The CD38/NAD/SIRTUIN1/EZH2 Axis Mitigates Cytotoxic CD8 T Cell Function and Identifies Patients with SLE Prone to Infections

Abstract

Patients with systemic lupus erythematosus (SLE) suffer frequent infections that account for significant morbidity and mortality. T cell cytotoxic responses are decreased in patients with SLE, yet the responsible molecular events are largely unknown. We find an expanded CD8CD38^high T cell subset in a subgroup of patients with increased rates of infections. CD8CD38^high T cells from healthy subjects and patients with SLE display decreased cytotoxic capacity, degranulation, and expression of granzymes A and B and perforin. The key cytotoxicity-related transcription factors T-bet, RUNX3, and EOMES are decreased in CD8CD38^high T cells. CD38 leads to increased acetylated EZH2 through inhibition of the deacetylase Sirtuin1. Acetylated EZH2 represses RUNX3 expression, whereas inhibition of EZH2 restores CD8 T cell cytotoxic responses. We propose that high levels of CD38 lead to decreased CD8 T cell-mediated cytotoxicity and increased propensity to infections in patients with SLE, a process that can be reversed pharmacologically.

Keywords: CD38; CD8 T cell; EZH2; Sirtuin1; cytotoxicity; infection; nicotinamide adenine dinucleotide; patients; systemic lupus erythematosus.

Figure 1.. Patients with SLE Prone to Infections Display an Increased Percentage of CD8CD38^high

(A) Representative histograms showing CD38 expression in CD4 and CD8 T cells from both a healthy subject and a patient with SLE. (B) Percentage of CD38 in CD4 and CD8 T cells by flow cytometry. The rectangle identifies the cutoff point separating patients with SLE in 2 subpopulations: those with high or normal percentages of CD8CD38^high (healthy subjects = 24, SLE = 42; Welch’s test). (C) MFI of CD38 in CD4 and CD8CD38^high T cells (healthy subjects = 24, SLE = 35: Kolmogorov-Smirnov) from healthy subjects and patients with SLE by flow cytometry. (D) Percentage of CD4CD38^high and CD8CD38^high T cells from patients with SLE sorted based on disease activity: inactive (SLEDAI < 4, n = 26) or active (SLEDAI ≥4, n = 15) by flow cytometry (Kolmogorov-Smirnov test). (E) Sensitivity and specificity of the percentage of CD8CD38^high T cells in defining patients with any infection (SLE = 35, chi-square test). In all figures, average data are represented as mean ± SD.

Figure 2.. CD8CD38^high T Cells Display Poor Cytotoxic Capacity

(A) Representative histograms showing the expression of CD107a on CD8CD38^low and CD8CD38^high T cells from healthy subjects by flow cytometry. To measure CD107a expression, cells were stimulated with plate-coated CD3 and CD28 antibodies with Golgiplug and anti-CD107a for 5 h and stained with proper surface markers. (B) Degranulation (%CD107a) of CD8 T cells from healthy subjects and patients with SLE evaluated by flow cytometry (healthy subjects = 8, SLE = 12; Welch’s test). (C) Correlation between percentage of CD38^high and degranulation in CD8 T cells (healthy subjects = 8, SLE = 12; closed circles, healthy subjects; opened circles, SLE; Pearson’s correlation). (D–F) Degranulation (%CD107a) of CD8 T cells sorted on CD38^low and CD38^high of both healthy subjects and patients with SLE in both groups (D) and separately in healthy subjects (E) and SLE (F) by flow cytometry (healthy subjects = 8, SLE = 12; closed circles, control; opened circles, SLE; healthy subjects = 8, SLE = 12; Welch’s test). (G) Degranulation (%CD107a) of the different subpopulations of CD8 T cells sorted on CD38^low and CD38^high by flow cytometry (n = 5 healthy donors; one-way ANOVA with multiple comparison). (H) Representative dot blots (left) and cumulative data (right) showing percentage of of Annexin V⁺ P815 cells after coincubation with CD8CD38^low or CD8CD38^high T cells from healthy subjects analyzed by flow cytometry (normal T cells = 4; Kolmogorov-Smirnov test). In all figures, average data are represented as mean ± SD.

Figure 3.. CD8CD38^high T Cells Express Low Levels of Cytotoxic Molecules

(A) Representative histograms showing granzyme A (GZMA), granzyme B (GZMB), perforin (PRF1), and interferon gamma (IFN-γ) in CD8CD38^low and CD38^high T cells from a healthy subjects by flow cytometry. (B) Percentage of GZMA, GZMB, perforin and IFN-γ in CD8CD38^low and CD8CD38^high T cells evaluated by flow cytometry (healthy subjects ≥3; paired t test). (C) Percentage of GZMA, GZMB, perforin, and IFN-γ in CD8CD38^low and CD8CD38^high T cells evaluated by flow cytometry (SLE ≥3; paired t test). (D) mRNA levels of GZMA, GZMB, PRF1, and CD38 from CD8CD38^low and CD8CD38^high T cells by quantitative PCR (normal T cells = 10; Kolmogorov-Smirnov test). Cells were sorted by fluorescence-activated cell sorting (FACS) aria. In all figures, average data are represented as mean ± SD.

Figure 4.. CD8CD38^high T Cells Express Decreased Levels of EOMES, RUNX3, and T-bet

(A) Scheme showing the transcription factors regulating various cytotoxic molecules. (B) Representative histograms showing EOMES, RUNX3, and T-bet in CD8CD38^low and CD38^high T cells from a healthy donor by flow cytometry. (C) MFI from EOMES, RUNX3, and T-bet in CD8CD38^low and CD8CD38^high T cells from by flow cytometry (healthy subjects ≥3; paired t test). (D) Percentage of EOMES, RUNX3, and T-bet in CD8CD38^low and CD8CD38^high T cells by flow cytometry (SLE = 6; paired t test). (E) mRNA levels of EOMES, RUNX3, and T-bet (TBX21) in CD8CD38^low and CD8CD38^high T cells by flow cytometry (normal T cells = 10; Kolmogorov-Smirnov test). Cells were sorted by FACS aria. In all figures, average data are represented as mean ± SD.

Figure 5.. CD38 Modulates the Epigenetic Profile of CD8 T Cells by Controlling the Sirtuin1/EZH2 Axis

(A) Schema showing the regulation of transcriptional factors by CD38 through the NAD/Sirtuin1 (SIRT1)/EZH2 axis. ADPR, adenosine ribose; cADPR, cyclic adenosine ribose. (B) NAD⁺ levels from the same number of CD8CD38^low and CD8CD38^high T cells, sorted by FACS aria. NAD⁺ levels were quantified by colorimetric measurements (normal T cells = 2). (C) Representative western blot (left) and cumulative data (right) showing lysine acetylation in total protein from CD8CD38^low and CD8CD38^high T cells sorted by FACS aria (normal T cells = 8; Kolmogorov-Smirnov test). (D) Percentage of EZH2 in CD8CD38^low and CD8CD38^high T cells from healthy subjects and patients with SLE by flow cytometry (healthy subjects = 7, SLE = 5, paired t test in each group). (E and F) Cell lysates from Jurkat CD38^WT and CD38^KO were subjected to immunoprecipitation with an EZH2 antibody followed by western blot analysis with a lysine acetylation or EZH2 antibodies (E). Same experiment as in (E) using cell lysates from Jurkat CD38^KO cells, after overnight treatment with 50 μM of EX527 (SIRT1 inhibitor) (F). (G and H) Representative histogram and cumulative data showing H3K27me3 in CD8CD38^low and CD8CD38^high T cells by flow cytometry (healthy subjects = 4; paired t test) (G). Same experiment using Jurkat CD38^KO cells treated with 50 μM of EX527 overnight (n = 4 in 3 independent experiments, paired t test) (H). In all figures, average data are represented as mean ± SD.

Figure 6.. Inhibition of Sirtuin1/EZH2 Axis Restores CD8 Cytotoxicity

(A) MFI from EOMES, RUNX3, and T-bet in CD8 T cells pretreated with 50 μM of EX527 (SIRT1 inhibitor) overnight by flow cytometry (healthy subjects = 7; paired t test). (B) MFI from EOMES and RUNX3 in TALL104 CD8 T cells pretreated with 1 μM of GSK126 (EZH2 inhibitor) overnight by flow cytometry (n = 4, 3 independent experiments; paired t test). (C) MFI from RUNX3 in CD8 T cells in patients with SLE pretreated with 1 μM of GSK126 overnight by flow cytometry (SLE = 4; paired t test). (D and E) Percentage of IFN-γ and GZMB in CD8CD38^low or CD8CD38^high T cells from healthy controls (D) and from patients with SLE (E) treated with 1 μM of GSK126 overnight by flow cytometry (healthy subjects = 5, shown in closed circle, SLE = 6, shown in opened circle; paired t test). (F and G) Degranulation (%CD107a) of CD8CD38^low or CD8CD38^high T cells from healthy subjects (F) and from patients with SLE (G) treated with GSK126 overnight and stimulated with plate-coated CD3/CD28 antibodies for 5 h with GSK126 at the indicated concentration (n = 4, one-way ANOVA with multiple comparisons).