OP0102 THE SYNOVIUM IN CHRONIC INFLAMMATORY JOINT DISEASES: COMPARISON OF CLINICAL, HISTOLOGICAL AND scRNA-SEQ Data BETWEEN RA, PsA, SpA AND UA

Abstract

BackgroundThe synovium is the primary location of inflammation in various rheumatic diseases. However, specific differences of joint inflammation have not been explored on a single-cell level so far.ObjectivesTo characterize the synovium of rheumatoid arthritis (RA), psoriatic arthritis (PsA), spondyloarthritis (SpA) and undifferentiated arthritis (UA) based on clinical and histological characteristics and single-cell RNA sequencing (scRNA-seq).MethodsWe performed ultrasound (US) guided synovial biopsy in patients fulfilling classification criteria: 10 RA, 4 PsA, 4 SpA and 3 UA. 3 osteoarthritis (OA) samples were obtained from surgery. Clinical data were collected at time of biopsy. Histological analysis of the synovium included Krenn score [1], synovial pathotype [2], vascularization [3] and presence of lymphoid follicles. OA histology was not available. We prepared scRNA-seq libraries with 10X Genomics and sequenced on NovaSeq 6000. ScRNA-seq data was analysed with Cell Ranger, Seurat and Harmony R packages. We selected overexpressed genes using log2 ratio (>0.25) and FDR adjusted p value < 0.05.ResultsPatients showed typical disease characteristics. In RA, 6/10 were seropositive, 8/10 were female and median age was 59 years (IQR 12). Patients with PsA, SpA and UA were seronegative. 2/4 SpA and 1/4 PsA patients were HLA-B27 positive. Median age was 53 years in PsA (IQR 14.2), 52.5 in SpA (IQR 14.8) and 53 in UA (IQR 3.5). In PsA 3/4, SpA 2/4 and UA 1/3 patients were male.RA joints had significantly higher B-Mode US scores compared to PsA and SpA joints (mean B-Mode Score: RA 2.7; PsA 2; SpA 2). Correspondingly, RA patients reported a significantly higher amount of swelling in the biopsied joint compared to PsA and SpA (mean “Swelling-Score”: RA 6.85; PsA 4.25; SpA 4).Histology showed no clear differences in the cell composition; most joints showed a lympho-myeloid pathotype. Mean Krenn Score was highest in RA (4.78) and lowest in SpA (3.33). PD-Mode US score showed a significant positive association with histological vascularization. Lymphoid follicles were significantly more seen in RA compared to all other diseases.Integration of scRNA-seq data revealed 16 cell clusters with respective marker gene: PRG4 synovial fibroblasts (SF), THY1 SF, ACTA2 smooth muscle cells, VWF endothelial cells, TPSB2 mastcells, IGHG plasma cells, CD79 B-cells, CD69 T-Cells, CCL5 natural killer cells, DMTN Treg-cells, CXC2 neutrophil granulocytes, TREM2 macrophages, CD48 macrophages, CLEC10A macrophages, CD4 dendritic cells, MKI67 proliferating cells. OA had the lowest and RA the highest number of inflammatory cells within the synovium. SpA synovium had the highest number of neutrophils. Re-Clustering of only SF revealed 4 subtypes: PRG4 SF (3578 cells), CXCL12 SF (4539 cells), POSTN SF (2387 cells) and MFAP5 SF (1734 cells). Top 5 marker genes are shown in the Table 1. Most differentially expressed genes in OA were found in PRG4 SF; the previously described OA specific gene CSN1S1 [3] was not only underexpressed in RA but also in other inflammatory joint diseases (Figure 1). PGF, a gene associated with pathological angiogenesis, was one of the top discriminator genes, highly expressed in PsA SF (mostly PRG4 SF). RA SF (mostly CXCL12 SF) showed a high expression of CHI3L1, which is a RA specific autoantigen. High expression of hemoglobin genes was found in SpA PRG4 SF; an unexplored but previously described finding [5, 6].Table 1.Top 5 synovial fibroblast (SF) marker genes.ClusterGenePRG4 SFPRG4CSN1S1MMP3CRTAC1HTRA1CXCL12 SFAPOECXCL12CCL2PTGDSCD74POSTN SFCOMPCOL1A1POSTNCILPPTNMFAP5 SFMFAP5SFRP2APODCXCL14CFDConclusionWe were able to compare the synovium of the most common chronic inflammatory joint diseases on various levels for the first time. The findings set the path for future diagnostic, prognostic, and therapeutic approaches in inflammatory joint diseases.