Renal and placental secretion of erythropoietin during anemia or hypoxia in the ovine fetus

OBJECTIVE: The source of the erythropoietin (EPO) that circulates in the fetus is unknown although it is known that EPO does not cross the placenta and that fetal kidneys, liver, and placenta express the EPO gene. This study tested to what extent in vivo EPO secretion by the fetal kidneys and placenta can be demonstrated under normoxic and hypoxic conditions. STUDY DESIGN: Renal arterial and venous EPO concentrations were determined in eight late-gestation chronically catheterized fetal sheep made progressively anemic by exchange transfusion with saline solution over 5 to 8 days. In a separate additional series of experiments, umbilical arterial and venous EPO concentrations were determined in nine normoxic fetuses and in nine fetuses subjected to 12 hours of hypoxia induced by lowering maternal-inspired oxygen content. Organ secretion rates were calculated as the product of plasma flow rate and the arteriovenous concentration differences. RESULTS: Renal vein plasma EPO concentration was higher than the arterial concentration in 36 of 40 paired samples (P < .0001) by 16.3% ± 2.7% (mean ± SE). This difference was concentration independent over a range of 12 to 4100 mU/mL. Renal EPO secretion rates were variable and averaged 155 ± 105 mU/min when hematocrit was 31.3% ± 1.6% (n = 5) and 1124 ± 300 mU/min postexchange transfusion when hematocrit was 15.6% ± 0.8% (n = 12). In contrast, umbilical venous and arterial EPO concentrations (range 9-35 mU/mL), although highly correlated (r = 0.94), were not different during normoxia (P_O2 = 21.6 ± 0.5 mm Hg, n = 9). Under hypoxic conditions (P_O2 = 15.6 ± 0.4 mm Hg, n = 9), umbilical vein EPO concentration (range 151-1245 mU/mL) was higher than arterial concentration (range 140-951 mU/mL) in eight of nine paired samples by 13.6% ± 3.3% (P < .01). Under these conditions, estimated umbilical EPO secretion rate was 27,900 ± 11,500 mU/min. CONCLUSION: Under nonanemic, normoxic basal conditions, the kidneys secreted EPO into the fetal circulation, whereas secretion by the placenta was not demonstrated. In the phlebotomy-induced fetal anemia experiments, the kidney demonstrated marked, progressive increases in the rate of EPO production. Similarly, in the fetal hypoxemia experiments, the placenta demonstrated progressive increases - albeit an order of magnitude greater than the kidneys - in EPO production rate. As an extension of these findings, we speculate that the hypoproliferative neonatal anemia that invariably occurs in the early weeks after birth is in part the result of loss of EPO production by the placenta.