Virus induced dysbiosis promotes type 1 diabetes onset

Abstract

Autoimmune disorders are complex diseases of unclear etiology, although evidence suggests that the convergence of genetic susceptibility and environmental factors are critical. In type 1 diabetes (T1D), enterovirus infection and disruption of the intestinal microbiota are two environmental factors that have been independently associated with T1D onset in both humans and animal models. However, the possible interaction between viral infection and the intestinal microbiota remains unknown. Here, we demonstrate that Coxsackievirus B4 (CVB4), an enterovirus that accelerates T1D onset in non-obese diabetic (NOD) mice, induced restructuring of the intestinal microbiome prior to T1D onset. Microbiome restructuring was associated with an eroded mucosal barrier, bacterial translocation to the pancreatic lymph node, and increased circulating and intestinal commensal-reactive antibodies. The CVB4-induced change in community composition was strikingly similar to that of uninfected NOD mice that spontaneously developed diabetes, implying a mutual "diabetogenic" microbiome. Notably, members of the Bifidobacteria and Akkermansia genera emerged as conspicuous members of this diabetogenic microbiome, implicating these taxa, among others, in diabetes onset. Further, fecal microbiome transfer (FMT) of the diabetogenic microbiota from CVB4-infected mice enhanced T1D susceptibility and led to diminished expression of the short chain fatty acid receptor GPR43 and fewer IL-10-expressing regulatory CD4⁺ T cells in the intestine of naïve NOD recipients. These findings support an overlap in known environmental risk factors of T1D, and suggest that microbiome disruption and impaired intestinal homeostasis contribute to CVB-enhanced autoreactivity and T1D.

Keywords: coxsackievirus; fecal microbiota transplant (FMT); microbiome; type 1 diabetes; viral-induced autoimmunity.

Figure 1

Microbial Dysbiosis is a Consequence of Coxsackievirus B4 (CVB4) Infection (A) Diabetes incidence in uninfected control (n = 20) vs. CVB4-infected (n = 26) mice as determined by blood glucose measurement. Data are combined from 3 independent experiments and analyzed using a log-rank (Mantel-Cox) test. (B) Degree of pancreatic insulitis as (at day 21 pi) determined by histology (n = 5-6 per group). (C) Timeline of microbial community profiling. Fecal samples were collected at days 0, 2, 3, 7, 14, and 21 days pi. (D) Copies of 168 genes per gram of fecal material as measured by qPCR (CVB4 = days 2-21 pi). P value calculated using paired Welch's t test (two-tailed). (E) a-diversity (Observed number of OTUs) in bacterial communities following CVB4 infection. Control group includes baseline samples from all mice in the experiment on day 0 prior to mock or CVB4 infection. P values were calculated using Welch's ANOVA with Dunnet's T3 multiple comparisons test. (F) Principal component analysis (PCoA) of microbial community compositions. Longitudinal changes in the microbiome of (G) uninfected control (n=20) and (H) CVB4-infected mice (n=26) at the phylum level throughout sampling (Days 0, 3, 7, 14, and 21 pi). Data include replicates from at least three independent experiments. Results in D and E are represented as mean ± range, G and H are mean ± standard error. *P≤0.05 was considered statistically significant; ***P≤0.001; ****P≤ 0.0001, ns, not significant.

Figure 2

Bacterial Indicators of CVB4-induced Dysbiosis Associated with T1D (A) The relative abundance of bacterial taxa within experimental groups. Inner rings indicate composition on a phylum level and each subsequent ring represents a lower taxonomic level (class, order, family). Uninfected (normoglycemic, days 0-21 pi, n = 72); Early infection, days 0-3 pi (n = 34); Late infection - Normoglycemic, days 7-21 pi with blood sugar <16.2 mg/dL (n = 38); Late infection - Diabetic, days 7-21 pi with blood sugar >16.2 mg/dL (n = 24); Spontaneous diabetic, no infection with blood sugar >16.2 mg/dL (n = 8). (B) Copies of Bifidobacteria genes per gram of fecal material over the first 6 days of infection as measured by qPCR. (Control n = 3, CVB4 n = 6). Data are from one experiment and analyzed by Two-Way ANOVA using Šidak's multiple comparison test. (C) Ratio of Bacteroidota to Firmicutes abundance in fecal pellets of control, spontaneously diabetic (SD) and CVB4-infected mice at days 0-3 pi (early infection) and 7-21 pi (late infection). Data are from 3 independent experiments and represented as mean ± range. P values were calculated using Welch's ANOVA with Dunnet's T3 multiple comparisons test. (D) Lefse analysis displaying indicator bacterial species associated with uninfected control samples vs. late infection timepoints (Day 7-21pi) (Uninfected n = 72, Late Infection n = 62 samples). All data are representative of at least three independent experiments unless stated otherwise. **P≤0.01 was considered statistically significant; ***P≤0.001

Figure 3

Altered intestinal physiology and bacterial translocation in CVB4-infected NOD mice NOD mice (11-12 weeks) were infected with 400 pfu CVB4 intra-peritoneally. (A) Intestinal permeability based on detection of orally gavaged FITC-Dextran (4 kD) in the serum. Data are combined from four independent experiments. (B) Gene expression levels of tight junction-related proteins claudin 1 and tight junction protein 1 in the proximal colon of naïve and CVB4-infected mice. Target gene expression was normalized to GAPDH and expressed relative to uninfected control mice. Data are combined from two independent experiments. (C) Representative images of Alcian Blue stained mucus (between arrows) from control and CVB4-infected mice (left) and quantification of mucus width (right). (D) LPS endotoxin levels in the serum. Samples collected from two independent experiments were run on single assay. (E) Detection of 16S rRNA in the MLN and PLN of control and CVB4-infected mice (day 7 pi) as determined by qPCR. Intestinal permeability is represented as median with range. All other results are expressed as the mean ± SEM. P values in A-D were calculated using Welch's ANOVA with Dunnet's T3 multiple comparisons test and E was calculated using paired Welch's r test (two-tailed). *P<0.05 was considered statistically significant; *** P≤ 0.001.

Figure 4

CVB4 modifies host responses to commensal bacteria (A) Abundance of bacterial antigen-specific IgM, IgG, or IgA in the serum as measured by ELISA (n = 5-6 mice per group). Each sample was run in duplicate and averaged. (B) Relative expression of polymeric Ig receptor in the small intestine as determined by RT-qPCR. (C) Colonic luminal IgA and IgG reactivity to bacterial antigen as measured by ELISA. (D) Flow cytometry of bacteria from the feces of infected mice indicating proportion of microbes coated with either IgG (top) or IgA (bottom) antibodies. Representative flow plots showing gating of IgG and IgA bound microbes (left), and quantification on the right (n = 4 mice per group). Results are expressed as the mean ± SEM. P values in A and D are calculated by Two-Way ANOVA with Sidák's multiple comparisons test while those in B and C were calculated via Welch's ANOVA with Dunnet's T3 multiple comparisons test. *P<0.05 was considered statistically significant; **P≤0.01.

Figure 5

The CVB4-modified microbiome is diabetogenic (A) Experimental design for antibiotic depletion of recipients followed by FMT from donor mice. (B) Microbial a-diversity (Observed OTUs) in reconstituted FMT recipients. P values calculated via Welch's ANOVA with Dunnet's T3 multiple comparisons test. (C) The relative abundance of bacterial taxa within FMT recipient mice at day 35 post-FMT. Inner ring represents phyla and concentric rings moving out represent lower taxonomic levels (class, order, family) (D) Diabetes incidence in mice receiving FMT from either control (n=11) or a previously infected donor (n = 18) collected at day 14 pi. P value calculated by log-rank (Mantel-Cox) test. (E) Insulitis levels in mice receiving FMTs at 5 weeks post FMT. All mice in (D) were included in microbiome analyses and insulitis measurements. All data in this figure is representative of four independent experiments. A total of four donor pairs (control vs CVB4-infected) were tested. statistically significant. *P<0.05 was considered statistically significant.

Figure 6

FMT modifies intestinal immunity (A) Expression of free fatty acid receptors in the proximal colon of FMT recipient mice. (B) Abundance of CD4 Foxp3+ regulatory T cells in the ileum of FMT mice. (C) Expression of IL-10 in the intestinal Tregs. (D) Expression of IL-22 in the proximal colon measured by RT-qPCR. Results were normalized to GAPDH and expressed as relative expression compared to age-matched naive controls. (E) Abundance of anti-commensal antibodies in the serum of FMT recipient mice as measured by ELISA. All samples were analyzed at day 35 post-FMT. All p values were calculated using paired Welch's test (two-tailed) from at least 2 independent experiments. *P<0.05 was considered statistically significant.