Enterococcus Induces MHC-II Expression in the Colonic Epithelium to Aggravate Graft-Versus-Host Disease

Intestinal domination by Enterococcus, a genus of Gram-positive facultative bacteria, has been associated with an increased risk of GVHD mortality in allo-HCT patients and in preclinical models. However, the immunological mechanisms underlying this association remain unclear. Several preclinical studies have demonstrated that MHC-II expression by intestinal epithelial cells (IECs) is involved in allo-antigen presentation and GVHD and can be modulated by the intestinal microbiota. MHC-II genes are constitutively expressed by professional antigen-presenting cells (APCs) but can be induced in non-professional APCs by IFN-γ. Here we reported that Enterococcus could potentially aggravate GVHD through upregulation of MHC-II expression by IECs in an IFN-γ-dependent manner.

Using an MHC-disparate mouse model of GVHD (C57BL/6J into BALB/cJ), we profiled the fecal microbiome of GVHD mice by 16S rRNA sequencing and the intestinal epithelium by flow cytometry on day 7 post-transplant. We found that endogenous Enterococcus relative abundance was positively correlated with MHC-II expression by colonic IECs (CD45- CD31- EpCAM+) during GVHD (Fig. 1a; p=0.0004; 2 replicates; n=19).

To determine whether this phenotype occurs in the absence of systemic inflammation, we utilized gnotobiotic mice treated with either E. faecalis, B. producta (a bacteria genus previously associated with improved GVHD outcomes) or PBS via oral gavage. Monocolonization with E. faecalis was sufficient to upregulate MHC-II expression by colonic IECs in non-transplanted (steady-state) animals compared to PBS-treated germfree controls (Fig. 1b; p=0.0027; 3 replicates; n=8-12/group). On the contrary, monocolonization by B. producta did not induce MHC-II expression by colonic IECs compared to germfree controls (p=0.96). We also observed that MHC-II upregulation was promoted by specific species within the genus Enterococcus, such that compared to PBS-treated germfree controls, only E. faecalis (p=0.0007; 3 replicates; n=8-12/group) was able to induce MHC-II expression by colonic IECs, but not E. durans (p=0.06), E. gallinarium (p=0.98) and E. hirae (p=0.99).

Using a Villin-Cre IFN-γR^fl/fl mouse model with genetic ablation of the IFN-γ receptor specifically in epithelial cells, we investigated whether MHC-II expression induced by Enterococcus was dependent on IFN-γ signaling. We colonized Villin-Cre^pos and Villin-Cre^neg littermate controls with E. faecalis via oral gavage after antibiotics decontamination (ampicillin/enrofloxacin) and profiled the intestinal epithelium. We observed that Villin-Cre^pos mice exhibited a significantly lower MHC-II expression in the small intestine compared to Villin-Cre^neg mice on day 3 post-Enterococcus colonization (p=0.012; 2 replicates; n=7/group), suggesting that IFN-γ signaling is necessary to mediate Enterococcus interaction with the intestinal epithelium.

Finally, to explore a potential intervention to reduce intestinal inflammation upon Enterococcus domination, we treated GVHD mice with either butyrate-producing B. producta or E. faecalis. Colonization with B. producta reduced fecal Enterococcus burden on day 7 post-transplant (p=0.007; 2 replicates; n=8/group) and improved survival in GVHD mice, compared to the E. faecalis-colonized group (p=0.002; 2 replicates; n=18-19/group). We and others have previously shown that butyrate administration can alleviate GVHD through its effects on epithelial homeostasis. To determine whether butyrate could regulate allo-antigen presentation by the intestinal epithelium, we employed the human colorectal adenocarcinoma HT-29 cell culture. We pre-treated the cells with IFNγ (100U/ml) to induce MHC-II expression, then added butyrate (5mM) to the culture for 24 hours. Butyrate treatment was sufficient to suppress MHC-II expression by HT-29 cells, compared to DMSO-treated controls (p=0.006; 3 replicates; n=6/group).

Altogether, our results suggest that one mechanism by which Enterococcus can aggravate GVHD involves the upregulation of MHC-II expression by colonic IECs, and treatment with butyrate might serve as an intervention to prevent intestinal inflammation. Further identification of the upstream mediators involved will provide additional mechanistic insight into the interaction between Enterococcus and the intestinal epithelium at steady state and during GVHD.