A specific microbial consortium enhances Th1 immunity, improves LCMV viral clearance but aggravates LCMV disease pathology in mice

Abstract

Anti-viral immunity can vary tremendously from individual to individual but mechanistic understanding is still scarce. Here, we show that a defined, low complex bacterial community (OMM¹²) but not the general absence of microbes in germ-free mice leads to a more potent immune response compared to the microbiome of specific-pathogen-free (SPF) mice after a systemic viral infection with LCMV Clone-13. Consequently, gnotobiotic mice colonized with OMM¹² have more severe LCMV-induced disease pathology but also enhance viral clearance in the intestinal tract. Mechanistically, single-cell RNA sequencing analysis of adoptively transferred virus-specific T helper cells and endogenous T helper cells in the intestinal tract reveal a stronger pro-inflammatory Th1 profile and a more vigorous expansion in OMM¹² than SPF mice. Altogether, our work highlights the causative function of the intestinal microbiome for shaping adaptive anti-viral immunity with implications for vaccination strategies and anti-cancer treatment regimens.

Fig. 1. OligOMM¹² colonized mice display a stronger body weight loss but improved viral clearance associated with an enhanced Th1 immune response.

SPF (gray), OligOMM¹²-colonized (OMM¹², blue) and germ-free (GF, pink) mice were intravenously infected with LCMV Clone-13 and analyzed at day 8 post infection (dpi). A Illustration of the experimental groups and setup. Created in BioRender (https://BioRender.com/q80p879). B Percentage of body weight loss of indicated groups at 4 and 8 dpi. Dotted line indicates 20% body weight loss (endpoint criteria). P values are as followed: 4 days post infection (SPF vs OMM¹² = 0.0052, OMM¹² vs GF = 0.0107) and 8 days post infection (SPF vs OMM¹² < 0.0001, OMM¹² vs GF = 0.0034). C Plots depict plaque forming units (PFU/g) in the small intestine and spleen of GF, SPF and OMM¹² colonized animals at 8 dpi. P values are as followed: small intestine (SPF vs OMM¹² = 0.0096, OMM¹² vs GF = 0.05) and spleen (SPF vs GF = 0.0061 and OMM¹² vs GF = 0.0048). D Representative flow cytometry plots of Th1 (T-bet^hi of CD4⁺ T cells (E) cell frequencies and absolute cell numbers in small intestine and spleen. P are as followed for small intestine (SPF vs OMM¹² = 0.0017 and SPF vs GF = 0.0032), for absolute Th1 counts in small intestine (SPF vs OMM¹² = 0.0010 and OMM¹² vs GF < 0.0001) and in spleen (SPF vs GF = 0.0161). F Representative flow cytometry plots and frequencies (G) of cytokine⁺ CD4⁺ T cells after restimulation with the virus-peptide gp61. P values are as followed: IFNγTNFα small intestine (SPF vs OMM¹² = 0.0013 and SPF vs GF = 0.0071), total IFNγ small intestine (SPF vs. OMM¹² = 0.0109), spleen (SPF vs OMM¹² = 0.0213) and total TNFα small intestine (SPF vs OMM¹² = 0.0064 and OMM¹² vs GF = 0.0037). For small intestine samples of two animals of the same group were occasionally pooled to increase the yield. Data from two independent experiments are shown. SPF n = 9, OMM¹² n = 6, GF n = 8 mice. Each dot represents an individual mouse (or occasionally pooled cells from two individual mice of the same group) and mean ± SD from at least two independent experiments is shown. Statistical analysis for (B, C, E, and G) was performed by using one-way ANOVA with Tukey correction for multiple comparison. P value of <0.05 was considered statistically significant. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 2. CD4⁺ T cell depletion prevents body weight loss and viral clearance in OligOMM¹² colonized animals.

SPF (gray) and OligOMM¹²-colonized (OMM¹², blue) mice were treated one day before, on the day of the clone-13 LCMV infection and 6 days post infection (dpi) with either a depleting anti-CD4 or an isotype antibody control as illustrated in (A) Illustration of the experimental groups and setup. Created in BioRender (https://BioRender.com/x96c559). B Percentage of body weight loss of indicated groups at 4 and 8 dpi. Dotted line indicates 20% body weight loss (endpoint criteria). P values are as followed: for 4 dpi (SPF isotype vs OMM¹² isotype = 0.0014, SPF anti-CD4 vs OMM¹² isotype = 0.0061 and OMM¹² isotype vs OMM¹² anti-CD4 = 0.0034) and for 8 days post infection (SPF anti-CD4 vs OMM¹² isotype = 0.0191). C Plots depict plaque forming units (PFU/g) in the small intestine and spleen of SPF and OMM¹² colonized animals treated with an isotype antibody control or anti-CD4 antibody at 8 dpi. P values are as followed: for small intestine (SPF isotype vs SPF anti-CD4 = 0.0160, SPF anti-CD4 vs OMM¹² isotype = 0.0042 and SPF anti-CD4 vs OMM¹² anti-CD4 = 0.0438). D Representative pictures of haematoxylin and eosin-stained sections of the small intestine for the indicated groups. Bar plots show (E) inflammation score, (F) epithelial apoptosis counts G) and infiltration of intraepithelial lymphocytes. P values are for small intestine (SPF isotype vs OMM¹² anti-CD4 = 0.0115). H, I, J) Colonization dynamics of OMM¹² mice during infection and after CD4 depletion: absolute abundance of individual OMM¹² strains. P values are as followed: for (H) d0 (OMM¹² isotype vs OMM¹² anti-CD4 = 0.00569) and for 1 dpi (OMM¹² isotype vs OMM¹² anti-CD4 = 0.0041) and OMM¹² anti-CD4 (1 dpi vs 8 dpi = 0.0363 and d0 vs 8 dpi = 0.03296, I) OMM¹² isotype (d0 vs 6 dpi = 0.0085 and 1 dpi vs 6 dpi = 0.0123) and OMM¹² anti-CD4 (d0 vs 8 dpi = 0.016 and 1 dpi vs 8 dpi = 0.0280) and J) OMM¹² isotype vs OMM¹² anti-CD4 (1 dpi = 0.0001 and d0 = 0.0350) and OMM¹² anti-CD4 (d0 vs 8 dpi = 0.0026 and 1 dpi vs 8 dpi = 0.0033). Data are plotted as 16S rRNA gene copy number of the individual strains per g of extracted gDNA. Data from two independent experiments are shown. Data from SPF (ctr) n = 10, SPF (anti-CD4) n = 10, OMM¹² (ctr) n = 8, OMM¹² (anti-CD4) n = 8 mice are shown (A–C). Histological analysis is shown from SPF (ctr) n = 8, SPF (anti-CD4) n = 8, OMM¹² (ctr) n = 6, OMM¹² (anti-CD4) n = 6 mice (D–G). Abundance of 16S rRNA copy number from OMM¹² (ctrl) n = max. 8 and OMM¹² (anti-CD4) n = max. 8 according to capacity of animals to donate feces (H–J). Each dot represents an individual mouse and mean ± SD from two independent experiments is shown. Statistical analysis for B), C), F) and G) was performed by using one-way ANOVA with Tukey correction for multiple comparison. Statistical analysis for (H–J) was done by performing multiple unpaired t-tests between each group. P value of < 0.05 was considered statistically significant. *p < 0.05, **p < 0.01. Source data are provided as a Source Data file.

Fig. 3. Virus-specific T helper cells (SMARTA) show an increased expansion and effector function in OligOMM¹²-colonized mice.

A 5×10³ SMARTA T cells were adoptively transferred into SPF (gray), OligOMM¹²-colonized (OMM¹², blue) and germ-free (GF, pink) mice 5 days prior infection with LCMV clone-13 and analyzed at 8 days post infection (dpi). Created in BioRender (https://BioRender.com/w55l495). B Percentage of body weight loss of indicated groups at 4 and 8 dpi. Dotted line indicates 20% body weight loss (endpoint criteria). P values are as followed: 4 days post infection (SPF vs OMM¹² = 0.0491 and OMM¹² vs GF = 0.00462), 8 days post infection (SPF vs OMM¹² = 0.0003). C Plots depict plaque forming units (PFU/g) in the small intestine and spleen of GF, SPF and OMM¹² colonized animals at 8 dpi. P values are as followed: small intestine (SPF vs OMM¹² = 0.0020 and OMM¹² vs GF = 0.0057). D Representative flow cytometry plots and (E) frequencies as well as absolute cell numbers of CD45.1⁺ SMARTA T cells in indicated organs. P values are as followed: for frequencies of SMARTA in small intestine (SPF vs OMM¹² = 0.0001) and spleen (SPF vs. GF = 0.0053); for the absolute SMARTA count in small intestine (SPF vs OMM¹² < 0.0001 and OMM¹² vs GF = 0.0027). Representative flow cytometry plots (F) and (G) frequencies of cytokine⁺ SMARTA T cells after restimulation with the virus-peptide gp61 in the indicated organs. P values are as followed: IFNγ small intestine (SPF vs OMM¹² = 0.0108), TNFα small intestine (SPF vs OMM¹² = 0.0236, OMM¹² vs GF = 0.0069) and spleen (SPF vs OMM¹² = 0.0009 and OMM¹² vs GF = 0.0003). H Frequencies of cytokine⁺ endogenous CD4⁺ T cells after restimulation with the virus-peptide gp61 in the indicated organs. P values are as followed for IFNγ small intestine (SPF vs OMM¹² = 0.0143) and spleen (SPF vs OMM¹² = 0.0081 and OMM¹² vs GF = 0.0196) and for TNFα small intestine (SPF vs OMM¹² = 0.0042 and OMM¹² vs GF = 0.0119) and spleen (SPF vs OMM¹² = 0.0081 and OMM¹² vs GF = 0.0196). For small intestine samples as well as cytokine stimulation samples of two animals of the same group were occasionally pooled to increase the yield. Each dot represents an individual mouse or two if pooled and mean ± SD from two independent experiments is shown. Data from two independent experiments with (B) SPF n = 12, OMM¹² n = 12, GF n = 10 mice, (C) SPF n = 9, OMM¹² n = 8, GF n = 5 or (E–G) SPF n = 5-10, OMM¹² n = 5-10, GF n = 5-7 mice are shown. Statistical analysis for (B, C, E, G and H) was performed using one-way ANOVA with Tukey correction for multiple comparison. P value of < 0.05 was considered statistically significant. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 4. SMARTA cells are transcriptionally different among SPF, OMM¹² and GF mice.

SPF (gray), OligOMM¹²-colonized (OMM¹², blue) and germ-free (GF, pink) mice received 5000 adoptively transferred SMARTA T cells 5 days prior infection with LCMV Cl-13. SMARTA T cells (live/dead⁻ CD4⁺CD45.1⁺) cells from small intestine lamina propria and spleen of 3–5 mice per group were re-isolated at 8 dpi and subjected to RNAseq analysis. A UMAP depicting Louvain cluster for all cells or (B) for each organ and microbiome condition individually. C Relative distribution of clusters in each sample group. Numbers inside the bars indicate the percentage of cells within a specific cluster. D Heatmap of expression levels for the topmost significant cluster-defining transcripts. E Feature plot of the average expression and percentage of cells from each cluster for selected genes.

Fig. 5. OMM¹² favors cytotoxic and proliferative Th1 cell clusters at the expense of Tfh cells.

SPF (gray, square), OMM¹² (blue, triangle) and GF (pink, circle) mice received 5.000 SMARTA cells 5 days prior infection with LCMV Cl-13 and were analyzed at 8 dpi. Small intestine lamina propria and spleen live⁺ CD4⁺ CD45.1⁺ cells were re-isolated at 8 dpi and subjected to RNAseq analysis. Expression of (A) Bcl6 and Cxcr5 on each individual cell for each sample group and organ represented over the reduced space of Violin plots. Representative (B) FACS plots and (C) frequencies of adoptively transferred Tfh SMARTA cells in spleen of SPF, OMM¹² and GF mice. P values are as followed: for adoptively transferred Tfh SMARTA in the spleen (SPF vs OMM¹² = 0.0043). Expression of (D) Plac8, (E) Ccr9, (F) Ly6c2 and Tbx21 and (G) GzmB on each individual cell for each sample group represented over the reduced space of UMAP. H Representative flow cytometry plots and I) frequencies of GzmB secretion of adoptively transferred SMARTA cells after restimulation with the virus-specific gp61 peptide. P values are for GzmB secretion of SMARTA in spleen (SPF vs OMM¹² = 0.0015 and OMM¹² vs GF = 0.0005). Each dot represents an individual mouse and mean ± SD from two independent experiments is shown. Data from two independent experiments with SPF n = 10, OMM¹² n = 8, GF n = 3 mice (C) or SPF n = 5, OMM¹² n = 5, GF n = 5 mice (I) are shown. Statistical analysis for (B, C, E, G and H) was performed using one-way ANOVA with Tukey correction for multiple comparison. P value of < 0.05 was considered statistically significant. *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as a Source Data file.