The hydroxyurea-induced small GTP-binding protein SAR modulates γ-globin gene expression in human erythroid cells

Hydroxyurea (HU), a ribonucleotide reductase inhibitor and an S-phase–specific cytotoxic agent, has been successfully used in the clinic to treat sickle cell disease because it augments the production of fetal hemoglobin (Hb F).^1,2  In a multicenter study of HU in sickle cell anemia (MSH), HU treatment dramatically reduced the frequency of hospitalization due to the incidence of pain, acute chest syndrome, and the blood transfusion requirements in patients with sickle cell disease.²  After 2 years of HU treatment, Hb F levels increased from 5% of total Hb to 9% in the majority of patients.²  Similarly, as Hb F levels became elevated after HU treatment, the rate and extent of sickle hemoglobin (Hb S) polymerization appeared to be reduced.¹  A 9-year follow-up of 233 patients in the original MSH cohort indicated, among other findings, that taking HU was associated with a 40% reduction in mortality.³ 

Several mechanisms have been proposed to account for the induction of Hb F by HU. Because HU directly kills late erythroid progenitor cells and triggers rapid erythroid regeneration,⁴  HU has been assumed to promote Hb F production by perturbing the maturation of erythroid precursors. Accelerated erythropoiesis increases the chance of premature lineage commitment and the induction of F-cell formation.⁵  HU also increases erythropoietin (EPO) production and enhances the proliferation of erythroid precursors and their ability to synthesize Hb F.⁶  HU could effectively stimulate erythropoiesis and production of Hb F through multiple mechanisms.⁷  It has been reported that induction of Hb F by HU is mediated by up-regulation of GATA-2 and down-regulation of GATA-1.^7,8  Recent studies from our laboratory have also demonstrated that, in addition to regulating GATA expression, HU also induces expression of genes related to the cell cycle and apoptosis; the actions of HU depend upon its concentration.⁷  Induction of peroxidation by HU can generate nitric oxide (NO),^9,10  a known inducer of soluble guanylate cyclase (sGC), which, in turn, can modulate the expression of γ-globin.¹¹  However, the molecular mechanisms by which HU regulates γ-globin gene expression need to be further elucidated. To elucidate the molecular mechanisms underlying the HU-induced elevation of Hb F and to explore new signal transduction pathways related to HU-mediated hemoglobin gene expression, we used mRNA differential display to identify HU inducible gene(s) in human adult erythroid cells. Here, we identify a new small guanosine triphosphate (GTP)–binding protein, secretion-associated and RAS-related (SAR) protein, as a specific HU-inducible gene.

The small GTP-binding protein superfamily comprises more than 100 members in eukaryocytes.^12,13  Members of the SAR/adenosine diphosphate (ADP) ribosylation factor (Arf) family are GTPases^14-16  active in protein trafficking and maturation.¹⁷  SAR itself functions in cargo selection and export of proteins from the endoplasmic reticulum to the Golgi via the cytosolic coat protein complex II (COPII) secretory pathway. It has been reported that several Ras superfamily members participate in hematopoietic cell proliferation and differentiation processes involving endo- and exocytosis or cell adhesion.^18-21  However, it is unknown whether SAR participates in hemoglobin gene regulation, and which signal transduction pathway SAR uses. Here, we demonstrate that HU-induced SAR plays a pivotal role in γ-globin gene induction by causing cell apoptosis and G₁/S-phase arrest through the reduction of phosphatidylinositol 3 (PI3) kinase and extracellular protein–related kinase (ERK) phosphorylation and increased p21 and GATA-2 expression. Localization of SAR in the endoplasmic reticulum was also confirmed.

Two-phase liquid cultures of human adult erythroid cells (HAECs) were prepared from the peripheral blood of healthy donors. Mononuclear cells were isolated by centrifugation over Ficoll-Hypaque (1.077 g/mL; Organon Teknika, Durham, NC), washed twice with Dulbecco phosphate-buffered saline (PBS), and cultured according to a previously published 2-phase liquid culture protocol.²²  Cells were treated with 100 μM HU on days 7 and 8, and RNA was collected on day 10 of phase II culture for differential display. K562 cells grew in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) plus 4 mM glutamine, 10 U/mL penicillin, and 10 μg/mL streptomycin. Growth curves were performed by seeding cells at a density of 2.5 × 10⁴ cells/mL, then counting them every 24 hours for 4 days. The morphology of K562 cells was observed after hematoxylin and eosin staining. Enriched primary bone marrow AC133⁺ cells and CD34⁺ cells were purchased from Poietic Technologies (Gaithersburg, MD) and cultured according to a previously described protocol.²³ 

Differential display was performed using an RNA Image Kit (GeneHunter, Nashville, TN). We used a MessageClean DNase I kit (GeneHunter) to eliminate chromosomal DNA from total RNA before differential display experiments. Reverse transcription–polymerase chain reaction (RT-PCR) followed the manufacturer's protocol. Reproducible differential bands were excised and eluted by boiling the gel slices for 15 minutes. Eluted DNA was reamplified using the same primer set, purified, and cloned into the PCR II TOPO vector (Invitrogen, Carlsbad, CA) for sequencing. The human SAR gene was cloned by 5′- and 3′-RACE (rapid amplification of cDNA ends) with the Marathon cDNA Amplification Kit using bone marrow Ready Marathon cDNA (Clontech, Palo Alto, CA). Homology searches of GenBank and expressed sequence tag (EST) databases were performed using GCG software (Genetics Computer Group, Madison, WI).

A full-length SAR cDNA was cloned into the pGEM-T Easy vector (Promega, Madison, WI), then transcribed and translated into a sulfur 35 (³⁵S)–labeled peptide using a TNT Quick T7-coupled transcription/translation system (Promega) with [³⁵S]-methionine (555 MBq/mL [15 mCi/mL], Amersham Pharmacia Biotech, Piscataway, NJ). The translation product was subjected to electrophoresis in 4%-12% Bis-Tris ([bis(2-hydroxyethyl)amino]tris(hydroxymethyl)methane) gels. The gels were exposed to film to determine the size of the translated product by comparison with known protein standards.

A SAR cDNA probe was hybridized to Multiple Tissue Northern (MT) Blots (Clontech) according to the manufacturer's protocol. Membranes were stripped between probes by incubating the blots in sterile 0.5% sodium dodecyl sulfate (SDS) solution at 95°C for 10 minutes. The Human Cell Cycle filter membrane containing various human cDNA sequences was purchased (Superarray, Frederick, MD). For hybridization probes, 5 μg of total RNA from K562 cells transfected with SAR or with vector alone was converted to cDNA by random hexamer labeling with Moloney murine leukemia virus (MMLV) reverse transcriptase and [³³P]–deoxycytidine triphosphate (dCTP) (370 MBq/mL; NEN Life Sciences, Boston, MA), according to the protocol provided by the supplier. The filter arrays were scanned in a Fuji BAS-1500 phosphoimager (Fuji, Stamford, CT). Before quantitative analysis, hybridization signals for each probe were normalized using GAPDH (glyceraldehyde-3-phosphate dehydrogenase) cDNAs as reference controls. For Western blots, cell lysates containing 50 μg protein were loaded on Tris-glycine gels (Invitrogen), transferred to nitrocellulose membranes (Invitrogen), and then immunoblotted with antibodies to PI3 kinase (BD Bioscience, Palo Alto, CA), phospho- and total ERK, GATA-1, and GATA-2 (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 hour. Washing and detection using an enhanced chemiluminescence (ECL) Western Blotting Analysis System (Amersham Pharmacia Biotech) was performed according to the manufacturer's protocol.

Proteins were separated on 4% to 15% gradient SDS–polyacrylamide gel and transferred onto Immobilon-P membrane (Millipore, Bedford, MA). After transfer, the membrane was dried and re-wet in methanol, then blocked for 1 hour with blocking buffer (LI-COR, Lincoln, NE). Next, the membrane was clamped with a Western blotting manifold that isolated 41 channels across the membrane. In each channel, a complex antibody cocktail was added and allowed to hybridize for 1 hour at 37°C. The blot was washed and hybridized for 30 minutes at 37°C with secondary goat antimouse conjugated to Alexa680 fluorescent dye (Molecular Probes, Eugene, OR). The membrane was washed, dried, and scanned at 700 nm using the Odyssey Infrared Imaging System (LI-COR). The bands were identified, and molecular masses were assessed using specialized software. Each observed band, where possible, was manually matched against the expected molecular mass of a protein recognized by an individual antibody in the mixture.

We cloned the SAR cDNA into pEF6/V5-His-TOPO (Invitrogen) and stably transfected it into K562 cells by electroporation (Amaxa Biosystems, Cologne, Germany). Stable SAR-expressing clones were selected by culturing the cells in 6 μg/mL blasticidin for 2 weeks. An antibody directed against the His tag was used to identify high and low levels of SAR by Western blotting; expression was confirmed by Northern blotting and RT-PCR.

The SAR coding region with a green fluorescent protein (GFP) tag at the 3′ end or GFP alone was subcloned into the murine stem cell virus (MSCV) retroviral expression vector (Clontech). The retroviruses were packaged into PT67 packaging cells. The viral titers were determined and high-titer viral clones were selected. For SAR transduction in CD34⁺ hematopoietic stem cells, 2 × 10⁶ viral particles were preloaded onto a RetroNectin-coated plate (Takara Shuzo, Kyoto, Japan) and incubated at 37°C for 5 hours. Just prior to infection, the viral supernatant was discarded and the plate was washed with PBS; 1 × 10⁵ CD34⁺ cells were then added to the preloaded viral plate with growth medium, and the plate was incubated at 37°C. This procedure was repeated twice at 24-hour intervals. On day 7, the infected CD34⁺ cells were subjected to GFP sorting, and GFP-positive CD34⁺ cells were selected and cultured. The sorted cells were harvested on day 10, day 14, and day 18 for further analysis.

RNA was isolated with Tri Reagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer's recommended protocol. RT-PCR was carried out using SuperScript reverse transcriptase and Taq DNA polymerase (Invitrogen) according to the manufacturer's instructions. PCR primer pairs were as follows: γ-globin sense (SN): 5′-AAGATGCTGGAGGAGAAAC-3′, antisense (ASN): 5′-TGCTTGCAGAATAAAG CC-3′; SAR SN 5′-CAATATCCAATGTGCCAATCCTTA-3′, ASN 5′-GAAAACCCTCGCC GTACCTT3′; GAPDH SN 5′-SN TGAAGGTCGGAGTCAACGGATTTGGT-3′, ASN 5′-CATGTGGGCCATGAGGTCCACCAC-3′; Glycophorin C (GlyC) SN: 5′-TGGGG AAGCAAGGAAATAAG-3′; ASN 5′-AGCTGGATCAATGGCAGAC-3′. Real-time PCR was accomplished with Platinum QPCR SuperMix-UDG (Invitrogen) according to the manufacturer's instructions. The γ-globin and β-globin gene primers and probe sequences were published previously²⁴ ; control β-actin primers and probe were ordered from Applied Biosystems (Foster City, CA).

GFP was tagged onto the 3′ end of the SAR coding region by cloning it into the pGFP-N1 vector (Clontech); the resulting construct was transfected into K562 cells by electroporation. After 8 hours or 18 hours of culture in fresh medium, transfectants were collected and fixed with 2% paraformaldehyde in PBS. Cells were then washed 3 times before being cytospinned onto glass slides. The cells were then incubated with an anticalreticulin antibody (Affinity Bioreagent, Golden, CO) for 4 hours and antirabbit antibody for 1 hour before being visualized under a Zeiss TV200 confocal fluorescence microscope with a × 100 objective (1.4 NA). Emission wavelengths were 488 nm and 567 nm for fluoroscein isothiocyanate (FITC) and rhodamine, respectively. The images were collected on an ORCA ER CCD (Hamamatsu, Bridgewater, NJ). Images were acquired and analyzed using Improvision Software (Improvision, Lexington, MA).

Cells were counted and centrifuged at 207 g for 5 minutes. The cell pellets were resuspended at a density of 0.5 to 1.0 × 10⁶ cells/mL; 400 μL Quickstain solution (FAST Systems, Gaithersburg, MD) was then added to each sample, and the samples were incubated in darkness for 30 minutes at 22 to 28°C. We used flow cytometry (EPICS-XL Model; Beckman Coulter, Hialeah, FL) with signal acquisition at 675 nm to determine cell-cycle distributions. Laser excitation was at 488 nm; emission filters for FITC and propidium iodide were 525 nm and 630 nm, respectively. The terminal deoxynucleotidyl transferase–mediated dUTP (deoxyuridine triphosphate) nick-end labeling (TUNEL) assay for cell apoptosis was performed according to the manufacturer's protocol (Promega). The images were acquired with a fluorescence microscope system (Axioplan 2; Carl Zeiss MicroImaging, Thornwood, NY) with a 20 ×/0.75 NA objective and a Zeiss Axiocam HRc real-color CCD camera. Acquisition software is AxioVision 4.

The statistical analysis of the results were obtained using the Student t test for significance of the difference of individual pairs and the Pearson of correlation coefficient for the closeness of linear relationship.²⁵ 

In vivo studies have shown that continuous infusion of 20 mg/kg HU into patients maintains a serum concentration of HU at 100 μM.²⁶  In order to obtain an exposure regimen for the differential display study that closely resembled the in vivo pharmacokinetics of HU treatment in patients, HAECs were treated with 100 μMHU on day 7 and day 8, and RNA was collected on day 10. A reverse dot blot was used to confirm differential expression, and one particular transcript was identified that was significantly induced in response to HU treatment. This gene had 92% similarity to the mouse small GTP–binding protein SAR1 by sequence analysis. We then obtained the full-length human SAR cDNA sequence (2878 base pair [bp]) by screening a bone marrow library with 5′-and 3′-RACE. The human SAR gene was mapped to chromosome 10q22 using a National Center for Biotechnology Information (NCBI) genomic data search. Alignment of the cDNA with the chromosome 10 genomic sequence (Celera Discovery System, Rockville, MD) revealed that the SAR gene contains 8 exons and 7 introns. Sequence analysis showed a 599-bp open reading frame that includes 3 GTP binding motifs.

Human multiple-tissue RNA blots probed with the SAR cDNA revealed transcripts of 3.0 and 0.8 kilobase (kb) in most tissues. SAR mRNA was abundant in glandular tissues, such as prostate, thyroid, and adrenal gland. In contrast, relatively low levels were found in spleen, small intestine, spinal cord, lymph node, and trachea, and very low levels were detected in leukocytes, bone marrow, and ovary (Figure 1A). SAR mRNA was barely detectable in primary cultures of HAECs, but increased by at least 4-fold after 48 hours of 100 μM HU treatment (Figure 1B). In vitro transcription and translation of full-length SAR cDNA produced a single 22-kDa peptide (Figure 1C).

To localize SAR expression within cells, we transiently transfected GFP-tagged SAR (SAR-GFP) cDNA into the K562 cells. Cells were fixed after transfection, and then incubated with an anticalreticulin antibody for comparison with the SAR-GFP localization; calreticulin is expressed exclusively in the endoplasmic reticulum. Under a confocal fluorescence microscope, the SAR-GFP signal appeared predominantly in the cytoplasmic and perinuclear areas (Figure 1D). K562 cells transfected with GFP alone showed primarily nuclear fluorescence (data not shown). The SAR-GFP colocalized with calreticulin, suggesting that SAR was present in the endoplasmic reticulum, as expected from SAR's known protein-trafficking function.

To determine whether and how SAR regulates γ-globin gene expression in human erythroid cells, we stably transfected SAR into K562 cells and then analyzed multiple SAR-expressing and vector-transfected clones. Northern blotting revealed significantly higher γ-globin mRNA levels in SAR-transfected cells than in vector-transfected cells (Table 1). To better clarify the correlation between SAR and γ-globin expression, we divided the SAR-expressing cells into 2 groups: clones with high γ-globin gene expression (> 2-fold induction) and clones with lower γ-globin gene expression (< 2 fold induction). As shown in Table 1, both groups had significant induction of γ-globin mRNA compared with the vector control (P < .001), ranging from a 1.5-fold to a 4.3-fold increase in γ-globin mRNA levels. On average, SAR-expressing clones had 3.5-fold higher levels of γ-globin mRNA than vector controls. Coexistence of SAR mRNA and high levels of γ-globin mRNA in K562 cells indicate that HU-induced SAR may be involved in γ-globin mRNA induction in erythroid cell development.

To investigate whether SAR is a common or specific inducer for the γ-globin gene, we designed 3 sets of experiments. First we used semi-quantitative RT-PCR to analyze and correlate SAR and γ-globin mRNA levels in SAR-expressing K562 cells; we found a correlation coefficient of 0.76 between SAR and γ-globin mRNA levels (Figure 2A). Second, when parental K562 cells were treated with various hemoglobin gene inducers (hydroxyurea, butyrate, and 5-azacytidine) for 48 hours, a Northern blot indicated that all 3 inducers up-regulated γ-globin mRNA, from 2.5-fold for butyrate to 3.0-fold for hydroxyurea. In the same experiment, SAR mRNA increased in parallel with the γ-globin mRNA, from 1.5-fold after butyrate treatment to 2.1-fold after hydroxyurea (Figure 2B-C). Third, the expression of SAR and γ-globin mRNA in cultured primary bone marrow AC133⁺ cells was determined by RT-PCR at different time points. In the presence of EPO, both SAR and γ-globin mRNA began to be expressed at day 10, reached a peak at day 21, and declined to baseline at day 28, reflecting a similar time course and pattern of gene expression (Figure 2D). These data further indicate that there is a close temporal relation between SAR and γ-globin mRNA expression in primary bone marrow cells, and strongly suggest that the SAR and γ-globin genes coexist in cells of this erythroid lineage.

We next investigated the effects of SAR using primary bone marrow stem cells. A SAR-GFP fusion construct was produced using the pMSCV retroviral expression system. A pure population of SAR-infected CD34⁺ stem cells was obtained after cell sorting. Both γ- and β-globin mRNA transcripts were measured dynamically using real-time PCR (Figure 3A-B). SAR enhanced both γ-globin and β-globin mRNA expression in the CD34⁺ stem cells, but the effects on γ-globin were larger and more sustained than the effects on β-globin (Figure 3C). SAR enhanced γ-globin gene expression 3-fold at day 10, 5.4-fold at day 14, and 15.8-fold at day 18 in SAR-infected CD34⁺ cells, compared with 5.7-fold at day 10, 1.6-fold at day 14, and 2.9-fold at day 18 for β-globin mRNA induction.

We also noted that SAR-transfected K562 cells grew more slowly than vector-transfected controls. Parental and vector-transfected cells doubled in approximately 18 hours, compared with 30 hours in SAR transfectants (Figure 4A). Hematoxylin and eosin staining of these cells showed the existence of many enlarged cells. The nuclear-to-cytoplasmic ratio was increased in some of these cells, which also exhibited macrocytic morphology, implying that SAR delayed erythroid maturation. HU-treated K562 cells had similar morphologic features, particularly after 4 days of HU treatment (Figure 4B).

HU not only causes apoptosis in K562 cells, it also up-regulates apoptosis-related molecules.⁷  As expected, the incidence of apoptosis was notably higher in SAR-expressing K562 clones than in vector controls. A TUNEL assay showed that a sizable fraction of SAR-expressing cells underwent apoptosis (Figure 4C). This fact and the prolonged doubling time in SAR-expressing cells indicate that the cell cycle might be impacted due to the forced SAR expression. As shown in Figure 4D, the percentage of cells in G₂/M decreased from 19.3 ± 0.2% in vector-transfected cells) to 5.3 ± 3.0% in SAR transfectants, while the G₁/S phase fraction increased to 21% in SAR-expressing cells, suggesting that SAR-expressing cells accumulate in G1 and S phase. By measuring the forward angle light scatter (FALS), an indicator of cell volume, we observed a significant increase in SAR-expressing cells (49.1 ± 1.4%) compared with vector-transfected cells (33.7 ± 0.9%) (P < .05). The effects on cell volume and G₁/S accumulation by SAR were again consistent with the known effects of HU on erythroid cells.

The erythroid transcription factors GATA-1 and GATA-2 are considered to be active modulators of γ-globin gene expression.^7,8  We therefore investigated the GATA-1 and GATA-2 proteins in SAR-expressing K562 cells with Western blots. GATA-2 protein levels, but not GATA-1, were significantly elevated in these cells compared with vector controls (P < .05; Figure 5A). In the same experiment, similar effects on GATA-1 and GATA-2 protein levels were observed in K562 cells after 3 days of HU treatment. Interestingly, p21 mRNA was also elevated 2.4-fold in SAR-expressing cells compared with the vector control (Figure 5B).

Because there was observable growth inhibition in SAR-expressing cells, overexpressing SAR may impact growth-related signals. It has been reported that the inhibition of Ras/ERK enhances erythroid cell differentiation.²⁷  Therefore, we examined mitogen- and stress-activated signaling molecules in the SAR-expressing cells and vector controls. Interestingly, phospho-ERK (pERK) levels, but not total ERK protein, decreased in all 3 SAR-transfected clones compared with vector-transfected cells (Figure 5C). In contrast, neither the phosphorylation state nor the total level of p38 protein was altered in response to SAR expression (data not shown). ERK phosphorylation transiently decreased after 3 hours of exposure to 100 μM HU in K562 cells, but returned to baseline by 24 hours. By using a PowerBlot screen (BD Biosciences Pharmingen, San Jose, CA), we also found that the most significant protein change was an 11-fold reduction of PI3 kinase in SAR-expressing clones compared with vector controls. This result was further confirmed by Western blot (Figure 5D).

Using differential display, we identified the small GTP-binding protein SAR as an HU-inducible gene. The induction of γ-globin mRNA by HU was correlated with an increase of SAR mRNA in human erythroid cells, leading us to hypothesize that SAR is a specific factor involved in γ-globin production after HU treatment. This hypothesis is supported by (1) the induction of SAR mRNA by HU in both HAECs and K562cells; (2) the induction of γ-globin in both K562 cells and CD34⁺ cells expressing SAR; (3) the correlation of SAR and γ-globin mRNA expression in SAR-transfected K562 cells; (4) parallel increases in mRNA levels for both genes in response to 3 γ-globin gene inducers in K562 cells; and (5) parallel induction of both γ-globin and SAR mRNA in AC133⁺ cells cultured with EPO. These observations indicate that the effects of SAR expression in hematopoietic cells not only closely resemble the effects of HU in these cells, but also that the induction of SAR is a specific event in the HU-induced enhancement of γ-globin production.

SAR has long been known to participate in protein trafficking from the endoplasmic reticulum to the Golgi complex.^16,17  The localization of SAR in the endoplasmic reticulum and its association with γ-globin gene expression demonstrated in this study suggest that SAR may also play a special role in hemoglobin regulation. Although the precise pathway(s) and molecules through which SAR regulates the γ-globin gene remain unknown, SAR may increase the transport of membrane-bound transcription factor precursors from the endoplasmic reticulum to the Golgi. In this model, the proteolytic cleavage of the precursor proteins in the Golgi would activate cytosolic fragments that enter the nucleus and regulate erythroid-specific transcription factors, such as GATA, eventually modulating γ-globin gene expression. Alternatively, increased SAR levels in erythroid cells may influence local concentrations and intracellular locations of GTP and cyclic guanosine monophosphate (cGMP). Increased availability of intracellular GTP could affect γ-globin expression either directly or indirectly. It has been reported that guanine, guanosine, and guanine ribonucleotides, but not ATP, CTP, and UTP, induce K562 differentiation, which is associated with the accumulation of γ-globin mRNA.²⁸  GTP also dramatically increases the half-life of γ-globin mRNA.²⁹  Activators of the sGC and protein kinase G (PKG) pathways have been implicated in the regulation of γ-globin gene expression in both erythroleukemic cells and in primary erythroblasts.¹¹  Expression of γ-globin and the sGC alpha subunit are correlated, indicating that GTP-binding proteins may participate in γ-globin induction. The finding of a direct interaction between SAR and erythroid-specific factors will strengthen these speculative models.

Stable transfection of SAR into K562 cells arrested the cells at a premature stage, which is associated with increased synthesis of γ-globin. K562 cells express γ-, ϵ-, and δ-globin but not β-globin, and HU is reported to promote both γ- and β-globin production.^7,30  We therefore transduced the SAR gene into CD34⁺ cells using a retroviral system to investigate its effects on β- and γ-globin production. We found that SAR and HU were similarly able to regulate hemoglobin production in the CD34⁺ stem cells, and SAR induced both γ- and β-globin production. SAR-mediated γ-globin induction was more sustained and profound than the induction of the β-globin gene. This indicates that γ-globin is an important target of SAR and HU during erythroid maturation.

Alterations in the p21, GATA-2, and ERK signal transduction pathways were seen in SAR-expressing cells. The transcription factors GATA-1 and GATA-2 are required for normal hematopoiesis. GATA-1 is essential for survival during primitive and definitive erythropoiesis and terminal maturation. GATA-2 is required for the expansion of multipotential hematopoietic progenitors, but is dispensable for terminal differentiation of erythroid cells and plays a critical role in both the proliferation and survival of early hematopoietic cells.³¹  Enforced expression of GATA-2 in pluripotent hematopoietic cells blocks both their amplification and differentiation.³²  GATA-1 and GATA-2 are important in hemoglobin regulation.⁸  In the regulatory loop between SAR and the γ-globin gene, GATA-1 and GATA-2 can interact to enhance γ-globin promoter activity. It has been shown that a decline in GATA-1 levels alone, without a substantial change in GATA-2, does not affect γ-globin gene expression.³³  Increasing GATA-2 levels increase ϵ-globin and γ-globin transcripts in K562 cells.⁸  In the present study, GATA-2 protein levels increased significantly in SAR-expressing cells without a significant change in GATA-1, consistent with the effect of HU on parental K562 cells. We posit that SAR may disturb erythroid maturation by up-regulating GATA-2, resulting in an increase in γ-globin production.

p21 plays a regulatory role within the G₁/S cell-cycle check-point. Forced GATA-2 expression inhibits cytokine-dependent growth of normal hematopoietic stem/progenitor cells by modulating p21 protein levels.³⁴  In p21-transgenic mice, p21 was found to regulate proliferation and differentiation of early hematopoietic progenitors and, to a lesser extent, of erythroid cells.³⁵  In this study, SAR-expressing cells not only up-regulated GATA-2, but also had a 2- to 3-fold increase in p21 levels. It is likely that p21 induction in SAR-transfected cells exceeded a growth-inhibiting threshold; this would have prevented cells from differentiating into terminal erythroid cells, thereby keeping them at a stage conducive to Hb F production. Our data are consistent with a recent report that GATA-2 inhibited the growth of hematopoietic stem cells as well as other hematopoietic cell lines.³⁴  The authors suggested that overexpression of GATA-2 may alter cytokine signals and shift the binding partners of GATA-2.

Small GTP-binding proteins interact with PI3 kinase binding proteins to facilitate vesicle-mediated vacuolar protein sorting. It has been reported that PI3 kinase regulates erythroid cell differentiation and hemoglobin production,^36,37  and inhibition of PI3 kinase activity could dephosphorylate ERK.³⁸  Enforced SAR expression in K562 cells down-regulated PI3 kinase and dephosphorylated ERK, which may imply that SAR blocks ERK phosphorylation through down-regulation of PI3 kinase. Inhibition of ERK has been shown to inhibit cell growth and gene transcription.³⁹  ERK activity can be elevated in response to proliferative factors, as well as in response to differentiating agents; related signal molecules can show differences depending on the genetic makeup of cells used in each experiment. ERK dephosphorylation might induce hemoglobin synthesis and inhibit cell proliferation. Therefore, our data suggest that HU-induced increases in γ-globin gene mRNA levels may involve the PI3 kinase-ERK signal transduction pathway.

Taken together, our findings suggest that SAR induces γ-globin gene expression in human erythroid cells. HU up-regulates SAR gene expression in erythroid cells, implying that SAR may play a role in γ-globin gene induction and erythroid cell maturation. We have shown that enforced SAR expression alone can replicate the known effects of HU on erythroid cells, including cytotoxicity, G₁/S-phase accumulation, macrocytosis, and γ-globin induction. The induction of SAR mRNA expression by HU also suggests that selective enhancement of membrane trafficking may play a role in erythrocyte maturation and proliferation. SAR is likely to be involved in a variety of normal erythrocyte functions.

This study represents the first evidence that the small GTP-binding protein SAR participates in cell-cycle control through up-regulation of GATA-2 and p21 in erythroid maturation, as depicted in Figure 6. Elucidation of the relationship among SAR's vesicular trafficking functions and γ-globin induction may allow new therapies for sickle-cell disease and β-thalassemia. Further studies will determine whether SAR is an obligatory mediator of these effects and will explore the specific molecular pathways involved. Direct interactions between SAR and associated binding partner(s) should be established to provide the molecular basis of SAR function in hemoglobin protein production.

We gratefully thank Dr Juan Bonifacino and Dr Ivan Ding for valuable discussion and helpful suggestions.