Harvard Medical School, Boston Masschusetts, Institute of T echnilogy , Cambridge
1 The Divison of Hematology-Oncology of the Department of Medicine, Children's Hospital Medical Center, the Sidney Farber Cancer Center
and the Department of Pediatrics, Harvard Medical School, Boston, Mass. 02115.
2 The Department of Medicine, Beth Israel Hospital and Harvard Medical School, Boston, Mass. 02115.
3 The Department of Biology and the Center for Cancer Research, Massachusetts Institute of Technology, Cambridge, Mass. 02139.
Abbreviations used : Hb: hemoglobin; mRNA: messenger RNA. RNase: ribonuclease. cDNA: DNA copy of globin mRNA synthesized
by RNA dependent DNA polymerase (reverse transcriptase) of avian myeloblastosis virus.

I. Introduction

The study of erythroid cell differentiation is pertinent to the study of human leukemia from two points of view. First, since a common stem cell gives rise to both erythroid and myeloid cells, it is not unexpected that disorders of myeloid cell proliferation and differentiation should be occasionally associated with abnormalities of erythroid cell differentiation. In fact in many cases of human leukemia, there are abnormalities of erythroid cells. Secondly, the study of erythroid cell differentiation can serve as an experimental model for the study of normal and abnormal gene expression, a topic of vital importance to the understanding of the etiology and pathogenesis of human leukemia. The erythoid cell provides a number of advantages as a model system for the study of the control of gene expression. This highly specialized cell devotes approximately 95 0/0 of its protein synthesis to the production of one protein, hemoglobin, and therefore only a limited number of the cell's genes are expressed. In addition, a number of biochemical techniques are currently available for the isolation, characterization and quantitation of globin messenger RNA (mRNA), the necessary intermediary between globin gene expression and globin chain synthesis. Erythroid cell differentiation can be considered from two points of view: 1) differences between fetal and adult mature red blood cells; and 2) differences between erythroid cells at different stages of morphologic maturation. We will discuss first the abnormalities of red cell differentiation, mainly the emergence of fetal erythropoiesis, which can occur during the course of various human leukemias. Then we will discuss experimental studies on the quantitation of heme synthesis, globin synthesis and globin messenger RNA content in murine erythroid cells at different stages of maturation. IV. Summary and Conclusions We have reviewed erythroid cell differentiation from two points of view: 1) differences between fetal and adult human red cells with particular reference to alterations which can occur in the normal pattern of erythroid cell development during the course of leukemia; 2) beochemical events which occur during erythroid cell maturation, as a model system for the study of the control of gene ex presslOn. During the course of many leukemias there is the synthesis of red cells containing fetal hemoglobin. In most cases this phenomenon is limited to a small population or clone of red cells and probably represents a nonspecific response of the bone marrow to a hematologic stress. However, in juvenile chronic myeloid leukemia and, in rare cases of erythroleukemia, there is a major reversion to fetal erythropoiesis, with progressive increase in fetal hemoglobin levels and synthesis of red cells which contain not only fetal hemoglobin but have a true fetal pattern of protein synthesis affecting proteins other than Hb F, namely Hb A2, carbonic anhydrase and the membrane antigens i and I. In this case, the fetal erythropoiesis may be a more specific manifestation of the leukemic process and may be related to the phenomenon of fetal protein synthesis (alfa-fetoprotein of carcinoembryonic antigen) observed in other types of neoplasia. Further information on the etiology and pathogenesis of abnormal cell proliferation and differentiation in the leukemias can be obtained by the study of experimental systems permitting the investigation of the regulation of gene expression in differentiating mammalian cells. Maturing erythroid cells provide a promsing system for such investigations for many reasons: differentiating erythroid cells can be obtained relatively free of other cell types; a large amount of a well characterized product, hemoglobin, is synthesized; techniques are now available that permit isolation of erythroid precursors at different stages of differentiation ( 5-8) ; and finally, highly sensitive methods of measuring globin mRNA levels by DNA-RNA hybridization are currently available (13, 26, 27). We have used such techniques to measure levels of globin mRNA in separated populations of murine erythroid cells at different stages of maturation. These studies demonstrated a correlation between globin mRNA content and degree of morphological maturation. In the least well differentiated cells, however, there appeared to be a disproportionate amount of mRNA for the level of hemoglobin synthesis in these cells. These results suggest the presence of some translational control of globin mRNA in the early stages of erythroid development, although the major control of globin gene expression in this system seems to be at the transcriptional level. Finally, when the immature erythraid cells were cultured in the presence of erythropoietin, de novo synthesis of ³H-uridine labeled globin mRNA was demonstrated by the specific RNA-cDNA hybridization assay. These results clearly demonstrate the utility of this model system and these techniques for the study of the interaction between a specific gene and the factors which regulate or modulate its expression.