Objective: Systemic sclerosis (SSc) is a multifactorial autoimmune fibrotic disorder involving complex rewiring of cell-intrinsic and cell-extrinsic signaling coexpression networks involving a range of cell types. However, the rewired circuits as well as corresponding cell-cell interactions remain poorly understood. To address this, we used a predictive machine learning framework to analyze single-cell RNA-sequencing data from 24 SSc patients across the severity spectrum as quantified by the modified Rodnan skin score (MRSS).
Methods: We used a least absolute shrinkage and selection operator (LASSO)-based predictive machine learning approach on the single-cell RNA-sequencing data set to identify predictive biomarkers of SSc severity, both across and within cell types. The use of L1 regularization helps prevent overfitting on high-dimensional data. Correlation network analyses were coupled to the LASSO model to identify cell-intrinsic and cell-extrinsic co-correlates of the identified biomarkers of SSc severity.
Results: We found that the uncovered cell type-specific predictive biomarkers of MRSS included previously implicated genes in fibroblast and myeloid cell subsets (e.g., SFPR2+ fibroblasts and monocytes), as well as novel gene biomarkers of MRSS, especially in keratinocytes. Correlation network analyses revealed novel cross-talk between immune pathways and implicated keratinocytes in addition to fibroblast and myeloid cells as key cell types involved in SSc pathogenesis. We then validated the uncovered association of key gene expression and protein markers in keratinocytes, KRT6A and S100A8, with SSc skin disease severity.
Conclusion: Our global systems analyses reveal previously uncharacterized cell-intrinsic and cell-extrinsic signaling coexpression networks underlying SSc severity that involve keratinocytes, myeloid cells, and fibroblasts.
© 2023 American College of Rheumatology.