Contextual control of skin immunity and inflammation by Corynebacterium

Abstract

How defined microbes influence the skin immune system remains poorly understood. Here we demonstrate that Corynebacteria, dominant members of the skin microbiota, promote a dramatic increase in the number and activation of a defined subset of γδ T cells. This effect is long-lasting, occurs independently of other microbes, and is, in part, mediated by interleukin (IL)-23. Under steady-state conditions, the impact of Corynebacterium is discrete and noninflammatory. However, when applied to the skin of a host fed a high-fat diet, Corynebacterium accolens alone promotes inflammation in an IL-23-dependent manner. Such effect is highly conserved among species of Corynebacterium and dependent on the expression of a dominant component of the cell envelope, mycolic acid. Our data uncover a mode of communication between the immune system and a dominant genus of the skin microbiota and reveal that the functional impact of canonical skin microbial determinants is contextually controlled by the inflammatory and metabolic state of the host.

Figure 1.

Dermal γδ T17 cells increase upon cutaneous C. accolens association. (A) Mean of absolute numbers (represented by the size of the circles) and frequencies (represented by the colors of the circles) of IL-17A–producing CD45⁺ CD90.2⁺ γδ TCR^low cells in the skin of mice previously associated or not with distinct skin commensal microbes. Data were collected after in vitro restimulation with PMA and ionomycin (Iono) in the presence of BFA. Results are representative of three independent experiments with four to six animals per group. (B) Frequencies (mean ± SEM) of CD45⁺ CD90.2⁺ TCRβ⁺ and γδ TCR^low cells from the skin of C. accolens–associated mice. (C–E) Frequencies (mean ± SEM) and absolute numbers of CD45⁺ CD90.2⁺ γδ TCR^low cells producing IL-17A in the skin (C) and ear-skin draining lymph nodes (cLN) (D) upon PMA/Iono restimulation in the presence of BFA or in the skin after treatment with BFA during tissue digestion (E). In B–E, data shown are representative of three independent experiments with three to eight mice per treatment group. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 as calculated by two-tailed, unpaired Student’s t test. (F and G) Absolute numbers of γδ TCR^low IL-17A⁺ cells (PMA/Iono restimulation in the presence of BFA) isolated from the ear skin of mice at different time points after the initial association. Data shown are representative of two independent experiments, with two to five animals per group. *, P < 0.05; **, P < 0.01 as calculated using one-way ANOVA with Holm-Šídák’s multiple comparison test. (H) Relative abundance of skin associated microbiota from either naive control or C. accolens–associated animals at different days after the initial association, as determined by 16S rRNA gene sequencing of V1-V3 hypervariable regions. Each bar represents the percentage of sequences in operational taxonomic units assigned to each phylum for an individual mouse. Data shown are representative of two independent experiments, with three to five animals per group. (I) Absolute numbers of IL-17A–producing γδ TCR^low cells from the skin of unassociated control or C. accolens–mono-associated GF mice. Data collected after in vitro restimulation with PMA/Iono in the presence of BFA. Data shown are representative of two independent experiments, with three to four animals per group. *, P < 0.05 as calculated by two-tailed, unpaired Student’s t test.

Figure 2.

C. accolens preferentially promotes the activation of Vγ4⁺ γδ T cells. Mice were topically associated with S. epidermidis (S. epi) or C. accolens (C. acco) or left unassociated (Control). (A) Frequencies of skin CD45⁺ CD90.2⁺ γδ TCR^low cells producing IL-17A (top) and flow cytometric analysis of TCR Vγ4 (Vγ4) and CCR6 expression by those cells (bottom) 14 d after the first association. Numbers are mean frequencies (±SEM). (B and C) Absolute numbers of skin γδ TCR^low IL-17A⁺ (B), γδ TCR^lowIL-17A⁺Vγ4⁺ CCR6⁺, and γδ TCR^low IL-17A⁺ Vγ4^neg CCR6⁺ (C) cells 14 d after the first association. In A–C, data shown are representative of two independent experiments with three animals per treatment group. **, P < 0.01 as calculated using one-way ANOVA with Holm-Šídák’s multiple comparison test. (D) Absolute number of IL-17A–producing γδ TCR^low Vγ4⁺ and γδ TCR^low Vγ4^neg in the skin of unassociated or C. accolens–associated C57BL/6 or B6.SJL (Sox13-deficient) mice 14 d after the first association (n = 4–5 mice per group). Data shown are representative of two independent experiments. ***, P < 0.001; NS, not significant (one-way ANOVA with Holm-Šídák’s multiple comparison test). (E) Heat map of relative gene expression analyzed by NanoString in γδ TCR^low Vγ4⁺ CCR6⁺ purified from unassociated control mice (Control, n = 4) or mice associated with S. epidermidis (S. epi, n = 3) or C. accolens (C. acco, n = 5; day 14 after association). Relative gene expression is clustered by sample (columns) and genes (rows) using Euclidean distance-based hierarchical clustering. Genes illustrated were significantly different after a two-way ANOVA with Tukey’s correction for multiple hypotheses.

Figure 3.

The ability to induce γδ T17 cells is evolutionally conserved in the Corybacterium genus. (A) Mean of absolute numbers (represented by the size of the circles) and frequencies (represented by the colors of the circles) of IL-17A–producing CD45⁺ CD90.2⁺ γδ TCR^low or γδ TCR^low Vγ4⁺ CCR6⁺ cells isolated from the skin of mice previously associated or not with various Corynebacterium species. (B) Flow cytometric plots illustrating the frequencies (mean ± SEM) of IL-17A–producing cells among skin CD45⁺ CD90.2⁺ γδ TCR^low cells from unassociated control mice or mice topically associated with C. accolens, C. jeikeium, or C. amycolatum. In A and B, results are representative of two independent experiments with four to six animals per group. (C) Schematic of the cell wall of most species of Corynebacterium bacteria. (D and E) Absolute numbers of IL-17A–producing CD45⁺ CD90.2⁺ γδ TCR^low cells (D) and flow cytometric analysis of TCR Vγ4 (Vγ4) and CCR6 expression by skin CD45⁺ CD90.2⁺ γδ TCR^low IL-17A⁺ cells (E) isolated from SPF mice topically associated with the WT strain (C. acco WT) or mycolic acid synthase–deficient mutant strain (C. acco Δ0503) of C. accolens. (F) Absolute numbers of skin IL-17A–producing γδ TCR^low Vg4⁺ CCR6⁺ cells from GF mice topically associated with the WT strain (C. acco WT) or mycolic acid synthase–deficient mutant strain (C. acco Δ0503) of C. accolens. In D–F, data shown are presentative of two independent experiments with four to nine mice per group. *, P < 0.05 (one-way ANOVA with Holm-Šídák’s multiple comparison test). (G) γδ TCR⁺ Vγ4⁺ CCR6⁺ cells purified from the skin and/or ear-skin draining lymph nodes of C. accolens–associated mice were co-cultured with splenic DCs untreated (Control) or preincubated with one of the following: hk C. accolens WT strain (C. acco WT), hk mycolic acid synthase–deficient C. accolens mutant strain Δ0503 (C. acco Δ0503), hk C. amycolatum (C. amy), hk C. striatum WT strain (C. striatum WT), or LAMs/lipomannans-deficient C. striatum (C. stria Δ647); liposomes containing TEEs from C. accolens WT strain (Liposomes + TEE) or LAMs (Liposomes + LAM); or empty liposomes. Graph illustrates the frequencies of IL-17A–producing Vγ4⁺ CCR6⁺ cells in overnight co-cultures. Graph is a compilation of the results from three to four experiments (each dot represents one culture well). ****, P < 0.0001 (two-tailed, unpaired Student’s t test).

Figure 4.

IL-23 is important for γδ T17 expansion and activation in the skin after C. accolens topical association. (A) Absolute numbers of total γδ TCR^low cells and IL-17A–producing γδ TCR^low Vγ4⁺ cells in the skin of unassociated control and associated WT C57BL/6 (WT) or C57BL/6 Il1r1^−/− (Il1r1^−/−, backcrossed to C57BL/6 for at least 10 generations) mice. (B) Frequencies of CD45⁺ CD90.2⁺ γδ TCR^low cells in the skin of unassociated (Control) and C. accolens–associated (C. acco) Il23r^−/− and littermate control (Il23r^+/─) mice 2 wk after the first topical association. Flow plots illustrate cells from associated mice. (C) Absolute numbers of skin γδ TCR^low Vγ4⁺IL-17A⁺ cells in Il23r^−/− and littermate control (Il23r^+/−) mice 2 wk after the C. accolens topical association. Control mice were left unassociated. In A–C, data shown are representative of three independent experiments with 3–11 mice per groups. ***, P < 0.001; ****, P < 0.0001; NS, not significant as calculated using two-way ANOVA with Holm-Šídák’s correction for multiple hypothesis. (D) γδ TCR⁺ Vγ4⁺ CCR6⁺ cells purified from the skin and ear-skin draining lymph nodes of naive (Naive γδ) or C. accolens–associated mice (C. acco γδ) were either co-cultured with splenic DCs preincubated with hk C. accolens (C. acco DCs) or not (Control DCs) or treated with the supernatant from DCs preincubated with hk C. accolens (DCs sup). Flow plots and graph illustrate the frequencies of IL-17A–producing Vγ4⁺ CCR6⁺ cells in overnight cultures. Flow plots are representative of three independent experiments. The graph is a compilation of the results from those three experiments (each dot represents one culture well). (E and F) Frequencies of IL-17A–producing Vγ4⁺ CCR6⁺ cells (E) and concentration of IL-17 in the supernatant (F) of overnight cultures of γδ TCR⁺ Vγ4⁺ CCR6⁺ cells (purified from the skin and ear draining lymph nodes of C. accolens–associated mice) treated with various concentrations of recombinant mouse IL-1β or IL-23. Each bar graph represents the mean frequency or mean concentration (±SD) of triplicate cultures. Data shown are representative of two independent experiments. **, P < 0.01; ***, P < 0.001 as calculated by two-tailed, unpaired Student’s t test.

Figure 5.

C. accolens has a contextual proinflammatory effect. (A) IMQ was applied daily on the ears of unassociated mice (Control) or mice topically associated with C. accolens (C. acco, every other day for 1 wk) starting 2 mo after the first topical association. The daily ear-skin thickness measurement is reported as the change in ear-skin thickness (mean ± SEM) relative to baseline at day 0 (first day of IMQ application). *, P < 0.05 (two-way ANOVA with Holm-Šídák’s correction for multiple hypotheses). (B) Absolute numbers of γδ TCR^low Vγ4⁺ IL-17A⁺ cells at day 5 after the start of IMQ application in the skin of mice from A. (C) IMQ was applied daily on the ears of unassociated mice (Control), C. accolens–associated mice (C. acco), or C. amycolatum–associated mice (C. amy) starting 2 wk after the first topical association. The daily ear-skin thickness measurement is reported as the change in ear-skin thickness (mean ± SEM) relative to baseline at day 0 (first day of IMQ application). In A–C, data shown are representative of two to three independent experiments with five to eight mice per group. *, P < 0.05; ***, P < 0.005 as calculated using two-way ANOVA with Holm-Sidak’s correction for multiple hypothesis. (D–J) 3-wk-old mice were fed either a high-fat (HF) or a control (Ctrl) diet for 1 mo before topical association with C. accolens. Mice were treated with anti-Vγ4 (αVγ4) or isotype control antibodies (Iso) starting on the first day of C. accolens topical association. (D) The ear-skin thickness measurement is reported as the change in ear-skin thickness (mean ± SEM) relative to baseline at day 0 (first day given the high-fat or control diet). (E) Absolute numbers of skin IL-17A-producing γδ TCR^low Vγ4⁺ and γδ TCR^low Vγ4^neg cells in unassociated or associated mice (day 14 after the first topical association) fed high-fat or control diet. In D and E, data shown are representative of two to three independent experiments with three to five mice per group. *, P < 0.05; ***, P < 0.005 as calculated using two-way ANOVA with Holm-Šídák’s correction for multiple hypothesis. (F) Venn diagram analysis comparing all statistically up-regulated genes in the skin between mice given control and high-fat diets after C. accolens association. (G) Pathway analysis was performed using Enrichr based on all genes up-regulated in the skin of mice given the high-fat diet after C. accolens association. Top 9 GO pathways are shown based on enrichment score (−log10 [adjusted p-value]). (H) Representative histopathological comparison of the ear pinnae of unassociated and C. accolens–associated mice given a high-fat diet at day 14 after the first topical association. Bars, 50 µm. (I) Frequencies (mean ± SEM) of Ki67⁺ keratinocytes (CD45^neg CD31⁺ CD34⁺ Sca-1⁺) in the skin of unassociated and C. accolens–associated mice given a high-fat diet (day 14 after association). In H–I, data shown are representative of two independent experiments. (J) Heat map of immunology-related genes statistically differentially expressed in the skin (ear pinnae) of unassociated and C. accolens–associated mice given a high-fat diet. Each row of the heat map represents gene expression from the two ears (pooled) of one mouse (n = 5 per group). (K) 3-wk-old Il23r^−/− and littermate control (Il23r^+/−) mice were fed a high-fat (HF) diet for 1 mo prior being topically associated with C. accolens or not. The ear-skin thickness measurement is reported as the change in ear-skin thickness (mean ± SEM) relative to baseline at day 0 (first day given the high-fat diet). Data shown are representative of two independent experiments with three to four mice per group. *, P < 0.05 as calculated using two-tailed, unpaired Student’s t test. (L) 3-wk-old mice were fed either a high-fat (HF) or a control (Ctrl) diet for 1 mo before topical association with the WT strain (C. acco Wt) or corynomycolic acid synthase–deficient mutant strain (C. acco Δ0503) of C. accolens or with C. amycolatum (C. amy). The ear-skin thickness measurement is reported as the change in ear-skin thickness (mean ± SEM) relative to baseline at day 0 (first day given the high-fat or control diet). Data shown are representative of two independent experiments with five mice per group. ***, P < 0.005 as calculated using two-way ANOVA with Holm-Šídák’s correction for multiple hypotheses.