Sugar in erythropoietin: clinical and forensic implications.

Most proteins secreted into the plasma are heavily glycosylated. Complex branched polysaccharides are N-linked to certain asparagine residues and O-linked to certain serine or threonine residues. The carbohydrate content and structure of plasma proteins are important determinants of their half-lives in the circulation. The cleavage of terminal sialic residues generally results in the rapid clearance of the protein by “asialo” receptors in the liver. Erythropoietin (EPO) contains 40% carbohydrate, most of which is attached to 3 Asn sites. In the December 15, 2001 issue of Blood, Skibeli and colleagues (Blood. 2001;98:3626-3634) report on rather striking differences in the carbohydrate structure of endogenous plasma EPO versus that of recombinant human erythropoietin (rhEPO) produced by high-level expression of the human EPO gene in hamster cells. They show that 3 different commercial preparations of rhEPO (alfa, beta, and omega) have a higher molecular weight than endogenous plasma EPO and a high content of fully sialylated tetraantennary glycans versus none in plasma EPO. The analysis of endogenous EPO was done on protein purified from the plasma of 2 patients with aplastic anemia. The differences in carbohydrate structure between endogenous and rhEPO are likely due to the lack of certain monosaccharide transferases in the hamster rather than modification of rhEPO during circulation in vivo. These differences in carbohydrate structure are likely to have important clinical consequences. The development of rhEPO is arguably the most successful therapeutic application of recombinant DNA technology. Few pharmaceutical agents can match rhEPO’s combination of efficacy and safety. However, there is growing concern about case reports of red cell aplasia developing in patients on chronic rhEPO therapy. The neutralizing antibody that is induced by the recombinant product cross-reacts with endogenous EPO, thereby markedly suppressing erythropoiesis. Because the only structural difference lies in the carbohydrate, it is possible that this is the antigenic stimulus responsible for this rare but serious complication. In this report Skibeli and colleagues demonstrate subtle but significant differences in the carbohydrate structures of rhEPO alfa, beta, and omega. It will be of interest to learn whether these correlate with antigenicity and development of erythroid aplasia. Recently Amgen has developed a superglycosylated rhEPO, Aranesp, that has a markedly prolonged half-life in the circulation and therefore can be administered less frequently. In view of the possible relationship between carbohydrate structure and antigenicity, there will be heightened awareness regarding the development of erythroid aplasia in patients treated with this promising new agent. In recent years there have been an alarming number of reports of sudden deaths among athletes who have “doped” themselves with rhEPO in order to improve their performance in competition. Prevention of this illegal practice is thwarted by the difficulty in detecting surreptitious use of rhEPO. Radioimmune and ELISA assays in current clinical use cannot distinguish endogenous EPO from rhEPO. Moreover, because the half-life of rhEPO is approximately 5 hours, the levels of plasma EPO will be normal the day following selfadministration. The measurement of reticulocytes or surface markers of a cohort of young erythrocytes offers a longer window of time in which the abuse of rhEPO can be detected. Alternatively, the development of a sensitive immunologic assay that can distinguish rhEPO from endogenous EPO would also enable the detection of surreptitious drug use. —H. Franklin Bunn Harvard Medical School