Preclinical trial of a MAP4K4 inhibitor to reduce infarct size in the pig: does cardioprotection in human stem cell-derived myocytes predict success in large mammals?

Abstract

Reducing infarct size (IS) by interfering with mechanisms for cardiomyocyte death remains an elusive goal. DMX-5804, a selective inhibitor of the stress-activated kinase MAP4K4, suppresses cell death in mouse myocardial infarction (MI), human pluripotent stem cell-derived cardiomyocytes (hPSC-CMs), and 3D human engineered heart tissue, whose fidelity to human biology is hoped to strengthen the route to clinical success. Here, DMX-10001, a soluble, rapidly cleaved pro-drug of DMX-5804, was developed for i.v. testing in large-mammal MI. Following pharmacodynamic studies, a randomized, blinded efficacy study was performed in swine subjected to LAD balloon occlusion (60 min) and reperfusion (24 h). Thirty-six animals were enrolled; 12 were excluded by pre-defined criteria, death before infusion, or technical issues. DMX-10001 was begun 20 min before reperfusion (30 min, 60 mg/kg/h; 23.5 h, 17 mg/kg/h). At all times tested, beginning 30 min after the start of infusion, DMX-5804 concentrations exceeded > fivefold the levels that rescued hPSC-CMs and reduced IS in mice after oral dosing with DMX-5804 itself. No significant reduction occurred in IS or no-reflow corrected for the area at ischemic risk, even though DMX-10001 reduced IS, expressed in grams or % of LV mass, by 27%. In summary, a rapidly cleaved pro-drug of DMX-5804 failed to reduce IS in large-mammal MI, despite exceeding the concentrations for proven success in both mice and hPSC-CMs.

Keywords: Cardioprotection; Drug development; Human pluripotent stem cell-derived cardiomyocytes; Infarct size.

Fig. 1

Predicted and verified levels of the active agent DMX-5804, following infusion of DMX-10001. a Chemical structures, highlighting the phosphate moiety of the pro-drug. b Rat PK after i.v. injection of 5 mg/kg DMX-10001, showing rapid disappearance of the pro-drug and generation of DMX-5804 even at the earliest time-points. Data are the geometric mean ± SD; n = 3. c Predicted levels in pigs from three- and two-step infusion protocols (Left: 60 mg/kg/h, 0.5 h + 20 mg/kg/h, 2.5 h + 16 mg/kg/h, 21 h. Right: 60 mg/kg/h, 0.5 h + 17 mg/kg/h, 23.5 h.) DV, dependent variable (ng/mL); IVAR, independent variable (h); green line, target concentration. d Empirically determined levels of DMX-5804 in the efficacy study, using two-step infusion as in (c). The geometric mean ± SD (red) and individual levels in all 12 treated swine (white or gray) are shown, by comparison with the serial levels (black) and time-averaged concentration in mice (C_av) that reduced infarct size by 50%. Mouse PK results are from ref. [11]. Results are shown relative to the time of first dosing, which differed between the species (pig, 20 min before reperfusion; mouse, 60 min after reperfusion)

Fig. 2

Experimental protocol. Swine underwent 60 min of LAD occlusion, followed by 24 h of reperfusion, and were treated with vehicle versus DMX-10001 using a two-step infusion protocol. Hemodynamics and LV function were measured as shown beneath the timeline. Animals were then euthanized and analyzed for IS, NR, and AAR

Fig. 3

Effects of DMX-10001 infusion on IS and NR. a Data are shown as single data points plus the mean ± SD. Vehicle, n = 12; DMX-10001, n = 12; *P < 0.05; unpaired two-tailed t test. Absolute IS and the IS/LV ratio were reduced by DMX-10001, attributable to an underlying difference in AAR. b Data as in (a), omitting an outlier that was well beyond the standard deviation for DMX-10001-treated animals. The conclusions were identical in both analyses

Fig. 4

Analysis of covariance. The end-points were correlated with AAR, AAR/LV or IS/AAR, as shown, assuming a linear relationship. Each data point represents one animal; vehicle (open circles), n = 12; DMX-10001 (closed squares), n = 11. No significant difference in slope was seen between groups in any of these relationships