The feed-back regulation of erythropoietin production in healthy humans.

The proposed oxygen-dependent feed-back loop regulation of EPO production (Fogh 1978, Erslev 1991, LeHir et al 1991, Nielsen 1991) is mainly supported by data from studies in animals and cell cultures. The feed-back loop and its dependence on oxygen was therefore challenged by studies in healthy humans: Exposure of humans to different levels of acute and continued altitude hypobaria provided evidence for an oxygen dependency of the EPO response. This response is consistent with the proposed feed-back loop regulation of EPO production. Exposure to continued altitude hypobaria demonstrated that the decline in human EPO production is initiated before an EPO-induced erythropoiesis is detectable, and that this decline is related to a concomitant decrease in the haemoglobin-oxygen affinity. Contrary to the feed-back loop, this time-relation indicate that the feed-back regulation of EPO production during continued hypobaric hypoxia is exerted primarily through a decrease in the haemoglobin-oxygen affinity, rather than by the effects of an EPO-stimulated erythropoiesis. A laboratory experiment demonstrated that the lowering of PaO2 by short-term breathing of hypoxic gas could trigger human EPO production. Since the possible influence of spontaneous non-hypoxic variations in EPO production was taken into account, the magnitude of the human EPO response to hypoxia was found to be less than formerly observed in humans and rodents. Contrary to former observations in rodents, the triggering of human EPO production was not related to a hypoxic or normoxic increase of the haemoglobin-oxygen affinity, and normocapnia, without changes in pH or haemoglobin-oxygen affinity, sufficed to suppress the human EPO response to hypoxia. Together with the altitude studies, these results have demonstrated that human EPO production is dependent on, and triggered by, oxygen deprivation, and that changes in the haemoglobin-oxygen affinity play a role mainly in the feed-back regulation rather than in the triggering of human EPO production. Increased circulating levels of the proinflammatory cytokine IL-6 was found in healthy humans during four days of altitude exposure as compared with sea level. The other proinflammatory cytokines IL-1 beta, and TNF alpha remained unchanged, and the increased serum IL-6 did not induce production of c-reactive protein. However, a relation was shown between serum IL-6, SaO2, and serum EPO. These observations in humans confirm earlier in vitro studies in that IL-6 may have a co-stimulatory role on EPO production in the presence of hypoxia. Comparable circadian variations in human EPO production were shown in sedentary subjects, athletes, and healthy but hypoxaemic subjects. Human EPO production could not be triggered by one hour of high-intensity exercise, whereas longitudinal changes in exercise showed a trend of relation between human EPO production, serum concentration of free testosterone, and indices of body composition. These results have demonstrated an endogenous, probably hormonal, and oxygen-independent regulation of human EPO production, which is at variance with the oxygen dependent feed-back loop regulation of EPO production. Conclusively, the present investigations have shown that the feed-back loop regulation of EPO production, as described by studies in rodents and cell-culture models, cannot be applied to the physiology of human EPO production without modifications.