The Integrin-linked Kinase Regulates the Cyclin D1 Gene through Glycogen Synthase Kinase 3β and cAMP-responsive Element-binding Protein-dependent Pathways*

The cyclin D1 gene encodes the regulatory subunit of a holoenzyme that phosphorylates and inactivates the pRB tumor suppressor protein. Cyclin D1 is overexpressed in 20–30% of human breast tumors and is induced both by oncogenes including those for Ras, Neu, and Src, and by the β-catenin/lymphoid enhancer factor (LEF)/T cell factor (TCF) pathway. The ankyrin repeat containing serine-threonine protein kinase, integrin-linked kinase (ILK), binds to the cytoplasmic domain of β1 and β3integrin subunits and promotes anchorage-independent growth. We show here that ILK overexpression elevates cyclin D1 protein levels and directly induces the cyclin D1 gene in mammary epithelial cells. ILK activation of the cyclin D1 promoter was abolished by point mutation of a cAMP-responsive element-binding protein (CREB)/ATF-2 binding site at nucleotide −54 in the cyclin D1 promoter, and by overexpression of either glycogen synthase kinase-3β (GSK-3β) or dominant negative mutants of CREB or ATF-2. Inhibition of the PI 3-kinase and AKT/protein kinase B, but not of the p38, ERK, or JNK signaling pathways, reduced ILK induction of cyclin D1 expression. ILK induced CREB transactivation and CREB binding to the cyclin D1 promoter CRE. Wnt-1 overexpression in mammary epithelial cells induced cyclin D1 mRNA and targeted overexpression of Wnt-1 in the mammary gland of transgenic mice increased both ILK activity and cyclin D1 levels. We conclude that the cyclin D1 gene is regulated by the Wnt-1 and ILK signaling pathways and that ILK induction of cyclin D1 involves the CREB signaling pathway in mammary epithelial cells.

The cyclin D1 gene encodes the regulatory subunit of a holoenzyme that phosphorylates and inactivates the pRB tumor suppressor protein. Cyclin D1 is overexpressed in 20 -30% of human breast tumors and is induced both by oncogenes including those for Ras, Neu, and Src, and by the ␤-catenin/lymphoid enhancer factor (LEF)/T cell factor (TCF) pathway. The ankyrin repeat containing serine-threonine protein kinase, integrinlinked kinase (ILK), binds to the cytoplasmic domain of ␤ 1 and ␤ 3 integrin subunits and promotes anchorageindependent growth. We show here that ILK overexpression elevates cyclin D1 protein levels and directly induces the cyclin D1 gene in mammary epithelial cells. ILK activation of the cyclin D1 promoter was abolished by point mutation of a cAMP-responsive element-binding protein (CREB)/ATF-2 binding site at nucleotide ؊54 in the cyclin D1 promoter, and by overexpression of either glycogen synthase kinase-3␤ (GSK-3␤) or dominant negative mutants of CREB or ATF-2. Inhibition of the PI 3-kinase and AKT/protein kinase B, but not of the p38, ERK, or JNK signaling pathways, reduced ILK induction of cyclin D1 expression. ILK induced CREB transactivation and CREB binding to the cyclin D1 promoter CRE. Wnt-1 overexpression in mammary epithelial cells induced cyclin D1 mRNA and targeted overexpression of Wnt-1 in the mammary gland of transgenic mice increased both ILK activity and cyclin D1 levels. We conclude that the cyclin D1 gene is regulated by the Wnt-1 and ILK signaling pathways and that ILK induction of cyclin D1 involves the CREB signaling pathway in mammary epithelial cells.
The cyclin D1 gene encodes a regulatory subunit of a serinethreonine kinase that phosphorylates and inactivates the tumor suppressor pRB (1). The abundance of cyclin D1 was shown to be rate-limiting in cellular proliferation induced by diverse signaling pathways in fibroblasts and breast epithelial cells, including MCF7 cells (2,3). Homozygous deletion of the cyclin D1 gene in mice results in defects in mammary gland development (4,5) and in serum-induced cellular proliferation (6). The abundance of cyclin D1 is increased in more than 30% of human breast tumors, and overexpression of cyclin D1 under control of the MMTV 1 promoter in transgenic mice induces mammary adenocarcinoma (7). The majority of breast cancer cell lines and mammary tumors induced by transgenic overexpression of either pp60 v-src or ErbB-2 oncogenes overexpress cyclin D1, suggesting that the induction of cyclin D1 may play an important role in mammary tumorigenesis (8). The cyclin D1 gene is activated by mitogenic stimuli induced by G-protein-coupled receptors (9), tyrosine kinase receptors (10) and by small monomeric GTPases of the Ras/Rac family (11)(12)(13). Recent studies implicated components of the ␤-catenin/T cell factor (TCF)/lymphoid enhancer factor 1 (LEF-1) pathway in activating the cyclin D1 gene in colon cancer cell lines (14,15) and HeLa cells (15).
Wnt signaling plays an important role in development of both vertebrates and invertebrates (16,17), and deregulated Wnt signaling is considered a significant factor in the development of several human cancers (18,19). The first member of the Wnt family was isolated as the product of the cellular oncogene, wnt-1, and is activated by proviral insertion in murine mammary carcinomas (20). Targeted overexpression of Wnt-1 under control of the MMTV promoter was also shown to induce mammary adenocarcinoma (21). Control of ␤-catenin stability, which is central to Wnt signaling, is regulated by the serine-threonine kinase, glycogen synthase kinase-3␤ (GSK-3␤). GSK-3␤ inhibits diverse signaling pathways including AP-1 (22), CREB (23), and TCF/LEF (24) activity, in part through phosphorylation-dependent events that reduce transcription factor binding to their cognate DNA sequences. Recent studies have provided evidence for an important role for the CRE in Wnt signaling as the CREB site can collaborate with a LEF site in regulating transcription of TCR␣ (25) and the CRE site of the proto-oncogene WISP-1 gene was required for induction by ␤-catenin (26).
Recent studies of components of the Wnt signaling pathway have demonstrated a role for the ankyrin repeat containing serine-threonine kinase, integrin-linked kinase (ILK) (27). ILK was first identified as a protein that binds to the cytoplasmic domain of the ␤ 1 integrin subunit (28), and whose activity is induced by integrin clustering, by interactions with the extracellular matrix, and by growth factors (27,29,30). ILK activates protein kinase B (PKB/AKT) (31) and inhibits GSK-3␤ activity (32). The anchorage-independent growth induced by ILK overexpression is associated with the induction of cyclin D1 expression (33). Since cyclin D1 levels are reduced when cells become contact inhibited at higher density (34) and when cells are not anchored to the extracellular matrix (35), it is possible that the cyclin D1 gene is regulated by integrin and ILK-dependent signaling pathways.
In this study, we show that activation of the Wnt pathway in transgenic mice is associated with induction of both ILK activity and cyclin D1 expression. ILK overexpression activated the cyclin D1 promoter through a CREB/ATF-2 binding site and involved the phosphatidylinositol (PI) 3-kinase/AKT pathway. ILK enhanced both CREB transactivation and binding of CREB to the cyclin D1 promoter CRE. GSK-3␤ inhibited ILKinduced cyclin D1 promoter activity and CREB binding to the cyclin D1 CRE. The identification of the cyclin D1 promoter as a direct target of ILK underscores a likely mechanism by which integrin signaling may enhance DNA synthesis.

Construction of Reporter Genes and Expression Vectors-
The human cyclin D1 promoter reporter constructions (9, 10) PALUC, which contains 7 kilobase pairs of the human cyclin A promoter sequence (36), and the reporter gene pA 3 LUC, was previously described. The reporter (UAS) 5 E1BTATALUC consisting of the (UAS) 5 E1BTATA sequences from Gal 5 CAT cloned in the reporter pA 3 LUC were described previously (37). The integrity of the reporter constructions used was confirmed by sequence analysis.
PCR-based Subtractive Hybridization-A PCR-based subtractive hybridization was performed by using the PCR-Select cDNA subtraction kit from CLONTECH (Palo Alto, CA) following the manufacturer's protocol. Tester double-stranded cDNA was synthesized from 2 g of poly(A) ϩ RNA isolated from the C57MG cells stably transfected with Wnt-1 retrovirus (C57MG/Wnt-1) and driver cDNA from 2 g of poly(A) ϩ RNA from the control LNCX retrovirus-transfected C57MG cells (C57MG/LNCX). The subtracted cDNA library was subcloned with the pGEM-T Easy Vector Systems (Promega, Madison, WI) for further analysis. For Northern blot analysis, total RNA was isolated by guanidium thiocyanate-phenol-chloroform using a single-step extraction (45). RNA samples were resolved on a 1% agarose, 3% formaldehyde gel. Transfer, hybridization, and washing were performed as described by Sambrook et al. (46). Probes were labeled with [␣-32 P]dCTP by use of a random priming kit (Ambion, Austin, TX).
Reporter Assays and Cell Culture-Cell culture, DNA transfection, and luciferase assays were performed as described previously (10). The MCF7 cells and human embryonic kidney cells (HEK-293; obtained from M. Moran, University of Toronto, Toronto, Ontario, Canada) (32) were maintained in Dulbecco's modified Eagle's medium with 10% (v/v) calf serum and 1% penicillin/streptomycin. In transient expression studies, cells were transfected either by calcium phosphate precipitation or the use of Superfect transfection reagent (Qiagen, Valencia, CA) as described by the manufacturer. The medium was changed after 6 h and luciferase activity determined after an additional 24 h. The effect of an expression vector was compared with the effect of an equal amount of vector cassette. Treatments with the MAP kinase/ERK kinase inhibitor PD98059 (10 -20 M) (47), the PI 3-kinase inhibitor LY294002 (100 pM to 50 nM), the p38 MAP kinase inhibitor SB203580 (10 -20 M) (48), and rapamycin (100 pM to 50 nM) (49), were performed for 24 h, and results were compared with Me 2 SO vehicle treatment. Luciferase assays were performed at room temperature using an AutoLumat LB 953 (EG&G Berthold, Bad Wildbad, Germany). Luciferase content was measured by calculating the light emitted during the initial 10 s of the reaction, and the values are expressed in arbitrary light units (9). Statistical analyses were performed using the Mann-Whitney U test with significant differences established as p Ͻ 0.05.
Oligodeoxyribonucleotides and Electrophoretic Mobility Shift Assays (EMSA)-The wild type CRE/ATF site of the cyclin D1 promoter, CD1CREwt, was synthesized as complementary oligodeoxyribonucleotide strands for EMSA (6). The sequence of the cyclin D1 promoter CRE/ATF site oligodeoxyribonucleotide (CD1CREwt) was 5Ј-AAC AAC AGT AAC GTC ACA CGG AC-3Ј. EMSA were performed using nuclear extracts as described previously (32). Cells were washed with ice-cold PBS and harvested by scraping in 1.5 ml of PBS. Cells were pelleted and resuspended in 400 l of buffer A (10 mM HEPES (pH 7.9), 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT, 0.2 mM PMSF). After 10 min of incubation on ice, nuclei were pelleted and then resuspended in 50 l of buffer B (20 mM HEPES-KOH (pH 7.9), 25% glycerol, 420 mM sodium chloride, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM DTT, 0.2 mM PMSF). Proteins were incubated for 20 min on ice and then centrifuged to clear cellular debris. Gel shift assays were performed by incubating 2 g of the nuclear extracts for 20 min at room temperature with the ␥-32 Plabeled oligonucleotides (50 fmol, 100,000 counts/min). The protein-DNA complexes were analyzed by electrophoresis through a 5% polyacrylamide gel, with 0.5ϫ Tris borate, EDTA buffer (TBE: 0.045 M Tris borate, 0.001 M EDTA) and 3.5% glycerol at 180 V for 3-4 h. The specificity of the DNA-protein interaction was established by competition experiment and supershift assays as described (8). The gels were dried and exposed to autoradiographic film (Labscientific Inc., Livingston, NJ).
Immune Kinase Assays on Transgenic Mice Tumors-The genotype of Wnt-1 transgenic mice was determined by isolating tail DNA and Southern blot hybridization with Wnt-1-labeled DNA probes as described previously (50,51).
The ILK kinase assays were performed as described previously (29) using a rabbit immunoaffinity-purified ILK antibody (06-592) (Upstate Biotechnology, Lake Placid, NY) and myelin basic protein as substrate. Tissues from MMTV-Wnt-1 transgenic mice were homogenized by Dounce in lysis buffer (150 mM NaCl, 50 mM HEPES (pH 7.2), 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0.1% Tween 20, 0.1 mM PMSF, 2.5 g/ml leupeptin, and 0.1 mM sodium orthovanadate (Sigma)), at 4°C. Lysates were centrifuged at 10,000 ϫ g for 5 min, and protein concentrations were determined using a modified Bradford assay protocol (Bio-Rad). The supernatants (100 g) were precipitated for 12 h at 4°C with protein A-agarose beads precoated with saturating amounts of the antibody. Immunoprecipitated proteins on beads were washed twice with 1 ml of lysis buffer and twice with kinase buffer (50 mM HEPES (pH 7.0), 10 mM MgCl 2 , 5 mM MnCl 2 , 1 mM DTT). The beads were then resuspended in 40 l of kinase buffer containing the protein substrate (2 g of myelin basic protein), 10 mM ATP, and 5 mCi of [␥-32 P]ATP (6000 Ci/mmol; 1 Ci ϭ 37 GBq, Amersham Pharmacia Biotech). The samples were incubated for 30 min at 30°C with occasional mixing. The samples were boiled in polyacrylamide gel sample buffer containing sodium dodecyl sulfate and separated by electrophoresis. Phosphorylated proteins were quantified after exposure to autoradiographic film (Labscientific, Inc., Livingston, NJ) by densitometry using ImageQuant version 1.11 (Molecular Dynamics Computing Densitometer, Sunnyvale, CA).
Western Blots-The abundance of cyclin D1 protein was determined by Western analysis as described previously (9, 10), using a monoclonal cyclin D1 antibody DCS-6 (NeoMarkers, Fremont, CA) and a guanine nucleotide dissociation inhibitor (GDI) antibody (a generous gift from Dr. Perry Bickel, Washington University, St. Louis, MO) (52) as internal control for protein abundance. Cell homogenates (50 g) were electrophoresed in an 12% SDS-polyacrylamide gel and transferred electrophoretically to a nitrocellulose membrane (Micron Separations Inc., Westborough, MA). After transfer, the gel was stained with Coomassie Blue as a control for blotting efficiency. The blotting membrane was incubated for 2 h at 25°C in T-PBS buffer supplemented with 5% (w/v) dry milk to block nonspecific binding sites. Following a 6 h incubation with primary antibody at a 1:1000 dilution (cyclin D1) or 1:2500 (GDI) in T-PBS buffer containing 0.05% (v/v) Tween 20, the membrane was washed with the same buffer. For detection of cyclin D1, the membrane was incubated with goat anti-mouse horseradish peroxidase secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and washed again. The cyclin D1 protein was visualized by the enhanced chemiluminescence system (Kirkegaard and Perry Laboratories, Gaithersburg, MD).

ILK Induces the Cyclin D1 Gene through a CREB/ATF-2 Binding Site in Mammary Epithelial Cells-Previous studies have
shown that overexpression of ILK induces anchorage-independent growth and elevates cyclin D1 levels (33). To determine whether ILK overexpression induces the cyclin D1 gene, transient expression studies were conducted in MCF7 cells. ILK overexpression induced the full-length cyclin D1 promoter linked to the luciferase reporter (Ϫ1745CD1LUC) by 4.5-fold (Fig. 1A).
Recent studies have shown GSK-3␤ is inhibited by induction of ILK signaling (32). Wild type ILK activates protein kinases B (PKB/AKT) and inhibits GSK-3␤ activity in a PI 3-kinase-dependent manner (31). GSK-3␤ regulates several downstream signaling events, including the function of the CREB transcription factor (23,53). Since the cyclin D1 promoter contains a CREB binding site (37), we examined the effect of wild type ILK on a cyclin D1 promoter containing a point mutation in the CRE/ATF binding site. This point mutation, shown to abolish binding of CREB/ATF-2 to the cyclin D1 promoter sequence (8) (Ϫ1745CD1LUC CREmut), was not induced by ILK (Fig. 1A). Point mutation of the TCF site did not affect regulation of the cyclin D1 promoter by wild type ILK (data not shown). The activation of the cyclin D1 promoter by ILK was specific as reporter plasmids containing the Rous sarcoma virus, cyclin A, and MMTV promoters, and the empty luciferase vector pA 3 LUC, were not induced by ILK overexpression (Fig. 1B). These results suggest that the CREB/ATF-2 binding site of the cyclin D1 promoter is required for its induction by ILK.
To determine the intracellular signaling pathway(s) by which ILK induced cyclin D1 promoter activity, several chemical inhibitors and dominant negative expression vectors were employed. In these studies, the level of the cyclin D1 promoter activity in the presence of ILK was set as 100% ( Fig. 2A). The activity of ILK was previously linked to the PI 3-kinase pathway (31), and ILK contains a phosphoinositide binding motif (30). The addition of the PI 3-kinase inhibitor LY294002 to cells transfected with ILK and the cyclin D1 promoter reporter reduced the ILK-induced cyclin D1 promoter activity by 35% ( Fig. 2A). In contrast, the basal activity of the cyclin D1 promoter in the absence of ILK expression was not inhibited by LY294002 (Fig. 2B). These results are consistent with previous observations showing that the basal activity of the cyclin D1 promoter is not induced by the PI 3-kinase pathway in NIH3T3 cells (54). Addition of rapamycin, which prevents induction of pp70 S6K by its known agonists, also inhibited ILK-induced cyclin D1 promoter activity by 45% (at 100 pM) ( Fig. 2A), without affecting basal promoter activity (Fig. 2B). Addition of the p38 inhibitor SB203580 reduced ILK-induced and basal cyclin D1 promoter activity to the same extent (25%) at concentrations shown previously to inhibit p38 MAP kinase activity (Fig.  2, A and B). The inhibitor PD98059 (at 10 M), shown previously to selectively inhibit the MAP kinase/ERK kinase-ERK pathway (47), did not reduce ILK-induced cyclin D1 promoter activity, but rather increased it by 75%. Since previous studies demonstrated that pp60 v-src induction of cyclin D1 in MCF7 cells is inhibited by PD98059 (8), our results imply that ILK and pp60 v-src activate distinct signaling pathways.
GSK-3␤ Inhibits ILK-induced Activation of Cyclin D1-Since activation of the Wnt (55) and the PI 3-kinase pathways (56), or ILK overexpression, all lead to phosphorylation and inactivation of GSK-3␤ (31), we examined whether GSK-3␤ inhibited ILK-induced cyclin D1 transcription. When GSK-3␤ was cotransfected with ILK, the ILK-induced cyclin D1 promoter activity was inhibited by GSK-3␤ (Fig. 3A), without inhibiting basal cyclin D1 promoter activity (data not shown), suggesting that GSK-3␤ is involved in the ILK-mediated activation of the cyclin D1 promoter. The expression of GSK-3␤ did not affect the expression of ILK from its expression vector (32). To investigate the role of the Jun kinase pathway in ILK induction of the cyclin D1 promoter activity, we employed an expression vector encoding JIP-1. JIP-1 functions as a powerful dominant negative inhibitor of JNK-dependent activation through cytoplasmic retention of JNK of both c-Jun and ATF-2 transcriptional activities (57). Transfection of the Jun kinase-interactive protein-1 (JIP-1) inhibited ILK-induced cyclin D1 promoter activity by 45% (Fig. 3B). However, JIP-1 overexpression reduced basal cyclin D1 promoter activity by 51%. The magnitude of reduction in ILK-induced promoter activity was therefore similar to that observed on basal promoter activity, suggesting that ILK-induced activation of cyclin D1 is not JNK-dependent (Fig. 3B). These studies provide support for a role of JNK in regulating basal cyclin D1 promoter activity.
ILK Induction of Cyclin D1 Requires CREB in MCF7 Cells-Since the ILK-induced activation of the cyclin D1 promoter required the cyclin D1 CRE site (Fig. 1A), and CREB and ATF-2 proteins bind to the cyclin D1 CRE in MCF7 cells (8), we next examined the role of CREB and ATF-2 in the ILK-induced cyclin D1 promoter activation. Dominant negative mutants of CREB or ATF-2 were co-expressed with the cyclin D1 promoter either in the presence or absence of ILK (Fig. 4). The CREB and ATF-2 dominant negative mutants reduced ILK-induced cyclin D1 promoter activity by 40% and 60%, respectively (Fig. 4A). In contrast, neither mutant inhibited basal activity of the cyclin D1 promoter (Fig. 4B).
Because ILK activation of cyclin D1 required CREB/ATF-2 function and the CRE binding site, we hypothesized that ILK may either regulate CREB/ATF-2 binding to the cyclin D1 CRE site or enhance transactivation by CREB or ATF-2. We first examined the possibility that ILK may directly induce the activity of CREB or ATF-2. The transactivation domains of CREB and ATF-2 linked to the GAL4 DNA binding domain were introduced into cells together with the GAL4 DNA binding sequence, (UAS) 5 E1BTATALUC (Fig. 5A). The CREB and ATF-2 transactivation domains conveyed a basal enhancer activity in MCF7 cells as described previously (8). In agreement with studies using F9 cells (42), overexpression of the catalytic subunit of PKA (PKAc) in MCF7 cells enhanced CREB activity by 6-fold and the point mutation in Ser 133 of CREB abolished induction by PKAc (Fig. 5A). We found that CREB activity was induced by PKB/AKT overexpression in MCF7 cells (by 2-fold; p Ͻ 0.05), and ILK overexpression also induced CREB activity 2-fold (p Ͻ 0.05). In contrast, the activity of ATF-2 was not induced by either AKT or ILK (Fig. 5B). These findings are consistent with recent observations that CREB activity is induced by PKB (PKB/AKT) in 293T cells and that this induction is also dependent on Ser 133 of CREB (58).
EMSA were performed to determine whether ILK and GSK-3␤ regulated binding to the cyclin D1 CRE. Comparison was made using equal amounts of nuclear extracts derived from cells transfected with expression vectors encoding either ILK alone or with the addition of expression plasmids for GSK-3␤ or kinase-dead AKT (Fig. 6A). The intensity of the previously described band, consisting of CREB proteins bound to the cyclin D1 CRE (Fig. 6A, lane 1) (8,37) was increased in cells transfected with ILK (Fig. 6A, lane 2). In addition a nonspecific band (NS) was noted in 293 cells. Co-expression of GSK-3␤ or kinase-dead AKT (AKTmt) reduced the amount of the cyclin D1 CRE binding (Fig. 6A, lanes 3 and 4). Together, these studies suggest ILK overexpression can enhance both the transactivation and binding of proteins at the cyclin D1 CRE.
To confirm that the transcription factor complex binding to the cyclin D1 CRE in MCF7 cells included CREB, EMSA were performed using supershifting antibodies. MCF7 cells were plated either on plastic plates (Fig. 6B, lanes 1-3) or collagen 1-coated plates (Fig. 6B, lanes 4 -9) to enhance basal ILK activity (32,33). A complex bound to the cyclin D1 CRE that was supershifted by the addition of the CREB antibody, but not by equal amounts of IgG control (Fig. 6B, lanes 1-3) or antibodies to c-Fos or c-Jun (data not shown). The abundance of the supershifted complex was increased in the cells plated on collagen 1 (Fig. 6B, lane 1 versus lane 4). Cells treated with the ILK inhibitor KP-SD1 (59) (100 M) showed reduced binding of this complex suggesting ILK activity may contribute to CREB binding to the cyclin D1 CRE site.
Induction of Wnt Signaling Induces ILK Activity and Cyclin

D1 Abundance in Mammary Epithelial Cell Tumors-Recent
studies with cultured cells demonstrated that ILK overexpression activates downstream components of the Wnt signaling pathway (27). Since overexpression of Wnt in mammary gland epithelial cells causes tumors (21), we examined whether activation of the Wnt pathway induces ILK activity and elevates cyclin D1 protein levels in vivo in mammary tissue of MMTV-Wnt-1 tumors. Equal amounts of mammary tissue derived from non-tumorous tissue of strain-matched female mice (NBT) and of MMTV-Wnt-1 tumors were probed for cyclin D1 protein (Fig.  7A, upper panel). The membrane was reprobed with the GDI antibody to normalize for transfer and loading. ILK activity was determined using equal amounts of tissue from the mammary tumors and from normal mammary glands from strainmatched control mice. ILK assays were performed using myelin basic protein as substrate and an ILK-specific antibody for immunoprecipitation (31,32). The results in Fig. 7 indicate that the abundance of cyclin D1 protein varied between tumors up to 5-fold. For each tumor the ILK activity and the cyclin D1 protein levels were plotted (Fig. 7B). In general, the tumors with increased ILK activity displayed increased cyclin D1 protein levels. These results suggest that activation of the Wnt pathway in the mammary epithelium in vivo is associated with both an induction of ILK activity and elevation of cyclin D1 protein levels.
To examine the effect of Wnt-1 overexpression on cyclin D1 abundance, C57MG cells stably overexpressing Wnt-1 and a cell line in which the empty vector was stably integrated were employed. mRNA derived from these cell lines was subjected to PCR-based subtractive hybridization and one of the 24 distinct PCR products identified (data not shown) was identical to the cyclin D1 cDNA. Northern blot analysis (Fig. 8A) showed that cyclin D1 mRNA was increased 3-fold in the Wnt-1 overexpressing C57MG cells when normalized for glyceraldehyde-3phosphate dehydrogenase abundance.
The cell lines were next treated with the ILK inhibitor KP-SD1 (59) using increasing doses from 10 to 100 M, and Western blot analysis was performed for cyclin D1 protein levels. Cyclin D1 protein levels were increased in the C57MG-Wnt-1 cells, and the addition of the ILK inhibitor reduced cyclin D1 protein levels 60 -70% (Fig. 8B). These studies suggest Wnt-1 overexpression is associated with the induction of cyclin D1 mRNA and protein abundance and that the increased abundance of cyclin D1 is at least in part dependent upon ILK activity. DISCUSSION Although various integrin signaling pathways collaborate in the induction of cellular proliferation and anchorage-independent growth, the specific cell-cycle targets affected by integrin signaling are only partially defined. In this study, we have used MCF7 cells in which the abundance of cyclin D1 is rate-limiting for S-phase entry (2,3) and showed that the serine-threonine kinase ILK, which activates integrin signaling upon binding the ␤ 1 and ␤ 3 integrin cytoplasmic domain, induced the cyclin D1 gene through a CREB/ATF-2 binding site. Induction of the cyclin D1 promoter by ILK in MCF7 cells involved the PI 3-kinase, but not the p38 or ERK signaling, pathways. The cyclin D1 gene is thus a direct transcriptional target of ILK through CREB, linking integrin signaling to nuclear cell cycle events.
The induction of cyclin D1 expression by ILK was reduced by inhibitors of PI 3-kinase and rapamycin, but not by inhibitors of the p38 or ERK pathways, consistent with previous studies in which integrin-dependent signaling pathways involved PI 3-kinase (60,61). Integrins signal through several distinct pathways including the focal adhesion kinases, Src family kinases, ERK, Abl, and ILK (62). ERK2/MAP kinase activity is induced by integrin-mediated cell adhesion through c-Src and focal adhesion kinases (63). In contrast ILK activates protein kinase B (PKB) and inhibits GSK-3␤ activity in a PI 3-kinasedependent manner (31,32). In the current studies, LY203580 (50 M) reduced ILK-induced activity 30% and at 100 M reduced activity by 80%. PI 3-kinase inhibitors were previously shown to reduce ILK activity (31,32), and ILK contains a putative phosphoinositide binding domain between residues 180 and 212. Our studies demonstrate that ILK activates the cyclin D1 promoter via a PI 3-kinase pathway. These studies are in agreement with findings that the PI 3-kinase pathway regulates cyclin D1 abundance through both transcriptional (64) and post-transcriptional processes (65). Rapamycin (100 pM), which inhibited ILK-induced cyclin D1 promoter activity by 45%, operates by affecting FRAP/mTOR and is induced by Akt/PKB (66). This is consistent with a model in which an ILK/Akt/FRAP pathway is involved in inducing the cyclin D1 promoter. Since rapamycin was reported to inhibit endothelial cell growth (67), our results imply that cyclin D1 may be a target of the FRAP/mTOR pathway involved in regulating cellular growth. We found that the ERK inhibitor PD98059 enhanced ILK-induced cyclin D1 promoter activity. ERK has previously been shown to inhibit JNK (68) and ␤-catenin/TCF signaling in Caenorhabiditis elegans (69). The current studies suggest that ERK can also antagonize ILK activity in the context of the cyclin D1 promoter.
Wild type ILK activates protein kinases B (PKB/AKT) and inhibits GSK-3␤ activity in a PI 3-kinase-dependent manner (31,32). In the current study, ILK induction of the cyclin D1 promoter required the CRE site previously shown to bind CREB proteins (8,37) and was inhibited by a dominant negative CREB mutant consistent with the important role for GSK-3␤ in regulating CREB function. In addition, ILK and AKT induced CREB activity requiring Ser 133 , suggesting that these kinases may function as components of a signal transduction pathway regulating CREB activity (58). The Ser 133 of CREB is required for the induction of CREB by several cellular kinases including PKA, pp90 rsk and AKT (53,58,70). The increased rate of germ cell apoptosis in CREB knockout mice suggests CREB conveys an important function in cell survival (71). This is supported by studies showing that T cell apoptosis and a G 1 phase cell-cycle proliferative defect occur in transgenic mice harboring a dominant negative CREB (72). The target genes involved in cell survival that are activated by CREB remain to be determined, however. It is of interest that mice in which both alleles of the cyclin D1 gene were deleted display increased retinal apoptosis (4, 5) and mouse embryo FIG. 7. Cyclin D1 and ILK expression is induced in the mammary gland by activation of the Wnt-1 pathway. A, MMTV-Wnt-1 tumors and normal control mammary glands were analyzed for cyclin D1 protein, GDI, and ILK protein levels. B, the cyclin D1 protein levels are plotted versus ILK activity from the mammary tumors. NBT, normal breast tissue. fibroblasts derived from such mice have increased basal and UV-induced apoptosis (73). It will be of interest to determine the relationship between CREB and cyclin D1 in AKT-mediated cell survival.
In the current studies, GSK-3␤ inhibited ILK-induced cyclin D1 promoter activity and reduced CREB binding to the cyclin D1 CRE. Binding of nuclear proteins to the cyclin D1 CRE, predominantly immunoreactive with the CREB supershifting antibody, was increased with the plating of cells on collagen 1 and was decreased by the addition of the ILK inhibitor KP-SD-1. The cyclin D1 CRE is a complex site involved in regulation by diverse oncogenic and mitogenic signals, binding distinct complexes in different cell types. This cyclin D1 gene CRE is involved in activation by serum, pp60 v-src , Rac1, and SV40 small t antigen (6,8,13,37). Although the cyclin D1 CRE binds CREB and ATF-2 proteins in MCF7 cells (8), in human trophoblastic JEG-3 cells the complex consists only of CREB (37) and in mouse embryo fibroblasts the complex is composed predominantly of c-Fos/FosB (6). The site is also capable of binding in vitro synthesized c-Jun (6). GSK-3␤ can regulate the binding and activity of each of these transcription factors and may thereby regulate diverse signaling pathways through this site. CREB phosphorylation by GSK-3␤ inhibits CREB binding to the somatostatin CRE (23), reduces c-Jun binding to DNA through phosphorylating the carboxyl terminus of c-Jun (22), and inhibits ␤-catenin signaling by enhancing its degradation (24,74,75). GSK-3␤ can reduce cyclin D1 protein abundance through phosphorylation-dependent degradation (76) and can also inhibit cyclin D1 transcription, GSK-3␤ apparently functions at multiple levels to regulate cyclin D1 expression. As GSK-3␤ is involved in controlling diverse signaling pathways related to growth control, survival, integrin signal transduction, and cellular metabolism (77), it is likely that transcriptional repression of cyclin D1 by GSK-3␤ may play a key role in coordinating many of these functions.