The Intestinal Microbiome Restricts Alphavirus Infection and Dissemination through a Bile Acid-Type I IFN Signaling Axis

Abstract

Chikungunya virus (CHIKV), an emerging alphavirus, has infected millions of people. However, the factors modulating disease outcome remain poorly understood. Here, we show in germ-free mice or in oral antibiotic-treated conventionally housed mice with depleted intestinal microbiomes that greater CHIKV infection and spread occurs within 1 day of virus inoculation. Alteration of the microbiome alters TLR7-MyD88 signaling in plasmacytoid dendritic cells (pDCs) and blunts systemic production of type I interferon (IFN). Consequently, circulating monocytes express fewer IFN-stimulated genes and become permissive for CHIKV infection. Reconstitution with a single bacterial species, Clostridium scindens, or its derived metabolite, the secondary bile acid deoxycholic acid, can restore pDC- and MyD88-dependent type I IFN responses to restrict systemic CHIKV infection and transmission back to vector mosquitoes. Thus, symbiotic intestinal bacteria modulate antiviral immunity and levels of circulating alphaviruses within hours of infection through a bile acid-pDC-IFN signaling axis, which affects viremia, dissemination, and potentially transmission.

Keywords: Clostridium; alphavirus; bile acid; chikungunya; interferon; microbiome; monocyte; pathogenesis; plasmacytoid dendritic cell.

Figure 1.. Depletion of the intestinal microbiome increases CHIKV burden.

A. Three-week-old male mice received oral AV for 3 days and then water for 3 days prior to CHIKV inoculation in the foot. B. Fecal samples were harvested prior to the start of AV (day −6), after AV cessation (day −3), and at CHIKV infection (day 0). Bacterial 16S rRNA copy number was determined (2 experiments, n = 8–10). C. Viral infection was measured by focus-forming assay (FFA) at 1 dpi from indicated tissues after treatments (2 experiments, n = 8). D. Viral titers at 3 dpi from indicated tissues after treatments (2 experiments, n = 9–10). E. CHIKV RNA copies at 28 dpi in tissues following indicated treatments or reconstitutions (4 experiments, n = 20–21). F. Joint swelling following vehicle, AV, or AV + FMT treatment and CHIKV infection. Swelling was normalized to day 0 of each group (3 experiments, n = 15 per group). Data are represented as mean + the standard error of the mean (SEM). G. CHIKV RNA in situ hybridization of spleens from uninfected mice or at 3 dpi of AV- or vehicle-treated CHIKV-infected mice: low (top; scale bars, 100 pm) and medium (middle; scale bars, 100 pm) magnification with a high magnification inset (scale bars, 10 pm) (representative images from n = 3 per group). H. GF mice or GF mice colonized with FMT from conventionally-housed Abx-naive mice were inoculated with CHIKV. Viral titers at 1 dpi (2 experiments, n = 10). B, C and E: one-way ANOVA with Dunnett’s test; D, H: Mann-Whitney test (ns, not significant; * P < 0.05; ** P < 0.01; *** P < 0.001); F: two-way ANOVA with Sidak’s post-test (ns, not significant; *, P < 0.05; ***, P < 0.001, ****, P < 0.0001). Dotted lines indicate the limit of detection (LOD). See also Figure S1.

Figure 2.. Circulating monocytes of Abx-treated or GF mice are more permissive to CHIKV.

A. CHIKV RNA levels in peripheral blood leukocytes following vehicle or AV treatment after normalization to Gapdh levels. The fold-change between each group is indicated (2 experiments, n = 8–10). B. CHIKV RNA levels in plasma after vehicle or AV treatment (2 experiments, n = 8–10). C. Cell surface expression of CHIKV envelope protein at 1 dpi following vehicle or AV treatment on indicated immune cells (2 experiments, n = 9). D-E. Viral RNA copies in peripheral blood leukocytes (D) and plasma (E) at 1 dpi in GF mice or GF mice colonized with FMT (2 experiments, n = 10). F. Peripheral blood leukocytes were stained for immune cell markers and CHIKV envelope proteins at 1 dpi in GF mice or GF mice colonized with FMT (2 experiments, n = 10). G-H. AV or vehicle-treated mice were inoculated with CHIKV-GFP. At 2 dpi, peripheral blood leukocytes were stained for immune cell markers and CHIKV envelope proteins. Representative flow cytometry plots (G) and summary (H) of Ly6C^hi monocytes showing surface viral antigen and GFP signal (2 experiments, n = 6–8). I-K. CHIKV RNA copies in serum of AV- or vehicle-treated mice at 1 dpi following depletion with anti-Ly6C/Ly6G (I), anti-CCR2 (J), or anti-Ly6G (K) or their isotype controls mAbs (2 experiments, n = 8–9). A-F, H-K: Mann-Whitney test (ns, not significant, ** P < 0.01; *** P < 0.001; **** P < 0.0001). See also Figure S2–S4.

Figure 3.. ISG responses in Ly6C^hi monocytes are impaired in the absence of the intestinal microbiome.

A-C. Peripheral blood leukocytes were collected from AV or vehicle-treated mice (uninfected or 1 dpi) from three biological replicates per group and pooled. Leukocytes were subjected to CD4/CD8/CD19 depletion and then microfluidic-based single-cell RNA-seq sequencing (~ 7,000 cells per condition with ~50,000 RNA reads per cell). A. tSNE plots of all groups merged with cell type clusters. Clusters were annotated using known cellular markers. B. ISG responses (top 30 differentially expressed ISGs) projected onto tSNE plots separated by treatment condition and time relative to CHIKV infection. C. Violin plots showing differential expression of selected ISGs from monocyte clusters 1, 2, 5, and 8 at 1 dpi from AV-treated (AV) and vehicle-treated (V) mice. A MAST test with a Bonferroni correction compared expression between groups at 1 dpi (** P < 0.01; **** P < 0.0001). D. Expression of ISGs normalized to Gapdh in sorted LyC6^hi monocytes from peripheral blood of AV- or vehicle-treated mice at day 1 following MAYV infection (2 experiments, n = 5–6). MAYV, a BSL2 agent, was used since it can be sorted safely in our flow cytometry facility. E. Donor CCR2+GFP+ monocytes were harvested from AV- or vehicle-treated naïve 4-week-old male CCR2+GFP+ transgenic mice. Sorted GFP+ monocytes were transferred to recipient AV- or vehicle-treated WT mice. Recipient mice were inoculated with CHIKV 16 h later, and viral antigen staining was measured (see Fig 2C). F-G. Flow cytometry plots (F) and data summary (G) of viral antigen staining on CCR2+GFP+ transferred Ly6C^hi monocytes or GFP⁻ Ly6C^hi recipient monocytes (3 experiments, n = 6). H. IFN-α/β levels in serum following indicated treatments and reconstitutions. Addition of a blocking anti-Ifnar1 mAb confirmed specificity (2 experiments, n = 9). I. IFN-α/β levels in serum measured in GF or GF+FMT mice (2 experiments, n = 10). J. CHIKV viremia following AV or vehicle treatment in 4–5-week-old male WT, Ifnar1^−/−, Stat1^−/−, Ifnar1^−/− + anti-IFN-γ mAb, or WT + anti-IFN-γ blocking mAb (2 experiments n = 6–10). K. IFN-γ levels in serum following AV or vehicle treatment (3 experiments, n = 12). L. Viremia following AV or vehicle treatment in 4–5-week-old male LysM-Cre⁻ Stat1^fl/fl or LysM-Cre⁺ Stat1^fl/fl (3 experiments, n = 4–5). D, I-K: Mann-Whitney test; G, H, L: one-way ANOVA with Dunnett’s post-test (ns, not significant, * P < 0.05; ** P < 0.01; **** P < 0.0001). See also Figure S5.

Figure 4.. IFN-α responses in pDCs are impaired following microbiome depletion

A-B. CHIKV viremia (A) or peripheral blood leukocytes (B) in AV or vehicle-treated mice following administration of anti-PDCA-1 or isotype control mAb (2 experiments, n = 8). C. Surface CHIKV antigen staining and CHIKV-GFP signal in LyC6^hi monocytes from AV or vehicle-treated mice at 2 dpi following administration of anti-PDCA-1 or isotype control mAb (2 experiments, n = 8). D. IFN-α/β levels in serum of AV or vehicle-treated mice at 1 dpi following administration of anti-PDCA-1 or isotype control mAb (2 experiments, n = 8 per group). E-H. Transcriptional analysis of sorted splenic pDCs from AV-or vehicle-treated mice at 0 or 1 day post-MAYV infection (n = 3 per group). E. Principal component analysis. F. Volcano plot of differentially expressed genes at 1 dpi. Fold-change is represented as AV/Vehicle with up- and down-regulated genes displayed in red and blue, respectively. G. Unbiased hierarchal clustering at 1 dpi of selected genes involved in viral sensing, type I IFNs, IFN signaling, or ISGs. H. Expression of Ifna4, Ifna5, and Ifna6 normalized to Gapdh in sorted splenic pDCs at 1 dpi (2 experiments, n = 8 per group). I. CHIKV viremia following AV or vehicle treatment in Irf7^−/− mice (2 experiments, n = 6–7). J-K. Splenic and BM pDCs sorted from naive AV or vehicle-treated mice were stimulated with imiquimod (IMQ), CpG-A, or CpG-B. Levels of IFN-α (J) or IL-6 (K) in the supernatant 16 h post-stimulation (2 experiments, n = 6). A-D: one-way ANOVA with Dunnett’s post-test; H, I-K: Mann-Whitney test (ns, not significant; * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001). See also Figure S6.

Figure 5.. Restriction of CHIKV dissemination by the microbiome requires MyD88 signaling in pDCs.

A-C. CHIKV RNA levels in serum (A), peripheral blood leukocytes (B), or spleen (C) of AV or vehicle-treated Myd88^−/− mice or WT littermate controls (3 experiments, n = 6–8). D. Type I IFN levels in serum of AV or vehicle-treated Myd8^−/− mice or WT littermate controls mice (2 experiments, n = 5 per group). E. BM was mixed 1:1 from CD45.2 Myd88^−/− or CD45.2 WT controls and BDCA-DTR (CD45.1/2) and used to reconstitute irradiated male WT CD45.1 recipients. F. Flow cytometric analysis of blood from mice reconstituted with CD45.2 WT + BDCA-DTR CD45.1/2 BM (BDCA-DTR:WT) or CD45.2 Myd88^−/− + BDCA-DTR CD45.1/2 BM (BDCA-DTR:Myd88^−/−). G. Flow cytometric analysis of splenic pDCs from mice reconstituted with CD45.2 WT + BDCA-DTR CD45.1/2 BM (BDCA-DTR:WT) or CD45.2 Myd88^−/− + BDCA-DTR CD45.1/2 BM (BDCA-DTR: Myd88^−/−) following DT administration. H-I. CHIKV RNA levels in peripheral blood leukocytes (H) and spleen (I) of AV or vehicle-treated chimeric BDCA-DTR:WT or BDCA-DTR:Myd88^−/− mice (2 experiments, n = 6–7). J. CHIKV antigen staining on LyC6^hi monocytes from AV or vehicle-treated chimeric BDCA-DTR:WT or BDCA-DTR: Myd88^−/− mice at 1 dpi (2 experiments, n=6–7). K-N. CHIKV viremia in AV or vehicle-treated (K) Tlr2^−/− (L) Tlr4^−/−, (M) Tlr7^−/−or (N) Tlr9^−/− or corresponding WT littermate control mice (3 experiments, n = 4–10 per group). A-D, H-N: Mann-Whitney test (ns, not significant; * P < 0.05; ** P < 0.01; *** P < 0.001).

Figure 6.. Colonization with a Clostridial bacterium restricts CHIKV infection.

A. CHIKV viremia in mice following indicated treatments and reconstitutions (2 experiments, n = 7–9). B. Phylogenetic tree of bacterial members closely related to an identified mouse Clostridium ASV that was differentially abundant in AV+FMT and AV+ANA mice. Scale bar represents 2% nucleotide sequence difference. C-D. CHIKV RNA levels in serum (C) or peripheral blood leukocytes (D) in mice following indicated treatments and colonizations with C. scindens (AV+C. scindens), E. faecalis (AV+E. faecalis), or C. clostridiforme (AV+C. clostridiforme) (2 experiments, n = 7–9). E. Type I IFN levels in serum at 1 dpi following indicated treatments and colonizations (2 experiments, n = 7–9). F. CHIKV viremia in Stat1^−/− mice after indicated treatments and colonization (2 experiments, n = 3–4). G. At 1 dpi, CHIKV antigen staining on LyC6^hl monocytes from Myd88^−/− mice or WT littermate controls that received indicated treatments and colonization (3 experiments, n = 4–6). H. CHIKV viremia in mice after indicated treatments and colonizations and administration of anti-PDCA-1 or isotype control mAb (2 experiments, n = 8). I. Sorted splenic pDCs from naïve mice following indicated treatments and colonization were stimulated with imiquimod (IMQ) or CpG-A. IFN-α levels in the supernatant 16 h post-stimulation (2 experiments, n = 7–10). J-K. CHIKV RNA in the heads of Aedes albopictus mosquitos (J) or proportion of heads positive for viral RNA (K) after blood meals from vehicle-, AV-, or AV+C. scindens- treated mice at 7 days-post-feeding. In K, positive was defined as an FFU equivalent ≥ 10 per head (2 experiments, n = 34–39 mosquitoes per group). A, C-E, I, J: one-way ANOVA with Dunnett’s post-test; F-H: Mann-Whitney test (ns, not significant; * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001 ); K: Fisher’s exact t-test (ns, not significant; *** P < 0.001; **** P < 0.0001). See also Figure S7.

Figure 7.. Secondary BA promotes type I IFN responses and limits CHIKV infection

A. DCA levels in feces of naive mice following indicated treatment and reconstitutions (2 experiments, n = 8). B-C CHIKV RNA in (B) serum or (C) peripheral blood leukocytes of mice treated with vehicle, AV, or receiving DCA during AV treatment and for 3 additional days prior to CHIKV infection (AV+DCA) (3 experiments, n = 12–13) D. Type I IFN levels in serum following indicated treatments (2 experiments, n = 10). E-F CHIKV RNA at 1 dpi in (E) serum or (F) peripheral blood leukocytes or CHIKV Ag staining of Ly6C^hi monocytes (G) of GF mice or GF mice treated with DCA (2 experiments, n = 8). H. Type I IFN levels in serum at 1 dpi following indicated treatments of GF mice (2 experiment, n = 8). I. Viremia in Myd88^−/− or WT littermate mice after indicated treatments and reconstitutions (3 experiments, n = 3–6). J. Expression of Ifit1, Ifit2, Rsad2, Irf7, Ifna5, and Ifna6 normalized to Gapdh in sorted pDCs from AV-treated, AV+DCA, or vehicle-treated mice at 1 dpi (3 experiments, n = 4–5). K-M. Abx-naïve mice received water (control) or water with DCA for one week prior to CHIKV inoculation. Viral infection in tissues (2 experiments, n = 8). N. A model by which the microbiome restricts alphavirus dissemination by promoting systemic type I IFN responses in pDCs through a TLR7-MyD88-dependent pathway. In B-D, J: one-way ANOVA with Dunnett’s post-test; in E-I, K-M: Mann-Whitney test (ns, not significant; * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001 ).