Comprehensive T-Cell Receptor Repertoire Analysis Using Deep Sequencing and Single Cell Cloning Reveals Extreme Oligoclonality of Ex Vivo Expanded Cytomegalovirus-Reactive Cytotoxic T-Cells

Background: Identification of functional T-cell receptors (TCRs) to their cognate antigens is the key to the development of effective anti-viral or anti-tumor T-cell therapy. Deep sequencing of rearranged complementarity-determining region 3 (CDR3) regions of TCRA and TCRB gene segments is an emerging technology that facilitates high-throughput and semi-quantitative analysis of TCR repertoire with high resolution and accuracy. However, this method alone does not yield data for correct pairs of TCRA and TCRB sequences required for the structural determination of TCRαβ heterodimers expressed in a single T cell. "Human TCR efficient cloning within 10 days" (hTEC10) is a powerful novel technology that enables concurrent sequencing of paired TCRA and TCRB gene segments at a single cell level (Nat Med. 2013;19:1542-6). In this study, we attempted to elucidate the comprehensive TCR repertoire of cytomegalovirus (CMV)-reactive cytotoxic T-cells (CTLs) by combining quantitative deep sequencing and hTEC10.

Methods: 20 ml peripheral blood samples were collected from healthy adult volunteers who gave written informed consent for the study. All donors were screened for CMV serostatus and typed for HLA-A. CMV-specific CD8+ T-cells were isolated and flow-sorted from peripheral blood (PB) mononuclear cells using an HLA-tetramer, NLV/A2, specific for HLA-A2-restricted CMV pp65-derived epitope (NLV peptide: NLVPMVATV). In some experiments, sorted NLV/A2-positive cells underwent one or two rounds of expansion with autologous PB mononuclear cells depleted of CD8+ and CD4+ T cells in the presence of NLV peptide. Using these samples, massive parallel sequencing of TCRA and TCRB V-D-J segments was performed after unbiased amplification of the target sequences by adaptor-ligation PCR. Concurrent TCRA and TCRB sequencing of TCRs expressed in a single sorted cell was performed by the hTEC10 protocol.

Results: Of 20 donors who participated in the study, 8 were found to be CMV-seropositive and HLA-A2-positive (CMV+A2+). The samples obtained from 2 of CMV+A2+ donors, V001 and V004, were subjected to comprehensive TCR analysis by quantitative deep sequencing and single cell cloning: the respective frequencies of NLV/A2 tetramer-positive cells in PB of these donors were 0.36% and 0.25% of a CD8+ T-cell fraction. After two rounds of NLV-peptide stimulation, NLV/A2-tetramer-positive cells obtained from V001 and V004 were enriched to 84.3% and 90.5% of CD8+ T-cells, respectively. By deep sequencing analysis, the total number of unique TCRA/TCRB reads in PB samples from these donors was 1472/5787 in V001 and 2054/9179 in V004, while that of the enriched NLV/A2 tetramer-positive fractions was 178/62 in V001 and 100/104 in V004. However, in both donors, TCR repertoire of the expanded tetramer-positive cells was extremely skewed and the number of the most abundant top 3 reads comprised more than 90-95% of total reads for both TCRA and TCRB.

To confirm the correct pairing of these TCRA and TCRB clonotypes at a single cell level, we also examined TCR sequences of the enriched tetramer-positive cells by hTEC10. A total of 180 cells each obtained from V001 and V004 were subjected for analysis. We identified 3 TCRA and TCRB clonotypes in samples from V001 and 6 clonotypes in those from V004. Importantly, 7 of 9 clones identified by hTEC10 were not listed in the clonotype determined by deep sequencing. To evaluate the effector functions of the cloned TCRs, paired TCRA and TCRB gene segments obtained by hTEC10 were transduced into PHA blasts established from T cells derived from CMV seronegative HLA-A2-negative donor. We confirmed that these PHA blasts transduced with HLA-A2-restricted CMV-specific TCRA and TCRB genes were NLV/A2-tetramer positive. Furthermore, these cells secreted INF-γ in response to NLV peptide as measured by ELISA.

Conclusions: Ex vivo expanded anti-CMV CTLs reactive with NLV/A2 tetramers were extremely oligoclonal and consisted of only a few dominant clones. Determination of CDR3 sequences of these clones was feasible by quantitative deep sequencing combined with single cell cloning of TCRA and TCRB gene segments, although the results were complementary and not identical. This method could be useful for the efficient screening of the highly functional TCRs for adoptive T-cell immunotherapy.