Regulation of human fetal and adult globin genes in mouse erythroleukemia cells

We have examined whether transfected mouse erythroleukaemia (MEL) cells can be used to examine differential expression of human gamma- and beta-globin genes. These cells, which express only their adult globin genes, will transcribe the human adult beta gene but not the fetal gamma genes when they are introduced on an intact human chromosome 11 by cell fusion. However, MEL cells stably transfected with the human A gamma gene attached to one of the active elements (HS2) of the beta-globin locus control region (LCR) readily produce gamma-globin mRNA in amounts equivalent to those seen with a comparable beta gene insert. When both beta and gamma genes are attached to HS2, equal amounts of beta A gamma mRNAs are produced, irrespective of the gene order. Furthermore, when HS2 is inserted into the 5' end of a 40-kb cosmid containing the G gamma A gamma-117 delta beta genes in their normal chromosomal organization (but with the Greek HPFH -117 A gamma gene mutation), it directs expression of readily detectable amounts of G gamma A gamma and beta-globin mRNAs in MEL cells. Therefore, under these circumstances we have observed no competition between beta and gamma genes for expression in MEL cells. These findings suggest that MEL cells are capable of perpetuating regulatory information involved in developmental control when it is provided by an intact chromosome, but are incapable of reconstructing such information on transfected DNA.     

Linkage of genes for adult alpha-globin and embryonic alpha-like globin chains.

In alpha-thalassemia, the genetic locus for the alpha chains of adult hemoglobin is not expressed. We have examined the hemoglobins of a number of individual mouse embryos heterozygous for a particular alpha-thalassemia (Hbath-J) and find no decrease in the proportion of hemoglobins containing the alpha chain as compared to the hemoglobin containing the alpha-like embryonic globin chain. This result suggests that the locus for this embryonic alpha-like chain is inactivated or deleted in these embryos as well. Because a single mutational event inactivated adult and embryonic loci, we conclude that they are probably closely linked to one another on the same chromosome. We also present evidence that an unusual hemoglobin in the blood of these embryos is composed only of an embryonic beta-like chain, and is thus analogous to the hemoglobin H (beta 4 tetramer) of adult alpha-thalassemics.     

Cytosine arabinoside plus hemin treatment of a human erythroid cell line, KMOE, strongly induces embryonic, fetal, and adult beta-like globin genes

A variety of regimens were utilized on KMOE cells to maximally raise globin mRNA levels for the purpose of improving the usefullness of this line for globin gene studies. Steady-state mRNA levels of embryonic (epsilon), fetal (gamma) and adult (beta) globin genes were assayed by the S1-nuclease protection method before and after exposure to inducing compounds. Exposure of KMOE cells to cytosine arabinoside and hemin leads to over 20-fold increases in beta- and gamma-globin mRNA steady-state levels, and an over 60-fold increase in epsilon-globin mRNA level. Exposure to cytosine arabinoside alone induced beta- and epsilon-globin but not gamma-globin gene expression. The alpha-like globin genes (zeta and alpha) were also monitored but found to be poorly expressed and not significantly inducible. The presence of epsilon-globin mRNA and the lack of alpha-globin mRNA distinguishes this line, KMOE-EL, from the KMOE sublines previously described.     

Stochastic expression of fetal hemoglobin in adult erythroid cells.

The expression of fetal hemoglobin, Hb F, in the adult cells is cellularly restricted both in vivo and in culture. Because, in cultures of erythroid progenitors, subclones that express or fail to express Hb F are derived from the same erythroid stem cell, a mechanism must exist whereby Hb F expression segregates in the progeny of erythroid progenitors during their differentiation. We present mathematical analyses of experimental data which suggest that expression of Hb F in human adult erythroid cells occurs on a stochastic basis. We quantified Hb F expression among the subclones of erythroid bursts (clones) in vitro by labeling subclones with fluorescent anti-Hb F antibodies. The observed data were compared with predictions from a stochastic model with the assumption that the expression of Hb F in a subclone occurs with a probability P equal to the frequency of Hb F-expressing subclones in an experiment. There was good fit between the observed and predicted data with respect to: (i) the relative frequencies of monomorphic F+, monomorphic F-, and bimorphic F+/F- bursts, respectively; (ii) the size distributions among F+, F- and F+/F- bursts; and (iii) the proportions of subclones expressing Hb F within bimorphic F+/F- bursts. Given the hypothesis that erythroid progenitors which have an active Hb F program are less differentiated than cells which do not proceed to express Hb F, the stochastic event indicated by our analyses may be the probability that adult erythroid progenitor cells undergo terminal differentiation at an earlier stage than usual.     

High performance liquid chromatography of rat globin chains of fetal liver during the switch from embryonic to adult hemoglobin

A reversed-phase high performance liquid chromatography (HPLC) method for the direct separation of globin chains from cell lysates of peripheral blood and fetal liver from rat fetuses is described. Partial amino acid analysis of the globins eluted from the HPLC columns as well as comparison with the known elution positions of the adult globin chains in carboxymethyl cellulose chromatography indicated that alpha-chains are eluted first, followed by adult beta chains. The last chains to be eluted are considered as embryonic globins because of their absence in adult rats and their rapid disappearance from the liver after the 14th day of gestation. Liver erythroid cells isolated from rat fetuses on day 14 of gestation mainly synthesized alpha-chains and embryonic globin chains, whereas cells prepared from 16-day old fetuses synthesized almost exclusively alpha-chains and adult beta chains. When the fetal rat liver cells were cultured for 20h with erythropoietin there was a significant stimulation in the synthesis of alpha-globins and adult beta chains but not on the synthesis of embryonic globin chains. It is concluded that HPLC can be useful for the study of rat globin chain synthesis during fetal development, because it separates the globin chains in the three groups of globins, namely alpha, beta and embryonic chains which are important in the switch occurring in the liver.     

Transcriptional regulation of fetal to adult hemoglobin switching: new therapeutic opportunities

In humans, embryonic, fetal, and adult hemoglobins are sequentially expressed in developing erythroblasts during ontogeny. For the past 40 years, this process has been the subject of intensive study because of its value to enlighten the biology of developmental gene regulation and because fetal hemoglobin can significantly ameliorate the clinical manifestations of both sickle cell disease and β-thalassemia. Understanding the normal process of loss of fetal globin expression and activation of adult globin expression could potentially lead to new therapeutic approaches for these hemoglobin disorders. Herein, we briefly review the history of the study of hemoglobin switching and then focus on recent discoveries in the field that now make new therapeutic approaches seem feasible in the future. Erythroid-specific knockdown of fetal gene repressors or enforced expression of fetal gene activators may provide clinically applicable approaches for genetic treatment of hemoglobin disorders that would benefit from increased fetal hemoglobin levels.     

Human fetal to adult hemoglobin switching: changes in chromatin structure of the beta-globin gene locus.

We have investigated the chromatin structure of the chromosomal DNA regions containing the human G gamma-, A gamma-, delta-, and beta-globin structural genes in both fetal and adult erythropoietic tissues and in two human erythroleukemia cells lines before and after induction. Our results indicate that DNase I introduces specific cuts into the beta-globin gene cluster in erythroid cells but not in leukocytes. The predominant sites are located at the 5' sides of the G gamma-, A gamma-, delta-, and beta-globin genes, within 200 base pairs of the respective cap sites. Examination of fetal liver cells has revealed the presence of hypersensitive sites at the 5' side of all four genes, whereas analysis of adult bone marrow has revealed the characteristic sites near the delta- and beta-globin genes but no hypersensitive sites at the 5' termini of the G gamma- or A gamma-globin genes. The presence of delta and beta hypersensitive sites in fetal cells suggests that the increment in expression of the delta and beta genes during development most likely involves the modulation of another pathway to gene expression. Using isolated nuclei from HEL and K562 cells, we have found that the G gamma, A gamma, delta, and beta genes are preferentially sensitive [relative to the pro-alpha2(I) collagen gene] to mild digestion with DNase I, whereas these genes are as resistant as collagen genes in cells that do not express globin. These findings are discussed within the context of chromatin structural correlates of hemoglobin switching.     

Bromodeoxyuridine treatment of normal adult erythroid colonies: an in vitro model for reactivation of human fetal globin genes

We report that bromodeoxyuridine (BrdU) addition in semi-solid cultures of normal adult erythroid progenitors causes a sharp rise of gamma-globin gene expression in erythroid colonies. Control studies were carefully carried out to exclude the possibility of toxic effects exerted by the drug in these experimental conditions. In particular, BrdU addition induces a sharp increase in the level of relative gamma-globin synthesis and content in pooled BFU-E-derived colonies: this rise is clearly observed in single bursts of the mature type (largely composed of late erythroblasts) but not in immature ones (essentially comprising early erythroblasts). Furthermore, it is associated with an increase of the G gamma/G gamma + A gamma synthetic ratio from adult up to fetal like values. Reactivation of gamma-synthesis was observed even if BrdU was added to colonies composed essentially of early erythroblasts, ie, when BrdU was added to either bursts at day 10 of culture or late CFU-E-derived clones at day 1. These in vitro observations indicate modulation of gamma-synthesis at the stage of erythroblasts from normal adults. At the molecular level we suggest that BrdU, by replacing thymidine in DNA, may inhibit the switch from a fetal-like biosynthetic program expressed in early erythroblastic differentiation to the adult program expressed in later stages of maturation.     

Distribution of fetal and embryonic hemoglobins in fetal erythroblasts enriched from maternal blood

BACKGROUND AND OBJECTIVES: To determine the distribution of embryonic and fetal hemoglobin chains in fetal erythroblasts isolated from maternal blood in the first trimester of pregnancy and establish the feasibility of using these chains as markers for fetal cell identification. DESIGN AND METHODS: Maternal blood was obtained from 187 singleton pregnancies at 11-14 weeks of gestation immediately before fetal karyotyping by chorionic villus sampling. In all cases included in this study the fetal karyotype was normal. Fetal erythroblasts were isolated using triple density gradient separation and anti-CD71 magnetic cell sorting techniques. The enriched erythroblasts were stained with Kleihauer-Giemsa and with fluorescent antibodies for the zeta (z), epsilon (e) and gamma (g) globin chains. The percentage of fetal cells positive for each stain was calculated. Fluorescent in situ hybridization (FISH) for X and Y chromosomes was also performed. Comparison was made with the percentage of cells with positive Y-signal FISH in pregnancies with male fetuses. RESULTS: The percentage of fetal erythroblasts stained positive was 37% for the z and 95% for both e and g globin chains, as well as the Kleihauer-Giemsa staining. There was a significant association between the Kleihauer-Giemsa stained cells and those stained with e and g globin chains. There was also an association between cells with Y-signals and those stained with e and g globin chains. INTERPRETATION AND CONCLUSIONS: Embryonic hemoglobin chains can be detected in the enriched fetal erythroblasts, with higher percentages of the e rather than the z globin chains. These chains are therefore potentially unique markers to be used in the identification of cells of fetal origin from maternal blood for prenatal diagnosis of genetic and chromosomal abnormalities.     

Definitive-like erythroid cells derived from human embryonic stem cells coexpress high levels of embryonic and fetal globins with little or no adult globin

Human embryonic stem cells are a promising tool to study events associated with the earliest ontogenetic stages of hematopoiesis. We describe the generation of erythroid cells from hES (H1) by subsequent processing of cells present at early and late stages of embryoid body (EB) differentiation. Kinetics of hematopoietic marker emergence suggest that CD45+ hematopoiesis peaks at late D14EB differentiation stages, although low-level CD45- erythroid differentiation can be seen before that stage. By morphologic criteria, hES-derived erythroid cells were of definitive type, but these cells both at mRNA and protein levels coexpressed high levels of embryonic (epsilon) and fetal (gamma) globins, with little or no adult globin (beta). This globin expression pattern was not altered by the presence or absence of fetal bovine serum, vascular endothelial growth factor, Flt3-L, or coculture with OP-9 during erythroid differentiation and was not culture time dependent. The coexpression of both embryonic and fetal globins by definitive-type erythroid cells does not faithfully mimic either yolk sac embryonic or their fetal liver counterparts. Nevertheless, the high frequency of erythroid cells coexpressing embryonic and fetal globin generated from embryonic stem cells can serve as an invaluable tool to further explore molecular mechanisms.     

Developmental switching of messenger RNA expression from the human alpha-globin cluster: fetal/adult pattern of theta-globin gene expression.

The alpha-globin gene cluster contains four functional globin genes, zeta, alpha 2, alpha 1, and theta. The developmental regulation of the embryonic zeta and fetal/adult alpha 2- and alpha 1-globin genes is well characterized at the level of protein synthesis. The developmental pattern of the theta-globin gene is not well characterized due to the inability to detect its encoded protein. Direct analysis of the globin switching at the steady-state messenger RNA (mRNA) level has been hampered by the difficulty in obtaining quantities of embryonic and early fetal mRNA sufficient for analysis. We analyzed the relative levels of the steady-state zeta-, alpha-, and theta-globin mRNAs in yolk sac in 5-, 6-, 7-, and 8-week postconception embryonic liver, and in cord and adult blood reticulocytes. We show that the switch in the alpha-globin gene cluster from the embryonic to fetal/adult pattern of expression begins at 5 to 6 weeks of gestation. Both the theta- and alpha-globin genes show similar patterns of developmental control that are reciprocal to zeta. alpha-globin RNA is barely detectable or undetectable at 5 weeks, and increases in the 6- to 8-week period, while theta-globin mRNA shows a parallel increase at 5 to 8 weeks postconception and is expressed in cord blood and adult reticulocytes. These data show that the theta-globin gene represents a fetal/adult gene, albeit expressed at a low level.