Abatacept increases T cell exhaustion in early RA individuals who carry HLA risk alleles

Abstract

Exhausted CD8 T cells (T_EX) are associated with worse outcome in cancer yet better outcome in autoimmunity. Building on our past findings of increased TIGIT⁺KLRG1⁺ T_EX with teplizumab therapy in type 1 diabetes (T1D), in the absence of treatment we found that the frequency of TIGIT⁺KLRG1⁺ T_EX is stable within an individual but differs across individuals in both T1D and healthy control (HC) cohorts. This TIGIT⁺KLRG1⁺ CD8 T_EX population shares an exhaustion-associated EOMES gene signature in HC, T1D, rheumatoid arthritis (RA), and cancer subjects, expresses multiple inhibitory receptors, and is hyporesponsive in vitro, together suggesting co-expression of TIGIT and KLRG1 may broadly define human peripheral exhausted cells. In HC and RA subjects, lower levels of EOMES transcriptional modules and frequency of TIGIT⁺KLRG1⁺ T_EX were associated with RA HLA risk alleles (DR0401, 0404, 0405, 0408, 1001) even when considering disease status and cytomegalovirus (CMV) seropositivity. Moreover, the frequency of TIGIT⁺KLRG1⁺ T_EX was significantly increased in RA HLA risk but not non-risk subjects treated with abatacept (CTLA4Ig). The DR4 association and selective modulation with abatacept suggests that therapeutic modulation of T_EX may be more effective in DR4 subjects and T_EX may be indirectly influenced by cellular interactions that are blocked by abatacept.

Keywords: HLA risk alleles; T cell exhaustion; abatacept; autoimmunity; rheumatoid arthritis.

Figure 1

TIGIT⁺KLRG1⁺ CD8 T cells are a stable cell type that varies across individuals. TIGIT⁺KLRG1⁺ CD8 T cells were measured by flow cytometry in longitudinal samples from (A) Individuals with type 1 diabetes (T1D; n = 66) and (B) healthy control subjects (HC; n = 99). T1D samples were collected at 6-month intervals over 2 years. HC samples were collected over a median of 8.7 months (inter-quartile range 7.2 to 14.2). Multiple samples (data points) isolated from individual subjects, each shown on a line, are graphed for both cohorts. Individuals are ordered by mean % TIGIT⁺KRLG1⁺ and annotated for CMV seropositivity by color. ICC, Intraclass correlation coefficient.

Figure 2

Across disease settings, co-expression of TIGIT and KLRG1 marks memory CD8 T cells with an EOMES-associated transcriptional signature. (A) Bulk RNA-seq data from sorted TIGIT⁺KLRG1⁺ and TIGIT^-KLRG1^- CD8 memory (not CD45RA⁺CCR7⁺ naïve) T cells isolated from healthy control subjects (HC) (n = 4) following 16-hour anti-CD3/CD28 stimulation. Selected markers of cell cycle, inhibitory receptor and transcription factor expression are annotated. (B) Comparison of sorted TIGIT⁺KLRG1⁺ and TIGIT^-KLRG1^- memory CD8⁺ T cells across multiple disease settings: age- and gender-matched HC; type 1 diabetes (T1D); renal cell carcinoma (RCC); cytomegalovirus infection (CMV-pentamer positive cells); and influenza infection (FLU-pentamer positive cells). (C) Correlation of EOMES protein expression and TIGIT⁺KLRG1⁺ in memory (not CD45RA+CCR7+ naïve) CD8 T cells in HC (n = 29), Spearman test with 95% confidence interval (dotted lines). Gating for sorts and analyses are shown in Supplementary Figure 2.

Figure 3

TIGIT⁺KLRG1⁺ memory CD8 T cells are dysfunctional in healthy control subjects and increased in terminal cell subsets and chronic viral reactive cells. (A) Proliferation following 3-day anti-CD3/CD28 stimulation of TIGIT⁺KLRG1⁺ (T⁺K⁺) memory cells relative to total memory CD8⁺ T cells from healthy control (HC) subjects (n = 12). Proliferation of memory CD45RO⁺ cells was measured by percentage of divided cells using a flow cytometry dye dilution assay. (B) Pro-inflammatory cytokine production (TNF-α and IFN-γ) following 24-hour anti-CD3/CD28 stimulation by gated T⁺K⁺ memory cells relative to total memory CD8⁺ T cells in HC subjects (n = 56). Cytokine production was measured by intracellular cytokine staining. (C) Effector cell surface marker expression (CD226 and CD127) and (D) Inhibitory receptor expression in the absence of T cell activation in gated T⁺K⁺ cells relative to total memory CD8⁺ T cells from HC subjects (n = 28). Wilcoxon matched-pairs signed-rank test was used in all comparisons. (E) Distribution of TIGIT⁺KLRG1⁺ cells within naïve (CD45RO^-CCR7⁺), central memory (CM: CD45RO⁺CCR7⁺), effector memory (EM: CD45RO⁺CCR7^-) and RA⁺ effector memory (EMRA: CD45RO^-CCR7^-). One representative HC sample shown from C. (F) TIGIT⁺KLRG1⁺ distribution in a subset of HLA-A2 subjects stained with Flu-, CMV- and EBV-specific Class I Pentamer (Pmr). Kruskal-Wallis test with Dunn’s correction for multiple tests. Gating for memory TIGIT⁺KLRG1⁺ is shown in Supplementary Figure 2 , gating for activating and inhibitory markers is shown in Supplementary Figure 4 . **=0.05, ***=0.005, ****=0.0005 p-values.

Figure 4

The frequency of TIGIT⁺KLRG1⁺ T_EX is influenced by RA HLA risk alleles. (A) Whole blood RNA-seq data of age- and sex-matched HC (n = 114) and RA (n = 97) subjects were parsed by RA-associated HLA risk genotype (DRB1*0401, 0404, 0405, 0408, 1001). Dark blue, EOMES module; light blue, EOMES module overlap; gray, no overlap with EOMES module. (B) Frequency of TIGIT⁺KLRG1⁺ memory CD8 T cells in age- and sex-matched HC and RA subjects (n = 10/cohort) selected for top or bottom tercile EOMES signature: Left, HC versus RA; Right, Risk RA HLA versus non-risk RA HLA. Mann-Whitney test. Gating shown in Supplementary Figure 2 .

Figure 5

TIGIT⁺KLRG1⁺T_EX are selectively increased with abatacept therapy in RA subjects carrying RA HLA risk alleles. (A) Frequency of TIGIT⁺KLRG1⁺ T_EX in memory CD8⁺ T cell compartment in individuals at risk for T1D treated with teplizumab (anti-CD3) stratified by DR4 risk and DR4 non-risk. (B) Frequency of TIGIT⁺KLRG1⁺ T_EX in memory CD8⁺ T cell compartment in individuals with new onset rheumatoid arthritis (RA) treated with abatacept (CTLA4-Ig) stratified by risk RA HLA and non-risk RA HLA. (C) Frequency of TIGIT⁺KLRG1⁺ T_EX in memory CD8⁺ T cell compartment in individuals with new onset RA adalimumab (anti-TNF) stratified by risk RA HLA and non-risk RA HLA. In B and C, risk RA HLA carried either DRB1*0401, 0404, 0405, 0408, or 1001 and non-risk RA HLA did not. In all three trials, age and CMV status (where available) did not differ between HLA groups. Wilcoxon matched-pairs signed-rank test was used for each risk group in all studies. Gating shown in Supplementary Figure 2 .