Poster 116: Integrated scRNA-seq and spatial transcriptomics analysis uncovers distinct cellular populations and transcriptomes in human hip synovium between patients with femoroacetabular impingement and osteoarthritis.

OBJECTIVES: Osteoarthritis (OA) is a common degenerative joint disease marked by joint dysfunction and chronic pain. Previous studies have outlined a variety of complex factors associated with cartilage degradation, inflammation, and pathological changes in knee synovium. However, hip and knee OA have been established as immunologically distinct types of joint diseases due to unique cytokine profiles of synovial fluid, and data regarding hip OA synovium is scarce(1). Evidence suggests that primary hip OA may result from cam-type femoroacetabular impingement (FAI). Thus, cam-type FAI has been considered a unique early-phase model of hip OA, allowing scientists to identify regulators implicated in hip OA development. Integrated single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (Spatial-seq) provides a powerful means to better understand unique genetic markers underlying signaling pathways and cellular interactions that may be involved in disease progression (i.e., FAI to hip OA). In this study, we hypothesize that FAI and hip OA patients have distinct transcriptomic profiles and signaling pathways in their synovial tissues. We aim to identify the underlying molecular mechanisms by using scRNA-seq, Spatial-seq, and bioinformatic analyses. METHODS: De-identified human hip synovial tissues were harvested from patients diagnosed with cam- type FAI or with OA (and accompanying cam morphology) according to approved IRB protocols at the University of Rochester (UR). Tissue samples were digested for scRNA-seq and sectioned for Spatial-seq separately and submitted to UR Genomics Research Center (10X Genomics). Distinct cell populations and differentially expressed genes (DEGs) were identified using bioinformatic tools including Seurat(2), Monocle(3), RcisTarget(4), and CellChat(5) R packages. Functional analysis for identified cell subsets was annotated by Gene Ontology term(6). Differentiation trajectories of non-hematopoietic cells and cell-cell crosstalk were also evaluated. RESULTS: Unsupervised clustering of FAI and OA hip synovial cells identified four major conserved groups: NK cells, endothelial cells, myeloid cells, and non-hematopoietic cells. Sub-clustering of integrated FAI and OA non-hematopoietic cells further yielded 5 distinct cell types including lining, sublining and transitional fibroblast-like synoviocytes (FLS) as well as endothelial cells and pericytes (Fig. 1A). In comparison to FAI synovium, OA synovium exhibited 4.5-, 5- and 8.5-fold increase in IL11(+)/CD82(+) sublining FLS (cluster 5), pericytes (cluster 7) and SPP1(+)/KDR(+) capillary endothelial cells (cluster 9), respectively (Fig. 1B). To further investigate the origin of these populations, we performed pseudotime analysis and revealed that pericytes may differentiate from C34(+)/C3(+)/CXCL14(+) sublining perivascular FLS (cluster 6) (Fig. 2A)(6). To determine the putative transcription factors (TFs) regulating the differentiation process between these two cell populations, we identified that KLF4, SOX5 and CREB3L2 (CAMP Responsive Element Binding Protein 3 Like 2) are potential key regulators using DNA-binding motif analysis. We also observed an opposite trend in expression of KLF4 and CREB3L2, with the highest expression levels of KLF4 in sublining perivascular FLS (Fig. 2B). Importantly, by integrating scRNA-seq and Spatial-seq datasets, we were able to map spatial location of identified cell clusters in the hip joint synovium, allowing us to further interrogate cell-cell crosstalk. For example, we observed that pericytes are in close proximity to SPP1(+)/KDR(+) capillary endothelial cells (cluster 9) and identified that VISFATIN- (ITGA5+ITGB1) and SPP1-CD44 are critical signaling pathways between these two cell populations in OA synovium (Fig. 3B). CONCLUSIONS: OA is the most prevalent degenerative joint disorder and is a primary cause of worldwide disability without any available disease modifying therapies. We believe that our findings will reveal key genetic markers related to hip OA progression and facilitate the development of targeted drug deliveries for future therapeutic applications. Our scRNA-seq analysis revealed unique transcriptomic patterns in the synovial cells between FAI and hip OA patients. Integration of non-hematopoietic cells revealed a significant increase in the number of pericytes and endothelial cells present in OA vs. FAI synovium. This is supported by the evidence of increased vasculature in OA, which in turn may promote enhanced infiltration of myeloid cells, cytokines and chemokines into the synovium, leading to an advanced joint inflammation and disease progression(7). Furthermore, we identified that both pericytes and capillary endothelial cells can express visfatin (an adipokine) and may modulate synovial inflammation in OA via integrin-mediated signaling pathways(8). This observation aligns with previous studies that highlight a significant contribution of endothelial cell-pericyte interactions to OA pathogenesis(8). In addition, our finding of decreased expression of KLF4 in pericytes is consistent with previous reports showing that KLF4 inactivation enhances the expression of pericyte markers(9). Our next step is to integrate our scRNA- seq and Spatial-seq datasets of non-hematopoietic cells with myeloid cells to further elucidate how altered cell-cell crosstalk and signaling pathways in synovium may lead to hip OA progression. Indeed, we have analyzed myeloid cells separately, and observed elevated accumulation of “hybrid” macrophages expressing pro- and anti-inflammatory markers (IL10(+)/IL1 (+)/IL1 (+)) in the hip OA compared to FAI synovium.