Dermal injury drives a skin to gut axis that disrupts the intestinal microbiome and intestinal immune homeostasis in mice

Abstract

The composition of the microbial community in the intestine may influence the functions of distant organs such as the brain, lung, and skin. These microbes can promote disease or have beneficial functions, leading to the hypothesis that microbes in the gut explain the co-occurrence of intestinal and skin diseases. Here, we show that the reverse can occur, and that skin directly alters the gut microbiome. Disruption of the dermis by skin wounding or the digestion of dermal hyaluronan results in increased expression in the colon of the host defense genes Reg3 and Muc2, and skin wounding changes the composition and behavior of intestinal bacteria. Enhanced expression Reg3 and Muc2 is induced in vitro by exposure to hyaluronan released by these skin interventions. The change in the colon microbiome after skin wounding is functionally important as these bacteria penetrate the intestinal epithelium and enhance colitis from dextran sodium sulfate (DSS) as seen by the ability to rescue skin associated DSS colitis with oral antibiotics, in germ-free mice, and fecal microbiome transplantation to unwounded mice from mice with skin wounds. These observations provide direct evidence of a skin-gut axis by demonstrating that damage to the skin disrupts homeostasis in intestinal host defense and alters the gut microbiome.

Fig. 1. Skin influences gene expression in the intestine and increases expression of Muc2 and Reg3.

Single-cell RNA sequencing defines 17 cell clusters in the mouse colon. a UMAP plot. b Differential percent abundance in K14/HYAL1 and control mice. c Violin plot of Reg3 expression in cluster 10. d, e Spatial sequencing landmark slide with proximal, transverse, and distal colon delineated by color. f UMAP plot of spatial sequencing from Control, DSS, K14/Hyal1, and K14/Hyal1 DSS mouse colon. Clusters are indicated by color and number. g The spatial representation of clusters in Control, DSS, K14/Hyal1, and K14/Hyal1 DSS mouse colon. h Percentage abundance of each cluster in the sample. i Violin plot of Muc2 expression. j Spatial plot of Muc2 localization and abundance. k Violin plot of Reg3b expression. l Spatial plot of Reg3b.

Fig. 2. Mucin and Reg3 protein expression increases in the intestine in response of skin wounding or exposure to hyaluronan.

a Representative image of periodic acid–Shiff (PAS) staining in the transverse colon. (scale bar: 100 micron, arrows point to Mucin staining in the intestinal crypt). b Western blotting of Reg3g extracted from transverse colon. c Quantification of staining intensity of Reg3g in the colon. (n = 7 independent biological replicates per group). d Immunofluorescent staining of the colon. (Muc2: Green, Reg3g: Red DAPI: Blue. scale bar: 1000 micron). e High magnification of crypt structure. (Scale bar: 25 µm). f Immunofluorescent staining of the colon of germ-free mice with and without skin wounds. (Muc2: Green, Reg3g: Red DAPI: Blue). g mRNA expression level of Reg3g in colon treated ex-situ with hyaluronan 6.8kDA fragments (n = 6 independent biological replicates per group). h Concentration of Reg3A in medium of colon epithelial cells (HT29) treated in culture with hyaluronan 6.8kDA fragments(n = 6 biologically independent cells in each group). i Comparison of mRNA expression of Reg3a in colon epithelial cells following treatment with low molecular weight (LMW) hyaluronan and hyaluronan 6.8kDA fragments(n = 5 biologically independent cells in each group). Statistical significance was determined using Student’s unpaired two-sided t-test (g and h), ordinary one-way ANOVA and Tukey’s multiple comparison two-sided test (c and i). Error bars indicate mean ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001. Each experiment was repeated at least 3 times. Source data are provided as a Source Data file.

Fig. 3. Skin injury changes the composition of the gut microbiome.

a Alpha diversity analysis by the Shannon index and significance testing between groups by the Mann Whitney U test. Box plots represents mouse type, with the center line indicating the median, the box bounds representing the 25 and 75 percentiles, and the whiskers extending to minima and maxima, 1.5 times the interquartile range from the 25th and 75th percentiles, respectively. N = 32 control, 28 vancomycin, and 32 wound. b Beta diversity analysis using robust Aitchison PCA and significance testing between groups by PERMANOVA. c Relative abundance of top 34 bacterial species (top 18 shown in the legend). d Metagenomic differential abundance by Songbird of bacterial species plotted by their log-ratio associated with wound (red) and control (blue). e GO terminology functional differential abundance by Songbird.

Fig. 4. Skin injury or dermal hyaluronidase expression alters bacterial survival and epithelial penetration in the intestine.

a Proportion of live bacteria in feces as measured by Flow cytometer (control, skin wound: n = 11, K14/HYAL1: n = 6, Vancomycin: n = 5 independent biological replicates per group). b Bacterial morphology as measured by flow cytometer. c qPCR measurement of relative abundance of 16 S rDNA per mg feces (control, skin wound: n = 9, K14/HYAL1: n = 6 independent biological replicates per group). d–e. Gram staining of bacteria in the transverse colon at (d). Low magnification (scale bar: 50 micron) and e Intensity of gram staining in crypt or muscle layers from control, skin wound, K14/HYAL1. (n = 5 independent biological replicates per group). f In situ hybridization assay of bacterial 16 S rDNA in the colon. (scale bar: 50 micron, arrows). g Intensity of 16 S signal normalized by DAPI. (n = 7 independent biological replicates per group). h Concentration of FITC-labeled dextran sulfate entering the plasma after oral gavage (SPF: n = 6, GF: n = 4 independent biological replicates per group). Statistical significance was determined using ordinary one-way ANOVA and Tukey’s multiple comparison two-sided test. Error bars indicate mean ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001. Each experiment was repeated at least three times.

Fig. 5. The capacity of the skin to promote DSS colitis is dependent on bacteria in the gut.

a TNF mRNA expression in the colon of mice after DSS challenge to control or K14/HYAL1 mice with or without pretreatment by oral vancomycin (50 mg/kg) (No treatment: n = 5, DSS treatment: control n = 8, K14/HYAL1 n = 13). b TNF mRNA expression in the colon of mice after DSS challenge following skin wounding with or without pretreatment by oral vancomycin (n = 8). c TNF mRNA expression in the colon of germ-free mice after DSS challenge following skin wounding (n = 6). d Percent change in body weight normalized to weight at day 0. e Colon length at the 14 days after beginning of DSS treatment (n = 8 independent biological replicates per group). f, g Histological images of the distal colon from control and K14/HYAL/1 mice treated with or without vancomycin and Disease activity index (n = 6). h Survival rate of control of K14/HYAL1 mice over time after administration of DSS (n = 8. Scale bar: 50 µm). i Schematic of FMT experiments created with BioRender.com. j Il6 mRNA expression in the colon of mice following FMT from mice with or without skin wounds and challenge by DSS (n = 8). k Colon length of mice following FMT and 5 days after the beginning of DSS treatment (n = 8). l, m Histological images of the distal colon of mice following FMT and 5 days after the beginning of DSS treatment and Disease activity index (n = 8. Scale bar: 50 µm). Statistical significance was determined using Student’s unpaired two-sided t test (k), ordinary one-way ANOVA and Tukey’s multiple comparison two-sided tests (a–c, e, g, j, k and m) and ordinary two-way ANOVA and Sidak’s multiple comparisons two-sided test (d). Error bars indicate mean ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001. Each experiment was repeated at least three times. Source data are provided as a Source Data file.