Fecal Microbiota Transplantation Stimulates Type 2 and Tolerogenic Immune Responses in a Mouse Model

Abstract

Objectives

Clostridioides difficile infection (CDI) is the leading hospital-acquired infection in North America. While previous work on fecal microbiota transplantation (FMT), a highly effective treatment for CDI, has focused on colonization resistance mounted against C. difficile by FMT-delivered commensals, the effects of FMT on host gene expression are relatively unexplored. This study aims to identify transcriptional changes associated with FMT, particularly changes associated with protective immune responses.

Methods

Gene expression was assessed on day 2 and day 7 after FMT in mice after antibiotic-induced dysbiosis. Flow cytometry was also performed on colon and mesenteric lymph nodes at day 7 to investigate changes in immune cell populations.

Results

FMT administration after antibiotic-induced dysbiosis successfully restored microbial alpha diversity to levels of donor mice by day 7 post-FMT. Bulk RNA sequencing of cecal tissue at day 2 identified immune genes, including both pro-inflammatory and Type 2 immune pathways as upregulated after FMT. RNA sequencing was repeated on day 7 post-FMT, and expression of these immune genes was decreased along with upregulation of genes associated with restoration of intestinal homeostasis. Immunoprofiling on day 7 identified increased colonic CD45+ immune cells that exhibited dampened Type 1 and heightened regulatory and Type 2 responses. These include an increased abundance of eosinophils, alternatively activated macrophages, Th2, and T regulatory cell populations.

Conclusion

These results highlight the impact of FMT on host gene expression, providing evidence that FMT restores intestinal homeostasis after antibiotic treatment and facilitates tolerogenic and Type 2 immune responses.

1. Introduction

Clostridioides difficile is a Gram-positive spore-forming obligate anaerobe that causes mild to severe diarrheal disease. C. difficile is ranked as an urgent threat to human health (1) and is associated with nearly half a million infections, 83,000 recurrences, and 29,000 deaths annually in the United States alone (2). The increased prominence of C. difficile infection (CDI) is due in part to the 2005 emergence of hypervirulent strains (3), which cause more severe disease and longer hospitalizations (4, 5). Due to the significant health impact and cost of CDI, development of highly effective treatments is critical. CDI is associated with disruption of the intestinal microbiota through antibiotic exposure (6, 7), allowing C. difficile to bloom and produce toxins, resulting in diarrhea, damage to the epithelium, and disruption of the intestinal barrier (8). Thus, maintaining or restoring microbiota homeostasis is critical for minimizing the risk of CDI.

While the standard treatment for CDI remains antibiotics (9), one in five patients experience recurrence (10). Because of this, fecal microbiota transplantation (FMT), the administration of a defined or whole microbial community from the stool of healthy donors into the patient’s gastrointestinal tract, has been explored as therapeutic for CDI. While FMT has been tested in a wide range of disorders (11), clinical trials have recently demonstrated limited success of FMT for the treatment of recurrent CDI: SER-109, a drug composed of purified Firmicutes spores reduced recurrence by 70% (12, 13); RBX2660, a pooled human donor microbiota preparation was successful in 50% (14, 15). The success of these trials makes understanding FMT’s mechanisms of action a priority.

Because CDI is initiated by antibiotic disruption of the intestinal microbiota (6, 7), research into FMT’s mechanism of action has primarily focused on its ability to restore the gut microbiome (16). While recurrent CDI is characterized by decreased microbial diversity (17) that can be reversed by FMT (18), it has recently been appreciated that the type of host immune response can also impact CDI severity (19, 20). While C. difficile toxins cause much of the virulence and epithelial barrier damage seen in CDI, an especially robust immune response and the resulting inflammation have been implicated in severe disease and worse clinical outcomes independently of bacterial burden (21–23). Specific types of immune responses have also been implicated in differential CDI outcomes: tissue-regulated Type 2 immunity has been associated with tissue repair and protection from severe disease (24–26), while Type 3 responses have been implicated in increased disease severity (27). The importance of Type 2 immunity in protection from acute CDI was validated in patients, where IL-25 was suppressed by dysbiosis and induced by fecal microbiota transplant (28), peripheral eosinophils were associated with survival (29), and blockade of IL-33-induced Type 2 immunity due to the decoy receptor sST2 was associated with severe CDI (26). Therefore, a balance of inflammatory and protective immune responses is critical.

While the ability of FMT to increase microbial diversity has been well established (30–32), its effect on the host transcriptome remains poorly understood. Further limiting our understanding of FMT is that many studies have investigated FMT in the context of other intestinal disorders, such as intestinal colitis models (33–35). In addition, studies often look well after FMT administration (36), making it difficult to distinguish the effects of FMT from the effects of gradual microbiota recovery in the absence of antibiotics. This study aims to address these gaps in knowledge by investigating the impact of FMT over the first week post-administration, while the intestinal microbiota is still recovering from antibiotic depletion.

Here we utilize a mouse model to explore the impact of FMT on host mucosal gene expression. FMT elicited significant changes in microbial composition and gene expression within 48 hours of administration. Transcriptional changes were initially enriched in immune activation genes, particularly those associated with pro-inflammatory and resolving Type 2 immune signatures. Immune gene upregulation was reversed by day 7, leading to suppression of immune activation and upregulation of genes associated with intestinal homeostasis and neuropeptide signaling. Immunoprofiling identified significant changes in immune cell populations post-FMT, confirming a more tolerogenic immune environment associated with increased Th2 and regulatory immune responses. This work highlights the impact of microbiota signaling on intestinal immune and homeostatic responses, identifying a potential mechanism by which FMT could protect against CDI.

2. Materials and Methods

2.1. Murine FMT Model

Murine experiments were carried out on eight-week-old male C57BL6/6J mice (Jackson Laboratory). Antibiotic treatment was adapted from a well-established CDI model (37) and has been previously published (26, 38). Starting at Day −6, mice were given an antibiotic cocktail in drinking water consisting of 45mg/L vancomycin (Mylan), 35mg/L colistin (Sigma), 35mg/L gentamicin (Sigma), and 215mg/L metronidazole (Hospira) for three days. Mice were then changed over to normal drinking water before a single IP injection (0.016mg/g) of clindamycin (Hospira) was administered on Day −1. All animal work complied with all relevant ethical regulations for animal testing and research procedures approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Virginia.

Mice were separated into either FMT recipient or vehicle control groups with an equal number of mice per group (n=10 for the Day 2 experiment; n=7 for the Day 7 experiment). On Day 0–1, FMT mice were given an oral gavage of 100μl FMT while control mice were gavaged with anaerobic PBS. FMT was prepared from fecal pellets from age- and sex-matched nonantibiotic-treated donor mice that were collected and immediately homogenized by vortexing in 1ml sterile anaerobic PBS. Mice were sacrificed on day 2 in the Day 2 experiment and on day 7 in the Day 7 experiment. In both experiments, cecal contents were collected for 16S V4 rRNA sequencing and harvested cecal tissue was immediately placed in RNAlater (Sigma) and stored at −80°C. For the Day 7 experiment only, colonic tissue and mesenteric lymph nodes were collected and immediately processed for flow cytometry.

2.2. 16S Sequencing and Data Analysis

DNA extraction and 16S V4 rRNA sequencing were performed as previously described (26, 31). DNA was extracted using the MagAttract PowerMicrobiome kit (Qiagen) with the EpMotion liquid handling system (Eppendorf). The V4 region of the 16S rRNA gene was amplified from each sample using a dual index sequencing strategy (39) and then samples were sequenced on the MiSeq platform (Illumina) using the MiSeq Reagent Nano V2 Kit according to manufacturer’s protocol with modifications found in the Schloss Wet Lab SOP (https://github.com/SchlossLab/MiSeq_WetLab_SOP). All analyses were performed in R (version 4.1.1) (40) using the DADA2 (v1.22.0) (41), phyloseq (v1.38.0) (42), and tidyverse packages (v2.0.0) (43) for read processing/taxonomic assignment, diversity analysis, and data organization, respectively.

2.3. RNA Isolation, Sequencing, and Analysis

RNAlater-stabilized cecal tissue was removed using sterile needles and placed in 1ml TRIzol per tissue. Tissues were homogenized using the TissueLyser II (Qiagen) and the aqueous phase was collected and purified using the RNeasy kit (Qiagen) according to manufacturer’s protocol. Sequencing of isolated RNA was performed by Novogene using a polyA enrichment and sequencing on a HiSeq platform (Illumina) with paired-end 150bp reads. Unprocessed FASTQ files were assessed for quality using FastQC (v0.12.1) (44) and MultiQC (v1.14) (45). Reads were pseudomapped to the murine genome (v109 from EnsemblDB) using Kallisto (v0.44.0) (46). Count tables were imported into R using TxImport (v1.22.0) (47) and the DESeq2 package (v1.34.0) (48) was used to exclude genes with low counts, normalize data, estimate dispersions, and fit counts using a negative binomial model. Differentially expressed genes were ranked according to Wald’s statistic from most upregulated to most downregulated post-FMT, and this ranked list was used for Gene Set Enrichment Analysis (GSEA) using the fgsea package (v1.20.0) (49) with Hallmark (50) or Gene Ontology (51) databases.

2.4. Flow Cytometry

Flow cytometry analysis was conducted on murine colonic and mesenteric lymph node (MLN) tissue. The MLN was carefully extracted and placed in 5ml of complete RPMI media with 10% FBS, pre-allocated into 6 well plates. Samples were kept on ice until the harvesting process was completed. Samples were subsequently passed through a 40μm strainer into a 50ml tube, then washed with 5ml PBS, which was also passed through the strainer into the sample tube. Samples were spun by centrifugation at 1500rpm for 7 minutes and resuspended in 1ml RPMI media. Resuspended samples were kept on ice until the initiation of the staining process for flow cytometry.

Colons were longitudinally dissected and rinsed in buffer containing HBSS, 25mM HEPES, and 5% FBS. Epithelial cells were isolated by placing tissue in dissociation buffer (HBSS with 15mM HEPES, 5mM EDTA, 10% FBS, and 1mM DTT) and incubating at 37°C and 122rpm agitation for 40 minutes. Manual dicing of the lamina propria was followed by digestion using RPMI 1640 with Liberase TL and DNase. Single-cell suspensions were obtained by passing digested cells through 100μM and 40μM cell strainers. Intracellular and extracellular staining markers are summarized in Supplemental Table 1. Surface staining incorporated Fc-blocking and LIVE/DEAD blue. Flow cytometry was performed on an Aurora Borealis 5 laser Spectral Flow Cytometer. All data analysis was performed using Omiq software.

2.5. Statistics

Statistical differences in alpha and beta diversity were calculated using a Wilcoxon Rank Sum test with a Benjamini-Hochberg False Discovery Rate (FDR) correction for multiple comparisons and PERMANOVA with pairwise comparisons, respectively. Differentially expressed genes from RNA sequencing results were calculated using Wald’s test with FDR correction for multiple comparisons. Statistics for transcriptomics and flow cytometry boxplots were performed using a Student’s t-test on normalized counts or cell counts per 100,000 cells, respectively.

2.6. Data Sharing

The raw 16S rRNA and bulk RNA sequencing data have been deposited in the Sequence Read Archive (SRA) of the National Center for Biotechnology Information (NCBI) under Bioproject accession number PRJNA1078834. Code for the organization and analysis of this data is available at https://github.com/petrilab-uva/2024-FMT-transcriptomics.

3. Results

3.1. FMT Restores Microbial Diversity After Antibiotic Treatment

To assess the effect of FMT on microbiome composition and host gene expression, we first induced dysbiosis using a previously described antibiotic regimen (26, 38) followed by administration of either FMT or PBS as a control (Fig. 1A). Mice were sacrificed on either day 2 or 7 in independent experiments using different batches of mice. Because of this, analysis focused primarily on differences between control and FMT-treated mice rather than changes across different days.

Figure 1: FMT restores the intestinal microbiome by day 7.

A) Overview of experimental design. C57BL6/6J mice were given an antibiotic cocktail in drinking water one week prior to treatment, followed by an IP injection of Clindamycin one day pre-FMT. On Day 0 and Day 1, mice were administered either FMT or vehicle (PBS) by oral gavage. In separate experiments, mice were sacrificed on either Day 2 or Day 7, at which time cecal contents, cecal tissue, and colon / MLN (Day 7 only) were harvested for downstream analysis. B,D) Alpha diversity (Total number of ASVs) on B) Day 2 or D) Day 7 for FMT donor (blue), FMT recipient (red), and PBS control (gray) groups. Statistics were calculated using a pairwise Wilcoxon Rank test with Bonferroni correction for multiple comparisons. C,E) Non-metric multidimensional scaling (NMDS) plot of Bray-Curtis Dissimilarity on C) Day 2 or E) Day 7 from FMT donor, FMT recipient, and PBS control groups. All groups were significantly different from each other at both timepoints using a PERMANOVA with pairwise comparison for multiple comparisons. Statistical significance is indicated as *p < 0.05, **p < 0.01, and ***p < 0.001.

The effect of FMT on microbiome alpha (within-sample) and beta (between-sample) diversity was assessed using 16S V4 rRNA sequencing. As expected, antibiotics disrupted the microbial community of control mice, resulting in a significant reduction in the number of Amplicon Sequence Variants (ASVs), or unique bacterial sequences (Fig. 1B). FMT treatment partially restored alpha diversity by day 2, with significantly increased ASVs compared to control mice. These trends were also observed in beta diversity, where FMT-treated mice had an intermediate diversity between controls and FMT donors (Fig. 1C). Three mice in the FMT group did not display increased alpha diversity and clustered with control mice via beta diversity, leading us to hypothesize that the FMT failed to engraft in these mice. Many biological factors have been associated with failure of FMT engraftment, including host immune status (52), donor microbial diversity (53), and differences in antibiotic regimen (54). However, these mice we co-housed in the same cage, leading us to hypothesize that the lack of microbial recovery may be due to a technical cage effect. On day 7, alpha diversity had fully recovered in FMT-treated mice, who exhibited no significant difference in observed ASVs compared to donor mice (Fig. 1D). Control mice maintained significantly decreased alpha diversity at day 7 compared to either donor or FMT-treated mice, indicating that microbiome disruption persisted at this timepoint. While beta diversity profiles for FMT-treated mice remained significantly different from donor mice at day 7, these groups clustered more closely along NMDS1 than control mice (Fig. 1E). Overall, these results confirm that FMT restores the intestinal microbiome after antibiotic treatment.

3.2. FMT-Treated Mice Upregulate Immune Activation Genes on Day 2

To better understand the effect of FMT on the host, bulk RNA sequencing (RNAseq) was performed on cecal tissue from a subset of mice (n=4 per group) sacrificed on Day 2. Control and FMT-treated samples clustered separately by Principal Component Analysis (PCA) (Fig. 2A), indicating that groups exhibited distinct transcriptional profiles. This was confirmed using differentially expressed gene (DEG) analysis, which identified 720 DEGs across conditions, with 351 upregulated and 369 downregulated in the FMT group (Supplemental Table 2).

Figure 2: FMT treatment is associated with activation of immune genes within one week post-FMT.

A) Principal component analysis (PCA) of log transformed count data from bulk RNA sequencing of control (gray) or FMT-treated (red) mouse cecal tissue at day 2. B) Gene Set Enrichment Analysis (GSEA) ranking of significantly upregulated or downregulated Hallmark gene sets in the FMT-treated conditions according to Normalized Enrichment Score (NES) at day 2. C) GSEA enrichment plot for the Reactome “Interleukin-4 and Interleukin-13 Signaling” gene set. Lines along the x axis represent the position of genes in this gene set within a ranked list of RNAseq-identified genes. D-F) Boxplots of normalized counts for leading edge genes associated with enrichment of D) immune activation, E) tissue remodeling, or F) Type 2 immunity gene sets. P values represent the results of a Student’s t-test between normalized counts across groups. Statistical significance is indicated as *p < 0.05, **p < 0.01, and ***p < 0.001.

Next, Gene Set Enrichment Analysis (GSEA) was performed to identify biological pathways with significantly altered gene expression. GSEA compares a list of RNAseq genes, which has been ranked from most upregulated to most downregulated after FMT treatment, to publicly available gene sets, allowing for enrichments to be identified (55). This analysis was performed using the Hallmark gene set, which provides concise gene sets representing core biological functions for enrichment analysis (50). Immune activation was identified as a major upregulated biological process following FMT, with significant enrichment of IFNγ Response, IL-6 Jak/Stat3 Signaling, and TNFα Signaling (Fig. 2B) gene sets. Pathways associated with Epithelial-Mesenchymal Transition and Allograft Rejection were also highly enriched after FMT treatment. Previous work from our group has identified Type 2 immune responses as protective during CDI (24, 26), and several genes associated with these responses (Il4ra, Il13ra1) were prominent drivers of enrichment in these pathways. However, because the Hallmark gene set did not contain a formal set of Type 2 immune genes, we could not determine whether this process was enriched in our data. We therefore took a targeted approach and examined a manually curated IL-4/IL-13 pathway gene set from the Reactome database (56). GSEA using this gene set identified Type 2 immune genes as significantly enriched post-FMT (Fig. 2C), suggesting that activation of immune genes was not limited to pro-inflammatory immune responses.

Leading edge genes, those most responsible for the enrichment of a gene set in GSEA, were next examined to identify the specific genes driving enrichment. Immune activation enrichment was driven by upregulation of a core set of pro-inflammatory signaling genes, including Bcl3, Myd88, and Stat3 (Fig. 2D). Enrichment of these gene sets as well as the Epithelial-Mesenchymal Transition gene set was also driven by genes associated with tissue remodeling and wound repair, including the matrix metalloprotease Mmp3 and inhibitors Timp1 and Serpine1 (Fig. 2E). While Mmp3 and Timp1 contributed to enrichment of the IL-4/IL-13 Signaling gene set, this pathway was also driven by genes encoding key members of the Type 2 immune response, including both components of the IL-4/IL13 receptor (IL4ra and IL13ra1) and the Type 2 cytokine Il33 (Fig. 2F). These results indicate that FMT-upregulated genes are significant mediators of immune activation, tissue remodeling, and Type 2 immunity.

3.3. Transcriptional Changes at Day 7 Suggest Intestinal Recovery Post-FMT

Upregulation of pro-inflammatory pathways in the FMT-treated group was unexpected, as FMT is commonly associated with regulatory immune responses (34). We hypothesized that upregulation of these genes was due to the acute activation of inflammation by Microbial Associated Molecular Patterns (MAMPs) or host-associated Damage Associated Molecular Patterns (DAMPs) from within the FMT and that these genes would be downregulated at later timepoints. To determine whether these changes in gene expression persisted over time, RNAseq was repeated in an independent experiment from mice sacrificed at day 7. Control and FMT-treated samples clearly separated by PCA (Fig. 3A), indicating that transcriptional differences between groups persist at day 7. These differences were more modest than on day 2, as only 233 genes were differentially expressed, with 103 upregulated and 130 downregulated in the FMT-treated group (Supplemental Table 3).

Figure 3: FMT promotes transcriptional changes in immune and intestinal homeostasis genes at day 7.

A) Principal component analysis (PCA) of log transformed count data from bulk RNA sequencing of control (gray) or FMT-treated (red) mouse cecal tissue at day 7. B) Gene Set Enrichment Analysis (GSEA) ranking of significantly upregulated or downregulated Hallmark gene sets in the FMT-treated conditions according to Normalized Enrichment Score (NES) at day 7. C) Boxplots of normalized counts for leading edge genes associated with immune activation pathways. P values represent the results of a Student’s t test between normalized counts across groups. D) Ranking of the Top 10 most enriched gene sets in the FMT-treated group by GSEA using Gene Ontology: Biological Processes gene sets. E-G) Boxplots of normalized counts in control (gray) or FMT-treated (red) samples for leading edge genes associated with E) intestinal homeostasis, F) synapse assembly, and G) neuropeptide signaling. P values represent the results of a Student’s t-test between normalized counts across groups. Statistical significance is indicated as *p < 0.05, **p < 0.01, and ***p < 0.001.

GSEA using the Hallmark gene set revealed significant downregulation of immune pathways at day 7 (Fig. 3B), including many of the same pathways upregulated at day 2 (TNFα signaling, IL-6 Jak/Stat3 Signaling, Inflammatory Response, and IFNγ Response). Leading edge gene analysis identified genes associated with immune signaling and chemoattraction of immune cells, which were all significantly decreased post-FMT by multivariate modeling (Fig. 3C). These genes represent mediators of the inflammatory signaling cascade, including the G-CSF receptor Csf3r, the prostaglandin-endoperoxide synthase 2 (COX-2) Ptgs2, and the phosphodiesterase 4 beta subunit Pde4b, which catalyzes the breakdown of cAMP. Expression of the pro-inflammatory cytokines Cxcl9 and Il1b was also decreased. Finally, expression of the epidermal growth factor receptor (EGFR) ligands epiregulin (Ereg) and amphiregulin (Areg) along with the transcription factors Egr1 and Egr2, which promote T-bet expression and inflammatory cytokine production (57, 58), were also decreased. Together, these results indicate that mediators of inflammatory signaling are downregulated on Day 7 post-FMT.

GSEA results using Hallmark genes identified only three significantly upregulated gene sets in the FMT-treated group: Myogenesis, Oxidative Phosphorylation, and genes downregulated by kRAS Signaling. In addition, several pathways associated with regulation of cell proliferation, including mTORc1, p53, and kRAS signaling, were downregulated. Based on these data, we hypothesized that FMT promoted expression of cell proliferation genes. As the output of GSEA is dependent on the gene sets tested, we repeated GSEA using the Gene Ontology Biological Processes (GO:BP) gene set (51), postulating that this would provide more specific pathway enrichment results.

GSEA using GO:BP identified 801 gene sets as significantly enriched, including 131 upregulated and 670 downregulated by FMT (Supplemental Table 4). Because this GSEA also identified immune signaling pathways as downregulated by FMT, we focused on novel pathways upregulated by FMT, primarily the top 10 most enriched by normalized enrichment score. These included many novel gene sets, with functions including morphogenesis and differentiation of neuronal tissue, regulation of synapse assembly, and neuropeptide signaling (Fig. 3D). As hypothesized, leading edge gene analysis identified several genes associated with cell proliferation and intestinal homeostasis. These include Denr and Mettl3, genes driving Cytosolic Translation enrichment, along with tRNA processing genes, which have been associated with intestinal cell proliferation (59). The Forelimb Morphogenesis gene set was driven by genes including the transcriptional repressor Zbtbl6 and the hedgehog family protein Ssh (Fig. 3E), which have both been identified as important for the renewal of intestinal stem cells (60, 61). These results indicate that FMT stimulates transcriptional changes that promote intestinal cell proliferation.

GSEA also identified gene sets associated with neuron development and signaling after FMT. Several of these gene sets were associated with development of neuronal tissue, including Cerebellar Cortex Morphogenesis, Hindbrain Morphogenesis, Cell Differentiation in Hindbrain, and Cerebellar Cortex Development. These gene sets shared a common set of leading edge genes, which were associated with both cell proliferation (Serpine1) as well as neuronal development (Grid2, Ttll1). A distinct set of genes associated with synapse assembly and neuron chemotaxis were also upregulated (Fig. 3F). Neuropeptide signaling, which consists of the generation of neuropeptide precursors and the enzymes required for their processing into mature signaling molecules (62), was also upregulated (Fig. 3G). These genes represent major signaling pathways that have been associated with intestinal homeostasis and protection from inflammation (63–65), highlighting a potential mechanism by which FMT restores homeostasis.

3.4. FMT Promotes a Tolerogenic Immune Environment in the Colon and MLN

Based on the importance of inflammatory genes at both timepoints, we investigated the effect of FMT on immune cell populations in the colon and mesenteric lymph nodes (MLN). Flow cytometry was performed on day 7 to quantify immune cell populations, with Uniform Manifold Approximation and Projection (UMAP) analysis visualizing differences in several cell populations (Fig. 4A). CD45+ immune cells were significantly increased in FMT-treated samples (Fig. 4B), which was primarily driven by CD11c+ dendritic cells (Fig. 4E) and B lymphocytes (Fig. S1E). Consistent with the downregulation of pro-inflammatory genes at day 7, pro-inflammatory immune populations, including neutrophils and Ly6c-hi monocytes made up a decreased percentage of cells post-FMT (Fig. S2B), with similar cell counts (Fig. S1B–C) despite significantly higher CD45+ cells in the FMT group. In contrast, innate immune cell populations associated with Type 2 and tolerogenic immune responses were significantly increased after FMT, including eosinophils, alternatively activated macrophages, and CD11c+ dendritic cells (Fig. 4C–E). These trends were also observed within T cell populations, where TCRγδ, Th2, and T regulatory cells were significantly increased post-FMT (Fig. 4F–H). While Th17 populations were not differentially abundant between groups (Fig. 4I), Foxp3+ RORγt+ double-positive T cells, which are associated with controlling inflammation and promoting tolerance (66), were significantly increased (Fig. 4J). These data indicate that while FMT increases total immune cell counts, their immune profiles are more tolerogenic.

Figure 4: Fecal microbiota transplantation induces Type 2 and tolerogenic immune responses in the colon and MLN.

Colonic lamina propria A-J) and mesenteric lymph node (MLN) K-O) immune cells were analyzed by flow cytometry on day 7. A) Cell cluster visualization using Uniform Manifold Approximation and Projection (UMAP) B-O) Cell counts per 100,000 live cells of B) CD45+ immune cells, C) eosinophils D) alternatively activated macrophages, E) dendritic cells (CD11b− CD11C+ MHCII+), F) γδT cells, G) Th2 CD4+ T cells, H) Th17 cells, I) FoxP3+ Treg cells, J) RORγt+ Treg cell, K) Th2 cells, L) ILC3s, M) Th17 cells, N) FoxP3+ Treg cells, O) RORγt+ Treg cell cells. Statistical significance is demarked as *P < 0.05, **P < 0.01, and ***P < 0.001. The error bar indicates SEM.

Immune profiles in the MLN were more similar between groups, with similar numbers of CD45+ immune cells (Fig. S3A) and innate immune cells (Fig. S3B–F). However, adaptive immune cells showed similar trends to those in the colon, with significantly increased Th2 and T regulatory cells in FMT-treated samples (Fig. 4K–M). In addition, pro-inflammatory Type 3 immune cells, namely innate lymphoid cell 3 (ILC3s) and Th17 T cells were significantly decreased (Fig. 4N–O). No changes were observed in other MLN cell populations investigated (Fig. S3). These results show that, like the intestinal immune environment, FMT promotes tolerogenic immune responses in the MLN.

4. Discussion

The primary finding from this work was that FMT promotes significant changes to the mucosal immune system as evidenced by enrichment in immune activation genes and changes in immune cell profiles within the colon and MLN. These FMT-induced changes in mucosal immunity were broadly more tolerogenic and less pro-inflammatory.

Transcriptional changes at day 2 post-FMT were enriched in pro-inflammatory immune pathways, which is consistent with prior evidence that disruption of the intestinal microbiota was associated with disruption of Type 1 immune responses that were restored through administration of FMT or MAMPs (67, 68). At day 7, many of these pro-inflammatory signaling pathways were downregulated. Indeed, the strongest drivers of immune gene set enrichment were pro-inflammatory mediators that were upregulated at day 2 but downregulated at day 7, including signaling molecules such as Il1b and Ccl2. Dampened inflammatory responses at day 7 were also observed in immunoprofiling data, as innate immune cells associated with inflammation (neutrophils and Ly6c-high monocytes) made up a lower relative proportion of total immune cells within the colon and Type 3 immune responses were significantly decreased in the MLN. Type 3 immune responses are associated with increased disease severity during CDI (27) and ILC3s are a driver of pathogenesis in inflammatory bowel disease (69, 70) that are decreased by FMT in a DSS-induced colitis model (34).

In addition to pro-inflammatory signatures, genes associated with Type 2 immunity were also upregulated at day 2. These changes persisted in the day 7 immunoprofiling data, as Th2 T cells were significantly increased in both the colons and MLN of FMT-treated mice. Innate immune cells associated with Type 2 and tolerogenic immune responses (eosinophils and alternatively activated macrophages) were also significantly increased in the colon. Type 2 immune responses are associated with tissue repair and return to homeostasis (71) as well as protection against severe disease during CDI (20, 24, 26).

Previous studies have observed that FMT promotes tolerogenic immune responses (33, 72, 73). Our results confirm these findings, identifying intestinal γδ T cells, T regulatory cells, and RORγt+ Foxp3+ T cells as upregulated after FMT treatment. Intestinal γδ T cells play a pivotal role in maintaining intestinal homeostasis, preserving mucosal tolerance, and molding the gut microbiota (74) and mice lacking γδ T cells (γδ −/−) exhibit heightened susceptibility to colitis (75–77). Characterization of RORγt+ Foxp3+ T cells has shown that they exhibit a distinctive hybrid phenotype, showcasing transcriptional and epigenetic profiles reminiscent of both Th17 and Treg cells (78). Our transcriptional data further supports these tolerogenic immune responses, as evidenced by increased expression of the retinoic acid synthase Aldh1a2. Retinoic acid has been identified as a key player for immune tolerance in the gut (79), and Aldh1a2 expression on intestinal dendritic cells is associated with the induction of T regulatory cell populations (80, 81). Of note, Il10 expression was not significantly changed in our study, despite its role in promoting tolerance during FMT (34, 82). Il10 expression trended higher in FMT treated mice at day 2 but showed no difference at day 7 (Figure S4), a trend that was also observed with other cytokines in this study and may be due to the broad decrease in immune processes at day 7 (Fig. 3B). One limitation of this study is that cytokines were not measured at the protein level, which may show differences between groups. Immunophenotyping identified significant increases in T regulatory cells, alternatively activated macrophages, and CD11c+ dendritic cells (Fig. 4), which are associated with intestinal tolerance and capable of producing IL-10 (83–85). Based on these data, IL-10 may play a role during the first week post-FMT, but further investigation is necessary.

Transcriptional signatures associated with intestinal homeostasis were also upregulated by FMT. These changes include genes associated with cell proliferation and restoration of intestinal stem cells, which are consistent with previous studies showing an increase in cell proliferation (as indicated by Ki67+ cells) after FMT treatment (82). FMT has also been shown to upregulate Ki67 in the context of Campylobacter jejuni infection (86) and LPS administration (87). Upregulation of genes associated with neuron development and signaling also supports the role of FMT in restoration of the gut, as these neuropeptide signaling pathways play key roles in maintaining intestinal homeostasis and protecting against inflammatory disease models (63–65). Crosstalk between intestinal immune cells and enteric neurons is critical in both homeostatic and dysbiotic conditions (88), and several of these genes are regulated by inflammatory signaling (89–91).

In conclusion, our results identified transcriptional changes in genes associated with immune activation and intestinal homeostasis over the first week post-FMT. These findings provide genetic evidence underlying tolerogenic changes in immune cell populations after FMT while also highlighting the role of neuronal cell development and signaling in this process. These results provide a potential mechanism through which FMT facilitates intestinal restoration after antibiotic-induced dysbiosis.

Supplementary Material

1

Supplemental Figure 1: Flow cytometry of colonic immune cells. Colonic lamina propria immune cells were analyzed by flow cytometry on day 7 A) Cell counts of myeloid immune cells, B) Ly6C high monocytes, C) Neutrophils, D) Th1 populations, E) B cells, F) ILC1 cells, G) ILC2, and H) ILC3 cell counts. Statistical significance is demarked as *P < 0.05, **P < 0.01, and ***P < 0.001. The error bar indicates SEM.

2

Supplemental Figure 2: Cell marker intensities in the colon and MLN. Colonic lamina propria A-H) and mesenteric lymph node (MLN) I-J) immune cells were analyzed by flow cytometry on day 7 A) Dot plots of CD45+ immune cells, B) eosinophils, C) alternatively activated macrophages, D) dendritic cells (CD11b− CD11C+ MHCII+), E) γδT cells, F) Th2 CD4+ T cells, G) Th17 cells and FoxP3+ Treg cells, H) MLN Th2 populations, I) MLN Th17 cells and FoxP3+ Treg cells, and J) MLN ILC3s.

3

Supplemental Figure 3: Flow cytometry of MLN immune cells. Immune cells from MLN were analyzed by flow cytometry on day 7 A-L) Cell count of A) CD45+ immune cells, B) myeloid cells, C) Ly6C high monocytes, D) Neutrophils, E) dendritic cells, F) alternatively activated macrophages, G) γδT cells, H) Th1 populations, I) B cells, J) CD8 cells, K) ILC1 cells, and L) ILC2 cell counts. Statistical significance is demarked as *P < 0.05, **P < 0.01, and ***P < 0.001. The error bar indicates SEM.

4

Supplemental Figure 4: Expression of Il10 in cecal tissue after FMT treatment. Normalized count data for the Il10 gene from cecal tissue of control (gray) or FMT-treated (red) samples at either day 2 (A) or day 7 (B). P values represent the results of a Student’s t-test.

Highlights.

• FMT restores microbiome alpha diversity within 7 days after antibiotic-induced dysbiosis.

• Pro-inflammatory and Type 2 immune genes are upregulated within 2 days of FMT treatment.

• Immune genes are downregulated at day 7, with increased expression of proliferation and signaling genes.

• FMT promotes increases in Type 2 and tolerogenic immune cell populations.

Data statement

The datasets generated or analyzed in the current study are available from the corresponding author upon reasonable request.