Clonal Trafficking of Individual T Cells between Peripheral Blood and Inflamed Synovial Tissue in Patients with Immune-Mediated Refractory Arthritis

Background

Rheumatoid arthritis (RA) is the most prevalent form of immune mediated arthritis, which is characterised by immune-mediated destruction of synovial joints, leading to synovitis, acute joint pain, reduced motility, and mortality. Enrichment of CD4+ memory T cell subsets with local clonal expansions in the RA synovial tissue (ST) has been extensively reported in the literature. In addition, emerging evidence suggests that CD8+ T cells with inflammatory phenotypes are also present in increased numbers in the ST. However, little is known about whether these populations represent autoantigen-experienced migratory T cells or locally proliferated populations. Here, with the help of paired single cell T cell receptor (scTCR) and RNA sequencing recognizing individual T cell clones, we aimed to gain insights into the trafficking of T cells between peripheral blood (PB) and inflamed target tissue.

Methods

We collected ST and PB samples from patients with immune mediated arthritis during surgical synovectomy. The patients had been diagnosed with seronegative juvenile arthritis (Pt1-JIA), seropositive RA (Pt2-SP) and seronegative RA (Pt3-SN), and at the time of sampling they were refractory to treatment. We flow sorted T cells and CD45+ leukocytes from ST and PB and profiled them with 10x Genomics 5´-scRNA+V(D)J-seq. Altogether, we captured 68,610 cells passing quality control thresholds. We integrated the data from all individual samples using scVI and visualised the subpopulation structure using UMAP dimensionality reduction as implemented in the Seurat package.

Results

Altogether we analysed 34134 CD4+ T cells and 23385 CD8+ T cells from PB and ST. In concordance with previous reports, we observed strong transcriptomic differences between the tissue of origin for both CD4+ and CD8+ T cells.

We pooled all CD4+ and CD8+ T cells separately and further grouped all cells with identical CDR3b aa sequences into unique clones. In CD8+ T cells, a high number of clones (n=101, pertaining to 7409 cells) was shared between PB and ST. 8 of these clones were enriched to the ST as compared to PB. Interestingly, Pt2-SP harboured a large clone in the ST consisting of 27% of the CD8+ T cells (CDR3b sequence - CASRGGTSITDTQYF). This clone was also detected in PB with a much smaller frequency of only 2% and in a phenotypically different cluster, indicating local proliferation of specific T cell clones in the synovium. By performing a DE gene analysis between cells with the same TCR but between different tissues of origin, we found that the top upregulated genes in ST were CCL4 and CREM suggesting more activated and inflamed phenotype in ST.

In the CD4+ compartment, we found only 12 intersecting TCR clones pertaining to 637 cells. Majority of these clones were coming from PB (458 cells from PB and 179 from ST) and were hyperexpanded in PB as compared to ST. No matches for common viral or bacterial epitopes in the existing VDJ database for any of the intersecting clones was found. We then pooled all T cell clones that were enriched in the ST and found 113 clones (p < 0.05, Fisher's exact test). DE gene analysis showed that ST enriched clones have a more activated profile. Pathway enrichment analysis with hypergeometric testing against the GO Biological Process database identified T cell activation and positive regulation of cell adhesion as the top hits. Using GLIPH2, we used publicly available data from a cohort of JIA patients to check TCR level similarity between the two cohorts. We found common motifs between these two cohorts in some of the hyperexpanded clones from PB and ST, and interestingly, TCRs which were predicted to target similar antigens, were enriched to both expanded and intersecting clones. By performing DE gene analysis, we found that predicted TCRs with similarity were more cytotoxic ( GZMH, GNLY, NKG7, GZMB, GZMA) and they overexpressed the TRBV7-6 gene.

Conclusions

Our results suggest that the cytokine signalling environment in the ST leads to a widespread activation and inflammatory signature within the T cell compartment. This shift in the phenotype occurs within clonal families and is not due to the selective recruitment of activated cells to inflamed tissue. Additional investigations may elucidate the cellular interactions that promote expansion of specific disease promoting clones in ST and provide a more detailed understanding of the spatiotemporal aspects of this process.