A two-step activation mechanism enables mast cells to differentiate their response between extracellular and invasive enterobacterial infection

Abstract

Mast cells localize to mucosal tissues and contribute to innate immune defense against infection. How mast cells sense, differentiate between, and respond to bacterial pathogens remains a topic of ongoing debate. Using the prototype enteropathogen Salmonella Typhimurium (S.Tm) and other related enterobacteria, here we show that mast cells can regulate their cytokine secretion response to distinguish between extracellular and invasive bacterial infection. Tissue-invasive S.Tm and mast cells colocalize in the mouse gut during acute Salmonella infection. Toll-like Receptor 4 (TLR4) sensing of extracellular S.Tm, or pure lipopolysaccharide, causes a modest induction of cytokine transcripts and proteins, including IL-6, IL-13, and TNF. By contrast, type-III-secretion-system-1 (TTSS-1)-dependent S.Tm invasion of both mouse and human mast cells triggers rapid and potent inflammatory gene expression and >100-fold elevated cytokine secretion. The S.Tm TTSS-1 effectors SopB, SopE, and SopE2 here elicit a second activation signal, including Akt phosphorylation downstream of effector translocation, which combines with TLR activation to drive the full-blown mast cell response. Supernatants from S.Tm-infected mast cells boost macrophage survival and maturation from bone-marrow progenitors. Taken together, this study shows that mast cells can differentiate between extracellular and host-cell invasive enterobacteria via a two-step activation mechanism and tune their inflammatory output accordingly.

Fig. 1. Mast cells are found in the Salmonella-infected cecum and come in close contact with invading bacteria.

a, b Toluidine blue-stained tissue sections of cecum from uninfected mice and 48 h after S.Tm^wt SL1344 infection. Arrows indicate different MC locations such as submucosa (a, bottom) and mucosa (a, top, b). Scale bars: 50 µm. c–e Representative IF images for cecum of uninfected mice, or infected mice in different stages of inflammation as indicated in the panel headings. Arrows indicate the position of MCs in healthy mucosa and submucosa (c), close to the epithelial layer (d) or close to invading bacteria (e). Scale bars: 50 µm. f–h Quantification of avidin+ cells per cecum section as total numbers (f), submucosal MCs (g) or mucosal MCs (h). Every dot indicates the mean of at least three sections for one mouse, n = 8 (Control), 9 (Infected). Horizontal lines display median, and the significance of two-sided Mann–Whitney U test is shown. **P < 0.01; ns nonsignificant. Exact P values given in source data. i RT-qPCR analysis of total cecum tissue for mast cell protease transcripts relative to Hprt (2^-ΔCq). A threshold of expression derived from Cq values < 38 was chosen. Note that some values fall under the threshold and are therefore not visible. One dot represents transcript levels in one mouse, n = 6. Bars indicate median ± 95% confidence intervals. No significance (P > 0.05) was detected for any comparison by two-sided Mann–Whitney U test. j Avidin+ cells present in the lumen of infected cecum tissue sections. Representative overview image (scale bars: 50 µm) and magnified images from avidin and anti-S.Tm-co-staining. Images are representative from sections of all mice; the number of MCs in the lumen are however higher in mice with strong inflammation, such as in the selected images. Scale bars: 10 µm.

Fig. 2. Mast cells mount a potent immunomodulatory response to Salmonella, which is triggered by TTSS-1 effectors.

a Representative TEM images of BMMCs infected with S.Tm^wt SL1344 for 4 h, as well as uninfected BMMCs. Arrows indicate intracellular bacteria. Scale bar: 5 µm. b, c Representative 25 ×25 µm images (b) and quantification by flow cytometry (c) of BMMCs, infected with MOI 50 of S.Tm^wt SL1344 or the indicated TTSS-mutants for 4 h. GFP signal and quantification show vacuolar S.Tm within BMMCs. n = 4. d–g Similar conditions as above but analysis of secreted IL-6 after 4 h (n = 6) (d) and 24 h (n = 6) (e). f Similar setup as in (d), but S.Tm 14028 strains were used (n = 5). g Similar setup as in (d), but PCMCs were used (n = 4). h, i Heatmap for RT-qPCR-quantified transcript levels for Il6, Il13, Tnf and Nr4a3 in BMMCs (h) and PCMCs (i), infected for 4 h by the indicated S.Tm SL1344 strains. j Secreted IL-6 levels from BMMCs infected with MOI 50 of S.Tm^wt and S.Tm^∆invG SL1344 as well as E. coli DH10B, E. coli MG1655 and Y. pseudotuberculosis for 24 h (n = 4). Every experiment was performed 2–3 times and mean ± SEM of pooled biological replicates is shown. Uninfected cells were used for statistical comparisons by one-way ANOVA and Dunnett’s posthoc test to all other groups. ***P < 0.001; ns—nonsignificant. Exact p values given in source data. For S.Tm ATCC 14028 infections, a ΔmalX strain was used as wt.

Fig. 3. The TTSS-1 effectors SopB, SopE, and SopE2 induce mast cell cytokine expression and secretion upon Salmonella infection.

a–g BMMCs (a–e) or PCMCs (f, g), infected with MOI 50 of S.Tm^wt SL1344 or the indicated TTSS-mutants for 4 h. a (n = 4), d (n = 4), and f (n = 4) show IL-6 secretion, b shows a heatmap for RT-qPCR-quantified transcript levels of Il6 and Tnf in BMMCs, g shows Il6 transcript levels in PCMCs (n = 4). c (n = 7, 8, 7, 8, 5, 4, 3, respectively, for groups from left to right) and e (n = 6) show percentage of BMMCs harboring vacuolar (ssaG-GFP + ) S.Tm. Every experiment was performed 2–4 times and mean ± SEM of pooled biological replicates is shown. S.Tm^wt-infected cells were used for statistical comparisons by one-way ANOVA and Dunnett’s posthoc test to all other groups. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. Exact P values given in source data.

Fig. 4. Salmonella-invaded mast cells are the main source of cytokines.

a BMMCs were infected with S.Tm^wt SL1344 carrying the pssaG-GFP reporter for 4 h and sorted to enrich the ssaG-GFP- population. Il6 transcript levels for both fractions as well as uninfected and unsorted S.Tm-infected BMMCs are shown (n = 4). b BMMCs were infected with S.Tm^wt SL1344 for 4 h and stained for Salmonella LPS. Relative population sizes of MCs positive for LPS and/or vacuolar (ssaG-GFP + ) S.Tm (n = 4 for “Δ4”, n = 6 for others). c Experimental setup for analysis of the source of soluble factors secreted by MCs. d–f BMMCs were infected with S.Tm^wt in one compartment while separated from BMMCs in the other compartment not coming in direct contact with the bacteria. RT-qPCR-quantified transcript levels for Tnf, Il6 and Il13, respectively, in the aforementioned transwell compartments (n = 6 for “Bystander”, n = 5 for others). Experiments were performed 2–3 times and mean ± SEM of pooled biological replicates is shown. a, c–f S.Tm^wt-infected cells were used for statistical comparison by one-way ANOVA and Dunnett’s post hoc test to all other groups. b Two-way ANOVA with Dunnett’s posthoc test was used to compare S.Tm^wt-infected cells within each subpopulation to all other groups. **P < 0.01; ***P < 0.001; ns nonsignificant. Exact p values given in source data. Panel c was created with BioRender.com.

Fig. 5. TLR4 and Sop effectors drive cytokine secretion from Salmonella-invaded mast cells.

a BMMCs were pretreated for 1 h with the Cyto D concentrations indicated, and infected with MOI 50 of S.Tm^wt SL1344 carrying the ssaG-GFP reporter. After 4 h, BMMCs harboring vacuolar S.Tm were quantified (n = 4). b Similar setup as in (a), but IL-6 secretion was measured (n = 4). c BMMCs were left uninfected or infected with MOI 50 of S.Tm^wt SL1344 or the indicated TTSS-mutants. After 1 h, cells were harvested and analyzed by immunoblot for P-Akt and Akt. d BMMCs were pretreated with MK-2206 for 1 h, infected with MOI 50 of S.Tm^wt, and analyzed as in (c). L = ladder. e, f Quantification of C-D (n = 5 for (e), except “Δ4” where n = 4, n = 4 for (f)). g BMMCs were pretreated with TAK-242 and/or MK-2206 for 30–45 min and infected with MOI 50 of S.Tm^wt. After 4 h, IL-6 secretion was measured (n = 4). h, i BMMCs from TLR4 KO mice and corresponding WT BMMCs were left uninfected, or infected with MOI 50 of S.Tm^wt SL1344 or the indicated TTSS-mutants. After 4 h, BMMCs harboring vacuolar S.Tm were quantified (n = 4) (h), and after 24 h IL-6 secretion was measured (BMMC WT n = 4 for “ΔinvG”, n = 6 for “Δ4”, n = 8 for others; TLR4 KO n = 10 for “Δ4”, n = 6 for “ΔinvG”, n = 9 for others) (i, j). j depicts an enlargement of (i) with independent statistical analysis (two-way ANOVA and Sidak’s posthoc test for both). Experiments were performed 2–3 times and mean ± SEM of pooled biological replicates is shown. Data was statistically analyzed with one-way ANOVA and Dunnett’s posthoc test, using S.Tm^wt-infected cells for comparisons to all other groups (a, b, e–g). h, i TLR4 KO BMMCs were compared with corresponding WT BMMC groups by two-way ANOVA with Sidak’s posthoc test. **P < 0.01; ***P < 0.001; ns nonsignificant. Exact P values given in source data.

Fig. 6. Salmonella induces a broad transcriptional and cytokine secretion response in mast cells with functional consequences on myeloid cells.

a RNA sequencing of BMMCs left uninfected, or infected with MOI 50 of S.Tm^wt or S.Tm^∆invG SL1344 for 4 h, presented as the top 30 significantly upregulated genes between S.Tm^wt-infected and uninfected control, displayed as a z-score-transformed heatmap. Genes are sorted by the formula “S.Tm^wt—S.Tm^∆invG” to highlight differences between those two groups. b Heatmap of relative log2-fold changes between the indicated groups, derived from a cytokine array of 24-h supernatants from BMMCs infected with MOI 50 of S.Tm^wt or S.Tm^∆invG SL1344. c Representative IF images of MCs in close contact to CD45+ and CD18+ immune cells in the S.Tm^wt SL1344 -infected intestinal mucosa and lumen at 48 h p.i. Arrows indicate MCs, scale bars: 10 µm. d Trypan blue-based live cell counts of bone marrow nucleated cells, cultured for 7 days in base medium supplemented with either medium alone, 24 h uninfected BMMC supernatant, S.Tm inoculum-conditioned supernatant, or supernatant of BMMCs infected with S.Tm^wt for 24 h. e, f Similar setup as in (d), but quantification of the total number of CD45⁺ CD11b⁺ CD11c⁻ cells (containing monocytes) (e) and CD45⁺ F4/80⁺ cells (containing macrophages) (f) in bone marrow cultures treated with the indicated supernatants, or base medium alone. d–f Data were statistically analyzed with one-way ANOVA and Dunnett’s post hoc test, using “BMMC + S.Tm” for comparisons to all other groups. Experiments were performed two times and mean ± SEM of pooled biological replicates is show, n = 4 (n = 3 for BMMCs and BMMCs + S.Tm), derived from bone marrow cultures of individual mice. ***P < 0.001. Exact P values given in source data.