A functional glucocorticoid-responsive unit composed of two overlapping inactive receptor-binding sites: evidence for formation of a receptor tetramer.

An unusual glucocorticoid-responsive element (called GRE A) was found to mediate the induction of the cytosolic aspartate aminotransferase gene by glucocorticoids and was bound by the glucocorticoid receptor in a DNase I footprinting assay. GRE A consists of two overlapping GREs, each comprising a conserved half-site and an imperfect half-site. The complete unit was able to confer glucocorticoid inducibility to a heterologous promoter (delta MTV-CAT). Mutation of any of the half-sites, including the imperfect ones, abolished inducibility by the hormone, demonstrating that each of the isolated GREs was inactive. In electrophoretic mobility shift assays, purified rat liver glucocorticoid receptor (GR) formed a low-mobility complex with GRE A, presumably containing a GR tetramer. When purified bacterially expressed DBD was used, low-mobility complexes as well as dimer and monomer complexes were formed. In inactive mutated oligonucleotides, no GR tetramer formation was detected. Modification of the imperfect half-sites in order to increase their affinity for GR gave a DNA sequence that bound a GR tetramer in a highly cooperative manner. This activated unit consisting of two overlapping consensus GREs mediated glucocorticoid induction with a higher efficiency than consensus GRE.

Congenital dyserythropoietic anemia type III. studies on erythroid differentiation of blood erythroid progenitor cells (BFUE) in vitro.

In order to investigate whether the morphological abnormalities observed in congenital dyserythropoietic anemia type III (CDA III) have a cellular or an environmental origin; BFUE from the blood of a patient exhibiting a CDA type III were grown in vitro. The progeny derived from these BFUE were subsequently studied at light and electron microscopic level. Giant multinuclear erythroblasts which represent the most prominent finding of CDA III in bone marrow were also found in culture. Nuclear clefts found in vivo were also observed by electron microscopic studies performed on the erythroblasts growing in vitro. In each erythroid colony, morphologically normal and giant multinuclear erythroblasts were intermingled. This finding indicates that the two populations of erythroblasts derive from the same defective stem cell. The studies by indirect immunofluorescence of i antigen was preferentially expressed in the immature erythroblasts as in culture from normal subjects but not in the giant mature erythroblasts. This finding suggests that the excess of i antigen expression of CDA III in vivo is rather the indirect consequence of a stimulation of erythropoiesis than result of the disease.

Normal human serum contains a factor(s) capable of inhibiting megakaryocyte colony formation.

Growth of megakaryocyte colonies from human bone marrow progenitors has been achieved in plasma clot culture. Megakaryocyte colonies were identified either by cytological examination or by immunofluorescent labelling using a monoclonal antiplatelet antibody (J 15). No significant differences were observed in the quantitation of the colonies by these two methods. In the absence of a stimulating factor, MK colonies were detectable when high cellular concentrations were seeded. Among the numerous conditioned media tested for their ability to stimulate MK colony formation, conditioned medium from leukocytes stimulated by PHA (PHA-LCM) was the most effective. However, under standard conditions of culture corresponding to 20% normal human sera, colony formation was not related linearly to seeding density even when cultures were stimulated by PHA-LCM. Upon reduction of the concentration of serum (2.5-5%) in the culture medium, colony formation displayed a linear relationship seeding density only when PHA-LCM was used as the stimulating factor. At the same time, the size of the colonies increased. Such inhibition was observed with all the human sera tested but it varied in extent from one batch to another. Replacement of serum by albumin, iron-saturated transferrin, alpha-thioglycerol and low density lipoproteins at physiological concentration but in the presence of bovine plasma provided a culture system whose ability to support colony formation equalled that of low concentrations of whole serum; spontaneous MK colony formation still occurred. Our results provide evidence for the presence of an inhibitor(s) of MK colony formation in normal human sera, and demonstrate the role of cellular factors in stimulating MK colony formation.

Clonal T-cell colony formation in agar culture: an attractive assay to test the T-cell depletion from bone marrow.

Current studies suggest that the depletion of T-lymphocytes from donor marrow is an effective method for preventing acute graft-versus-host disease (GVHD) after allogeneic bone marrow transplantation in man. To deplete the T-lymphocytes from bone marrow cells we use either monoclonal anti-T-cell antibodies and complement or T101 ricin A-chain immunotoxin. Residual T-lymphocytes are analyzed by their capacity to form clonal T-cell colonies in the presence of phytohemagglutinin (PHA), accessory cells, and recombinant interleukin 2. The method is compared to immediate indirect immunofluorescence (iF) and thymidine incorporation by marrow cells stimulated by PHA. IF is not suitable for evaluating the depletion by immunotoxin, and the interpretation of thymidine incorporation is generally questionable. The results of the colony formation show that the sensitivity of the colony assay is close to that of iF when T cells are depleted by complement lysis, and the sensitivity of the colony assay is not dependent upon the depletion procedure. Therefore, the T-cell colony assay is a simple functional control for the quality of bone marrow T-cell depletion, especially for T-cell depletion by immunotoxin.

Effects of ethanol, dexamethasone and RU 486 on expression of cytochromes P450 2B, 2E, 3A and glutathione transferase pi in a rat hepatoma cell line (Fao).

The aim of our study was to investigate the suitability of Fao cells, derived from the Reuber H35 rat hepatoma as a tool for studying regulation of drug-metabolizing enzymes and drug metabolism. Fao cells express P450 2B, 2E, 3A and GST pi and were used to study the effects different inducers on these enzymes. Ethanol considerably increased the amounts of P450 2E and, to a lesser extent, P450 2B and GST pi mRNA and protein. Dexamethasone decreased the amounts of P450 2B, 3A and GST pi mRNAs, but had no appreciable effect per se upon the protein concentration of these enzymes. However, it antagonized the induction of P450 2E, 2B and GST pi by ethanol, even at the protein level. RU 486 decreased P450 2B protein and P450 2E mRNA and protein levels without effecting P450 3A and GST pi expression. RU 486 did not antagonize the dexamethasone effects, suggesting that at least some of these effects are not mediated by the glucocorticoid receptor. These data indicate that these cells constitute a suitable tool for studying the regulation of drug-metabolizing enzyme expression and drug metabolism.

Activation by PHA of CD8 lymphocytes into clonal colony forming cells. Role of interleukin-1.

Monoclonal T cell colonies can be grown in agar culture from quiescent T lymphocytes under PHA stimulation, provided that (1) a low number of T lymphocytes (less than or equal to 5 X 10(4)/ml) is seeded, (2) IL-2 is added to the culture, and (3) a high number of accessory B cells (greater than or equal to 5 X 10(5)/ml) is present in contact with the T lymphocytes. Under these culture conditions the colony progenitors can be ascribed to the CD4 subset, whereas CD8 lymphocytes do not generate colonies. This finding is surprising since both CD4 and CD8 lymphocytes may be cloned in liquid culture. We now report the appropriate conditions required to grow cytotoxic CD8 lymphocyte colonies in agar. CD8 colony growth is dependent upon IL-2-IL-2 receptor interaction and is inhibited by anti-IL-2 receptor antibodies. In addition to PHA, accessory B cells and IL-2, an additional signal provided by recombinant IL-1 is necessary for CD8 colony formation. Exogenous IL-1 can be replaced by irradiated CD4 lymphocytes which stimulate the expression of membrane IL-1 activity in the accessory B cells. In addition, colony growth from quiescent but not preactivated CD8 lymphocytes is inhibited by anti-IL-1 antibodies. Altogether, the data show that an IL-1 signal is required for the induction of IL-2 responsive IL-2 receptors on quiescent CD8 colony forming cells.

Glucocorticoid hormones prevent the induction of gamma-glutamyl transpeptidase by ethanol in a rat hepatoma cell line.

The increase in serum gamma-glutamyl transpeptidase (GGT) is a well known marker of chronic alcoholism in man. We have previously shown that ethanol (180 mM) induces GGT activity 2-3-fold in the C2 rat hepatoma cell line. In this study, we have analyzed the interaction of ethanol with steroid hormones and drugs in this well defined cell culture system. Dexamethasone (100 nM), a synthetic glucocorticoid agonist, completely prevented the induction of GGT by ethanol, but had no effect when added alone. This inhibitory effect was also observed with other corticosteroids, but not with sex steroids; it was prevented by RU 486, a glucocorticoid antagonist. These observations suggest that dexamethasone acts through a high affinity glucocorticoid receptor. Conversely, ethanol did not interfere with the glucocorticoid induction of alanine aminotransferase in the same cell. We have analyzed the metabolism of ethanol in the C2 cells. These cells lack significant alcohol dehydrogenase activity as well as any cytochrome P-450 Alc immunoreactivity. Dexamethasone did not modify the disappearance of ethanol in the culture medium of those cells. We conclude that glucocorticoid hormones interact with ethanol at the cellular level, and that this interaction does not involve a modification of alcohol metabolism.

Fenofibrate modifies transaminase gene expression via a peroxisome proliferator activated receptor alpha-dependent pathway.

Fibrates modify the expression of genes implicated in lipoprotein and fatty acid metabolism via the peroxisome proliferator-activated receptor alpha(PPARalpha), leading to reductions in serum triglycerides and cholesterol. The expression of certain genes regulated by PPARalpha have been shown to be modified in a species dependent manner. Aspartate aminotransferase (AspAT or GOT) and alanine aminotransferase (AlaAT or GPT) are enzymes involved in intermediate metabolism in all cells and in hepatic gluconeogenesis. These enzymes are also widely used as serum markers of possible tissue damage. This study investigated whether fenofibrate could modify the expression of liver AspAT and/or AlaAT and thus possibly alter transaminase levels independently of a cytotoxic effect. In human Hep G2 cells, fenofibrate increased cytosolic AspAT (cAspAT) activity by 40% and AlaAT activity by 100%, as well as both mRNAs. Nuclear run on assays showed that this effect was, at least in part, transcriptional. Increases in mRNA were also observed in human hepatocyte cultures at concentrations of the drug attained in patients. In C57BL/6 mice, fenofibrate decreased cAspAT and cAlaAT mRNA, while these effects were abolished in PPARalpha knock-out mice. In conclusion, fenofibrate has been shown to modify cAspAT and AlaAT gene expression in a species and PPARalpha dependent manner. This is the first demonstration that cAspAT and AlaAT activities may be pharmacologically altered, independently of a toxic phenomenon.

Contribution of a nuclear factor 1 binding site to the glucocorticoid regulation of the cytosolic aspartate aminotransferase gene promoter.

Two regions of the cAspAT gene promoter mediate the glucocorticoid regulation of this gene in the Fao hepatoma cell line. The proximal region was localized by deletion studies and stable transfections in the Fao cells to the sequence -553/-398. This region includes the glucocorticoid-responsive element (GRE) A sequence, which consists of two overlapping GREs and which can mediate the glucocorticoid regulation of a heterologous promoter. DNase I footprinting studies have shown that a site 80 base pairs upstream of the GRE A was protected by liver and brain nuclear extracts (site P8). The binding was displaced by an excess of an oligonucleotide containing a typical NF1 binding site and by NF1-specific antibodies. In electrophoretic mobility shift assay using the P8 oligonucleotide as a probe, several complexes were formed. Most complexes were common to liver and brain but were less abundant when testis extracts were used. At least one complex was specific to the liver. All complexes, with the exception of two, were competed for by the NF1 oligonucleotide. Furthermore, the sequence of the P8 site showed a 7/9-base pair homology with a typical NF1 site. A mutation of the P8 site, which prevents the binding of NF1-like proteins to it, considerably decreases the regulation of the cAspAT promoter fragment by glucocorticoids. Surprisingly, the basal activity of the mutant promoter was increased 2-fold. Thus, the regulation of the cAspAT gene promoter is mediated by a regulatory unit comprising the GRE A and a NF1 binding site.

Congenital dyserythropoietic anaemia type I: absence of clonal expression in the nuclear abnormalities of cultured erythroblasts.

Erythroid colonies derived from the circulating early erythroid precursor (BFU-E) of a patient with congenital dyserythropoietic anaemia type I (CDA I) have been grown in plasma clot and studied by electron microscopy. The number of circulating BFU-E was in the normal range with a roughly normal appearance at the light microscopic level. However, investigation of individual colonies by electron microscopy has always shown a mixture of normal and abnormal erythroblasts exhibiting the typical nuclear aberrations found in vivo. The proportion of normal erythroblasts varied from one colony to another. After the release of the cells from the clot in order to permit new cellular interactions, macrophages were observed to phagocytose abnormal erythroblasts but also a few erythroblasts with normal nuclei. These findings demonstrate that CDA I is a disorder which results from a defective erythroid stem cell but that the progeny of each BFU-E may vary considerably in the extent to which they express the morphological defects. Based on studies of cultures of BFU-E, similar conclusions were previously made for CDA II.

Megakaryocyte colony formation from human bone marrow precursors.

We report the growth in plasma clot culture of megakaryocyte colonies from adult bone marrow cells with the use of four different sources of erythropoietin (Ep) as the stimulating factor. A major proportion of the megakaryocyte colonies (75%) were pure, while the others were mixed, involving erythroblasts and megakaryocytes. Ultrastructural studies have shown that the maturation of megakaryocytes was essentially normal and that platelet shedding occurred. Megakaryocyte colony formation required a large number of plated cells (greater than 3 X 10(5)/ml). In the absence of erythropoietin, rare spontaneous megakaryocyte colonies could be observed, while no erythroid colonies were present. However, erythropoietin induced a fivefold increase in the total number of colonies. These data suggest that erythropoietin is involved in the differentiation of human megakaryocytes, but that it does not act alone, since another factor related to the number of seeded cells appears essential for the formation of human megakaryocyte colonies.

Phorbol esters inhibit the glucocorticoid-mediated stimulation of cytosolic aspartate aminotransferase gene transcription.

The regulation of cytosolic aspartate aminotransferase (cAspAT) gene expression by phorbol esters was investigated in the highly differentiated hepatoma cell line Fao. Phorbol 12,13-dibutyrate (PdBu) had no effect on basal activity but partially inhibited the induction of cAspAT by dexamethasone. The extent of inhibition (40%) was similar to that obtained with insulin or vanadate. The inhibitory effects of PdBu and vanadate were additive. In the case of PdBu, the inhibitory effects could be eliminated by first incubating the cells with PdBu, which down-regulates protein kinase C. In contrast, inhibition by insulin was not modified by this treatment. The molecular mechanism of PdBu action was investigated. Northern blot analysis showed that the steady-state mRNA levels of cAspAT were decreased by PdBu in the presence of dexamethasone. In addition, the transcription rate, as measured by run-on experiments, was also decreased under the same conditions. Finally, a 2.4 kb promoter fragment driving the chloramphenicol acetyltransferase gene was stably transfected into the Fao cells. The regulation of the activity of this promoter fragment by dexamethasone and PdBu was similar to the regulation of the endogenous cAspAT activity. We conclude that PdBu acts by regulating the promoter activity of the cAsPAT gene.

Nucleotide sequence and glucocorticoid regulation of the mRNAs for the isoenzymes of rat aspartate aminotransferase.

Cytosolic and mitochondrial aspartate aminotransferase cDNAs were cloned from a lambda gt11 rat liver cDNA library. The complete coding sequence and the 3' non-coding sequence of the cytosolic isozyme mRNA were obtained from two overlapping cDNA clones. Partial sequences of the mitochondrial enzyme cDNAs were found to be identical to the recently published complete sequence (Mattingly, J. R., Jr., Rodriguez-Berrocal, F. J., Gordon, J., Iriarte, A., and Martinez-Carrion, M. (1987) Biochem. Biophys. Res. Commun. 149, 859-865). A single mRNA (2.4 kb (kilobase pair] hybridizing to the mitochondrial cDNA probe was detected by Northern blot analysis, whereas the cytosolic cDNA probe labeled one major (2.1 kb) and two minor (1.8 and 4 kb) mRNAs. The 1.8-kb and the 2.1-kb cytosolic aspartate aminotransferase mRNAs differ in their 3' ends and probably result from the use of either of the two polyadenylation signals present in the 3' noncoding region of the major cytosolic aspartate aminotransferase mRNA. Glucocorticoid hormones increased the activity of cytosolic but not mitochondrial aspartate aminotransferase in both liver and kidney. The increase in the enzyme activity was accompanied by an increase in the amount of the three corresponding mRNAs, while the mitochondrial enzyme mRNA was not significantly modified.

Regulation of cytosolic aspartate aminotransferase mRNAs in the Fao rat hepatoma cell line by dexamethasone, insulin and cyclic AMP.

Glucocorticoid hormones increase the activity of cytosolic aspartate aminotransferase (cAspAT) in the Fao rat hepatoma cell line. Maximal increase (6-10-fold) was observed 48 h following the addition of the glucocorticoid agonist dexamethasone at a concentration of 0.1 microM. The effect of dexamethasone was specific since it was not mimicked by sex steroids and was inhibited by the glucocorticoid antagonist RU 486. Insulin (0.1 microM) inhibited by more than 50% the induction of cAspAT by glucocorticoids. The cAMP analog, 8-bromoadenosine 3',5'-monophosphate (Br8cAMP, 0.5 mM), potentiated the effect of dexamethasone (2-3-fold) and partially relieved the inhibitory effect of insulin on the induction by dexamethasone. Both insulin and Br8-cAMP had no significant effect on basal activity. The mitochondrial isoenzyme was insensitive to the various hormonal treatments. Northern blot analysis revealed the presence of two major (2.1-kb and 1.8-kb) and one minor (4-kb) mRNA species hybridizing with a rat cAspAT probe. The regulation of these mRNAs by glucocorticoids, insulin and cAMP correlated with the variation of the cAspAT activity, suggesting that these hormones act at the pretranslational level. We compared the regulation of cAspAT mRNAs with those of tyrosine aminotransferase mRNA. Both were similarly increased by dexamethasone but the latter was also increased by cAMP even in the absence of the glucocorticoid agonist. In addition, the increase in tyrosine aminotransferase mRNA was inhibited by cycloheximide whereas the increase in cAspAT mRNAs was not. These results show that there are significant differences in the regulation of cAspAT and tyrosine aminotransferase by glucocorticoids and other hormones, although both enzymes probably contribute to the same metabolic pathway.

Insulin down-regulates cytochrome P450 2B and 2E expression at the post-transcriptional level in the rat hepatoma cell line.

Cytochromes P450 (P450s) are inducible drug-metabolizing enzymes involved in the metabolism of numerous endogenous and exogenous substrates. The regulation of some of these enzymes during experimental diabetes has been reported, but the direct involvement of insulin and the mechanism of its action remain unclear. The aim of our work was to study the effects of insulin on P450 2B and 2E expression in differentiated Fao hepatoma cells. Exposure of the cells to 0.1 microM insulin caused 60% and 80% decreases in the steady state levels of P450 2B and 2E proteins, respectively, within 24 hr. Before this, a rapid decrease in the corresponding messages was observed. Indeed, 5-6 hr of insulin treatment produced 80 and 50% decreases in P450 2B and 2E mRNA levels, respectively. Nuclear run-on transcription and mRNA turnover studies were performed to determine the mechanism (transcriptional and/or post-transcriptional) by which insulin modulated these mRNA levels. From our results, it can be concluded that insulin down-regulates the expression of P450 2B by shortening the half-life of its mRNA (half-lives of 6.9 hr without insulin and 3.6 hr with insulin), whereas it down-regulates the expression of P450 2E both by weak repression of the transcription rate (-30%) and, in particular, by acceleration of its mRNA turnover (half-lives of 8.5 hr without insulin and 3.3 hr with insulin).

Phenotype study with monoclonal antibodies of T lymphocyte colonies in normal individuals and in patients with chronic OKT8+ lymphocytic leukaemia.

The lymphocyte colony forming capacity of peripheral blood mononuclear cells from normal controls and from two patients with chronic OKT8+ lymphocytic leukaemia was determined in agar culture under PHA stimulation. The number and size of the colonies in patients were reduced compared to normal. The lymphocytic phenotype of colony cells was studied with monoclonal antibodies in colonies harvested from agar culture and in colonies expanded in liquid culture in the presence of TCGF. This study was performed in individual colonies and in pooled colonies. Colonies from normal controls contained a mixture of the OKT4+ and OKT8+ lymphocyte subsets. In contrast, colonies from the two patients contained essentially OKT4+ lymphocytes. The data indicate that, in the patients, progenitors of the OKT8+ subset are unresponsive to normal proliferative and/or differentiative stimuli under the present culture conditions.

Heterogeneous accessory cell requirement for human peripheral blood T lymphocyte activation by PHA into IL-2-responsive colony-forming cells.

Mitogen-driven T cell proliferation in liquid culture requires accessory cells that cooperate in interleukin 2 production. We have investigated the accessory cell requirement for human lymphocyte colony formation under PHA stimulation. Semisolid medium limits cell-to-cell contact emphasizing the role of cooperating cells both in growth factor production and in triggering events. Culturing at high T cell density demonstrates that accessory cells can be substituted for colony formation by exogenous IL-2. Culturing at low T cell density in the presence of IL-2 also demonstrates that accessory cells are required for activation of a subset of progenitors into IL-2 responsive colony-forming cells. Consequently, T colony progenitors, contained in the E-rosetting cell fraction of peripheral blood, are heterogeneous in their triggering signals: a minor subset is directly inducible by PHA, and a major subset is inducible by PHA in the presence of accessory cells. We found that monocytes and some leukemic B cells support effective accessory function in both colony growth factor production and colony progenitor sensitization.

Cell-specific regulation of cytosolic aspartate aminotransferase by glucocorticoids in the rat kidney.

The basal expression and hormonal regulation of cytosolic aspartate aminotransferase (cAspAT) were investigated in the rat kidney. In adrenalectomized animals, the basal activity was highest in the renal cortex and in the inner stripe of the outer medulla (0.1-0.15 U/mg protein). The glucocorticoid analogue dexamethasone increased cAspAT activity about twofold in the cortex and in the inner stripe of the outer medulla but not in the papilla. A half-maximal increase in the activity was achieved at doses of approximately 5 micrograms/100 g body wt. The mineralocorticoid aldosterone did not modify the cAspAT activity. The cell specificity of the hormonal regulation was analyzed by in situ hybridization. In untreated adrenalectomized rats, a cAspAT cRNA probe labeled mainly the inner stripe of the outer medulla. After dexamethasone or hydrocortisone treatment, labeling was uniformly increased in this part of the medulla and was heterogeneously increased in the renal cortex. The specific increase in labeling within the cortex was shown to be confined to the distal convoluted tubule and the thick ascending limb. We conclude that, in addition to widespread basal expression, cAspAT is regulated by glucocorticoids in a highly cell-specific manner in the renal cortex. The enzyme may thus participate in the increased energy metabolism elicited by these hormones in these cells.

Acinar zonation of the hormonal regulation of cytosolic aspartate aminotransferase in the liver.

The zonation of the expression and regulation of the cytosolic aspartate aminotransferase (cAspAT) mRNAs in the liver acinus was investigated in diabetic and/or adrenalectomized rats. Dexamethasone increased cAspAT activity two- to threefold alone and up to sixfold in combination with streptozotocin-induced diabetes. Northern blot analysis showed that the cAspAT mRNAs were increased by those treatments; the effect of streptozotocin was reversed by the administration of insulin. In situ hybridization experiments showed that basal cAspAT mRNAs were uniformly distributed within the liver acinus. However, cAspAT mRNAs were induced by glucocorticoids specifically in the periportal zone and by streptozotocin in a larger area including the periportal and intermediary zone. The alpha 2u-globulin mRNAs which are specifically expressed in the perivenous hepatocytes are also induced by glucocorticoids in this zone, suggesting that the specific regulation of the cAspAT gene by glucocorticoids in the periportal zone is not due to the absence of functional glucocorticoid receptors in the other zones. We conclude that the regulation of the cAspAT housekeeping gene is zone specific in the liver. Furthermore, this zonation depends on the gene and on the type of hormonal or pharmacological treatment.

Cellular regulation of i and I antigen expressions in human erythroblasts grown in vitro.

In order to study the cellular regulation of i and I antigen expressions n erythroid cells, burst-forming-unit erythroids were grown in methylcellulose from the blood of normal human adults. Erythroid colonies were harvested on days 9, 11 or 14 of culture. Individual colonies or pooled colonies were subsequently cytocentrifuged and then i or I antigens were studied by immunofluorescence using a monoclonal anti-i or I antibody. Labelled cells were firstly located, their erythroid origin and their stage of maturation were subsequently ascertained after May-Grünwald-Giemsa staining. This approach shows that: (1) a marked heterogeneity occurred from one colony to another, most of the colonies being i-negative while the others contained a mixed population composed by a minority of i-positive erythroid cells; (2) the proportion of both i colonies and i cells decreased with the aging of culture; (3) i antigen was preferentially expressed on immature erythroblasts; however, immaturity per se was not sufficient to determine i expression; (4) I antigen expression was also related to the process of erythroid maturation but in an inverse way; its expression was higher in mature erythroblasts than in immature ones.