Haploinsufficiency of GATA-2 perturbs adult hematopoietic stem-cell homeostasis

The zinc finger GATA family of transcription factors influences cell fate in multiple tissues through recognition of a GATA consensus sequence in target genes. Among these, GATA-1, -2, and -3 are primarily expressed in blood cells,¹  and gene-targeting experiments have illustrated that each has a salient function in hematopoiesis.^2-5  GATA-1 and -3 are widely acknowledged as lineage-specific transcription factors affecting erythroid cells,^2,3,6  megakaryocytes,^7,8  and eosinophils,^9,10  and T cells,⁴  respectively. In contrast, GATA-2 is predominantly expressed within adult and developing stem-cells^11,12  and mast cells¹³  and is a pivotal regulator of the stem-cell compartment during ontogeny.⁵  Mice engineered to be deficient in GATA-2 succumbed during gestation at embryonic day 10 to 11 (day e10-11) with severe hematopoietic defects. GATA-2 null/wild-type chimeric mice generated to define the function of GATA-2 in definitive hematopoiesis showed an absence of null myeloid cell output and Rag-2^–/– adult chimeras produced only very low levels of null lymphoid cells. Additionally, in vitro culture of null embryonic stem (ES) cells and yolk-sac cells from E9.5 null embryos revealed pervasive defects in definitive hematopoiesis, including altered responsiveness to stem-cell factor (SCF) and decreased cell survival.^5,14  Taken together, these findings imply a central function for GATA-2 in the proliferation and survival of the developing stem-cell pool.

Enforced expression experiments have been used to address the impact of GATA-2 level in adult stem-cells and has primarily been achieved using 2 systems: (1) GATA-2 internal ribosomal entry site (IRES)–murine stem-cell virus (MSCV) retrovirus,¹⁵  and (2) estrogen receptor inducible GATA-2/ER vector.^16,17  The GATA-2/MSCV retroviral system demonstrated that a 2-fold increase in GATA-2 level in murine primitive cells (Sca1⁺Lin^–) was sufficient to reduce their number by 40-fold and cause their failure to differentiate appropriately in vivo. Similar levels of enforced expression were achieved using GATA-2/ER into hematopoietic cell lines and primary adult hematopoietic cells and demonstrated that elevated levels of GATA-2 restricted primitive cell entry into the cycle.¹⁶  This change in cell cycling was associated with increasing levels of p21 and p27,¹⁷  cell-cycle inhibitors that have been shown to control the proliferation of the stem and progenitor cell pools, respectively.^18,19  Although GATA-2/ER may not entirely reflect authentic GATA-2 effects,²⁰  these experiments collectively suggest that the level of GATA-2 within a cell may be a critical determinant of cell function. Enforced expression, however, is a controversial model for evaluating the physiologic effects of a gene product.

In this study, we have analyzed hematopoiesis in GATA-2^+/– and GATA-2^+/+ adult animals. GATA-2^+/– mice survive through adulthood without overt abnormalities and provide a tractable system to assess the impact of GATA-2 level on adult hematopoiesis. We showed a diminution in Lin^–c-kit⁺Sca-1⁺CD34^– (LK-SCD34^–) frequency from GATA-2^+/– marrow. Furthermore, GATA-2^+/– marrow exhibited a reduced capacity to engraft an irradiated host and decreased week-5 cobblestone area–forming cell (CAFC) units, suggesting a decreased functional stem-cell pool in the GATA-2^+/– animal. This was accompanied by changes in primitive cell apoptosis and cell cycling, but not the relative ability of primitive cells to form progenitors or to self-renew.

We generated heterozygote 129/B6 GATA-2 mice as previously described,⁵  and maintained them under conditions ratified by the Subcommittee on Research Animal Care of the Massachusetts General Hospital (MGH) and the United Kingdom Home Office. The heterozygote mice (+/–) were backcrossed for 6 generations into the 129/SV strain and bred into a wild-type 129/SV background to produce heterozygote and wild-type offspring. Genotyping was achieved by isolating genomic DNA from tail biopsy and DNA polymerase chain reaction (PCR) as previously described.⁵  Gender-matched littermates were used in experiments.

Mouse bone marrow was obtained from 6- to 12-week-old animals. Marrow cells were obtained by flushing femurs and tibias with Iscove modified Dulbecco medium (IMDM) supplemented with 5 IU/mL penicillin, 5 μg/mL streptomycin, 2 mM l-glutamine, and 10% fetal calf serum (Mediatech, Herndon, VA). The marrow cell suspension was filtered with 30micron nylon mesh (Sefar America, Kansas City, MO), and stored on ice until use. A 10-μL aliquot of marrow cell suspension was diluted 20-fold with phosphate-buffered saline (PBS) solution (Mediatech) containing 3% acetic acid. After 3 minutes, the sample was vigorously mixed, loaded onto a hemocytometer, and the nucleated cells counted.

RNA was prepared from bone marrow Sca-1⁺c-kit⁺Lin^– cells using the Picopure Kit (Arcturus Bioscience, Mountain View, CA). The amount of RNA was quantified using RNA-specific fluorescent dye Ribogreen (Molecular Probes, Eugene, OR). cDNA was synthesized using the Promega ImProm II Reverse Transcription System (Promega, Madison, WI). Approximately 100 ng RNA was added to a reaction containing 1x ImProm-II reaction buffer, 0.5 mM deoxynucleoside triphosphate (dNTP), 0.5 μg random primers, 20 U recombinant RNasin ribonuclease inhibitor, and 1 μL ImProm-II reverse transcriptase. Reverse transcription was performed as follows: 5 minutes annealing at 25°C followed by incubation for 1 hour at 40°C, inactivation at 70°C for 15 minutes. The cDNA was used as a template for real-time PCR. The PCR reaction contained 1x SYBR green reaction mix (Applied Biosystems), 0.5 μM GATA-2 forward primer (GCACCTGTTGTGCAAATTGT) and GATA-2 reverse primer (GCCCCTTTCTTGCTCTTCTT) and 2 ng template in a 20 μL reaction volume. Amplification was performed on an ABI 7700 with incubation times of 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 to 50 cycles of 95°C for 15 seconds, and 60°C for 1 minute. The mRNA content was measured relative to that of rodent glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (TaqMan Rodent GAPDH Control Reagents; Applied Biosystems).

Blood was sampled from a lateral tail vein and collected in a Microtainer Brand Tube with EDTA (ethylenediaminetetraacetic acid; Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ). Samples were analyzed on a Hemavet 850 (Drew Scientific, Oxford, CT).

This assay was used to measure the progenitor cell frequency as previously described,^18,19  but with the following modifications. M3434 (Stem Cell Technologies, Vancouver, BC, Canada) was used to enumerate total colony-forming cells (CFCs). Granulocyte macrophage–CFCs (CFC-GMs), erythroid burst-forming units (BFU-Es), and pre–B-cell CFCs were scored after culture in M3534, M3334, and M3630 (Stem Cell Technologies), respectively. All colony assays were established in 6-well plates at cell densities recommended by the manufacturer.

To quantify primitive cells in vitro, we adapted the week-5 cobblestone area–forming cell (CAFC) and week-6 long-term culture-initiating cell (LTC-IC) assay as previously described.^18,19  Briefly, stromal layers were prepared from the marrow of 129/SV mice and cultured at 33°C, 5% CO₂ in 25 cm² flasks in M5300 long-term culture medium (alpha-MEM with 12.5% horse serum, 12.5% fetal bovine serum, 0.2 mM I-inositol, 20 mM folic acid, 10^–4 M 2-mercaptoethanol, 2 mM l-glutamine, and 10^–6 M hydrocortisone; Stem Cell Technologies). After 2 to 3 weeks, confluent layers were trypsinized, irradiated at 15 Gy, and subcultured in 96-well flat-bottomed plates at a density of 2 × 10⁴ cells per well. After 1 to 7 days, cultures were seeded at 2-fold dilutions (10⁵-195 per well) of nucleated marrow cells from each genotype. CAFCs were scored at weeks 2, 3, and 5. For the LTC-IC assay, cultures were established as described for the CAFC assay. At week 5, M5300 medium was removed from cultures and replaced with M3434 (Stem Cell Technologies). Cultures were maintained at 37°C, 5% CO₂ and LTC-ICs were analyzed at week 6.

To test hematopoietic cell cycling in vivo, a single 150 mg/kg dose 5-FU (Pharmacia and Upjohn, Peapack, MI) was administered intraperitoneally; mice were killed 24 hours later and nucleated marrow cells were prepared for long-term culture at limiting dilution.

Flow cytometry was used to quantify the cell-cycle status in the stem-cell compartment. Bone marrow nucleated cells were labeled with antilineage biotin antibodies (CD3, CD4, CD8, B220, Gr-1, Mac-1 [Caltag, Burlingame, CA]; TER-119 [Pharmingen, San Diego, CA]) and stem-cell marker (Sca-1, c-kit; Pharmingen). An enriched stem-cell phenotype (Lin^–c-kit⁺Sca-⁺⁺) was gated and a DNA dye, Hoechst 33 342 (Molecular Probes) was used to stain the antibody-bound cells simultaneously to measure the cycling cell percent in the populations. To measure stem-cell quiescence within a Lin^–c-kit⁺Sca-1⁺ gate, Pyronin Y (Polysciences, Warrington, PA) and Hst dyes were used as described previously.^18,19,21  To detect apoptotic cells, annexin-V (Caltag) in conjunction with the DNA dye, 4,6-diamidino-2-phenylindole (DAPI) (Molecular Probes), was used to stain Lin^–c-kit⁺Sca-1⁺ cells, which was subsequently analyzed by flow cytometry. Cells excluding DAPI and binding annexin-V were considered apoptotic. To examine intracellular Bcl-xL expression within the stem-cell compartment, bone marrow cells were stained for Sca-1, c-kit, and lineage antibodies, treated with a Cytofix/Cytoperm kit (Pharmingen) according to manufacturer's instructions, and subsequently stained with Bcl-xl antibody (Santa Cruz Biotechnology, Santa Cruz, CA). For cell sorting experiments of Lin^–c-kit⁺Sca-1⁺ and Lin^–c-kit⁺Sca-1⁺CD34^– cell populations, marrow cells were labeled with lineage antibodies as described and tagged with antibiotin microbeads (Miltenyi, Surrey, United Kingdom) before depletion on an AutoMacs column (Miltenyi). Lineage-depleted cells were stained with strepavidin–phycoerythrin (PE)–Cy5 (Pharmingen) for remaining lineage-positive cells, Sca-1 and c-kit (Lin^–c-kit⁺Sca-1⁺ sorts) or strepavidin-PE-Cy5, Sca-1, c-kit, and CD34 (Pharmingen) (Lin^-c-kit⁺Sca-1⁺CD34^-sorts) and analyzed/sorted by flow cytometry.

Serial bone marrow transplantation was used to evaluate the ability of stem-cells to self-renew as described previously.^18,19 

Three 6- to 8-week-old GATA-2^+/– and GATA-2^+/+ mice were killed. Two million nucleated bone marrow cells from each genotype were admixed in a 1:1 ratio and transplanted into lethally irradiated 129/SV recipients. After 7 weeks, transplant recipients were killed and bone marrow was assayed in CFC and LTC-IC assays as described previously.¹⁹  The marrow of recipients was similarly evaluated in CFC assays after 4 and 8 months. After 11 months, peripheral blood obtained from recipients was sorted for granulocytes (Gr-1⁺), monocytes (Mac-1⁺), B lymphocytes (B-220⁺), and T lymphocytes (CD3⁺), and quantitative PCR was performed. Briefly, DNA was prepared using the QIAamp DNA Blood Kit (Qiagen, Valencia, CA). The amount of DNA was quantified and standardized using DNA-specific fluorescent dye Picogreen (Molecular Probes). The PCR reaction contained 1x SYBR green reaction mix (Stratagene, La Jolla, CA), 0.5 μM neomycin (NEO) forward primer (AGACAATCGGCTGCTCTGAT) and 0.5 μM of NEO reverse primer (CTCGTCCTGCAGTTCATTCA) and 10 ng of template in a 25 μL reaction volume. Amplification was performed on a Mx3000P real-time PCR system (Stratagene) with incubation times of 95°C for 10 minutes, followed by 35 cycles of 95°C for 30 seconds, 63°C for 1 minute, and 72°C for 1 minute.

All experiments were analyzed using a 2-tailed Student t test unless otherwise stated. Graphical data were expressed as mean plus or minus standard error of the mean (SEM).

To evaluate whether GATA-2^+/– mice may provide a model to study the effect of GATA-2 level on adult bone marrow hematopoiesis, we measured GATA-2 mRNA levels by real-time reverse transcriptase PCR in whole bone marrow from GATA-2^+/– and GATA-2^+/+ animals. GATA-2 expression was only reduced by approximately 20% in the GATA-2^+/– genotype in whole marrow (Figure 1A). However, in the more stem-cell–enriched Lin^–ckit⁺Sca-1⁺ (LKS) population, where GATA-2 is predominantly expressed, GATA-2 mRNA from GATA-2^+/– animals was decreased to approximately 50% of the wild-type value (Figure 1B). The GATA-2^+/– animal is therefore a valid model to assess reduced GATA-2 dose in adult hematopoiesis.

The hematopoietic potential of GATA-2^+/– and GATA-2^+/+ mice was examined under homeostatic conditions. Mature blood populations and overall bone marrow cellularity were comparable between the 2 genotypes (Table 1). Using the CFC assay to quantitate marrow progenitor populations, a decrease in total CFCs was observed in the GATA-2^+/– mouse (Figure S1A, available on the Blood website; see the Supplemental Figures link at the top of the online article). By analyzing specific classes of lineage-committed progenitor cell in GATA-2^+/– marrow, we noted a decrease of a similar magnitude in CFC-GMs (Figure S1B), but not in BFU-Es (Figure S1C) or CFC-Bs (data not shown), suggesting that the level of GATA-2 selectively affects GM colony formation.

To evaluate if hematopoietic defects were evident in GATA-2^+/– primitive cells, we performed limiting dilution CAFC assays²²  and enumerated CAFCs appearing at 2, 3, or 5 weeks after plating. CAFC production from GATA-2^+/– marrow was decreased at these respective time points (Figure 2A-C). Particularly noteworthy was the observed decrease in GATA-2^+/– week-5 CAFCs; at this time point, the CAFC assay is thought to imperfectly correlate with stem-cell function.²²  Further evaluation of primitive cells was performed by immunophenotyping. The CD34^– subset of LKS cells (LKSCD34^–) has been shown to be highly enriched for stem-cells and isolates a population capable of repopulating 21% of lethally irradiated hosts at the single cell level.²³  Flow cytometry of bone marrow from GATA-2^+/– animals revealed a lower frequency of LKSCD34^– cells compared with wild-type controls (Figure 2D-E). Week-5 CAFC analysis of LKSCD34^– demonstrated the reduced functional capacity of GATA-2^+/– stem-cells (Figure 2F).

To more stringently evaluate stem-cell activity in vivo, competitive transplantation of a 1:1 admixture of GATA-2^+/– to GATA-2^+/+ nucleated cells into wild-type recipients was performed.²⁴  To specifically examine the stem and progenitor cell contribution from the 2 genotypes in transplant recipients, bone marrow from individual recipients was analyzed in the LTC-IC and CFC assays, respectively. Single colonies emerging from each of these cultures were microisolated and underwent genotyping PCR. At 7 weeks, the proportion of GATA-2^+/– stem-cells (LTC-ICs) and progenitor cells (CFCs) in recipients was lower than that of the GATA-2^+/+ genotype (Figure 3B). Given the similar representation of GATA-2^+/– stem-cells and GATA-2^+/– progenitor cells in recipients at 7 weeks (Figure 3B), GATA-2^+/– stem-cell output was analyzed by genotyping PCR of progenitor cells at longer time intervals. At 4 and 8 months after transplantation, pronounced reductions in GATA-2^+/– progenitor cell contribution were noted, suggesting that the GATA-2^+/– stem-cell compartment was impaired in competitive repopulating ability (Figure 3C). At 11 months after transplantation, peripheral blood chimerism in granulocytes, monocytes, T lymphocytes, and B lymphocytes was assessed using quantitative DNA PCR. In peripheral blood pooled from 4 animals that underwent transplantation, the contribution of the GATA-2^+/– genotype to monocytes (Mac-1⁺), T cells (CD3⁺), and B cells (B-220⁺) was reduced compared to the GATA-2^+/+ genotype (monocytes: GATA-2^+/–, 27%; GATA-2^+/+, 73%; T cells: GATA-2^+/–, 39%; GATA-2^+/+, 61%; B cells: GATA-2^+/–, 36%; GATA-2^+/+, 64%). Unexpectedly, the representation of GATA-2^+/– and GATA-2^+/+ genotype in granulocytes (Gr-1⁺) was similar (granulocytes: GATA-2^+/–, 48%; GATA-2^+/+, 52%). Blood cell counts from transplant recipients were normal (data not shown). Y chromosome analyses and genotyping PCR of colonies demonstrated 100% donor cell engraftment (data not shown). Taken together, these data indicate GATA-2^+/– marrow stem-cells are compromised in their ability to contribute to select mature cell lineages in competitive transplantation.

Postulating that the observed GATA-2^+/– stem-cell phenotype may be due to the reduced capacity of primitive cells to proliferate, we assessed the cell-cycle status of an enriched stem-cell pool from the 2 genotypes using the DNA dye Hoechst 33342 (Hst).²⁵  The proportion of cells in S/G₂/M within LKS cells was unchanged between GATA-2^+/– and GATA-2^+/+ genotypes (Figure 4A). We next used the RNA dye Pyronin Y (PY) to measure cell dormancy within Hst^low cells²¹  and noted a significant increase in the PY^low Hst^low staining LKS cells in the GATA-2^+/– genotype (Figure 4B-C). The mitochondrial dye rhodamine (Rho) was used in conjunction with Hst, as a further assessment of primitive cell quiescence.²⁶  In agreement with the results of PY/Hst analysis, an increased proportion of GATA-2^+/– LKS Rho^low Hst^low cells was observed (data not shown). These data indicate a higher fraction of quiescent cells within the primitive cell compartment of GATA-2^+/– mice.

To define whether this functionally corresponded to alterations in in vivo cell cycling, we challenged GATA-2^+/– and GATA-2^+/+ mice with the antimetabolite 5-fluorouracil (5-FU), which selectively depletes cycling hematopoietic cells due to cell-cycle–dependent killing.^27,28  Marrow was harvested from GATA-2^+/– and GATA-2^+/+ animals 1 day after 5-FU injection and limiting dilution CAFC assay was performed. After 5-FU challenge, a substantially greater proportion of week-5 CAFCs was scored in the GATA-2^+/– group (Figure 4D). These data provide functional validation that GATA-2^+/– primitive cells cycle less actively than wild-type controls.

The observed increase in quiescence could render GATA-2^+/– primitive cells insensitive to cytokine stimulation, thereby altering their capacity to differentiate into progenitor cells. To address this possibility, we sorted LKS cells from GATA-2^+/– and GATA-2^+/+ marrow, plated them into methylcellulose medium supplemented with prodifferentiative cytokines, and scored CFC and CFC-GM frequency. There was relative parity in total CFC and GM colony frequency between each genotype (data not shown), suggesting that GATA-2^+/– primitive cells retain the ability to differentiate to progenitor cells.

Next we assessed whether the smaller GATA-2^+/– primitive cell pool may also be caused by augmented cell death and evaluated apoptosis on LKS cells using the annexin-V assay.²⁹  The LKS compartment of GATA-2^+/– freshly isolated marrow had a consistently higher proportion of cells in the annexin high/DAPI low population compared with GATA-2^+/+ marrow, indicating that GATA-2^+/– primitive cells had an increased propensity to undergo apoptosis (Figure 5A). To explore potential molecular mechanisms underlying enhanced apoptosis in GATA-2^+/– primitive cells, we surveyed the expression of molecules associated with apoptosis. Of the molecules examined, intracellular flow cytometric analyses revealed attenuated expression of the antiapoptotic protein Bcl-xL in GATA-2^+/– LKS cells (Figure 5B-C). Together, these data imply that altered cell survival contributes to the reduced size of the GATA-2^+/– functional stem-cell pool and suggest that the increased predisposition of GATA-2^+/– primitive cells to undergo apoptosis may be related to decreased Bcl-xL expression.

To assess if functional changes within the heterozygous stem-cell compartment cause alteration of self-renewal potential in vivo, we performed serial transplantation experiments.^30,31  Minimal differences were observed in the survival of animals from both genotypes after 4 sequential transplantations. After the fifth transplantation, however, differences were noted in mortality between the 2 genotypes (Figure 6A). This result could be due to either altered stem-cell self-renewal in GATA-2^+/– mice or the paucity of stem-cell number initially transplanted from GATA-2^+/– mice. To discern between these 2 possibilities, limiting dilution week-5 CAFC assays were performed. Despite a lower abundance of stem-cells transplanted initially (BMT 0) in the heterozygous group, the 2 genotypes display similar kinetics of week-5 CAFCs during the course of sequential transplantations (Figure 6B). Y-chromosome analysis of recipient marrow at the fourth serial transplantation demonstrated the presence of donor genotype (data not shown). Furthermore, low cell dose transplantation (2 × 10⁵ cells/recipient) in the fourth serial transplantation resulted in uniform fatality in both genotypic groups, arguing against significant host cell contribution to hematopoiesis (data not shown). From these data we infer that the self-renewal capacity of heterozygous stem-cells remains intact.

In this report we used GATA-2^+/– mice to examine the effect of GATA-2 level on the stem-cell compartment and to broadly probe the function of GATA-2 in adult mammalian hematopoiesis. Our data show that heterozygosity of GATA-2 has an impact on stem and progenitor cell homeostasis. The most notable phenotype was observed within the primitive cell compartment, where LKSCD34^– frequency was reduced, and week-5 CAFC assays and competitive transplantation revealed attenuation in the size of the GATA-2^+/– functional stem-cell pool. Of particular note, there was a striking reduction in GATA-2^+/– genotype contribution to the marrow of long-term recipients of competitively transplantations. These data jointly support the conclusion that adult stem-cells are compromised in the setting of reduced GATA-2 level.

Less marked reductions were also noted in GATA-2^+/– contribution to monocytes, T cells, and B cells in the peripheral blood of transplant recipients. Unexpectedly, GATA-2^+/– representation to the mature granulocytic lineage was comparable to that of the GATA-2^+/+ genotype. Why GATA-2 expression in stem-cells should affect distinct mature cell lineages in the setting of transplantation is unclear but such an effect is consistent with studies demonstrating that manipulating GATA-2 level in stem-cells alters outgrowth of their lineage-specific progeny in transplantation.^15,32 

Progenitor cells from GATA-2^+/– animals, particularly those yielding CFC-GMs, were also decreased, an observation that parallels previous results suggesting that GATA-2 affects the GM pathway.¹⁶  Yet the capacity of primitive cells to form CFC-GMs was unimpaired, raising the possibility that the GATA-2^+/– genotype may perturb the maintenance and propagation of GM-committed progenitors (N.P.R., P.V., and T.E., unpublished observations, 2004). The normal abundance of granulocytes and monocytes in the peripheral blood of GATA-2^+/– animals, however, implies the presence of GATA-2 independent compensatory mechanisms in vivo.

Decreased primitive cell numbers in GATA-2^+/– mice was associated with evidence for several potentially contributing mechanisms. Cell cycling of the primitive cell compartment was reduced in the GATA-2^+/– animals. Although other studies have suggested that GATA-2 may restrict stem-cell entry into the cycle,^14-17  our experiments define that GATA-2 regulates primitive cell quiescence. It was, however, unanticipated that GATA-2^+/– mice would have a larger fraction of quiescent primitive cells. Higher levels of GATA-2 cause failure of stem-cell proliferation and the relatively high levels of GATA-2 observed in wild-type bone marrow have been hypothesized to sustain stem-cell dormancy.¹⁵  It was therefore expected that lower levels of GATA-2 would abate primitive cell quiescence. Yet we observed that the primitive cell compartment of GATA-2^+/– marrow was less proliferative than that of GATA-2^+/+ marrow. Cell cycling therefore appears to be sensitive to the level of GATA-2 in either direction. The basis for such effects in GATA-2^+/– stem-cells were not addressed here, but may involve modulation of p21 and p27, cell-cycle inhibitors that have been shown to be responsive to alterations in GATA-2 level.¹⁷ 

In addition to altering the prevalence of actively cycling cells, the GATA-2^+/– primitive cell pool was more prone to apoptosis than its wild-type counterpart. Furthermore, depressed cell survival from GATA-2^+/– stem-cells was marked by alteration in the antiapoptotic gene Bcl-xL expression. Parenthetically, we have noted activation of Bcl-xL after enforced expression of GATA-2 in Paterson factor-dependent cell (FDCP)–mix cells (T.E., unpublished observations, 2004). Relatively high levels of Bcl-xL have been observed within the primitive cell pool,³³  yet the function of Bcl-xL in regulating this compartment is unclear. Our data suggest that GATA-2 and Bcl-xL act in concert to modulate stem-cell pool size. It is possible that GATA-2 directly affects Bcl-xL expression as GATA sites are present on the promoter regions of Bcl-xL; however, we did not assess binding of GATA-2 to the Bcl-xL promoter in this study.³⁴  Alternatively, GATA-2 may secondarily affect Bcl-xL levels through effects on other gene products. Whether the changes in cell cycling and apoptosis are related is at present unclear, but a number of other studies have indicated a link between these processes in other cell types.^35-38  For example, Bcl-2 family members have been implicated in both apoptotic and cell-cycle regulation.^39-42  Future studies will be of interest to determine whether the effects of GATA-2 on quiescence and apoptosis are based on a shared biochemical mechanism.

Although select features of GATA-2^+/– stem-cell pool homeostasis were altered, no difference was noted in self-renewal potential between the 2 genotypes after serial transplantation. Diminished survival of GATA-2^+/– transplant recipients was apparent after 5 serial transplantations. However, we interpret this as reflective of the lower fraction of GATA-2^+/– functional stem-cells transplanted initially. The fact that the change in survival in GATA-2^+/– group was observed after 5 transplantations, and not at an earlier stage, further suggests that the GATA-2^+/– stem-cell self-renewal remains largely intact during transplantation.

The data presented here were obtained within the context of a single deleted GATA-2 allele. It is largely presumed that in diploid organisms genetic information is expressed from both alleles and that if loss of a single allele were to arise, there would be compensation from the intact allele. However, single allele mutations have previously been shown to induce haploinsufficiency syndromes in both mouse models and human disease.⁴³  Although it is not known how this occurs in the GATA-2 heterozygote mouse, various mechanisms have been proposed to account for the occurrence of haploinsufficiency. Mathematical analyses of gene expression have suggested that haploinsufficiency may relate to stochastic gene initiation and regulation.⁴⁴  Alternatively, it is possible that GATA-2 may be subject to monoallelic regulation, a controversial phenomenon documented for Toll-like receptor-4.⁴⁵ 

How altered expression of a single transcription factor itself can affect cell phenotype is not entirely clear, though precedent for this is well established with other transcription factors in hematopoiesis.^46-48  Possible mechanisms include altered stochiometry of interactions with associated transcription factors, competition for target sites, and/or multiprotein complex production/stability. Defining GATA-2 partner proteins and target genes in stem-cells should afford insights into the process.

In addition to the role ascribed to GATA-2 during hematopoietic stem-cell development,^5,14  the data here implicate GATA-2 as a critical regulator of the adult stem-cell pool. While other transcription factors such as SCL and AML-1 have been shown to be essential for hematopoietic specification during ontogeny, they appear to be expendable for maintenance of the adult stem-cell compartment.^49-53  In marked contrast, GATA-2 has both a critical role in hematopoietic stem-cell development and the maintenance of normal adult stem-cell homeostasis.

The authors wish to thank Drs Louise Purton, Heather Fleming (Scadden laboratory), Boris Guyot (Vyas laboratory), Gill May, Alex Tipping (Enver laboratory), and Yuko Fujiwara (Orkin laboratory) for helpful discussions and technical assistance.