REGULATION OF ERYTHROPOIETIN MESSENGER RNA

Abstract

Erythropoietin (Epo), the glycoprotein hormone which regulates red blood cell formation, is chiefly controlled through modulation of its mRNA accumulation. In this thesis, an RNase protection assay was used to study Epo gene expression In vivo and in isolated perfused rat kidneys. The sensitivity of the assay was optimised to detect very low levels of expression, and a system for highly reproducible quantitation of Epo mRNA, over 4 orders of magnitude, was developed and validated. The organ sites of Epo gene expression were clearly established. In unstimulated adult rats, Epo mRNA was found mainly in the kidney, but was also detected in the liver and, for the first time, in the testis, brain and lung. Following severe hyporxia (7.5% oxygen for 4 hours), Epo mRNA levels were highest in the kidneys but hepatic Epo mRNA also increased dramatically to account for 40-50% of the total, a contribution much higher than previous estimates in bilaterally nephrectomised rats. Experiments with the latter model, however, showed that the surgical procedure itself may compromise hepatic Epo mRNA production. In normal rats, hypoxia also induced an increase in Epo mRNA in the testis, brain and spleen, indicating that oxygen-sensing occurs at these sites. The control of Epo gene expression in the liver and kidney in vivo was investigated. Developmental regulation was demonstrated in experiments which showed a 'switch' from predominant hepatic to mainly renal Epo mRNA formation at 3-4 weeks of postnatal life in the rat. At all ages, Epo production was increased by hypoxic stimulation, but the relative magnitude of the response increased with age. Comparison of total Epo mRNA levels with serum Epo concentrations suggested that the translational efficiency of Epo mRNA also increased progressively with age. In adult rats exposed to graded degrees of normobaric hypoxia or haemorrhagic anaemia, differences in the feedback loop regulating renal and hepatic Epo mRNA were observed. With severe hypoxia or anaemia, the liver accounted for >33% of the total Epo mRNA. However, the liver was much less responsive than the kidney to milder stimulation, particularly with normobaric hypoxia where hepatic Epo mRNA only constituted 2% of the total. The reasons for this difference are unclear, but could be related to physiological differences in local oxygenation during hypoxic stimulation, and/or to intrinsically different control mechanisms for gene expression in the 2 organs. Since different transcriptional initiation sites in liver and kidney have been reported in mice bearing a human Epo transgene, hepatic and renal Epo mRNA was examined with respect to the transcriptional start and polyA-addition site. Mapping studies, using RNase protection, of Epo mRNA from the liver and kidney of normal and transgenic mice showed that the same transcriptional start site was utilized in both organs under unstimulated and hypoxic conditions. Similarly, S1 nuclease mapping of the polyA addition site in hypoxic murine liver and kidney revealed usage of a single, identical site in both organs. To determine the role of extrarenal Epo production in conditions where renal synthesis Is compromised, studies were performed in rats rendered uraemic by subtotal nephrectomy. Although normoxic uraemic rats were anaemic (haematocrlt 0.32�04-i vs 0.43�04 in controls), serum Epo concentrations were not significantly different from controls (32�vs 24�mU/ml). In uraemic animals, the Epo mRNA content of the remnant kidney was only -50% of that in control kidneys, and while hepatic Epo mRNA levels were slightly increased, the rise was far less than expected for the degree of anaemia and the combined Epo mRNA content of the liver and remnant kidney was very similar to that in controls. In uraemic rats exposed to severe hypoxia, however, Epo mRNA in both organs increased dramatically: hepatic Epo mRNA increased to levels comparable to that in controls, but production in the remnant kidneys was only -32% of control kidneys. These data suggest that in uraemia, the kidney and liver are less sensitive to hypoxic stimulation, but the reduced hepatic sensitivity can be overcome by severe hypoxia. The mechanisms of renal oxygen sensing were investigated in the isolated perfused rat kidney since this is the best available in vitro renal model. Perfusion conditions were optimised to obtain oxygen-regulated Epo mRNA production of comparable amplitude to normal kidneys in vivo. In this preparation, Epo mRNA production was closely correlated with renal oxygen delivery, but not with renal oxygen consumption. Inhibition of oxidative phosphorylation using a range of doses of 3 separate agents (cyanide, antimycin and oligomycin) which block the process at different sites, did not stimulate Epo mRNA production despite causing graded reductions in renal oxygen consumption. Similarly, administration of 4 compounds which activate the cyclic AMP system in different ways (dibutyrl-cAMP, forskolin, 5'N-ethylcarboxamidoadenosine and salbutamol), did not significantly increase Epo mRNA levels over the wide range of perfusate pAO2 tested. These data strongly suggest that oxygen-regulated renal Epo mRNA accumulation is not mediated through hypoxic compromise of oxidative phosphorylation or through a cAMP-dependent mechanism.