An altered cytotoxic program of CD8+ T-cells in HIV-infected patients despite HAART-induced viral suppression

Abstract

Despite the suppression of viral replication induced by the highly active anti-retroviral therapy (HAART), an increased immune activation and inflammatory state persists in HIV-infected patients, contributing to lower treatment response and immune reconstitution, and development of non-AIDS conditions. The chronic activation and inflammation affect the functionality and differentiation of CD8+ T-cells, particularly reducing their cytotoxic capacity, which is critical in the control of HIV replication. Although previous studies have shown that HAART induce a partial immune reconstitution, its effect on CD8+ T-cells cytotoxic function, as well as its relationship with the inflammatory state, is yet to be defined. Here, we characterized the functional profile of polyclonal and HIV-specific CD8+ T cells, based on the expression of cell activation and differentiation markers, in individuals chronically infected with HIV, under HAART. Compared with seronegative controls, CD8+ T-cells from patients on HAART exhibited a low degranulation capacity (surface expression of CD107a), with consequent low secreted levels and high intracellular expression of granzyme B and perforin. This degranulation defect was particularly observed in those cells expressing the activation marker HLA-DR, which were further characterized as effector memory cells with high expression of CD57. The expression of CD107a, but not of granzyme B and perforin, in CD8+ T-cells from HIV-infected patients on HAART reached levels similar to those in seronegative controls when the treatment duration was higher than 25 months. In addition, the expression of CD107a was negatively correlated with the expression of exhaustion markers on CD8+ T-cells and the plasma inflammatory molecule sCD14. Thus, despite HAART-induced viral suppression, CD8+ T-cells from HIV-infected patients have an alteration in their cytotoxic program. This defect is associated with the cellular activation, differentiation and exhaustion state, as well as with the inflammation levels, and can be partially recovered with a long and continuous treatment.

Fig 1. Kinetics of activation of circulating CD8⁺ T-cells.

Kinetics of the expression of HLA-DR and CD38 (A), PD-1 (B), Annexin V (C), perforin and/or granzyme B (D), CD107a alone or together with granzyme B or perforin (E) and IFN-γ alone or together with granzyme B or CD107a (F) in CD8⁺ T-cells from seronegative individuals (n = 3) after stimulation with PMA-Ionomycin (at 50 and 500 ng/mL, respectively). In B, the expression of PD-1 in total and HLA-DR/CD38-expressing CD8⁺ T-cells is shown. *P = 0.04 vs 0 hours; &P = 0.04 vs 2 hours; #P = 0.04 vs 6 hours; +P = 0.04 vs 12 hours.

Fig 2. Alteration in the cytotoxic program of CD8⁺ T-cells from HIV-infected patients.

Frequencies of granzyme B⁺ perforin⁺ (A), granzyme B⁺ IFN-γ⁺ (B) and CD107a⁺ (C) CD8⁺ T-cells in seronegative (n = 15) and HIV-infected (n = 30) individuals after polyclonal stimulation. D. Correlation between the frequency of granzyme B⁺ perforin⁺ and CD107a⁺ CD8⁺ T-cells in seronegative and HIV-infected individuals (n = 15 and n = 30, respectively). E. Levels of granzyme B in supernatant of cultured PBMC from seronegative and HIV-infected individuals (n = 7 and n = 15, respectively). The dashed line indicates the assay limit of detection. F. Correlation between the levels of granzyme B in supernatant of cultured PBMC and the frequency of CD107a⁺ CD8⁺ T-cells in seronegative and HIV-infected individuals after polyclonal stimulation (n = 7 and n = 15, respectively). In A-C and E, the P value of the Mann-Whitney test is shown. In D and F, the rho and P value of the Spearman test are shown.

Fig 3. HLA-DR⁺ CD8⁺ T-cells exhibit an effector memory profile.

A. Expression of CCR7, CD28, CD45RA, CD45RO, CD95, CD57, and CTLA-4 in HLA-DR/CD38-expressing CD8⁺ T-cells from a representative seronegative and HIV-infected individual. The numbers indicate the percentage of positive cells for each marker. The summary of the results is shown in the right panels (n = 5 in both groups of individuals). B. Expression of HLA-DR and CD38 in CD107a⁺ CD8⁺ T-cells from a representative seronegative individual after 12 hours treatment with PMA-ionomycin. The summary of the results is shown in the lower panel (n = 15). *P≤0.01 vs all other subsets. C. Frequencies of HLA-DR/CD38-expressing CD8⁺ T-cells that are CD107a⁺ (upper panel), and granzyme B⁺ perforin⁺ (lower panel), in seronegative individuals (n = 15). The P value of the Wilcoxon test is shown. D. Expression of perforin and CD107a in CD57^hi, CD57^lo and CD57^- cells among HLA-DR/CD38-expressing CD8⁺ T-cells from a representative seronegative individual after PMA-ionomycin stimulation, from a total of 5 individuals. NS: Not statistically significant.

Fig 4. Alteration in the cytotoxic program of HLA-DR/CD38-expressing CD8⁺ T-cells from HIV-infected patients.

Frequencies of HLA-DR/CD38-expressing CD8⁺ T-cells that are CD107a⁺ (A), granzyme B⁺ perforin⁺ (B) and granzyme B⁺ IFN-γ⁺ (C) in seronegative (n = 15) and HIV-infected (n = 30) individuals. The P value of the Mann-Whitney test is shown. NS: Not statistically significant. D. Cell ACCENSE t-SNE plots of HLA-DR, CD38, perforin, granzyme B, CD107a and IFN-γ expression in merged stimulated CD8⁺ T-cells from representative seronegative and HIV-infected individuals (n = 4 in each group; M = 60,000 cells). The circles indicate the cells positive for the indicated molecules. E-H. Correlation between the frequency of granzyme B⁺ perforin⁺ or CD107a⁺ cells and the frequency of their respective total HLA-DR⁺ CD38^- (E and F) or HLA-DR^- CD38⁺ (G and H) CD8⁺ T-cells from HIV-infected individuals (n = 30). The rho and P value of the Spearman test are shown.

Fig 5. Cytotoxic program of HIV-specific CD8⁺ T-cells.

A-B. Expression of CD107a alone or together with perforin, granzyme B or IFN-γ, and granzyme B plus IFN-γ in CD8⁺ T-cells treated with medium or a pool of HIV Gag peptides (n = 30). The P value of the Wilcoxon test is shown. C. Levels of granzyme B in supernatant of cultured unstimulated or Gag peptides-stimulated PBMC from HIV-infected individuals (n = 15). The dashed line indicates the assay limit of detection. D. Expression of HLA-DR and CD38 in CD107a⁺ CD8⁺ T-cells from a representative HIV-infected individual, after stimulation with a pool of HIV Gag peptides. The summary of the results is shown in the lower panel (n = 30). E-F. Frequencies of HLA-DR/CD38-expressing CD8⁺ T-cells that are CD107a⁺ (E), and granzyme B⁺ IFN-γ⁺ (F) in HIV-infected individuals (n = 30), after stimulation with a pool of HIV Gag peptides. G. Correlation between the levels of granzyme B in supernatant of cultured PBMC and the frequency of CD107a⁺ CD8⁺ T-cells in HIV-infected individuals (n = 15) after stimulation with a pool of HIV Gag peptides. The rho and P value of the Spearman test are shown. In D, E and F, the P value of the Dunn’s post-hoc test is shown.

Fig 6. Association between the treatment duration, the exhaustion/activation state and inflammation levels with the cytotoxic program of CD8⁺ T-cells from HIV-infected patients.

A. Heat map of the weight of the variables included in the Component 1 of the Principal Component Analysis performed, including the phenotypic and functional characterization of CD8⁺ T-cells, as well as other variables such as the length of the treatment, the CD4:CD8 ratio, and plasma levels of sCD14. B. Spatial distribution of seronegative (n = 15) and HIV-infected individuals, the latter classified in those with <24 months (n = 17) or ≥25 (n = 13) months of therapy, after Principal Component Analysis. C-D. Frequencies of CD107a⁺ (C) and granzyme B⁺ perforin⁺ (D) CD8⁺ T-cells in seronegative (n = 15) and HIV-infected individuals, the latter classified in those with <24 months or ≥25 months of therapy (n = 17 and n = 13, respectively). E. Correlation between the frequency of CD107a⁺ CD8⁺ T-cells after Gag peptides stimulation and the treatment time in months in HIV-infected patients (n = 30). F. Percentage of specific lysis after in vitro cytotoxicity assay in cells from HIV-infected patients with <24 or ≥25 months of therapy (n = 3 in each group, run in duplicate in each case). G. Correlation between the frequency of HLA-DR⁺ CD38⁺ CD8⁺ T-cells that are CD107a⁺ after PMA-ionomycin stimulation and that of HLA-DR⁺ CD38⁺ CD8⁺ T-cells that are PD-1⁺ in resting peripheral blood, in HIV-infected patients (n = 30). H. Correlation between the frequency of HLA-DR⁺ CD38⁺ CD8⁺ T-cells that are CD107a⁺ after Gag peptides stimulation, the frequency of HLA-DR⁺ CD38⁺ CD8⁺ T-cells that are CD107a⁺ after PMA-ionomycin stimulation and the levels of plasma sCD14, in HIV-infected patients (n = 30). In C and D, The P value of the Dunn’s post-hoc test is shown. In E, G and H, the rho and P value of the Spearman test are shown.