Sulfide is a keystone metabolite for gut homeostasis and immunity

Abstract

Hydrogen sulfide is a gaseous, reactive molecule specifically enriched in the gastrointestinal tract. Here, we uncover a non-redundant role for sulfide in the control of both microbial and immune homeostasis of the gut. Notably, depletion of sulfide via both pharmaceutical and dietary interventions led to a profound collapse of CD4 T cells in the ileum of the small intestine lamina propria and significant impact on microbial ecology. As a result, mice with reduced sulfide within the gut were deficient in their ability to mount T cell dependent antibody responses to oral vaccine. Mechanistically, our results support the idea that sulfide could act directly on CD4 T cells via enhanced AP-1 activation, leading to heightened proliferation and cytokine production. This study uncovers sulfides as keystone components in gut ecology and provides mechanistic insight between diet, gut sulfide production and mucosal immunity.

Figure 1.. Gut CD4 T cells are reduced after depletion of gut sulfides using bismuth or diet.

(A) Schematic of bismuth treatment protocol. Mice were treated with 5mg BSS every 12 hours for 72 hours, then assessed at day 10 for immune phenotyping. (B) Measurements of free sulfide at various locations in the small intestine, as noted on the left. Right, free sulfide concentrations as measured by microsensor in the small intestinal lumen in vehicle or BSS treated mice, at 4 hours post-treatment. (C) Cell counts for various immune cells in the small intestine lamina propria (SILP) at day 10. Size of circles indicates the fold reduction in cells of BSS treated mice compared to vehicle and color indicates p value (n=5). (D, E) Frequency (D) and counts (E) of CD4 T cells from days 3 to 10 post vehicle or BSS treatment (n=5). (F) CD4 T cell counts in the duodenum, jejunum and ileum lamina propria of the small intestine. (G) Ileal Peyer’s patches of vehicle or BSS treated mice at 10 days post treatment, stained with CD4 (white) B220 (teal) Bcl6 (blue) RANKL (red) and CD35 (pink). Scale bar indicates 100μm. (H-K) Total counts in ileal Peyer’s patches of CD4 T cells (H), Tfh frequency (I) CD4 T cell subsets (J) or germinal center B cells (K) at day 10 of Vehicle or BSS treatment. (L) Cell counts for various immune cells in the small intestine intraepithelial lymphocytes (SIIEL), mesenteric lymph node (MLN), large intestine lamina propria (LILP) and intraepithelial lymphocytes (LIIEL). Size of circles indicates the fold reduction in cells of BSS treated mice compared to vehicle (n=5). (M) Schematic of SAA diet treatment protocol, mice were placed on a defined amino acid diet (teal), SAA deficient diet (orange) or SAA deficient diet plus NaHS in drinking water, for 21 days before quantification of immune cells in gut. (N) Fecal sulfide concentrations after 21 days on dietary interventions in (M). (O,P) Frequency (O) and total counts (P) of CD4 T cells in ileum lamina propria. (Q) Total counts of CD4 T cells per Peyer’s patch. n.s. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001 as calculated by two-tailed t-test, with Benjamini-Hochberg correction for multiple comparisons (C, L).

Figure 2.. Sulfide depletion modulates gut microbiome composition and function, which does not drive CD4 T cell loss.

(A) Schematic of microbiome sampling, mice were treated with BSS as in Fig. 1A, then microbiome compartments were sampled weekly for longitudinal samples. (B) Longitudinal tracking of fecal microbiome beta diversity in vehicle and BSS treated mice, as measured by Axis 2 of PCA of unweighted unifrac distance from 16S rDNA sequencing. (C) Microbiome sampling at day 3 post treatment of small intestinal lumen, mucosa, and feces, as measured by PCA plots of unweighted unifrac beta diversity with indicated axes. Adjusted p values calculated by PERMANOVA. (D) Relative abundance of top 16 identified taxa at the genus (or order/family) level in each gut compartment. (E, F) Flow cytometry of SILP in conventional (SPF) and germ free mice treated with vehicle or BSS, with the total numbers of CD4 T cells (E) and percent Ki67+ CD4 T cells (F) shown. n.s. not significant, * p < 0.05, ** p < 0.01, as measured by two-tailed test (E,F) or Mann-Whitney (B).

Figure 3.. Gut CD4 T cell activation and AP-1 expression is impaired in the absence of gut sulfide.

(A) Representative flow cytometry plots of CD4 effector T cells in SILP of vehicle and BSS treated mice at day 10 after start of treatment, with T-bet and Ki67 markers shown. (B) Frequency of Ki67+ CD4 T cells in SILP. (C) Total counts of Ki67+T-bet+ CD4 T cells in SILP. (D) Frequency of T-bet+ CD4 T cells in SILP. (E) Median fluorescence of T-bet in T-bet+ CD4 T cells in SILP. (F) Frequency of IFNγ+ CD4 T cells in SILP. (G) Frequency of Ki67+ CD4 T cells in SILP of mice on SAA diet or SAA diet + NaHS supplemented drinking water on day 21 of treatment. (H) Median fluorescence of Nur77 in SILP of vehicle or BSS treated mice CD4 T cells, with representative histograms shown, including isotype control (IgG). (I) Schematic of scRNAseq procedure, mice were treated with vehicle or BSS for 3 days and ileal CD4 T cells were isolated on day 3 or day 10 for scRNA sequencing and bulk ATAC seq. (J) UMAP of all CD4 T cells with identified clusters shown. Four T-bet+ T_H1 clusters are noted with dotted oval. (K) Total number of significant differentially expressed genes in each cluster when comparing vehicle and BSS at day 3 or day 10, using Wilcox Rank Sum test. (L) Volcano plot of T-bet+ clusters comparing vehicle and BSS at day 3, with teal circles indicating significantly decreased and red significantly increased genes in BSS treated mice. (M) Violin plot of Fos expression among all CD4 T cells in vehicle, and days 3 and 10 of BSS treatment. (N) Gene ontogeny enrichment analysis of significantly upregulated genes in T-bet+ clusters, comparing vehicle vs BSS day 3. (O) Select ATAC-seq differentially enriched motifs, and their matching transcription factor identities, as analyzed by motif enrichment analysis of vehicle vs. BSS day 3 enriched peaks (TF, transcription factor). n.s. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, as measured by two-tailed t test (B-H) or Mann-Whitney U test (M).

Figure 4.. Sulfide acts directly on CD4 T cells to modulate MAP kinase phosphorylation.

(A) In vitro culture procedure, naïve CD4 T cells isolated from splenocytes and lymph nodes were cultured for 2 days with plate bound αCD3/αCD28. Cells were then moved to new wells with or without 100μM sulfide donor GYY4137 for 5 days without stimulation. (B) Total counts of CD4 T cells at days 2 and 5 of sulfide treatment. (C) Mean number of cell divisions for CD4 T cells at days 2 and 5 of sulfide treatment. (D,E) CD4 T cells at day 2 and 5 of sulfide treatment assessed for percent Ki67+ (D) or Nur77 median fluorescence (E). (F) TCR signaling leading to Fos and Jun transcription via MAPK cascade. (G) CD4 T cells cultured for 5 days with sulfide were restimulated with crosslinked αCD3 for 0–10 minutes and assessed for Erk phosphorylation by flow cytometry. Phospho-ERK+ cells shown as defined by isotype control antibody. (H) CD4 T cells treated with GYY4137 for 3 days were subjected to the ProPerDP protocol to enrich for persulfidated proteins. Volcano plot shows significantly enriched proteins in GYY4137 treated cells, with proteins of interest highlighted. * p < 0.05, as measured by two-tailed t test

Figure 5.. Gut sulfides are required for productive mucosal vaccine response.

(A) Vaccination of BSS treated mice with oral gavage of cholera toxin and ovalbumin one day after 3 days of BSS treatment, with vaccine response analyzed 10 days later. (B) Absolute counts of Peyer’s patch CD4 T cells in unvaccinated mice, or vaccinated mice treated with vehicle or BSS. (C) Frequency of Ki67+ CD4 T cells in Peyer’s patches. (D,E) Absolute counts of Tfh cells (D) and GC B cells (E) in Peyer’s patches. (F, G) ELISPOT of cholera toxin specific IgA secreting cells in the ileum of vaccinated mice, as shown by representative ELISPOT wells (F) and total frequency of antibody secreting cells (G). (H) Total counts of OVA specific transferred OT-II cells in ileum of vaccinated mice. (I) Vaccination of SAA diet fed mice after 2 weeks on SAA or defined AA diet, then assessed for vaccine response 10 days later. (J-L) ELISPOT of cholera toxin specific IgG and IgA secreting cells in the mLN of vaccinated mice, as shown by representative ELISPOT wells (J) and total frequency in mLN for cholera toxin specific IgG (K) and IgA secreting cells (L). (M) Cholera toxin specific gut IgA titers as measured by ELISA of gut luminal content. n.s. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, as measured by two-tailed t test.