T cell costimulation blockade blunts pressure overload-induced heart failure

Abstract

Heart failure (HF) is a leading cause of mortality. Inflammation is implicated in HF, yet clinical trials targeting pro-inflammatory cytokines in HF were unsuccessful, possibly due to redundant functions of individual cytokines. Searching for better cardiac inflammation targets, here we link T cells with HF development in a mouse model of pathological cardiac hypertrophy and in human HF patients. T cell costimulation blockade, through FDA-approved rheumatoid arthritis drug abatacept, leads to highly significant delay in progression and decreased severity of cardiac dysfunction in the mouse HF model. The therapeutic effect occurs via inhibition of activation and cardiac infiltration of T cells and macrophages, leading to reduced cardiomyocyte death. Abatacept treatment also induces production of anti-inflammatory cytokine interleukin-10 (IL-10). IL-10-deficient mice are refractive to treatment, while protection could be rescued by transfer of IL-10-sufficient B cells. These results suggest that T cell costimulation blockade might be therapeutically exploited to treat HF.

Figure 1. The inflammatory signature in hypertrophic left ventricle of mice.

Gene expression analysis (TaqMan real-time qPCR) of mediators of inflammation within the left ventricle of C57BL6/J mice. Relative mRNA expression in sham-operated control mice (white bars) and TAC-operated mice (black bars) at 1 and 4 weeks after surgery, internally normalized to 18 s ribosomal RNA expression. Tnfa, Il6, Tgfb1, Ccl2, Ccl4, Ccl5, Cxcl10, Cxcl11 and the innate cell marker Itgam (CD11b) were significantly increased in the TAC group compared with sham, 1 week after TAC. Four weeks after the operation, Il4 and the T cell marker Cd3e were significantly increased. Values are mean±s.e.m. (n=7–9). Two-way analysis of variance (ANOVA), Bonferroni post-test: *P value<0.05; **P value<0.01; ***P value<0.001.

Figure 2. T cells in the ailing left ventricle.

(a) Representative immunohistochemical (IHC) staining of left ventricles for CD3e (brown) in sham/TAC mice at 4 weeks. Original magnification 10 × ; bars=200 μm. (b) Summary of CD3e IHC. Mean±s.e.m. (n=6). Unpaired t-test. (c) Staining for CD3e (brown) in TAC-operated mice, 1 week post-operation. Original magnification 10 × ; bar=200 μm. (d) Representative fluorescence-activated cell sorting (FACS) analysis of CD3e⁺ cells from cardiac single cell suspension of TAC-operated mice 1 week post-operation. (e) FACS analysis of mediastinal (heart-draining) lymph nodes, inguinal lymph nodes and spleens 2 days post-operation. Mean fluorescence intensities of CD25 on CD3e⁺ cells. Mean±s.e.m.; sham (white bars), TAC (black bars) (n=4). Unpaired t-test. (f) Representative Azan's trichrome collagen staining (blue) of cardiac biopsies from healthy ventricle tissues (n=3), patients with severe dilated cardiomyopathy (DCM) due to mutation in lamin A/C, before placement of a left ventricular assist device (HF LVAD 1M) (n=4), and patients with more severe DCM due to mutation in lamin A/C and mutation in titin, before placement of a LVAD (HF LVAD 2M) (n=2) patients. Original magnification, 20 × ; bar=100 μm. (g) Statistical analysis of collagen deposition in ten identical regions of interest (ROIs), applied to all samples. Mean±s.e.m. Fisher's exact test for presence versus absence of fibrosis. Amount of collagen was also positively associated with disease severity (one-way analysis of variance (ANOVA); post-test for linear trend: P<0.001). (h) Representative staining for CD3e (brown) on the same samples as f. Bar=100 μm. (i) Statistical analysis of CD3e IHC analysis. Mean±s.e.m. One-way ANOVA with Dunn's post-test. (j) Statistical analysis of collagen deposition in cardiac biopsies from healthy ventricle tissues (n=3) and patients with HF from aortic stenosis (n=2) stained as in f. Mean±s.e.m. Healthy tissues (white bar), HF (black bars). Fisher's exact test for presence versus absence of fibrosis. (k) Statistical analysis of CD3e IHC analysis on the same samples as j. Healthy tissues (white bar), HF (black bars). Values are mean±s.e.m. Mann–Whitney test. For all tests *P value<0.05; **P value<0.01; ***P value<0.001.

Figure 3. Abatacept blunts progression of cardiac dysfunction in pressure-overloaded mice.

Mice underwent TAC or sham operation; 2 days post-operation, the mice were treated with three intraperitoneal injections per week of 200 μg of abatacept or PBS, for 4 weeks. (a) Fractional shortening (%FS), (b) ejection fraction (%EF), (c) left ventricle internal dimension in diastole (LVIDd) and (d) left ventricle internal dimension in systole (LVIDs) in TAC- and sham-operated mice at baseline and at time points 1, 3 and 4 weeks after operation, with and without abatacept administration. Data show the mean %FS, %EF, LVIDd and LVIDs for each experimental group at all time-points±s.e.m. (n=7–9). Two-way analysis of variance (ANOVA) with Bonferroni post-test: P values shown in the panel. Abatacept ameliorates pressure overload-induced cardiac fibrosis in mice. (e) Representative macroscopic images of the heart of untreated, PBS-treated and abatacept-injected mice 4 weeks post-sham- or TAC (scale bar=2mm). (f) Cardiac sections of untreated, PBS-treated or abatacept-treated, TAC- or sham-operated mice, at 4 weeks post-operation were stained with Azan's trichrome (n=2). Five identical regions of interest (ROIs) were applied to all samples. The collagen staining intensity was quantified by image acquisition software; plot points indicate the % of collagen pixels in each ROI. Red bars indicate the mean % collagen in each experimental group. ROIs with a collagen signal higher than zero were considered fibrotic. Fisher's exact tests for the presence or absence of fibrosis were applied to sham versus TAC-operated groups for each treatment category. The dotted red line separates fibrotic from non-fibrotic ROIs. *P value<0.05. (g,h) Mice underwent TAC, 2 weeks post-operation, the mice were treated with three intraperitoneal injections per week of 200 μg of abatacept or PBS, for 2 weeks. (g) Fractional shortening (%FS) and (h) ejection fraction (%EF) were measured at baseline and at 2 and 4 weeks after operation. Data show mean of %FS and %EF for each experimental group at all time-points±s.e.m. (n=7). Two-way ANOVA with Bonferroni post-test: ***P value<0.001.

Figure 4. Abatacept administration suppresses the immune response in TAC-operated mice.

(a) Mediastinal (heart-draining), inguinal lymph nodes and spleens were collected 1 week after TAC or sham-operation, stained and analysed by flow cytometry. Percentage of CD25⁺ out of CD3e⁺ cells are plotted as mean±s.e.m.; sham (white bars), TAC abatacept (grey bars) and TAC PBS (black bars) (n=3). One-way analysis of variance (ANOVA) with Tukey's post-test: *P value<0.05; **P value<0.01, ***P value<0.001. (b) Statistical analysis of immunohistochemical staining of left ventricles for the T cell marker CD3e in TAC mice at 4 weeks post-operation, treated with abatacept or PBS, and representative images of the staining (brown colouration; original magnification 40 × ; scale bar=50 μm). Number of CD3e⁺ cells is plotted as mean±s.e.m.; TAC abatacept (white bars); TAC PBS (black bars). Unpaired t-test; *P value<0.05 (n=2). (c) Statistical analysis of immunohistochemical staining of left ventricles for the macrophage marker AIF-1 in TAC mice at 1 week post-operation, treated with abatacept or PBS, and representative images of the staining (brown colouration; original magnification 20 × ; scale bar=100 μm). AIF-1 density plotted as mean±s.e.m.; TAC abatacept (white bars); TAC PBS (black bars). Unpaired t-test; **P value<0.01 (n=2). (d,e) Cardiac single cell suspensions of TAC operated mice, 1 week after the operation, were stained and analysed by flow cytometry. Percentage of F4-80⁺ Ly6C⁺ out of CD11b⁺ CD45⁺ live cells (d) and F4-80⁺ Ly6C^- out of CD11b⁺ CD45⁺ live cells (e) are plotted as mean±s.e.m.; TAC abatacept (black circles); TAC PBS (black squares). Unpaired t-test; *P value<0.05; **P value<0.01 (n=4, 3). (f) Gene expression analysis (TaqMan real-time qPCR) of the left ventricle of C57BL6/J mice, 1 week after TAC or sham operation, with abatacept or PBS treatment. Bars show relative mean Il6 and Il10 expression, internally normalized to 18 s ribosomal RNA expression. Values are mean±s.e.m. (n=5, 8). One-way ANOVA, Dunn's post-test: *P value <0.05; n.s., not significant.

Figure 5. Abatacept attenuates HF through the action of IL-10.

(a) Immunohistochemical staining of left ventricles for CD3e in TAC-operated Il10 KO mice treated with abatacept or PBS, 4 weeks post-operation. Mean±s.e.m. (n=2). Unpaired t-test; ns, not significant. Representative staining for CD3e (brown; original magnification 20 × ; scale bar=100 μm). (b–e) Heart functionality is not preserved in Il10 KO TAC-operated mice after abatacept treatment. TAC/sham-operated mice, starting 2 days post-operation, were treated with three intraperitoneal injections per week of abatacept or PBS, for 4 weeks. (b) Fractional shortening (%FS). (c) Ejection fraction (%EF). (d) Left ventricle internal dimension in diastole (LVIDd). (e) Left ventricle internal dimension in systole (LVIDs). Mean±s.e.m. (n=5–9). Two-way analysis of variance (ANOVA) with Bonferroni post-test; open circle, P value<0.05 versus TAC WT abatacept; open four pointed star, P value<0.01 versus TAC WT abatacept; *P value<0.001 versus TAC WT abatacept; ⁺P value<0.05 versus sham not-treated; closed circle, P value<0.01 versus sham not-treated; ^#P value<0.001 versus sham not-treated; ^§P value<0.01 versus TAC WT PBS. (f) Abatacept treatment in the presence but not absence of IL-10 reduces cardiomyocyte apoptosis in TAC-operated mice. TUNEL assay staining in slides for cardiomyocyte apoptosis on hearts of treated mice 4 weeks post-TAC, in wild-type and Il10 KO mice. Mean±s.e.m. of TUNEL-positive cells (n=2); white bars, abatacept-treated TAC-operated mice; black bars, PBS-treated TAC-operated mice. Two-way ANOVA with Bonferroni post-test; *P value<0.05. (g,h) Wild-type B cell but not T cell transfer in Il10 KO TAC-operated mice restores abatacept therapeutic effects. Il10 KO mice received wild-type T or B cells. Subsequently, they underwent TAC or sham operation and then treated with abatacept as in b–e. (g) %FS and (h) %EF at baseline and 1 week after operation. Mean %FS and %EF for each experimental group at all time-points±s.e.m. (n=3–7). Two-way ANOVA with Bonferroni post-test, *statistics for Il10 KO TAC abatacept; +WT B cells; #statistics for WT TAC abatacept; §statistics for sham not treated.

Figure 6. Abatacept blunts cardiac dysfunction by suppressing the immune response.

Schematic cartoon of the mechanism of action of abatacept in heart failure. In pathological hypertrophy, T cells are activated (through their TCR) and receive costimulation via CD28 from CD80/CD86-expressing antigen presenting cells (macrophages, B cells, dendritic cells). The full activation of T cells, identified by high levels of CD25, enhances the chronicity of the cardiac inflammatory response. This also involves the proinflammatory action of cardiac macrophages. As a result, there is increased cardiomyocyte apoptosis, fibrosis and reduced heart functionality. During abatacept treatment, the drug blocks CD80/CD86-mediated costimulation by macrophages and B cells, leading to inhibition of T cell activation, proliferation and/or infiltration. The effects on macrophages (which may be both direct and indirect) lead to lower maturation and infiltration. Direct effects on B cells lead to production of anti-inflammatory cytokine IL-10, which may also be produced to a lesser extent by T cells. As a consequence of the effect on T cells, B cells and macrophages, the progression of cardiac pathology is blocked, even if the drug is administered at a late stage. The protective effect is dependent on IL-10 presence.