TCF-1 regulates HIV-specific CD8+ T cell expansion capacity

Abstract

Although many HIV cure strategies seek to expand HIV-specific CD8+ T cells to control the virus, all are likely to fail if cellular exhaustion is not prevented. A loss in stem-like memory properties (i.e., the ability to proliferate and generate secondary effector cells) is a key feature of exhaustion; little is known, however, about how these properties are regulated in human virus-specific CD8+ T cells. We found that virus-specific CD8+ T cells from humans and nonhuman primates naturally controlling HIV/SIV infection express more of the transcription factor TCF-1 than noncontrollers. HIV-specific CD8+ T cell TCF-1 expression correlated with memory marker expression and expansion capacity and declined with antigenic stimulation. CRISPR-Cas9 editing of TCF-1 in human primary T cells demonstrated a direct role in regulating expansion capacity. Collectively, these data suggest that TCF-1 contributes to the regulation of the stem-like memory property of secondary expansion capacity of HIV-specific CD8+ T cells, and they provide a rationale for exploring the enhancement of this pathway in T cell-based therapeutic strategies for HIV.

Keywords: AIDS/HIV; Adaptive immunity; Immunology; T cells.

Figure 1. TCF-1 expression is elevated in HIV- and SIV-specific CD8⁺ T cells from controllers.

(A) Frequency of peripheral blood multimer⁺ HIV–specific CD8⁺ T cells. (B) Proliferation of HIV-specific CD8⁺ T cells in response to 6-day in vitro cognate peptide stimulation as measured by dilution of cell-trace violet (CTV). (C) Gating strategy (left: green, multimer⁺ from controller; gray, bulk CD8⁺ T cells) and distribution (right) of effector-memory phenotypes amongst multimer⁺ cells (TN, naive [gate includes, and likely primarily contains, CD95⁺ stem-cell memory cells, TSCM]; TCM, central memory; TTM, transitional memory; TEM, effector memory; TEMRA, effector memory-RA, separated by level of CD27 expression). (D) Gating (top left; TCF-1⁺ population gated based on CD8⁺ TN population from an HIV-uninfected participant, blue), representative flow plots (top right; median [range]), and summary data (bottom) showing TCF-1 expression in multimer⁺ HIV–specific CD8⁺ T cells from viremic (VIR; magenta), ART-suppressed (ARTs; black), and controller (C; green) individuals of all multimer specificities (left) and within the HIV Gag/HLA-A*02:SL9 multimer⁺ population (right). (E) TCF-1 expression in the SIV Gag/Mamu-A*01:CM9 multimer⁺ population from viremic and controller macaques. Phenotypes assessed by flow cytometry. FMO, fluorescence-minus-1 control. Box plots: median ± IQR. The human studies included data from a maximum of n = 13 viremic, 10 ART-suppressed, and 12 controller participants (as indicated in each figure), some with 2 multimer specificities. The macaque studies included n = 6 viremic and 4 controller animals. Linear mixed effects models to account for clustering within participants (A, C, D), Kruskal-Wallis followed by Dunn’s multiple comparison testing (B), Wilcoxon’s rank sum (E) were used.

Figure 2. High TCF-1 expression in HIV-specific CD8⁺ T cells is associated with a memory-like phenotype.

(A) Phenotype of HIV-specific multimer⁺ CD8⁺ T cells from viremic, ART-suppressed, or controller individuals (compared with naive CD8⁺ T cells from an HIV negative individual). (B and C) TCF-1 expression is higher in less differentiated naive and effector-memory subsets of bulk (B) and HIV-specific (C) CD8⁺ T cells. (D) Correlation between the expression of TCF-1 and other phenotypic markers in multimer⁺ CD8⁺ T cells (GranB, Granzyme B; black dots, all participants; green triangles, controllers only). (E) Expression of CD127, Granzyme B, and T-bet within TCF-1⁺ versus TCF-1^– HIV-specific CD8⁺ T cells. These studies included data from a maximum of n = 13 viremic, 10 ART-suppressed, and 12 controller participants (as indicated in each figure), some with 2 multimer specificities. Linear mixed effects models to account for clustering within participants (A and B), Wilcoxon’s signed-rank test (C and E), Spearman’s correlation (D).

Figure 3. HIV-specific CD8⁺ T cells have lower expression of PD-1 but not other coinhibitory receptors.

(A) Expression of PD-1 in multimer⁺ HIV–specific multimer⁺ CD8⁺ T cells (annotated with median and range; MFI, median fluorescence intensity). (B) Correlation between PD-1 and TCF-1 expression in multimer⁺ HIV-specific CD8⁺ T cells (black dots, all participants; green triangles, controllers only). (C) Expression of PD-1 within TCF-1⁺ versus TCF-1^– HIV-specific CD8⁺ T cells. (D) Expression of TIGIT, CD160, and 2B4 in multimer⁺ HIV–specific multimer⁺ CD8⁺ T cells. These studies included data from a maximum of n = 13 viremic, 10 ART-suppressed, and 12 controller participants (as indicated in each figure), some with 2 multimer specificities. Linear mixed effects models to account for clustering within participants (A, D), Wilcoxon’s signed-rank (C), and Spearman’s correlation (B).

Figure 4. TCF-1 expression is negatively correlated with antigen exposure.

(A) TCF-1 expression in HIV-specific CD8⁺ T cells after 6 days of in vitro peptide stimulation in divided versus undivided cells (n = 12 biological replicates). (B) TCF-1 expression in multimer⁺ CD8⁺ T cells (black) in each division peak after 6 days of in vitro peptide stimulation (naive CD8⁺ T cells shown in blue for reference); data are representative of pattern observed in n = 12 biological replicates. (C) TCF-1 expression in multimer⁺ CMV-specific CD8⁺ T cells from individuals with HIV (ART-suppressed [n = 12] versus controller [n = 4]). (D) Negative correlation between TCF-1 expression and plasma HIV viral load (VL; n = 15). (E) Expression of TCF-1 in HIV-specific CD8⁺ T cells in ART-suppressed individuals, depending on the duration of ART (n = 9 individuals). Wilcoxon’s signed-rank (A), Wilcoxon’s rank sum (C), Spearman’s correlation (D), and linear mixed effects models to account for clustering within participants (E).

Figure 5. TCF-1 KO impairs human CD8⁺ T cell proliferation.

(A) Correlation between TCF-1 expression in HIV-specific CD8⁺ T cells and the proportion that divide after 6-day in vitro peptide stimulation. (B) Correlation between TCF-1 expression in CMV-specific CD8⁺ T cells and their expansion after 6-day in vitro peptide stimulation (fold change [FC] in the frequency of CMV-specific CD8⁺ T cells [gated on total CD8⁺ T cells], stimulated versus unstimulated cells; n = 14). (C) CRISPR-Cas9–mediated deletion of TCF7 (Scr, scrambled guide RNA [gRNA]; RNP, ribonucleoprotein; n = 13 biological replicates). (D) TCF-1 protein downregulation in CD8⁺ T cells after TCF7 KO (red) compared with electroporation with Scr gRNA (gray). (E) CD8⁺ T cell phenotype after TCF7 KO. (F) Proportion of divided CD8⁺ T cells (%CTV^lo) after 5-day stimulation with αCD3/CD28. (G) Correlation between the reduction in TCF-1 expression (MFI) and the reduction of CD8⁺ T cell proliferation (%CTV^lo after stimulation). Spearman’s correlation (A, B, and G) and Wilcoxon’s signed-rank (D–F).

Figure 6. TCF-1 overexpression enhances HIV-specific CD8⁺ T cell expansion after peptide stimulation.

(A) Generation of HIV-specific T cell receptor (TCR) T cells using CRISPR-Cas9 knock-in (KI) of HIV-specific TCR and TCF7 (versus truncated Nerve Growth Factor Receptor [tNGFR]); T2, HLA-A*02-expressing T2 cell line; SL9, SLYNTVATL peptide. (B) TCF-1 protein expression after tNGFR (black) or TCF7 (blue) KI (left, representative flow plots; right, summary TCF-1 MFI from n = 6 biological replicates; lines connect control and TCF-1-expressing samples generated from the same donor). (C) Frequency of TCF7 KI TCR T cells after 6-day in vitro stimulation with SL9 peptide loaded on T2 cells (FC versus tNGFR KI; n = 6 biological replicates). (D) CTV tracings of TCR T cells (unstimulated or stimulated). (E) Granzyme B expression in TCF7 KI TCR T cells. Wilcoxon’s signed-rank (B, C, and E).