T cell receptor costimulation blockade effectively prevents acute graft versus host disease without reducing the size or diversity of the T cell immune repertoire

Abstract

Abstract Background: Both graft versus host disease (GVHD) and infections remain major causes of morbidity and non-relapse mortality after hematopoietic cell transplant (HCT). As acute GVHD (aGVHD) is a T cell mediated alloreactive process, several T cell depleting, suppressing, and modulating strategies have been employed to prevent GVHD. Abatacept modulates T cell activity through CD28:CD80/86 costimulation blockade. In the ABA2 trial, a 4-dose regimen of abatacept (10mg/kg on days -1, +5, 14, 28) when added to CNI/MTX (the ‘ABA’ arm) significantly reduced aGVHD without increasing infections, when compared to CNI/MTX/Placebo (‘PBO‘), after unrelated-donor HCT. Understanding the biology underpinning ABA's success in preventing aGVHD, as well as its potential immunologic trade-offs, is crucial to advance prevention and treatment of GVHD. Here, we studied the size and diversity of the T cell repertoire in the ABA2 trial. Methods: We performed T cell receptor (TCR) sequencing using TCRB Immunosequencing (Adaptive Biotechnologies) of samples from ABA2's 185 participants. We analyzed peripheral blood TCRs at baseline (prior to conditioning) and post-HCT on days 28, 63, 100, 180, and 365. In addition, TCR sequencing data from select transplant product bags was available. A total of 889 samples were analyzed with a median of 3,725,249 TCRs profiled per time point. To assess TCR diversity, we calculated Shannon clonality, for which 0 represents a completely even sample and 1 a monoclonal sample. We also assessed richness (the number of unique TCRs in a sample) and singleton TCRs after performing clone abundance-based down-sampling. Finally, to assess repertoire stability and turnover, we calculated Morisita indices, which also range 0-1, with 1 representing complete repertoire overlap between timepoints. Results: Both PBO and ABA patients demonstrated high TCR diversity, which was evident early and persisted throughout the first year post-HCT. Notably, rather than decreasing TCR diversity vs CNI/MTX, as has been demonstrated for PT-Cy/Tac/MMF, CNI/MTX/ABA actually increased TCR diversity compared to PBO: For example, on day 28, median Shannon Clonality for PBO = 0.095 vs ABA = 0.036 (p <0.001); on day 60 median Shannon Clonality for PBO = 0.155 vs ABA = 0.074 (p = 0.04). Down-sampled singleton TCR analysis at Day 28 similarly revealed a significantly higher number of singletons in those receiving ABA vs PBO (median singleton TCR count of 730 for PBO vs 863 for ABA when data statistically down-sampled to 1000, p = 0.003; median singleton count of 2882 for PBO vs 3846 for ABA when data statistically down-sampled to 5000, p<0.001). Similarly, richness was higher in ABA compared to PBO at Day 28 (median richness 799 for PBO vs 904 for ABA when data statistically down-sampled to 1000, p = 0.003; 3381 for PBO vs 4205 for ABA when data statistically down-sampled to 5000, p<0.001). Together, these measurements provide a consistent illustration of intact TCR repertoire diversity in ABA patients despite significant control of aGVHD. Analysis of TCR sharing between two successive time points demonstrated no difference in the achievement of repertoire stability between PBO and ABA, with no difference in Morisita indices when comparing graft vs day 28, day 28 vs 60, and day 60 vs 100. Notably, ABA patients demonstrated more repertoire turnover between Days 100-180 compared to PBO (median Morisita index 0.79 for PBO vs 0.54 for ABA, p=0.01), suggesting increased T cell dynamics after abatacept levels waned. Conclusions: Here we demonstrate that ABA did not reduce either the size or diversity of the T cell repertoire in HCT recipients, despite its ability to control aGVHD. These results are consistent both with the body of work establishing the mechanism of ABA as being immune modulating rather than T cell depleting, and with the intact protective immunity observed in ABA2. Notably, the higher number of singletons and higher richness at day 28 in those receiving ABA is consistent with our previous results demonstrating that ABA preserves naïve T cells. This naïve T cell preservation, along with repertoire turnover after Day 100, is consistent with the control of aGVHD with ABA without a beneficial impact on chronic GVHD.Together, these findings underscore the ability of ABA to preserve the TCR repertoire, providing a key biologic explanation for its ability to control aGVHD without increasing the risk of infectious complications after HCT.