Identification of AKT-regulated genes in inducible MERAkt cells

IRENE KUHN1,2, MARTY F. BARTHOLDI2, HUGH SALAMON3, RICHARD I. FELDMAN1, RICHARD A. ROTH4 and PAUL H. JOHNSON1

1 Departments of Cancer Research
2 Genomics and Gene Therapy
3 Immunology, Berlex Biosciences, Richmond 94804-0099
4 Department of Molecular Pharmacology, Stanford University Medical School, Stanford, California 94305-5174


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
AKT/protein kinase B plays a critical role in the phosphoinositide 3-kinase (PI3-kinase) pathway regulating cell growth, differentiation, and oncogenic transformation. Akt1-regulated genes were identified by cDNA array hybridization analysis using an inducible AKT1 protein, MERAKT. Treatment of MERAkt cells with estrogen receptor ligands resulted in phosphorylative activation of MERAKT. Genes differentially expressed in MERAkt/NIH3T3 cells treated with tamoxifen, raloxifene, ICI-182780, and ZK955, were identified at 3 and 20 h. AKT activation resulted in the repression of c-myc, early growth response 1 (EGR1), transforming growth factor ß receptor III (TGF-ßr III), and thrombospondin-1 (THBS1). Although c-myc induction is often associated with oncogenic transformation, the c-myc repression observed here is consistent with the anti-apoptotic function of AKT. Repression of THBS1 and EGR1 is consistent with the known pro-angiogenic functions of AKT. AKT-regulated genes were found to be largely distinct from platelet-derived growth factor-ß (PDGFß)-regulated genes; only T-cell death-associated gene 51 (TDAG51) was induced in both cases. In contrast to their repression by AKT, c-myc, THBS1, and EGR1 were induced by PDGFß, indicating negative interference between elements upstream and downstream of AKT1 in the PDGFß signal transduction pathway.

protein kinase B; cDNA array hybridization; estrogen inducible expression; oncogenesis; gene regulation; signal transduction networks


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
AKT1 IS A SERINE/THREONINE protein kinase whose activity is positively regulated by phosphoinositide 3-kinase (PI3-kinase) and negatively regulated by the phosphatases PTEN and PP2A. AKT transduces signals from a number of extracellular stimuli (including insulin, cytokines, and mechanical stress) that affect glucose metabolism, cell cycle progression, protein synthesis, tumor neovascularization, and apoptotic susceptibility (9).1

Overexpression, hyperactivation, or constitutive activation of AKT is oncogenic. Akt1 was first identified as the transforming gene, v-Akt, encoded by the AKT8 retrovirus. v-Akt differs from its cellular homolog, Akt1, in that it has a 5' fused viral gag sequence that causes constitutive activation of the AKT1 kinase. Constitutively active AKT causes cellular transformation in vitro, and cells transfected with constitutively active AKT1 form tumors in nude mice. In vivo, mice transgenic for constitutively activated AKT1 develop T cell lymphomas rapidly (33). Cotransfection with dominant-negative AKT blocks transformation, indicating that AKT activity is necessary and sufficient for oncogenic transformation. AKT activity is elevated in samples of ovarian, breast, prostate, lung, and glioblastoma tumors (4, 36).

The purpose of this study was to identify Akt1-regulated genes to aid in understanding the mechanism of AKT1-mediated oncogenesis. The use of an inducible form of the AKT1 protein, MERAKT, made it possible to study the pattern and kinetics of Akt1-regulated gene expression using a single cell line. Activity of the MERAKT kinase is rapidly stimulated in the presence of estrogen receptor (ER) ligands, and repressed in their absence. To create MERAkt, a mutant murine ER-hormone binding domain (ER-HBD) was fused to the 3' end of the kinase domain of Akt1, and a myristoylation sequence substituted for the pleckstrin homology (PH) domain at the 5' end of Akt1. Activation of the gene product, MERAKT, like AKT1, requires membrane localization and phosphorylation at two sites, Thr308 and Ser473 (23). The NH2-terminal myristoylation sequence localizes the MERAKT protein to the plasma membrane, but the ER-HBD blocks the phosphorylative activation sites until ER ligands are added to the culture (23). Mirza et al. (31) recently used MERAkt to show that induction of AKT1 activity is sufficient to transform Rat1 or (to a lesser extent) NIH3T3 fibroblasts, and to increase the apoptotic resistance and cell cycle progression of Rat1 fibroblasts. Increases in apoptotic resistance and cell cycle dysregulation contribute to oncogenic transformation (6).

Akt plays a role in the inhibition of apoptosis (9, 33, 36), increased cell-cycle progression (22), enhancement of tumor neovascularization (29), and stimulation of glycolysis (2), all of which contribute to tumor formation and maintenance. However, there are few reports of genes regulated by Akt to carry out these AKT-dependent functions. The few reported Akt-regulated genes are GLUT-1 and phosphoenolpyruvate carboxykinase (PEPCK), whose regulation increases glycolysis (3, 26); vascular endothelial growth factor (VEGF), which promotes angiogenesis (19); the anti-apoptotic factor Bcl-2 (34); and p27kip, a cell cycle inhibitor repressed by AKT (5, 11, 30, 39). In contrast, there are numerous reports of Akt-regulated transcription factors, including AP-1 (12), glucocorticoid receptor (35), E2F (7, 15), forkhead transcription factor FKHRL1 (11, 30, 32, 39), hepatocyte inducing factor 1{alpha} (HIF1{alpha}; 29), cAMP response element binding proteins (CREB), and nuclear factor {kappa}B (NF-{kappa}B) (20). Each of these transcription factors affects a number of characterized target genes, yet determination of which of these is Akt-regulated has not been reported. Investigation of Akt-regulated genes is complicated by the small magnitude of the changes in gene expression due to activation of AKT (2- to 3-fold) (34) and by the repeated observation that AKT-dependent signaling requires costimulation of other signaling pathways (12, 15, 20). Activation of other signaling pathways may also eliminate AKT-dependent signals through negative interference.

Previous studies using the inducible MERAkt gene have shown that nuclear signaling due to MERAKT activation is indistinguishable from signaling initiated by insulin-activated endogenous AKT1. Transcriptional regulation of PEPCK and GLUT-1 by 4-hydroxytamoxifen (Tam)-induced MERAKT mimics that of insulin-induced AKT1 (3, 26). We report here more than 20 AKT-regulated genes identified at 3 and 20 h after induction of AKT1 activity in MERAkt cells (retrovirally derived from NIH3T3 cells). Six chemically distinct ER ligands were used to activate MERAKT in replicate experiments. The distinct chemical structures and activation potencies of these ligands allowed AKT-dependent regulation to be distinguished from AKT-independent regulation. The known functions of a few of the genes contrast with the oncogenic phenotype induced by activation of AKT1. The regulation of many of the genes, however, is consistent with oncogenesis. Among these genes are anti-apoptotic factors; promoters of tumor neovascularization, invasion, and metastasis; and regulators of cell cycle progression. Our results identify for the first time genes regulated by activation of the oncogene Akt1 and also indicate that there is significant negative interference between AKT-dependent signals and signals from elements of the platelet-derived growth factor-ß (PDGFß) pathway upstream of AKT.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Cell culture and sample preparation.
MERAkt cells [referred to as myrAKT {Delta}4-129-ER by Kohn et al. (22)] were incubated in DMEM (GIBCO), 10% FBS (Hyclone). Cells for expression analysis were plated at 7.5 x 105 per P150 plate, grown to 95% confluence, washed, and incubated in DMEM, 0.5% FBS for an additional 48 h.

Both phenol red-free DMEM and charcoal-adsorbed FBS were tested in MERAkt cell proliferation assays and found to have no effect on generation time or cellular morphology.

To minimize any differences due to culture conditions, all treated and untreated MERAkt cultures were grown, serum-starved, treated, and handled in parallel in a given experiment. All experiments were repeated at least once on separate occasions, and untreated reference culture samples were prepared at each time point in triplicate. Samples were harvested at 3 and 20 h after drug treatment.

ER ligands and growth factors for activation of MERAKT.
17ß-Hydroxyestradiol (E2) and Tam (Sigma) were dissolved in ethanol at 1 mM and stored at -20°C. Raloxifene (Ral), ICI-182780 (ICI), ZK427, ZK819, and ZK955 were obtained from the Molecular Pharmacology Division of Berlex Biosciences, reconstituted to 10 mM in DMSO, and diluted to 1 mM in ethanol. The ER ligand concentration at which intensity of the phosphorylated MERAKT signal was half-maximal is defined as the EC50. All ER ligands were used at a final concentration of 0.1 µM. PDGFß (R&D Systems) was reconstituted according to the manufacturer’s instructions (fatty acid-free BSA, 0.1%, was used to avoid fatty acid stimulation of Akt activity) and used at 10 ng/ml.

Flow cytometric analysis.
The percentage of cells in each phase of the cell cycle under a given set of culture conditions was determined by flow cytometric analysis of propidium iodide-stained nuclei. MERAkt or NIH3T3 cells were harvested by trypsinization, washed in Dulbecco’s phosphate-buffered saline with calcium and magnesium (GIBCO), and counted, and 3 million cells were resuspended in 0.25 ml 4°C phosphate-buffered glucose/EDTA solution on ice. Then, 0.75 ml of 95% nondenatured alcohol (-20°C) was added dropwise. Nuclei were pelleted, resuspended in 0.5 ml PBS containing 50 µg/ml RNase, and incubated at 37°C for 20 min; then 25 µg propidium iodide was added to each sample, and the samples incubated at ambient temperature for 30 min. Samples were analyzed on a Becton-Dickinson FACScan apparatus.

Protein analysis by Western blot.
Cell samples were lysed in anti-phosphatase buffer (1% Nonidet P-40, 15% glycerol, 137 mM NaCl, 20 mM Tris, pH 7.5, 2 µg/ml leupeptin, 2 µg/ml aprotinin, 2 mM benzamidine, 20 mM NaF, 10 mM NaPPi, 25 mM ß-glycerolphosphate, with fresh 1 mM Na3VO4 and 2 mM PMSF). Equivalent amounts of total protein (determined by bicinchoninic acid assay; BCA, Pierce) were analyzed by Western blot using Novex 10% acrylamide Tris-glycine gels (Novex) following the manufacturer’s instructions. Primary antibody was used at 1:1,000; secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit antisera, New England Biolabs) was diluted 1:2,000, and developed with Luminol (New England Biolabs).

Antibodies for Western analysis.
Phosphorylated AKT and MERAKT were detected by anti-peptide sera purchased from New England Biolabs (NEB). NEB rabbit sera raised against an AKT1 decapeptide containing phosphoserine-473 or raised against a similar peptide containing AKT1 phosphothreonine-308 are referred to here as anti-phospho473AKT and anti-phospho308AKT sera, respectively. Phosphorylation of glycogen synthase kinase-3 (GSK3) was similarly detected using rabbit anti-phosphorylated-GSK3-peptide sera from NEB. Horseradish peroxidase-conjugated goat anti-rabbit sera, as well as the similar anti-mouse sera, were purchased from NEB. A mouse monoclonal antibody (P67920) from Transduction Laboratories was used to distinguish AKT1 from the other isoforms of AKT.

cDNA array hybridization studies.
The mouse arrays (Atlas 1.2, 7853-1; Clontech Laboratories, Palo Alto, CA) containing 1,176 genes representing all major cellular pathways and functions were employed in these studies. To prepare probe for hybridization, total RNA was purified from each sample using Qiagen RNeasy Midi kits according to manufacturer’s instructions. Probe was prepared from 4 µg of total RNA according to Clontech’s November 1999 protocol, except that G-50 TE spin columns (Boehringer-Mannheim) were used to purify labeled complex probe away from unincorporated [{alpha}-32P]dATP. Size distribution of the 32P-labeled complex probe was determined by electrophoresis on 6% acrylamide urea denaturing gels (Novex) followed by autoradiography.

Data quantification and background determination.
Fuji imaging screens were exposed to the hybridized gene array membranes for 72 h and read into a Storm PhosphorImager (Molecular Devices, Sunnyvale, CA) at 200 µm resolution. The hybridization signal intensities at each cDNA spot were quantified using Array Vision (Imaging Research, St. Catherines, Ontario, Canada). The grid definition protocol used an automated alignment algorithm to finely adjust the grid over the spots. The signal intensity was calculated as the mean pixel value minus a local regional background and reported for each spot in the array. The signal intensities were then normalized to the mean of all spots on the array.

Identification of differentially expressed genes by GEPView.
GEPView is a proprietary software, from Berlex Laboratories, for analysis of cDNA hybridization array data. Output of Array Vision (above) was entered into GEPView for determining changes in signal intensity and z score determination. GEPView calculates the fold change of the intensity of the hybridization signal for each gene (cDNA) in a treated sample to the signal of the same gene in the matched untreated reference sample. For fold change calculations, if the signal intensity value fell below a threshold set at background/2, then the intensity value for that gene was reset to a value of background/2. The criteria set for differential expression was that the fold change in expression comparing treated to untreated signals be greater than 2 or less than -2 and that the z score be at least 0.3. The z score reflects the difference between two intensity values being evaluated compared with the spread of the data for the entire array (mean/standard deviation). The algorithm used for z score determination was developed in Array Vision. Differentially expressed genes were then sorted using Microsoft Excel to identify those genes differentially expressed in response to multiple types of treatment.

Hybridization intensity normalization and statistical analyses.
Analyses were accomplished using in-house software implementing plotting techniques based on Student’s t-test calculations, as in Callow et al. (8). P values reported are uncorrected for multiple tests. A conservative (strict) correction for multiple tests was also applied, where the correct value was calculated as 1 - (1 - P)N, with P being the uncorrected value and N the number of expression levels tested. The null hypothesis for each gene’s level of expression was that treated and untreated sets of hybridization intensities were the same; the mean values for the treated samples and untreated samples were assumed to be equal. The t-test evaluates this, and the hypothesis may be rejected at the reported alpha level or "P value."

Quantile-quantile plots of statistics were used to investigate the distribution of statistics calculated. The quantile-quantile plot compares the t-statistics calculated for each gene, testing the difference between the mean values for treated and untreated samples, with the quantiles of the t-distribution derived from the number of samples tested and rank of the statistic in the dataset. Points on the plot which show much more extreme values in the calculated t statistic than in the quantile value may be outliers of note.

Prior to calculating t statistic values and plotting, each array of hybridization intensities was normalized to a vector consisting of the mean values for each gene using Locfit for the statistical package R (http://cm.bell-labs.com/cm/ms/departments/sia/project/locfit/index.html). Trends in values following mean intensities across arrays for expression levels were thus removed; if an array showed an idiosyncratically low set of values for highly expressed genes, for example, this trend was removed prior to statistical calculations.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Activation of MERAkt kinase activity.
Figure 1 illustrates the structure of recombinant AKT1 kinase, MERAKT, encoded by the MERAkt retrovirus used in this study. The AKT1 kinase is activated by phosphorylation at threonine-308 and serine-473 (23). The binding of heat shock protein HSP90 to the mutant murine ER-HBD of MERAKT blocks these two activation sites of MERAKT unless an ER ligand such as Tam (also known as 4-HT) is present. Binding of ER ligands such as Tam to the ER-HBD domain results in the release of HSP90 and subsequent phosphorylation of the activation sites T308 and S473. It was previously shown that MERAKT is enzymatically inactive in untreated cells and that AKT1 activity is rapidly induced by Tam treatment. Tam treatment results in phosphorylative activation of MERAKT’s AKT1 activity (22). We show in Fig. 2A that MERAKT is fully phosphorylated 30 min after addition of Tam to cultures of MERAkt cells and remains phosphorylated for at least 3 h. Figure 2B shows that the total amount of MERAKT does not increase as a result of Tam treatment, suggesting that posttranslational modification, rather than de novo synthesis, is responsible for the observed increase in the amount of phosphorylated MERAKT. The same samples were analyzed by Western blot using anti-phospho308AKT serum and anti-phosphoGSK3 serum. The results demonstrate that Tam treatment resulted in phosphorylation of MERAKT at position 308 in addition to position 473 (Fig. 2C) and concomitant phosphorylation of the AKT substrate GSK3 (Fig. 2D), indicating the induction of AKT1 activity in MERAkt cells by Tam treatment.



View larger version (5K):
[in this window]
[in a new window]
 
Fig. 1. Structure of the MERAKT inducible kinase system [Kohn et al. 1998 (22)]. Diagram is oriented with the NH2 terminus on the left. The NH2 terminus was engineered to encode a myristoylation site, the pleckstrin homology region (PH) was deleted, and a heme agglutinin (HA) epitope was fused to the COOH terminus of the AKT1 kinase, followed by the G525R mutant hormone binding domain of the estrogen receptor (ER-HBD) (27).

 


View larger version (40K):
[in this window]
[in a new window]
 
Fig. 2. Western analysis of MERAKT activation kinetics. Nearly confluent MERAkt cells were incubated 48 h in 0.5% FBS, then treated for various times with 0.2 µM tamoxifen (Tam). Lane 1, untreated control; lane 2, 30 min; lane 3, 60 min; lane 4, 120 min; lane 5, 180 min; and lane 6, control cells incubated in 10% FBS medium. A: Western developed using anti-phospho473AKT serum. B: Western shown in A was stripped and redeveloped using anti-AKT1 monoclonal antibody which is specific for the AKT1 isoform. C: Western developed using anti-phospho308AKT serum. D: Western developed using anti-phosphoGSK3 serum. (Rabbit sera raised against an AKT1 decapeptide containing phosphoserine-473 or raised against a similar peptide containing Akt1 phosphothreonine-308 are referred to here as anti-phospho473AKT and anti-phospho308AKT sera, respectively.)

 
Flow cytometric analysis of serum-deprived cultures.
MERAkt cells were selected from a population of NIH3T3 cells infected with a retrovirus encoding MERAkt (22). The retrovirus integrates randomly in the cellular genome, and the encoded protein, MERAKT, is inactive (lanes 1 and 6 of Fig. 2A; see Ref. 22). It was therefore expected that MERAkt cells would not differ phenotypically from the parental NIH3T3 cells. However, flow cytometric analysis of MERAkt and NIH3T3 nuclei demonstrated that MERAkt cells are more resistant to serum deprivation than the parental NIH3T3 cells. Cells were grown in reduced serum media for various time periods, and nuclei were harvested, stained with propidium iodide, and analyzed by flow cytometry. The results (data not shown) indicated that 24 h in 0.5% FBS is sufficient to reduce the percent NIH3T3 cells in S phase to 2%, whereas after 48 h in 0.5% FBS the percent MERAkt cells in S phase does not drop below 8%. A significant percentage of MERAkt cells continue to progress through the cell cycle despite serum deprivation. This observation cannot be attributed to MERAKT-dependent AKT1 activity, because we have shown that MERAKT is inactive in the absence of ER ligands. We note that phenotypic changes acquired during retroviral transfection and selection that appear to be independent of retrovirally encoded genes are not unusual but do require characterization. This observed difference between the parental and retrovirus-derived cell lines led us to determine whether growth factor signal transduction in MERAkt cells correlated with published results obtained using NIH3T3 (parental) cells.

PDGFß-regulated genes.
MERAkt cells stimulated with 10 ng/ml PDGFß were analyzed by Western blot using anti-phospho473AKT and anti-phosphoGSK3 sera (Fig. 3, A and B). PDGFß stimulated transient activation of endogenous AKT1 with an ~50-fold increase in phosphorylation by 15 min, which then decreased 10-fold by 3 h (lanes 3 and 4 of Fig. 3A). There was a coordinate increase in phosphorylation of GSK3, particularly of GSK3-{alpha} at 15 min, returning to basal levels by 3 h (lanes 3 and 4 of Fig. 3B). We note that handling cultures was sometimes sufficient to activate MERAKT, as evident in lane 2 of Fig. 3A, but this activation was 20- to 50-fold less than the level of phosphorylative activation induced by treatment with any of the potent MERAKT activators such as Tam (Fig. 2C, lane 3; Fig. 4, lane 1). This relatively low-level activation of MERAKT did not result in phosphorylation of GSK3 (Fig. 3B, lane 2), indicating that signal transduction from MERAKT was undetectable under these conditions. It follows that the low level of MERAKT phosphorylation seen in lane 3 of Fig. 3A results from handling of the culture during PDGFß treatment but does not contribute to the increase in GSK3 phosphorylation observed in lane 3 of Fig. 3B.



View larger version (46K):
[in this window]
[in a new window]
 
Fig. 3. Western analysis of platelet-derived growth factor-ß (PDGFß)-dependent phosphorylation of endogenous AKT1 and GSK3. Nearly confluent MERAkt cells were grown 48 h in 0.5% FBS, then treated with 10 ng/ml human PDGFß or vehicle control (VC) and harvested at 15 min or 3 h. A: developed using anti-phospho473AKT. B: developed using anti-phosphoGSK3 serum. Lane 1, VC at 15 min; lane 2, VC at 3 h; lane 3, PDGFß at 15 min; and lane 4, PDGFß at 3 h.

 


View larger version (61K):
[in this window]
[in a new window]
 
Fig. 4. ER ligand-induced activation of MERAKT. MERAkt cells were incubated in 0.5% FBS for 48 h prior to stimulation for 1 h with the indicated ER ligand, twofold serial diluted from 0.2 µM to 6.25 nM (lanes 1–6). The Western blots were developed using anti-phospho473AKT serum.

 
PDGFß-regulated genes were identified by cDNA array hybridization analysis of 3-h stimulated MERAkt samples (Table 1). The results are in good agreement with PDGFß-regulated genes identified in NIH3T3 cells overexpressing a chimeric macrophage colony-stimulating factor (MCSF)-PDGFß receptor by Fambrough et al. (13). They report that nine of the genes represented on the cDNA arrays we used are induced at 2–4 h after PDGFß stimulation; we identified eight of these nine genes. Based on the similarity of our results with those of Fambrough et al. (13), we conclude that Akt-dependent cytokine signaling was not detectably perturbed by introduction of the MERAkt retrovirus into NIH3T3 cells. We also identified 11 additional PDGFß-induced genes and 9 genes that were repressed by PDGFß stimulation.


View this table:
[in this window]
[in a new window]
 
Table 1. PDGF-ß-regulated genes (3 h)

 
Activation of MERAKT by ER ligands.
Previous work had shown that E2 binds the mutant murine ER-HBD that controls the activation of MERAKT several logs more weakly than Tam binds to the same ER-HBD (27), but the relative binding affinity of the other ER ligands was unknown. Western blot analysis was used to determine the EC50 values for activation of MERAKT by each of seven ER ligands: Tam, Ral, ICI, ZK955, ZK819, ZK427, and E2. Serum-starved MERAkt cultures were treated in a twofold serial dilution series with each of the ER ligands for 2 h (Fig. 4). The ER ligands tested fall into two classes, potent and weak MERAKT activators, based on the magnitude of their EC50 values (Table 2).


View this table:
[in this window]
[in a new window]
 
Table 2. Activation of MERAKT by estrogen receptor ligands

 
AKT-regulated genes.
Replicate sets of data derived from MERAkt cultures treated with each of the four potent MERAKT activators (Tam, Ral, ICI, and ZK955) and the two weak MERAKT activators (E2 and ZK427) were analyzed by comparison to untreated reference cultures to identify genes differentially regulated by treatment with these ER ligands. The 0.1 µM concentration of ER ligand used to stimulate AKT1 activity was about twice the EC50 of the four potent MERAKT activators. Higher concentrations of ER ligands were not used, to avoid nonphysiological conditions.

Data were analyzed by two methods: GEPView, which reports the fold change in the intensity of the signals in matched treated vs. untreated samples for each gene and each treatment, and by a statistical analysis based on a modification of Student’s t-test. GEPView analysis facilitates comparison of the gene regulation effects of the individual treatments used to deduce the mechanism of gene regulation. Lists of genes identified as differentially expressed, using a twofold change cutoff, in response to each treatment were compiled and sorted to identify genes differentially regulated by the greatest number of treatments. The results of this sorting are presented in Tables 3 and 4. No genes were differentially regulated by all six ER ligand treatments. Those genes most frequently regulated by the treatments were commonly regulated by the four potent MERAKT activators and infrequently regulated by the weak MERAKT activators, E2 and ZK427. No genes were differentially regulated by both E2 and ZK427. Fold changes in gene signal intensity between zero and twofold are indicated by a "•". Genes are listed in order of descending confidence that they are AKT regulated. Confidence that a given gene is AKT regulated is proportional to the number of experiments in which the gene’s expression was affected by any of the four potent MERAKT activators and is inversely proportional to the number of times the gene was affected by either of the weak MERAKT activators.


View this table:
[in this window]
[in a new window]
 
Table 3. AKT-regulated genes (3 h)

 

View this table:
[in this window]
[in a new window]
 
Table 4. AKT-regulated genes (20 h)

 
The second analytical method for evaluating gene expression changes relied on statistical analysis based on Student’s t-test. Only data sets from potent MERAKT activator-treated samples were analyzed. The analysis generated inclusive lists of AKT-regulated genes because the twofold change criterion was not applied. The complete results of this statistical analysis are available from us upon request. The results in Tables 5 and 6 show only those genes most likely to contribute to AKT-mediated oncogenesis (discussed below). The P values reported are uncorrected for multiple tests. The corrected P values remained below 0.05 for a single gene, c-myc (P = 0.0059). Early growth response-3 (EGR3) gave a corrected P value of 0.053. The quantile-quantile plot presented in Fig. 5 identifies outlying t-test statistic values. The gene c-myc is a clear outlier that also shows a strictly significant P value after correction for multiple tests; c-myc is expressed at different levels upon treatment. Analysis with GEPView was compared with the t statistics calculated; the rank order of confidence that these genes are AKT-regulated was found to be consistent between the two analytical methods.


View this table:
[in this window]
[in a new window]
 
Table 5. AKT-regulated genes (3 h)

 

View this table:
[in this window]
[in a new window]
 
Table 6. AKT-regulated genes (20 h)

 


View larger version (13K):
[in this window]
[in a new window]
 
Fig. 5. Quantile-quantile plot showing results of statistical analysis comparing the expression levels of each of the 1,176 genes in treated vs. untreated samples at 3 h. Strict corrected P values are in gray, and the cutoff values are indicated with dashed lines. The data point shown in gray corresponds to the gene c-myc, a gene showing a significant t statistic value.

 
Insulin-like growth factor binding protein 1 (IGFBP-1) and prolactin receptor are strongly induced (5- to 30-fold) by ER ligands in many ER+ cell types (41). It is noteworthy that neither IGFBP-1 nor prolactin receptor, both represented on the arrays used, was identified as differentially regulated by ER ligand treatment. Such weak or undetectable ER-dependent gene regulation in MERAkt cells is consistent with the report that NIH3T3 cells (from which MERAkt cells are derived) express only low levels of ER (21).

To determine whether any of the genes identified in Tables 3 or 4 could be ER ligand regulated by an AKT-independent mechanism, we analyzed data from MERAkt cultures treated with E2 or ZK427. E2 regulates ER{alpha}- and ERß-dependent genes on average one log more strongly than any of the four potent MERAkt activators, and ZK427 regulates ER{alpha}-dependent genes (41). As shown in Tables 3 and 4, none of the genes commonly regulated by the four potent MERAKT activators was consistently or strongly regulated by either E2 or ZK427, indicating that the mechanism of their regulation in unlikely to be ER dependent.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Akt-regulated genes were identified at 3 and 20 h by cDNA hybridization array analysis of replicate samples stimulated with each of four potent MERAKT activators. The four potent MERAKT activators are chemically distinct and therefore are expected to affect gene regulation in independent manners. They are similar only in their potent activation of MERAKT, i.e., AKT1 activity. We hypothesize that AKT1 regulation is the primary mechanism responsible for the changes in gene expression observed for those genes differentially expressed in response to treatment by all four potent MERAKT activators.

This study depended on the use of ER ligands to induce AKT1 activity via binding to the mutant ER-HBD of MERAKT. Therefore, we investigated whether regulation of any of the genes commonly regulated by the four potent MERAKT activators could be ER dependent and AKT1 independent. We did not expect to observe ER-dependent regulation by any of the four potent MERAKT activators because 1) all four compounds are very weak inducers of ER-dependent gene expression, at least one log less potent than E2 (41), and 2) the NIH3T3 cell line used is reported to express only low levels of ER (21). We showed that ER-dependent genes (IGFBP-1 and prolactin receptor) were not differentially expressed in our system in response to E2 treatment, indicating that the functional level of ER in the cells used for this study is below that required for ER-dependent gene regulation. Additionally, we showed that the genes commonly regulated by the four potent MERAKT activators are not regulated by treatment with E2 or ZK427. E2 and ZK427 are strong and moderate inducers of ER-dependent gene expression respectively, but both are relatively weak MERAKT activators. If the genes commonly regulated by the four potent MERAKT activators were regulated by an ER-dependent mechanism, then they would have been more strongly regulated by E2 or ZK427 than by the four potent MERAKT activators. The opposite was observed. Last, ER-independent regulation common to the four potent MERAKT activators is unlikely given the dissimilarity of their chemical structures. Ral is a benzothiophene, ICI is a 7{alpha}-substituted E2, and Tam and ZK955 are triphenylethylenes (41). These compounds are unlikely to have similar effects on gene expression independent of their common ability to bind to the mutant ER-HBD of MERAKT. We therefore conclude that genes commonly regulated by the four potent MERAKT activators are regulated by AKT1 activity.

This is the first time that global expression analysis has been used to identify Akt-regulated genes. The small-magnitude changes are consistent with those reported for Akt regulation of Bcl-2 (34), one of the only Akt-regulated genes for which quantitative data has been reported. Similarly small magnitude changes were recently reported for PTEN-regulated genes identified by global expression analysis. PTEN ("Phosphatase with tensin homology") is a negative regulator of AKT activation, acting immediately upstream of AKT in the PI3-kinase pathway. Forty-two of the 43 genes reported changed two- to threefold. Only one gene was greater than threefold affected by PTEN (18). Small magnitude changes in gene expression may be common for these pathways.


Akt regulation and oncogenesis.
The function of many of the Akt-regulated genes is consistent with tumor initiation or maintenance through promotion of cell cycle progression, inhibition of apoptosis, and enhancement of angiogenesis. The function of other Akt-regulated genes contrasts with Akt’s oncogenic phenotype, and the explanation for their regulation by Akt remains unclear.

Akt-regulated induction of 78-kDa glucose-regulated protein (GRP) and BAG-1 are consistent with AKT’s anti-apoptotic function (40). Induction of MECP-2 (represses tumor suppressor gene expression), matrix metalloproteinase 2, extracellular matrix metalloproteinase inducer, 58-kDa inhibitor of RNA-activated protein kinase (an oncogene) (1), and repression of N-cadherin are all consistent with tumor formation and invasion or metastasis. Most classes of cathepsins (L, B1, D, H, W), proteases whose expression correlates with aggressive tumor growth (38), are induced by AKT at both 3 and 20 h.

C-myc, N-myc, and L-myc are all repressed by AKT1 activity. This finding contrasts with frequent reports of c-myc upregulation in cancers, most clearly illustrated in colorectal cancer (16). There are several ways in which these data may be reconciled. AKT increases translation of 5'-terminal oligo pyrimidine track (5'-TOP) mRNAs, including that of c-myc (22). The increased C-MYC levels may negatively regulate c-myc transcription, a testable hypothesis that may reconcile our results with those of He et al. (16). However, the kinetics of c-myc mRNA repression upon activation of AKT1 argues against this explanation. A second hypothesis is that the mechanism of AKT oncogenesis relies on inhibition of apoptosis rather than induction of mitogenesis (e.g., by C-MYC). Upregulation of c-myc can induce apoptosis (10); repression of c-myc may therefore be part of Akt’s anti-apoptotic mechanism. Finally, our observation that c-myc is induced by PDGFß stimulation (which activates AKT1) indicates that AKT1 repression of c-myc expression may be physiologically irrelevant. AKT was activated in our study in isolation of upstream signaling. AKT is unlikely to be activated in isolation under physiological conditions. Our results comparing PDGFß and AKT1 gene regulation indicate that signaling from upstream pathway elements may negatively interfere with c-myc signaling initiated by AKT1, discussed below.

Tumor neovascularization may be promoted by AKT-regulated coordinate repression of EGR1, thrombospondin-1 (THBS1), and several transforming growth factor ß (TGF-ß) family ligands and receptors. EGR1 suppresses tumor formation by stimulating the TGF-ß1 and fibronectin promoters, leading to enhanced cell attachment and regulated growth (28). Repression of EGR1 would allow anchorage-independent growth, loss of regulation, and reduced TGF-ß1 levels. TGF-ß1 normally cooperates with THBS1 in suppressing endothelial cell proliferation (37). Reductions in TGF-ß and TGF-ß receptor levels combined with reductions in THBS1 levels are therefore expected to promote angiogenesis and tumor neovascularization. Tumor neovascularization would be further aided by AKT-regulated induction of VEGF mRNA (observed at 20 h) and other HIF1{alpha}-regulated genes, and phosphorylative activation of endothelial constitutive nitric oxide synthase (ecNOS) (14).

There are a number of genes whose regulation by AKT1 is inconsistent with AKT’s anti-apoptotic or oncogenic functions. Apoptotic factors include AKT1-regulated induction of BAX, T-cell death-associated gene 51 (TDAG51), BAK1, and repression of SMN ("survival of motor neurons"). Oncogenic factors include repression of oncogenes pim1, Erb-B2, N-ras, and Met, and induction of TIMP3 ("tissue inhibitor of metalloproteinases 3"). We have no explanation for these apparent inconsistencies but note similar findings in a recent study of Raf-1-regulated genes. Many of the Raf-1-regulated genes functioned to inhibit, rather than promote, oncogenesis (17).

Akt regulation is cell-type specific.
Observations of gene regulation depend on the cell type studied. The following are examples of genes whose AKT regulation has been reported in other systems but which were not identified as AKT-regulated genes in our study. It has been reported that AKT regulates cell survival and induces cell cycle progression by phosphorylative inhibition of the transcription factor FKHRL1, consequently reducing p27kip1 expression (24, 25). MERAkt cells express low levels of FKHRL1 (unpublished results), possibly explaining why significant changes in p27kip1 mRNA levels were not observed. Similarly, AKT regulates VEGF expression under hypoxic conditions, by regulating the transcription factor HIF1{alpha} (29). Under the normoxic conditions here, we did not observe AKT regulation of VEGF expression at 3 h and observed only a minimal increase at 20 h. Additionally, very low steady-state levels of HIF1{alpha} mRNA were measured in these samples. GLUT-1 mRNA levels are affected by AKT1 activity (3), but GLUT-1 regulation was not observed in MERAkt cells, presumably because components of required signaling pathways are not expressed in these cells. The PEPCK gene was not represented on the arrays used in this study.

This is the first time that signaling from a pathway element (AKT1) has been compared with signaling initiated at the cell surface (PDGFß stimulation), allowing detection of interference of signals originating in a single pathway. Comparison of AKT1-regulated and PDGFß-regulated genes reported here shows that AKT1-regulated genes are not a subset of PDGFß-regulated genes. The only two genes that were similarly regulated by AKT1 and PDGFß were TDAG51 and 78-kDa GRP. Genes that are regulated by AKT1 but not by PDGFß are evidence of signals originating upstream of AKT that block, or negatively interfere with, AKT signaling regulating expression of that gene. More striking, there are a number of genes, including c-Myc, EGR1, and THBS1 that are oppositely regulated by AKT1 and PDGFß, indicating that signals from PDGFß elements upstream of AKT are dominant over those originating from AKT.

In conclusion, a number of genes were identified as AKT1-regulated under the conditions tested. Many of these genes may contribute to AKT1-mediated oncogenic transformation, tumor growth, or maintenance. Further work is needed to determine whether protein levels or activities increase or decrease in concert with the changes in steady-state mRNA levels reported here. To determine which of these changes in mRNA levels contribute to AKT-mediated oncogenesis, overexpression or antisense methods could be employed.


    ACKNOWLEDGMENTS
 
This work was supported in large part by the Berlex Laboratories Advanced Degree Program (to I. Kuhn), by the Biochemistry and Molecular Biology Graduate Program of University of California, Davis (to I. Kuhn and P. H. Johnson), and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-34976 (to R. A. Roth).

Present address of P. H. Johnson: EpiGenx Pharmaceuticals, 5385 Hollister Ave., Santa Barbara, CA 93111.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: I. Kuhn, Berlex Biosciences, 15049 San Pablo Ave., Richmond, CA 94804-0099 (E-mail: Irene_Kuhn{at}Berlex.com).

1 By convention, in this report, gene names are indicated by lower case abbreviations, and protein product names are indicated by upper case abbreviations. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Barber GN, Thompson S, Lee TG, Strom T, Jagus R, Darveau A, and Katze MG. The 58-kilodalton inhibitor of the interferon-induced double-stranded RNA-activated protein kinase is a tetratricopeptide repeat protein with oncogenic properties. Proc Natl Acad Sci USA 91: 4278–4282, 1994.[Abstract]
  2. Barthel A, Kohn AD, Luo Y, and Roth RA. A constitutively active version of the Ser/Thr kinase Akt induces production of the ob gene product, leptin, in 3T3-L1 adipocytes. Endocrinology 138: 3559–3562, 1997.[Abstract/Free Full Text]
  3. Barthel A, Okino ST, Liao J, Nakatani K, Li J, Whitlock JP Jr, and Roth RA. Regulation of GLUT1 gene transcription by the serine/threonine kinase Akt1. J Biol Chem 274: 20281–20286, 1999.[Abstract/Free Full Text]
  4. Bellacosa A, de Feo D, Godwin AK, Bell DW, Cheng JQ, Altomare DA, Wan M, Dubeau L, Scambia G, and Masciullo V. Molecular alterations of the AKT2 oncogene in ovarian and breast carcinomas. Int J Cancer 64: 280–285, 1995.[ISI][Medline]
  5. Biggs WH III, Meisenhelder J, Hunter T, Cavenee WK, and Arden KC. Protein kinase B/Akt-mediated phosphorylation promotes nuclear exclusion of the winged helix transcription factor FKHR1. Proc Natl Acad Sci USA 96: 7421–7426, 1999.[Abstract/Free Full Text]
  6. Blume-Jensen P and Hunter T. Oncogenic kinase signalling. Nature 411: 355–365, 2001.[ISI][Medline]
  7. Brennan P, Babbage JW, Burgering BM, Groner B, Reif K, and Cantrell DA. Phosphatidylinositol 3-kinase couples the interleukin-2 receptor to the cell cycle regulator E2F. Immunity 7: 679–689, 1997.[ISI][Medline]
  8. Callow MJ, Dudoit S, Gong EL, Speed TP, and Rubin EM. Microarray expression profiling identifies genes with altered expression in HDL-deficient mice. Genome Res 10: 2022–2029, 2000.[Abstract/Free Full Text]
  9. Chan TO, Rittenhouse SE, and Tsichlis PN. AKT/PKB and other D3 phosphoinositide-regulated kinases: kinase activation by phosphoinositide-dependent phosphorylation. Annu Rev Biochem 68: 965–1014, 1999.[ISI][Medline]
  10. Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, and Greenberg ME. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91: 231–241, 1997.[ISI][Medline]
  11. del Peso L, Gonzalez VM, Hernandez R, Barr FG, and Nunez G. Regulation of the forkhead transcription factor FKHR, but not the PAX3-FKHR fusion protein, by the serine/threonine kinase Akt. Oncogene 18: 7328–7333, 1999.[ISI][Medline]
  12. Eder AM, Dominguez L, Franke TF, and Ashwell JD. Phosphoinositide 3-kinase regulation of T cell receptor-mediated interleukin-2 gene expression in normal T cells. J Biol Chem 273: 28025–28031, 1998.[Abstract/Free Full Text]
  13. Fambrough D, McClure K, Kazlauskas A, and Lander ES. Diverse signaling pathways activated by growth factor receptors induce broadly overlapping, rather than independent, sets of genes. Cell 97: 727–741, 1999.[ISI][Medline]
  14. Fulton D, Gratton JP, McCabe TJ, Fontana J, Fujio Y, Walsh K, Franke TF, Papapetropoulos A, and Sessa WC. Regulation of endothelium-derived nitric oxide production by the protein kinase Akt. Nature 399: 597–601, 1999.[ISI][Medline]
  15. Gille H and Downward J. Multiple ras effector pathways contribute to G(1) cell cycle progression. J Biol Chem 274: 22033–22040, 1999.[Abstract/Free Full Text]
  16. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, and Kinzler KW. Identification of c-MYC as a target of the APC pathway. Science 281: 1509–1512, 1998.[Abstract/Free Full Text]
  17. Heinrich J, Bosse M, Eickhoff H, Nietfeld W, Reinhardt R, Lehrach H, and Moelling K. Induction of putative tumor-suppressing genes in Rat-1 fibroblasts by oncogenic Raf-1 as evidenced by robot-assisted complex hybridization. J Mol Med 78: 380–388, 2000.[ISI][Medline]
  18. Hong TM, Yang PC, Peck K, Chen JJ, Yang SC, Chen YC, and Wu CW. Profiling the downstream genes of tumor suppressor PTEN in lung cancer cells by complementary DNA microarray. Am J Respir Cell Mol Biol 23: 355–363, 2000.[Abstract/Free Full Text]
  19. Jung YD, Nakano K, Liu W, Gallick GE, and Ellis LM. Extracellular signal-regulated kinase activation is required for up-regulation of vascular endothelial growth factor by serum starvation in human colon carcinoma cells. Cancer Res 59: 4804–4807, 1999.[Abstract/Free Full Text]
  20. Kane LP, Shapiro VS, Stokoe D, and Weiss A. Induction of NF-{kappa}B by the Akt/PKB kinase. Curr Biol 9: 601–604, 1999.[ISI][Medline]
  21. Kato K, Ueoka Y, Hachiya T, Nishida J, and Wake N. Contribution of enhanced transcriptional activation by ER to [12Val] K-Ras mediated NIH3T3 cell transformation. Oncogene 15: 3037–3046, 1997.[ISI][Medline]
  22. Kohn AD, Barthel A, Kovacina KS, Boge A, Wallach B, Summers SA, Birnbaum MJ, Scott PH, Lawrence JC Jr, and Roth RA. Construction and characterization of a conditionally active version of the serine/threonine kinase Akt. J Biol Chem 273: 11937–11943, 1998.[Abstract/Free Full Text]
  23. Kohn AD, Takeuchi F, and Roth RA. Akt, a pleckstrin homology domain containing kinase, is activated primarily by phosphorylation. J Biol Chem 271: 21920–21926, 1996.[Abstract/Free Full Text]
  24. Kops GJ and Burgering BM. Forkhead transcription factors: new insights into protein kinase B (c-akt) signaling. J Mol Med 77: 656–665, 1999.[ISI][Medline]
  25. Kops GJ, de Ruiter ND, De Vries-Smits AM, Powell DR, Bos JL, and Burgering BM. Direct control of the Forkhead transcription factor AFX by protein kinase B. Nature 398: 630–634, 1999.[ISI][Medline]
  26. Liao J, Barthel A, Nakatani K, and Roth RA. Activation of protein kinase B/Akt is sufficient to repress the glucocorticoid and cAMP induction of phosphoenolpyruvate carboxykinase gene. J Biol Chem 273: 27320–27324, 1998.[Abstract/Free Full Text]
  27. Littlewood TD, Hancock DC, Danielian PS, Parker MG, and Evan GI. A modified oestrogen receptor ligand-binding domain as an improved switch for the regulation of heterologous proteins. Nucleic Acids Res 23: 1686–1690, 1995.[Abstract]
  28. Liu C, Yao J, de Belle I, Huang RP, Adamson E, and Mercola D. The transcription factor EGR-1 suppresses transformation of human fibrosarcoma HT1080 cells by coordinated induction of transforming growth factor-ß1, fibronectin, and plasminogen activator inhibitor-1. J Biol Chem 274: 4400–4411, 1999.[Abstract/Free Full Text]
  29. Mazure NM, Chen EY, Laderoute KR, and Giaccia AJ. Induction of vascular endothelial growth factor by hypoxia is modulated by a phosphatidylinositol 3-kinase/Akt signaling pathway in Ha-ras-transformed cells through a hypoxia inducible factor-1 transcriptional element. Blood 90: 3322–3331, 1997.[Abstract/Free Full Text]
  30. Medema RH, Kops GJ, Bos JL, and Burgering BM. AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404: 782–787, 2000.[ISI][Medline]
  31. Mirza AM, Kohn AD, Roth RA, and McMahon M. Oncogenic transformation of cells by a conditionally active form of the protein kinase Akt/PKB. Cell Growth Differ 11: 279–292, 2000.[Abstract/Free Full Text]
  32. Nakae J, Park BC, and Accili D. Insulin stimulates phosphorylation of the forkhead transcription factor FKHR on serine 253 through a Wortmannin-sensitive pathway. J Biol Chem 274: 15982–15985, 1999.[Abstract/Free Full Text]
  33. Pekarsky Y, Koval A, Hallas C, Bichi R, Tresini M, Malstrom S, Russo G, Tsichlis P, and Croce CM. Tcl1 enhances Akt kinase activity and mediates its nuclear translocation. Proc Natl Acad Sci USA 97: 3028–3033, 2000.[Abstract/Free Full Text]
  34. Pugazhenthi S, Nesterova A, Sable C, Heidenreich KA, Boxer LM, Heasley LE, and Reusch JE. Akt/protein kinase B up-regulates Bcl-2 expression through cAMP-response element-binding protein. J Biol Chem 275: 10761–10766, 2000.[Abstract/Free Full Text]
  35. Rogatsky I, Waase CL, and Garabedian MJ. Phosphorylation and inhibition of rat glucocorticoid receptor transcriptional activation by glycogen synthase kinase-3 (GSK-3). Species-specific differences between human and rat glucocorticoid receptor signaling as revealed through GSK-3 phosphorylation. J Biol Chem 273: 14315–14321, 1998.[Abstract/Free Full Text]
  36. Shayesteh L, Lu Y, Kuo WL, Baldocchi R, Godfrey T, Collins C, Pinkel D, Powell B, Mills GB, and Gray JW. PIK3CA is implicated as an oncogene in ovarian cancer. Nat Genet 21: 99–102, 1999.[ISI][Medline]
  37. Sheibani N and Frazier WA. Thrombospondin 1 expression in transformed endothelial cells restores a normal phenotype and suppresses their tumorigenesis. Proc Natl Acad Sci USA 92: 6788–6792, 1995.[Abstract]
  38. Soderstrom M, Ekfors T, Bohling T, Aho A, Aro HT, and Vuorio E. Cysteine proteinases in chondrosarcomas. Matrix Biol 19: 717–725, 2001.[ISI][Medline]
  39. Tang ED, Nunez G, Barr FG, and Guan KL. Negative regulation of the forkhead transcription factor FKHR by Akt. J Biol Chem 274: 16741–16746, 1999.[Abstract/Free Full Text]
  40. Yu Z, Luo H, Fu W, and Mattson MP. The endoplasmic reticulum stress-responsive protein GRP78 protects neurons against excitotoxicity and apoptosis: suppression of oxidative stress and stabilization of calcium homeostasis. Exp Neurol 155: 302–314, 1999.[ISI][Medline]
  41. Zajchowski DA, Kauser K, Zhu D, Webster L, Aberle S, White FA, III, Liu HL, Humm R, MacRobbie J, Ponte P, Hegele-Hartung C, Knauthe R, Fritzemeier KH, Vergona R, and Rubanyi GM. Identification of selective estrogen receptor modulators by their gene expression fingerprints. J Biol Chem 275: 15885–15894, 2000.[Abstract/Free Full Text]