Overexpression of the Integrin-linked Kinase Promotes Anchorage-independent Cell Cycle Progression*

(Received for publication, November 25, 1996, and in revised form, February 28, 1997)

Galina Radeva , Teresa Petrocelli , Elke Behrend , Chungyee Leung-Hagesteijn , Jorge Filmus Dagger , Joyce Slingerland § and Shoukat Dedhar

From the Department of Medical Biophysics, University of Toronto and Cancer Biology Research, Sunnybrook Health Science Centre, Toronto, Ontario M4N 3M5, Canada

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENT
REFERENCES


ABSTRACT

Cell adhesion to substratum has been shown to regulate cyclin A expression as well as cyclin D- and E-dependent kinases, the latter via the up-regulation of cyclin D1 and the down-regulation of cyclin-Cdk inhibitors p21 and p27, respectively. This adhesion-dependent regulation of cell cycle is thought to be mediated by integrins. Here we demonstrate that stable transfection and overexpression of the integrin-linked kinase (ILK), which interacts with the beta 1 and beta 3 integrin cytoplasmic domains, induces anchorage-independent cell cycle progression but not serum-independent growth of rat intestinal epithelial cells (IEC18). ILK overexpression results in increased expression of cyclin D1, activation of Cdk4 and cyclin E-associated kinases, and hyperphosphorylation of the retinoblastoma protein. In addition, ILK overexpression results in the expression of p21 and p27 Cdk inhibitors with altered electrophoretic mobilities, with the p27 from ILK-overexpressing cells having reduced inhibitory activity. The transfer of serum-exposed IEC18 cells from adherent cultures to suspension cultures results in a rapid down-regulation of expression of cyclin D1 and cyclin A proteins as well as in retinoblastoma protein dephosphorylation. In marked contrast, transfer of ILK-overexpressing cells from adherent to suspension cultures results in continued high levels of expression of cyclin D1 and cyclin A proteins, and a substantial proportion of the retinoblastoma protein remains in a hyperphosphorylated state. These results indicate that, when overexpressed, ILK induces signaling pathways resulting in the stimulation of G1/S cyclin-Cdk activities, which are normally regulated by cell adhesion and integrin engagement.


INTRODUCTION

Normal, untransformed epithelial cells require anchorage to a substratum for cell growth and survival. Adhesion to the extracellular matrix (ECM)11through the G1 and into the S phase of the cell cycle. When forced to remain in suspension, such cells arrest in the G1 phase of the cell cycle and undergo apoptosis (1-3). Oncogenic transformation frequently induces anchorage-independent growth, in vitro, and is a specific correlate of tumor growth in vivo (4, 5).

In fibroblasts, cell adhesion has recently been demonstrated to regulate cell cycle progression by inducing the expression of cyclin D1 (6), the activation of cyclin E-Cdk2 (6, 7), and phosphorylation of the retinoblastoma protein (Rb) (6). Fibroblast adhesion also results in the down-regulation of expression of the Cdk inhibitor proteins, p21 and p27 (6, 7). The combined, adhesion-dependent elevation in cyclin D1 and the decrease in the expression of p21 and p27 result in the stimulation of cyclin D-Cdk4 and cyclin E-Cdk2 activities, both of which can phosphorylate Rb. This latter event relieves the restriction of the entry of cells into S phase, presumably by the release of the transcription factor E2F from phosphorylated Rb (8, 9). In some cell types the expression of cyclin A is also regulated in an anchorage-dependent manner (3, 10, 11), and anchorage-independent growth induced by activated Ras has been shown to depend on cyclin A expression (11). However, in these latter experiments cyclin D1 expression (12) and cyclin E-dependent kinase activity (11) were also dependent on Ras activation. Although mitogens can also activate cyclin D- and cyclin E-dependent kinases, cell adhesion per se can regulate these activities. The regulation of G1 Cdks, therefore, requires the convergence of signals from both growth factors as well as from the ECM.

Anchorage of cells to the ECM is mediated to a large extent by integrins, a large family of heterodimeric cell surface receptors (13, 14). The interaction of integrins with ECM ligands results in the transduction of intracellular signals leading to stimulation of tyrosine phosphorylation (15, 16), turnover of phosphoinositides (17), and activation of the Ras-mitogen-activated protein kinase (MAPK) pathways (18-21). The activation of MAPK by cell adhesion is dependent on the presence of an intact actin cytoskeleton (22), as well as activated p21rho (23). Presumably the adhesion-dependent stimulation of cyclin A expression and of cyclin D1- and cyclin E-associated Cdk activities is also mediated via integrins, although it is not clear as yet whether this requires the activation of MAPK. The cytoplasmic domain of the integrin beta 1 subunit is required for many of the integrin mediated signaling events (13, 24, 25).

Integrin-proximal events involved in the initiation of integrin-mediated signal transduction are still poorly understood. However, a novel ankyrin-repeat containing serine-threonine protein kinase (ILK) has recently been demonstrated to associate with the integrin beta 1 and beta 3 subunit cytoplasmic domains (26) and may be involved in regulating integrin-mediated signaling. Overexpression of ILK in intestinal epithelial cells results in an altered cellular morphology, reduction in cell adhesion to ECM, and also the stimulation of anchorage-independent growth in soft agar (26). Such constitutively ILK-overexpressing cells are also tumorigenic in nude mice.

We now report that overexpression of ILK in rat intestinal epithelial cells (IEC18) increases the expression of cyclin A, cyclin D1, and Cdk4 proteins. The activities of both cyclin D1-Cdk4 and cyclin E-Cdk2 kinases are also elevated, resulting in hyperphosphorylation of the Rb protein. In addition both p21 and p27, inhibitors of cyclin-Cdks, have altered electrophoretic mobilities and p27 from ILK-overexpressing cells has reduced inhibitory activity as compared with the p27 from the parental IEC18 cells. Furthermore, whereas cyclin A and cyclin D1 protein expression, and Rb phosphorylation, are down-regulated upon transfer of IEC18 cells to suspension culture, they are constitutively up-regulated in ILK-overexpressing cells kept in suspension. ILK overexpression in these epithelial cells thus overrides the adhesion-dependent regulation of cell cycle progression through G1 and into S phase, indicating that ILK maybe a key regulator of integrin-mediated cell cycle progression.


MATERIALS AND METHODS

Cell Culture

Three cell lines were used throughout this study: IEC18, ILK13, and ILK14. IEC18 is an immortalized non-tumorigenic rat intestinal epithelial cell line (27), cultured in alpha -minimal essential medium supplemented with 2 mM L-glutamine (Life Technologies, Inc.), 3.6 mg/ml glucose (Sigma), 10 µg/ml insulin (Sigma), and 5% fetal calf serum (Life Technologies, Inc.). ILK13 cells were engineered to overexpress ILK by stable transfection into the parental IEC18 as described previously (26). ILK14 cells are the control transfectants (26). Both ILK13 and ILK14 cell lines were grown under the same conditions as the parental IEC18, with addition of 200 µg/ml G418 (Geneticin, Life Technologies, Inc.) to maintain a selection pressure for ILK or control vector, respectively. Two independently derived clones of each ILK13 (A1a3 and A4a) and ILK14 (A2c3 and A2c6) were used.

Growth Curves

IEC18, ILK13 (ILK-overexpressing cells) and ILK14 (control transfectants) cells were harvested from tissue culture, counted and 104 cells from each cell line were plated in 35-mm tissue culture plates (Nunc). Cells were grown in alpha -minimal essential medium as described above under different serum concentrations (fetal calf serum, Life Technologies, Inc.) for various number of days. At each time point, adherent cells were harvested with 5 mM EDTA/PBS (phosphate-buffered saline, pH 7.6) and viable cells were then quantitated by trypan blue exclusion.

Suspension-maintained Cells

Asynchronously growing cells were harvested from monolayer culture using 5 mM EDTA/PBS and washed two times in PBS. Cells were then resuspended in alpha -minimal essential medium containing 5% fetal calf serum and transferred to 50-ml tubes. A short burst of CO2 was given to the cells before tubes were capped. Suspension cells were incubated for 12 h, rotating on a nutator at 37 °C in 5% CO2. Thereafter, cells were either fixed for fluorescence-activated cell sorting (FACS) analysis or alternatively cell pellets were recovered, washed twice in ice-cold PBS, and then lysed in Nonidet P-40 lysis buffer.

Cell Cycle Analysis

Cells were collected, washed in ice-cold PBS (pH 7.6), fixed in 70% ethanol for 1 h on ice, rinsed with PBS, and DNA stained with 50 µg/ml propidium iodide in PBS containing 10 µg/ml RNase for 30 min at room temperature. Cell cycle profiles were analyzed by FACS using a Becton Dickinson FACScan analyzer, and the percentage of cells in the various phases of cell cycle was calculated using CellFit software.

Immunoblotting

Cells grown in monolayer or in suspension were lysed in ice-cold Nonidet P-40 lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris, pH 7.6, 1 mM EDTA) plus inhibitors (1.0 mM PMSF, 20 µg/ml aprotonin, 20 µg/ml leupeptin) or in ice-cold Tween 20 lysis buffer (0.1% Tween 20, 50 mM Hepes, pH 7.5, 150 mM NaCl, 2.5 mM EGTA, 1 mM EDTA) plus inhibitors (1 mM dithiothreitol, 1.0 mM PMSF, 20 µg/ml aprotonin, 10 mM beta -glycerophosphate, 0.1 mM sodium vanadate, 1 mM sodium fluoride).

Total protein extracts or immune complexes were resolved on SDS-PAGE and then transferred to Immobilon-P (Millipore). The membrane was first blocked in 5% milk in TBST (0.05% Tween 20 (Sigma) in Tris-buffered saline, pH 7.4) and then incubated with primary antibody. The following antibodies were used: anti-cyclin D1 (DCS-6, mouse monoclonal, from Dr. J. Bartek, Danish Cancer Society, Copenhagen, Denmark), anti-cyclin E (rabbit polyclonal, Santa Cruz), anti-cyclin A (rabbit polyclonal, Santa Cruz), anti-Cdk4 (rabbit polyclonal, Santa Cruz), anti-Cdk2 (rabbit polyclonal, Santa Cruz), anti-PSTAIRE (mouse monoclonal, a gift from Dr. S. Reed, The Scripps Research Institute, La Jolla, CA), anti-p27 (mouse monoclonal, Transduction Laboratories), anti-p21 (rabbit polyclonal, Santa Cruz), anti-retinoblastoma (mouse monoclonal, Pharmingen), and anti-ILK (affinity-purified rabbit polyclonal). Protein detection was carried out using secondary antibody (either anti-mouse-HRP (Jackson Laboratories or Pharmingen), anti-rabbit-HRP (Jackson Laboratories), or protein A-HRP (Amersham Life Science)) and the enhanced chemiluminescence (ECL) detection system (Amersham Life Science).

Kinase Assays

For Cdk4-associated kinase activity, asynchronous cells growing in monolayer culture were lysed in ice-cold Tween 20 lysis buffer (0.1% Tween 20, 50 mM Hepes, pH 7.5, 150 mM NaCl, 2.5 mM EGTA, 1 mM EDTA), containing the following inhibitors (1 mM dithiothreitol, 0.1 mM PMSF, 20 µg/ml aprotonin, 10 mM beta -glycerophosphate, 0.1 mM sodium vanadate, 1 mM sodium fluoride). Cell lysates were then sonicated. Protein A-Sepharose beads (Sigma Immunochemicals Co.) precoated with Cdk4 antibody (rabbit polyclonal, Santa Cruz) were used to immunoprecipitate Cdk4. Cdk4-associated kinase activity was assayed using the protocol of Matsushime et al. (28).

For cyclin E kinase assays, cells from asynchronous monolayer culture were lysed in 1% Nonidet P-40 lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris, pH 7.6) plus protease inhibitors (0.1 mM PMSF, 20 µg/ml aprotonin, 20 µg/ml leupeptin). Cyclin E was immunoprecipitated with polyclonal anti-cyclin E serum (gift from Dr. Steve Reed, The Scripps Research Institute, La Jolla, CA and also from Dr. D. Agrawal, M. Lee Moffit Cancer Center, Tampa, FL) and complexes collected on protein A-Sepharose beads (Sigma Immunochemicals Co.). Cyclin E-associated kinase reactions were carried out as described previously (29).

For both cyclin D1-Cdk4 and cyclin E-Cdk2 assays, kinase reaction products were resolved by SDS-PAGE and the incorporation of radioactivity in substrate was visualized by autoradiography (X-Omat AR (Eastman Kodak Co.) or REFLECTIONTM (DuPont)) and quantitated by PhosphorImager (Molecular Dynamics).

p27 Inhibitory Assay

Cell lysates (100 µg of protein) were recovered from asynchronously growing IEC18 or ILK-overexpressing cells using lysis buffer as for cyclin E-Cdk2 kinase assays with Nonidet P-40 at 0.1% concentration. Lysates were boiled for 5 min and clarified by centrifugation. p27 was immunoprecipitated (rabbit polyclonal anti-p27 serum provided by Dr. T. Hunter, Salk Institute, La Jolla, CA) from boiled lysates. Immune complexes were collected on protein A-Sepharose beads and then washed five times in 0.1% Nonidet P-40 lysis buffer. To release bound p27, the beads were resuspended in 200 µl of 0.1% Nonidet P-40 lysis buffer, containing protease inhibitors (1 mM PMSF and 20 µg/ml each aprotonin, leupeptin, and pepstatin), boiled for 5 min and supernatants recovered. Cyclin A-Cdk2 complexes immunoprecipitated from asynchronous ILK14 cells (control-transfected cells) were used as test substrate for inhibition by p27. Heat-stable p27 released from immune complexes was incubated at 30 °C for 30 min together with cyclin A-Cdk2. Cyclin A-Cdk2 kinase activity was assayed using histone H1 (Boehringer Mannheim) as a substrate and compared with the activity of cyclin A-Cdk2 complexes without added immunoprecipitated p27. As a negative control, non-immune serum immunoprecipitates were collected, boiled, and supernatant added to active cyclin A-Cdk2 test complexes. The p27 antiserum used in these assays does not cross-react with p21. Detection of radioactivity in kinase substrate was carried out as described for kinase assays.


RESULTS

ILK Overexpression Induces Adhesion-independent Cell Growth and Survival but Not Serum-independent Growth

We have shown previously that overexpression of the ILK in normal rat intestinal epithelial cells (IEC18) results in a less adherent phenotype and in anchorage-independent growth in soft agar (26). When maintained in suspension, IEC18 cells have been demonstrated to undergo programmed cell death (30), which is suppressed by mutant c-Ha-ras oncogene expression (30). Since we have found that ILK overexpression in IEC18 cells induces anchorage-independent growth as well as tumorigenicity in nude mice,2 we wanted to determine whether ILK overexpression also suppresses suspension-induced cell cycle arrest and cell death. ILK-overexpressing cell clones (ILK13) are capable of anchorage-independent cell growth in soft agar (26). This increased cell survival is reflected in the greater proportion of ILK13 cells that are present in S phase after 12 h in suspension, as compared with the control ILK14 cells, in which the percentage of cells in S phase falls to 5% (Fig. 1A). Furthermore, a sub-G1 (<2 N) population is present in the control ILK14 cells after 12 h in suspension, consistent with the presence of apoptotic cells. This population of cells is completely absent in the ILK-overexpressing ILK13 clones (Fig. 1A). We next wished to address whether ILK overexpression also induces serum-independent growth in monolayer-adherent cultures. As shown in Fig. 1B, the growth rate of ILK13 cells is not elevated when compared with the IEC18 or the control ILK14 cells. In fact, the ILK-overexpressing clones grow slightly more slowly than the parental IEC18 and the ILK14 control-transfected cells (Fig. 1B). In addition, ILK13 cells fail to survive in serum-free conditions similar to the IEC18 and control ILK14 cells. These data demonstrate that ILK overexpression selectively induces anchorage-independent growth but not serum (mitogen)-independent growth.


Fig. 1. A, cell-cycle profiles of ILK13 and ILK14 cells maintained in suspension or monolayer culture. Asynchronously growing ILK-overexpressing (ILK13) and control-transfected (ILK14) cells (26) were transferred from monolayer to suspension culture for 12 h as described under "Materials and Methods." After that the cell cycle profiles of the cells in suspension (S) or in monolayer (A) were analyzed by FACScan (see "Materials and Methods") and compared. The numbers on the left represent the percentage of cells in each phase of the cell cycle. B, growth rates of IEC18, ILK13 and ILK14 cells at various serum concentrations. 104 cells from each cell line were plated on 35-mm tissue culture plates under various serum concentrations. At different time points, adherent cells were harvested and number of viable cells was determined by trypan blue exclusion. Cell lines correspond as follows: black-square, IEC18; bullet , ILK14 (A2c3); triangle , ILK14 (A2c6); black-down-triangle , ILK13 (A1a3); diamond , ILK13 (A4a).
[View Larger Version of this Image (18K GIF file)]

ILK Overexpression Alters the Expression of Cell Cycle Regulators

Adhesion of fibroblasts to ECM has been shown to induce the expression of cyclin D1 (6). Since overexpression of ILK in epithelial IEC18 cells induces cell survival and cell cycle progression in the absence of adhesion, we wanted to determine whether ILK overexpression altered the expression and/or activity of cell cycle regulators. The expression of various cell cycle regulators was examined in IEC18, ILK13, and ILK14 cells growing under standard tissue culture conditions. As shown in Fig. 2A, ILK-overexpressing cell clones (ILK13) (26) express substantially higher levels of cyclin D1 protein than the parental IEC18, or control-transfected ILK14 cells. In contrast, the level of expression of cyclin E is not altered in ILK13 cells. The expression of cyclin A was examined as well and was found to be elevated in ILK13 cells (data not shown in Fig. 2; see Fig. 4). Since the cyclins function in complex with the cyclin-dependent kinases, Cdks, we also determined the expression of Cdk4 and Cdk2 kinases that complex with cyclin D1 and cyclin E, respectively. Surprisingly the level of Cdk4 protein is also elevated in the ILK13 cells, whereas Cdk2 is not altered (Fig. 2A). The kinase activities of Cdk4 and Cdk2 are also regulated by inhibitor proteins, p21 and p27, and the expression of these inhibitors is known to be enhanced in non-adherent (suspension) cells and decreases upon cell adhesion (6, 7). In ILK-overexpressing cells, both p21 and p27 are increased (Fig. 2A), and their electrophoretic mobilities are clearly altered. The faster migrating forms of p21 and p27 in ILK-overexpressing cells may reflect covalent modification, or in the case of p27, the product of partial proteolytic cleavage (31). For p27, at least, this alteration correlates with a decreased inhibitory potential (see Fig. 3b). Such faster migrating forms have also been observed after exposure of fibroblasts to UV irradiation (32).


Fig. 2. A, alteration in the expression levels of the constituents of the G1/S cyclin-Cdk complexes. Immunoblot analysis of the various cell cycle regulators was carried out as described under "Materials and Methods." Two independently derived ILK-overexpressing clones (ILK13: A1a3 and A4a) (26) and the control transfectants (ILK14: A2c3 and A2c6) (26) were tested and compared with the parental IEC18 (rat intestinal epithelial cell line). The levels of cyclin D1 and Cdk4 proteins were increased in the ILK-overexpressing cells, while no difference in the amount of cyclin E and Cdk2 proteins was observed. Cyclin-Cdk inhibitory proteins, p21 and p27, were found to have an altered mobility in the ILK-overexpressing cells. B, cyclin D1 overexpression is specifically triggered by ILK overexpression. IEC18 cells were transfected with a metallothionine-inducible construct (33), containing the complete ILK gene inserted in-frame (MT-ILK-1). Transfection was performed using Lipofectin, as per manufacturer's instructions (Life Technologies, Inc.). Transfected cells were cloned by limiting dilution in 96-well tissue culture plates (Nunc). Metallothionine inductions were performed in the presence of serum supplemented with 100 µM ZnSO4 and 2 µM CdCl2 for 18-24 h. After induction, cells were lysed in Nonidet P-40 lysis buffer and ILK and cyclin D1 levels screened by immunoblot analysis. ILK expression was induced in the MTILK1 clone (containing plasmid vector with cDNA encoding for ILK) following treatment of the cells with Zn2+/Cd2+. Concomitantly, the expression of cyclin D1 protein was also induced. ILK and cyclin D1 protein levels were quantified by densitometric analysis using an LKB laser densitometer (model 2222-020) using Gelscan XL Software (Pharmacia Biotech Inc.). C, immunoblot analysis of cyclin D1-Cdk4 complex. Cdk4 was immunoprecipitated from each of the cell lines described, and associated cyclin D1, Cdk4, and p27 were detected by immunoblotting. ILK13 cells show a higher content of cyclin D1, Cdk4, and p27 in the immunoprecipitated cyclin D1-Cdk4 complex. D, immunoblot analysis of cyclin E-Cdk2 complex. Cyclin E was immunoprecipitated from each of the cell lines described, and then associated Cdk2 and p27 were detected by immunoblotting.
[View Larger Version of this Image (33K GIF file)]


Fig. 4. Adhesion-independent overexpression of cyclin D1 and cyclin A, and hyperphosphorylation of Rb in ILK-overexpressing cells. Expression of cyclin D1 and cyclin A and phosphorylation of Rb were analyzed in response to cell attachment. Adherent ILK13 and ILK14 cells were harvested, transferred into 50-ml tubes, and maintained in suspension (S) for 12 h. Cell lysates were then recovered from cells in suspension (S) and cells growing in monolayer culture (A). Cyclin D1, cyclin A, and Rb proteins were analyzed by immunoblotting. A, each cell line was found to have elevated cyclin D1 protein upon adhesion to substratum in comparison to cells kept in suspension. However, the level of cyclin D1 is constitutively higher in ILK13 cells kept in suspension. B, cyclin A protein was higher in ILK13 adherent cells than in ILK14 adherent. After transferring cells in suspension, ILK13 cells continue to maintain high cyclin A, while in ILK14 cells cyclin A expression falls dramatically. C, the retinoblastoma protein is hyperphosphorylated in suspension ILK13 cells, but not in control suspension ILK14 cells.
[View Larger Version of this Image (25K GIF file)]


Fig. 3. Panel a, effect of ILK overexpression on the kinase activities of the G1/S cyclin-Cdk complexes. A, cyclin D1-Cdk4 kinase assay. Following immunoprecipitation of Cdk4 from IEC18, ILK13, and ILK14 cells, an in vitro kinase assay was performed using Rb (QED Bioscience Inc.) as substrate. The incorporation of radioactivity in Rb substrate is severalfold higher in ILK13 clones, indicating higher kinase activity of cyclin D1-Cdk4 in the ILK-overexpressing cells. B, cyclin E-Cdk2 kinase assay. Cyclin E was immunoprecipitated from IEC18, ILK13, and ILK14 cells, and the associated Cdk2 histone H1 kinase activity was assayed in vitro. Cyclin E-Cdk2 from the ILK13 cells showed higher kinase activity then that of cyclin E-Cdk2 from IEC18 or ILK14 cells. The lower panels in A and B represent IgG from Coomassie Blue-stained gels to confirm equal loading. C, immunoblot of retinoblastoma protein immunoprecipitated from the three different cell lines: IEC18, ILK13, and ILK14. ILK13 cells (ILK overexpressors) show an increase in the hyperphosphorylated form of Rb, as compared with IEC18 and ILK14, in which no difference between the two forms of Rb is detected. Panel b, p27 inhibitory activity and immunoprecipitation. A, inhibitor activity: increasing amounts of p27 were immunoprecipitated from asynchronous IEC18 cells (1x, 100 µg of lysate; 2x, 200 µg of lysate; 3x, 300 µg of lysate), recovered on protein A-Sepharose beads, and released by boiling. Heat-stable p27 was mixed with cyclin A-Cdk2 immunoprecipitated from 100 µg of lysates from asynchronously growing cells, and the ability of p27 to inhibit the test cyclin A-Cdk2 kinase activity was assayed. (Cyclin A-Cdk2 kinase activity, without any added p27, is shown in first lane.) No Cdk2 inhibitory activity was recovered from boiled preimmune (PI) serum immunoprecipitates. B, p27 inhibitory activity in IEC18 and ILK13 cells. Equal quantities of p27 (1x) were immunoprecipitated from IEC18 and ILK13 cells. p27 was released from protein A-Sepharose beads by boiling, added to test cyclin A-Cdk2 from 50 µg of lysate, and kinase activity was assayed on H1 as a substrate. H1 kinase reactions were resolved by SDS-PAGE, gels dried, and radioactivity in histone H1 bands was quantitated by PhosphorImager. The activity of test cyclin A-Cdk2 without inhibitor was compared with that with added p27, and results are presented as percent maximum kinase activity in uninhibited cyclin A-Cdk2. Comparison of p27 inhibitory activity from equal amounts of p27 from IEC18 and ILK13 (1x and 2x) shows greater inhibition by p27 from IEC18 cells. C, levels of p27 used in the inhibitor assays (B). To show that the amounts of p27 added to test cyclin A-Cdk2 in the inhibitor assays shown in B were equivalent, p27 was immunoprecipitated from IEC18 (1x, 50 µg of lysate) and ILK13 cells (1x, 15 µg; 2x, 30 µg). The quantity of p27 used is shown by resolving complexes by SDS-PAGE and immunoblotting with p27 antibody.
[View Larger Version of this Image (37K GIF file)]

To demonstrate that the observed changes are mediated by ILK, we transfected IEC18 cells with an ILK expression vector under the control of a metal inducible promoter (33). As shown in Fig. 2B, induction of ILK expression with Zn2+/Cd2+ results in the stimulation of expression of ILK. Concomitantly, the expression of cyclin D1 is also induced in these cells (Fig. 2B). The data shown were confirmed in two independent clones, and the treatment of the parental IEC18 cells with Zn2+/Cd2+ had no effect on ILK or cyclin D1 expression (data not shown). These data demonstrate that increased ILK expression can induce the expression of cyclin D1 protein.

We next determined whether the complex formation between the cyclins, Cdks, and the p21/p27 inhibitors was also altered upon ILK overexpression. As shown in Fig. 2C, both cyclin D1 and Cdk4 are elevated in Cdk4 immunoprecipitates from ILK13 cells as compared with the parental IEC18 and control ILK14 cells. Although the amount of p27 is also higher in Cdk4 immunoprecipitates from ILK13 cells, quantification clearly demonstrates that the ratio of p27 to cyclin D1-Cdk4 is much higher in IEC18 and ILK14 cells than it is in the ILK13 cells (Table I). Furthermore, the p27 in Cdk4 immunoprecipitates from ILK13 clones has the faster electrophoretic mobility (Fig. 2C). The amount of cyclin E-associated Cdk2 did not differ between the parental IEC18 and ILK-overexpressing (ILK13) cells. However, in the ILK13 cells, although cyclin E-associated p27 was increased, p27 manifested the altered mobility seen in the Cdk4 complexes (Fig. 2D).

Table I. Ratios of p27/cyclin D1 in Cdk4 immunoprecipitates

The amounts of cyclin D1 and p27 proteins in the cyclin D1-cdk4 complexes were quantitated by densitometry using a LKB Laser Densitometer (model 2222-020) and Gelscan XL Software (Pharmacia). The densitometric values for a given protein were obtained after subtracting the value present in the negative control (antibody alone lane). The ratios of intensities of p27/cyclin D1 were calculated for each cell line. The exposure of the film on which the scanning was done was in the linear range of ECL. The amounts of cyclin D1 and p27 proteins in the cyclin D1-cdk4 complexes were quantitated by densitometry using a LKB Laser Densitometer (model 2222-020) and Gelscan XL Software (Pharmacia). The densitometric values for a given protein were obtained after subtracting the value present in the negative control (antibody alone lane). The ratios of intensities of p27/cyclin D1 were calculated for each cell line. The exposure of the film on which the scanning was done was in the linear range of ECL.

IEC18 ILK13 (A1a3) ILK13 (A4a) ILK14 (A2C3)

p27/cyclin D1 Ratio 2.44:1 0.73:1 0.94:1 4.3:1

ILK Overexpression Leads to the Stimulation of Cyclin D1-Cdk4 and Cyclin E-Cdk2 Kinase Activities

Cyclin D1 and Cdk4 proteins are increased upon ILK overexpression, but those of cyclin E and Cdk2 are not (Fig. 2). To determine whether this translates into increased kinase activities, we carried out immune complex in vitro kinase assays for both cyclin D1-Cdk4 and also cyclin E-Cdk2 using recombinant Rb and histone H1 as substrates, respectively. As shown in Fig. 3a, the kinase activity of Cdk4 is dramatically increased in the ILK-overexpressing clones (ILK13). Although protein levels of cyclin E and Cdk2 are not elevated (Fig. 2), cyclin E-Cdk2 kinase activity is also increased in these cells (Fig. 3a). Cell adhesion in fibroblasts has been shown to stimulate Cdk2 activity (6, 7), without elevations in cyclin E or Cdk2 levels (6). This is thought to be brought about by the down-regulation of the Cdk inhibitors p21 and p27. In the ILK-overexpressing cells, the increased Cdk2 activity could result, at least in part, from the decreased inhibitory activity of p27 (see Fig. 3b). The net effect of the increased activities of cyclin D1-Cdk4 and cyclin E-Cdk2 in ILK13 cells is an increase in the retinoblastoma protein phosphorylation (Fig. 3a).

p27 from ILK-overexpressing Cells Is Altered and Has a Lower Cdk Inhibitory Potential

It has been demonstrated previously that nonadherent fibroblasts express high levels of the Cdk inhibitor p27 and that down-regulation of p27 upon cell substratum adhesion increases cyclin E-Cdk2 activity (6, 7). Although ILK overexpression elevates cyclin E-Cdk2 activity, the levels of Cdk inhibitors, p21 and p27, are not decreased. In fact, they appear to be elevated as compared with the IEC18 and ILK14 control cells (Fig. 2). However, both p21 and p27 from ILK13 cells have an altered electrophoretic mobility (Fig. 2). We therefore determined whether the altered electrophoretic mobility of p27 correlated with an altered inhibitory potential of this protein and hence might contribute to the increased cyclin E-Cdk2 kinase activity. To analyze p27 activity, we immunoprecipitated p27 from IEC18 and ILK13 cells and assayed its ability to inhibit test cyclin A-Cdk2 kinase complexes. As shown in Fig. 3b, p27 from IEC18 cells inhibits cyclin A-Cdk2 in a dose-dependent manner. When compared with the activity of p27 from IEC18 cells, equivalent amounts of ILK13-derived p27 (Fig. 3b) showed significantly less inhibitory activity in this type of assay (Fig. 3b). This decreased p27 inhibitory activity could contribute to the higher cyclin E-Cdk2 activity present in the ILK13 cells. Thus although, p27 can complex with cyclin E-Cdk2 (Fig. 2) in the ILK13 cells, its inhibitory potential is reduced resulting in a net higher level of cyclin E-Cdk2 kinase activity.

Adhesion-independent Up-regulation of Cyclin D1 and Cyclin A Expression, and Rb Hyperphosphorylation in ILK-overexpressing Cells

Non-adherent fibroblasts express low levels of cyclin D1 and have low cyclin D1-Cdk4 and cyclin E-Cdk2 activities. Untransformed fibroblasts and epithelial cells are also growth-inhibited in suspension and arrest in the G1 phase (1-3). Since ILK overexpression in IEC18 cells induces cell survival and promotes cell cycle progression in suspension, we wanted to determine whether the increased levels of cyclin D1 and Rb protein hyperphosphorylation were maintained in suspension. Furthermore, since the expression of cyclin A is regulated in an anchorage-dependent manner in some cells, we also examined adhesion-dependent regulation of cyclin A protein expression in IEC18 and ILK-overexpressing (ILK13) cells. Exponentially growing adherent cultures of ILK13 and the control, ILK14 cells were placed in suspension for 12 h. The cells were then lysed, and the expression of cyclin D1 and cyclin A and Rb phosphorylation were determined by immunoblotting. As expected, cyclin D1 and cyclin A proteins fall with increased duration in suspension in control (ILK14) cells (Fig. 4). However, in the ILK13 cells, the elevated cyclin D1 and cyclin A expression is maintained in suspension. Similarly, whereas Rb is rapidly dephosphorylated in control (ILK14) cells in suspension, a substantial proportion of Rb remains hyperphosphorylated in suspension ILK13 cells (Fig. 4). These data indicate that overexpression of ILK overcomes the adhesion-dependent regulation of cyclin D1 and cyclin A protein expression and Rb phosphorylation, suggesting that ILK is in the signaling pathway that mediates integrin-dependent regulation of the cell cycle.


DISCUSSION

Cell adhesion to components of the extracellular matrix is a requirement for cell growth and survival for a wide variety of cell types (1, 2, 4). Inhibition of cell adhesion results in growth arrest, and many epithelial and endothelial cells also undergo apoptosis (1-3). Cell adhesion to the ECM results in the activation of signaling pathways, which maintain cell cycle progression from G1 to S phase. The key components of the cell cycle machinery known to be regulated by cell adhesion to ECM are cyclin D1 and cyclin A expression, activation of cyclin D-Cdk4 and cyclin E-Cdk2 kinases, and retinoblastoma protein phosphorylation (3, 6-8, 10, 11). Determination of the molecular basis of this regulation is clearly important and may be central to our understanding of anchorage-independent cell growth and oncogenic transformation.

It is highly likely that integrins, as receptors for ECM components, initiate signaling events that activate the above mentioned cell cycle regulators. Integrin activation and ligation have been shown to activate MAPK via p21ras-dependent (18, 20, 34) and -independent (21) pathways. Activation of MAPK, in turn, can regulate the transcription (35), and translation of cyclin D1 mRNA, the latter by regulating the activity of PHAS-1 (36). The adhesion-dependent increase in cyclin D1 expression is also regulated, in part, at the level of mRNA translation (6), and therefore activation of MAPK may be crucial in adhesion-dependent cell cycle control. Anchorage-dependent expression of cyclin A has been shown to be regulated at the level of gene transcription (10). The integrin-proximal events responsible for the activation of downstream signaling pathways still need to be fully characterized.

We have recently identified a novel serine/threonine protein kinase (ILK), which can associate directly with the cytoplasmic domain of integrin beta 1 and beta 3 (26). Overexpression of this kinase in epithelial cells induces anchorage-independent growth (26) and oncogenic transformation.2 In this paper, we have demonstrated that, when overexpressed, human ILK induces adhesion-independent cell survival of rat intestinal epithelial cells, increases both cyclin D1 and cyclin A protein levels, and stimulates the activation of cyclin-dependent kinases. Specifically, we have shown that the induction of ILK expression by stable transfection (ILK13), or by inducible transfection, results in the elevation cyclin D1 protein. Furthermore, the activity of Cdk4 is substantially elevated in ILK13 clones when compared with parental IEC18 cells or control ILK14 clones. In contrast, although the expression of cyclin E and Cdk2 are unchanged, cyclin E-Cdk2 activity is increased in the ILK13 clones. The combined activation of Cdk4 and Cdk2 activities results in the hyperphosphorylation of the retinoblastoma protein (Rb), the phosphorylation of which regulates the entry of cells into S phase (8, 9). Surprisingly, ILK overexpression also seems to increase the levels of both p21 and p27 Cdk inhibitors. Since Cdk inhibition by KIP family proteins relies on an increase in the molar ratio of p21 or p27 in the Cdk complex, the ratio of p27:target cyclin-Cdk is important (37). The ratio of p27 to cyclin D1 in complex with Cdk4 is substantially higher in the IEC18 and ILK14 cells than it is in the ILK-overexpressing (ILK13) cells. Thus despite the ILK-mediated increase in KIP proteins, the molar ratio (shown for p27) of KIP:cyclin D-Cdk4 in ILK-13 cells is not increased.

Another interesting consequence of ILK induction is the expression of altered forms of both p21 and p27. These altered forms have faster electrophoretic mobilities as compared with p21 and p27 from the parental IEC18 cells and the control-transfected (ILK14) clones. The nature of this alteration is not clear as yet but could result from altered phosphorylation (38) or proteolytic degradation (31). However, the expression of different isoforms, for example, by alternative splicing cannot be ruled out. A potential functional consequence of this alteration appears to be decreased inhibitory activity, as demonstrated for p27. This decreased inhibitory activity could account for the increased cyclin E-Cdk2 activity observed in ILK-overexpressing cells.

Of significant importance to the oncogenic properties of ILK and its role in integrin-mediated signal transduction is the finding that ILK-overexpressing cells (ILK13) continue to cycle in serum-containing suspension cultures, whereas the control transfectant clones (ILK14) undergo cell cycle arrest and apoptosis, as described previously (30). In IEC18 and control ILK14 cells, inhibition of adhesion to ECM results in a rapid down-regulation of expression of both cyclin D1 and cyclin A proteins, Rb dephosphorylation, and G1 arrest. This is in marked contrast to ILK13 clones, in which cyclin D1 and cyclin A expression as well as Rb phosphorylation are maintained upon transfer to suspension cultures and there is no inhibition of cell cycle progression. ILK, like Ras, stimulates the expression of cyclin A and cyclin D1 resulting in Rb phosphorylation. However, unlike Ras, ILK does not induce serum-independent cell growth, indicating that anchorage-independent cell growth can be stimulated independently of serum-independent cell growth. Preliminary data indicates that overexpression of ILK does not activate Ras, but can activate MAPK,3 thus suggesting that ILK can activate a Ras-independent pathway capable of altering cell cycle control resulting in anchorage-independent cell growth. Ras activates, on the other hand, other cellular functions, which result in both anchorage-independent and serum-independent cell growth. Although one must be cautious in interpreting data from overexpression studies, our results indicate that ILK may have an important role in modulating anchorage-independent cell cycle progression. Whether the kinase activity of ILK is required for this regulation remains to be determined.

Elevated cyclin D1 expression is quite common in certain types of cancers, especially breast and esophageal carcinomas (39-41). Although in some cases the increased cyclin D1 expression is due to gene amplification (41), for the majority of the cases, the molecular basis of this increased expression is unclear (42). Since ras mutations are infrequent in breast carcinomas, it is unlikely that Ras plays an important role in the elevation of cyclin D1. Our results suggest that the altered expression of ILK might be involved in the elevated cyclin D1 expression seen in some cancers, and this will be the subject of future studies. Finally, our results suggest a role for ILK in specifically coupling anchorage-dependent growth and cell cycle regulation. Altered expression, and/or kinase activity, of ILK could have an important role in uncoupling cell cycle regulation by cell adhesion and may play a crucial role in pathogenesis of cancer and cardiovascular diseases.


FOOTNOTES

*   This work was supported in part by grants from the National Cancer Institute of Canada and the Medical Research Council of Canada.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Supported by the Medical Research Council of Canada.
§   Clinician scientist supported by the Ontario Cancer Treatment and Research Foundation.
   Terry Fox Cancer Scientist of the National Cancer Institute of Canada. To whom correspondence should be addressed: Dept. of Medical Biophysics, University of Toronto and Cancer Biology Research, Sunnybrook Health Science Centre, 2075 Bayview Ave., Rm S-218, Toronto, Ontario M4N 3M5, Canada.
1   The abbreviations used are: ECM, extracellular matrix; MAPK, mitogen-activated protein kinase; Rb, retinoblastoma; ILK, integrin-linked kinase; PBS, phosphate-buffered saline; FACS, fluorescence-activated cell sorting; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; HRP, horseradish peroxidase.
2   C. Wu, S. Y. Keightley, C. Leung-Hagesteijn, G. Radeva, J. McDonald, and S. Dedhar, submitted for publication.
3   D. Hackam, E. Behrend, and S. Dedhar, unpublished observations.

ACKNOWLEDGEMENT

We thank Mina Viscardi for secretarial assistance.


REFERENCES

  1. Frisch, S. M., and Francis, H. (1994) J. Cell Biol. 124, 619-626 [Abstract]
  2. Meredith, J., Bazeli, B., and Schwartz, M. (1993) Mol. Biol. Cell 4, 953-961 [Abstract]
  3. Boudreau, N., Sympson, C. J., Werb, Z., and Bissell, M. J. (1995) Science 267, 891-893 [Medline] [Order article via Infotrieve]
  4. Shin, S.-I., Freedman, V. H., Risser, R., and Pollack, R. (1975) Proc. Natl. Acad. Sci. U. S. A. 72, 4435 [Abstract]
  5. Ruoslahti, E., and Reed, J. (1994) Cell 77, 477-478 [Medline] [Order article via Infotrieve]
  6. Zhu, X., Ohtsubo, M., Bohmer, R. M., Roberts, J. M., and Assoian, R. K. (1996) J. Cell Biol. 133, 391-403 [Abstract]
  7. Fang, F., Orend, G., Watanabe, N., Hunter, T., and Ruoslahti, E. (1996) Science 271, 499-502 [Abstract]
  8. Johnson, D. G., Schwarz, J. K., Cress, W. D., and Newins, J. R. (1993) Nature 365, 349-352 [CrossRef][Medline] [Order article via Infotrieve]
  9. Sherr, C. J. (1996) Cell 79, 551-555
  10. Schulze, A., Zerfan-Thome, K., Berges, J., Middendorp, S., Jansen-Dur, P., and Henglein, B. (1996) Mol. Cell. Biol. 16, 4632-4638 [Abstract]
  11. Kang, J., and Kraus, R. S. (1996) Mol. Cell. Biol. 16, 3370-3380 [Abstract]
  12. Filmus, J., Robles, A. I., Shi, W., Wong, M. J., Colombo, L. L., and Conti, C. J. (1994) Oncogene 9, 3627-3633 [Medline] [Order article via Infotrieve]
  13. Schwartz, M. A., Schaller, M. D., and Ginsberg, M. H. (1995) Annu. Rev. Cell Biol. Dev. Biol. 11, 549-600 [CrossRef][Medline] [Order article via Infotrieve]
  14. Hynes, R. O. (1992) Cell 69, 11-25 [Medline] [Order article via Infotrieve]
  15. Clark, E. A., and Brugge, J. S. (1995) Science 268, 233-239 [Medline] [Order article via Infotrieve]
  16. Dedhar, S., and Hannigan, G. E. (1997) Curr. Opin. Cell Biol. 8, 657-669 [CrossRef]
  17. McNamee, H. P., Ingber, D. E., and Schwartz, M. A. (1993) J. Cell Biol. 121, 673-678 [Abstract]
  18. Kapron-Bras, C., Fitz-Gibbon, L., Jeevaratnam, P., Wilkins, J., and Dedhar, S. (1993) J. Biol. Chem. 268, 20701-20704 [Abstract/Free Full Text]
  19. Chen, Q., Kinch, M. S., Lin, T. H., Burridge, K., and Juliano, R. (1994) J. Biol. Chem. 269, 26602-26605 [Abstract/Free Full Text]
  20. Clark, E. A., and Hynes, R. O. (1996) J. Biol. Chem. 271, 14814-14818 [Abstract/Free Full Text]
  21. Chen, Q., Lin, T. H., Der, C. J., and Juliano, R. L. (1996) J. Biol. Chem. 271, 18122-18127 [Abstract/Free Full Text]
  22. Zhu, X., and Assoian, R. K. (1995) Mol. Biol. Cell 6, 273-282 [Abstract]
  23. Hotchin, N. A., and Hall, A. (1995) J. Cell Biol. 131, 1857-1865 [Abstract]
  24. Akiyama, S. K., Yamada, S. S., Yamada, K. M., and LaFlamme, S. E. (1994) J. Biol. Chem. 269, 15961-15964 [Abstract/Free Full Text]
  25. Lukashev, M. E., Sheppard, D., and Pytela, R. (1994) J. Biol. Chem. 269, 18311-18314 [Abstract/Free Full Text]
  26. Hannigan, G. E., Leung-Hagesteijn, C., Fitz-Gibbon, L., Coppolino, M., Radeva, G., Filmus, J., Bell, J., and Dedhar, S. (1996) Nature 379, 91-96 [CrossRef][Medline] [Order article via Infotrieve]
  27. Quaromi, A., and Isselbacher, K. J. (1981) J. Natl. Cancer Inst. 67, 1353-1362 [Medline] [Order article via Infotrieve]
  28. Matsushime, H., Quelle, D. E., Shurtleff, S. A., Shibuya, M., Sherr, C. J., and Kato, J. Y. (1996) Mol. Cell. Biol. 14, 2066-2076 [Abstract]
  29. Dulic, V., Lees, E., and Reed, S. I. (1996) Science 257, 1958-1961
  30. Rak, J., Mitsuhashi, Y., Erdos, V., Huang, S. N., Filmus, J., and Kerbel, R. S. (1995) J. Cell Biol. 131, 1587-1598 [Abstract]
  31. Loda, M., Cukor, B., Tam, S. W., Lavin, P., Fiorentino, M., Draetta, G. F., Jessup, M., and Pagano, M. (1997) Nature Med. 231-234
  32. Poon, R. Y. C., Jiang, W., Toyoshima, H., and Hunter, T. (1996) J. Biol. Chem. 271, 13283-13291 [Abstract/Free Full Text]
  33. Filmus, J., Remani, J., and Klein, M. H. (1992) Nucleic Acids Res. 20, 2755-2760 [Abstract]
  34. Schlaepfer, D. D., Hanks, S. K., Hunter, T., and van der Geer, P. (1994) Nature 372, 786-791 [Medline] [Order article via Infotrieve]
  35. Lavoie, J. N., L'Allemain, G. L., Brunet, A., Müller, R., and Pouyssegur, J. (1996) J. Biol. Chem. 271, 20608-20616 [Abstract/Free Full Text]
  36. Lin, T. A., Kong, X., Haystead, T. A. J., Panse, A., Balsham, G., Sonnenberg, N., and Laurence, J. J. C. (1994) Science 266, 653-656 [Medline] [Order article via Infotrieve]
  37. Sherr, C. J., and Roberts, J. M. (1995) Genes Dev. 9, 1149-1163 [CrossRef][Medline] [Order article via Infotrieve]
  38. Gu, Y., Rosenblatt, J., and Morgan, D. O. (1992) EMBO J. 11, 3995-4005 [Abstract]
  39. Bartkova, J., Lukas, J., Muller, H., Liitzhof, T., Strauss, M., and Bartek, J. (1994) Int. J. Cancer 57, 353-361 [Medline] [Order article via Infotrieve]
  40. Motokuva, T., and Arnold, A. (1993) Biochim. Biophys. Acta 1155, 63-78 [CrossRef][Medline] [Order article via Infotrieve]
  41. Keyomasi, K., and Paredee, A. B. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 1112-1116 [Abstract]
  42. Buckley, M. F., Sweeney, K. J. E., Hamilton, J. A., Sini, R. L., Manning, D. L., Nicholson, R. I., DeFagio, A., Watts, C. K. W., Musgrove, E. A., and Sutherland, R. L. (1993) Oncogene 8, 2127-2133 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.