The altered landscape of the human skin microbiome in patients with primary immunodeficiencies

Abstract

While landmark studies have shown that microbiota activate and educate host immunity, how immune systems shape microbiomes and contribute to disease is incompletely characterized. Primary immunodeficiency (PID) patients suffer recurrent microbial infections, providing a unique opportunity to address this issue. To investigate the potential influence of host immunity on the skin microbiome, we examined skin microbiomes in patients with rare monogenic PIDs: hyper-IgE (STAT3-deficient), Wiskott-Aldrich, and dedicator of cytokinesis 8 syndromes. While specific immunologic defects differ, a shared hallmark is atopic dermatitis (AD)-like eczema. We compared bacterial and fungal skin microbiomes (41 PID, 13 AD, 49 healthy controls) at four clinically relevant sites representing the major skin microenvironments. PID skin displayed increased ecological permissiveness with altered population structures, decreased site specificity and temporal stability, and colonization with microbial species not observed in controls, including Clostridium species and Serratia marcescens. Elevated fungal diversity and increased representation of opportunistic fungi (Candida, Aspergillus) supported increased PID skin permissiveness, suggesting that skin may serve as a reservoir for the recurrent fungal infections observed in these patients. The overarching theme of increased ecological permissiveness in PID skin was counterbalanced by the maintenance of a phylum barrier in which colonization remained restricted to typical human-associated phyla. Clinical parameters, including markers of disease severity, were positively correlated with prevalence of Staphylococcus, Corynebacterium, and other less abundant taxa. This study examines differences in microbial colonization and community stability in PID skin and informs our understanding of host-microbiome interactions, suggesting a bidirectional dialogue between skin commensals and the host organism.

Figure 1.

Representative clinical images of disease severity in the different patient groups. (A) Non-PID atopic dermatitis, (B) Hyper IgE syndrome, (C) Wiskott-Aldrich, and (D) DOCK8 deficiency.

Figure 2.

Bacterial taxonomic classifications show colonization with unique taxa and altered representation of diverse taxa. Relative abundances of 14 major phyla-family taxonomies in the antecubital fossa (Af) are shown in A and retroauricular crease (Ra) in B. Shown are seven representative healthy controls and all Sanger-sequenced STAT3-HIES patients. Full versions of all skin and nares sites are in Supplemental Figure S4. (C) Inlaid pie charts compare the mean relative abundances of major skin genera across patient groups. Mean ± SEM for C are shown in Supplemental Figure S5 and Supplemental Tables S9 and S10. Patient identifiers are in Supplemental Table S13.

Figure 3.

Taxonomies associated with variation within and between individuals. Principal coordinates analysis (PCoA) of the Yue-Clayton theta coefficient, which calculates the similarity between two samples based on (1) number of species in common between two samples, and (2) their relative abundances. Samples that have similar principal coordinates appear closer together, i.e., are more similar. Biplot lines indicate the most significant unique consensus taxonomies contributing to variation along axis 1; Spearman correlations (ρ) are with axis 1. Length of biplot lines reflects the contribution of that taxa to the top three axes. Associated P-values < 2.2 × 10⁻¹⁶. Percentage variation attributed to an axis is indicated. Antecubital fossa (Af) and retroauricular crease (Ra) are shown in A and B, respectively; nares and volar forearm are shown in Supplemental Figure S6.

Figure 4.

Community-wide metrics suggest altered permissivity in niche specificity in colonization in PID individuals. Site codes: (Af) antecubital fossa; (N) nares; (Ra) retroauricular crease; (Vf) volar forearm. (A) Mean ± SEM for Shannon diversity is plotted for each patient group (colored as indicated) at all sites. (B) Site specificity is measured by the Yue-Clayton theta coefficient as in Figure 3, calculated between moist (antecubital fossa), dry (volar forearm), and sebaceous (retroauricular crease) skin sites and nares. (C) Longitudinal stability was assessed by calculating the theta coefficient between two and four timepoints for controls (light blue) and STAT3-HIES patients (red).

Figure 5.

Taxonomic association with clinical features. (A) Pie charts comparing the mean relative representation of major staphylococcal species across patient groups at all sites. Site codes: (Af) antecubital fossa; (N) nares; (Ra) retroauricular crease; (Vf) volar forearm. (B) Correlation of taxonomy with clinical markers of disease severity for the antecubital fossa. Unsupervised hierarchical clustering of Spearman correlation coefficients for the 34 taxa whose mean abundance was >0.25% across STAT3-HIES patients. Correlations are calculated against blood levels of lactate dehydrogenase, lymphocytes, IgE, and eosinophils, and a score of skin disease severity (scoring atopic dermatitis [SCORAD]). Hemoglobin was a control. Reds indicate correlation; blues indicate anti-correlation.

Figure 6.

Characterization of fungal skin communities suggest increased permissivity and elevated levels of opportunistic pathogenic fungi. (A) Boxplots showing fungal richness (species observed) of STAT3-HIES versus controls at all sites; black bars indicate median. (B) Select ITS1 fungal taxonomic classifications for the antecubital fossa, grouped by patient category. Full versions of all skin and nares sites are shown in Supplemental Figure S8. Patient identifiers are in Supplemental Table S13.