A T Cell Receptor Sequencing-Based Assay Identifies Cross-Reactive Recall CD8+ T Cell Clonotypes Against Autologous HIV-1 Epitope Variants

Abstract

HIV-1 positive elite controllers or suppressors control viral replication without antiretroviral therapy, likely via CTL-mediated elimination of infected cells, and therefore represent a model of an HIV-1 functional cure. Efforts to cure HIV-1 accordingly rely on the existence or generation of antigen-specific cytotoxic T lymphocytes (CTL) to eradicate infected cells upon reversal of latency. Detecting and quantifying these HIV-1-specific CTL responses will be crucial for developing vaccine and T cell-based immunotherapies. A recently developed assay, called MANAFEST, uses T cell receptor (TCR) Vβ sequencing of peptide-stimulated cultures followed by a bioinformatic pipeline to identify neoantigen-specific T cells in cancer patients. This assay is more sensitive than conventional immune assays and therefore has the possibility to identify HIV-1 antigenic targets that have not been previously explored for vaccine or T cell immunotherapeutic strategies. Here we show that a modified version of the MANAFEST assay, called ViraFEST, can identify memory CD8 + T cell responses against autologous HIV-1 Gag and Nef epitope variants in an elite suppressor. Nine TCR Vβ clonotypes were identified and 6 of these were cross-reactive for autologous variants or known escape variants. Our findings are a proof of principle that the ViraFEST assay can be used to detect and monitor these responses for downstream use in immunotherapeutic treatment approaches.

DeepTCR is a deep learning framework for revealing sequence concepts within T-cell repertoires

“…Furthermore, when applying this trained repertoire classifier for TCR-level inference, we noted that 17/18 (Fisher’s exact test: p < 1e−10) of the originally reported experimentally validated TCR-peptide pairs were correctly predicted to be cognate binders (Supplementary Fig. 14 ) 31 .

Fig.

…”

“…We present DeepTCR, a platform of both unsupervised and supervised deep learning that is able to be applied at the level of individual T cell receptor sequences as well as at the level of whole T cell repertoires, which can learn patterns in the data that may be used for both descriptive and predictive purposes. In order to demonstrate the utility of these algorithms, we collected a variety of TCR-Seq datasets including samples sorted by antigen specificity 20 – 22 , samples collected from single-cell RNA-seq experiments (10x_Genomics), and samples collected from a novel experimental assay used in detecting functional expansion of T cells 31 (full dataset details in Supplementary Fig. 1 ).…”

“…experimental exposures, therapies, clinical outcomes) apply to an entire repertoire of TCR sequences and not to any individual sequence. To test the utility of this approach, we collected data from published TCR-Seq data of an assay where T cells from an elite suppressor (ES8) of HIV were cultured with autologous HIV-1 Gag and Nef epitope variants and sequenced after culture to determine the immune repertoire against each epitope 31 , 44 – 48 . In the original work, TCR sequences were deemed to be antigen specific if they met certain statistical requirements based on the read count of a given sequence, a proxy for clonal expansion.…”

“…Arguably, methods such as the ones shown in this paper may be uniquely able to discover these changes for the first time. Additionally, while the data that exist to train these models comes through high-throughput methods such as tetramer sorting 21 , 22 or T cell stimulation assays 31 , these methods often introduce some level of noise due to non-specific binding or stimulation. Unfortunately, gold standard methods that would require isolation and cloning of T cell receptors to specifically interrogate the specificity of a given TCR are highly laborious and low throughput.…”

“…All data analyzed in this manuscript can be found at https://github.com/sidhomj/DeepTCR as well as in the original publications that generated the data 20 – 22 , 31 . The 10X Genomics dataset used to train the supervised sequence regression can be found at https://www.10xgenomics.com/resources/application-notes/a-new-way-of-exploring-immunity-linking-highly-multiplexed-antigen-recognition-to-immune-repertoire-and-phenotype/ .…”