Glucocorticoid-induced tumour necrosis factor receptor family-related protein (GITR) drives atherosclerosis in mice and is associated with an unstable plaque phenotype and cerebrovascular events in humans

Abstract

Aims: GITR-a co-stimulatory immune checkpoint protein-is known for both its activating and regulating effects on T-cells. As atherosclerosis bears features of chronic inflammation and autoimmunity, we investigated the relevance of GITR in cardiovascular disease (CVD).

Methods and results: GITR expression was elevated in carotid endarterectomy specimens obtained from patients with cerebrovascular events (n = 100) compared to asymptomatic patients (n = 93) and correlated with parameters of plaque vulnerability, including plaque macrophage, lipid and glycophorin A content, and levels of interleukin (IL)-6, IL-12, and C-C-chemokine ligand 2. Soluble GITR levels were elevated in plasma from subjects with CVD compared to healthy controls. Plaque area in 28-week-old Gitr-/-Apoe-/- mice was reduced, and plaques had a favourable phenotype with less macrophages, a smaller necrotic core and a thicker fibrous cap. GITR deficiency did not affect the lymphoid population. RNA sequencing of Gitr-/-Apoe-/- and Apoe-/- monocytes and macrophages revealed altered pathways of cell migration, activation, and mitochondrial function. Indeed, Gitr-/-Apoe-/- monocytes displayed decreased integrin levels, reduced recruitment to endothelium, and produced less reactive oxygen species. Likewise, GITR-deficient macrophages produced less cytokines and had a reduced migratory capacity.

Conclusion: Our data reveal a novel role for the immune checkpoint GITR in driving myeloid cell recruitment and activation in atherosclerosis, thereby inducing plaque growth and vulnerability. In humans, elevated GITR expression in carotid plaques is associated with a vulnerable plaque phenotype and adverse cerebrovascular events. GITR has the potential to become a novel therapeutic target in atherosclerosis as it reduces myeloid cell recruitment to the arterial wall and impedes atherosclerosis progression.

Keywords: Atherosclerosis; Carotid artery; Co-stimulation; GITR; Monocyte.

Graphical Abstract

Figure 1

GITR is expressed in human carotid artery plaques and associated with cerebrovascular events. Expression of GITR (blue chromogen in A and D, pink in B and C) co-localized with expression of CD3 (A; pink), CD68 (B; blue), CD31 (C, blue), and with α-smooth muscle actin (D; pink). Co-localization is marked by black arrows, GITR expression by white arrows, and CD68/CD31/α-smooth muscle actin expression by arrowheads. Higher levels of soluble GITR (sGITR) were measured in plasma from subjects with cardiovascular disease compared with healthy controls (E). Expression of GITR was higher in endarterectomy plaques from symptomatic than asymptomatic patients (F). Representative immunohistochemical detection of GITR is shown in plaques from asymptomatic (G, n = 100) and symptomatic (G, n = 93) patients. Scale bars represent 50 µm in (A–D) and 100 µm in (F, G)—all insets are 2 mm. Statistical comparisons were performed using the unpaired t-test in (E) and the Mann–Whitney U test in (F).

Figure 2

GITR is associated with a vulnerable plaque phenotype. Table showing Spearman correlations between plaque GITR expression (visualized by immunohistochemistry) and plaque components in human endarterectomy samples (B, n = 193, with the exception of necrotic core, n = 32). Representative histology shown for asymptomatic (B–E, J–M) and symptomatic plaques (F–I, N–Q) stained for GITR (B, F, J, N) Oil Red O (C, G), α-smooth muscle actin (D, H), cleaved collagen (E, I), CD68 (K, O), glycophorin A (L, P), and collagen type III (M, Q) in consecutive sections of endarterectomy plaques. Scale bars represent 1 mm. Statistical comparisons were performed using the Mann–Whitney U test.

Figure 3

GITR expression and plaque phenotype in murine aortic root plaques. Expression of GITR in aortic root plaques (brown chromogen in A, pink in B, blue in C, D) co-localized with expression of CD3 (A; blue), CD68 (B; blue), CD31 (C, pink), and with α-smooth muscle actin (D; pink). Co-localization (dark brown in A, purple in B–D) is marked by black arrows, GITR expression by white arrows, and CD68/CD31/α-smooth muscle actin expression by arrowheads. In Gitr^−/−Apoe^−/− mice aortic root plaque size (E; representative plaques stained by haematoxylin and eosin, n = 10 mice) and CD68⁺ macrophages (F; n = 9 mice) was reduced. Necrotic regions of aortic root plaques were smaller in Gitr^−/−Apoe^−/− mice compared to Apoe^−/− mice (G; Apoe^−/−, n = 16 plaques from six mice, Gitr^−/−Apoe^−/−, n = 20 plaques from eight mice). The vulnerability-index was lower in aortic root plaques of Gitr^−/−Apoe^−/− than in Apoe^−/− mice (H; Apoe^−/−, n = 6, Gitr^−/−Apoe^−/−, n = 8 mice). Scale bars represent 50 µm in (A–D) (with 200 µm in insets), 200 µm in (E, F) and 1 mm in (G). Analyses were performed on female mice and statistical comparisons were performed using the unpaired t-test.

Figure 4

Mitochondrial dysfunction in non-classical monocytes Gitr^−/−Apoe^−/−. Dysregulated canonical pathways in Gitr^−/−Apoe^−/− mice compared to Apoe^−/− mice identified by IPA (A). Genes of the mitochondrial dysfunction pathway shown as a heatmap with each column representing one sample/mouse (B). Top affected functions in Gitr^−/−Apoe^−/− mice acquired via IPA downstream effects analysis are shown in (C). Activation Z-score is calculated by the IPA software and predicts whether a specific function is increased (positive z-score) or decreased (negative z-score) based on the experimental dataset. Production of reactive oxygen species (ROS) in classical (D) and non-classical (E) blood monocytes from Gitr^−/−Apoe^−/− compared to Apoe^−/− mice as measured via flow cytometry. Statistical comparisons were performed using the unpaired t-test in (D, E). n = 3 mice in (A–C), and n = 5 replicates from a total of three mice in (D, E).

Figure 5

Altered leucocyte recruitment in GITR-deficient mice. Adherence to carotid arteries ex vivo was reduced in Gitr^−/−Apoe^−/− leucocytes to both GITR^+/+ and GITR^−/− endothelium (A, Apoe^−/− cells are green and Gitr^−/−Apoe^−/− cells are red, n = 5 mice/10 carotid arteries per group, experiment repeated twice). Flow cytometric analysis showing altered expression of activation markers on classical and non-classical monocytes among circulating cells in the blood, in the spleen and in the bone marrow in Gitr^−/−Apoe^−/− compared to Apoe^−/− mice (B, n = 3 mice). Scale bar represents 100 µm. Statistical comparisons were performed using Student’s unpaired t-test (A) or a multiple t-test (B).

Figure 6

RNA sequencing of bone marrow-derived macrophages show mitochondrial dysfunction and altered cell migration. Dysregulated canonical pathways in Apoe^−/− mice compared to Gitr^−/−Apoe^−/− mice as identified via IPA (A). Top affected functions in DTA-stimulated GITR-deficient bone marrow-derived macrophages acquired via IPA downstream effects analysis are shown in (B). Activation z-score is calculated by the IPA software and predicts whether a specific function is increased (positive z-score) or decreased (negative z-score) based on the experimental dataset. Mitochondrial membrane potential (ΔΨm, C), mass (D), and nitric oxide (NO) production (E) was decreased in Gitr^−/−Apoe^−/− mice compared to Apoe^−/− mice. Statistical comparisons were performed using the unpaired t-test in (C–E). n = 3 mice (A, B). n = 4 replicates in (C, D), and n = 3 replicates in (E) (each from three mice in total).

Take home figure

Our data reveal a novel role for the immune checkpoint GITR in driving myeloid cell recruitment and activation in atherosclerosis, thereby inducing plaque growth and vulnerability in mice. In humans, elevated GITR expression in carotid plaques is associated with a vulnerable plaque phenotype and adverse cerebrovascular events. (This figure was created in part using templates modified from Servier Medical Art (Mountain View, CA, USA, www.servier.com, licensed under a Creative Commons Attribution 3.0 Unported Licence).