Fecal microbiota transplantation increases colonic IL-25 and dampens tissue inflammation in patients with recurrent Clostridioides difficile

Study Design

The study is registered in ClinicalTrials.gov (Identifier: NCT02797288). The goal of this study was to identify the immune mechanisms underlying successful FMT for rCDI. Specifically, we tested if FMT increased IL-25 signaling, which has been identified as initiating a protective type 2 immune response in the mouse model of CDI. Participants were recruited from patients undergoing FMT treatment for recurrent C. difficile infection as a part of their medical care at the University of Virginia Health System (UVAHS). All patients underwent standard vancomycin antibiotic treatment prior to FMT. Colonic biopsies and whole blood were collected immediately prior to FMT and 60 days after FMT. A total of 10 participants were recruited (Table 1). Biopsies were obtained from the sigmoid colon in subjects at the time of fecal transplantation. Follow-up biopsies were obtained from the sigmoid colon 60 days after FMT for six subjects. The primary outcome measure was colonic tissue IL-25 concentration. It was hypothesized that successful FMT would restore IL-25 and IL-4 in the colon. Biopsies taken for research purposes at each colonoscopy were analyzed to determine tissue levels of immune signaling proteins IL-25, IL-4, and IL-1, gene expression by RNA-sequencing, infiltrating and circulating immune cells using high dimensional flow-cytometry, and microbiota changes by 16s rDNA sequencing.

Luminex Immunoassay

Colonic biopsies were immediately stored at -80° C, dry, until protein extraction. The biopsies were lysed in MAP lysis buffer (Millipore) with HALT proteinase inhibitors (ThermoFisher) using 5 mm steel beads (Qiagen) with a bead-beater (Tissuelyzer II, Qiagen). After centrifugation, the supernatants were transferred to -80° C until use in immunoassays. The lysates were thawed and a 47-plex magnetic bead-based Luminex immunoassay kit (Milliplex, Millipore EMD) was used to measure protein concentrations. The mean fluorescence intensity (MFI) for each protein was read using Luminex MagPix and analyzed using Milliplex Analyst 5.1 (Millipore EMD), which uses a 5-log standard fit to calculate protein concentration from MFI. For MFIs outside the range of the standards, the concentrations were extrapolated using their respective standard curves (derived from the Milliplex Analyst, back calculated using Matlab, Mathworks). Differences in protein concentrations between pre and post FMT were tested for significance (alpha = 0.05) using a linear mixed effect model, where patient ID was included as a random variable (Table S2). IL-1a, IL-1b, IL-4, and IL-25 were defined a priori in the analysis plan so the p-values were not corrected for multiple comparisons. A Benjamini-Hochberg adjustment was used for the p-values calculated for the 43 remaining proteins since they were not pre-selected.

Messenger RNA-Seq

Colonic biopsies were obtained during colonoscopy and immediately placed in allprotect (Qiagen) and stored at -80 C until nucleic acid extraction (Qiagen AllPrep). A total amount of 1 μg RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using NEBNext Ultra RNALibrary Prep Kit for Illumina (NEB, USA) following manufacturer’s recommendations. Briefly, mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). First strand cDNA was synthesized using random hexamer primer and M-MuLV Reverse Transcriptase (RNase H-). Second strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of 3’ ends of DNA fragments, NEBNext Adaptor with hairpin loop structure were ligated to prepare for hybridization. In order to select cDNA fragments of preferentially 150∼200 bp in length, the library fragments were purified with AMPure XP system (Beckman Coulter, Beverly, USA). Then 3 μl USER Enzyme (NEB, USA) was used with size-selected, adaptor ligated cDNA at 37 °C for 15 min followed by 5 min at 95 °C before PCR. Then PCR was performed with Phusion High-Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer using the NEBNext Multiplex Oligos for Illumina. PCR products were purified (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 system. Clustering of index-coded samples was performed on a cBot Cluster Generation System using PE Cluster Kit cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on a MiSeq (Illumina) and 150 bp paired-end reads were generated.

mRNA-Seq Analysis

A total of 13 samples were processed, 1 from pre FMT and 1 from post FMT for each of 6 patients except for patient 3 which had 2 samples from pre FMT as a control for batch effects. Raw reads were processed through fastp (Chen, Zhou et al. 2018) to remove adapter, poly-N sequences, and reads with low quality. Q20, Q30 and GC content of the clean data were calculated, and only high-quality reads were preserved. Paired-end clean reads were aligned to the reference genome Homo Sapiens (GRCh37/hg19) using the Spliced Transcripts Alignment to a Reference (STAR) software (Dobin, Davis et al. 2013). FeatureCounts was used to quantify reads mapped to each gene (Liao, Smyth et al. 2014). Read counts were processed using the bioconductor package DESeq2 v1.30.1 in R (version 4.0.5) and normalized using the DESeq algorithm. Principal component analysis (PCA), hierarchical clustering, and density maps(Kolde and Kolde 2015) assessed overall similarity between samples and was used to determine the importance and influence of various factors, including the effects of FMT, patient, and persistent IBS symptoms following FMT. From the PCA, top positive and negative loadings for components 1 and 2 were reported. For the density maps, a heatmap was constructed from the top 30 most variable genes. For each gene, the standard deviation of counts across all samples was calculated as a measure of variability. Differentially expressed genes were calculated using the Wald test in DESeq2 (Love, Anders et al. 2014). We used two models, a grouped model and a patient-specific model. The grouped model contrasted between pre and post FMT while controlling for patient effects by adding patient ID as a separate variable to the model (design = ∼Patient+Visit) (n=13). The patient-specific model contrasted pre and post FMT for each paired patient (6 total) by using a term combining patient ID and whether the sample was pre or post FMT (design = ∼Patient_Visit). Genes with log₂fold change differences (FC) ≥ 1 and using the Benjamini–Hochberg false discovery rate correction (FDR, 0.1) were considered significant.

Gene Set Enrichment Analysis

Gene set enrichment analysis was performed on a list of all genes and their respective log2FCs in the grouped model. The package fgsea v3.13. (Korotkevich, Sukhov et al. 2021) in R was used to analyze the fold changes of the genes to identify enriched gene pathways and functions using KEGG pathways. P values are calculated using an enrichment score statistic (Korotkevich, Sukhov et al. 2019). Enrichment maps were constructed to determine functional modules using the emapplot function within the enrichplot v1.10.1 package () in R. The resulting enriched pathways and their corresponding gene lists were then selectively narrowed for each patient using the patient-specific model to measure the extent of enrichment for a pathway for each patient. As a measure of overall FC, the log FCs for the differentially expressed genes were summed for each enriched gene pathway of interest for each patient with paired RNA-seq data between pre and post FMT. This sum represented the extent of enrichment for a given pathway for that patient.

Flow Cytometry of LPMCs and PBMCs

Lamina propria mononuclear cells (LPMCs) were isolated from colonic biopsies as follows. Colonic biopsies were submerged in ice-cold HBSS without calcium and magnesium and immediately transported to the laboratory for cell disassociation. The biopsies were digested at room temperature for 20 minutes (Accumax, eBiosciences), filtered (70 µm), to isolate LPMCs. For PBMC isolation, whole blood was centrifuged, and the plasma removed. The PBMCs were isolated by adding Ficoll (Cytiva) and then centrifuged in SepMate tubes (Stemcell Technologies). Isolated LPMCs and PBMCs were cryopreserved in liquid nitrogen until staining.

Isolated LPMCs and PBMCs were thawed and stained with fluorochrome-conjugated antibodies (Table S1). Appropriate controls (unstained, single stained, and fluorescence minus one (FMO) stained PBMCs as controls. In addition, each sample also underwent a fixed viability stain (Zombie NIR, Biolegend). Cells were first stained with surface stains, then permeabilized (Foxp3 Transcription Factor Staining Buffer Set, Invitrogen) for intracellular staining. Samples were analyzed with the 5 Laser Cytek Aurora Borealis flow cytometer. All cells were collected for LPMC samples and 100,000 cells for PBMC samples. These fluorescence reads were then analyzed using FlowJo V.10.7.2. to phenotype immune cell subsets via tSNE clustering and traditional gating (Figure S2). Spectral deconvolution and gating were based on single stained and fluorescence minus one (FMO) stained PBMC control samples.

For each tSNE cluster or immune cell phenotype of any given sample, the raw counts and percentage of leukocyte population (viable CD45+ cells) was calculated. Differences in these raw counts and percentages between pre and post FMT were quantified using a linear mixed effect model, where patient ID was included as a random variable.

16 S rDNA Sequencing and analysis of epithelial-associated microbiota

Bacterial DNA from colonic biopsies was amplified using V4 specific primers and indexed using Nextera XT Index Kit, which was then sequenced by Illumina via MiSeq sequencing (Kozich, Westcott et al. 2013). The amplification was insufficient from one post FMT sample, narrowing the number of samples analyzed (10 pre FMT and 5 post FMT). A mock sample (ZymoBIOMICS) was included as a control. The resulting library was pre-processed through the DADA2 package (Callahan, McMurdie et al. 2016) in R to form a library of amplicon sequence variants (ASVs). Taxonomy was assigned with Silva (Quast, Pruesse et al. 2012) release 138 databases. Forward reads were truncated at 240bp while reverse reads were truncated at 160bp. A paired-end library was used with expected amplicon size of around 250bp. This length was confirmed after merging forward and reverse reads in the DADA2 pipeline.

Alpha diversity measures Shannon and Simpson diversity values were calculated using phyloseq (McMurdie and Holmes 2013) v.1.34.0 library in R. Differences in alpha diversity measures were assessed between pre and post FMT using the pairwise.wilcox.test base function in R with a Bonferroni correction for multiple comparisons. This function uses a paired, ranked, Mann-Whitney test to evaluate differences.

Principal coordinate analysis (PCoA) was performed to visualize the differences in relative abundance of pre and post FMT microbiota. Differences between pre and post FMT were evaluated by permutational multivariate analysis of variance (permANOVA) in the vegan package (Dixon 2003) v.2.5.7 of R.

Participants

We recruited 10 patients undergoing FMT to treat rCDI infections, of these 90% had at least 2 recurrences of rCDI before FMT. Of the 10 participants recruited for the study, 6 completed both visits, while 4 did not return for the 60-day follow-up visit after FMT. All patients had zero CDI recurrences and zero hospitalizations in the 60-day follow-up period post FMT. Additionally, FMT prevented further CDI recurrences for at least 2 years. However, though every patient was clinically cured with an absence of diarrhea and negative CDI test, of the 6 that completed both visits, two patients had persistent symptoms consistent with post infectious irritable bowel syndrome (IBS) as defined by Rome IV criteria (Drossman and Hasler 2016). The 6 participants that completed the second visit were female. The clinical characteristics are summarized in Table 1.

FMT Increased Colonic Type 2 Cytokines

To evaluate changes in immune cell signaling in response to FMT, we measured 47 cell signaling proteins in colonic biopsy tissue lysates (Supplemental Table 1). IL-25 and IL-4 cytokines were of particular interest since they were related to eosinophil recruitment and a type 2 immune response, which we predicted a priori to be restored by FMT (Buonomo, Cowardin et al. 2016). IL1 has also been shown to be important in determining type-1 vs type-2 immune responses, so IL-1a and IL-1b were also selected a priori for analysis in the trial outcomes.

We found that protein concentrations of IL-1a, IL-1b, and IL-25 were significantly increased after FMT (linear mixed effect model, p<0.05) with a trend to increased IL-4 concentrations after FMT (p = 0.058) (Figure 1). Taking out the two patients that developed IBS-like symptoms, we found that the increase in IL-1a and IL-1b were no longer significant (p>0.1, p>0.05, respectively) and IL-4 increase was still not significant (p>0.1), while the increase in IL-25 was still significant (p<0.05), suggesting IBS-like symptoms may influence the concentrations of specific cytokines.

• Download figure
• Open in new tab

Figure 1. FMT increased Th2 cytokine levels in the colon.

Levels of tissue cytokines were quantified by Luminex assay. The median value and IQR of tissue cytokines before and after FMT are shown. Linear mixed effect model with patient as a random variable was the statistical test used to calculate p-values. There were no significant differences in total protein concentration between pre and post FMT in the samples used for the Luminex assays (not shown). MFIs outside the standard curve range were extrapolated (triangles).

In the parallel transcriptomic analysis, while IL-25 transcripts could not be identified, FMT modestly increased the expression of IL1a (log₂FC = 0.83) and IL1b (log₂FC = 0.47) though these increases were not significant (Supplemental Figure S1A and B) and none of these cytokines were identified as differentially expressed genes. There was an increase in the transcript levels of the IL1 inhibitor IL1RA (encoded by IL1RN, log₂FC = 0.83) following FMT (Supplemental Figure S1C). IL1A transcripts were negatively correlated with IL-1a protein levels (r<-0.6), while IL1B transcripts were positively correlated with IL-1b protein levels (r>0.6) (Supplemental Figure S1D). IL1RN transcripts were positively correlated with IL1RA protein concentration (r> 0.7) (Supplemental Figure S1D).

Transcriptional Responses to FMT

To assess the transcriptional response in intestinal epithelial cells we compared gene expression in colonic biopsies obtained immediately prior to FMT to those obtained 60 days post FMT. PCA of the normalized gene counts from RNA-seq was done to investigate unsupervised groupings by visit, patient, or development of IBS-like symptoms post FMT (Figure 2A). Principal component 2 (PC2) separated the samples into pre vs post FMT while PC1 was related to patients that developed IBS-like symptoms post FMT-like symptoms. Negative loading unique to PC1 was driven by OLFM4 (Figure 2B). OLFM4 encodes olfactomedin-4, a selective marker of inflammation in the colonic epithelium (Liu and Rodgers 2016). The largest positive loading, LYZ, encodes for lysozyme which has been found previously to be expressed at lower levels in IBS patients relative to healthy controls (Aerssens, Camilleri et al. 2008). The top loadings uniquely driving PC2 were ST6GAL2 (negative) and HLA-DRB5 (positive). ST6GAL2 encodes a sialyltransferase. Specifically in relation to the gastrointestinal system, sialic acid catabolism has been shown to be related to intestinal inflammation and microbial dysbiosis (Krzewinski‐Recchi, Julien et al. 2003, Huang, Chassard et al. 2015, Kim 2020) though the specific role of ST6GAL2 has not yet been established. HLA-DRB5 encodes for an HLA class II protein which is specific to antigen presentation for immune cells and is related to ulcerative colitis (Silverberg, Cho et al. 2009). Three genes overlapped between PC1 and PC2 loadings: MTRNR2L12, MTRNR2L8, and MTRNR2L2, which encode isoforms of the mitochondrial peptide humanin, which has been shown to have cytoprotective effects (Bodzioch, Lapicka-Bodzioch et al. 2009).

• Download figure
• Open in new tab

Figure 2. FMT driven variation in the colonic transcriptome.

A) Principal component analysis of gene expression data identifies separation between pre (red) and post (green) FMT samples. Batch effects were minimal, as noted by how closely pre FMT patient 3 (P03) samples from 2 different batches grouped. PC1 captured some variance induced by the development of IBS-like symptoms post FMT in patients 2 (P02) and 5 (P05) clustered on the negative side of PC 1. PC 2 captured the variance related to FMT, where each patient consistently had a relatively higher PC2 value before FMT and a lower PC2 value after FMT. B) PC1 loadings were driven by OLFM4 (positive) and LYZ (negative) while PC2 loadings were driven by HLA-DRB5 (positive) and ST6GAL2 (negative). C) Volcano plot of differentially expressed genes after FMT. Orange indicates genes that met the threshold of log₂FC≥1, FDR≤0.1 (Wald test, Benjamini-Hochberg adjustment). Genes mentioned in the main manuscript are labeled here, which include homeobox genes, laminins, and some genes related to immune responses. D) Overrepresented pathways from pre to post FMT related to bile acids had overlapping genes. Using the FCs from DEGs from pre to post FMT, overrepresented pathways were found using KEGG gene lists (enrichment score test, Benjamini-Hochberg adjustment). E) Overlaps in pathways were mapped to find common themes.

We also performed hierarchical clustering of the 30 most variable genes before and after FMT (Figure S2). The dendrogram branched first between a group without any patients that developed IBS-like symptoms post FMT-like symptoms (right) and a group that included the two patients that developed IBS-like symptoms (left). For the no IBS-like symptom group, the second level of hierarchy was the split between pre and post FMT samples, before splitting by patient. The group that contained samples from patients that developed IBS-like symptoms were then split by patient before splitting into pre vs post FMT. The two patients that developed IBS-like symptoms post FMT did not cluster together suggesting patient-specific transcriptional responses vary even within patients with similar IBS-like symptoms.

FMT induces transcriptional programs of tissue repair and suppresses inflammatory responses

We analyzed differentially expressed genes (DEGs) between pre and post FMT using a grouped model comparing all patients before and after FMT. Based on the unsupervised analyses of PCA (Figure 2A) and hierarchical clustering (Supplemental Figure 2) we incorporated patient ID into the model to account for patient-to-patient variability. A total of 103 DEGs were identified (log₂FC≥1, FDR≤0.1) of which 62 were induced while 41 were repressed by FMT (Figure 2C). Some notable induced genes include homeobox genes (EVX2, HOXD10, HOXD12, HOXD13) and laminin genes (LAMC2 and LAMA3). Homeobox genes have been shown to be related to development and differentiation of the intestinal epithelium (Beck 2002, de Santa Barbara, Van Den Brink et al. 2003), while laminins are components of the basal lamina which have been shown to be absent in ulcerative colitis (Teller and Beaulieu 2001). Some notable repressed genes include HLA-DRB5, CCL11, and SLC9A3. HLA-DRB5 is a class II allele involved in antigen presentation to CD4 T cells and appeared in the above PCA analysis as being a positive loading to a component separating pre and post FMT. CCL11, also known as eotaxin-1, is pro-inflammatory and involved in eosinophil chemotaxis (Figure 2C) (Menzies-Gow, Ying et al. 2002). SLC9A3 is related to bile acid elimination and diarhea (Hegyi, Maléth et al. 2018).

FMT suppresses bile acid secretion and olfactory transduction pathways

To identify functional differences in gene expression due to FMT, the differentially expressed genes (DEGs) between pre and post FMT were analyzed for overrepresented pathways. The DEGs from the grouped model controlling for patient differences was used because this model was representative of significant gene expression differences common across all patients. Gene set enrichment analysis was analyzed in the KEGG human pathway database (Kanehisa and Goto 2000). A total of 7 pathways were significantly overrepresented (FDR < 0.1) among genes suppressed after FMT (Figure 2D). Functional overlaps in enriched pathways were identified for 6 of the 7 pathways found (Figure 2E). The overlapping pathways were (in order of decreasing gene set size): bile secretion, metabolism of xenobiotics by cytochrome P450, chemical carcinogenesis – DNA adducts, drug metabolism – cytochrome P450, steroid hormone biosynthesis, and pentose and glucuronate interconversions. Bile secretion and cytochromes P450 are related to oxidative stress, toxin elimination, and drug processing (Gonzalez and Gelboin 1994, Lorbek, Lewinska et al. 2012). Bile secretion can also be directly regulated by monooxygenases like cytochromes P450 (Lorbek, Lewinska et al. 2012). The repression of these pathways indicates a heightened transcriptional response against foreign invaders or the build-up of reactive oxygen species before FMT and a return to homeostasis after FMT. Of these overlapping pathways, the two most significant pathways were the bile secretion pathway (FDR = 0.074 enrichment score = -0.65) and the olfactory transduction pathway (FDR = 0.035, enrichment score = -0.74), which was not functionally linked to any of the other overrepresented pathways. Secondary bile acids have been shown to inhibit cell division of C. difficile (Kang, Myers et al. 2019). It has been previously shown that olfactory receptors are present all along the gastrointestinal tract and can be modulated by the intestinal microbiota (Priori, Colombo et al. 2015). The activation of odorant receptors has been linked to the inhibition of cell proliferation and apoptosis(Weber, Al-Refae et al. 2017).

Peripheral Th17 subsets are decreased following FMT

To complement the whole tissue RNA-seq analyses, we used high dimensional flow cytometry to profile immune cell subsets in colonic biopsies and blood from before and after FMT. Lamina propria mononuclear cells (LPMCs) were isolated from colonic biopsies, and peripheral blood mononuclear cells (PBMCs) were isolated from venous blood samples. Overall, we isolated an average of 3291+/-2481 (mean +/-SD) viable CD45+ cells from the LPMCs from each patient, and an average of 255397 +/-80841 (mean +/-SD) viable CD45+ cells from the PBMCs. t-SNE plots were used to visualize cell populations in LPMC and PBMC samples and did not reveal any immune cell subsets unique to pre or post FMT (Supplemental Figure S2). Statistical analyses could not be used to compare cluster to cluster between pre and post FMT as specific subsets of interest appeared in various proportions overlapping one another and tSNE does not account for differences in the relative number of cells. Therefore, we used traditional gating to test for differences in subsets of interest. This revealed a significant decrease (p = 0.05, linear mixed effect model with patient as random variable) in the peripheral Th17 (CD3+CD19/CD20-CD4+RORyt+) after FMT (Figure 3L), a similar trend was noted in LPMCs but this did not reach statistical significance (Figure 3E). Despite this increase, there were no statistically significant differences in Th1:Th2, Th17:Treg ratios in LPMCs or PBMCs (not shown).

• Download figure
• Open in new tab

Figure 3.

Peripheral Th17 cells were decreased after FMT (p < 0.05). A decrease in Th17 cells was observed for all patients except for patient 7, who had about the same number of Th17 cells pre and post FMT. The number of Th1 and cytotoxic T cells (CTL) were decreased post FMT, however these decreases were not statistically significant. There were no significant differences in the number of viable CD45+ cells between pre and post FMT nor were there significant differences in Th17 cells between patients that developed IBS-like symptoms post FMT and those that did not (p > 0.1). A linear mixed effect model with patient as a random variable was used to calculate p values.

FMT increased microbial diversity

To evaluate the local changes in the microbial population by FMT, 16s rDNA sequencing was performed using DNA isolated from biopsy samples. The use of a biopsy allows us to investigate the mucosal microbial population in close association with the colonic epithelium, as opposed to the lumen of the gastrointestinal tract. We found that the alpha diversity measures Shannon and Simpson indices significantly increased after FMT (Mann-Whitney, p < 0.01) (Figure 4A and B), which agreed with other studies on diversity changes after FMT treatment for rCDI (Chang, Antonopoulos et al. 2008, Staley, Kaiser et al. 2018). For the beta diversity measures, clear groupings between pre and post FMT samples were evident in a PCoA plot (Figure 4C). The PCoA axis 1 explained 26.4% of the variance while axis 2 explained 16.9% and separated the patients by FMT status (permANOVA, alpha = 0.01). Patients that developed post IBS-like symptoms following FMT did not appear distinct in the PCoA analysis.

• Download figure
• Open in new tab

Figure 4. FMT increased alpha diversity of the epithelial-associated microbiome.

(a)Increased Shannon and Simpson(B) Alpha Diversity measures were observed in 4 of the 5 patients analyzed (p < 0.01, Mann-Whitney). Patient 2 (P02) showed no change in diversity post FMT and developed IBS-like symptoms post FMT. (C) PCoA plot shows separation of relative abundance of microbiota populations before and after FMT. This separation was statistically significant (p < 0.01, permANOVA). One point that overlapped between pre and post FMT was from a pati1en1t (P5) that developed IBS-like symptoms post FMT.

The top 20 most abundant ASVs across all samples were grouped by taxonomic information and ranked. In the most abundant ASVs, the top families were Enterobacteriaceae, Akkermansiaceae, Bacteroidaceae, Clostridiaceae, Acidaminococacear, Rikenellaceae, Fusobacteriaceae, Lachnospiraceae, Ruminococcaceae, and Campylobacteraceae, in decreasing order of abundance. The top genera from the same list of ASVs were Escherichia/Shigella, Akkermansia, Bacteroides, Clostridium_sensu_stricto_13, Phascolartobacterium, Alistepes, Lachnoclostridium, Clostridium_sensu_stricto_1, Agathobacter, Faecalibacterium, Fusicatenibacter, Campylobacter and Cetobacterium, in decreasing order of abundance. Some notable taxa include the genera Escherichia/Shigella, Akkermansia and Bacteroides, which had different abundances between pre and post FMT. Escherichia/Shigella was found to be more abundant prior to FMT, which agrees with previous studies (Ling, Liu et al. 2014, Li, Gao et al. 2018). Akkermansia decreased after FMT while Bacteroides was increased after FMT. Akkermansia muciniphila plays a role in degrading mucins in the gut and is generally thought to have a beneficial effect on the gut to reduce inflammation (Schneeberger, Everard et al. 2015, Li, Lin et al. 2016, Zhao, Liu et al. 2017). Bacteroides has also been shown to have beneficial effects on the gut and has been shown to increase after FMT. C. difficile has also been shown to reduce Bacteroides growth (Khoruts, Staley et al. 2021). All patients were treated with vancomycin to treat C. difficile before FMT (Table 1) which may have had impacted the microbiota. A total of 15 ASVs were found to be differentially abundant between pre and post FMT (FDR < 0.05, Table 3). All differentially abundant ASVs were also present as one of the top 20 most abundance ASVs. The average transcriptional fold change of genes related to bile secretion was positively correlated (r>0.5) to differences in Akkermansia (Supplemental Figure S1E). Akkermansia abundance was negatively correlated (r <-0.7) with IL-25 protein concentration (Supplemental Figure S1F).