GATA-6 Induces p21Cip1 Expression and G1Cell Cycle Arrest*

GATA transcription factors represent a family of highly conserved zinc finger proteins with tissue-specific expression patterns. Previous studies have shown that GATA-6 is expressed in vascular smooth muscle cells (VSMCs) and rapidly down-regulated when VSMCs are induced to proliferate. Here we investigated whether the GATA-6 transcription factor can modulate cellular proliferation. Transient transfection with a GATA-6 expression vector inhibited S-phase entry in VSMCs and in mouse embryonic fibroblasts (MEFs) lacking both p53 alleles. The GATA-6-induced growth arrest correlated with a marked increase in the expression of the general cyclin-dependent kinase (Cdk) inhibitor p21. In contrast to p53-deficient MEFs and VSMCs, MEFs null for both p21 alleles were refractory to the GATA-6-induced growth inhibition. These data demonstrate that elevated GATA-6 expression can promote the quiescent phenotype in VSMCs.

The GATA transcription factors are required for establishing hematopoietic cell lineages and for the proper formation of the cardiovascular system. GATA-1/2/3 isoforms are predominantly expressed in hematopoietic cells, and disruption of each of these genes results in deficiencies in hematopoiesis and embryonic lethality (1). In contrast, the more recently identified GATA-4/5/6 genes represent a subfamily of factors that are expressed in cardiovascular tissues (2,3). Embryos that lack GATA-4 fail to develop a centralized heart tube and foregut (4,5), and antisense oligonucleotides to GATA-4 inhibit the differentiation of P19 embryonic carcinoma cells into beating cardiac myocytes (6). Little is known about the functions of GATA-5 and -6, but their expression patterns overlap with that of GATA-4 in the developing cardiovasculature. Both GATA-4 and -6 are expressed in heart throughout development, while GATA-5 is expressed during early heart development (7,8). Of note, GATA-4 and -6 are co-expressed in the vasculature during early development, but only GATA-6 is detected in vessels post-day 13.5 of mouse embryonic development (7)(8)(9). Moreover, GATA-6 mRNA expression is also detected in cultured rat and human VSMCs 1 (7,10).
Unlike cardiac and skeletal muscle cells, VSMCs are not terminally differentiated and can dedifferentiate and reenter the cell cycle in response to multiple stimuli (11,12). The proliferation of VSMCs contributes to neointimal lesion formation in response to chronic or acute vessel wall injury (13,14). Following acute injury, the quiescent VSMCs of the vessel wall up-regulate cyclin and Cdk expression and rapidly re-enter cell cycle in a synchronous manner (15). As lesion size increases the proliferative capacity of VSMCs decreases, and this correlates with an up-regulation of the Cdk inhibitors p21 Cip1 and p27 Kip1 (16 -18).
In cultured VSMCs, GATA-6 expression is rapidly and transiently down-regulated following mitogen stimulation, suggesting that GATA-6 might have a role maintaining VSMCs in a quiescent state (10). In this study, we analyzed the cell cycle regulatory properties of the human GATA-6 factor. The forced expression of GATA-6 inhibited S-phase entry in VSMCs and non-muscle cells. GATA-6 overexpression resulted in the induction of the Cdk inhibitor p21, and the cell cycle inhibitory activity of GATA-6 was markedly attenuated in p21-deficient cells. These data suggest that GATA-6 may function in VSMCs to coordinate cell cycle activity with the state of cellular differentiation.

EXPERIMENTAL PROCEDURES
Cell Culture-Cells were incubated at 37°C in Dulbecco's modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS) and penicillin/streptomycin. Primary cultures of rat smooth muscle cells were prepared from thoracic aortas of adult male Sprague-Dawley rats according to Mader et al. (19). COS1 cells were cultured in DMEM with 10% FBS. Mouse embryonic fibroblasts with a homozygous deletion of both p53 alleles designated p53Ϫ/Ϫ MEFs were a generous gift from Dr. Arnold Levine (20). Both p21 knockout mice and passage 28 mouse embryonic fibroblasts that contain a homozygous disruption of both p21 alleles (p21Ϫ/Ϫ MEFs) were generous gifts from Dr. Philip Leder (21). Early passage embryonic fibroblasts derived from p21 knockout mice (21) were prepared from day 13.5 embryos as described previously (22). p53Ϫ/Ϫ MEFs were induced to the quiescent state by serum starvation for 3 days in 0.5% FBS/DMEM, while early and late passage p21Ϫ/Ϫ MEFs cells were serum-starved in 0.2% FBS/DMEM for 4 days.
15 g of each construct was transiently transfected into COS1 cells by the calcium phosphate method. The cells were harvested 48 h after transfection and whole cell extracts were prepared.
The p21 cDNA (23) was subcloned into pCDNA/Amp expression vector and was used as a positive control for the cell cycle inhibition experiments. pCMV-␤-galactosidase was purchased from CLONTECH.
The rat Gax open reading frame was obtained by polymerase chain reaction amplification of a plasmid derived from a ZAP cDNA clone as described previously (24). This restriction fragment was inserted into the pCGN vector that contains the N-terminal epitope of the influenza virus hemagglutinin protein (25). Adenoviral infections in culture (24) and the adenoviral constructs Ad-␤-Gal (26) and Ad-p21 (27,28) have been described previously.
Electrophoretic Mobility Shift Assay-Whole-cell extracts were prepared from COS1 cells. In brief, cells were washed twice in PBS, removed from culture dishes by scraping, and collected by centrifugation. The pellet was resuspended in an equal volume of 2 ϫ lysis buffer (20 mM HEPES-KOH (pH 7.8), 0.6 M KCl, 1 mM dithiothreitol, 20% glycerol, 2 mM EDTA, 2 g/ml leupeptin) and subjected to three cycles of freezing and thawing. After centrifugation at 16,000 ϫ g for 10 min at 4°C, protein concentration was measured by the Bradford method according to the direction of the manufacturer (Bio-Rad). Electrophorectic mobility shift assays were carried out in reaction mixtures containing 5-20 g of extract, 20 fmol of probe, 1 g of poly(dL-dC), and when indicated 200 ng of single-stranded oligonucleotide as nonspecific DNA competitors. Electrophoresis was carried out on 5% nondenaturing polyacrylamide gels with 0.5 ϫ TBE (45 mM Tris, 45 mM boric acid, 1 mM EDTA) in a circulating water-cooled gel box. A probe to the GATA site was prepared from the mouse globin gene. Probes, and competitor DNAs were double-stranded synthetic oligonucleotides, and the codingstrand sequences are as follows: GATA, 5Ј-GCCGGGCAACTGATAAG-GATTCCCA-3Ј; GATA mutant, 5Ј-GACCGGGCAACTGcgAAGGATTC-CCA-3Ј.
Nucleotides in lowercase letters designate the mutated nucleotides. One pmol of each probe was labeled at the 5Ј terminus with [␥-32 P]ATP and used at a concentration of 10 fmol/l.
Antibodies-Three peptide sequences were synthesized based on the human GATA-6 sequence (amino acids 8 -22, 373, 387, and 428 -442) and used to raise antibodies. These peptides were conjugated to KLH and injected into New Zealand White rabbits subcutaneously. The immunoglobin fraction of the antisera was purified with protein A-Sepharose CL-4B according to the instructions of the manufacturer (Amersham Pharmacia Biotech).
Western Blot Analysis-Whole-cell extracts prepared from COS1 cells (50 g) were subjected to a 10% SDS-polyacrylamide gel electrophoresis, and transferred to Immobilon-P (Millipore) by semidry blotting. Filters were blocked for 1 h at room temperature in PBS, 0.2% Tween 20, 5% nonfat dry milk. The filters were then incubated with anti-human GATA-6 antibody (1 g of immunoglobin/ml) overnight at 4°C in PBS, 0.2% Tween 20, 2% non-fat dry milk. Filters were washed in PBS, 0.2% Tween 20, 2% non-fat dry milk and incubated with anti-rabbit antibody conjugated to horseradish peroxidase (Amersham Pharmacia Biotech). Visualization of the immunocomplexes was performed as recommended by the manufacturer (Enhanced Chemiluminescence kit; Amersham Pharmacia Biotech).
Growth Arrest Assays-Cultured rat VSMCs (A7r5) were seeded in six-well (20-mm) dishes for the assay. 2.5 g of pMSV-␤-galactosidase, which expresses the bacterial lacZ gene under transcriptional control of the murine sarcoma virus long terminal repeat and either GATA-6, ⌬ZF-GATA-6, Gax, or p21 expression plasmids at a ratio of 1:5 (w/w) were mixed, and 5 g/well of this mixture was used for transfection by the calcium phosphate method. After 24 h, the cells were washed three times with DMEM and incubated with 20 M 5-bromo-2-deoxyuridine (BrdUrd) for another 24 h in high serum medium. The cultures were then washed twice with ice-cold Dulbecco's phosphate-buffered saline (DPBS) and fixed for 5 min on ice with 2% formaldehyde, 0.2% glutaraldehyde in DPBS.
Fixed cells were washed three times with ice-cold DPBS and stained in 0.5 ml of solution containing 1 mM Red-Gal (6-choloro-3-indolyl-␤-Dgalactopyranoside, Research Organics), 10 mM sodium phosphate (pH 7.0), 1 mM MgCl 2 , 150 mM NaCl, 3.3 mM K 4 Fe(CN) 6 , 3.3 mM K 3 Fe(CN) 6, and 0.1% Triton X-100. Cells were stained in a humid chamber at 37°C until a light red color appeared as observed under a microscope with bright-field illumination. Following destaining in 3% dimethyl sulfoxide (MeSO 4 ) in DPBS for 30 s, the wells were washed twice with ice-cold DPBS. To stain for BrdUrd, the cells were permeabilized with 1 ml/well acetone/methanol (1:1) for 3 min on ice, rehydrated in 3 ml of DPBS for 5 min, and washed twice with ice-cold DPBS. Genomic DNA was denatured by incubating the cells with 2 N HCl for 1 h at 37°C in a humid chamber followed by neutralization with 0.1 M sodium tetraborate (pH 8.5) and equilibrated with DPBS. After blocking in 6 g of IgG/ml of normal goat serum for 20 min, anti-BrdUrd antibody (Boehringer Mannheim) was added for 1 h. The anti-BrdUrd IgG was detected using an ABC kit as directed by the manufacturer (Calbiochem-Novabiochem). The monolayer cells were mounted under 25-mm round glass coverslips with a glycerol-based mounting media (Kirkegaard & Perry Laboratories Inc.). Cells were visualized under bright-field microscopy (Diaphot; Nikon Inc.) with a neutral filter. Red cells expressing ␤-galactosidase with dark nuclei (BrdUrd-positive) and light nuclei (BrdUrd-negative) were counted.
In an alternate assay, VSMCs, p53Ϫ/Ϫ and p21Ϫ/Ϫ MEFs were cultured in 60-mm plates containing gelatin-coated coverslips at a density of 50,000 -130,000 cells/dish. Cells were made quiescent by serum starvation as described above. Cells were transfected in triplicate in Opti-MEM media for four hours with 5 g of construct using LipofectAMINE (1:6, w/w) (Life Technologies, Inc.). After 4 h of incubation cells were washed twice with PBS and subsequently cultured for 10 -12 h in 10% FBS DMEM. Cells were then labeled by the addition of 10 M BrdUrd for 10 -12 h. Cells were fixed in either methanol or 70% ethanol/50 mM glycine buffer and permeabilized with 0.1% Nonidet P-40 for 5 min. Cells were then blocked in 2% goat serum for 1 h and incubated with either 1 g/ml rabbit anti-GATA-6 antibody, 1 g/ml anti-Gax antibody (24), 1 g/ml rabbit anti-mouse ␤-galactosidase antibody (Calbiochem), or 5 g/ml rabbit anti-p21 antibody (Calbiochem) at 4°C overnight. Cells were then washed three times in PBS and incubated for 1 h with rhodamine-conjugated goat anti-rabbit-antibody (1:200 dilution, Kirkegaard & Perry Laboratories Inc.). Cultures were then incubated with anti-BrdUrd antibody (Amersham Pharmacia Biotech) which was detected with goat anti-mouse-fluorescein isothiocyante-conjugated antibody (1:800 dilution, Kirkegaard & Perry Laboratories Inc.). After washing nuclei were counterstained with Hoechst 33258 (0.5 g/ml, Sigma) for 5 min. Coverslips were mounted on glass slides (mounting medium for fluorescence, Kirkegaard & Perry Laboratories Inc.). Specimens were examined and photographed on a Diaphot microscope (Nikon Inc.) equipped for phase contrast and epifluorescence visualization.
Double Immunofluorescence-p53Ϫ/Ϫ MEFs were cultured on 60-mm plates containing 1.5% gelatin-coated glass coverslips and serum starved for 3 days in 0.5% FBS/DMEM. 5 g of plasmid was transfected using the LipofectAMINE procedure and, following trans- Western blot analyses of these cell extracts were performed with anti-GATA-6 antibody raised against peptides corresponding to regions outside of the zinc finger domains. C, preabsorption of the anti-GATA-6 antibody with the immunogenic peptides abrogates the Western blot signal. D, the wild-type GATA-6 factor, but not the ⌬ZF mutant, specifically binds to DNA. Extracts prepared from COS1 cells transfected with either the wild type or ⌬ZF mutant GATA-6 constructs were used for electrophoretic mobility gel shift assays using a double-stranded probe corresponding to the GATA-6 DNA-binding site in the mouse ␤-globin gene (7). The nucleoprotein complex was sensitive to competition by a 100-fold molar excess of the wild type, but not the mutant, GATA-6 probe.
fection, was serum-stimulated (10% FBS for 24 h). Cultures were then fixed in 4% neutral buffered formalin for 10 min, permeabilized for 5 min using 0.1% Nonidet P-40 and blocked in 2% goat serum. A mixture of 2.5 g/ml mouse anti-p21 antibody (Calbiochem) and 1 g/ml rabbit anti-GATA-6 was applied at 4°C overnight. Coverslips were washed and incubated with 1:200 dilution of fluorescein isothiocyanteconjugated goat anti-mouse antibody and 1:800 dilution of goat anti-rabbit rhodamine-conjugated antibody. Nuclei were counter stained with Hoechst 33258 and mounted on glass slides.
Colony Formation Assay-Triplicate 100-mm cultures of p21Ϫ/Ϫ MEFs (passage 3) were co-transfected with 10.8 g of pCDNA-p21 or pCDNA-GATA-6 expression vector and 1.2 g of pZEOSV2 expression vector (Invitrogen) using LipofectAMINE. 24 h following transfection cultures, 500 g of zeocin/ml was added to the medium to select for stably transfected cells. Two weeks following initial selection, cultures were fixed in methanol, stained with trypan blue, and quantified by visual inspection.

Deletion of the Zinc Finger Domain Abrogates GATA-6 DNA
Binding Activity-We compared the functionality of two plasmid constructs that express either wild type GATA-6 or a mutant GATA-6 that lacks the conserved zinc finger binding domain, ⌬ZF-GATA-6 (Fig. 1A). The wild-type and mutant GATA-6 proteins were comparably expressed in transfected COS1 cells as determined by Western blot analysis (Fig. 1B). The specificity of the anti-GATA-6 antibody was indicated by an abrogation of signal when the primary antibody was preabsorbed with a molar excess of immunogenic peptide (Fig. 1C).
Recombinant GATA-6 protein formed a nucleoprotein complex when nuclear extracts from transfected cells were incubated with a double-stranded radiolabeled oligonucleotide containing the consensus GATA site from the mouse globin gene, and this complex was sensitive to competition by a molar excess of nonlabeled wild type probe, while a mutant oligonucleotide had no effect (Fig. 1D). In addition, preincubation of rabbit anti-GATA-6 antibody with the nuclear extract induced a supershifted band (not shown). In contrast to wild type GATA-6, nuclear extracts from ⌬ZF-GATA-6-transfected cells were unable to bind to the wild-type oligonucleotide (Fig. 1D).
GATA-6 Overexpression Inhibits DNA Synthesis in VSMCs-Immunofluorescence was performed on rat VSMCs to determine the relative levels of endogenous and transiently-transfected GATA-6 protein. GATA-6 mRNA has been shown by immunoblot analysis to be highly expressed in quiescent VSMCs and down-regulated upon serum stimulation (10). As shown in Fig. 2A, GATA-6 protein is detected in the nuclei of quiescent VSMCs and is less abundant in proliferating VSMCs. However, serum-stimulated VSMCs transfected with the wildtype GATA-6 vector express GATA-6 protein at levels comparable with quiescent VSMCs.
To explore the growth regulatory properties of GATA-6, rat A7r5 VSMCs were transiently transfected with expression vectors encoding either the wild-type GATA-6 or the mutant, ⌬ZF-GATA-6, in combination with a ␤-galactosidase expression vec- tor (29). Transfected (␤-galactosidase-positive) cells were then scored for their ability to traverse S-phase as indicated by the incorporation of BrdUrd (Fig. 2B). Relative to the pCDNA vector alone, the GATA-6 expression vector inhibited BrdUrd incorporation by 60% (Fig. 2C). By comparison, a plasmid expressing the Cdk inhibitor p21 inhibited BrdUrd incorporation in A7r5 cells to a similar extent. The growth inhibition by GATA-6 was dependent on its DNA binding activity as indicated by the inability of ⌬ZF-GATA-6 construct to significantly inhibit BrdUrd incorporation.
Ectopic GATA-6 Induces Cell Cycle Arrest in Fibroblasts Lacking p53-Since mouse embryonic fibroblasts (MEFs) deficient for p53 do not express GATA-6, ectopic expression of this factor could be directly analyzed with anti-GATA-6 antibodies (Fig. 3A). In addition, analyses of growth control in this cell type are not confounded by p53-induced p21 expression (see below). In these experiments, quiescent p53Ϫ/Ϫ MEFs were transfected with GATA-6 expression vectors and labeled with BrdUrd following serum stimulation. Immunofluorescence analysis of both BrdUrd and GATA-6 at 24 h post-transfection revealed that wild type GATA-6 inhibited S-phase entry by 95 Ϯ 1% (Fig. 3B). This inhibition was not due to a delay in S-phase entry since similar results were obtained when GATA-6 transfected cultures were assessed at 48 h (not shown). In comparison, the VSMC transcription factor Gax also inhibited BrdUrd incorporation in this assay consistent with its known cell cycle inhibitory properties (24). As anticipated by its lack of activity in VSMCs, the ⌬ZF-GATA-6 mutant construct had no effect on cell cycle activity in the p53Ϫ/Ϫ MEFs (Fig.  3B).
Role of p21 in GATA-6-mediated Growth Arrest-Since p21 Cdk inhibitor has been implicated in G1 cell cycle arrest in VSMCs and other cell types (17, 24, 30), we examined the role of p21 in GATA-6-transduced cells. Double immunofluorescence assays revealed that nontransfected p53Ϫ/Ϫ MEFs do not express detectable levels of p21, presumably due the lack of p53 activation of the p21 promoter (23), but an intense p21 signal was detected in cells that were positive for wild type GATA-6 expression (Fig. 4A). High levels of p21 expression occurred in all GATA-6-positive cells examined (Ͼ100 transfected cells). In contrast, no detectable p21 expression was found in any of the ⌬ZF-GATA-6-positive cells examined.
To determine the functional significance of the p21 induction in GATA-6 expressing cells, p21-deficient MEFs (passage 3) were tested for their sensitivity to GATA-6-induced cell cycle inhibition. In contrast to the p53Ϫ/Ϫ MEFs, GATA-6-transfected p21Ϫ/Ϫ MEFs incorporated BrdUrd (Fig. 4B), with 91% co-localization of GATA-6 and BrdUrd (Fig. 4C). Late passage p21Ϫ/Ϫ MEFs (passage 28) were partially resistant to GATA-6-induced growth arrest (Fig. 4C), but they incorporated Br-dUrd at a much higher frequency than p53-deficient fibroblasts (63% for p21Ϫ/Ϫ MEF versus 5% for p53Ϫ/Ϫ MEFs). The resistance of late passage p21Ϫ/Ϫ MEFs to growth arrest by GATA-6 does not appear to be due to decreased susceptibility to cell cycle arrest, since adenovirus-mediated restoration of the p21 gene (Ad-p21) inhibited BrdUrd incorporation by 81% (Fig.  4D). Finally, stable colony formation assays revealed that p21deficient MEFs (passage 3) transfected with the GATA-6 plasmid formed colonies at a 2.5-fold higher frequency than cells transfected with a p21 expression vector (not shown), further documenting that p21 is essential for GATA-6-induced growth arrest. DISCUSSION VSMCs are of interest because of their ability to undergo a dedifferentiation process referred to as phenotypic modulation in response to mitogen stimulation, vascular injury, or other environmental stimuli (11,12). Phenotypically modulated VSMCs are proliferative and migratory and resemble their fibroblast-like precursor cells in that they express low levels of contractile proteins and high levels of extracellular matrix. Thus, it is likely that VSMCs contain mitogen-regulated transcription factors that function to coordinate the expression of genes required for cell cycle activity with those encoding smooth muscle contractile proteins (31,32). A number of mitogen-regulated transcription factors have been identified in VSMCs, including the homeoprotein Gax (33)(34)(35), MEF2 (36,37), and Ets-1 (38). Recently, GATA-6 expression was shown to be rapidly down-regulated when quiescent human or rat VSMCs were stimulated to proliferate with mitogens (10). This expression pattern is reminiscent of that described for the gas and gadd family genes (39,40), some of which have been shown to inhibit cell growth when overexpressed (41,42).
In this study we explored the growth regulatory properties of GATA-6 in VSMCs and fibroblasts. The forced expression of GATA-6 in VSMCs inhibited the incorporation of BrdUrd following mitogen stimulation, but a GATA-6 mutant lacking the zinc finger domains exhibited no detectable effect on DNA synthesis. The wild type GATA-6, but not mutant GATA-6, also inhibited the mitogen-induced proliferation of a fibroblast line that does not normally express GATA-6. These results demonstrate that GATA-6-induced growth arrest is not cell typespecific and that it may control cell cycle activity through the modulation of a ubiquitously expressed cell cycle regulatory protein.
Ectopic GATA-6 expression induced the expression of p21, a general Cdk inhibitor induced in skeletal myoblasts and other cell types during differentiation (reviewed in Ref. 30) and whose expression is modulated by the tumor suppressor, p53 (23). GATA-6-mediated growth arrest and p21 induction are independent of p53 function as indicated by the ability of ectopic GATA-6 to fulfill these functions in fibroblasts deficient for p53. Other factors that activate p21 expression in a p53independent manner include the transcriptional regulators MyoD (43), C/EBP-␣ (44), Gax (24), and the vitamin D3 nuclear hormone receptor (45) and the tumor suppressor BRCA-1 (46). Thus, up-regulation of p21 by these developmental regulators may represent a general mechanism by which cell growth and phenotypic differentiation are coordinated during embryogenesis and tumorigenesis. Similar mechanisms may also be involved in regulating VSMC phenotypic modulation in response to injury in adult vessels.
Our data indicate that the up-regulation of p21 is a functionally significant feature of growth arrest induced by GATA-6. In contrast to results with p53-deficient MEFs, p21-deficient low passage MEFs expressing GATA-6 incorporated the S-phase marker BrdUrd at a high frequency (91% for p21Ϫ/Ϫ MEFs versus 5% for p53Ϫ/Ϫ MEFs). These data suggest that the up-regulation of p21 by GATA-6 overexpression largely accounts for the G1 cell cycle arrest under these conditions. Consistent with this proposed mechanism GATA-6-transfected p21Ϫ/Ϫ MEFs form stable colonies at a high frequency, while FIG. 4. Role of the Cdk inhibitor p21 in GATA-6-mediated growth arrest. A, co-localization of ectopic GATA-6 and p21 protein. Triplicate cultures of p53Ϫ/Ϫ MEFs were transfected with either GATA-6 or ⌬ZF-GATA-6 expression plasmids. 24 h post-serum stimulation cultures were fixed in neutral-buffered formalin and analyzed for co-localization of p21 and GATA-6 or ⌬ZF-GATA-6. B, GATA-6 inhibition of the cell cycle is mediated by p21. A representative micrograph is shown for BrdUrd incorporation in p21Ϫ/Ϫ MEFs ectopically expressing GATA-6 or ␤-galactosidase (␤-Gal). Triplicate cultures of p21Ϫ/Ϫ MEFs were transiently transfected with either a GATA-6 or a ␤-galactosidase expression plasmid. At 12 h post-serum stimulation, 10 M BrdUrd was added, and incubation was continued for an additional 10 -12 h. GATA-6 and ␤-galactosidase expression and BrdUrd incorporation were detected by immunofluorescence. C, quantification of BrdUrd incorporation in early passage p21Ϫ/Ϫ MEFs cultures transfected with either pCMV-␤-galactosidase (n ϭ 169) or pCDNA-GATA-6 (n ϭ 86) and late passage p21Ϫ/Ϫ MEFs transfected with either pCDNA-GATA-6 (n ϭ 201) or empty pCDNA vector (n ϭ 260). D, restoration of p21 induces growth arrest in p21-deficient MEFs. Quadruplicate cultures of late passage quiescent p21Ϫ/Ϫ MEFs were transduced at 100 m.o.i. using a replication deficient adenoviral construct expressing either human p21 or ␤-galactosidase. Following infection, cells were serum-stimulated for 12 h before 10 M BrdUrd was added and incubation continued an additional 10 -12 h. the restoration of p21 decreases colony formation. In separate experiments, GATA-6 was unable to transactivate expression from a 2.4-kilobase pair fragment of the p21 promoter (not shown), and these data suggest that GATA-6 either acts on an unidentified p21 gene regulatory element at a distant site or increases p21 expression through a post-transcriptional mechanism. Post-transcriptional regulation of p21 was recently reported for C/EBP-␣, which increases the stability of the p21 protein (47).
In summary, we have shown that GATA-6 overexpression can up-regulate the expression of the p21 Cdk inhibitor and inhibit S-phase entry in VSMCs and fibroblasts. Though VSMC-specific gene targets for GATA-6 are currently unknown, it is reasonable to speculate that in VSMCs GATA-6 functions, like GATA-4 in heart, to activate the expression of contractile protein genes associated with the differentiated phenotype. Therefore, GATA-6 may serve multiple regulatory roles in VSMCs and function, along with other mitogen-regulated transcription factors, to coordinate VSMC proliferation and differentiation during normal vessel development. The identification of GATA-6 as a regulator of VSMC proliferation may also have implications in the development of molecular therapies to treat proliferative vessel wall disorders.