Dual-hit with sequential isoprenaline and PD-1 blockade in A/J Mice to model immune checkpoint inhibitor-induced myocarditis

Background Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment, but can induce severe immune-related adverse effects, including myocarditis. However, the pathogenesis and risk factors of ICI-related myocarditis remain poorly understood and would benefit from improved preclinical pharmacological models. Purpose We hypothesized that ICI-related myocarditis in mice involves a “dual-hit” mechanism with an ischemic index event priming the immune system, followed by loss of tolerance during ICI therapy. Methods We subjected A/J mice to a single isoprenaline injection (160 mg/kg IP) or saline (w1). Three weeks later, both groups received 5 injections of either anti-PD-1 (25 mg/kg IP, q3 days) or saline. Analysis of myocardial injury (plasma troponin I) and systemic inflammation (serum amyloid A, SAA) was performed in 4 groups: isoprenaline+anti-PD-1 (n=11), isoprenaline+saline (n=10), saline+anti-PD-1 (n=11), and saline+saline (n=10) at baseline, 1h after isoprenaline or saline injection, before anti-PD-1 or saline injection, and at w8. Cardiac magnetic resonance-based measurements (7T Brucker Biospec 70/30) of LV systolic function and T2-relaxation times were performed at baseline, prior to anti-PD-1, and at w8. We semi-quantitatively measured inflammation (CD45 positive cell count) and fibrosis (CSA, cross sectional area on Sirius red-stained sections) on 13 randomly selected high-power fields at midventricular level. Results Troponin I levels significantly increased 1h after isoprenaline (p<0.0001 vs saline) with no further rise after anti-PD1 treatment in contrast to SAA, which peaked after anti-PD-1 (p<0.0001 vs saline). Administration of isoprenaline, anti-PD1 or combination did not affect global LV systolic function as evidenced by stable EDVi, ESVi or LVEF measurements after 8w. However, at 8w mean T2 relaxation times, reflecting myocardial edema, were significantly higher compared to baseline in mice exposed to isoprenaline+anti-PD-1 (p<0.0001 vs saline). Consistent with SAA levels, cardiac CD45-positive leukocyte cell count was significantly higher in mice receiving anti-PD1 alone (p<0.01 vs saline) and in double-hit mice (p<0.0001 vs saline) in whom prior isoprenaline-induced injury was associated with a further rise in CD45+ cells [58.5 (30.4 – 104.7) cells/mm² vs 36.5 (27.2 – 41.8) median (IQR) in anti-PD-1 alone]. Consequently, mice in the double hit group had significant more fibrosis (7.8±4.7%, mean±SD) compared to mice who received saline (1.5±0.5%, p<0.0001) or either challenge alone (anti-PD1 1.1±0.4%, p<0.0001 or isoprenaline 3.2±1.9%, <0.001). Conclusion We successfully characterized a dual-hit mouse model for ICI-related myocarditis based on an initial ischemic myocardial injury and sequential anti-PD1 administration. Our data highlight the intricate interplay between ischemia and PD1-mediated immune activation and will facilitate future studies on targeted interventions for ICI-myocarditis.