Glycogen Synthase Kinase 3 Inactivation Drives T-bet-Mediated Downregulation of Co-receptor PD-1 to Enhance CD8(+) Cytolytic T Cell Responses

Abstract

Despite the importance of the co-receptor PD-1 in T cell immunity, the upstream signaling pathway that regulates PD-1 expression has not been defined. Glycogen synthase kinase 3 (GSK-3, isoforms α and β) is a serine-threonine kinase implicated in cellular processes. Here, we identified GSK-3 as a key upstream kinase that regulated PD-1 expression in CD8(+) T cells. GSK-3 siRNA downregulation, or inhibition by small molecules, blocked PD-1 expression, resulting in increased CD8(+) cytotoxic T lymphocyte (CTL) function. Mechanistically, GSK-3 inactivation increased Tbx21 transcription, promoting enhanced T-bet expression and subsequent suppression of Pdcd1 (encodes PD-1) transcription in CD8(+) CTLs. Injection of GSK-3 inhibitors in mice increased in vivo CD8(+) OT-I CTL function and the clearance of murine gamma-herpesvirus 68 and lymphocytic choriomeningitis clone 13 and reversed T cell exhaustion. Our findings identify GSK-3 as a regulator of PD-1 expression and demonstrate the applicability of GSK-3 inhibitors in the modulation of PD-1 in immunotherapy.

Graphical abstract

Figure 1

GSK-3 Inactivation via siRNA KD or Small Molecule Inhibitors Specifically Downregulated PD-1 Expression with Enhanced CTL Function (A) i: siRNA against GSK3 increases the cytolytic killing of OT-I CD8⁺ T cells. Scrambled siRNA (gray bars), siRNA against GSK3 (blue bars) (n = 5). ii: Flow cytometry profiles of receptor expression (gray line, isotype control; green line, scrambled siRNA; red line, GSK3 siRNA). iii: Anti-GSK3 blotting of cell lysates; FITC-tagged siRNA uptake. iv: Scatchard plot analysis. (B) i: OT-I T cells were incubated without (gray bars) or with SB415286 (blue bars) (n = 5). ii: Flow cytometry profiles (as above except red line: SB415286). iii: Scatchard plot analysis. (C) Flow cytometry profiles. Gray line, isotype control; green line, untreated cells; red line, cells incubated with SB415286. (D) i: Cytolytic assay using GSK-3α siRNA, scrambled control siRNA (gray bars), or siRNA against GSK3α (blue bars) (n = 5). ii: GSK-3α-specific blot. iii: Flow cytometry profiles of GSK-3α, GSK-3β, and PD-1. (E) i: Cytolytic assays using other GSK-3 inhibitors. ii: Flow cytometry profiles of PD-1 in the presence or absence of the specific inhibitor (gray line, isotype control; green line, scrambled siRNA; red line, inhibitor). Error bars based on triplicate values in individual experiments; data shown representative of five independent experiments.

Figure 2

GSK-3 Potentiates OT-I Cytolytic Killing of EL4-OVA Target Cells via the Downregulation of PD-1 (A) Percent target killing of EL4-OVA targets by OT-I CD8⁺ CTLs expressing scrambled or GSK-3 siRNA in the presence or absence of blocking anti-PD-1 or PDL1-Fc. (B) Percent killing of EL4-OVA targets by OT-I CD8⁺ CTLs incubated in the presence or absence of SB415286 with or without blocking anti-PD-1 or PDL1-Fc. OVA alone: light gray bars; anti-PD-1: dark blue bars; PD-L1-Fc: light blue bars (n = 4). (C) Relative Pdcd1 expression in the presence and absence of PD-1 siRNA and/or SB415286. (D) Percent target killing from conditions in (C). Error bars based on triplicate values in individual experiments; data shown representative of ≥4 independent experiments.

Figure 3

GSK-3 KD or Drug Inactivation Inhibits PD-1 and Increases Tbx21 Transcription and Binding to the Pdcd1 (A–C) Tbx21 and Pdcd1 transcription in response to anti-CD3 activation with or without SB415286 (A) (n = 4), siRNA GSK-3α/β KD in OT-I cells in response to OVA (B) (n = 3), and in OT-I T cells responding to OVA peptide with and without SB415286 (C) (n = 3). (D) ChIP using anti-T-bet antibodies 2-H10 and 4B10 followed by PCR analysis of Pdcd1 promoter. Mouse T cells (left); Jurkat T cells (right). (E) Tbx21-driven Ifng CNS-12 promoter activity in response to CD3 with or without SB415286. (F) Effect of T-bet siRNA on Pdcd1 and Tbx21 transcription (top). Flow cytometry of PD-1 expression on T cells expressing scambled siRNA or T-bet siRNA in the presence or absence of SB415286 (bottom). (G) CTL killing efficiency of T-bet siRNA-expressing CTLs in the presence of anti-PD-L1 blockade and/or SB415286 and on expression of GZMB and Lamp1 expression in T-bet siRNA-expressing OT-I cells (n = 3). (H) Killing efficiency of OT-I CTLs exposed to SB415286 followed by pPD-1 expression and flow cytometric profile of restored PD-1 expression (n = 3). Error bars based on triplicate values in individual experiments; data shown representative of ≥3 independent experiments.

Figure 4

GSK-3 Inactivation In Vivo Suppresses PD-1 and Increases T-bet Expression Concurrent with Enhanced OT-I CTL Function SB415286 administered in vivo on days 0 or 7 and cells purified ex vivo on days 7 or 14. Ex vivo purified T cells were then assessed for cytolytic activity (percent target killing) and qPCR for Pdcd1 expression on days 7 (A) and 14 (B). Mean and SD of six mice per group.

Figure 5

In Vivo GSK-3 Inhibition Increased Acute MHV-68 Viral Clearance via PD-1 Downregulation BALB/c mice were intra-nasally infected with MHV-68 with/without an i.p. injection of SB415286. (A) Histogram showing maximal radiance values on days 3, 5, and 7 (n = 20). (B) Histogram showing viral titers at days 7 and 14 (n = 5). (C) qPCR values of Pdcd1 and Tbx21. Gray bars, non-treated; blue bars, SB415286 treated. Each column represents a different mouse. (D) CTL killing of EL4-M2 cells (percent target killing). T cells isolated from spleen were assessed for ex vivo killing of EL4-M2 cells (n = 3). (E and F) BALB/c mice were infected with MHV-68 and treated with anti-PD1 with and without SB415286 undergoing the same regime as in (A) (n = 3). Radiance values on days 3, 5, and 7 (E); viral titers on days 7 and 14 (F). (G) Flow cytometry profile of Lamp1, GZMB, and IFN-γ expression (n = 3). (H) Flow cytometry profiles of CD44, CD62L, and CD4 expression. Gray line, isotype control; green line, T cells stimulated with OVA peptide alone; red line, T cells stimulated with OVA peptide in the presence of SB415286. (I) Absolute numbers of M2 peptide-specific CD8⁺ T cells. Mean and SD of at least five mice per group. Also see Figures S5A and S5B.

Figure 6

GSK-3 Inhibition In Vivo Can Enhance Viral Clearance in Chronic Infection by LCMV-Cl13 Mice were infected i.v. with LCMV-Cl13 and treated with/without SB415286. (A) Flow cytometry depicting PD-1 expression on D^bGP_33-41 CD8⁺ T cells in LCMV-Cl13-infected mice treated with and without SB415286. Right plot represents MFI values for PD-1 expression D^bGP_33-41 CD8⁺ T cells at 23–40 days after infection (n = 2). Blue box, DMSO-treated LCMV-Cl13-infected mice; red box, SB415286-treated LCMV-Cl13-infected mice; white box, LCMV-Arm control. (B) Inverse relationship between PD-1 and T-bet in mice treated with SB415286 (boxes as in A; each box represents a different mouse). (C) Viral load in serum of LCMV-Cl13-infected mice treated with or without SB415286. pfu/ml viral titers in serum (×10⁵) days 23, 30, and 37. (D) Percent infectivity at days 30 and 37 compared to day 23. (E) Number of D^bGP_33-41 CD8 T cells taken from the spleens of mice, treated with or without SB415286 over 30–40 days. (F and G) Number of CD4⁺ T cells (CD11a⁺CD49d⁺CD4⁺) (F) and total number of CD4⁺ cells (G) from mice infected with LCMV-Cl13 and treated with or without SB415286. (H) Flow cytometry profiles of IFN-γ- and TNF-α-expressing T cells from mice infected with LCMV-Arm control (day 37) or LCMV-Cl13 (day 25, 30, 37). Histograms (linear scale and log scale) showing the numbers of D^bGP_33-41 CD8⁺ T cells expressing IFN-γ and TNF-α after GP_33-41 peptide stimulation of PBMCs ex vivo from untreated and SB415286-treated mice. Blue boxes, DMSO-treated mice; red boxes, SB415286 treated; white boxes, LCMV-Arm control. Each box represents a different mouse. Data represent two combined independent experiments of n ≥ 4 mice per group. Statistics for (B) were done by 2-way ANOVA, (C) and (E) by non-parametric Mann-Whitney at each time point. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001.