We have investigated the question of erythropoietin (Epo) hypersensitivity versus Epo independence as the basis for the endogenous erythroid bursts (EEBs) that develop in cultures without added Epo from hematopoietic cells of polycythemia vera (PV) patients. Using an improved serum-free (SF) medium containing interleukin (IL)-3, but no insulin-like growth factor-1 (IGF-1), and devoid of contaminants that influence erythropoiesis, we compared circulating normal and PV early erythroid progenitors (BFU-E) with respect to their responses in vitro to recombinant human (rHu) Epo. Cultures were seeded with Ficoll- Hypaque density-separated peripheral blood (PB) mononuclear cells (MNCs), and erythroid bursts, together with their component colonies of > or = 50 cells, were scored in situ at 13 to 16 days of culture. The Epo dose-response curve of BFU-E from PV patients was found to be statistically indistinguishable from that of normal subjects. This observation provides compelling evidence against the Epo- hypersensitivity hypothesis. In the complete SF medium minus Epo, the sensitivity of BFU-E to IGF-1 was much greater in PV than in normals, the dose-response curve being shifted to the left by at least 2 orders of magnitude. These data show that the erythroid progenitor cell response in PV is hypersensitive to IGF-1, and independent of Epo. The data also emphasize the importance of truly SF medium conditions for assessment of progenitor cell sensitivities to recombinant growth factors. Depletion of adherent cells totally prevented erythroid burst formation by normal circulating progenitors, but did not prevent the hypersensitive response to IGF-1 of such cells from PV patients. Hence, again unlike its normal counterpart, the progenitor cell response in PV appears to be independent of adherent cell control.

Several culture media for the growth of human circulating erythroid burst-forming units (BFU-E) that have been claimed to be “serum-free” (“SF”) have actually included albumin preparations known to be contaminated with an undefined burst-promoting activity (BPA); a BPA has also been found in the preparations of other “SF” medium components. This has precluded reliable investigation of the growth factor (GF) requirements of these progenitors. Using a defatted, BPA- free bovine serum albumin (BSA) and the recombinant human growth factors (GFs) erythropoietin (rHu Epo), insulinlike growth factor 1 (rHu IGF-1), and interleukin-3 (rHu IL-3), we have developed an improved serum-free (SF) medium for the production of erythroid bursts from normal adult human peripheral blood mononuclear cells (PBMNC), which requires both hemin and retinyl acetate for its optimal performance. In the presence of BSA without IL-3 or Epo, no burst or colony formation was observed. With IL-3 and Epo alone, only a small number of day 14 erythroid colonies was obtained (12 +/- 1/10(5) PBMNC). Addition of hemin (0.1 mmol/L) allowed the direct scoring of day 14 hemoglobinized colonies and increased their number sevenfold (86 +/- 5). Inclusion of retinyl acetate at physiologic concentrations further augmented the number of colonies threefold to fourfold. Under these apparently optimal conditions, we found that IGF-I could entirely replace Epo. However, IGF-I required a 100-fold higher molar concentration than that of Epo to reach maximal stimulation. The combined effect of Epo and IGF-I was found to be less than the sum of their individual effects, suggesting an overlap in the sensitivities of erythroid progenitors to these GFs. The colony-forming efficiencies of erythroid progenitors in the improved SF medium was very high: 700 single, day 14 erythroid colonies/10(5) PB MNC (at 0.25 mmol/L hemin) distributed as 126 clusters (bursts), with a mean of 5.6 component colonies per burst. These findings show that IGF-I has an Epo-like activity that targets circulating early erythroid progenitors or their progeny, providing strong evidence for the existence of an Epo- independent pathway for normal human adult erythropoiesis, possibly operative when Epo levels are low.

Hematopoietic progenitor cells in 2 myeloproliferative disorders, juvenile chronic myelomonocytic leukemia and polycythemia vera, are known to be hypersensitive to cytokines that control normal progenitor cell proliferation, differentiation, and survival in their respective granulocyte/macrophage and erythroid lineages. Because thrombopoietin controls these functions in the normal megakaryocytic lineage, we asked the question: Are megakaryocytic progenitor cells in the myeloproliferative disorder essential thrombocythemia (ET) hypersensitive to thrombopoietin? Peripheral blood mononuclear cells from patients with ET, or secondary (reactive) thrombocytosis (2°T), or healthy volunteers were grown in strictly serum-free agarose culture containing interleukin 3 (IL-3) and all-trans-retinoic acid, with various concentrations of PEG-rHu megakaryocyte growth and development factor (MGDF). The concentration of cytokine at half-maximum colony number served as a measure of progenitor cell sensitivity. Hypersensitivity to PEG-rHu MGDF was found in circulating progenitors from 18 of 20 (90%) informative patients with presumptive diagnosis ET, 1 of 8 (12.5%) 2 °T patients, and none of the 22 healthy volunteers. Median MGDF sensitivity ratio in ET patients was approximately 53 times greater than in the controls. This hypersensitivity, which was also directed to rHu thrombopoietin, was highly specific with respect to cytokine, disease, and cell lineage. We propose that, despite their single pluripotential cell origin, the different clinicopathologic phenotypes in different chronic myeloproliferative disorders are determined by lineage-restricted hypersensitivities of hematopoietic progenitor cells to endogenous cytokines. This work emphasizes the importance of stringent serum-free conditions for revealing true sensitivities to cytokines. The findings also offer a basis for evolving a positive test for ET, a diagnosis now made essentially by exclusion.

Diamond-Blackfan anemia is a congenital disorder of erythropoiesis in humans, characterized by a macrocytic anemia often associated with physical anomalies. Mutations at either the W or Steel loci in the mouse also leads to a severe macrocytic anemia, as well as other developmental abnormalities. The W locus encodes the proto-oncogene c- kit, a member of the receptor tyrosine kinase family, while the Steel locus encodes a potent hematopoietic growth factor that is the ligand for c-kit. Growth of clonogenic marrow erythroid progenitor cells in vitro in the presence of the recombinant hematopoietic growth factors interleukin-3 (IL-3) and Steel was used to characterize this disease at the cellular level. Three patterns of in vitro marrow response to both recombinant IL-3 or Steel were observed among 10 Diamond-Blackfan patients: those that responded quantitatively and qualitatively almost as well as cells from normal marrow, those that responded at an intermediate level, and those that did not respond at all. These results provide evidence for cellular heterogeneity underlying the pathogenesis of this disorder and therefore raise the possibility that there may be more than one underlying molecular basis for the disease. No gross abnormalities in the structure of either the c-kit or Steel loci were observed in these patients. The normal response in culture of the progenitor cells from at least some patients to Steel with or without IL-3 raises the possibility of using this novel growth factor as a therapeutic agent in Diamond-Blackfan anemia.