Graphical abstract graphic file with name fx1.jpg [47]Open in a new tab Highlights * • Af6 in intestinal epithelial cells has a protective function in colitis * • Loss of Af6 in IECs impaired B-cell signaling and reduced IgA production during colitis * • Af6 regulates IECs' MHCII expression and mediates epithelial-immune interaction * • IgA supplementation improved colitis caused by Af6 deficiency __________________________________________________________________ Immunology; Cell biology Introduction IBD is a disorder of the gastrointestinal (GI) tract of unknown etiology, genetic predisposition, and extremely difficult to cure, which can have lifelong recurrences and may increase the risk of colorectal cancer.[48]^1 Crohn’s disease (CD) and ulcerative colitis (UC) are two types of IBD. There has been an alarming increase in the incidence of IBD over the last few decades. It has become a global disease and will continue to increase in the future.[49]^2 It is now accepted that IBD is the result of an extremely complex interaction between genetic and environmental factors, dysregulation of the immune response, changes in the microbiome, and possibly other risk factors.[50]^3^,[51]^4 However, there is no good treatment for this disease in the clinic. Understanding the pathogenesis and pathogenic factors of IBD and finding effective treatments is therefore urgent and necessary. Previous research has focused on the role of immune cells in IBD pathogenesis. While T-cells have been dominant in IBD pathology and treatment studies, there is increasing evidence that suggests that B-cells are also involved in IBD. In the gut, B cells shape the microbiome and regulate intestinal homeostasis by secreting immunoglobulin (Ig). Among them, immunoglobulin A (IgA) is the most abundant antibody (Ab) produced in mammals, more than all other Ab isotypes combined.[52]^5 DCs or M cells are able to present mucosal antigens, after which T cells and B cells are activated, and mucosal B cells mediate IgA class switch recombination (CSR). IgA CSR are divided into T-cell-dependent (TD) and T-cell-independent (TI) mechanisms according to the degree of T cell involvement in this process. However, humoral immune homeostasis in IBD is disrupted, and in particular, IgA production has been shown to be impaired in patients with IBD, and there was also an increase in the local production of the pro-inflammatory antibody IgG.[53]^6^,[54]^7^,[55]^8^,[56]^9 In addition, recent studies have also placed plasma cells at the center of research into new therapeutic avenues for UC through the disruption of circulation and intestinal B-cells.[57]^10 It is therefore clear that the role of B cells in IBD is attracting a lot of attention. Besides, IECs have emerged as key immunomodulatory factors and have been identified as a major target in IBD treatment.[58]^11 IECs are unspecialized antigen-presenting cells that present luminal antigen to host immune cells by Major Histocompatibility Class (MHC) I and II complexes expressed in it[59]^12 and the influence of the expression of MHC II in the IECs on intestinal homeostasis has received renewed attention lately. During active IBD, the expression of MHC-II in the colonic epithelium has been shown to be upregulated.[60]^13 MHC-II expression in IECs is mainly induced by interferon (IFN-γ) from immune cells invading the mucosa during active inflammation in IBD.[61]^14 Binding of IFN-γ to IFN-γ receptors leads to the activation of the JAK-STAT1 pathway. It then induces the transcription of IRF-1 and CIITA. CIITA then assembles a transcription factor complex at the MHC II promoter to regulate MHC II transcription.[62]^15^,[63]^16^,[64]^17 It has been reported that IECs can interact with T cells and myeloid cells through MHC II and restrict the composition of microorganisms.[65]^18^,[66]^19 The absence of MHC II in IECs reduces bacterial clearance by reducing IgA binding to symbiotic and pathogenic bacteria.[67]^20 IL18 promotes MHC II expression in IECs by IFN-γ.[68]^21 However, it is not clear how MHC II in IECs affects intestinal homeostasis/intestinal microbial composition and its related regulatory mechanisms. Cell polarity, defined as asymmetric cell shape and/or asymmetric distribution of proteins and functions within cells, is a key phenomenon in many biological processes that promote normal tissue integrity and development.[69]^22 As the largest mucosal surface in the human body, the IEC is composed of a simple columnar epithelium of polarized cells. IEC polarity is central for the homeostasis and immunity of the gut epithelium. Mechanistic defects in the polarity of IECs have been functionally linked to IBD and colorectal cancer,[70]^23 and some cell connexins associated with barrier integrity have also been used as markers for IBD. AF6, a polarity protein, localizes to AJs at polarized epithelial apices, is involved in a variety of biological processes in normal cells, including the formation of cell junctions, cell polarization, cell proliferation, migration, survival, differentiation, and the occurrence and development of tumors.[71]^24 Notably, AF6 has been shown to be a dual-resident protein that can be localized in both cell membranes and nuclei and is capable of interacting with many proteins, including cell adhesion molecules and their associated molecules, and signaling molecules to play a regulatory role.[72]^25^,[73]^26 Previous studies in our laboratory have also found this phenomenon.[74]^27 This suggests that the nuclear localization of AF6 may be involved in the regulation of other genes and signaling pathways to play an important role in disease. Although AF6 has been reported to play a role in inflammation,[75]^28^,[76]^29 the specific mechanism by which it participates in inflammation is still poorly understood, particularly in gastrointestinal disease. Here, we used the sodium dextran sulfate (DSS)-induced colitis model and C. rodentium infected model and found that AF6 has a protective effect on the occurrence and progression of colitis. AF6 intestinal epithelium-knockout mice (AF6^ΔIEC mice) had significant IgA deficiency and B cell signaling pathway damage during the development of IBD. Mechanically, AF6 regulates the expression of STAT1 by binding to IRF1, thereby influencing the expression of MHC II in IECs, thus participating in the crosstalk between epithelial cells and immune cells during inflammation to maintain intestinal homeostasis. AF6 also affects the IgA library by affecting the expression of IgA transporter polyimmunoglobulin receptor (pIgR) in intestinal epithelium. The loss of IgA altered the composition of the gut microbiota. Results AF6 levels are decreased in both mice and humans with colitis To explore the potential involvement of AF6 in colitis, we firstly examined the expression of AF6 in various tissues and organs by analysis of public databases at the National Center for Biotechnology Information (NCBI) and Genecards ([77]https://www.ncbi.nlm.nih.gov/gene/4301, [78]https://www.genecards.org/cgi-bin/carddisp.pl?gene=AFDN&keywords=af 6). We found that AF6 transcript accumulated to higher levels in the colon and small intestine compared to other tissues. To investigate whether AF6 might play a role in intestinal inflammation, we adopted the most widely used model of DSS-induced acute ulcerative colitis. Specifically, wild-type C57BL/6 mice were maintained for 7 days on drinking water containing 2.5% DSS; following euthanasia, colon tissues were collected and assessed for relevant biomarkers ([79]Figure S1A). Compared to control mice, animals maintained on DSS exhibited more severe weight loss; shorter colon length; and significant increased structural damage to, and inflammatory cell infiltration of, colon tissue ([80]Figures S1B–S1D), confirming the validity of this colitis model. Next, we examined the expression of Af6 in the colon, both at the mRNA and protein levels. Compared to the colon of control mice, that of DSS mice exhibited significantly lower levels of Af6 transcript ([81]Figure 1A). In particular, we found that Af6 protein levels were significantly lower in the intestinal epithelium of DSS mice compared to controls ([82]Figure 1B). Figure 1. [83]Figure 1 [84]Open in a new tab AF6 is decreased in the colitis of mice and human (A) qPCR was used to detect Af6 mRNA level in colon tissues of mice. (B) The protein expression of AF6 in IECs was detected by Western blot. (C) The mRNA level of AF6 in human colitis samples from GEO database. On the left is the comparison of AF6 expression in healthy controls (n = 21), patients with UC (n = 18), and patients with CD (n = 37) in the [85]GSE126124 dataset. On the right is a comparison of AF6 expression in healthy controls (n = 21), moderately active ulcerative colitis (n = 66), and severely active ulcerative colitis (n = 40) in the [86]GSE87473 dataset. (D) qPCR was used to detect the relative expression of AF6 mRNA in patients with UC and CD and healthy controls. The patient samples were collected from the Tenth People’s Hospital of Tongji University in Shanghai. n = 6–8 mice per group. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. To further verify the above results, we analyzed several public datasets. The results showed that the mRNA levels of AF6 were significantly lower in patients with UC compared to healthy controls, and the level of Af6 transcript exhibited a significant inverse correlation with disease progression (NCBI Gene Expression Sets [87]GSE126124 and [88]GSE87473) ([89]Figure 1C). We also detected significant decreases in AF6 mRNA levels in samples from patients with IBD compared to healthy controls ([90]Figure 1D). These results indicated that AF6 is involved in DSS-induced colitis and is decreased in samples of color from both humans and mice with colitis. AF6 deficiency exacerbates dextran sulfate sodium-induced colitis and C. Rodentium-induced colitis To directly evaluate the role of AF6 in colitis, we constructed a mouse strain harboring an IEC-specific knockout of the Af6 gene using the Cre-loxP system, which hybridized villin-cre mice with AF6^flox/flox mice ([91]Figures S2A and S2B), referred to hereafter as Af6^ΔIEC mice. After verifying that our mice exhibited good knockout efficiency ([92]Figures S2C and S2D), we induced colitis by maintaining male Af6^ΔIEC mice, as well as their “wild-type” (non-knockout) counterparts (Af6^f/f), for 7 days on drinking water containing 2.5% DSS before comparing the animals’ phenotypes ([93]Figure 2A). Notably, Af6^ΔIEC mice exhibited more severe colitis than did the control animals, as indicated by significant decreases in weight ([94]Figure 2B), increases in the disease activity index (DAI) score ([95]Figure 2C), and decreases in colon length ([96]Figure 2D). Histopathological analysis showed that the AF6 deficiency in IECs led to a more severe disruption of the mucosal epithelium in response to DSS exposure ([97]Figure 2E). Figure 2. [98]Figure 2 [99]Open in a new tab AF6 deficiency in IECs exacerbates DSS-induced colitis and C. rodentium infection-induced enteritis (A–F) (A) Schematic diagram of the ulcerative colitis model, which was conducted using both Af6^f/f and Af6^ΔIEC mice. Age matched Af6^f/f (n = 6) and Af6^ΔIEC (n = 6) mice (8–10 weeks old) were maintained for 7 days on drinking water containing 2.5% DSS, during which time body weight ((B); expressed as percent baseline) and conditions of the feces of the mice were recorded daily. After 7 days, the mice were euthanized, and the disease activity index was scored according to the changes in body weight, feces, and blood in the stool (C). The length of the colon tissue was measured at necropsy (D), and the tissue was then fixed for hematoxylin and eosin (HE) staining and histological scoring (E), Scale bar = 100um. (F) Schematic diagram of C. rodentium infection model in Af6^f/f (n = 6) and Af6^ΔIEC (n = 6) mice. (G) Changes in body weight over time in C. rodentium-infected mice. (H) Weights of ceca and colon (combined) and spleen of C. rodentium-infected mice at Day 11. (I) Quantification of C. rodentium in feces of C. rodentium-infected mice at Day 9. (J) Quantification of C. rodentium in the spleens of C. rodentium-infected mice at Day 11. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. To verify these findings, we also constructed an infectious model of enteritis induced by C. rodentium infection ([100]Figure 2F). The results showed that compared to control animals, Af6^ΔIEC mice experienced significant decreased body weight during infection ([101]Figure 2G) and significantly increased caecal and colon weight and spleen weight on Day 11 of infection ([102]Figure 2H). Infected Af6^ΔIEC mice also exhibited significantly increased C. rodentium colonization of both the colon and spleen compared to control animals, indicating decreased intestinal resistance to pathogenic bacteria ([103]Figures 2I and 2J). Together, these results suggested that loss of AF6 in the intestinal epithelium potentiates host susceptibility to C. rodentium infection. These observations were consistent with previous reports,[104]^29 and suggested that AF6 has a protective effect against colitis, both as induced by DSS and by bacterial infection. AF6 deficiency in inflammatory bowel disease mice leads to impaired immune signaling related to immunoglobulin A production We further investigated the role of AF6 in colitis using the DSS-induced model. Since DSS-induced colitis may lead to epithelial cell death and worsen colitis, to investigate whether AF6 has an effect on epithelial cells treated with DSS, markers of intestinal epithelial cells after DSS treatment were detected. We found no difference in the expression of E-cadherin and Epcam in Af6^f/f and Af6^ΔIEC mice ([105]Figures S2E–S2G). In order to further explore the mechanism of AF6 involvement in colitis, we performed RNA-seq analysis of colon tissue from DSS-induced Af6^f/f and Af6^ΔIEC mice. Gene ontology (GO) analysis showed that genes expressed differentially between colitis mice of the two genotypes were enriched for loci encoding components of the Ig-mediated immune response and of B cell-mediated immunity ([106]Figure 3A); similarly, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that significant differences between colitis mice of the two genotypes were detected in transcripts encoding components of the intestinal immune network for IgA production and the B cell receptor signaling pathway ([107]Figure S3A). Furthermore, Gene Set Enrichment Analysis (GSEA) based on GO and KEGG also highlighted the important role of AF6 in immunoglobulin production and B cell signaling ([108]Figures 3B and [109]S3B). These data suggested that AF6 in the intestinal epithelium may be contribute to the interaction between IECs and immune cells of the lamina propria (LP), especially B cells, to promote the activation of these B cells and subsequent IgA production. Consistent with this inference, flow cytometric analysis of the colon LP revealed that DSS-treated Af6^ΔIEC mice exhibited compared to DSS-exposed control mice a significant decrease in the proportion of IgA^+ B220^- plasma cells ([110]Figure 3C). We further demonstrated that the IgA content of the feces and serum of DSS-treated Af6^ΔIEC mice was significantly decreased compared to that of DSS-exposed control animals using enzyme-linked immunosorbent assay (ELISA) ([111]Figure 3D). Notably, Af6^ΔIEC mice showed a significant decrease in fecal IgA content compared to controls without DSS exposure ([112]Figure S3C). However, we did not observe differences in intestinal structure and body weight between Af6^f/f and Af6^ΔIEC mice ([113]Figures S3D–S3E). The latter result is consistent with the asymptomatic phenotype or mild consequences that have been reported in many patients with IgA deficiency.[114]^30 Figure 3. [115]Figure 3 [116]Open in a new tab AF6 deficiency in IECs impairs B cell signaling and IgA production in mice with colitis (A) GO pathway enrichment analysis of colonic tissue transcriptomes in Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. (B) GSEA analysis, based on GO, of colonic tissue transcriptomes in Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. NES, standardized enrichment scores. (C) Flow cytometric analysis was used to measure the proportion of IgA^+ B220^- cells in the colon lamina propria of Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. (D and E) (D) IgA content was measured by ELISA in feces (after 3 days on DSS-containing drinking water) and serum (after 7 days on DSS-containing drinking water) of Af6^f/f and Af6^ΔIEC mice. Flow cytometric analysis was used to measure the proportion of (E) MHCII^+ CD11c^+ cells and (F) CD44^+CD62L^− in the lamina propria of Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days, n = 5–8 mice per group. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. In the intestinal mucosa, IgA is induced in a T cell dependent and independent manner.[117]^31 The T cell-dependent pathway requires T cells to activate B cells, such that the activated B cells then differentiate into plasma cells and produce IgA; in contrast, the T cell-independent pathway does not require T cells, instead employing the direct stimulation of B cells by dendritic cells (DCs), IECs, or microbial metabolites. Since IECs have been reported to interact with LP myeloid cells, T cells, and B cells,[118]^18^,[119]^19^,[120]^32 we hypothesized that AF6 deficiency in the intestinal epithelium may affect DC- and/or T cell-mediated immune responses, leading in turn to impaired B cell signaling and deficiency for IgA. To test this theory, we used flow cytometry to analyze the LP immune cells of Af6^f/f and Af6^ΔIEC mice that had been maintained on DSS. Consistent with our speculation, the proportions of MHCII^+ CD11c^+ DC cells and CD44^+ CD62L^− T cells in the LP of the colon of Af6^ΔIEC mice were significantly decreased following DSS treatment ([121]Figures 3E and 3F). Furthermore, subsets of the T cells, including Th1 cells (which secrete primarily interferon (IFN) -γ) and Th17 cells (which secrete primarily interleukin (IL) −17), were also significantly depleted under these conditions ([122]Figures S3F–S3G). To further demonstrate these results and to examine the possibility that other immune cells are involved, we also examined the changes in neutrophils and macrophages, and cytokines involved. Immunofluorescence results showed that CD4 infiltration was significantly reduced in Af6^ΔIEC mice compared to the control group ([123]Figures S4A and S4D), which is consistent with our flow cytometry results ([124]Figure 3F). However, we did not find a difference between macrophages and neutrophils ([125]Figures S4B–S4F). In addition, we found that cytokines previously reported to be upregulated in colitis, such as IFN-γ, IL17, IL6, and TNF-α, were all reduced in our Af6^ΔIEC mice, although IL1 was significantly upregulated. In addition, we detected a significant up-regulation of IL10 levels in Af6^ΔIEC mice ([126]Figure S4G). More importantly, AF6 deletion did not result in changes in intestinal immunophenotype and immune cell infiltration in mice in homeostasis ([127]Figures S5 and [128]S6), suggesting that it performs a unique function in the context of inflammation. Taken together, these results suggested that AF6 deficiency in IECs affects T cell-dependent and -independent immune signaling, which may contribute to decreased IgA production. However, we cannot exclude a role for direct interaction between IECs and B cells in promoting IgA production. AF6 deficiency restricts the expression of major histocompatibility complex class II during colitis Based on these results, we further explored how AF6 in intestinal epithelial cells affects immune signaling in the lamina propria of the intestine. We revisited the previous RNA-seq results. Among them, the MHC II-mediated antigen processing and presentation pathway in GO analysis has attracted our attention ([129]Figure 3A). The ability of IECs to constitutively express the molecules required for antigen presentation (including major histocompatibility complex Class I and Class II proteins (MHC I and MHC II, respectively)) and to take up and process soluble antigens has been reported for decades.[130]^13^,[131]^33 There is increasing evidence that IECs are not only the target of immune response, but also involved in maintaining and regulating the mucosal immune response. Therefore, we postulated that AF6 contributes to the expression of MHC II in IECs, thereby regulating cross-talk between IECs and immune cells. And our RNA-seq data showed that the genes associated with antigen presentation and processing pathways were significantly decreased in DSS-exposed Af6^ΔIEC mice compared to DSS-exposed Af6^f/f animals ([132]Figures 4A, 4B, and [133]S7A). Similar results were observed for qPCR analysis of colonic tissue from C. rodentium-infected Af6^ΔIEC mice compared to infected Af6^f/f animals ([134]Figure S7B). Since this RNA-seq result was at the overall level of intestinal tissue, to further validate the downregulation of MHC II in intestinal epithelial cells, we isolated colon epithelial cells from DSS-exposed Af6^f/f and Af6^ΔIEC mice. Indeed, qPCR analysis of the RNA isolated from this subset of colon cells confirmed that the mRNA levels of genes encoding transcription factors that regulate MHCII expression, as well as those encoding several MHC II molecules, were significantly higher in Af6^f/f mice exposed to DSS compared to levels in non-exposed Af6^f/f mice, consistent with previous reports.[135]^13 However, this up-regulation effect disappeared when AF6 was absent from IECs ([136]Figure 4C). Flow cytometry and immunofluorescence staining of MHC II also confirmed the above results ([137]Figures 4D and 4E). Thus, these data indicated that the AF6 contributes to the expression of MHC II in IECs under inflammatory conditions. Figure 4. [138]Figure 4 [139]Open in a new tab AF6 deficiency in IECs is associated with decreased MHC II expression in mice with colitis (A) GSEA analysis, based on GO, of colonic tissue transcriptomes in Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. NES, standardized enrichment scores. (B) Transcriptome analysis of colon tissues from Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. Heat maps show a significant decrease in expression (in Af6^ΔIEC mice compared to Af6^f/f animals) of genes involved in antigen processing and presentation pathways. (C–E) (C) qPCR was used to measure the expression of genes encoding MHC II-related functions in intestinal epithelial cells (IECs) of Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. MHCII production in (IECs) of Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days was assessed by (D) flow cytometric analysis and (E) immunofluorescence staining, Scale bar = 100 μm. (F) Schematic diagram of the co-culture of organoids with T cells. (G) Flow cytometric analysis of T cell effector factors in co-cultures. n = 5–8 mice per group, and for co-culture, there were 2 organoids in each group. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. To further demonstrate that AF6-mediated MHC II expression in the intestinal epithelium is involved in epithelial-immune cell cross-talk, we established an Organoids-T cells co-culture system ([140]Figure 4F). We found that the absence of AF6 in colonic organoids resulted in decreased T cell effector function after co-culturing ([141]Figure 4G). These results indicated that AF6 may further mediate the T cell-dependent production of IgA. AF6 modulates MHC II expression by partly regulating STAT1 transcription Next, we explored the molecular pathways that might be involved in the IEC-specific regulation of MHCII expression by AF6. Cytokines (e.g., IFN-γ) produced by activated pro-inflammatory mucosal cells have been reported to induce the de novo synthesis of MHC II molecules.[142]^16^,[143]^17^,[144]^34 Previous work has also shown that the induction of MHCII expression by IFN-γ is mediated via the STAT1/IRF1/Class II transactivator (CIITA) signaling pathway, a regulatory pathway that is evolutionarily conserved from the bony fishes to mammals. We conjectured that the molecular mechanism of AF6 to regulate MHCII expression in IECs is also affected by IFN-γ-mediated signal transduction. To test this hypothesis, we first assessed the protein levels of components of the IFN-γ-mediated signaling pathway in IECs from Af6^f/f and Af6^ΔIEC mice exposed to DSS. Surprisingly, the levels of phosphorylated STAT1 and total STAT1 were significantly lower in IECs of DSS-exposed Af6^ΔIEC mice compared to those in the IECs of DSS-exposed treated Af6^f/f mice ([145]Figure 5A). Similar results were observed in the C. rodentium infection model ([146]Figure S7C). Interestingly, under normal conditions, there was no significant change in the IFN-γ pathway-related proteins in the two groups of mice ([147]Figure S7D). qPCR analysis of the RNA from the IECs further demonstrated that Stat1 transcript levels also were significantly lower in IECs of DSS-exposed Af6^ΔIEC mice compared to those in the IECs of DSS-exposed treated Af6^f/f mice ([148]Figure 5B), an observation that was consistent with the results of RNA-seq ([149]Figure S7E). Therefore, we inferred that AF6 regulates the transcription of STAT1. To verify this hypothesis, we isolated and cultured organoids from the colons of Af6^f/f and Af6^ΔIEC mice and treated them with IFN-γ. The results showed that the mRNA levels and protein levels of STAT1 were decreased when the loss of AF6, and the expression of STAT1 remained lower after IFN-γ treatment than that of the control group ([150]Figures 5C and 5D). Similar results were obtained for cultures of HT29 cells ([151]Figure S7F). These data suggest that AF6 acts by regulating STAT1 transcription. Figure 5. [152]Figure 5 [153]Open in a new tab AF6 regulates the expression of MHC II by modulating the expression of STAT1 in intestinal epithelial cells (IECs) (A) The protein levels of components of the IFN-γ-related signaling pathway were detected by immunoblotting of colon tissues from Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. (B) qPCR was used to assess STAT1 transcript levels in colon tissues from Af6^f/f and Af6^ΔIEC mice maintained on 2.5% DSS for 7 days. (C and D) (C) The mRNA levels of STAT1 and its downstream target genes (as assessed by qPCR) and (D) the protein levels of STAT1 and proteins encoded by its downstream target genes (as assessed by immunoblotting) were determined in organoids generated from colon tissues of Af6^f/f and Af6^ΔIEC mice, as measured before and after IFN-γ treatment. (E) Co-immunoprecipitation (co-IP) was used to detect endogenous interactions between AF6 and IRF1 in IECs. (F) 293T cells were co-transfected with constructs encoding hemagglutinin-tagged AF6 (HA-AF6) and Flag peptide-tagged IRF1 (Flag-IRF1); their interaction domains were detected by co-IP. (G) Immunoblotting was used to assess the expression of IRF1 and proteins encoded by its downstream genes in HT29 cells with or without IFN-γ exposure. (H) HT29 cells were transfected with constructs encoding AF6 with no or 3 nuclear localization signaling (NLS) domains (ΔNLS-AF6 and 3×NLS-AF6, respectively), and the expression and localization of IRF1 and proteins encoded by its downstream genes were assessed by immunoblotting following the separation of the nuclear and cytoplasmic fractions. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. This observation begs the question of how AF6 affects STAT1 transcription. Notably, AF6 has not been reported to exhibit transcription factor activity, although AF6 is known to localize both to the cell membrane and the nucleus, and to possess multiple domains that facilitate interactions with other proteins, including transcription factors.[154]^26^,[155]^27^,[156]^35 IRF1 is a protein with a similar profile: IRF1 is a STAT1-targeted gene but the IRF1 protein, a known transcription factor, also regulates STAT1 transcription.[157]^36^,[158]^37 We speculated that, in IECs, AF6 may interact with IRF1 to regulate STAT1 transcription, a role mediated by the nuclear localization function of AF6. To evaluate this proposal, we first tested the HT29 cell line to determine whether AF6 also has dual-localization characteristics in cells derived from the intestinal epithelium. Western blot results showed that AF6 localized to both the cell membrane and the nucleus in the HT29 cell line, although the AF6 proteins present at the two sites represented distinct major isomers ([159]Figure S7G). Next, we explored whether AF6 exhibited physical association with IRF1. Co-immunocoprecipitation (Co-IP) experiments revealed an endogenous interaction between AF6 and IRF1 in IECs isolated from mice ([160]Figure 5E). Further in vitro Co-IP experiments in HEK 293T cells employing truncated AF6 protein plasmids and IRF1 plasmid showed that AF6 binds to IRF1 via AF6’s N-terminal domain ([161]Figures 5F and [162]S7H). Given these results, we hypothesized that the binding of AF6 to IRF1 may affect the nuclear localization of IRF1 and result in reduced binding to target genes. To verify this proposal, we performed nuclear and cytoplasmic separation assays on the HT29 cell line. The results showed that the knockout of AF6 resulted in a decrease in the accumulation of IRF1 in the nucleus and an increase in the accumulation of IRF1 in the plasma compared to cells retaining Af6; concomitantly, the expression of IRF1’s downstream target genes, including STAT1, was decreased ([163]Figure 5G). We also showed in HT29 cells that the expression of 3XNLS-AF6, but not that of ΔNLS-AF6 [i.e., AF6 lacking the nuclear localization sequence (NLS) that promotes protein nuclear translocation[164]^27], promotes the nuclear accumulation of IRF1 and the expression of IRF1’s downstream target genes ([165]Figure 5H). Taken together, these results confirm our conjecture that AF6 interacts with IRF1, increasing the nuclear localization of this transcription factor. Additionally, we noted, as mentioned above, that the proportion of IFN-γ^+ CD4^+ Th1 cells in the intestinal LP of Af6^ΔIEC mice was significantly decreased compared to that in Af6^f/f mice; this phenomenon may have further contributed to the deficiency of MHCII in IECs ([166]Figure S3F). AF6 contributes to immunoglobulin A transport via the regulation of polyimmunoglobulin receptor expression IgA produced by intestinal B cells typically exists as a dimer, with monomers linked by a protein called the J-chain; the dimer then is transported to the mucus layer by binding to the polyimmunoglobulin receptor (pIgR), a transporter located on the basal side of intestinal epithelial cells, yielding a version of the antibody referred to as secretory IgA (SIgA). pIgR not only transports the polymeric IgA, but also provides secretory components that are in part covalently bound to the antibody, thereby serving as a component of the SIgA complex. Thus, SIgA production and secretion are influenced by pIgR expression and activity.[167]^38 The expression of Pigr was examined in IECs from Af6^f/f and Af6^ΔIEC mice that were exposed to DSS. We found that Pigr expression at both the mRNA and protein levels was significantly lower in IECs of DSS treated Af6^ΔIEC mice compared to Af6^f/f mice ([168]Figures 6A and 6B). This result suggested that AF6 is involved in the regulation of pIgR expression in the intestinal epithelium under inflammatory conditions. Previous work has reported that a pro-inflammatory cytokine, IL-17, enhances pIgR expression, and it acts through a nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) -binding element located in the first intron of the Pigr gene in both mouse and human IECs.[169]^39^,[170]^40 To explore whether AF6 regulates the expression of pIgR in the intestinal epithelium of mice by participating in IL17-NF-kB signaling, we isolated colon organoids from Af6^f/f and Af6^ΔIEC mice and exposed them or not to IL17. The result suggested that pIgR expression at both the mRNA and protein levels was indeed significantly upregulated in IL17-treated colon organoids of Af6^f/f mice compared to untreated. However, pIgR expression at both the mRNA and protein levels was significantly down-regulated in the colon organoids of Af6^ΔIEC mice compared to Af6^f/f mice, with or without IL17 treatment. Consistent with this, the expression of p-p65 (p-NF-kB) was significantly reduced in the colon organoids of Af6^ΔIEC mice compared to Af6^f/f mice with or without IL17 treatment, which supports our hypothesis ([171]Figures 6C and 6D). In addition, as mentioned above, we demonstrated that the proportion of Th17 cells expressing IL-17A in the intestinal LP of DSS-exposed Af6^ΔIEC mice was significantly lower than that in DSS-exposed Af6^f/f mice ([172]Figure S3G), which may further inhibit the IL17-NF-kb pathway. Previous studies also reported that signaling through the IFN-γ receptor causes the activation and nuclear translocation of STAT1 dimers, resulting in the de novo transcription of the Irf1 gene and association of the IRF1 protein with a cognate element in the first exon of the Pigr gene.[173]^41^,[174]^42 To investigate these effects in the context of our study, we repeated IFN-γ exposure in a derivative of HT29 in which AF6 was knockdown by shRNA lentivirus and in mice colon organoids derived from Af6^f/f and Af6^ΔIEC mice. In HT29 cells, the depletion of AF6 impeded IFN-γ-induced pIgR expression ([175]Figure 6E), consistent with the proposal that the IFN-γ-STAT1-induced expression of pIgR in human IECs is mediated by AF6. Interestingly, exposure of colon organoids of Af6^f/f mice to IFN-γ did not alter pIgR expression, which is consistent with previous reports.[176]^42 However, we observed that pIgR expression at both the mRNA and protein levels was reduced in the colon organoids of Af6^ΔIEC mice compared to Af6^f/f mice, regardless of exposure to IFN-γ ([177]Figures 6F and 6G). Thus, our data suggested a species-specific difference in the regulation of pIgR expression: in mice, AF6 appears to regulate pIgR expression primarily via IL-17, while in humans, both IL-17 and IFN-γ may be involved.[178]^40 Considered together, these results suggested that the loss of AF6 affects not only IgA production but also IgA transport, leading to low gastrointestinal levels of IgA and severe colitis. Figure 6. [179]Figure 6 [180]Open in a new tab AF6 deficiency in intestinal epithelial cells (IECs) is associated with the decreased expression of Pigr (A and B) The levels of (A) Pigr transcript (as assessed by qPCR) and (B) pIgR protein (as assessed by immunoblotting) were measured in IECS from Af6^f/f and Af6^ΔIEC mice maintained on DSS. (C) The mRNA levels of Pigr (as assessed by qPCR) and (D) the protein levels of p-NF-κB and pIgR (as assessed by immunoblotting) were determined in organoids generated from colon tissues of Af6^f/f and Af6^ΔIEC mice, as measured before and after IFN-γ treatment. (E) The protein levels of pIgR were determined (by immunoblotting) in an Af6-silenced HT29 cell line (generated by stable transfection with PLKO and shAF6 constructs) grown in the presence of various concentrations of IFN-γ. (F and G) (F) The mRNA levels of Pigr (as assessed by qPCR) and (G) the protein levels of pIgR (as assessed by immunoblotting) were determined in organoids generated from colon tissues of Af6^f/f and Af6^ΔIEC mice, as measured before and after IFN-γ treatment. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. AF6 deficiency in intestinal epithelial cells changes the composition of the intestinal microbiota IgA plays an important role in maintaining a healthy balance of gut microbiota. Given that gut IgA levels were attenuated in Af6^ΔIEC mice than those in Af6^f/f mice ([181]Figure S3C), we next performed a 16S rRNA-seq analysis of the fecal microbiota in Af6^f/f and Af6^ΔIEC mice under normal conditions (i.e., without the induction of colitis or enteritis). Community composition analysis showed that the gut bacterial composition of mice of the two genotypes differed significantly at the phylum level. Specifically, Af6^ΔIEC mice exhibited, compared to Af6^f/f mice, a significantly lower abundance of Actinomyces and significantly higher abundances of Firmicutes, Bacteroides, Proteobacteria, Dermobacillus, and Verrucobacteria ([182]Figure 7A). These data were then subjected to α diversity analysis, which defined the number and richness of species in the sample, using different methods. The results showed that the α diversity index of fecal bacteria in the Af6^ΔIEC mice was significantly higher than that in the control mice ([183]Figure 7B). To further validate the differences in microbial structure between animals of the two genotypes, β-diversity analysis was performed using principal component analysis (PCA) and principal coordinates analysis (PCoA) analysis; notably, both techniques revealed significant clustering between Operational Taxonomic Unit (OTUs) in Af6^f/f and Af6^ΔIEC mice ([184]Figure 7C). To further evaluate the significant differences in gut bacterial community dominance between the two genotypes, linear discriminant analysis effect size (LEfSe) was employed for high-dimensional comparison. By this approach, Af6^f/f mice showed a preponderance of Actinomycetes and Bifidobacteria, while Af6^ΔIEC mice showed a preponderance of Clostridiales, Deferribacteres, Verrucomicrobia, Akkermansia, and Oscillospiraceae ([185]Figure 7D). These results suggested that IgA deficiency resulting from the selective knockout of Af6 in IECs changed the composition of the intestinal microbiota, leading to a decrease in the abundance of probiotic microbes and an increase in symbiotic bacterial pathogens. Figure 7. [186]Figure 7 [187]Open in a new tab AF6 affects the composition of the intestinal microbiota (A) Bar chart of the phylum-level composition of the fecal microbiota in Af6^f/f and Af6^ΔIEC mice. (B) Boxplots of α diversity of fecal microbiota in Af6^f/f and Af6^ΔIEC mice, show the Shannon index, Simpson index, inverse Simpson index, richness index, and evenness index. (C) Plots show the β diversity analyses of fecal microbiota in Af6^f/f and Af6^ΔIEC mice. Upper panel: PCA analysis; lower panel, PCoA analysis based on the Bray-Curtis distance algorithm. (D) The LDA scores were calculated based on the differences in fecal microbiota abundances between Af6^f/f and Af6^ΔIEC mice. The feature selection criterion was log (LDA score) > 4. n = 4 mice per group. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. Immunoglobulin A administration and antibiotic treatment restore the sensitivity to dextran sulfate sodium-induced colitis and improve the gut microbiota IgA is known to play an important role in protecting intestinal integrity and regulating the intestinal flora. To confirm that the more severe colitis observed with Af6 deletion resulted from the depletion of gut IgA in the mutant mice, we provided the mice with free access to drinking water containing monoclonal IgA antibody (Isolate W27) for 3 days before the initiation of the administration of DSS in the drinking water ([188]Figure 8A). Our results showed that compared to Af6^ΔIEC mice drinking water lacking IgA, Af6^ΔIEC mice drinking water containing W27 exhibited significantly attenuation of the adverse effects of DSS-induced colitis, including the attenuation of weight loss ([189]Figure 8B), a reduced disease score index ([190]Figure 8C), longer colon length ([191]Figure 8D), and improved histopathology ([192]Figure 8E). In contrast, W27 supplementation did not affect the disease phenotype in DSS-exposed Af6^f/f mice ([193]Figures 8A–8E). These observations suggested that the gut secretion of IgA, which results from the IEC-immune cell interaction, is essential for inhibiting intestinal inflammation. Figure 8. [194]Figure 8 [195]Open in a new tab IgA administration in mice with DSS-induced colitis restores both the normal sensitivity to the disease and the abundances of the gut microbiota (A) Schematic diagram of the experimental W27 IgA supplementation model. Age matched Af6^f/f (n = 6) and Af6^ΔIEC (n = 6) mice (8–10 weeks old) were giving W27 supplemented water for 3 days then maintained for 7 days on drinking water containing 1.5% DSS, during which time body weight ((B); expressed as percent baseline) and conditions of the feces of the mice were recorded daily. After 7 days, the mice were euthanized, and the disease activity index was scored according to the changes in body weight, feces, and blood in the stool (C). The length of the colon tissue was measured at necropsy (D), and the tissue was then fixed for hematoxylin and eosin (HE) staining (E), Scale bar = 300um. (F) The effects of W27 supplementation on the intestinal microbiota of Af6^ΔIEC mice were detected by qPCR. Data are expressed as mean ± SEM. Pairwise comparisons between groups were conducted using two-tailed non-paired Student’s t tests. p-values were shown in the panel, and p < 0.05 indicates a significant difference. As noted above, the loss of IgA led to changes in the gut microbiota, in particular Verrucomicrobia and Akkermansia ([196]Figure 7D), a result that is consistent with a previous report.[197]^5 Therefore, we also examined whether IgA supplementation has a regulatory effect on the abundance of these microorganisms. As expected, we found that compared to Af6^ΔIEC mice maintained on unsupplemented drinking water, Af6^ΔIEC mice maintained on drinking water containing W27 demonstrated the attenuation of the expansion of Akkermansia muciniphila and potentiation of the abundance of Bifidobacterium ([198]Figure 8F). To further explore whether changes in gut flora caused by IgA loss were the cause of more severe colitis, we performed an antibiotic clearance assay ([199]Figure S8A). Interestingly, antibiotic clearance seemed to improve the colitis phenotype of Af6^f/f mice, as we observed longer colon length than Af6^f/f mice without antibiotic treatment, although there was no significant difference in other measures; antibiotic clearance significantly improved colitis severity in Af6^ΔIEC mice ([200]Figures S8B–S8E). However, similar to IgA supplementation, antibiotic clearance did not fully restore the phenotype of Af6^ΔIEC mice, as they still had a severe colitis phenotype compared to Af6^f/f mice. This suggests that AF6 is involved in colitis through other pathways besides affecting IgA. Discussion Several polarity proteins, including ZO-1, ZO-2, and AF6, have been shown to possess both NLSs and NESs; given the effect of these domains on the variable localization of these proteins, these molecules are also referred to as dual-residence proteins. Work in our laboratory and in those of others has shown that these molecules are able to co-localize to the nucleus, where these proteins contribute to the transcription of other genes. In the present study, we found that mice with an IEC-specific knockout of Af6 IEC exhibited a more severe colitis phenotype in the DSS-induced IBD model and in the C. rodentium infection-induced enteritis model. Surprisingly, RNA-seq analysis suggested that AF6 deficiency in IECs results in impaired B cell signaling in the LP and defects in IgA production during colonic inflammation. We then confirmed that nuclear AF6 binds to the IRF1 transcription factor via AF6’s N-terminal domain, with concomitant changes in the regulation of STAT1 expression and subsequent alterations in the expression of MHCII in IECs in IBD mice; we further showed that these changes influence the crosstalk between IECs and B cells, resulting in the modulation of intestinal IgA production. Additionally, we demonstrated that AF6 deficiency may also attenuate the transcription in IECs of Pigr, which encodes a protein responsible for the transport of IgA across the gut mucosal epithelium. Furthermore, we showed that AF6 deficiency in IECs changes the composition of intestinal microbiota, with subsequent effects on the severity of colitis. Together, these results suggest the molecular mechanism whereby the dual-residence protein AF6 serves as an essential arm of the adaptive immune system, providing a link between IECs and B cells of the LP. The host mucosal compartment is tightly regulated, serving both as a physical barrier and a source for the secreted molecules such as mucus and IgA.[201]^43 sIgA is the major mucosal antibody, preserving homeostasis with the intestinal immune system and mediating microbial homeostasis in the gut. The literature shows that IBD results in a disruption in the balance of antibody secretion by B cells, especially for the secretion of IgA.[202]^7^,[203]^38^,[204]^44^,[205]^45 Notably, patients with IBD have been shown to exhibit impaired production of IgA, as evidenced by the depletion of symbiotic bacteria in the gut of these individuals.[206]^6 IBD has also been associated with a relative decrease in the production of IgA2, a less-understood subclass of mucosal antibodies. The present study showed that the severity of IBD was potentiated in mice harboring an IEC-selective deletion of Af6, and further demonstrated that this effect is associated with the dysregulation of B cell-mediated humoral immunity and a deficiency in gut IgA. This work highlighted the important role of intestinal B cells and of the IgA secreted by these cells in the development and progression of IBD. While not all DSS-induced mice exhibit remarkable attenuation of AF6 levels, we did find that AF6 expression at both the mRNA and protein level is decreased in the intestinal epithelium of patients with IBD; however, some individuals with IgA deficiency are sometimes asymptomatic or have only mild symptoms of IBD, due to the partial functional compensation caused by the increased production of IgM.[207]^30^,[208]^46 Notably, the induction of IBD in Af6^ΔIEC mice consistently results in IgA deficiency; the oral administration of IgA in Af6^ΔIEC mice with IBD attenuated the disease severity and restored the intestinal microbiota balance to a state better resembling that of healthy animals. Previous work reported that the expression of MHC is associated with that of IgA.[209]^47^,[210]^48 MHC proteins are expressed not only in specialized antigen-presenting cells, but also in non-specialized antigen-presenting cells such as IECs. The expression of MHC II in IECs was first reported several decades ago, but only recently have studies begun to investigate the function of MHC II in intestinal inflammation and epithelial homeostasis.[211]^19^,[212]^49 Compared to control animals, mice harboring an IEC-specific deletion of the MHCII-encoding gene have a lower proportion of IgA-coated intestinal bacteria and a decreased intracavitary concentration of SIgA; furthermore, in a C. rodentium infection model, these animals exhibit an increased abundance of fecal pathogens and elevated mortality.[213]^20 Genome-wide Association Study (GWAS) analysis showed that a polymorphism in the human MHC II-encoding gene is associated with ulcerative colitis (UC).[214]^50^,[215]^51 As noted in a recent article, selective deletions and single-cell techniques have brought renewed attention to the potential role of IEC-expressed MHC II in mucosal immunity.[216]^52 Although the expression of MHC II in antigen-presenting cells is known to be stimulated by various cytokines and chemokines, little is known about the regulation of MHC II expression in IECs.[217]^53 Notably, we found that the loss of the intestinal epithelial expression of MHC II as a result of AF6 deficiency led to IgA deficiency and more severe disease in mice with DSS-induced colitis; IgA supplementation effectively counteracted this phenotype. We propose that the oral administration of IgA in patients with AF6-deficient colitis may be a valuable treatment strategy. The expression of MHC II is mainly controlled at the level of gene transcription. Class II transactivator (CIITA) has been recognized as the “master regulator” of MHC-II expression.[218]^16^,[219]^17^,[220]^54 Multiple proteins, including STAT1, IRF1, IRF2, and Upstream Stimulatory Factor 1 (USF1), have been implicated in the induction of CIITA activity following exposure to IFN-γ.[221]^15^,[222]^55^,[223]^56 IFN-γ has also been shown to induce the expression of IRF1 and IRF8, providing a time-dependent increase in the interaction of these proteins with STAT1 and in the binding of STAT1-IRF1 complexes to transcriptional regulatory elements.[224]^57 IRF1 also been reported to promote the transcription of STAT1 and phosphorylation of the encoded protein.[225]^36^,[226]^58 Notably, our study showed that nuclear AF6 binds to IRF1, thereby influencing IRF1-STAT1-CIITA signaling; this observation provides new insights into the understanding of this pathway. Intestinal IgA is produced by mature B cells upon the stimulation of these cells; this process involves interactions between T cells, B cells, DCs, and IECs; the interactions of IECs with various immune cells in the LP are known to be mediated via MHC II. Here, we did not explore whether IEC AF6 loss directly or indirectly affects B cell signaling, because we observed a broad intestinal immune deficiency, including the downregulation of DC and T cell subsets, and we prefer to think that it affects multiple signaling pathways and ultimately leads to the downregulation of B cell signaling. This also highlights the important role of IECs as non-immune cells involved in intestinal immunity by expressing MHC II. In the present work, we also demonstrated that the depletion of MHC II in epithelial cells as a result of the IEC-specific deletion of Af6 reduces T cell signaling; this effect may contribute to a decrease in the T cell-dependent production of IgA. This observation also highlights the important role of IECs as non-immune cells that contribute to intestinal immunity, a role that appears to depend on the expression of MHC II. IgA levels in the gut are influenced not only by B cell production, but also by changes in the expression in IECs of the pIgR transport receptor. pIgR expression is regulated by a variety of signals, and two transcription factors, IRF1 and NF-κB, have been reported to be involved.[227]^41 IL-17 has been reported to promote pIgR expression, in part by stimulating the NF-κB pathway.[228]^39 Our results suggested that, at least in mice, AF6 regulates pIgR expression by influencing the NF-κB pathway, although the specific regulatory mechanism remains unknown. IgA plays an important role in the maintenance of symbiotic microbial homeostasis, and IgA deficiency often results in gut dysbiosis.[229]^5^,[230]^59 We found that compared to control mice, Af6^ΔIEC mice were depleted for gut IgA, with associated changes in the gut microbiota, particularly an increase in the abundances of the Clostridiales, the Deferribacterales, and the Verrucomicrobia. Some members of the Clostridiales and of the Deferribacterales have been reported to be associated with intestinal inflammation and other diseases.[231]^60^,[232]^61^,[233]^62 In recent years, the Verrucobacteria, especially members of the genus Akkermansia, have been proposed as probiotics,[234]^63 but these bacteria have been reported to accumulate in the intestine of IgA-deficient mice in which colitis has been experimentally induced.[235]^5 Previous work has shown that IgA supplementation can effectively alleviate colitis and counteract disease-associated changes in the intestinal microbiota, while attenuating the expansion of Akkermansia.[236]^5 Thus, our study revealed the important role of AF6 expression in the intestinal epithelium, where the protein mediates epithelial-immune cell interactions and intestinal IgA homeostasis. To our knowledge, this work represents the first demonstration that a polarity protein contributes to these processes. Given the important and complex regulatory functions that AF6 plays in the gut, we expect that further exploration of the mechanism of AF6’s involvement in intestinal homeostasis will further both our understanding of this protein’s role in colitis and the development of therapies to treat this disease. and will open new horizons in our understanding of the immunologic mechanisms that contribute to gut inflammation. Limitations of the study One limitation of this study is that when exploring how epithelial cells affect B cell signaling, only epithelial cells were detected to affect B cell signaling through interaction with T cells, and whether epithelial cells directly affect B cells was not investigated due to limitations of experimental techniques. In addition, the specific mechanism of the AF6 regulation of pIgR expression remains to be further studied. Although our study found that AF6 expressed by intestinal epithelial cells plays an important role in colitis, we only found part of the regulatory mechanisms involved in colitis. Future studies need to explore other important pathways affected by AF6, which is expected to provide more profound guidance for the prevention and treatment of colitis. For the patient samples used in the article, due to the lack of race, ancestry, and ethnicity information, the reference significance for different regions is reduced. In the future, more patient samples need to be collected to verify the experimental conclusion, and the impact of race, ancestry, and ethnicity information on the conclusion should be focused on. Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Lixing Zhan (lxzhan@sinh.ac.cn). Materials availability This study did not generate any new reagents. Data and code availability * • RNA-seq data that support the findings of this study have been deposited in the NCBI under accession code PRJNA1190056. 16S rRNA seq data that support the findings of this study have been deposited in the NCBI under accession code PRJNA1191386. * • This article does not report original code. * • All other data supporting the findings of this study are available from the corresponding author on reasonable request. Acknowledgments