High-Throughput T Cell Receptor (TR) Repertoire Analysis of Virus-Specific T Cells: Implications for T Cell Immunotherapy and Viral Infection Risk Stratification

Viral infections, mainly by cytomegalovirus (CMV), Epstein Barr virus (EBV) and polyomavirus type I (BKV), are major causes of morbidity and mortality after allogeneic hematopoietic stem cell transplantation (allo-HSCT). As effective immune responses against human viruses rely on an armamentarium of T-cell receptor (TR) repertoire capable of recognizing a broad range of antigenic peptides of those pathogens, reconstitution of antiviral immunity, either by spontaneous generation of endogenous virus-specific T cells (VSTs) or by adoptive immunotherapy with VSTs, plays a critical role to fight infections.

We here evaluated the diversity and clonality of TR repertoire of functional tri-virus-specific T cell products generated from immunocompetent donors (n=10) and compared their TR gene repertoire to that of peripheral blood mononuclear cells (PBMCs) from patients who had undergone allo-HSCT (n=5). To generate tri-VSTs, PBMCs derived from 15-20ml of peripheral blood of normal donors, were exposed to EBV, CMV and BKV overlapping peptides and cultured in the presence of interleukin 4 (IL-4) and IL-7 for 10 days in G-rex bioreactors. Specificity of donor-derived VSTs and patient-derived PBMCs was measured by IFN-γElispot. TR diversity was investigated by next-generation sequencing on a MiSeq Sequencer, after amplification of TR beta chain gene rearrangements by RT-PCR with the BIOMED-2 protocol. Raw NGS reads were filtered based on their length and quality and the filtered-in sequences were submitted to IMGT/HighVQUEST. Metadata analysis and clonotype computation were performed using a validated in-house bioinformatics platform. As clonotype we defined sequences carrying the same TRBV gene and identical CDR3 amino acid sequence.

Tri-VSTs provided 947,298 productive TRBV-TRBD-TRBJ rearrangements and a polyclonal and highly diverse TR gene repertoire, consisting of a total of 169,502 unique clonotypes (average: 16,950/sample, range 4,057-45,602), 64,971 (38.3%) of which were expanded (corresponding to more than one sequence). In terms of clonality, the mean relative frequency of the major clonotype in all tri-VSTs was 12.6% (range 3.3-29.2%). Interestingly, among tri-VST cell lines, 637 clonotypes were shared (present in >2/10 samples), 80 were highly shared (present in >3/10 samples) while 7 were present in 6-8 different VST lines and largely expanded, accounting for up to 29.2% of all sequences.

Importantly, there were 65 of 96 major VST clonotypes shared, thus suggesting that they were potentially associated with recognition of the targeted viruses. Given that 4/10 VSTs cell lines were not specific for CMV, while being EBV-and BKV-specific, dominant TRs in those 4 cell lines can potentially be associated with EBV- or BKV-activity. By searching a public database of TR clonotypes with known reactivity against EBV and/or CMV (ShugayM, Nucleic Acids Research, 2018), we found 8 shared EBV-specific and 4 shared CMV-specific clonotypes among our VSTs and the 499 public clonotypes. When we compared the produced VSTs with PBMCs from 3 allo-grafted patients with circulating CMV-, BKV- and EBV-specific T cells and previous viral reactivation, we detected 163 shared clonotypes. Likewise, we observed 21 and 23 shared clonotypes in similar frequencies, between VSTs and PBMCs from 2 patients with CMV- or BKV-specific T cell immunity. These data identify clones that potentially expand in vivo and protect patients from viral infections.

Overall, our findings reveal high levels of TR clonality in cell lines enriched for T cells reactive against EBV and/or CMV and/or BKV and provide insights into the TR repertoire of ex vivo- or endogenously-generated VSTs. Our approach may help to identify optimal TRs for immunotherapy as well as TRs which can be used as a tool for risk stratification of viral infections.