Exclusive Enteral Nutrition Exerts Anti-Inflammatory Effects through Modulating Microbiota, Bile Acid Metabolism, and Immune Activities

Exclusive enteral nutrition (EEN) can induce remission in patients with pediatric Crohn’s disease (CD). This study aims to depict EEN’s modification of bile acid (BA) metabolism in pediatric CD and explores the effect of the EEN-enriched BA in inhibiting the inflammatory response. The twelve enrolled pediatric CD patients showed BA dysmetabolism, represented by decreased levels of fecal secondary and unconjugated BAs as determined by UPLC–TQMS, which were accompanied by gut microbiota dysbiosis and reduced BA-metabolizing bacteria including Eubacterium and Ruminococcus genera, assessed by shotgun metagenomic sequencing. EEN treatment induced remission in these patients at eight weeks, and nine patients remained in stable remission for longer than 48 weeks. EEN improved BA dysmetabolism, with some enriched BAs, including hyocholic acid (HCA), α-muricholic acid (αMCA), strongly associated with decreased severity of CD symptoms. These BAs were significantly correlated with the increased abundance of certain bacteria, including Clostridium innocuum and Hungatella hathewayi, which express 3β-hydroxysteroid dehydrogenase and 5β-reductase. HCA could suppress TNF-α production by CD4+ T cells in the peripheral blood mononuclear cells (PBMCs) of CD patients. Moreover, intraperitoneal injection of HCA could attenuate dextran sulfate sodium (DSS)-induced mouse colitis. Our data suggests that BA modification may contribute to the EEN-induced remission of pediatric CD.


Introduction
A steadily increasing incidence of inflammatory bowel disease (IBD) worldwide over the past four decades indicates its emergence as a global disease [1,2]. Crohn's disease (CD) is characterized by transmural inflammation, causing thickening and narrowing of the gastrointestinal tract (GI) wall and eventually leading to the disabling development of deep ulcerations, fistulae, strictures, and abscesses. A defect in barrier integrity is closely associated with profound alterations in the intestinal metabolome, including a shortage of short-chain fatty acids (SCFA) [3], altered concentrations of amino acids and lipid acids [4], and dysregulation of bile acid (BA) pool composition [5]. These metabolic abnormalities are proved to be partially driven by gut microbiota dysbiosis [6][7][8][9].

Analysis of Fecal Microbiota with Shotgun Metagenomic Sequencing
MicrobiomeAnalyst [28] was used to analyze the bacterial community composition data [29]. Data with low count (minimum count < 4 and prevalence in samples < 20%) and low variance (less than 10% inter-quantile range measure of variance) were filtered and were further normalized based on centered log-ratio (clr) transformation to account for the compositional nature of the metagenomic data. Microbiome alpha diversity was measured using the Shannon's, Chao1, Simpson's, and Inverse Simpson's indexes. The beta diversity among samples was calculated through principal coordinate analysis (PCoA) to Euclidean distance based on clr. Permutational multivariate analysis of variance (PERMANOVA) was carried out to test whether the gut microbiome structure was significantly different. Statistical differences in alpha diversity and abundances of taxa were assessed via Mann-Whitney U tests and Kruskal-Wallis tests, with p-values adjusted using the Benjamini-Hochberg correction procedure for controlling the false discovery rate (FDR).

Quantification of Fecal Bile Acids with Ultra-Performance Liquid Chromatography Coupled with Triple Quadrupole Mass Spectrometry
Fecal samples were kept frozen at −80 • C and then lyophilized overnight in a freeze dryer system (Labconco FreeZone 2.5 Plus, Kansas City, MO, USA). BAs were extracted using a two-step extraction procedure as previously described [30,31]. In brief, an aliquot of 10 mg of lyophilized feces from each sample was homogenized in 20 mL of deionized water. Then, 180 µL of acetonitrile:methanol = 80:20 containing the six internal standards (CA-d4, GCA-d4, GCDCA-d4, GDCA-d4, LCA-d4, and UDCA-d4 at 50 nM each) was added, homogenized for 5 min, and centrifuged at 13,200 rpm at 4 • C for 15 min. After transferring the supernatant to a 1.5 mL Eppendorf tube, the residue was further extracted by adding 180 µL acetonitrile:methanol = 80:20 containing the six internal standards as the first extraction solvent. The two supernatants were combined and vortexed for 3 min before centrifugation at 13,200 rpm at 4 • C for 15 min.
MetaboAnalyst 5.0 [32] was used for statistics and feature selection of the metabolomic data. Univariate statistical analysis was performed using Mann-Whitney U tests and Kruskal-Wallis tests with p-values adjusted using the Benjamini-Hochberg correction procedure for controlling FDR.

Peripheral Blood Mononuclear Cell Isolation and Stimulation
Heparin whole blood was collected from five adult treatment-naïve CD patients. PBMCs were isolated from the whole blood by centrifugation using Biocoll Separating Solution (BiochromGmbH, Berlin, Germany). All samples were incubated for 4 h in a medium consisting of RPMI with 10% fetal bovine serum and 1% penicillin-streptomycin (all three: Gibco, Carlsbad, CA, USA) at 37 • C in HERA Cell 150 (Thermo Fisher Scientific, Waltham, MA, USA) after adding CD3/CD28-coated magnetic beads (1 µg/mL). In addition, Brefeldin A (Sigma-Aldrich, St. Louis, MO, USA) dissolved in dimethyl sulfoxide (Sigma-Aldrich) was added to all flow cytometry samples after 2 h to prevent cells from secreting tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ).

Enzyme-Linked Immunosorbent Assay
The cytokine production from PBMCs was assessed using ELISA assay (BioGems, Westlake Village, CA, USA) following the manufacturer's instructions. Supernatant and standard controls (100 µL prepared according to kit instructions) were transferred to a pre-coated primary antibody (IFN-γ or TNF-α) stripwell microplate, incubated at 37 • C for 90 min, and washed. The avidin-biotin-peroxidase complex was added at 100 µL/well and incubated at 37 • C for 30 min. After additional washes, 90 µL of color developing reagent was added to each well and incubated in the dark for 30 min. The reaction was terminated by the addition of 100 µL stop solution, and O.D. absorbance was read with a microplate reader at 450 nm.

Dextran Sulfate Sodium Colitis Model
C57/BL6 male mice (6-7 weeks old) were purchased from GemPharmatech (Nanjing, China). For acute colitis induction, mice were administered 3% (w/v) DSS (36-50 kDa; MP Biomedicals, Santa Ana, CA, USA) in drinking water for nine days. For the treatment groups, HCA (30 mg/kg body weight, Sigma-Aldrich, St. Louis, MO, USA) was intraperitoneally injected from day four to day nine. Saline was injected as the vehicle in the control group. Body weight was monitored daily to assess the severity of intestinal colitis. The colon length was measured at sacrifice, which was performed on day nine. The animal experimental procedures were approved by the Laboratory Animal Ethics Committee of Shenzhen University.

Statistical Analysis
Continuous variables were presented as medians with interquartile range (IQRs, 25th to 75th percentiles), and categorical variables were summarized as absolute numbers with percentages. SPSS software (version 20.0) was used for statistical analyses. Nonparametric Mann-Whitney U tests and Kruskal-Wallis tests were used to analyze continuous variables between two groups and among multiple groups, respectively. The categorical variables were compared using the chi-squared test. The p-values were adjusted using the Benjamini-Hochberg correction procedure for controlling FDR. Statistical significance was defined by a p-value of less than 0.05.

Gut Microbiota Dysbiosis and Bile Acid Dysmetabolism in Pediatric CD Patients
The enrolled pediatric patients with CD showed an altered gut microbiota profile compared to healthy individuals, which was characterized by a reduction in biodiversity ( Figure 2). Decreased abundance at the bacterial species level, including Anaerostipes hadrus, Blautia wexlerae, Faecalibacterium prausnitzii, Bacteroides uniformis, Eubacterium

Gut Microbiota Dysbiosis and Bile Acid Dysmetabolism in Pediatric CD Patients
The enrolled pediatric patients with CD showed an altered gut microbiota profile compared to healthy individuals, which was characterized by a reduction in biodiversity ( Figure 2). Decreased abundance at the bacterial species level, including Anaerostipes hadrus, Blautia wexlerae, Faecalibacterium prausnitzii, Bacteroides uniformis, Eubacterium hallii, Ruminococcus bromii, Fusicatenibacter saccharivorans, Bifidobacterium longum, and Bifidobacterium pseudocatenulatum (Figure 3), and enrichment of genera Escherichia, Peptostreptococcus, and Morganella, which are known to contain a variety of opportunistic pathogens, were observed in pediatric CD patients (Supplementary Figure S2).

Gut Microbiota Dysbiosis and Bile Acid Dysmetabolism in Pediatric CD Patients
The enrolled pediatric patients with CD showed an altered gut microbiota profile compared to healthy individuals, which was characterized by a reduction in biodiversity ( Figure 2). Decreased abundance at the bacterial species level, including Anaerostipes hadrus, Blautia wexlerae, Faecalibacterium prausnitzii, Bacteroides uniformis, Eubacterium hallii, Ruminococcus bromii, Fusicatenibacter saccharivorans, Bifidobacterium longum, and Bifidobacterium pseudocatenulatum (Figure 3), and enrichment of genera Escherichia, Peptostreptococcus, and Morganella, which are known to contain a variety of opportunistic pathogens, were observed in pediatric CD patients (Supplementary Figure S2).

Figure 2.
Alpha diversity in healthy children and pediatric CD patients before and after four and eight weeks of EEN treatment. Alpha diversity was measured using the Chao1 index (a) and Shannon index (b). Statistical differences were assessed via Mann-Whitney U tests and Kruskal-Wallis tests with p-values adjusted with the Benjamini-Hochberg correction procedure for controlling false discovery rate. *** p < 0.001. HC, healthy controls; CD, Crohn's disease; BSL, baseline.

EEN Induced CD Remission Is Associated with Coordinated Modifications in Gut Microbiota Composition and Bile Acid Metabolism
Significant changes in gut microbiota composition, but not in alpha diversity, were observed after EEN treatment (Figures 2 and 6). EENtreated CD patients showed increased relative abundance of Hungatella, Parvimonas, Clostridioides, Solobacterium, Clostridium, and Enterococcus genera after four or eight weeks of EEN (Supplementary Figure S2). At the species level, increased abundance of Actinomyces sp oral taxon 414, Clostridium innocuum, Clostridium perfringens, Clostridioides difficile, Hungatella hathewayi, Parvimonas micra, and Solobacterium moorei, and decreased levels of Bacteroides stercoris, Haemophilus parainfluenzae, and Veillonella atypica were detected after four or eight weeks of EEN treatment ( Figure 6). It is worth noting that neither toxin (TcdA and TcdB) was detected in the stool samples, nor were diarrhea or infection symptoms observed in these CD patients. Conjugated and unconjugated BA levels, and ratio of unconjugated to conjugated BAs. Significance was determined via Mann-Whitney U test. * p < 0.05; ** p < 0.01; **** p < 0.0001. BAs, bile acids; HC, healthy controls; BSL, baseline.

EEN Induced CD Remission Is Associated with Coordinated Modifications in Gut Microbiota Composition and Bile Acid Metabolism
Significant changes in gut microbiota composition, but not in alpha diversity, were observed after EEN treatment (Figures 2 and 6). EENtreated CD patients showed increased relative abundance of Hungatella, Parvimonas, Clostridioides, Solobacterium, Clostridium, and Enterococcus genera after four or eight weeks of EEN (Supplementary Figure S2). At the species level, increased abundance of Actinomyces sp oral taxon 414, Clostridium innocuum, Clostridium perfringens, Clostridioides difficile, Hungatella hathewayi, Parvimonas micra, and Solobacterium moorei, and decreased levels of Bacteroides stercoris, Haemophilus parainfluenzae, and Veillonella atypica were detected after four or eight weeks of EEN treatment ( Figure 6). It is worth noting that neither toxin (TcdA and TcdB) was detected in the stool samples, nor were diarrhea or infection symptoms observed in these CD patients.
EEN treatment induced remarkable changes in fecal BA metabolism in pediatric CD patients. Both primary and secondary BA levels increased after EEN treatment, but statistical significance was only achieved for secondary BAs (Figure 4a). However, the ratio of secondary to primary BAs did not significantly change after EEN treatment. The level of unconjugated BAs as well as the ratio of unconjugated to conjugated BAs increased significantly at four weeks post EEN treatment and maintained throughout the eight weeks of the therapeutic period (Figure 4b). Among the significantly increased BAs, the levels of HCA, αMCA, and 6−keto−lithocholic acid (6−ketoLCA) were elevated after four weeks of EEN in the pediatric CD patients, and persisted for eight weeks after EEN therapy (Mann-Whitney U-test, q < 0.05 and fold change > 2.0; Figure 7). Most notable is the higher concentration of HCA compared to the other BAs at week eight (mean ± sd, 123.88 ± 47.57 nmol/g). In addition, HDCA, one of HCA species, was also significantly   EEN treatment induced remarkable changes in fecal BA metabolism in pediatric CD patients. Both primary and secondary BA levels increased after EEN treatment, but statistical significance was only achieved for secondary BAs (Figure 4a). However, the ratio of secondary to primary BAs did not significantly change after EEN treatment. The level of unconjugated BAs as well as the ratio of unconjugated to conjugated BAs increased significantly at four weeks post EEN treatment and maintained throughout the eight weeks of the therapeutic period (Figure 4b). Among the significantly increased BAs, the levels of HCA, αMCA, and 6-keto-lithocholic acid (6-ketoLCA) were elevated after four weeks of EEN in the pediatric CD patients, and persisted for eight weeks after EEN therapy (Mann-Whitney U-test, q < 0.05 and fold change > 2.0; Figure 7). Most notable is the higher concentration of HCA compared to the other BAs at week eight (mean ± sd, 123.88 ± 47.57 nmol/g). In addition, HDCA, one of HCA species, was also significantly enriched after eight weeks of EEN (Mann-Whitney U-test, q < 0.05 and fold change > 2.0; Figure 7).   . Significantly changed BAs in CD patients after EEN therapy. Heatmap demonstrating significantly changed BAs (all upregulated) in CD patients after EEN induction therapy with fold changes > 2.0. Statistical significance was determined using two-tailed Mann-Whitney U-tests, with FDR < 0.05 and fold difference > 2.0. bUDCA, 3β-ursodeoxycholic acid; UDCA, ursodeoxycholic acid; ACA, allocholic acid; βMCA, β-muricholic acid; HCA, hyocholic acid; CA, cholic acid; CDCA, chenodeoxycholic acid; 6-ketoLCA, 6-ketolithocholic acid; 6,7-diketoLCA, 6,7-diketolithocholic acid; 12-DHCA, 12-dehydrocholic acid; HDCA, α-hyodeoxycholic acid; aMCA, α-muricholic acid; NorCA, norcholic acid.
negatively correlated with FCP, WBC, PLT, CRP, CDEIS, and ESR, indicating their potential contributions in improving CD symptoms (Supplementary Figure S3a). The relative abundance of Eggerthella lenta, Clostridium innocuum, and Gordonibacter pamelaeae inversely correlated with levels of PCDAI, FCP, WBC, and ESR (Supplementary Figure  S3b). Peptostreptococcus stomatis and Anaerostipes caccae were positively correlated with ESR, and Peptostreptococcus stomatis was also positively correlated with PCDAI, suggesting a pathogenic/pro-inflammatory potential in these species. We next tried to identify associations between changed gut bacteria and BAs in CD patients using Spearman's rank correlation coefficient. The results show that certain gut We next tried to identify associations between changed gut bacteria and BAs in CD patients using Spearman's rank correlation coefficient. The results show that certain gut bacteria were correlated with eight kinds of BAs that increased in CD patients after EEN therapy. For example, Hungatella hathewayi showed positive correlations with HDCA, 6-ketoLCA, and αMCA, Haemophilus parainfluenzae showed negative correlations with αMCA and βMCA, and Actinomyces sp ICM47 positively correlated with bUDCA and αMCA. Notably, there was a relative abundance of gut bacteria that were negatively corrected with PCDAI, including Eggerthella lenta, Clostridium innocuum, and Gordonibacter pamelaeae, which were positively correlated with levels of HDCA (Figure 9).

HCA Suppresses TNF Production in CD4 T Lymphocytes within Peripheral Blood of CD Patients
We hypothesized that specific EEN-enriched BAs could contribute to CD remission by suppressing intestinal inflammatory immune responses. To test this hypothesis, we further analyzed the immunity modulation effects of HCA in vitro. PBMCs isolated from five CD patients, and their characteristics and clinical information were included in Supplementary  Table S1. The addition of HCA significantly suppressed TNF-α production in PBMCs, while not affecting the levels of IFN-γ (Figure 10a). Moreover, flow cytometry analysis demonstrated that the incidences of IFN-γ-and TNF-α-producing cells in the CD4+ T cell population were significantly decreased due to HCA intervention (Figure 10b), while their incidences in CD8+ T cells were not affected in most of the tested cases (Figure 10c). bacteria were correlated with eight kinds of BAs that increased in CD patients after EEN therapy. For example, Hungatella hathewayi showed positive correlations with HDCA, 6-ketoLCA, and αMCA, Haemophilus parainfluenzae showed negative correlations with αMCA and βMCA, and Actinomyces sp ICM47 positively correlated with bUDCA and αMCA. Notably, there was a relative abundance of gut bacteria that were negatively corrected with PCDAI, including Eggerthella lenta, Clostridium innocuum, and Gordonibacter pamelaeae, which were positively correlated with levels of HDCA (Figure 9). Figure 9. Interrelationships among bile acid profiles, PCDAI and gut microbiota. Sankey plot constructed using NAMAP with Spearman's rank correlations between PCDAI, gut microbiota (species level), and fecal BAs that were significantly enriched after EEN induction therapy. The network shows associations that are statistically significant with a cutoff of p < 0.05 and r > 0.7 based on bootstrapping of 100 iterations. Direct correlations are indicated as blue edges and inverse correlations as red edges. 6,7−ketoLCA, 6,7−ketolithocholic acid; CA, cholic acid; CDCA, chenodeoxycholic acid; 6−ketoLCA, 6−ketolithocholic acid; HDCA, α−hyodeoxycholic acid; PCDAI, Pediatric Crohn's Disease Activity Index; aMCA, α−muricholic acid; βMCA, β−muricholic acid; bUDCA, 3β−ursodeoxycholic acid.

HCA Suppresses TNF Production in CD4 T Lymphocytes within Peripheral Blood of CD Patients
We hypothesized that specific EEN-enriched BAs could contribute to CD remission by suppressing intestinal inflammatory immune responses. To test this hypothesis, we further analyzed the immunity modulation effects of HCA in vitro. PBMCs isolated from five CD patients, and their characteristics and clinical information were included in Supplementary Table S1. The addition of HCA significantly suppressed TNF-α production in PBMCs, while not affecting the levels of IFN-γ (Figure 10a). Moreover, flow cytometry analysis demonstrated that the incidences of IFN-γ-and TNF-αproducing cells in the CD4+ T cell population were significantly decreased due to HCA intervention (Figure 10b), while their incidences in CD8+ T cells were not affected in most of the tested cases (Figure 10c). Figure 9. Interrelationships among bile acid profiles, PCDAI and gut microbiota. Sankey plot constructed using NAMAP with Spearman's rank correlations between PCDAI, gut microbiota (species level), and fecal BAs that were significantly enriched after EEN induction therapy. The network shows associations that are statistically significant with a cutoff of p < 0.05 and r > 0.7 based on bootstrapping of 100 iterations. Direct correlations are indicated as blue edges and inverse correlations as red edges. 6,7-ketoLCA, 6,7-ketolithocholic acid; CA, cholic acid; CDCA, chenodeoxycholic acid; 6-ketoLCA, 6-ketolithocholic acid; HDCA, α-hyodeoxycholic acid; PCDAI, Pediatric Crohn's Disease Activity Index; aMCA, α-muricholic acid; βMCA, β-muricholic acid; bUDCA, 3β-ursodeoxycholic acid.

HCA Relieves DSS-Induced Colitis in Mice
We used the DSS-induced mouse colitis model to further access the potential of HCA against colonic inflammation. We induced colitis in C57BL/6 mice by administering 3% DSS in drinking water for nine days. HCA was administrated on day four via intraperitoneal injection for five days at a daily dose of 20 mg/kg body weight. DSS-induced body weight loss was relieved by HCA administration (Figure 11a). On day nine, the average colon length was longer in the HCA-treated mice than in the DSS group, but a statistical difference was not achieved (Figure 11b). Thus, HCA treatment could alleviate DSS-induced colonic colitis in mice to a certain extent, suggesting a Figure 10. HCA suppresses the inflammatory response in PBMCs isolated from CD patients. (a) TNFα and IFN-γ production was measured via ELISA and compared between HCA-and DMSO-treated PBMCs (isolated from seven CD patients) after stimulation with CD3/CD28-coated magnetic beads. Proportions of TNF-α+ and IFN-γ+ subsets in (b) CD4+ and (c) CD8+ T cells measured by flow cytometry. The paired samples t-test was used to compare the two groups. * p < 0.015; ** p < 0.01. CD, Crohn's disease; PBMCs, peripheral blood mononuclear cells; TNF-α, tumor necrosis factor-α; IFN-γ, interferon-γ.

HCA Relieves DSS-Induced Colitis in Mice
We used the DSS-induced mouse colitis model to further access the potential of HCA against colonic inflammation. We induced colitis in C57BL/6 mice by administering 3% DSS in drinking water for nine days. HCA was administrated on day four via intraperitoneal injection for five days at a daily dose of 20 mg/kg body weight. DSS-induced body weight loss was relieved by HCA administration (Figure 11a). On day nine, the average colon length was longer in the HCA-treated mice than in the DSS group, but a statistical difference was not achieved (Figure 11b). Thus, HCA treatment could alleviate DSS-induced colonic colitis in mice to a certain extent, suggesting a potential protective/anti-inflammatory effect of HCA. necrosis factor-α; IFN-γ, interferon-γ.

HCA Relieves DSS-Induced Colitis in Mice
We used the DSS-induced mouse colitis model to further access the potential of HCA against colonic inflammation. We induced colitis in C57BL/6 mice by administering 3% DSS in drinking water for nine days. HCA was administrated on day four via intraperitoneal injection for five days at a daily dose of 20 mg/kg body weight. DSS-induced body weight loss was relieved by HCA administration (Figure 11a). On day nine, the average colon length was longer in the HCA-treated mice than in the DSS group, but a statistical difference was not achieved (Figure 11b). Thus, HCA treatment could alleviate DSS-induced colonic colitis in mice to a certain extent, suggesting a potential protective/anti-inflammatory effect of HCA.

Discussion
EEN was first introduced to treat active IBD in the 1970s and became the frontline therapy in the 1990s [33]. Current evidence indicates that EEN improves outcomes in CD patients by ameliorating micronutrient and macronutrient deficiencies, reducing antigenic load, exerting direct anti-inflammatory effects, and correcting gut microbiota dysbiosis [16,34]. However, the mechanisms of action of EEN remain elusive. In this study, we reported the therapeutic outcomes of EEN in pediatric CD, characterized the differences in gut microbiota and the altered BA pool in pediatric CD patients during

Discussion
EEN was first introduced to treat active IBD in the 1970s and became the frontline therapy in the 1990s [33]. Current evidence indicates that EEN improves outcomes in CD patients by ameliorating micronutrient and macronutrient deficiencies, reducing antigenic load, exerting direct anti-inflammatory effects, and correcting gut microbiota dysbiosis [16,34]. However, the mechanisms of action of EEN remain elusive. In this study, we reported the therapeutic outcomes of EEN in pediatric CD, characterized the differences in gut microbiota and the altered BA pool in pediatric CD patients during EEN therapy, and explored the effect of EEN-enriched HCA in inhibiting the inflammatory response. After eight weeks of EEN therapy, 91.7% (11/12) of pediatric CD patients achieved clinical remission, and 75% (9/12) achieved endoscopic remission, which was similar to previous studies [10,13,15]. EEN treatment significantly decreased serum and fecal inflammatory markers, including ESR, WBC, PLT, CRP, and FCP. Although BMI and STAMP scores displayed no significant improvement, other nutritional indexes (i.e., HB, ALB, and PA) significantly increased after EEN. Overall, our data revealed that EEN could induce clinical remission and endoscopic remission, decrease inflammatory markers, and partially improve the nutrition status of pediatric CD patients.
IBD-associated gut microbiota dysbiosis and impaired microbiota enzymatics lead to dysmetabolism, including reduced SCFA production and modifications in luminal BA pool composition [5,35]. Gut microbiota dysbiosis was repeatedly reported in pediatric CD patients and recognized as a crucial factor involved in the inflammation process and affecting the responses to multiple therapeutics [36]. In line with previous findings, the gut microbiota of pediatric CD patients showed decreased biodiversity, reduced levels of beneficial taxa such Bifidobacterium spp. and Faecalibacterium prausnitzii, and increased levels of proinflammatory bacteria such as Escherichia. Moreover, a reduced abundance of multiple BA-metabolizing bacteria, such as Eubacterium (which possess 7α-dehydroxylation) and Ruminococcus (which exhibit both 7α-and 7β-hydroxysteroid dehydrogenase (HSDH)) was detected in accordance with declined fecal levels of secondary and unconjugated BAs, confirming impaired BA metabolism. Our previous study suggested that the effects of anti-TNF agents for CD might be partially mediated by upregulating gut bacteria that produce bile salt hydrolases (BSH), thereby restoring BA metabolism and inhibiting inflammation [35].
In the present study, we observed that a range of BA-metabolizing bacteria increased in the gut microbiota of pediatric CD patients after EEN treatment, including several species of Clostridium (C. innocuum, C. difficile, and C. perfringens). Notably, Clostridium spp. are carriers of a range of genes involved in BA deconjugation and dehydroxylation [37]. For example, C. innocuum carries genes encoding enzymes 3β-HSDH, 7α-HSDH, 7β-HSDH and 5β-reductase (5BR) [38,39], thereby participating in BA transformations such as the production of oxo (keto) BAs. Gordonibacter pamelaeae and Hungatella hathewayi, two additional bacterial species increased under EEN, were also suggested to carry 3α/3β-HSDH and 5BR [39], respectively. Thus, EEN increased certain bacteria with enzymes involved in the production of these enriched BAs.
Gut microbiota dysbiosis in IBD can result in a profound impact in BA metabolism, including deficiencies in the deconjugation, dehydrogenation, and dehydroxylation of primary BAs. Altered BA metabolism, especially a reduced level of secondary BAs, is suggested to contribute to a pro-inflammatory state [21]. Deficiencies in DCA, LCA, and their metabolites, such as oxo derivatives, were suggested to promote intestinal inflammation in UC patients [40]. Duboc et al. reported increased levels of fecal conjugated BAs in active IBD, while secondary BAs were decreased [5]. Previous studies showed that some secondary bile acids, such as isoalloLCA, LCA, UDCA, and tauroursodeoxycholic acid (TUDCA), play important roles in maintaining intestinal homeostasis [41][42][43]. Connors et al. noted that patients who achieved and sustained remission after EEN had a fecal BA profile dominated by secondary BAs [44]. Our analysis also indicated a significant increase in secondary BA and unconjugated BA concentrations after EEN therapy. In addition, there were multiple EEN-enriched BAs that showed close associations with improvements in CD patients' inflammatory and nutritional indexes. Interestingly, we found that the concentration of HCA increased and persisted after eight weeks of EEN in pediatric CD patients, and HDCA also significantly increased after eight weeks EEN.
Previous studies suggested that some BAs could regulate host immune responses by modulating the Th17 and Treg balance [22,45]. On this basis, we attempted to explore the role of HCA in suppressing intestinal inflammatory immune responses. HCA was recently found to simultaneously activate TGR5 while inhibiting FXR in enteroendocrine cells [46], and both receptors were found to modulate the inflammation occurring in IBD [47]. CD has been associated with an inflammatory CD4+ T cell phenotype [48]. Our data showed an anti-inflammatory effect of HCA through inhibiting TNF-α production in human PBMCs stimulated with anti-CD/anti-CD28 monoclonal antibodies, which revealed the role of HCA in attenuating the inflammatory response. Moreover, the administration of HCA alleviated DSS-induced acute colitis in mice, suggesting a potential protective/anti-inflammatory effect of HCA. Further experiments are required to access the anti-inflammatory capacities of other EEN-enriched BAs, and the results may lead to targets for the development of new therapies for IBD.
Recent studies indicate that BA receptors are involved in regulating both innate and adaptive immunity [49][50][51]. We found that some EEN-enriched BAs in CD patients are agonists/antagonists of some BA receptors, such as FXR, TGR5, VDR, and RORγt (Supplementary Table S2). TGR5 activation promotes the differentiation of macrophages towards M2 phenotypes, thereby enhancing the development of Treg cells via IL-10 production and inhibiting the M1 phenotype, which is known to produce pro-inflammatory cytokines [52]. Activating FXR via its agonist was also shown to prevent gut barrier dysfunction [53]. FXR knockout mice developed an IBD-like phenotype and exhibited pronounced hepatic inflammation and fibrosis [54]. Among the BAs enriched after EEN treatment, CA and CDCA are agonists of FXR while the MACs including αMCA, βMCA, and HCA were found to inhibit FXR as antagonists [46,49]. In our previous study, B. longum CECT 7894 improved the efficacy of infliximab for DSS-induced colitis and increased the expression of FXR via regulating the gut microbiota composition and bile acid metabolism [55]. Vitamin D deficiency is associated with the onset and activity of IBD, mainly because vitamin D exerts a regulatory role in mucosal immunity and host defenses via VDR [56]. As a subset of CD4+ T cells , Th17 cells produce IL-17A, IL-17F, IL-21, IL-22, IL-26, and the chemokine CCL20, which are critically involved in the mucosal inflammation observed CD. The oxo-(or keto-) derivatives of BAs are shown to be antagonists of RORγt [21], and the EENenriched 6-keto-LCA and 6,7-diketo-LCA may inhibit the proinflammatory activities of Th17 cells in a RORγt-dependent manner. Taken together, these EEN-enriched BAs and the BA receptors they activate/inhibit are versatile in modulating the development and functions of the innate and adaptive immune system and thereby contributing to colitis remission and mucosa repair in pediatric CD patients.
This study has some limitations. First, this study was conducted at a single center with a relatively small cohort. Multicenter randomized controlled trials could be performed to minimize confounding factors. Second, the assessment of gut microbiota and BAs was not performed in sustained and non-sustained remission CD patients after EEN therapy. Third, for research on the mechanism by which EEN enriches BAs, we only demonstrated that HCA could inhibit pro-inflammatory cytokine production by regulating CD4 + T cells in vitro. Thus, further experiments are required to explore and detect which BA receptors HCA stimulates/inhibits to perform its immune regulatory functions.

Conclusions
In summary, we showed that EEN therapy altered microbiota composition and the BA pool in pediatric CD patients. EEN treatment induced remission in children with active CD partially by increasing a number of secondary BAs, such as HCA, which could suppress TNF-α production by CD4+ T cells in PBMCs of CD patients and attenuate DSS-induced acute colitis in mice. These BAs are likely produced by increased levels of certain bacteria, including Clostridium innocuum and Hungatella hathewayi, which express 3β-hydroxysteroid dehydrogenase and 5β-reductase. Thus, certain BAs may provide a possibility for improving anti-inflammatory and immunity modulation effects in pediatric CD patients in the future.  Figure S1. EEN-induced endoscopic improvement in pediatric CD. Endoscopic images showing the transverse colon in two patients and the ileocecal junction in one patient at baseline and eight weeks after EEN treatment. Supplementary Figure S2. Metagenomic analysis of the gut microbiota profiles in CD patients before and after EEN induction therapy. Heatmap demonstrates changes in the relative abundances of gut bacteria at the genus level during eight weeks of EEN induction therapy. Significance was determined by Mann-Whitney U test with p < 0.05 based on the clr-transformed counts. # Significantly different between CD and HC at baseline. * Significantly changed after either 4 w or 8 w of EEN therapy. Supplementary Figure S3. EEN-induced changes in gut microbiota and bile acid metabolism are closely associated with CD symptom improvement. Heatmap showing the Spearman correlations between the serum inflammatory markers and severity indexes of pediatric CD (PCDAI and CDEIS), and (a) BAs and (b) gut bacterial species. Significance was determined by t-test. * p < 0.05; ** p < 0.01; *** p < 0.001.
Author Contributions: X.G. and T.Z. designed the study; F.X. and X.G. analyzed the omics data and created figures; F.X., H.H. and J.L. collected and interpreted the clinical data; Y.C. and X.S. performed the animal experiments; Y.X. and Z.L. performed the in vitro experiment and flow cytometry analysis; F.X. and X.G. interpreted the experimental data and drafted the manuscript; Y.W. and T.Z. revised the manuscript; all authors read and approved the final manuscript. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement:
This study conformed to the ethical guidelines of the 1964 Declaration of Helsinki and its later amendments. The Institutional Review Board of Shanghai Children's Hospital approved the study protocol (No. 2020R034-E02, approval date 19 June 2020). The animal experimental procedures were approved by the Laboratory Animal Ethics Committee of Shenzhen University.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data that support the findings of this study are available from figshare at the DOI https://doi.org/10.6084/m9.figshare.19665972.v1. The raw data from this study will be published after acceptance of this paper for publication.

Conflicts of Interest:
The authors declare that they have no conflict of interest.