Debaryomyces is enriched in Crohn's disease intestinal tissue and impairs healing in mice

Abstract

Alterations of the mycobiota composition associated with Crohn's disease (CD) are challenging to link to defining elements of pathophysiology, such as poor injury repair. Using culture-dependent and -independent methods, we discovered that Debaryomyces hansenii preferentially localized to and was abundant within incompletely healed intestinal wounds of mice and inflamed mucosal tissues of CD human subjects. D. hansenii cultures from injured mice and inflamed CD tissues impaired colonic healing when introduced into injured conventionally raised or gnotobiotic mice. We reisolated D. hansenii from injured areas of these mice, fulfilling Koch's postulates. Mechanistically, D. hansenii impaired mucosal healing through the myeloid cell-specific type 1 interferon-CCL5 axis. Taken together, we have identified a fungus that inhabits inflamed CD tissue and can lead to dysregulated mucosal healing.

Fig. 1.. Broad-spectrum antibiotic–mediated inhibition of colonic crypt regeneration was reversed by treatment with amphotericin B.

(A and B) WT mice were pretreated for 4 weeks with antibiotics (VNAM) or vehicle, biopsy injured, and analyzed at day 12 after injury. Representative hematoxylin and eosin (H&E) images (A) and wound bed lengths (B) are shown (n = 11 to 17 wounds per group; seven or eight mice per group). (C and D) VNAM-pretreated mice were biopsy injured and treated with NS-398 (PGE₂ synthesis inhibitor) or vehicle (control) twice daily from days 4 to 10 after injury. Representative H&E images (C) and wound bed lengths (D) are shown (n = 11 to 12 wounds per group; five mice per group) at day 12 after injury. (E and F) VNAM-pretreated mice were administered amphotericin B (AmpB) or vehicle (control) and then biopsy injured. Representative H&E images (E) and wound bed lengths (n = 12 or 13 wounds per group; five to seven mice per group) at day 12 after injury are shown. Significance (unpaired Student’s t test): ****P < 0.0001; N.S., not significant. All values in (B), (D), and (F) are means ± SEM. Dashed black lines in (A), (C), and (E) represent wound bed lengths (largest distance between the crypts). Scale bars, 100 μm.

Fig. 2.. Debaryomyces was enriched in wounds of VNAM-pretreated mice; its administration to conventionally raised or gnotobiotic mice led to impaired crypt regeneration.

(A to E) WT mice, pretreated for 4 weeks with VNAM or Kool-Aid (control), were biopsy injured and wounds were analyzed at day 8 after injury. A schematic representation (A) and relative abundance of fungi in the wounds (B) are shown. C, control; V, VNAM treated. The representative cultures (C) and counts of viable colonies (D) are for 14 to 16 wounds per group with four to six mice per group. Significance (unpaired Student’s t test): **P < 0.01. In (E), we show the identification of fungi by ITS sequencing of the pooled colonies per mouse in (C). (F to H) WT mice gavaged with phosphate-buffered saline (PBS; controls), D. hansenii (B6A1), or S. cerevisiae were biopsy injured and wounds were analyzed at day (d) 12 after injury. The experimental scheme (F), representative H&E images (G), and wound bed lengths are shown for 13 to 19 wounds per group with six to eight mice per group. (I to K) WT mice were gavaged with PBS (control), D. hansenii, or C. tropicalis and analyzed after DSS recovery. The experimental setup (I), representative H&E images (J), and percent colonic crypt loss (n = five to seven mice per group) are shown. Significance [one-way analysis of variance (ANOVA) and Tukey’s post hoc test]: ****P < 0.0001. All values in (D), (H), and (K) are means ± SEM. Dashed black line in (G) represents wound bed length. Scale bars, 100 μm.

Fig. 3.. D. hansenii impairs crypt regeneration through type I IFN and CCL5.

(A) Representative plots of live cells isolated from mouse colons after DSS recovery. (B) Cells from (A) were lysed and cultured on SDA (n = 4 to 6 mice per group). Significance (two-way ANOVA with Tukey’s post hoc test): ***P < 0.001. (C) Fold change in D. hansenii–infected wounds (black circles) or macrophages (blue circles) compared with respective controls. (D and E) Representative H&E images (D) and wound bed lengths (E) at day 12 after injury (n = 12 to 17 wounds per group; five to seven mice per group) in Ccl5^−/− and WT mice gavaged with D. hansenii (B6A1). Significance (unpaired Student’s t test): ****P < 0.0001. (F) Percent colonic crypt loss in mice after DSS recovery in the absence or presence of D. hansenii (n = 6 mice per group). Significance (two-way ANOVA and Tukey’s post hoc test): **P < 0.01. (G) CCL5 in BMDM supernatant 24 hours after D. hansenii stimulation (n = four experiments). N.D., not detected. (H and I) Representative H&E images (H) and wound bed lengths (I) at day 12 after injury (n = 16 wounds; six to seven mice per group) in mice gavaged or not with D. hansenii as indicated. Significance (unpaired Student’s t test): ****P < 0.0001. (J) CCL5 in BMDM supernatant 24 hours after D. hansenii stimulation (n = 4 experiments). Significance (unpaired Student’s t test): ****P < 0.0001. (K and L) Representative H&E images (K) and percent colonic crypt loss (L) in LysMCre Stat1^fl/fl and Stat1^fl/fl mice after DSS recovery in the presence of D.hansenii (n = 6 to 10 mice per group). Significance (unpaired Student’s t test): ****P < 0.0001. All values in (B), (E), (F), (G), (I), (J), and (L) are means ± SEM. Dashed black lines in (D) and (H) represent wound bed lengths. Scale bars, 100 μm.

Fig. 4.. D. hansenii was enriched in inflamed intestinal tissue of patients with CD, and the human isolate impaired crypt regeneration in vivo.

(A and B) Ileal biopsy tissue from healthy individuals and CD patients, cultured on SDA, shown in a schematic representation (A) and as the relative abundance of fungi assessed by ITS sequencing of the pooled colonies for each individual (B). (C and D) ITS sequencing of DNA isolated from inflamed and noninflamed ileal regions of CD surgical resections, as shown in a schematic representation (C) and as the relative abundance of fungi (D). Only genera with a combined mean across both groups of >1.5% relative abundance are plotted. All other genera are combined into “Other.” (E) Relative fungal abundance of colonies cultured from the surgical resections in (C). (F) D. hansenii abundance was analyzed by qPCR and normalized to human Gapdh DNA. Significance (paired Student’s t test): *P < 0.05; **P < 0.01. WU, Washington University; CS, Cedars-Sinai. (G to I) WT mice were gavaged with D. hansenii (CDA1) in the presence or absence of amphotericin B (AmpB) and biopsy injured. A schematic representation (G), representative H&E images (H), and wound bed lengths (I) at day 12 after injury are shown (n = 14 to 17 wounds per group; five to seven mice per group). Significance (one-way ANOVA and Tukey’s post hoc test): ****P < 0.0001. All values in (I) are means ± SEM. The dashed black line in (H) represents the wound bed. Scale bars, 100 μm.