Elevated miR‐124‐3p in the aging colon disrupts mucus barrier and increases susceptibility to colitis by targeting T‐synthase

Abstract The risk of colitis and colorectal cancer increases markedly throughout adult life, endangering the health and lives of elderly individuals. Previous studies have proposed that bacterial translocation and infection are the main risk factors for these diseases. Therefore, in the present study, we aimed to identify the underlying mechanism by focusing on the mucus barrier function and mucin‐type O‐glycosylation. We evaluated alterations in the colon mucus layer in 2‐, 16‐, and 24‐month‐old mice and aged humans. Aged colons showed defective intestinal mucosal barrier and changed mucus properties. The miR‐124‐3p expression level was significantly increased in the aged distal colonic mucosa, which was accompanied by an increase in pathogens and bacterial translocation. Meanwhile, T‐synthase, the rate‐limiting enzyme in O‐glycosylation, displayed an age‐related decline in protein expression. Further experiments indicated that miR‐124‐3p modulated O‐glycosylation by directly targeting T‐synthase. Moreover, young mice overexpressing miR‐124‐3p exhibited abnormal glycosylation, early‐onset, and more severe colitis. These data suggest that miR‐124‐3p predisposes to senile colitis by reducing T‐synthase, and the miR‐124‐3p/T‐synthase/O‐glycans axis plays an essential role in maintaining the physiochemical properties of colonic mucus and intestinal homeostasis.

Crohn's disease (CD), has evolved into a global disease, with rising incidence and prevalence on every continent (Ashwin et al., 2020;Chi, 2016;Kaplan, 2015). As the global population ages, the number of older people living with IBD increases markedly, and these individuals often suffer from a high risk of developing colorectal cancer (Ha & Katz, 2014). The precise etiology of IBD remains elusive, but gut dysfunction is strongly correlated with the pathogenesis of IBD.
Our recent work  demonstrated that the number of interstitial cells of Cajal and enteric neurons in the entire digestive tract is reduced with increasing age, which might account for GI motility malfunction and attenuated digestion and absorption functions. Further experiments indicated that the age-related changes could be due to the inflammatory state in their microenvironment , suggesting that aging is correlated with the genesis and development of IBD and associated tumors. Mounting evidence has shown that the breakdown of intestinal homeostasis, which depends on the intestinal mucosal barrier and dynamic crosstalk among the intestinal microbiota, intestinal epithelial cells, and the host immune system, majorly contributes to the pathogenesis of inflammation (Maloy & Powrie, 2011;Sartor & Wu, 2017;Shin & Kim, 2018); however, the underlying mechanism is still not fully clarified.
The intestinal mucosal barrier integrity is essential for maintaining normal intestinal homeostasis. Physiologically, the thick mucus covering the epithelium, which is synthesized by goblet cells, is placed at the center of interactions between the epithelial immune system and luminal pathogens and is the primary defensive layer separating microbiota from epithelial cells. In a healthy colon, mucus comprises two layers; the dense inner layer is virtually free of bacteria and prevents bacteria from penetrating due to acting as a size-exclusion filter, while the loose outer layer is heavily colonized by the gut microbiota (Johansson & Hansson, 2016). Gel-forming mucin2 (MUC2), which is the major building block of mucus, is a large glycoprotein characterized by abundant and variable O-linked oligosaccharides (O-glycans), which account for up to 80% of the mass of mucin molecules and are responsible for the gel-like properties of mucus (Fu et al., 2011;Johansson et al., 2008). Recent studies have revealed that the biosynthesis of mucin-type O-glycans is initiated by polypeptide α-N-acetylgalactosaminyltransferase (ppGal-NAcT), which transfers GalNAc to specific serine or threonine (Ser/ Thr) residues to generate intermediate Thomsen-nouvelle (Tn) antigen (GalNAcα1-Ser/Thr), which is then normally extended to form the main type of O-glycans, core 1-derived structures (also called Thomsen-Friedenreich antigen or T antigen). T-synthase (core 1 β1,3-galactosyltransferase) is the key enzyme in the process of core 1-derived glycans, and the active form of T-synthase requires the co-expression of the unique molecular chaperone Cosmc (Fu et al., 2011;Ju & Cummings, 2002). Therefore, factors regulating the expression of T-synthase and Cosmc are of interest because they can lead to an inability to synthesize core-1 O-glycan and the deposition of Tn antigens, which potentially contribute to several autoimmune diseases, including IgA nephropathy and Tn syndrome (Berger, 1999;Serino et al., 2012). The Tn antigen is also recognized as a tumor-associated antigen in pancreatic cancer and colitis-associated colon cancer (Bergstrom et al., 2016;Chugh et al., 2018;Hofmann et al., 2015). However, it has remained unclear whether the aberrant glycosylation of mucins is associated with the increased susceptibility of elderly individuals to microbiota dysbiosis and IBD compared to that in younger individuals.
Since the discovery of microRNAs (miRNAs), extensive research has shown that miRNA molecules are important regulators of gene expression as a part of the epigenetic machinery (Moutinho & Esteller, 2017). Clinically, individual miRNAs are dysregulated in almost all diseases, and much attention has been paid to the role of miRNAs in colitis and inflammation-related gut cancers (Tili et al., 2017). To date, great effort has focused on deciphering functions of miRNAs on validated target transcripts in the gut immune system (regulatory T cells, macrophages, inflammasomes, etc.) and the epithelial barrier (Heinsbroek et al., 2016;Neudecker et al., 2017;Wang et al., 2016;Yang et al., 2016). Nevertheless, the miRNAs implicated in regulating the properties of intestinal mucus and bacterial invasion remain obscure.
Here, to explore the relationships among miRNAs, colonic mucus properties and intestinal homeostasis during aging, we performed studies in both naturally aging mice and human tissues. We focused on determining the mucus properties in aging and the interaction between miR-124-3p and T-synthase gene. To further assess the role of miR-124-3p, we used miR-124-3p agomir-treated mice in conjunction with oral dextran sulfate sodium (DSS) administration to model colitis.

| Aged colon exhibits a defective mucosal barrier
We first evaluated the intestinal mucosal barrier function during aging. At 24 months (mo) of age, we found marked translocation of E. coli inhabiting the mucus layer to the mesenteric lymph nodes (MLNs) and liver (Figure 1a), accompanied by increased proinflammatory cytokines, including tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, and IL-6, in the colonic mucosa in both old mice and elderly population, we also observed increases in TNF-α in the mouse liver (Figure 1b-d). Moreover, aging of the distal colonic mucosa led to a prominent increase in the numbers of adherent total bacteria based on bacteria analysis, and the numbers of enteropathogenic bacteria E. coli, Helicobacter hepaticus (H. hepaticus), and Helicobacter pylori (H. pylori) increased, while the levels of genus Lactobacillus and Bifidobacterium as probiotics decreased in the elderly ( Figure 1e). Nevertheless, the epithelial microstructure was intact, and there was no epithelial damage or ulceration associated with aging ( Figure 1f). However, the mucus thickness was greatly reduced in old mice as shown by Alcian blue/periodic acid-Schiff (AB/ PAS) staining (Figure 1g, Figure S1A) and MUC2 immunofluorescence assay (Figure 1h, Figure S1B), but the number of goblet cells in the colon was unchanged with increasing age ( Figure S1C).

| Aged distal colon epithelium exhibits low levels of T-synthase and a breached inner mucus gel layer
The bacterial translocation in the aged distal colon was evaluated by fluorescence in situ hybridization (FISH). Surprisingly, the aged (24 month) colonic mucus layer allowed bacteria to penetrate. One of the 16-month-old mice also showed a defective inner layer, but the inner mucus layer was intact and sterile in all 2-month-old mice and the other four 16-mo-old mice (Figure 2a). It is known that fucose is a common capping structure on O-glycans, and immature O-glycans truncated to Tn antigen does not carry any fucose; thus, we used lectin Ulex europaeus agglutinin-1 (UEA1, binds Fucα1,2 Gal epitopes) staining to detect O-glycan structures. The reduction in UEA1 in goblet cells that occurred with aging is shown in Figure 2b and Figure S2A, suggesting the loss of O-glycans. The Tn antigen is a frequently occurring truncated immature O-glycan that expressed at high levels in many diseases. As expected, the specific expression of the Tn antigen was observed in intracellular space in aged mice, and it was barely detectable in young tissues (Figure 2c, Figure   S2B). The Western blot analysis of the Tn antigen and its sialylated form, sialyl-Tn (sTn) antigen, also showed increased expression dur-  Figure S2D). In contrast, neither the protein nor mRNA expression of Cosmc changed in the aged colon ( Figure 2g, Figure   S2E). Owing to these findings, the emphasis here is on T-synthase.

| MiR-124-3p is a direct regulator of the C1GALT1 and downregulates T-synthase expression
The T-synthase is known to be directly involved in mucin-type Oglycosylation; however, the mechanism of its decreased function in F I G U R E 1 Aged colons exhibit microbiota dysbiosis and a thinner mucus layer. (a) Real-time PCR analysis of pathogenic E. coli detected in the MLNs and liver of 24-mo-old mice but not 2-mo-old mice. (b-d) Relative mRNA fold changes or levels of TNF-α, IL-1β, and IL-6 in young and old mouse livers (b) and the distal colonic mucosa of mice (c) and humans (d). (e) Relative fold changes of common bacteria in the distal colonic mucosa in old mice compared to young mice (standardized to 18S rRNA). (f) Representative H&E staining images of the distal colon from 2-and 24-mo-old mice. (g, h) Images showing AB/PAS staining (g) and MUC2 immunofluorescence staining (h) in the distal colons of young and old mice. Data are presented as mean ± SD. ***p < 0.001 compared with 2-mo-old mouse group or young people group. Student's t test or Mann-Whitney U test for (b-d). n = 8 mice or 7 human samples per group. Abbreviations: Ms, mouse; Hu, human.
aged mucosa is still unknown. In recent years, miRNAs have attracted considerable attention as regulators of gene expression (Lewis et al., 2005); therefore, we integrated the results from public databases endowed with prediction algorithms, such as PicTar (https:// pictar.mdc-berlin.de/), TargetScan (http://www.targe tscan.org/ vert_71/), and miRanda (http://www.micro rna.org/micro rna/home. do) (Figure 3a). Among 8 overlapping presumable miRNAs based on the consensus from the three algorithms, miR-124-3p caught our attention because of its high expression in senescent skin, UC, and CD (Harada et al., 2017;Qin et al., 2017;Zhao et al., 2016). Then, we found that the expression of miR-124-3p in the distal colonic mucosa was significantly increased in old mice and elderly persons, but no noticeable alterations were observed in the proximal colonic mucosa (c) Immunofluorescence staining of 2-, 16-, and 24-mo-old mouse colon sections with anti-Tn. (d) Western blot analysis showing the high expression of Tn antigen and sTn antigen in the old mouse distal mucosa compared to young. (e) T-synthase activity in the mouse distal colon and human colon. (f, g) Relative mRNA expression of C1GALT1 (f) and COSMC (g) and Western blot analysis of T-synthase (f) and Cosmc (g) protein expression in the mouse and human colon. Data are presented as mean ± SD. *** p < 0.001 compared with 2-mo-old mouse group or young people group; ## p < 0.01, ### p < 0.001 compared with 16-mo-old mouse group. Student's t test for (e-g Ms PC and Hu), and ANOVA and LSD test for (e-g Ms DC). n = 8 mice or 7 human samples per group. Abbreviations: Ms, mouse; Hu, human; PC, proximal colon; DC, distal colon.

| Mice overexpressing miR-124-3p show aberrant O-glycans and defective mucus layer integrity
To investigate whether the overexpression of miR-124-3p affects the properties of the intestinal mucus in vivo, young mice (aged 2 mo) received tail-vein intravenous injections of miR-124-3p agomir ( Figure 4a). The expression of T-synthase was efficiently F I G U R E 3 MiR-124-3p downregulates T-synthase expression. (a) Presumable miRNAs that regulate C1GALT1 as identified from the public databases PicTar, TargetScan, and miRanda. (b) MiR-124-3p expression was higher in old mice and aged human colon. Data are presented as mean ± SD. *** p < 0.001 compared with 2-mo-old mouse group or young people group; ### p < 0.001 compared with 16-mo-old mouse group. Student's t test for (Ms PC and Hu), and ANOVA and LSD test for (Ms DC). n = 8 mice or 7 human samples per group. (c) Putative miR-124-binding sites in the C1GALT1 3'UTR based on TargetScan analysis and mutated 3'UTR sequences of C1GALT1. (d) The results of the dualluciferase reporter assay. Data are presented as mean ± SD. *** p < 0.001 compared with miR-control group (ANOVA and LSD test). n = 3 per group. (e) Real-time PCR analysis of miR-124-3p in cells transfected with miR-124-3p mimic and NC. Data are presented as mean ± SD. *** p < 0.001 compared with mimic NC group (ANOVA and LSD test). n = 3 per group. (f) Relative expression of C1GALT1, and Western blotting showing low expression of T-synthase protein in cells transfected with miR-124-3p mimic. Data are presented as mean ± SD (ANOVA and LSD test). n = 3 per group. Abbreviations: Ms, mouse; Hu, human; PC, proximal colon; DC, distal colon. FISH revealed that the distal colon epithelium in the control group was covered by an intact inner layer that was devoid of bacteria.
In contrast, the epithelium overexpressing miR-124-3p was in direct contact with bacteria, and the inner layer was less defined

| Mice overexpressing miR-124-3p show earlyonset and more severe DSS-induced colitis
We showed that the mucus barrier was defective in aged individuals because of miR-124-3p-induced aberrant O-glycosylation. To further determine whether increased miR-124-3p enhanced the susceptibility to colitis, an acute colitis mouse model was established by treatment with 5% DSS (Figure 4a). The mice pretreated with miR-124-3p agomir showed a tendency to lose more bodyweight and exhibited an increased disease activity index (DAI) score (Figure 5a,b), which was associated with shortening of the colon (Figure 5c group led to superficial inflammation in the colon, which was characterized by substantial inflammatory cellular infiltration and larger colonic mucosa collapses (Figure 5g,h, Figure S5A). Moreover, the expression of TNF-α and IL-1β in the agomir-treated mice showed an increasing tendency compared to the control group (Figure 5i).
Consistent with these findings, the protein level of T-synthase was reduced, which was consistent with the high degree of miR-124-3p expression in the distal colon in the mice treated with agomir ( Figure 5j, Figure S5B). These data together indicated that miR-124-3p markedly exacerbated DSS-induced colitis.

| DISCUSS ION
The human gut harbors trillions of microbes, which are separated from the inner milieu of the host by the intestinal barrier. The function of this barrier mainly relies on crosstalk among the following 3 components: the commensal microbiota, the mucus layer, and the intestinal epithelium. The latter two significantly affect commensal bacterial colonization, and their defects might trigger inflammation, which is highly prevalent in elderly people. Our study showed the appearance of a gut pathobiont, E. coli, in the MLNs and liver, mirroring previous findings (Manfredo Vieira et al., 2018). These results in young mice, while bacteria were notably in direct contact with the colonic epithelium in mice aged 24 months, which was accompanied by aging-associated gut dysbiosis, such as the shift toward proinflammatory commensals and reduction in beneficial microbes, as well as enhanced bacterial adhesion. These results illustrate that dysbacteriosis and "inflammaging" may be partly due to inadequacies in the barrier function of mucus layer, but the potential and detailed dynamic changes in gut mucus in aging people are still largely unexplored.
The impaired expression of MUC2 O-glycans and deposition of Tn/sTn antigen have recently been reported to cause colonic mucus barrier breach and contribute to colitis and associated cancer (Bergstrom et al., 2016(Bergstrom et al., , 2017Zhao et al., 2018). The present study showed that aged distal colons also exhibited blockade of Oglycosylation, the high level of Tn/sTn antigen, and leakiness of the F I G U R E 4 Mice overexpressing miR-124-3p show aberrant O-glycans and defective integrity of the mucus layer. (a) Mouse models used in the study. (b) High expression of miR-124-3p and low T-synthase protein expression in the proximal and distal colonic mucosa in the miR-124-3p agomir-treated mice. (c) Representative AB/PAS staining of colonic tissues from the high miR-124-3p expression and control groups.

(d) Fluorescent images showing O-glycans by using UEA1 staining. (e) FISH images showing bacterial penetration into the inner mucus layer
after miR-124-3p agomir administration. The magnified images below showing boxed regions. White arrows point to bacteria direct contact with the mucosal epithelium. (f) Microbiota dysbiosis in the distal colonic mucosa in miR-124-3p agomir-treated mice. (g) High levels of TNF-α and IL-6 in the distal colonic mucosa and liver in the agomir-treated group. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 compared with agomir NC group (Student's t test or Mann-Whitney U test). n = 6 mice per group. Abbreviations: Ms, mouse; Hu, human; PC, proximal colon; DC, distal colon.
inner mucus. It is well known that normal O-glycosylation of MUC2 relies on both the critical rate-limiting enzyme T-synthase and its molecular chaperone Cosmc (Fu et al., 2011;Ju & Cummings, 2002). Although Kudelka et al. (2016) reported that Cosmc was an IBD risk factor and was implicated in regulating dysbiosis in IBD, no changes in Cosmc expression levels were observed in aged humans or mice, resulting in the speculation that Cosmc might not be associated with inactive T-synthase during aging. Importantly, we found lower enzymatic activity and decreased protein expression of T-synthase in the aged colon. Hence, the molecular mechanism Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 compared with agomir NC group. Student's t test for (a,d,e,f,i,j), Mann-Whitney U test for (b,g). n = 5 mice per group.
behind the impaired functioning of T-synthase attracted our attention.
Accumulating evidence has shown the function of miRNAs as gene regulators and their involvement in the pathogenesis of gut diseases. In the present study, we found that miR-124-3p expression was significantly higher in the aged distal colon. Using the prediction algorithms TargetScan, PicTar, and miRanda, it was predicted that miR-124-3p might be bound to the 3'-UTR of C1GALT1.
Accordingly, we further confirmed that miR-124-3p directly targeted C1GALT1 and epigenetically downregulated its protein levels in both in vivo and in vitro experiments. As for the reason that C1GALT1expression level was also significantly decreased in the aged distal colonic mucosa, we have identified that there exist other factors downregulating T-synthase mRNA and protein expression in the aging process (unpublished data). Intriguingly, all age-related changes in T-synthase and miR-124-3p were observed exclusively in the distal colon, whereas the proximal segment was essentially unaltered. It has been documented that the proximal and distal colon differ in many respects, including embryological origin, local luminal environment, excretory function, and microbial flora (Flemer et al., 2017;Neumann et al., 2014;Pekow et al., 2015). There was a distinct regional expression pattern of miR-124-3p/T-synthase in the proximal colon versus distal colon, adding to the evidence that physiological mucus heterogeneity between the proximal and distal colon exists, and is contributed to the pathological involvement of the distal colon in age-associated diseases. Our results highlight a possible reason for this finding: The level of T-synthase fell sharply in the aged distal colon, where numerous microorganisms live, leading to impaired glycosylation of mucins.
Importantly, the present study also serves as an additional affirmation of the relationship between miR-124-3p and the penetra- This viewpoint is also supported by recent literature showing that miR-124 might aggravate colitis by inhibiting the aryl hydrocarbon receptor in CD .
In addition to the changes in mucus properties, a thinner AB/ PAS-and MUC2-positive mucus layer was observed overlying the mucosal surface of aged mice, which was in line with the results reported by van Beek et al. (2016). Goblet cells are the main source of intestinal gel-forming mucins. Given that the number of goblet cells in the colon was unchanged both in the aged mice and in the miR-124-3p agomir-treated mice, and the latter mice showed a normal inner mucus layer in terms of thickness, we believe that miR-124-3p or T-synthase was not responsible for the thinner mucus in the aged colons, at least not directly. Our previous study indicated that SCF/c-KIT signaling promoted intestinal mucus secretion and development of mucinous colorectal adenocarcinoma by activating PKCδ-MARCKS , therefore, future studies of this mechanism underlying the reduction in mucus thickness during aging are necessary. A previous study on the thickness of the mucus gel layer focused on colitis indicated that mucus thickness alone was not a useful indicator of mucus barrier function (Johansson et al., 2014).
But recent research (Ahmadi et al., 2020) demonstrated that metformin increased goblet cell mass and mucin production in the older obese mouse gut, which in turn ameliorated aging-related leaky gut and inflammation. Whatever the case is, the thinner mucus layer together with the abnormal mucus properties would make it considerably easier for microbes, food antigens, and toxins to attach to the epithelium. The long-term and direct contact between bacteria and the mucosal epithelium might prompt some microbial species to establish themselves as residents on the surface of the mucosa and further weaken gut function during aging.
In conclusion, in the present study, the actual linkage between miR-124-3p and intestinal homeostasis is proposed. We demonstrated that miR-124-3p expression was significantly increased along with gut bacterial translocation in the distal colon in aged mice and people. Furthermore, we identified miR-124-3p as a negative regulator of C1GALT1 and discovered that the miR-124-3p/T-

| Animals
Male C57BL/6 mice were purchased from Capital Medical University and divided into three groups according to age, 2, 16, and 24 mo old (n = 8 per group), to simulate the life phases at approximately 20, 50, and 70 years of age in humans (Flurkey, 2009).
An additional 22 male mice (aged 2 months) were used to develop the animal model. All mice were sacrificed by appropriate anesthesia and/or cervical dislocation prior to sample collection. The colon, liver, and MLNs were surgically and aseptically removed from the mice. All animal procedures were carried out under protocols approved by the Animal Care and Use Committee of Capital

| Human tissue samples
The clinical tissue samples, including 14 normal mucosa tissue samples locating at descending colon from different colorectal cancer patients, were collected immediately after surgical resection prior

| Administration of miR-124-3p agomir
Young (2-mo-old) mice were randomly assigned to two groups (n = 6 per group) and injected via the tail vein with either miR- RiboBio,China) or agomir NC (Cat.No.miR04101-1-10) at a dose of 5 nmol/day for 2 consecutive days. Mice were sacrificed at 96 h after the last injection, and tissues were collected for further analyses (Figure 4a).

| Model of DSS-induced colitis
The procedure used to induce the DSS-colitis model is shown in Figure 4a. Briefly, young (2-mo-old) mice received tail-vein injections of miR-124-3p agomir or NC (n = 5 per group) daily for 2 consecutive days. This treatment was applied twice at a 4-day interval to maintain high expression of miR-124-3p. Simultaneously, acute colitis was induced by the administration of 5% DSS (Cat.No.0216011080, MP Biomedicals, USA) in drinking water continuously for 4 days, and the mice were euthanized on day 8. The severity of colitis was assessed by body weight, disease activity index (DAI), and histological damage scores, as previously described with slight modification (Table 1).
The assessment of the disease severity in the mice with DSS-induced colitis is described in the Supplementary experimental procedures.
TA B L E 1 Assessment of disease severity in DSS-colitis mice DAI scoring system.

Score
Diarrheal stool score Bloody stool score Note 0 Normal stool Normal colored stool The sum of the scores of two parameters was defined as the DAI score. The scoring system ranged from 0 to 6, with 6 indicating the worst injury. The sum of the scores of three parameters was defined as the mucosal damage score. The scoring system ranged from 0 to 10, with 10 indicating the worst injury.

| T-synthase activity assay
The assessment of the T-synthase activity was conducted using a sensitive fluorescence assay method (Ju et al., 2011). In brief, tissue extracts were prepared using homogenization buffer (

| Data analysis
Images were acquired with a fluorescence microscope (Nikon Ni, Japan) or light microscope (Leica DM LB2) and analyzed using Image-Pro Plus 6.0 software (Media Cybernetics Inc). For the determination of mucus thickness using AB/PAS or MUC2-stained tissue sections, 20 measurements were taken from each mouse colon with 5 mice per group. Each measurement, the vertical distance between the cell surface and the luminal mucus surface, was calculated, and the mucus thickness is clearly delineated when this layer is well preserved. The numbers of AB/PAS-positive goblet cells were counted in 10 colonic crypts per section, and three random sections were counted per mouse, with 5 mice per group.

| Statistical analysis
The obtained data were analyzed using one-way analysis of variance (ANOVA) followed by LSD post hoc tests for multiple groups, and Student's t test (parametric data) or Mann-Whitney U test (nonparametric data) for two groups using SPSS 21.0 software (IBM Corporation). We analyzed the normality of data distribution and homogeneity of variance using Shapiro-Wilk's and Levene's test, respectively. Significance was defined when the p-value was < 0.05.

| Other methods
Please see details in the Supplementary experimental procedures for assessment of disease severity in mice with DSS-induced colitis, hematoxylin and eosin (H&E) staining, AB/PAS staining, immunofluorescence staining, Western blot analysis, and RNA and genomic DNA extraction and real-time PCR. The primers used are listed in Table S1.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in the supplementary material of this article.