The results of a new study suggest that subtypes of the idiopathic inflammatory myopathies can be distinguished on the basis of their unique gene expression patterns in muscle biopsy specimens. Moreover, these patterns also identified pathological processes that are highly relevant in each subtype.

In the study, the researchers first identified all differentially expressed genes in muscle tissue specimens from 119 patients with inclusion body myositis (IBM) or myositis-specific antibody (MSA)-positive myositis (dermatomyositis, anti-synthetase syndrome or immune-mediated necrotizing myopathy (IMNM)) as well as in 20 specimens from healthy individuals that were used as comparators. They then compared the ability of ten different machine learning models to classify the specimens into each subtype of myositis on the basis of the gene expression data; one model, the linear support vector machine (SVM) model, was able to do so with >90% accuracy.

Credit: Springer Nature Limited

The recursive feature elimination technique was then used to determine which genes used by the linear SVM model were most important for differentiating the myositis subtypes. Using this approach, the researchers identified several genes that were overexpressed only in one type of myositis; for example, CAMK1G, EGR4 and CXCL18 were overexpressed only in anti-synthetase syndrome samples. Other genes were expressed at high levels only in certain MSA-defined myositis subtypes, such as APOA4 and MADCAM1, which were upregulated in anti-HMGCR-positive IMNM and anti-Mi2-positive dermatomyositis samples, respectively.

Consistent with previous observations related to myositis pathogenesis, genes associated with interferon signalling were among the top differentially expressed genes in dermatomyositis, and genes related to muscle regeneration were prominently overexpressed in IMNM.

the researchers identified several genes that were overexpressed only in one type of myositis

Although the study did not include specimens from patients with polymyositis or MSA-negative patients, and the analysis was restricted to gene expression (and not protein expression), the findings suggest that analysis of the gene expression profiles of myositis subtypes could provide pathologically relevant information.