1 European Molecular Biology Laboratory, Postfach 10.2209,6900 Heidelberg, FRG
2 Present address: University of New York at Stonybrook, Dept. of Microbiology, Life Sciences Bldg., Stonybrook, NY 11790

A. Introduction

An emerging, important characteristic of many leukemic cell types is their altered dependence on and/or response to hematopoietic growth factors (6, 17]. In mammals, many of these growth-regulatory proteins have been purified and the respective genes molecularly cloned (16, 26], but the mechanism by which they regulate growth and differentiation of normal hematopoietic precursors is still poorly understood. This is due partly to the fact that such hematopoietic precursors do not self-renew in vitro and constitute only a minor fraction of bone marrow cells, precluding their purification in large numbers (25]. One particularly successful approach to circumventing this problem was the use of avian retroviral oncogenes that transform hematopoietic precursors (10]. For instance, avian retroviruses containing tyrosine kinase oncogenes such as verbB, v-sea, or v-src, as well as the v-Ha-ras oncogene, readily transform avian erythroid progenitor cells (late BFU-E (burst-forming unit erythroid) to early CFU-E (colonyforming unit erythroid)] from chick bone marrow (8, 11]. By transformation, these precursors are induced to self-renew and thus to grow into mass cultures ofimmature, precursor-like cells. The transformed cells, however, retain the ability to undergo terminal differentiation at low frequency (3,7 , 12]. At the same time, they become independent of an activity present in anemic chicken serum that induces CFU-E-like colonies in chicken bone marrow and probably represents avian erythropoietin (EPO) (1,21]. Recently, our laboratory described the use of erythroblasts transformed with temperature-sensitive mutants of v-erbB (ts AEV) and v-sea (ts S13) containing retroviruses as novel systems for studying differentiation of normal and leukemic erythroid cells. Upon a shift to the nonpermissive temperature, ts-AEV and ts-SI3 erythroblasts are induced to differentiate synchronously into erythrocytes. At the same time, the cells regain their dependence on a factor(s) from anemic chicken serum (1-3] (H. Beug et al., unpublished work). Since little information was available on the nature of this avian erythropoietin-like factor(s) and its possible relationship to mammalian erythropoietin (5, 20], we were interested in using ts-oncogene-transformed erythroblasts to attempt its purification and characterization. Here we describe two simple assay systems for avian erythroid growth factors and their use in partially purifying and characterizing chicken erythropoietin, which is shown to be a glycoprotein of 38 kd, resembling mammalian EPO in many respects.

D. Conclusions

In this paper we have shown that ts-oncogene-transformed erythroleukemic cells can be successfully used to assay, purify, and characterize avian erythroid growth factors. Although these erythroid leukemic cells appear to be completely growth-factor independent at the permissive temperature, when the oncogene is fully active, they become dependent for survival, growth, and differentiation on specific erythroid growth factors as soon as the oncogene product is temperature inactivated. Our studies also clearly show that chicken EPO does not induce or modulate the erythroid differentiation program, but rather controls the cell's ability to undergo a series of preprogrammed differentiation events. Since no growth factor detectable in the two assay systems described or in CFU -E assays is secreted by the leukemic cells [2, 3], the oncogene seems to induce factor independence by an intrinsic mechanism, for instance by producing a constitutive signal normally generated by the EPO receptor after ligand binding [4].

Although it appears difficult at present to achieve a complete purification of chicken EPO, due to the limited availability of anemic chicken serum, purification ofother erythroid growth factors secreted by cell lines is currently underway, using the assay systems described. In the long run, we hope to identify the respective growth factor receptors in order to be able to study how tyrosine kinase oncogenes such as v-erbB or v-sea can bypass signaling pathways employed by erythroid-specific growth factor receptors, with the outcome of fatal leukemia.