In an attempt to stimulate Hb F synthesis in baboons by means other than erythropoietic stress, we considered the possibility that an agent that inhibits methylation of CpG sequences in DNA may be effective. 5-Azacytidine, a cytosine analogue that cannot be methylated, is such an agent. Animals whose packed red cell volume was maintained at approximately 20% by bleeding were given 10 daily intravenous injections of the drug (6 mg/kg) in 12 days. Hb F levels in these animals started to increase on day 5 of this regimen and peak levels, which were 6-30 times higher than those produced by bleeding alone, occurred 5-7 days after the last dose of the drug. In animals previously identified as genetically "high" or "low" Hb F responders, the maximal

5-Azacytidine (azaC) has previously been shown to raise Hb F levels in the repeatedly phlebotomized baboon (PCV: around 20%). The administration of tetrahydrouridine (THU), an inhibitor of the enzymatic conversion of azaC to 5-azauridine, made it possible to reduce the amount of azaC and also of 2-deoxy-5-azacytidine (d-azaC) by more than 90% and still achieve maximal Hb F elevations. However, the granulocytopenia, usually occurring after 5-azaC, was not altered by the lowering of the dosages in the presence of THU. Thus, the granulocytopenia is not due to 5-azauridine or other catabolic products resulting from deamination. It is also unlikely that it is caused by a direct influence of azaC on RNA since d-azaC also causes granulocytopenia. The persistence of reticulocytosis throughout the treatment with azaC or d-azaC makes it appear likely that the observed increase in Hb F levels to more than 60% of total hemoglobin is not due to a cytotoxic effect on erythropoiesis resulting in a shift of cell populations toward greater immaturity, but to a direct influence of the drug on the regulation of gamma globin chain production.

Maximal fetal hemoglobin (Hb F) elevations in the baboon subsequent to phenyl hydrazine-induced hemolytic anemia, bleeding, bleeding plus hydroxyurea (HU), or cytosine arabinoside were two to three times lower than those achieved with bleeding plus 5-azacytidine (azaC). Because, in the baboon, maximal elevations in F cell numbers occurred with bleeding alone, changes in the levels of Hb F in hemolysates and in Hb F per F cell could be considered to be due to the administered drugs. Erythropoietic toxicity of azaC was minimal, making it unlikely that the marked elevations of Hb F were due to shifts in the population of erythroid progenitors and precursors and more likely that they were related to a biochemical effect of the drug on DNA. The data indicate marked DNA hypomethylation. This was also found to be associated, but to a much lesser extent, with the modest Hb F elevations after bleeding, hemolysis, and treatment with HU. This drug had greater erythroid toxicity than azaC, and it appeared that the Hb F elevations occurred mainly on the rebound from the early cytotoxicity. The explanation of the molecular DNA changes with this drug and in erythropoietic stress alone remains unknown.

In a previous study we have reported that stored erythrocytes have an augmented tendency to agglomeration in 0.25 M sucrose, at pH 7.2. In the present study we determined the role of protein coating of the erythrocyte membrane and the effect of pH in the agglomeration phenomenon. Erythrocytes, which did not agglomerate and had no protein coating when fresh, agglomerated with storage; although no membrane gamma globulin was detectable, their surface contained fibrinogen. When fresh erythrocytes were coated with albumin, gamma globulin, fibrinogen or hemoglobin, they did not agglomerate. The incubation of fresh washed erythrocytes in ACD or PBS at pH 5.5, in the presence or absence of plasma, induced agglomeration, and the cells reacted with antibody to fibrinogen. Fresh cells from patients with immune disorders, including multiple myeloma, immune‐hemolytic anemia, and rheumatoid arthritis, did not agglomerate. Erythrocytes of these patients were stored in the presence and absence of their plasma, at different buffer systems. Agglomeration after storage was independent of the protein level and the pH of the system appeared to be the major factor. Agglomeration also occurred after several weeks of storage at constant pH of 7.1 and absence of proteins. The findings are consistent with the hypothesis that traces of fibrinogen are adsorbed to circulating erythrocytes, but are only accessible to the antibody reagent after metabolic membrane alterations had occurred after storage. These changes can also be induced by lowering the pH. It appears that the interaction of these factors is mainly responsible for the agglomeration phenomenon.