Roles of erythropoietin, insulin-like growth factor 1, and unidentified serum factors in promoting maturation of purified murine erythroid colony-forming units.

We have used 75% to 90% pure murine erythroid colony-forming units (CFU-E) to delineate the processes and factors underlying their maturation. These CFU-E form 32 cell colonies and are drawn from what we term generation I of a six-generation long maturation sequence (Landschulz et al, Blood 79:2749, 1992). Applying assays of 59Fe-heme biosynthesis and colony numbers as measures of maturation and analyses of DNA degradation as an index of programmed cell death, we find that (1) erythropoietin (Epo) enhances maturation throughout most of its course; (2) Epo first seems able to forestall DNA degradation when CFU-E reach generation II; (3) the processes that Epo elicits thereafter start to persist when Epo is withdrawn; (4) insulin-like growth factor I (IGF-I) also forestalls DNA breakdown, but later loses effectiveness; (5) IGF-I adds little to maturation when Epo levels are high, but when Epo levels are low, enhances it substantially; and (6) for maturation to be entirely optimal, an unidentified serum factor(s) is probably required when Epo levels are high and is certainly needed when Epo levels are like those in normal animals. Quantitatively, about 40% of optimal in vitro erythropoiesis at normal Epo levels depends on Epo alone, another 30% or less on the addition of IGF-I, and the remaining 30% or more on the addition of unidentified serum factor(s). Applied together, these three or more factors lead to two-thirds of the maximum maturation realized with saturating Epo levels. Because we also find that heme accumulated in CFU-E culture can closely approach levels in red blood cells, we suppose that our conclusions apply as well to CFU-E maturation in vivo.

Onset of erythropoietin response in murine erythroid colony-forming units: assignment to early S-phase in a specific cell generation.

Murine erythroid colony-forming units (CFU-E) representing successive cell generations in a six-generation long in vitro maturation sequence were tested for their response to erythropoietin (Epo) by measurement of Epo-exposure times necessary to stimulate heme biosynthesis. Generation I CFU-E, which produce mainly 32-cell erythroid colonies, were isolated in 82% average purity from spleens of thiamphenicol-treated anemic animals via differential centrifugation. Generation II CFU-E, which produce mainly 16-cell colonies, were similarly isolated in 51% average purity. Although both types of CFU-E had equivalent dose sensitivity to and affinity for Epo, generation II CFU-E responded to shorter pulses of Epo than did generation I. Correlations between DNA cell-cycle profiles and 59Fe-heme biosynthesis resulting from pulsed exposures established that appreciable Epo response only begins when CFU-E attain early S-phase of generation II. Because CFU-E did not require Epo or other serum factors to pass from generation I to II and because the onset of Epo responsiveness coincided with the beginning of DNA replication in generation II, we suppose that differentiation has reprogrammed one or more of the events associated with generation II S-phase in CFU-E and that these alterations allow Epo to act. Further comparisons between CFU-E from generation I and II may allow us to identify the alterations in question and the nature of their interaction with Epo.

Erythropoietin receptors on murine erythroid colony-forming units: natural history.

Erythropoietin (Epo) response and binding was assessed in purified murine CFU-E and their descendants. Several features emerged. First, Epo on CFU-E is in rapid flux: Half-time for 125I-Epo internalization is approximately four to five minutes. Second, computer-aided Scatchard analyses indicate that greater than 70 high-affinity Epo-receptor sites on anemic animal CFU-E are sometimes already occupied by Epo acquired in vivo. When this is removed, 40% of greater than or equal to 370 sites per CFU-E belong to a high-affinity class (dissociation constant, kd: 73 pmol/L +/- 15 [SE]) and 60% belong to a low-affinity class (kd: 813 pmol/L +/- 246). Third, the few small colonies that develop from CFU-E in the absence of Epo are shown, by serial assay of 59Fe-heme biosynthesis, to stem from contaminating erythroblasts: a result consistent with our finding that, after eight-hour CFU-E culture, most erythroblasts no longer require appreciable Epo for growth. Thus, although the early need for Epo by CFU-E is nearly absolute, this need is not met by the often substantial Epo already on board. The inference is that repeated occupancy of the rapidly turning over Epo receptors is required. Fourth, Epo bound and/or internalized by CFU-E descendants decreases to 40% of zero-time levels after 14 hours in Epo-supplemented culture and disappears after 28 hours. Scatchard analyses indicate that 73 pmol/L kd receptor sites become undetectable at seven to eight hours, whereas 813 pmol/L kd sites are undiminished and only one-third less by 16 hours. This apparent disappearance of high-affinity sites and persistence of low-affinity sites suggests that (a) at least two gene products mediate Epo binding, eg, two different receptor polypeptides or one receptor and one cofactor which modulates affinity; (b) high-affinity sites mediate the growth function of Epo during the first eight hours of culture; and (c) lingering low-affinity receptors may mediate some unrecognized Epo function. Fifth, the efficiency with which 106- and 91-Kd CFU-E membrane polypeptides can be cross-linked to 125I-Epo is two- to threefold higher for cells labeled at high Epo concentrations than at low ones, which suggests that these polypeptides largely reflect low-affinity site reactions.

Isolation of a rat liver delta-aminolevulinate dehydrase (ALAD) cDNA clone: evidence for unequal ALAD gene dosage among inbred mouse strains.

We have isolated several cDNA clones encoding delta-aminolevulinate dehydrase [ALAD; porphobilinogen synthase; 5-aminolevulinate hydro-lyase (adding 5-aminolevulinate and cyclizing), EC 4.2.1.24], the second enzyme in the heme biosynthetic pathway. We used a rabbit polyclonal antibody developed against the purified 35-kDa subunit of rat liver ALAD to screen a lambda gt11 cDNA expression library constructed from rat liver mRNA. A prototype clone (ALAD-1) contained a 680-base-pair insert and expressed a 140-kDa beta-galactosidase gene cDNA insert fusion protein immunoreactive with both polyclonal and monoclonal anti-ALAD. Identity of ALAD-1 was verified by hydridization to ALAD mRNA that had been enriched via immunopurification of rat liver polysomes with anti-ALAD. Intensity of such hybridization to a 1500-nucleotide-long mRNA was approximately equal to 200-fold greater than that realized with whole liver mRNA, a result consistent with the degree of immunoenrichment of ALAD mRNA found independently by analysis of cell-free translation products. A second ALAD cDNA clone (ALAD-3) was isolated when the rat liver cDNA expression library was rescreened with ALAD-1. The identity of both ALAD cDNA clones was established by correspondence between their nucleotide sequence and the reported amino-terminal protein sequence of bovine ALAD. Hybridization of ALAD cDNA to mouse genomic DNA indicates that the previously unexplained incremental differences in ALAD enzymatic activity among inbred mouse strains has arisen through alterations in ALAD gene dose.

Production of F cells in sickle cell anemia: regulation by a genetic locus or loci separate from the beta-globin gene cluster.

Levels of fetal hemoglobin (HbF) bearing reticulocytes (F reticulocytes) range from 2% to 50% in patients with sickle cell (SS) anemia. To learn whether any portion of such variation in F cell production is regulated by loci genetically separable from the beta-globin gene cluster, percentages of F reticulocytes were compared in 59 sib pairs composed solely of SS members, including 40 pairs from Jamaica and 19 from the United States. We reasoned that differences in F reticulocyte levels might arise (1) from any of several kinds of artifact, (2) via half-sib status, or (3) because one or more genes regulating F cell production segregate separately from beta S. We minimized the role of artifact by assay of fresh samples from 84 SS individuals, including both members of 38 sib pairs. In 78 of the 84 subjects, serial values for percent F reticulocytes fell within 99.9% confidence limits or were alike by t test (P greater than or equal to .05). This left 32 sib pairs for which F reticulocyte levels in each member were reproducible. When sib-sib comparisons were limited to these 32 pairs, percentages of F reticulocytes were grossly dissimilar within 12 Jamaican and 3 American sibships. Within them, the probability that sibs were alike was always less than or equal to .005 and usually less than or equal to 10(-4). We next minimized the contribution of half-sibs among Jamaicans by a combination of paternity testing and sib-sib comparison of beta-globin region DNA restriction fragment length polymorphisms, especially among discordant pairs. We thereafter concluded that at least seven to eight Jamaican pairs were composed of reproducibly discordant full sibs. There is thus little doubt that there are genes regulating between-patient differences in F cell production that are separate from the beta-globin gene cluster. Still unanswered is (1) whether or not these genes are actually linked to beta S, (2) why F reticulocyte levels in Americans tend to be lower than in Jamaicans, and (3) whether or not differences in F cell production among SS patients are regulated by several major loci or by only one.

Developmental changes in human erythrocyte carbonic anhydrase levels: coordinate expression with adult hemoglobin.

In order to bolster the argument that parallel developmental changes in erythrocyte adult hemoglobin (HbA) and carbonic anhydrase (CA) content provide a potentially suitable model for the dissection of coordinate gene expression, the magnitude of fetal vs adult differences in CA I and CA II levels was examined in human red cell subpopulations obtained after varying periods of exposure to CA-dependent, NH4Cl-HCO-3-mediated, acetazolamide-modulated hemolysis. When content of CA I and CA II was immunologically assessed in cohorts surviving successively longer periods of hemolysis, cord blood red cells were divisible into two populations. Fifteen to thirty percent are rapidly disrupted and have CA I and CA II concentrations similar to those in adult blood erythrocytes. The remaining 70 to 85% have CA I concentrations which are 100-fold less and CA II concentrations which are 5- to 20-fold less than those found in adults. Thus, contrary to past reports, the magnitude of the developmental change in CA I concentration closely resembles the magnitude of change in HbA levels.

Adjuvants to immunological methods for mRNA purification. Application to the isolation of mRNA for carbonic anhydrase I from rabbit reticulocytes.

In order to generate the molecular probes needed to investigate the seemingly coordinate expression of carbonic anhydrase (CA I) and beta-globin within erythrocytes during human development, CA I-containing polyribosomes have been isolated from rabbit reticulocytes by reaction with purified antibodies to CA I followed by immunoadsorption of immune complexes with formalin-fixed protein A-bearing bacteria. In the course of such isolation, a series of maneuvers were seen to have a markedly favorable influence on the level of purity attained. These maneuvers include the use of 5 mg/ml of heparin concentrations to attenuate nonspecific binding and entrapment of unwanted polyribosomes, the addition of 200 units/ml of placental ribonuclease inhibitor to augment recovery in reactions which by test already appeared RNase-free, and the preadsorption of polyribosomes with formalin-fixed bacteria prior to immunological reaction so as to remove a subset of polyribosomes seemingly predisposed to nonspecific binding. In the absence of all of the maneuvers, attained purity was no greater than a few per cent. When all were employed, CA I-mRNA derived from immunopurified polyribosomes was recovered with more than 80% purity and 20% yield, as evident from both immunoassays and electrophoresis of its cell-free products.

Enrichment of erythrocytes of fetal origin from adult-fetal blood mixtures via selective hemolysis of adult blood cells: an aid to antenatal diagnosis of hemoglobinopathies.

Red cell lysis in isotonic solutions containing NH4Cl, NH4HCO3, and a carbonic anhydrase enzyme inhibitor (acetazolamide) is a function of erythrocyte enzyme activity and permeability of cells to the inhibitor. Erythrocyte carbonic anhydrase activity is at least fivefold greater and acetazolamide permeability about tenfold less for adults than for newborns. In this setting, greater than 99.9% of red cells from adults can be hemolyzed at a time when greater than 25% of those from newborns remain intact. This easily applied method may be useful when antenatal diagnosis of hemoglobinopathies is otherwise precluded by contaimination with maternal erythrocytes. The feasibility of differential hemolysis via NH4Cl--HCO3-mediated, acetazolamide-modulated reactions is shown by the successful isolation of the few fetal-origin erythrocytes present in grossly nonbloody amniotic fluids and, in one instance, by approximately 3300-fold enrichment of apparently authentic fetal-origin red cells from the arm blood of a woman in her 18th wk of pregnancy.

Variations in the frequency of fetal hemoglobin-bearing erythrocytes (F-cells) in well adults, pregnant women, and adult leukemics.

Fetal hemoglobin (HbF) in adult life is usually restricted to a few erythrocytes called F-cells. Their presence can be detected and the quantity of HbF per F-cell estimated by pericellular immunodiffusion reactions with anti-HbF. In normal adults F-cell frequency (range: 1 F-cell per 25 erythrocytes to 1 per 6800) and the mean amount of HbF per F-cell (range: approximately 14 to 28% of mean cell hemoglobin), although they are variables, remain constant over many months. Frequencies are similar in men and women. For a given individual, F-cell lifetimes are probably similar to those of other erythrocytes. F-cell production is not appreciably influenced by short term exposure to high altitude (approximately 4300 m) hypoxia. F-cell frequencies are briefly but substantially increased in many women during midpregnancy. In some women, presumptive 5- to 15-fold increases in F-cell production result in observed 3- to 7-fold increases in F-cell frequency during the 23rd to 31st weeks of gestation. These increases in F-cell frequency arise from selective alterations in maternal erythropoiesis and not from transplacental bleeding from the fetus. Substantial increases in F-cell frequency also occur in most adults with acute leukemia. In both pregnancy and leukemia, F-cell contributions of HbF are sufficient to account for modest elevations found in hemolysate HbF levels.

Fetal hemoglobin restriction to a few erythrocytes (F cells) in normal human adults.

During adult life, the quantity of fetal hemoglobin (HbF) present in F cells--that is, rare erythrocytes which are reactive with rabbit antiserum to human HbF during microscopic immunodiffusion--is sufficient to account for all of the small quantity (less than 0.7 percent) of HbF normally present in whole blood. Thus, erythrocytes are normally heterogeneous with respect to the presence of HbF.

Silent hemoglobin alpha genes in apes: potential source of thalassemia.

Small quantities of unusual hemoglobins were found in 1 of 37 chimpanzees and 2 of 6 gorillas. In each genus these hemoglobins contain unique alpha chains that differ from the ordinary by eight to nine scattered amino acid changes. The unusual chains arise from a hitherto undetected hemoglobin (3)alpha locus. No (3)alpha products are found in most apes; accordingly, (3)alpha is considered synthetically inactive in all but a few reversion mutants. Indirect evidence that the inactive (3)alpha locus is juxtaposed to an active alpha locus together with the supposition that (3)alpha exists in man provides a setting wherein thalassemia might be produced by nonhomologous recombination between two loci.

Hemoglobins A and A2 in New World primates: comparative variation and its evolutionary implications.

Hemoglobin A(2) (alpha(2)delta(2)) in New World primates represents about 1/160 to 1/16 of total hemoglobin and, by virtue of this low proportion, is presumed to be functionally unimportant. Nonetheless, A(2) exhibits genetic polymorphism by electrophoresis in three out of five genera, whereas the major component, hemoglobin A (alpha(2)beta(2)), is electrophoretically invariant. Moreover, in four genera, including man, the evolutionary accumulation of mutations has been greater in delta than in beta Such findings suggest that both polymorphism and evolutionary changes can accrue to an effectively functionless and thus selectively nearly netutral gene.