The erythroleukemia induced by the Friend virus complex in the mouse is a well studied model of multi-step leukemogenesis. The early phase of the disease is characterized by the polyclonal expansion of erythroblasts in the spleen of infected animals. This pre-leukemic phase is due to the constitutive and/or abnormal activation of the EpoR which results from its interaction with the gp55 viral envelope glycoprotein. During the late stage of the disease, leukemic clones emerge from this polyclonal population which are characterized by two recurrent genetic alterations. The first is the inactivation of the wild type function of p53 ; the second is the overexpression of Spi-1/PU.1 by retroviral insertional mutagenesis. Several experimental approaches indicate that each of these alterations is involved in the deregulation of the normal proliferation and differentiation control of erythroblasts.
An heterologous system which uses avian primary erythroblasts is described that allows to study, at a clonal level, the contribution to erythoblastic transformation of the alterations which are specific of the early and late stages of the Friend disease. We show that the expression of Spi-1/PU.1 profoundly modifies the normal response of primary erythroblasts to both Epo stimulation and deprivation and induces their proliferation in response to SCF. The activity of Spi-1/PU.1 was found to depend upon its cooperation with an activated form of the erythropoietin receptor [Epo R (R129C)] which mimics the constitutive or abnormal activation of the EpoR by its interaction with gp55. Although devoid of detectable effect by itself, expression of a mutant p53 was found to reinforce the effects of Spi-1/PU.1 on survival and differentiation of these cells. These observations suggest that evolution of Friend erythroleukemia is a highly integrated process in which the function of Spi-1/PU.1 or events controled by this protein during the late stage of the Friend disease depend upon specific signals emanating from the forms of the EpoR which are activated in the pre-leukemic phase. The availability of this system offers the opportunity to further dissect at the molecular level the exact nature of this cooperation.