The Non-Coding Genome in Dermatomyositis: Alu Element Overexpression as a Potential Driver of the Interferon Signature

Abstract

Dermatomyositis (DM) has long been recognized as a distinct clinicopathological entity within the spectrum of idiopathic inflammatory myopathies (IIMs), primarily characterized by its profound type I interferon (IFN) signature [1]. As recent discussions on IIM classification have emphasized, conventional clinicopathological categories are increasingly being complemented by serological, genetic, and molecular frameworks that better capture disease heterogeneity [2]. However, the precise “upstream” trigger for this IFN surge has remained elusive. While viral mimicry has been hypothesized, recent breakthroughs in genomic sequencing—specifically the Telomere-to-Telomere (T2T) reference genome—have shifted our focus toward the “dark matter” of our own DNA: the noncoding genome [3].

A landmark study reveals that the pathogenic driver in DM may not be an external pathogen, but rather the uncontrolled expression of Alu elements [3]. These short interspersed elements (SINEs), which occupy approximately 10% of the human genome, have a high propensity to form endogenous double-stranded RNA (dsRNA). Under normal physiological conditions, the enzyme ADAR (adenosine deaminase acting on RNA) “edits” these dsRNA structures—converting adenosine to inosine (A-to-I)—to prevent them from triggering innate immune sensors [4].

In DM muscle, however, a global elevation of Alu expression—including 577 unique Alu elements—appears to overwhelm the cellular editing capacity. This leads to an accumulation of unedited or insufficiently edited dsRNA that triggers MDA5-mediated inflammation [5]. MDA5 itself is also a known autoantigen in DM, thereby linking endogenous dsRNA sensing with the characteristic IFN cascade [6]. This internal “viral mimicry” model provides a compelling explanation for the prominent type I IFN signature of DM, supported by the observation that Alu overexpression appears relatively specific to DM and is not similarly enriched in other myositis subtypes such as inclusion body myositis (IBM) or antisynthetase syndrome (ASyS).

The clinical implications of these genomic findings are substantial but still emerging. The correlation between Alu expression, interferon-stimulated genes, and inflammatory cell signatures suggests that the “Alu editome” may capture an important layer of immune activation in DM and could eventually contribute to molecular subtyping or disease monitoring. Furthermore, the enrichment of edited Alu elements on chromosome 19 highlights a nonrandom genomic architecture that may intensify dsRNA sensing and IFN activation in DM [3].

However, translating the Alu signature into clinical practice will require validation in larger, longitudinal, and disease-subtype–stratified cohorts. Future studies should carefully account for treatment exposure, disease duration, biopsy site, autoantibody status, and baseline disease activity to avoid confounding and bias when linking genomic signatures to clinical outcomes [7].

Currently, the clinical application of these findings lies in refining our diagnostic categories, improving molecular phenotyping, and generating testable hypotheses regarding therapeutic responses. The relative specificity of the Alu signature in DM supports the concept that molecular signatures may complement conventional clinicopathological and autoantibody-based classification frameworks in IIMss [2].

The discovery of the Alu-driven IFN signature marks a pivotal moment in myositis research. By integrating advanced genomic insights with rigorous clinical validation, we are moving closer to a future where dermatomyositis classification, monitoring, and treatment are guided by the molecular mechanisms that drive disease.

Tang-Kai Cheng and Teng-Chieh Hsu drafted the manuscript. James Cheng-Chung Wei conceptualized the study, provided supervision, and critically revised the manuscript.

The authors declare no conflicts of interest.

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.