The Tumor Suppressor Gene PTEN Can Regulate Cardiac Hypertrophy and Survival*

Cardiac hypertrophy is a complex process involving the coordinated actions of many genes. In a high throughput screen designed to identify transcripts that are actively translated during cardiac hypertrophy, we identified a number of genes with established links to hypertrophy, including those coding for Sp3, c-Jun, annexin II, cathepsin B, and HB-EGF, thus showing the general utility of the screen. Focusing on a candidate transcript that has not been previously linked to hypertrophy, we found that protein levels of the tumor suppressor PTEN (phosphatase andtensin homologue on chromosome ten) were increased in the absence of increased messenger RNA levels. Increased PTEN expression by recombinant adenovirus in cultured neonatal rat primary cardiomyocytes caused cardiomyocyte apoptosis as evidenced by increased caspase-3 activity and cleaved poly(A)DP-ribose polymerase. Expression of PTEN was also able to block growth factor signaling through the phosphatidylinositol 3,4,5-triphosphate pathway. Surprisingly, expression of a catalytically inactive PTEN mutant led to cardiomyocyte hypertrophy, with increased protein synthesis, cell surface area, and atrial natriuretic factor expression. This hypertrophy was accompanied by an increase in Akt activity and improved cell viability in culture.

tor-1 (IGF-1) 1 and calcineurin (2,3), can also influence cell survival. Hence, extrinsic stimuli on the cardiomyocyte surface are no longer thought to necessarily evoke a linear signaling cascade to the nucleus or contractile machinery but rather to feed into a "signaling web," with many components and nodal points of control (4). One protein that has received recent attention and functions at such a convergent signaling nodal point, is PTEN.
The human tumor suppressor gene PTEN/MMAC1/TEP1 (PTEN, phosphatase and tensin homologue on chromosome ten; MMAC1, mutated in multiple advanced cancers-1; TEP1, TGF-␤-regulated, epithelial cell-enriched phosphatase) is either deleted or inactivated in a high percentage of breast, endometrial, brain, and prostate cancers (5)(6)(7). PTEN is also linked to two dominantly inherited disorders, Cowden disease and Bannayan-Zonana syndrome, which lead to multiple defects, including excessive developmental growth of specific structures such as digits, formation of many benign outgrowths called hamartomas, and increased occurrence of cancer (8,9). Tumor suppressor function has been confirmed by several gene ablation studies in mice, which show that mice with only one functional copy of the gene are more likely to develop tumors of multiple origins, and that homozygous loss of PTEN is embryonic lethal (10 -12).
PTEN is a dual-specificity phosphatase with homology to the focal adhesion-associated protein tensin (13). In vitro, PTEN can dephosphorylate acidic polypeptides, focal adhesion kinase (FAK), and the adaptor protein Shc. However, the major in vivo substrate for PTEN may be phosphatidylinositol 3,4,5-triphosphate (PIP 3 ), because embryonic fibroblasts taken from PTEN null mouse strains have abnormally high levels of PIP 3 and are resistant to apoptosis (12). These fibroblasts, as well as several breast cancer and glioblastoma-derived cell lines, show extremely high levels of activated Akt, a serine/threonine kinase that is regulated by PIP 3 and phosphatidylinositol 3,4-biphosphate (14,15). Akt is an important regulator of both cell survival and cell growth (16). In work spanning worms to mammals, PTEN has been defined genetically and biochemically to act as a negative regulator of Akt in opposition to the evolutionarily con-served IGF-1/PI3K/Akt signaling pathway (17)(18)(19).
Transcriptome and proteome analyses of general biological processes are beginning to yield a general picture of normal and altered cell states (20,21). We have used a modified transcriptome approach to examine those transcripts that are actively translated during cardiac hypertrophy. In the course of this screen, PTEN emerged as a strong candidate. In this study, we demonstrate that PTEN is differentially expressed during cardiac hypertrophy, and show that altered expression can impact on cardiomyocyte viability. PTEN can also function as a critical regulator of cardiomyocyte hypertrophy and survival. A PTEN mutant, H123Y, exhibits dominant-negative activity in cardiomyocytes, leading to increased Akt activation and cardiomyocyte hypertrophy. To our knowledge this is the first demonstration of a true dominant negative activity for a PTEN mutant in cultured cardiomyocytes and should contribute toward understanding the potential consequences of human tumorigenic PTEN mutations in other cell types.
Polysome Analysis of Isoproterenol-treated versus Vehicle-treated Hearts-FVB/N mice were infused with 60 mg/kg/day isoproterenol or 0.02% acetic acid (vehicle only as control) for 14 days. Crude heart homogenates from either six control or isoproterenol-treated mice were fractionated on 10 -40% sucrose gradients, and fractions were continuously monitored for absorbance at 260 nm. The material in the polysome fractions, which sedimented at Ͼ80 S was defined as "translated", while the RNA present in the fractions containing the single ribosomes, ribosomal subunit fractions, and non-sedimented material was defined as "untranslated." The resulting RNA fractions were then reversetranscribed, 32 P-labeled, and hybridized to four individual CLONTECH Atlas 1.2 Macroarray filters, and the signals were quantified by using a PhosphorImager. For each array position that had a signal greater than 0.5ϫ the median filter value and a signal on all four membranes, a ratio was calculated as shown below in Fig. 1.
PTEN Western Analysis-Total mouse heart or brain homogenates were prepared in radioimmune precipitation buffer (1ϫ PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 l/ml PMSF, 10 l/ml sodium orthovanadate) containing a protease inhibitor mixture (Complete protease inhibitor tablet, Roche Molecular Biochemicals). Protein concentration was determined by the Bio-Rad assay in triplicate, and the proteins were separated by discontinuous SDS-PAGE before transfer to polyvinylidene difluoride membrane for Western blot analysis. Rabbit polyclonal anti-PTEN antibody (Upstate Biotechnology Inc.) at a 1:2000 dilution or monoclonal anti-GAPDH antibody (Chemicon International, Inc.) at a 1:5000 dilution were used to probe the membranes before detection with horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnologies) and the ECL-Plus chemiluminescence kit (Amersham Pharmacia Biotech). The luminescence signals were quantified on a Storm 860 device (Molecular Dynamics). The experiment was repeated four separate times with similar results, and results were analyzed by Student's t test.
Cloning of PTEN cDNA-Two hundred nanograms of total RNA from adult FVB/N mouse heart was used in a single-tube reverse transcriptase-polymerase chain reaction (Titan, Life Technologies) to amplify PTEN cDNA. The sense primer was 5Ј-AGGCTCCCAGACATGA-CAGCCATC-3Ј (bases 937-960 of PTEN cDNA, GenBank accession number U92437), and the antisense primer was 5Ј-ACTTTTGTAATT-TGTGAATGCTGATC-3Ј (bases 2156 -2131). To place a peptide tag on the COOH terminus of expressed protein to allow detection, amplified PTEN cDNA was cloned in-frame into the pUni/V5-His TOPO Echo expression plasmid (Invitrogen) to create pWTPTENV5TOPO. This allowed fusion at the COOH terminus to a 14-amino acid peptide derived from the SV5 Paramyxovirus (V5 epitope) and a 6xHis tag. Site-directed mutagenesis to create the H123Y substitution was per-formed using the QuikChange kit (Stratagene) and the following complementary oligomers spanning bases 1300 -1332: 5Ј-CAT GTT GCA GCA ATT TAC TGT AAA GCT GGA AAG-3Ј and 5Ј-CTT TCC AGC TTT ACA GTA AAT TGC TGC AAC ATG-3Ј. Plasmids pWTPTENV5TOPO (WTPTEN) and pH123YPTENV5TOPO (H123YPTEN) were sequenced to verify integrity.
Cardiomyocyte Cultures and Adenovirus Infection-Rat neonatal cardiomyocytes were isolated using the Worthington Cardiomyocyte Isolation System with minor modifications as follows. Cardiomyocytes from 1-day-old Harlan Sprague-Dawley rat pup hearts were isolated by overnight incubation with trypsin at 4°C followed by a 50-min, 37°C collagenase digestion as per the manufacturer's instructions. Cells were preplated for 1 h at 37°C in DMEM/F-12 medium, 15 mM HEPES, 5% fetal bovine serum, and 1 g/ml gentamicin to allow adherence of non-cardiomyocytes. Cells remaining in the supernatant were pelleted and resuspended in 1:1 medium consisting of DMEM/F-12, 15 mM HEPES, 5% horse serum, 2.5 g/ml insulin, 2.5 g/ml transferrin, 2.5 ng/ml sodium selenite, 30 g/ml bromodeoxyuridine, and 1 g/ml gentamicin. Viable cardiomyocytes, as determined by trypan blue exclusion, were plated at 250 cells/mm 2 on ProNectin-F-coated plastic cultureware (4 g/ml coating solution in 1ϫ PBS; BIOSOURCE International) and grown for 12-18 h in 1:1 medium at 37°C in 5% CO 2 . For virus infection, cells were infected at a multiplicity of infection (m.o.i.) of 10 infectious units (ifu)/cell for 1 h in DMEM/F-12 at 37°C. Cells were allowed to recover for 12 h in 1:1 medium before being placed into DMEM/F-12 for the remainder of the culture period, usually 24 -48 h. For immunofluorescence, cells were grown in 4-well-chambered Permanox slides (Nalge Nunc) coated with ProNectin-F. Primary antibodies were used at 1:100 dilution on cells fixed in 4% paraformaldehyde/1ϫ PBS and permeabilized in 40 mM HEPES (pH 7.4), 50 mM Pipes (pH 6.9), 10 mM EGTA, 5 mM MgCl 2 , and 0.1% Triton X-100.
Replication-deficient recombinant adenoviruses were made using the AdEasy system described by He et al. (1998). XhoI-XbaI restriction fragments from either pWTPTENV5TOPO or pH123YPTENV5TOPO, spanning the cloned insert and tag, were ligated into XhoI-XbaI-digested pShuttleCMV. To create AdLacZ, a NotI-NotI restriction fragment from pCMV␤ (22) was ligated to NotI-digested pShuttleCMV. Recombination into the pAdEasy-1 viral backbone was accomplished in bacteria as described previously (23). Recombinant pAdEasy plasmids containing CMV-cDNA inserts were purified over Qiagen columns and 4 g of PacI-digested DNA was used to transfect HEK293 cells by LipofectAMINE (Life Technologies). Cells were seeded at 2 ϫ 10 6 cells per 25-cm 2 flask 24 h prior to transfection. Lysis of transfected cells indicating adenoviral growth occurred by 4 days. Following amplification, lysates containing clonal recombinant adenovirus were prepared from 10 75-cm 2 flasks and purified by CsCl gradient centrifugation. Recovered virus was aliquoted and stored at Ϫ20°C in 5 mM Tris (pH 8.0), 50 mM NaCl, 0.05% bovine serum albumin, and 25% glycerol. Virus was titered by serial dilution infection of HEK293 cells and counting of plaques under 0.3% agarose/10% fetal bovine serum/1ϫ Dulbecco's modified Eagle's medium overlay.
PTEN Phosphatase Assays-Duplicate 100-mm plates seeded with 1.96 ϫ 10 6 cardiomyocytes were either mock infected or infected with AdlacZ, AdWTPTEN, or AdH123YPTEN at an m.o.i. ϭ 10 ifu/cell. After 48 h, cells were lysed in cold lysis buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10 mM DTT, 1 mM PMSF, and protease inhibitors (Complete protease inhibitor tablet, Roche Molecular Biochemicals). Recombinant PTEN proteins were immunoprecipitated from lysates with anti-V5 antibody for 1 h, and the precipitate was treated with protein A/G-agarose beads (Santa Cruz Biotechnologies) for 3 h at 4°C. The beads were then washed in a buffer containing 20 mM HEPES (pH 7.7), 50 mM NaCl, 0.1 mM EDTA, and 2.5 mM MgCl 2 , followed by a wash in phosphatase assay buffer lacking PIP 3 (100 mM Tris-HCl (pH 8), 10 mM DTT). The phosphatase reactions were done as described previously (24). Briefly, immunoprecipitated PTEN proteins were incubated in a 50-l reaction with phosphatase assay buffer and 200 M water-soluble diC 8 -PIP 3 (Echelon) for 40 min at 37°C in a 96-well plate. Released phosphate was determined by Biomol Green Reagent (Biomol) according to the manufacturer's instructions, and absorbance was determined at 650 nm. The amount of free phosphate released was calculated from a standard curve analysis. The positive control used was 2.5 g of recombinant purified PTEN protein (Upstate Biotechnology Inc.).

Cell Surface Area Determination and [ 3 H]Leu
Assays-Cardiomyocytes were seeded at 250 cells/mm 2 in 6-well culture dishes 12 h prior to viral infection. Twenty-four h following infection (or mock treatment) cardiomyocytes transduced with AdlacZ, AdWTPTEN, or AdH123YPTEN were incubated for 6 h in leucine-free RPMI 1640 supplemented with 5 Ci/ml [ 3 H]leucine. Cells were washed, and the proteins were precipitated with ice-cold 10% trichloroacetic acid then washed again with cold 95% ethanol, and the lysates were collected in 0.5 N NaOH for scintillation counting. Genomic DNA was extracted from duplicate samples in 10 mM Tris-HCl (pH 8.0), 0.1 M EDTA (pH 8.0), 20 g/ml RNase, and 0.5% SDS, incubated at 37°C for 1 h, then extracted in phenol:chloroform before ethanol precipitation. Recovered nucleic acid was quantified by absorbance at 260 nm and used to control for cell number. Relative cell surface area was calculated from digitized cell images taken from random fields of view using National Institutes of Health IMAGE software (version 1.62). At least 100 cells were counted per sample, and statistical analysis was carried out using Student's t test.
Caspase-3 Assay-Four 100-mm plates were seeded with 1.96 ϫ 10 6 cardiomyocytes for each sample. Cells were mock infected, infected with AdlacZ, AdWTPTEN, AdH123YPTEN, or treated with 0.5 M staurosporine for 6 h prior to harvest. Trypsinized cells were scraped from the plates and counted. 1 ϫ 10 6 cells from each sample (in duplicate) were lysed and used in a 100-l reaction with 50 M of the colorimetric DEVD-pNA substrate in 1ϫ reaction buffer as per the manufacturer's instructions (Caspase-3 Colorimetric Assay kit, CLONTECH). Five M of the irreversible caspase-3 inhibitor, DEVD-fmk, was used as a control for specificity. The colorimetric reaction products were determined at 400 nm.
Akt Assay-Duplicate 100-mm plates seeded with 1.96 ϫ 10 6 cardiomyocytes were either mock infected or infected with AdlacZ, Ad-WTPTEN, or AdH123YPTEN at an m.o.i. ϭ 10 infectious units/cell. For a positive control mock infected cells were treated with 10 nm of IGF-1 for 15 min prior to harvest. Where indicated AdWTPTEN-infected cells were also treated with 10 nM IGF-1 in the same manner. For a negative control, cells were treated with 10 M of the PI3K inhibitor LY294002 (Cell Signaling Technology) for 45 min prior to and during IGF-1 treatment. Cells were harvested 48 after infection in 1 ml of 1ϫ lysis buffer (20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerolphosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, 1 mM PMSF). Two hundred microliters of lysate was then used for immunoprecipitation of Akt using 20 l of immobilized anti-Akt antibody slurry (Akt kinase assay kit, Cell Signaling Technology). Pellets were washed twice in lysis buffer, then twice in kinase buffer (25 mM Tris (pH 7.5), 5 mM ␤-glycerolphosphate, 2 mM DTT, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 ), before initiating the reaction with the addition of 200 M ATP and 1 g of GSK-3 fusion peptide. Reactions were incubated at 30°C for 30 min, then terminated with 3ϫ SDS sample buffer, and the reaction products were separated by discontinuous SDS-PAGE. The gels were transferred to polyvinylidene difluoride membrane, and the phosphorylated GSK-3␣/␤ fusion peptide was detected with anti-pGSK-3 antibody (1:1000 dilution, Cell Signaling Technology).
MTT Cell Survival Assay-Cardiomyocytes were seeded into 96-well plates at a density of 250 cells/mm 2 18 h prior to infection with adenovirus. Cells were mock infected or infected with AdlacZ, AdWTPTEN, or AdH123YPTEN at an m.o.i. of 10 ifu/cell for 1 h at 37°C, allowed to recover for 12 h in 1:1, then placed into 100 l of DMEM/F-12 without phenol red. At indicated time points, 10 l of 5 mg/ml MTT was added to the cells, which were then incubated at 37°C for 4 h. Solubilization of the converted purple formazan dye was accomplished by placing cells in 100 l of 0.01 N HCl/10% SDS and incubating overnight at 37°C. The reaction product was quantified by absorbance at 570 nm. All samples were done in triplicate, and data were analyzed by Student's t test.

Detection of Transcripts That Shift to
Polysomes during Hypertrophy-We initially sought to define genes that are differentially expressed during cardiac hypertrophy. Chronic isoproterenol (ISO) infusion in mice is a simple and wellcharacterized protocol that results in a characteristic 20 -30% cardiac hypertrophy as measured by heart to body weight ratios (25). To increase the probability of finding genes that were altered at the level of protein expression, polysomes derived from mouse ventricles were loaded on sucrose gradients and size-fractionated using velocity density centrifugation (Fig.  1A). Fractions were taken and the ventricular RNA isolated to select for the transcripts that were loaded onto polysomes in response to ISO (26). Four CLONTECH Atlas 1.2 macroarray filters were simultaneously hybridized to radiolabeled cDNA probes that were derived from either vehicle-treated (control) free or polysome-bound RNA, or from ISO-treated free or polysome-bound RNA (Fig. 1). For each array element, a numerical value for the shift to polysomes was calculated by taking the ratio of ISO bound/free signal divided by the ratio of vehicle bound/free signal. Signals were normalized to the median filter value to correct for differences in probe specific activities. A value of Ͼ1 indicates an increase for that transcript in the polysome fraction due to chronic ISO infusion. A typical block of array elements from the four filters (Fig. 1B) shows that the tumor suppressor PTEN (arrow) and the collagen IX ␣-2 subunit (arrowhead) undergo a 3.3-fold and 6.8-fold shift to polysomes, respectively (Table I). Beta actin and GAPDH were invariant. Interestingly, although the candidate genes exhibited a wide range of polysome shifts (ϳ2-to 90-fold) almost all of the candidates exhibited only modest changes in total RNA amounts between control and hypertrophic hearts (0.6-to 5-fold) (Fig. 2, A and B). The lone exception was the serine/ threonine kinase pim-1, which normally is most highly expressed in myeloid cells (27). The candidate genes cover a spectrum of cellular functions such as signaling, cell architecture, transcription, proteolysis, and phosphorylation (Table I).
Increased PTEN Expression Causes Cardiomyocyte Apoptosis-Apparent transcript loading into heavy polysome complexes could be the result of either a block to translation elongation and hence decreased protein expression, or an increase in translation initiation and a resultant rise in protein production. One of the candidates, PTEN, has been linked to growth and metabolism of many cell types. Therefore, we hypothesized that it may play a similar role in the heart. Although the data in Fig. 2 showed that a number of transcripts exhibited a more robust response to isoproterenol stimulation, we undertook a detailed examination of PTEN to obtain a better understanding of the screening's general value, and because of the involvement of PTEN in cell cycling in other systems. Quantitative Western blot analyses showed that PTEN protein expression is, in fact, induced in ISO-treated mouse hearts relative to vehicletreated hearts (Fig. 3), in agreement with the polysome-derived data. Taken together with the shift of PTEN mRNA into the heavy polysome fractions during hypertrophy and the minimal change of total PTEN mRNA, this finding is consistent with regulation of PTEN expression by increased translational initiation. To address the consequences of altered PTEN expression in the heart, recombinant adenoviruses expressing wildtype PTEN or the catalytically inactive H123YPTEN mutant (13) were used to infect neonatal rat primary cardiomyocytes in culture (Fig. 4). In uninfected cells or cells infected with a control virus expressing lacZ, endogenous PTEN exhibits a punctate pattern in the cytosol or plasma membrane and also is present in the nucleus (Fig. 4A). Overexpression of wild-type PTEN resulted in fewer cells remaining on the plate 48 h post infection, with many of the cells detached and floating (Fig.  4B). Within 48 h nearly 80% of the cells had died as determined by an MTT survival assay (Fig. 5A). To address whether cell death was due to apoptosis, caspase-3 enzymatic activity was measured in cell lysates 24 h post recombinant virus infection of neonatal cardiomyocytes. Caspase-3 is a key "executioner" of apoptosis and is responsible for many proteolytic cleavage events in apoptotic cells (28). Although expression of lacZ or H123YPTEN did not alter caspase-3 activity, expression of wild-type PTEN led to a significant increase in caspase-3 activity that could be blocked with a specific caspase-3 inhibitor (Fig. 5B). PARP (poly (ADP-ribose) polymerase), a nuclear DNA repair enzyme that is cleaved and inactivated as a late apoptotic event (29,30), also accumulates in wild-type PTENexpressing cells (Fig. 5C).
Expression of H123YPTEN Leads to Cardiomyocyte Hypertrophy-Surprisingly, expression of the H123YPTEN mutant led to cardiomyocyte hypertrophy (Fig. 4C), with a well-ordered sarcomeric structure being conserved in the cardiomyocytes as shown by ␣-actinin staining (Fig. 4F). To determine if any molecular markers of hypertrophy were activated, we examined whether AdH123YPTEN-infected cells express atrial natriuretic factor (ANF), a readily identified marker of cardiomyocyte hypertrophy and stress, which accumulates in a perinuclear ring. Cells infected with either WTPTEN or H123YPTEN showed increases in perinuclear accumulations of ANF (Fig. 6), but the dominant negative mutant induced a much more robust response than the wild-type protein.
Other hallmarks of hypertrophy include increased cell volume and protein synthesis. Expression of H123YPTEN caused a Ͼ2.5-fold increase in cardiomyocyte cell surface area and a significant increase in [ 3 H]leucine incorporation in the H123YPTEN-infected cells (Fig. 6, G and H). These data are consistent with the hypothesis that H123YPTEN can act in a dominant negative manner to enhance growth signaling in cardiomyocytes.
PTEN Alters the Activation of Akt in Cardiomyocytes-PTEN has been implicated genetically and biochemically as a negative regulator of insulin and insulin-like growth factors (12, 17-19, 31, 32). This is primarily accomplished by dephospho-  Fig.  1 and under "Experimental Procedures." The gene numbers correspond to Table I. B, total RNA changes were calculated by summing signals from all four array membranes. With the exception of the serine/threonine kinase, pim-1 (gene number 23), the candidate genes exhibited minimal RNA fluctuation. rylation of 3Ј-phosphoinositides generated by PI3K stimulation. In this capacity, PTEN can decrease the activity of the important kinase, Akt, and indeed, PTEN null cells show an elevated basal activation of Akt (12,32). Active Akt is phosphorylated on several key residues, and subsequently migrates from the cytosol to the plasma membrane and then to the nucleus (33,34). We wished to determine the effects of expressing the wild-type and dominant negative forms of PTEN on the subcellular localization of Akt. As expected, expression of WTPTEN in cardiomyocytes blocks the nuclear accumulation of phosphorylated Akt (pAkt) that is seen under basal, nonstimulated conditions (compare Fig. 7A with 7B). Surprisingly, expression of the H123YPTEN mutant leads to an increase in nuclear accumulation of pAkt and apparent Akt activation (Fig. 7C). This was confirmed by an in vitro kinase assay to measure active Akt in cardiomyocytes (Fig. 7J). Compared with either untreated cells or to those that were infected with lacZ or WTPTEN-expressing virus, expression of H123YPTEN leads to an increase in Akt enzymatic activity in cardiomyocytes. Because H123YPTEN can cause cardiomyocyte hypertrophy, and this is accompanied by an increase in Akt activation, we conclude that H123YPTEN acts in a true dominant negative manner to effectively reduce the function of endogenous PTEN.
PTEN Blocks Signaling by the Hypertrophic Agonist IGF-1-Because PTEN is a downstream negative regulator of PI3Kmediated growth factor signaling, we next tested the ability of WTPTEN to block IGF-1 activation of Akt (Fig. 8). As expected, WTPTEN blocks nuclear accumulation of pAkt in response to 10 nM IGF-1. This effect also extends to estrogen, which activates Akt in vascular endothelial cells (35)(36)(37) and in cardiomyocytes. 2 These observations are corroborated by an in vitro Akt kinase assay using lysates derived from cells overexpressing WTPTEN (Fig. 8G), confirming the ability of PTEN to block growth factor stimulation of the PI3K pathway in cardiomyocytes. H, recombinant WTPTEN is a functional PIP 3 phosphatase, but H123YPTEN is not. Recombinant PTEN or H123YPTEN was immunoprecipitated using anti-V5 antibody on protein A/G beads for 4 h. The phosphatase assay was performed as described under "Experimental Procedures" with release of free phosphate measured colorimetrically. 1-4 correspond to the lanes in the Western blot (G).

FIG. 5. Expression of wild-type PTEN causes cardiomyocyte apoptosis.
A, cell survival as determined by MTT assay. Living cells convert MTT to an insoluble purple precipitate that can be measured by absorbance at 570 nm. Primary rat neonatal cardiomyocytes were infected with either lacZ, wild-type PTEN, or H123YPTEN expressing virus and allowed to recover for 12 h in regular growth medium before being placed into DMEM/F-12 (t ϭ 0). Expression of H123YPTEN led to a significant increase in cell viability (*, p Ͻ 0.05 at 48 h), but expression of wild-type PTEN led to increased cell death at both 24 and 48 h (**, p Ͻ 0.01 at both points). B, to determine if cell death was caused by apoptosis in the PTEN-infected cells, caspase-3 activity was measured in lysates from infected cells via an enzymatic colorimetric assay. Activities were normalized to protein amounts and are reported relative to uninfected cell lysates. Inclusion of the caspase-3 inhibitor, DEVD-fmk at 5 M, significantly reduces enzymatic activity of wild-type PTEN lysates (*, p Ͻ 0.05 and **, p Ͻ 0.05 relative to uninfected control, respectively). 0.5 M staurosporine for 6 h on uninfected cells was used as a positive control for apoptosis. C, cleaved PARP, another indication of apoptosis, was determined by Western blot analysis. Expression of WTPTEN leads to the accumulation of cleaved PARP, whereas H123YPTEN does not. sequently be used to narrow an imposing (and growing) list of potentially interesting molecules to those that may be biologically active under a given set of conditions. Several of the genes obtained in our screen are associated with cardiac hypertrophy, including those encoding HB-EGF (38), c-Jun (39), annexin II (40), cathepsin B (41), and Sp3 (42). Interestingly, annexin II can complex with the protease cathepsin B on outer cell membranes to effect extracellular matrix remodeling (43), a process that occurs during cardiac hypertrophy. Annexin II can also bundle the actin filaments, modulating the cytoskeletal architecture (44), a process regulated by members of the multigenic Ca 2ϩ -binding S100 proteins such as calcyclin (S100A6) in association with the calcyclin binding protein obtained in this screen (Table I ( 45,46)). Two genes encoding the interleukin-1 receptor and lecithin:cholesterol acetyltransferase (LCAT), which were identified in the screen (Table I), are also induced by isoproterenol in other cell types (47,48). The transcription factor Epas-1 (Table I) is normally associated with hypoxia and can induce the expression of enzymes involved in glycolysis (49). Epas-1 may therefore play a role in the transition from fatty acid oxidation to glucose metabolism observed in hypertrophic myocardium (50).
In preliminary Western blot experiments, about half of the proteins that were identified by the polysome-based screen increase in expression (eg Sp3, c-Jun, and HB-EGF), some proteins decrease in expression (cathepsin D), whereas others do not change (A-raf). 3 Therefore, screening for altered gene expression during hypertrophy by polysome fractionation appears to be a reasonable protocol for whole tissue samples. Its advantage with respect to fractionation via staining of twodimensional electrophoretic gels is that low abundance proteins such as transcription factors and kinases will be detected (51). An advantage over traditional arrays or microchips is that candidate genes are derived, based on polysome status, and thus are being actively translated (26).
Cardiac hypertrophy, a general adaptive response of the heart, can be stimulated in a number of ways, including the IGF-1/PI3K/Akt pathway. The above data show that PTEN, a known antagonist to PI3K signaling, can serve as a critical determinant of cardiomyocyte growth: It can block growth factor signaling, prevent hypertrophy, and, under certain circumstances, cause cell apoptosis. In tissue culture, expression of a dominant negative PTEN leads to increased cardiomyocyte viability and significant hypertrophy.
That expression of the H123YPTEN mutant caused the cultures of cardiomyocytes to hypertrophy was unexpected. Our original hypothesis was that the mutant would serve as a negative control. The H123Y mutation was originally isolated from a human endometrial tumor (13) and the mutation maps within the conserved "P-loop" that defines the catalytic core of dual specificity phosphatases (52). The mutant protein does not possess in vitro phosphatase activity (13) and, unlike the wild- Duplicate plates were processed in parallel for determination of DNA content as an internal control. All assays were done in triplicate, and the values reported are relative to uninfected cells. *, p Ͻ 0.05 relative to uninfected control cells.

FIG. 7. Expression of H123YPTEN increases activity of Akt.
A-C, primary cardiomyocytes treated as above (Fig. 4) were analyzed for the presence of phosphorylation of Akt at serine 473, the activated form of the kinase. Active pAkt is first targeted to the plasma membrane and then accumulates in the nucleus. B, representative cells from the WT PTEN-infection show partial redistribution of pAkt out of the nucleus. C, expression of H123YPTEN increases pAkt in the nucleus. D-F, expression of recombinant PTEN proteins was confirmed by immunofluorescence using anti-V5 antibody against the V5-COOH-terminal fusion epitope placed in-frame on virally expressed PTEN proteins. G-I, the total Akt (tAKT) localization pattern was similar to that for pAkt, but note that nuclear Akt appears to be predominantly phosphorylated (compare A to G, and B to H). J, expression of H123YPTEN causes increased Akt kinase activity. Akt was immunoprecipitated from cell lysates and used to phosphorylate a GSK-3␣/␤ fusion peptide that was then subjected to SDS-PAGE and Western blot analysis to detect the phospho-GSK-3 product. 10 nM IGF-1 stimulation of cardiomyocytes for 15 min was used as a positive control. The PI3K inhibitor LY29004 (10 M for 45 min prior to and during IGF-1 incubation, INH.) was used to block IGF-1 stimulation. The increase in Akt activity is consistent with H123YPTEN acting in a dominant negative fashion. type PTEN, is unable to arrest MCF-7 tumor cells at the G 1 phase of the cell cycle (53). PTEN has been clearly established as a potent tumor suppressor, but the precise mode and scope of PTEN function may depend on cellular context. Mice that are heterozygous for PTEN inactivation by gene ablation develop tumors in multiple tissues such as prostate, thyroid, and endometrium, as well as non-neoplastic hyperplasia of the lymph nodes (10 -12). Tissue and developmental cues probably play a role in gene dosage effects on neoplastic growth as not all of the tumors examined from heterozygous PTEN mice suffer a loss of both alleles; i.e. loss of heterozygosity. Analysis of human tumors shows that PTEN expression may be silenced in other ways such as missense RNA-mediated decay (54) or biallelically repressed by unknown epigenetic mechanisms (55). Our data are consistent with the hypothesis that certain PTEN mutations, such as H123Y described here and C124S (56), may contribute directly to the observed phenotype. Although it is not known whether both alleles of PTEN are active in the primary Harlan Sprague-Dawley rat cells used in this study, sequencing of the endogenous cardiac PTEN cDNA revealed complete amino acid identity to rat PTEN sequence determined by others (GenBank accession number AF017185).
There are several possible mechanisms by which H123YPTEN could act as a dominant suppressor of endogenous PTEN, including the sequestration of PTEN binding partners needed for full activity. Yeast two-hybrid studies show that PTEN can interact with members of the membrane-associated guanylate kinase family (57,58). These proteins, MAGI-2 and MAGI-3, are proposed to assemble multiprotein complexes on the inner surface of the plasma membrane at regions of cell-cell contact. Truncated PTEN mutants that do not bind to MAGI-2 or -3, because they lack the necessary COOH-terminal PDZ domain, do not function as efficiently as wild-type PTEN in suppressing Akt activation (57). Another PTEN binding partner is FAK, and Tamura and colleagues (59, 60) described a substrate trapping D92A PTEN mutant that can stabilize FAK-PTEN association. In related studies, they showed that PTEN binds to FAK directly, leading to its dephosphorylation and inactivation. Consistent with the ability of H123YPTEN to sequester FAK from endogenous PTEN, we have observed an increase in FAK tyrosine phosphorylation in AdH123Y-infected cells. 3 We are currently analyzing whether the H123Y mutation stabilizes the interaction with FAK in cardiomyocytes.
Recently, Weng and colleagues (56) showed that PTEN can block insulin-stimulated MAPK activation, a result that may explain genetic studies in Caenorhabditis elegans that implicate PTEN in a PI3K-independent signaling pathway (19,61). This inhibition is mediated through a block in IRS-1 phosphorylation by an unknown mechanism. Although not examined in the current study, expression of WT PTEN would be expected to block insulin/IGF-1-mediated MAPK activation as well as Akt activation. Conversely, the H123Y PTEN mutant may stimulate MAPK and thus contribute to the observed hypertrophic phenotype.
Little is known of how PTEN expression is controlled. Li and Sun (6) first noted that TGF-␤1 could reduce PTEN mRNA in human keratinocytes and others (62) later showed that TGF-␤1 could decrease PTEN protein expression in a dose-dependent manner in human hepatocarcinoma cells. There is some evidence that TGF-␤1 can induce PTEN expression as well (63). However, no direct link connecting TGF-␤1 receptor activation to PTEN expression has been established. TGF-␤1 may play a role in extracellular matrix deposition during cardiomyopathy and be an important marker for end-stage failure (64). Interestingly, in preliminary experiments 3 we observed decreased PTEN expression in genetic models of cardiomyopathy generated by the cardiac-specific expression of tropomodulin (65) or calcineurin (66). Unlike the isoproterenol-treated mice studied in this report, the transgenic models all exhibited severe cardiac hypertrophy or dilation. It may be that decreased expression of PTEN is a marker for heart failure and may correlate with increased expression of TGF-␤. We are currently investigating whether TGF signaling plays a role in cardiac expression of PTEN.
Recent studies have suggested a direct link from G-proteincoupled receptors, such as the ␤-receptors that bind to isoproterenol, to activation of PI3K and Akt (67)(68)(69). Because steadystate levels of PTEN are increased in the hearts of mice treated by isoproterenol infusion, and PTEN can block signaling mediated by PI3K, this suggests a novel negative feedback mechanism to blunt adrenergic stimulation. G-proteins themselves are heterotrimeric and consist of ␥␤ and ␣ subunits. G␥␤ heterodimers can activate Akt in a PI3K-dependent fashion, but G␣ subunits inhibit this activation (69). It is intriguing to speculate that this inhibition may be mediated through the induction of PTEN expression.  Fig.  7. Expression of PTEN blocked nuclear accumulation of pAkt in response to 10 nM IGF-1 (compare A to B) or 10 nM 17-␤-estradiol (compare D to E), known activators of the PI3K pathway. Total Akt was distributed in a similar manner to pAkt. 3 C and F, cells expressing recombinant protein are shown by anti-V5 immunofluorescence. Arrowheads denote cells with a pronounced lack of nuclear pAkt. G, expression of wild-type PTEN blocks Akt kinase activity. WTPTEN-infected cells were incubated with 10 nM IGF-1 for 15 min, 48 h post-infection, and compared with cells infected with AdLacZ in the presence (ϩ) or absence (Ϫ) of IGF-1; the PI3K inhibitor LY29004 was used as a negative control. Note the difference in Akt activation due to IGF-1 in the LacZ-infected cells versus the wild-type PTEN-expressing cells.