Gut microbiota derived metabolites 3-idoleacetic acid with LPS induces generation of IL-35+B cells

rongcun Yang (  ryang@nankai.edu.cn ) Nankai University https://orcid.org/0000-0002-5826-4493 Xiaomin Su Nankai University Houbao qi Nankai University Minying Zhang Nankai University Yunhuan Gao Nankai University Yazheng Yang Nankai University Huan Yun Nankai University Qianjing Zhang Nankai University Xiaorong Yang Nankai University Yuan Zhang Nankai University Jiangshan He Nankai University Yaqi Fan Nankai University Yuxue wang Nankai University Pei Guo Nankai University Chunze Zhang nankai university


Abstract
Background IL-35 producing B and Treg cells are critical regulators in worldwide chronic illnesses, which are related to disruptions of gut microbiota composition. Whether gut microbiota regulate these IL-35 + cells maintains elusive. Here we investigated the regulation of gut microbiota in the IL-35 + cells through genetically modi ed mouse models against obesity.

Results
Here, we rst found that gut Reg4 can promote resistance to high-fat diet induced obesity. Using 16S rRNA gene-based microbiota analysis combined with LC-MS (Liquid chromatography-mass spectrometry)/MS, we demonstrate that gut Reg4 associated microbiota such as lactobacillus can

Background
The gut microbiota in uences essential human functions including in ammation, digestion and energy metabolism by modulating immune pathways, neural and endocrine system of the host [1][2][3][4]. Disruptions of the microbiota composition and function by factors such as genetics have been suggested to be critical contribution to the growth of the worldwide chronic illness such as metabolic diseases [5][6][7].
However, that microbiota alterations in uence the development and outcome of metabolic diseases is incompletely characterized.
Gut microbiota/metabolites can affect the differentiation and development of immune cells. Multiple transcription factors such as aryl hydrocarbon receptor (AhR) [8,9], Foxp3 and RORγ [10][11][12] etc. are involved in this process such as AhR contributes to IL22 transcription [8] through an AhR ligand indole-3aldehyde, which is produced by that lactobacillus reuteri. Studies in animal models and in humans have demonstrated that gastrointestinal bacteria/metabolites also participate in B cell differentiation, maturation, and activation [13,14] such as that aryl hydrocarbon contributes to the transcriptional Results Reg4 promotes resistance to high-fat diet induced obesity We previously reported that Reg4 could kill E. coli through complement-dependent pathway [35]. Since increased E. coli in colon are related to occurrence and development of multiple diseases such as obesity [36][37][38], we investigated the role of Reg4 in high-fat diet (HFD) mediated obesity using Reg4 knockout (KO) mice. We found Reg4 KO mice showed more sensitivity to HFD-induced obesity, including more body weight, more weight of fat pad tissues, decreased sensitivity to insulin and reduced tolerance to glucose ( Fig.1a-d) although these Reg4 KO mice had not remarkable changes as compared to control cohoused wild-type (WT) mice under normal chow ( Fig. S1a and b). Histochemical staining showed larger adipose cells in Reg4 KO mice than in WT mice (Fig. 1e). In ammation is related to development of obesity [39].
In ammatory cytokines such as TNFa, IL-6 and MCP-1 were also increased in the adipose tissues of Reg4 KO mice (Fig. 1f). To further understand the role of gut Reg4 in HFD-mediated obesity, we generated intestine epithelial cell human REG4 transgenic mice (huREG4 IECtg ) mice (Fig. S1c-e). These huREG4 IECtg mice did not have markedly changes as compared to WT control mice from their littermates under feeding a normal chow ( Fig. S1f and g). However, while these mice were fed with HFD, huREG4 IECtg mice showed markedly resistance to HFD-induced obesity, including less body weight and weight of fat pad tissue, increased sensitivity to insulin and tolerance to glucose (Fig. 1g-j). Histochemical staining showed smaller adipose cells in huREG4 IECtg mice than those in WT mice (Fig. 1k). In ammatory cytokines such as TNFa, IL-6 and MCP-1 were also less in the adipose tissues of huREG4 IECtg mice than control WT mice (Fig. 1l). Thus, gut derived Reg4 (REG4 in human) is involved in sensitivity to HFD mediated obesity.
Reg 4 is related to accumulation of IL-35 + cells Since chronic in ammation play a critical role in occurrence and development of obesity [39]. Proin ammatory cells such as M1 macrophages, Th1 cells, CD8 T cells, and mature B cells are often found in adipose tissues of individuals with obesity; Other anti-in ammatory cells including M2 macrophages, regulatory T cells (Tregs), Th2 cells, and eosinophils are also found in the adipose tissues of thin individuals [39,40]. Indeed, the number and proportion of M1 and M2 macrophages, Treg and Th1 cells had remarkable changes in the adipose tissues of huREG4 IECtg and Reg4KO mice as compared to their control WT mice. Typically, M2 macrophages and Treg cells were markedly increased in the adipose tissues of huREG4 IECtg mice, which are resistance to HFD-mediated obesity; whereas in Reg4 KO mice, which are more sensitive to HFD-mediated obesity, these cells were signi cantly decreased ( Fig. 2a and  b). These data support that M2 macrophages and Treg cells are involved in the resistance to HFDmediated obesity [39,40]. We next look for the factor(s) which could cause increased M2 macrophages and Treg cells. IL-35 can potentially induce regulatory T cell population and M2 macrophages [41,42] and reduce gene expression of pro-in ammatory cytokines and Th1 associated transcription factors [41].
Thus we next investigated IL-35 + cells. The numbers of both IL-35 + B and IL-35 + CD4 cells were markedly reduced in the adipose tissues (fat pad and subcutaneous adipose tissues) of Reg4 KO mice; whereas the IL-35 + cells were signi cantly increased in the adipose tissues in huREG4 IECtg mice ( Fig. 2c and d; Fig.   S2a and b). Accumulation of IL-35 + B cells in the adipose tissues of huREG4 IECtg mice were further con rmed using immunohistochemical staining (Fig. 2e). The transcription levels of Epi3 and P35 also were higher in adipose tissues of huREG4 IECtg mice but lower in Reg4 KO mice (Fig. 2f). Cytokine analyses also showed higher levels of IL-35 in the peripheral blood of huREG4 IECtg mice; whereas it was lower in the peripheral blood of Reg4 KO mice (Fig. 2g). We further characterized the IL-35 + B cells. Data showed that increased IL-35 + B cells in the adipose tissue of huREG4 IECtg mice with resistance to HFDmediated obesity were belong to immature CD19 + IgD + B cells ( Fig. S2c and d), which are different from IgG producing B cells found in the adipose tissues of mice with sensitivity to obesity [43,44].
Reg4 associated gut microbiota/metabolites are related to IL-35 + cells Gut microbiota /metabolites play a critical role in the formation of immune system [1][2][3]. Reg4 can not only kill E. coli through complement-dependent pathway [35] but also induce damage to bacterial cell wall [33,34], implying that Reg4 might cause changed gut microbiota in gut. Indeed, ow cytometry showed that WGA + bacteria increased but LPS + bacteria reduced in the fresh stool of huREG4 IECtg mice; whereas WGA + bacteria decreased but LPS + bacteria increased in Reg4 KO mice as compared to their control ( Fig.   4a), indicating the changed proportion of Gram + and Grambacteria.16S ribosomal RNA (V3-V4 variable region) sequencing analyses of the ileum and colonic contents showed marked changes in the composition of gut microbiota, especially lactobacillus. The proportion of lactobacillus was markedly higher in huREG4 IECtg mice (Fig. 4b), consistent with that Reg4 affects the proportion of gut lactobacillus [45]. We also used LC-MS (Liquid chromatography-mass spectrometry)/MS combined with ELISA to analyze metabolites in peripheral blood and stool of huREG4 IECtg mice. There was markedly increased 3idoleacetic acid (IAA) in the peripheral blood and fresh stool of huREG4 IECtg mice; whereas lower levels of IAA could be found in peripheral blood and stool of Reg4 KO mice as compared to their control ( Fig. 4c and d).
Since gut microbiota/metabolites play a critical role in forming immune system and in maintaining gut immune homeostasis [1,2], we hypothesized that accumulation of IL35 + cells in the adipose and gut tissues was derived from changed gut microbiota. To found the relationship between gut microbiota and IL-35 + cells, we performed feces exchanged experiments, huREG4 IECtg mouse feces transferred into WT mice could increase IL-35 cells; whereas Reg4 KO mouse feces transferred into WT mice reduced the number of IL-35 cells in colonic LP tissues ( Fig. S5a and b). Furthermore, CD19 + p35 + cells markedly increased in germ-free (GF) mice infused by huREG4 IECtg mouse feces but not in GF mice infused by Reg4 KO mouse feces ( Fig. S5c and d). All of these suggest that Reg4 associated gut microbiota/metabolites play a role in the generation of IL-35 + cells.
IAA with LPS promotes binding of PXR with NFkB p65 and RXR IAA is a potent bioactive metabolites to activate pregnane X receptor (PXR) or AhR [8]. There exist multiple potential binding sites of PXR on the promoter region of p35 and Ebi3, two subunits of IL-35 (https://biogrid-lasagna.engr.uconn.edu/lasagna_search/). However, alone IAA did not cause signi cantly increased IL-35 + cells in vitro culture (Fig. 5a), suggesting that other transcription factors may be included. The generation of IL35 + B cells needs TLR4 [16], a receptor of LPS. A certain levels of plasma LPS were also often observed in humans and animals with metabolic syndrome [36,46,47] (Fig. S6a).
Thus, we tried to use IAA plus LPS to induce IL-35 + cells. Both IAA and LPS could markedly induce more IL-35 + B cells than LPS or IAA alone in vitro ( Fig. 5a and b). Furthermore, there existed markedly a dose response after exposed to different concentration of IAA in the presence of LPS (Fig. 5c). Since IAA can activate PXR or AhR [8], we used PXR or AhR inhibitor to observe their effects on the generation of IL-35 + cells. PXR inhibitor could produce stronger suppressive role in both IAA and LPS mediated IL-35 + B cells than AhR inhibitor ( Fig. 5a and b), suggesting that PXR may play a main role in inducing generation of IL-35 + cells. Meanwhile, NFkB inhibitor also produced marked suppression on the generation of IL-35 + cells ( Fig. 5a and b). Notably, IAA with LPS could effectively induce IL-35 + B cells but not IL-35 + T cells in vitro p35, a subunit of IL-35 can expanse IL-35-expressing regulatory B cells [19]. CHIP-sequencing and CHIP-PCR exhibited that PXR could markedly be enriched in the promoter region of p35 (a subunit of IL-35) in both IAA and LPS but not in IAA or LPS alone treated WEHI B cells ( Fig. 5d and e), implying that PXR may be a critical factor to regulate the expression of p35. After exposed to IAA plus LPS, PXR markedly increased in the cytoplasma and also in the nucleus (Fig. S7a), suggesting that IAA with LPS promotes the activity of PXR. While cells were exposed to IAA plus LPS, they also promoted both NFkBp65 and RXR to enter nucleus (Fig. S7a). Since that PXR exerts its transcriptional regulatory functions is through dimerization with retinoic X receptor RXR [48, 49], we used anti-PXR antibody to perform immunoprecipitation (IP) and check their binding. There indeed existed increased binding of PXR with RXR in nucleus after exposed to IAA plus LPS but not IAA or LPS alone ( Fig. 5f; Fig. S7b). These bindings could be impeded by PXR or NFkB inhibitor ( Fig. 5f), suggesting that not only PXR but also NFkB is also necessary for the complex. We also used anti-NFkB p65 to perform IP, results also exhibited markedly increased PXR in nucleus after exposed to IAA plus LPS (Fig. 5f). However, IAA with LPS also promoted binding of NFkBp65 to RXR (Fig. 5f), suggesting that NFkBp65 did not only bind with PXR but also with RXR after exposed to both IAA and LPS. Immunostaining also showed the increased binding of PXR, NFkB and RXR after exposed to LPS plus IAA (Fig. 5g). Thus, complexes composed of PXR, NFkB and RXR may be necessary for the expression of IL-35. In addition, inactive PXR is predominantly sequestered in the cytoplasm [50,51] by cytoplasmic androstane receptor (CAR) retention protein and heat shock protein 90 complexes [52,53]. IP using anti-PXR also exhibited the binding of PXR with CAR in cytosolic extract (Fig. 5h). Furthermore, this binding decreased in cytoplasma after exposed to IAA or IAA+LPS (Fig.  5h), indicating that active PXR increases. Immunostaining also showed increased PXR in the cytoplasma and nucleus after IAA plus LPS (Fig. 5i). All of these suggest that IAA can induce PXR to enter nucleus to promote IL-35 expression through binding with NFkB p65 and RXR.

IAA with LPS mediated IL-35 + B cells need PXR and TLR4
To further demonstrate that IAA and LPS mediated IL-35 cells is dependent on PXR and NFkB p65, we employed PXR KO and TLR4 KO mice (TLR4 can activate NFkB p65). Since active PXR is regulated by the CAR retention protein [52,53], the effects of CAR KO on IL-35 + cells are also observed. Meanwhile, we also observed AhR KO mice, in which AhR is potentially recognized by IAA [8]. The spleen cells from WT, PXR KO, TLR4 KO, CAR KO and AhR KO mice were cultured in the medium with IAA plus LPS. IAA plus LPS did not effectively induce IL-35 + B cells from spleen cells of PXR KO or TLR4 KO mice; whereas splenic cells from WT mice could produce more IL-35 + B cells in the presence of IAA with LPS ( Fig. 6a; Fig. S8a).
Notably, CAR KO had also signi cant effects on the generation of IL-35 + B cells as compared to WT mice ( Fig. 6a; Fig.S8a), consistent with above ndings (Fig. 5). After in vivo injecting IAA plus LPS into PXR KO, CAR KO, TLR4 KO and also AhR KO mice, IAA plus LPS could promote generation and accumulation of IL35 + cells in WT mice but not in PXR KO and TLR4 KO mice ( Fig. 6b and c; Fig. S8b and d). Higher levels of IL-35 could be detected in the spleen, Payer patch (PP) and peripheral blood of WT but not in PXR KO and TLR4 KO mice (Fig. 6d). Again there were less IL-35 in the spleen, PP and peripheral blood of CAR KO mice (Fig. 6d). Immunostaining also further con rmed the increased IL-35 + CD19 cells in the spleen of WT mice injected by IAA with LPS (Fig. S8c). In addition, AhR KO also affected generation of IL-35 + cells in vitro and in vivo in some degree (Fig. 6a-d), suggesting that AhR is partly involved in IAA plus LPS mediated IL-35 + cells.
Since the proportion of lactobacilli was signi cantly higher in the ileum and colon of huREG4 IECtg mice, we isolated one strain of dominant lactobacillus from fresh stool of huREG4 IECtg mice, Lactobacillus reuteri, which can produce IAA ( Fig. S9a and b). We also generated mutated lactobacillus (lactobacillus DiaaM ), which could not produce IAA. While these lactobacilli or lactobacilli DiaaM were infused into mice, lactobacilli could induce the production of IL-35 in the presence of LPS; whereas lactobacilli DiaaM. did not ( Fig. 6e; Fig. S9c and d). Notably, lactobacilli did not induce the generation of IL-35 + B cells in PXR KO or TLR4 KO mice ( Fig. 6e; Fig. S9c and d). There were higher levels of IL-35 cytokine in the spleen, PP and peripheral blood of WT but not PXR KO or TLR4 KO mice infused lactobacillus (Fig.   6f). Unlike to lactobacillus, infusion of lactobacillus DiaaM did not affect the levels of IAA in spleen, PP and peripheral blood of mice (Fig. 6g). Thus, IAA generated by Reg4 associated lactobacillus could promote the generation of IL-35 + cells in WT mice but not in PXR KO or TLR4 KO mice. Taken together, IAA can induce the generation of IL-35 + B cells through PXR and TLR4.
Lower levels of IAA are detected in the peripheral blood of individuals with obesity Individuals with obesity are also associated with gut microbiota and metabolites [54,55]. The systematic review indicates that probiotics, which could produce IAA, may have bene cial effects on weight loss in overweight adults [56]. Thus, we investigated the relevance of gut microbiota-derived IAA with body weight in 208 individuals with different body mass index (BMI). The concentrations of microbiota metabolites IAA were lower in individuals with increased BMI (Fig 7a). Indeed, individuals with obesity have lower IAA (Fig 7b). Since increased IAA could potentially induce generation of IL-35 + cells in human, we also detected the levels of IL-35 in peripheral blood of individuals with overweight or obesity. Unexpectedly, IL-35 was very low in all individuals with different BMI. Thus, there exist lower levels of IAA in the individuals with overweight or obesity.

Discussion
Here we found that Reg4 derived from gut epithelial cells is related to HFD-mediated obesity. There exist a markedly accumulation of IL-35 + cells in the adipose tissues and gut tissues and other organs such as spleen in huREG4 IECtg mice, which are resistant to HFD-mediated obesity; whereas the number of IL-35 + cells in Reg4 KO mice is less than their control WT mice. Altered gut microbiota in huREG4 IECtg or Reg4 KO mice have direct relationship with IL-35 + cells. We also found the accumulation of IL35 + cells is related to lactobacillus derived metabolites IAA. IAA can induce generation and accumulation of IL35 + B cells in the presence of LPS through activating PXR. In addition, lower levels of IAA are also found in the individuals with obesity. These data suggest that there may exist an axis of Reg4 /gut microbiota / IAA/IL-35 + B cells/resistance to HFD-mediated obesity.
Gastrointestinal bacteria/metabolites participate in B cell differentiation, maturation and activation [13,14]. Mice depleted of endogenous bacteria following administration of broad-spectrum antibiotics do not develop Il-10 producing Bregs [57]. We here demonstrate that IAA produced by Reg4 associated contributes to the transcriptional program of IL-10-producing regulatory B cells [15]. We found that IAA can induce the IL-35 + B cells through activating PXR receptors.
Our results suggest that lactobacillus producing IAA may have a negative relationship with development of obesity. The peripheral blood of individuals with metabolic syndrome exhibits lower levels of IAA in human and mice. Others also found that the abundance of IAA and body mass index (BMI) show strong negative correlation [7].

Conclusion
We here found that gut Reg4 can promote resistance to high-fat diet induced obesity and accumulation of IL-35 + cells in adipose tissues. We demonstrate that gut Reg4 associated microbiota such as lactobacillus can promote generation of IL-35 + B cells through producing 3-idoleacetic acid (IAA) in the presence of LPS. Lower levels of IAA could be detected in the peripheral blood of individuals with obesity.

Methods
All regents and oligoes used in this study were listed in supplementary Table S1.

Mice
Four-to six-week-old male or female C57BL/6 mice were obtained from Nanjing Animal Center, Nanjing, China; AhR KO mice were from the Third Military Medical University, Chongqing, China; PXR KO mice and CAR KO mice from Chinese Academy of Inspection and Quarantine ,Tianjin, China; TLR4 KO mice were from Shanghai Model Organisms Center, Shanghai, China. All experimental litters were bred and maintained under speci c pathogen-free conditions in Nankai University. Experiments were carried out using age-and gender-matched mice. All the procedures were conducted according to the Institutional Animal Care and Use Committee of the Model Animal Research Center. Animal experiments were approved by the Institute's Animal Ethics Committee of Nankai University. All experimental variables such as husbandry, parental genotypes and environmental in uences were carefully controlled. Human REG4 transgenic mice ( huREG4IEC tg) mice were prepared by Nanjing Animal Center. HD5 promoter, which may speci cally promote the REG4 expression in gut Paneth cells, was conjugated into Pinsulator-pHD5-promoter-CDS-poly plasmid. The fragment of REG4-CDS and polyA were cloned into HD5 promoter-Pinsulator. This conjugation was demonstrated using primers (REG4-HD5-tF: CATCCAACTCCAGGACGGAGTC and REG4-HD5-tR: CACCTGTAACATTGGCACTTTG) and sequencing using the primers (Promoter-cx tF: GTTTGCTGGGTCAGAACTGA and Promoter-cx tR: GTAATTTAGGTGCGTACAGCC). HuREG4IEC tg mice were identi ed using the following primers (HD5-Test-Tf:CTGTCTCAGGTCTTCTCCCAG and REG4 -PA-F: GATCTTTTTCCCTCTGCCAAA, which produced 323bp band in tg mice and no band in WT mice). Human REG4 tg mice or WT siblings were from a cross between WT mice and human REG4 tg mice (heterozygous mice).

Human samples
For human serum collection, 208 adult participants, among which 42 with BMI<18.5 kg/cm2, 41 with BMI 18.5-24 kg/cm2, 43 with BMI 24-27 kg/cm2, 42 with BMI 27-30 kg/cm2 and 40 with BMI ≥30 kg/cm2 were selectively recruited. More than half (54.3%) were males and the mean age was 42 years (SD=13 years). The mean BMI was 25.05±5.07 kg/cm2. All participants were free of acute stress conditions such as fever and diarrhea. Height and weight were measured to the nearest 0.1cm and 0.1 kg without shoes or heavy clothing using a calibrated stadiometer (GL-310, Seoul, Korea). Participants were instructed to fast for ≥ 12 hours before blood sampling the next morning. This study was conducted with approval from the Institutional Review Board at Nankai University, Tianjin Union Medical Center and Tianjin First Central Hospital. Participants were recruited from the health screening centers of Tianjin Union Medical Center and Tianjin First Central Hospital. All participants provided written informed consent.

Mouse models
For high-fat diet model, 6-8 week old male mice and their control littermates were fed using high-fat diet (D12492, protein 26.2%, carbohydrate, 26.3% and Fat 34.9%) and control diets (D12450B(60% of calories may be derived from fat), which was from Research Diets, Inc..(NJ, USA).
For microbiota transplantation, 6-to 8-week-old mice were treated with pan-antibiotics (ampicillin (A, 1 g/L, Sigma), vancomycine (V, 0.5g/L), neomycin sulfate (N, 1 g/L), and metronidazole (M, 1g/L)) via the drinking water. Water containing antibiotics was exchanged every three days. To con rm the elimination of bacteria, stool was collected from antibiotic-treated and untreated mice and cultured in anaerobic and aerobic condition. The bacteria were counted under microscope. Then cecal contents from detergent treated mice or bacteria (1×109) were suspended in 1 ml PBS with 30% glycerol. Mice (4 weeks old) were removed from the isolator and were orally administered 200 ml of fecal suspension or bacteria made using glycerol stocks.
For lactobacillus transplantation, mice were treated with pan-antibiotics via the drinking water for two weeks. After con rming the elimination of bacteria, 1×109 CFU/0.2ml PBS Lactobacillus was infused into mice with intraperitoneal injection with 2mg/kg LPS (once/ week for three weeks). After infusing lactobacillus, mice were fed with enriched (0.48%) tryptophan diets.
For in vivo administration of IAA plus LPS, mice were randomly divided into 4 groups, normal group with 0.2 ml PBS only, IAA group i.p with 500 mg/kg IAA diluted in DMSO, LPS group with 2 mg/kg O111:B4 in 0.2ml PBS, and IAA plus LPS group with 500mg/kg IAA and 2 mg/kg O111:B4. After administration for 24 hrs, spleen and PP tissues were separated for further analyses.
For WEHI231 B cells stimulation, WEHI231 B cells were seeded in 24-wells plate. The cells were then stimulated with IAA and LPS with or without PXR inhibitor or NF-κB inhibitor for 3hrs or 6 hrs, and harvested for WB, CHIP-SEQ and CHIP-PCR analyses.

Metabolism experiments
For glucose and insulin tolerance, after 5 hrs fasting, baseline blood glucose levels were measured using a Nova Max Plus GlucoseMeter. Mice were then injected intraperitoneally with 2 g glucose per kg body weight in sterile PBS or with 0.5U insulin per kg body weight (Sigma, St. Louis, Missouri), and blood glucose levels were measured at different times after injection.

Gut microbiota analyses
Previous protocol was used in gut microbiota analyses (Qi et al, 2019). Bre ely, Gut microbiota was analyzed by Majorbio Biotechnology Company (Shanghai, China) using primers that target to the V3-V4 regions of 16S rRNA. Once PCR for each sample was completed, the amplicons were puri ed using the QIAquick PCR puri cation kit (Qiagen Valencia, CA), quanti ed, normalized, and then pooled in preparation for emulsion PCR followed by sequencing using Titanium chemistry (Roche, Basel Switzerland) according to the manufacturer's protocol. Operational Taxonomic Unit (OTU) analysis was performed as follows: sequences were processed (trimmed) using the Mothur software and subsequently clustered at 97% sequence identity using cd-hit to generate OTUs. The OTU memberships of the sequences were used to construct a sample-OTU count matrix. The samples were clustered at genus and OTU levels using the sample-genus and sample-OTU count matrices respectively. For each clustering, Morisita-Horn dissimilarity was used to compute a sample distance matrix from the initial count matrix, and the distance matrix was subsequently used to generate a hierarchical clustering using Ward's minimum variance method. The Wilcoxon Rank Sum test was used to identify OTUs that had differential abundance in the different sample groups.
For lactobacillus isolation, 100 mg fresh stool samples were collected and diluted in 2 ml BPS solution, and cultured on Rogosa SL selective medium (Sigma-Aldrich) for Lactobacillus enumeration, and then colonies were identi ed and puri ed using 16s ribosomal DNA sequence analyses for the speciation of colonial genotype. The lactobacilli were cultured in deMan, Rogosa, Sharpe (MRS; 3 M Health Care, St. Paul, MN) media and also grown on MRS agar containing either 10% sucrose. Anaerobic conditions were generated with sachets of AnaeroPack-Anaero (Mitsubishi Gas Chemical, Japan) in an air-tight jar. At 24 h of cultivation in liquid media, isolated lactobacillus Ruteri could reach 1 × 109 CFU/ml).

IaaM gene deletion in lactobacillus
For lactobacillus iaaM gene deletion, the upstream and downstream fragments of the iaaM gene of lactobacillus were rst ampli ed. The puri ed upstream and downstream homologous fragments were inserted into the Xoh I, Pem I and Sac I, Bgl II digestion sites of the pNZ5319 plasmid, respectively.
Receptive cells of lactobacillus were prepared and the recombinant pNZ5319 plasmid was electrotransferred into the receptive cells of lactobacillus with the electrotransfer parameters of 1.7 kV (2 mm electrode cup), 200 Ω resistance and 25 μF capacitance. Single colone of chloramphenicol-resistant lactobacillus were selected, and the iaaM-up-F/R, iaaM-down-F/R, CM-F/R primers were used to validate the strains with single exchange. The single-exchange strains were cultured for 3 generations per day at 30°C, and the suspensions were taken at 40 generations for double-exchange strain screening. The colonies that could grow normally on plates containing chloramphenicol resistance but not on solid plates containing erythromycin resistance were selected. The screened double exchangers were veri ed using iaaM-F/R PCR. Double-exchanger positive bacteria were prepared as receptor cells and electrotransfected with pNZTs-Cre plasmid for 3 generations per day to about 10 generations, eliminating the chloramphenicol resistance gene from the genome and veri ed using CM-F/R PCR. The heat-sensitive plasmid pNZTs-Cre was eliminated by incubation in an incubator at 42°C for 3-5h.

LC-MS (Liquid chromatography-mass spectrometry)/MS
The samples were thawed at room temperature. All samples were extracted with 300 L of methanol, and For absolute cell number count, spleen tissue and whole PP nodes of mice were weighed and single-cell suspensions were prepared for ow staining analysis. The total number of cells per gram of spleen or PP node was counted and then multiplied by the proportion of positive cells to obtain the absolute number of cells.

H & E staining
For hematoxylin/eosin (H&E) staining, previously reported methods were used in this experiment (Su et al, 2014). Brie y, adipose tissues were xed in 4% (w/v) paraformaldehyde buffered saline and embedded in para n, 5 µm sections colon sections were cut and stained with H&E.

Immunostaining
For immune staining, previous reported protocol was used(Shang et al, 2019). Brie y, colon tissues or spleen were embedded in OCT compound (Tissue-Tek, Torrance, CA) and frozen over liquid nitrogen. 5μm-thick sections were prepared from frozen tissue and xed in acetone (−20°C) for 10 min. After rehydration in PBS for 5 min and further washing in PBS, tissue sections were blocked with 1% (w/v) BSA and 0.2% (w/v) milk powder in PBS (PBS-BB). The primary antibody was added in PBS-BB and incubated overnight at 4°C. After washing (three times, 5 min each), tissue was detected with DAB kit or uorescence labeled second antibody. Nuclei were stained by DAPI.

Immunoprecipitation and immunoblot
Immunoprecipitation and immunoblot were performed according to previous methods (Su et al, 2014).
The cells were lysed with cell lysis buffer (Cell Signaling Technology), which was supplemented with a protease inhibitor 'cocktail' (Calbiochem). The protein concentrations of the extracts were measured using a bicinchoninic acid assay (Pierce). For the immunoblot, hybridizations with primary antibodies were conducted for 1 h at room temperature in blocking buffer. The protein-antibody complexes were detected using peroxidase-conjugated secondary antibodies (Boehringer Mannheim) and enhanced chemiluminescence (Amersham).   WT mice fed HFD for three months (mixed sample from 14 mice). g ELISA of IL-35 in huREG4IECtg WT(WT1) and Reg4 KO or WT (WT2) mice with (HFD) or without (Normal) feeding HFD for three months.

CHIP-seq and CHIP-PCR
Scale bars=40 µm Student's t-test in all panels, mean ±SD; Data are a representative of three independent experiments; R. E, relative expression. *P < 0.05, **P < 0.01, and ***P < 0.001; NS, No signi cance. Student's t-test in all panels, mean ±SD; Data are a representative of three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001; NS, No signi cance.  IAA with LPS promotes binding of PXR with NFkB p65 and RXR. a Flow cytometry of p35+CD19+ cells in the spleen cells after exposed to IAA, LPS, LPS+IAA LPS+IAA+AhR inhibitor, LPS+IAA+PXR inhibitor, LPS+IAA+NFkB inhibitor for 24 hrs. NC, negative control. b Flow cytometry of p35+Ebi35+CD19+ cells in the spleen cells after exposed to IAA, LPS, LPS+IAA LPS+IAA+AHR inhibitor, LPS+IAA+PXR inhibitor, LPS+IAA+NFkB inhibitor for 24 hrs.NC, negative control. c ELISA of IL-35 in the supernatants of spleen cells after exposed to different concentration of IAA in the presence of LPS or different concentration of LPS in the presence of IAA. d Chip-sequencing analyses of B lymphocyte line WEHI 231 B cells after exposed different stimulators LPS, IAA or LPS +IAA for 6 hrs. e Chip-PCR of PXR binding site on the promoter region of p35 in the WEHI 231 cells after exposed to different stimulators LPS, IAA or LPS +IAA for 6 hrs. f Immunblotting of p65, RXRa or PXR in the cytosolic or nuclear extract of immunoprecipitation by anti-PXR or anti-p65. Iso, isotypic antibody. g Immunostaining of PXR, p65 and RXRa in the WEHI231 B cells after exposed to IAA+LPS. Ctr., no stimulators. h Immunoblotting of CAR and PXR of cytosolic extracts and nuclear extracts after immunoprecipitation using anti-PXR in the WEHI231 B cells after exposed to IAA or IAA+LPS. i Immunostaining of PXR and CAR in the WEHI231 B cells after exposed to IAA+LPS. Student's t-test in a, b and e, mean ±SD; Analysis of variance test in c. *P < 0.05, **P < 0.01, and ***P < 0.001, NS; No signi cance.