IL18-containing 5-gene signature distinguishes histologically identical dermatomyositis and lupus erythematosus skin lesions

Abstract

Skin lesions in dermatomyositis (DM) are common, are frequently refractory, and have prognostic significance. Histologically, DM lesions appear similar to cutaneous lupus erythematosus (CLE) lesions and frequently cannot be differentiated. We thus compared the transcriptional profile of DM biopsies with CLE lesions to identify unique features. Type I IFN signaling, including IFN-κ upregulation, was a common pathway in both DM and CLE; however, CLE also exhibited other inflammatory pathways. Notably, DM lesions could be distinguished from CLE by a 5-gene biomarker panel that included IL18 upregulation. Using single-cell RNA-sequencing, we further identified keratinocytes as the main source of increased IL-18 in DM skin. This study identifies a potentially novel molecular signature, with significant clinical implications for differentiating DM from CLE lesions, and highlights the potential role for IL-18 in the pathophysiology of DM skin disease.

Keywords: Autoimmune diseases; Autoimmunity; Cytokines; Dermatology; Skin.

Figure 1. DM skin lesions demonstrate a strong IFN signature.

(A) Principal components analysis of differentially expressed genes (DEGs) between lesional DM biopsies and healthy control skin. (B) IPA identifying key pathways in the DM DEG list. (C) Graphical representation of log₂ mRNA expression values of IFN genes from DM lesional skin microarrays. Bars depict SD. (D) IFN score comparison between DM and healthy control biopsies. The box plots depict the minimum and maximum values (whiskers), the upper and lower quartiles, and the median. The length of the box represents the interquartile range.

Figure 2. IFN-β and IFN-κ protein expression are increased in DM skin.

IHC of healthy control or 2 DM lesional biopsies (representative out of 8 patients) for the IFN-regulated protein MX1 and IFN-α, IFN-β, and IFN-κ. Original magnification, ×20.

Figure 3. DM shares IFN genes with CLE lesions.

(A and B) Comparison of DEGs in DM lesional skin (y axis) versus DEGs in SCLE (A) and DLE (B). Shared DEGs in the same direction are denoted in blue. (C) Venn diagram showing shared and unique DEGs between DM, SCLE, and DLE. (D) Genomatix Pathway analysis of shared 251 DEGs between DM and CLE subtypes.

Figure 4. DM exhibits fewer T cell–related genes compared with CLE.

(A) Heatmap of overlap of DM, SCLE, and DLE lesional DEGs with cytokine signatures generated via stimulation of keratinocytes with indicated cytokines followed by RNA-Seq. (B) xCell analysis shows no increase in total T cell score (driven by CD4⁺ central memory, effector memory, memory, and naive cells and CD8⁺ central memory, effector memory, and naive cells) in DM lesions versus healthy control (HC) whereas CLE has a high T cell score. Bars depict SD.

Figure 5. DM lesions can be distinguished by a 5-gene score and exhibit increased IL-18 in dermal inflammation.

(A) Box plots of DEGs in DM but not CLE (IL18 and LCE2D) or DEGs in CLE but not DM (CCL7 and CD2). (B) RNA was isolated from 9 DM and 9 CLE lesional samples from the independent validation cohort and subjected to real-time PCR with the indicated primers. Data are presented as the fold change calculated as 2^-ΔΔCT of DM versus CLE. Statistical significance was calculated via multiple 2-tailed t tests using false discovery rate to account for multiple comparisons of delta CT values normalized to GAPDH expression. The q values are denoted on the graph as follows: *q < 0.05; **q < 0.01. (C) IHC for IL-18 or isotype control in healthy control (representative of 3 controls), DM (representative of 3 patients), and CLE skin (representative of 4 patients). (D) xCell enrichment score for macrophage-derived transcripts in HC, DM, and CLE lesions. (E) scRNA-Seq analysis of DM lesional and nonlesional skin compared with healthy control and lupus skin. Graph represents gene expression in keratinocytes for percentage of cells expressing the indicated gene (by circle size) and degree of expression (by color).