CD8 T cells cross-restricted by HLA-B*57 and HLA-E*01 recognize HIV Gag with different functional profiles [dataset 1]

Few non-classical HLA-E restricted HIV-specific epitopes have been described, and even less is known about the functional profile of responding CD8 T cells (CD8s). This study evaluates the functional characteristics of CD8s targeting the Gag epitope (KAFSPEVIPMF or KF11) based on their restriction by either HLA-E (E-CD8s) or HLA-B57 (B57-CD8s). CD8s from 8 people with HIV (PWH) were cocultured with KF11 peptide presented by cell lines expressing HLA-B*57:01, HLA-E*01:01 or E*01:03. CD8s were analyzed through single-cell (sc) RNA and TCR sequencing. Additionally, supernatants were analyzed for soluble proteomics using a Luminex assay. B57-CD8s secreted higher levels of cytotoxic cytokines such as IFN?, while E-CD8s produced more chemotactic cytokines, including RANTES, CXCL10 (IP-10), and IL27confirmed through scRNAseq. Despite distinct cytokine profiles, TCR clonotypes stimulated by KF11 were cross-restricted by HLA-B*57 and HLA-E*01/03. In vitro T cell reporter assays clearly demonstrated this cross-restriction. A TRAV5-containing metaclonotype cluster was seen in PWH with lower viral loads. These findings demonstrate that HIV-specific CD8s in PWH exhibit cross HLA-B*57 and HLA-E*01/03 restriction, resulting in functionally distinct immune responses that may contribute to HIV control. Overall design: CD8+ T cells were isolated from PBMC's of 1 HIV+ controller (C) Ci10004, noncontroller (NC) Ci10076, and ART+ noncontroller (ART+ NC) Ci10074, or just PBMC's were used. These CD8+ T cells were co-cultured with HLA-null cell line 721.221?E transduced with nothing (null), HLA-B*57, E*01:01, or E*01:03, either loaded with no HIV-1 gag KF11 (unstimulated) or loaded with KF11 (stimulated). Alternatively, PBMC's from these same people with HIV (PWH) were unstimulated or stimulated with peptide. This stimulation (co-culture or PBMC) was performed for 18 hours before staining with oligo-tagged antibodies (hashtagging), then CITE-seq and fluorophore-tagged antibody cocktail. An equal number of Activated and unactivated CD8+ T cells in each of these conditions were then sorted into the same tube for each of the 3 PWH (3 tubes of 10 multiplexed conditions each). These were then used for library construction. There were 10 conditions (5 HLA/PBMC conditions - null, E01, E03, B57, PBMC X 2 stimulation conditions (unstimulated, stimulated with KF11)) per PWH, for 30 conditions total. After this, GEX, CITE-seq, and TCR data generated 2 separate sets of files (GEX and CITE-seq were both in GEX_FB) for each multiplexed of 3 samples. Demultiplexing was performed as described in the manuscript to generate 30 conditions X 3 sample types (but GEX and feature barcoding were demultiplexed to the same GEX_FB file from which CITE-seq and GEX information was obtained) = 60 files. Through the pipeline and based on folder structure, metadata was applied holistically and 30 sets of conditions, each with TCR, GEX, and CITE-seq data were assembled into a single Seurat file at which point downstream analysis was performed. Though CD8 T cells were sorted with high purity, some of the co-cultured cell line was also incorrectly sorted and formed a distinct cluster far from all other clusters which was also used for analysis of the source of Luminex-identified cytokines - thus, the "cell line" column indicates the co-cultured HLA-expressing cell line as well, as this may be relevant for that cluster. Please note that processed data files generated from both GEX and FB raw files are linked to the correpsonding GEX records.