Alterations in skin microbiome mediated by radiotherapy and their potential roles in the prognosis of radiotherapy-induced dermatitis: a pilot study

Abstract

Radiotherapy-induced dermatitis (RID) is an inflammatory cutaneous disorder that is acquired as an adverse effect of undergoing radiotherapy. Skin microbiome dysbiosis has been linked to the outcomes of several dermatological diseases. To explore the skin microbiota of RID and deduce their underlying impact on the outcome of RID, cutaneous microbiomes of 78 RID patients and 20 healthy subjects were characterized by sequencing V1-V3 regions of 16S rRNA gene. In total, a significantly apparent reduction in bacterial diversity was detected in microbiomes of RID in comparison to controls. Overall, the raised Proteobacteria/ Firmicutes ratio was significantly linked to delayed recovery or tendency toward the permanence of RID (Kruskal Wallis: P = 2.66 × 10^-4). Moreover, applying enterotyping on our samples stratified microbiomes into A, B, and C dermotypes. Dermotype C included overrepresentation of Pseudomonas, Staphylococcus and Stenotrophomonas and was markedly associated with delayed healing of RID. Strikingly, coexistence of diabetes mellitus and RID was remarkably correlated with a significant overrepresentation of Klebsiella or Pseudomonas and Staphylococcus. Metabolic abilities of skin microbiome could support their potential roles in the pathogenesis of RID. Cutaneous microbiome profiling at the early stages of RID could be indicative of prospective clinical outcomes and maybe a helpful guide for personalized therapy.

Figure 1

Phylum level analysis of cutaneous microbiota. Bar charts illustrate the relative proportions of the dominant phyla in skin microbiome of all studied groups. (a) Cutaneous microbiomes of rapidly recovered RID. (b) Cutaneous microbiomes of RID that showed delayed healing or turned to chronic ulcer. X-axis denotes the microbiome profile of each enrolled subjects, while Y-axis indicates the relative abundance of each phyla.

Figure 2

Alpha diversity of cutaneous microbiome was represented by box plots of the following indices: (a) the number of observed species and (b) Chao1 index, and (c) Shannon diversity index. Study groups were plotted on the X-axis and alpha indices were plotted on the Y-axis. The median is defined by the line in each box, the interquartile range (IQR) between the 25th and 75th percentile were delimited by the box, and the range was represented by the whisker. Statistical significance of pairwise comparisons was defined using the nonparametric Wilcoxon rank-sum test. Only significant differences were displayed with either * (p < 0.05), ** (p < 0.01) or *** (p < 0.001).

Figure 3

Principal coordinates analysis (PCoA) of cutaneous microbiomes based on weighted UniFrac distance matrices. PCoA plot represents the similarity distances between cutaneous microbiota of (a) healthy and diseased groups, (b) only RID microbiomes. Ellipses indicate the significant clustering (Adonis: r2 = 0.047, P = 0.00092) that was based on health state (a) and time required for healing (b).

Figure 4

The relative abundance of the most predominant genera in cutaneous microbiomes of healthy and patients with RID groups. (a) Bar plots depict the mean proportion and differences in mean proportions with 95% confidence intervals. (b) The core genera of the entire dataset are indicated by green boxes. (c) Colored boxes indicate the core microbiome of healthy and patients with RID, and strikes represent the statistically different genera between study groups. Statistical significance of pairwise comparisons was defined using the nonparametric Wilcoxon rank-sum test. Significant differences were displayed with either ns, p-value > 1; *p < 0.05, **p < 0.01 or ***p < 0.001. (d) Potential biomarkers for each group were inferred using LEfSe and the numbers denote LDA scores.