Transcriptome profiling of osteoclast subsets associated with arthritis: A pathogenic role of CCR2hi osteoclast progenitors

Abstract

Introduction: The existence of different osteoclast progenitor (OCP) subsets has been confirmed by numerous studies. However, pathological inflammation-induced osteoclastogenesis remains incompletely understood. Detailed characterization of OCP subsets may elucidate the pathophysiology of increased osteoclast activity causing periarticular and systemic bone resorption in arthritis. In our study, we rely on previously defined OCP subsets categorized by the level of CCR2 expression as circulatory-like committed CCR2^hi OCPs, which are substantially expanded in arthritis, and marrow-resident CCR2^lo OCPs of immature phenotype and behavior.

Methods: In order to perform transcriptome characterization of those subsets in the context of collagen-induced arthritis (CIA), we sorted CCR2^hi and CCR2^lo periarticular bone marrow OCPs of control and arthritic mice, and performed next-generation RNA sequencing (n=4 for each group) to evaluate the differential gene expression profile using gene set enrichment analysis with further validation.

Results: A disparity between CCR2^hi and CCR2^lo subset transcriptomes (863 genes) was detected, with the enrichment of pathways for osteoclast differentiation, chemokine and NOD-like receptor signaling in the CCR2^hi OCP subset, and ribosome biogenesis in eukaryotes and ribosome pathways in the CCR2^lo OCP subset. The effect of intervention (CIA) within each subset was greater in CCR2^hi (92 genes) than in CCR2^lo (43 genes) OCPs. Genes associated with the osteoclastogenic pathway (Fcgr1, Socs3), and several genes involved in cell adhesion and migration (F11r, Cd38, Lrg1) identified the CCR2^hi subset and distinguish CIA from control group, as validated by qPCR (n=6 for control mice, n=9 for CIA mice). The latter gene set showed a significant positive correlation with arthritis clinical score and frequency of CCR2^hi OCPs. Protein-level validation by flow cytometry showed increased proportion of OCPs expressing F11r/CD321, CD38 and Lrg1 in CIA, indicating that they could be used as disease markers. Moreover, osteoclast pathway-identifying genes remained similarly expressed (Fcgr1) or even induced by several fold (Socs3) in preosteoclasts differentiated in vitro from CIA mice compared to pre-cultured levels, suggesting their importance for enhanced osteoclastogenesis of the CCR2^hi OCPs in arthritis.

Conclusion: Our approach detected differentially expressed genes that could identify distinct subset of OCPs associated with arthritis as well as indicate possible therapeutic targets aimed to modulate osteoclast activity.

Keywords: RNA sequencing; chemokine receptor; flow cytometry; gene expression; mouse arthritis; osteoclast differentiation; osteoclast progenitor (OCP); qPCR.