T cell response to intact SARS-CoV-2 includes coronavirus cross-reactive and variant-specific components

Abstract

SARS-CoV-2 provokes a robust T cell response. Peptide-based studies exclude antigen processing and presentation biology, which may influence T cell detection studies. To focus on responses to whole virus and complex antigens, we used intact SARS-CoV-2 and full-length proteins with DCs to activate CD8 and CD4 T cells from convalescent people. T cell receptor (TCR) sequencing showed partial repertoire preservation after expansion. Resultant CD8 T cells recognize SARS-CoV-2-infected respiratory tract cells, and CD4 T cells detect inactivated whole viral antigen. Specificity scans with proteome-covering protein/peptide arrays show that CD8 T cells are oligospecific per subject and that CD4 T cell breadth is higher. Some CD4 T cell lines enriched using SARS-CoV-2 cross-recognize whole seasonal coronavirus (sCoV) antigens, with protein, peptide, and HLA restriction validation. Conversely, recognition of some epitopes is eliminated for SARS-CoV-2 variants, including spike (S) epitopes in the Alpha, Beta, Gamma, and Delta variant lineages.

Keywords: Antigen-presenting cells; Dendritic cells; Infectious disease; T cells.

Figure 1. AIM detection and enrichment of SARS-CoV-2–specific T cells in response to whole virus.

(A) PBMC from participant W003 incubated with inactivated cell-associated SARS-CoV-2 or mock antigen. Gating scheme at top. Lower panels show expression of activation markers CD137 and CD69 in response to 18-hour stimulation among CD4 or CD8 T cells. (B) Similar layout for subject W005, specimen 1 PBMC stimulated with autologous moDC pretreated with SARS-CoV-2 or mock antigen. Gating scheme at top. For both stimulation methods, numbers are percentages of gated T cells expressing dual activation markers. (C) CD69⁺/CD137⁺ CD4 T cells from the pathway in A were expanded and tested for reactivity with inactivated cell-associated SARS-CoV-2 or mock antigen. Gated, live, responder, CD3⁺/CD4⁺ CD8 cells are shown. Numbers are percent of cells accumulating the indicated cytokines.

Figure 2. Recognition of infected respiratory epithelial cells by SARS-CoV-2–specific CD8 T cells.

(A) HBEC3-KT-A cells with or without infection with SARS-CoV-2 were cocultivated with tetramer-enriched, HLA-A*03:01-restricted, S aa 378-386–specific CD8 T cells and activation measured by IFN-γ secretion. S, S-specific T cells; M, control MCPyV-specific CD8 T cells. (B) Both T cell populations specifically recognized HBEC3-KT-A cells treated with relevant viral peptide. Triplicate raw data points are shown with the bar representing the mean of triplicate. Results are representative of 3 repeat experiments.

Figure 3. SARS-CoV-2 CD8 T cell antigens and epitopes from PBMC stimulation with whole SARS-CoV-2 antigen.

(A) Subject W005, specimen 1 CD8 TCL were assayed with aAPC expressing each SARS-CoV-2 protein and relevant HLA-A and -B. Four reactivities were noted. (B) Subject W010 CD8 TCL is reactive with HLA-A*01:01 aAPC cotransfected with NSP3. For A and B, negative controls are at right. (C) CD8 TCL from B assayed against column (C) and row (R) pooled NSP3 peptides with autologous EBV-LCL as APC. (D) Tetramer stain of CD8 TCL before and after sorting and expansion of tetramer⁺ cells. Percentages of tetramer⁺ cells shown. (E) Reactivity of tetramer-enriched cells for aAPC transfected with the indicated plasmids or treated with the indicated peptides. (F) Dose response for HLA-A*01:01 aAPC with the indicated concentrations of NSP3 aa 1637–1646. Duplicate or triplicate IFN-γ release assays show raw data as bars (A) or dots for each value and means as bars (B, C, E, and F). Results are representative of 1–3 repeat experiments per panel.

Figure 4. Summary of SARS-CoV-2 proteome-level T cell reactivity for PBMC after COVID-19 illness, using AIM enrichment with whole viral antigen.

Rows indicate donors and days between recovery from illness and PBMC sampling. Upper 5 rows were studied without DC boosting; lower 12 rows used this procedure. Columns are individual SARS-CoV-2 proteins. (A) CD8 TCL scoring positive (purple). (B) CD4 TCL scoring positive (red). At right are number of proteins recognized and, at bottom, are number of subjects with reactivity at one or more time points. Each positive cell represents replicates, as detailed in Methods. M, membrane protein of SARS-CoV-2.

Figure 5. CD4 T cell coronavirus crossreactivity.

(A) CD4 TCL from PBMC stimulated with moDC and whole SARS-CoV-2 (subjects W001, W005, W012) or PBMC stimulated with whole SARS-CoV-2 (subject W003) recognize whole SARS-CoV-2, and sCoVOC43 or 229E antigens (V), but not mock (M) antigens. (B) CD4 TCL from subject W001 recognizes full-length S protein from OC43 but not empty vector control. (C) CD4 TCL recognize homologous S peptides from SARS-CoV-2 and OC43 in an HLA-DR–restricted fashion as indicated by inhibition with locus-specific mAb. An overlapping SARS-CoV-2 peptide shows DRB1*15:01 restriction at right. Conserved aa are underlined. Duplicate or triplicate raw data and mean bars are shown. Results are representative of 1–2 repeat experiments per panel.

Figure 6. CD8 TCL recognition by SARS-CoV-2 variant peptides.

aAPC transfected with HLA-A*02:01 and treated with WT but not variant peptides were recognized by polyclonal CD8 TCL lines. Lineage B.1.351 is also known as Beta and P.1 is also known as Gamma. Triplicate raw data and mean bars are shown.

Figure 7. CD4 TCL recognition by SARS-CoV-2 variant peptides.

Summary of recognition of S strain Wu-1 peptides and variants. Donors indicated at top. Each Wu-1 peptide recognized by 1 or more subject is numbered at left. Coordinates and aa substitutions or deletions of variants tested are listed at right. Variants in WHO VOC are indicated with blue squares. Color codes at left summarize level of recognition of variant peptide by each TCL. VBM, variants being monitored; VOC, variants of concern per US CDC October 2021. Each cell represents data from a duplicate or triplicate experiment, coded as detailed in Methods.