data are shown in . Consistent with both the MS data described earlier, and the structure reported by Zähringer and colleagues [bib_ref] Cytochrome b 558/566 from the archaeon sulfolobus racidocaldarius has a unique Asn-linked..., Zähringer [/bib_ref] , the analysis showed Man, Glc, and GlcNAc as the only observable sugars. No GalNAc was present, confirming the chitobiose core sequence. The Man:Glc ratio is consistent with the : GC-MS analysis of TMS methyl glycoside sugar derivatives obtained from S. acidocaldarius S-layer glycoprotein. Note that for experimental reasons GlcNAc recoveries are always poor in sugar analysis experiments. We consider it likely that this is the reason for the GlcNAc:Mannose ratio being much lower than expected from the LC-ES/MS/MS data, although we cannot rule out the possibility that other mannose-containing polymers are present in the sample.
[formula] G A G V V E F L L T A Y 1221 1517 P K V V N E A I F W P P A [/formula]
# Discussion
In this study we have employed glycoproteomic methodologies to map the C-terminal domain of the S. acidocaldarius Slayer, spanning residues L 1004 -Q 1395 . This domain has eleven potential N-linked glycosylation sites, nine of which we have shown to be glycosylated [fig_ref] Figure 7: Sequence of the C-terminal domain of the S-layer which was mapped in... [/fig_ref] , including Asn 1390 which is only six residues from the C-terminus. The two sites which were not identified, Asn 1120 and Asn 1154 , fall within a 66 residue tryptic peptide (T-29) that has three consensus sequences, the third of which (Asn 1168 ) was found to be occupied via analysis of a chymotryptic digest of the S-layer. Unfortunately the other two sites were not revealed by this digest. It is noteworthy that we were successful in obtaining good quality data on T-31, which is a 54 residue tryptic glycopeptide carrying two N-glycans (see [fig_ref] Figure 7: Sequence of the C-terminal domain of the S-layer which was mapped in... [/fig_ref] and [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] whose size is not very different from the predicted value for monoglycosylated T-29. We think it is likely, therefore, that our failure to detect signals attributable to the T-29 glycopeptide is probably due to either or both Asn 1120 and Asn 1154 being glycosylated, in addition to Asn 1168 , thus moving this glycopeptide outside the observable m/z range of the glycoproteomics experiment. Irrespective of whether these two sites are indeed occupied, the glycosylation density in the C-terminal domain is quite remarkable, averaging one glycosylation site for each stretch of 30-40 residues.
Each of the observed glycosylation sites was found to be heterogeneously glycosylated with a family of glycans, the largest of which has a composition consistent with it having the sequence of the tribranched hexasaccharide found in cytochrome b 558/566 of S. acidocaldarius, [bib_ref] Cytochrome b 558/566 from the archaeon sulfolobus racidocaldarius has a unique Asn-linked..., Zähringer [/bib_ref]. The other members of the family appear to be biosynthetic precursors of this glycan. The most abundant is Man 1 GlcNAc 2 (see annotations on [fig_ref] Figure 4: Summed mass spectra recorded between 46 and 52 min from an on-line... [/fig_ref]. In addition, a nonextended GlcNAc is a minor component at some sites [fig_ref] Figure 6: Summed mass spectra of T24 recorded between 48 [/fig_ref] , and a pentasaccharide lacking one of the hexoses of the mature glycan was also observed [fig_ref] Figure 4: Summed mass spectra recorded between 46 and 52 min from an on-line... [/fig_ref]. Interestingly Man 2 GlcNAc 2 was only observed as a very minor component at a single site (T-34, [fig_ref] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from... [/fig_ref] , suggesting that either the second mannose is added after the 6-sulfoquinovose, or addition of the latter is rapid compared with the second mannosylation. It is not surprising that the S-layer and the cytochrome share a glycan sequence because conservation of N-glycosylation within a species appears to be characteristic of Archaea [bib_ref] Protein glycosylation in Archaea: sweet and extreme, Calo [/bib_ref]. Moreover, the presence of a family of biosynthetically related glycans is not unexpected since the S-layer of H. volcani exhibits a similar phenomenon [bib_ref] Haloferax volcanii AglB and AglD are involved in Nglycosylation of the S-layer..., Abu-Qarn [/bib_ref].
With the exception of the two halophiles described in the Introduction, whose glycans are linked via glucose, all known archaeal glycans are attached to Asn via GalNAc or GlcNAc. Interestingly, S. acidocaldarius is the only species characterized so far whose glycans are linked via chitobiose (GlcNAcβ1-4GlcNAc), the core disaccharide shared by the N-linked glycans of Eukarya. Moreover the tribranched topology of the Sulfolobus glycan is reminiscent of eukaryotic glycans which are usually multiantennary.
Sulfolobus spp. are developing into model organisms for studies on Eukarya-like mechanisms of transcription, translation, and cell division, and the vast amount of recently established genetic tools now make these organisms a prime choice to study these phenomena. The glycan Archaea structural information reported in this paper will facilitate the application of genetic tools to the elucidation of the N-glycosylation pathway in S. acidocaldarius. It will be very interesting to establish whether the commonalities in core structure between the glycans of S. acidocaldarius and those of Eukarya are mirrored in the biosynthetic pathways. Moreover, identification of the glycosylation enzymes of S. acidocaldarius could lead to interesting biotechnological applications.
[formula] M F P Y A K G N L V I A L T Y V A F S T K T L S S P I Q V V P V I T F Q L S G T F G A A L E T L T P L V N G A P V Q F P A L R I P E A T I P V Q T T S T G V G D S V T I I A A G I A S G T Y T V F P L L F L G A P G L S P S G T H V N F Q S A E Y V F A Q I Q F E Q P Y P L I S T V Y I K L V Q Q Q P T I V I Y E Y K L K P A N P L P V Q S L V P G G A F P V A S L V I V Y K G A P L N S I G Q L Q V Y V L L N P V S T G I G L A P Y A R V A S Y S F Y V P A A T Q T L T F I L Q A P Y E L Q Q V N L N L L [/formula]
[fig] Figure 1: (a) Structure of the 6-sulfoquinovose-containing tribranched hexasaccharide identified in cytochrome b 558/566 [/fig]
[fig] Figure 3: SDS PAGE gel of the S. acidocaldarius S-layer showing the band that was cut out for in-gel digestion. [/fig]
[fig] Figure 4: Summed mass spectra recorded between 46 and 52 min from an on-line nanoLC-ES-MS/MS analysis of an in-gel tryptic digest of the S-layer. The quadruply charged signals at m/z 1014.07, 1111.59 and 1151.09 are molecular ions of glycopeptides of sequence GAGVVEFLLTAYPYTGNITFAPPWFIAENVVK carrying the glycans shown in the annotations. Unassigned signals are molecular ions of peptides from elsewhere in the S-layer. [/fig]
[fig] Figure 5: MS/MS of m/z 1151.09 (a) and m/z 1014.07 (b) (see mass spectrum in Figure 4). Note the diagnostic fragment ions for HexNAc in both spectra and for HexNAc-QuiS in (a). The peptide sequence ions are identical in (a) and (b), and their assignments are shown in (c).Archaea [/fig]
[fig] Figure 6: Summed mass spectra of T24 recorded between 48.5 and 50.6 min from an on-line nanoLC-ES-MS/MS analysis of an in-gel tryptic digest of the S-layer. The triply charged signals at m/z 1256.38, 1378.09, 1507.28, and 1561.45 are molecular ions of glycopeptides of sequence LLNLNVQQLNNSILSVTYHDYVT-GETLTATTK carrying the glycans shown in the annotations. Unassigned signals are molecular ions of peptides from elsewhere in the S-layer.mixture of glycans that we found in the glycoproteomics experiments. The unusual quinovose sugar was not observed because sulfonated sugars are not recovered under the experimental conditions used for sugar analysis. [/fig]
[fig] Figure 7: Sequence of the C-terminal domain of the S-layer which was mapped in the glycoproteomics experiments. Consensus sites (in bold) which were shown to be glycosylated are underlined. [/fig]
[table] Table 1: Glycopeptides observed in LC-MS/MS analyses of proteolytic digests of the S-layer from S. acidocaldarius. Tryptic glycopeptides are designated T24, T26 and so forth based on the in silico digest shown inFigure 2. Two chymotryptic glycopeptides derived from the C-terminal domain of the S-layer are designated C-1 and C-2. [/table]
|
Evidence for impaired glucose metabolism in the striatum, obtained postmortem, from some subjects with schizophrenia
Studies using central nervous system tissue obtained postmortem suggest pathways involved in energy and metabolism contribute to the pathophysiology of schizophrenia; neuroimaging studies suggesting glucose metabolism is particularly affected in the striatum. To gain information on the status of pathways involved in glucose metabolism in the striatum, we measured levels of glucose, pyruvate, acetyl-CoA and lactate as well as the β subunit of pyruvate dehydrogenase, a rate limiting enzyme, in the postmortem tissue from subjects with schizophrenia and age/sex-matched controls. The subjects with schizophrenia were made up of two subgroups, which could be divided because they either had (muscarinic receptor deficit schizophrenia (MRDS)), or did not have (non-MRDS), a marked deficit in cortical muscarinic receptors. Compared to controls, levels of β subunit of pyruvate dehydrogenase were lower (Δ mean = − 20%) and levels of pyruvate (Δ mean = +47%) and lactate (Δ mean = +15%) were significantly higher in the striatum from subjects with schizophrenia. Notably, in subjects with non-MRDS, striatal levels of β subunit of pyruvate dehydrogenase were lower (Δ mean = − 29%), whereas levels of pyruvate (Δ mean = − 66%), acetyl-CoA (Δ mean = − 28%) and glucose (Δ mean = -27%) were higher, whereas levels of lactate (Δ mean = +17%) were higher in MRDS. Finally, discriminate analyses using levels the β subunit of pyruvate dehydrogenase and glucose, or better still, β subunit of pyruvate dehydrogenase and glucose in combination with pyruvate, lactate or acetyl-CoA could separate subjects with non-MRDS from controls with high levels of specificity (up to 93%) and selectivity (up to 91%). Our data show the benefit of being able to study defined subgroups within the syndrome of schizophrenia as such an approach has revealed that changes in glucose metabolism may be a significant contributor to the pathophysiology of non-MRDS.
# Introduction
Studies of gene expression in human central nervous system suggest that dysregulated metabolic pathways contribute the pathophysiology of the schizophrenia. [bib_ref] Gene expression profiling reveals alterations of specific metabolic pathways in schizophrenia, Middleton [/bib_ref] The notion that changes in metabolic pathways are disturbed in subjects with schizophrenia has gained further supported from studies of the human proteome 2 and metabolome 3 in CNS tissue from subjects with schizophrenia. Significantly, studies on the human CNS metabolome broadly implicate changes in glucose metabolism in the pathophysiology of schizophrenia. [bib_ref] Metabonomic studies of schizophrenia and psychotropic medications: focus on alterations in CNS..., Ma [/bib_ref] This is important because the CNS does not have extensive energy reserves in the form of glycogen and therefore relies heavily on producing its energy requirements from available glucose. [bib_ref] Sugar for the brain: the role of glucose in physiological and pathological..., Mergenthaler [/bib_ref] Thus, any change in glucose metabolism could affect nearly all cellular process in neurons and glia, thus impacting on processes including neurotransmission and rates of cell death. Moreover, changes in glucose metabolism in the CNS could particularly impact neuronal function because, on a per cell basis, neurons make the highest energy demand and therefore require ready access to glucose for metabolism. [bib_ref] Updated energy budgets for neural computation in the neocortex and cerebellum, Howarth [/bib_ref] As in any cell, neurons metabolized glucose through a number of steps, the end product of which is the generation of cellular energy in a number of forms including ATP, which can also act as a neurotransmitter in its own right. [bib_ref] Sugar for the brain: the role of glucose in physiological and pathological..., Mergenthaler [/bib_ref] Thus, any significant changes in glucose metabolism in neurons, in particular, would have profound effect on CNS function which could lead to disorders such as schizophrenia.
Two pathways that have a major role in glucose metabolism are the Embden-Meyerhof pathway and the tricarboxylic acid cycle (TCA).The Embden-Meyerhof pathway metabolize glucose to pyruvate with a net production of two ATP molecules and two molecules of reduced nicotinamide adenine dinucleotide (NADH), both of which represent energy currency for a cell. Pyruvate can then be converted to acetyl-CoA by the pyruvate dehydrogenase complex with acetyl-CoA being a substrate for the TCA cycle. When one molecule of substrate, acetyl-CoA, is processed through the TCA there is a release of electrons from NADH and reduced flavin adenine dinucleotide (FADH 2 ) to facilitate the production of 4 molecules of ATP. It has been reported that the activity of aconitase, α-ketoglutarate dehydrogenase complex and succinate thiokinase is lower in the dorsolateral prefrontal cortex from subjects with schizophrenia. [bib_ref] Abnormalities in the tricarboxylic acid (TCA) cycle in the brains of schizophrenia..., Bubber [/bib_ref] These data are significant because these enzymes catalyse the first three reactions of the TCA.In addition, a gene expression array study has reported lower levels of mRNA for succinate dehydrogenase complex subunit A and lactate dehydrogenase A in the hippocampus from subjects with schizophrenia. [bib_ref] Deficient hippocampal neuron expression of proteasome, ubiquitin, and mitochondrial genes in multiple..., Altar [/bib_ref] Succinate dehydrogenase is another enzyme in the TCA cycle, whereas lactate dehydrogenase can convert pyruvate to lactate, a product that does not enter the Embden-Meyerhof pathway.Hence, these data suggest that changes in glucose metabolism in the CNS from subjects with schizophrenia may be widespread across a number of regions.
Neuroimaging studies have been used to assess glucose metabolism across the human CNS and such studies have reported changes in glucose utilization in the CNS of subjects with schizophrenia. [bib_ref] Positron emission tomography studies of abnormal glucose metabolism in schizophrenia, Buchsbaum [/bib_ref] Significantly, changes in glucose utilization in the striatum in subjects with schizophrenia have been reported in a number of neuroimaging studies. [bib_ref] Positron emission tomography and subcortical glucose metabolism in schizophrenia, Resnick [/bib_ref] [bib_ref] Striatal metabolic rate and clinical response to neuroleptics in schizophrenia, Buchsbaum [/bib_ref] [bib_ref] Hypermetabolic pattern in frontal cortex and other brain regions in unmedicated schizophrenia..., Soyka [/bib_ref] [bib_ref] Cerebral glucose utilization and platelet mitochondrial complex I activity in schizophrenia: A..., Ben-Shachar [/bib_ref] [bib_ref] Differential brain glucose metabolic patterns in antipsychotic-naive first-episode schizophrenia with and without..., Horga [/bib_ref] These data suggest the striatum may be particularly affected by changes in glucose metabolism in subjects with schizophrenia. Despite such neuroimaging data, postmortem studies do not seem to have focussed on striatum to discover the mechanisms that might underlie changes in glucose metabolism in subjects with schizophrenia.
One approach to defining the mechanisms by which glucose metabolism is altered in the CNS from subjects with schizophrenia can be by measuring the activity, protein levels or levels of expression of enzymes in pathways such as the Embden-Meyerhof pathway or the TCA. However, another approach is to measure levels of individual metabolites as surrogate markers of changes in the activity of glucose utilizing metabolic pathways. Using this latter approach, lower levels of lactate have been reported in the anterior limb of the internal capsule from subjects with schizophrenia, [bib_ref] Metabolic abnormalities in fronto-striatal-thalamic white matter tracts in schizophrenia, Beasley [/bib_ref] but higher levels of that metabolite were measured in the cerebellum [bib_ref] Increased lactate levels and reduced ph in postmortem brains of schizophrenics: medication..., Halim [/bib_ref] and cortex 2 from subjects with the disorder. In addition, compared with controls, lower levels of pyruvate have been reported in the mediodorsal thalamus from subjects with schizophrenia. [bib_ref] Proteome analysis of the thalamus and cerebrospinal fluid reveals glycolysis dysfunction and..., Martins-De-Souza [/bib_ref] In aerobic conditions, which likely dominate most of the time in the CNS, lower levels of pyruvate could be indicative of an increase in processing of this substrate by the pyruvate dehydrogenase complex or a reduced production of pyruvate by the Embden-Meyerhof pathway. By contrast, the conversion of pyruvate to lactate and lactic assay is mostly associated with anaerobic metabolism, a process that is known to begin to dominate sometime after death. [bib_ref] Increased lactate levels and reduced ph in postmortem brains of schizophrenics: medication..., Halim [/bib_ref] To differentiate between these processes it is important to note that pyruvate dehydrogenase complex is the rate limiting step in the conversion of pyruvate to acetyl-CoA; hence, it would be expected that changes involving aerobic metabolism would be apparent because of changes in the activity or levels of the pyruvate dehydrogenase complex that would be accompanied by inverse changes in the enzymes substrate (pyruvate) and its product (acetyl-CoA) without changes in the anaerobic product (lactate; [fig_ref] Figure 1: Schematic diagram showing critical components glycolytic pathways controlling the conversion of glucose... [/fig_ref].
The pyruvate dehydrogenase complex is made up of three enzymes, pyruvate dehydrogenase, dihydrolipoyl transacetylase and dihydrolipoyl dehydrogenase. [bib_ref] The pyruvate dehydrogenase complex in cancer: An old metabolic gatekeeper regulated by..., Saunier [/bib_ref] Pyruvate dehydrogenase is a quadripartite complex made up of 2 α and 2 β subunits with each subunit having roles in regulating the activity of pyruvate dehydrogenase and the pyruvate dehydrogenase complex. Any change in the levels of either of the subunits of the pyruvate dehydrogenase complex would be expected to change the activity of the pyruvate dehydrogenase complex and hence levels of pyruvate and acetyl-CoA. Therefore, the report of changes in levels of pyruvate 17 in the CNS of subjects with schizophrenia would be consistent with changes in the metabolism of pyruvate by the pyruvate dehydrogenase complex.
The growing body of data from postmortem CNS studies, plus the data from neuroimaging studies and our expression microarray study, [bib_ref] Changed gene expression in subjects with schizophrenia and low cortical muscarinic M1..., Scarr [/bib_ref] underpinned our decision to measure levels of the β subunits of pyruvate dehydrogenase (PDHB) in the striatum of subjects with schizophrenia and age/sex-matched controls. In addition, to gain further insight into the extent of perturbation of glucose metabolism in the striatum subjects with schizophrenia we measured levels of glucose, acetyl-CoA, lactate and pyruvate [fig_ref] Figure 1: Schematic diagram showing critical components glycolytic pathways controlling the conversion of glucose... [/fig_ref]. We chose to focus our studies in the striatum because a number of neuroimaging studies have reported changes in glucose utilization in that CNS region in subjects with schizophrenia. [bib_ref] Positron emission tomography and subcortical glucose metabolism in schizophrenia, Resnick [/bib_ref] [bib_ref] Striatal metabolic rate and clinical response to neuroleptics in schizophrenia, Buchsbaum [/bib_ref] [bib_ref] Hypermetabolic pattern in frontal cortex and other brain regions in unmedicated schizophrenia..., Soyka [/bib_ref] [bib_ref] Cerebral glucose utilization and platelet mitochondrial complex I activity in schizophrenia: A..., Ben-Shachar [/bib_ref] [bib_ref] Differential brain glucose metabolic patterns in antipsychotic-naive first-episode schizophrenia with and without..., Horga [/bib_ref] It is now recognized that the diagnoses of schizophrenia defines a syndrome of disorders and that studying the syndrome as a whole is impeding the identification of the core pathophysiologies of the disorders within the syndrome. [bib_ref] Accelerating new knowledge in schizophrenia, Tamminga [/bib_ref] We have shown that it is possible to distinguish a subset of individuals within the disorder based on a marked loss of cortical muscarinic receptor binding; we have termed this subset as muscarinic receptor deficit schizophrenia (MRDS). Subjects with MRDS have been shown to have a widespread loss of cortical muscarinic receptors [bib_ref] Widespread decreases in cortical muscarinic receptors in a subset of people with..., Gibbons [/bib_ref] [bib_ref] An investigation of the factors that regulate muscarinic receptor expression in schizophrenia, Seo [/bib_ref] similar to what was reported in some individuals with the disorder in a neuroimaging study of muscarinic receptors in schizophrenia. [bib_ref] In vivo determination of muscarinic acetylcholine receptor availability in schizophrenia, Raedler [/bib_ref] Recently we have reported that subjects with MRDS also have decreased radioligand binding to striatal muscarinic receptors and higher levels of post-synaptic density protein 95, [bib_ref] Changes in cholinergic and glutamatergic markers in the striatum from a sub-set..., Dean [/bib_ref] showing changes in the molecular cytoarchitecture of the CNS in subjects with MRDS extends beyond the cortex. It has long been recognized that muscarinic receptors have been shown to act to modulate glucose uptake in the CNS [bib_ref] Reversal of scopolamineinduced amnesia and alterations in energy metabolism by the nootropic..., Piercey [/bib_ref] and that components of the glucose metabolism pathways, such as PDH, can modulate levels of muscarinic receptors. [bib_ref] Reduction of cortical muscarinic receptors after inhibition of pyruvate dehydrogenase complex (PDHc)..., Frölich [/bib_ref] Hence, given these recognized interactions between muscarinic receptors and energy and metabolism, we decided to measure levels of markers related to glucose metabolism in the striatum from subjects with schizophrenia and age-and sex-matched controls, with the schizophrenia cohort consisting of subjects with MRDS and subjects with no apparent loss of cortical muscarinic receptors (non-MRDS).
# Materials and methods
## Tissue collection
CNS was initially collected from subjects with a potential history of schizophrenia and subjects with no apparent history of psychiatric disorders who were matched for age and sex. CNS was only collected from people who had been seen alive up to 5 h prior to being found dead and where cadavers had been refrigerated within 5 h to ensure slowing of any autolysis of the CNS. [bib_ref] Brain protein preservation largely depends on the postmortem storage temperature: implications for..., Ferrer [/bib_ref] A postmortem assessment was completed by reviewing histories, discussions with relatives and treating clinicians which, for non-psychiatric controls, was to exclude any history of significant psychiatric symptoms. For each subject with history of schizophrenia relevant clinical and neuropsychopharmacological data were recorded. Subsequently, duration of illness (DI) was calculated as the time from first presentation to a psychiatric service until death and postmortem interval (PMI) was calculated as the time between witnessed death and autopsy or the midpoint between the subject being found and being last seen alive and autopsy. The final recorded dose of antipsychotic drug (FRADD) was noted and converted to chlorpromazine equivalents using algorithms as proposed in the literature, [bib_ref] Neuroleptic equivalence, Foster [/bib_ref] as was total lifetime exposure to such drugs (LEAP). The diagnosis of schizophrenia was agreed by consensus between two senior psychiatrists and the person completing the assessment according to DSM IV criteria using the Diagnostic Instrument for Brain Studies. [bib_ref] Confirmation of the diagnosis of schizophrenia after death using DSM-IV: a Victorian..., Roberts [/bib_ref] To ensure optimum tissue preservation each left CNS hemisphere was processed in a standardized manner to ensure the tissue was frozen to − 70°C within 30 min of autopsy. [bib_ref] The localisation and quantification of molecular changes in the human brain using..., Dean [/bib_ref] For each case, CNS pH was measured as an indicator of the quality of tissue preservation 31 as this measure is recognized as giving a good indication as to the biochemical integrity of tissue, which is often not related simply to PMI. [bib_ref] Human postmortem tissue: what quality markers matter?, Stan [/bib_ref] For this study, striatum was cut from frozen slices of tissue from subjects with schizophrenia and age-and sex-matched controls.
In this study, we divide subjects with schizophrenia into two subgroups based on levels of [ 3 H]pirenzepine binding in Brodmann's area 9. [bib_ref] Decreased cortical muscarinic receptors define a subgroup of subjects with schizophrenia, Scarr [/bib_ref] Thus, subjects with MRDS have levels of [ 3 H]pirenzepine binding in BA 9 of o90 fmol mg − 1 estimated tissue equivalents whilst non-MRDS have levels of 4120 fmol mg − 1 estimated tissue equivalents.
## Measurement of pdhb
Measurement of PDHB was by western blotting based on established methodologies involving an internal control (IC) but optimized using an anti-human PDHB antibody (Abcam, Melbourne, VIC, Australia; Cat# ab55574). Thus, striatum was homogenized into 10 mM Tris HCl (pH 7.4) containing 1% SDS and 1 mM sodium vanadate on ice using a Potter-Elvehjem grinding chamber and Teflon-coated pestle. The protein concentration of each homogenate was measured prior to the sample being frozen at-70°C until required. On the day of assay, homogenates were diluted in 0.5 M Tris HCl (pH 6.8) containing 40% SDS, 20% glycerol, 10% β-mecaptoethanol and 0.05% bromophenol blue to give a final 5 μg of protein to be loaded onto a 12% polyacrylamide mini-gel. Following resolving the proteins across the PAGE gel, proteins were transferred to a nitrocellulose membrane.
To measure the intensity of PDHB, which appeared as a 39.1 kDa immunogenic band, the nitrocellulose membrane containing the transferred proteins was blocked with 5% non-fat milk powder in Tris-buffered saline containing 0.1% Tween 20 (0.1% TBST). The blocked membranes were then exposed to 1/100 dilution of an anti-human PDHB antibody (Abcam, Cambridge, UK) for 1hr at r.t. on a rocking platform. Membranes were then washed with 0.1% TBST before being exposed to a goat antimouse IgG horse radish peroxidase conjugated antibody (DAKO, North Sydney, NSW, Australia; 1/2000 dilution; P0448) in 0.1% TBST for 1 h at r.t. on a rocking platform. Membranes were again washed with 0.1% TBST and then exposed to Supersignal West Pico Chemiluminescent Substrate (Termo Fisher, Scoresby, Australia) for 5 min at r.t. on a rocker. Excess solution was then drained and blotted off each membrane and the intensity of the chemiluminesence of the PDHB band measure using a Kodak Image Station 440CF.
It is clear that the notion of a reference gene, which is a gene whose expression does not vary during neurodevelopment, between CNS regions or between disorders of the CNS, is not tenable for studies of human CNS at the level of mRNA 34 or protein. [bib_ref] Total protein analysis as a reliable loading control for quantitative fluorescent Western..., Eaton [/bib_ref] As suggested, [bib_ref] Total protein analysis as a reliable loading control for quantitative fluorescent Western..., Eaton [/bib_ref] we have taken an approach to minimize the impact of variation in tissue processing as well as separating and measuring protein levels using Western blotting that does not use a reference protein as a loading control. Hence, de-identified tissue is allocated to the experimenter in batches that contain tissue from subjects with schizophrenia and their matched control. The experimenter processes each batch of tissue so that each Western blot contains an equal number of controls and subjects with schizophrenia; this process ensures the experimenter remains blind to diagnoses. The protein level in each sample is measured for a second time on the morning they are loaded onto a gel to ensure the protein levels loaded onto each gel are accurate. Great care is taken on loading each case, in duplicate, onto the loading gel and following protein transfer, the integrated level of the intensity of Ponceau stained protein bands in each lane is measured to ensure the amount of protein transferred is constant. Finally, to further control for any gel to gel variation an IC homogenate is prepared so that it can be run in every lane of two gels to establish both intra-and interblot variation for measuring the level of the protein of interest in the IC. [bib_ref] Regionally specific changes in levels of cortical S100beta in bipolar 1 disorder..., Dean [/bib_ref] The IC is run (in duplicate) on each subsequent gel containing the cases. These gels are exposed so that the sum intensity of the IC falls within the range established as the mean ± 2 s.d. of the intensities of the two master gels measurements. Finally, the results from all subjects were standardized by expressing them as a ratio of IC. Using this approach we have shown intraand intergel variation is o 10 and o15%, respectively, for measuring any protein of interest.
## Measurement of metabolites
All experimental measures were made in a way that ensured each experimenter was blind to diagnoses.
To measure pyruvate and lactate, striatal tissue was dissected and its wet weight recorded. [ 14 C]lactate was added to each tissue sample to give a final concentration of 1.0 nM prior to the tissue being homogenized, on ice, into 540 μl of 0.6% perchloric acid in a Potter-Elvehjem grinding chamber using a Teflon-coated pestle. [bib_ref] 1H NMR metabolite fingerprinting and metabolomic analysis of perchloric acid extracts from..., Kruger [/bib_ref] Each homogenate was centrifuged at 20 000 g for 30 min at 4°C, the supernatant was collected and adjusted to pH 7.0 with 1.5 M potassium carbonate. The neutralized homogenates were centrifuged at 20 000 g for 30 min at 4°C, the supernatant collected and the level of radioactivity in 20 μl of supernatant measured. The measured radioactivity was divided by the total radioactivity added to the tube before extraction to calculate the extraction efficiency for each sample. Homogenates were frozen at − 70°C until required.
To measure acetyl-CoA and glucose tissue was extracted according to the manufacture's instructions and hence, 40 μl [ 3 H]-acetyl-coA was added to 5 ml sample buffer containing 10% PCA. Approximately 100 mg of striatum was then homogenized, on ice, into 5 × volume of sample buffer containing the [ 3 H]-acetyl-coA in a Potter-Elvehjem grinding chamber using a Teflon-coated pestle. The homogenate was centrifuged at 12 500 g for 10 min at 4°C and the supernatant was collected and adjusted to pH 6-8 using 3 M K 2 CO 3 . After pH adjustment the homogenate was centrifuged at 12 500 g for 2 min at 4°C. The radioactivity in 3 × 20 μl aliquots of homogenate was counted and the measured radioactivity/total radioactivity added prior to extraction were calculated as the extraction efficiency and each homogenate. Each homogenate was stored at − 80°C until required.
Levels of pyruvate (Abnova, Taipei City, Taiwan), lactate (Sigma-Aldrich, Castle Hill, NSW, Australia), acetyl-CoA (Sigma-Aldrich) and glucose (Arbor Assays, Ann Arbor, MI, USA) were measured using commercially available kits using their recommended protocols. The measurement of pyruvate involved the oxidization of pyruvate by pyruvate oxidase, which causes a change in assay reagent color that can be measured at OD 570 nm. Levels of lactate were measure by a prepared enzymatic assay, which again caused a change in a reagent that could be measured at 570 nm. Acetyl-CoA was measured using a highly sensitive assay in which levels of the metabolite can be determined by a coupled enzyme assay reaction that results in a fluorometric read out (lex = 535/lem = 587 nm). Finally, glucose was measured by mixing samples with a Colorimetric Substrate containing horse radish peroxidase with the reaction being initiated by addition of glucose oxidase. This reaction produces hydrogen peroxide which, in the presence of horse radish peroxidase, reacts with the Colorimetric Substrate to convert the colorless substrate into a pink-colored product that can be read at 560 nm. Increasing levels of glucose cause a linear increase in color. All measurements for each metabolite were carried out using triplicate samples. Each measurement of pyruvate, lactate, acetyl-CoA and glucose was correct by multiplying by the relevant extraction efficiency. For all methodologies, interassay variation was o3% and intra-assay variation o10%.
## Glucose metabolism in schizophrenia b dean et al
## Statistics
The D'Agostino & Pearson omnibus normality test was used to analyze data distribution as this is the best approach when data sets are small. [bib_ref] A suggestion for using powerful and informative tests of normality, D'agostino [/bib_ref] Unpaired student t-tests and one-way analysis of variance with post hoc analyses using Dunnett's multiple comparisons test to identify the source of any variance were used to compare numeric demographic, tissue collection and treatment data between the diagnostic cohorts. Nonnumeric data relating to gender and suicide completion were compared using χ 2 -tests. Levels of PDHB and acetyl-CoA in the striatum from subjects with schizophrenia and age/sex-matched controls were normally distributed and therefore compared using an unpaired student t-test, levels of the remaining analytes were compared using the Mann-Whitney test. Levels of PDHB, acetyl-CoA and glucose in striatum from subjects with MRDS, non-MRDS and controls were normally distributed and were therefore compared using one-way analysis of variance with post hoc analyses using Dunnett's multiple comparisons test to identify the source of any variance. Levels of pyruvate and lactate across the three diagnostic groups were compared using Kruskal-Wallis and post hoc analyses using Dunnett's multiple comparisons test.
Relationships between age, PMI, DI, CNS pH, FRADD and LEAP and experimental data were determined using appropriate linear regression and the resulting coefficient of determination. The small sample sizes in this study meant that only strong relationships, r 2 ⩾ 0.49; ρo 0.05, warranted further consideration as potential confound. Where there were strong correlations non-experimental data were included as covariates when analyzing experimental data. Finally, discriminate analyses were used to determine the power of metabolic measurements to separate MRDS, non-MRDS and controls with sensitivity and specificity of the analyses determined as described previously. [bib_ref] Diagnostic tests. 1: Sensitivity and specificity, Altman [/bib_ref] In these analyses, diagnosis was used to define the groups, whereas the experimental measures were used as the predictors. The discrimination function was linear and no crossvalidation was performed.
Statistical analyses were completed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA) and Minitab 15 (Minitab, Sydney, NSW, Australia).
# Results
## Study cohorts
Initial study design was to measure markers of glucose metabolism in the striatum from 39 subjects with schizophrenia and 20 controls based on tissue availability; the cohort from subjects with schizophrenia containing 20 subjects with MRDS. As tissue was used for sequential study insufficient tissue was available for all measurements in all individuals and hence cohort sizes varied between studies (Supplementary [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref].
Pyruvate dehydrogenase-β PDHB was measured in the striatum from 39 subjects with schizophrenia that included 20 subjects with MRDS and 20 ageand sex-matched controls. The were no significant differences in [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref] ; similarly, there was no difference in gender frequency, age, CNS pH, PMI or brain weight between subjects with MRDS, non-MRDS and controls or between rates of suicide completion, DI, FRADD and LEAD between MRDS and non-MRDS. Compared with controls, levels of PDHB were lower (−20%) in the striatum from subjects with schizophrenia [fig_ref] Figure 2: Levels of pyruvate dehydrogenase subunit β [/fig_ref]. Notably, PDHB levels were lower in the striatum from subjects with in non-MRDS (−29%) but not MRDS. Levels of PDHB did not vary with gender (PDHB Ratio IC mean ± s.e.m.: male = 0.85 ± 0.07 vs female = 0.80 ± 0.07; P = 0.63) but there was a trend to lower levels of PDHB between suicide completers (0.68 ± 0.10) compared with those who died by other causes (0.85 ± 0.04; P = 0.07). The regression lines describing the relationships between PDHB and age, CNS pH, PMI, FRADD, LEAD, brain weight, pyruvate, lactate and acetyl-CoA did not deviate significantly from a slope = 0 [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref].
## Pyruvate
Pyruvate and lactate was measured in the striatum of 20 control subjects and 36 subjects with schizophrenia, which included 18 subjects with MRDS. Demographic and tissue collection data did not vary with diagnoses and rates of suicide, duration of illness or treatment data varied with diagnoses did not vary between MRDS and non-MRDS [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref].
Compared with controls, levels of pyruvate were higher (Δ median = 31%; Δ mean = 47%) in the striatum from subjects with schizophrenia [fig_ref] Figure 2: Levels of pyruvate dehydrogenase subunit β [/fig_ref]. The higher levels of pyruvate were in the striatum from non-MRDS subjects (Δ median = 34%; Δ mean = 66%), but not in the striatum from subjects with MRDS. Levels of pyruvate did not vary with gender (pyruvate (nmol g − 1 per wwt) mean ± s.e.m.: male = 0.74 ± 0.16 vs female = 0.95 ± 0.31; P = 0.56) or between suicide completers (0.88 ± 0.39) compared to those who died by other causes (0.76 ± 0.14; P = 0.71). In control subjects levels of pyruvate were related to levels of glucose whereas in schizophrenia levels of pyruvate were related to levels of lactate and glucose [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref]. The regression lines describing the relationships between pyruvate and age, CNS pH, PMI, FRADD, LEAD, and brain weight did not deviate significantly from a slope = 0.
Lactate Compared with controls, levels of lactate were higher (Δ median = 17%; Δ mean = 15%) in the striatum from subjects with schizophrenia [fig_ref] Figure 2: Levels of pyruvate dehydrogenase subunit β [/fig_ref]. Notably, higher levels of lactate appear to be limited to the striatum from subjects with MRDS subjects (Δ median = 23%; Δ mean = 17%), not non-MRDS. Levels of lactate did not vary with gender (lactate (mmol g − 1 per wwt) mean ± s.e. m.: male = 1.09 ± 0.05 vs female = 1.27 ± 0.08; P = 0.14) or between suicide completers (1.06 ± 0.05) compared with those who died by other causes (1.15 ± 0.06; P = 0.40). In controls levels of lactate were related to levels of acetyl-CoA whereas in schizophrenia lactate levels were related to levels of acetyl-CoA and glucose [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref]. There was a moderate relationship between CNS pH and lactate in both controls and subjects with schizophrenia but the regression line describing the relationships between lactate age, PMI, FRADD, LEAD, brain weight did not deviate significantly from a slope = 0 [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref].
Acetyl-CoA Levels of acetyl-CoA were measured in the striatum of 14 controls and 25 subjects with schizophrenia including 13 subjects with
## Mrds.
There was a trend towards a difference in CNS pH between subjects with schizophrenia and controls [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref]. There were no significant differences between other demographic, CNS-collection related or treatment variables with diagnoses. There were no significant differences in any demographic, CNS collection or treatment parameters between MRDS, non-MRDS and controls. Compared with controls, there was a strong trend to higher levels of acetyl-CoA (+14%) in the striatum from subjects with schizophrenia [fig_ref] Figure 2: Levels of pyruvate dehydrogenase subunit β [/fig_ref]. Further analysis showed that levels of acetyl-CoA were higher in the striatum of the non-MRDS subjects (+28%) but not in striatum from subjects with MRDS. Acetyl-CoA levels did not vary with gender (acetyl-CoA (nmol g − 1 per wwt): mean ± s.e.m.: male = 24 ± 0.19 vs female = 22 ± 0.96; P = 0.30) or between suicide completers (21 ± 1.5) compared with those who died by other causes (21 ± 1.5; P = 0.42). In schizophrenia, levels of acetyl-CoA were related to levels of glucose [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref]. The regression lines describing the relationships between acetyl-CoA and age, CNS pH, PMI, FRADD, LEAD and brain weight did not deviate significantly from a slope = 0.
Glucose Glucose levels were measured in the striatum of 13 controls and 25 subjects with schizophrenia including 14 subjects with MRDS. Compared with controls, there was a trend levels CNS pH being lower for subjects with schizophrenia, but there was no differences in other demographic, CNS-collection related or treatment variables with diagnoses [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref]. There were no significant differences in any demographic, CNS-collection related or treatment variables differing between subjects with MRDS, non-MRDS or controls.
Levels of glucose did not differ in the striatum from subjects schizophrenia compared with controls [fig_ref] Figure 2: Levels of pyruvate dehydrogenase subunit β [/fig_ref]. However, compared with controls, levels of glucose were higher in the striatum of the non-MRDS subjects (+27%) but not in striatum from subjects with MRDS. Levels of striatal glucose did not vary with gender (Glucose (mg g − 1 per wwt): mean ± s.e.m.: male = 0.24 ± 0.01 vs female = 0.26 ± 0.01; P = 0.48) or between suicide completers (0.23 ± 0.02) compared with those who died by other causes (0.24 ± 0.01; P = 0.54). The regression lines describing the relationships between glucose and age, CNS pH, PMI, FRADD, LEAD and brain weight did not deviate significantly from a slope = 0 [fig_ref] Table 1: Relationships levels of different analytes in human striatum and the same analytes... [/fig_ref].
Discriminate analyses Discriminate analyses of individual analytes did not separate subjects with schizophrenia, MRDS, non-MRDS or controls. By contrast, using striatal levels of PDHB and glucose separated subjects with non-MRDS from controls with a specificity of 93%, selectivity 83% [fig_ref] Figure 3: Levels of striatal pyruvate dehydrogenase subunit β [/fig_ref]. Using levels of PDHB, pyruvate and glucose [fig_ref] Figure 4: Levels of pyruvate dehydrogenase subunit β [/fig_ref] : specificity 93%, selectivity 82%) or PDHB, lactate and glucose [fig_ref] Figure 4: Levels of pyruvate dehydrogenase subunit β [/fig_ref] : specificity 93%, selectivity 83%) did not increase the separation between controls and non-MRDS. However, using levels of PDHB, acetyl-CoA and glucose gave a slight improvement in selectivity [fig_ref] Figure 4: Levels of pyruvate dehydrogenase subunit β [/fig_ref] : specificity 93%, selectivity 91%). More complex combinations of levels of metabolites did not improve separations between subjects with non-MRDS or controls.
# Discussion
This study presents the data that supports the hypothesis that abnormal striatal glucose metabolism is contributing to the pathophysiology of schizophrenia. This proportion comes from our findings of low levels of PDHB and higher levels of pyruvate, along with a trend to higher levels of acetyl-CoA, in the striatum from subjects with the disorder. These findings would be consistent with decreased activity of the TCA cycle and/or the Embden-Meyer pathway [fig_ref] Figure 1: Schematic diagram showing critical components glycolytic pathways controlling the conversion of glucose... [/fig_ref] in subjects with schizophrenia.
Our study design was novel as it was based on the premise that studying the pathophysiology of the syndrome of schizophrenia as a whole may be hindering the identification of the underpinning biochemical bases of individual disorders within the syndrome. [bib_ref] Accelerating new knowledge in schizophrenia, Tamminga [/bib_ref] Our data would support that hypothesis because they show that changes glucose, pyruvate, PDHB and acetyl-CoA are limited to subjects with non-MRDS, whereas changes in lactate occur in MRDS. Significantly, using different combinations of measures of glucose metabolisms in discriminate analyses allows the separation of subjects with non-MRDS, but not MRDS, from controls with a high degree of specificity and selectivity. These data argue that changes in striatal glucose metabolism may be a major contributor to the genesis of symptoms associated with subjects with non-MRDS. As changes in markers of glucose metabolism are only detectable in subjects with non-MRDS, but not MRDS who have markedly lower levels of muscarinic receptors in their cortex [bib_ref] Decreased cortical muscarinic receptors define a subgroup of subjects with schizophrenia, Scarr [/bib_ref] and striatum, [bib_ref] Changes in cholinergic and glutamatergic markers in the striatum from a sub-set..., Dean [/bib_ref] it may be significant that some data suggests there is bi-direction links between the activity and/or levels of muscarinic receptors and levels of glucose metabolism. [bib_ref] Reversal of scopolamineinduced amnesia and alterations in energy metabolism by the nootropic..., Piercey [/bib_ref] [bib_ref] Reduction of cortical muscarinic receptors after inhibition of pyruvate dehydrogenase complex (PDHc)..., Frölich [/bib_ref] Hence, our data may indicate that low levels of CNS muscarinic receptors may be protective against changes in glucose metabolism in the subset of subjects with MRDS.
Glucose metabolism is a source of energy for all cells in the CNS but neurons are particularly dependent on glucose as they have little or no glycogen stores that can be mobilized as an alternate source of energy. [bib_ref] Neurons have an active glycogen metabolism that contributes to tolerance to hypoxia, Saez [/bib_ref] Moreover, it has been recently suggested that neuronal activity may be a highly glucose dependent. [bib_ref] Direct neuronal glucose uptake heralds activity-dependent increases in cerebral metabolism, Lundgaard [/bib_ref] Our data using homogenates of human striatum does not allow any conclusions to be proposed on which cells may have increased levels of glucose in subjects with non-MRDS. It will therefore be important to determine which cells in the striatum have decreased levels of PDHB as low levels of the enzyme would appear to be an important measurable marker that could be used in immunohistochemistry studies to identify which cells may have altered levels of glucose metabolism in subjects with non-MRDS. In addition, in using cell homogenates we cannot distinguish whether the high levels of striatal glucose in non-MRDS are intra-or extracellular. If the high levels of glucose are in the extracellular milieu this would suggest there is an inability of striatal cells to take up glucose which would be consistent with neuroimaging studies in schizophrenia. [bib_ref] Positron emission tomography studies of abnormal glucose metabolism in schizophrenia, Buchsbaum [/bib_ref] [bib_ref] Positron emission tomography and subcortical glucose metabolism in schizophrenia, Resnick [/bib_ref] [bib_ref] Striatal metabolic rate and clinical response to neuroleptics in schizophrenia, Buchsbaum [/bib_ref] [bib_ref] Hypermetabolic pattern in frontal cortex and other brain regions in unmedicated schizophrenia..., Soyka [/bib_ref] [bib_ref] Cerebral glucose utilization and platelet mitochondrial complex I activity in schizophrenia: A..., Ben-Shachar [/bib_ref] [bib_ref] Differential brain glucose metabolic patterns in antipsychotic-naive first-episode schizophrenia with and without..., Horga [/bib_ref] Glucose uptake by the cells in the striatum is dependent on a number of glucose transporters and therefore determining whether there is a deficiency in glucose transporters in striatum from subjects with non-MRDS may shed light on whether these subjects could have lost the ability to internalize glucose into cells.
As in any study using subjects with schizophrenia treated with antipsychotic drugs, it is possible that changes in glucose metabolism may have resulted from such treatments. Against this argument is the finding that subjects with MRDS and non-MRDS appear to have had similar antipsychotic drug treatment regimes. [bib_ref] Decreased cortical muscarinic receptors define a subgroup of subjects with schizophrenia, Scarr [/bib_ref] This would suggest the change in glucose metabolism in one of these two groups is not likely to be simply an effect of antipsychotic drug treatment. Moreover, subjects with schizophrenia in this study were treated with typical antipsychotic drugs. It is therefore significant that it has been shown that treating rats with chlorpromazine suppresses glycolysis in the CNS, but not the periphery. [bib_ref] Suppression of glycolysis in rat brain in vivo by chlorpromazine, reserpine, and..., Gey [/bib_ref] In addition, it has been shown that treatment with haloperidol lowered whole brain glucose and pyruvate, caused no change in levels of acetyl-CoA 43 and increased levels of brain lactate. [bib_ref] Increased lactate levels and reduced ph in postmortem brains of schizophrenics: medication..., Halim [/bib_ref] [bib_ref] The effect of the combination of lithium and haloperidol on brain intermediary..., Guynn [/bib_ref] Finally, it has been reported that haloperidol did not change striatal metabolic rates measured using [18F]fluorodeoxyglucose in haloperidol-responsive patients with schizophrenia. [bib_ref] Effect of a haloperidol challenge on regional brain metabolism in neurolepticresponsive and..., Bartlett [/bib_ref] These data argue that treatment with typical antipsychotic drugs per se do not alter CNS glucose metabolism.
A limitation of our studies is that, owing to limited tissue availability, not all markers could be measured in every case meaning cohort sizes become relatively small. That been said, our smallest cohort sizes are equal to those in many other postmortem CNS studies. In addition, the findings on lactate need to be treated with caution as levels of that metabolite were related to measures of CNS pH. As pH is a measure of CNS tissue preservation [bib_ref] Human postmortem tissue: what quality markers matter?, Stan [/bib_ref] we cannot rule out that the higher levels of lactate in the striatum from subjects with schizophrenia is a postmortem artefact, particularly in the light of reports that lactate in human blood [bib_ref] Biochemistry changes that occur after death: potential markers for determining post-mortem interval, Donaldson [/bib_ref] and cerebellum [bib_ref] Increased lactate levels and reduced ph in postmortem brains of schizophrenics: medication..., Halim [/bib_ref] increase after death and these increases can be related to falling pH. It has been suggested that CNS glucose increases with PMI in mouse CNS, [bib_ref] Evaluation of the postmortem glucose and glycogen levels in hepatic, renal, muscle,..., Gumus [/bib_ref] which is in marked contrast to our findings where there was no relationship between levels of glucose in the striatum from subjects with schizophrenia (r 2 = 0.010, P = 0.74) or controls (r 2 = 0.013, P = 0.59) and PMI. This difference may be due to changes after death in mice CNS not modeling changes in human CNS due to the different sizes of the CNS between the two species. Using rat, a rapid rise of CNS acetyl-CoA has been reported minutes after death after which the levels of CNS acetyl-CoA appear to stabilize. [bib_ref] Acetyl coenzyme A in the brain: radioenzymatic determination, use of microwaves, and..., Ricny [/bib_ref] Although these data could suggest that the higher levels of acetyl-CoA we report in the striatum from non-MRDS could be a postmortem artefact, for that to occur levels of acetyl-CoA must have increased to a greater degree in a subset of subjects with schizophrenia (non-MRDS) compared with another subset of subjects with schizophrenia (MRDS) and controls. It is difficult to provide a convincing hypothesis as to how this selected change in acetyl-CoA could have occurred. Hence, other than the measure of lactate, our data would suggest that the changes we report in striatum from subjects with non-MRDS are unlikely to be solely due to some postmortem artefact.
We have previously reported that subjects with MRDS have been shown to have a widespread loss of cortical muscarinic receptors, [bib_ref] Widespread decreases in cortical muscarinic receptors in a subset of people with..., Gibbons [/bib_ref] [bib_ref] An investigation of the factors that regulate muscarinic receptor expression in schizophrenia, Seo [/bib_ref] have disruptions in muscarinic M1 receptor signaling, [bib_ref] Altered M(1) muscarinic acetylcholine receptor (CHRM1)-Galpha(q/11) coupling in a schizophrenia endophenotype, Salah-Uddin [/bib_ref] have a distinct pattern of muscarinic M1 receptor gene promoter methylation and higher levels of a microRNA that acts to reduce receptor expression, [bib_ref] Decreased cortical muscarinic M1 receptors in schizophrenia are associated with changes in..., Scarr [/bib_ref] and have a reduced response to a muscarinic M1-receptor-positive allosteric modulator. [bib_ref] Changes in BQCA allosteric modulation of [3H] NMS binding to human cortex..., Dean [/bib_ref] In addition, we have reported that subjects with MRDS also have decreased radioligand binding to striatal muscarinic receptors and higher levels of post-synaptic density protein 95, [bib_ref] Changes in cholinergic and glutamatergic markers in the striatum from a sub-set..., Dean [/bib_ref] showing changes in the molecular cytoarchitecture of the CNS in subjects with MRDS extends beyond the cortex. Thus, to our knowledge, this study is different in that we have shown the majority of changes in markers of glucose metabolism are present in the striatum from subjects with non-MRDS.
In conclusion, we have shown a change in levels of a number of markers in the striatum of subjects with schizophrenia that are consistent with changes in glucose metabolism, and that these changes are restricted to subjects with non-MRDS. These findings reinforce the argument that studying subgroups of subjects within the syndrome of schizophrenia does lead to being able to better understand the pathophysiology of the different disorders in the syndrome. These systematic changes allow the differentiation of subjects with non-MRDS from controls with a high degree of selectivity and specificity suggesting changes in glucose metabolism is an important component of the pathophysiology of non-MRDS but not MRDS. Recently, there has been a major focus on potential abnormalities in peripheral glucose metabolism in subjects with schizophrenia that are either caused or worsened by treatment with certain antipsychotic drugs. [bib_ref] Investigation of molecular serum profiles associated with predisposition to antipsychotic-induced weight gain, Schwarz [/bib_ref] Our data would suggest that a continued focus on changes in glucose metabolism in the CNS from subjects with schizophrenia will also be important in better understanding the pathophysiology of the disorder and whether treatment with the antipsychotic drugs peripheral glucose homeostasis may be detrimental to CNS homeostasis.
[fig] Figure 1: Schematic diagram showing critical components glycolytic pathways controlling the conversion of glucose to acetyl-CoA or lactate. FAD, flavin adenine dinucleotide; NAD, nicotinamide adenine dinucleotide; TPP, thiamine pyrophosphate. [/fig]
[fig] Figure 2: Levels of pyruvate dehydrogenase subunit β (a: mean ± s.e.m.), pyruvate (b: median ± IQR), lactate (c: median ± IQR), acetyl-CoA (d: mean ± s.e.m.) and glucose (e: mean ± s.e.m.) in the striatum from controls, subjects with schizophrenia, subjects with schizophrenia and a deficit in cortical muscarinic receptors (MRDS) and subjects with schizophrenia without a deficit in cortical muscarinic receptors (non-MRDS). IQR, interquartile range. MDRS, muscarinic receptor deficit schizophrenia.Glucose metabolism in schizophrenia B Dean et al gender frequency, age, CNS pH, PMI or brain weight between the subjects with schizophrenia and the controls (Supplementary [/fig]
[fig] Figure 3: Levels of striatal pyruvate dehydrogenase subunit β (PDHB) and glucose in non-MRDS subjects and controls. IC, internal control; MRDS, muscarinic receptor deficit schizophrenia. [/fig]
[fig] Figure 4: Levels of pyruvate dehydrogenase subunit β (PDHB) compared to levels of pyruvate and glucose (a), lactate and glucose (b) and acetyl-CoA and glucose (c) in the striatum from subjects with non-MRDS and controls. MRDS, muscarinic receptor deficit schizophrenia. [/fig]
[table] Table 1: Relationships levels of different analytes in human striatum and the same analytes and demographic, CNS collection and pharmacological data from the cases from whom the striatum was collected Abbreviations: CNS, central nervous system; DI, duration of illness; FRADD, final recorded drug dose expressed as chlorpromazine equivalents per day; LEAD, lifetime exposure to antipsychotic drugs expressed as chlorpromazine equivalents per year × 10 − 3 ; PDHB, pyruvate dehydrogenase; PMI, postmortem interval. Relationships where the linear regression line deviated significantly from zero are shown in bold. [/table]
|
Compound Library Screening Identified Cardiac Glycoside Digitoxin as an Effective Growth Inhibitor of Gefitinib-Resistant Non-Small Cell Lung Cancer via Downregulation of α-Tubulin and Inhibition of Microtubule Formation
Non-small-cell lung cancer (NSCLC) dominates over 85% of all lung cancer cases. Epidermal growth factor receptor (EGFR) activating mutation is a common situation in NSCLC. In the clinic, molecular-targeting with Gefitinib as a tyrosine kinase inhibitor (TKI) for EGFR downstream signaling is initially effective. However, drug resistance frequently happens due to additional mutation on EGFR, such as substitution from threonine to methionine at amino acid position 790 (T790M). In this study, we screened a traditional Chinese medicine (TCM) compound library consisting of 800 single compounds in TKI-resistance NSCLC H1975 cells, which contains substitutions from leucine to arginine at amino acid 858 (L858R) and T790M mutation on EGFR. Attractively, among these compounds there are 24 compounds CC 50 of which was less than 2.5 µM were identified. We have further investigated the mechanism of the most effective one, Digitoxin. It showed a significantly cytotoxic effect in H1975 cells by causing G2 phase arrest, also remarkably activated 5 1 adenosine monophosphate-activated protein kinase (AMPK). Moreover, we first proved that Digitoxin suppressed microtubule formation through decreasing α-tubulin. Therefore, it confirmed that Digitoxin effectively depressed the growth of TKI-resistance NSCLC H1975 cells by inhibiting microtubule polymerization and inducing cell cycle arrest.
# Introduction
According to the World Health Organization (WHO), lung cancer is becoming the major reason of cancer-caused death, and accounts for 20% of all cancer death cases and causes millions of deaths per 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay was used to detect cell inhibition rate of 800 candidate compounds on H1975 cells and A549 cells which harbors EGFR wild type (WT). All 800 compounds were tested in both cell lines for 72 h as preliminary screening at the concentration range of 0, 2.5, 5 and 10 µM and only 24 compounds showed CC 50 values less than 2.5 µM in both cell lines, which were shortlisted in ascending order in [fig_ref] Table 1: CC 50 values of twenty-four shortlisted candidate compounds in H1975 and A549... [/fig_ref]. As shown in [fig_ref] Table 1: CC 50 values of twenty-four shortlisted candidate compounds in H1975 and A549... [/fig_ref] , Digitoxin has the highest cytotoxicity in H1975 cells, whose CC 50 value was 0.19˘0.06 µM. These data implied that low dose of Digitoxin strongly effected on cells regardless of EGFR type, suggesting although Digitoxin had no selectivity for EGFR wild type and mutated NSCLC cells, is still useful in killing Gefitinib-resistance NSCLC cells. We further determined the cytotoxic effect of Digitoxin on normal lung fibroblast CCD-19Lu cells. Surprisingly, we found that the CC 50 value of Digitoxin in H1975 cells was more than 25-fold lower than that of CCD-19Lu cells, which suggested that Digitoxin has strong inhibition selectivity in NSCLC cells [fig_ref] Figure 1: Cytotoxicy of Digitoxin [/fig_ref]. In our result [fig_ref] Figure 1: Cytotoxicy of Digitoxin [/fig_ref] , the EC 50 value of Digitoxin was 0.78 µM, demonstrating that Digitoxin was an effective Na + /K + -ATPase inhibitor, which was consistent with previous studies [bib_ref] On the reversibility of binding of cardiotonic steroids to a partially purified..., Yoda [/bib_ref] [bib_ref] Influence of 16 beta formylation on Na, K-ATPase inhibition by cardiac glycosides, De Pover [/bib_ref]. that Digitoxin has strong inhibition selectivity in NSCLC cells [fig_ref] Figure 1: Cytotoxicy of Digitoxin [/fig_ref]. In our result [fig_ref] Figure 1: Cytotoxicy of Digitoxin [/fig_ref] , the EC50 value of Digitoxin was 0.78 μM, demonstrating that Digitoxin was an effective Na + /K + -ATPase inhibitor, which was consistent with previous studies [bib_ref] On the reversibility of binding of cardiotonic steroids to a partially purified..., Yoda [/bib_ref] [bib_ref] Influence of 16 beta formylation on Na, K-ATPase inhibition by cardiac glycosides, De Pover [/bib_ref].
## Digitoxin induced cell cycle arrest in h1975
The mechanism of the anti-cancer effect of Digitoxin on H1975 has not yet been defined. To determine the mechanism of Digitoxin, we examined whether Digitoxin has a cell cycle arrest effect in H1975 cells. Cells were treated with the indicated concentrations of Digitoxin for 12 and 24 h, and stained with Propidium iodide (PI). The samples were analyzed by flow cytometry. The results showed that after 12 h treatment, the percentage of cells in G0/G1 phase slightly decreased while increased in G2/M phase [fig_ref] Figure 2: Digitoxin induced cell cycle arrest in H1975 cells [/fig_ref]. Moreover, a more apparent trend was found after treatment for 24 h. Digitoxin-treated cells exhibited a 15% of increase in G2/M phase accompanied by a significant reduction in G0/G1 phase at high dose [fig_ref] Figure 2: Digitoxin induced cell cycle arrest in H1975 cells [/fig_ref]. These results indicated that Digitoxin-induced inhibition mechanism is likely due to causing G2/M cell cycle arrest in H1975 cells in dose dependent manners.
## Digitoxin induced cell cycle arrest in h1975
The mechanism of the anti-cancer effect of Digitoxin on H1975 has not yet been defined. To determine the mechanism of Digitoxin, we examined whether Digitoxin has a cell cycle arrest effect in H1975 cells. Cells were treated with the indicated concentrations of Digitoxin for 12 and 24 h, and stained with Propidium iodide (PI). The samples were analyzed by flow cytometry. The results showed that after 12 h treatment, the percentage of cells in G0/G1 phase slightly decreased while increased in G2/M phase [fig_ref] Figure 2: Digitoxin induced cell cycle arrest in H1975 cells [/fig_ref]. Moreover, a more apparent trend was found after treatment for 24 h. Digitoxin-treated cells exhibited a 15% of increase in G2/M phase accompanied by a significant reduction in G0/G1 phase at high dose [fig_ref] Figure 2: Digitoxin induced cell cycle arrest in H1975 cells [/fig_ref]. These results indicated that Digitoxin-induced inhibition mechanism is likely due to causing G2/M cell cycle arrest in H1975 cells in dose dependent manners. All data was presented as mean ± SEM (n = 3, * p < 0.05, ** p < 0.01, *** p < 0.001).
## Effects of digitoxin on cell cycle regulatory proteins in h1975
To further clarify the underlying mechanism of Digitoxin in inducing cell cycle arrest in H1975, we examined the effect of Digitoxin on the expression of several cell cycle regulatory proteins. As shown in [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref] ,B, Digitoxin significantly decreased the protein content of cyclin B1 (CCNB1) and cyclin A1 (CCNA1) resulting in G2/M phase arrest, which were consistent with the results of cell cycle arrest data detected by flow cytometry.
We also determined the effect of Digitoxin on modulating p21, p27 and phosphor-AMPK (p-AMPK) proteins. Western blotting results showed that Digitoxin remarkably down-regulated the expression of p21 and p27, both of which have been defined as cyclin-dependent kinase inhibitors (CKIs). It has been published that overexpression of p21 and p27 suppressed DNA synthesis, which arrested cell cycle in G1 phase and thus inhibited cell proliferation [bib_ref] The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent..., Harper [/bib_ref] [bib_ref] Inhibition of CDK2 activity in vivo by an associated 20K regulatory subunit, Gu [/bib_ref]. Therefore, the inhibition effect of Digitoxin was not caused by induction of G1-arrest. As the key regulator of cellular metabolism and energy homeostasis, AMPK was also associated with cell cycle and cell proliferation. Our results showed that Digitoxin was found to activate p-AMPK in a dose-dependent manner [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref].
## Effects of digitoxin on cell cycle regulatory proteins in h1975
To further clarify the underlying mechanism of Digitoxin in inducing cell cycle arrest in H1975, we examined the effect of Digitoxin on the expression of several cell cycle regulatory proteins. As shown in [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref] ,B, Digitoxin significantly decreased the protein content of cyclin B1 (CCNB1) and cyclin A1 (CCNA1) resulting in G2/M phase arrest, which were consistent with the results of cell cycle arrest data detected by flow cytometry.
We also determined the effect of Digitoxin on modulating p21, p27 and phosphor-AMPK (p-AMPK) proteins. Western blotting results showed that Digitoxin remarkably down-regulated the expression of p21 and p27, both of which have been defined as cyclin-dependent kinase inhibitors (CKIs). It has been published that overexpression of p21 and p27 suppressed DNA synthesis, which arrested cell cycle in G1 phase and thus inhibited cell proliferation [bib_ref] The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent..., Harper [/bib_ref] [bib_ref] Inhibition of CDK2 activity in vivo by an associated 20K regulatory subunit, Gu [/bib_ref]. Therefore, the inhibition effect of Digitoxin was not caused by induction of G1-arrest. As the key regulator of cellular metabolism and energy homeostasis, AMPK was also associated with cell cycle and cell proliferation. Our results showed that Digitoxin was found to activate p-AMPK in a dose-dependent manner [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref]. Altogether, our data demonstrated that Digitoxin affected the cell cycle arrest in G2/M phase by altering the expression of CCNB1 and CCNA1 and may be also associated with AMPK activation.
To verify the inhibition of cell growth by Digitoxin in H1975, we performed western blotting to examine the expression level of c-Myc. It has been reported that c-Myc plays a major role in Myc family to target DNA transactivation and promote cell proliferation [bib_ref] Evaluation of the antitumor effects of c-Myc-Max heterodimerization inhibitor 100258-F4 in ovarian..., Wang [/bib_ref]. As shown in [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref] , by 24 h, c-Myc expression was significant reduced by Digitoxin. Loss of c-Myc suggested that Digitoxin has suppressive effect on the proliferation of H1975 cells. Altogether, our data demonstrated that Digitoxin affected the cell cycle arrest in G2/M phase by altering the expression of CCNB1 and CCNA1 and may be also associated with AMPK activation.
To verify the inhibition of cell growth by Digitoxin in H1975, we performed western blotting to examine the expression level of c-Myc. It has been reported that c-Myc plays a major role in Myc family to target DNA transactivation and promote cell proliferation [bib_ref] Evaluation of the antitumor effects of c-Myc-Max heterodimerization inhibitor 100258-F4 in ovarian..., Wang [/bib_ref]. As shown in [fig_ref] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells [/fig_ref] , by 24 h, c-Myc expression was significant reduced by Digitoxin. Loss of c-Myc suggested that Digitoxin has suppressive effect on the proliferation of H1975 cells. The protein of p-AMPK were determined by western blotting, and GAPDH was considered as a loading control; (B,D) Statistical analysis of CCNB1, CCNA1, p21, p27, c-Myc and p-AMPK. All data was presented as mean ± SEM (n = 3, * p < 0.05, *** p < 0.001). At least three independent experiments were performed.
## Digitoxin inhibited microtubule formation
Microtubule is known as a key regulator of cell cycle progression. Therefore, we determined the effect of Digitoxin on α-tubulin protein content and intensity. As measured by western blotting and flow cytometer, α-tubulin was remarkably reduced [fig_ref] Figure 4: Digitoxin suppressed tubulin polymerization in H1975 cells [/fig_ref]. Furthermore, the intensity of α-tubulin was visualized by immunofluorescence. As shown in [fig_ref] Figure 4: Digitoxin suppressed tubulin polymerization in H1975 cells [/fig_ref] , the control group displayed a normal microtubule network that fully extended to the whole cell, while Digitoxin-treated cells weaken the microtubules incorporation and showed a severely disruption and disorganized structure. Collectively, these results suggested that Digitoxin might target on α-tubulin, suppress microtubule formation, prevent mitotic spindle to break chromosome and consequently interfere with mitosis.
## Effect of digitoxin on cell proliferation
Since Digitoxin is able to modulate cell cycle arrest, we try to determine whether it regulates cell proliferation. We used colony formation assay to detect the role of Digitoxin on clonogenic survival. As shown in [fig_ref] Figure 5: Digitoxin inhibited cell proliferation [/fig_ref] , Digitoxin treatment significantly attenuated cells in a dose-dependent manner. When cells were treated with 0.5 μM of Digitoxin, the percentage of cell number remarkably decreased by 95% approximately compared with control cells, suggesting that Digitoxin reduced cell proliferation and colony growth. The protein of p-AMPK were determined by western blotting, and GAPDH was considered as a loading control; (B,D) Statistical analysis of CCNB1, CCNA1, p21, p27, c-Myc and p-AMPK. All data was presented as mean˘SEM (n = 3, * p < 0.05, *** p < 0.001). At least three independent experiments were performed.
## Digitoxin inhibited microtubule formation
Microtubule is known as a key regulator of cell cycle progression. Therefore, we determined the effect of Digitoxin on α-tubulin protein content and intensity. As measured by western blotting and flow cytometer, α-tubulin was remarkably reduced [fig_ref] Figure 4: Digitoxin suppressed tubulin polymerization in H1975 cells [/fig_ref]. Furthermore, the intensity of α-tubulin was visualized by immunofluorescence. As shown in [fig_ref] Figure 4: Digitoxin suppressed tubulin polymerization in H1975 cells [/fig_ref] , the control group displayed a normal microtubule network that fully extended to the whole cell, while Digitoxin-treated cells weaken the microtubules incorporation and showed a severely disruption and disorganized structure. Collectively, these results suggested that Digitoxin might target on α-tubulin, suppress microtubule formation, prevent mitotic spindle to break chromosome and consequently interfere with mitosis.
## Effect of digitoxin on cell proliferation
Since Digitoxin is able to modulate cell cycle arrest, we try to determine whether it regulates cell proliferation. We used colony formation assay to detect the role of Digitoxin on clonogenic survival. As shown in [fig_ref] Figure 5: Digitoxin inhibited cell proliferation [/fig_ref] , Digitoxin treatment significantly attenuated cells in a dose-dependent manner. When cells were treated with 0.5 µM of Digitoxin, the percentage of cell number remarkably decreased by 95% approximately compared with control cells, suggesting that Digitoxin reduced cell proliferation and colony growth. (E) Cells were treated with or without Digitoxin for 24 h and processed with immunofluorescence staining. Images were captured by Delta Vision Live Cell Imaging System (20×, 60× objective magnification). All data was presented as mean ± SEM (n = 3, * p < 0.05, *** p < 0.001).
# Discussion
TCM has been widely used in treatment of complex diseases, including diabetes, neurodegenerative disease and cancer [bib_ref] Sustained antidiabetic effects of a berberine-containing Chinese herbal medicine through regulation of..., Zhao [/bib_ref] [bib_ref] A systems pharmacology approach to decipher the mechanism of danggui-shaoyao-san decoction for..., Luo [/bib_ref] [bib_ref] Impact of arsenic trioxide in the treatment of acute promyelocytic leukemia, Lengfelder [/bib_ref]. As such, it could be used as important resources for new drug discovery [bib_ref] From traditional Chinese medicine to rational cancer therapy, Efferth [/bib_ref].
Digitalis containing cardiac glycosides has been used for the treatment of heart diseases for over 200 years [bib_ref] Contemporary use of digoxin in the management of cardiovascular disorders, Gheorghiade [/bib_ref]. An ancient Chinese remedy that employs an extract of Bufo toad secretions contains several cardiac glycosides and is still being used today for managing cancerous conditions [bib_ref] The cooperative interaction of two different signaling pathways in response to bufalin..., Watabe [/bib_ref] [bib_ref] Effects of bufalin and cinobufagin on the proliferation of androgen dependent and..., Yeh [/bib_ref] [bib_ref] Anti-tumor activities and apoptosis-regulated mechanisms of bufalin on the orthotopic transplantation tumor..., Han [/bib_ref]. Additionally, in 1980s, digitalis was firstly found to have anti-cancer effect in breast cancer [bib_ref] Cardiac glycosides and breast cancer, revisited. New Engl, Stenkvist [/bib_ref]. Since then, more and more researchers put their efforts on investigating the anti-cancer mechanism of digitalis including Ouabain, Digoxin, Digitoxin and so on. For instance, it has been reported that ouabain inhibited cell proliferation and induced cell death by activation of c-Jun N-terminal kinase (JNK) pathway [bib_ref] Cardiac glycoside ouabain induces autophagic cell death in non-small cell lung cancer..., Trenti [/bib_ref]. Digitoxin was described as a target of Na + /K + -ATPase pump inhibitor in the last century. In addition, its cytotoxicity was reported colsely related inhibition effect on Na + /K + -ATPase [bib_ref] On the reversibility of binding of cardiotonic steroids to a partially purified..., Yoda [/bib_ref] [bib_ref] Influence of 16 beta formylation on Na, K-ATPase inhibition by cardiac glycosides, De Pover [/bib_ref]. It indicated that Na + /K + -ATPase not only signals the ion pump, but also participates as a signaling transducer in several downstream pathways, including EGFR, mitogen-activated protein kinases (MAPK), and phosphoinositide 3-kinase (PI3K) pathways [bib_ref] Inhibition of epidermal growth factor signaling by the cardiac glycoside ouabain in..., Wolle [/bib_ref]. It has been investigated that digitalis-induced activation of EGFR-MAPK pathway by binding with Na + /K + -ATPase. It led to cell cycle arrest by resulting in up-regulation of p21, also known as CKI1 [bib_ref] Digitalis-induced signaling by Na + /K + -ATPase in human breast cancer..., Kometiani [/bib_ref]. In addition, digoxin has been reported for suppressing cancer through inhibiting multiple proto-oncogene tyrosine-protein kinase (Src)-related signaling pathways in NSCLC cell lines [bib_ref] Digoxin Suppresses Tumor Malignancy through Inhibiting Multiple Src-Related Signaling Pathways in Non-Small..., Lin [/bib_ref]. Digitoxin, a member of digitalis family, showed that it can increase level of intracellular calcium by inhibiting the sodium-potassium adenosine triphosphatase (Na + /K + -ATPase) complex in myocardial cells [bib_ref] Cardiac glycosides stimulate Ca 2+ increases and apoptosis in androgen-independent, metastatic human..., Mcconkey [/bib_ref]. It has been reported that a low dose of Digitoxin resulted in activation of cell apoptosis pathway in H460 cells, but no less sensitive on primary human lung epithelial cells and non-tumorigenic human lung epithelial cells [bib_ref] Digitoxin and a synthetic monosaccharide analog inhibit cell viability in lung cancer..., Elbaz [/bib_ref]. Nevertheless, the mechanism of it in the treatment of cancer has not been fully defined yet.
Although plenty of reports about the anti-cancer function of digitalis have been published, few articles about EGFR T790M/L858R were on digoxin [bib_ref] Digoxin Suppresses Tumor Malignancy through Inhibiting Multiple Src-Related Signaling Pathways in Non-Small..., Lin [/bib_ref] and ouabain [bib_ref] Cardiac glycoside ouabain induces autophagic cell death in non-small cell lung cancer..., Trenti [/bib_ref]. In NSCLC, many studies were mostly focused on A549 [bib_ref] Cardiac glycoside ouabain induces autophagic cell death in non-small cell lung cancer..., Trenti [/bib_ref] [bib_ref] Digoxin Suppresses Tumor Malignancy through Inhibiting Multiple Src-Related Signaling Pathways in Non-Small..., Lin [/bib_ref] [bib_ref] Cardenolide aglycones inhibit tumor necrosis factor alpha-induced expression of intercellular adhesion molecule-1..., Okina [/bib_ref] [bib_ref] Digoxin downregulates NDRG1 and VEGF through the inhibition of HIF-1alpha under hypoxic..., Wei [/bib_ref] [bib_ref] Src mediates extracellular signal-regulated kinase 1/2 activation and autophagic cell death induced..., Wang [/bib_ref] [bib_ref] Inactivation of Src-to-ezrin pathway: A possible mechanism in the ouabain-mediated inhibition of..., Shin [/bib_ref] [bib_ref] Inhibition of cell migration by ouabain in the A549 human lung cancer..., Liu [/bib_ref] [bib_ref] Odoroside A and ouabain inhibit Na + /K + -ATPase and prevent..., Takada [/bib_ref] [bib_ref] Cardiac glycosides induce autophagy in human non-small cell lung cancer cells through..., Wang [/bib_ref] [bib_ref] Reduction of doxorubicin cytotoxicity by ouabain: Correlation with topoisomerase-induced DNA strand breakage..., Lawrence [/bib_ref] [bib_ref] Influence of ouabain on cell inactivation by irradiation, Verheye-Dua [/bib_ref] and H460 cells with WT EGFR [bib_ref] Digoxin Suppresses Tumor Malignancy through Inhibiting Multiple Src-Related Signaling Pathways in Non-Small..., Lin [/bib_ref] [bib_ref] Digitoxin and a synthetic monosaccharide analog inhibit cell viability in lung cancer..., Elbaz [/bib_ref] [bib_ref] Digitoxin and its synthetic analog MonoD have potent antiproliferative effects on lung..., Yakisich [/bib_ref] [bib_ref] Ouabain mediates integrin switch in human lung cancer cells, Ninsontia [/bib_ref]. To our knowledge, no research report has addressed the effect of Digitoxin on Gefitinib resistant EGFR T790M/L858R mutated H1975 cells. In our study, we have applied cell-based method to large-scale screening of a TCM compound library consisting of 800 compounds and used NSCLC H1975 and A549 cell lines as screening tools, we have shortlisted 24 candidates with CC 50 value less than 2.5 µM, one of the compounds, Digitoxin, was found effective in Gefitinib resistance H1975 with EGFR T790M/L858R double mutation.
By using quantitative flow cytometry analysis, we found that Digitoxin remarkably increased the percentage of cells at G2/M phase coherent with a decrease at G1 phase. Moreover, these results indicated that Digitoxin arrested cell cycle at G2/M phase consequently inhibited mitosis. As a promising target, regulation of cell cycle checkpoints significantly contributed to the treatment of cancer [bib_ref] Cell cycle regulation in the G1 phase: A promising target for the..., Owa [/bib_ref]. In eukaryotic cells, cell cycle evolved an integrated protective mechanism, which regulated by a group of cyclin, cyclin-dependent kinases (CDK), and CKIs [bib_ref] The history and future of targeting cyclin-dependent kinases in cancer therapy, Asghar [/bib_ref] [bib_ref] Regulation of the DNA damage response by cyclin-dependent kinases, Trovesi [/bib_ref] [bib_ref] Cyclin dependent kinase inhibitors, Harper [/bib_ref]. Besides that, the cell cycle checkpoint is a cell survival monitoring system, which inspects DNA damage and defective spindle formation, controlling gene expression and maintaining genomic stability [bib_ref] BRCA1: Cell cycle checkpoint, genetic instability, DNA damage response and cancer evolution, Deng [/bib_ref] [bib_ref] Checkpoint kinase 1 in DNA damage response and cell cycle regulation, Patil [/bib_ref]. Whether cells can pass the G2/M checkpoint depends on the level of CCNB1, which is the initial activator and pivotal regulator, leading to regulation and activation of CDK1/CCNB1 complexes. Phosphorylation of these complexes further enforced cell cycle through the checkpoint and facilitate transition from G2 phase into mitosis [bib_ref] Chk1-induced CCNB1 overexpression promotes cell proliferation and tumor growth in human colorectal..., Fang [/bib_ref] [bib_ref] Cyclin B1 overexpression induces cell death independent of mitotic arrest, Eichhorn [/bib_ref]. Therefore, restraining the expression of CCNB1 arrested cell cycle at G2/M phase prevents cell cycle progression. Yuan J and others performed the experiments to reduce CCNB1 expression by RNA interference, which caused inhibition of cell proliferation by arresting cells at G2 phase, suggesting that the expression level of CCNB1 played an important role in initial regulation of mitosis [bib_ref] Cyclin B1 overexpression induces cell death independent of mitotic arrest, Eichhorn [/bib_ref]. Besides, different from CCNB1, CCNA1 functions in S and G2/M phases, which is due to its two different phosphorylated substrates, CDK1 and CDK2, subsequently conducing tumor cells proliferation. Our data demonstrated that Digitoxin arrested the H1975 cell cycle. We further observed that Digitoxin significantly down-regulated CCNB1 and CCNA1, which could explain our flow cytometry results of G2 phase cell cycle arrest. By western blotting, we also found that both p21 and p27 were suppressed by Digitoxin.
In addition, activation of AMPK provides an inhibition effect on CDK, p21 and p27 [bib_ref] AMP-activated protein kinase (AMPK) beyond metabolism: A novel genomic stress sensor participating..., Sanli [/bib_ref]. It has been reported that AMPK effected ataxia telangiectasia mutated (ATM) signals to mediate p53, p21 and p27, consequently targeted to cell cycle checkpoints [bib_ref] The AMPK signalling pathway coordinates cell growth, autophagy and metabolism, Mihaylova [/bib_ref] [bib_ref] LKB1 and AMP-activated protein kinase control of mTOR signalling and growth, Shaw [/bib_ref] [bib_ref] The protein kinase landscape of the DNA damage response, Bensimon [/bib_ref] [bib_ref] Metformin inhibits growth and enhances radiation response of non-small cell lung cancer..., Storozhuk [/bib_ref]. The activity of AMPK also varies at different stages of the cell cycle, it progressively increases from G1 phase to G2/M phase, reaching the peak at G2/M, and then decreases when cells entering G1/S stage [bib_ref] AMP-activated protein kinase (AMPK) beyond metabolism: A novel genomic stress sensor participating..., Sanli [/bib_ref] [bib_ref] Energy-dependent regulation of cell structure by AMP-activated protein kinase, Lee [/bib_ref]. Our study observed that Digitoxin induced cell cycle arrest at G2/M phase, which could be a combined mechanistic effect of cell cycle regulators and AMPK activation to maintain cell cycle at G2/M. Furthermore, overexpression of AMPK is also associated with a lack of mammalian target of rapamycin (mTOR) level, resulting in suppression of cell growth, proliferation and protein synthesis [bib_ref] LKB1 and AMP-activated protein kinase control of mTOR signalling and growth, Shaw [/bib_ref] [bib_ref] From growth signal integration to cancer, diabetes and ageing, Zoncu [/bib_ref] [bib_ref] Metformin: Taking away the candy for cancer?, Jalving [/bib_ref]. Thus, it may explain the colony formation suppression effect by Digitoxin.
Moreover, AMPK is also defined as a member of microtubule affinity-regulating kinase (MARK)/PAR kinase subfamily, leading to phosphorylation of microtubule-associated proteins (MAPs), which controls microtubule dynamics [bib_ref] AMPK attenuates microtubule proliferation in cardiac hypertrophy, Fassett [/bib_ref]. The microtubule is a key component of the cytoskeleton, formed by the polymerization of a dimer of two globular proteins, alpha and beta tubulin [bib_ref] Microtubule formation in vitro in solutions containing low calcium concentrations, Weisenberg [/bib_ref]. As structural elements, microtubules support cell stabilization and function as mitotic spindle in eukaryotic cells to segregate their chromosomes during cell division [bib_ref] Structure and function of polarity-inducing kinase family MARK/Par-1 within the branch of..., Marx [/bib_ref] [bib_ref] Cellular stress responses: A balancing act, Chen [/bib_ref] [bib_ref] Dynamic instability of microtubule growth, Mitchison [/bib_ref] [bib_ref] Posttranslational acetylation of alpha-tubulin constrains protofilament number in native microtubules, Cueva [/bib_ref]. By α-tubulin tyrosination and acetylation, microtubules maintain a dynamic balance [bib_ref] Tubulin tyrosination is a major factor affecting the recruitment of CAP-Gly proteins..., Peris [/bib_ref] [bib_ref] Motor-dependent microtubule disassembly driven by tubulin tyrosination, Peris [/bib_ref] [bib_ref] Reactive oxygen species, AMP-activated protein kinase, and the transcription cofactor p300 regulate..., Mackeh [/bib_ref] [bib_ref] Cytoskeletal role in the contractile dysfunction of hypertrophied myocardium, Tsutsui [/bib_ref] , loss of microtubule also leads to cell cycle arrest. Our studies showed that Digitoxin-induced G2 phase arrest is associated with AMPK activation and loss of α-tubulin and microtubles.
Despite the fact that Digitoxin had strong cytotoxic effect on H1975, it showed no selectively between A549. Similar to our findings, it has been published that two derivatives of the digitalis, digoxin and ouabain, also induced moderate G2/M arrest in A549 and H460, by activation of AMPK-mTOR pathway [bib_ref] Cardiac glycosides induce autophagy in human non-small cell lung cancer cells through..., Wang [/bib_ref]. It seems that digitalis families could influence the cell cycle both in the presence or absence of the EGFR mutant. Nevertheless, since cell mitotic, migration and growth strongly relied on microtubule, nowadays, anti-microtubule polymerization is emerging as a new anti-cancer therapy. For example, a novel microtubule-targeting agent, 7-deazahypoxanthines, showed statistically significant tumor size reduction in a colon cancer mouse model [bib_ref] Novel Microtubule-Targeting 7-Deazahypoxanthines Derived from Marine Alkaloid Rigidins with Potent in Vitro..., Medellin [/bib_ref]. Denning et al. also refereed anti-tubulin drugs as life-saving chemotherapeutics that can kill cancer cells by stabilizing or disrupting microtubules [bib_ref] Anti-tubulins DEPendably induce apoptosis, Denning [/bib_ref]. Our novel discovery of Digitoxin as a promising anti-tubulin formation agent on Gefitinib resistance NSCLC has opened new anti-genfitnib resistance intervention option.
# Materials and methods
# Materials
The natural product compound library consisting of 800 single purified compounds were purchased from MicroSource Discovery Systems/Topscience (Shanghai, China). Digitoxin powder was purchased from Sigma Aldrich (St. Louis, MI, USA). MTT powder and Dimethyl sulfoxide (DMSO) was purchased from Acros Organics (Morris Plains, NJ, USA). Adenosine 5 1 -Triphosphatase was purchased from Sigma Aldrich. Radioimmunoprecipitation (RIPA) lysis buffer (10ˆ) and the primary antibodies of CCNB1, CCNA1, p21, p27, c-Myc and GAPDH were purchased from Cell Signaling Technology (Danvers, MA, USA). The primary antibodies of phospho-AMPK, α-tubulin were purchased from Santa Cruz (Dallas, TX, USA). The secondary antibodies of anti-rabbit and anti-mouse were purchased from Odyssey (Belfast, ME, USA). Fluorescein-conjugated goat anti-rabbit and mouse anti-bodies were purchased from Odyssey (Belfast, ME, USA) and Invitrogen (Waltham, MA, USA). Propidium iodide (PI) staining kit was purchased from BD Biosciences (San Jose, CA, USA). RNase A was purchased from Sigma Aldrich.Ten percent fetal bovine serum (FBS), 100 U/mL penicillin and 100 µg/mL streptomycin were purchased from Gibco (Oklahoma, ME, USA). A complete mini, EDTA-free tablet was from Roche (Mannheim, Germany). DCTM protein assay kit was purchased from Bio-Rad (Hercules, CA, USA). Nitrocellulose (NC) membrane was purchased from GE Healthcare (Waukesha, WI, USA). Crystal violet was purchased from Amresco (Solon, OH, USA).
## Cell culture
H1975, A549 NSCLC cell lines, and CCD-19Lu the normal lung fibroblast cells of lung, all of these cell lines were purchased from ATCC. H1975 and A549 cell lines were cultivated with RPMI 1640 medium supplemented with 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin. The CCD-19Lu cells were cultivated with MEM medium supplemented with 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin. All cells were cultivated at 37˝C in a humidified atmosphere of 5% CO 2 .
## Mtt assay
Cells were seeded with 3ˆ10 3 cells/well in a 96-well plate and allowed to adhere after overnight incubation. Cells were treated with various concentrations of compounds (0, 0.625, 1.25, 2.5, 5 µM) and with vehicle control, DMSO, for 72 h. Ten µL of MTT solution were added to each well and incubated at 37˝C for 4 h. Then, 100 µL of resolved solution (10% SDS and 0.1 mM HCL) was added to each well and incubated at 37˝C for 4 h to solubilize the formazan crystals. Absorbance of plates was measured at 570 nm (absorbance) and 650 nm (reference) with Tecan microplate reader (Morrisville, NC, USA). Cell viability (%) was calculated by percentage of the absorbance of the treated group divided by control group. It was expressed as = [OD of treated group/OD of control group]ˆ100%, and the viability of control group was considered as 100%.
## Na + /k + atpase enzymatic activity assay
The enzymatic activity of Na + /K + ATPase was measured as the rate of release of inorganic phosphate during ATP hydrolysis. Indicated concentrations (0, 0.25, 0.5, 1 µM) of Digitoxin were mixed with 67.5 µL Tris HCl Buffer, which contains Ethylenediaminetetraacetic Acid and Magnesium Chloride and made as testing buffer. Ten µL Na + /K + ATPase and 2.5 µL KCl/NaCl were gently mixed with the testing buffer and incubated for 30 min at 37˝C. After that, 5 µL ATP was added and the reaction was started by incubated again for 15 min at 37˝C. After the final incubating, spin the mixture carefully, followed by centrifugation at 1000 rpm for 3 min. Transferred 50 µL supernatant of each sample to a 96-well plate with 100 µL Taussky-Shorr Reagent to terminate reaction. Absorbance of plates was measured at 660 nm with Tecan microplate reader.
# Cell cycle analysis
H1975 cells were plated with 2.0ˆ10 5 cells/well at a 6-well plate and cultured overnight for attachment. Cells were treated with Digitoxin at 0, 0.0625, 0.125, 0.25, 0.5 µM for 12 h and 24 h respectively. After treatment, all cells were harvested by trypsinization, and collected by centrifugation. After removing all suspension, cells were washed by PBS. Cells pellets were re-suspended in 70% ethanol at 4˝C for 30 min. Cells were centrifuged at 1000 rpm for 5 min to remove all the ethanol. Each cell pellet was re-suspended in 500 µL PI staining solution at 37˝C for 30 min in dark. Then, cells were washed in PBS twice. Cells were re-suspended in 300 µL 1ˆbinding buffer and transferred to the flow cytometer (BD FACS Aria III).
# Western blot analysis
H1975 cells were plated with 2.0ˆ10 5 cells/well at a 6-well plate and cultured overnight for attachment. Cells were lysed in 1ˆRIPA lysis buffer with proteinase inhibitor and phosphatase inhibitors added, and were scraped off from the plate by a plastic cell scraper. All the lysate was mixed well and kept in ice then transferred to a new tube. The lysate was centrifuged at 4˝C 12,000 rpm for 5 min. After centrifugation, the suspension with protein lysate was kept on ice. Protein concentration was quantitatively measured by DCTM protein assay kit, the supernatant was transferred into a new tube and mixed with 5ˆloading buffer. Each sample was boiled at 100˝C for 5 min. Twenty five µg of each protein samples were loaded into the well of a 10% SDS-PAGE gel with one lane of 3 µL protein molecular weight marker. The gel was run for 20 min at 80 V for stacking, and then added to 120 V for protein separation. After separation, the proteins from the gel were transferred to a NC membrane for 2 h at 300 mA, the membrane was blocked with 5% non-fat milk diluted with 1ˆTBST (0.1% Tween 20 in Tris-buffered saline) at room temperature for 1 h and washed with 1ˆTBST for three times. Membranes were incubated with primary antibodies at 1:1000 dilution at 4˝C overnight. After washing the membrane 3 times with 1ˆTBST, the membranes were incubated with secondary antibodies at 1:10,000 dilutions for 1 h at room temperature. GAPDH was used as endogenous loading control for normalization. The signal intensity of the membranes was detected by LI-COR Odyssey scanner (Belfast, ME, USA).
## Immunofluorescence flow cytometry
To identify the quantity of α-tubulin caused by Digitoxin, H1975 cells were cultured in 6-well plate, each well of 1.5ˆ10 5 cells. After exposure to Digitoxin of 24 h. the cells were collected by centrifugation and washed once with PBS. After being fixed with 4% paraformaldehyde in PBS, the cells were permeabilized by ice-cold 100% methanol for 30 min on ice. Centrifuged after having been washed by incubation buffer (0.5 g Bovine Serum Albumin (BSA) in 100 mL 1ˆPBS). α-tubulin was stained with α-tubulin antibody overnight at 4˝C. Washed by centrifugation in 2-3 mL incubation buffer. Resuspend cells in fluorescein isothiocyanate (FITC)-conjugated secondary antibody, diluted in incubation buffer at 1:250 dilution. Incubate for 1 h at room temperature in dark. Cells were washed by centrifugation in 2-3 mL incubation buffer. Resuspend cells in PBS and measured by flow cytometer.
## Indirect immunofluorescence microscopy
H1975 cells were grown on glass coverslips in 6-well plate, each well of 1.5ˆ10 5 cells. After exposure to Digitoxin, cells were washed once with PBS and then fixed with 4% paraformaldehyde in PBS for 15 min. The fixed cells were washed twice with PBS and then permeabilized with methanol for 2 min. Then rehydrated with PBS and incubated with α-tubulin (1:500) primary antibody in 5% bovine serum albumin at 4˝C overnight. Excess antibodies were removed by multiple washings with PBS. Added conjugated anti-mouse IgG (1:500) antibody as secondary antibody. After incubation for 2 h at room temperature, cells were stained by hoechst (1:10,000) for 10 min. Remove the hoechst washed with PBS twice, coverslips were mounted onto microscope slides with fluor save regent. Images were captured with Delta Vision Live Cell Imaging System by 20ˆ, 60ˆobjective respectively.
## Colony formation assay
H1975 cells were seeded at a density of 5ˆ10 3 cells/well in a 6-well plate, and cultured overnight for attachment. Exposed on Digitoxin with concentration (0, 0.03125, 0.0625, 0.125, 0.25, 0.5 µM) for 7 days. Then it was changed medium at every 72 h. At the 10th day, the colonies were washed by PBS once, and fixed with 4% PFA. The colonies were washed by PBS twice, stained with crystal violet solution for 20 min, and washed. The number of colonies formed were counted in each group.
# Statistical analysis
All data represent mean values of at least three independent experiments and were expressed as mean˘SEM. The statistical significant differences were analyzed by one-way ANOVA followed by Bonferroni for comparison tests, using Graph Prism Version 6.0 software (GraphPad Software, Inc., San Diego, CA, USA). * p < 0.05, ** p < 0.01, *** p < 0.001 were considered as significant.
# Conclusions
In conclusion, by using drug screening, in vitro functional assays and mechanism studies, we discovered that Digitoxin can lead to G2 phase arrest in H1975, a Gefitinib-resistant NSCLC cells with EGFR double mutations. We innovatively found that Digitoxin could directly suppress microtubule formation by reducing α-tubulin. Finally, we considered that Digitoxin has potential as the anti-cancer therapy with multifunction.
[fig] Figure 1: Cytotoxicy of Digitoxin. (A) Chemical structure of Digitoxin; (B) MTT assay results of Digitoxin on H1975 cells, A549 cells, and CCD-19Lu cells after 72 h treatment, respectively; (C) In vitro enzymatic assay of Na + /K + -ATPase; (D) SI values of H1975 cells, A549 cells, and CCD-19Lu cells respectively. All data were presented as mean ± SEM (n = 4, ** p < 0.01, *** p < 0.001) vs. vehicle control. [/fig]
[fig] Figure 2: Digitoxin induced cell cycle arrest in H1975 cells. (A) H1975 cells were treated with Digitoxin at different concentrations (0, 0.0625, 0.125, 0.25, 0.5 μM) for 12 h. Cells were stained with PI and cell cycle arrest was detected by flow cytometry; (B) Cells were stained and collected for cell cycle arrest assay at 24 h after treatment with Digitoxin in the indicated concentrations; (C) Statistical analysis of cell cycle distribution in 12 h; (D) Statistical analysis of cell cycle distribution in 24 h. [/fig]
[fig] Figure 3: Digitoxin significantly regulated cell cycle-related proteins in H1975 cells. (A) H1975 cells were treated with Digitoxin at different concentrations (0, 0.0625, 0.125, 0.25, 0.5 μM) for 24 h. Protein levels of CCNB1, CCNA1, p21, p27, c-Myc and GAPDH by western blotting; (C) [/fig]
[fig] Figure 4: Digitoxin suppressed tubulin polymerization in H1975 cells. (A) H1975 cells were treated with Digitoxin at different concentrations (0, 0.0625, 0.125, 0.25, 0.5 μM) for 24 h. Western blotting was used to detect the protein content of α-tubulin; (B) H1975 cells were treated with Digitoxin for different concentration as indicated, and then cells were incubated with primary α-tubulin antibodies and stained with anti-mouse-FITC secondary antibodies. The fluorescence intensity was analyzed by flow cytometry; (C) Statistical analysis of α-tubulin; (D) Statistical analysis of the relative decrease in α-tubulin intensity;(E) Cells were treated with or without Digitoxin for 24 h and processed with immunofluorescence staining. Images were captured by Delta Vision Live Cell Imaging System (20×, 60× objective magnification). All data was presented as mean ± SEM (n = 3, * p < 0.05, *** p < 0.001). [/fig]
[fig] Figure 5: Digitoxin inhibited cell proliferation. (A) Data were shown as representative photomicrographs of colony formation assay after treatment with Digitoxin at different concentration (0.03125, 0.0625, 0.125, 0.25, 0.5 μM); (B) Statistical analysis of colony formation assay. Statistical analysis was conducted using one-way analysis of variance (ANOVA). All data was presented as mean ± SEM (n = 3, ** p < 0.01). [/fig]
[table] Table 1: CC 50 values of twenty-four shortlisted candidate compounds in H1975 and A549 cell lines. [/table]
|
The Role of Brassica Bioactives on Human Health: Are We Studying It the Right Way?
Brassica vegetables and their components, the glucosinolates, have been suggested as good candidates as dietary coadjutants to improve health in non-communicable diseases (NCDs). Different preclinical and clinical studies have been performed in the last decade; however, some concerns have been posed on the lack of established and standardized protocols. The different concentration of bioactive compounds used, time of intervention or sample size, and the lack of blinding are some factors that may influence the studies' outcomes. This review aims to analyze the critical points of the studies performed with Brassica-related biomolecules and propose some bases for future trials in order to avoid biases.
# Introduction
Nowadays, the incidence of non-communicable diseases (NCDs), including obesity, diabetes, cancer, and other chronic conditions, is increasing and showing high mortality indexes worldwide (https://www.who.int/nmh/topics/es/). These diseases are multifactorial, but it has been described that lifestyle, such as smoking or sedentary life, contributes to the prevalence of the NCDs. As a result, different prevention strategies have been developed, mainly the promotion of healthy habits, recommending avoiding a sedentary life, quitting smoking and the consumption of alcoholic drinks, and keeping a healthy diet. To emphasize the importance of these strategies, it has been proven that these habits have an important role in the prevention of cardiovascular diseases and type-II diabetes.
Since one of the main points of intervention for a healthy lifestyle is diet, global health recommendations are always encouraging the message of eating a minimum of five portions of fruits and vegetables daily. Vegetables are not only a natural source of amino acids and minerals, but also of phytochemicals. In particular, cruciferous vegetables, like broccoli, cabbage, red radish, or Brussels sprouts, are consumed worldwide not only because of their culinary value being popular in many countries with plates being part of the culture legacy, but also because of their diverse content in phytochemicals with high health-promoting benefits. Between them vitamin C, polyphenols, and minerals can be found. However, one of the most relevant biomolecules are glucosinolates (GSLs), mainly found in plant species and varieties of the order Brassicales, mainly Brassicaceae as the most Another abundant GSL present in cruciferous vegetables is glucobrassicin (GB), whose main degradation product is ascorbigen. However, its corresponding indole-3-carbinol (I3C) and its condensation product 3,3-diindolylmethane (DIM)have been the main focus of medical research because they were readily commercially available. Glucobrassicin was discovered as the precursors of the already known ascorbigenI3C and DIM were found to be formed in vitro, but not in vivo. From the very fast formation of ascorbigen at plant pH, it is well established that I3C is neither an intermediate in the in vivo formation of ascorbigen, which must be due to reaction with ascorbic acid with an earlier intermediate in the glucobrassicin breakdown. Early authors (e.g., Bjeldanes, in the 1980s) suggested indoles to be cancer protecting; later authors noted Janus properties, being either protectors or carcinogens depending on the timing of the carcinogen and the indole, as reviewed by, e.g., Holst and Williamsonand Agerbirk et al.. Although several studies showed that DIM interferes with diverse signal transduction pathways implied in tumorigenic and inflammatory processes, such as AKT kinase, phosphoinositide 3 kinase (PI3K), NFκB pathway, or EGFR/ERK, the exact interaction has not been yet elucidated. Furthermore, it has been studied that the growth and expansion of certain cancer types, as colorectal, are promoted by the presence of pro-inflammatory interleukins, like TNFα and IL-6. These molecules activate pathways regulated by NFκB and transcription factor STAT3, inducing and maintaining in time a pro-inflammatory microenvironment. In this way, carcinogenic cell proliferation, tumor invasion, angiogenesis, and immunosuppression are promoted. The work of Zou et al.showed that when human ovarian cancer cells (SKOV3 and A27809) were treated with DIM, a downregulation in STAT3 and a subsequent inhibition in cell adhesion and invasion was observed.
All these studies performed in vitro suggest that cruciferous vegetables mainly from Brassica spp. (either as foods or ingredients rich in bioactive substances) are good candidates as dietary coadjutants to improve human health in NCDs. As a consequence, some clinical and pre-clinical studies have been developed during the last six years. Epidemiological studies have linked the intake of GSLs with the risk of coronary heart disease or type 2 diabetes. Nevertheless, this information from large cohorts is usually based in food frequency questionnaires, making it difficult to know the specific effects of the GSLs in the intake. On the other hand, promising anti-cancer results have been obtained from interventional studies but, due to the reduced sample size, cannot be interpreted as a generalization. In general, the studies analyzed in our review did not present an established and standardized protocol, differing in the concentrations of the component, time of the intervention, and sample size. Moreover, not all of them presented a double-blind design. For all that, this review aims to analyze the critical points of the studies performed with Brassica-related biomolecules with the aim to establish a basis for future trials and avoid biases.
# Methodology
In the present work we pursued the objective of critically reviewing the clinical studies in which ingredients and bioactive compounds from cruciferous vegetables (Brassicaceae family) had been used with different pathologies. Based on the anti-cancer properties attributed to GSLs and isothiocyanates (ITCs), we opted to search according to these pathologies since they were the most studied in the past decades. Additionally, one of the main target pathways of the SFN in the organism is the Nrf2-Keap1-ARE system, with a wide role in the development and progression of chronic diseases (e.g., cancers, respiratory problems), and, finally, based on their anti-inflammatory characteristics we opted also to review the more recent research data on the effects of the consumption of GSLs in the development and management of metabolic diseases, mediated by systemic inflammatory conditions. Diverse bibliographical searches were carried out in databases including PubMed, Scopus, and SpringerLink using the following terms and keywords: cancer, metabolic disease, respiratory tract disease, sulforaphane, and glucosinolates. In order to make a more concise selection, the results were limited to clinical trials. From the obtained records, articles including cruciferous foods and derived products were the selected items, and we reduced the timeframe to the last years, from 2012 to the present as well. This selection gave us 53 articles of studies developed with human subjects, with granted access to the full text of the given publication, and contributed with clinical data of the compounds and the pathologies present in the studies. The articles based on in vitro or in vivo (animal models) experiments or trials were not considered because we focused our work on the data of studies carried out with human adults with different pathologies, to back the decision-making when looking for nutritional recommendations in the daily clinical practice with such patients. Nor were articles using low-caloric diets, low-fat diets, or Mediterranean-type diets considered, because these are healthy options for the human subjects in any intervention study, and they did not provide us with specific information about consumption of cruciferous foods or ingredients derived from Brassicaceae products. Additionally, articles where the isothiocyanates (ITCs) were used as biomarkers in intra-hospitalar tumor resections or documents based on healthy volunteers were also not included in the analysis, because they were patients with pathologies unrelated to our objective of study, and therefore, 36 articles were discarded.
According to all these selected criteria, 15 clinical studies were selected to carry out this critical study of clinical evidence using bioactives from Brassica and also one cohort study because of its ample sample size and long period of follow-up of 22 years. We also carried out the analysis of a "letter to editor"that was submitted as a response to this cohort study.
## The clinical evidences of health-promotion with brassica bioactives from a critical point of view
In order to proceed with the critical analysis of the articles, they were organized and grouped according time of duration of the study and type of pathology, in short-term or acute studies (Supplementary Table S1), medium duration or length , and long-term studies (Supplementary .
## Short-time studies
Within the short-term studies (Table S1), we found two documents on prostateand melanoma, three papers on respiratory pathologies, and one on type 2 diabetes. The main limitation observed in these studies is the length or duration of the study, of 1 month or less, because the validity of the data in terms of changes in bioavailability of the intake of compounds could be acceptable, but for the evaluation of data from tissues, the time of exposition is too low. In the article of Gee et al., a Phase Ib study, studied the concentration of DIM in prostate tissue after its administration in three groups: 200 mg, 400 mg, or placebo. They found a positive increase of DIM in plasma, but not concomitant or related with the presence of DIM in the prostate tissues, and the initial hypothesis was not corroborated. On the other hand, the urinary 2/16-hydroxysterone increased significantly by 68% (p = 0.030) in the group administrated with 400 mg of DIM. The levels of PSA taken separately, were also non-significant, but comparing between the group of 400 mg versus placebo, the reduction was of 1.8 times (p = 0.1). The gene expression of the CYP1A1 in blood after 15 days in the Group of 200 mg was significantly reduced by 31% (p = 0.032) and more importantly for the reduction for the CYP2B6 was highly significant by 87% (p = 0.040).
In Tahata et al., patients with more than two nevus or cell-proliferation structures of ≥4 mm Ø were selected and showed significant increases (ratio of concentration >1.5) in 14 out of the 92 proteins identified as tumor suppressors during the time of the study. On the other hand, it was not possible to relate the positive data on the significant reduction of proinflammatory cells such as IP-10, MCP-1, MIG, and MIP-1b to the supplementation of broccoli sprout extract (BSE) enriched in sulforaphane (SFN) because of different problems, such as reduced size in terms of number of participants, samples that were too small, and of a different nevus, which made the analysis difficult and induced higher inter-lesion variability. From the size of the nevus, it was not possible to obtain significant differences, only a general increase in size that was less patent in the patients treated with 100 µg BSE-SFN. Against these results it is worth highlighting that the changes in the size of the nevus in a period of one month are unusual and the comparisons between pictures depends in the illumination characteristic of the images, being unreliable data when analyzing the results. One of the strengths of this work actually are the improvements suggested for the future to use a different technique for the biopsies (in patch), to reduce the stress in the area of lesion, and to use a dermatoscope to substitute the pictures because of their great variability according to the illumination conditions. The short-time studies included also the three studies on respiratory problems. In the work of Sudini et al., the first observation was the extremely short time of intervention of three days, the shortest of all the analyzed studies. The intervention was carried out using food, not pills or supplements, because the interest was to study if the intake of cruciferous foods with high content in glucosinolate, especially glucoraphanin (sulforaphane), could have the same effect that was found in the interventions with broccoli sprout extract (BSE). The final objective of this work was to provide evidence to make recommendations of public health to increase the consumption of broccoli sprouts and germinates as a preventive factor in asthma and related pathologies. The study established very clearly the amount of sprouts to be ingested, the method of preparation, the temperature of the rest of foods, the time of chewing the sprouts, as well as the age of the sprouts, in order to reduce the variability in the concentration of glucoraphanin (sulforaphane) ingested. Despite this good design, the very short time of exposure resulted in non-significant results in any of the studied parameters: lung inflammation, oxidative stress, gene expression, antioxidant level, symptoms or lung function.
The work of Wise et al., compared to the previous one, increased the time of intervention, but again used a supplement-capsules with different concentrations of SFN. The pathology under study was chronic obstructive pulmonary disease (COPD), and in both cases, asthma and COPD were obstructive conditions. In the exclusion criteria, this one was the only paper that considered that the patients could be under treatment with oral anticoagulants. Other studies mentioned the cardiac pathologies as exclusive and were not as specific with the medication. It is indeed an important exclusion factor, because these drugs are anti-vitamin K and are affected by the dietary intake of green-leafy vegetables and other dietary sources of vitamin K, which may put at risk the correct availability of the drug, affecting the coagulation of blood, and requiring more control of the dosage administrated to the patients. The results obtained, with an statistical potency close to the 80%, were not significant for the expression of the NRF2 factor. The only significant difference in the treated groups was for the AKR1C1, which increased 1.45 times in the placebo versus 1.08 and 0.79, in the groups of 25 µmol and 150 µmol, respectively. Nevertheless, none of these changes was significant with respect to the basal level. The results were not very encouraging, but the authors opened a door to the analysis of the absorption via nasal application of bioactive compounds and its medical implications. Unlike the previous article, Brown et al., found significant results. The strong points of this work worth highlighting are the detailed preparation of the lyophilized material used and the characteristics of the diet description and instructions for the patients specifying the vegetables to be avoided. With respect to the experimental data, the induction with methacholine produced a reduction of 28.7% ± 7.2% in forced expiratory volume FEV1 (measure of lung disease); after consuming the SFN, 60% of the patients with asthma blocked the bronchoconstriction (BC) mediated by methacholine, in 20% of the patients the situation worsened, and in the last 20% there was no response. Besides, a significant reduction in the resistance to the entrance of air was detected (p = 0.03), and a small but significant increase in the average aerial way lumen. The FEV1 results and the NRF2 values were positively correlated between them. The results of bronchoconstriction (BC) and bronchoprotection (BP) were a paradox that needed a multivariable regression model to be understood (R2 = 0.67, p = 0.0003), considering the bronchoprotection as a resulting variable and using as independent values the reduction of FEV1 induced by methacholine, and the changes in the genes GCLM, GST1, and NQO1. Using this model, a negative correlation between BP and FEV1, and a positive correlation between BP and NQO1, were found. The authors concluded that SFN increased the expression of the NQO1 gene (phase II enzyme) associated with the increase in FEV1 as a response to the induction of methacholine and the improvement associated independently in the bronchoprotective capacity of deep breathing. The last work in this group of short-time interventions analyzed is about metabolic pathologies, corresponding to the work of Bahadoran et al., with an intervention on 81 patients of type 2 diabetes mellitus within the first year after diagnosis of the disease. A number of participants dropped-out of the study, and at the end, 63 patients (80%) were analyzed for fasting glucose, showing a significant reduction (reduction in both A-225 µmol SFN and B-112 µmol SFN groups), and a significant reduction in the A-group of insulin and insulin resistance.
## Medium-length studies
The analysis of the articles of medium-length duration between 2 and 8 months indicated that three were dedicated to studies on cancer in different localizations, two studies were on prostate cancer, and one study was on breast cancer.
In both studies of Alumkal et al.and Cipolla et al., the main factor studied was the change in the PSA (prostate specific antigen) parameter. In their study, Alumkal et al. ambitiously established the objective of a reduction of 50% in the level of PSA after 20 weeks of intervention, an objective that was not achieved. The obtained reductions were significant and in the range of 3% to 20%. In both studies it was also positive to find an increased time for doubling the amount of PSA after the intervention was found (pre-and post-intervention in placebo from 6.1 to 16.6 months; 9.6 to 31.9 months, in the treated group).
Cipolla et al.were focused in the changes in PSA over time, which were not significant when adjusting the data to the months (m0, m1, m3, and m6); but, when eliminating the m1, the reduction was of 0.0180 ng/mL/month. Besides, the median of log PSA was 38.5% lower in the intervention group than in the placebo, and the progression of the PSA was by 71.8% in the placebo versus the 44.4% in the SFN-treated group. Looking into the changes of the phase II enzymes, the GSTM1, and the inhibition of the HDAC in the work of Alumkal et al., non-significant differences were found. In both studies, the time for the intervention was similar, namely 20 and 24 weeks, respectively, but the amount of SFN was much higher in the work of Cipolla et al., and that probably favored the better more encouraging final results in this study. In any case, Atwell et al.planned to evaluate the efficacy of the consumption of the broccoli sprouts extract to modify the activity of the HDAC (histone deacetylase) and the breast tumor marker CA (carcinogenic antigen) in order to improve the diagnosis in women with benign disease and non-invasive CA. They proposed an intervention using six pills with a theoretical content of SFN of 30 mg in three administrations per day. A point in favor of this article is that they carried out an internal analysis of the composition of the capsules, and determined the total daily dosage of 224 mg of SFN instead of the declared 180 mg of SFN as designated by the supplier of the supplement. When evaluating the results, the levels of the HDAC before and after administration between groups were not statistically different. Only when comparing the two groups, treated and untreated, after the intervention was a significant difference observed, with a reduction in the HDAC in the group with the treatment. It should be noted that the authors also carried out a stratified data analysis considering the consumption of NSAIDs, using this variable as independent, and they found a significant reduction in the HDAC in the patients that were non-regular consumers of NSAIDs (p = 0.04), and it was not possible to find significant differences in the group of consumers of NSAIDs. Additionally, the biomarkers H3K18ac, H3K9ac, HDAC3, HDAC6, Ki-67, and P21 did not present statistically significant differences in the pre-or post-analysis. When comparing the levels of pre-and post-treatment within each group, a significant reduction was found in Ki-67 and HDAC3 in benign tissues in the SFN treated group, and a significant reduction of H3K9ac in the tissular DCIS in the placebo group. As a conclusion of this work, the authors did not plan as a hypothesis that the NSAIDs inhibited the synthesis of prostaglandins, and this could suppress the expression of regulator proteins through the involvement of the recruitment of the HDAC. The increased recruitment of the HDAC in the chromatin could prevent the inhibition associated with the consumption of SFN, and that could be a potential explanation for the observed data, and for the future studies using interventions with SFN, longer studies will be required to confirm these findings.
## Long-term studies
Finally, we proceeded with the analysis of the articles studying long-term interventions . In these works, we found various publications on different types of cancers: prostate, breast, ovarian, and pancreas (pilotand development, and one work on type 2 diabetes mellitusand a letter to the editoras a result of this study.
The works of Paltsev et al.and Traka et al.were both dedicated to prostate cancer, for a period of 12 months, but with different interventions and different parameters of the stage of development. In the first one, the intervention was carried out using capsules with 900 mg of DIM (3,3-diindolylmethane) and considered data of morphological changes (MI formula to calculate presence of cancer) in the prostate, and registration of the prostatic intraepithelial neoplasias (PIN) of low and high degree as well as the number of foci of cancer and urological dynamic parameters (maximum, average, and residual urinary flow) and erectile function. In the second work, with a higher dosage of glucoraphanin (parental glucosinolate of SFN) through the intake of different broccoli soups (based on broccolis with different contents of glucoraphanin), specific parameters included the change in the genetic expression of the RNA, changes in metabolites after biopsy from prostate tissue, and the plasma PSA levels. In the results of Paltsev et al., a significant reduction in the cancerous and precancerous cells was obtained in the treated group (from MI = 50 before treatment to MI = 0.08 after treatment), obtaining a complete regression in 45.5% of the patients (5 out of 11). Despite the results, the extremely reduced sample size in this study made it impossible to establish a significant difference between groups. On the other hand, Traka et al.observed changes in the genetic expression in all the patients, considering as reference the group that took the soup with the unmodified broccoli (not enriched). The changes were more striking in the patients that took the soups with enriched broccoli (with much higher glucoraphanin than regular broccoli, through the presence of one or two alleles of the Myb28 gene), but the study did not reach a significant result, because the number of patients or participants was below the minimum necessary of 78 participants, in order to detect significant changes of p < 0.02. The authors concluded that a diet rich in glucoraphanin may attenuate the transcriptional changes that take place in the prostate, but the data were not conclusive because of that limitation. It became clear, as in other studies with natural products and interventions for different types of diseases, that the size and number of participants should be enough to generate sound results that would help in the elaboration of recommendations for dietary options with food rich in bioactives such as Brassica vegetables, microgreens, and sprouts.
The works of Thomson et al.on breast cancer and Kiselev et al.on ovarian cancer were studies with the highest number of participants and the longest periods of evaluation. Thomson et al.studied the effects of diindolylmethane (DIM) + tamoxifen administration for 12 months. Previously, a tolerance test was carried out with the 10 first participants to establish a 150 mg dosage of DIM. From the aims of the study, the main objective was to analyze the breast density, and it had to be discarded because of the high incidence of participants with bilateral mastectomy, and therefore, they analyzed the effects on 2-OHE/16-α-OHE (alpha hydroxylated estrogen), the levels of strone, stradiol, and SHBG (sex hormone binding globulin). The statistical power of this study was established under 88% because it needed at least 55 participants per group, and in this case, the study was finished by 51 patients in the placebo group and 47 in the intervention group. The results support the hypothesis that the DIM (from Brassica) induced an increase in the 2-OHE/16-α-OHE in urine (p = 0.001) and increased the circulating SHGB in the patients with the combined treatment with tamoxifen. Besides that, the pharmacokinetic of the tamoxifen was changed, reducing the circulating forms of this drug. The results are positive but cannot be used for generalization.
On the other hand, the work of Kiselev et al.provided data on the survival and the survival without progression of ovarian cancer at 5 years in the patients with the combined treatment of EGCG (epigallocatechin gallate of green tea) and indole-3-carbinol (I3C)-the precursor of DIM. The combined I3C + EGCG demonstrated a 1.5 times increase in the survival free of cancer, a significant reduction in recurrences of the ovarian cancer with ascites in the treated group (by 8%-9%) when compared with the untreated group (by 60%-63%). The first results were observed in the pre-surgery stage because the majority of the patients (81%-85%) in the therapeutic branches of maintenance 1-3 could take a successful reduction surgery to eliminate all the visible tumor focus. On the contrary, for the participants in branches 4-5, it was not possible to carry out such procedure. Additionally, the levels of CA-125 were statistically lower in the 1-3 pre-surgery groups and after treatment than in the control group. Based on these data, the authors concluded that these were preliminary results, but they hypothesized that the treatment of I3C + EGCG, administered as part of the maintenance therapy during and after the combined treatment, inhibited the cancer precursor cells at the ovarian level, and therefore, diminished significantly the incidence of recurrence (ascites), and this was also redundant for higher rate of survival, global survival, and free-of-cancer survival.
From the article of Lozanovski et al., a pilot study, we found that one of their inclusion criteria made it difficult to extrapolate the results to all the participants, because it only considered patients without gastrointestinal symptoms, which are common symptoms of this pathology and in the chemotherapy treatments. With respect to the objective of the study, they aimed to analyze the effects of the intake of encapsulated broccoli sprouts as co-adjuvants in the chemotherapy and proposed the intake of 15 capsules of broccoli sprouts at a time, giving 90 mg of SFN and 180 mg of glucoraphanin, a never before tested dosage, which was established because of the poor life expectancy of the patients and the aggressive characteristics of this type of cancer. The results of Lozanovski et al., later on, established that the secondary effects of the chemotherapy, the lack of appetite, nausea, vomits, diarrhea, mouth sores, etc., were factors that made it very difficult for the patients to intake 15 pills at once, as the study initially planned. Besides, the masking of the bioactive agent was not possible because the pills were easily distinguishable. All these raised inconveniences, the progression of the pancreatic cancer and the GI symptomatology, led to a high rate of drop-off of 72% in the treatment group and 55% abandonment in the control group, and therefore, the results were not significant.
In order to finish this critical analysis, the only study of cohorts included in this review, from the group of Ma et al., analyzed data from 200,907 people free of disease (cancer, diabetes mellitus, or cardiovascular disease) at the beginning of the study. To establish this, they used semi-quantitative polls of food frequency: vegetables, fruits, whole grains, sugar drinks, juices, nuts, legumes, red meat, processed foods, trans-fats, omega-3 fatty acid rich foods, polyunsaturated fats, salt, and alcohol consumption. They also included the risk factors for type 2 diabetes mellitus (T2DM), body weight, smoking, pharmaceutical treatments or vitamin complex intake, family history of diabetes, chronic diseases, hypertension, dyslipidemia, and physical activity calculated on a weekly basis. With all this information, the authors presented the results that there was a risk for T2DM with the intake of cruciferous foods and glucosinolates. However, in agreement with the work of Oliviero et al., this study presented big limitations because it did not review or consider the cooking procedures, it did not quantify the amount of vegetables consumed on a daily or weekly basis, and it did not quantify the glucosinolate contents in the different vegetables consumed. Besides, the population sample was homogenous, because all of the participants were health-area workers, and the results are not extrapolatable to the general population. Finally, it should be highlighted that other nutrients and cofounding factors were not considered in the study, and they could affect the development of these pathologies.
## Future perspectives and challenges
The assessment of the health effects of bioactive compounds from plant-derived foods requires several aspects to be taken into consideration and a different perspective from that when evaluating single compounds of any chemical entity. In the case of Brassica spp. and other cruciferous foods and bioactive compounds obtained from them, it is crucial to control the concentration of glucosinolates ingested or delivered in the administration, and therefore to control the factors that affect this composition in the natural matrix, namely the production practices, handling, cooking procedures, time of chewing, age of sprout/mature plant, amount ingested, and concomitant ingestion with other foods, etc. In this context, well-researched studies are being carried out that include and analyze all these factors and standardize the Brassica-derived products by their SFN content, and efforts must continue in this line. The exact calculation of the amount of raw broccoli needed to produce a supplement that delivers a specific amount of SFN and its metabolites within the body remains to be established. It is important to point out that a food product should be tested in its pattern of consumption on a daily basis or with a frequency of consumption clearly reported.
One point that deserves attention is the unreasonable but common focus in the literature on glucoraphanin and glucobrassicin that merits mention, and contrasted to the high number of other GSLs occurring in vegetables. There are evidences presented for exciting protective effects of many other GSLs, but they are not commercially available. The same could be said for the indoles, with the literature focusing to an absurd degree on compounds that happen to be commercially available, although they are not necessarily the most widespread in vegetables, or when comparing a non-substituted and a substituted indole.
Similarly, efforts on health-oriented breeding (e.g., manipulation of the activity of nitrile specifier protein in broccoli by Roman et al.as well as the very promising bioactivities of exotic GSLs like glucomoringin and its ITC moringin from a tropical tree not from the Brassicaceae family but within the order Brassicales. This area of work together with additional efforts on the discovery of brassicaceous phytochemicals is another perspective: the widespread occurrence of epithionitriles in cabbage and new GSL products in these plants; the poorly known chemistry of some vegetables, e.g., commercial chemotype of watercress in the USA; and other derivatives including seleno-glucosinolates.
Additionally, the potential for positive synergistic effects of combinations of GSLs have been little studied, even though some pioneering work can be found in this subject. These are all examples of the many open lines of work for the future of understanding the potential benefits of the consumption of the bioactive compounds present in cruciferous foods from the farm to the foods for health.
The CONSORT (Consolidated Standards of Reporting Trials) statement includes a check list as a guidance for reporting randomized controlled trials and hence, it serves indirectly as a guidance for the design of human intervention studies to evaluate the beneficial effects of foods. The report of the trial design must include the eligibility criteria and a participant flow diagram, complete data of the intervention administered, and well-defined primary and secondary outcome measures, among others. It is highly advisable to refer to this check list before performing an intervention to ensure that we take into account all the factors that may later influence the final result.
In human studies, sample size is an important factor to reduce inter-individual variability, especially in markers subjected to a great variability, as the expression of particular genes, but for logistical reasons it is not always possible to have a large and appropriate number of participants. However, to help to reduce the variability in clinical trials performed in parallel groups, it is essential to control variables (the so-called covariables) that can influence in the behavior of other variables. Sex, age, and BMI are covariables that greatly influence responses linked to inflammatory processes or gene expression. It is important to randomize intervention groups stratified by these covariables, in order to reduce inter-individual variability, so that significant differences are more likely to be detected, and some clinical trials do not take these aspects into account. Absorption, distribution, metabolism, and excretion (ADME) processes are also influenced by age, sex, dietary habits, or the microbiome. The COST Action POSITIVe working group has recently published the determinants involved in the ADME process of different bioactive compounds, and they include the concept of metabolic phenotype (metabotype) or the similar pharmacokinetics profile common in some individuals, which may result in similar response patterns. This concept should be applied to the human host as well as to the gut microbiota metabolisms and is dependent on gene polymorphisms as well as microbiome composition. This perspective can help to understand the differences in responses observed between individuals. Stratification of individuals by metabotype is an attractive challenge for future research.
It is crucial that the outcome measure is of biological relevance; biomarkers or risk factors measured must be associated to the latter development of a disease. Validation of appropriate markers is a pending issue as very few are recognized as valid in terms of specificity and applicability to a range of population.
Biomarkers of effect are early stage end-points, for instance the modulation of phase 2 enzymes by glucosinolates. They need more time than biomarkers of exposure to be influenced by the dietary treatment. Hence, length or duration of the study must be defined according to the biomarker measured to be modified, that is, to define perfectly the time of exposure to observe changes in relevant parameters. Gene expression is one important target for glucosinolates, and it requires a sufficient period of exposure to (de)activate signaling pathways involved. It is crucial to find appropriate biomarkers of effect that are linked to later disease outcomes, and more investigation is needed in this sense. Post-study follow-up can be of great value in assessing the persistence of certain effects, or in discovering those that appear more long-term.
In case of clinical trials performed with patients and not healthy subjects, the appropriate timing of treatment with the nutraceutical or food product is also a pending issue. The degree of progression of the disease greatly influences the response observed. In some cases, a bioactive compound is no longer sufficient, and in other cases it can have the opposite effect. It has been suggested that the effect of SFN on Nrf2 activation is desirable in the early stages of tumorigenesis, whilst it may have a deleterious effect in later stages. It can behave as an oncogenic transcription factor and increase the cytoprotective activity in the cancer cell and cause resistance to chemotherapy. More studies are needed on the potential effects of this molecule on patients with diagnosed cancer before making any recommendations of supplements with these phytochemicals. The need for studies with larger populations is clear, as well as longer follow-ups, in order to find significant results with enough strength to support recommendations for public health in favor of increasing the consumption of cruciferous foods on a daily basis.
Supplementary Materials: The following are available online. : Short-time studies; : Medium-length studies; : Long term studies. Funding: This research was partially supported by Fundación Seneca, Murcia Regional Agency for Science and Technology, Project 20855/PI/18. |
Intrinsic and Extrinsic Thymic Adrenergic Networks: Sex Steroid-Dependent Plasticity
The thymus is sexually differentiated organ providing microenvironment for T-cell precursor differentiation/maturation in the major histocompatibility complex-restricted self-tolerant T cells. With increasing age, the thymus undergoes involution leading to the decline in efficacy of thymopoiesis. Noradrenaline from thymic nerve fibers and "(nor) adrenergic" cells is involved in the regulation of thymopoiesis. In rodents, noradrenaline concentration in thymus and adrenoceptor (AR) expression on thymic cells depend on sex and age. These differences are suggested to be implicated in the development of sexual diergism and the age-related decline in thymopoiesis. The programming of both thymic sexual differentiation and its involution occurs during the critical early perinatal period and may be reprogrammed during peripubertal development. The thymic (re) programming is critically dependent on circulating levels of gonadal steroids. Although the underlying molecular mechanisms have not yet been elucidated fully, it is assumed that the gonadal steroid action during the critical perinatal/peripubertal developmental periods leads to long-lasting changes in the efficacy of thymopoiesis partly through (re) programming of "(nor)adrenergic" cell networks and AR expression on thymic cells.
The thymus is organ in which T cells are continually generated in a highly dynamic process comprising T-cell receptor (TCR) gene rearrangement, lineage commitment, and selection [bib_ref] T cell development by the numbers, Krueger [/bib_ref]. These processes are linked to distinct rates of proliferation and cell death by apoptosis [bib_ref] T cell development by the numbers, Krueger [/bib_ref]. With increasing age, the thymus atrophies and declines in functions, the phenomenon termed involution [bib_ref] Thymus and aging: morphological, radiological, and functional overview, Rezzani [/bib_ref]. Consequently, thymic generation of naïve T cells declines [bib_ref] It's not all equal: a multiphasic theory of thymic involution, Aw [/bib_ref]. This leads to the shrinkage of peripheral TCR repertoire and the expansion of memory T cell compartment, i.e., to the changes covered by the canopy term immunosenescence [bib_ref] Age-related changes in lymphocyte development and function, Linton [/bib_ref] [bib_ref] Aging and T-cell diversity, Goronzy [/bib_ref]. At the clinical level, the immunosenescence is associated with a greater susceptibility to infections [bib_ref] Ageing and life-long maintenance of T-cell subsets in the face of latent..., Nikolich-Žugich [/bib_ref] , an impaired response to vaccinations [bib_ref] How sex and age affect immune responses, susceptibility to infections, and response..., Giefing-Kröll [/bib_ref] , and an increased propensity for malignant diseases [bib_ref] Potential role of immunosenescence in cancer development, Fulop [/bib_ref] [bib_ref] The aging immune system and its relationship with cancer, Foster [/bib_ref]. In addition, according to the U.S. Center for Disease Control, approximately 80% of aged individuals are afflicted with at least one chronic disease as a result of a declination of immune function. Consequently, factors contributing to the thymic involution and mechanisms of their action are becoming the subject of increased interest in the scientific and healthcare communities alike. It should be emphasized that understanding of the mechanisms underlying thymic involution is important not only for moderating the deleterious effects of immunosenescence, but also for envisaging strategies to "rejuvenate" the immune system. It is noteworthy that even in a significant thymic involution thymopoiesis does not cease completely, so it may be enhanced [bib_ref] Thymic function in young/ old chimeras: substantial thymic T cell regenerative capacity..., Mackall [/bib_ref]. The thymic "rejuvenation" becomes particularly important after exposure to chemotherapy, ionizing radiation, and some infective agents (e.g., HIV-1) [bib_ref] Thymic involution and immune reconstitution, Lynch [/bib_ref].
There is evidence that (i) early perinatal programming of the thymus is crucial for the development of thymic involution, and consequently the efficacy of immune responses from early life through adulthood [bib_ref] Thymus size and age-related thymic involution: early programming, sexual dimorphism, progenitors and..., Gui [/bib_ref] , and (ii) this phenomenon is sexually dimorphic [bib_ref] Thymus size and age-related thymic involution: early programming, sexual dimorphism, progenitors and..., Gui [/bib_ref] [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref]. Consistently, sex differences in the organ size, structural organization, and thymopoiesis [bib_ref] Thymus size and age-related thymic involution: early programming, sexual dimorphism, progenitors and..., Gui [/bib_ref] [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref] [bib_ref] Sex differences in the phenotypic characteristics of rat thymocytes, Leposavić [/bib_ref] [bib_ref] Differential effects of gonadectomy on the thymocyte phenotypic profile in male and..., Leposavić [/bib_ref] [bib_ref] Both age and gender affect thymic output: more recent thymic migrants in..., Pido-Lopez [/bib_ref] [bib_ref] Sex, the aging immune system, and chronic disease, Bupp [/bib_ref] [bib_ref] Sex hormones have pervasive effects on thymic epithelial cells, Dumont-Lagacé [/bib_ref] [bib_ref] Tracing thymic output in older individuals, Mitchell [/bib_ref] , and consequently T-cell immune response [bib_ref] Gonadal steroids and immunity, Olsen [/bib_ref] [bib_ref] Sex differences in immune responses, Klein [/bib_ref] , have been observed. A more rapid thymic involution was found in male compared with female mice [bib_ref] Gender-related differences in the rates of age associated thymic atrophy, Aspinall [/bib_ref]. Consequently, adult females have the ability to reject allografts more efficiently, a greater ability to combat various viral and bacterial infections, and superior antitumor responses [bib_ref] Gonadal steroids and immunity, Olsen [/bib_ref].
There are data indicating that the early programming of the thymus development/involution is shaped by genetic, environmental, and hormonal factors [bib_ref] Thymus size and age-related thymic involution: early programming, sexual dimorphism, progenitors and..., Gui [/bib_ref]. The role of genetic factors has been shown in both mice and rats [bib_ref] Genetically determined thymus enlargement in the early life in BUF/Mna rats, Saito [/bib_ref] [bib_ref] Strain differences in thymus weight in mice with different predispositions to spontaneous..., Belyaev [/bib_ref] [bib_ref] Genetic control of thymus size in inbred mice, Peleg [/bib_ref] [bib_ref] Age-related thymic involution in C57BL/6J x DBA/2J recombinant-inbred mice maps to mouse..., Hsu [/bib_ref] [bib_ref] MCL1 increases primitive thymocyte viability in female mice and promotes thymic expansion..., Gui [/bib_ref]. These genetically based differences are suggested to be connected to strain differences in susceptibility to various pathologies involving immune mechanisms [bib_ref] Genetically determined thymus enlargement in the early life in BUF/Mna rats, Saito [/bib_ref] [bib_ref] Strain differences in thymus weight in mice with different predispositions to spontaneous..., Belyaev [/bib_ref] [bib_ref] Genetic control of thymus size in inbred mice, Peleg [/bib_ref] [bib_ref] Age-related thymic involution in C57BL/6J x DBA/2J recombinant-inbred mice maps to mouse..., Hsu [/bib_ref]. Environmental factors, such as malnutrition, and exposure to endocrine disruptors, in early postnatal life are also shown to influence the pace of thymic involution [bib_ref] Early nutrition and immunity-progress and perspectives, Calder [/bib_ref]. Alterations in circulating levels of sex steroids in the critical early postnatal developmental "window" may influence not only sexual dimorphism in structural and functional thymic parameters, but also the timing of thymic involution [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref] [bib_ref] Age-associated remodeling of thymopoiesis: role for gonadal hormones and catecholamines, Leposavić [/bib_ref]. Furthermore, gonadal steroids may influence sexual dimorphism in thymopoiesis, and the age-related decline in its efficacy through: (i) modulating thymic extrinsic (encompassing noradrenergic nerve fibers) and intrinsic [composed of noradrenaline-synthesizing cells, i.e., "(nor)adrenergic" cells] adrenergic regulatory networks, in terms of their density/noradrenaline content and (ii) adrenoceptor (AR) expression on thymic cells [bib_ref] Early postnatal castration affects thymic and thymocyte noradrenaline levels and beta-adrenoceptor-mediated influence..., Leposavić [/bib_ref] [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref]. In addition, it should be pointed out that the ablation of gonadal steroids during the peripubertal developmental "window" leads not only to short-term increase in thymic weight and enhancement of thymopoiesis, but also to the long-lasting thymic "rejuvenation" [bib_ref] Age-associated remodeling of thymopoiesis: role for gonadal hormones and catecholamines, Leposavić [/bib_ref].
The central goal of this mini review is to summarize recent findings and current knowledge related to the mechanisms of indirect (nor)adrenaline-mediated action of gonadal steroids on the programming/reprogramming of thymic involution, as its action may be easily controlled by many drugs in use for nonimmune indications.
## Thymic extrinsic and intrinsic (nor)adrenergic regulatory networks
Thymic extrinsic (nor)adrenergic network
The thymus receives extensive noradrenergic innervation [bib_ref] Sympathetic innervation of murine thymus and spleen: a comparative histofluorescence study, Williams [/bib_ref] [bib_ref] Components of sympathetic innervation of the rat thymus during late fetal and..., Leposavić [/bib_ref]. The varicose noradrenergic fibers terminate in close proximity to thymocytes [bib_ref] Components of sympathetic innervation of the rat thymus during late fetal and..., Leposavić [/bib_ref] [bib_ref] Neurochemical, electrophysiological and immunocytochemical evidence for a noradrenergic link between the sympathetic..., Vizi [/bib_ref] , and various subsets of thymic non-lymphoid (stromal) cells [bib_ref] Neurochemical, electrophysiological and immunocytochemical evidence for a noradrenergic link between the sympathetic..., Vizi [/bib_ref] [bib_ref] β-Adrenoceptor mediated effects in rat cultured thymic epithelial cells, Kurz [/bib_ref] [bib_ref] Thymic innervation in the rat: a light and electron microscopical study, Novotny [/bib_ref] [bib_ref] Adrenergic nerve fibres and mast cells: correlation in rat thymus, Artico [/bib_ref]. In rodents, noradrenergic fibers appear in the thymus in late embryonic period, and their density increases during prepubertal development [bib_ref] Immunomodulatory role of noradrenergic neural and non-neural system: role for β-adrenoceptor-mediated mechanisms..., Leposavić [/bib_ref] [bib_ref] Sympathetic innervation of fetal mouse thymus, Singh [/bib_ref]. The data on postpubertal changes in their density are inconsistent [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref] [bib_ref] Maintenance of noradrenergic sympathetic innervation in the involuted thymus of the aged..., Bellinger [/bib_ref] [bib_ref] Relationships between monoaminergic and cholinergic innervation of the rat thymus during aging, Mićić [/bib_ref] [bib_ref] Alterations in sympathetic innervation of thymus and spleen in aged mice, Madden [/bib_ref] [bib_ref] Occurrence of adrenergic nerve fibers and of noradrenaline in thymus gland of..., Cavallotti [/bib_ref] [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref]. In advanced age, in rodents of distinct (sub)strains has been observed decrease, increase and lack of changes in thymic noradrenergic nerve fiber density compared with young adult (sub)strain-matched ones [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref] [bib_ref] Maintenance of noradrenergic sympathetic innervation in the involuted thymus of the aged..., Bellinger [/bib_ref] [bib_ref] Relationships between monoaminergic and cholinergic innervation of the rat thymus during aging, Mićić [/bib_ref] [bib_ref] Alterations in sympathetic innervation of thymus and spleen in aged mice, Madden [/bib_ref] [bib_ref] Occurrence of adrenergic nerve fibers and of noradrenaline in thymus gland of..., Cavallotti [/bib_ref] [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref]. This inconsistency is most likely linked to (sub)strain and sex-dependent differences in the kinetics of postpubertal changes in thymic noradrenergic innervation. It has also been suggested that the noradrenaline content in thymic nerve fibers, and consequently thymic noradrenaline concentration vary with age [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref] [bib_ref] Maintenance of noradrenergic sympathetic innervation in the involuted thymus of the aged..., Bellinger [/bib_ref] [bib_ref] Relationships between monoaminergic and cholinergic innervation of the rat thymus during aging, Mićić [/bib_ref] [bib_ref] Alterations in sympathetic innervation of thymus and spleen in aged mice, Madden [/bib_ref] [bib_ref] Occurrence of adrenergic nerve fibers and of noradrenaline in thymus gland of..., Cavallotti [/bib_ref] [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref]. In addition, both thymic parameters were found to be greater in male than in age-matched female rats (51) .
## Thymic intrinsic (nor)adrenergic network
Many types of mature immune cells synthesize and secrete catecholamines [bib_ref] Discovery of endogenous catecholamines in lymphocytes and evidence for catecholamine regulation of..., Bergquist [/bib_ref] [bib_ref] HPLC-ED measurement of endogenous catecholamines in human immune cells and hematopoietic cell..., Cosentino [/bib_ref] [bib_ref] Expression of tyrosine hydroxylase in lymphocytes and effect of endogenous catecholamines on..., Qiu [/bib_ref]. The investigations of the expression of tyrosine hydroxylase (TH), the key rate-limiting enzyme in catecholamine synthesis in freshly isolated thymic cells, cultured thymocytes and cells from adult thymic organ culture revealed that thymic cells, including thymocytes, synthetize noradrenaline [bib_ref] Early postnatal castration affects thymic and thymocyte noradrenaline levels and beta-adrenoceptor-mediated influence..., Leposavić [/bib_ref] [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref] [bib_ref] Effects of catecholamines on thymocyte apoptosis and proliferation depend on thymocyte microenvironment, Radojević [/bib_ref]. TH-immunoreactive cells were found across all thymocyte subsets delineated by CD3 expression levels, but their frequency was highest among the most mature CD3 high thymocytes [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref]. In addition, TH-immunoreactive cells were observed in various thymic non-lymphoid cell subpopulations [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref]. Their density varies across distinct thymic microenvironments. They are frequent at the medullary side of the corticomedullary junction, whereas their density is moderate and poor in the subcapsular cortex, and intracortically/intramedullary, respectively [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref]. This is important as various thymic non-lymphoid cell subsets are strategically positioned in particular thymic microenvironments to orchestrate thymocyte differentiation/ maturation [bib_ref] Journey through the thymus: stromal guides for T-cell development and selection, Takahama [/bib_ref]. TH immunoreactivity was observed in thymic epithelial cells (TECs) [bib_ref] β-Adrenoceptor mediated effects in rat cultured thymic epithelial cells, Kurz [/bib_ref] [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref] [bib_ref] Immuno-characterisation of neuroendocrine cells of the rat thymus gland in vitro and..., Botham [/bib_ref] [bib_ref] Immunoreactivity of neural crest-derived cells in thymic tissue developing under the rat..., Jones [/bib_ref] [bib_ref] Ontogeny of intrinsic innervation in the human thymus and spleen, Anagnostou [/bib_ref] , macrophages, and dendritic cells [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Catecholamines as immunomodulators: a role for adrenoceptor-mediated mechanisms in fine tuning of..., Leposavić [/bib_ref]. In TEC population, TH immunoreactivity was found in neural crest-derived thymic nurse cells [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref] [bib_ref] Immuno-characterisation of neuroendocrine cells of the rat thymus gland in vitro and..., Botham [/bib_ref] [bib_ref] Immunoreactivity of neural crest-derived cells in thymic tissue developing under the rat..., Jones [/bib_ref] , type 1 (subcapsular/perivascular), and type 5 (located mainly in corticomedullary region) cells [bib_ref] β-Adrenoceptor mediated effects in rat cultured thymic epithelial cells, Kurz [/bib_ref] [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref] [bib_ref] Ontogeny of intrinsic innervation in the human thymus and spleen, Anagnostou [/bib_ref]. The density of both lymphoid and non-lymphoid TH-immunoreactive cells was shown to be higher in male than in female rats (51) . In addition, the overall noradrenaline content in thymocytes was found to be greater in male compared with female adult rats (51) . Although studies in rat adult thymic organ and thymocyte cultures suggested that noradrenaline from thymic "(nor)adrenergic" cells is implicated in the fine tuning of thymopoisesis (55), a role for thymic intrinsic adrenergic network in thymic homeostasis is still far from being understood. It is noteworthy that intrinsic (nor)adrenergic cellular networks: (i) have also been identified in some other tissues and (ii) suggested to be particularly important under specific conditions, e.g., following sympathectomy, gonadectomy, chronic stress [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref] [bib_ref] Neuroendocrine properties of intrinsic cardiac adrenergic cells in fetal rat heart, Huang [/bib_ref] [bib_ref] An intrinsic adrenergic system in mammalian heart, Huang [/bib_ref] [bib_ref] Norepinephrine from synovial tyrosine hydroxylase positive cells is a strong indicator of..., Miller [/bib_ref] [bib_ref] Catecholaminergic systems in stress: structural and molecular genetic approaches, Kvetnansky [/bib_ref] [bib_ref] Neural crest-derived resident cardiac cells contribute to the restoration of adrenergic function..., Tamura [/bib_ref] , as it allows for mainly local regulation of the catecholamine influence [bib_ref] Catecholaminergic systems in stress: structural and molecular genetic approaches, Kvetnansky [/bib_ref] [bib_ref] First appearance and location of catecholaminergic cells during experimental arthritis and elimination..., Capellino [/bib_ref].
## Ar expression on thymic cells
To corroborate modulatory role for noradrenaline in the thymus is the expression of ARs on both thymocytes and thymic FiGURe 1 | Influence of alterations in circulating ovarian steroid levels in critical developmental periods on programming/reprogramming of thymic extrinsic and intrinsic adrenergic networks. This figure indicates (middle schemes) sex differences in noradrenaline content in noradrenergic nerve fibers and thymocytes, density of tyrosine hydroxylase (TH)-expressing ("adrenergic") cells, density of β2-adrenoceptor (AR) + thymic cells and thymocyte β2-AR surface density in young adult rat thymus, and influence of (upper scheme) single injection of testosterone on the third postnatal day to female rats and (lower scheme) ovariectomy in peripubertal period on noradrenergic nerve fiber and thymic "adrenergic" cell density and their noradrenaline content, as well as the density of AR-expressing thymic cells and thymocyte AR surface density in young adult rats.
non-lymphoid cells. Thymic cells express β2and α1-AR [bib_ref] The thymus gland as a major target for the central nervous system..., Marchetti [/bib_ref] [bib_ref] Sympathetic nervous system control of thymus gland maturation: autoradiographic characterization and localization..., Marchetti [/bib_ref] [bib_ref] Characterization, expression, and hormonal control of a thymic beta 2-adrenergic receptor, Marchetti [/bib_ref] [bib_ref] Expression of alpha1-adrenoceptors on thymic cells and their role in fine tuning..., Pesić [/bib_ref]. Their expression is reciprocally regulated during thymocyte maturation [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref] [bib_ref] β-Adrenergic receptors on murine lymphocytes: density varies with cell maturity and lymphocyte..., Fuchs [/bib_ref] [bib_ref] Regulated expression of alpha-1 adrenergic receptors in the immune system, Kavelaars [/bib_ref]. The most mature CD3 high thymocytes predominantly express β2-AR, whereas α1-AR expression is predominant on the most immature CD3 − thymocytes [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref] [bib_ref] Catecholamines as immunomodulators: a role for adrenoceptor-mediated mechanisms in fine tuning of..., Leposavić [/bib_ref] [bib_ref] Expression of alpha1-adrenoceptors on thymic cells and their role in fine tuning..., Pesić [/bib_ref] [bib_ref] Regulated expression of alpha-1 adrenergic receptors in the immune system, Kavelaars [/bib_ref]. There is sexual diergism in the expression of β2-AR on thymocytes. Immunophenotyping showed the higher frequency of β2-AR-expressing cells among thymocytes from female compared with male young adult rats, but lower density of the receptor on their surface (73) . In addition, autoradiographic studies indicated a sexually dimorphic pattern of postnatal changes in the density of β-AR in rat thymus [bib_ref] Characterization, expression, and hormonal control of a thymic beta 2-adrenergic receptor, Marchetti [/bib_ref]. There are no data on sex differences in α1-AR expression on thymocytes.
The expression of β2-AR was also demonstrated on cortical (aminopeptidase A + ), and medullary (UEA-1 + ) TECs, CD68 + macrophages, and OX62 + dendritic cells [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref]. In addition, α1-AR-immunoreactive cells were observed among TECs and macrophages located predominantly in subcapsular/subtrabecular and corticomedullary thymic regions [bib_ref] Catecholamines as immunomodulators: a role for adrenoceptor-mediated mechanisms in fine tuning of..., Leposavić [/bib_ref]. Thymic dendritic cells also express α1-AR [bib_ref] Age-associated remodeling of neural and nonneural thymic catecholaminergic network affects thymopoietic productivity, Leposavić [/bib_ref]. The subsets of β2-AR + and α1-AR + non-lymphoid cells were shown to co-express TH (60). Thus, not only paracrine, but also autocrine noradrenaline action may be expected in the thymus.
## Gonadal steroids and programming/reprogramming of the thymic (nor)adrenergic networks and ar expression early postnatal thymic programming
The thymus is sexually differentiated organ [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref]. The sexual differentiation in the thymus, as in the brain areas controlling gonadotropin release, occurs during the critical perinatal period, and is governed by sex steroid-dependent mechanisms [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref]. In addition, the widely accepted organizational/activational hypotesis of the bran development is extended to encompass the thymic differentiation [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref]. According to the original hypothesis, in the absence of testicular androgens during the critical period (starting at the late prenatal period and continuing, at least, to day 5 postpartum), the areas controlling gonadotropin release develop in a primarily female manner [bib_ref] Sexual differentiation of the brain, Mackinnon [/bib_ref] [bib_ref] Organizing action of prenatally administered testosterone propionate on the tissues mediating mating..., Phoenix [/bib_ref] [bib_ref] Sexual differentiation of the hypothalamus, Gorski [/bib_ref] [bib_ref] Synaptogenesis and neuronal plasticity to gonadal steroids: implication for the development of..., Arai [/bib_ref]. Conversely, the presence of testicular androgens leads to their defeminization/masculinization, a phenomenon known as neonatal androgenization [bib_ref] Sexual differentiation of the hypothalamus, Gorski [/bib_ref] [bib_ref] Synaptogenesis and neuronal plasticity to gonadal steroids: implication for the development of..., Arai [/bib_ref] [bib_ref] Evidence for a morphological sex difference within the medial preoptic area of..., Gorski [/bib_ref]. This postpones sexual maturation and leads to development of non-ovulatory ovaries with estrogen hyporesponsiveness [bib_ref] Synaptogenesis and neuronal plasticity to gonadal steroids: implication for the development of..., Arai [/bib_ref] [bib_ref] Effect of estrogen and progesterone on the development of endometrial hyperplasia in..., Tang [/bib_ref] [bib_ref] Production of anovulatory, sterile rats by single injections of testosterone propionate, Barraclough [/bib_ref]. The mechanisms of testosterone action in the brain and thymus are extremely complex, as in both organs it converts into estrogen [bib_ref] Evidence for sexual dimorphism of estrogen receptors in hypothalamus and thymus of..., De Fougerolles Nunn [/bib_ref] [bib_ref] Sexual differentiation of the brain, Mackinnon [/bib_ref] [bib_ref] Organizing action of prenatally administered testosterone propionate on the tissues mediating mating..., Phoenix [/bib_ref] [bib_ref] Sexual differentiation of the hypothalamus, Gorski [/bib_ref] [bib_ref] Synaptogenesis and neuronal plasticity to gonadal steroids: implication for the development of..., Arai [/bib_ref] , and consequently does not act only through androgen receptors [bib_ref] Identification and characterization of membrane androgen receptors in the ZIP9 zinc transporter..., Thomas [/bib_ref]. The binding of estradiol to classical estrogen receptor (ER)α or ERβ in the cytoplasm of target cells causes the receptor dimerization and translocation in nucleus, where the dimer associates with various coactivators to enable binding to the estrogen response elements (EREs) in or near the promoters of target genes [bib_ref] Estradiol and the developing brain, Mccarthy [/bib_ref]. Estradiol can also influence expression of genes that do not harbor EREs in their promoter regions. In this case, ligand-activated ERs do not bind DNA directly, but through protein-protein interactions with other classes of transcription factors at their respective response elements in promotor region of their target genes [bib_ref] Minireview: G protein-coupled estrogen receptor-1, GPER-1: its mechanism of action and role..., Filardo [/bib_ref]. In addition, estradiol may act through membrane G protein-coupled ER (GPER, previously termed GPR30) [bib_ref] Minireview: G protein-coupled estrogen receptor-1, GPER-1: its mechanism of action and role..., Filardo [/bib_ref]. This involves mobilization of diverse signaling pathways and may depend on a number of conditions, like the availability of signal transduction molecules and downstream targets [bib_ref] Minireview: G protein-coupled estrogen receptor-1, GPER-1: its mechanism of action and role..., Filardo [/bib_ref].
It was shown that a single injection of testosterone on the third postnatal day enhanced thymic growth and postponed thymic involution in female rats, which normally starts around puberty [bib_ref] Neonatal androgenization affects the intrathymic T-cell maturation in rats, Leposavić [/bib_ref] [bib_ref] Neonatal testosterone imprinting affects thymus development and leads to phenotypic rejuvenation and..., Leposavić [/bib_ref]. Accordingly, long-lasting changes in thymopoiesis, mirrored in the enhanced thymocyte differentiation/maturation in adult animals were observed [bib_ref] Neonatal testosterone imprinting affects thymus development and leads to phenotypic rejuvenation and..., Leposavić [/bib_ref]. In addition, neonatal androgenization facilitated the generation of CD4 − CD8 + TCRαβ high cells, and consequently shifted CD4 + /CD8 + recent thymic emigrant ratio in peripheral blood toward the latter [bib_ref] Neonatal testosterone imprinting affects thymus development and leads to phenotypic rejuvenation and..., Leposavić [/bib_ref]. The thymopoietic changes were ascribed to thymocyte overexpression of Thy-1, as its overexpression reduces thymocyte negative selection and favors maturation of CD8 + T cells [bib_ref] Thymocytes in Thy-1-/-mice show augmented TCR signaling and impaired differentiation, Hueber [/bib_ref]. Considering CD8 + T cell dominance in the periphery of males when compared with females [bib_ref] Gonadal steroids and immunity, Olsen [/bib_ref] [bib_ref] Gender differences in the development and function of the immune system, Shames [/bib_ref] , the previous findings indicate defeminization/masculinization of T-cell compartment in adult neonatally androgenized rats, i.e., speak in favor of a sex steroid role in the sexual differentiation of thymus.
Although aware of the complexity of changes in neuroendocrine-thymic communications in neonatally androgenized rats, in this review we focused on those mediated by catecholamines. Neonatal androgenization was shown to increase thymic noradrenaline concentration in adult rats [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref]. This mainly reflected the increase in nerve fiber noradrenaline content [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref]. Consistent with the so-called transsynaptic action of sex steroids on neurotransmitter synthesis [bib_ref] Gonadal steroid modulation of neuroendocrine transduction: a transynaptic view, Alonso-Solís [/bib_ref] , the previous finding may be explained by an augmented sympathetic tone in neonatally androgenized rats [bib_ref] Hormonal control of the development and regulation of tyrosine hydroxylase expression within..., Simerly [/bib_ref] [bib_ref] Gender differences in hypothalamic tyrosine hydroxylase and alpha(2)-adrenoceptor subtype gene expression in..., Plut [/bib_ref]. However, the higher noradrenaline concentration partly reflected the greater density of noradrenaline-synthesizing cells and noradrenaline content per cell [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref]. Considering that the circulating level of testosterone was elevated in neonatally androgenized rats [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref] , this could be associated with data indicating that androgens prominently transactivate TH promoter [bib_ref] Regulation of the transcriptional activity of the tyrosine hydroxylase gene by androgen..., Jeong [/bib_ref]. In light of data from other studies (51), the previous findings suggest thymic defeminization/masculinization in neonatally androgenized rats .
As additional sign of defeminization/masculinization (73), the frequency of β2-AR-expressing cells within thymocytes [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref] was diminished in neonatally androgenized rats . In addition, neonatal androgenization decreased β2-AR density on thymocytes (35) . Given that in many cell types estrogen, acting through classical ERs, upregulates β2-AR expression [bib_ref] The effect of estrogen and progesterone on beta-adrenergic receptor activity in rabbit..., Moawad [/bib_ref] [bib_ref] Oestrogen changed cardiomyocyte contraction and beta-adrenoceptor expression in rat hearts subjected to..., Wu [/bib_ref] , the alterations in β2-AR density could reflect estrogen hyporesponsiveness [bib_ref] Effect of estrogen and progesterone on the development of endometrial hyperplasia in..., Tang [/bib_ref] [bib_ref] Effects of testosterone on fat cell lipolysis. Species differences and possible role..., Arner [/bib_ref]. This hyporesponsiveness most likely emerged from the ER interaction with an excess of estrogen (as a result of testosterone aromatization) during the critical period [bib_ref] Effect of the absence of neonatal testosterone imprinting on the activity of..., Csaba [/bib_ref] [bib_ref] Effect of a single treatment (imprinting) with genistein or combined treatment with..., Csaba [/bib_ref]. The interaction of receptor with excess ligand in the critical period is shown to cause misprinting substantiated in diminished receptor binding capacity and responsivity in later life [bib_ref] Effect of the absence of neonatal testosterone imprinting on the activity of..., Csaba [/bib_ref] [bib_ref] Effect of a single treatment (imprinting) with genistein or combined treatment with..., Csaba [/bib_ref]. The elevation of thymic noradrenaline concentration following the testosterone injection could also impair the efficacy of β2-AR signaling (through the hormonal misprinting) (35), leading to the diminished noradrenaline action as the ultimate effect. In favor of this assumption is the increase in Thy-1 expression in adult neonatally androgenized rats [bib_ref] Estradiol and the developing brain, Mccarthy [/bib_ref]. Namely, the incubation of murine thymocytes with noradrenaline causes time-and concentration-dependent decreases in the Thy-1 mRNA levels, which are completely preventable by propranolol [bib_ref] Mechanism of catecholamine-mediated destabilization of messenger RNA encoding Thy-1 protein in T-lineage..., Wajeman-Chao [/bib_ref] [bib_ref] Thy-1 mRNA destabilization by norepinephrine a 3′ UTR cAMP responsive decay element..., Lajevic [/bib_ref]. Moreover, given that: (i) noradrenaline upregulates α1-AR expression (100) and (ii) long-lasting α1-AR blockade facilitates thymocyte differentiation/maturation toward CD4 + CD8 − TCRαβ high cells [bib_ref] Expression of alpha1-adrenoceptors on thymic cells and their role in fine tuning..., Pesić [/bib_ref] , the contribution of an augmented α1-AR signaling (reflecting its increased density and/or noradrenaline concentration) to the thymocyte maturation skewed toward CD4 − CD8 + TCRαβ high cells in adult neonatally androgenized rats cannot be ruled out.
In favor of the role of sex steroids in perinatal programming of thymic noradrenergic networks are also data showing that orchidectomy in the critical perinatal period lowers levels of both neurally-and thymocyte-derived noradrenaline in adult rats and thereby contributes to the deceleration of the thymic involution [bib_ref] Early postnatal castration affects thymic and thymocyte noradrenaline levels and beta-adrenoceptor-mediated influence..., Leposavić [/bib_ref]. This is consistent with data indicating that not only in presence of excess ligand in the critical periods, but also in its absence the ligand-receptor connection changes for life [bib_ref] The biological basis and clinical significance of hormonal imprinting, an epigenetic process, Csaba [/bib_ref].
To summarize, the previous findings indicate that alterations in circulating levels of sex steroids in the critical perinatal period may affect the programming of the sexually dimorphic (nor) adrenaline influence on thymopoiesis. However, the molecular mechanisms standing behind this phenomenon remain to be elucidated.
## Peripubertal thymic reprogramming
It has been suggested that the hormonal changes occurring at the time of puberty lay the framework for biological differences that persist throughout life [bib_ref] Sex hormones and glucocorticoids: interactions with the immune system, Silva [/bib_ref]. In addition, the original organizational/activational hypothesis of sexual differentiation of the brain has been extended to include puberty [bib_ref] Organizing action of prenatally administered testosterone propionate on the tissues mediating mating..., Phoenix [/bib_ref] [bib_ref] Back to the future: the organizational-activational hypothesis adapted to puberty and adolescence, Schulz [/bib_ref]. Namely, ovariectomy in peripubertal period leads to a long-lasting postponement/alleviation of the postpubertal decline in thymopoiesis [bib_ref] Age-associated remodeling of thymopoiesis: role for gonadal hormones and catecholamines, Leposavić [/bib_ref] [bib_ref] Peripubertal ovariectomy provides long-term postponement of age-associated decline in thymic cellularity and..., Perišić [/bib_ref]. This could be partly related to ovariectomy-induced changes in thymocyte proliferation [bib_ref] Neonatal androgenization affects the efficiency of β-adrenoceptor-mediated modulation of thymopoiesis, Radojević [/bib_ref]. Given that the agerelated decline in thymopoiesis has been partly related to the rise in the thymic noradrenaline level [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Age-associated plasticity of α1-adrenoceptor-mediated tuning of T-cell development, Leposavić [/bib_ref] [bib_ref] Age-associated alterations in sympathetic noradrenergic innervation of primary and secondary lymphoid organs..., Thyagarajan [/bib_ref] , one may assume that the peripubertal ovariectomy affects thymic adrenergic networks. Indeed, it was shown that it diminishes the thymic noradrenaline level in young adult (2-month-old) rats [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. This reflected the decrease in the density of noradrenergic nerve fibers and noradrenaline content in both noradrenergic nerve fibers and non-lymphoid cells, as thymocyte noradrenaline content increased (45) . These changes were preventable by estrogen supplementation [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. This could be explained by the following facts: (i) estrogen represents the key factor in remodeling of noradrenergic innervation in some other tissues [bib_ref] Differential effects of oestrogen on developing and mature uterine sympathetic nerves, Chávez-Genaro [/bib_ref] and (ii) is implicated in the regulation of TH expression [bib_ref] Sex differences in the regulation of tyrosine hydroxylase gene transcription by estrogen..., Thanky [/bib_ref]. Estrogen is suggested to regulate TH gene expression through direct genomic effects, as the TH promoter contains several elements, including the activation protein 1 and Sp1/Egr1 motifs that might mediate estrogen action on TH gene [bib_ref] Interactions between Egr1 and AP1 factors in regulation of tyrosine hydroxylase transcription, Nakashima [/bib_ref] [bib_ref] Sequences that direct rat tyrosine hydroxylase gene expression, Fung [/bib_ref]. The thymic cell type-specific effects of peripubertal ovariectomy on TH expression could be explained by data indicating that estrogen may regulate TH transcription in opposite direction through ERα and ERβ [bib_ref] Connections and regulation of the human estrogen receptor, Mcdonnell [/bib_ref]. Given that estrogen may influence TH expression trough extragenomic and indirect genomic effects, it may also be supposed that estrogen, through the same ER, may produce opposing effects by interacting with proteins with distinct action on gene transcription in distinct cells [bib_ref] Response of tyrosine hydroxylase and GTP cyclohydrolase I gene expression to estrogen..., Serova [/bib_ref] [bib_ref] Effects of sex and estrogen on tyrosine hydroxylase mRNA in cultured embryonic..., Raab [/bib_ref]. In peripubertally ovariectomized rats, the density of noradrenergic nerve fibers and TH-expressing non-lymphoid cells remained lower than in age-matched controls until the age of 11 months [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. On the other hand, thymocyte noradrenaline, which was elevated in 2-month-old peripubertally ovariectomized rats, continued to rise until the age of 11 months [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. In 11-month-old peripubertally ovariectomized rats it was comparable with controls [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. Thus, it seems that the ovariectomy-induced changes are long lasting [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref].
In addition, peripubertal ovariectomy in young adult rats diminished the average thymocyte surface density of β2-AR, but it increased that of α1-AR (reflecting estrogen, and estrogen and progesterone deficiency, respectively) (45) . These changes, despite the rise in circulating estrogen level postovariectomy because of extragonadal synthesis [bib_ref] Extragonadal aromatization increases with time after ovariectomy in rats, Zhao [/bib_ref] , remained stable until the age of 11 months [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. This could be related to a decreased sensitivity to estrogen action, as a consequence of peripubertal hormone misprinting. Finally, it is noteworthy that the increased noradrenaline content in thymocytes and diminished frequency of β2-AR + thymocytes in young adult ovariectomized rats suggested that peripubertal ovariectomy instigates some signs of thymic defeminization/masculinization [bib_ref] Sexual dimorphism in the catecholamine-containing thymus microenvironment: a role for gonadal hormones, Pilipović [/bib_ref] [bib_ref] Cellular and nerve fibre catecholaminergic thymic network: steroid hormone dependent activity, Leposavić [/bib_ref].
The putative role of peripubertal orchidectomy in longlasting reprogramming of the thymic adrenergic networks has not been examined. However, 1 month following peripubertal orchidectomy the changes in both extrinsic and intrinsic noradrenergic networks were similar to those described 1 month following ovariectomy in the same age [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] [bib_ref] Peripubertal ovariectomy influences thymic adrenergic network plasticity in adult rats, Pilipović [/bib_ref]. In addition, an impaired β-AR-mediated influence on thymus led to more efficient thymocyte positive selection/less efficient negative selection, and preferential differentiation/maturation of thymocytes into mature CD4 + CD8 − TCRαβ high cells in orchidectomized rats [bib_ref] Gonadal hormone dependent developmental plasticity of catecholamine:β2-adrenoceptor signaling complex in male rat..., Pilipović [/bib_ref] , i.e., to a more "feminine" pattern of T-cell development [bib_ref] Gonadal steroids and immunity, Olsen [/bib_ref].
# Conclusion
In summary, a growing body of evidence indicates that both thymic sexual differentiation and involution are, at least partly, "controlled" during the critical developmental periods by gonadal steroids. In addition, it suggests that the gonadal steroid-mediated thymic (re)programming involves extrinsic and intrinsic noradrenergic regulatory networks and AR expression on thymic cells. The challenge remains to elucidate the molecular mechanisms underlying these gonadal steroidinduced effects. Nonetheless, it may be assumed that (i) alterations in circulating levels of gonadal steroids during the critical developmental periods (either induced endogenously or by endocrine disruptors in the environment) lead to long-lasting effects on thymopoiesis and (ii) pharmacological manipulation with (nor)adrenaline action on thymus may be useful means in preventing/moderating deleterious effects of aging on thymopoiesis. Gonadal Steroids and Thymic (Re)programming Frontiers in Endocrinology | www.frontiersin.org January 2018 | Volume 9 | Article 13
# Author contributions
GL wrote the draft version, and IP designed the figure, whereas GL and IP equally contributed to the editing of this manuscript.
## References
Conflict of Interest Statement:The authors declare that they have not any commercial or financial relationships that could be construed as a potential conflict of interest.Copyright © 2018 Leposavić and Pilipović. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication inthis journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
From incomer to insider: The development of the TRANSPEC model – A systematic review of the factors influencing the effective rapid and early career TRANsition to a nursing SPECiality in differing contexts of practice
ObjectiveShortages in the speciality nursing workforce, both nationally and internationally are driving the need for the development of an evidence-based model to inform recruitment and retention into speciality nursing practice. This study aimed to identify the factors influencing rapid and early career transition into speciality nursing practice.MethodsA comprehensive systematic review of the literature was undertaken using a convergent qualitative synthesis design where results from qualitative, quantitative and mixed methods studies were transformed into qualitative findings. Databases included CINAHL, Medline, Scopus and PsycINFO. Search terms were: nurse, early career, rapid career, transition, specialty, and Medical Subject Heading terms included: professional development and educational, nursing, and continuing. Using validated tools, papers were independently assessed by a minimum of two reviewers.ResultsTwenty-three research articles were included. There were no randomized control trials. Through thematic analysis and matrix mapping of the results, the TRANSPEC model was PLOS ONE | https://doi.
# Introduction
The need to re-examine the numbers and the types of nurses, and their role within the health workforce arises due to many factors including population drivers; economic considerations such as increased costs of medical advances and the use of technology to provide services; and problems with health workforce supply and mal-distribution [bib_ref] Challenges in health and health care for Australia, Armstrong [/bib_ref]. Population drivers include an increased demand on health services driven by the increased life expectancy of populations, particularly in developed countries, alongside the rising number of people living with chronic conditions who require more complex and long term treatment [bib_ref] Digital reablement-a personalised service to reduce admissions and readmissions to hospitals and..., Doughty [/bib_ref] [bib_ref] Health workforce design for the 21st century, Duckett [/bib_ref]. In Australia, a major driver of the change in health service delivery has been the 50% of the population with at least one chronic condition, as 40% of all preventable hospital admissions are related to chronic conditions. Servicing the need for this population alone has driven up the cost of provision of services. Described as a "growing health crisis", the increased cost for treatment has resulted in an increased acuity of patients in the acute care system, with a flow on effect of earlier discharge in an attempt to reduce the length of hospitals stays. The flow on effect has meant that patients who are sicker are being cared for in their homes by multidisciplinary teams from numerous specialty areas and service providers. The cost of medical devices, particularly the increase in availability and use of technology for diagnosis and treatment [bib_ref] Challenges in health and health care for Australia, Armstrong [/bib_ref] and Tele-health provision, particularly to populations in rural and remote areas, has also driven the cost of delivery [bib_ref] Digital reablement-a personalised service to reduce admissions and readmissions to hospitals and..., Doughty [/bib_ref]. These changes have resulted in a need to re-examine traditional ways that health professionals provide services within hospitals and the community.
Complicating this role redefinition, has been workforce supply and mal-distribution. Similar to its population, Australia has an ageing health workforce, many of which have reduced their working hours as they prepare for retirement. The preference of many health professionals, particularly medical practitioners, to work in metropolitan areas, has resulted in a mal-distribution of that workforce. While this mal-distribution has been the focus of the Australian Government for many years, it is now increasingly evident that, while the mal-distribution of medical practitioners may continue, driving a shortage in rural and remote areas, there is now a growing overall shortage of nurses in all geographical areas as well as in specific areas of speciality nursing practice [bib_ref] Specialist outreach services in regional and remote Australia: key drivers and policy..., O'sullivan [/bib_ref]. The overall response of the Australian government to address the mal-distribution and shortage has been to increase supply. Similar shortages have occurred in other countries such as the United States of America and the United Kingdom, for several decades [bib_ref] Qualities of exemplary nurse leaders: perspectives of frontline nurses, Anonson [/bib_ref] [bib_ref] The nursing shortage in the United States of America: an integrative review..., Goodin [/bib_ref]. However, Buchan and Aiken [bib_ref] Solving nursing shortages: a common priority, Buchan [/bib_ref] also suggest that the reasons for the shortages are complex and not limited to actual numbers entering the profession.
In Australia there are three levels of nurses. Two levels are regulated by the National Body: Australian Health Practitioner Agency (AHPRA),-the registered nurse (RN) and the enrolled nurse (EN). The RNs complete a university degree (Master or Bachelor), while the ENs mostly graduate with a diploma obtained from a technical college. The third level are unregulated care providers known as assistants in nursing (AIN) or Personal Care Workers (PCWs), who may or may not have certificate qualifications. The three levels of nurses work across all care settings, with the registered nurse held responsible for the care that is delegated to ENs/AINs. There are also nurse practitioners (NP) who were introduced to the workforce in 2001. NPs are experienced RNs who work as advanced clinicians following the completion of a specialised master's degree. NPs are endorsed through AHPRA, have limited prescribing rights and manage their own caseload.
In Australia, it is predicted that there will be a shortfall of 85,000 nurses by 2025 mostly due to increased work intensification coupled with recruitment and retention issues [bib_ref] Digital reablement-a personalised service to reduce admissions and readmissions to hospitals and..., Doughty [/bib_ref]. The shortages have been noted in all areas of nursing, and in particular at the RN level, in specialities such as critical care, neonatal, aged care, emergency nursing, rural and remote areas and mental health nursing. Internationally, the World Health Organization has predicted a global shortage of 7.6 million by 2030.The focus on supply has resulted in more newly graduated RNs than places for their employment, and, as several years of nursing experience is normally a pre-requisite for entry into specialised practice this creates a delay in recruitment in areas where there are already shortages [bib_ref] Solving nursing shortages: a common priority, Buchan [/bib_ref]. Thus, there is a need to examine the factors influencing recruitment and retention into speciality nursing practice.
RNs in Australia, similar to other countries, work as either generalists or specialists. For the purpose of this study, generalist RN practice has been defined as:
Encompassing a comprehensive spectrum of activities. It is directed towards a diversity of people with different health needs. It takes place in a wide range of health care settings and it is reflective of a broad range of knowledge and skills. Generalist practice may occur at any point on a continuum from beginning to advanced.
Speciality practice has been defined as a practice that follows and builds on generalist practice..
## It focuses on a specific area of nursing and is directed towards a defined population or a defined area of activity and it is reactive of depth of knowledge and relevant skills. specialist practice may occur at any point on a continuum from beginning to advanced
To enable a RN to move from generalist to specialist, transition programs have been developed. These programs are taken post-registration as either formal degrees or as continuing development programs. The latter may articulate into a formal degree. The programs can be taken within five years of completing a pre-registration or after some time within the workforce [bib_ref] Retention of early career registered nurses: the influence of self-concept, practice environment..., Mills [/bib_ref]. These programs are designed to: "assist the newly graduated or transferred nurse/ midwife to acquire further general and speciality knowledge and skills in a logical, sequenced supported approach to effectively transition to work expectations".
Transition programs are often delivered by employers as a way of 'growing their own' specialists. In the current climate of projected and real shortages, many employers have therefore designed and delivered their own programs. Thus, a transition program into peri-operative nursing, for example, can vary considerably from one employer to another as there are no national standards to regulate the content of the program.
This study was commissioned by the State Government of Queensland's Health Department. The need for this study was a growing shortage of experience registered specialist nurses in several areas, specifically pre-operative, paediatrics, neonatal, critical care, mental health and rural and remote area nursing. To overcome the current and projected future shortages, this study sought to understand the factors that influence recruitment into and retention within these nursing specialities from both an early career (defined as within 5 years of registration) and rapid transition (after 5 years) perspectives [bib_ref] A systematic review of the factors influencing the effective early career transition..., Hegney [/bib_ref]. This study was designed, therefore, to gather evidence-based data to support the rapid and early career transition of a RN from a generalist to a specialist role.
# Methods
The overall study aimed to provide an evidence-based model for early career and rapid specialisation to nursing specialities. The project contained three studies: a systematic review of the literature, interviews with RNs who had participated in a recent transition to speciality practice program and a Delphi study to confirm findings of stages 1 and 2 and to provide a model on which to base policy and education programs. The aim of study one, the systematic review, was to identify, assess and summarize available evidence relating to early or rapid career transitioning into a specialty area of nursing practice. The specialty areas considered for this review were peri-operative, mental health, neonatal, critical care, rural and remote area and emergency nursing. The primary outcome of the review was to establish a model or framework for successful early and rapid transition into speciality practice. As a secondary outcome, the review was designed to inform the healthcare and education industry of the inhibitors and enablers, influencing an effective transition into specialty areas.
## Design
This systematic review was designed and reported in accordance with international guidelines: the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref]. Full details of the overall search strategy can be found in the research protocol, registered with the International Prospective Register of Systematic Reviews (PROSPERO) [bib_ref] A systematic review of the factors influencing the effective early career transition..., Hegney [/bib_ref]. A convergent qualitative synthesis design, as outlined by Pluye and Hong, was used where results from qualitative, quantitative and mixed methods studies were transformed into qualitative findings.
## Search strategy
The search of the literature was undertaken from February to May, 2018. Articles were identified through a search of CINAHL, Medline, Scopus and PsycINFO databases. The search terms employed were nurse, early career, rapid career, transition, specialty, and Medical Subject Heading (MeSH) terms related to professional development, educational, nursing and continuing. Studies undertaken within Australia, New Zealand, United Kingdom, Canada, United States and Europe were included in the search. The results from the search were imported into EndNote X8 to remove duplicates and manage the references through the different stages of the review (full title, abstract, and full text). Titles and/or abstracts of studies were retrieved using the search strategy. Hand searching was also used to screen and identify studies that may have been missed. All papers were screened independently by two review authors to identify studies that potentially met the inclusion criteria. Inclusion criteria were: critical care, peri-operative, mental health, rural and remote, neonatal and emergency nursing, early career, rapid transition, nursing, published in English from 2007-2018. While it is recognised that nursing shortages have been a foci for several decades, the ten year time frame was selected as the shortages within nursing specialities, particularly in Australia, have accelerated, particularly influenced by the anticipated retirement of the aged workforce. Systematic or literature reviews were excluded from the study as were articles from the grey literature and research papers involving transition to general nursing (specifically new-graduate transition to generalist nursing) and all other nursing specialities.
## Risk of bias (quality) assessment
The quality of the relevant literature was appraised by two authors who independently assessed for risk of bias. Due to the lack of randomized controlled trials, and the predominance of surveys, quantitative papers were critically appraised using the Critical Appraisal Questions for Surveys (CAQS) tool. Qualitative papers selected for retrieval were assessed by two independent reviewers for methodological validity prior to inclusion in the review using standardized critical appraisal instruments from the Joanna Briggs Institute Qualitative Assessment and Review Instrument (JBI-QARI). Any disagreements that arose between the reviewers were resolved through discussion, or with a third reviewer. Three qualitative manuscripts that did not demonstrate Human Research Ethical or International Review Board Clearance were excluded from the study.
To avoid publication bias, each of these studies were assessed for overlap between sub-studies, however, none of the data appeared to repeat. Disagreements between appraisal scores were resolved through discussion and an acceptable quality appraisal. A cut-off score of 6 out of 10 for quantitative studies using the CAQS tool was adopted. Reasons for exclusion were: insufficient information on recruitment processes; measurement bias unclear; confounding factors in design and/or analysis unclear; results flawed; insufficient discussion of implications for transferability or generalizability. For qualitative studies a cut-off score of 7 out of 10 was adopted. Reasons for exclusion were: lack of congruity between the research methodology and the methods used to collect the data; lack of congruity between the research methodology and the interpretation of results; the influence of the researcher on the research and vice versa were addressed; participants and their voices are adequately represented; conclusions drawn in the research report flow from the analysis, or interpretation, of the data [bib_ref] Checklist for qualitative research, Briggs Institute [/bib_ref]. Mixed methods studies were included if the quantitative and qualitative sub-study scored above the identified appraisal cut-off. The results of the literature search uncovered 23 articles to be included in the review (12 quantitative, 8 qualitative, 3 mixed methods) [fig_ref] Fig 1: The PRISMA flowchart [/fig_ref] [fig_ref] Table 1: Articles included in the systematic review [/fig_ref]. The PRISMA flowchart [fig_ref] Fig 1: The PRISMA flowchart [/fig_ref] was used to summarise the selection process with reasons for exclusion of the study [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref]. The data were analysed using matrix code analysis.
## Data synthesis
The quantitative and mixed method full text papers could not be used for a quantitative pooled analysis due to the diverse range of results and small sample sizes. We used a qualitative coding template thematic analysis [bib_ref] Doing template analysis, King [/bib_ref] where extracts were coded into broad themes and uploaded electronically into NVIVO 11for matrix code analysis. This revision of coding produced templates that were flexibly structured as a result of emergent and unanticipated datainto a matrix code analysis and frequencies. This formed a framework for the model, which was confirmed and then expanded by the qualitative meta-synthesis.
Independently to the matrix code analysis and frequencies, qualitative research findings were pooled using the JBI Qualitative Assessment and Review Instrument (JBI-QARI). Overall job satisfaction was significantly higher in the post-internship group compared to the preinternship group. No statistically significant difference was found between job satisfaction and birth year. Statistically significant differences were found between night shift and 3 areas of job satisfaction: ability to identify work resources, ability to manage demands of job and having information to perform job effectively. Improved job satisfaction was also reflected in lower turnover rates (12% in post-implementation and 20% in pre-implementation groups). . Nurse graduate nurses (NGNs) who were younger (<23 years) were more satisfied with the practice environment compared with those who were older (p = 0.024). Three variables-unite satisfaction, satisfaction with clinical supervision and being assigned to critical-care areas were forerunners to be independent and statistically significant predictions of NGNs satisfaction. 32.5% of the variance in PES-AUS scores was explained by these.
Klingbeil, C., Schiffman, R. F., Ziebert, C., Totka, J. P., Schmitt, C. A., Doyle, L., . . . & Johnson, N.2016 Workplace empowerment is defined as 4 contextual factors-opportunity, information, support and resources. There was a statistically significant relationship between certification and overall perception of empowerment. The 3 patient outcome variables were: rate of central line catheter-associated blood stream infection, rate of Ventilator-Associated Pneumonia (VAP) and prevalence of pressure ulcers as likely to be sensitive to nursing practice. No statistically relationship was detected between proportion of certified nurses on a unit and patient outcomes.
(Continued )
This involved the synthesis of findings to generate a set of statements that represented aggregation, through assembling the findings rated according to their quality, and categorizing these findings on the basis of similarity in meaning. These categories were then subjected to a metasynthesis in order to produce a single comprehensive set of synthesized findings that were formulated into an over-arching theme, three phases and three concepts. The early findings were discussed by the quantitative and qualitative teams and from this discussion, the over-arching theme, the three phases and three concepts further emerged. The over-arching theme, phases and concepts were then plotted conceptually and from this aggregation, the TRANSPEC Model was developed which is discussed in detail in the following section.
# Results
## Overall
As described in [fig_ref] Table 1: Articles included in the systematic review [/fig_ref] , there were twelve quantitative papers , eight qualitative papers [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] [bib_ref] Using eDelphi to identify capability requisites for postgraduate certificate in Neonatal Intensive..., Bromley [/bib_ref] [bib_ref] Valuing teamwork: Insights from newlyregistered nurses working in specialist mental health services, Cleary [/bib_ref] [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] [bib_ref] Meeting the needs of new graduates in the emergency department: A qualitative..., Glynn [/bib_ref] [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] and two mixed method studies [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref].
## Model development
The thematic analysis and matrix mapping that identified the over-arching theme (context of practice), the three phases (pre-entry, incomer and insider) and three concepts (the self, transition processes and sense of belonging) were then mapped conceptually to form the TRANS-PEC model (see [fig_ref] Fig 2: The effective early career and rapid TRANSition to a Nursing SPECiality in... [/fig_ref]. As we can see from [fig_ref] Fig 2: The effective early career and rapid TRANSition to a Nursing SPECiality in... [/fig_ref] central to successful transition are the concepts of the self (both personal and professional) and the development of a sense of belonging within the speciality and the context of practice. The formal and informal transition processes such as mentoring, shadowing, formal curricula, continuing professional educational short courses are a necessary part of the transition as this is the theoretical and clinical content required for speciality practice. Influencing all of this is the over-arching theme of context of practice, which can be a combination of: the geographical location; the health service type, policies and procedures; the speciality; and the links that the nurse has to other members of the multi-disciplinary team and the community in which the health service is located.
In the context of this model, the transitioning clinician goes through three distinct phases: pre-entry, incomer and insider. At each phase of the transition, the enablers and inhibitors influence the transition and therefore recruitment and retention. For example, pre-entry experiences such as a positive clinical placement in pre-registration programs may affect how attractive the speciality appears to the nurse and will affect steps to move into or away from the specific speciality. The TRANSPEC model [fig_ref] Fig 2: The effective early career and rapid TRANSition to a Nursing SPECiality in... [/fig_ref] demonstrates the interactive nature of the overarching theme, the phases and concepts.
The findings from Tables 2, 3 and 4 are now discussed. [fig_ref] Table 2: Template and matrix coding analysis with the frequency of codes per study... [/fig_ref] presents the matrix mapping which resulted in the over-arching theme and the three concepts being identified by individual papers. Within these papers, it is noted at what phase or phases these arose. Additionally, [fig_ref] Table 2: Template and matrix coding analysis with the frequency of codes per study... [/fig_ref] shows the count or frequency that the over-arching theme and/or the concept or concepts were identified within each paper. [fig_ref] Table 3: Context of Practice [/fig_ref] add to the quantitative findings demonstrating the over-arching theme [fig_ref] Table 3: Context of Practice [/fig_ref] and concepts [fig_ref] Table 4: (Continued) [/fig_ref] and findings within each phase. The enablers and inhibitors identified in each qualitative manuscript are also provided from this analysis [fig_ref] Table 3: Context of Practice [/fig_ref]. Thus the qualitative data have further informed the quantitative findings allowing for an understanding of specific inhibitors and enablers that are influencing the concepts at each phase as well the context of practice.
## Overarching theme: context of practice
In this model, the term context of practice is used to describe the interconnected factors, experiences, and opportunities, which enable or inhibit a nurse's progress across the continuum of speciality practice within the working environment. Therefore, the context of specialty nursing practice is framed by professional and organisational elements. The professional element is based on a core body of nursing knowledge that is continually developed and refined by practice, research and innovation. Organisational elements impacting on specialty nursing practice include the geographical location, size and capability of the health service, policies and procedures that influence the scope of practice, the community in which the health care facility is located and the diversity of health care delivered by the health service to that and other communities [bib_ref] Code of Conduct for Nurses, Nursing [/bib_ref].
The frequency of codes for the over-arching theme 'context of practice' ranged from 17 to 178 [bib_ref] Knowledge and learning in speciality practice, Spence [/bib_ref]. Spence et al. [bib_ref] Knowledge and learning in speciality practice, Spence [/bib_ref] maintain that a context of practice underpinned by an active learning environment, strong educational processes, and contemporary engineered models of care is important for specialist practice. Further, nurses prefer to acquire specialist knowledge and skills in the practice environment and need support to optimise this learning [bib_ref] Knowledge and learning in speciality practice, Spence [/bib_ref].
At the pre-entry phase, there were no identified inhibitors or enablers in the literature. As an incomer, there were several inhibitors including: economic cutbacks that limited the scope [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] (Continued ) of experience [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] and the information and communications technology (ICT) environment, which were seen as a major challenge to care delivery in a new context. Specifically, nurses noted how they struggled with the technology and different ICT systems. This was particularly the case in critical care environments. One enabler at the incomer phase was the presence of trust of the incomer [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] , and this was particularly important in the mental health environment. As an insider, the major influence, seen as an inhibitor was the lack of positions to retain nurses who had completed their transition process/program [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref].
## Concept one: the self
The concept of 'the self' is seen as a complex system steeped in philosophical, psychological and sociological arguments. Thagard argues that the self is:
A multilevel system consisting of social, individual, neural and molecular mechanisms. . . Each of these levels can be understood as a sub-system consisting of environmental influences,
## Component parts, interconnections between parts, and regular changes in the properties and relations of the parts.
Thagard [bib_ref] The self as a system of multilevel interacting mechanisms, Thagard [/bib_ref] goes on to note that each of these levels are influenced by additional phenomena such as 'self-presentation, self-esteem, self-enhancement; self-regulation; self-expansion and self-development'. For the purpose of this study the concept of 'the self' is defined as an individual who is constantly changing and adjusting. These changes are driven by self-preservation, self-esteem, self-enhancement, self-regulation, self-expansion and self-development at two levels. For the professional nurse, these can be described as changes to the personal and the professional self.
Frequency of the codes for the concept 'the self' ranged from four [bib_ref] Education and experience make a difference: results of a predictor study, Gillespie [/bib_ref] to 120 [bib_ref] The career development year: responding to the emergency nursing shortage in Australia, Morphet [/bib_ref] [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref]. As reported by Morphet et al. [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] , enhancement of the personal and professional self occurs by the use of an evidence-based transition to speciality practice framework to guide the structure, process and outcomes of speciality practice. Morphet et al. [bib_ref] The career development year: responding to the emergency nursing shortage in Australia, Morphet [/bib_ref] reported that a dedicated "career development year" enriches the development of the personal and professional self and translates into greater recruitment and retention of speciality nurses.
Personal self. At the pre-entry phase, confidence was seen as both an enabler and an inhibitor [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref] , while at the incomer stage willingness to learn [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] , self-care [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] , and confidence [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] [bib_ref] Valuing teamwork: Insights from newlyregistered nurses working in specialist mental health services, Cleary [/bib_ref] were seen as both inhibitors and enablers. Coping [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] , being overwhelmed by the environment [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] and concealing lack of experience in the speciality [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] were inhibitors. In contrast, motivation and commitment [bib_ref] Valuing teamwork: Insights from newlyregistered nurses working in specialist mental health services, Cleary [/bib_ref] and job satisfaction [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] were seen as enablers. There were no enablers or inhibitors identified for the personal self at the insider phase.
Professional self. There were no enablers or inhibitors identified in the pre-entry phase. At the incomer phase, willingness to learn [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] was an enabler and the level of knowledge was seen as an inhibitor [bib_ref] Meeting the needs of new graduates in the emergency department: A qualitative..., Glynn [/bib_ref]. At the insider phase, competence [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] [bib_ref] Using eDelphi to identify capability requisites for postgraduate certificate in Neonatal Intensive..., Bromley [/bib_ref] was seen as both an inhibitor and enabler. Keeping up to date [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] , willingness to learn [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] and confidencewere seen as enablers.
## Concept two: transition processes
For the purpose of this paper, transition processes refer to the processes or period of changing from one phase to another. The time to complete the transition processes will vary from one nurse to another as will the transition processes they undertake. Influences will include the context of practice, the beginning level knowledge and skills, the nature of the processes being undertaken and the availability of support. Transition processes can be formal and/or informal. In some instances, both formal and informal processes were used, and in other cases only one was used.
Formal processes. Formal transition processes are defined as processes that involve educational and clinical support and are designed to facilitate the transition. The inhibitors and enablers identified related to these formal processes and the nurse's ability to access them.
This concept had the highest frequencies of quantitative matrix coding as most of extracts described outcomes from transition programs or educational interventions. In the highest matrix coding frequency extract, 154 codes [bib_ref] Knowledge and learning in speciality practice, Spence [/bib_ref] , found that nurses with less than one year speciality experience typically had the lowest 'critical' knowledge essential for safe practice, with the majority of novice nurses showing a preference for experiential 'on-the-job' learning. Gillespie's [bib_ref] Education and experience make a difference: results of a predictor study, Gillespie [/bib_ref] paper with a matrix coding frequency of 141, reported that although both years of experience and specialty education contributed to nurses' perceived competence, experience was the strongest contributor.
The transition processes [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] were regarded as an enabler at the pre-entry phase. At the incomer phase, preparation processes embedded in the reality of practice, the way the program prepared the nurse for further education and feedback on performance were identified as enablers [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref]. Additionally, the quality of the processes was also seen as an enabler [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref]. The volume of knowledge to be learnt [bib_ref] Using eDelphi to identify capability requisites for postgraduate certificate in Neonatal Intensive..., Bromley [/bib_ref] [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] and the ability to access the transition processes [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] were both enablers and inhibitors. Insider enablers were the way that they processes had enhanced successful transition [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] , the accreditation of the processes into post-graduate programs [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] , embedding learning into practice [bib_ref] Using eDelphi to identify capability requisites for postgraduate certificate in Neonatal Intensive..., Bromley [/bib_ref] were all enablers. There were no inhibitors found at the insider phase.
Informal processes. No inhibitors or enablers were found at the pre-entry phase. For insiders, supportive staff [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] and role models [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] were both enablers and inhibitors. Observational learning [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] was an enabler and conflicting information was an inhibitor. Support for continuing professional education was both an enabler and inhibitor [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] at the incomer phase.
## Concept three: sense of belonging
Baumeister and Leary [bib_ref] The need to belong: Desire for interpersonal attachments as a fundamental human..., Baumeister [/bib_ref] define a sense of belonging as the degree to which an individual perceives as being accepted, included and supported by others in their environment. When people feel they belong, positive emotional and cognitive outcomes occur [bib_ref] The role of a sense of school belonging in understanding the effectiveness..., Prince [/bib_ref]. A sense of belonging can be experienced in many contexts [bib_ref] The ASPIRE principles and pedagogy for the implementation of social and emotional..., Roffey [/bib_ref]. It relates to the inclusiveness of all participants where a culture of connection is built through relationships to promote a sense of belonging [bib_ref] Why do women opt out? Sense of belonging and women's representation in..., Good [/bib_ref]. The within this concept three sub-concepts emerged as important to a successful transition. These were: were the culture of the organisation, the working conditions and the level of support. The over-arching theme of the context of practice, also influenced this concept and the enablers and inhibitors within it.
Culture. Culture is defined as the totality of the relationships formed within the culture, and shapes how those relationships are structured [bib_ref] Reflecting on Holistic Relationships: A definition of culture, Oehlert [/bib_ref]. The highest frequency of matrix codes was found in Morphet et al. [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] , 35 codes [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] , followed by Halfer [bib_ref] The organizational impact of a new graduate pediatric nurse mentoring program, Halfer [/bib_ref] and Hussein [bib_ref] Predictors of new graduate nurses' satisfaction with their transitional support programme, Hussein [/bib_ref] , both 29 codes. From these studies, it was reported that speciality graduates are eager to establish goals for lifelong learning, specialty-nursing certification, postgraduate study, clinical ladder advancement, clinical governance participation, and professional organizational membership [bib_ref] The organizational impact of a new graduate pediatric nurse mentoring program, Halfer [/bib_ref]. Also, it is imperative for nursing management to create opportunities for new speciality nurses and more experienced staff to develop closer working relationships, and improve retention through team projects and other department 'bonding' activities as these engender a sense of 'belongingness' among the new speciality nurses [bib_ref] The organizational impact of a new graduate pediatric nurse mentoring program, Halfer [/bib_ref]. There were no identified inhibitors or enablers at the pre-entry phase. As an incomer, job satisfaction [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] and resource investment [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] were enablers while feedback on performance was an inhibitor [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref]. Staff personalities [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] and team support [bib_ref] Valuing teamwork: Insights from newlyregistered nurses working in specialist mental health services, Cleary [/bib_ref] [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] were both inhibitors and enablers. The culture of the organisation at the insider phase was enabled by positive feedback [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] team support/acceptance [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref] [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] and promoting professional development [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref].
Working conditions. Nurses, in the majority, are in paid work. Working conditions for paid work normally are considered to include working time (hours of work, rest periods and work schedules), remuneration and the physical conditions and mental demands that exist in the workplace. Working conditions for the purpose of this study include: skill-mix, staffing ratios, remuneration, role clarity, support, sufficient staff, workload allocation and role clarity. There were no matrix codings for this sub-concept. In the pre-entry phase the inability of organisation to pre-empt staff shortages was identified [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref]. At the incomer phase, remuneration and staffing levels [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] were inhibitors, whereas workload allocation was an enabler [bib_ref] Meeting the needs of new graduates in the emergency department: A qualitative..., Glynn [/bib_ref]. There were no inhibitors or enablers identified at the insider phase.
Level of support. Support, for the purpose of this study, was defined as support within the workplace and was provided by colleagues and management. There were no matrix codes for this sub-concept. At the pre-entry phase there were no enablers or inhibitors. At the incomer phase, support from colleagues [bib_ref] Valuing teamwork: Insights from newlyregistered nurses working in specialist mental health services, Cleary [/bib_ref] [bib_ref] A framework for transition to speciality practice programmes, Morphet [/bib_ref] was an enabler, while perceptions of competent management of a ward or unit [bib_ref] Using eDelphi to identify capability requisites for postgraduate certificate in Neonatal Intensive..., Bromley [/bib_ref] was an inhibitor. Management support for their role [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] was both an inhibitor and an enabler. At the insider phase, the opportunities to become a leader [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] were seen as an inhibitor and being supported, valued and accepted [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] and empowered [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref] were enablers.
# Discussion
The main concepts identified in the transition process included: the Self (personal and professional), the Transition Processes (formal and informal), and a Sense of Belonging. The data also clearly identified three phases to the transition: Pre-entry, Incomer and Insider. Influencing successful transition at each phase and within each theme were the identified enablers and inhibitors. It was apparent from the data analysis that at each phase, the three major concepts of the Self, Transition Processes and the Sense of Belonging occurred. It was possible therefore to map these conceptually (hence TRANSPEC model emerged). It was also apparent that the over-arching theme of the Context of Practice, particularly the geographical location and the speciality, influenced the transition. Thus it was possible to enclose the phases and concepts within an over-arching theme of the Context of Practice. To inform the development of a transition program, it was possible also to identify enablers, inhibitors or both in each phase and concept. Thus the model developed from the data analysis.
## Context of practice
The model acknowledges the importance of the context of practice in both early career and rapid transition to specialty practice. There are many ways that the context can impact on the scope of practice on a nurse and therefore the level of professional and personal experience, skills, knowledge and competence required for practice. For example, geographical location of the health service has been found previously to influence: a) access to health services; and b) the scope of practice of the nurse [bib_ref] Characteristics of chemotheraphy practice in rural and remote health facilities in Queensland, Mccarthy [/bib_ref]. Additionally, while there were many similarities in the influence of the context of practice across specialities, there were also differences that needed to be taken into consideration. For example, mental health specialities saw the need for personal and team safety [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] , whereas nurses working in the critical care environment highlighted the complexity of practice caused by the technology used. A third major influence on the context of practice, which has been well documented previously, are the organisational influences (e.g. funding, scope of practice restrictions, management styles).
The TRANSPEC model acknowledges the importance of context of practice and allows for each speciality within a discipline to identify specific enablers, inhibitors or both that limit recruitment and then retention. It therefore allows for the different contexts of speciality practice.
## The self (professional and personal)
The professional self. Nationally and internationally, speciality practice has been described in terms of the movement from novice to expert. Benner's model in particular has provided clear guidance for what is needed to advance professional practice. However, to date research has focused on specific specialities, for example, critical care, emergency department, mental health, neonatal, paediatric, rural and remote and so on. This has resulted in research and standards of practice for specialities being developed in silos.
Additionally, transition to speciality practice research has focused on nurses at the incomer and insider phase. The model demonstrates that entry to professional speciality practice begins when the speciality practice is being considered as a career path (pre-entry). The importance of experience during pre-registration clinical placements and its impact on recruitment into speciality practice has been highlighted by this review [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref] [bib_ref] The developing role of transition to practice programs for newly graduated mental..., Procter [/bib_ref]. This review and one recent study demonstrate the importance of enablers and inhibitors at pre-entry. Yet, these have received limited attention in research [bib_ref] Impact of clinical placements on graduates' choice of first staff-nurse post, Wareing [/bib_ref].
The personal self. The findings demonstrate how the personal self is influenced by confidence, anxiety about a new working environment and coping within it [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref]. Attention to self-care was seen as an important way to cope with the new environment [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref]. Previous research and policies have focused on building the competence (skills and knowledge) of the nurse as they transition within that speciality. Recent work highlights the importance of individual well-being of nurses and how factors such as burnout and resilience can not only influence retention within the workforce, but also the quality of patient care delivered [bib_ref] Workloads and contributing factors, and their impact on implicit care rationing, Hegney [/bib_ref].
## Transition processes (formal and informal)
Nursing has been acknowledged as a practice profession, with a combined theoretical knowledge (science of nursing) and the practice (art of nursing), being equally important [bib_ref] Recovering the Art of Nursing to Enhance Nursing Practice, Henry [/bib_ref]. Experiential learning, where the application of the theoretical knowledge informs practice, is important, with most health-related professions acknowledging the importance of on-the-job-training, mentorship and ongoing learning [bib_ref] Professional identity development: Learning and journeying together, Bridges [/bib_ref]. Life-long-learning (LLL) has become an important element of nursing practice through a requirement by AHPRA that all registered health practitioners provide evidence of continuing practice development, stipulating the annual minimum hours required for the continuance of registration. LLL is a combination of self-directed learning, organisational learning through in-service sessions, required competence sign-offs, and required skills and knowledge acquisition in accordance with the profession or practice [bib_ref] Self-directed learning competence assessment within different healthcare professionals and amongst students in..., Cadorin [/bib_ref]. It requires a practitioner to take control of their learning, to engage with their profession and to actively support quality improvement initiatives that support personal and service adaptation and flexibility through evidencebased practice [bib_ref] Self-directed learning competence assessment within different healthcare professionals and amongst students in..., Cadorin [/bib_ref]. When a practitioner is transitioning to practice, whether it be a new area of practice, or gaining experiential knowledge as a new practitioner, both the practitioner and the service employing them have an obligation to access and provide learning opportunities within the field of practice the employment occurs. These can be divided into formal and informal processes.
Formal processes. The formal curricula, on the whole, began when the nurse started in the new role. Few curricula focus on developing the nurse before they enter the speciality-in the recruitment or pre-entry phase. Procter et al. [bib_ref] The developing role of transition to practice programs for newly graduated mental..., Procter [/bib_ref] note the importance of a positive experience during entry-to-practice clinical placements as this experience has the potential to influence how attractive the student may see the speciality in the longer term.
Formal processes include targeted educational programs that are offered at tertiary institutions, and may be at Certificate, Diploma, Bachelor or higher degree level. For specialty practice areas, these are general postgraduate qualifications that build on existing general practice knowledge to support transition to speciality practice [bib_ref] The effect of Transition to Specialty Practice Programs on Australian emergency nurses'..., Morphet [/bib_ref]. An important feature identified was that of processes of being embedded in the reality of practice [bib_ref] Orientation to emergency nursing: perceptions of new graduate nurses, Patterson [/bib_ref] [bib_ref] The developing role of transition to practice programs for newly graduated mental..., Procter [/bib_ref]. The importance of the preceptor and/or mentor was emphasised with other studies into transition, for example Hooper et al. [bib_ref] Graduate nurses' experiences of mental health services in their first year of..., Hooper [/bib_ref] noted that they were important to assist with building and maintaining resilience and self-development. Conversely, the quality and level of support from preceptors and mentors can vary and thus impede the transition process [bib_ref] Transitioning into new clinical areas of practice: an integrative review of the..., Kinghorn [/bib_ref].
## Informal processes.
Informal processes noted in this study, employed during the transition processes included: role modelling which could be an enableror an inhibitor [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] ; observational learning [bib_ref] New nurses' experience of their role within interprofessional health care teams in..., Schwartz [/bib_ref] and general supportive staff (nurses or other health care professionals) [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref]. Similar to pre-entry exposure, the behaviour of staff already established in the area could be either an enabler or an inhibitor depending upon the characteristics of the individual staff member [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref]. Importantly, encouragement to pursue further education in the speciality was seen as an enabler [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref].
## Sense of belonging
The TRANSPEC model emphasises the importance of the development of a sense of belonging within the speciality, within the health service, within the multidisciplinary team and within the community. Hospitals, just like schools and other organizations do not exist in a vacuum. They respond to the social contexts in which they are immersed. Opportunities for participation, respect for and from participants in the work as well as general community as well as being valued and accepted, can aid in realising an individual's psychological need to belong. Such support provides a sense of belonging and a connectedness, in which social and cultural facets shape the individual and community identity [bib_ref] The role of a sense of school belonging in understanding the effectiveness..., Prince [/bib_ref]. When people feel they belong, positive emotional and cognitive outcomes occur [bib_ref] The role of a sense of school belonging in understanding the effectiveness..., Prince [/bib_ref]. There were three sub-concepts that impacted on the sense of belonging: the culture, the working conditions and the level of support provided. Additionally, was the over-arching theme of the context of the specialisation. The personalities of current employees within the organisation and their interaction with the new nurse were both inhibitors and enablers. The culture of the workplace, particularly leadership, has been shown in many previous studies to impact on transition entry to practice as well as retention within practice [bib_ref] The employee retention triad in health care: Exploring relationships amongst organisational justice,..., Perreira [/bib_ref]. The importance of other staff in the transition processes have also been highlighted by previous studies noting that promotion of a working environment that encouraged 'inter-professional collaboration' [bib_ref] Graduate nurses' experiences of mental health services in their first year of..., Hooper [/bib_ref] and was seen to be an enabler, while 'maltreatment, abuse and neglect' were seen to be inhibitors [bib_ref] Graduate nurses' experiences of mental health services in their first year of..., Hooper [/bib_ref]. In line with these findings, this study has highlighted the importance of positive feedback and team support and acceptance [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] [bib_ref] The relationship of religious self-identification to cultural adaptation among Iranian immigrants and..., Saghafi [/bib_ref]. This is at the core of a sense of belonging as this increases mutual trust and therefore respect, wellbeing and self-esteem. The impact of working conditions such as remuneration, workload allocation, staffing levels and lack of role clarity [bib_ref] Distinguishing the Clinical Nurse Specialist from other graduate nursing roles, Mohr [/bib_ref] have been raised previously as well as in this study [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref] [bib_ref] Meeting the needs of new graduates in the emergency department: A qualitative..., Glynn [/bib_ref] [bib_ref] Graduate nurses' experiences of mental health services in their first year of..., Hooper [/bib_ref] [bib_ref] Transitioning into new clinical areas of practice: an integrative review of the..., Kinghorn [/bib_ref]. The impact of workload and the working environment on the personal self has also been documented previously [bib_ref] The Effects of Occupational Violence on the Well-being and Resilience of Nurses, Rees [/bib_ref]. The time spent in the speciality is also seen as an enabler with insiders having a successful transition if allowed time to do so [bib_ref] Transitioning into new clinical areas of practice: an integrative review of the..., Kinghorn [/bib_ref]. Support from colleagues and management also impacted on the sense of belonging. Leadership was demonstrated by either example [bib_ref] Capability: How is it recognised in student nurses undertaking postgraduate studies in..., Bromley [/bib_ref] or by opportunities [bib_ref] A qualitative descriptive exploration of the educational and career plans of early..., Coughlan [/bib_ref]. Successful leadership within healthcare organisations highlights the importance of emotional intelligence for insight into the "self' of the nurse transitioning through change. Emotional intelligence involves skills that are learnt and enhanced. Developing a sense of belonging and competence in the speciality happens over time, and depends on the individual (self).
# Limitations
The review was limited to the specialities of critical care, mental health, neonatal, peri-operative and it is acknowledged that the inhibitors and enablers outlined in this review may vary from one speciality to another. However, the findings of this study are similar to those undertaken in a parallel transition to community and midwifery practice systematic review which has further developed this model for that specific contentand suggests that the TRANS-PEC model can be applied to other nursing and non-nursing speciality contexts.
# Conclusions
The TRANSPEC model provides a foundation for effective recruitment and retention for both early career and rapid transitions to speciality practice. It shows that transition begins prior to entry to the speciality (pre-entry) and then continues to be developed over time. A successful transition relies on, at each phase, enablers to the professional and personal self, the formal and informal transition processes and a sense of belonging developing. This transition is influenced by the context of practice and the transition time will vary from one nurse to another.
The TRANSPEC model brings new knowledge to speciality transition in that it acknowledges the importance of the pre-entry phase, which has not been a focus previously. It also acknowledges that recruitment and retention into speciality nursing practice involves both the professional and personal self. The importance of personal well-being is increasingly being recognised, particularly its effects on quality patient care and patient safety.
The TRANSPEC model is designed to be used as a basis from which inhibitors and enablers can be considered. While this review has outlined some inhibitors and enablers, and recognising that there will be many of these that are not unique to a particular speciality, it is recommended that when developing a transition program using this model, the focus should be on enablers at all three phases for each individual speciality.
## Recommendations
Further testing of the TRANSPEC model should occur to ascertain its applicability to successful rapid and early career transition to a nursing speciality.
## Supporting information
[fig] Fig 1: The PRISMA flowchart. https://doi.org/10.1371/journal.pone.0216121.g001 TRANSPEC model: Supporting transition to speciality nursing practice [/fig]
[fig] Fig 2: The effective early career and rapid TRANSition to a Nursing SPECiality in differing contexts of practice-the TRANSPEC model. https://doi.org/10.1371/journal.pone.0216121.g002 [/fig]
[fig] S1: Fig. PRISMA 2009 checklist. (DOC) [/fig]
[table] Table 1: Articles included in the systematic review. [/table]
[table] Table 2: Template and matrix coding analysis with the frequency of codes per study related to each theme and transition time point. [/table]
[table] Table 3: Context of Practice: Phases and enablers and inhibitors.https://doi.org/10.1371/journal.pone.0216121.t003Table 4. Concepts and Phases: Enablers and inhibitors. So we are going to be even more short staffed. It's going to be more pressure on everyone. So I think all of this should have been addressed months before now. For the good of the unit . . . maybe it should be a continuous thing that they're always doing." Interview 2 [31] Neonatal [/table]
|
Usefulness of Mobile Devices in the Diagnosis and Rehabilitation of Patients with Dizziness and Balance Disorders: A State of the Art Review
Objective:The gold standard for objective body posture examination is posturography. Body movements are detected through the use of force platforms that assess static and dynamic balance (conventional posturography). In recent years, new technologies like wearable sensors (mobile posturography) have been applied during complex dynamic activities to diagnose and rehabilitate balance disorders. They are used in healthy people, especially in the aging population, for detecting falls in the older adults, in the rehabilitation of different neurological, osteoarticular, and muscular system diseases, and in vestibular disorders. Mobile devices are portable, lightweight, and less expensive than conventional posturography. The vibrotactile system can consist of an accelerometer (linear acceleration measurement), gyroscopes (angular acceleration measurement), and magnetometers (heading measurement, relative to the Earth's magnetic field). The sensors may be mounted to the trunk (most often in the lumbar region of the spine, and the pelvis), wrists, arms, sternum, feet, or shins. Some static and dynamic clinical tests have been performed with the use of wearable sensors. Smartphones are widely used as a mobile computing platform and to evaluate the results or monitor the patient during the movement and rehabilitation. There are various mobile applications for smartphone-based balance systems. Future research should focus on validating the sensitivity and reliability of mobile device measurements compared to conventional posturography. Conclusion: Smartphone based mobile devices are limited to one sensor lumbar level posturography and offer basic clinical evaluation. Single or multi sensor mobile posturography is available from different manufacturers and offers single to multi-level measurements, providing more data and in some instances even performing sophisticated clinical balance tests.
# Introduction
The aging of the balance system presents as dizziness, disequilibrium, and vertigo. It causes stress and isolation, and it results in a reduction of physical activities in the older adults due to an increased risk of falling. Community surveys show dizziness affects more than 30% of people over 70 and it is one of the most commonly reported medical complaints, affecting 15-35% of the adult population.Saber Tehrani et al 2 stated that in 2011, there were an estimated 3.9 million visits to U. S emergency departments for complaints of dizziness or vertigo. In a review of 16 studies on the economic evaluation of vertigo, up to 9.6 visits were reported annually at the primary care provider, up to 7.2 visits to a specialist, and there were up to 6 instrumental diagnostic procedures.Epidemiological data indicate a higher prevalence among women (27%) than men (14%), and an increase in prevalence with age.In a US study, the national costs for patients presenting with dizziness to emergency departments were estimated to exceed $4 billion per year. Of those, 25.7% were of otologic/ vestibular origin.For many years, posturographic examination was considered the golden standard for objective body posture examination. It is based on an inverted pendulum model and the Center of Pressure (COP) is assumed to be the free end of the pendulum. Tracking the displacement of the COP projection in a two-dimensional plane is done by registering the pressure of the feet on a posturographic platform, which is read by detectors, converted into a digital signal, and transmitted to a computer. There is no international consensus as to which of the parameters describing the displacement of COP have significant clinical value in the diagnosis of balance disorders.The literature emphasizes the value of the Center of Mass (COM) test and determining the relationship of COM to COP. There is also doubt regarding the significance of the COP evaluation on a posturographic platform in static conditions. A further development of this research method is computerized dynamic posturography, in which the researcher can force ground motions and change the visual reference points, thereby enabling the clinician to assess the patient's compensatory mechanisms by analysing COP displacements. This type of posturography allows for a more functional assessment of the range of postural stability as well as the dependence of the equilibrium system on information from the vision organ.
In recent years, new technologies have been applied in many disciplines for health monitoring, especially in physical activity assessment, specifically miniaturized wearable, body-fixed motion devices such as switches, pedometers, actometers, goniometers, accelerometers, and gyroscopes.These new inventions contributed to the development of mobile posturography, which enables COM testing during complex dynamic activities -both everyday life tasks and in clinical/laboratory conditions. Mobile posturography can also be utilized in assessing balance disorders and patient rehabilitation.It is especially aimed at older patients, as the aging of the world's population and the increasing number of active old people demand new non-invasive methods to evaluate balance and gait disorders. New wearable devices could predict falls and make it possible to introduce training programs tailored to older adults. It could not only improve the wellbeing of the aging population, but it could also have additional economic and social benefits.
The aim of this state of the art review is to present the usefulness of various mobile devices as cutting edge technology which could replace conventional posturography in the diagnosis and rehabilitation of dizziness and balance instability. We focused only on devices which had at least trunk worn sensor, as the trunk is proposed to be the most optimal region for single sensor mobile posturography. We also analysed the possibility of performing static and dynamic clinical tests with the use of mobile devices.
## Types of mobile devices
Various types of mobile devices have been described in balance disorders diagnosis and treatment. We have summarized the information on mobile devices, with at least one trunk worn sensor, available from literature inA commercially available device, the VertiGuard ® D/RT system (VestiCure GmbH, Germany, now Otocure GmbH, Metzingen, Germany), was frequently used. It is a waistworn vibrotactile system that consists of two gyrometers and a battery-driven main unit fixed on a belt at the COM. It records the angular acceleration of the patient's body in the sagittal and horizontal planes during the patient's daily tasks.The system is available in a diagnostic (D) version and a more complex therapeutic (RT) version, which has additional vibrators on the front, back, left, and right sides that can induce vibrotactile feedback signals.The VertiGuard ® device was used in balance rehabilitation in the older adults population and to assess body sway in patients with chronic uncompensated vestibular disorders. 7,10 Its use was further extended to balance training with neurofeedback in patients with different balance disorders (like canal paresis, otolith disorder, peripheral vestibular impairment after removal of an acoustic neuroma, microvascular compression syndrome, Parkinson's disease (PD), or presbyvertigo) and in identifying fallers in PD.The SwayStar™ system (Balance International Innovations GmbH, Iseltwald, Switzerland, and BESTecetp GmbH, Germany) is another commercially available mobile device, which is mounted on a belt at the back. It has two angular velocity transducers oriented in roll and pitch planes, and it directly measures the angular deviations of the trunk near the COM (around L3-L5).can be equipped with an additional vibrotactile and auditory biofeedback unit -Balance Freedom™ -in which signal transducers are mounted on a head-band.Faraldo-Garcia et al 14 established population age-adjusted reference values for the Sway Star system. The test results can be summarized with an option called balance control summary, which is an analysis of 14 balance tests.
Some researchers have used Opal, another commercial wearable system with inertial sensors, with the associated Mobility Lab™ Software (APDM, Inc. Portland, OR, USA, https://apdm.com).The system consists of up to six sensors, each containing 3-axis accelerometers, 3-axis gyroscopes, and 3-axis magnetometers. The sensors are attached to the wrists, arms, sternum, lumbar region, feet, and shins. Clinical tests, like the instrumented Timed Up and Go, Two-minute Walk, and Postural Sway tests can be performed using Mobility Lab™.Sankarpandi et al 18 demonstrated good to excellent test-retest reliability of the instrumented tests (iTUG, iSway) using the Opal-Mobility Lab™ system in 27 patients with uni/bilateral vestibular loss. The system was validated for gait analysis in young children. Compared with 3D motion capture, the temporal parameters were similar between the two systems, but the spatial ones may be used with a correction.The Mobility Lab™ system was also used in monitoring balance and gait in PD.A few authors have used Dynaport MiniMod Hybrid ® (McRoberts BV, The Hague, The Netherlands, https:// www.mcroberts.nl/). This device is fixed on a belt at the lower back (at the level of L4-5), and it consists of a triaxial accelerometer and a tri-axial gyroscope.It was tested in older adults to assess the strategies used to counteract imbalance during unsupported standing and to evaluate mobility performance and its correlation with Parkinsonian gait. 20,21 Iluz et al 22 recorded data in 40 patients with PD using Dynaport, and they developed an algorithm for objective and automated detection of missteps in these patients' daily lives, which may contribute to the early identification of fallers.
Another solution used in research is the Xsens device (Xsens, Enschede, The Netherlands www.xsens.com), which is composed of six body-worn sensors, each consisting of a 3-dimensional gyroscope and a tri-axial accelerometer.This device was used in a study by Spain et al 24 who recorded balance in MS patients with normal walking speed compared to healthy controls. They DovePress observed that body-worn sensors can detect early mobility changes in contrast to stopwatch timed tests. Physilog ® (GaitUp, Lausanne, Switzerland https://gaitup.com/), an inertial measurement unit (IMU) sensor in its latest version as Physilog ® 6, is a 10D sensor that includes a high-quality 3D accelerometer, 3D gyroscope, 3D magnetometer, and a barometric pressure sensor. The manufacturer developed additional research software, the PhysiGait Lab, specifically for gait analysis (assessing 26 parameters) in movement science and neurology (PD, stroke, SM). These sensors are attached to each wrist, one on each shank, and one on the trunk, complemented with one device on each foot and one on the back (lumbar 3); they were used in post-stroke patients to verify the validity of the iTUG test.In balance diagnostics and rehabilitation, many different prototype mobile devices have been developed.Kingma et alproposed the BalanceBelt (Elitac Wearables, Utrecht, The Netherlands) worn around the waist. It is a vibrotactile system with sensors (6 DOFaccelerometer and gyroscope) fixed at the back and 12 tactors positioned in a belt equally distributed around the belt. An inclination of more than 2.5 degrees in any direction will activate a tactor, sending a vibration to the patient. The system was tested on patients with severe bilateral vestibular loss, and a significant improvement in daily life was observed in 23 out of 39 patients. Alessandrini et al 33 used a Movit ® sensor (Captiks Srl, Rome, Italy; http://www.captiks.com/) with 3D accelerometers mounted on the posterior trunk near the COM. Movit System G1 is a wearable wireless system for motion capture and analysis. The authors confirmed the correlation between the parameters measured by mobile devices and forceplate-based posturography in patients with unilateral vestibular failure and in healthy subjects. 33
## Types and location of sensors
Wearable sensors used for body motion assessment may be categorized as inertial motion sensors and plantar force sensors.Many devices use accelerometers (linear acceleration measurement), gyroscopes (angular acceleration measurement), and magnetometers (heading measurement relative to the Earth's magnetic field).They are often based on microelectromechanical systems (MEMS), which enabled the development of miniaturized inertial sensorsa key technology in the process of making wearable devices smaller, lighter, and more comfortable.In a review of 26 balance studies in PD, 69% of researchers used triaxial accelerometers, 27% -inertial sensors, and 4% -other types of sensors (1D and 2D gyroscopes).Similar data were reported by other authors; thus, the accelerometer was the most often used sensor (70%) in the group of single-sensor devices.It was used as the only inertial sensor in 70% of the 40 studies reviewed by Howcroft et alfor fall risk assessment in people aged 60 and over, while accelerometers and gyroscopes were used in 27.5% of studies. Rucco et al 38 pointed out that some features of the accelerometer, such as its low cost, small size and weight, long-lasting batteries, and the number of commercially available solutions, have contributed to it being used the most frequently in mobile devices.
A tendency to reduce the number of sensors in mobile devices has been observed, so the devices can be more easily adopted by older patients. Regarding the number of sensors, in a review of 42 articles on the type and location of wearable sensors in healthy aged people, one sensor was used in 16 papers, two sensors were used in 17 papers, and three or more simultaneous sensors were used in only nine papers.Sensors in mobile devices are mounted on various parts of the body, but most often on the trunk (the lumbar region of the spine or the pelvis), and less often on the lower limbs, the head, upper limbs, and the chest. The lower back was the only location of sensors in 60-65% of studies.However, in 30%, sensors were placed on the feet (as pressure sensors within shoes), using two or three sensors.In a study by Özdemir,14 healthy young volunteers wearing six sensors attached to different body parts participated in 16 types of daily activities and 20 types of falls (so, the dataset consists of 2520 movements). It was demonstrated that the waist is the best region for a single sensor location on the body for wearable fall detection systems, with an average 98.42% accuracy by six machine learning techniques. It was followed by a thigh sensor (97.89% accuracy), ankle sensor (97% accuracy), head sensor (96.61% accuracy), and chest sensor (96.5% accuracy). The least accurate was the wrist-mounted sensor (94.92% accuracy). The trunk is considered the optimal place for the location of the sensors because, in this position, the sensor is the closest to the COM.There are some doubts as to the position of sensors on the head because although the vestibular system is located in this region, the accuracy of the sensors was lower than for the trunk.However, in older people, the head-based model submit your manuscript | www.dovepress.com
## Dovepress
Clinical Interventions in Aging 2020:15 was better at classifying fallers, and multi-sensor models were better at classifying non-fallers. 42
## Instrumented static and dynamic clinical tests
Clinicians are constantly looking for objective methods to quantitively evaluate balance deficit for diagnosis and therapy. Computerized systems such as posturography and wearable inertial sensors could provide more sensitive, specific, and repetitive testing in clinical practice. A serious disadvantage of such complex systems, such as CDP and gait laboratories, is that they are very expensive, and they require a large room and trained staff to perform the tests, which limits their availability to tertiary reference centres.
The Romberg test is one of the most commonly used clinical bedside examinations of postural stability. The iSway is an instrumented Romberg test obtained from wearable signals (accelerometers or gyroscopes) placed on, eg, the trunk, which measure the patient's sway. It provides useful quantitative information on postural balance, which can substitute clinical balance scales. It is used for screening for balance problems and vestibular disorders. The most frequent variation is the double leg stance for 30 s with eyes open and closed, although some authors also include different surface conditions, such as standing on a firm and foam surface, or a different stance, eg, tandem, semi-tandem, or single-leg stance. There is evidence that the results of sway tests from iSway using body-worn accelerometers (ACC) are useful and comparable with posturographic evaluation. Mancini et alperformed tests on subjects with early PD and on controls, in which they measured and compared the postural sway from the force plate and iSway using a body-worn ACC. They suggest that iSway measurements are more sensitive, reliable, and valid for characterizing posture control in PD subjects. Sankarpandi et al 18 recorded iSway results (Mobility Lab™) for 27 patients with vestibular deficit and concluded that this wearable system identified and distinguished fallers from non-fallers.
Another standing balance test performed with wearable inertial sensors is the balance error scoring system (BESS). The standard protocol includes three different stances: feet together, one foot, and tandem stance on two different types of surface -firm and foam (six conditions in total). A modified, shorter version of the BESS (mBESS) includes three different stances but limited only to a firm surface. In the BESS test, the examiner notes the failures according to the standardized balance errors list, in each 20-second trial, and scores them appropriately.Brown et alvalidated an algorithm to objectively measure BESS using a wireless inertial sensor on the forehead in healthy subjects. Their preliminary results suggest that an instrumented mBESS test may provide enough balance information to predict total BESS scores. King et alalso analysed the BESS test in adolescent athletes after mild traumatic brain injury, using an inertial sensor (Opal by APDM, Inc) placed at L-5 with an elastic belt. The mBESS results had the highest diagnostic accuracy for distinguishing post-traumatic patients from healthy individuals. The National Institutes of Health (NIH) Balance Toolbox suggests using inertial sensors to perform the Standing Balance Test for general balance screening (the measurement of anteroposterior postural sway).A different, complex balance test is the Clinical Test of Sensory Integration for Balance (iCTSIB), which in some ways resembles the sensory organization test (SOT) in CDP. Patients stand on a firm surface or on foam, with eyes open or closed, to quantitively assess the use of surface, visual, and vestibular information for postural sway control. Whitney et al 46 examined healthy adults with a wide age range to find a relationship between ACC measured at the pelvis and the center of pressure obtained from CDP (EquiTest ® ) in a SOT test. The subjects completed three consecutive trials for each of the six SOT conditions. They found an association between COP and ACC measurements, even when using the first trial. The iCTSIB was recommended as a screening test for sensory system performance in neurological patients.
One of the most useful clinical tests to evaluate both static and dynamic balance and mobility is the Timed Up and Go test (TUG). Measured with a stopwatch, it uses the time that a person takes to rise from a chair, walk three meters, turn around, walk back to the chair, and sit down. However, a limitation of TUG is that it focuses only on the total time, without focusing on different subphases of the subject's movements. Wearable inertial sensors could be used to record and analyse each of the gait subcomponents (walking, turning, rising from the chair, and sitting down) of the TUG test. Salarian et al 47 compared the results of a conventional TUG with stopwatch measurements and the iTUG, which employed seven portable inertial sensors in 20 subjects in the early stages of PD and in a control group. The authors concluded that iTUG showed Clinical Interventions in Aging 2020:15 submit your manuscript | www.dovepress.com DovePress significant differences between early PD and healthy individuals in the number of steps per minute, the angular velocity of arm-swing, turning duration, and the time to perform turn-to-sits. In their opinion, out of all the iTUG subcomponents, the gait, turning, and turn-to-sit were most sensitive. Similarly, Zampieri et alshowed that turning might be the most challenging movement for PD patients, and that the turning phase in TUG seems to be the most sensitive for detecting impairments. Gerhardy et al 49 used iCTSIB with inertial sensors and correlated these results with the smartphone-based version of the iTUG. One of their conclusions was that the performance of the vestibular and somatosensory system is strongly correlated with iTUG total time.
The utility of wearable systems in distinguishing fallers from non-fallers was confirmed by Sankarpandi et alwho recorded the results of the iTUG and iSway in two sessions in three consecutive measurements, and they confirmed good to excellent reliability of both systems' parameters. O'Sullivan et al 50 also stated that an accelerometer could distinguish between fallers and non-fallers, show different balance conditions, and possibly correlate with clinical tools such as the Berg Balance Scale (BBS) and TUG.
Other applications of mobile devices included mobility tests like the Stand and Walk (ISAW) test and the Long Walk plugin (IWalk). Patients wear a Mobility Lab™ system, which makes it possible to calculate gait variability, phase coordination index, turning, and asymmetry.This kinematic data can be used to quantitively measure progress in rehabilitation. The vast amount of data and parameters measured by mobile devices and their sensors required the employment of machine-learning techniques to develop algorithms capable of interpreting these datasets. The authors mentioned above recommend using wearable sensors to perform iTUG, Sway (iSway), Stand and Walk (ISAW), and Long Walk plugin (IWalk) tests.The instrumented tests allow for an objective, low-cost quantification of balance performance, and they are more sensitive than clinical examinations.
## Patient notification and telephone applications (apps)
Mobile phone technology has had a major impact on the development of monitoring systems that are based on wearable sensors.Smartphones (SP) are widely-used electronic devices that comprise several sensors like an accelerometer, magnetometer, gyroscope, global positioning system, microphone, and camera. SP operating systems are capable of multitasking and can run processes in the background, therefore allowing them to function as a mobile computing platform, which could be used in fall detection and prevention.Rey-Martinez et al 52 created the RombergLab application for the iPhone's operating system, which consists of full posturography software based on a wireless IMU. Lee et al 53 designed, developed, and assessed a smartphone-based vibrotactile feedback system for balance rehabilitation training at home. The tool included an iPhone, with its tri-axial linear accelerometer, and a "tactor bud" plugged into the SP's audio jack. This instrument was used to measure body acceleration, and anterior-posterior and medial-lateral body tilt. Using variations of Romberg test (with eyes open/closed, semi-tandem, and tandem), healthy subjects and those with vestibular deficits corrected their posture in response to vibration feedback.
The smartphone balance training system could be used at home without difficulty. Another smartphone-based system for balance training and rehabilitation in older people was proposed by Fleury et al.An application called iBalance-ABF sends audio feedback to the user through an earphone when the tilt angle of the trunk exceeds an adjustable and predetermined threshold. The audio assistance allowed the patient to decrease trunk sway in the tandem stance with closed eyes and to maintain this destabilizing position.
Mellone et alpresented the results of the FARSEEEING group (an EU consortium to develop fallrelated technologies, databases, and telemedical service model) regarding research on using SPs for fall detection and fall prevention. For fall detection, the uFall application was developed, where real-time fall-detection algorithms already proposed in the literature have been used. The app continuously records the signals from inertial sensors and runs another process that simultaneously acts as a fall/event detector. When a fall is suspected, an audio alarm is generated, and the user has to manually disable it. The uFall app is configured by a clinician before the patient uses it for the first time, and then, during the motor activities, its use is completely transparent to the user. There is still no validation on the uFall fall-detection algorithm. Another app for fall detection is the iFall, which continuously records data from the SP's accelerometer.The user has to respond to the app when a fall is suspected, otherwise the system triggers an alarm. submit your manuscript | www.dovepress.com
## Dovepress
Clinical Interventions in Aging 2020:15
The first attempt to instrument a clinical functional test was called uTUG.The uTUG is performed with an SP on the lower back. The app records and processes the signals from the accelerometer and gyroscope and then sends a report to a remote server. The uTUG proved to be a pocket-sized tool for fast screening, assessment, and follow-up.
Cohen et al 56 used an inertial motion Bluetoothenabled IMU (iPod, Apple, Inc.) unit to measure kinematic variables in a version of the Romberg test with the head still or moving in healthy controls and in patients with benign paroxysmal positional vertigo, postoperative acoustic neuroma resection, and chronic peripheral unilateral weakness. The subjects were instructed to move their heads in time with a specific type of tone.
Only Romberg tests on foam with eyes closed, with the head still or moving, were useful in screening for vestibular impairments, especially in older people. Another application for monitoring human mobility and falls is the Mover.This app categorizes activities as "Sleeper, Sitter, Lagger, Walker, Mover, Hyper." It also employs an experimental fall-detection algorithm. When a fall is suspected, the Mover will play an alert, which will notify an emergency contact unless disarmed.
## The use of mobile devices in the assessment and rehabilitation of balance disorders
Mobile devices make it possible to assess posture and gait in both clinical tests and everyday life activities. They are portable, lightweight, and less expensive than classic balance assessment systems (static and dynamic posturography). For these reasons, there is increased interest in their use in diagnostics and rehabilitation.They were used to evaluate posture and gait disorders in healthy people, especially in the aging population, and to detect falls in the older adults.They were employed in the rehabilitation of various neurological diseases, like PD, SM, Alzheimer's disease, and stroke, but also in osteoarticular and muscular system diseases, in diabetes mellitus, and in vestibular disorders.In PD, the use of various types of mobile devices to assess posture and gait has been studied to differentiate patients at high risk of falls, different types of disease, carriers of the LRRK2 gene, and people at high risk of developing this disease.Rovini et al 67 identified five major applications for wearable devices in PD: early diagnosis, tremor detection, analysis of motor performance, analysis of motor fluctuations (on/off phases), and home and long-term patient monitoring. Some authors reported that postural instability and some features of gait analysis, such as jerk, harmonic stability, oscillation range measured by accelerometers on the pelvis and head, and asymmetry between the right and left side, could be biomarkers for the prodromal phase of PD.Hasegawa et alreported that in PD patients, the most sensitive domains of balance dysfunction were anticipatory postural adjustments prior to gait initiation, and dynamic balance while walking (gait); those measures were significantly correlated with the severity of the disease.
Mobile devices have also been used in rehabilitation and balance training with biofeedback.Brugnera et alobserved that that vibro-tactile stimulation from the Vertiguard™ device improved the body balance of patients who did not respond well to conventional vestibular rehabilitation. In PD, a significantly higher increase in the SOT score was reported in patients using the VertiGuard ® -RT vibrotactile neurofeedback system in comparison to CPD training.Numerous studies have attempted to establish criteria (eg, the location of sensors, the most useful variables to assess fall risk, classification models) to identify people at high risk of falling (fallers/non-fallers) using wearable sensors.This problem was researched by an international team in the ProFaNE (Prevention of Falls Network Europe) network, which was part of a broader project, "STate of the Art Robot-Supported assessments," or STARS; however, no consensus regarding parameters that are most useful to objectively assess the risk of falls was established. 71
# Limitations
Limitations on the use of mobile devices in clinical practice result from the fact that many studies describe prototype devices with different locations and number of sensors, and no consensus has been reached regarding the test protocols or derived parameters.The methodologically poor quality of research is emphasized. There is a lack of standardization in data acquisition, mathematical models, and algorithms used to process data. Hubble et al 37 in a review of 26 papers on the use of mobile systems in PD only three papers were characterized by research of high methodological quality. Alessandrini et al 33 also underline the importance of validation, sensitivity, and the reliability of mobile device measurements compared Clinical Interventions in Aging 2020:15 submit your manuscript | www.dovepress.com DovePress to conventional posturography. In healthy subjects, Faraldo-Garcia et alreported that results of computed dynamic posturography and the SwayStar system were significantly related only when the measurements were performed simultaneously and under the same sensory stimulation conditions. Future research should focus on expanding the studied populations and selecting more appropriate patients. Research protocols lack information on patients' assessment for vision, sensation, neuro-motor skills, cognitive functions, and other systemic diseases that may affect the diagnosis and rehabilitation of balance disorders.It is especially important in older patients, in whom balance disorders are caused by ultrastructure damage, such as the loss of sensory cells, otoconia, and vestibular neurons.These lesions change the postural strategy and cause greater dependence on visual and proprioceptive signals in the older adults.
# Conclusions
Smartphone based mobile posturography applications are easy-to-use and affordable solutions, but are limited to one sensor lumbar level posturography and offer only basic clinical evaluation (TUG) and limited biofeedback capabilities. Therefore, more research is needed to develop solutions to provide continuous monitoring, improve biofeedback necessary to implement these solutions into home rehabilitation regimens and to decrease the number of false alarms.
Single or multi sensor, commercially available mobile posturography devices offer multi-level measurements, which seem to provide more possibilities for balance monitoring and testing in clinical settings. However the cost and size of these devices makes them impractical to use at home, therefore significantly limiting their capability for continuous monitoring and home rehabilitation.
## Abbreviations
COP, center of pressure; COM, center of mass; CDP, computerized dynamic posturography; IMU, inertial measurement unit; SP, smartphone; PD, Parkinson disease; MS, multiple sclerosis; BESS, balance error scoring system; iCTSIB, Clinical Test of Sensory Integration for Balance; SOT, sensory organization test; TUG, Timed Up and Go test; BBS, Berg Balance Scale; app, smartphone application. |
Progress in Research on SARS-CoV-2 Infection Causing Neurological Diseases and Its Infection Mechanism
COVID-19 has spread rapidly worldwide since its outbreak and has now become a major public health problem. More and more evidence indicates that SARS-CoV-2 may not only affect the respiratory system but also cause great harm to the central nervous system. Therefore, it is extremely important to explore in-depth the impact of SARS-CoV-2 infection on the nervous system. In this paper, the possible mechanisms of SARS-CoV-2 invading the central nervous system during COVID-19, and the neurological complications caused by SARS-CoV-2 infection were reviewed.
# Introduction
The novel coronavirus is a previously unknown β-coronavirus, which is a single-stranded positive-strand RNA virus. The World Health Organization named it 2019-nCoV and the International Committee on Virus Taxonomy named it SARS-CoV-2. The virus belongs to a branch of the sarcoma virus subfamily of the coronavirus subfamily. SARS-CoV-2 is the seventh member of the coronavirus family that infects humans. SARS-CoV-2 is the virus that caused COVID-19 in 2019. SARS-CoV-2 infection can cause severe acute respiratory syndrome. SARS-CoV-2 has a high potential to spread and infect humans all over the world. Since the first case of COVID-19 was diagnosed in Wuhan, the number of SARS-CoV-2 infections worldwide has increased exponentially in the past few months. COVID-19 was originally described as a respiratory infection, but now it is increasingly regarded as a multi-organ disease, including nervous system manifestations. An updated version of the new guidelines for the diagnosis and treatment of coronary pneumonia issued by the National Health Council of China (China NHCotPsRo, 2020) points out that histopathological samples from some COVID-19 patients showed that SARS-CoV-2 invasion involved multiple organs, including lung, spleen and hilar lymph nodes, heart and blood vessels, liver and gallbladder, kidney, brain, adrenal gland, esophagus, stomach, and intestine. In particular, edema, and partial neuronal degeneration were observed in brain tissue (China NHCotPsRo, 2020) Neurodegenerative changes observed in cells infected with SARS-CoV-2, including cell death and hyperphosphorylation, as well as dislocation of Tau protein, these changes can be observed in conditions such as hyperthyroidism or Alzheimer's disease (4) However, the specific mechanism of neurodegenerative changes induced by SARS-CoV-2 remains to be further studied in the future. The central nervous system may serve as a reservoir for SARS-CoV-2, some groups observed that viral particles gradually accumulated within the neuronal cells of the brain organs from 6 to 72 h after SARS-CoV-2 infection, indicating that the virus replicated actively and effectively in the neuronal cells within the first few days of infection. However, some groups observed that viral infection did not replicate effectively in the first few days and suggested that the central nervous system might serve as a long-term reservoir of the virus. More and more evidence shows that SARS-CoV-2 has a potential neuroinvasive effect. It is estimated that more than 1/3 of COVID-19 patients will have nervous system symptoms, including central nervous system symptoms (dizziness, headache, disturbance of consciousness, acute cerebrovascular disease, ataxia, epilepsy). Peripheral nervous system symptoms (taste disorder, olfactory disorder, visual impairment, neuralgia).
So far, although the epidemic in China has been effectively controlled, the COVID-19 epidemic is still very serious worldwide. According to statistics, there are more than 54 million confirmed cases worldwide, and more than 15 million existing confirmed cases. In this global public health emergency, we are still facing a very serious situation. In the face of SARS-CoV-2, understanding the impact of SARS-CoV-2 infection on the nervous system and its invasion mechanism is of great significance for the reasonable treatment of patients. In this paper, the effects of SARS-CoV-2 on nervous system are systematically analyzed and reviewed.
## Three mechanisms of sars-cov-2 invading the nervous system
The central nervous system is protected by a highly complex brain barrier system, which is the first line of defense against virus invasion. The brain barrier is composed of the bloodbrain barrier, blood-cerebrospinal fluid barrier, and braincerebrospinal fluid barrier. The blood-brain barrier has a maximum surface area that can be used for communication between the brain and blood. It consists of cerebral capillary endothelial cells, extracellular matrix, and astrocyte podocytes. The blood-cerebrospinal fluid barrier is located in the choroid plexus of the ventricle of the brain. The epithelial cells of the choroid plexus are mainly responsible for the barrier function of the blood-cerebrospinal fluid barrier. The blood-brain barrier and the blood-cerebrospinal fluid barrier can inhibit paracellular diffusion, protect the central nervous system from the influence of the constantly changing blood environment, infections and toxins, and are crucial for maintaining the homeostasis of the central nervous system.
To cause a central nervous system infection, the virus must first successfully cross the protective barrier of the brain. Crossing the blood-brain barrier or the blood-cerebrospinal fluid barrier requires special adaptation of the virus. Despite this, SARS-COV-2 can still enter the nervous system rapidly in some special ways after infection.
The following three ways may be the main ways for SARS-COV-2 to invade the central nervous system: (1) SARS-CoV-2 directly infects vascular endothelial cells by means of angiotensin converting enzyme 2, thus directly crosses the blood-brain barrier. (2) SARS-CoV-2 enters the central nervous system through synaptic connections via the olfactory nerve. (3) SARS-CoV-2 Induces Inflammation to destroy the Brain Barrier System and enter the central nervous system.
## Sars-cov-2 directly infects vascular endothelial cells and crosses the blood-brain barrier
Studies have shown that similar to SARS-COV, SARS-COV-2 can use ACE2 to enter the cell interior. The spike protein (S protein) in SARS-CoV-2 is a trimer projecting from the viral membrane and contains a receptor binding domain (RBD) in each monomer. Through it, the viral protein can directly interact with ACE2 receptors on the surface of many host cells. S protein is the main tool for SARS-CoV-2 to bind to the ACE2 receptor (infect cells), and it can strongly bind to ACE2. Therefore, the S protein may be used as a key target for the treatment of COVID-19 and vaccine development.
In addition, host cell protease also plays an important role in virus entry and infect cells. The S1 subunit of S protein on the surface of SARS-CoV-2 first binds to neuron ACE2 receptor and adheres to the surface of target cells; then, the S2 subunit of the virus is activated by the serine protease TMPRSS2 of host cells, and the virus can enter the nerve cells. Therefore, TMPRSS2 activity is very important for the infection and transmission of SARS-CoV-2 in host cells and is important pathogenesis of neurological complications .
Although the S protein activity of SARS-CoV-2 is weaker than that of previous coronaviruses. However, the binding affinity of SARS-CoV-2 S protein to ACE2 is 10-20 times higher than that of SARS-CoV S protein to ACE2. This is due to the fact that the receptor binding domain of SARS-CoV2 is different from that of previous coronaviruses on several key amino acid residues. It contains 4 positively charged residues and five negatively charged residues, so SARS-CoV-2 S protein is slightly more positively charged than SARS-CoV S protein.
Although the charge difference between the S proteins of SARS-CoV-2 and SARS-CoV is quite small, this effect can be amplified by the presence of a large number of S proteins on the virus particles.
At the interface between the SARS-CoV-2 RBD and the cellular ACE2 receptor, the difference in the amino acid content of the S protein can lead to a more specific interaction between the S protein and the host cell receptor. Therefore, compared with SARS-CoV, SARS-CoV-2 is more likely to establish interactions with different targets in the human body through non-specific and specific interactions. All of these can ultimately increase the ability of SARS-CoV-2 to enter human cells, which may explain why COVID-19 has stronger pathogenicity, transmissibility, and greater global influence.
This charge difference between SARS-CoV-2 and SARS-CoV S proteins can have a significant impact on endothelial cell adhesion and crossing the blood-brain barrierso that SARS-CoV-2 infected with vascular endothelial cells of the blood-brain barrier has a higher efficiency of reaching the brain and can cross the blood-brain barrier directly into the central nervous system.
Through the study of the distribution of ACE2 in the nervous system, it was found that ACE2 was expressed in different brain regions, such as subfornical organs, nucleus tractus solitarius and ventrolateral region of the medullary head, as well as regions such as motor cortex and raphe. According to spatial distribution analysis, ACE2 was also expressed in the substantia nigra. Due to the existence of ACE2 receptors in glial cells and neurons, it has become a potential target for SARS-CoV-2 causing brain injury and neurological symptoms.
In addition, ACE2 is widely attached to the extracellular surfaces of the lungs, arteries, heart, kidneys, intestines, and brain (cell membranes). In addition to respiratory system involvement, SARS-COV-2 infection may also cause multiorgan dysfunction. Despite the predominance of respiratory symptoms, there is post-infection damage to the myocardium, kidneys, intestines, and liver, perhaps ACE2 provides a crucial link between immunity, inflammation, and cardiovascular disease.
The autopsy of novel coronavirus pneumonia showed brain edema and partial degeneration of neurons. However, there is not enough autopsy evidence to prove that SARS-CoV-2 exists in neurons and glial cells. Further studies are needed to prove this.
## Olfactory nerve pathway
Among the 12 pairs of cranial nerves, the olfactory nerve is not a real nerve but a conduction bundle of the central nervous system. It can directly contact the brain (31), coupled with the special location and structure of the olfactory nerve itself, we speculate that perhaps in the mechanism of SARS-CoV-2 invading the central nervous system, the virus entering the central nervous system through the olfactory nerve is also one of the main ways. And the olfactory nerve provides this way to enter FIGURE 2 | SARS-CoV-2 invades the brain through the olfactory nerve. the central nervous system, successfully bypassing the bloodbrain barrier, effectively making it a channel between nasal epithelium and central nervous system. In the early stage of respiratory transmission, SARS-CoV-2 can enter the brain through the olfactory nerve (37) .
In the nasal cavity, the special olfactory neuroepithelium has an apical surface mainly composed of support cells, supporting the dendritic processes of neurons containing olfactory cilia. The dendrites of olfactory neurons are directly exposed in the airway of the nose. Although the olfactory system is very effective in controlling viral nerve invasion under normal conditions, however, data from several studies indicate that nasal respiratory epithelial cells express ACE2 and TMPRSS2. As mentioned above, SARS-CoV-2 can enter the cell with the help of ACE2 and the serine protease TMPRSS2 of the host cell.
The expression of ACE2 and TMPRSS2 located in the olfactory neuroepithelium indicates a potential entry point for SARS-CoV-2 into the central nervous system, so that the virus can invade the olfactory nerve. The retrograde or anterograde transport of neurons is realized by kinesin and kinesin. During the infection process, SARS-CoV-2 uses the olfactory nerve to pass through the cribriform plate of the ethmoid bone to cause brain invasion, which results in a rapid, transneuronal spread of the SARS-CoV-2 to relevant regions of the brain, which in turn interacts with ACE2 expressed on the surface of brain neurons. This cell tropism may be the reason why SARS-CoV-2 is highly infectious and related to olfactory dysfunction.
Because SARS-CoV-2 can directly act on nasal respiratory epithelial cells in the nasal cavity, olfactory dysfunction often occurs in the early stages of the disease. In mild to moderate cases, sudden loss of smell and taste is considered to be the strongest predictive symptom of early infection with the SARS-CoV-2 virus, and this mild, non-specific symptom can become asymptomatic. Or the only manifestation of a mildly infected person. The report of symptoms related to anosmia should be regarded as a sign of SARS-CoV-2 infection and a sign of COVID-19. If it is accompanied by transient brain edema and other neurological diseases, the first consideration should be the neuroinvasiveness of SARS-CoV-2.
The olfactory dysfunction caused by SARS-CoV-2 may be explained by the following four mechanisms: x Viral infections of the nasal mucosa can trigger inflammation of the nasal tissue, including the olfactory mucosa, thereby creating an obstructive barrier between odor chemicals and olfactory receptors; y direct damage to olfactory receptors could prevent odor signals from being transmitted; z the virus, being neurotropic, can attack the area of the brain responsible for smell along the path of the olfactory nerve; { Loss of sense of smell may actually be a sequela of brain edema and partial neurodegeneration. Any or all of these four mechanisms may lead to loss of sense of smell in COVID-19. Therefore, exploring the relationship between early loss of sense of smell and a long-term sense of smell has special clinical and prognostic value.
Although there have been many studies proving that the olfactory neuroepithelium expresses ACE2 and TMPRSS2, there are no sufficient and strong studies to prove that ACE2, TMPRSS2, and SARS-CoV-2 are widely present in the transmission bundle from the olfactory bulb to the CNS. In the future, further research is needed to resolve these inconsistencies, and more autopsy reports are needed to prove the presence of SARS-CoV-2 in the olfactory tract. Finally, the distribution of SARS-CoV-2 infected cells in human olfactory nerve conduction tracts was determined.
## Coronavirus induces inflammatory responses to disrupt the blood-brain barrier system
The destruction of the blood-brain barrier through inflammation is also one of the ways for SARS-CoV-2 to enter the central nervous system. Infection with SARS-CoV-2 can destroy the blood-brain barrier by producing a large number of inflammatory mediators in the following three ways.
## Sars-cov-2 directly induces the release of cytokines by immune cells
When SARS-CoV-2 enters the human body, the virus activates immune cells, such as monocytes/macrophages, neutrophils, T cells, natural killer cells, and mast cells. The activated immune cells kill the virus by synthesizing and releasing cytokines. These cytokines mainly include interferon (IFN), interleukin (IL), chemokine, and tumor necrosis factor (TNF). Some of their functions are to promote the inflammatory response and some to inhibit the inflammatory response. These cytokines are maintained in a balanced state in the healthy human body. Among them, pro-inflammatory factors can activate and recruit other immune cells, immune cells can secrete more cytokines, activate, and recruit more immune cells, thus forming a positive feedback cycle.
SARS-CoV-2 can cause immune cells to produce excessive immunity, cytokines are uncontrolled, a large number of cytokines are released, amplifying positive feedback, breaking the balance, marking an uncontrolled and dysfunctional immune response, leading to systemic inflammation, further aggravating the inflammatory response and increasing the severity of the disease. Although this excessive immune response can kill the virus, it can also cause some additional damage. Among them, the blood vessels suffered the most damage. Cytokine storms make the vessel wall more easily penetrated, and the blood-brain barrier is disrupted, causing neocoronavirus to enter the brain, inducing corresponding central nervous system symptoms.
## Activation of glial cells releases proinflammatory cytokines
Some neurotropic viruses can induce the pro-inflammatory state of glial cells and make them secrete cytokines. As mentioned earlier, glial cells express ACE2, We speculate that SARS-CoV-2, which enters the central nervous system through olfactory nerve, blood-derived, and other pathways, may also activate glial cells and induce a pro-inflammatory state. In addition, experiments have confirmed that glial cells secrete a large number of inflammatory factors after being infected with coronavirus, such as interleukin-6, interleukin-12, interleukin-15, and tumor necrosis factor α, etc.. These cytokines can also damage the blood-brain barrier, further promote coronavirus to enter the brain, causing symptoms of central nervous system disease.
So far, the autopsy report on the glial cells of SARS-CoV-2 patients is still insufficient. We propose this possible mechanism based on the existing literature. In future studies, more autopsy reports are needed to prove whether SARS-CoV-2 infects glial cells through ACE2, or whether there are other receptors that can bind to SARS-CoV-2 in glial cells.
## Vascular endothelial growth factor induces inflammation
Vascular endothelial growth factor (VEGF) is widely distributed in the central nervous system. In addition, the combination of SARS-CoV-2 and ACE2 can activate the renin-angiotensin system which is involved in inflammation response, and then further promote the synthesis of VEGF through the binding of angiotensin II (AngII) and angiotensin II type 1 receptor (AT1R). In fact, in brain diseases, VEGF not only promotes angiogenesis but also destroys the blood-brain barrier by inducing inflammatory responses.
Inflammation is the precursor and companion of blood vessel formation, manifested by increased vascular permeability and recruitment of inflammatory cells. ACE2 is a key enzyme that catalyzes Ang I and Ang II to Ang 1-9 and Ang 1-7, respectively. When SARS-CoV-2 attacks ACE2, the inactivation of this enzyme can lead to the enhancement of the ACE/AngII/AT1R axis signal, followed by excessive AngII production. In the brain infected with SARS-CoV-2, the cumulative feedback of Ang II promoted the increase of ACE2. VEGF in turn reversely enhances Ang II, thus forming a vicious cycle in the release of pro-inflammatory cytokines, including TNF-α, IL-1β, IL-6, IL-8, and ICAM-1. In addition, among these cytokines, interleukin-6 (IL-6) is an important
## Clinical manifestations of nervous system diseases caused by sars-cov-2
The neurological diseases possibly caused by SARS-CoV-2 can be divided into three major categories:
x Nervous system consequences of related lung and systemic diseases, such as cerebrovascular disease; y The virus directly invades the central nervous system, such as encephalitis; z Potential immunemediated complications after infection, such as Guillain-Barre syndrome (GBS) and other types of demyelinating diseases.
Cerebrovascular disease refers to a group of diseases that occur in the blood vessels of the brain and cause brain tissue damage due to the disturbance of intracranial blood circulation.
Although the main manifestation of patients infected with SARS-CoV-2 is a lung disease, there are also cerebrovascular diseases. When the virus proliferates in lung tissue, it causes diffuse alveolar and interstitial inflammatory exudates, and even hyaline membrane formation. This will lead to abnormal alveolar gas exchange, hypoxia of the central nervous system, an increase of anaerobic metabolism of brain tissue, induction of intercellular edema, obstruction of cerebral blood flow, causing ischemia of cerebral circulation, with the increase of intracranial pressure, the brain function deteriorated gradually. It may even induce the occurrence of acute cerebrovascular disease, such as acute ischemic stroke.
On the other hand, cerebral hemorrhage caused by elevated blood pressure may also be the result of the expression of ACE2 receptor (69, 70). ACE2 is one of the cardio-cerebrovascular protective factors, which plays an important role in regulating blood pressure and anti-atherosclerosis mechanism. SARS-CoV-2 may cause an imbalance of the renin-angiotensin system (RAS) by acting on the ACE2 receptor, leading to microcirculation disorder, affects cerebral blood flow regulation, leading to an abnormal increase of blood pressure, and increases the risk of cerebral hemorrhage and ischemic stroke.
Encephalitis refers to the inflammatory lesions of brain parenchyma caused by pathogens, including neuronal damage and nerve tissue damage. It is characterized by acute episodes and common symptoms include headache, fever, nausea, vomiting, fatigue, convulsions, and disturbance of consciousness. At present, the presence of viral encephalitis in many patients infected with SARS-CoV-2 further speculated the existence of this neurological complication.
The treatment team of Beijing Ditan Hospital confirmed the presence of SARS-CoV-2 in the cerebrospinal fluid of patients infected with SARS-CoV-2 through genome sequencing, thereby clinically confirming viral encephalitis. This provides a solid foundation for SARS-CoV-2 to cause encephalitis. However, no signs of inflammation were found in brain tissue images of patients infected with SARS-CoV-2. We guess that SARS-CoV-2 will produce virion vacuoles like MERS-CoV and SARS-CoV, if this hypothesis holds, then vacuolation may be a defense against infection. This problem needs further study and more pathological cases to clarify.
In patients with SARS-CoV-2, SARS-CoV-2 can stimulate immune cells to produce a variety of cytokines, resulting in an immune response process that causes nerve demyelination. For example, SARS-CoV-2 can cause Guillain-Barré syndrome. Guillain-Barre syndrome, also known as acute idiopathic polyneuritis or symmetrical polyradiculitis, is an acute polyradiculoneuropathy. Clinical manifestations are progressive ascending symmetrical paralysis, quadriplegia, and varying degrees of sensory disorders. The exact pathogenesis of nerve demyelination caused by SARS-CoV-2 is not clear and remains to be further studied.
## Future prospects
COVID-19 is a challenge to the world. At present, there is sufficient evidence that SARS-CoV-2 can invade the central nervous system and induce nervous system diseases. The possible pathways of SARS-CoV-2 invasion into the central nervous system include direct invasion of infected endothelial cells, invasion through the olfactory nerve, and invasion by inducing inflammation to destroy the brain barrier system. These pathways are all related to ACE2 receptors, so the relationship between SARS-CoV-2 and ACE2 should be further studied so as to take better measures to protect the central nervous system in patients with COVID-19. In fact, human respiratory viruses may also enter the central nervous system through other different ways, including the trigeminal nerve, cerebrospinal fluid, lymphatic system, and so on. The three mechanisms discussed in this article may be applicable to SARS-CoV-2, but we must be alert to other invasion mechanisms of SARS-CoV-2 until there is conclusive pathological evidence. In addition, it can be inferred from the existing data that encephalitis, cerebrovascular disease, nerve demyelination symptoms, and olfactory changes in COVID-19 patients are all likely to be related to SARS-CoV2 infection. These symptoms can be used as potential indicators of patient severity and prognosis. Understanding these knowledge is very important for the prevention and treatment of central nervous system symptoms and the rehabilitation of COVID-19 patients.
In addition, recent studies point out that other proteins expressed in nerve cells, such as Nrp1, may also become receptors for SARS-CoV-2. Future work needs to verify whether SARS-CoV-2 can invade the central nervous system using alternative receptors.
# Author contributions
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. |
A quantitative method to decompose SWE differences between regional climate models and reanalysis datasets
## Note: uncertainties in the quantitative decomposition of model-reference difference in swe
## A. uncertainties in snsr swe
The SNSR derived SWE was compared with 108 snow pillow and 202 snow course in-situ measurement sites throughout the Sierra Nevada. For the 202 snow course sites, the mean error in SNSR was found to be less than 3 cm and the root-mean-square error was 13. In the 108 snow pillow measurement sites, the mean error in SNSR was 1 cm and the root-mean-square error was 11 cm. Averaged over all 310 measurements, the mean error is 2.3 cm, indicating the systematic error of SNSR is 2.3 cm; root-mean-square error is averaged 12.3 cm indicating a random error of 12.3 cm at each site. The random error at each site results in the standard error of the mean by 12.3/√(310) = 0.7 cm, or 1.4 cm with a 95% confidence interval. Therefore, the total uncertainty is √(2.3 2 + 1.4 2 ) @ 3 cm over the survey time and area when and where the mean SWE is 1.05 m, indicating a 3% uncertainty. This uncertainty affects the decomposition of model-reference difference of SWE where SNSR SWE is used in the calculation, i.e., row and in . In , we list each of the datasets used in and include the true magnitudes of each variable along with their uncertainties. For example, reference SWE is 190 mm with 3% uncertainty, i.e., 190 (±6) mm.
## B. uncertainties in prism daily temperature and precipitation
The PRISM derived T were compared with in-situ measurements in the leeward side of the Sierra Nevada [bib_ref] Testing the daily PRISM air temperature model on semiarid mountain slopes, Strachan [/bib_ref]. A random error with standard deviation of 1.67 °C was found. Our surveyed spatial-temporal extent covers about 3000 PRISM grids point over 3000 days. Therefore the random error only contributes to negligible error of the mean, which is 1.67/√(3000 x 3000) < 0.001 °C. However, a cold bias associated with topography was reported in PRISM daily T, with -0.75 °C bias in daily minimum T (Tmin) and -1.95 °C bias in daily maximum T (Tmax), or -1.35 °C in mean T. Further, we compared the PRISM and Livneh 2015 (L15) reanalysis dataset and found PRISM to be warmer than L15 by 1.1 °C in the windward side of Sierra Nevada. It is reasonable to believe the 95% confidence interval is below the upper limit of the systematic error (1.35 °C), which would result in a ± 69 mm uncertainty in snowfall. This uncertainty should be considered when comparing snowfall calculated from reference T and from modeled T. In , this is included in row REF in column $ , $ , $ " . PRISM derived P was shown to have an estimated seasonal bias of 0-3% in North Carolina [bib_ref] High-resolution precipitation mapping in a mountainous watershed: Ground truth for evaluating uncertainty..., Daly [/bib_ref]. In the Sierra Nevada, a -1% and -4% seasonal P bias was found for two reanalysis datasets that rescaled their estimates to the PRISM climatology [bib_ref] High-Elevation Precipitation Patterns: Using Snow Measurements to Assess Daily Gridded Datasets across..., Lundquist [/bib_ref]. Given these two in-situ validation attempts of PRISM derived P, we assume a ±3% uncertainty due to systematic error in our surveyed area, which would lead to ±3% (± 10 mm) uncertainty in snowfall. This uncertainty should be considered when comparing snowfall calculated from reference P and from modeled P. Similar to PRISM daily T, the random error of PRISM daily P make negligible contribution to the standard error of the mean due to the large sample size. In , uncertainties in PRISM daily P is included in row REF in
[formula] column $ , $ , $ ) , $ * . [/formula]
c. Uncertainties about when precipitation occurs We use daily total P and daily mean T to calculate the daily total snowfall in both reference and model datasets. The underlying assumption is that all precipitation occurs at daily mean T. However, precipitation could occur at any time of the day, and we do not know the exact distribution of P with T from our datasets. Therefore, we use a Monte Carlo method to estimate the associated uncertainty in snowfall, by assuming P occurs at a random T between Tmax and Tmin. The random T follows uniform distribution between Tmax and Tmin. It should be noted that the time of day when P occurs is not completely independent between grid cells, which means a reduced degree of freedom in random samples. To reduce the degree of freedom accordingly, we assume the random T has the same relative distance between Tmin and Tmax on the same day at all grid points. Therefore, snowfall estimates are calculated using daily total P and the generated random T. We repeat the generation of random T and calculation of snowfall until the mean snowfall converges. The associated uncertainties in snowfall equals twice the standard deviation and represents a 95% confidence interval. This results in an 8 mm uncertainty in reference snowfall, and a 4-9 mm uncertainty (dependent on the amount of P and diurnal variability of T) in NA-CORDEX snowfall in column $ , $ , $ ) , $ * , $ , $ + , $ " and $ in .
d. Uncertainties about rain-snow partitioning Snowfall in the Sierra Nevada has been shown to occur between 0 -3 °C, with 90% precipitation falling as snow at 0 °C, and 10% at 3 °C [bib_ref] Rain versus snow in the Sierra Nevada, California: Comparing Doppler profiling radar..., Lundquist [/bib_ref]. To assess the associated uncertainty with this observed estimate of rain-snow partitioning, we assume a random rain-to-snow percentage following uniform distribution between 0 and 1 at each grid cell and each day. We then apply a Monte Carlo method and found that the resultant uncertainty in the total snowfall is below 1 mm. This is assumed to be negligible.
## E. uncertainties in lapse rate
Lapse rates on the windward side of Sierra Nevada were found to range between 3.5 and 5.0 °C/km. Assuming the lapse rate is uniformly distributed between 3.5 and 5.0 °C/km, then 95% confidence interval for lapse rate is also between 3.5 to 5.0 °C/km. If we modify the lapse rate to 3.5 or to 5.0 °C /km, this changes $ " and $ + by 1 to 11 mm in addition to the uncertainties caused by other factors. The total uncertainties are shown in .
## F. propagation of uncertainties
For variables influenced by more than one uncertainty, total uncertainty is the root sum square of each uncertainty. Therefore, the total uncertainty is a combined estimate of systematic errors and random errors with a 95% confidence interval.
We calculate the uncertainties in ablation , column M) and in decomposed model-reference difference using the upper/lower bound in the 95% confidence interval of upstream features, so the uncertainties propagate to determine the uncertainty in the decomposition presented in .
[fig] Figure S1: Land surface topography (in km), and 20-years mean winter precipitation (P, in mm) and temperature (T, in °C) from the reference dataset (REF) and CRCM5 model running at 0.44° (CRCM5-44), 0.22° (CRCM5-22) and 0.11° (CRCM5-11) spatial resolution. Data is linearly interpolated to the Sierra Nevada transect line as shown in Figure 1. P and T are temporally averaged over 20 winters between November and March from year 1985 to 2005. [/fig]
[fig] Figure S2: Air temperature from the NA-CORDEX regional climate model (RCM) simulations forced by global climate model (GCM), compared against the PRISM reference dataset. The plotted climatological mean spans 20-years from OCT 1985 to SEP 2005, averaged over the 10 California headwater regions. Simulations are named in the format of "GCM_RCM_resolution". Line-style indicates the RCM and color indicates the GCM used for boundary forcing. [/fig]
|
Depressive Symptoms and Risk of New Cardiovascular Events or Death in Patients with Myocardial Infarction: A Population-Based Longitudinal Study Examining Health Behaviors and Health Care Interventions
Background: Depressive symptoms is associated with adverse cardiovascular outcomes in patients with myocardial infarction (MI), but the underlying mechanisms are unclear and it remains unknown whether subgroups of patients are at a particularly high relative risk of adverse outcomes. We examined the risk of new cardiovascular events and/or death in patients with depressive symptoms following first-time MI taking into account other secondary preventive factors. We further explored whether we could identify subgroups of patients with a particularly high relative risk of adverse outcomes.Methods and Results:We conducted a prospective population-based cohort study of 897 patients discharged with firsttime MI between 1 January 2009 and 31 December 2009, and followed up until 31 July 2012. Depressive symptoms were found in 18.6% using the Hospital Anxiety and Depression Scale (HADS-D$8). A total of 239 new cardiovascular events, 95 deaths, and 288 composite events (239 new cardiovascular events and 49 deaths) occurred during 1,975 person-years of follow-up. Event-free survival was evaluated using Cox regression analysis. Compared to the 730 patients without depressive symptoms (HADS-D,8), the 167 patients with depressive symptoms (HADS-D$8) had age-and sex-adjusted hazard ratios [HR] (95% confidence interval [CI]) of 1.53 (95% CI, 1.14-2.05) for a new cardiovascular event, 3.10 (95% CI, 2.04-4.71) for death and 1.77 (95% CI, 1.36-2.31) for a composite event. The associations were attenuated when adjusted for disease severity, comorbid conditions and physical inactivity; HR = 1.17 (95% CI, 0.85-1.61) for a new cardiovascular event, HR = 2.01 (95% CI, 1.28-3.16) for death, and HR = 1.33 (95% CI, 1.00-1.76) for a composite event. No subgroups of patients had a particularly high risk of adverse outcomes.Conclusions: Depressive symptoms following first-time MI was an independent prognostic risk factor for death, but not for new cardiovascular events. We found no subgroups of patients with a particularly high relative risk of adverse outcomes.
# Introduction
Major depression following myocardial infarction (MI) affects 16% to 27% of patients within 2 weeks after MI [bib_ref] Prevalence of depression in survivors of acute myocardial infarction. Review of the..., Thombs [/bib_ref]. Post-MI depression is associated with about a doubling of the risk of new cardiovascular events or death [bib_ref] Prognostic association of depression following myocardial infarction with mortality and cardiovascular events:..., Meijer [/bib_ref] [bib_ref] Prognostic association of depression following myocardial infarction with mortality and cardiovascular events:..., Van Melle [/bib_ref]. The explanatory mechanisms remain unclear [bib_ref] Depression as a risk factor for cardiac mortality and morbidity: a review..., Carney [/bib_ref] even if several mechanisms have been suggested, among others poor adherence to recommended lifestyle and secondary prophylactic medication advice [bib_ref] Patients with depression are less likely to follow recommendations to reduce cardiac..., Ziegelstein [/bib_ref] , poor social support [bib_ref] Living alone after myocardial infarction, Case [/bib_ref] [bib_ref] Social support, depression, and mortality during the first year after myocardial infarction, Frasure-Smith [/bib_ref] , severe cardiac disease [bib_ref] Selfreported depressive symptoms, diagnosed clinical depression and cardiac morbidity and mortality after..., Zuidersma [/bib_ref] , low heart rate variability [bib_ref] Low heart rate variability and the effect of depression on post-myocardial infarction..., Carney [/bib_ref] , inflammatory processes [bib_ref] Depression, C-reactive protein and two-year major adverse cardiac events in men after..., Frasure-Smith [/bib_ref] , and fewer invasive cardiovascular procedures [bib_ref] Mental disorders and use of cardiovascular procedures after myocardial infarction, Druss [/bib_ref]. In patients with stable coronary heart disease, Whooley et al. [bib_ref] Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with..., Whooley [/bib_ref] found that the association between depressive symptoms and new cardiovascular events or death was largely explained by health behavior, especially physical inactivity. They therefore suggested that the adverse effect of depression on the prognosis of coronary heart disease might be prevented through behavioral modification. Exercise may be just as effective at reducing depressive symptoms in patients with coronary heart disease as antidepressants [bib_ref] Exercise and Pharmacological Treatment of Depressive Symptoms in Patients With Coronary Heart..., Blumenthal [/bib_ref] , but it is unknown whether these results apply to patients with MI [bib_ref] Selfreported depressive symptoms, diagnosed clinical depression and cardiac morbidity and mortality after..., Zuidersma [/bib_ref] [bib_ref] Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with..., Whooley [/bib_ref]. It also remains unknown whether subgroups of MI-patients with depressive symptoms are at a particularly high risk of adverse outcomes and whether treatment of post-MI depression improves the adverse cardiovascular outcomes in these persons. Zuidersma et al. [bib_ref] Selfreported depressive symptoms, diagnosed clinical depression and cardiac morbidity and mortality after..., Zuidersma [/bib_ref] found that the association between post-MI depressive symptoms and new cardiovascular events or death was largely explained by cardiac disease severity, but they did not take into account physical activity, treatment of depression, or cardiac rehabilitation.
In the present prospective population-based cohort study of 897 participants, we examined the association between depressive symptoms following first-time MI and new cardiovascular events and/or death, taking into account disease severity, health behavior, use of health care interventions, and social and demographic characteristics. We further explored whether we could identify any subgroups of patients with a particularly high relative risk of adverse outcomes.
# Methods
# Ethic statement
The study was approved by the Danish Data Protection Agency (J.nr. 2009-41-3018), the Scientific Research Evaluation Committee of the Danish Academy of General Practitioners (ref no. 03-2009), and written informed consent was obtained from the patients.
## Study design and participants
We conducted a population-based cohort study comprising people in the Central Denmark Region (1,250,000 inhabitants) with a first-time MI based on data from nationwide registers and questionnaires. The establishment of the cohort is described in detail elsewhere [bib_ref] Rehabilitation status three months after first-time myocardial infarction, Larsen [/bib_ref]. Briefly, we consecutively invited all patients discharged from hospital with a first-time MI from 1 January 2009 to 31 December 2009. Data on patients discharged with an MI (International Classification of Diseases (ICD-10) code I21) [bib_ref] Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish..., Joensen [/bib_ref] were received from the Danish National Patient Register on a monthly basis. Patients who had been discharged with an MI between 1994 and 2008 were excluded to identify first-time cases. Information on name, address, and vital status was obtained from the Civil Registration System, which also provided the unique personal identification number used to link data between the registers and questionnaires.
## Data collection
A questionnaire was mailed to all patients 12 to 14 weeks after their discharge from hospital. The questionnaire was pilot-tested and non-responders received two reminders [bib_ref] Rehabilitation status three months after first-time myocardial infarction, Larsen [/bib_ref].
## Depressive symptoms
We assessed depressive symptoms using the Hospital Anxiety and Depression Scale (HADS) [bib_ref] The hospital anxiety and depression scale, Zigmond [/bib_ref]. The HADS is a self-report instrument that consists of a depression scale (HADS-D) with seven items that are each answered on a four-point verbal rating scale with a score of 0-3 (total score 0-21). The HADS was designed to be valid in clinical populations with symptoms of physical disease. It thus avoids items that might be endorsed by physical rather than mental states [bib_ref] The hospital anxiety and depression scale, Zigmond [/bib_ref]. The HADS has been validated in MI patients [bib_ref] A confirmatory factor analysis of the Hospital Anxiety and Depression Scale in..., Martin [/bib_ref] [bib_ref] Performance characteristics of depression screening instruments in survivors of acute myocardial infarction:..., Thombs [/bib_ref] and has been proved to have satisfactory reliability (HADS-D Cronbach's a < 0.80) [bib_ref] A confirmatory factor analysis of the Hospital Anxiety and Depression Scale in..., Martin [/bib_ref]. Compared with a physician-administered structured clinical interview for the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV), which is the golden standard, a HADS-D$8 identified possible cases of depression in a general practitioner population with a sensitivity of 80% and a specificity of 88% [bib_ref] The Hospital Anxiety and Depression Rating Scale: a cross-sectional study of psychometrics..., Olsson [/bib_ref] and among MI patients with a sensitivity of 65% and a specificity of 90% [bib_ref] Performance characteristics of depression screening instruments in survivors of acute myocardial infarction:..., Thombs [/bib_ref].
## Co-morbidity and cardiac disease severity
Information on co-morbidity was retrieved from the Danish National Patient Register, the Danish National Diabetes Register, and the prescription database covering the entire Central Denmark Region. The Danish National Patient Register provided information on stroke (ICD-10: I61, I63, I64), transient ischemic attack (ICD-10: DG45, DG46), heart failure (ICD-10: I11.0, I13.0, I13.2, I42.0, I42.6, I42.7, I42.9, I50.0, I50.1, I50.9), and revascularization (ICD-10: KFN, KFW) from 1994 to 2008. The Danish National Diabetes Register provided information on diabetes from 1990 to 2008 according to an algorithm developed on the basis of information from four nationwide registers [bib_ref] The National Diabetes Register, Carstensen [/bib_ref]. The prescription database provided information on all reimbursed drugs according to the Anatomical Therapeutic Chemical Classification System (ATC), dispensing dates, and the total number of tablets dispensed. Persons were categorized with hypertension if they had redeemed a combination therapy with at least two antihypertensive drugs (ATC: C02A-C, C02D, C03A-E, C03X, C04, C05, C07, C08, C09) 0 to 180 days before the index MI, as validated previously [bib_ref] Validation of risk stratification schemes for predicting stroke and thromboembolism in patients..., Olesen [/bib_ref]. Persons were categorized with depression before MI if they had redeemed an antidepressant (ATC: N06A) 0 to 180 days before the index MI.
The cardiac disease severity was measured using the British Medical Research Council (MRC) dyspnea score, a self-report instrument [bib_ref] Standardised questionnaire on respiratory symptoms: a statement prepared and approved by the..., Fletcher Cm [/bib_ref]. A score $3 has been shown to provide a simple and valid method for predicting overall mortality [bib_ref] Should we continue using questionnaires on breathlessness in epidemiologic surveys?, Vestbo [/bib_ref].
## Health behavior, health status, and health care interventions
Smoking, alcohol use, physical activity, intake of fruit and vegetables, intake of fish, intake of fish oil supplement, height, and weight (body-mass index = weight [kg] per height [m 2 ]) were selfreported and classified according to the general recommendations from the Danish National Board of Health [bib_ref] Rehabilitation status three months after first-time myocardial infarction, Larsen [/bib_ref]. To assess physical activity, we asked, ''How many days per week are you physically active for at least 30 minutes per day? Include any physical activity at your work or in your spare time where you feel that your pulse rate increases''. Participants chose from one of the following eight categories: no days, 1-6 days per week, every day. We defined cardiac rehabilitation [bib_ref] Secondary prevention through cardiac rehabilitation: position paper of the Working Group on..., Giannuzzi [/bib_ref] in the questionnaire and asked the patients whether they had participated in hospital-based phase II cardiac rehabilitation. Those who responded ''yes, and I took part'' were classified as ''participants''; those who responded ''yes, but I didn't take part'' or ''no'' were classified as ''nonparticipants'' [bib_ref] Rehabilitation status three months after first-time myocardial infarction, Larsen [/bib_ref].
Drug prescription data were obtained from the prescription database. Data on aspirin (ATC: B01AC06), clopidogrel (ATC: B01AC04), statins (ATC: C10AA), b-blockers (ATC: C07), ACEinhibitors/angiotensin 2 receptor blockers (ATC: C09), and antidepressants (ATC: N06A) were collected. We calculated whether the patient had tablets available on the day that we sent the questionnaire (the number of tablets on the last redeemed prescription before the questionnaire was sent $ the number of days to the questionnaire was sent) and defined the patient as receiving treatment if tablets were available. We defined the patient as receiving secondary prophylactic medication if the patient was receiving treatment with three or more of the following drugs: aspirin, clopidogrel, statins, and b-blockers.
## Other patient characteristics
Data on age at MI and sex were obtained from the Civil Registration System. Each patient's social and demographic characteristics from the year before MI (2008) were retrieved from the Danish Integrated Database for Labor Market Research.
## Cardiovascular events and death
Outcome events were new cardiovascular events (MI, heart failure, stroke, or transient ischemic attack), all-cause mortality, and a composite endpoint comprising a new cardiovascular event or death. Information on outcomes was collected from baseline to the last day of follow-up. The Danish National Patient Register provided information on cardiovascular events from all hospital contacts. Vital status (dead or alive) was obtained from the Civil Registration System.
# Statistical analysis
Baseline differences in characteristics between MI patients with and without depressive symptoms were compared using t tests and x2 tests. We calculated the event-free incidence-time as the time from 3 months after the MI (baseline evaluation of depressive symptoms) to the first cardiovascular event or death. If no event or death occurred, the patient was censored at 31 July 2012. Two persons emigrated during the time of follow-up and they were censored at the time of their emigration. Owing to the use of nationwide registers, we had complete follow-up of all patients. The risk of cardiovascular events or death associated with depressive symptoms was estimated using Cox proportional hazards models. We evaluated whether the hazard ratios (HR) of depressive symptoms following MI varied by subgroup by testing for interaction using Wald test in an age-adjusted model. The covariates for the multivariate model (age, sex, history of stroke, diabetes, or heart failure, cardiac disease severity, smoking, secondary prophylactic medication, and physical activity) were chosen on the basis of previous studies. No variable had more than 3.1% missing data. P,.05 was considered statistically significant.
# Results
Among a total of 1,288 eligible patients with first-time MI, 897 (69.6%) completed a questionnaire and 167 (18.6%) of those had depressive symptoms (HADS-D$8). Non-responders were more often women, older, with fewer socioeconomic resources, and had more comorbid conditions than responders. Compared with MI patients without depressive symptoms, MI patients with depressive symptoms had more severe cardiac disease, more comorbid conditions, used antidepressants more frequently, and received secondary prophylactic medication less frequently. They were also more likely to be female, living alone, outside the work force, smokers, less physically active, and have a smaller intake of fish than MI patients without depressive symptoms . Adjustments for cardiac disease severity and comorbid conditions attenuated the HRs. The associations remained statistically significant for death (HR, 2.29; 95% CI, 1.48-3.53; P,0.001) and for a composite event (HR, 1.41; 95% CI, 1.07-1.86; P = .015) but not for a new cardiovascular event (HR, 1.25; 95% CI, 0.91-1.71; P = .164) [fig_ref] Table 3: Association Between Baseline Depressive Symptoms and Subsequent Cardiovascular Events or Death [/fig_ref].
Additional adjustment for physical activity further attenuated the associations. In the final adjusted model, MI patients with depressive symptoms had a 35% higher rate of a new cardiovascular event or death (HR, 1.35; 95% CI, 1.02-1.79; P = .034; [fig_ref] Table 3: Association Between Baseline Depressive Symptoms and Subsequent Cardiovascular Events or Death [/fig_ref] than MI patients without depressive symptoms. In the final model, depressive symptoms were associated with an increased rate of death (HR, 2.07; 95% CI, 1.32-3.25; P = .001), but not of new cardiovascular events (HR, 1.18; 95% CI, 0.86-1.62; P = .302; [fig_ref] Table 3: Association Between Baseline Depressive Symptoms and Subsequent Cardiovascular Events or Death [/fig_ref].
We found no statistically significant difference in the HR between subgroups of MI patients characterized by sex, marital status, education, labor market status, body mass index, comorbidity, history of depression, cardiac disease severity, antidepressant use, secondary prophylactic medication, alcohol consumption, smoking, intake of fruit and vegetables, intake of fish, intake of fish oil supplement, physical activity, or participation in phase 2 cardiac rehabilitation [fig_ref] Figure 2: Association Between Depressive Symptoms and Subsequent Cardiovascular Events or Death for Patients... [/fig_ref]. However, the association between depressive symptoms and new cardiovascular events or death tended to be smaller among person who took antidepressants (P value for interaction = 0.35) or were physical active (P value for interaction = 0.12).
# Discussion
Depressive symptoms following MI was associated with an increased risk of a new cardiovascular event and/or death. However, the associations were confounded by the severity of the underlying heart disease and physical inactivity but not by other secondary preventive factors. After adjusting for these confounders, post-MI depressive symptoms remained an independent prognostic risk factor for death but not for new cardiovascular events.
A recent meta-analysis [bib_ref] Prognostic association of depression following myocardial infarction with mortality and cardiovascular events:..., Meijer [/bib_ref] found that patients with depressive symptoms following MI had a 2.25 (95% CI, 1.73-2.93) times higher risk of all-cause mortality and a 1.59 (95% CI, 1.37-1.85) times higher risk of new cardiac events than patients without such symptoms. Only eight studies [bib_ref] Gender, depression, and one-year prognosis after myocardial infarction, Frasure-Smith [/bib_ref] [bib_ref] The impact of negative emotions on prognosis following myocardial infarction: is it..., Frasure-Smith [/bib_ref] [bib_ref] Affective disorders and survival after acute myocardial infarction. Results from the postinfarction..., Ladwig [/bib_ref] [bib_ref] Depression and prognosis following hospital admission because of acute myocardial infarction, Lauzon [/bib_ref] [bib_ref] Depression predicts mortality and hospitalization in patients with myocardial infarction complicated by..., Rumsfeld [/bib_ref] [bib_ref] Depressive symptoms predict 12-month prognosis in elderly patients with acute myocardial infarction, Shiotani [/bib_ref] [bib_ref] The association of cognitive and somatic depressive symptoms with depression recognition and..., Smolderen [/bib_ref] [bib_ref] Depression assessed over 1-year survival in patients with myocardial infarction, Sorensen [/bib_ref] provided adjusted estimates and they were too heterogeneous to pool into a common estimate. In these studies, the estimates were on average 21% lower after adjustment for cardiac disease severity and comorbidity. In a recent study, Zuidersma et al. [bib_ref] Selfreported depressive symptoms, diagnosed clinical depression and cardiac morbidity and mortality after..., Zuidersma [/bib_ref] found that one third to half of the association between post-MI depressive symptoms and cardiovascular events or death was explained by cardiac disease severity and previous MI. Whooley et al. [bib_ref] Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with..., Whooley [/bib_ref] found that physical inactivity partly explains the association between depressive symptoms and new cardiovascular events or death in patients with stable coronary heart disease. Our study adds that it also seems to be the case for patients with first-time MI. These interpretations assume that physical inactivity act as a confounder (physical inactivity increases the risk of post-MI depression and increases the risk of adverse outcome) and therefore should be adjusted for. If physical inactivity act as a mediator (a step on the causal pathway from depressive symptoms to the adverse outcome) it should not be adjusted for. In sub-analyses, we adjusted for other potential mediators (marital status, education, labor market status, body mass index, antidepressant use, and participation in phase-2 cardiac rehabilitation), but this did not change the estimates. In a sub-analysis, we also excluded patients with the more severe underlying physical disease (MRC$3, previous stroke, heart failure or diabetes mellitus) and found that patients with depressive symptoms had a 1.91 (95% CI, 0.76-4.78; P = .166) times higher risk of death than patients without these symptoms, albeit the results did not reach statistical significance. These findings support that post-MI depressive symptoms seem to be an independent prognostic risk factor for death. We examined the association between post-MI depressive symptoms and adverse outcome in several subgroups, but identified no factors that modified the risk. However, the sample size was low in some of the subgroups, and we found a tendency towards a lower association with increasing physical activity and among users of antidepressants. Larger studies are needed to clarify the impact of these potential modifiers of the association and to evaluate how such modifiers may be catered for in the treatment of post-MI patients with symptoms of depression. MI is a severe life event that even increases the risk of suicide [bib_ref] Myocardial infarction and risk of suicide: a population-based case-control study, Larsen [/bib_ref] and general recommendations are that clinicians should recognize and treat post-MI depression [bib_ref] Depression and coronary heart disease: recommendations for screening, referral, and treatment: a..., Lichtman [/bib_ref] [bib_ref] Infarction Depression Clinical Practice Guideline Panel (2009) AAFP guideline for the detection..., Post-Myocardial [/bib_ref] [bib_ref] Screening for depression in patients with myocardial infarction by general practitioners, Larsen [/bib_ref]. Randomized trials have found that antidepressants and cognitive behavioral therapy reduce depressive symptoms in persons with MI [bib_ref] Psychological and pharmacological interventions for depression in patients with coronary artery disease, Baumeister [/bib_ref] [bib_ref] Does Evidence Support the American Heart Association's Recommendation to Screen Patients for..., Thombs [/bib_ref]. Physical exercise reduces depressive symptoms in patients with stable coronary heart disease [bib_ref] Exercise and Pharmacological Treatment of Depressive Symptoms in Patients With Coronary Heart..., Blumenthal [/bib_ref] , but no interventional studies have examined the effect of exercise on depressive symptoms in patients with a recent MI. In these patients, physical exercise should be supervised and preceded by a treadmill stress test for patient security and to reassure the patients that their hearts are strong enough to withstand regular exercise training. Furthermore, it remains unknown whether treatment of depression improves the adverse cardiovascular outcomes of persons with coronary heart disease. Most trials have been underpowered to detect a potential effect on cardiovascular events or death, but one observational study has suggested that antidepressants reduce death and recurrent MI in patients with post-MI depression [bib_ref] Effects of antidepressant medication on morbidity and mortality in depressed patients after..., Taylor [/bib_ref]. Mounting evidence shows that comprehensive and collaborative care is effective in managing persons with depression and co-existing physical illness [bib_ref] Collaborative care for patients with depression and chronic illnesses, Katon [/bib_ref] , including MI [bib_ref] Enhanced depression care for patients with acute coronary syndrome and persistent depressive..., Davidson [/bib_ref]. Such care includes components like education about the condition, interventions to encourage physical exercise, and systematic monitoring of the patient's adherence to medication [bib_ref] A pilot randomised controlled trial of personalised care for depressed patients with..., Tylee [/bib_ref]. Initiatives should be taken to implement comprehensive and collaborative care, and future studies should evaluate whether these strategies also improve the overall prognosis.
Previous studies suggested that the association between post-MI depressive symptoms depended on whether the depression was present before the onset of the MI or if it was a new depression that arose after MI [bib_ref] Depression following myocardial infarction: first-ever versus ongoing and recurrent episodes, Spijkerman [/bib_ref] [bib_ref] Onset of major depression associated with acute coronary syndromes: relationship of onset,..., Glassman [/bib_ref]. In a sub-analysis, patients with a history of depression were excluded, but this did not change the estimates.
The major strengths of our study were its population-based nature, the homogenous study population of patients with firsttime MI, the high response rate, the complete follow-up, and the opportunity to analyze the importance of several suggested mechanisms. Our information on MI was registered prospectively and did not rely on participants' or relatives' memory. The MI diagnosis in the National Patient Register was based on current European Society of Cardiology criteria for MI, coded by the physician in charge of the discharge, and is known to have a high sensitivity (90%) and specificity (92%) that is unrelated to age [bib_ref] Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish..., Joensen [/bib_ref]. The specificity of our study was even higher because we confirmed the MI diagnosis by reviewing the discharge summaries [bib_ref] Rehabilitation status three months after first-time myocardial infarction, Larsen [/bib_ref] , which reduced the risk of information bias. A diagnosis of depression should ideally be based on a diagnostic interview. However, the diagnostic usefulness of a HADS$8 has a high quality with a specificity of 90% in MI patients [bib_ref] Performance characteristics of depression screening instruments in survivors of acute myocardial infarction:..., Thombs [/bib_ref] , and it hence performs well compared with a physician-administered structured clinical interview for DSM-IV. With a cut-off of a HADS-D$11, the specificity approximates 100% [bib_ref] Performance characteristics of depression screening instruments in survivors of acute myocardial infarction:..., Thombs [/bib_ref] [bib_ref] The Hospital Anxiety and Depression Rating Scale: a cross-sectional study of psychometrics..., Olsson [/bib_ref] and the use of this cutoff did not change the estimates of our analyses. Moreover, selfreported depressive symptoms are a more accurate predictor of cardiac morbidity and mortality than clinical depression [bib_ref] Selfreported depressive symptoms, diagnosed clinical depression and cardiac morbidity and mortality after..., Zuidersma [/bib_ref].
HADS has a relatively low sensitivity (65%), probably because it does not include somatic symptoms of depression because they could be confused with symptoms of the cardiac disease. We may therefore have underestimated the number of patients with depressive symptoms leading to an attenuation of the association between depressive symptoms and new cardiac events or death. However, the prevalence rate (18.6%) of depressive symptoms in our sample corresponds to prevalence rates of other studies [bib_ref] Prevalence of depression in survivors of acute myocardial infarction. Review of the..., Thombs [/bib_ref] [bib_ref] Depression and coronary heart disease: recommendations for screening, referral, and treatment: a..., Lichtman [/bib_ref]. We also reduced the risk of information bias by using previously translated and validated scales, pilot-testing the questionnaire among MI-patients, and using high-quality register data [bib_ref] Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish..., Joensen [/bib_ref] [bib_ref] The National Diabetes Register, Carstensen [/bib_ref]. Comorbid conditions and social and demographic characteristics were evaluated before the MI. Even though we had a high response rate of 70%, we found that non-responders tended to be frailer than responders when we compared their social resources and comorbid conditions, which could affect our estimates. In order to address the potential risk of selection bias, we used antidepressant consumption as a proxy for depressive symptoms similarly to previous studies. [bib_ref] Epidemiology of multimorbidity and implications for health care, research, and medical education:..., Barnett [/bib_ref] The estimates of the association between antidepressant consumption and new cardiovascular events or death in responders (HR, 1.49; 95% CI, 1.09-2.05) and in non-responders (HR, 1.40; 95% CI, 1.05-1.87) were similar, indicating that the possibility of selection bias was limited. Another strength is that we evaluated depressive symptoms 3 months post-MI in contrast to many other studies that evaluated depression during hospital admission. This allowed natural recovery of depressive symptoms after a stressful life event.
We were unable to evaluate whether the association is explained by potential biological mechanisms such as heart rate variability and inflammatory mechanisms because we had no information on this. In the study by Whooley et al. [bib_ref] Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with..., Whooley [/bib_ref] , inflammation as measured by C-reactive protein explained a small part of the association, but it is unclear whether inflammation acted as a mediator between depressive symptoms and cardiovascular events or death or as a marker of more severe disease. Our study evaluated depressive symptoms, cardiac disease severity, behavioral factors, and treatment strategies at the same time. We therefore cannot determine whether they were the cause or result of depressive symptoms. Lifestyle behavior was self-reported, and it is possible that patients with depressive symptoms were more likely to underreport adverse lifestyle including physical inactivity. However, patients with depressive symptoms in our study did report adverse lifestyle, and Whooley et al. [bib_ref] Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with..., Whooley [/bib_ref] found no difference when substituting self-reported physical activity with an objective measure of physical fitness.
In conclusion, we found that depressive symptoms following first-time MI was associated with an increased risk of new cardiovascular events and/or death. The association was partly explained by disease severity and physical inactivity. Depressive symptoms remained an independent prognostic risk factor for death but not for new cardiovascular events. Mounting evidence shows that comprehensive and collaborative care is effective in managing persons with depression and co-existing physical illness [bib_ref] Collaborative care for patients with depression and chronic illnesses, Katon [/bib_ref] [bib_ref] Collaborative care for depression: a cumulative meta-analysis and review of longer-term outcomes, Gilbody [/bib_ref] , including MI [bib_ref] Enhanced depression care for patients with acute coronary syndrome and persistent depressive..., Davidson [/bib_ref]. Such care includes components like education about the condition, interventions to encourage physical exercise, and systematic monitoring of the patient's adherence to medication [bib_ref] A pilot randomised controlled trial of personalised care for depressed patients with..., Tylee [/bib_ref]. Our results support the relevancy of examining whether collaborative care also improves the overall prognosis of patients with depressive symptoms following MI.
## Supporting information
Table S1 Comparison of responders and non-responders. (DOCX)
[fig] Figure 2: Association Between Depressive Symptoms and Subsequent Cardiovascular Events or Death for Patients with Myocardial Infarction by Specific Characteristics. doi:10.1371/journal.pone.0074393.g002 [/fig]
[table] Table 1: A total of 239 new cardiovascular events, 95 deaths, and 288 composite events (239 new cardiovascular events and 49 deaths) occurred during 1,975 person years of follow-up (mean follow-up: 2.2 years; SD 1.0); 59 (35.3%) new cardiovascular events, 37 (22.2%) deaths and 76 (45.5%) composite events occurred among persons with a HADS$8, and 180 (24.7%) new cardiovascular events, 58 (8.0%) deaths and 212 (29.0%) composite events occurred among persons with a HADS,8 (Table 2). Compared to the MI-patients without depressive symptoms, the MI-patients with depressive symptoms had age-and sex-adjusted hazard ratios [HR] (95% confidence interval [CI]) of 1.53 (95% CI, 1.14-2.05; P = .005) for a new cardiovascular event, 3.10 (95% CI, 2.04-4.71; P,.001) for death and 1.77 (95% CI, 1.36-2.31; P,.001) for a composite event (Table 2, [/table]
[table] Table 3: Association Between Baseline Depressive Symptoms and Subsequent Cardiovascular Events or Death. [/table]
|
A phase 2 study of an oral mTORC1/mTORC2 kinase inhibitor (CC-223) for non-pancreatic neuroendocrine tumors with or without carcinoid symptoms
Second-generation mammalian target of rapamycin (mTOR) inhibitors such as CC-223 may have theoretical advantages over first-generation drugs by inhibiting TOR kinase in mTOR complex 1 (mTORC1) and 2 (mTORC2), potentially improving clinical efficacy for well-differentiated neuroendocrine tumors (NET).Enrolled patients had metastatic, well-differentiated NET of non-pancreatic gastrointestinal or unknown origin, with/without carcinoid symptoms, had failed �1 systemic chemotherapy, and were taking a somatostatin analog (SSA). Oral once-daily CC-223 was administered in 28-day cycles starting at 45 mg (n = 24), with a subsequent cohort starting at 30 mg (n = 23). Objectives were to evaluate tolerability, preliminary efficacy, and pharmacokinetic and biomarker profiles of CC-223. Forty-seven patients completed the study, with mean treatment duration of 378 days and mean dose of 26 mg; 26% of patients remained on the starting dose. Most frequent grade �3 toxicities were diarrhea (38%), fatigue (21%), and stomatitis (11%). By investigator, 3 of 41 evaluable patients (7%) showed partial response (PR) and 34 (83%) had stable disease (SD) for a disease control rate (DCR) of 90% (95% confidence interval [CI] 76.9-97.3%). Duration of PR was 125-401 days; median SD duration was 297 days (min-max, 50-1519 days). Median progression-free survival was 19.5 months (95% CI 10.4-28.5 months). Carcinoid symptoms of flushing, diarrhea, or both improved in 50%, 41%, and 39% of affected patients, respectively. For the first time, this study describes that a second-generation mTOR pathway PLOS ONE | https://doi.org/10.1371/journal.pone.
# Abstract
Second-generation mammalian target of rapamycin (mTOR) inhibitors such as CC-223 may have theoretical advantages over first-generation drugs by inhibiting TOR kinase in mTOR complex 1 (mTORC1) and 2 (mTORC2), potentially improving clinical efficacy for well-differentiated neuroendocrine tumors (NET).Enrolled patients had metastatic, well-differentiated NET of non-pancreatic gastrointestinal or unknown origin, with/without carcinoid symptoms, had failed �1 systemic chemotherapy, and were taking a somatostatin analog (SSA). Oral once-daily CC-223 was administered in 28-day cycles starting at 45 mg (n = 24), with a subsequent cohort starting at 30 mg (n = 23). Objectives were to evaluate tolerability, preliminary efficacy, and pharmacokinetic and biomarker profiles of CC-223. Forty-seven patients completed the study, with mean treatment duration of 378 days and mean dose of 26 mg; 26% of patients remained on the starting dose. Most frequent grade �3 toxicities were diarrhea (38%), fatigue (21%), and stomatitis (11%). By investigator, 3 of 41 evaluable patients (7%) showed partial response (PR) and 34 (83%) had stable disease (SD) for a disease control rate (DCR) of 90% (95% confidence interval [CI] 76.9-97.3%). Duration of PR was 125-401 days; median SD duration was 297 days (min-max, 50-1519 days). Median progression-free survival was 19.5 months (95% CI 10.4-28.5 months). Carcinoid symptoms of flushing, diarrhea, or both improved in 50%, 41%, and 39% of affected patients, respectively. For the first time, this study describes that a second-generation mTOR pathway
# Introduction
Neuroendocrine tumors (NET), although generally believed to be uncommon, are the second most common gastrointestinal (GI) malignancy, superseded only by colorectal neoplasias [bib_ref] One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine..., Yao [/bib_ref]. This malignancy is reported to be increasing in prevalence, with the greatest increase in lung, small intestine, and rectal NETs [bib_ref] One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine..., Yao [/bib_ref]. Although survival has improved with introduction of newer therapies, prognosis for patients with advanced disease remains poor [bib_ref] One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine..., Yao [/bib_ref]. The phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K)-protein kinase B (AKT)-mammalian target of rapamycin (mTOR) signalling pathway, which is inappropriately activated in many cancers, plays a central role in the genesis and progression of NET. The mTOR complex 1 (mTORC1) inhibitor everolimus has become well-established as an effective therapy for NET arising from the pancreas, lung, or GI tract, with an associated median progression-free survival (PFS) of 11.0 months [bib_ref] Everolimus for the treatment of advanced, non-functional neuroendocrine tumours of the lung..., Yao [/bib_ref]. mTOR kinase is a serine/threonine kinase that serves as a core component of two complexes; mTORC1 and mTORC2. The former modulates cell proliferation via phosphorylation of eukaryotic translation initiation factor 4E binding protein 1 (4E-BP1) and p70 ribosomal S6 kinase 1 (S6K1), whereas the latter acts via phosphorylation of AKT [bib_ref] Potentiation of the anticancer effects of everolimus using a dual mTORC1/2 inhibitor..., Kim [/bib_ref]. Everolimus inhibits the mTORC1 complex but may upregulate mTORC2, which thus maintains AKT activation. Consequently, inhibitors that can target both complexes may offer clinical advantages in terms of improved efficacy [bib_ref] Potentiation of the anticancer effects of everolimus using a dual mTORC1/2 inhibitor..., Kim [/bib_ref] [bib_ref] Doubling Down on mTOR Inhibition: Harnessing ZEBRA for Insights, Chan [/bib_ref] [bib_ref] Inhibition of mTOR's Catalytic Site by PKI-587 Is a Promising Therapeutic Option..., Freitag [/bib_ref] [bib_ref] Dual mTOR inhibitor MLN0128 suppresses Merkel cell carcinoma (MCC) xenograft tumor growth, Kannan [/bib_ref].
CC-223 is an inhibitor of TOR kinase in mTORC1 and mTORC2 that is known to overcome mTORC2 feedback pathway upregulation and thus potentially minimize development of therapy resistance due to increased phosphorylation of AKT [bib_ref] CC-223, a Potent and Selective Inhibitor of mTOR Kinase: In Vitro and..., Mortensen [/bib_ref]. In non-clinical studies, inhibition of mTORC1 and mTORC2 leads to more effective inhibition of cancer cell proliferation than does blocking mTORC1 alone [bib_ref] CC-223, a Potent and Selective Inhibitor of mTOR Kinase: In Vitro and..., Mortensen [/bib_ref]. In mice bearing human carcinoid xenografts that secrete serotonin, tumors that progressed on rapamycin remained stable or decreased in volume from baseline after treatment with CC-223 [bib_ref] mTOR Kinase Inhibition Effectively Decreases Progression of a Subset of Neuroendocrine Tumors..., Orr-Asman [/bib_ref].
In a first-in-human study, CC-223 was evaluated as a treatment for patients with advanced solid tumors, and the 45 mg/d dose was established as the maximum tolerated dose (MTD) [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref]. Here, we report results of an expansion cohort in patients with non-pancreatic GI or unknown primary NET treated at the previously-established MTD or lower dose if needed.
# Materials and methods
## Study design and participants
This open-label phase 1/2 study comprised two parts: dose-finding to determine maximum tolerated dose (phase 1) in a variety of solid and hematologic malignancies, the results from which have been previously published [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref] , and cohort expansion (phase 2). In phase 2, adults (�18 years) with locally unresectable or metastatic well-differentiated (grade 1 or 2) non-pancreatic GI NET or NET of unknown primary origin, with or without carcinoid syndrome, who had failed at least one prior systemic anticancer therapy were the principal eligibility criteria. Concurrent therapy with a somatostatin analog (SSA), evidence of radiologic disease progression prior to study start with no receptor-targeted radio-labelled or liver-directed therapy within 3-4 months, and an Eastern Cooperative Oncology Group (ECOG) performance status score no higher than 1 were also requirements for enrolment. Specified laboratory results for inclusion were a hemoglobin level of �9 g/dL, absolute neutrophil count �1.5×10 9 /L, platelet count �100×10 9 /L, serum potassium within normal limits or correctable with supplements, aspartate and alanine aminotransferases �2.5×upper limit of normal (ULN) or �5.0×ULN if liver tumor was present, bilirubin �1.5×ULN or �2×ULN if liver tumor was present, creatinine �1.5×ULN or 24-hour clearance �50 mL/min, and a negative serum or urine pregnancy test within 48 hours before starting CC-223 for females of childbearing potential. Patients with pancreatic NET, pulmonary carcinoid, high-grade poorly differentiated, or rare variant tumor types were excluded. Using the optimal CC-223 dose of 45 mg established in phase 1 [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref] , cohort expansion objectives were to determine the tolerability of CC-223, characterize the pharmacokinetic (PK) and pharmacodynamic (biomarker) profiles, and identify a preliminary efficacy signal in seven parallel cohorts of various preselected tumor types; only the NET cohorts are reported here. Patient flow is shown in [fig_ref] Fig 1: CONSORT diagram. [/fig_ref]. The protocol and supporting CONSORT checklist are available as supporting information (S1 Checklist and S1 Protocol). The first patient was enrolled into the study on July 18, 2011 and the last patient completed the study on October 25, 2016.
## Procedures
CC-223 capsules were taken orally once daily in continuous 28-day cycles. An initial cohort started CC-223 at a dose of 45 mg; this was the maximum tolerated dose (MTD) established in phase 1 but, after observing a high rate of dose adjustments due to toxicity among patients with NET, a second cohort started CC-223 at 30 mg. To mitigate toxicity, dosage could be adjusted with either short interruptions, dose level reductions, or every-other-day administration. Treatment was permanently discontinued for disease progression, unmanageable toxicity, patient withdrawal of consent, or other reasons specified by the protocol as requiring study withdrawal.
Baseline assessments were completed within 28 days prior to first dose of CC-223. Drug tolerability was continuously evaluated until 28 days after the last dose, using clinical event and laboratory test monitoring comprising a conventional hematologic and coagulation (if taking anti-coagulants) panel; serum electrolytes; fasting blood glucose (this also included daily selfmonitoring with finger stick samples for the first month and for longer if necessary), HbA 1c , insulin, and c-peptide; liver enzymes; serum creatinine; and protein. Others included serum uric acid, amylase and lipase, lipoproteins, thyroid function tests, creatine kinase, and immunoglobulins including T-cell subsets (CD4+ and CD8+). We also performed serial triplicate electrocardiograms (ECGs), periodic left ventricular ejection fraction estimations (echo or multiple gated acquisition [MUGA]), regular physical exams and vital sign measurements (heart rate, blood pressure, respiratory rate, and temperature). Adverse events were coded with Medical Dictionary for Regulatory Activities (MedDRA) version 19.0. Severity of adverse events was classified using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. A safety review committee provided study oversight.
Tumors were restaged by the principal investigator at each site using computed tomography (CT), positron emission tomography (PET)-CT, or magnetic resonance imaging scans after every 2 cycles through cycle 6, and every 3 cycles thereafter, using Response Evaluation Criteria In Solid Tumors (RECIST) version 1.1 criteria for response [bib_ref] New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1), Eisenhauer [/bib_ref].
Metabolic changes were assessed by percent change in total of standardized uptake value (SUV) of 2-18 F-fluoro-2-deoxyglucose (FDG) with PET scanning from baseline to day 15 in cycle 1. Carcinoid symptoms of flushing and diarrhea were documented monthly with a non- https://doi.org/10.1371/journal.pone.0221994.g001 mTORC1/C2 kinase inhibitor (CC-223) for non-pancreatic NET validated symptom questionnaire that was completed at each study visit. Patients reported whether they had experienced specific events, and the severity and/or frequency of these events. Symptom improvement was defined as �50% reduction in frequency and/or a 1-grade reduction in severity for each from baseline.
Central laboratories conducted serial stimulated blood biomarker assays in monocytes (CD14+) for p4E-BP1 and pPKB (also known as pAKT) for assessment of mTORC1 and mTORC2 inhibition, respectively. Monthly changes from baseline for chromogranin A (CgA), and for other serum hormones if elevated, were assayed locally.
Blood for PK analysis was drawn on day 1 pre-dose and for up to 48 hours post-dose, and then on day 15 up to 8 hours post-dose. Urine on day 1 was collected within 30 minutes prior to dosing and at intervals up to 24 hours post-dose. Samples were assayed for CC-223 and the principal active metabolite (M1) using validated chiral liquid chromatography mass spectrometry. CC-223 PK parameters were calculated using dosing and sample collection times. Noncompartmental PK analysis was performed used WinNonlin Enterprise version 5.2 Model 200 and Model 210 (Certara, Princeton, NJ).
## Outcomes
The primary objectives were to investigate the safety and tolerability and determine the preliminary PK of CC-223 administered orally. Secondary objectives were to characterize the PK of M1, to evaluate the inhibition of mTORC1 and mTORC2 in peripheral blood using p4E-BP1 and pAKT as biomarkers, and to evaluate the preliminary efficacy of CC-223 in terms of tumor response, PFS, and overall survival (OS).
# Statistical analysis
This phase 1/2 study was not powered for inferential statistics and mostly descriptive statistics were used. All patients dosed with CC-223 were included in the safety analysis. The efficacy analysis included all patients completing at least one cycle of drug who had both baseline and one post-baseline tumor efficacy assessment, unless otherwise specified. Categorical baseline variables including tumor type-specific characteristics were summarized using frequency counts and percentage. Continuous demographic and baseline characteristic variables were summarized by descriptive statistics. A Wilcoxon signed rank test was used to analyse change from baseline in serum hormone levels at selected scheduled visits.
Best overall tumor response by investigator assessment was summarized using frequency tabulation. Waterfall plots were used to show best percentage change from baseline in the sum of length (longest diameter) of target lesions.
Relationship between PFS and NET-specific assessments such as serum hormone levels and carcinoid symptoms were investigated. The Kaplan-Meier estimate of median PFS with its two-sided 95% confidence interval [CI] was provided for each baseline category of the selected biomarkers. Kaplan-Meier plots of PFS were created by category. The raw P value of the log rank test comparing survival distribution of PFS between categories was also provided.
Approximately 40 evaluable patients with NET were planned for further evaluation of safety and preliminary antitumor activity. Patients could receive CC-223 at the MTD and/or a lower dose level based on the safety, PK, and PD data from the dose escalation phase [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref].
Enrollment and tumor response rate monitoring were based on Wald's sequential analysis [bib_ref] Sequential tests of statistical hypotheses, Wald [/bib_ref]. Assuming a target response rate of 20% for the tumor type of interest (NET), based on the sequential design, with 80% power to identify the group with 20% or more response rate at 5% significance level when the response rate is at 10%, the enrollment could be stopped for futility when there was no responder out of up to 14 evaluable patients.
# Results
The study was conducted between July 2011 and October 2016 at 15 investigational sites. From a total of 58 screened patients, 47 met all eligibility criteria. Of those, the first 24 started CC-223 at the MTD of 45 mg/day and the subsequent 23 started at 30 mg/day. Median age was 63 years (min-max: 35-79 years) [fig_ref] Table 1: Patient demographics and baseline characteristics [/fig_ref]. Men comprised 40% of the population; 92% of patients were white, and 64% had ECOG performance status score of 1 on entry. Most (91%) had received prior SSA therapy and 43% had received additional systemic treatment; no patient received prior mTOR inhibitor therapy; however, one patient received concomitant treatment with everolimus. The majority (57%) had midgut, followed by other GI (23%) tumors, and 53% reported NET-related symptoms of either flushing or diarrhea. Eighteen patients (38%) withdrew from the study due to progression of their disease, 9 (19%) due to withdrawal of consent, 8 (17%) due to adverse events, 2 patients (4%) died, and 10 (21%) discontinued for other reasons.
Median treatment duration was 252 days, or 9 cycles (maximum of 1517 days, or 54 cycles), but only 12 (26%) patients remained on the dose initially assigned. Dose adjustment was Of the 46 (98%) patients who reported at least one treatment-related adverse event (TRAE) by investigator determination, 37 (79%) experienced grade 3 or 4 events. Overall, 8 of 47 patients (17%) experienced one or more serious adverse event related to CC-223. A total of 40 (85%) patients required dose adjustments due to toxicity and 6 (13%) discontinued CC-223 because of toxicity. Most common TRAEs were diarrhea (77%), fatigue (66%), stomatitis (53%), pruritus (40%), rash and nausea (38% each), hyperglycemia (28%), decreased appetite and maculopapular rash (26% each), and asthenia (23%) . Most frequent grade 3 or higher toxicities were diarrhea (38%), fatigue (21%), and stomatitis (11%); there were no grade 4 or 5 toxicities reported for the most common TEAEs. Diarrhea (9%), decreased appetite . Most common treatment-related adverse events in �10% of patients (n = 47). (6%), and nausea (4%) were the more common toxicities causing C-223 discontinuation, whereas diarrhea (40%), stomatitis and fatigue (17% each), and increased blood creatinine (13%) required the most dose adjustments. No patient died on study due to treatment-related toxicity. The PK results for patients with NET were consistent with results previously reported for phase 1 [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref]. The sample size for patients with NET who had serial PK sampling was small [fig_ref] Table 3: Pharmacokinetic characteristics for CC-223 and principal metabolite [/fig_ref] and included patients who had dose adjustments; both were limitations to these analyses. Overall, regardless of starting dose, CC-223 was rapidly absorbed with maximum plasma concentrations occurring at a median T max of 1.5-3 hours and, for M1, 3-8 hours. CC-223 and M1 exposure (C max and AUC t ) increased from 30 mg to 45 mg and aligned with doseproportional increases in exposure observed in the overall dose range evaluated in phase 1 [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref]. Significant accumulation of M1 was observed after repeated dosing.
## Any dose (n = 47) cc-223 30 mg (n = 23)
## Cc-223 45 mg (n = 24)
## Adverse event, n (%) any grade, n (%) grade 3, a n (%) any grade, n (%) any grade, n (%)
## Gastrointestinal disorders
Across both dose level cohorts in this study, substantial median biomarker inhibition relative to baseline was achieved for pAKT (>70%) and p4E-BP1 (>26%) at 1.5 hours after multiple dosing on day 15 regardless of whether patients started at 45 mg or 30 mg [fig_ref] Table 4: Regulation of pAKT and p4E-BP1 in monocytes [/fig_ref]. Among the 34 of 47 (72%) patients with NET who had evaluable biomarker results, 13 and 21 patients were dosed initially with 45 mg and 30 mg CC-223, respectively. At 45 mg, blood pAKT showed a median inhibition of 83% 1.5 hours after a single dose (day 1) and 76% after multiple doses (day 15) of CC-223. Inhibition of p4E-BP1 was more moderate and showed a median inhibition of 37% 1.5 hours after a single dose and 47% inhibition after multiple doses. At the lower starting dose of 30 mg, weaker inhibition was observed after the first dose for both biomarkers, but the difference became less pronounced by day 15 after repeated dosing and disappeared completely for pAKT. At 30 mg, pAKT showed a median inhibition from baseline of 59% after the first dose and of 79% at day 15 after multiple doses, while p4E-BP1 showed a median inhibition of 9% following a single dose and of 26% after multiple doses. Taken together, these results demonstrate good inhibition of pAKT and less but meaningful 4E-BP1 inhibition regardless of CC-223 dose after repeat dosing at these dose levels. However, as noted above, the majority of patients could not tolerate the starting doses, and doses were https://doi.org/10.1371/journal.pone.0221994.t003 mTORC1/C2 kinase inhibitor (CC-223) for non-pancreatic NET reduced for extended periods to levels as low as 15 mg for 45% of patients. The pharmacodynamic effects at lower doses were not evaluated. Due to challenges obtaining optional paired biopsies in this population, the effects of CC-223 in tumor tissue could not be evaluated.
Mean percent decrease from baseline to day 15 of treatment in [ 18 F]-FDG PET SUV in 31 patients was 32.0% (standard deviation 35.2%; P<0.0001, Wilcoxon signed rank test). Twentyone patients (44.7%) exhibited at least a 15% reduction in uptake and 18 (38.3%) had at least a 25% reduction in SUV. These metabolic changes did not correlate strongly with subsequent tumor response outcomes using RECIST criteria .
Of the 47 patients, 41 completed at least one cycle of CC-223, had at least one restaging, and were thus evaluable for efficacy. By investigator assessment, no patient had a complete response (CR), 3 patients (7.3%) had a partial response (PR), 34 (82.9%) showed stable disease (SD), 1 (2.4%) had progressive disease (PD), and for 3 (7.3%) the assessment was not done. Thus, the objective response rate (CR+PR) was 7.3% (95% CI 1.5-19.9%) and the disease control rate (CR+PR+SD) was 90.2% (95% CI 76.9-97.3%). It is noteworthy that tumor shrinkage of any magnitude relative to baseline was observed in 73.2% patients (95% CI 57.1-85.8%) . Duration of response for the three patients with PR was 125, 253+ (no PD reported), and 401 days. The median duration of SD was long and lasted 297 days (min 50 days, max 1519 days). Duration of SD was longer than a year in 17 patients. Median PFS for the treated population was 19.5 months (95% CI 10.4-28.5 months) with rates at 6 and 10 months of 85% and 64%, respectively; median OS was not assessable (OS rate at 48 weeks was 0.97) .
On enrollment, carcinoid-related flushing and diarrhea were reported in 16 (34%) and 22 (47%) patients, respectively, with both reported in 13 patients (28%). Symptom reduction at any time during the study was observed for flushing in 8 (50%) patients, for diarrhea in 9 (41%) patients, and for both symptoms in 5 (39%) patients. An ad hoc exploratory analysis suggested that patients with carcinoid symptom improvement had a higher PFS than patients without improvement Among the 35 patients with elevated baseline levels in at least one hormone, a �50% decline from baseline was reported for CgA (7 patients), gastrin (6 patients), glucagon (3 patients), and serotonin (2 patients). Sample sizes were small and hormones fluctuated considerably over time; none of the changes from baseline were statistically significant except for CgA (cycle 21 and end of treatment; P�0.039 [log rank text] for each cycle) and serotonin in some later cycles (cycles [bib_ref] CC-223, a Potent and Selective Inhibitor of mTOR Kinase: In Vitro and..., Mortensen [/bib_ref] [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref] [bib_ref] Sequential tests of statistical hypotheses, Wald [/bib_ref] [bib_ref] Efficacy and Safety of Everolimus in Extrapancreatic Neuroendocrine Tumor: A Comprehensive Review..., Faggiano [/bib_ref] , and end of treatment; P�0.037 [log rank test] for each cycle). One responder had a >50% decline in CgA, and one responder had a >50% decline in gastrin. Too few patients with any elevated baseline levels had sufficient tumor response data for meaningful correlative analysis.
# Discussion
This study provides encouraging first results on response of patients with well-differentiated non-pancreatic GI NET to a TOR kinase inhibitor in combination with an SSA. The extended times on treatment and durable PFS were notable (median PFS of 19.5 months for CC-223), considering that previously published data for everolimus reported a median PFS of 11.0 months [bib_ref] Everolimus for the treatment of advanced, non-functional neuroendocrine tumours of the lung..., Yao [/bib_ref]. These findings support preclinical data indicating that dual inhibition by CC-223 may have advantages over everolimus, for example in everolimus-refractory patients [bib_ref] mTOR Kinase Inhibition Effectively Decreases Progression of a Subset of Neuroendocrine Tumors..., Orr-Asman [/bib_ref]. The frequency of carcinoid symptom reduction, which started early in patients who were refractory to an SSA alone, was also of interest. Since symptom monitoring using a non-validated questionnaire was added partway through the study in response to patients spontaneously reporting early-onset marked relief in both flushing and diarrhea, further evaluation of this observation is needed. The selection of an optimal dose of CC-223 at the conclusion of the dose-finding phase of the study was based on considerations of drug tolerability, magnitude and duration of mTORC1 and mTORC2 biomarker inhibition, and preliminary efficacy observed in a mixed tumor population [bib_ref] A phase I dose-escalation study to assess safety, tolerability, pharmacokinetics, and preliminary..., Bendell [/bib_ref]. The 45 mg CC-223 once daily was not well-tolerated in the NET cohort. Dose reductions were also reported for >50% of patients in tumor-specific cohorts of nonsmall cell lung cancer and hepatocellular carcinoma dosed at 45 mg [bib_ref] Phase I expansion trial of an oral TORC1/TORC2 inhibitor (CC-223) in advanced..., Varga [/bib_ref]. Most dose reductions occurred during the first 1-2 cycles of treatment. Additionally, while most adverse events occurred at lower rates at the 30 mg/day dose in patients with NET , the 30-mg cohort also required frequent dose adjustments. Longer treatment duration was achieved at significantly lower doses for both cohorts. This may be partly explained by the relatively less aggressive disease characteristics of NET and the reluctance of these highly functioning patients to tolerate the inconvenience of adverse effects associated with their cancer treatment. Further consideration is needed to determine whether it is better to start treatment at 30 mg/day (ie, close to the conventional mean dose administered [26 mg]) or to start at a much lower dose and titrate up to an optimally tolerable level in order to improve compliance and quality of life. To minimize toxicity and maintain compliance, a number of patients required dose reductions to as low as 15 mg per day, with 3 patients at this dose every other day. At the 15 mg dose level, target biomarker inhibition would be predicted (not tested) to be modest, yet these patients benefitted from durable tumor shrinkage for extended periods of time. These findings suggest it may be appropriate to re-evaluate the traditional practice in dose-finding studies of dosing close to MTD for cohort expansion, and instead consider using the dose achieving maximal biomarker inhibition. This is particularly relevant for oral, daily administered drugs where cumulative toxicities often occur beyond the "classical" four-week dose-limiting toxicity period.
The PET SUV findings approximately 2 weeks into therapy showed a statistically significant reduction in glucose uptake from baseline. However, among patients who achieved at least some metabolic response (at least a 15% reduction from baseline) there were no strong correlations between early metabolic improvements and subsequent tumor response assessed by RECIST . This outcome may have been confounded by CC-223-related hyperglycemia. Furthermore, it was anticipated that well-differentiated NET would have generally low FDG-PET uptake because of its low metabolic rate, as was confirmed in this cohort. Reductions in CgA and other abnormal serum hormones were generally not statistically significant, but patients with carcinoid symptom improvement showed a numerically higher PFS probability than patients without symptom improvement (median PFS 108 weeks vs 48 weeks). Correlations between carcinoid symptom improvement and reduction in baseline serum hormone levels were inconclusive, possibly because of the small sample sizes and the multiplicity of analyses (S1 The safety profile of CC-223 appears to be comparable with the previously characterized class of approved mTOR inhibitors [bib_ref] Everolimus for the treatment of advanced, non-functional neuroendocrine tumours of the lung..., Yao [/bib_ref] [bib_ref] Efficacy and Safety of Everolimus in Extrapancreatic Neuroendocrine Tumor: A Comprehensive Review..., Faggiano [/bib_ref] [bib_ref] Strategies for the management of adverse events associated with mTOR inhibitors, Kaplan [/bib_ref] , notwithstanding the increased incidence and severity of certain TRAEs. However, toxicity generally emerged early into treatment and was well-managed either by aggressive dose adjustments or standard treatment interventions. The commonly experienced on-target class effect of hyperglycemia correlated with drug exposure and was managed effectively with dose-reduction or with hypoglycemic agents [bib_ref] Hyperglycemia Associated With Targeted Oncologic Treatment: Mechanisms and Management, Goldman [/bib_ref]. No lateonset new toxicity was identified even after extended treatment with CC-223. Overall, neither unexpected nor new major safety signals were identified for CC-223 compared with currently available mTOR inhibitors.
In conclusion, treatment with CC-223 showed encouraging preliminary evidence of antitumor activity in patients with advanced metastatic non-pancreatic GI NET. The durable tumor responses, coupled with meaningful symptom relief in patients reporting carcinoid symptoms, support further evaluation of CC-223 in this patient population. Dual mTORC1 and mTORC2 kinase inhibitors may have advantages over everolimus, for example in everolimus-refractory patients, as has been shown pre-clinically [bib_ref] mTOR Kinase Inhibition Effectively Decreases Progression of a Subset of Neuroendocrine Tumors..., Orr-Asman [/bib_ref]. In addition, CC-223 efficacy may be enhanced by combination with other cancer therapies, which will be worthy to explore.
## Supporting information
[fig] Fig 1: CONSORT diagram. [/fig]
[fig] Fig 2 g002 Fig 3: Waterfall plot of tumor response according to RECIST version 1.1 in patients treated with CC-223, by starting dose. Best percent change from baseline in the sum of the longest diameters of target lesions (n = 39). PR, partial response; RECIST, Response Evaluation Criteria In Solid Tumors; SD, stable disease. https://doi.org/10.1371/journal.pone.0221994.Progression-free survival (A) and overall survival (B) for the treated population. $, censored; CI, confidence interval; NA, not assessable; OS, overall survival; PFS, progression-free survival. https://doi.org/10.1371/journal.pone.0221994.g003 mTORC1/C2 kinase inhibitor (CC-223) for non-pancreatic NET PLOS ONE | https://doi.org/10.1371/journal.pone.0221994 September 17, 2019 10 / 14 [/fig]
[fig] S1: File. CONSORT checklist. (PDF) S2 File. Trial protocol. (PDF) S3 File. Supporting Information. (DOCX) Fig. CC-223 dose adjustments and duration, best overall RECIST version 1.1 and target lesion response, and carcinoid symptom improvement, for patients started at 45 mg/day (A) and 30 mg/day (B) CC-223. Cycle, total treatment cycles completed; RECIST, best overall response; TL change (%), best target lesion change from baseline. $, carcinoid symptomatic improvement; NE, not evaluable; PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria In Solid Tumors; SD, stable disease. (TIF) [/fig]
[table] Table 1: Patient demographics and baseline characteristics (n = 47). [/table]
[table] Table 3: Pharmacokinetic characteristics for CC-223 and principal metabolite (M1) in patients with NET initially treated with 30 mg or 45 mg CC-223. [/table]
[table] Table 4: Regulation of pAKT and p4E-BP1 in monocytes. Inhibition (percent change from baseline) for pAKT and p4E-BP1 in monocytes of patients initially dosed with 45 mg or 30 mg CC-223. [/table]
|
Probiotics as Beneficial Dietary Supplements to Prevent and Treat Cardiovascular Diseases: Uncovering Their Impact on Oxidative Stress
The gut microbiota, the ecosystem formed by a wide symbiotic community of nonpathogenic microorganisms that are present in the distal part of the human gut, plays a prominent role in the normal physiology of the organism. The gut microbiota's imbalance, gut dysbiosis, is directly related to the origin of various processes of acute or chronic dysfunction in the host. Therefore, the ability to intervene in the gut microbiota is now emerging as a possible tactic for therapeutic intervention in various diseases. From this perspective, evidence is growing that a functional dietary intervention with probiotics, which maintain or restore beneficial bacteria of the digestive tract, represents a promising therapeutic strategy for interventions in cardiovascular diseases and also reduces the risk of their occurrence. In the present work, we review the importance of maintaining the balance of the intestinal microbiota to prevent or combat such processes as arterial hypertension or endothelial dysfunction, which underlie many cardiovascular disorders. We also review how the consumption of probiotics can improve autonomic control of cardiovascular function and provide beneficial effects in patients with heart failure. Among the known effects of probiotics is their ability to decrease the generation of reactive oxygen species and, therefore, reduce oxidative stress. Therefore, in this review, we specifically focus on this antioxidant capacity and its relationship with the beneficial cardiovascular effects described for probiotics.
# Introduction
Literature searches on "gut microbiota" performed by the guest editors via the PubMed platform revealed extensive publications, consisting of approximately 880 reviews in 2018 (~2/day). Approximately 380 (~1/day) reviews on "probiotics" were published in 2018. Restricting the search with filters to published reviews using the combination of terms "probiotics and cardiovascular diseases" reduced the number to. Interestingly, when the term "oxidative stress" was included in the combination "probiotics and cardiovascular diseases (CVD) and oxidative stress," no published reviews were identified. It is known that CVD remains the leading cause of death and disability in developed countries. These two scenarios associated with exciting original articles on probiotics and CVD in the prior 3 years motivated us to promote the publication of this special issue.
The gut microbiota is a broad symbiotic community of nonpathogenic microorganisms composed primarily of anaerobic bacteria (although some gut bacteria preferentially grow under microaerophilic conditions) and fungi. One of its functions is the maintenance of a barrier via enterocytes covered with a brush border of mucus, which is produced by goblet cells and nonpenetrable tight junctions between enterocytes, a layer of luminal mucus and tight adherens junctions between enterocytes that allows the control of absorption and metabolism and the maturation and stimulation of the immune system, which are essential functions for an effective mechanism of defense against pathogens in the host. Signals generated in the gut microbiota communicate with distant organs by crossing the intestinal epithelium and triggering diverse signaling processes located at the epithelial cell border, subsequently reaching the systemic circulatory system. Different pathways are responsible for the bidirectional interaction between the gut microbiota and systemic organs in healthy individuals .
Currently, there is a growing body of evidence that an abnormal predominance of pathogenic over commensal (nonpathogenic) microorganisms, a condition termed gut dysbiosis, can initiate or worsen the dysfunction of diverse target systemic organs. Gut dysbiosis is also a confirmed cause of increased oxidative stress in the body. In fact, frequent consumption of fats and refined sugars in the Western-type diet produces an increase in reactive oxygen species (ROS) production and inflammatory processes. Additionally, the gut microbiota regulates the production of mitochondrial ROS. In recent years, various studies have revealed that gut dysbiosis may contribute to the development and progression of CVD and other related diseases.
On the other hand, several studies have demonstrated that probiotics may be beneficial in reestablishing the microbiota through different mechanisms such as appropriate intestinal homeostasis. The name probiotic is applied to those live microorganisms which, when administered in adequate amounts, confer a health benefit on the host. This definition, established by the joint Food and Agriculture Organization of the United Nations (FAO) and World Health Organization (WHO) Working Group, is the one used today and accepted by the International Scientific Association for Probiotics and Prebiotics (ISAPP).
Therefore, the purpose of this review is to update our knowledge concerning the contribution of the exacerbated production of oxidative stress to the development of cardiac and vascular dysfunctions in the clinic and in experimental models of arterial hypertension, as well as the possible beneficial effects of dietary supplementation with probiotics, in an attempt to prevent or reverse these cardiovascular disturbances.
## Probiotics as promising coadjuvants for prevention/treatment of arterial hypertension
Arterial hypertension constitutes a main risk factor for the development of severe pathologies, such as acute myocardial infarction, heart failure, stroke, and renal failure, as well as for premature death worldwide. Primary (or essential) hypertension is a multifactorial process that involves genetics, demographics, comorbid disorders, and environmental influences. Approximately 8% of cases exhibit secondary hypertension, which has a known origin, including endocrine diseases, drugs, cancer, or hyperactivation of the renin-angiotensin system, among others. Antihypertensive therapy used in clinical practice has been shown to be effective in maintaining blood pressure (BP) at safer levels, thereby reducing the morbidity and mortality associated with this disease. Several international reports, mainly those prepared by the Eighth Joint National Committee (JNC 8), have established the guidelines for the treatment of hypertensive patients and different reference values for those over 60 years old.
In addition to pharmacological treatments, it is necessary to establish a series of nonpharmacological measures for the control of the disease. In this sense, the relevance of a good diet is paramount, and the contribution of probiotics can be fundamental. It has been demonstrated that the gut microbiota participates in an important manner in the control of BP by several mechanisms, such as exerting control at the level of the central and autonomic nervous system or protecting endothelial function (see illustration in . Additionally, gut dysbiosis has been described in animal models of hypertension and hypertensive patients.
In agreement with the above discussion and based on the state-of-the-art role of the gut microbiota and its interaction with different organs, a working group at the National Heart, Lung, and Blood Institute recently discussed the current status and future directions for the treatment and prevention of high BP, considering the use of probiotics. Therefore, functional foods that contribute to the maintenance of the intestinal flora can be very useful when avoiding excessively high BP levels, as previously reviewed in detail by several authors (see.
Among the foods that have been shown to provide various cardiovascular benefits, kefir has been reported to effectively lower BP. Chronic consumption of this synbiotic attenuates the abnormal increase in BP in spontaneously hypertensive rats (SHR), which has been the most commonly used genetically hypertensive animal for a better understanding of several cardiovascular abnormalities. Kefir has been tested in the protection of vascular endothelial dysfunctionand in the correction of impaired autonomic cardiovascular function, including its inhibitory effects on angiotensin-converting enzyme (ACE). Therefore, there is growing evidence that probiotics could be a promising natural coadjuvant in the prevention/treatment of CVD, including the hypertensive process. In another study using SHR, animals fed with Minas Frescal probiotic cheese showed a significantly lower BP compared with the control group, in addition to an improvement in other indicators of cardiovascular health, such as blood levels of triglycerides and cholesterol.
Several clinical studies in humans have also demonstrated the ability of probiotics to reduce abnormally high BP levels. For example, an extract of Lactobacillus casei, which has been shown to reduce BP in SHR, was able to induce a reduction in systolic/diastolic BP and heart rate in hypertensive patients. In 2002, an interesting study showed that food supplementation with Lactobacillus plantarum produced a significant decrease in systolic BP in heavy smokers. A Norwegian study showed, in 2011, that the incidence of preeclampsia, which is associated with hypertension and inflammation, is decreased by chronic intake of probiotics. Additionally, in a randomized double-blind clinical trial with type II diabetes mellitus, probiotic soy milk containing Lactobacillus plantarum significantly decreased systolic/diastolic BP, and in a study with prediabetic patients, there was a significant tendency to reduce hypertension in those patients receiving a multispecies probiotic. In 2014, a meta-analysis carried out based on the results of nine clinical trials found that consumption of probiotics slightly reduced BP and that this effect was more marked if the basal BP was elevated. The authors also concluded that several species of probiotics used together provided enhanced effects. Finally, the duration of the intervention must be ≥8 weeks, and the dose of daily consumption of probiotics should be ≥10 11 colony-forming units.
In contrast, several studies have questioned the role of some probiotics in producing low BP. A treatment for 4 weeks by dietary supplementation of Lactobacillus plantarum either together with fermented blueberry or with three synthesized phenolic compounds did not lower BP in NG-nitro-L-arginine methyl ester-(L-NAME-) induced hypertensive rats. In a clinical trial, probiotic strains of Lactobacillus acidophilus and Bifidobacterium animalis, provided in either yogurt or capsule form, did not improve cardiovascular risk factors since they did not modify BP or concentrations of total cholesterol LDL-C, HDL-C, or triglycerides in overweight or obese individuals. Additionally, a study of postmenopausal women with metabolic syndrome showed that administration of milk supplemented with Lactobacillus plantarum produced several beneficial effects, but it did not provide a significant decrease in BP. Furthermore, long-term treatment with Lactobacillus helveticus-fermented milk containing bioactive peptides reduced arterial stiffness in hypertensive subjects but did not induce statistically significant differences between the effects of the probiotic and placebo treatment on BP.
In view of these reports, it will still be necessary to carry out more studies to verify the possible role of probiotic foods : Summary of the interaction between the gut microbiota and neuroimmune and neuroendocrine systems and the interaction between the gut microbiota and microbiota target organs of the host. The gut microbiota provides (i) a mucosal barrier through tight and adherens junctions between enterocytes, (ii) immunomodulation and anti-inflammation through recruitment of immune cells, and (iii) energy metabolism via metabolites/short-chain fatty acids (SCFAs), vitamins, and hormones. The brain-intestine axis acts through both an integrative autonomic nervous system, including the sympathetic/parasympathetic (vagal) afferent/efferent nerve pathways, associated with the neural myenteric network, and a neuroendocrine system, including the hypothalamus-adrenocortical gland system. as coadjuvants in the treatment of arterial hypertension. In any case, the results of the different studies suggest that the complicated mechanisms of the development of hypertension, the choice of different bacterial strains, the different types of patients, and the previous state of their microbiome can be decisive in terms of obtaining satisfactory results for the reduction of BP.
## Endothelial dysfunction: the role played by oxidative stress
The vascular endothelium is a single layer of smooth, thin cells that constitutes the first barrier between the bloodstream and the vascular muscle. Among its functions is to act as a selective membrane through which fluid and solutes, as well as trafficking of inflammatory cells, interchange between the plasma and tissue spaces. The endothelium also contributes to the regulation of vascular tone by synthesizing and releasing a huge number of vasodilating substances, both vasodilators such as nitric oxide (NO), prostacyclin, and endothelium-derived hyperpolarizing factor (EDHF) and vasoconstrictors such as endothelin (via ET A ), angiotensin II (via AT 1 receptors), and ROS. In addition, its action is the key in the control of platelet aggregation and blood hemostasis, regulating the antithrombotic/prothrombotic balance, and it also participates in the inflammatory and immune response (for a detailed review, see.
Due to this multifunctional role of the endothelium, it is easy to understand that its alterations may lie at the origin and/or in the development of various diseases. Therefore, endothelial dysfunction is recognized as a risk factor for the onset of CVD and appears in the early stages and during the development of hypertension, cardiac ischemia, atherosclerosis, stroke, or peripheral vascular disease. Other diseases such as diabetes, kidney failure, infectious diseases, and tumor progression also have a component of endothelial dysfunction.
Endothelial dysfunction can be caused by inflammatory processes, leading to a decrease in endothelial NO synthase (eNOS) enzyme activity, thereby decreasing the NO bioavailability and culminating in hypertension. Moreover, oxidative stress also contributes to the development of endothelial dysfunction, reducing the availability of NO. In fact, the generation of ROS caused by hypertension, hypercholesterolemia, diabetes, or other cardiovascular risk factors causes a decrease in the release of endothelial NO.
As mentioned above, there is an important relationship between dysbiosis and the development of hypertension (see also, which could involve the impaired endothelial function due to alterations of the gut microbiota during the chronic rise in BP. In fact, fecal microbiota transplantation from SHR to normotensive WKY rats caused a chronic impairment of endothelial function, accompanied by greater vascular oxidative stress and increased systolic BP. In contrast, transplantation of fecal microbiota from WKY to SHR provoked the opposite effects with an improvement of endothelial dysfunction in hypertensive animals. Accordingly, several studies have suggested that probiotics could lead to an improvement in endothelial function. Rashid et al.reported that the endothelial dysfunction of mesenteric artery rings in rats with common bile duct ligation is mediated in part by oxidative stress, possibly due to bacterial translocation inducing a proinflammatory response, and that this effect is improved by the ingestion of a probiotic formulation.
Endothelial dysfunction can be identified physiologically by means of NO-dependent mechanisms. In this situation, blood vessels show a reduced vasodilator response to agents that contribute to the release of NO, such as acetylcholine and, conversely, an exacerbated response to vasoconstrictor agents, such as α 1 -adrenergic agonists or thromboxane A2 analogues. Using this method, chronic probiotic treatment with Lactobacillus coryniformis reversed the endothelial dysfunction observed in obese mice and improved the endothelial dysfunction and vascular oxidative stress induced by lipopolysaccharides (LPS) in control mice.
In a similar way, our group evaluated the effects of the probiotic kefir on endothelial dysfunction in SHR. Our results suggested that kefir treatment for eight weeks (even at a low dose) could attenuate endothelial dysfunction in the large vessels in hypertensive rats, and the main mechanism for this beneficial effect was exerted through a repair of the vascular endothelial architectureand a reduction of the oxidative stress, together with an increase in NO bioavailability as well as endothelial progenitor cell recruitment. These beneficial effects of kefir on vascular endothelial function have recently been reviewed.
This effect of probiotics was also confirmed in a study in which lactic acid bacteria partially reversed the relaxation deficit of the aorta in SHR. In addition, it also increased the NO level, which is abnormally decreased in SHR serum. Both effects are indicative of a probiotic-induced improvement in endothelial function due to a reduction of vascular oxidative and inflammatory status.
In addition, using SHR and WKY rats for comparison, a study by Gomez-Guzman et al.demonstrated that chronic oral administration of the probiotic Lactobacillus fermentum or Lactobacillus coryniformis plus Lactobacillus gasseri restored gut eubiosis and improved endothelial dysfunction as a result of a reduced vascular proinflammatory and prooxidative status. Some studies in humans or human cells have also shown an improvement in endothelial function due to probiotic treatment. In endothelial cells, soy milk fermented with Lactobacillus plantarum or Streptococcus thermophilus stimulated NO production and eNOS activity, suggesting their effectiveness for the improvement of endothelial function. A 6-week supplementation with Lactobacillus plantarum in men with stable coronary artery disease improved endothelial function for both conduit and resistance vessels through increasing NO bioavailability while concomitantly reducing systemic inflammation, as measured by brachial artery flow-mediated dilation. These results suggest that the intestinal microbiota is mechanistically linked to systemic inflammation and vascular endothelial function. Another clinical trial showed that a multispecies probiotic supplement improved both functional and biochemical parameters of endothelial dysfunction, including systolic BP, vascular endothelial growth factor, pulse wave velocity (PWV) and its augmentation index, interleukin-6, tumor necrosis factor alpha (TNFα), and thrombomodulin in obese postmenopausal women. In contrast, in a study of subjects with metabolic syndrome receiving supplementation with the probiotic strain Lactobacillus casei Shirota, no significant changes in parameters used to assess low-grade inflammation or endothelial dysfunction were observed.
In general, studies both in vivo and in vitro, as well as clinical studies in humans, suggest that supplementation with several types of probiotics contributes to an improvement in endothelial function through various mechanisms. Although further research is needed, the role of probiotic supplementation in the prevention of CVD by correcting endothelial dysfunction is promising. In addition, the multifunctional role of the endothelium extends this potential use of probiotics to all diseases, not only cardiovascular, in which its pathophysiology may be related to endothelial dysfunction.
## Evidence of the beneficial effects of probiotics on the autonomic control of cardiovascular function
Prebiotics, probiotics, and synbiotics are some of the best evidenced ways of manipulating the microbiota, and their potential role in the prevention and treatment of multiple diseases has recently garnered a significant interest. Recent data from experimental time-course studies have shown that long-term treatment with kefir (at least 30 to 60 days), in addition to the antihypertensive effect, attenuated cardiac hypertrophy in SHR. Considering the relationship between the gut microbiota and the target systemic organs, it is important to highlight studies that relate the influence of these microorganisms to cardiovascular function. Those findings led the authors to investigate whether the benefits of kefir supplementation could also include the autonomic neural control of BP (baroreflex function) and the cardiac pacemakers controlling the chronotropic rhythm under the neural efferent pathways from the brainstem integrative areas. They observed that administration of kefir (for at least 60 days) attenuated and partially reversed the abnormal cardiac sympathetic predominance over the parasympathetic tone in SHR, raising the following question: "by which mechanisms can probiotics and synbiotics affect brain areas?" As illustrated in, there is a consistent and wellrecognized neuroendocrine gut-brain axis connection, which includes the hypothalamus-pituitary-suprarenal gland axis and the autonomic sympathetic/parasympathetic afferent/ efferent pathways. Others have attempted to address the question by proposing relevant interactions between gut endocrine cells and vagal afferents through gut chemosensing mechanisms.
Other exciting findings in recent years has been the demonstration of a marked association between the effects of probiotics and decreased production of intravascular ROS and augmented NO bioavailability. It seems that the mechanisms underlying the beneficial actions of probiotics on cardiac autonomic control could occur through its capability to decrease the production of cytokines and ROS in the hypothalamic paraventricular nucleus. In turn, it could attenuate hypertension and end-organ damage by upregulating antiinflammatory and antioxidant molecules, therefore restoring the normal balance between parasympathetic and sympathetic activity to the heart, as recently observed by our group. In a similar way and as expected, the investigators observed that SHR treated with probiotics presented a partial recovery of the baroreceptor sensitivity, which is characterized in this experimental model by a high variability of the resting BP. The SHR exhibited diminished reflex tachycardia or bradycardia to induce hypotensive or hypertensive changes in the resting BP. Probiotic supplementation was able to partially repair this BP variability and baroreflex sensitivity, and this could occur because kefir repairs the normal gut microbiota and, consequently, restores the production of neuroactive compounds in the intestinal lumen. Therefore, the above findings corroborate the knowledge that probiotics have a modulatory action on the integrative central or peripheral components of the gut-brain axis.
Recently, Brasil et al.assessed whether the soluble nonbacterial fraction of kefir (bioactive compounds) and not the probiotic effects would improve cardiovascular hemodynamics, enhancing the baroreflex sensitivity, which could include the ACE inhibitory properties. Therefore, an important mechanism by which probiotics decrease high BP and repair endothelial dysfunction and cardiac autonomic tones could be achieved through probiotic bioactive compounds. In addition, it has been observed that probiotic supplementation caused a decrease in ACE activity measured in the serum of SHR treated with the soluble nonbacterial fraction of kefir, supporting ACE inhibition as a likely mechanism for kefir's beneficial cardiovascular effects during hypertension. These effects indicate that the improvement in baroreflex gain cannot be attributed to the probiotic effect of kefir but rather to other bioactive compounds produced by microbial action. Clearly, how these different fractions (e.g., probiotic bacteria or bioactive compounds) influence the baroreflex and other cardiovascular risk factors are still poorly understood. To our knowledge, very few publications have evaluated the effects of probiotics on baroreflex function and autonomic control of heart rate. These previous studies used the fermented food kefirand its bioactive compounds.
In conclusion, the benefits of probiotics in the cardiovascular system in models of hypertension include the reversion of cardiac dysautonomia, which is characterized in hypertensive subjects by an inverted predominance of sympathetic over vagal tone, including a significant attenuation of the high variability of BP and heart rate and their effectiveness to partially revert the decreasedbaroreflex sensitivity. Nonetheless, probiotics attenuate disturbances in the neural control of cardiovascular function in a similar manner to that achieved with physical exercise, therapy with flavonoids, and pharmacological medication. Therefore, there is clearly a need for more mechanistic studies that would help to identify the missing links to explain the protective effects of fermented foods, such as pre-, pro-, and synbiotics, as well as their bioactive compounds on the neural control of BP. summarizes the possible sites of action of probiotics.
## Heart failure: a target for the benefits of functional diets
Heart failure (HF) patients experience some changes in the gut microbiome during disease. Some reports have described increased levels of pathogenic microbes that could have potential deleterious effects on cardiac function. This phenomenon might be explained by the so-called "gut hypothesis," in which the reduced cardiac pumping function and congestion observed in HF patients would be responsible for an intestinal ischemia, favoring bacterial translocation and increases in circulating endotoxins that elicit inflammation. In fact, the intestinal blood flow is reduced in HF patients, contributing to juxtamucosal bacterial growth. Kummen et al.reported that the gut microbiota in HF patients is related to persistent Tcell activation. In fact, the removal of Gram-negative intestinal bacteria by antibiotics reduces the monocyte CD14 expression, along with reduced levels of endotoxins and cytokines, with improved flow-mediated dilation in patients with severe HF. The association between gut dysbiosis and CVD was highlighted when 60 stable HF patients were selected to test whether the characteristics of the gut microbiota would correlate with their cardiovascular functional status. The authors evidenced that HF patients had more colonies of pathogenic bacteria than control participants, along with an increased intestinal permeability that favored bacterial translocation. In addition, severe HF was associated with more pathogenic types of bacteria than mild HF. The contribution of the gut microbiota to the pathogenesis of CVD has been supported by the discovery that some dietary products that are metabolized by gut microbes produce toxic metabolites that could have negative impact on the cardiovascular system. Changes in the gut microbiota can lead to increases in trimethylamine N-oxide (TMAO), which is a major contributor to cardiovascular and renal diseases. TMAO is an endotoxin that is produced via the metabolism of trimethylamine from the carnitine molecule, which is absorbed into the blood and converted into TMAO in the liver by flavincontaining monooxygenases. In fact, intestinal microbes participate in the phosphatidylcholine metabolism and in the increased TMAO levels, which was independently associated with major cardiovascular eventsand the incidence of chronic kidney diseases. The findings also showed that TMAO could contribute to the risk prediction scores of deaths in acute HF patients, revealing a poor 1-year prognosis.
Experimental studies have demonstrated potential therapeutic actions of probiotics in different animal models of HF. Using a rat model of acute myocardial infarction by permanent coronary occlusion, Gan et al.showed improved ventricular function and structure after treatment with the probiotic Lactobacillus rhamnosus. The possible mechanism by which probiotics act in the infarcted heart has been described by Lam et al., in which the probiotic Lactobacillus plantarum decreased the leptin levels and, thus, reduced the infarct size in rats. Additionally, the antiapoptotic effect of probiotic-fermented purple sweet potato yogurt was evidenced in a rat model of hypertensive HF. Despite the increasing number of experimental studies on probiotics in HF models, only one study has addressed the prognostic effects of probiotics in HF patients. In this pilot trial, patients with HF class II or III and LVEF <50% were randomized to Saccharomyces boulardii or placebo for 3 months in a double-blinded fashion. Patients treated with the probiotic showed a significant reduction in the left atrial diameter, uric acid, CRP, and creatinine levels. An important find of this study was that the treatment was safe and well tolerated, without reports of side effects or adverse events.
# Conclusions
Throughout this review, we have presented evidence in the literature indicating that a habitual consumption of probiotics, which restore the balance of the intestinal microbiota, could present cardiovascular benefits based, at least in part, on its ability to reduce oxidative stress. Although many points remain to be clarified and many of the published results are contradictory, it is evident that consumption of probiotics constitutes a promising complement to more conventional cardiovascular therapies, as well as to nonpharmacological measures that are commonly used to counteract the onset and progression of CVD. Further studies are needed to clarify the interaction between the gut microbiota and the neuroimmune system, as well as the endocrine system, to create nutrigenetic profiles that may aid in achieving individual homeostasis. It will also be necessary to improve knowledge concerning the different bacterial strains present in probiotics and how they should be consumed to take full advantage of their potential beneficial effects for each specific situation. Finally, studies of the great variety of enzymes, peptides, and biochemical pathways generated by the intestinal microbiota, which differ from the resources of the host, could constitute an innovative strategy for the design of new drugs for the treatment of CVD. : Diagram illustrating the pathways of neural control of cardiac and vascular function through the autonomic neural system and possible sites of action of the probiotics (Pb) based on recent evidence. Mechanosensitive baroreceptors are located in the carotid sinus and aortic arch, and they discharge at BP systole-by-systole bursts of action potentials to the brainstem. The autonomic vagal and sympathetic efferents innervate the cardiac pacemakers and the myocardium (sympathetic ends). The resistance and conductance vessels are innervated by sympathetic efferent components. Modified from Vasquez et al., with permission. |
Contribution of BCR-ABL molecular variants and leukemic stem cells in response and resistance to tyrosine kinase inhibitors: a review
Chronic myeloid leukemia (CML) is a myeloproliferative neoplasm generated by reciprocal chromosomal translocation, t (9; 22) (q34; q11) in the transformed hematopoietic stem cell. Tyrosine kinase inhibitors (TKIs) target the mature proliferating BCR-ABL cells, the major CML driver, and increase overall and disease-free survival. However, mutant clones, pre-existing or due to therapy, develop resistance against TKIs. BCR-ABL1 oncoprotein activates various molecular pathways including the RAS/RAF/MEK/ERK pathway, JAK2/STAT pathway, and PI3K/AKT/mTOR pathway. Stimulation of these pathways in TKI resistant CML patients, make them a new target. Moreover, a small proportion of CML cells, leukemic stem cells (LSCs), persist during the TKI therapy and sustain the disease in the patient. Engraftment of LSCs in the bone marrow niche and dysregulation of miRNA participate greatly in the TKI resistance. Current efforts are needed for determining the reason behind TKI resistance, identification, and elimination of CML LSC might be of great need for cancer cure.
# Introduction
Leukemias are group of life-threatening blood and bone marrow disorders that affect both children and adults of a major four groups, acute lymphocytic leukemia (ALL), acute myeloid (AML), chronic lymphocytic leukemia (CLL), and CML. [bib_ref] Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia..., Yeoh [/bib_ref] The BCR-ABL gene results from a reciprocal translocation t (9; 22) (q34; q11) in the transformed hematopoietic stem cell (HSC). 1-3 BCR-ABL is considered the most common genetic abnormality in chronic myeloid leukemia (CML) (95%), followed by acute lymphocytic leukemia (ALL) (35%), and rarely presented in acute myelogenous leukemia (AML) (1%). [bib_ref] Impact of cytogenetics on the outcome of adult acute lymphoblastic leukemia: results..., Pullarkat [/bib_ref] [bib_ref] Risk-and response-based classification of childhood B-precursor acute lymphoblastic leukemia: a combined analysis..., Schultz [/bib_ref] [bib_ref] The P190, P210, and P230 forms of the BCR/ABL oncogene induce a..., Li [/bib_ref] The most detected variant of BCR-ABL in CML is P210 BCR-ABL1 that occasionally presents in ALL and AML. The P210 BCR-ABL1 variant expresses in about 48% of Ph+ CML cases, while 52% co-express P210 BCR-ABL1 and P190 BCR-ABL1 variants. [bib_ref] Chronic myeloid leukemia patients with the e13a2 BCR-ABL fusion transcript have inferior..., Lucas [/bib_ref] Approximately 50% of CML patients present without symptoms and are diagnosed incidentally after routine laboratory tests. [bib_ref] The biology of chronic myeloid leukemia, Faderl [/bib_ref] Cytogenetics is the standard diagnostic tool of CML; the Ph chromosome is detected in 90% of the cases. However, 5% of CML cases have a cryptic Ph chromosome that could not be detected by karyotyping. The cryptic Ph chromosome can be detected by fluorescence in situ hybridization (FISH) and/or real-time polymerase chain reaction (PCR). [bib_ref] Translocation of BCR to chromosome 9: a new cytogenetic variant detected by..., Hagemeijer [/bib_ref] Clinically CML is divided into three different stages: chronic phase (CP), accelerated phase (AP), and blast phase (BC). Without intervention treatment, the disease progresses from the chronic phase to blast phase which is most likely accompanied by multiple genomic abnormalities. Here we present a review of the BCR-ABL response and resistance to TKI therapy.
## Bcr-abl gene & protein
BCR-ABL results in the juxtaposition of 3 0 sequences of ABL (Abelson proto-oncogene) on chromosome 9 and the 5 0 sequences of the truncated BCR (breakpoint cluster region) on chromosome 22. Consequently, the BCR/ABL fusion gene forms on chromosome 22 that encodes BCR-ABL1 oncoprotein, a constitutively active tyrosine kinase. [bib_ref] Chromosome abnormalities in CML, Hagemeijer [/bib_ref] [bib_ref] Letter: A new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by..., Rowley [/bib_ref] [bib_ref] The molecular genetics of Philadelphia chromosome-positive leukemias, Kurzrock [/bib_ref] The constitutive tyrosine kinase activity of the chimeric BCR-ABL1 oncoprotein, deregulates the downstream signaling of many pathways which results in uncontrolled proliferation, arresting the differentiation of the HSC, and limits apoptosis. Consequently, the normal HSC transforms into leukemic stem cells (LSC), which accumulates in the bone marrow (BM) and thus replaces the normal stem cells. [bib_ref] Primitive, quiescent, Philadelphia-positive stem cells from patients with chronic myeloid leukemia are..., Graham [/bib_ref] [bib_ref] Low or undetectable numbers of Philadelphia chromosome-positive leukemic stem cells (Ph(+)CD34(+)CD38(neg)) in..., Mustjoki [/bib_ref] Three different BCR-ABL chimeric proteins emerge from the fusion of mRNA molecules of different lengths of the ABL gene with the BCR gene. The most common breakpoint in BCR occurs in intron 13 or intron 14, then exon 13, or 14 exons (M-BCR) fused to the ABL1 gene at exon a2, which is referred to as e13a2, e14a2. These fusions result in a BCR-ABL1 protein with 210 kilodaltons molecular mass (p210 BCR-ABL1), present in 95% of CML. The second alternative (<1%) is that the e19a2 fusion transcript produces a larger fusion protein with 230 kilodalton weight (p230 BCR-ABL1), a diagnostic marker for neutrophilic-chronic myeloid leukemia. The third alternative is p190 BCR-ABL1 protein resulting from e1a2 fusion transcript, most common in B cell ALL, less so in AML, and rarely presented in CML. BCR-ABL oncoprotein possesses different domains from BCR and ABL. The BCR portion includes the NH2-terminal coil-coil (CC) domain (amino acid 1-63), an oligomerization domain, followed by a serine/threonine kinase domaincontaining Tyr 177, a binding site for growth factor receptor-binding protein 2 (GRB2), and Ras homolog gene/guanine nucleotide exchange factor (Rho/GEF) kinase domain, a GTPase activating protein that gets activated when bind to GTP and turned off when bind to GDP, and thus control the activation of Ras protein. Whereas ABL domains include Srchomology (SH) SH3, SH2, and SH1 (Kinase) domains, followed by proline-rich domain and binding domain (BD), and facilitate the nuclear protein, actin, and DNA binding. CC domain and Y177 are required for the efficient activation of ABL kinase. Tetramerization of BCR-ABL is essential for constitutive kinase activity. Targeting the CC domain deregulates the tetramerization of BCR-ABL, reduces the kinase activity, and increases sensitivity to imatinib (tyrosine kinase inhibitor). [bib_ref] The NH(2)-terminal coiled-coil domain and tyrosine 177 play important roles in induction..., Zhang [/bib_ref] [bib_ref] Targeting of the N-terminal coiled coil oligomerization interface by a helix-2 peptide..., Beissert [/bib_ref] In BCR-ABL, Y177 plays a crucial role in leukemic progenitor cell expansion, proliferation, and survival through activation of the Ras and PI3K/Akt pathway. Interestingly, mutated 177Y fail to develop CML and increase sensitivity to REVISED Amendments from Version 1 I have added a paragraph about leukaemia in general and its different categories and removed figure one from the article. I have replaced figure 2 with a table and did revision for English.
Any further responses from the reviewers can be found at the end of the article imatinib, suggesting a target of Y177 to eliminate the leukemic stem cells. [bib_ref] BCR-tyrosine 177 plays an essential role in Ras and Akt activation and..., Chu [/bib_ref] Rho/GEF retained in p230 and p210 BCR-ABL but not in p190, thought to facilitate calcium-dependent lipid binding, restrict proliferation and support survival. [bib_ref] Contributions of the RhoGEF activity of p210 BCR/ABL to disease progression, Tala [/bib_ref] ABL protein has a tyrosine kinase function that is regulated by SH2 and SH3 domains in the N-terminal end. SH3 is considered as a negative regulator of the tyrosine kinase activity, therefore inhibition of SH3 leads to an abnormal increase in the enzyme activity and enhances the malignant transformation. The SH1 domain of ABL has an autophosphorylation site that activates kinase as conjugates to ATP. However, ATPbinding inhibitors, compete with ATP on its binding site and impeding the downstream intracellular signaling pathway. 20,21 SH2, the protein-protein interaction domain, activates the SH1 by establishing a tight junction with the N-terminal lobe of SH1, resulting in downstream activation of the intracellular signaling pathways. Interestingly, the open conformation of SH1 allows interacting with the SH2 domain, whereas the close conformation prevents this binding. The SH3 of ABL binds to SH1 through N-terminal myristoyl modification, which induces conformational changes on SH1 that allow SH2-SH3 docking onto it. 22
In a normal physiological state, tyrosine kinase (TK) regulates the activities of the cell cycle. Constitutive activation of TK interrupts the normal state of the cell cycle, resulting in inhibition of cell differentiation, uninterrupted activation of the cell proliferation, and escapes cell apoptosis. Thus, BCR-ABL deregulates many pathways. 23 ABL protein shuttles between the nucleus and cytoplasm; however, the chimeric BCR-ABL oncoprotein stops this shuttle and retains ABL protein in the cytoplasm, where it interacts with many different proteins that are involved in the malignant transformation. 24
In BCR-ABL fusion, the N-terminal oligomerization domain of BCR activates the ABL tyrosine kinase by blocking the SH3 on the N-terminal of ABL1. 25 Accordingly, BCR-ABL oncoprotein activates diverse signaling pathways including the RAS pathway, JAK2/STAT pathway, and PI3K-AKT-mTOR pathway. BCR-ABL oncoprotein directly activates JAK2/STAT pathway and continuously enhances cell survival in CML. Response and resistance to TKI TKIs have dramatically changed the treatment and prognosis of CML. TKI targets the proliferating mature BCR-ABL cells which are considered the major driver of CML. However, a minority of subclones with genetic aberration and quiescent leukemic stem cells are not eliminated by TKI. 27,28 Molecular resistance is measured by quantitative real-time polymerase chain reaction (qRT-PCR) of BCR-ABL1 and cytogenetic resistance (Ph+ persistence). These procedures are considered the main standard for monitoring the patient response to therapy, predicting relapse, and guide treatment decisions. Binding of ATP to tyrosine kinase active site of BCR-ABL oncoprotein phosphorylate tyrosine residue of substrate results in progression of CML. TKI impedes the ATP-tyrosine kinase binding and thereby halts the constitutive tyrosine kinase activity of the BCR-ABL1 oncoprotein and thus inhibits the CML progression. There are three types of TKI inhibitors. Type I is ATP competitive inhibitors like imatinib and nilotinib which compete with ATP in the ATP binding site of the kinase domain. This, consequently, prevents autophosphorylation, substrate phosphorylation and thus inhibits proliferation and induces apoptosis for BCR-ABL cells. Type II is ATP competitive inhibitors like dasatinib, which bind to the ATP-binding site as well as the hydrophobic adjacent binding site which is accessible only when the kinase is in an inactive configuration. The third type is allosteric inhibition which involves the binding of TKI to the allostericmyristate binding pocket, induces a conformational change in the kinase domain and renders the kinase-inactive. 31 It has been estimated that about 25% of CML patients develop resistance to TKI that necessitates to switch TKIs at least once during their lifetime. 32
Resistance to imatinib occurs in approximately 17% of patients with 5 years of follow-up. The mechanism of TKI resistance is subdivided into ABL1 mutation-dependent (acquired resistance) which means loss of response, and ABL1 mutation-independent (primary resistance), which results in conformational changes of the TKI binding site due to mutation in SH1 of ABL1. 33 Therefore, a new target therapy or combining TKI is required to cure CML. As a result, the new version of TKIs, dasatinib, and bosutinib, have been approved as first-line therapy for imatinib resistance patients. The dominant mechanism of imatinib resistance is acquired mutation in the kinase domain (KD) of BCR-ABL with greater frequency in AP, BP CML patients than CP. 35,36 These mutations reside in phosphate binding loop at positions M244, G250, Q252H, Y253H/F, and E255K/V, gatekeeper (T315, F317), the activation lobe (H396), SH2, and C-lobe (M351, F359) that ultimately impaired TKI binding by affecting the essential residue for direct contact or by preventing the inactive conformation of ABL kinase domain appropriate for imatinib binding. 37 Nilotinib shares the same binding site of imatinib with significantly higher affinity that overcomes many imatinib-resistant mutations except for Y253, E255, T315I, and F359 mutations. 38,39 About 25% of CML patients develop T315I mutation that inhibits TKI binding and develop resistance in all TKIs except ponatinib. [bib_ref] A phase 2 trial of ponatinib in Philadelphia chromosome-positive leukemias, Cortes [/bib_ref] T315I mutation, which occurs in gatekeeper residue of the ABL domain, affects the binding affinity of imatinib, dasatinib, bosutinib, and nilotinib to the ATP-binding site. Mutations within the ABL kinase domain and overexpression SRC are among the imatinib resistance mechanisms in CML. Dasatinib and bosutinib have dual inhibitory action against SRC and ABL kinase that address most of the BCR-ABL mutations associated with TKI resistance except V299, E255, and T315I. [bib_ref] Targeting ABL and SRC kinases in chronic myeloid leukemia: experience with dasatinib, Quintas-Cardama [/bib_ref] [bib_ref] Activity of bosutinib, dasatinib, and nilotinib against 18 imatinib-resistant BCR/ABL mutants, Redaelli [/bib_ref] Ponatinib, a pan-BCR-ABL inhibitor, has a great efficacy against native and mutated BCR-ABL1 including the T315I variant. [bib_ref] AP24534, a pan-BCR-ABL inhibitor for chronic myeloid leukemia, potently inhibits the T315I..., O'hare [/bib_ref] However, a dose-dependent increased risk of developing arterial occlusion, stroke, limb ischemia, and other vascular events has been found. [bib_ref] Ponatinib efficacy and safety in Philadelphia chromosome-positive leukemia: final 5-year results of..., Cortes [/bib_ref] [bib_ref] Ponatinib: A Review of Efficacy and Safety, Massaro [/bib_ref] Asciminib is an allosteric inhibitor, unlike other TKI inhibitors, that binds the myristoyl pocket of BCR-ABL. Asciminib is a potent and highly selective inhibitor of both native and mutated BCR-ABL including the gatekeeper T315I mutant. Asciminib is considered a relatively safe option for CP or AP CML with resistance to the currently available TKIs 46,47 [fig_ref] Table 1: The types of kinase domain mutations with their response and resistance to... [/fig_ref].
BCR-ABL-independent resistance might develop through upregulation of the PI3K/AKT/mTOR pathway. [bib_ref] Compensatory PI3-kinase/Akt/mTor activation regulates imatinib resistance development, Burchert [/bib_ref] At this point, PI3K/AKT/mTOR pathway inhibitors might be a new target for TKI resistance in CML patients. [bib_ref] Nilotinib and MEK inhibitors induce synthetic lethality through paradoxical activation of RAF..., Packer [/bib_ref] NVP-BEZ235 is a dual inhibitor which shows efficacy against imatinib-resistance CML by arresting the cell cycle at G0/G1 and inactivating the PI3K/AKT/mTOR pathway. [bib_ref] Efficacy of the dual PI3K and mTOR inhibitor NVP-BEZ235 in combination with..., Xin [/bib_ref] CML leukemic stem cells CML leukemia stem cells (LSCs) are subpopulations of CML cells that persist through TKI therapy and sustain the disease in patients. LSCs are quiescent cells that reside in the microenvironment of BM and remain a potential reservoir for disease relapse. LSCs are characterized by high heterogeneity of genetic, epigenetic, and transcriptional mechanisms. [bib_ref] BCR-ABL Independent Mechanisms of Resistance in Chronic Myeloid Leukemia, Loscocco [/bib_ref] The persistent LSCs result from resistance to targeted therapy that might happen primarily as a lack of initial response to treatment or acquired as disease relapses after initial response to the therapy. [bib_ref] Single-cell transcriptomics uncovers distinct molecular signatures of stem cells in chronic myeloid..., Giustacchini [/bib_ref] Of crucial importance, determination, and elimination of LSC might be of great need for cancer cure.
CD34+/CD38-is a stem cell marker of both normal stem cells and LSC. 53 CD34+/CD38-LSC in CML patients was found to express higher levels of CD33 compared to normal CD34+/CD38-stem cells. [bib_ref] CD34(+)/CD38(-) stem cells in chronic myeloid leukemia express Siglec-3 (CD33) and are..., Herrmann [/bib_ref] Recently, Herrmann et al. demonstrated that CD34+/CD38-CML-LSC express higher levels of CD 25, CD33, and CD 125 compared to normal BM stem cells. Moreover, immunological targeted therapy against these markers resulted in depletion of CML-LSC and reduce LSC engraftment. [bib_ref] Delineation of target expression profiles in CD34+/CD38-and CD34+/CD38+ stem and progenitor cells..., Herrmann [/bib_ref] In a study of single-cell gene expression analysis combined with an immunophenotypic screening of CML-LSC, an aberrant expression of CML-LSC surface markers CD25, CD26, and interleukin1 receptor accessory protein (IL1RAP) was detected. These markers were found to be downregulated after TKI administration. [bib_ref] Single-cell transcriptomics uncovers distinct molecular signatures of stem cells in chronic myeloid..., Giustacchini [/bib_ref] In contrast to normal stem cells, IL1RAP is a tight marker of BCR-ABL positive cells, and CD176 is a hematopoietic stem cell marker, co-expressed on CD34+/BCR-ABL+ cells of peripheral blood from CML patients. Interestingly, a bivalent antibody specific to both IL1PRAP and CD176 was found to target the cell population of CML but not their corresponding normal cells. Another study demonstrated that BCR-ABL CML CD34+/CD38-/CD26+LSC cells are resistant to TKI. [bib_ref] Residual Peripheral Blood CD26(+) Leukemic Stem Cells in Chronic Myeloid Leukemia Patients..., Bocchia [/bib_ref] Interestingly, it turned out that the pathways of TGF-ß and TNF-α were dysregulated in both BCR-ABL-and BCR-ABL+ CML-LSC with increased LSC quiescence and resistance to TKI. Moreover, increased serum levels of TGF-ß and TNF-α at diagnosis were associated with low treatment response in CML that might have an important role as a clinical predictive biomarker. [bib_ref] Single-cell transcriptomics uncovers distinct molecular signatures of stem cells in chronic myeloid..., Giustacchini [/bib_ref] Codavarthy et al. demonstrated that expression of CD44 on BCR-ABL + LSC and E-selectin on bone marrow endothelium participates greatly in engraftment of LSC in bone marrow niche which protects the LSC from imatinib treatment. Interestingly, GMI-1271, an E-selectin inhibitor, showed that releasing of BCR-ABL+ from the bone marrow niche increase responsiveness to Imatinib by decreasing the leukocyte counts and BCR-ABL cells.Bone morphogenetic proteins (BMP) are soluble growth factors that regulate stem cell fate and proliferation. The concentration of BMP2 and BMP4 was found to be higher in the bone marrow of CML patients in CP corresponding to normal bone marrow donors. [bib_ref] Primitive CML cell expansion relies on abnormal levels of BMPs provided by..., Laperrousaz [/bib_ref] BMP bind to its receptor, BMP receptor type B1 (BMPR1B), on LSC in the tumor niche and thus sustain the survival of BMPR1B+ LSC and progenitor cells. Upon treatment, LSC and progenitor cells increase expression of BMPR1B, which progressively activate the BMP4 autocrine loop and result in TKI resistance. [bib_ref] Immature CML cells implement a BMP autocrine loop to escape TKI treatment, Grockowiak [/bib_ref] The bi-directional signal of BMP4-BMPR1B LSC alongside the Jak2-stat pathway has a crucial role in the persistence of quiescent LSC in the microenvironment. Targeting BMP and the Jak2-stat pathway has been proposed to release the LSC from quiescent state and eradicate it by TKI 58 [fig_ref] Figure 1: Schematic illustrates various mechanisms of CML-LSC [/fig_ref].
miRNA role in CML TKI-resistance miRNA is a small non-coding single-stranded RNA that controls gene expression. [bib_ref] Attenuation of miR-126 activity expands HSC in vivo without exhaustion, Lechman [/bib_ref] Normally, miRNA-126 is expressed at a high level in hematopoietic stem cells and progenitor cells which conserve quiescence and restrain cell cycle progression. An increased level of miRNA-126 support the quiescent and engraftment of CML LSC in the bone marrow niche. However, its level found to be downregulated in CML LSC compared to normal hematopoietic stem cells.
Although BCR-ABL downregulates miR-126 in CML LSC, endothelial cells of bone marrow express a high level of miR-126 that supports CML LSC and thus sustain quiescence and leukemic growth. [bib_ref] Bone marrow niche trafficking of miR-126 controls the self-renewal of leukemia stem..., Zhang [/bib_ref] Moreover, quiescent CML LSC expresses a higher level of miR-126 and more engraftment capacity than proliferating CML LSC. Interestingly, a combination of TKI and the CpG miR-126 inhibitor resulted in the elimination of LSC. [bib_ref] Attenuation of miR-126 activity expands HSC in vivo without exhaustion, Lechman [/bib_ref] miRNA-409-5p expression is lower in peripheral blood of CML compared to healthy children. NUP43, the targeted gene of miRNA-409-5p, has shown to be overexpressed in CML children's patients. [bib_ref] MiRNA-409-5p dysregulation promotes imatinib resistance and disease progression in children with chronic..., Liu [/bib_ref] Overexpression of miRNA-409-5p downregulates the expression of NUP43 thereby inhibiting the proliferation and arresting the cell cycle. Furthermore, overexpression of miRNA-409-5p improves imatinib-induced proliferation inhibition and cell cycle arrest. [bib_ref] MiRNA-409-5p dysregulation promotes imatinib resistance and disease progression in children with chronic..., Liu [/bib_ref] Recently, in a study of miRNome profiling of LSC from CML-CP patients revealed a nine-fold increase of miR-196a-5p in CD34 +CD38ÀCD26+ (BCR-ABL+) compared to CD34+CD38ÀCD26-(BCR-ABL-). [bib_ref] miRNome profiling of LSC-enriched CD34(+)CD38(-) CD26(+) fraction in Ph(+) CML-CP samples from..., Ruiz [/bib_ref] Amplification of BCR-ABL and increased Jak2 signaling activate ADAR1, a double-stranded RNA protein that mediates adenosine to inosine (A-to-I) editing, thereby enhance the capacity of CML-LSC renewal by impairing the biogenesis of the Let-7miRNA family.miR-451, a tumor suppressor, has demonstrated downregulation in some cancer types. In CML, miR-451 expression is associated with favorable prognosis, as this miR targets ABL and BCR-ABL directly. [bib_ref] Clinical significance of microRNAs in chronic and acute human leukemia, Yeh [/bib_ref] In contrast, miR-21 is abundantly expressed in various solid and hematologic tumors. [bib_ref] Emerging role of microRNA-21 in cancer, Feng [/bib_ref] Investigating miR-21 and miR-451 levels at diagnosis of CML is of great importance as a predictor of optimal TKI response. A high level of miR-451 at diagnosis was significantly correlated to optimal response to TKI after 6 and 12 months whereas a high level of miR-21at diagnosis predicts a lower probability of reaching the optimal response to therapy. [bib_ref] MicroRNA signature refine response prediction in CML, Alves [/bib_ref] Inhibitors of miR-21 and PI3K together with imatinib significantly decrease AKT phosphorylation and C-MYC expression, thereby enhancing imatinib-induced apoptosis in CML-LSC. [bib_ref] Silencing of miR-21 sensitizes CML CD34+ stem/ progenitor cells to imatinib-induced apoptosis..., Wang [/bib_ref] Pellicano et al., demonstrated that has-miR183 is highly expressed in the BCR-ABL dependent pathway that mediates inhibition early growth response 1 (EGR1) leading to upregulation of E2F1, transcription factor, and consequently, enhance the proliferation of CML-LSC. 70 CML patients express a low level of miR30a compared to normal control individuals. Also, decreased level of miR30a has revealed an increase in the level of BCR-ABL that consequently enhances cell proliferation. [bib_ref] Decreased microRNA-30a levels are associated with enhanced ABL1 and BCR-ABL1 expression in..., Liu [/bib_ref] Moreover, targeting miR30a enhances imatinib activity against CML mediated by Becline1and autophagy protein 5. [bib_ref] Targeting microRNA-30a-mediated autophagy enhances imatinib activity against human chronic myeloid leukemia cells, Yu [/bib_ref] Upregulation of miR29a-3p and miR660-5p target and downregulate TET2 and EPAS1, respectively, that in turn confer TKI-resistance to CML-LSC. Furthermore, downregulation of miR-494-39 in LSC induces c-MYC overexpression, which is strongly connected to TKI-resistance BCR-ABL. 73
# Conclusion
TKIs have completely changed the treatment and improved the overall survival of CML patients. Dasatinib and bosutinib, new versions of TKIs, are considered as first-line therapy for CML patients who develop resistance to TKI (imatinib) because of KD mutation. CML LSC persists through TKI therapy and sustains the disease in patients. So, the determination and elimination of LSC might be of great need for cancer cure. Moreover, engraftment of CML LSCs in the bone marrow niche protect them from imatinib treatment. Therefore, inhibitors that are involved in releasing BCR-ABL+ from the bone marrow niche might increase responsiveness to TKIs. In CML, miRNA profile has a potential role as an effective diagnostic biomarker, monitoring disease progression and drug response. Furthermore, a biological understanding of the role of the miRNA in TKIs response and resistance still needs to be researched.
## Data availability
No data are associated with this article.
## Pinar arslan
Biology Department, Faculty of Science, Çankırı Karatekin University, Çankırı, Turkey
The review is about chronic myeloid leukemia (CML) connections with stem cells. First, the BCR-ABL gene locus is mentioned in the chromosome and then proteins are described. During the treatment of CML, TKI has been changed. TKI mainly target the BCR-ABL genes. The CML stem cells are some subpopulations of CML cells and are found in the patients' tissues in the diagnosis and treatment procedure. Therefore, CML is introduced very well. Besides, most of the figures are clear. The author has made the corrections suggested by the reviewer. However, the manuscript may be enhanced by the below points: In the introduction, the author is recommended to add information regarding other types of Leukemias with BCR-ABL translocation.
1. "expression" is removed and replaced by "miRNA" profile. 4.
Revision for English is done. 5.
[fig] Figure 1: Schematic illustrates various mechanisms of CML-LSC (chronic myeloid leukemia-leukemic stem cell) resistance and response to TKI (tyrosine kinase inhibitor).19. Barila D, Superti-Furga G: An intramolecular SH3-domain interaction regulates c-Abl activity. Nat. Genet. 1998; 18(3): 280-282. PubMed Abstract|Publisher Full Text 20. Schindler T, et al.: Structural mechanism for STI-571 inhibition of abelson tyrosine kinase. Science. 2000; 289(5486): 1938-1942. PubMed Abstract|Publisher Full Text 21. Soverini S, et al.: Chronic myeloid leukemia: the paradigm of targeting oncogenic tyrosine kinase signaling and counteracting resistance for successful cancer therapy. Mol. Cancer. 2018; 17(1): 49. PubMed Abstract|Publisher Full Text 22. Nagar B, et al.: Structural basis for the autoinhibition of c-Abl tyrosine kinase. Cell. 2003; 112(6): 859-871. PubMed Abstract|Publisher Full Text 23. Sattler M, Griffin JD: Molecular mechanisms of transformation by the BCR-ABL oncogene. Semin. Hematol. 2003; 40(2 Suppl 2): 4-10. Publisher Full Text 24. Pendergast AM, et al.: BCR-ABL-induced oncogenesis is mediated by direct interaction with the SH2 domain of the GRB-2 adaptor protein. Cell. 1993; 75(1): 175-185. PubMed Abstract|Publisher Full Text 25. Smith KM, Yacobi R, Van Etten RA: Autoinhibition of Bcr-Abl through its SH3 domain. Mol. Cell. 2003; 12(1): 27-37. PubMed Abstract|Publisher Full Text|Free Full Text 26. Warsch W, Walz C, Sexl V: JAK of all trades: JAK2-STAT5 as novel therapeutic targets in BCR-ABL1+ chronic myeloid leukemia. Blood. 2013; 122(13): 2167-2175. PubMed Abstract|Publisher Full Text 27. Apperley JF: Chronic myeloid leukaemia. Lancet. 2015; 385(9976): 1447-1459. Publisher Full Text 28. Lebecque B, et al.: DNA Methylation and Intra-Clonal Heterogeneity: The Chronic Myeloid Leukemia Model. Cancers (Basel). 2021; 13(14). PubMed Abstract|Publisher Full Text 29. An X, et al.: BCR-ABL tyrosine kinase inhibitors in the treatment of Philadelphia chromosome positive chronic myeloid leukemia: a review. Leuk. Res. 2010; 34(10): 1255-1268. Publisher Full Text 30. O'Brien SG, et al.: Imatinib compared with interferon and lowdose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N. Engl. J. Med. 2003; 348(11): 994-1004. PubMed Abstract|Publisher Full Text 31. Zhang J, et al.: Targeting Bcr-Abl by combining allosteric with ATP-binding-site inhibitors. Nature. 2010; 463(7280): 501-506. PubMed Abstract|Publisher Full Text [/fig]
[fig] Figure 2: is not clear to me. I would replace it with a table clarifying the various mutations and their responses to TKIs.2.I would suggest the English revision of the article. 3.Is the topic of the review discussed comprehensively in the context of the current literature? YesAre all factual statements correct and adequately supported by citations? YesIs the review written in accessible language? YesAre the conclusions drawn appropriate in the context of the current research literature? YesMohammad Al Hamad, Imam Abdulrahman Bin Faisal University, Dammam, Saudi ArabiaA paragraph about leukemia in general and its different categories is added. 1. [/fig]
[table] Table 1: The types of kinase domain mutations with their response and resistance to imatinib, dasatinib, nilotinib, bosutinib, ponatinib, and asciminib. [/table]
[table] 32: Steegmann JL, et al.: European LeukemiaNet recommendations for the management and avoidance of adverse events of treatment in chronic myeloid leukaemia. Leukemia. 2016; 30(8): 1648-1671. PubMed Abstract|Publisher Full Text 33. Ma L, et al.: A therapeutically targetable mechanism of BCR-ABLindependent imatinib resistance in chronic myeloid leukemia. Sci. Transl. Med. 2014; 6(252): 252ra121. Publisher Full Text 34. Tokarski JS, et al.: The structure of Dasatinib (BMS-354825) bound to activated ABL kinase domain elucidates its inhibitory activity against imatinib-resistant ABL mutants. Cancer Res. 2006; 66(11): 5790-5797. PubMed Abstract|Publisher Full Text 35. Branford S, et al.: Detection of BCR-ABL mutations in patients with CML treated with imatinib is virtually always accompanied by clinical resistance, and mutations in the ATP phosphatebinding loop (P-loop) are associated with a poor prognosis. Blood. 2003; 102(1): 276-283. PubMed Abstract|Publisher Full Text 36. Hughes T, et al.: Monitoring CML patients responding to treatment with tyrosine kinase inhibitors: review and recommendations for harmonizing current methodology for detecting BCR-ABL transcripts and kinase domain mutations and for expressing results. Blood. 2006; 108(1): 28-37. PubMed Abstract|Publisher Full Text 37. Quintas-Cardama A, Kantarjian HM, Cortes JE: Mechanisms of primary and secondary resistance to imatinib in chronic myeloid leukemia. Cancer Control. 2009; 16(2): 122-131. PubMed Abstract|Publisher Full Text 38. Weisberg E, et al.: Characterization of AMN107, a selective inhibitor of native and mutant Bcr-Abl. Cancer Cell. 2005; 7(2): Reviewer Report 28 February 2022 https://doi.org/10.5256/f1000research.120985.r125621 © 2022 Arslan P. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. [/table]
|
Physicians’ legal knowledge of informed consent and confidentiality. A cross-sectional study
Background: Only a few studies have been conducted to assess physicians' knowledge of legal standards. Nevertheless, prior research has demonstrated a dearth of medical law knowledge. Our study explored physicians' awareness of legal provisions concerning informed consent and confidentiality, which are essential components of the physicianpatient relationship of trust.Methods:A cross-sectional study assessed attending physicians' legal knowledge of informed consent and confidentiality regulations. The study was conducted in nine hospitals in Dolj County, Romania. Physicians were given a questionnaire with ten scenarios and instructed to select the response that best reflected their practice. We assessed the responses of physicians who claimed their practice to be entirely legal. Their legal knowledge was evaluated by comparing their answers to applicable laws. We also calculated a score for the physicians who admitted to committing a legal breach.Results:Of the 305 respondents, 275 declared they never committed any law violation. However, their median correct answer score was 5.35 ± 1.66 out of 10. The specialty was the strongest predictor of legal knowledge, with emergency physicians rating the lowest and non-surgical physicians scoring the highest. Physicians who worked in both private and public sectors were better knowledgeable about legal issues than those who worked exclusively in the public sector. Results indicate that physicians are aware of the patient's right to informed consent but lack comprehensive understanding. While most physicians correctly answered simple questions, only a tiny minority identified the correct solution when confronted with ethical dilemmas. The physicians who acknowledged breaching the law, on the other hand, had a slightly higher knowledge score at 5.45 ± 2.18.Conclusion:Legal compliance remains relatively low due to insufficient legal awareness. Physicians display limited awareness of legal requirements governing patient autonomy, confidentiality, and access to health data. Law should be taught in all medical schools, including undergraduate programs, to increase physicians' legal knowledge and compliance.
# Introduction
The presumption of legal knowledge is a jurisprudential postulate indicating that individuals are assumed to know the law. Although ignorance of the law is not punishable, it may result in sanctions if it manifests as illegal conduct.
## Open access
*Correspondence: [email protected] Plaiasu et al. Fundamentally, even if the state does not compel legal knowledge, it does impose conformity with the law. On the other hand, conformity does not always entail legal knowledge but may derive from personal convictions or serendipitous alignments. Nevertheless, knowledge of the law is a prerequisite for ensuring compliance and solving ethical dilemmas. Even though the law should generally be accessible to everyone, physicians may encounter various access and content accessibility obstacles. In several countries, medical education has included the study of law for some years to overcome these barriers. The primary objective is to educate physicians on the legal framework and better prepare them to deal with anticipated legal dilemmas. Contemporary medicine expects physicians to perform at a higher level than experienced practitioners and go beyond the bare minimum of a level of care. Physicians must incorporate legal knowledge into their clinical practice on an equal footing with scientific techniques in disease etiology and treatment [bib_ref] Law as clinical evidence: a new constitutive model of medical education and..., Parker [/bib_ref]. Although integrating ethics and legal education into medical curricula is not a simple endeavor, numerous universities are currently undertaking it.
Only a few studies have been conducted to assess physicians' knowledge of legal standards. Nevertheless, prior research has demonstrated a dearth of medical law knowledge among physicians. According to Australian physicians are insufficiently aware of their legal responsibilities regarding informed consent. The study indicates that physicians are unaware of the court order requiring them to inform patients of any problem they may attribute significance to [bib_ref] Doctors' understanding of consent law, Skiba [/bib_ref]. Riley's research discovered that despite physicians' attitude in favor of confidential treatment for adolescent patients, nearly half of the respondents were ignorant of the Michigan legislation governing a teen's right to consent to confidential therapy [bib_ref] Physician knowledge and attitudes around confidential care for minor patients, Riley [/bib_ref]. Al-Busaidi et al. concluded that the majority of respondents were uninformed of the Omani medical laws applied to the sharing of confidential information with third parties [bib_ref] Development and validation of an instrument to measure physician awareness of bioethics..., Al-Busaidi [/bib_ref].
There is no previous study to assess legal knowledge in Romania, but some previous research indicated low legal compliance among Romanian physicians [bib_ref] Locul şi relevanţa prevederilor legale în contextul practicii medicale din România, Nanu [/bib_ref]. Although prior research indicated that most physicians' practices contravened the legal criteria for informed consent and confidentiality, it did not investigate the causes. As physicians' behavior may be influenced not only by their legal knowledge, but also by a variety of other factors, such as insufficient training, lack of time, laziness, and hospital habits [bib_ref] Assessment of doctors' knowledge and attitudes towards confidentiality in hospital care, Beltran-Aroca [/bib_ref] we believe it is important to assess physicians' awareness of the laws governing patient rights.
Our study offers an insight into physicians' legal knowledge in a post-communist country with relatively new and inconsistent medical legislation. Due to the essential role in the physician-patient relationship of trust, we decided to explore physicians' legal knowledge of informed consent and confidentiality. Physicians' awareness and compliance in this sector are the most important since patients lack the knowledge and power to participate in the shared decision-making [bib_ref] Knowledge is not power for patients: a systematic review and thematic synthesis..., Joseph-Williams [/bib_ref].
We researched physicians' awareness of the regulations regarding the duties related to informed consent and confidentiality as stipulated by the Romanian medical legislation (Law number 46/2013, Law number 85/2016, and Minister of Health Order number 1411/2016). Since Romania ratified the European Convention from Oviedo and subsequently joined the European Union, the rights of Romanian patients are consistent with those guaranteed by the European Union's legislation. According to the law, physicians are required to inform patients before performing any medical act and acquire their written consent for all procedures that may pose risks. The law also acknowledges that patients with decisionmaking capacity have full decisional autonomy, whereas a medical arbitration commission has the final authority to determine the appropriate therapy for patients without decision-making capacity. In an emergency, physicians are compelled to intervene without the consent of a proxy. Additionally, physicians are required to maintain confidentiality with third parties and provide complete information to patients.
In the past, bioethics addressed these issues, leading to the regulation of patient rights. As a result, physicians are now required by law to perform specific duties, and failure to do so may result in civil or criminal sanctions. Since current Romanian attending physicians had received no formal education in law throughout their undergraduate studies, we anticipated a lack of legal understanding. In 2017, the University of Medicine and Pharmacy of Bucharest included the study of legal components pertinent to patients' rights and medical malpractice, and it is the only university in the country to have done so. Consequently, no current resident or attending Romanian physician has studied medical law.
# Materials and methods
## Setting
We conducted a cross-sectional study to determine attending physicians' understanding of laws regarding informed consent and confidentiality. The study was conducted in nine hospitals in Dolj County, Romania between November 2021 and March 2022. Dolj county is located in the southwestern part of Romania, with over 700,000 residents (7th place among Romanian counties) and a monthly net income of approximately 630 euros (16th place in Romanian counties). The local healthcare system includes thirteen public hospitals (one of which is a psychiatric facility that was excluded from the study due to different legislation applicable to patients). The healthcare system in Romania is built on a modified version of the Bismark model, and although it includes both public and private facilities, it remains highly centralized. Because resources and employees are primarily concentrated in the public sector, we focused our research on the public hospitals in the region. Six out of the twelve public hospitals agreed to our invitation to participate in the study. We also sent invitations to the three private hospitals with the largest number of physicians in the region, of which two participated in the study. To ensure a diverse range of healthcare practices and county coverage, we included different local hospitals: one tertiary university hospital, one secondary community hospital, two specialty hospitals, two primary hospitals, and two private day hospitals.
## Participants and data collection
The study was conducted on a convenience sample consisting of physicians working in hospital settings from Dolj county on a voluntary base. To approach the participants, we sought administrative consent from the hospital management. On the basis of their affirmative responses, we approached the hospital ward heads to seek second approval for each ward. Due to the pandemic, the ward heads deposited printed copies of the questionnaire in on-call rooms, where physicians could easily access them. After a week or two, depending on the circumstances, we collected all of the filled-in questionnaires. We have distributed 418 printed questionnaires in accordance with the information provided by the hospital ward heads. We included all available attending physicians. However, psychiatric physicians were not included due to their distinct legal responsibilities. Additionally, physicians who worked solely in laboratories and had no contact with patients were also excluded from the cohort. Considering the particularities and similarities across specialties, we opted to group physicians into five categories: surgical specialties (cardiovascular, general, oral and maxillofacial, pediatric, plastic, thoracic, neurosurgery, ophthalmology, orthopedic, otolaryngology, urology), non-surgical specialties (allergy and immunology, dermatology, gastroenterology, hematology, internal medicine, neonatology, neurology, nephrology, radiotherapy, pediatrics, physical medicine and rehabilitation, pulmonology, rheumatology, oncology), obstetrics-gynecology, emergency, and anesthesia and intensive care.
## Questionnaire
The questionnaire was adapted from one previously published by Nanu et al. [bib_ref] Locul şi relevanţa prevederilor legale în contextul practicii medicale din România, Nanu [/bib_ref] in light of the findings of the previously cited research. We added three more scenarios regarding informed consent and one self-assessment question to the initial questionnaire, validated in 2007. All additional questions were validated by a panel of five experts. We asked the experts to evaluate if the additional questions were essential, relevant, and intelligibly formulated. After calculating the index of content validity and ratio, we retained all additional questions.
We divided the questionnaire into three sections: Section 1: Information on the research participants concerning medical specialty, age, years in practice, workplace sector, and location. Section 2: Multiple choice scenario-based questions on informed consent and confidentiality (n = 10 questions). We provided physicians with ten distinct scenarios, seven of them referring to informed consent and three to confidentiality. To each scenario, we offered three alternative answers, only one of which was in accordance with the law. We instructed the participants to choose the responses that most accurately reflected their regular practice. Respondent's knowledge of informed consent was tested by scenarios number 3, 4, 5, 6, 7, 9, and 10, while confidentiality knowledge was addressed by questions number 1, 2, and 8.
Section 3: Self-assessment (n = 1 question). Finally, we asked them to self-assess their practice and respond to the question of whether they engaged in any medical acts that could be considered a breach of medical laws in the preceding 3 years.
The questionnaire was applied in the Romanian language to the Romanian physicians. An additional file presents the English translation of the questionnaire in more detail (see Additional file 1).
# Data analysis
The question asking physicians to self-assess their compliance with the law also represented the exclusion criteria. To calculate physicians' legal knowledge ratings, we included those respondents who answered negatively to the self-assessment question and declared that they did not engage in potentially illegal medical acts. Their responses to the ten scenario-based questions were compared to the applicable legal standards. The correct answer was the legally acceptable one, and each scenariobased question was weighted equally (one point for each correct answer and zero points for incorrect answers). Additionally, we calculated the compliance score for the physicians who admitted to having committed a legal breach and compared the results. We decided that a high Kuder-Richardson Reliability Coefficient was not desirable, as the study relates to education knowledge, and each question evaluates an aspect on an entirely different scale, loading onto different factors [bib_ref] The Use of Cronbach's Alpha When Developing and Reporting Research Instruments in..., Taber [/bib_ref].
We used Microsoft Excel (Microsoft Corp., Redmond, WA, USA), together with the XLSTAT add-on for MS Excel (Addinsoft SARL, Paris, France) and IBM SPSS Statistics 20.0 (IBM Corporation, Armonk, NY, USA) for processing the data. Descriptive analysis of the study group was performed with MS Excel. Normality tests (Anderson-Darling) and complex statistical tests (Chi-Square, Kruskal-Wallis, Friedman, etc.) were performed using the XLSTAT add-on or SPSS. We used the Anderson-Darling test to verify the normality of the data. None of the numerical variables investigated had a normal distribution of data, globally or inside each studied group. Because the study involved numerical comparisons between more than two groups that did not have a standard (Gaussian) distribution, the nonparametric Kruskal-Wallis test was primarily used, followed by a posthoc analysis using the Dunn method for multiple pairwise comparisons with Bonferroni correction. Categorical data were compared using the Chi-square test (χ 2 ), which is a statistical test that shows if there is a connection (association or influence) between two factors. It was used to interpret incidence tables generated by cross-tabulation of two categorical variables recorded in the study.
# Results
305 out of the 418 questionnaires sent out to attending physicians were returned resulting in a response rate of 72%. All the 305 questionnaires were completely filled in. To calculate the legal knowledge score, we excluded the 31 physicians who participated in the poll but said they had engaged in medical practices that might be considered unlawful. Among the remaining 274 clinicians who indicated they had committed no illegal medical act, 87% were under 60 years old, and 50% had worked for less than 15 years. [fig_ref] Table 1: Physicians'general characteristics [/fig_ref] summarizes the cohort's demographic characteristics, including physicians' specialty, work sector, age, and years in practice. The average age of physicians was 43.2 ± 10.9 years, and the average number of years in practice was 16 ± 10.8.
The median correct answer score in the ten scenariobased questions was 5.35 ± 1.66 out of 10. The specialty was the strongest predictor of legal knowledge, with emergency physicians rating the lowest and non-surgical physicians scoring the highest. Dunn's post hoc test with Bonferroni correction revealed significant differences (p < 0.001) between emergency physicians and surgeons, non-surgery physicians, and obstetrics-gynecology physicians, but not between emergency physicians and anesthesia or intensive care physicians (p < 0.0235) [fig_ref] Table 2: Physicians' Median Knowledge and Standard Deviation total, informed consent, and confidentiality scores [/fig_ref].
The analysis of additional variables revealed that physicians who worked in both private and public sectors were more knowledgeable about legal issues than those who worked exclusively in the public sector (p < 0,005). On the other hand, our research indicated that respondents' age and the number of years in practice had no impact on their legal knowledge [fig_ref] Table 3: Physicians' Mean Score and Standard deviation detailed according to demographics *Number of... [/fig_ref].
## Knowledge of informed consent legislation
The median correct answer score in the seven scenario-based questions regarding informed consent was 3.82 ± 1.22 out of 7. We discovered disparities in physicians' knowledge of informed consent legislation across specialties. Surgeons were more knowledgeable of informed consent regulations, whereas emergency physicians were the least familiar. Additionally, Dunn's post hoc test with Bonferroni correction revealed significant differences (p < 0.001) between emergency physicians and the other specialties [fig_ref] Table 2: Physicians' Median Knowledge and Standard Deviation total, informed consent, and confidentiality scores [/fig_ref].
## Scenarios pertaining a patient with medical decision-making capacity
Most physicians (98.9%) declared they always informed the patient in detail before a potentially risky maneuver. However, fewer physicians (94.16%) were knowledgeable that written consent from the patient is mandatory. Even fewer physicians (55.11%) knew they should seek written permission from patients before collecting and examining biological samples.
Once confronted with an ethical dilemma, most physicians offered the incorrect response. When asked to imagine a hypothetical life-threatening scenario in which a patient declines emergency assistance, just 32.48% of respondents recognized the law required them to accept the patient's decision.
In another scenario, 70.8% of physicians believed they had the right to terminate a patient's therapeutic relationship if the patient refused to adhere to the prescribed therapy.
## Scenarios pertaining a patient without medical decision-making capacity
A slight majority (52.19%) knew that if an incompetent patient's relatives refused to consent to proper medical care, the legislation required an arbitrary medical committee to select the most appropriate treatment for the patient.
Additionally, only one of the sixteen emergency physicians rightly indicated that when physicians intervene in an emergency without the consent of a substitute decision-maker, they are legally required to document the circumstance. The total percentage of respondent physicians correctly answering the question was low (32.12%).
Between emergency physicians and the other groups, there were statistically significant differences.
## Knowledge of conconfidentialityd health data disclosure legislation
The median correct answer score in the three scenariobased questions was 1.53 ± 0.85 out of 3. Two questions addressed the issue of disclosing information about patients' health to third parties, close relatives included. We asked physicians whether health data could be disclosed to third parties in the first question. We offered as an alternative the option of obtaining the patient's consent prior to disclosure. Most physicians (70.8%) selected the correct response. There were differences between specialties; for example, only one in two anesthesia and intensive care (47.06%) and emergency physicians (50%) knew the information about a patient's health was confidential. In the following case, we asked physicians if a patient's treatment details could be divulged to third parties. This question did not include the patient's permission as a viable response. Only 31.75% of respondents knew that the law prohibited disclosure in this circumstance.
In the scenario centered on physicians' attitudes regarding sharing health information with patients, almost half of respondents (49.64%) were unaware that patients had the right to acquire complete health records.
The confidentiality general knowledge score was not influenced by physicians' specialty, age, years in practice, or work sector [fig_ref] Table 3: Physicians' Mean Score and Standard deviation detailed according to demographics *Number of... [/fig_ref].
## Physicians who declared legal breaches
In terms of demographic features and legal compliance, there was no statistically significant difference between physicians who admitted committing a law violation and those who did not. In this group, the average age of physicians was 41.6 ± 7.5 years, and the average number of years in practice was 14.7 ± 8.1. Within this group, there was no significant variance in legal compliance attitude based on specialty, work sector, age, or the number of years in practice. The compliance score for these responders was 5.45 ± 2.18.
# Discussion
A deeper understanding of medical legislation is critical since it significantly impacts legal compliance and consequently the integrity of the patients' rights. Non-compliance with the law may expose physicians to significant risks in the event of malpractice claims.
Our study indicates that physicians have insufficient knowledge of informed consent and confidentiality laws. Compared to other studies regarding physicians' knowledge of various aspects of the law, our respondents' overall score was marginally lower. The confidentiality score of 1.53 out of 3 is consistent with Karasneh's (7 out of 14) but lower than Beltran Aroca's (6.8 out of 10) and Tegegne's (3.91 out of 7) [bib_ref] Assessment of doctors' knowledge and attitudes towards confidentiality in hospital care, Beltran-Aroca [/bib_ref] [bib_ref] Physicians' knowledge, perceptions, and attitudes related to patient confidentiality and data sharing, Karasneh [/bib_ref] [bib_ref] Health professionals' knowledge and attitude towards patient confidentiality and associated factors in..., Tegegne [/bib_ref].
Our findings corroborated prior studies indicating that although specialty was a predictor of legal awareness, age and years of practice were not [bib_ref] Development and validation of an instrument to measure physician awareness of bioethics..., Al-Busaidi [/bib_ref] [bib_ref] Doctors' knowledge of the law on withholding and withdrawing life-sustaining medical treatment, White [/bib_ref] [bib_ref] Knowledge and practices of obtaining informed consent for medical procedures among specialist..., Jukic [/bib_ref]. While we cannot explain the disparity in legal knowledge within specialties, a plausible explanation for the homogeneity of knowledge scores across age and years of practice groups is the abiding absence of legal or ethical education from university curricula. Additionally, we identified a gap between physicians who work entirely in the public sector and those who work in both the public and private sectors, implying that the private sector exposes physicians to additional legal standards. This could be attributed to the fact that private-sector Romanian physicians are more prone to be accused of medical malpractice, according to a recent study [bib_ref] Socio-demographic, professional and institutional characteristics that make Romanian doctors more prone to..., Hanganu [/bib_ref]. Results indicate that physicians are aware of the patient's right to informed consent but lack comprehensive understanding. Informed consent is a complex process. It is more than obtaining patients' signatures on a consent form [bib_ref] Informed consent and Italian physicians: change course or abandon ship-from formal authorization..., Turillazzi [/bib_ref]. While physicians are aware of their duties to inform, obtain and document consent, they demonstrate low legal understanding in more complex scenarios.
According to the law, physicians ought to be aware of the regulations regarding the respect for the right to autonomy, which include the obligation to inform the patient prior to any medical procedure, to obtain written permission for medical procedures that involve risks, to respect the decision of patients with decision-making capacity, and to make the appropriate medical decision for patients without decision-making capacity. In addition, physicians should be aware of their obligation to respect patient confidentiality and to grant patients complete access to their medical data.
In essence, physicians' responses suggest a fundamental incomprehension of the core idea of informed consent, which includes the patients' right to accept or refuse any medical procedure or treatment according to their beliefs and values without jeopardizing the patientdoctor relationship. For instance, there is a prevalent misconception among physicians that they are permitted to end their relationship with a patient in the event of treatment refusal. A previous survey found a better result, with slightly more than half of the family physicians responding that they would not discontinue therapy if patients declined their consultation request. In direct contrast to our findings, Craig's research revealed that 98.7% of respondents acknowledged a patient's right to refuse treatment [bib_ref] Clinicians' consent law knowledge: the case for education, Craig [/bib_ref].
Our research identified more situations where physicians' insufficient legal awareness resulted in arbitrary decision-making and low legal compliance. In one such example, roughly a third of the physicians acknowledged a patient's refusal to undergo potentially life-saving surgery. The low compliance rate matches White's study (32%), which presented physicians with a similar scenario [bib_ref] The role of law in decisions to withhold and withdraw life-sustaining treatment..., White [/bib_ref]. Although White's findings support the theory that there is a correlation between legal knowledge and legal compliance, he argues that there may be additional grounds for low legal compliance in the case of a life-threatening event. Moreover, he relates it to ethical considerations. According to his findings, the physicians' appraisal systems are built on a hierarchy of decision-making criteria, in which legislation is subordinate to clinical variables connected to the patient's health [bib_ref] The role of law in decisions to withhold and withdraw life-sustaining treatment..., White [/bib_ref] [bib_ref] The legal role of medical professionals in decisions to withhold or withdraw..., Willmott [/bib_ref]. Craig's study partially supports those findings, implying that physicians occasionally evaluate a patient's motivation and disregard his decision accordingly if they have different viewpoints or values [bib_ref] Clinicians' consent law knowledge: the case for education, Craig [/bib_ref]. His research, however, shows that after receiving educational training, physicians demonstrate higher levels of legal compliance. According to our findings, physicians are more predisposed to make medical rather than ethical decisions when ignorant of applicable laws. Nonetheless, we believe that additional research is required in order to gain a comprehensive understanding of the topic.
Furthermore, the analysis of responses regarding patients without decision-making capacity indicated that physicians have a limited understanding of the applicable laws, both in emergency and non-emergency scenarios.
According to Romanian legislation, a substitute decisionmaker must be found if a patient lacks the legal capacity to make medical decisions. A patient's substitute can be a parent, spouse, family member, or a conventional or court-appointed representative. They may seek and accept medical interventions for the patient but cannot veto a necessary medical procedure. Nearly half of the respondents in this instance were unaware that they were legally required to request an arbitrary committee decision. Additionally, physicians had little knowledge of their legal duty to complete a report when unable to obtain consent from a substitute decision-maker due to an emergency.
Moreover, the average confidentiality score indicates an insufficient awareness of the minimal legal standards for disclosing health information to third parties and assuring the patients' complete access to their data. Despite earlier research indicating that specialty [bib_ref] Confidentiality breaches in clinical practice: What happens in hospitals, Beltran-Aroca [/bib_ref] , age, and years in practice [bib_ref] Assessment of doctors' knowledge and attitudes towards confidentiality in hospital care, Beltran-Aroca [/bib_ref] [bib_ref] Physicians' knowledge, perceptions, and attitudes related to patient confidentiality and data sharing, Karasneh [/bib_ref] , are predictors of confidentiality breaches, our data revealed no variations among physicians. However, we did not assess other variables considered in prior research, such as gender, ethics education, patient volume, and the frequency of ethical issues physicians encounter in their field of practice [bib_ref] Health professionals' knowledge and attitude towards patient confidentiality and associated factors in..., Tegegne [/bib_ref].
Previous studies revealed multiple causes for confidentiality violations. Some are the product of negligence, and others result from poor infrastructure [bib_ref] Confidentiality breaches in clinical practice: What happens in hospitals, Beltran-Aroca [/bib_ref] , while others are the consequence of legislation unawareness. In our study, physicians demonstrated some unfamiliarity with legal requirements. On the good side, data indicate that most physicians know they may disclose confidential information to third parties with their patients' express permission. Nonetheless, when unaware of the patient's attitude, a sizable majority believed the law permits disclosure of health data.
Additionally, we noticed paternalistic behavior in supplying patients with access to their health records. A sizable proportion of respondents incorrectly assumed they were legally required to provide only minimal diagnostic and treatment information to patients.
We noticed a substantial discrepancy between our findings and other previously cited studies, especially those from Spain and Australia. We believe this is a result of several circumstances, including a relatively new Romanian medical legislation, a large number of laws, and the social and political context.
Comparing the two groups of physicians, those who have admitted to breaking the law and those who have not, we observed no significant differences in their interactions with patients. One of the plausible explanations could be attributed to the minority's urge to adhere to the attitude of the majority, as identified in prior research [bib_ref] Studies of independence and conformity: I. A minority of one against a..., Asch [/bib_ref].
## Limitations of the study
The primary study's limitations originate from biases imposed by our method of assessing legal knowledge. First, when physicians' opinions and laws align, compliance with the law may not result from actual knowledge. Second, self-perception biases might have influenced physicians to assess their practice as legal since individuals frequently convince themselves that their inadvertent unethical actions are legitimate.
Additionally, there are some limitations due to the scarcity of physicians in emergency, anesthesia, and intensive care units. Furthermore, despite our best efforts, we could not conduct our research in several tertiary hospital wards due to hospital ward heads' reluctance. Finally, our respondents were selected from a single county, and, despite the homogeneity in national educational curricula, the results may not be generalizable to all regions.
# Conclusion
Legal compliance remains relatively low due to insufficient legal awareness. Although physicians assess their activity as lawful, almost half of the cases resulted in violations of legal requirements. Physicians struggle with both procedural legal requirements and substantive aspects of the law. In our experience, physicians display limited awareness of patient autonomy and bodily integrity and hence are not fully capable of acknowledging patients' physical boundaries as sovereign. Additionally, physicians are insufficiently aware of their duties regarding patients' confidentiality and health data access.
Our study supports previous research [bib_ref] A qualitative interview study of Australian physicians on defensive practice and low..., Ries [/bib_ref] [bib_ref] Good Samaritan laws and graduate medical education: a tristate survey, Adusumalli [/bib_ref] [bib_ref] Knowledge, attitudes, and perceptions about medicolegal education: a survey of OB/GYN residents, Mathew [/bib_ref] [bib_ref] Treatment decisions for terminally Ill patients: physicians' legal defensiveness and knowledge of..., Van Mccrary [/bib_ref] , indicating legal education should be increased among physicians. Law should be taught at all medical schools, including undergraduate programs, to increase physicians' legal knowledge and compliance. Integrating legal education into medical curricula is not a simple task, but immediate interventions should be taken where an insufficient level of compliance is identified.
[table] Table 1: Physicians'general characteristics [/table]
[table] Table 2: Physicians' Median Knowledge and Standard Deviation total, informed consent, and confidentiality scores [/table]
[table] Table 3: Physicians' Mean Score and Standard deviation detailed according to demographics *Number of scenarios which equals total score available [/table]
|
Pelvic floor dysfunction one year after first childbirth in relation to perineal tear severity
The aims of this study were to evaluate pelvic floor dysfunction symptoms one year after delivery and investigate whether adverse functional outcomes after childbirth were related to the degree of perineal injury. A prospective cohort of 776 primiparas were included. Self-reported pelvic floor function data were obtained using a web-based questionnaire. Women with no/first-degree injuries, second-degree injuries, third-/fourth-degree injuries (obstetric anal sphincter injury, OASI) and cesarean section were compared. A total of 511 women (66%) responded. Second-degree tears were a risk factor for stress incontinence (aOR 2.6 (95% CI 1.3-5.1)). Cesarean section was protective against stress incontinence (aOR 0.2 (95% CI 0.1-0.9)). OASI was a risk factor for urge incontinence (aOR 4.8 (95% CI 1.6-15)), prolapse (aOR 7.7 (95% CI 2.1-29)) and pelvic pain (OR 3.3 (95% CI 1.1-10)). Dyspareunia was reported by 38% of women, 63% of women in the OASI group (aOR 3.1 (95% CI 1.1-9.0)). Women with OASI reported that the injury affected daily life (OR 18 (95% CI 5.1-59)). Pelvic floor dysfunction is common after childbirth, even in women with moderate injury. Women with OASI had significantly higher risks of symptoms of prolapse, urge urinary incontinence, pain, dyspareunia and impacts on daily life.Approximately 80% of primiparas suffer from perineal laceration. An estimated 40-50% of lacerations involve the perineal muscles, and up to 7% of these women suffer from severe obstetric anal sphincter injury (OASI) 1,2 . The prevalence rates of minor (first-degree) and muscular (second-degree) lacerations are estimated according to clinical experience, and these data are not always recorded in patient charts or registers. OASI is the most common cause of anal incontinence among women. Grade two and OASI can lead to perineal pain and dyspareunia and increase the risk of concomitant pelvic floor disorders 3 . The association between previous birth-related pelvic floor muscle trauma and future signs of pelvic organ prolapse has been shown in long-term follow-up studies 4 . The odds of fecal or urinary incontinence four years after delivery are markedly higher among women who experience pelvic disorder symptoms in the first year after birth 5 . Perineal trauma has been shown to relate to the frequency and severity of postpartum dyspareunia 6 .Research has primarily focused on the risk factors and complications of OASI, although vast areas in the field remain unexplored 7 . There is also a lack of knowledge about the extent of short-term complications following moderate spontaneous or iatrogenic injuries to the perineal muscles.The aims of this study were to evaluate symptoms of prolapse, urinary and anal incontinence, and perineal pain as well as sexual function one year after birth and investigate whether adverse functional outcomes were related to the degree of perineal injury. We hypothesized that the prevalence of symptoms of postpartum pelvic floor dysfunction would be related to the degree of perineal trauma.ResultsA flowchart of participant retention in the study is presented inFig. 1. The one-year follow-up questionnaire was distributed to 776 women, and 511 participants completed the questionnaire, accounting for a response rate of 66%. Of these participants, 74 (14.2%) were excluded from the analysis of symptoms due to de novo pregnancy, and 12 (2.3%) datasets were illegible/incomplete.Table 1shows the descriptive statistics of the responders and nonresponders. There were no differences between the women who participated in the follow-up and the nonresponders regarding rates of perineal trauma. The respondents were older, and among them, the vacuum delivery rate was higher. Epidural use was more common among nonresponders than among responders(Table 1).
www.nature.com/scientificreports/ The demographic and obstetric characteristics of the responders in the three different perineal trauma groups and the cesarean section group are presented and compared in . There were no differences among the three injury groups regarding maternal age, body mass index, infant head circumference, gestational age at delivery, induction rate, labor augmentation or epidural use. The duration of the active secondary stage of labor was significantly longer in patients with second-degree injuries than in those with an intact perineum or first-degree injuries. There was a significant difference in infant birthweight between patients with no or minor perineal trauma and OASI, with the highest infant birthweight among women affected by OASI. Occiput posterior presentation was more common in the OASI group than in the other groups, but the difference was not significant after post hoc testing.
Obstetrical vacuum extraction was most common in those with second-degree injuries. Mediolateral episiotomy was performed in 20.0% (85/424) of all women, resulting in iatrogenic injury in 45.5% (80/175) of the women with second-degree tears. Among women with second-degree lacerations, episiotomy was more common in deliveries by vacuum extraction (73.5% 25/34) than in spontaneous deliveries (38.7%, 55/142, p < 0.001). In the OASI group, the episiotomy rate was 26.3%, and none of the deliveries were vacuum assisted. There were no forceps deliveries.
Symptoms of prolapse were reported in 8.3% of the primiparas one year after delivery. OASI was a risk factor for developing symptoms of prolapse (aOR 7.7, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref]. In total, 6.2% of the patients had to insert a finger in the vagina to assist in emptying their bowels.
Urinary stress incontinence was present in 31.0% of women, and 18.0% suffered from urge incontinence. Second-degree trauma was a risk factor for stress incontinence (aOR 2.6, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref] , and giving birth by cesarean section was protective against stress incontinence (aOR 0.2, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref]. The risk for urge (aOR 4.7, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref] incontinence was elevated in the group with the largest injuries. An impact on lifestyle was reported in 12.1% of the women with urinary incontinence, and in women with OASI, the risk of urinary incontinence that affected their lifestyle was significantly elevated [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref].
Anal incontinence was experienced by 13.9% of women. Severe incontinence with leakage of solid stool was reported after only vaginal delivery. The severity of the symptoms was more prominent among women in the OASI group, who also had an increased risk of anal incontinence affecting their lifestyle, than among women in the other groups (OR 23, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref].
Most of the women were sexually active, although 9.7% of the women had not resumed sexual relations. Four women chose not to answer the question. Dyspareunia was experienced by a large proportion, accounting for 38.3% of the women who were sexually active. The rate of dyspareunia ranged from 31.3-41.4% in the groups with first/second-degree injuries and cesarean section and reached 62.5% in the group with OASI. The risk of . Demographic and obstetric characteristics of analyzed responders in the three different groups of perineal trauma and cesarean section. Data are the mean ± SD or n (%). *na = not analyzed. www.nature.com/scientificreports/ experiencing dyspareunia was elevated in women with OASI (OR 3.1, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref]. The feeling of being too tight (14.6%) was more common than the feeling of being too loose (8.6%). Perineal pain was experienced by 11.6% of women, and 21 women reported that pain that was severe enough to prevent most activities in the last 3 months. This type of severe pain was present in all groups, with an elevated risk in the OASI group (OR 3.3, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref]. OASI was a risk factor for symptoms originating from the injury that had an impact on daily activities (OR 18, [fig_ref] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence,... [/fig_ref]. Patients with OASI had the highest rate of complications.
# Discussion
In this study, we evaluated pelvic floor function one year after birth in a nonselected cohort of primiparas and found that adverse functional outcomes were experienced in women with perineal lacerations of all grades as well as those with an intact perineum. Dyspareunia and urinary incontinence were the most common symptoms of pelvic floor dysfunction one year after childbirth. Those who suffered from OASI had a significantly higher risk of experiencing symptoms of prolapse, urinary urge incontinence, anal incontinence, dyspareunia and pain. Patients with OASI also had a significant risk of their injury impacting daily life one year after giving birth. Notably, women with moderate perineal lacerations (second-degree and episiotomy injuries) reported various symptoms of perineal dysfunction one year after delivery, and second-degree trauma was a risk factor for urinary stress incontinence.
Women who underwent cesarean section generally had a low incidence of dysfunction apart from dyspareunia (31%), for which the prevalence was in line with that in women with an intact perineum. Cesarean section was also protective against stress incontinence. The results may have been affected by the fact that this was an allencompassing cesarean section group including women who underwent elective (35%) and emergency cesarean sections; nevertheless, the prevalence of perineal dysfunction seemed to be lower in this group than in the other groups. Similar findings have been reported in evaluations of pelvic floor function 5 and 10 years after birth, in which cesarean section deliveries were associated with significantly lower hazards for stress urinary incontinence, overactive bladder, and pelvic organ prolapse 8 . However, birth by cesarean section is associated with a risk of pelvic floor morbidity coupled to the surgery itself [bib_ref] Unintentional transvesical caesarean section: Incidence, risk factors, surgical technique and post-operative management, Franchi [/bib_ref].
Pain during intercourse was reported in 38% of the women, which is in line with data from previous studies (22.4-37.3%) [bib_ref] Postpartum sexual functioning and its relationship to perineal trauma: A retrospective cohort..., Signorello [/bib_ref] [bib_ref] Degree of bother from pelvic floor dysfunction in women one year after..., Lipschuetz [/bib_ref]. The etiology of dyspareunia is multifactorial; scar formation after perineal laceration is one possible cause; other causes include inflammatory states, infections, and lack of sexual arousal. The incidence of dyspareunia has been reported to increase after childbirth, from 12% before pregnancy to 31% six months after birth 12 . Breastfeeding has traditionally been used by midwives and gynecologists as an explanation for postpartum dyspareunia and was found to be a risk factor for dyspareunia at 6 months after birth in a recent study [bib_ref] Risk factors for dyspareunia after first childbirth, Alligood-Percoco [/bib_ref]. In Sweden, more than 70% of mothers have ceased breastfeeding at 12 months after birth 2 . To obtain in-depth knowledge of sexual function after childbirth, one needs to assess prepregnancy function and perform repeated follow-ups with a validated assessment tool, such as the pelvic organ prolapse/incontinence sexual questionnaire-IUGA revised (PISQ-IR) [bib_ref] The pelvic organ prolapse incontinence sexual questionnaire, IUGA-revised (PISQ-IR). Int. Urogynecol, Rogers [/bib_ref]. www.nature.com/scientificreports/ In our study, the frequency of urinary incontinence was lowest in the subgroup who underwent cesarean section and was highest among patients with major injuries. Symptoms of prolapse were most prominent in patients with OASI. A previous study showed that vaginal delivery was associated with higher risks of prolapse and urinary incontinence than cesarean delivery [bib_ref] Pelvic floor disorders 5-10 years after vaginal or cesarean childbirth, Handa [/bib_ref].
A similar trend was observed for anal incontinence. Excluding patients with OASI, severe anal incontinence with leakage of stool was found in only women with vaginal deliveries, which might indicate that they had undiagnosed sphincter injuries [bib_ref] Anal sphincter damage after vaginal delivery using three-dimensional endosonography, Williams [/bib_ref] [bib_ref] Prevalence of anal sphincter injury in primiparous women, Rojas [/bib_ref]. Various studies have postulated that the frequency of undiagnosed OASI is high, at 11-29% [bib_ref] Anal sphincter damage after vaginal delivery using three-dimensional endosonography, Williams [/bib_ref] [bib_ref] Prevalence of anal sphincter injury in primiparous women, Rojas [/bib_ref] [bib_ref] Occult anal sphincter injuries-myth or reality, Andrews [/bib_ref]. Women with OASI reported a significant impact of symptoms on daily life. Other studies have also shown that anal incontinence results in a decrease in overall quality of life [bib_ref] Determining levels of fecal incontinence in the community: A New Zealand cross-sectional..., Sharma [/bib_ref].
We recognize that there are limitations to our study due to the inclusion of measurements of rather subtle pelvic floor dysfunction characterized by symptoms occurring as little as once a month. There were no objective measures, such as in-person examinations, during follow-up. On the other hand, functional impairment is an easy way to identify patients who need further evaluation and treatment. In the aftermath of childbirth-related trauma, early identification of pelvic floor dysfunction enables us to intervene with effective strategies (such as pelvic floor rehabilitation) to prevent subsequent aggravation of pelvic floor dysfunction requiring surgery. Predicting and preventing long-term morbidity due to injury to the pelvic floor will also decrease healthcare costs [bib_ref] Costs of outpatients with fecal incontinence, Deutekom [/bib_ref].
There are a number of biases inherent to investigations utilizing questionnaires that we tried to minimize by using a shortened version of an already existing and validated national questionnaire 21 . One possible bias is that respondents experiencing symptoms were more likely to respond, but those experiencing symptoms were comparable to the nonrespondents in the majority of baseline characteristics. The responders were older than the nonresponders, representing a potential source of responder bias given that age is a risk factor for pelvic floor dysfunction, but responder bias is unlikely due to the minor differences. Confidence intervals were wide in the OASI group due to an OASI prevalence of 4.5%, leading to a small sample size (n = 19).
The strengths of this study include its prospective cohort design and the nonselected study population since the hospital is the only one in the area. Higher fetal weight, a longer secondary active state and occiput posterior position are well-known risk factors for both OASI [bib_ref] Modifiable risk factors of obstetric anal sphincter injury in primiparous women: A..., Jango [/bib_ref] and levator trauma [bib_ref] Intrapartum risk factors for levator trauma, Shek [/bib_ref]. Patients who stated that they were pregnant at the time of the questionnaire were excluded from the analysis of symptoms to ensure that the results were not affected by the impact of pregnancy on function; consequently, the population eligible for analysis was substantially reduced.
The average rate of vacuum extraction in the country is 9.5%; in this study, the prevalence was 11.1%. Responders had a higher rate of vacuum delivery than nonresponders, creating possible bias due to more traumatic birth experiences. The episiotomy rate was markedly higher than the national average, at 20% in this study compared to 10% nationally 2 . On the other hand, vacuum delivery and episiotomy were not risk factors for pelvic floor dysfunction in this study. Vacuum extraction is associated with an increased prevalence of OASI and second-degree tears [bib_ref] Risk factors for perineal and vaginal tears in primiparous women: The prospective..., Jansson [/bib_ref]. Episiotomy was more common in vacuum deliveries than in other types of deliveries, probably because clinical practice has been influenced by studies suggesting the reduced risk of OASI when performing episiotomy in vacuum deliveries in nulliparas [bib_ref] High episiotomy rate protects from obstetric anal sphincter ruptures: A birth register-study..., Raisanen [/bib_ref]. The benefits and risks of routine episiotomy have been widely debated [bib_ref] Variations in rates of severe perineal tears and episiotomies in 20 European..., Blondel [/bib_ref]. Recent studies have concluded that selective episiotomy is recommended when normal delivery is anticipated [bib_ref] Selective versus routine use of episiotomy for vaginal birth, Jiang [/bib_ref] [bib_ref] Selective use of episiotomy: What is the impact on perineal trauma? Results..., Franchi [/bib_ref]. A reduction in the episiotomy rate in unassisted vaginal deliveries led to an increase in second-degree injuries of anatomic damage less than or equal to that caused by episiotomy without and increased rate of severe second-degree injuries [bib_ref] Selective use of episiotomy: What is the impact on perineal trauma? Results..., Franchi [/bib_ref]. There is continuing support for the use of episiotomy in vacuum delivery in nulliparas 29 although the number needed to treat was 27, and the results from randomized controlled trials to validate these recommendations are awaited [bib_ref] Lateral episiotomy versus no episiotomy to reduce obstetric anal sphincter injury in..., Bergendahl [/bib_ref]. Episiotomy is an obstetric technique associated with levator injury [bib_ref] Intrapartum risk factors for levator trauma, Shek [/bib_ref]. Trauma to the levator ani during vaginal childbirth is associated with female pelvic organ prolapse 31 and has been suggested but not confirmed to have an association with symptoms of urinary and anal incontinence [bib_ref] Levator ani muscle injury and risk for urinary and fecal incontinence in..., Mathew [/bib_ref] [bib_ref] Levator ani muscle morphology and function in women with obstetric anal sphincter..., Volloyhaug [/bib_ref]. Regarding sexual function, a recent study found no association between postpartum levator hiatus and sexual dysfunction.
In conclusion, OASI is an evident risk factor for pelvic floor dysfunction after childbirth, but symptoms of pelvic floor disorder were found to be common, even in women with mild to moderate perineal laceration. Second-degree lacerations have a wide range of perineal tear extension and a subclassification may be helpful further to stratify the risk of subsequent pelvic floor dysfunction [bib_ref] Selective use of episiotomy: What is the impact on perineal trauma? Results..., Franchi [/bib_ref]. An elevated risk associated with a larger injury was observed in our study, and women with OASI reported a significant impact of their symptoms on daily life. This indicates that strategies should remain focused on preventative measures and improved diagnostics for large perineal lacerations.
Since pelvic floor disorders are stigmatizing conditions causing great physical and emotional suffering, measures aiming to reduce postpartum morbidity and to improve women's sexual health are of great importance. The association between perineal trauma and levator trauma 35 and the increased risk of long-term pelvic dysfunction with composite trauma constitutes an important area for further research to increase confidence in vaginal births and mitigate the increased demand for cesarean sections.
# Methods
This study was part of a large interdisciplinary project involving birth-related perineal injuries, endoanal ultrasonography and postpartum pelvic floor disorders. The present study is a follow-up study in the cohort of the aforementioned project. The project aimed to prospectively include all nulliparous women who gave birth from January 2016 to January 2018 at Östersund Hospital. This is the only hospital in the county of Jämtland-Härjedalen, Sweden, and approximately 98% of pregnant women residing in the area give birth at this hospital. Written consent to participate in the study was obtained at the routine fetal anomaly scan at 18-20 weeks of pregnancy. www.nature.com/scientificreports/ Women who agreed to participate in the study were clinically examined immediately after birth to diagnose perineal injury. The clinical examination was performed by a midwife or, in eligible cases, by a doctor, and consisted of inspection and bidigital palpation in the lithotomy position. The laceration grade was documented by the midwife or doctor according to the International Classification of Diseases Tenth Revision (ICD-10) World Health Organization (WHO) guidelines [fig_ref] Table 4: World Health Organization [/fig_ref]. Only mediolateral episiotomy was performed. The findings were documented in a study protocol. Data on maternal, obstetric and neonatal characteristics were obtained from medical records.
Participants who were one year postpartum received an invitation to complete a follow-up questionnaire by mail. An information sheet that explained the purpose of the study, security standards, data storage, and the consent process and provided the contact information of the responsible investigators was included with the letter. Participants were asked to complete a web-based questionnaire regarding symptoms of prolapse, urinary incontinence, anal incontinence, and perineal pain as well as sexual function. The questionnaire consisted of 15 questions and contained the most important items of the questionnaire for the postoperative evaluation of perineal injury provided by the National Swedish Society of Obstetrics and Gynecology 21 . The scored items of the questionnaire were validated. The questionnaire was available in only Swedish (native language). All questions had "no" or "never" as response options, and for questions about sexual function, patients had the option of not answering without being excluded from the analysis. Adaptive questioning processes were used in cases where supplementary questions were required to obtain further details on occurring symptoms, such as anal incontinence. Completion of the survey was voluntary. The patients provided informed consent by logging in with their individual codes and submitting the results. Participants received two reminders by text message and/ or telephone call. Answers were automatically recorded in a database to which only the responsible investigators had access.
The participants were divided into three groups depending on the extent of perineal injury (no/first-degree injury, second-degree injury, and OASI). Women who underwent cesarean section comprised the fourth group. The period prevalence of self-reported symptoms of prolapse, urinary and fecal incontinence, dyspareunia and perineal pain were calculated based on the proportion of women who reported symptoms that matched the study definitions divided by the total number of women with available data for follow-up. In cases of multiple birth, the study considered the infant with the highest birthweight. Prolapse was defined as the feeling of something bulging out of the vagina or the need to press against the posterior vaginal wall to empty the bowel once a month or more. Stress urinary incontinence was defined as incontinence once a week to daily, urge incontinence as once a month to daily. Impact on daily activity was defined as noticeable symptoms once a month or more often. Involuntary passing of flatus was defined as the involuntary passing flatus once a month or more. Leakage of stool was defined as stool leakage once a month to daily. Sexually active was defined as intercourse within the last three months. Perineal pain was defined as pain that is impossible to ignore and that affects concentration on daily tasks and activities. The statement "very satisfied" or "satisfied" was used to determine satisfaction with suturing.
Statistical analyses were performed by analysis of variance for continuous data (ANOVA). The chi-square test or Fisher's exact test was used as needed for categorical data, and the independent-sample t test was used for normally distributed continuous data. When analysis with the chi-square test was performed with more than two groups and statistical significance was found, pairwise testing and post hoc testing was performed to verify the finding. Odds ratios were calculated with binary logistic regression analysis, with 95% confidence intervals (CIs). Odds ratios were adjusted for age, body mass index (BMI) and fetal weight when appropriate. After ANOVA analysis, the posthoc test was performed to identify the groups that demonstrate statistically significant differences. Patients with no/first-degree injuries were compared with patients with second-degree injuries and OASI, as well as with participants who gave birth by cesarean section. Significance was set at p < 0.05. Analysis of power was performed in the initial study of the project; because the current study is a follow-up study involving the same cohort, power analysis was not applicable. All the data were compiled in an encoded database using Microsoft Access for Office 365, Version 1908 (Microsoft Corp., Redmond, WA, USA). Analyses were performed using SPSS Version 24 (IBM SPSS Statistics, Armonk, NY).
Ethical approval was granted by the Regional Ethical Review Board of Umeå (Nr. 2016-458-31 M). Informed consent was obtained from all participants. All data were deidentified prior to analysis. All experiments were performed in accordance with relevant guidelines and regulations. Patients could choose at any stage of the study to terminate their participation.
[fig] Figure 1: Flow chart of the study population. [/fig]
[table] Table 1: Demographic and obstetric characteristics of the study population. Data are presented as the mean ± standard deviation (SD) or n (%). * na = not analyzed. [/table]
[table] Table 3: Degree of trauma and odds ratios for symptoms of prolapse, urinary incontinence, anal incontinence, sexual function and other outcomes. † Cesarean section was compared to vaginal delivery. Data are OR (95%CI) ¤ Adjusted for age and fetal weight and BMI na = not analysed. [/table]
[table] Table 4: World Health Organization (WHO) classification of perineal tears. Grades 3-4 are regarded as obstetric anal sphincter injury (OASI). Beyond perineal skin into muscles of perineal body 3 A. External anal sphincter (less than 50%) B. External anal sphincter (more than 50%) C. External and internal anal sphincter 4 External and internal anal sphincter; anal mucosa www.nature.com/scientificreports/ 35. Shek, K. L., Green, K., Hall, J., Guzman-Rojas, R. & Dietz, H. P. Perineal and vaginal tears are clinical markers for occult levator ani trauma: A retrospective observational study. Ultrasound Obstet. Gynecol. 47, 224-227. https:// doi. org/ 10. 1002/ uog. 14856 (2016). [/table]
|
Beyond the Warburg Effect: N-Myc Contributes to Metabolic Reprogramming in Cancer Cells
Cancer cells generate large amounts of lactate derived from glucose regardless of the available oxygen level. Cancer cells finely control ATP synthesis by modulating the uptake of substrates and the activity of enzymes involved in aerobic glycolysis (Warburg effect), which enables them to adapt to the tumor microenvironment. However, increasing evidence suggests that mitochondrial metabolism, including the tricarboxylic acid (TCA) cycle, oxidative phosphorylation (OXPHOS), and glutaminolysis, is paradoxically activated in MYCN-amplified malignancies. Unlike non-amplified cells, MYCN-amplified cancer cells significantly promote OXPHOS-dependent ATP synthesis. Furthermore, tumor cells are differentially dependent on fatty acid β-oxidation (FAO) according to N-Myc status. Therefore, upregulation of FAO-associated enzymes is positively correlated with both N-Myc expression level and poor clinical outcome. This review explores therapeutic strategies targeting cancer stem-like cells for the treatment of tumors associated with MYCN amplification.
# Introduction
N-Myc contains a C-terminal basic region that can bind to DNA and a basic-helix-loop-helixleucine zipper domain that is responsible for the physical interaction with its counterpart MAX. Myc/MAX heterodimers bind to the DNA sequence 5 ′ -CACGTG-3 ′ , which is termed the consensus E-box. N-Myc represses transglutaminase 2 (TG2) transcription in cooperation with specific protein 1 (SP1) and the subsequent recruitment of histone deacetylase 1 (3); however, N-Myc also directly induces the transcription of a specific subset of ATP-binding cassette (ABC) transporter genes. These examples strongly suggest a double-edged sword role for N-Myc in transcriptional regulation according to cell context and tumor microenvironment. The Myc family is essential for normal development of the central nervous system, and the expression pattern of Myc changes from N-Myc to c-Myc during differentiation to transit-amplifying progenitors.
N-Myc is overexpressed in malignant neoplasms of the nervous system, including neuroblastoma, medulloblastoma, glioblastoma multiforme, retinoblastoma, and astrocytoma, as well as in non-neuronal tumors, including hematologic malignancies, small cell lung cancer, neuroendocrine prostate cancer, rhabdomyosarcoma, and Wilms tumors. N-Myc is expressed in self-renewing and quiescent hematopoietic stem cells, and expression changes to c-Myc upon differentiation to transit-amplifying progenitors. This finding suggests that N-Myc plays a role in the activation of stem cell-like properties characterized by self-renewal potential. Consistently, enforced expression of N-Myc, but not c-Myc, in murine bone marrow cells causes rapid development of acute myeloid leukemia in vivo. This is supported by the effect of N-Myc upregulation on driving the formation of a neuroendocrine tumor type that differs from c-Myc-driven prostate cancer in histology and response to androgen receptor (AR) signalingtargeted therapies. N-Myc suppresses AR signaling and induces polycomb repressive complex 2 (PRC2) target gene repression irrespective of phosphatase and tensin homolog deleted on chromosome 10 (PTEN) status. N-Myc binds to AR enhancers and forms a complex with AR in a manner dependent on its interaction with enhancer of zeste homolog 2 (EZH2). Furthermore, the catalytic activity of EZH2 promotes N-Myc/AR/EZH2-PRC2 complex formation. Thus, N-Myc might play a fundamental role in lineage switching from an epithelial origin to a neuroendocrine prostate carcinoma.
Despite its therapeutic potential, targeting N-Myc directly using small molecules remains challenging. As of this writing, there are very few reports demonstrating the efficacy of any compound targeting the binding between N-Myc and MAX, and effective small molecules capable of interfering with the N-Myc/MAX heterodimer in vivo have not been identified. Several indirect strategies to target N-Myc-driven malignancy are currently being explored, such as impeding MYCN transcription with inhibitors of bromodomain and extraterminal (BET) proteins such as JQ1; targeting proteins that increase N-Myc stability such as allosteric Aurora-A inhibitors; and exploiting synthetic lethal interactions with agents that deregulate N-Myc such as checkpoint kinase 1 (CHK1) inhibitors. CHK1 is involved in DNA repair, which is modulated by c-Myc-induced replicative stress. CHK1 transcription is markedly elevated in patients with MYCNamplified neuroblastomas. In addition, the MRN complex is composed of RAD50, meiotic recombination 11 (also referred to as MRE11), and NBS1 (also known as nibrin). MRN plays a critical role in processing, sensing, and repairing the doublestrand breaks of DNA. Petroni and Giannini reported that N-Myc-dependent proliferation of neuroprogenitor cells is accompanied by DNA replication stress, which is attenuated by the MRN complex, a direct transcriptional target of N-Myc. MRN inhibition via mirin also results in the accumulation of DNA damage response (DDR) markers and replication stress-associated DNA foci in an N-Myc-dependent manner. The functional inactivation of the MRN complex mediated by mirin in N-Myc-expressing neural cells fails to induce CHK1 Abbreviations: ACR, acyclic retinoid; α-KG, α-ketoglutarate; AMPK, AMPactivated protein kinase; AR, androgen receptor; BET, bromodomain and extraterminal; CD44v, CD44 variant; CHK1, checkpoint kinase 1; CPT1C, carnitine palmitoyltransferase 1; CSCs, cancer stem-like cells; EpCAM, epithelial cell adhesion molecule; ER, endoplasmic reticulum; EZH2, zeste homolog 2; FAO, fatty acid β-oxidation; GLDC, glycine decarboxylase; HCC, hepatocellular carcinoma; LDH-A, lactate dehydrogenase A; MCT, monocarboxylate transporter; mTOR, mammalian target of rapamycin; NASH, non-alcoholic steatohepatitis; OXPHOS, oxidative phosphorylation; PRC2, polycomb repressive complex 2; PTEN, phosphatase and tensin homolog deleted on chromosome 10; ROS, reactive oxygen species; SCD1, stearoyl-CoA desaturase-1; TCA, tricarboxylic acid; UFA, unsaturated fatty acid. phosphorylation and S phase arrest, whereas it activates both p53 and ATM to trigger apoptotic cell death. CCT244747, which is a highly selective and orally active CHK1 inhibitor, has shown therapeutic effects in an N-Myc-driven transgenic murine model of neuroblastoma. demonstrated that the compound 10058-F4, which is thought to disrupt the interaction between c-MYCN-Myc and MAX, impairs respiratory chain and FAO, resulting in apoptosis. A recent study showed in vitro that 10058-F4 is effective against acute promyelocytic leukemia and acute lymphoblastic leukemia with c-Myc overexpression.
## Metabolic reprogramming through the regulation of amino acid transporters by n-myc
Amino acid transporters contribute to metabolic reprogramming and maintain cancer stem-like phenotypes. xCT (SLC7A11) takes up cystine in exchange for glutamine, which is used for the synthesis of reduced glutathione (GSH), whereas ASCT2 (solute carrier family 1 member 5; SLC1A5) simultaneously takes up glutamine. The heterodimer composed of LAT1 (SLC7A5) and CD98 heavy chain (SLC3A2) is broadly and highly expressed in cancer cells and provides essential amino acids characterized by leucine, thereby activating the mammalian target of rapamycin (mTOR) complex1. Oncogenic c-Myc and hypoxia-induced factor 2α (HIF2α) upregulate LAT1 in a coordinated manner, whereas miR-126 suppresses LAT1 expression. The leucine influx mediated by LAT1 is associated with another amino acid antiporter, ASCT2. Pharmacological inhibition of ASCT2 suppresses LAT1-mediated leucine uptake, which leads to mTOR signaling inactivation in malignancy. Glutamine contributes to the synthesis of α-ketoglutarate (α-KG) via its conversion to glutamate, thereby promoting the tricarboxylic acid (TCA) cycle and the synthesis of nucleotides required for cellular proliferation. CD44 variant (CD44v)-positive cancer stem-like cells (CSCs) express high levels of xCT and ASCT2, which promote GSH synthesis from cysteine and α-KG from glutamine, respectively. Because c-Myc regulates amino acid transporters such as ASCT2 (23), c-Myc is likely to induce metabolic reprogramming in CD44v-positive CSCs. Collectively, metabolic reprogramming, which is orchestrated by the increased expression and interplay of amino acid transporters, results in glutamine addiction and protects CSCs from redox stress.
Ren et al.reported that MYCN-amplified neuroblastoma cells prominently depend on ASCT2 to maintain sufficient level of glutamine to activate TCA cycle. MYCN amplification is present in ∼30-40% of high-grade neuroblastoma patients and is a poor prognostic factor. MYCN-amplified neuroblastoma cells need an efficient machinery to meet the metabolic demands to keep enough amount of glutamine, which is a process which strictly relies on the interaction of specific amino acid transporters. High expression levels of Myc are necessary to maintain the glutaminolytic phenotype and addiction to glutamine as a bioenergetic substrate. Induction of key regulatory genes encoding glutamine transporters, glutaminase, and lactate dehydrogenase A (LDH-A) is important for glutaminolysis correlated with the Myc-induced increase in glutamine uptake and glutaminase flux. Glutamine addiction is associated with the activity of Myc in redirecting glucose carbon utilization away from mitochondria as a result of LDH-A activation. Consequently, Myc-transformed cells become dependent on glutamine anapleurosis to maintain mitochondrial integrity and TCA cycle function. Indeed, MYCN-amplified neuroblastoma cells display addiction to glutamine metabolism (36). Ren et al.identified a well-conserved N-Myc binding site in the ASCT2 promoter region. Although N-Myc and ATF4 cooperatively transactivate ASCT2, this amino acid transporter is significantly upregulated in response to glucose and/or glutamine deprivation and endoplasmic reticulum (ER) stress .
## Mitochondrial metabolic reprogramming in n-myc-driven cancer cells
Alptekin et al. proposed an alternative therapeutic strategy against MYCN-amplified neuroblastoma. These authors demonstrated that upregulation of genes associated with the serine-glycine-one-carbon (SGOC) metabolic pathway underlies the excessive dependence on glycine decarboxylase (GLDC). N-Myc activates GLDC transcription and is essential for maintaining high levels of GLDC expression in MYCN-amplified neuroblastoma cells, suggesting that GLDC and other SGOC pathway genes are cooperatively upregulated. Xia et al. reported that the SGOC pathway is inhibited by N-Myc-dependent metabolic vulnerability. Genes encoding SGOC-associated enzymes including PHGDH, PSAT1, PHSH, and SHMT are direct transcriptional targets of ATF4. The SGOC pathway provides potential targets for preventing therapeutic resistance in neuroblastoma. Increasing evidence supports the pathophysiological significance of GLDC upregulation. Increased expression of GLDC in CSCs from non-small-cell lung cancer is essential for their proliferation and tumorigenic ability by driving glycolysis, pyrimidine biosynthesis, and sarcosine production. In addition, GLDC upregulation is essential for the viability and growth of glioblastoma cells expressing high levels of serine hydroxymethyltransferase 2, which changes serine into glycine in mitochondria. GLDC contributes to metabolic reprogramming exclusively in MYCN-amplified neuroblastoma cells, as demonstrated by the effect of GLDC knockdown on central carbon metabolism pathways, including glycolysis and the TCA cycle, as well as lipid synthesis. GLDC knockout decreases the levels of glucose-6-phosphate, 3-phosphoglycerate, and lactate, suggesting that GLDC drives aerobic glycolysis. In addition, GLDC knockdown inhibits the glutamine-dependent reductive carboxylation pathway. Glutaminolysis and the reductive carboxylation pathway contribute to the synthesis of acetyl-CoA for lipid synthesis in tumor cells. Consistent with this finding, GLDC knockdown considerably decreases levels of fatty acids (palmate and myristic acid) and sterols (lanosterol and cholesterol).
MYCN-amplified neuroblastoma cells exhibit enhanced expression of genes and proteins involved in aerobic glycolysis (Warburg effect), oxidative phosphorylation (OXPHOS), detoxification of reactive oxygen species (ROS), and FAO. In MYCN-amplified tumor cells, glycolytic enzymes including hexokinase isoform 2 are upregulated, and enzymes associated with the TCA cycle and the electron transport chain, such as citrate synthase and isocitrate dehydrogenase isoform 2, are expressed at high levels. Increased expression of N-Myc-induced respiratory subunit genes is correlated with adverse clinical outcome in patients with neuroblastoma. A moderate increase in ROS in malignant neoplasms modulates cancer cellular signaling via the oxidation of cysteine. H 2 O 2 inactivates PTEN, a widely-known tumor suppressor, by oxidizing cysteine residues in the active site; this results in the formation of a disulfide bond, which prevents PTEN from inactivating the phosphatidylinositol-3-kinase (PI3K) signaling pathway. Therefore, cancer cells finely maintain ROS level within a narrow window which stimulates cellular proliferation. Because most mitochondrial proteins are overexpressed in MYCN-amplified neuroblastoma, and OXPHOS results in the production of intracellular ROS in mitochondria, the members of the peroxiredoxin ROS scavenger system, such as peroxiredoxin 6, are upregulated. Cancer cells tend to produce large amounts of lactate from glucose, regardless of the available oxygen level, and they activate glycolytic metabolism even before exposure to hypoxic conditions. However, neoplastic cells highly depend on the TCA cycle and OXPHOS in mitochondria, rather than on aerobic glycolysis in cytoplasm. The metabolic symbiosis between tumor cells and cancer-associated fibroblasts (CAFs) needs the expression of different subtypes of monocarboxylate transporter (MCT) by each cell population. Epithelial tumor cells typically express MCT1, which promotes the uptake of lactate generated and provided by MCT4expressing and caveolin1-negative CAFs. Cancer cells synthesize pyruvate from lactate, providing the TCA cycle and subsequent OXPHOS with an intermediate metabolite. demonstrated that MYCN-amplified neuroblastoma cells depend on both aerobic glycolysis and the TCA cycle; however, metabolic reprogramming driven by N-Myc may rely mostly on mitochondrial metabolism. Furthermore, N-Myc overexpression is related to the FAO process. Indeed, acyl-CoA dehydrogenase is involved in regulating the first step of FAO and is negatively correlated with clinical outcome of patients with MYCN-amplified malignancy. Accordingly, the FAO inhibitor etomoxir, which inhibits the rate-limiting FAO enzyme carnitine palmitoyltransferase 1 (CPT1C), suppresses the growth of xenograft tumors derived from MYCN-amplified neuroblastoma. This finding suggests that neuroblastoma cells are differentially susceptible to FAO inhibition according to N-Myc expression. FIGURE 1 | N-Myc and ATF4 act together to upregulate ASCT2, thereby enhancing mitochondrial metabolism. In MYCN-amplified neuroblastoma, the N-Myc/MAX heterodimer binds to the ASCT2 promoter region and aberrantly regulates its transcriptional level. Endoplasmic reticulum (ER) stress induces ATF4 via protein kinase-like ER kinase (PERK) and phosphorylated eukaryotic initiation factor 2α (p-EIF2α). ATF4 activates amino acid and glucose metabolism, and promotes protein translation to support increased biosynthetic activities. As such, ATF4 upregulates ASCT2 under ER stress conditions in N-Myc-overexpressing cancer cells. ASCT2 upregulation thus promotes robust uptake of glutamine, which is converted into glutamate and subsequently α-ketoglutarate (α-KG), a substrate of the TCA cycle.
## Lipid metabolic reprogramming in n-myc-driven tumor cells
High expression levels of CPT1C, a brain-specific metabolic enzyme, in N-Myc-positive neuroblastoma cells suggest that increased FAO might be an important metabolic feature in this malignancy. While CPT1A and CPT1B exist on the cytosolic surface of mitochondria, CPT1C is localized in endoplasmic reticulum; CPT catalyzes the synthesis of acyl-carnitine from acyl-CoA and promotes lipid conveyance through the membrane of mitochondria . The production of acyl-carnitine represents a pivotal regulatory step before FAO because both CPT1A and CPT1C are likely to be inhibited by malonyl-CoA, a final metabolite of the FAO of acyl-CoA molecules. Malonyl-CoA, a fatty acid precursor, simultaneously promotes fatty acid biosynthesis and inhibits fatty acid catabolism, thereby regulating the balance of intracellular fatty acids. MYCN-amplified neuroblastoma cells show high levels of CPT1C expression, and inhibition of both CPT1C and peroxisomal β-oxidation leads to lipid accumulation. Furthermore, impairment of bromodomain-containing protein 4 (BRD4) due to JQ1 significantly downregulates N-Myc in association with the formation of lipid droplets. These findings strongly suggest that N-Myc contributes to lipid metabolic reprogramming, and N-Myc inhibition is responsible for lipid accumulation. Qin et al.reported that signaling networks that regulate ER stress, characterized by the endocannabinoid cancer inhibition pathway, are regulated by stearoyl-CoA desaturase-1 (SCD1) in hepatic tumor-initiating cells with high N-Myc expression. SCD1 is a rate-limiting enzyme which contributes to the synthesis of monounsaturated fatty acids. Deletion of the gene encoding SCD1 increases the rate of FAO mediated by the AMPactivated protein kinase (AMPK) pathway and the upregulation of enzymes necessary for FAO. AMPK phosphorylation significantly increases when the function of SCD1 is inhibited, and this inhibition of SCD1 activity has favorable effects for lipid metabolism, such as attenuated lipogenesis and/or increased FAO; these effects are partly attributed to an increase in AMPK activation. Changes in AMPK phosphorylation caused by SCD1 up and downregulation affect NAD + levels following changes in NAD + -dependent deacetylase sirtuin-1 activity and epigenetic alterations characterized by histone H3 residue 9 acetylation and methylation status.
Deregulated N-Myc requires MondoA for lipid metabolic reprogramming in Myc-driven tumors. MondoA is associated with the outer membrane of mitochondria, where it can sense both glycolytic intermediate metabolites characterized by glucose 6-phosphate and mitochondrial metabolites necessary for glutaminolysis. Metabolites drive the nuclear localization of MondoA, which activates the transcription of genes related to glucose metabolism. Depletion of MondoA inhibits N-Myc-induced glutamine uptake, glutaminolysis, and glutamine-derived lipid biosynthesis, which is why apoptosis occurs in the absence of MondoA. There is a collaboration network between the nutrientutilizing N-Myc and the nutrient-sensing MondoA, and the orchestrated interaction between N-Myc and MondoA is critical for amino acid transport, lipid metabolism, nucleotide biosynthesis, and mitochondrial biogenesis. MondoA depletion is responsible for the significant downregulation of fatty acid synthase (FASN), stearoyl-CoA desaturase (SCD), and sterol regulatory element-binding protein-1 (SREBP-1). Metabolite set enrichment analysis indicates that N-Myc activation broadly alters the cellular metabolic characteristics, promotes fundamental changes in amino acid metabolism, and results in an increased amount of precursors for de novo lipid and purine biosynthesis., and N-Myc plays a pathological role in recurrence of de novo hepatocellular carcinoma (HCC). N-Myc is highly expressed in hepatic CSCs compared with non-CSCs. Furthermore, MYCN amplification occurs in at most 2.5% of HCC patients. In de novo HCC, there is a positive correlation between the expression of N-Myc and that of canonical Wnt signaling target molecules such as EpCAM, suggesting that N-Myc contributes to the stemness of HCC in cooperation with Wnt/β-catenin signaling. There is no correlation between the expression levels of c-Myc and N-Myc in de novo HCC. Alterations of fatty acid metabolism associated with metabolic reprogramming play a pivotal role in facilitating carcinogenesis in the liver. HCC cells import fatty acids and other lipids from the bloodstream; however, HCC cells upregulate enzymes involved in the biosynthesis of fatty acids, including SREBP-1-regulated genes including ATP citrate lyase (ACLY), acetyl-CoA carboxylase (ACC), FAS, and SCD-1, leading to the de novo synthesis of fatty acids. Lipid droplets are considered to be intracellular spherical organelles which are surrounded by a phospholipid single layer. The generation of lipid droplets is promoted by hypoxia related to HIF1-and HIF2-mediated repression of CPT1A, an essential enzyme involved in mitochondrial FAO. In addition, lipid-modifying enzymes that convert saturated fatty acid (SFA) to monounsaturated fatty acids (MUFA) are responsible for carcinogenesis. This is partly because the ratio of long chain n6-polyunsaturated fatty acids to n3polyunsaturated fatty acids is associated with a higher risk of HCC development in non-alcoholic steatohepatitis (NASH) model mice.
Qin et al. investigated the therapeutic effect of acyclic retinoid (ACR), a synthetic retinoid X receptor α-ligand, against hepatic CSCs expressing high levels of N-Myc. Hepatic CSCs with high expression levels of both EpCAM and N-Myc are more susceptible to ACR than non-CSCs negative for N-Myc expression. In general, inhibition of retinoid signaling in the liver is associated with the rapid progression of HCC associated with ROS. Thus, retinoids play crucial roles in mediating lipid metabolism in normal hepatic cells, and altered retinoid-inducing signaling is related to the progression of non-alcoholic fatty liver disease/NASH (NAFLD/NASH). Sp1 may promote the transcription of N-Myc in collaboration with E2F, whereas ACR-induced nuclear translocation of transglutaminase 2 inhibits N-Myc expression in association with the inactivation of Sp1. ACR is a promising vitamin A-like compound for the chemoprevention of HCC because it selectively kills N-Myc-overexpressing CSCs. ACR has chemopreventive effects on HCC mediated by the inhibition of the hyper-phosphorylation of retinoid receptors and lipid metabolic reprogramming. Indeed, lipid metabolic reprogramming is required for the initiation step of HCC tumorigenesis. Because lipid biogenesis as well as glutaminolysis are essential for the proliferation of N-Myc-driven cancer cells, inhibitors of fatty acid synthesis show specific toxicity to malignancy with high expression level of N-Myc. The growth-suppressive activity of ACR in HCC cells involves upregulation of pyruvate dehydrogenase kinase 4, which decreases the flux of glycolytic carbon into OXPHOS in mitochondria. This response switches the energy source from glucose to fatty acids to maintain the stable ATP production. Although synthetic inhibitors of unsaturated fatty acids (UFAs) such as oleic acid are toxic to N-Myc-overexpressing cells, UFA treatment partially rescues apoptosis induced by knockdown of MondoA, a nutrient-sensing transcription factor. Collectively, these findings indicate that disruption of N-Mycinduced lipid metabolic reprogramming may be responsible for the specific toxicity of ACR to hepatic CSCs.
# Conclusions and perspectives
N-Myc enables metabolic reprogramming of cancer cells, which cannot be simply explained by constitutive aerobic glycolysis (Warburg effect). However, MYCN-amplified cells depend on the TCA cycle and OXPHOS as well as lipid metabolism, rather than the Warburg effect. N-Myc upregulates ASCT2, the amino acid transporter contributing to glutamine addiction. MondoA, a nutrient-sensing transcription factor associated with Myc signaling, plays an important role in maintaining N-Mycinduced glutaminolysis and glutamine-derived lipid biosynthesis. ACR, a leading compound of vitamin A, was recently shown to specifically kill EpCAM-positive liver CSCs expressing high levels of N-Myc. ACR holds much promise for preventing de novo HCC recurrence. Such CSC population is enriched in enzymes necessary for lipid desaturation including FADS1/2 and SCD1. Considering the complexity of mitochondrial metabolism, further investigation is warranted to design novel therapeutic strategies targeting metabolic reprogramming triggered by N-Myc.
# Author contributions
GY searched the articles, wrote the manuscript, and submitted the paper to the journal.
# Funding
This review article was financially supported by the Japan Society for the Promotion of Science (19K23898). |
Acute intermittent porphyria: Diagnostic dilemma and treatment options
Acute intermittent porphyria (AIP) presents with diverse group of symptoms making its early diagnosis difficult. Delaying diagnosis and treatment of AIP can be fatal or can cause long term or permanent neurological damage. We present here a case report of AIP where the diagnosis was missed. The diversity of symptoms and details concerning the treatment options for AIP are discussed.
# Introduction
The porphyrias are uncommon and complex metabolic conditions caused by deficiencies in the activities of the enzymes of the heme biosynthetic pathway. Acute intermittent porphyria (AIP) is inherited as an autosomal dominant disorder due to deficiency of the enzyme porphobilinogen (PBG) deaminase located in the hydroxymethylbilane synthase (HMBS) gene in chromosome no. 11q23.3. [bib_ref] Intermittent acute porphyria -Demonstration of a genetic defect in porphobilinogen metabolism, Meyer [/bib_ref] Case Report A female patient aged 18 years weighing 35 kg was admitted to the Intensive Care Unit (ICU) in our hospital with complains of diffused dull aching abdominal pain, not related to food habit with occasional nausea and vomiting from 2 months. She was treated with proton pump inhibitors (tablet pentoprazole 40 mg twice daily) and antispasmodic medication (dicyclomin). Gradually she also developed diarrhea, vomiting and progressive weakness of her lower limbs and finally flaccid quadriparesis. At the time of presentation to ICU, she was conscious, oriented, Glasgow coma score (GCS) 15/15, pulse rate 100/min, blood pressure (BP)-130/80 mmHg, normal respiratory and cardiovascular system and flaccid quadriplegia. There was also a history of two episodes of seizure last night. She gave no history of accidental or intentional ingestion of any poisonous substances. Her family history was unremarkable. Initial laboratory investigations showed total leucocyte count 12.0 × 10 9 /L with neutrophil 80%, sodium 128 meq/L, chloride 96 meq/L and normal magnetic resonance imaging cervical spine and cerebrospinal fluid analysis. Treatment was started with anticonvulsants (injection phenytoin 100 mg 3 times a day), antibiotics (injection ceftriaxone 1 g twice daily) and correction of the electrolyte imbalance. Over the next 15 days, she developed respiratory muscle paralysis, was intubated and put on mechanical ventilation. Despite all measures she continued to have seizures and labile BP. Gradually she also became delirious with deterioration of GCS to 10. No improvement was observed in her laboratory parameters also. Finally urine PBG was done which was found positive. Electromyography results also demonstrated the presence of severe sensory motor axonal polyneuropathy. Immediately, all porphyrogenic drugs were stopped, phenytoin was replaced by gabapentin and a high-carbohydrate diet (400 to 450 g/day) was started. Treatment with hematin could not be started, due to the difficulty in obtaining the drug. Even after 1-week of therapy there was no improvement in her clinical condition and a decision was made to try hemodialysis. However, eventually she developed sepsis requiring high-ionotropic support which precluded further leading to her demise.
# Discussion
Acute intermittent porphyria also known as "Swedish porphyria" is one of the porphyrias which involves defects in heme metabolism resulting in excessive secretion of porphyrins and porphyrin precursors like amino levulinic acid (ALA) and PBG. [bib_ref] Acute intermittent porphyria, Herrick [/bib_ref] AIP is a rare autosomal dominant metabolic disorder characterized by a deficiency of the enzyme PBG deaminase [bib_ref] Intermittent acute porphyria -Demonstration of a genetic defect in porphobilinogen metabolism, Meyer [/bib_ref] (also known as HMBS or uroporphyrinogen 1 synthetase).
Its prevalence is 2-3 cases per 100000 persons per year. [bib_ref] Acute intermittent porphyria associated with respiratory failure: A multidisciplinary approach, Menegueti [/bib_ref] Though it is not much reported in India, it is much prevalent in some parts of the country like Kumar and Maheshwari communities of western Rajasthan. [bib_ref] Acute Intermittent Porphyria in a Kumhar community of Western Rajasthan, Sachdev [/bib_ref] The patients generally present with neurovisceral symptoms such as pain abdomen (85-95%), vomiting (50%), constipation (50%), peripheral neuropathy (42-68%), seizures (10-16%), delirium, coma and depression. [bib_ref] Acute Intermittent Porphyria in a Kumhar community of Western Rajasthan, Sachdev [/bib_ref] Autonomic disturbances may manifest as urinary retention, paralytic ileus, restlessness, tremor, excessive sweating, tachycardia, and labile BP. [bib_ref] Acute Intermittent Porphyria in a Kumhar community of Western Rajasthan, Sachdev [/bib_ref] Complications like bradycardia and sudden death have also been reported. [bib_ref] Intensive care management of patients with acute intermittent porphyria: Clinical report of..., Mehta [/bib_ref] The attack is often precipitated by environmental factors, reduced calorie intake, medications (barbiturates, calcium channel blockers, antibiotics, antifungals, and hormones), large alcohol intake, nicotine abuse, infection, surgery and psychiatric illness. [bib_ref] Acute intermittent porphyria associated with respiratory failure: A multidisciplinary approach, Menegueti [/bib_ref] [bib_ref] Intensive care management of patients with acute intermittent porphyria: Clinical report of..., Mehta [/bib_ref] Endo and exogenous steroids have a role in precipitating an acute attack which might be the reason for its frequent finding in females after puberty as in our case. [bib_ref] Acute intermittent porphyria with SIADH and fluctuating dysautonomia, Nabin [/bib_ref] Diagnosis of AIP is confirmed by detection of porphyrin or porphyrin precursors in freshly voided urine by Watson-Schwartz test using Ehrlich's aldehyde reagent. [bib_ref] Peripheral Neuropathy with severe intractable hyponatremia as a presentation of acute intermittent..., Arora [/bib_ref] Classic burgundy red discoloration of long stored urine is also a clue. [bib_ref] Peripheral Neuropathy with severe intractable hyponatremia as a presentation of acute intermittent..., Arora [/bib_ref] Quantitative measurements of PBG and ALA in urine or erythrocyte HMBS enzyme test are more reliable confirmatory tests. Heavy metal poisoning (i.e., arsenic and lead and thallium) may simulate AIP making their screening a mandatory requirement. [bib_ref] Peripheral Neuropathy with severe intractable hyponatremia as a presentation of acute intermittent..., Arora [/bib_ref] Porphyric neuropathy typically presents as a motor neuropathy of the axonal type preferentially affecting the proximal musculature with occasional sensory involvement. The predominantly motor neuropathy associated with weakness and areflexia in AIP can mimic Guillain-Barré syndrome. [bib_ref] Peripheral Neuropathy with severe intractable hyponatremia as a presentation of acute intermittent..., Arora [/bib_ref] Nerve conduction studies help in differentiating these two conditions as porphyric neuropathy which presents as axonal degeneration lacks the features of demyelinating neuropathy.
Treatment of AIP during acute attack is directed toward decreasing the heme synthesis and production of porphyrin precursors. Removal of precipitating factors, treatment of underlying infection, a carbohydrate diet and intravenous (i.v) heme remains mainstay of treatment of an acute attack. I.V. dextrose in high doses (300-500 g/day) blocks induction of the enzyme and prevents accumulation of precursors. However in severe cases especially if paresis occurs, heme therapy is indicated at 3-5 mg/kg/day for 3-5 days. [bib_ref] Acute intermittent porphyria -Diagnostic and treatment traps, Cojocaru [/bib_ref] Heme acts by depressing the ALA synthetase enzyme. [bib_ref] Acute intermittent porphyria -Diagnostic and treatment traps, Cojocaru [/bib_ref] The symptoms improve readily on heme therapy generally within 24 h. It is available in the form of hematin (Abott laboratories), heme albumin or heme arginate (Leiras Oy, Turku, Finland). [bib_ref] Acute Intermittent Porphyria in a Kumhar community of Western Rajasthan, Sachdev [/bib_ref] However, as these drugs are not easily available in India we could not use them. Charcoal hemoperfusion and hemodialysis has been used successfully. [bib_ref] Hemodialysis: A therapeutic option for severe attacks of acute intermittent porphyria in..., Prabahar [/bib_ref] Seizure can be controlled by correction of hyponatremia [bib_ref] Acute intermittent porphyria with SIADH and fluctuating dysautonomia, Nabin [/bib_ref] and gabapentine, vigabetin. [bib_ref] Acute intermittent porphyria with SIADH and fluctuating dysautonomia, Nabin [/bib_ref] Long-term monitoring is required in AIP patients as they are at increased risk for development of chronic renal failure and hepatocellular carcinoma. Screening of the family members even if asymptomatic to detect the genetic defect is important for prevention of an acute attack. [bib_ref] Acute intermittent porphyria -Diagnostic and treatment traps, Cojocaru [/bib_ref] |
Assessing a possible vulnerability to dental caries in individuals with rare genetic diseases that affect the skeletal development
Background: Individuals diagnosed with a rare genetic disease that affects skeletal development often have physical limitations and orofacial problems that exert an impact on oral health. The aim of the present study was to analyze the possible vulnerability to dental caries in individuals with rare genetic diseases that affect skeletal development.Methods: A paired cross-sectional study was carried out with a sample of 140 individuals [70 with rare genetic diseases affecting skeletal development: mucopolysaccharidosis (MPS) (n = 29) and osteogenesis imperfecta (OI) (n = 41) and 70 without rare diseases] and their parents/caregivers. The participants in the first group were recruited from two reference hospitals specialized in rare genetic diseases in the city of Belo Horizonte, Brazil. All participants were examined for the evaluation of breathing type, malocclusion, dental anomalies, oral hygiene and dental caries. The parents/caregivers answered a structured questionnaire addressing the individual/behavioral characteristics and medical/dental history of the participants. Statistical analysis involved the chi-square test and multiple logistic regression analysis for the dependent variable (dental caries) (α = 5%). This study received approval from the Human Research Ethics Committee of the Universidade Federal de Minas Gerais.Results:The mean age of the individuals was 10.34 ± 6.55 years (median: 9.50 years). Individuals with inadequate oral hygiene were 4.70-fold more likely to have dental caries (95% CI: 2.13-10.40) and those with the rare genetic diseases (MPS/OI) were 2.92-fold more likely to have dental caries (95% CI: 1.38-6.17).Conclusion: Individuals with inadequate oral hygiene and those with MPS and OI had a greater chance of belonging to the group with dental caries. Based on the present findings, individuals with the rare genetic diseases may be considered vulnerable to caries.
# Background
The World Health Organization (WHO) defines rare diseases as all diseases for which the prevalence is less than 65 cases per 100,000 inhabitants. Rare diseases are characterized as debilitating and chronically degenerative and require continuous medical follow up.
Affected individuals often have impaired physical, mental, sensorial and behavioral capacities, which can compromise their autonomy with regard to performing activities of daily living.
Mucopolysaccharidoses (MPS) and osteogenesis imperfecta (OI) are two rare genetic diseases that compromise skeletal development and affect general health. The two diseases lead to dental problems. Studies show that malocclusion, tooth agenesis, tooth rotation and microdontia are common in this population. These diseases are also associated with alterations in genes that regulate the formation of tooth enamel and dentin. Indeed developmental defects of enamel (DDE) are common in individuals with MPS and both dentinogenesis imperfecta (DI) and DDE are common in individuals with OI.
These dental problems make oral hygiene more difficult. The low mineral content in dental tissues in interaction with environmental factors may favor the occurrence of dental caries. Moreover, studies have suggested that access to dental services is more difficult for individuals with special needs. The difficulty in adequately performing oral hygiene due to the limitations imposed by disease and a lack of information on the part of parents or caregivers about the importance of oral health care can place individuals with special needs in a vulnerable position with regard to dental caries.
The concept of vulnerability in health is based on the understanding of susceptibility to illness. Illness is considered to arise from a set of individual, collective and contextual factors. This concept also involves the potential for coping with health problems in order to promote strategies for healthcare actions. In the present study, vulnerability is approached from the perspective of the expanded concept of health, exploring the more complex factors of biopsychosocial frailty that expose individuals with rare diseases to dental caries. The investigation of vulnerability provides more integrated means of assessing dental care needs taking into account the abstract and subjective elements associated with the process of becoming ill.
Therefore, the aim of the present study was to analyze the possible vulnerability to dental caries in individuals diagnosed with rare genetic diseases that affect skeletal development.
# Methods
A paired cross-sectional study was carried out with a sample of 140 individuals (70 with rare diseases and 70 without rare diseases) between two and 27 years of age and their parents/caregivers. A convenience sample was selected of individuals with two rare genetic diseases affecting skeletal development: MPS (n = 29) and OI (n = 41).
The group with rare diseases was recruited from two public hospitals in the city of Belo Horizonte, Brazil. The hospitals are reference centers for the treatment of these two diseases. Individuals without rare diseases were recruited from outpatient clinics at the same two hospitals. All hospitals belong to the The PS program (Power and Sample Size Calculation, version 3.0, Nashville, TN, USA) was used to calculate the test power. Considering the data obtained, the probability of exposure to dental caries among the controls was 34.3% and the correlation coefficient for exposure between matched cases and controls was 0.5. The odds ratio for dental caries among the individuals with rare diseases compared to the control group was 4.1. Thus, the test power was 100%, with a 5% margin of type I error.
## Data collection
Data collection involved oral examinations of the participants and the administration of a questionnaire to parents/ caregivers addressing sociodemographic and behavioral aspects of the participants (based on Oliveira et al., 2008a; 2008b. The type of rare disease was identified by the patient's medical record. Economic status was determined based on the Brazilian Economic Classification Criteria (ABEP), which addresses the purchasing power and general situation of households and classifies families in A1 (highest), B1, B2, C1, C2, D and E (lowest). Classes were categorized into high (A1, B1 and B2), middle (C1 and C2) and low (D and E). Ethnicity was categorized using the criteria established by the Brazilian Institute of Geography and Statistics (IBGE) for skin color: white, black, brown or yellow.
The oral examinations were performed by two examiners with the patient sitting in a chair under artificial light (Petzl Zoom head lamp®, Petzl America, Clearfield, UT, USA). The examiners used a mouth mirror (Duflex® n°5), Community Periodontal Index probe (Golgran®, São Paulo, SP, Brazil) and appropriate personal protective equipment to avoid cross infection. Radiography was not employed.
Breathing type was determined using the oral mirror test. A double-faced oral mirror was placed under the patient's nose. If the patient was a mouth breather (either alone or with nasal breathing), the mirror would be fogged on the lower portion; if the patient had only nasal breathing, the mirror would be fogged only on the upper portion.
The following malocclusions were investigated: overjet (increased/protrusion, anterior crossbite, absent), overbite (increased/deep bite, anterior open bite, absent, top) and posterior crossbite. The following dental anomalies were investigated: conical tooth, tooth agenesis, tooth rotation, DDE and DI. Tooth agenesis was considered a possible diagnosis, because the oral examination was performed only clinically. The diagnostic criteria for malocclusion and dental anomalies were based on Seow (2014),and the WHO (2013b).
The Simplified Oral Hygiene Index (OHI-S) was used to evaluate oral hygiene and was scored as follows: 0 = Absence of dental plaque/dental calculus; 1 = little dental plaque/dental calculus, less than 1/3 of dental surface covered; 2 = dental plaque / dental calculus covering more than 1/3 and less than 2/3 of the dental surface; and 3 = dental plaque/dental calculus covering more than 2/3 of dental surface. Plaque and dental calculus were evaluated separately. The final result of the OHI-S was obtained from the sum of the codes divided by the total number of teeth examined and classified as satisfactory (0 to 1), fair (1.1 to 2), deficient (2.1 to 3) or poor (≥3.1). The classification was dichotomized as adequate (satisfactory and fair) or inadequate (deficient and poor).
Dental caries was assessed according to the WHO diagnostic criteria. The number of decayed (presence of cavitated lesion) primary and permanent teeth was recorded.
## Training and calibration process
Training and calibration exercises were performed prior to the main study and were divided into theoretical and practical steps. The theoretical step involved the analysis of images of malocclusions, dental anomalies, different levels of oral hygiene and dental caries. The practical step performed at one of the hospitals selected for the main study. Due to the limited number of individuals with MPS and OI, only individuals without rare diseases were examined during the calibration process. These individuals were not included in the final sample. The results of the examiners were compared to the findings of an experienced epidemiologist (gold standard) using the Kappa statistic. Agreement was very good, with Kappa coefficients between 0.76 and 0.98 for all conditions examined.
## Pilot study
A pilot study was carried out after the calibration process involving 10 individuals with rare diseases and their respective parents/caregivers at the previously selected public hospitals. The pilot study indicated that no changes to the methodology were required. The participants in the pilot study were included in the main study.
## Directed acyclic graph
Prior to the data analysis, a directed acyclic graph (DAG) was used to select the covariates for the statistical adjustments. This is a theoretical method with visual representations of causal assumptions that is increasingly used in modern epidemiology to help identify confounding factors for the causal question at hand.
To identify possible confounding variables in the association between the rare diseases and dental caries, individual factors (diet, oral hygiene, physical and mental disability)and clinical factors (malocclusion, dental anomalies and breathing type)were included in the DAG model. Variables related to contextual and collective factors (previous dental experience, professional advice to consult a dentist, access to dental services, oral health policies, lack of experienced professionals and dental insurance)were also included. Based on the model, there were no confounding factors in the association between the rare diseases and dental caries. As the individuals in the different groups were matched for age, sex and economic status, these variables were not incorporated into the DAG.
# Statistical analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS for Windows, version 21.0, IBM Inc., Amonk, NY, USA). The chi-square test was used to determine the association between the exposure (dental caries) and the independent variables (p < 0.05). Multiple logistic regression analysis was performed to identify the independent impact of each variable studied. The independent variables were included in the decreasing logistic model in accordance with their statistical significance (p < 0.25; backward stepwise procedure).
# Results
The age of the 140 subjects examined ranged from two to 27 years (mean: 10.34 ± 6.55 years; median 9.50 years). The mean age of the parents/caregivers was 37.93 ± 9.00 years (median: 37.00 years).
The distribution of the 70 individuals with rare diseases is represented in . The most frequent types of dental anomalies in the population studied were DDE, DI and tooth rotation.
No significant differences between groups were found regarding sex (p = 1.000), age (p = 0.723), skin color (p = 0.859), parent's/caregiver's schooling (p = 0.205) and economic status (p = 0.301).
Rare disease (p < 0.001) and oral hygiene (p < 0.001) were significantly associated with dental caries.displays the results of the multiple logistic regression analysis. The variables "oral hygiene" and "rare disease" remained in the final model. Individuals with rare diseases had a 2.92-fold greater chance of belonging to the group diagnosed with dental caries (95% CI: 1.37-6.17; p = 0.005) and those with inadequate oral hygiene had a 4.70-fold greater chance of belonging to the group diagnosed with dental caries (95% CI: 2.13-10.40; p < 0.001).
# Discussion
The vulnerability concept discusses the health-disease process taken in account the more complex causes associated with it. The analysis of vulnerability to dental caries is an important point of reflection for formulating measures to ensure greater protection for vulnerable individuals. In contrast, isolated emergency actions do not modify causality.
The limitations imposed by some rare diseases can negatively impact quality of life and the occurrence of oral problems exacerbates this situation. Dental caries can cause acute or chronic pain, the formation of fistulas and abscesses, and tooth loss. The consequences of untreated tooth decay can affect different aspects of life, such as activities of daily living, sleep, speech, eating, social relations and self-esteem. The rare genetic diseases MPS and OI were chosen for the present investigation because both affect the development of the skeletal system. Malocclusions and dental anomalies are also commonly found in this population. Furthermore, care for the patients of these conditions is offered a university hospital that is considered to be a reference center for the treatment of rare genetic diseases.
In the present study, individuals with a poor oral hygiene (with or without a rare disease) had a greater chance of belonging to the group with dental caries. The individuals with rare diseases were also more likely to belong to the group with dental caries. The influence of oral hygiene on the prevalence of dental caries has been widely discussed. The increased chance of individuals with rare diseases having dental caries is the result of a set of factors. According to some authors, not only individual factors, but also collective and contextual factors lead to greater susceptibility to dental caries.
A physical limitation or motor impairment in a disabled individual can lead to dependence with regard to performing activities of daily living, such as oral hygiene. Moreover, parents and caregivers often have difficulties with this activity, performing it improperly, infrequently or even not at all. In the present study, the results did not prove that dental anomalies and malocclusion were associated with a greater chance of dental caries vulnerability in individuals with MPS and OI or those without rare genetic diseases. However, is important to consider that previous studies showed that the presence of malocclusion and dental anomalies can lead to the retention of food scraps and the accumulation of dental plaque and can also hamper tooth brushing.
Depending on its extent and the involvement of the organism, a rare disease can lead to a stressful routine of constant medical appointments, therapies and hospitalizations. As a consequence of this and also due to a lack of information and guidance, parents/ caregivers of affected children often fail to prioritize oral health, which constitutes a barrier to early and preventive dental care. It is therefore important for the medical team that cares for the patient with a rare disease to advise the parents/guardians to take him/ her to the dentist.
Previous studies suggest that the lack of oral health policies and programs targeting this portion of the population as well as limited knowledge and experience regarding the peculiarities of rare diseases make many oral health professionals feel uncomfortable and unprepared to treat these individuals . This constitutes another barrier to preventive dental care and consequently increases the risk of dental caries among individuals with rare diseases.
Some limitations of the present study should be considered. First, the factors associated with greater vulnerability to dental caries in individuals with MPS and OI were not investigated. Further research is needed to better understand such association. The cross-sectional this study design impedes the inference of causal relationships. Moreover, the use of questionnaires is always accompanied by the risk of recall bias on the part of respondents. However, this study also has strengths that should be highlighted. The use of a matched control group without rare diseases minimizes the possible influence of the matched characteristics on the association between the dependent and independent variables. The authors also used a DAG to identify possible confounding factors and explore the influence of individual, collective and contextual factors on vulnerability to dental caries.
It is certainly of great importance to reflect on vulnerability to dental caries among individuals with rare diseases, since such patients are a small part of the general population and do not enjoy the same visibility in terms of preventive measures implemented by public health programs. The greater commitment to general health care often leads to the neglecting of dental care needs. The lack of referral and orientation from other health professionals can aggravate the oral health status of these individuals, who also occupy an unfavorable position in terms of access to health services and they are less likely to receive dental care compared to the general population. This is presumably due to their physical limitations as well as difficulties oral health professionals face when treating these patients.
The study of vulnerability to dental caries allows theoretical approximations that are not restricted to individual behavior and the biomedical approach. It is necessary to strengthen integral care for individuals with rare diseases, ensuring access to dental services through public policies and professional training. Professionals who treat patients with rare diseases should advise parents and caregivers regarding the importance of caring for their children's oral health.
# Conclusion
Individuals with inadequate oral hygiene and individuals with MPS and OI had a greater chance of belonging to the group with dental caries. Based on these findings, individuals with the rare genetic diseases studied may be considered vulnerable to caries. |
Expression Profiling of Exosomal miRNAs Derived from Human Esophageal Cancer Cells by Solexa High-Throughput Sequencing
Cellular genetic materials, such as microRNAs (miRNAs), mRNAs and proteins, are packaged inside exosomes, small membrane vesicles of endocytic origin that are released into the extracellular environment. These cellular genetic materials can be delivered into recipient cells, where they exert their respective biological effects. However, the miRNA profiles and biological functions of exosomes secreted by cancer cells remain unknown. The present study explored the miRNA expression profile and distribution characteristics of exosomes derived from human esophageal cancer cells through Solexa high-throughput sequencing. Results showed that 56,421 (2.94%) unique sequences in cells and 7727 (0.63%) in exosomes matched known miRNAs. A total of 342 and 48 known miRNAs were identified in cells and exosomes, respectively. Moreover, 64 and 32 novel miRNAs were predicted in cells and exosomes, respectively. Significant differences in miRNA expression profiles were found between human esophageal cancer cells and exosomes. These findings provided new insights into the characteristics of miRNAs in exosomes derived from human esophageal cancer cells and the specific roles of miRNAs in intercellular communication mediated by exosomes in esophageal cancer.
# Introduction
Exosomes are small (40-100 nm) membrane vesicles that are derived from the invagination of endosomal compartments called multivesicular bodies and then released into the extracellular environment by most cell types [bib_ref] Extracellular organelles important in intercellular communication, Mathivanan [/bib_ref]. Recent studies have reported that exosomes exist in the culture supernatant of numerous eukaryotic cells, including cytotoxic T-lymphocytes [bib_ref] CD4 + T cell-released exosomes inhibit CD8 + cytotoxic T-lymphocyte responses and..., Zhang [/bib_ref] , B-lymphocytes [bib_ref] Exosomes from human lymphoblastoid B cells express enzymatically active CD38 that is..., Zumaquero [/bib_ref] , epithelial cells [bib_ref] Tissue factor-bearing exosome secretion from human mechanically stimulated bronchial epithelial cells in..., Park [/bib_ref] and tumor cells [bib_ref] Comparison of ultracentrifugation, density gradient separation, and immunoaffinity capture methods for isolating..., Tauro [/bib_ref] [bib_ref] Isolation of extracellular nanovesicle microRNA from liver cancer cells in culture, Kogure [/bib_ref] [bib_ref] Molecular characterization of exosome-like vesicles from breast cancer cells, Kruger [/bib_ref]. Exosomes have also been found present in physiological and pathological fluids, including plasma, urine, saliva, breast milk, malignant effusions and bronchoalveolar lavage fluid [bib_ref] Human saliva, plasma and breast milk exosomes contain RNA: Uptake by macrophages, Lässer [/bib_ref] [bib_ref] Surface-bound TGF-β1 on effusion-derived exosomes participates in maintenance of number and suppressive..., Wada [/bib_ref] [bib_ref] Altered microRNA profiles in bronchoalveolar lavage fluid exosomes in asthmatic patients, Leväne [/bib_ref] [bib_ref] Urinary exosomes as a source of kidney dysfunction biomarker in renal transplantation, Alvarez [/bib_ref]. Exosomes mediate the disposal of obsolete membranes from original cells [bib_ref] Small RNA transcriptomes of two types of exosomes in human whole saliva..., Ogawa [/bib_ref] , aid in intercellular communication, modulate selected cellular activities [bib_ref] Exosomes: A common pathway for a specialized function, Van Niel [/bib_ref] or modify their growth and mobility capacities [bib_ref] Analysis of exosome release and its prognostic value in human colorectal cancer, Silva [/bib_ref] and stimulate the proliferation, survival and adhesion of target cells [bib_ref] Blood/plasma secretome and microvesicles, Inal [/bib_ref]. However, the biological functions of exosomes remain unclear. A recent study has shown that cellular gene products, such as proteins, mRNAs and microRNAs (miRNAs), are packaged inside exosomes and are delivered into recipient cells, where they exert their respective biological effects [bib_ref] Let-7 microRNA family is selectively secreted into the extracellular environment via exosomes..., Ohshima [/bib_ref]. The expression of peripheral blood miRNAs from tumor cells can remain stable, because the extracellular nucleic acids in the bloodstream are highly protected by exosomes, microvesicle-like structures that can resist the degradation of various enzymes [bib_ref] Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in..., Arroyo [/bib_ref].
Esophageal cancer is a common type of cancer worldwide. However, esophageal cancer with distant metastasis and local invasion is still associated with a poor prognosis. To date, endoscopic and radiologic examinations are still being used as early detection methods. Simple and non-invasive diagnostic methods, such as blood and urine esophageal squamous cell carcinoma (ESCC) screening, are currently unavailable [bib_ref] Clinical impact of serum exosomal microrna-21 as a clinical biomarker in human..., Tanaka [/bib_ref]. Tumor-derived exosomes may become systemic by distribution through the blood stream; hence, exosomes in the blood or other body fluids of tumor patients can be used for clinical testing [bib_ref] Systemic presence and tumor-growth promoting effect of ovarian carcinoma released exosomes, Keller [/bib_ref]. The miRNA profiles of tumor-derived circulating exosomes isolated from body fluids may be used as novel diagnostic biomarkers when tumor cell materials are inaccessible. This strategy can also aid in the development of novel tumor therapeutics.
Next-generation RNA sequencing technology has rapidly progressed in recent years and is now widely used in various biological applications. This technology allows the rapid, sensitive and precise procurement of all RNA subtypes, as well as unannotated novel miRNA candidates or low-abundance RNAs that are weakly expressed [bib_ref] Identification and characterization of microRNAs involved in growth of blunt snout bream..., Yi [/bib_ref] [bib_ref] Genome-wide analysis of small RNA and novel microRNA discovery in human acute..., Zhang [/bib_ref] [bib_ref] Characterization of RNA in exosomes secreted by human breast cancer cell lines..., Jenjaroenpun [/bib_ref]. On the basis of small RNA digital analysis through Solexa high-throughput sequencing, the commercially available high-throughput sequencing technology allows the identification of numerous short RNAs in a sample. miRNA sequences are 19-25 nt long, thus, the Solexa platform is apparently a suitable choice for miRNA discovery [bib_ref] Deep sequencing discovery of novel and conserved microRNAs in trifoliate (Citrus trifoliata), Song [/bib_ref]. This technology can be used to explore the specific miRNA expression and distribution characteristics in tumor-derived exosomes and to discover novel miRNAs [bib_ref] Genome-wide analysis of small RNA and novel microRNA discovery in human acute..., Zhang [/bib_ref]. In the present study, Solexa high-throughput sequencing was used to explore the miRNA expression profile and distribution characteristics of exosomes derived from human esophageal cancer cells.
We utilized the next-generation RNA sequencing technology to construct two small-RNA cDNA libraries from human esophageal cancer cells and esophageal cancer-derived exosomes. This technology can describe the expression of miRNAs in both intracellular and extracellular environments of esophageal cancer. We identified miRNAs in the two libraries and analyzed differentially-expressed miRNAs through high-throughput sequencing and bioinformatics analysis.
We validated several miRNAs and their expression profiles in esophageal cancer cells and esophageal cancer-derived exosomes through stem-loop RT-PCR. This study provided insights into the involvement of exosome-mediated miRNAs in regulating an esophageal cancer microenvironment. We identified and profiled the RNA species present in exosomes and original cells; our results suggested that exosomal miRNAs can be used as potential esophageal cancer-specific biomarkers.
# Results
## Characterization of exosomes released by esophageal cancer cells
Based on the unique size and density of exosomes, we isolated exosomes from the culture supernatant of EC9706 cells following a classical ultracentrifugation protocol. [fig_ref] Figure 1: Characterization of exosomes through transmission electron microscopy and Western blot [/fig_ref] shows that the purified small vesicles had diameters ranging from approximately 30-60 nm and a lipid bilayer. Western blot analysis confirmed the presence of the known exosomal membrane protein, CD63, which is an exosomal marker; in addition, the cells tested positive for CD63 and β-actin [fig_ref] Figure 1: Characterization of exosomes through transmission electron microscopy and Western blot [/fig_ref].
## Overview of small rna sequencing data
To increase the coverage of cellular and exosomal miRNAs through Solexa sequencing, two small RNA libraries were constructed from cell and exosome RNA samples collected from the human esophageal cancer cell line, EC9706. After removing low-quality reads, contaminants, adaptors and sequence reads (length < 15 nt), we obtained 13,088,424 clean reads that represent 1,919,950 unique small RNAs in the cells and 7,736,476 clean reads that represent 1,226,905 unique small RNAs in the cotyledon. To simplify the sequencing data, all identical sequence reads in the small RNA library were grouped and converted into sequence tags. In total, 9,595,761 (cells) and 7,193,132 (exosomes) unique small RNAs with sizes ranging from 15-32 nt were detected from the two constructed libraries.
## Exosomes cells
[formula] CD63 (53kD) β-actin (43kD) [/formula]
To guarantee the accuracy of sequence data, only the sequences that were observed more than twice were selected as small RNA molecules [fig_ref] Table 1: Summary of high-throughput sequencing data detected in cells and exosomes [/fig_ref]. Further analysis indicated that several reads were annotated, including miRNA, mRNA, rRNA, tRNA, snRNA, other small RNAs and genomic repeats. The number of miRNAs was significantly small; however, the most abundant class of small RNAs in the cells and exosomes was unannotated small RNA [fig_ref] Table 1: Summary of high-throughput sequencing data detected in cells and exosomes [/fig_ref]. This finding indicated that a significant number of unannotated small RNAs that regulate gene expression in EC9706 cells and exosomes were secreted by the cells. The unannotated reads were used to identify novel miRNAs in subsequent analyses. The length distribution of unique small RNA sequences in the cells and exosomes remarkably varied. The lengths of unique small RNAs in the cells ranged from 21-24 nt. The most abundant size class was 22 nt, followed by 23 nt and then by 24 nt. Moreover, the lengths of unique small RNAs in the exosomes ranged from 19-22 nt; the most abundant size class was 28 nt [fig_ref] Figure 2: Length distribution of small RNAs in [/fig_ref]. This result can be attributed to the fact that the isomiRs (miRNA mature variants) in the exosomes were more clearly identified than the other small RNAs. The above phenomenon illustrates that the size class is miRNA-focused in the small RNAs and ranges from 15-32 nt in both cells and exosomes.
## Identification of known mirnas
To identify the conserved miRNAs, all unique small RNA clean reads from the small RNA libraries of the cells and exosomes were compared with the known human miRNAs in miRBase 18.0 (University of Manchester, Manchester, UK). We identified 342 and 48 types of published mature miRNAs in cells and exosomes, respectively. The lengths of the identified mature miRNAs ranged from 17-26 and 20-23 nt in cells and exosomes, respectively. Among all of the identified mature miRNAs, 22 nt-long miRNAs accounted for the highest proportions of 57.91% (cells) and 56.06% (exosomes) [fig_ref] Figure 3: Length distribution of mature miRNAs in cells and exosomes compared with known... [/fig_ref]. The absolute sequence reads were transformed into transcript abundance by data normalization. The clean read counts of miRNAs can represent the corresponding expression levels of each set of miRNAs. The miRNA reads varied from two to 382,634, indicating that the reads can still be detected through Solexa high-throughput, sequencing regardless of the miRNA expression level. Accordingly, miRNAs, such as mir-21-5p and let-7f-5p, were highly expressed, whereas mir-9-3p and mir-9-5p were lowly expressed [fig_ref] Figure 4: Expression abundance of known miRNA families in cells and exosomes [/fig_ref]. The clean read counts ranged from <10 to >100,000, which is across five orders of magnitude . hsa-let-7f-5p hsa-miR-21-5p hsa-miR-27b-3p hsa-miR-19a-3p hsa-miR-148a-3p hsa-miR-19b-3p hsa-miR-320a hsa-let-7g-5p hsa-let-7c hsa-miR-181a-5p hsa-miR-30c-5p hsa-miR-9-3p hsa-miR-151a-3p hsa-miR-31-5p hsa-miR-542-3p hsa-miR-152 hsa-miR-196a-5p hsa-miR-374a-3p hsa-miR-10a-5p hsa-miR-1180 hsa-miR-155-5p hsa-miR-20a-5p hsa-miR-30d-5p hsa-miR-98 expression abundance cells exosomes . Overall expression levels of known miRNAs in cells and exosomes.
## Ec9706 and their corresponding exosomes contain a subset of dysregulated mirna
We summarized the common and specific sequences between the cells and exosomes. Utilizing the isolated miRNA from EC9706 and their corresponding exosomes, we profiled the miRNAs and found 342 known miRNAs in the cells, 48 of which were detected in the exosomes. [fig_ref] Figure 6: Profiling of miRNAs isolated from EC9706 and their corresponding exosomes [/fig_ref] shows the number of shared and specific miRNAs between the EC9706-derived exosomes and cells. The cells contained 294 miRNAs that were not detected in the exosomes. Meanwhile, 64 and 32 novel miRNAs were detected in the cells and exosomes, respectively. Moreover, 12 of these novel miRNAs were detected in the two samples. Remarkably, several high-number miRNAs were selectively identified in the exosomes, but not in the cells. This result may be attributed to certain miRNAs that were transferred into the exosomes at low levels within the cells. This phenomenon may result in missing miRNAs when the cDNA library in the course of human miRBase establishment was constructed. [fig_ref] Table 2: Common transcripts in intracellular samples in the high category of more than... [/fig_ref] show the top-ranking miRNA transcripts in the intracellular and extracellular spaces, respectively. Noticeably, the let-7 family of miRNAs is predominantly represented in the top-ranking miRNAs in both intracellular and extracellular samples of esophageal cancer. In addition to the let-7 family, other miRNAs, such as miR-21-5p and miR-26a-5p, were also highly expressed in the intracellular and extracellular esophageal cancer samples. All known miRNAs in exosomes were detected at lower levels compared with their corresponding cells, whereas several novel miRNAs in exosomes were detected at higher levels compared with their corresponding cells [fig_ref] Table 3: Common transcripts in extracellular samples that belong to the mid-range category with... [/fig_ref]. Among the 12 common novel miRNAs, eight miRNAs were upregulated by more than two-fold in the exosomes, and four miRNAs were down-regulated by less than two-fold in the cells. Overall, these results show that the miRNA profiles in the cells and exosomes are notably different, which agree with the observations that miRNA is sorted and released in exosomes through unknown mechanisms [bib_ref] Selective release of microRNA species from normal and malignant mammary epithelial cells, Pigati [/bib_ref].
## Prediction of novel mirnas
To identify potential novel miRNAs from the small RNA library of EC9706 cells and exosomes, the following approaches and criteria were used: (i) reads that matched known miRNAs (miRBase 18.0) and other non-coding RNAs were excluded; (ii) non-conserved unique reads that were only sequenced once were removed; and (iii) to be considered as new miRNAs, reads must be entirely within the arm of the hairpin, and the hairpin must not possess large internal loops and bulges [bib_ref] Solexa sequencing analysis of chicken pre-adipocyte microRNAs, Yao [/bib_ref]. Based on the criteria for miRNAs used in the present study, we detected 64 and 32 novel miRNA candidates in cells and exosomes, respectively, using miRDeep2 software [bib_ref] Rajewsky, N. miRDeep2 accurately identifies known and hundreds of novel microRNA genes..., Friedländer [/bib_ref]. The 12 miRNAs found in both libraries can be divided into two categories, namely, miRNAs with only a single locus and those with multiple loci [fig_ref] Table 5: Candidate miRNAs with only a single locus [/fig_ref]. The expression levels of the candidate miRNAs in the cells and exosomes were significantly different. In addition, the length of these newly identified miRNA sequences varied from 17-25 nt, in which the 17, 18 and 22 nt miRNAs were dominant.
## Sequence variations in mirnas
IsomiRs are miRNA variants commonly reported in next-generation RNA sequencing studies. IsomiRs are encoded by the same pre-miRNAs and exhibit sequence variations from the reference miRNAs in miRBase [bib_ref] Dynamic isomiR regulation in Drosophila development, Fernandez-Valverde [/bib_ref]. Such a phenomenon was also observed in the present study. The results of Solexa sequencing revealed that most of the identified miRNAs showed length and sequence heterogeneity. Multiple unique reads that map the same position of a hairpin sequence (isomiRs) were further represented by the sequence of the most abundant read, because the read counts of the most abundant isomiR, rather than of the miRBase reference sequences, provided the most robust approach for comparing the expression levels between samples . In 158 cases, the most abundant sequence did not exactly correspond to the current human miRBase 18.0 reference sequences. For instance, the size of has-miR-7 is 24 (the read number is 30,883) in our library, but 23 (the read number is 8329) in miRBase 18.0. The size of another miRNA has-miR-222 is 23 (the read number is 16,121) in our library but 21 (the read number is 1047) in miRBase 18.0. This difference also suggests that either the relative abundance of isomiRs may vary across different studies or that the original submission of this miRNA to miRBase is incorrect [bib_ref] Discovery of porcine microRNAs in multiple tissues by a Solexa deep sequencing..., Xie [/bib_ref].
## Figure 7.
Example of has-miR-21 annotation for sample cells and exosomes. Various isomiRs are observed for several known miRNAs. The reference miRNA sequence from miRBase is, in a few cases, not the most frequently observed isomiR. The repertoire of has-miR-21 and its isomiRs contains the following: structure and sequence information of has-miR-21 pre-miRNA in miRBase, including ID, genomic location, folding free energy, and its secondary structure; sequence and read counts of all isomiRs aligned to has-miR-21 pre-miRNA hairpin; sequence and read counts of mature has-miR-21 annotated in miRBase; sequence and read counts of has-miR-21 as the most abundant isomiR; and total read counts found at the 5p or 3p arm of has-miR-21 pre-miRNA. "+" means positive sense strand in genomic DNA.
## Mirna validation assays by qrt-pcr
To determine whether or not the changes that occurred within exosomes also occurred in cells, qRT-PCR analysis of miRNA was performed on EC9706 and exosomes released from the cells. In agreement with the sequence data, this analysis confirmed that the expression levels of the five miRNAs that showed the highest expression in exosomes were relatively similar to those in cells [fig_ref] Figure 8: Expression profiles of five miRNAs in cells and exosomes [/fig_ref].
# Discussion
miRNAs are important members of small, well-conserved and non-coding RNAs. Their function in cancer pathogenesis as oncogenes or tumor suppressors has been investigated and attracted attention for several years [bib_ref] A review of the current understanding and clinical utility of miRNAs in..., Sakai [/bib_ref]. With the development of next-generation RNA sequencing technology, an increasing number of miRNAs in tumor-derived exosomes has been discovered [bib_ref] Characterization of RNA in exosomes secreted by human breast cancer cell lines..., Jenjaroenpun [/bib_ref] [bib_ref] Profiling of microRNAs in exosomes released from PC-3 prostate cancer cells, Hessvik [/bib_ref]. However, the roles of miRNAs in esophageal cancer-derived exosomes remain unknown. Serum miR-18a level was significantly higher in patients with esophageal, pancreatic or hepatocellular cancer. The areas under the receiver-operating characteristic curve (ROC) were 0.944, 0.936 and 0.881 in esophageal, pancreatic and hepatocellular, respectively, which showed that serum miR-18a may serve as a screening biomarker. Researchers have characterized the miRNA profiles of colorectal cancer (CRC) serum exosomes using microarray analyses to identify several special miRNAs up-regulated in CRC patients and found that the levels of seven miRNAs are significantly down-regulated after surgical resection of tumors. Through ROC analysis, miR-23a and miR-1246 were detected to have higher sensitivities of 95% and 90%, respectively, for Stage I samples compared with known tumor markers (CA19-9 and CEA). This finding supports the promising application of miR-23a and miR-1246 as diagnostic biomarkers [bib_ref] Circulating exosomal microRNAs as biomarkers of colon cancer, Ogata-Kawata [/bib_ref]. Another study on the serum microRNA expression profile of esophageal cancer showed that the expression level of miR-1246 is higher in ESCC patients than in controls and that this expression level decreases after surgical resection. To assess the diagnostic and prognostic values of serum miR-1246 in ESCC, the ROC curve assay of miR-1246 was conducted. Results showed that the sensitivity and specificity of miR-1246 were 71.3% and 73.9%, respectively, in distinguishing ESCC patients from healthy controls. The abundant exosomal circulating miR-1246 in serum significantly correlates with the ESCC tumor-node-metastasis stage, and it is the most potent independent risk factor for poor survival [bib_ref] Serum microRNA expression profile: miR-1246 as a novel diagnostic and prognostic biomarker..., Takeshita [/bib_ref]. Tanaka et al. [bib_ref] Clinical impact of serum exosomal microrna-21 as a clinical biomarker in human..., Tanaka [/bib_ref] analyzed the exosomes in serum from patients who suffered from ESCC and confirmed that exosomal miR-21 expression is significantly higher in the serum of ESCC patients than in that of benign controls. Exosomal miR-21 expression is related to clinicopathological stage, tumor classification, positive lymph node status and metastatic has-let-7a has-let-7f has-miR-26a has-miR-27b 2 −ΔCt EC9706 Exosomes status; therefore, miR-21 has been suggested to be a biomarker for detecting ESCC progression. In the present study, we developed a workflow for analyzing next-generation RNA sequencing of data that focus on miRNAs in human esophageal cancer cells and their corresponding exosomes. Using this technology, we detected 9,595,761 (cells) and 7,193,132 (exosomes) unique small RNAs. This result indicates the presence of a significant number of small RNA in exosomes, among which, miRNAs predominate. This workflow has revealed several trends in known and novel miRNA expression profiles for esophageal cancer-derived exosomes. A total of 80 miRNAs, including 48 known and 32 novel miRNAs, were firstly reported in esophageal cancer cell-derived exosomes. In addition, comparison with the miRBase 18.0 revealed several common miRNAs that were expressed in the cells and exosomes. The expression levels of most known miRNAs were significantly lower in the exosomes than in the cells. We also found several novel miRNAs that presented different profiles between the cells and their corresponding exosomes. Several novel miRNAs were expressed at higher levels in the exosomes than in the cells, implying that these novel miRNAs may be uniquely packaged into exosomes with specialized, but unknown functions. This phenomenon may be also caused by the fact that most studies only focused on miRNAs with high read counts in cells. The studies based on miRNAs are also generally from high to low read counts to establish a miRNA database. Some novel miRNAs are highly expressed in exosomes, but lowly expressed in cells. Taylor et al. [bib_ref] MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer, Taylor [/bib_ref] studied the miRNA expression in ovarian tumor cells and their corresponding exosomes. Among the 218 mature miRNAs that were positive in both cells and exosomes, 12 and 31 were elevated in cells and exosomes, respectively. Bellingham et al. [bib_ref] Small RNA deep sequencing reveals a distinct miRNA signature released in exosomes..., Bellingham [/bib_ref] proposed a selective mechanism for the incorporation and release of miRNAs in exosomes that contain limited or no 18S and 28S cellular ribosomal species. Not all mRNAs and miRNAs contained within cells can be directly targeted and packaged in exosomes. Therefore, the differential expression profiles of miRNAs in the cells and exosomes probably have different biological functions.
In the present study, has-miR-21, has-let-7 family, miR-26a and miR-27b were abundant in EC9706 cells and their corresponding exosomes. Known miRNA species have been previously sorted into exosomes in other cell lines. Chiba et al. [bib_ref] Exosomes secreted from human colorectal cancer cell lines contain mRNAs, microRNAs and..., Chiba [/bib_ref] revealed that exosomes derived from the colorectal cancer cell lines contain miRNAs, such as miR-21, and can be delivered into recipient cells through exosomes. Moreover, studies on miR-21 and esophageal cancer revealed that miR-21 is post-transcriptionally regulated by phosphatase and tensin (PTEN) via binding to the 3'-UTR of PTEN mRNA. Consequently, PTEN inhibits tumor cell growth and invasion by blocking the PI3K/AKT pathway [bib_ref] Down-regulation of PTEN expression modulated by dysregulated miR-21 contributes to the progression..., Li [/bib_ref] [bib_ref] PI 3-kinase and cancer: Changing accents, Vogt [/bib_ref]. Yan et al. [bib_ref] Knockdown of miR-21 in human breast cancer cell lines inhibits proliferation, in..., Yan [/bib_ref] found, through locked nucleic acid silencing combined with microarray technology, that miR-21 knockdown can inhibit the growth and migration of breast cells in vitro and tumor growth in nude mice. A previous study found that the let-7 miRNA family is abundant in both the intracellular and extracellular fractions of a metastatic gastric cancer cell line (AZ-P7a); this study suggested that AZ-P7a cells release let-7 miRNAs via exosomes into the extracellular environment to maintain their oncogenesis [bib_ref] Let-7 microRNA family is selectively secreted into the extracellular environment via exosomes..., Ohshima [/bib_ref]. Kobayashi et al. [bib_ref] Ovarian cancer cell invasiveness is associated with discordant exosomal sequestration of Let-7..., Kobayashi [/bib_ref] determined that the miRNA transcripts of the let-7 family exist in both ovarian cancer cell lines and their exosomes; they established that the release and miRNA content of exosomes significantly differ between ovarian cancer cell lines and correlate with their invasive potential. Colamaio et al. [bib_ref] Let-7a down-regulation plays a role in thyroid neoplasias of follicular histotype affecting..., Colamaio [/bib_ref] demonstrated that the let-7a overexpression in the follicular thyroid carcinoma cells increases cell adhesion and reduces cell migration; however, let-7a silencing in normal rat thyroid cells induces the opposite effects through stable transfections. Has-miR-26a and has-miR-27b were both found in ovarian tumor cells and their corresponding exosomes; however, their expression levels significantly differ between the two samples [bib_ref] MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer, Taylor [/bib_ref]. Several studies on breast cancer have shown that miR-26a can inhibit cell proliferation, colony formation and migration, as well as promote apoptosis by regulating several carcinogenesis-related processes, including several mechanisms that involve the targeting of MCL-1, MTDH and EZH2 [bib_ref] miR-26a inhibits proliferation and migration of breast cancer through repression of MCL-1, Gao [/bib_ref] [bib_ref] Pathologically decreased miR-26a antagonizes apoptosis and facilitates carcinogenesis by targeting MTDH and..., Zhang [/bib_ref]. The altered expression levels of different miRNAs involved in multiple signaling transduction pathways can determine cancer development and/or progression by interacting with one another. Namwat et al. [bib_ref] Expression profiles of oncomir miR-21 and tumor suppressor let-7a in the progression..., Namwat [/bib_ref] showed that the oncogenic factor miR-21 is up-regulated, while the tumor suppressor let-7a is down-regulated in human cholangiocarcinoma tissues. The high level of IL-6 in such a chronic inflammation-related cancer possibly stimulates miR-21 expression, and RAS activation is partly correlated with let-7a down-regulation. Kida et al. [bib_ref] PPARα is regulated by miR-21 and miR-27b in human liver, Kida [/bib_ref] found that the PPARα protein level in the human hepatocellular carcinoma cell line HuH7 significantly decreases when the levels of miR-21 or miR-27b are over-expressed and inhibited, respectively. This result suggests that the two miRNAs may serve as important regulators for the fatty acid catabolism of liver cells. Another study on miRNAs in oral squamous cell carcinoma (OSCC) reported that increasing miR-125b and miR-100 levels reduces cell proliferation; however, co-transfecting the two miRNAs significantly influences proliferation than transfecting OSCC cells individually. Researchers proposed that the reduced proliferation observed in cells co-transfected with miR-125b and miR-100 is caused by the additive effect of the two miRNAs on gene expression rather than a synergistic change [bib_ref] Decreased expression of miR-125b and miR-100 in oral cancer cells contributes to..., Henson [/bib_ref]. miRNA expression profiles were previously confined to tissues and cells; today, circulating miRNAs reportedly exist in the peripheral blood with exosomes as carriers. The effective delivery of miRNAs via exosomes in the bloodstream has provided a new mechanism for cell-to-cell communication. Investigating the mechanisms by which these candidate miRNAs carried by exosomes function in esophageal cancer development is important.
Our study revealed a significant number of isomiRs derived from almost all detected miRNAs. In total, 17,111 known and 5212 novel isomiRs were detected in our two libraries. Most isomiRs showed variability at their 5' and/or 3' ends, likely resulting from the variability in either Dicer or Drosha cleavage positions within the pre-miRNA hairpin; an example of which is given in [fig_ref] Figure 8: Expression profiles of five miRNAs in cells and exosomes [/fig_ref] , which shows sequenced isomiRs for miR-21 with high read counts in cells and exosomes. This result indicates that the majority of esophageal cancer-derived miRNA nucleotide variants resulted from post-transcriptional modifications. IsomiRs have several biological functions, such as the amelioration of the signal-to-noise ratio in miRNA-mRNA communication by targeting miRNA-controlled genetic networks. A close correlation exists between the expression characteristics of isomiRs and canonical miRNAs. These isomiRs are functional, and their biological roles likely involve the improvement of the signal-to-noise ratio in miRNA-mRNA communication by targeting miRNA-controlled genetic networks [bib_ref] MicroRNAs and their isomiRs function cooperatively to target common biological pathways, Cloonan [/bib_ref]. This finding suggests that isomiRs may have various functions. However, further studies need to be conducted for verification.
Recent studies have suggested that tumor-derived exosomes present a novel mechanism for intercellular communication [bib_ref] Exosomes/microvesicles as a mechanism of cell-to-cell communication, Camussi [/bib_ref] [bib_ref] Secretory mechanisms and intercellular transfer of microRNAs in living cells, Kosaka [/bib_ref]. The miRNA profiles of circulating exosomes from the body fluids of patients with tumors have also been proposed to be used as signatures for disease diagnosis. Furthermore, small non-coding RNA signatures exist in microvesicles isolated from glioblastoma multiforme (GBM) patient serum. A study has shown that exosomal miRNA expression levels are significantly higher in 25 patients than in age-and sex-matched healthy controls [bib_ref] A small noncoding RNA signature found in exosomes of GBM patient serum..., Manterola [/bib_ref]. A study on pancreatic adenocarcinoma (PC) revealed that serum exosomal miR-17-5p and miR-21 levels are significantly higher in PC patients than in healthy participants and non-primary carcinoma patients; this result suggests that serum exosomal miRNAs may serve as potential biomarkers of PC [bib_ref] Analysis of serum exosomal microRNAs and clinicopathologic features of patients with pancreatic..., Que [/bib_ref]. Furthermore, a study on exosomal microRNA in lung cancer indicated that a significant difference in total exosome and exosomal miRNA levels exists between lung cancer patients and controls; this study also noted that exosomal miRNA profiling has the potential to be used as a screening tool for cancer detection because of its minimal invasiveness [bib_ref] Exosomal microRNA: A diagnostic marker for lung cancer, Rabinowits [/bib_ref].
This study is the first to report that exosomes released from esophageal cancer cells contain a series of miRNAs and that the profiles of several exosomal miRNAs resemble those of their parent cells. However, several novel miRNAs in exosomes still need to be investigated. In conclusion, the discovery of these known and novel miRNAs using Solexa high-throughput sequencing provided new insights into the expression profiles and distribution characteristics of miRNAs in esophageal cancer-derived exosomes and afforded significant values to further explore miRNAs in intercellular communication mediated by exosomes in esophageal cancer.
## Experimental section
## Cell culture
The human esophageal cancer cell line EC9706 was purchased from Shanghai Tiancheng Science and Technology Co., Ltd. EC9706 cells were cultured in RPMI-1640 medium (Hyclone, Logan, UT, USA) supplemented with 10% heat-inactivated fetal bovine serum (Hyclone), 100 U/mL penicillin and 100 U/mL streptomycin under a humidified 5% CO 2 atmosphere at 37 °C. The cells were centrifuged at 10,000× g for 30 min to remove exosomes in fetal bovine serum and then ultracentrifuged (Beckman Coulter, Brea, CA, USA) at 200,000× g for 6 h to remove bovine-derived exosomes.
## Exosome isolation
EC9706 cells (40 mL) at a density of 1 × 10 8 cells per 175-cm 2 flask were cultured in complete RPMI-1640 medium at 37 °C and 5% CO 2 . After 48 h, exosomes were isolated and purified from the culture medium of EC9706 using a sequential centrifugation protocol. Briefly, the culture medium was collected and centrifuged at 300× g for 10 min, 800× g for 10 min, 1200× g for 20 min and 10,000× g for 30 min to remove lifted cells and cellular debris. The supernatant was then ultracentrifuged at 100,000× g for 3 h using a 70Ti rotor (Beckman Coulter) to pelletize the exosomes. The exosome pellets were washed with filtered phosphate-buffered saline (PBS) and then recentrifuged at 100,000× g for 2 h. The supernatant was removed, and the final exosomal pellet was resuspended in 100 μL of PBS. All centrifugation steps were conducted at 4 °C.
## Transmission electron microscopy
A 20-μL aliquot of the suspension was loaded onto a carbon-coated grid for 2 min at room temperature. The grid was positioned with the coating side facing the drop containing exosomes.
The samples were fixed by covering the grid with 10 μL of 1% aqueous phosphotungstic acid for 1 min and then observed under a transmission electron microscope (Hitachi, Shiga, Japan).
## Western blot
Samples of cells and exosomes were washed and resuspended in RIPA lysis buffer (Bi Yuntian, Hefei, China) with protease inhibitor mixture. After centrifuging at 14,000× g and for 5 min at 4 °C, the supernatant was collected, and the protein concentration was determined using a BCA Protein Assay Kit (Generay, Shanghai, China). Up to 67 μg of proteins were denatured by boiling in sodium dodecyl sulfate (SDS) loading buffer, loaded onto SDS-polyacrylamide gels and then transferred onto polyvinylidene difluoride membranes (0.45 μm pore size, Millipore, Billerica, MA, USA). The blots were blocked overnight with 5% non-fat dry milk in Tris-buffered saline containing Tween 20. After blocking, the membranes were incubated with primary mouse monoclonal antibody anti-CD63 (1:1000) (sc-51662, Santa Cruz, CA, USA) or mouse anti-β-actin (1:400, internal standard) (BM0627, Boster, Wuhan, China) in TBST for 1.5 h at room temperature. The membranes were washed six times for 5 min each before incubating with the secondary antibody. The secondary antibody conjugated with horseradish peroxidase was incubated at a dilution ratio of 1:5000 for 1 h at room temperature. Immunoreactive bands were visualized using enhanced chemiluminescence (Thermo, Pittsburgh, PA, USA) and Image Scanner III (EPSON, Tokyo, Japan).
## Rna isolation and analysis
Total RNAs from exosomes and cultured cells were isolated using the mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA) according to the manufacturer's total RNA isolation procedure. The RNA quality and concentration were assessed with a 260/280 ratio using a NanoDrop spectrophotometer (NanoDrop ND-1000 Technologies, Inc., Wilmington, DE, USA).
## Small rna libraries construction and solexa sequencing
Approximately 1 μg of total RNA from each sample was used to prepare the miRNA sequencing library. The 3'-adapter was ligated to the isolated small RNAs using T4 RNA ligase (Epicentre, San Diego, CA, USA). The 5'-adapter was ligated to the isolated small RNAs using T4 RNA ligase. The adapter-ligated sRNAs were used as templates for RT-PCR to amplify single-stranded cDNA templates to double-stranded cDNAs. Then, ~110-130-bp PCR-amplified fragments (correspond to ~15-32 nt small RNAs) from the 15% denaturation PAGE gels were extracted and then purified. After the completed libraries were quantified using Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA), the DNA fragments in the libraries were denatured with 0.1 M NaOH to generate single-stranded DNA molecules, captured using Illumina flow cells, amplified in situ and then finally sequenced for 36 cycles using a Genome Analyzer IIx (Illumina, San Diego, CA, USA) according to the manufacturer's instructions.
## Bioinformation analysis of small rna data
After generating the sequence reads from the Genome Analyzer IIx, image analysis and base calling were performed using Off-Line Basecaller software (OLB V1.8.0, Illumina, San Diego, CA, USA). Subsequently, 3' adapter sequences were trimmed from clean reads (reads that passed Solexa CHASTITY quality filter), and reads shorter than 15 nt were discarded. After removing the vector sequences, the modified sequences from 15-32 nt were counted and used for further analysis. The length distribution and common/specific clean reads of cells and exosomes were then analyzed. In addition, all unique sequences were used to search the ncRNA data Rfam, (European Molecular Biology Laboratory, Heidelberg, Germany) for the removal of non-miRNA sequences (rRNA, tRNA and snRNA). Clean reads that matched these sequences were excluded from further analysis. Subsequently, the unique small RNA sequences were used to perform a BLASTN search against the miRNA database miRBase 18.0 and to identify the conserved miRNAs in cells and exosomes. The unannotated clean reads were used to predict novel miRNAs and to analyze the single nucleotide variants of known miRNAs.
## Expression patterns of known and novel mirnas
The clean reads were analyzed against the Rfam database using BLAST. The count of clean reads that originated from each miRNA can indicate the expression level of the corresponding miRNA. To understand the differential expression of conserved and novel miRNAs between cells and exosomes, the frequency of each miRNA was normalized to the same order of magnitude using the formula: normalized expression = actual miRNA count/total count of clean reads × 1,000,000 [bib_ref] Identification and characterization of microRNAs involved in growth of blunt snout bream..., Yi [/bib_ref]. If the normalized expression level of an miRNA had read counts <2, this miRNA was removed from future differential expression analyses. Only the perfectly matched sequences were considered to be conserved miRNAs. Furthermore, the prediction of novel miRNAs of cells and exosomes was conducted using miRDeep 2 software.
## Mirna variants
IsomiRs are mature variants of miRNAs from their miRBase reference sequences. The Solexa sequencing results in the present study revealed that most of the identified miRNAs had length and sequence heterogeneity. The considerable degree of isomiR variability can be attributed to the variability in either Dicer or Drosha cleavage positions within the pre-miRNA hairpin . To characterize the isomiR variability, sequences that matched the miRNA precursors in the mature miRNA region ±4 nt (not more than one mismatch) were considered as mature miRNA isomiRs and were grouped according to the 5' (5p) or 3' (3p) arm of the precursor hairpin.
## Measurement of mirna expression level by qrt-pcr
Total RNA was isolated from exosomes and cultured cells using a mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA). qRT-PCR was performed using SYBR assay according to the manufacturer's instructions. Each reaction mixture comprised 1 μL of RT-primer, 2.5 μL of 5× reverse-transcription buffer, 1 μL of 2.5 mM dNTPs, 0.25 μL of RNase-Inhibitor, 0.25 μL of MMLV reverse transcriptase (Promega, Madison, WI, USA) and 5 ng of RNA (exosomes) or 500 ng of RNA (cells). ddH 2 O was added to a mixed volume of 12.5 μL. The RT reactions were conducted under the following conditions: 42 °C for 60 min, 70 °C for 10 min and then held at 4 °C. For the exosomes, the total RNA content was extremely low. The RT product was pre-amplified prior to the RT-PCR step to enhance sensitivity [bib_ref] Analysis of circulating microRNA biomarkers in plasma and serum using quantitative reverse..., Kroh [/bib_ref]. The reaction product was then setup for RT-PCR with the ABI PRISM 7300 real-time PCR system (Applied Biosystems, Carlsbad, CA, USA). Briefly, each reaction was performed with a final volume of 10 μL containing 1 μL of DNA, 3.75 μL of SYBR Green, 0.75 μL of Plus, 1 μL of 5 µM forward primer, 1 μL of 5 μM reverse primer and 2.5 μL of ddH 2 O. The reaction mixtures were incubated at 95 °C for 5 min, followed by 40 cycles of 95 °C for 15 s, 60 °C for 30 s and 72 °C for 30 s. Fluorescence measurements were then conducted. To calculate the relative fold change values, the C t value data were normalized to U6 and cel-miR-39 as the endogenous controls for cells and exosomes, respectively. The relative quantification (fold change) for the miRNA expression of the host cells and exosomes was determined using 2 −ΔCt (where ΔC t = (C t of miRNA of interest) − (C t of endogenous control gene (U6 or cel-miR-39)). Three technological replicates were used to ensure the reliability of the quantitative analysis. The melting curves in each experiment were analyzed to examine the sensitivity and specificity of the primers. . The predicted structure of pre-miRNA is presented as a graph. The experimentally inferred Drosha and Dicer cleavage positions are indicated with red and blue arrows, respectively. Long arrows represent cleavage sites for the most abundant isomiR, and short arrows indicate other isomiRs.
# Conclusions
This study explored the miRNA expression profile and distribution characteristics of exosomes derived from human esophageal cancer cells using the next-generation RNA sequencing technology. These findings provided new insights into the characteristics of miRNAs in exosomes derived from human esophageal cancer cells. These specific miRNAs expressed in exosomes may play important roles in intercellular communication mediated in esophageal cancer.
[fig] Figure 1: Characterization of exosomes through transmission electron microscopy and Western blot. (A) Exosomes isolated from the culture supernatant of EC9706 cells have a saucer-like shape that was limited by a lipid bilayer and diameters ranging from ~30-60 nm. Black arrowheads point to exosomes. Scale bar = 100 nm; (B) Both exosomes and cells are positive for CD63, as determined through Western blot analysis. β-actin was used as a positive control. [/fig]
[fig] Figure 2: Length distribution of small RNAs in (A) cells and (B) exosomes. [/fig]
[fig] Figure 3: Length distribution of mature miRNAs in cells and exosomes compared with known human miRNAs in miRBase 18.0. [/fig]
[fig] Figure 4: Expression abundance of known miRNA families in cells and exosomes. [/fig]
[fig] Figure 6: Profiling of miRNAs isolated from EC9706 and their corresponding exosomes. (A) Pie chart of detected known miRNAs in cells and exosomes; and (B) Venn diagram of detected novel miRNAs in cells and exosomes. [/fig]
[fig] Figure 8: Expression profiles of five miRNAs in cells and exosomes. Total RNA isolated from cells and exosomes were analyzed through qRT-PCR. ∆C t values normalized against RNU6B (cells) and cel-miR-39 (exosomes) levels were calculated, representing the relative expression levels of five target miRNAs. [/fig]
[table] Table 1: Summary of high-throughput sequencing data detected in cells and exosomes. [/table]
[table] Table 2: Common transcripts in intracellular samples in the high category of more than 10,000 transcripts. [/table]
[table] Table 3: Common transcripts in extracellular samples that belong to the mid-range category with five to 100 transcripts. No transcripts with counts higher than 100 were observed in the extracellular space. [/table]
[table] Table 4: Top 10 novel miRNAs expressed in exosome libraries. [/table]
[table] Table 5: Candidate miRNAs with only a single locus. +, positive-sense strand in genomic DNA; −, antisense strand in genomic DNA. [/table]
[table] Table 6: Candidate miRNAs with multiple loci. [/table]
|
The assessment of in vivo somatic mutations in survivors of childhood malignancy.
The assessment of chromosomal mutations in children may provide information about aetiology and risk of second malignancies. A somatic cell mutation assay which determines variant erythrocytes lacking expression of an allelic form of the sialoglycoprotein, glycophorin A, was applied to samples from children before and after receiving potentially genotoxic therapy. Fifty-six children who had received treatment for their malignancy, 15 with malignancy but prior to treatment and 43 control children were assessed for the presence of N0 and NN mutant variant red cells. Control children had mean (s.d.) N0 and NN variant frequencies (Vf) of 9.5 (7.0) and 5.8 (3.3) x 106 erythrocytes respectively. Comparison between pre-treatment
and control groups demonstrated that prior to chemotherapy, patients with paediatric malignancy do not have mutant frequencies significantly different from the normal population. Children who had received chemotherapy, with or without radiotherapy, showed a significant elevation of both N0 and NN variants over 10 years from the end of treatment. Exposure of children to radiotherapy or known chemical mutagens leads to an increased frequency of variant erythrocytes which is probably the result of in vivo somatic cell mutations. The long term implications have yet to be determined.
The improved survival of children with malignancies over the last 30 years is clearly the result of more intensive treatment schedules. Radiotherapy and many chemotherapeutic drugs, however, are known to be genotoxic and chromosomal damage in treated patients has been documented [bib_ref] Sister chromatid exchanges and chromosome aberrations in children after treatment for malignant..., Haglund [/bib_ref] [bib_ref] Cytogenetic studies of long-term survivors of childhood acute lymphoblastic leukemia, Robison [/bib_ref]. An increased risk of second malignancies in treated patients has also been reported [bib_ref] Incidence of second primary tumours among childhood cancer survivors, Hawkins [/bib_ref] and may be related to the increased number of mutations, although the relationship between mutagenesis and carcinogenesis remains to be clarified. Survivors of childhood malignancy have a longer life span than their adult counterparts and the lifetime risk of second malignancy is therefore greater. Assessment of chromosomal damage may help to predict those children at an increased risk of second malignancy following treatment. Current methods of measuring somatic mutations, such as the HPRT assay, require large volumes of blood (20 ml) and are therefore difficult to apply in paediatric practice. The glycoprotein A (GPA) assay requires small amounts of blood (100 pl), provides rapid results (within 36 h) and does not require immediate preparation of samples. Consequently it is ideal for assessing mutations in the paediatric population or where samples have to be transported from peripheral centres. The principles of the assay have been described previously but briefly it enumerates variant red cells which have undergone phenotypic change of the M and N antigens resident on the glycoprotein molecule. Heterozygotes, who express both M and N antigens can be assessed for the absence of one or other antigen (N0 or M0 variant cells) or their reduplication (NN or MM variant cells) on some red cells. Fluorescently labelled monoclonal antibodies are used to visualise these antigens and a flow cytometer to determine the variant cell frequency (Vf). This provides an assessment of chromosomal damage in the erythroid stem cells at the GPA locus -4q28-q31 . The assay has been applied to normal adults and the mean variant frequencies established as N0 = 6.2 and NN = 8.7 per million red cells . Adults with a known increased risk of developing malignancy have been assessed with the GPA assay survivors of the atomic bomb , individuals with Bloom's syndrome and ataxia telangiectasia ) -and have been found to have raised Vfs. Adult patients receiving chemotherapy have also been studied and demonstrate Vfs which rise through treatment, plateau and then fall back into the normal range. The current study used the GPA assay to evaluate the genotoxic effects of radiation and chemotherapy in the paediatric age group.
# Materials and methods
Patients and samples Samples of blood for both control and study groups were obtained from children attending the Royal Hospital for Sick Children, Bristol or regionally supported clinics. Forty-three children (mean age 6.3 years; range 0.9-16.4 years) attending the hospital for routine surgery provided control samples. In 15 of the patient group, the samples were obtained before chemotherapy was started (mean age 5.4 years; range 1.6 -16.6 years) and these children are referred to as the pretreatment group. Fifty-six patients with paediatric malignancies (mean age 13.4 years; range 2.2-32.9 years) provided samples 1 to 16.5 years after diagnosis. These children, who were aged 0.93 to 16.58 years at the time of diagnosis, formed the post-treatment group. No child in this part of the study provided more than one sample. The study was approved by the Bristol and Weston Health Authority Ethical Committee.
The criteria for selection of the patients were that they were MN phenotypes, that they had not received any blood product transfusions within the preceding 160 days and that those in the post-treatment group had received radiotherapy and chemotherapy containing at least one mutagenic drug.
Children currently receiving intensive courses of chemotherapy were invariably excluded from the study as they were usually dependent upon transfused blood products. The children who had received treatment had the following maligancies (numbers in parentheses): leukaemia (22), lymphoma (eight), bone tumours (six), renal tumours (five), rhabdomyosarcomas (six), undifferentiated round cell tumours (four), neuroblastoma (three) and malignant teratoma (one). The following tumours were found in the children who formed the pre-treatment group: leukaemia (eight), bone tumours (two), renal tumours (two), rhabdomyosarcomas (one), malignant teratoma (one) astrocytoma (one) and malignant nerve sheath tumour (one). The chemotherapeutic drugs included: actinomycin D, asparaginase, doxorubicin, daunorubicin, epirubicin, bleomycin, etoposide, cyclophosphamide, carboplatin, cisplatin, cytosine, ifosfamide, methotrexate, prednisolone, procarbazine, chlorambucil, mercaptopurine, thioguanine and vincristine. Thirty-six of the 56 patients received radiotherapy as part of their treatment protocol. Preparation and assay of samples Samples were collected in EDTA tubes and were stored at 4°C for up to 1 week before being prepared. MN status was confirmed using commercial anti-M and anti-N typing sera (Biotest). The method of preparing the erythrocytes has been described but briefly 100 ftl of donor cells are fixed in 10% formalin and made spherical with a 0.005% solution of sodium dodecylsulphate (SDS). The Nspecific antibody, BRIC157, was a gift of Dr D. Anstee, Blood Group Reference Laboratory, Bristol and was purified by chromatography over Protein G Sepharose (Pharmacia) before beign labelled with fluorescein (Sigma) (abbreviated to BRIC157-). The M-specific antibody, 6A7, was a gift of Dr R. Jensen, Lawrence Livermore Laboratories, California and was supplied biotinylated (abbreviated as 6A7b). Both antibodies are now available from the Commercial Department, Blood Products Reference Laboratory, Dagger Lane, Elstree. London.
Approximately 25 x 106 erythrocytes were incubated with both antibodies at room temperature for 1 h, washed and then mixed with streptavidin-phycoerythrin (suffix-PE) (Vecta Labs, 5il ml1'). Propidium iodide (Sigma, 10llml-') was later added to exclude any nucleated cells from analysis by the cytometer. All samples were analysed on a BD Facscan within 12 h of labelling. Control mixtures of red cells were used on each run to establish standard gain and compensation settings and to set the variant cell windows. Cells were analysed at a rate of 3,000-4,000 per second. Only N0 and NN variants can be assessed using the BRIC157/6A7 antibody combination . For each sample, two runs of 106 cells were analysed and the Vfs (N01, NN1; N02, NN2) determined by dividing the number of events in the respective variant cell window by the total number of events in the flow distribution. Mean N0 and NN values (mN0 and mNN) were calculated from these two runs. At least one control was included with each batch of samples run on the cytometer.
# Statistical analysis
The data from the patient samples was not normally distributed and therefore a non-parametric analysis comparing the three groups was undertaken with the Kruskall-Wallis and Mann-Whitney U tests. Samples were grouped according to time from the end of chemotherapy and comparison with the control group was undertaken to assess persisting elevation from the normal range over time. Comparison of variant frequences between patients who had or had not received radiotherapy was also performed using the Mann-Whitney-U test.
# Results
The mean Vfs (and s.d.) for the control group were N0 9.5 (7.0) and NN 5.8 (3.3) and an example of a scattergram from one such patient is shown in [fig_ref] Figure 1: Bivariate scattergram from normal control [/fig_ref]. Linear regression analysis of the N0 Vfs against age in the control group did not suggest a positive relationship in this group of children (data not shown). The results for the three groups, along with inter-group statistical analyses, are summarised in [fig_ref] Table I: Summary statistics for the three study groups Values represent variant frequencies x... [/fig_ref].
Both mN0 and mNN variant values were plotted against time elapsed from the end of chemotherapy [fig_ref] Figure 2: Distribution of N0 variant frequencies against time elapsed from the end of... [/fig_ref]. Variant frequencies rose through the latter parts of some of the prolonged courses of chemotherapy but then fell towards normal over the subsequent years. Comparison between groups of patients and the control group showed that there was a significant, and clearly, persisting elevation of both N0 and NN variants over 10 years from from the end of treatment (P = 0.01 and 0.02 respectively), [fig_ref] Table I: Summary statistics for the three study groups Values represent variant frequencies x... [/fig_ref]. A detailed assessment of the whole group for the possible mutagenic effect of individual drugs was not possible due to the large number, the range of doses and the variety of drug combinations used. It is hoped, however, that as more patients are analysed this type of information may become available. Comparison of Vfs between those patients who had received radiotherapy and those who had not showed no statistical difference between age at diagnosis, age at providing sample and variant frequency.
Four children provided samples whilst still receiving oral chemotherapy. All had acute lymphoblastic leukaemia (ALL) and were being treated on the UK ALL (X) protocol and, consequently, were receiving maintenance treatment of daily 6-mercaptopurine, weekly methotrexate and thrice weekly co-trimoxazole (trimethoprim/sulphamethoxazole). Three of these patients showed very high Vfs on numerous occasions, with characteristic bivariate scattergrams . This scattergram pattern was similar to that seen in a patient with ataxia telangiectasia and a non Hodgkin's lymphoma [fig_ref] Figure 5: Bivariate scattergram of patient with ataxia telangiectasia [/fig_ref] who was not included in the main study group.
# Discussion
Although the precise relationship between mutagenesis and carcinogenesis remains unclear there is much circumstantial Bivariate scattergram from patient with high Vfs. Labelling details as for [fig_ref] Figure 1: Bivariate scattergram from normal control [/fig_ref]. In this sample there is a tail of partially labelled variants extending from the main MN peak down to the N0 variant window. There is also a small tail running parallel to the x-axis from the MN peak which would represent partially labelled M0 variants. The patient consistently produced this pattern with repeated analysis. He relapsed on treatment and died shortly afterwards. evidence to support the two are related [bib_ref] The molecules of cancer, Sikora [/bib_ref]. Assessment of mutations, therefore, may help clarify aetiological factors in the development of tumours. The outlook for children with malignancy has improved substantially but the risk of developing second tumours has also become significant. It is likely that this increased incidence is related to the use of radiation and high dose chemotherapy during the primary treatment and is presumably mediated through the induction of mutations. Monitoring mutations in the years after treatment may help to clarify this point and determine'those children at increased risk of further malignancies.
Like most other methods for assessing chromosomal mutations, the GPA assay looks specifically at alterations in protein expression at one particular genetic locus and from this, the likelihood of generalised DNA damage is extrapolated. Clearly such generalised damage may not have occurred in * * * * XL 30 Anti-N 145 L : I all patients since isolated mutations, involving only a few base pairs, could be responsible for the malignant change in some cases.
In the current study, the GPA assay showed mutation frequencies for a group of normal children of around nine N0 variants and six NN variants per million red cells. These values are similar to those determined by the GPA assay and the HPRT assay [bib_ref] Longitudinal study of the in vivo HPRT mutant frequency in human T-lymphocytes..., O'neill [/bib_ref] in normal adults. The origin of N0 variants is likely to be deletion mutations involving the GPA locus on chromosome 4 whereas the homozygous variants, NN, probably arise by chromosomal missegregation or mitotic recombination [bib_ref] Expression of recessive alleles by chromosomal mechanisms in retinoblastoma, Cavenee [/bib_ref]. Both of the latter events are independent of the deletions which give rise to the hemizygous variants. Consequently the determination of both variants may ultimately provide information on different mechanisms on DNA damage.
Comparison of the pre-treatment group with controls demonstrates that before exposure to chemotherapy most patients with paediatric malignancy do not have mutant frequencies different from the normal population. This implies that they are neither more prone to spontaneous mutations nor have they been exposed to any obvious mutagens. This has also been established in adults by others [bib_ref] Increased mutation frequency following treatment with cancer chemotherapy, Dempsey [/bib_ref].
It is recognised that chromosomal mutation frequencies increase with age [bib_ref] HPRT somatic mutation data, Cole [/bib_ref] and the current study does have an age imbalance between the control and treatment groups which arose from ethical constraints of obtaining blood from normal children. One has to consider the possibility that the observed differences in variant frequencies between these groups simply reflects a difference in age composition between the groups. Analysis of the data, however, showed this not to be the case, as Vfs from control patients did not demonstrate a positive relationship with age and the high Vfs seen in many of the post treatment groups lay well outside the expected range suggested by [bib_ref] In vivo somatic mutations in the glycophorin A locus of human erythroid..., Jensen [/bib_ref].
Analysis of the post-treatment group demonstrates that exposure to radiotherapy or known chemical mutagens leads to an increased frequency of variant erythrocytes which are presumably the result of in vivo somatic cell mutations. This finding confirms those in other studies which have looked at chemotherapy induced mutations using both the GPA and the HPRT assays [bib_ref] Increased mutation frequency following treatment with cancer chemotherapy, Dempsey [/bib_ref] in adults.
The plot of Vfs against time from end of chemotherapy demonstrates that the induced high levels of mutations eventually fall back towards normal. This decay, however, is protracted and well beyond the lifespan of circulating red cells suggesting that some mutations at the GPA locus have been stably incorporated into the erythroid stem cells. The elevations persisted over 10 years from the end of treatment which seems at variance with the report of where the high Vfs induced by chemotherapy in adults fell back to the normal range within 6 months .
Persistently elevated Vfs have been reported in adults after exposure to high doses of radiation but comparison of Vfs between those children who had received radiotherapy and those who had not failed to show a significant difference between the two. An alternative explanation is that these childhood patients received substantially more intensive chemotherapy than the adults previously reported.
If the evolution of cancers is due to chromosomal mutations, then the persistence of such induced mutations many years after the end of chemotherapy may have implications for the development of second malignancies. Further long term studies, however, are necessary to determine whether the increase in mutation frequency is of any value in predicting which patients develop second tumours.
The grossly elevated Vfs observed in three of the four children who provided samples whilst still receiving chemotherapy merits further investigation. Clearly the number of patients is small but the bivariate scattergram which was observed was most striking. The long tail of cells from the main peak to the hemizygous variant loss window was repeated in all three of the high count patients. These cells with partial loss variants have been observed before [bib_ref] Detection of somatic mutations at the glycophorin A locus in erythrocytes of..., Kyoizumi [/bib_ref]. Obviously this pattern could be artefactual and might represent interference with the attachment of the monoclonals by drug metabolites. If this was the case, however, one would expect all four patients receiving the same chemotherapy to demonstrate the same pattern. Furthermore, a generalised disruption in antibody attachment would lead to diffuse distribution of cells on the bivariate scattergram. In this small group of patients, the partial loss variants are expressing normal amounts of the N antigen and a range of M antigen, from complete expression (i.e. MN cells) to complete absence (i.e. N0 cells). The bivariate pattern seen is similar to that produced in patients with known DNA repair defect syndromes such as ataxia telangiectasia [fig_ref] Figure 5: Bivariate scattergram of patient with ataxia telangiectasia [/fig_ref] and Bloom's syndrome. Affected patients with both these conditions are susceptible to chemotherapy/radiotherapy induced DNA damage. The three patients presented here may have an increased tendency to chromosomal mutation or a decreased ability to repair DNA defects and this may have some bearing on the fact that two of the three patients with very high Vfs have relapsed either on treatment or shortly after. Clearly more detailed investigation of such patients is needed.
The patients in this study had a heterogeneous group of malignancies and consequently received a wide variety of chemotherapeutic agents. This must obviously influence the interpretation of the comparative data. The study does, however, show that chemotherapy with or without radiotherapy given to children with malignant disease leads to DNA mutations. In some patients these mutations persist for many years after the end of treatment and this may influence the long term prognosis.
[fig] Figure 1: Bivariate scattergram from normal control. Fixed erythrocytes from the donor are labelled with BRIC157-f (anti-N, FLI) and 6A7b-PE (anti-M, FL2) and one million events analysed and shown in the scattergram. Axes represent 64 linear channels equally distributed over a four decade range of fluorescent intensity. The main peak of MN cells has channel coordinates of x = 42 and y = 46. [/fig]
[fig] Figure 2: Distribution of N0 variant frequencies against time elapsed from the end of chemotherapy. Dotted line marks the end of therapy. Vfs for control A, and pre-treatment B, groups are shown to the left of the dotted line. [/fig]
[fig] Figure 3: Distribution of NN variant frequencies against time elapsed from the end of chemotherapy. Annotations as for Figure 2. [/fig]
[fig] Figure 5: Bivariate scattergram of patient with ataxia telangiectasia. This sample from a 10 year old boy with ataxia telangiectasia shows two long tails of variant erythrocytes arising from the main MN peak and running towards the N0 and M0 variant windows. These patients have very high variant cell frequencies. [/fig]
[table] Table I: Summary statistics for the three study groups Values represent variant frequencies x 106 red blood cells. aMean Vfs for each donor were used to determine range. bMann Whitney U test. [/table]
|
Subsets and Splits