Plasma from hypertransfused and normal sheep and experimentally induced anemic and normal goats was fractionated by ultrafiltration. Fractions obtained were assayed for erythropoiesis stimulatory factor (ESF) or erythropoiesis inhibitory factor (EIF) activity (or both) in the posthypoxic polycythemic mouse assay. The most potent sheep plasma-inhibitor fraction was found in the retentate on a membrane with a cutoff at mol wt 50,000. The most potent EIF fraction from anemic goats passed into the ultrafiltrate, but the comparable EIF from normal goats remained in the retentate on a membrane with a cutoff at mol wt 500. The yield of the most potent EIF fraction was higher from anemic goats than was the yield from normal goats. During thin-layer chromatography, EIF extracted from goat plasma fractions had the same mobility as did prostaglandin F2 alpha. Cohn rabbit plasma fraction IV-4 had an inhibitory factor, and a benzene extract of the fraction contained a component that had the same mobility as did prostaglandin F2 alpha. Cohn rabbit plasma fraction V contained a component that potentiated an erythropoietin-generating factor.

Concentrates of urinary erythropoietin, from two patients with paroxysmal nocturnal hemoglobinuria, were electrofractionated exhaustively. Erythropoietin (ESF1) was collected (biweekly for 8 weeks) from the anode section between Amicon membranes with molecular weight (MW) cutoffs at 20,000 to 30,000. An erythropoietin generating factor(s) (EGF)s was collected from the cathode section between Amicon membranes with MW cutoffs at 30,000 to 50,000. A relatively large erythropoiesis stimulating fraction (ESF2) remained in the center section for several more weeks. The electrofractionation process was continued, however, and after 12 weeks all activity moved to the anode (ESF1) and cathode sections (EGF).

A comparison was made of two erythropoietin preparations distributed by the National Heart and Lung Institute with two preparations from our laboratory. One NIH preparation contained a mixture of at least two erythropoiesis regulatory factors, an erythropoietin-generating factor and erythropoietin. A second NIH preparation contained erythropoietin but had no detectable generating factor.

The bioassay for erythropoietin-stimulating factor (ESF) was compared in 3 strains of mice. One strain was made polycythemic by hypertransfusion and the other 2 by hypoxia. An erythropoiesis-generating factor (EGF) measured in the 3 strains generated approximately the same amounts of ESF, determined with log dose-log response curves for the respective techniques, after incubation with normal rabbit serum (NRS). The amount of ESF generated by EGF without prior incubation with NRS was also approximately the same in the 3 strains of mice.

A mixture of an erythropoiesis stimulating factor (ESF) and an ESF-generating factor (EGF), isolated from the urine of a patient with paroxysmal nocturnal hemoglobinuria (PNH), was separated by selective membrane filtration. The optimum pH for activity of the EGF fraction from the urine of the PNH patient was about 7.4, the same as that previously found for the EGF from the urine of a patient with an anemia secondary to multiple myeloma (MM). The urine from the PNH patient contained an ESF that appeared larger than EGF, whereas the urine from the MM patient did not; the urine from both patients, however, contained an ESF that appeared smaller than EGF as previously noted.

Erythropoiesis regulatory factors, isolated from the urines of a male patient with paroxysmal nocturnal hemoglobinuria (PNH) and a female patient with an anemia secondary to multiple myeloma (MM), were treated with an ethanol: acetone solution for chemical characterizations. The described treatment appeared to inactivate reversibly an erythropoiesis generating factor (EGF, molecular weight <50,000 >30,000) and irreversibly an erythropoiesis stimulating factor (ESF2 mol wt >50,000) from the PNH patient. The inactivated EGF appeared to be reactivated in the presence of several adrenocorticosteroids or batyl alcohol. The treatment with ethanol: acetone did not inactivate EGF or ESF1 (mol wt <30,000 >20,000) from the MM patient nor an inhibitor of erythropoiesis (mol wt <30,000 >10,000).

A urine concentrate from a patient with paroxysmal nocturnal hemoglobinuria contained four regulators of erythro-poiesis that could be isolated by an electro-fractionation technique combined with selective membrane permeability. An erythropoie-sis-stimulating factor (ESF) passed toward the anode through a membrane with a cutoff at a molecular weight of 30,000, but was retained by a membrane with a cutoff at 20,000, while another ESF was retained by a membrane with a cutoff at 50,000. An ESF-generating factor passed toward the cathode through a membrane with a cutoff at 50,000 but was retained by a membrane with a cutoff at 30,000.

Two hypoxia techniques for producing polycythemia in 2 strains of mice used in the bioassay for erythropoietin were compared. Erythropoiesis regulatory factors (erf) brought about a larger incorporation of 59Fe into the hemoglobin of strain D-1 mice than of strain AKR mice, then both strains of mice were exposed to the same technique. The uptake of 59Fe into heme was a hyperbolic function of the packed cell volume (pcv).

A urine concentrate of erythropoietin, obtained by chromatography on diethylamino‐ethyl (DEAE)‐cellulose, contained regulators of erythropoiesis that could be fractionated by selective membrane permeability. Two erythropoiesis stimulating factors (ESF) were obtained after exhaustive dialysis, one that diffused through the membrane and one in the retentate. A biochemical study of the ESF in the diffusate showed no appreciable sialic acid or fucose, a trace of hexosamine and a relatively small amount of protein. The remainder of the complex appeared to be an adrenocorticosteroid(s). The protein appeared to be a fragment of a glycoprotein. After exhaustive dialysis the retentate‐ESF had the characteristics of an ESF‐generating factor. The optimum pH for activity was about 7.4. It is suggested that an ESF(s) remains inactive when bound to a glycoprotein such as α 1 ‐acid glycoprotein or the corticosteroid‐binding globulin. A role of an ESF‐generating factor would be to act on the complex to produce a fragment of the glycoprotein steroid complex, thereby activating the hormone.

An ESF-generating factor (EGF), erythropoietin (ESF) and a mixture of both were incubated with dithiothreitol (DTT, a protective reagent for sulfhydryl groups). Ten μM DTT completely prevented the production of 0.26 ± 0.01 IU of ESF by EGF but did not inactivate ESF. DTT was used to indicate the relative amounts of EGF and ESF activity in a mixture of both.

Erythropoietin levels in both serum and urine have been measured in seven patients with various haematologic disorders and three normal male subjects before and after the short term administration of androgen. Increase in erythropoietin was seen in the urine of the normal subjects. Increase was also seen either in the urine or serum of all 7 patients, 5 of whom had already elevated erythropoietin titers.

SUMMARY Non‐neutralizing antiserum, normal rabbit, human and mouse sera potentiate the erythropoietic activity of anaemic human urinary Fraction II + III. This substantiates the concepts introduced by Kuratowska, Lewartowski and Lipinski. Orosomucoid (alpha‐1 acidic‐glycoprotein) similarly potentiates the erythropoietic activity of anaemic human urinary Fraction II + III. The erythropoietic activity of Standard B was not enhanced by the addition of serum proteins. The erythropoietic activity of human urinary concentrates was enhanced by the addition of normal serum proteins and suggestive enhancement of anaemic human serum and plasma was noted.

E rythropoietin has been thought to possess only slight antigenicity (Jacobson, Goldwasser and Gurney, 1959; Lowy, Keighley, Borsook and Graybiel, 1959). Schooley and Garcia (1962a), however, recently reported the production of neutralizing antibodies to human urinary erythropoietin. The present report describes the production of antibodies against sheep plasma erythropoietin and confirms and extends the observations of Schooley and Garcia.