Commensal-derived tryptophan metabolites fortify the skin barrier: Insights from a 50-species gnotobiotic model of human skin microbiome

Abstract

The epidermal barrier defends the body against dehydration and harmful substances. The commensal microbiota is essential for proper differentiation and repair of the epidermal barrier, an effect mediated by the aryl hydrocarbon receptor (AHR). However, the microbial mechanisms of AHR activation in skin are less understood. Tryptophan metabolites are AHR ligands that can be products of microbial metabolism. To identify microbially regulated tryptophan metabolites in vivo, we established a gnotobiotic model colonized with fifty human skin commensals and performed targeted mass spectrometry on murine skin. Indole-related metabolites were enriched in colonized skin compared to germ-free skin. In reconstructed human epidermis and in murine models of atopic-like barrier damage, these metabolites improved barrier repair and function individually and as a cocktail. These results provide a framework for the identification of microbial metabolites that mediate specific host functions, which can guide the development of microbe-based therapies for skin disorders.

Keywords: aryl hydrocarbon receptor; keratinocytes; skin barrier; skin microbiota; tryptophan.

Figure 1:. Cultivation of unique human skin commensal bacteria.

(A) Swabs obtained from healthy subjects were cultured on specified solid agar media. Distinct colonies, discerned by their unique morphologies, were selected, and re-streaked on blood agar plates for further isolation and purification. (B) Phylogenetic tree of the 50 members of the microbial consortium (FF50) based on multiple sequence alignments of 16S rRNA gene sequences. Bacterial species names are color-coded based on their respective phyla: Actinobacteria (black), Deinocotta (magenta), Proteobacteria (purple), and Firmicutes (orange). Colored circles at the terminal nodes denote different genera as shown in legend. The outer edge features photographs of blood agar plates, showcasing the diverse morphologies across the collection. The scale bar provides a reference for the phylogenetic distance. See also Figure S1 for enlarged pictures of bacterial morphologies.

Figure 2:. Colonization with human skin microbiome restores skin barrier function in germ-free mice.

(A) Schematic illustrating colonization of GF mice with FF50 consortia. (i) Microbes previously identified (as described in detail in Fig. 1) and biobanked as glycerol stocks were streaked on blood agar plates. (ii) 48 hours later, individual colonies were picked, after confirming morphologies and planktonic cultures were grown as per optimized conditions. (iii) Cultures were combined at 10⁷ CFUs/microbe to (iv) inoculate bedding of GF mice every other day (M,W,F) for 1 month to (v) obtain gnotobiotic mice colonized with humanized skin microbiome. Whole metagenomics shotgun sequencing was performed on skin swabs collected from mice at the end of one month and relative abundance was calculated for (B) phylum, (C) genus and (D) species levels. Y-axis indicates relative abundance (calculated as percentage) for each mouse and controls (indicated on X-axis). (E) Barrier repair kinetics were compared between age-matched mice colonized with FF50 community and GF mice (n=5 male mice/group). (i) Schematic depicts experimental design for assessing barrier recovery. (ii) Following comparable insults (~20g/m²/h), transepithelial water loss (TEWL) was measured up to 24 hours and recovery kinetics were compared after fitting a linear regression model [F (1,36) =42.82, ***p<0.001]. (F) Barrier permeability was compared between FF50 colonized mice and GF mice in a model of ovalbumin-induced sensitization. (i) Mice (n=5 male mice/group) were tape-stripped (~1cm² area) to TEWL levels of ~20g/m²/h and treated with ovalbumin continuously for 5 days and (ii) barrier permeability was assessed [two-sided T-test, ***p<0.001]. (iii) Skin biopsies were harvested from non-sensitized dorsal skin (left panel) and ovalbumin sensitized regions (right panel) for histopathology. Scale bar (bottom right) indicates 100 μm and black dashed line shows dermal-epithelial boundary.

Figure 3:. Skin microbes possess tryptophan-metabolizing enzyme motifs.

(A) The bacterial tryptophan metabolizing pathways are illustrated, with intermediate metabolites represented in rectangular boxes and enzymes denoted in italics. (B) Open reading frames (ORFs) were systematically derived from the entire genomic sequences of bacteria, and the PFAM profiles of enzymes illustrated in Figure 2A were subsequently queried within these bacterial ORFs. The counts indicate the frequency of the specified motif in the genome. Consistent with Figure 1, species names are color-coded based on their respective phyla: Actinobacteria (black), Deinocotta (magenta), Proteobacteria (purple), and Firmicutes (orange). Colored circles at the terminal nodes denote different genera. A similar approach was used to mine human skin metagenomes for enzyme motifs, shown in accompanying supplemental Figure S3.

Figure 4:. Identification of tryptophan metabolites in the skin.

(A) Mice were colonized with FF50 community for one month and skin biopsies were collected for targeted mass spectrometry. (B) Key tryptophan metabolites detected using liquid chromatography/mass spectrometry are indicated. The amounts of (C) tryptophan, (D) tryptamine, (E) kynurenine, (F) indole propionic acid, (G) indole pyruvate, (H) indole-3-aldehyde, and (I) indole acetic acid detected per milligram of tissue were compared between germ-free (GF) and FF50 colonized mice using a two-sided T-test, with Benjamini-Hochberg correction applied for multiple comparisons, and adjusted p-values are reported (*p< 0.05).

Figure 5:. Tryptophan metabolites improve epithelial barrier.

(A) Therapeutic efficacy of metabolites in preventing increased barrier permeability was tested by pre-treating mice topically with metabolites followed by OVA-induced epicutaneous sensitization. (B) Epithelial barrier permeability was assessed at endpoint by measuring TEWL in age-matched male/female mice (as denoted by shape; n=3-5 mice/gender within each group). ‘Mock’ group indicates mice that were not treated with any chemicals; ‘OVA only’ group indicates mice that were pre-treated with vehicle. (C) N/TERT-1, (D) NIKS, and (E) primary neonatal foreskin keratinocytes were grown on transwells in an air-liquid interface in the presence of 100 μM of indicated treatments and transepithelial electrical resistance (TEER) was measured. Increase in TEER (computed as percentage) over vehicle (dotted line in graph at 100%) is reported. (F) Effect of increasing metabolite concentration (10 μM, 100 μM, 1000 μM) on TEER in N/TERT-1 cells was assessed. Average vehicle TEER value is indicated by the dotted line. Raw TEWL (panel B) and TEER values (panels C-D) were analyzed using one-way ANOVA, compared to vehicle using Dunnett's multiple comparisons test, and adjusted p-value is reported (*p<0.01, **p<0.001, ***p<0.0001).

Figure 6:. Tryptophan-aryl hydrocarbon receptor signaling cascade in skin barrier repair.

Reporter assay to assess AHR-activation in HaCaT cells using CYP1A1-luciferase reporter (A, B). Transfected cells were treated with (A) individual metabolites [FICZ (10 nM) and others (100 μM)] or (B) metabolite cocktail consisting of indole acetic acid, indole-3-aldehyde and indole pyruvate (each 100 μM) in presence or absence of AHR inhibitor (10 μM). Fold-change compared to vehicle is reported. (C) Primary neonatal foreskin keratinocytes were grown on transwells in an air-liquid interface. Transepithelial electrical resistance (TEER) was measured following treatment with FICZ (10 nM) or cocktail (100 uM each) in presence or absence of AHR inhibitor (10 μM) and. Each circle represents an independent donor. (D) Mice were topically pre-treated with indicated metabolite in presence or absence of AHR inhibitor, and barrier permeability was assessed following OVA-induced epicutaneous sensitization. (E) Germ-free mice were colonized with FF50 community for two weeks, tape-stripped and treated with ovalbumin. The mice were treated topically with AHR inhibitor throughout the course of the experiment. (F) Barrier permeability was assessed following OVA-induced epicutaneous sensitization and (G) Colony forming units (CFUs) were assessed by collecting punch biopsies. Effect of inhibitor on AHR activation (Panels A, B) was compared using a two-sided T-test. Effect of inhibitor on TEER was assessed using a multiple paired-T-test (Panel C). TEWL values were analyzed using a two-way ANOVA (Panel D), T-test (Panel F) and adjusted p-value is reported. (*p<0.05, **p<0.001, ***p<0.0001).