Mycobiota-induced IgA antibodies regulate fungal commensalism in the gut and are dysregulated in Crohn's disease

Abstract

Secretory immunoglobulin A (sIgA) plays an important role in gut barrier protection by shaping the resident microbiota community, restricting the growth of bacterial pathogens and enhancing host protective immunity via immunological exclusion. Here, we found that a portion of the microbiota-driven sIgA response is induced by and directed towards intestinal fungi. Analysis of the human gut mycobiota bound by sIgA revealed a preference for hyphae, a fungal morphotype associated with virulence. Candida albicans was a potent inducer of IgA class-switch recombination among plasma cells, via an interaction dependent on intestinal phagocytes and hyphal programming. Characterization of sIgA affinity and polyreactivity showed that hyphae-associated virulence factors were bound by these antibodies and that sIgA influenced C. albicans morphotypes in the murine gut. Furthermore, an increase in granular hyphal morphologies in patients with Crohn's disease compared with healthy controls correlated with a decrease in antifungal sIgA antibody titre with affinity to two hyphae-associated virulence factors. Thus, in addition to its importance in gut bacterial regulation, sIgA targets the uniquely fungal phenomenon of hyphal formation. Our findings indicate that antifungal sIgA produced in the gut can play a role in regulating intestinal fungal commensalism by coating fungal morphotypes linked to virulence, thereby providing a protective mechanism that might be dysregulated in patients with Crohn's disease.

Extended Data Fig. 1. Identification of gut fungi from fecal material by flow cytometry and anti-C. albicans sIgA dynamics.

a, Microbes in fecal material from SPF WT WCM-CE mice were distinguished as a Sybr^hi population that is absent in GF mouse feces. b, Fungi (Sybr^hiCFW⁺) were enriched from bacteria (Sybr^hiCFW⁻) through size separation by 900 × g centrifugation and calcofluor white (CFW) staining of the resulting pellet. c, C. albicans cultured for 18 hours in hyphae-inductive media was stained with fecal supernatant from C. albicans-colonized GF mice (N = 6) collected at 0, 2-, 4-, 8- and 14-days post colonization, followed by sIgA staining. Analysis of IgA binding representative of two independent experiments, one-way ANOVA, followed by Sidak’s test. d, Representative flow cytometry plots of frequency of B220⁺IgA⁺ among Live CD45⁺CD4⁻ cells in the PP of germ-free (GF) mice orally gavaged with PBS (GF) or colonized for two weeks with C. albicans (+Ca). Data in (c) represents mean ± SEM.

Extended Data Fig. 2. CFW⁺Sybr^hi FSC^hiSSC^hi C. albicans population in feces represents hyphal/pseudohyphal fungal morphologies that are preferentially bound by sIgA.

a, CFW⁺Sybr^hi fungal population from feces of SPF mice colonized with CAF2-RFP C. albicans was sorted into FSC^hiSSC^hi and FSC^loSSC^lo fractions. Constitutive expression of RFP in this strain allows for high visibility and resistance to signal quenching upon prolonged light exposure during flow cytometry and microscopy on the same material. b, Immunofluorescence microscopy of sorted material from (a). Composite images at 20X magnification of FSC^hiSSC^hi and FSC^loSSC^lo shown in left and right panels, respectively. Scale bar represents 25μm. Data representative of two independent experiments. c, CFW⁺Sybr^hi fungal population from feces of SPF mice colonized with CAF2-RFP was sorted into IgA⁺ and IgA⁻ populations. Gray histograms represent IgA-isotype control staining used to distinguish sorted populations. d-e, Area (d) and perimeter length (e) of CAF2-RFP were compared between IgA⁺ and IgA⁻ sorted populations. Data represents two independent experiments, mean ± SEM. Two-sided Mann-Whitney test. N = 5.

Extended Data Fig. 3. Assessment of Ca-dREP C. albicans double reporter strain upon IgA staining and hyphae forming deficiency of efg1Δ/Δ cph1Δ/Δ C. albicans strain.

a-b, Immunofluorescence microscopy of Ca-dREP incubated with human fecal supernatant as a source of sIgA and stained with DAPI and anti-human IgA-APC (a) or an APC isotype control (b). Single channel staining of 2 samples shown. Left to right: DAPI, constitutive ENO1-GFP expression, hyphae-specific HWP1-RFP expression, and anti-human IgA-APC (a) or APC isotype control (b). Top rows in a and b correspond to composite images in Fig. 2d–e, representing three independent experiments. Scale bar represents 50μm. c, Hyphae-competent (WT), but not hyphae-deficient (yeast-locked; efg1Δ/Δ cph1Δ/Δ) strains of C. albicans forms hyphae upon hyphae-inducing stimuli in vitro. Scale bar represents 25μm.

Extended Data Fig. 4. Flow cytometry gating strategy in PPs, LP and in feces.

a-b, Cell gating startegy for assessment of IgA⁺ GC B cell in PPs (a) and IgA⁺ plasmablasts in lamina propria (b). c, gating strategy of C.albicans cells in feces pre- and post- C.albicans (C.a) colonization.

Extended Data Fig. 5. Graphical abstract for the model of antifungal IgA induction by and regulation of intestinal fungal commensalism.

(Credit: Created with BioRender.com)

Figure 1.. Under homeostatic conditions, the majority of the mouse and human gut mycobiota is coated by secretory IgA primarily induced by C. albicans.

a-b, Frequency of IgA-, IgG-, and IgM-bound fungi within the Sybr^hiCFW⁺ fraction of feces collected from SPF WT WCM-CE mice, assessed by flow cytometry. Gray curves in (a) represent corresponding Ig-isotype control staining. Data representative of three independent experiments. N = 5. c-d, Frequency of IgA-bound fungi assessed in WCM-CE SPF mouse feces from Rag1^−/− and B cell-deficient μMT^−/− mice in comparison with WT mice. Filled-in gray curve in (c) represent IgA-isotype control staining. Data represents two independent experiments. Kruskal-Wallis test followed by Dunn’s multiple comparisons test. WT, N = 13; μMT^−/−, N = 8; Rag1^−/−, N = 5. e-f, Feces collected from healthy human individuals assessed for IgA- and IgG-bound fungi by flow cytometry. Gray curves in (e) represent IgA-isotype control staining. Two-sided Mann Whitney test. N = 8. g, Total free sIgA levels in the feces of germ-free (GF) mice monocolonized with common commensal and dietary fungal species. Data representative of two independent experiments. Kruskal-Wallis test followed by Dunn’s multiple comparisons test. GF, N = 12; C. albicans, N = 15; S. cerevisiae, N = 4; S. fibuligera, N = 4; A. amstelodami, N = 4; W. sebi, N = 5. h, Germ-free (GF) mice were orally gavaged with C. albicans (+Ca) and sIgA binding to intestinal C. albicans was observed by flow cytometry in feces collected at days 2-, 4-, 8-, and 14-days post-colonization. Analysis of IgA binding representative of two independent experiments; one-way ANOVA, followed by Sidak’s test, N=6. i, Germ-free (GF) mice were orally gavaged with PBS (GF) or colonized for two weeks with C. albicans (+Ca). PP of all mice were harvested at day 14 for flow cytometry analysis of frequency of B220⁺IgA⁺ among Live CD45⁺CD4⁻ cells (i). Pooled from two independent experiments; two-sided Mann Whitney test (i). GF, N = 9; +Ca, N = 14. j-m, Representative plots and flow cytometry analysis of frequency of B220⁺IgA⁺ in PP B cells (j-k) and germinal center B cell (GC-B, l-m) subset in the PP of mycobiota-free altered Schaedler flora (ASF) mice orally gavaged with PBS (ASF) or colonized for two weeks with C. albicans (+Ca). ‎ Pooled from two independent experiments; two-sided Mann Whitney test. ASF, N = 8; +Ca, N = 10. Dots, fecal samples/gut-associated lymphoid cells (GALT) of individual mice (a-d, g, i, j, k l, m) or healthy humans (e-f); error bars, SEM. ns-p ≥ 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 2.. Secretory IgA antibodies preferentially bind fungal hyphae and influence C. albicans morphotypes in the gut.

a-b, Representative plots of forward scatter (a, left) and side scatter (a, right), analysis (b) of all CFW⁺Sybr^hi fungi (All), and the IgA⁺ / IgA⁻ fungal fractions, collected from healthy human fecal samples. Mixed effects analysis with Geisser-Greenhouse correction. N = 6. c, SPF-WT JAX mice were fed with 1×10⁸ C. albicans and feces were collected four days later. Average forward scatter (left) and side scatter (right) values of all CFW⁺Sybr^hi fungi (All), as well as the IgA⁺ and IgA⁻ CFW⁺Sybr^hi fractions. Mixed effects analysis with Geisser-Greenhouse correction. N = 8. d-f, Immunofluorescence microscopy of Ca-dREP incubated with human fecal supernatant as a source of sIgA, stained with DAPI and anti-human IgA-APC (d) or an APC isotype control (e). GFP⁺RFP⁻ yeast, GFP⁺RFP⁺ hyphae, and IgA⁺ events for each were counted calculation of IgA-coating frequency by morphotype (f). Composite images of 20X magnification and zoomed-in dotted regions of each are shown in left and right panels, respectively. Single channels are shown in Extended Data Fig. 3 (top rows). Scale bar represents 50μm in 20X panel and 20μm in zoomed-in panel. Data representative of three independent experiments. Two-tailed Wilcoxon matched pairs signed rank test. g-h, Ca-dREP were cultured for 18 hours in hyphae-inductive media and stained with fecal supernatant from C. albicans-colonized IgA-sufficient WT or IgA-deficient Rag1^−/− mice, followed by sIgA staining. Representative plots (g) and analysis (h) of IgA binding from two independent experiments. Gray curves in (g) represent Ig-isotype control staining. Multiple two-tailed t tests. WT, N = 4. Rag1^−/−, N = 4. i-j, SPF-WT (i) and Rag1^−/− (j) mice were treated with cefoperazone and colonized with a 1:1 mix of 1×10⁸ C. albicans strain CAF2-RFP (WT, black) and efg1^Δ/Δ (pink) C. albicans, after which CFUs of each were tracked for 20 days. For each mouse genotype, competition indices (CI) were generated for each strain, calculated as: CI = log₂(recovered CFUs/original inoculum CFU). Multiple two-tailed t tests. WT, N = 4; Rag1^−/−, N = 4. Data representative of two independent experiments. k-l, Representative flow cytometry plots (k) and analysis (l) of RFP⁺ hyphae in WT, μMT^−/−, and Rag1^−/− feces of mice 4 days after cefoperazone-aided colonization with 1×10⁸ Ca-dREP. Plots gated on Ca-dREP. Data representative of three independent experiments. Kruskal-Wallis test followed by Dunn’s multiple comparisons test. WT, N = 10; μMT^−/−, N = 5; Rag1^−/−, N = 8. m-o, Quantification of IgA-coating and Ca-dREP morphotypes in SPF WT JAX mice feces 4 and 14 days after cefoperazone-aided colonization. Plots in (m) represent colon contents gated on CFW⁺ events, with GFP⁺RFP⁻ yeast and GFP⁺RFP⁺ hyphae Ca-dREP populations easily distinguished. Connected dots in (o) represent % IgA⁺ fungal cells within the hyphal (red) and yeast (green) Ca-dREP gates in a single fecal sample collected at day 4 or 14 respectively. Two-tailed Wilcoxon matched pairs signed rank test. N = 7. p-q, sIgA from healthy human feces was applied (5μg/well) at 0 and 6 hours to a two-partner system comprising of Ca-dREP and Caco-2 cells upon Ca-dREP infection. The frequency of hyphal (p) and yeast (q) morphologies was measured by flow cytometry using the gating strategy described in (m). Each dot represents a well. Multiple unpaired two-tailed t tests followed by Sidak’s test. Error bars, SEM.

Figure 3.. C. albicans hyphal morphotypes induce potent sIgA responses.

a-c, Feces collected from cefoperazone treated SPF-WT JAX mice 4 days after oral gavage with a 1:1 mixture of 1×10⁸ CAF2-RFP and GFP⁺ efg1^Δ/Δ/cph1^Δ/Δ C. albicans. Representative flow cytometry plots of WT (black gate) and efg1^Δ/Δ/cph1^Δ/Δ (pink gate) (a) and the IgA⁺ gates within each (b), quantified as the IgA-bound fraction for each strain. Gray curves in (b) represent Ig-isotype control staining (iso). C. albicans in feces plotted in (a) is gated on CFW⁺Sybr^hi population. Data from two independent experiments. Dots represent analyses of IgA binding in each C. albicans strain found in the same sample. Two-tailed Wilcoxon matched pairs signed rank test. N = 8. d-g, Flow cytometry representative plots (d), analysis (e) of frequency of B220⁺IgA⁺ among Live CD45⁺CD4⁻ cells in the PP of ASF-WT mice and C. albicans burdens (f) two weeks after oral colonization with either WT C. albicans (+Ca) or efg1^Δ/Δcph1^Δ/Δ yeast-locked C. albicans (efg1^Δ/Δcph1^Δ/Δ). Free fecal sIgA reactivity was compared between WT C. albicans hyphae and yeast lysates by ELISA (g). Values graphed = OD450[hyphae] – OD450[yeast]. Pooled from two independent experiments; Kruskal-Wallis test followed by Dunn’s multiple comparisons test (e) and two-tailed Wilcoxon Signed Rank Test (f). ASF, N = 8; WT, N = 14; Δefg1/Δcph1, N = 13. h-k, sIgA binding to intestinal C. albicans in colon and small intestinal (SI) contents was measured (h), and Peyer’s patches (PP) B cells and colonic lamina propria (cLP) plasmablasts were assessed for IgA⁺ class-switch recombination (CSR) in SPF-WT JAX mice after two weeks of cefoperazone-aided colonization with WT C. albicans (+Ca) relative to untreated controls (NT). Representative plots and analysis of PP B220⁺IgA⁺ B cells (i), Fas⁺GL7⁺ germinal center B cells (GC-B) and IgA⁺ GC-B cells (j). B220⁻IgA⁺ plasmablasts in the colonic lamina propria (cLP) are shown (k). Two samples were excluded from the +Ca experimental group due to failed fluorescent staining. Two-tailed Mann-Whitney test. NT, N = 10; +Ca, N = 8. Box in (g) shows 25^th and 75^th percentiles with middle horizontal line and whiskers representing the median and min/max, respectively. Error bars, SEM.

Figure 4.. C. albicans sIgA responses are mediated through interaction with DC2 and CX3CR1⁺ MNPs.

Peyer’s patches (PP) B cells and colonic lamina propria (cLP) plasmablasts were assessed for IgA⁺ class-switch recombination (CSR) after two weeks of cefoperazone-aided colonization with WT C. albicans in Cd11c-cre^+/− × IRF4^fl/fl (ΔIRF4; a-c) or Cd11c-cre^+/− × Cx3cr1^DTR (ΔCX3CR1; e-g) SPF mice relative to Cd11c-cre^−/− littermates (Litt.). Representative plots and analysis of PP B220⁺IgA⁺ B cells (a, e), Fas⁺GL7⁺ germinal center B cells (GC-B), and IgA⁺ GC-B cells (b, f). Representative plots and analyses of B220⁻IgA⁺ plasmablasts in the lamina propria (cLP) are shown in (c, g). d, h. ELISA-based characterization of luminal anti-C. albicans sIgA in small intestines of ΔIRF4 (d), ΔCX3CR1(h) mice and the respective littermates after two weeks of intestinal colonization by WT C. albicans. Representative plot of side and forward scatter (i, k) and respective analysis (j, l) of CFW⁺Sybr^hi C. albicans (gated as shown in Extended Figure 4c) in luminal content of ΔIRF4 (i, j), ΔCX3CR1(k, l) mice and the respective littermates. Mixed effects analysis with Geisser-Greenhouse correction. Data represents 2 or 3 independent experiments. Two-tailed Mann-Whitney test. ΔIRF4, N = 12*; Litt., N = 7; ΔCX3CR1, N = 8; Litt., N = 12. *One sample was excluded from the ΔIRF4 experimental group due to failed fluorescent staining. Error bars, SEM.

Figure 5.. C. albicans-induced sIgA that target hyphae-associated virulence factors are decreased in CD patients.

a-e, Characterization of sIgA in feces of ASF-WT mice with (+Ca) or without (NT) two weeks of intestinal colonization by WT C. albicans. Reactivity assessed by ELISA against lysates from WT C. albicans hyphae and yeast (a), fungal cell wall mannan (b), hyphae-associated virulence factors Sap6 (c) and candidalysin (Ece1-III) (d), and bacterial flagellin (e). Data represents 3 experiments, two-tailed Mann-Whitney test. NT, N = 10; +Ca, N = 11. f-g, Characterization of ASCA sIgA and IgG in mucosal washings (f) and serum (g) of healthy individuals and CD patients. Two-tailed Mann-Whitney test. Healthy, N = 9; CD, N = 12. h-i, Characterization of sIgA reactivity from mucosal washings of healthy individuals and CD patients against Sap6 (h), and candidalysin (i). Two-tailed Mann-Whitney test. Healthy, N = 9; CD, N = 12. j, Representative plot of side scatter (j, left) and respective analysis (j, right) exploring granularity of CFW⁺ Sybr^hi fungi in mucosal washings of healthy individuals and CD patients. Healthy, N = 9; CD, N = 12. Each dot represents an individual human subject, analysis performed with Two-tailed Mann-Whitney test followed by Benjamini-Hochberg (BH) correction. Error bars, SEM.