ALBERT

Description: It is well known that the globin genes are organized within the human genome as two gene clusters on chromosomes 16 and 11 (α cluster: ζ2-ζ1-α2-α1-? and β cluster: ε-Gγ-Aγ-δ-β). Switching in the expression patterns of those genes during human ontogeny generally follows a pattern determined by their genomic arrangement, but the genetic mechanism is not defined. In an effort to better understand the γ-to-β globin switching phenomenon, gene expression patterns from reticulocyte RNA were studied. cDNA were generated from the reticulocytes of 28 separate donors (14 cord blood; 14 adult blood) and analyzed using Affymetrix HG-U133 arrays. Quantitative PCR was used to confirm the array expression patterns. Among those cDNA identified as having expression patterns that were relatively decreased among adult reticulocytes, we were surprised by expression associated with a cluster of expressed sequence tags (EST) located within the α-globin locus in the region between the ζ1 and α2 genes. In the context of the 44,229 arrayed probes, high-level expression from that EST cluster was detected (ranked 〈 25th). Informatic analysis revealed alignment with the second and third exons previously described for the Ψα2 globin gene, but no significant similarity with the first exon of Ψα2. Based upon these comparisons, we used adult human reticulocyte RNA to amplify a cDNA clone for further study (GenBank: AY698022; named μ-globin). Sequencing of that clone revealed an insert that encodes a predicted 423 nt. open reading frame that is one amino acid residue shorter than the α2-globin gene. While the mature μ-globin mRNA results from splicing of two introns, it lacks a canonical Kozak sequence. BLASTP and CLUSTALW analyses revealed the highest level of homology with the avian αD-globin protein with a sequence identity of 55% (78/141 amino acids). In addition, the predicted heme- and globin-binding amino acids of μ-globin and avian αD-globin are largely conserved. Experimentally, we analyzed the expression level of the μ-globin transcripts using quantitative real-time PCR. Non-erythroid tissues including whole brain, white blood cells, and the Jurkat cell line demonstrated only background levels. The expression level of μ-globin (copies per ng cDNA) was 3.04x10(4) ± 1.68x10(3) in fetal liver, 1.71x10(5) ± 9.51x10(4) in cord reticulocytes, 1.15x10(4) ± 1.19x10(3) in bone marrow, and 2.17x10(4) ± 6.84x10(3) in adult reticulocytes. Based upon the consistent decrease in μ-globin expression between fetal and adult erythroid tissues, we measured the expression of γ-globin in those tissues for comparison. The expression level of γ-globin (copies per ng cDNA) was 2.49x10(6) ± 1.48x10(5) in fetal liver, 1.49x10(8) ± 7.88x10(7) in cord reticulocytes, 3.84x10(3) ± 3.83x10(2) in bone marrow, and 5.70x10(5) ± 5.80x10(5) in adult reticulocytes. To determine if μ-globin expression is regulated during erythropoiesis, we assayed primary human erythroid progenitor cells cultured in erythropoietin. The expression pattern of μ-globin was similar to the other globin genes increasing from background levels on day 0 to maximum levels in committed erythroblasts (3.56x10(5) ± 1.99x10(4) copies per ng cDNA). These results suggest that the human genome encodes a previously unrecognized α-globin that most closely resembles avian αD-globin. Expression of that gene is decreased during the fetal-to-adult transition of human ontogeny and regulated during erythropoiesis.

Description: Iron overload causes considerable morbidity associated with thalassemia due to inappropriate suppression of the iron regulator, hepcidin. One possible explanation for this phenomenon is that protein(s) that are normally secreted into the marrow microenvironment by erythroblasts become endocrine signals in patients with thalassemia due to the myeloproliferative nature of the disease. To test this hypothesis, progenitor and precursor cell transcriptional profiles were generated with Affymetrix GeneChip Human Genome U133 Plus 2.0 Arrays using primary erythroblasts cultured from 15 healthy human donors. For this study, informatic analyses were focused upon 54 members of the TGF–B/BMP superfamily. Among the subset of genes identified by this screen, evidence for high-level expression of a gene named growth differentiation factor 15 (GDF15) was discovered in the precursor cell profiles. Quantitative PCR, Western, and ELISA analyses confirmed expression and secretion of GDF15 during erythroblast maturation. GDF15 is an apoptosis-associated protein expressed primarily by the placenta. To determine whether GDF15 serves as a regulator of hepcidin expression, hepcidin expression assays were performed using a human hepatoma cell line (HuH-7). BMP2 and BMP4 were studied for comparison. Addition of BMP2 and BMP4 (range 10,000–500,000 pg/ml) resulted in dosed increases in hepcidin mRNA (5–30 fold). At the concentrations of GDF15 normally found in human blood (500 pg/ml), a 2-fold increase in the expression of hepcidin was measured compared to matched cultures containing no supplemental GDF15. However, GDF15 dosed to levels above 5,000 pg/ml resulted in a significant reduction in hepcidin expression. Next, plasma levels of GDF15 were measured in the peripheral blood of 162 donors (donor groups included healthy controls, sickle-cell syndromes, thalassemia syndromes, and other causes of anemia) to determine whether plasma GDF15 levels are dysregulated in thalassemia. Plasma samples from 21 hereditary hemochromatosis donors provided evidence that significantly elevated GDF15 expression is not associated with iron overload in the absence of erythroid pathology. Compared with mean plasma GDF15 levels of 536±222 pg/ml among the control samples, only patients with thalassemia intermedia, thalassemia major, and HbE-beta thalassemia showed significantly elevated plasma levels of GDF15 (9 donors; mean GDF15 24,600 pg/ml; range 8,980–75,100 pg/ml; p370 ug/L) were noted in each of those patients regardless of their transfusion or chelation history. These novel findings suggest that GDF15 is secreted from human erythroblasts, released into the circulation at extremely high levels in thalassemia patients, and contributes to iron overloading in those patients by suppressing hepcidin expression.

Description: Therapeutic regulation of globin genes is a primary goal of translational research aimed toward hemoglobinopathies. Signal transduction was used to identify chromatin modifications and transcription factor expression patterns that are associated with globin gene regulation. Histone modification and transcriptome profiling were performed using adult primary CD34+ cells cultured with cytokine combinations that produced low versus high levels of gamma-globin mRNA and fetal hemoglobin (HbF). Embryonic, fetal, and adult globin transcript and protein expression patterns were determined for comparison. Chromatin immunoprecipitation assays revealed RNA polymerase II occupancy and histone tail modifications consistent with transcriptional activation only in the high-HbF culture condition. Transcriptome profiling studies demonstrated reproducible changes in expression of nuclear transcription factors associated with high HbF. Among the 13 genes that demonstrated differential transcript levels, 8 demonstrated nuclear protein expression levels that were significantly changed by cytokine signal transduction. Five of the 8 genes are recognized regulators of erythropoiesis or globin genes (MAFF, ID2, HHEX, SOX6, and EGR1). Thus, cytokine-mediated signal transduction in adult erythroid cells causes significant changes in the pattern of globin gene and protein expression that are associated with distinct histone modifications as well as nuclear reprogramming of erythroid transcription factors.

Description: The highly variable levels of fetal hemoglobin (HbF) reactivation among patients with sickle cell disease is not well understood. We previously determined that pancellular reversal of gamma-globin gene silencing is achievable through the activation of specific signal transduction pathways using stem cell factor (SCF) and transforming growth factor beta (TGF-B). Based upon the reliability of this culture assay for HbF reactivation, the culture system is additionally being utilized to screen for inhibitors of HbF reactivation. As an initial screening strategy, published reports, high-throughput sequencing efforts, and array-based transcription profiles were examined to identify erythroblast-expressed receptors with known ligands that may serve a signaling role in the regulation of globin expression. A dose-escalation strategy was used to screen seventeen ligands that bind to a subset of G-protein coupled receptors (GPCR). Among those ligands, oleoyl-lysophosphatidic acid (OLPA) and neurokinin A (NKA) were identified as potential inhibitors of HbF reactivation. Lysophosphatidic acid is a lipid metabolite that is released after tissue injury and may play a role in vascular remodeling. Neurokinin A is a small neuropeptide mainly involved in pain signaling. OLPA inhibited HbF reactivation in a dose-dependent manner (HbF/HbF+HbA ratios: EPO: 0.7±0.01%; EPO+SCF+TGF-B (EST): 36.3±0.1%; EST+10nM OLPA: 36.5±0.1%; EST+100nM OLPA: 31.6±0.1%; EST+1uM OLPA: 12.7±0.1%, p=1.9E-03). NKA also inhibited HbF reactivation, but the magnitude of the NKA effect was donor-specific. Cells from one donor demonstrated a robust response (HbF/HbF+HbA ratios: EPO: 1.7%; EST: 26.7%; EST+10nM NKA: 24.5%; EST+100nM NKA: 8.8%; EST+1uM NKA: 7.8%). Since OLPA and NKA may both act through adenylate cyclase, the role of this enzyme was examined further using the activator, forskolin (0.8 uM-20uM) and inhibitor, SQ22536 (40uM – 600uM). While we were unable to demonstrate a significant increase in HbF among cultures grown in forskolin, the addition of SQ22536 resulted in a profound, dose-dependent inhibitory effect on the expression of HbF (EPO: 1.3±1.2%; EST: 36.7±11.3%; EST+40uM SQ: 30.3±10.2%; EST+200 uM SQ : 14.5±4.5%, p=1.9E-02; EST+600uM SQ: 2.3±2.7%, p=4.1E-03). Flow cytometry revealed that the inhibitory effects of SQ22536 on HbF were pancellular. Importantly, inhibition of the main downstream target of adenylate cyclase (protein kinase A) with H89 (10nM-1.0 uM) did not result in a significant reduction in HbF. However, phosphorylation studies demonstrated that SQ22536 altered the kinetics of MEK activation. MEK is a kinase that helps mediate the HbF-activating effect of SCF [Blood 2004, 1;103(5):1929–33]. These data suggest that inhibitory effects upon HbF reactivation in adult erythroblasts can be mediated by endogenous signaling molecules. Based upon these results, we propose that patient-specific responses to tissue injury and pain may play a negative role in HbF reactivation in sickle-cell disease.

Description: Adenosine signal transduction in human cells is achieved through four distinct G-protein-coupled receptors named A1, A2A, A2B and A3. Adenosine signaling events affect several physiological processes including growth regulation and cellular responses to hypoxia. Despite extensive studies of native and pharmacologically-targeted adenosine signal transduction, the potential for adenosine to regulate erythropoiesis remains largely unexplored. Here we systematically studied adenosine receptor expression, activation, and inhibition among CD34+ human erythroid progenitor cells cultured in the presence of erythropoietin. This 14 day culture system has been studied extensively in the context of erythroblast commitment and terminal differentiation (BLOOD 99(8):3005–13). Using real-time RT-PCR, we demonstrated that the A2A, A2B and A3 adenosine receptors are expressed in highly regulated patterns. The expression of A2A and A2B were highest early in the culture period and later declined as the cells underwent terminal differentiation. Receptor A3 gene expression gradually increased during the terminal differentiation. Expression of the A1 receptor gene was not detected above background levels at any stage of differentiation (compared with the positive control). Based upon those receptor expression patterns, several adenosine receptor signaling agonists and antagonists were screened for possible effects on erythroblast growth and differentiation. Among those screened molecules, the selective A3 adenosine receptor agonist, CI-IB-MECA, had effects on the proliferation and differentiation of the cells without overt toxicity in the dose range of 10–100uM. CI-IB-MECA significantly inhibited the proliferation of the cells with resulting cell counts being only 10% of those in matched controls after 14 days. Cell cycle analyses demonstrated that the growth inhibitory effects of CI-IB-MECA were largely due to a significant reduction in the percentage of S-phase proerythroblasts (S-phase: 16 ± 5% in CI-IB-MECA versus 51 ± 2% in matched controls; p80%) in the matched controls. Consistent with the morphological examination, HPLC analyses revealed barely detectable levels of hemoglobin in the presence of CI-IB-MECA after 14 days. These finding suggest that selective activation of adenosine A3 receptors permits erythroid commitment, but profoundly inhibits the proliferation of those committed cells and retards their terminal differentiation. As a novel regulator of erythropoiesis, adenosine A3 receptor signaling should be explored in patients with growth-related erythroid diseases.

Description: A previously undefined transcript with significant homology to the pseudo-α2 region of the α-globin locus on human chromosome 16 was detected as part of an effort to better define the transcriptional profiles of human reticulocytes. Cloning and sequencing of that transcript (GenBank AY698022; named μ-globin) revealed an insert with a 423-nucleotide open reading frame. BLASTP and ClustalW and phylogenetic analyses of the predicted protein demonstrated a high level of homology with the avian α-D globin. In addition, the heme- and globin-binding amino acids of μ-globin and avian α-D globin are largely conserved. Using quantitative real-time polymerase chain reaction (PCR), μ-globin was detected at a level of approximately 0.1% that measured for α-globin in erythroid tissues. Erythroid-specific expression was detected by Northern blot analysis, and maximal expression during the erythroblast terminal differentiation was also detected. Despite this highly regulated pattern of μ-globin gene transcription, μ-globin protein was not detected by mass spectrometry. These results suggest the human genome encodes a previously unrecognized globin member of the avian α-D family that is transcribed in a highly regulated pattern in erythroid cells. (Blood. 2005;106:1466-1472)

Description: Hemoglobin switching patterns during human ontogeny are highly correlated with transcription of the alpha and beta globin genes on chromosomes 16 and 11, but the molecular and genetic mechanisms responsible for the switching phenomenon are not yet defined. For a better understanding of post-natal globin gene switching and silencing, the expression profiles from purified reticulocyte mRNA (28 separate clinical samples; 14 cord, 14 adult) were studied using Affymetrix HG-U133 arrays (44,229 probe sets). To validate the transcriptome expression levels derived from the microarrays and to discriminate the differences between cord blood and adult blood reticulocytes, we performed quatitative real-time PCR. Among 21 tested control genes, the ratios of adult-to-cord blood (AB/CB) array hybridization intensities versus PCR copy numbers demonstrated a correlation coefficient of 0.978. Hence, the AB/CB array hybridization intensities provide a reliable description of erythroid transcriptome patterns associated with post-natal hemoglobin switching. Stringent screening criteria were then used to identify genes that may be involved in fetal-to-adult hemoglobin switching. All 44,229 probe sets were filtered to identify genes that demonstrated greater than a five fold difference in AB and CB mean signal intensities as well as consistent differences among the 14 samples from each group. In addition, the screening criteria required high-level signals in either AB or CB (〉1000 FU), and t-test p-values between AB and CB