TCR repertoire profiling revealed antigen-driven CD8+ T cell clonal groups shared in synovial fluid of patients with spondyloarthritis

Abstract

Spondyloarthritis (SpA) comprises a number of inflammatory rheumatic diseases with overlapping clinical manifestations. Strong association with several HLA-I alleles and T cell infiltration into an inflamed joint suggest involvement of T cells in SpA pathogenesis. In this study, we performed high-throughput T cell repertoire profiling of synovial fluid (SF) and peripheral blood (PB) samples collected from a large cohort of SpA patients. We showed that synovial fluid is enriched with expanded T cell clones that are shared between patients with similar HLA genotypes and persist during recurrent synovitis. Using an algorithm for identification of TCRs involved in immune response we discovered several antigen-driven CD8+ clonal groups associated with risk HLA-B*27 or HLA-B*38 alleles. We further show that these clonal groups were enriched in SF and had higher frequency in PB of SpA patients vs healthy donors, implying their relevance to SpA pathogenesis. Several of the groups were shared among patients with different SpAs that suggests a common immunopathological mechanism of the diseases. In summary, our results provide evidence for the role of specific CD8+ T cell clones in pathogenesis of SpA.

Keywords: TCR repertoire; ankylosing spondylitis; psoriatic arthritis; spondyloarthritis; synovial fluid.

Figure 1

Characteristics of clonal T cell repertoire of synovial fluid from patients with SpA. Comparison of clonal diversity (species richness) (A) and oligoclonality (B) between PB and SF repertoires of SpA patients (by Wilcoxon signed-rank test). Oligoclonality is measured by Gini index that equals 0 in even distribution of frequencies and 1 in monoclonal sample. (C, D) Normalized number of shared clonotypes (i.e., having identical CDR3 amino acid sequence and TRBV-segment) per donor in total (C) or depending on the number of matched HLA alleles (D) between donors (Wilcoxon rank sum test). Each dot represents an individual sample. For patients with relapses only repertoires of time point 1 were included in the analysis. Values below the box plots denote the number of donors. The middle line of a box corresponds to the median, the lower and upper hinges correspond to the 25th and 75th percentiles, respectively. The upper and lower whiskers extends from the hinges to the largest and smallest values no further than 1.5 * IQR from the corresponding hinge. Top, middle and bottom panels correspond to the repertoires of total, CD8+ and CD4+ T cells respectively. To assess diversity and clonal sharing repertoires were in silico equalized to the same sequencing depth. PB - peripheral blood, SF - synovial fluid, HLA - human leukocyte antigen. *p < 0.05, ***p < 0.001, ****p < 0.0001.

Figure 2

T cell clonotypes involved in immune response in synovial fluid of SpA patients. (A), Similarity network of SF CD8+ clonotypes identified by the ALICE. Only 29 clusters shared by ≥3 patients are shown. Each node represents a unique amino acid clonotype. Edges connect clonotypes differing in one CDR3 amino acid residue. Node’s size corresponds to the number of donors sharing this clonotype. Color highlights clusters associated with HLA-B*27, HLA-B*38, or found in VDJdb. (B), Occurrence of the clusters in CD8+ SF samples of HLA-B*27+ (n=15) vs HLA-B*38+ (n=7) patients. (C), T cell frequency of HLA-B*27-associated (upper, n(PB)=12, n(SF)=16) or HLA-B*38-associated (bottom, n(PB)=7, n(SF)=8) clusters in CD8+ repertoires of patients. Occurrence (D) and T cell frequency (E) of HLA-B*27-associated (upper, n(SpA)=29, n(HD)=51) or HLA-B*38-associated (bottom, n(SpA)=14, n(HD)=37) clusters in total PB repertoires of SpA patients and HD. Black circles mark clusters significantly enriched by Fisher’s exact test with Benjamini-Hochberg adjusted p<0.05. On the box plots the middle line corresponds to the median, the lower and upper hinges correspond to the 25th and 75th percentiles, respectively. The upper and lower whiskers extends from the hinges to the largest and smallest values no further than 1.5 * IQR from the corresponding hinge. Each dot represents an individual sample. For patients with relapses only repertoires of time point 1 were included in the analysis. PB - peripheral blood, SF - synovial fluid, SpA - spondyloarthritis, HD - healthy donors. Wilcoxon rank sum test with Benjamini-Hochberg adjustment *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 3

Characteristics of identified TCRbeta clusters. Sequence logo of CDR3 regions (A) and details (B) of the 9 identified clusters associated with HLA-B*27 or HLA-B*38. Values in parentheses represent the number of distinct clonotypes. PB, peripheral blood; SF, synovial fluid; NA, not applicable; ND, no data.