Neuronal nitric oxide synthase required for erythropoietin modulation of heart function in mice

Introduction: Erythropoietin (EPO) acts primarily in regulating red blood cell production mediated by high EPO receptor (EPOR) expression in erythroid progenitor cells. EPO activity in non-erythroid tissue is evident in mice with EPOR restricted to erythroid tissues (ΔEPORE) that become obese, glucose-intolerant, and insulin-resistant. In animal models, nitric oxide synthase (NOS) contributes to EPO activities including erythropoiesis, neuroprotection, and cardioprotection against ischemia-reperfusion injury. However, we found that extended EPO treatment to increase hematocrit compromised heart function, while the loss of neuronal NOS (nNOS) was protective against the deleterious activity of EPO to promote heart failure. Methods: Wild-type (WT) mice, ΔEPORE mice, and nNOS-knockout mice (nNOS-/-) were placed on a high-fat diet to match the ΔEPORE obese phenotype and were treated with EPO for 3 weeks. Hematocrit and metabolic response to EPO treatment were monitored. Cardiac function was assessed by echocardiography and ultrasonography. Results: ΔEPORE mice showed a decrease in the left ventricular outflow tract (LVOT) peak velocity, ejection fraction, and fractional shortening, showing that endogenous non-erythroid EPO response is protective for heart function. EPO treatment increased hematocrit in all mice and decreased fat mass in male WT, demonstrating that EPO regulation of fat mass requires non-erythroid EPOR. EPO treatment also compromised heart function in WT mice, and decreased the pulmonary artery peak velocity (PA peak velocity), LVOT peak velocity, ejection fraction, and fractional shortening, but it had minimal effect in further reducing the heart function in ΔEPORE mice, indicating that the adverse effect of EPO on heart function is not related to EPO-stimulated erythropoiesis. ΔEPORE mice had increased expression of heart failure-associated genes, hypertrophic cardiomyopathy-related genes, and sarcomeric genes that were also elevated with EPO treatment in WT mice. Male and female nNOS-/- mice were protected against diet-induced obesity. EPO treatment in nNOS-/- mice increased the hematocrit that tended to be lower than WT mice and decreased the PA peak velocity but did not affect the LVOT peak velocity, ejection fraction, and fractional shortening, suggesting that nNOS is required for the adverse effect of EPO treatment on heart function in WT mice. EPO treatment did not change expression of heart failure-associated gene expression in nNOS-/- mice. Discussion: Endogenous EPO has a protective effect on heart function. With EPO administration, in contrast to the protective effect to the cardiac injury of acute EPO treatment, extended EPO treatment to increase hematocrit in WT mice adversely affected the heart function with a corresponding increase in expression of heart failure-associated genes. This EPO activity was independent of EPO-stimulated erythropoiesis and required EPOR in non-erythroid tissue and nNOS activity, while nNOS-/- mice were protected from the EPO-associated adverse effect on heart function. These data provide evidence that nNOS contributes to the negative impact on the heart function of high-dose EPO treatment for anemia.