TIGIT acts as an immune checkpoint upon inhibition of PD1 signaling in autoimmune diabetes

Abstract

Introduction: Chronic activation of self-reactive T cells with beta cell antigens results in the upregulation of immune checkpoint molecules that keep self-reactive T cells under control and delay beta cell destruction in autoimmune diabetes. Inhibiting PD1/PD-L1 signaling results in autoimmune diabetes in mice and humans with pre-existing autoimmunity against beta cells. However, it is not known if other immune checkpoint molecules, such as TIGIT, can also negatively regulate self-reactive T cells. TIGIT negatively regulates the CD226 costimulatory pathway, T-cell receptor (TCR) signaling, and hence T-cell function.

Methods: The phenotype and function of TIGIT expressing islet infiltrating T cells was studied in non-obese diabetic (NOD) mice using flow cytometry and single cell RNA sequencing. To determine if TIGIT restrains self-reactive T cells, we used a TIGIT blocking antibody alone or in combination with anti-PDL1 antibody.

Results: We show that TIGIT is highly expressed on activated islet infiltrating T cells in NOD mice. We identified a subset of stem-like memory CD8+ T cells expressing multiple immune checkpoints including TIGIT, PD1 and the transcription factor EOMES, which is linked to dysfunctional CD8+ T cells. A known ligand for TIGIT, CD155 was expressed on beta cells and islet infiltrating dendritic cells. However, despite TIGIT and its ligand being expressed, islet infiltrating PD1+TIGIT+CD8+ T cells were functional. Inhibiting TIGIT in NOD mice did not result in exacerbated autoimmune diabetes while inhibiting PD1-PDL1 resulted in rapid autoimmune diabetes, indicating that TIGIT does not restrain islet infiltrating T cells in autoimmune diabetes to the same degree as PD1. Partial inhibition of PD1-PDL1 in combination with TIGIT inhibition resulted in rapid diabetes in NOD mice.

Discussion: These results suggest that TIGIT and PD1 act in synergy as immune checkpoints when PD1 signaling is partially impaired. Beta cell specific stem-like memory T cells retain their functionality despite expressing multiple immune checkpoints and TIGIT is below PD1 in the hierarchy of immune checkpoints in autoimmune diabetes.

Keywords: CD8+ T cell; NOD mouse; TIGIT; immune checkpoint; type 1 diabetes.

Figure 1

TIGIT and CD226 expression on islet-infiltrating T cells. (A–C) The frequency of PD1 and TIGIT expressing islet-infiltrating T cells (n= 7–9 mice/group). Data show CD4+ T cells (B) and CD8+ T cells (C) in islet CD45+ cells. Representative data (A) and pooled data showing mean ± SD and individual mice from two to three independent experiments (B, C). ***p<0.001 and *p<0.05 unpaired Student’s t-test. (D–F) The frequency of IGRP_206-214-specific CD8+ T cells expressing PD1 and TIGIT in islets (D, E) and spleen (D, F). Representative data (D) and pooled data show mean ± SD and individual mice from two independent experiments (n=8 mice/group) for islets and (n=4–6 mice/group) for spleen. ****p<0.0001 unpaired Student’s t-test. (G, H) Frequency of SLAMF6 and TIGIT-expressing PD1+CD8+ T cells from islets (n=4 mice/group). Representative data (G) and pooled data (H) showing mean ± SD of the frequency of cells in each quadrant in panel (G), with individual mice shown. **p<0.01 and ***p<0.001 one-way ANOVA Tukey’s multiple comparison test. (I, J) Frequency of CD226 and TIGIT-expressing PD1+CD8+ T cells from islets (n=4 mice/group). Representative data (I) and pooled data showing mean ± SD and individual mice (J). **p<0.01, ns=not significant one-way ANOVA Tukey’s multiple comparison test. All mice in Figure 1 were 16–20-week-old female NOD mice.

Figure 2

Tigit, Cd226, and Eomes expression in islet-infiltrating stem-like memory CD8+ T cells using single-cell RNA sequencing. (A) Uniform Manifold Approximation and Projection (UMAP) plot of islet-infiltrating PD1+CD44+ CD8+ T cells from 14–16-week-old NOD mice showing clusters of stem-like memory (Tscm1 and Tscm2), effector, terminally differentiated, and transitioning T cells. (B) Feature plot showing expression of indicated genes in various subsets of T cells. (C) Heatmap showing differentially expressed genes in Tscm1, Tscm2, terminally differentiated, and effector T cells. (D) The proportions of cells expressing the indicated genes in Tscm1, Tscm2, and terminally differentiated T cells. (E, F) Frequency of CD226 and SLAMF6 expressing PD1+CD8+ T cells from islets of 14–18-week-old female NOD mice. (E) Representative plot and (F) pooled data showing mean ± SD of individual mice (n=8), ***p<0.001 and ****p<0.0001, one-way ANOVA Tukey’s multiple comparison test. (G) Heatmap showing gMFI expression of cell surface markers using flow cytometry for the identification of T-cell subsets in the islets.

Figure 3

EOMES expression on islet-infiltrating T cells. (A, B) Frequency of EOMES+ and TIGIT+ cells among PD1+CD8+ T cells from islets. Representative plot (A) and pooled data showing mean ± SD and individual mice (B). ***p<0.001, unpaired Student’s t-test. (C, D) Frequency of EOMES+ and CD226+ cells among PD1+CD8+ T cells from islets. Representative plot (C) and pooled data showing mean ± SD and individual mice (D). One-way ANOVA Tukey’s multiple comparison test. (E, F) Frequency of SLAMF6 and EOMES-expressing cells among PD1+CD8+ T cells from islets, representative plot (E) and pooled data showing mean ± SD and individual mice (F). One-way ANOVA Tukey’s multiple comparison test. (G, H) Frequency of IFN-γ producing CD8+ T cells from islets. Representative plots (G) and pooled data showing mean ± SD and individual mice (H). Unpaired student’s t-test. (I, J) Frequency of EOMES-expressing IFN-γ producing cells among PD1+CD8+ T cells from islets. Representative plot (I) and pooled data showing mean ± SD and individual mice (J). Unpaired Student’s t-test. (K, L) Frequency of CD226 expressing IFN-γ producing cells among PD1+CD8+ T cells from islets. Representative plots (K) and mean ± SD and individual mice (L). Unpaired Student’s t-test. Female NOD mice 14–18 weeks old (n=8 mice/group) were used. p-values **p<0.01 and ****p<0.0001, ns=not significant. (M, N) Frequency of Ki-67+ cells among PD1+ cells gated on CD226+TIGIT− or CD226-TIGIT+ subsets. Representative plots (M) and mean ± SD and individual mice (N). Unpaired Student’s t-test. Female NOD mice 14–18 weeks old (n=4 mice/group) were used.

Figure 4

CD155 expression on beta cells and islet-infiltrating immune cells. (A, B) CD155 expression on B cells and CD11c+CD11b+ cells. Representative plots (A) and pooled data showing mean ± SD and individual mice of CD155 mean fluorescence intensity (MFI) (n=4 mice). Unpaired Student’s t-test. (C, D) CD155 expression on beta cells of young (4–6 weeks) or old (14–16 weeks) NOD mice. Representative plots (C) and pooled data showing mean ± SD and individual mice of CD155 MFI (D) (n=3 mice/group). Beta cells were identified as CD45− cells with high autofluorescence. Unpaired Student’s t-test. (E, F) Frequency of CD45+ cells in the islets of young and old NOD mice. Representative plots (E) and pooled data showing mean ± SD and individual mice (n=3 mice/group). Unpaired Student’s t-test. **p<0.01, ****p<0.0001, ns = not significant

Figure 5

The role of TIGIT as an immune checkpoint in autoimmune diabetes. (A) Schematic of the treatment strategy for anti-TIGIT and anti-PDL1 antibodies. (B) Diabetes induction by anti-TIGIT antibody given on days 1, 3, 5, and 7, or anti-PDL1 antibody given on days 1, 3, and 5 to 12–13-week-old female NOD mice (n=6 mice/group). Statistical analysis was done by log rank (Mantel–Cox test). (C) Schematic of the treatment strategy to evaluate the impact of anti-TIGIT and anti-PDL1 on islet-infiltrating T cells. (D) The total number of CD8+PD1+ T cells and (E) the frequency of Ki-67 expressing cells among CD8+ T cells from the islets of untreated (UT), anti-TIGIT, and anti-PDL1 treated 12–14-week-old female NOD mice. Representative plots (E) and pooled data showing mean ± SD and individual mice (F) (n=4–5 mice/group). One-way ANOVA Tukey’s multiple comparison test. (G, H) The frequency of PD1+ and SLAMF6+ among PD1+ CD8+ T cells from islets after anti-TIGIT or anti-PDL1 treatment of NOD mice. Representative plots (G) and pooled data showing mean ± SD and individual mice (H) (4–5 mice/group). One-way ANOVA Tukey’s multiple comparison test. (I, J) The frequency of SLAMF6-Ki67+ cells among PD1+CD8+ T cells from the islets after anti-TIGIT or anti-PDL1 treatment. Representative plots (I) and pooled data showing mean ± SD and individual mice (J) (n=4-7 mice/group). p-values *p<0.05,**p<0.01,***p<0.001, and ****p<0.0001, ns=not significant.

Figure 6

TIGIT acts as an immune checkpoint in the absence of PD1 signaling. (A, B) Frequency of CD226 and TIGIT-expressing PD1+CD8+ T cells from the islets of NOD mice after two doses of anti-PDL1 treatment. Representative plots (A) and pooled data showing mean ± SD and individual mice (n=3–5 mice/group). Unpaired Student’s t-test. (C) Schematic of the treatment strategy for combining anti-TIGIT and anti-PDL1 antibodies. (D) Diabetes incidence after one dose of anti-PDL1 alone or in combination with anti-TIGIT as detailed in the schematic in (C). Statistical analysis was done by log rank (Mantel–Cox test). p-values **p<0.01. (E) Schematic of effect on diabetes as a result of full or weak PD1 blockade or weak PD1 blockade together with TIGIT blockade.