Unmasking inflammation in juvenile dermatomyositis: myokine profiles of patients and bioengineered human muscle.

INTRODUCTION: Juvenile dermatomyositis (JDM), a rare autoimmune disease characterized by a type I interferon (IFN) gene signature and muscle weakness, lacks robust biomarkers and disease models. Myokines-muscle-derived cytokines-are potential biomarkers and therapeutic targets that may clarify muscle's role in JDM. We characterized myokine profiles in treatment-naïve JDM patients and compared them to an IFN-stimulated human tissue-engineered muscle model (myobundles) to identify biomarkers and validate the model. METHODS: Myobundles from four healthy pediatric donors were treated with IFNα, IFNβ, or IFN-stimulant poly(I:C). Sera from treatment-naïve JDM patients (n=21) and controls (n=9) were analyzed. A myokine panel (e.g., IL-6, IL-8, IL-17A, IL-18, CXCL9, CXCL10, TNFα, RANTES, IFNα-2a, IFNβ) was quantified in myobundle media and patient sera, with gene expression assessed by RNA sequencing in myobundles and JDM muscle biopsies. Serum myokines were correlated with Childhood Myositis Assessment Score (CMAS), and myobundle profiles were compared to patient signatures. RESULTS: Poly(I:C) triggered the strongest myokine response in myobundles, significantly increasing IL-6, IL-8, IL-15, IL-18, CXCL9, CXCL10, RANTES, and IFNβ. IFNα treatment increased TNFα, while IFNβ upregulated IL-15. JDM sera also showed elevations in IL-6, IL-15, IL-18, CXCL9, CXCL10, and IFNβ, with additional increases in IL-17 and IFNα (p(adj) = 0.0001-0.03). CXCL9, CXCL10, and IL-6 were significant independent predictors of CMAS, unlike conventional muscle enzymes. RNA sequencing confirmed elevated CXCL9 and CXCL10 expression in both IFN-treated myobundles and JDM biopsies. The myokine signature of IFNα-treated myobundles most closely reflected the JDM patient profile. CONCLUSION: JDM patients have a pro-inflammatory myokine profile in blood and muscle that can be recapitulated in IFN-stimulated myobundles. CXCL9 and CXCL10 are promising biomarkers, as are IL-6, IL-15, and IL-18, for JDM muscle activity. Our findings validate the myobundle model as a platform for studying JDM and support muscle as a key source of pathologic inflammation.