The tumor suppressor PTEN negatively controls the phosphoinositide 3-kinase pathway for cell survival by dephosphorylating the phospholipid substrates phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate. PTEN has been proposed to dephosphorylate focal adhesion kinase and is implicated in the regulation of cell spreading and motility. We analyzed the role of PTEN in invasion using the two highly infiltrative glioma cell lines U87MG (which lacks functional PTEN) and LN229 (wild-type PTEN). After constitutive overexpression of wild-type and phosphatase-deficient (C124S) PTEN, we found significant inhibition of invasion (50–70%) independent of the PTEN status of the cell and of the catalytic core domain of PTEN. Although wild-type but not mutant (C124S) PTEN decreased PKB/Akt phosphorylation and induced a stellate morphology in U87MG cells, an accompanying reduction of focal adhesion kinase phosphorylation was not seen. We conclude that phosphatase-independent domains of PTEN markedly reduced the invasive potential of glioma cells, defining a structural role for PTEN that regulates cell motility distinct of the PKB/Akt pathway.
Two diseases (Cowden and Lhermitte-Duclos) and the Bannayan-Zonana syndrome are rare inherited, phenotypically overlapping tumor syndromes presenting with hamartomas of the skin, breast, thyroid, and intestine and occasionally presenting with macrocephaly. Affected individuals are prone to develop thyroid and breast carcinoma, malignant gliomas, and meningiomas. The variable phenotypes are caused by germ-line mutations in the tumor suppressor gene PTEN/MMAC1/TEP-1 (hereafter PTEN), which is a dual specificity phosphatase located on human chromosome 10q23.3 (1, 2, 3, 4, 5). Furthermore, somatic PTEN mutations occur during progression of sporadic glial, prostate, endometrial, and kidney tumors, of small cell lung carcinoma, melanoma, and breast cancer (3, 4). The tumor suppressor role of PTEN has been substantiated in knock-out mice (6).
It is now accepted that PTEN is a lipid phosphatase dephosphorylating the 3-position of both PtdIns-3,4,5-P33 and PtdIns-3,4-P2, which are both involved in the PI-3K signaling pathway (7, 8, 9). One important downstream target of PI-3K is PKB/Akt, which protects cells from apoptosis (10). In addition, PTEN is thought to have protein phosphatase activity, and a potential cellular target for this activity is FAK (11). Given that FAK mediates communication between the extracellular matrix and cellular signaling pathways, this would implicate PTEN in processes such as cell spreading and motility (11). Consistent with this, in the PTEN-deficient glioma cell line U87MG, exogenous PTEN reduces invasion and is correlated with reduced phosphorylation of FAK (12). Single point mutations within the phosphatase domain consensus motif are sufficient to abrogate both the protein phosphatase and lipid phosphatase activity of PTEN (13). Although a high frequency of mutations identified within the PTEN gene are clustered around the phosphatase domain (8), numerous frameshift and nonsense mutations are also located at the COOH terminus (14, 15).
Glioblastomas are highly invasive tumors diffusely infiltrating the surrounding normal brain parenchyma (16). Mutations of PTEN are associated with the progression of glioblastomas (17, 18). We analyzed the role of PTEN in invasion using the two highly infiltrative glioma cell lines U87MG and LN229 that both form tumors in nude mice (19). Constitutive expression of wild-type PTEN, but not of a catalytically inactive mutant (C124S), reduced phosphorylation on Ser-473 of PKB/Akt in U87MG cells. Although wild-type PTEN induced a stellate morphology in U87MG cells, this was not associated with a reduction of FAK phosphorylation. Unexpectedly, both wild-type and mutant PTEN reduced cell invasion by 50–70% in PTEN-deficient U87MG cells but also in LN229 cells that express endogenous PTEN. We conclude that overexpression of PTEN is sufficient to reduce the invasive potential of glioma cells independent of the phosphatase activity and implies that other domains distinct from the catalytic motif have a structural role in regulating cell motility.
Materials and Methods
The human brain tumor cell lines LN229 and U87MG derived from adult patients with de novo glioblastoma were selected for this study. These cell lines display different genetic backgrounds with respect to the tumor suppressor genes p53, p16, p14ARF, and PTEN (19). LN229 cells are mutant for p53, deleted for p16 and p14ARF, and wild-type for PTEN, whereas U87 MG cells are wild-type for p53, deleted for p16 and p14ARF, and mutant for PTEN. The glioma cell line LN229 was established in Lausanne, Switzerland, and the cell line U87MG was obtained from the American Type Culture Collection (Rockville, MD). Glioma cell lines were cultured using DMEM supplemented with 10% FCS and standard antibiotics.
Expression Plasmids and Transfections.
cDNA encoding PTEN was synthesized from total RNA derived from glioma cell lines LN229, which expresses endogenous wild-type PTEN (20) and U343MG cells harboring the PTEN point mutant C124S (19). Total RNA was extracted using Trizol (Life Technologies, Inc., Basel, Switzerland), and first-strand cDNA synthesis was performed using oligo(dT) and reverse-transcribed with Superscript RNaseH− RT, according to the manufacturer’s protocol (Life Technologies, Basel, Switzerland). PTEN cDNA was amplified by PCR using Pfu DNA polymerase (Stratagene, La Jolla, CA) and cloned into the pEGFP-C1 vector from Clontech Laboratories (Palo Alto, CA) to generate pEGFP-PTEN-WT and pEGFP-PTEN-C124S. All constructs were sequenced entirely to exclude PCR-generated artifacts. For transfection experiments, glioma cell lines LN229 and U87 MG have been transiently transfected with the pEGFP-PTEN fusion constructs using the GenePORTER transfection system (Gene Therapy Systems, San Diego, CA). Transfection efficiency was >40%, as determined by fluorescence microscopy counting the number of transfected EGPF-positive cells that were costained with the nuclear Hoechst dye 33342 (Molecular Probes, Eugene, OR).
Immunoblotting, Immunoprecipitations, and Cell Adhesion Assays.
Cellular lysates or FAK immunoprecipitates from glioma cell lines LN229 and U87MG were immunoblotted with antibodies directed against GFP (Clontech Laboratories), PKB/Akt (UBI, Lake Placid, NY), phospho-PKB/Akt (Ser-473; UBI), phospho-tyrosine (4G10; UBI), FAK (UBI), and PTEN. Monoclonal antibodies against PTEN were generated by immunizing BALB/c mice with a peptide representing amino acids 388–400 of the PTEN protein (4) and the sequence of the peptide was ENEPFDEDQHTQIC (the cysteine at the COOH terminus was added for coupling of the hapten to the carrier molecule limpet hemocyanin). Immunodetection and immunoprecipitations was performed as described previously (21). Total lysates or immunoprecipitates of FAK from LN229 and U87MG cells transfected either with pEGFP, pEGFP-PTEN-WT, or pEGFP-PTEN-C124S were prepared 48 h after transfection. For cell adhesion assays, tissue culture plates were initially coated with 10 μg/ml fibronectin overnight at 4°C and then blocked with 10 mg/ml BSA in PBS for an additional 1 h at 37°C. Forty-eight h after transfection, LN229 and U87MG cells were harvested and washed once with serum-free DMEM, and 1 × 105 cells were resuspended in DMEM (1% BSA) and plated on 35-mm dishes coated with fibronectin and allowed to spread for 1 h.
Cell Invasion Assay.
Cell invasion assays were carried out using Transwell membrane filter inserts with 8-μm pore size (Corning Costar Corp., Cambridge, MA). The upper surface of the Transwell membrane was coated with 50 μg/ml Matrigel matrix (Collaborative Biomedical Products, Bedford, MA; Becton Dickinson Labware, Franklin Lakes, NY) overnight at 4°C, rehydrated once with 1% BSA in DMEM for 1 h at room temperature, and then placed into 24-well tissue culture plates containing 400 μl of DMEM supplemented with 10% FCS. Sixteen h after transfection, LN229 and U87MG cells were trypsinized and washed once in PBS, and 1 × 104 cells in serum-free medium (1% BSA) were added to each Transwell chamber and allowed to invade toward the underside of the membrane for 48 h. Cells on the upper surface of the filter were then removed by wiping with a cotton swab, and the relative number of invasion was determined by counting the number of invaded EGFP-positive cells. The number of invaded cells that carried the empty vector was set to 1 in each case. Ten fields/membrane were counted for each assay. Each determination represents the average of five individual experiments, and error represents SE. Ps were calculated by repeated measures of ANOVA (Graphpad Software, San Diego, CA), followed by a Student-Newman-Keuls test.
For immunofluorescence microscopy, LN229 and U87MG cells were harvested 48 h after transfection, trypsinized, and washed once with serum-free DMEM medium. Cells (1 × 105) were resuspended in DMEM (1% BSA), plated on fibronectin-coated glass coverslips (Labtec; Life Technologies. Inc.), and allowed to adhere for 1 h. Cells were then fixed with 4% paraformaldehyde for 5 min, washed twice with PBS, and subsequently blocked with 5% goat-serum containing 0.1% Triton X-100 in PBS for 1 h. After washing twice with PBS, coverslips were incubated with antibodies against phosphotyrosine (4G10; 1 μg/ml) and GFP (0.5 μg/ml; Clontech Laboratories). Cells were then washed three times with PBS for 20 min and incubated with antimouse and antirabbit IgGs conjugated to Cy3 and to FITC (Molecular Probes, Eugene, OR), respectively. Coverslips were washed with PBS and mounted with FluorSave reagent (Calbiochem, San Diego, CA) and analyzed under a confocal microscope (Leica TCS-SP; Z-steps, 0.8–1.2 μm), and pictures were further processed using IMARIS software (Bitplane, Zürich, Switzerland).
Restoration of PTEN Activity in U87MG Cells Induces a Stellate Phenotype That Is Not Associated with a Reduction of FAK Phosphorylation.
To examine PTEN function in glioma cells, the glioblastoma cell lines U87MG (PTEN mutant) and LN229 (PTEN wild-type) were transiently transfected with pEGFP-PTEN-(WT), pEGFP-PTEN-(C124S), or pEGFP. The catalytically inactive mutant C124S of PTEN has recently been detected in the glioma cell line U343MG (19) and is known to impair lipid phosphatase activity (7). To analyze the levels of exogenous PTEN expression, the success of transfection was documented by expression of EGFP in 40% of the cells and examined by Western blot analysis. As shown in Fig. 1, the monoclonal anti-PTEN antibody recognized endogenous PTEN protein in LN229 cells with the expected molecular mass of ∼56 kDa and the EGFP-PTEN fusion proteins of 82 kDa, which are expressed in both cell lines. The exogenous EGFP-PTEN fusion proteins were highly expressed in comparison with endogenous PTEN protein levels, whereas no endogenous PTEN protein was seen in U87MG cells. We first determined the involvement of PTEN with regard to its proposed function to negatively regulate PKB/Akt activity in glioma cell lines with known PTEN mutational status (19). Only cellular lysates derived from glioma cells mutant for PTEN displayed high basal activity of PKB/Akt, as assessed with the antibody that specifically recognizes phosphorylated PKB/Akt on residue Ser-473 (data not shown). Transient overexpression of wild-type PTEN substantially reduced the phosphorylation of PKB/Akt in U87MG cells, which was not observed in cells expressing mutant PTEN, which lacks phosphatase activity [Fig. 2, pPKB/Akt (Ser-473)]. The reduction in PKB/Akt phosphorylation signal seen in U87MG cells expressing wild-type PTEN was not attributable to a decrease in PKB/Akt protein levels, as revealed by reprobing the immunoblot with antibody recognizing unphosphorylated forms of PKB/Akt (Fig. 2, PKB/Akt). In addition to its function to regulate PKB/Akt, a marked reduction in the number of positive transfected cells was observed in U87MG cells expressing wild-type PTEN after 5 days in culture, which further implies that the catalytic activity is required to suppress growth in these cells because this numeric reduction of positively staining cells was not observed after expression of mutant (C124S) PTEN (data not shown).
PTEN has also been implicated to regulate spreading and motility by dephosphorylating FAK, which has been proposed to be a substrate of PTEN (11). Therefore, we used standard cell adhesion assays to further examine the role of PTEN in this PTEN-deficient cell line. FAK immunoprecipitates from each transfectant, which were allowed to spread on a fibronectin matrix for 1 h, were tested with antiphosphotyrosine antibody. No change in the amount of FAK phosphorylation could be observed in U87MG cells expressing either wild-type or mutant PTEN when compared with the vector control (Fig. 2, YP-FAK and FAK). We could not detect coimmunoprecipitation of EGFP-tagged PTEN in FAK precipitates using an anti-GFP antibody, suggesting that FAK is not a substrate of PTEN in this experimental system (data not shown). When these cells were analyzed by immunohistochemistry, wild-type PTEN was found to induce an altered cell morphology characterized by a stellate phenotype, whereas cells transfected with the catalytically inactive mutant (C124S) of PTEN displayed a round cell shape similar to untransfected cells (Fig. 3, A and C). In addition, tyrosine phosphorylation levels at focal adhesions were not altered in either cells expressing wild-type or mutant PTEN (Fig. 3, B and D). Thus, transfection of U87MG cells with pPTEN-WT, but not pPTEN-C124S, resulted in decreased phosphorylation of PKB/Akt and changes in cellular morphology, although these changes were not accompanied by a reduction in the phosphorylation of FAK.
Wild-Type and Catalytically Inactive Mutant (C124S) PTEN Inhibit Cell Invasion in U87MG and LN229 Cells.
To explore whether PTEN inhibits cell invasion, U87MG cells, which are highly infiltrative (data not shown), were transfected with pPTEN-WT or pPTEN-C124S and examined in invasion assays as described in “Materials and Methods.” As shown in Fig. 4,A, cell invasion was markedly reduced by 74% (0.26 ± 0.095) in U87MG cells expressing wild-type PTEN when compared with cells carrying the empty pEGFP vector. Surprisingly, inhibition of cell invasion was also observed in cells expressing the catalytically inactive mutant (C124S) of PTEN by 65% (0.35 ± 0.16; Fig. 4,A), suggesting that this inhibition of invasion was independent of the phosphatase activity of PTEN. To further test this hypothesis, LN229 cells, which express endogenous PTEN and are also highly infiltrative, were also assessed using the same constructs. We found that wild-type PTEN inhibits cell invasion by 69% (0.31 ± 0.097), and mutant (C124S) PTEN inhibits cell invasion by 57.5% (0.43 ± 0.16; Fig. 4 B). These data indicate that wild-type and mutant PTEN inhibited cell invasion, even in cells expressing endogenous PTEN, and suggest that this inhibition was mediated in a phosphatase-independent manner.
We show that wild-type and the catalytically inactive mutant (C124S) of PTEN both markedly inhibited cell invasion by up to 70% in the highly infiltrative glioma cell line U87MG. The inhibition of cell invasion mediated by the catalytically inactive mutant (C124S) of PTEN suggests that the phosphatase domain was not required for this activity. Accordingly, constitutive overexpression of wild-type and mutant (C124S) PTEN also reduces invasion to a similar extent in LN229 cells, which express endogenous PTEN. In contrast, the induction of a stellate morphology was clearly dependent on the phosphatase activity of PTEN, which was not associated with a reduction of FAK phosphorylation in U87MG cells. Furthermore, FAK did not coimmunoprecipitate PTEN, which excludes an interaction with PTEN in these cells. Likewise, a decrease of tyrosine phosphorylation levels at focal adhesions was not observed. Tamura et al. (12) showed that overexpression of PTEN in the PTEN-deficient glioma cell line U87MG reduces invasion in a phosphatase-dependent manner and correlates with reduced phosphorylation of FAK. In the light of other results, it was surprising that we found no decrease in FAK phosphorylation (12). We cannot exclude that this merely reflects the transient transfection of PTEN in U87MG cells. However, two observations do not support this interpretation: (a) PTEN clearly reduces phosphorylation of PKB at Ser-473; and (b) we found a significant decrease in FAK tyrosine phosphorylation after transfection of U87MG with FAK-related non-tyrosine kinase.4 The PI-3K signaling pathway has also been implicated in cell migration. Therefore, the effects of PTEN on migration, spreading, and focal adhesion assembly may not be attributable to tyrosine dephosphorylation but instead linked to the lipid phosphatase activity of PTEN (22).
Some of the PTEN-mediated effects could be attributable to the stable binding to its target substrates PtdIns-3,4,5-P3 and PtdIns-3,4-P2 by the mutant catalytic domain (C124S), because (C124S) PTEN mutant represents a potential “substrate-trapping” mutant (23). However, by overexpressing this specific mutant in U87MG cells, we could not detect any changes with respect to PKB/Akt and FAK phosphorylation levels and in addition, morphology alterations were not observed either. Inhibition of cell invasion was the only operational effect we noticed when this mutant was overexpressed in U87MG and LN229 cells. This suggests that the C124S PTEN mutant might not represent a trapping mutant because glioma cells harboring this “trapping” mutant (U343MG) displayed elevated PKB/Akt levels phosphorylated on Ser-473, as did PTEN-null cell lines (Fig. 2). These specific phosphorylation levels could be reduced after expression of wild-type but not mutant (C124S) PTEN (Fig. 2), a finding in agreement with previous studies (9, 24, 25). PTEN modulates PKB/Akt phosphorylation levels in a phosphatase-dependent manner, which correlated with the observed growth suppression in U87MG cells (24, 25).
PTEN displays domains other than the catalytic core motif, such as the NH2-terminal region, which shares high homology to tensin (3, 4), and the COOH-terminal part encoding a PDZ binding motif, which does not appear to be essential for suppression of anchorage-dependent growth (26). Interestingly, COOH-terminal deletion mutants of PTEN were demonstrated to be defective in catalyzing the release of phosphate from either the protein substrate poly-Glu-phospho-Tyr or 3-phosphoinositides, suggesting that the COOH terminus is important for the function of PTEN (24). Indeed, deletion mutants of PTEN were only poorly expressed, suggesting COOH-terminal sequences of PTEN might contribute to its stability.4
We conclude that a phosphatase-independent domain of PTEN substantially impairs the invasive potential of glioma cells independent of FAK tyrosine phosphorylation, defining a structural role for PTEN in regulating cell motility distinct of the PKB/Akt pathway.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grants 31-053746.98 (to A. M.), 4037-055167/1 (to A. M. and G. J.), and 31-49194.96 (to E. G. V. M.) from the Swiss National Science Foundation and grants from the Théodore Ott-Fonds (to A. M.) and the Di Bernardi and Novartis Foundations and the Swiss Cancer League UFS 614-2-1998 (to A. M.).
The abbreviations used are: PtdIns-3,4-P2, phosphatidylinositol 3,4-bisphosphate; PtdIns-3,4,5-P3, phosphatidylinositol 3,4,5-trisphosphate; PI-3K, phosphoinositide 3-kinase; FAK, focal adhesion kinase; EGFP, enhanced green fluorescent protein.
G. Jones, unpublished observation.
We would like to thank Annie-Claire Diserens, Marie-France Hamou, and Frank Hirth for technical assistance.