Loss of PTEN/MMAC1/TEP in EGF Receptor-Expressing Tumor Cells Counteracts the Antitumor Action of EGFR Tyrosine Kinase Inhibitors

Home , PTEN (gene)

Loss of PTEN/MMAC1/TEP in EGF receptor-expressing tumor cells counteracts the antitumor action of EGFR tyrosine kinase inhibitors

Roberto Bianco1, Incheol Shin1, Christoph A Ritter1, F Michael Yakes1, Andrea Basso2, Neal Rosen2, Junji Tsurutani3, Phillip A Dennis3, Gordon B Mills4 and Carlos L Arteaga*,1,5,6

1Department of Medicine, Vanderbilt University School of Medicine, Nashville, TN, USA; 2Departments of Medicine and Cell Biology and Genetics, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 3Center for Cancer Research, National Cancer Institute, Bethesda, MD, USA; 4Department of Molecular Therapeutics, MD Anderson Cancer Center, Houston, TX, USA; 5Department of Cancer Biology, Vanderbilt University School of Medicine, Nashville, TN, USA; 6Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, TN, USA

We have examined the possible mechanisms of resistance ade of the EGFR tyrosine kinase and Akt should be to the epidermal growth factor receptor (EGFR) inhibitors considered as a therapeutic approach. in tumor cells with variable levels of EGFR. ZD1839 Oncogene (2003) 22, 2812–2822. doi:10.1038/sj.onc.1206388 (Iressa) is a small-molecular-weight, ATP-mimetic that specificallyinhibits the EGFR tyrosinekinase. A431 cell Keywords: epidermal growth factor receptor; ZD1839; growth was markedlyinhibited byZD1839 (IC 50p0.1 lm) PTEN; tyrosine kinase inhibitors; Akt whereas the MDA-468 cells were relativelyresistant (IC50>2 lm). Low doses of ZD1839 delayed cell cycle progression and induced apoptosis in A431 cells but not in MDA-468 cells. In both cell lines, 0.1 lm ZD1839 Introduction eliminated EGFR phosphorylation. However, the basal activityof the phosphatidylinositol-3 kinase (PI3 K) target The epidermal growth factor receptor (EGFR, HER1, Akt was eliminated in A431 but not in MDA-468 cells, erbB1) is a 170-kDa protein composed of an extra- implying that their Akt activity is independent of EGFR cellular ligand-binding domain, a short transmembrane signals. A431 cells express PTEN/MMAC1/TEP, a domain, and an intracellular domain with intrinsic phosphatase that can dephosphorylate position D3 of tyrosine kinase activity. Overexpression and/or hyper- phosphatidylinositol-3,4,5 trisphosphate, the site that activity of the EGFR has been shown to play a causal recruits the plecstrin-homologydomain of Akt to the cell role in the progression of several epithelial neoplasms membrane. On the contrary, MDA-468 cells lack the (reviewed in Prenzel et al., 2001; Yarden and Sliwkows- phosphatase and tensin homolog (PTEN), potentially ki, 2001). The EGFR is activated by binding of ligand(s) setting Akt activityat a high threshold that is unrespon- to its extracellular domain which leads to receptor sive to EGFR inhibition alone. Therefore, we reintroduced homodimerization or heterodimerization with any of the (PTEN) byretroviral infection in MDA-468 cells. In other three members of this family of transmembrane MDA-468/PTEN but not in vector controls, treatment tyrosine kinases: HER2 (erbB2), HER3 (erbB3), and with ZD1839 inhibited P-Akt levels, induced relocaliza- HER4 (erbB4). This results in binding of ATP to the tion of the Forkhead factor FKHRL1 to the cell nucleus, receptor’s catalytic site, activation of the receptor’s and increased FKHRL1-dependent transcriptional activ- tyrosine kinase, and autophosphorylation on C-terminal ity. ZD1839 induced a greater degree of apoptosis and cell tyrosine residues, which, in turn, recruit several cyto- cycle delay in PTEN-reconstituted than in control cells. plasmic signal transducers. These effector molecules These data suggest that loss of PTEN, bypermitting a include PLC-g1, Ras-MEK-MAPK, phosphatidylinosi- high level of Akt activityindependent of receptor tyrosine tol-3 kinase (PI3K) and its target Akt, p70S6 kinase, kinase inputs, can temporallydissociate the inhibition of Src, and STATs, among others (Olayioye et al., 2000; the EGFR with that of Akt induced byEGFR inhibitors. Yarden and Sliwkowski, 2001). Thus, in EGFR-expressing tumor cells with concomitant The oncogenic potential of the EGFR and its high amplification(s) of PI3K-Akt signaling, combined block- level of expression in tumor tissues provides a rationale for targeting this oncoprotein with novel molecular therapeutics. Indeed, anti-EGFR molecules inhibit the progression of EGFR-dependent preclinical models and *Correspondence: CL Arteaga, Division of Oncology, Vanderbilt have recently demonstrated clinical efficacy against University School of Medicine, 2220 Pierce Ave., 777 Preston Res. several human carcinomas (reviewed in Arteaga, 2001). Bldg., Nashville, TN 37232-6307, USA; E-mail: [email protected] The antitumor effect of these inhibitors requires the Received 22 August 2002; revised 10 January 2003; accepted 13 January subversion of key postreceptor signaling pathways and 2003 cell cycle/apoptosis regulatory molecules that mediate RH Loss of PTEN/MMAC1/TEP R Bianco et al 2813 the transforming effects of the EGFR network. In- kinase inhibitors alone. We have examined this hypoth- activation of constitutively active MAPK and the serine/ esis in MDA-468 human breast cancer cells which threonine kinase Akt, a target of PI3K, have been overexpress EGFR and carry a deletion and frame-shift reported in EGFR-dependent tumor cells treated with mutation at codon 70 of the PTEN protein (Lu et al., EGFR inhibitors (Fan et al., 1997; Busse et al., 2000; 1999) as well as in PTEN-expressing and PTEN-null Lenferink et al., 2000; Albanell et al., 2001; Nelson and NSCLC cell lines. Fry, 2001). In addition, inhibition of the homologous tyrosine kinase HER2 in tumor cells results in inhibition Results of PI3K and Akt (Lane et al., 2000; Hermanto et al., 2001; Lenferink et al., 2001; Moulder et al., 2001; Yakes EGFR-overexpressing cancer cells with mutant PTEN are et al., 2002). The antitumor effect of HER2 inhibitors less sensitive to ZD1839 has been shown to require the subversion of Akt function (Yakes et al., 2002). We initially examined the effects of ZD1839 (Iressa) on EGFR-mediated activation of Akt requires the A431 and MDA-468 human cancer cells. ZD1839 is an activation of PI3K. This can occur via dimerization of ATP-competitive inhibitor of the EGFR catalytic EGFR with HER3, which is able to couple to PI3K activity; it inhibits the EGFR tyrosine kinase (isolated directly (Fedi et al., 1994), or by interaction of the from A431 cell membranes) in vitro with an IC50 of receptor with the intracellular adaptor Gab1 (Rodrigues 0.033 mm (Wakeling et al., 2002). A431 and MDA-468 et al., 2000). Upon activation, PI3K converts phospha- cells exhibit EGFR gene amplification and secrete TGFa, tidylinositol-4,5 bisphosphate (PI4,5P2) to phosphati- thus expressing autoactivated EGFR in the absence of dylinositol-3,4,5 trisphosphate (PI3,4,5P3); this lipid exogenous ligands (Ennis et al., 1989; Van de Vijver recruits the plecstrin-homology (PH) domain of Akt to et al., 1991). ZD1839 inhibited the growth of (wild-type the plasma membrane where its serine/threonine kinase (WT) PTEN) A431 cells with an approximate IC50 of is activated by phosphoinositide-dependent kinase 1 0.1 mm whereas the PTEN-null MDA-468 cells were (PDK1)- and PDK2-mediated phosphorylation (Chan relatively resistant to concentrations as high as 1 mm et al., 1999). Active Akt phosphorylates a number of (Figure 1a). Similar data were obtained with the substrates involved in apoptosis, cell cycle regulation, humanized anti-EGFR IgG2 C225 (Goldstein et al., protein synthesis, and glycogen metabolism, which 1995) and the EGFR tyrosine kinase inhibitor OSI-774 include the Bcl-2 family member Bad, Forkhead (Moyer et al., 1997) (data not shown). As measured by transcription factors, caspase-9, IkB kinase, MDM2, TUNEL and flow cytometry of labeled nuclei, ZD1839 GSK-3b, p21Waf1, p27Kip1, mTOR, and nitric oxide treatment resulted in apoptosis and cell cycle delay in synthase (Datta et al., 1999; Zhou et al., 2001a, b; Shin A431 cells. These effects were not seen in MDA-468 cells et al., 2002). Via phosphorylation, Akt functionally (Figure 1b, c). To test the integrity of the PI3K pathway, inactivates these proapoptotic and cell-cycle-regulatory we used LY294002, a specific inhibitor of the p110 molecules, thus enhancing tumor cell survival and catalytic subunit of PI3K (Vlahos et al., 1994). In both proliferation. In human tumors, Akt activity has been cell lines, LY294402 induced apoptosis and cell cycle shown to be upregulated by several alterations which arrest; the latter was more marked in the PTEN-null include PI3K isoform gene amplification, activating MDA-468 than in A431 cells (Figure 1b, c). mutations of p85, Akt gene amplification and overexpression, as well as loss of function of the phosphatase ZD1839 inhibits EGFR phosphorylation and coupling to and tensin homolog (PTEN) phosphatase (Vivanco and p85 in A431 and MDA-468 cells Sawyers, 2002). PTEN dephosphorylates the D3 position of membrane PI3,4,5P3 providing negative regula- Consistent with its EGFR specificity, sub-micromolar tion of PI3K and Akt activities (Cantley and Neel, 1999; concentrations of ZD1839 induced a similar degree of Simpson and Parsons, 2001). Mutations and/or dele- inhibition of basal EGFR tyrosine phosphorylation in tions in PTEN occur with variable frequency in both A431 and MDA-468 cells without a change in advanced cancers including glioblastoma multiforme, receptor protein levels (Figure 2a). As a result of the melanoma, endometrial, breast, ovarian, renal cell, published effect of EGFR inhibitors on PI3K/Akt thyroid, and a small subset of small-cell and nonsmall- signaling and the potent ability of HER3 to couple to cell lung cancers (NSCLC) (Ali et al., 1999; Vivanco and PI3K directly (Fedi et al., 1994), we examined the effect Sawyers, 2002). Reconstitution of PTEN expression in of ZD1839 on this signaling pathway and on EGFR/ PTEN-null cells has been shown to repress Akt and HER3 interactions. In A431 cells, treatment with inhibit tumor growth via induction of apoptosis or ZD1839 markedly reduced the constitutive association repression of cell proliferation (Li and Sun, 1998; Lu of EGFR with HER3 and with the regulatory subunit of et al., 1999; Sun et al., 1999). PI3K, p85a. MDA-468 cell lysates exhibited low levels The data summarized above suggest that inhibition of of HER3 by immunoblot (not shown). Compatible with Akt is required for the antitumor effect of EGFR this observation, p85a, but not HER3, was detectable in inhibitors. Hence, we speculate that EGFR-overexpres- EGFR precipitates from these cells. ZD1839 markedly sing tumor cells with coexisting amplification of the reduced the basal coupling of EGFR with p85 in MDA- PI3K-Akt pathway might harbor Akt activity at a high 468 cells (Figure 2a). Interestingly, Akt activation state threshold and, therefore, not respond to EGFR tyrosine as assessed by immunoblot analysis using antibodies

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2814

Figure 1 PTEN-mutant MDA-468 breast cancer cells are relatively insensitive to ZD1839. (a) A431 and MDA-468 cells were seeded on 12-well plates in IMEM/10% FCS at a density of 3 Â 104 cells/well. ZD1839 (0.1–1 mm) was added the next day and after 72 h, adherent cells were trysinized and counted using a Coulter counter. Each bar represents the mean7s.d. of four wells. (b, c) Cells were seeded at the density of 2 Â 105 cells/well in 60-mm dishes; the following day, 1 mm ZD18399 or 40 mm LY294002 were added. After 72 h, both adherent and floating cells were harvested and assayed for evidence of apoptosis by Apo-BrdU analysis. (b) Percentages indicate the proportion of FITC-positive apoptotic cells in the gated area as determined by flow cytometry. Identically treated cells were harvested by trypsinization after 48 h of treatment, fixed in methanol, labeled with 50 mg/ml propidium iodide containing 125 U/ml protease-free RNAse, and analysed by flow cytometry for cell cycle distribution as indicated in Materials and methods (c)

speciﬁc to S473 P-Akt was reduced in A431 but not in region (Figure 3a). Since both PTEN and GFP are MDA-468 cells without a change in total Akt protein expressed from a single viral mRNA, coexpression of levels following treatment with ZD839 (Figure 2b). both gene products is 100% and GFP can be used to LY294002 induced almost complete elimination of P- sort PTEN-positive cells. MDA-468 cells tolerated Akt levels in both cell lines (Figure 2b), indicating that transduction of the PTEN expression vector and P-Akt phosphorylation is dependent on PI3K activity in selection with no gross evidence of cell death or change MDA-468 cells. These results also imply that, in the in morphology (Figure 3b, c). By immunoblot analysis, PTEN-null MDA-468 cells, Akt activity is independent both MDA-468/PTEN and A431 cells exhibited a of EGFR autocrine signaling. robust level of PTEN protein whereas parental MDA- 468 and cells transfected with the pBMN-IRES-GFP empty vector did not (Figure 3d). Reconstitution of PTEN restores ZD1839-mediated Retroviral transduction and selection did not change inhibition of Akt EGFR content in MDA-468 cells. Treatment of MDA- To restore PTEN expression in MDA-468 cells, we 468/vector and MDA-468/PTEN cells with 1 mm introduced full-length PTEN into the pBMN-IRES- ZD1839 reduced EGFR phosphorylation as determined EGFP retroviral expression vector (Grignani et al., by P-Tyr immunoblot of EGFR precipitates (Figure 4a). 1998). In this vector, PTEN was subcloned immediately Reintroduction of PTEN eliminated basal PIP3-forming upstream an internal ribosomal entry sequence (IRES) (PI3K) activity (Figure 4b) and reduced P-Akt levels linked to the green ﬂuorescent protein (GFP) coding (Figure 4c) in MDA-468/PTEN cell lysates. Treatment

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2815

Figure 2 ZD1839 inhibits EGFR phosphorylation and association of EGFR and p85 in A431 and MDA-468 cells. (a) Cells were treated with 0.1–1 mm ZD1839 (in IMEM/1% FCS) and harvested after 24 h. Whole-cell lysates were prepared and 1 mg of total protein precipitated with an EGFR antibody. Immune complexes were next divided into four equal parts, each separated by 8% SDS–PAGE, and ﬁnally subjected to Immunoblot analysis for EGFR, P-Tyr, HER3, and p85. (b) Cells were treated with ZD1839 or LY294002 at the indicated concentrations and harvested after 24 h; 60 mg of total cell lysates were separated by 10% SDS–PAGE followed by Immunoblot analysis for S473 P-Akt and total Akt

Figure 3 Stable reconstitution of PTEN in MDA-468 cells. (a) Plasmid map of pBMN-IRES-GFP retroviral vector used to infect MDA-468 cells. (b) DIC image of MDA-468 cells stably expressing PTEN. (c) Green ﬂuorescence image of MDA-468 cells expressing PTEN and GFP. (d) Protein (60 mg) from total cell lysates from parental, vector control, and PTEN-infected MDA-468 cells were separated by 10% SDS–PAGE and immunoblotted with a PTEN antibody. A431 cells were included as a positive control of the stably transduced cells with ZD1839 or LY294002 cells (Figure 4c), suggesting that reconstitution of PTEN for 24 h reduced basal PI3K activity by approximately function restored at least in part the dependence of Akt 50% (Figure 4b). Consistent with this last result, activation on EGFR signaling in these cells. P-MAPK ZD1839 treatment induced a reduction in levels of P- was equally inhibited in both control and MDA-468/ Akt in PTEN-reconstituted but not MDA-468 control PTEN cells by treatment with ZD1839 (Figure 4c),

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2816

Figure 4 Reconstitution of PTEN inhibits basal PI3K activity and restores ZD1839-mediated inhibition of Akt in MDA-468 cells. Subconﬂuent MDA-468/vector control and MDA-468/PTEN cells were treated with 0.1–1 mm ZD1839 in IMEM/1% FCS and harvested after 24 h. Cell lysates were prepared in lysis buffer as described in Materials and methods. (a) Total cell lysates (500 mg) were precipitated with an EGFR antibody, divided into two equal parts, separated by 8% SDS–PAGE, and subjected to EGFR or P-Tyr Immunoblot analysis. (b) Lipids extracted from 32P-orthophosphate-labeled cells were analysed by TLC as described in Materials and methods. Numbers below indicate the intensity of the 32P-PIP3 band relative to the one from untreated MDA-468/vector cells. (c) Total cell lysates (60 mg) were separated by 10% SDS–PAGE and tested in immunoblot procedures for P-Akt, total Akt, P-MAPK, and total MAPK

further implying that constitutive activation of the Akt, had been restored in the PTEN-transduced cells. PI3K-Akt pathway, but not the Ras-MEK-MAPK Cells were transiently transfected with an FHRE-Luc pathway, is associated with this relative resistance to vector in which a Forkhead response element is linked to the EGFR inhibitor. a luciferase reporter gene. FHRE contains the binding To examine Akt function in response to ZD1839, we site of the FasL promoter. FKHRL1 binds this site and utilized a hemaglutinin (HA)-tagged FKHLR1 expres- enhances transcription of the FasL gene (Brunet et al., sion vector. FKHRL1 is a Forkhead transcription 1999). Treatment with ZD1839 did not induce Forkhead factor that has been shown to induce the transcription transcriptional activity in parental and vector-trans- of cell-death-related genes (Kops and Burgering, 1999). fected MDA-468 cells. However, consistent with its Upon phosphorylation by Akt, FKHRL1 translocates ability to inhibit Akt, treatment with 1 mm ZD1839 for from the nucleus to the cytoplasm where 14-3-3 proteins 24 h increased FHRE-dependent luciferase expression in sequester it and prevent its function (Datta et al., 1999; both A431 and MDA-468/PTEN cells (Figure 5b). With Shin et al., 2001). In A431, MDA-468, MDA-468/ the FKHRL1 localization data (above), these data vector, and MDA-468/PTEN cells transiently trans- further support that the ability of ZD1839 to inhibit fected FKHRL1 localized predominantly in the cyto- Akt function had been restored by reintroduction of plasm (Figure 5a, column 1 of each panel), consistent PTEN. with the constitutive Akt activity in these cells. Treat- ment with ZD1839 resulted in nuclear localization of Proliferation and survival of MDA-468/PTEN cells are FKHRL1 in A431 but not in PTEN-negative parental inhibited by EGFR inhibitors and vector control MDA-468 cells. However, blockade of the EGFR kinase with ZD1839 in MDA-468/PTEN We ﬁnally examined if re-expression of PTEN restored cells resulted in nuclear distribution of FKHRL1 sensitivity of MDA-468 to EGFR kinase inhibitors. (Figure 5a, bottom row of right panel). MDA-468/PTEN cells were markedly more sensitive We next asked whether Forkhead (FHRE)-dependent to ZD1839 compared to MDA-468/vector cells (IC50 transcriptional activity, as a function of inhibition of approximately 0.1 vs >1 mm, respectively) (Figure 6a).

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2817

Figure 5 ZD1839 induces nuclear localization and transcriptional activity of FKHLR1 in MDA-468/PTEN cells. (a) A431, parental MDA-468, MDA-468/vector, and MDA-468/PTEN cells were grown in 60-mm dishes (5 Â 105 cells/dish) and transfected with an expression vector encoding HA-tagged FKHRL1 for 16 h as described in Materials and methods. The cells were next transferred to coverslips on 12-well plates, incubated with 1 mm ZD1839 for 24 h, and subjected to immunostaining with anti-HA ﬂuorescein or Hoechst 33258. Coverslips were examined by laser scanning confocal microscopy. (b) The indicated cell lines were transiently transfected with FHRE/Luc and pCMV/R1 followed by treatment with 0.1% DMSO or 1 mm ZD1839 for 24 h. Fireﬂy and Renilla reniformis luciferase activities were determined using the Dual Luciferase Assay System as described in Materials and methods. Each data point represents the mean7s.d. of four wells

In addition, a 72-h incubation with ZD1839 resulted in cytometry of propidium-iodide-labeled cell nuclei, was marked dose-dependent cell death in MDA-468/PTEN modestly greater in PTEN-reconstituted than in MDA- cells, which was much less obvious in the PTEN- 468/vector cells (Figure 6c). Finally, the EGFR inhibi- negative controls (Figure 6b). Furthermore, the inhibitors C225 and OSI-774 also induced a greater degree of tion of cell cycle progression, as measured by ﬂow apoptosis in MDA-468/PTEN than in control cells

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2818

Figure 6 Reconstitution of PTEN in MDA-468 cells enhances the sensitivity to EGFR inhibitors. (a) MDA-468/vector and MDA- 468/PTEN cells were seeded on 12-well plates in IMEM/10% FCS at a density of 3 Â 104 cells/well. ZD1839 was added the next day. After 72 h, the cells were trypsinized and counted using a Coulter counter. Each bar represents the mean7s.d. of four wells. (b, c) Cells were seeded at the density of 2 Â 105 cells/well in 60-mm dishes; the following day, 0.1–1 mm ZD1839 was added. After 72 h, both adherent and floating cells were harvested and assayed for evidence of apoptosis by Apo-BrdU analysis (b). Percentages indicate the proportion of FITC-positive apoptotic cells in the gated area as determined by flow cytometry. Identically treated cells were harvested by trypsinization after 48 h of treatment, fixed in methanol, labeled with 50 mg/ml propidium iodide containing 125 U/ml protease-free RNAse, and analysed by flow cytometry for cell cycle distribution as indicated in Materials and methods (c). (d) MDA-468/vector and MDA-468/PTEN cells were seeded as in (b) and incubated with the indicated concentrations of C225 or the small molecule OSI-774. After 72 h, adherent and floating cells were assayed for evidence of apoptosis by Apo-BrdU analysis. Each data point represents the mean7s.d. of three 60-mm dishes

(Figure 6d), implying that re-expression of PTEN had (Figure 7a). Treatment with ZD1839 inhibited prolif- mainly restored or enhanced the ability of EGFR eration and delayed cell cycle progression in H1355 inhibitors to induce tumor cell apoptosis. PTEN-positive but not in H157 PTEN-null cells (Figure 7b, c).

PTEN-null NSCLC cells are less sensitive to ZD1839 Discussion To determine if this differential sensitivity to EGFR inhibitors as a function of PTEN status was not limited In this report, we have examined the potential mechan- to A431 and MDA-468 cells, we performed similar isms of resistance to EGFR inhibitors in EGFR- experiments in H1355 and H157 NSCLC lines. H1355 overexpressing human carcinoma cells. Studies in cells have WTPTEN (Kohno et al., 1998). The H157 transgenic mice and in human tumors have established cell line lacks PTEN protein and exhibits a nonsense that epithelial cancers rely on multiple genetic abnorm- mutation of the PTEN gene (Forgacs et al., 1998). Both alities resulting in several aberrant signaling pathways cell lines contain a mutation of K-ras (Brognard et al., that, in turn, mediate cancer maintenance and progres- 2001). They exhibited equal levels of cell surface EGFR sion. Thus, because of the potential for compensation, it by ﬂow cytometry of cells labeled with an EGFR is anticipated that interruption of a single signaling antibody against the receptor’s extracellular domain network or transforming molecule will not be curative in

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2819

Figure 7 PTEN-null NSCLC cells are relatively resistant to ZD1839. H1355, H157, and A431 (positive control) cells were incubated with anti-EGFR 528 mAb or a nonspeciﬁc IgG2a followed by FITC-labeled goat anti-mouse IgG2a. FITC-labeled cells were detected by ﬂow cytometry as indicated in Materials and methods. The dark areas mark the IgG2a (control)-stained cells whereas the clear peaks mark the EGFR-positive, 528 mAb-labeled cells. (b) Cells were seeded in DMEM/10% FCS in 12-well plates at a density of 3 Â 104 cells/well. The following day, they were changed to DMEM/0.1% FCS7ZD1839. Cell number in each well was determined 72 h later. Each data point represents the mean7s.d. of four wells. (c) Cells were plated in triplicate as in (b) and then changed to DMEM/0.1% FCS7 the indicated concentrations of ZD1839. After 24 h, cell cycle distribution was assessed as indicated in Materials and methods. Data are represented as the mean G1/S ratio7s.d. from three wells

late invasive cancers. For example, in a recent report, suggest that tumor cells may be endowed with multiple regulated overexpression of c-Myc in the mouse signaling molecules that could potentially counteract the mammary gland resulted in invasive mammary tumors blockade of HER2 or EGFR function at the transmem- that regressed upon repression of c-Myc. Eventually, brane receptor level. however, a large proportion of tumors did not regress Inhibition of EGFR and HER2 has been shown to when c-Myc was downregulated and the majority of inhibit PI3K and Akt kinase and their downstream these tumors were found to harbor activating mutations targets in several tumor cell lines (Busse et al., 2000; of K-ras and N-ras (D’Cruz et al., 2001). The seminal Lane et al., 2000; Lenferink et al., 2001; Moasser et al., studies with the HER2-blocking antibody Herceptin in 2001; Moulder et al., 2001). Forced expression of active patients with HER2-ampliﬁed metastatic breast cancer Akt prevents the antitumor effect of Herceptin against provide another example. Although some tumors HER2-overexpressing SKBR-3 and BT-474 human regress with antibody therapy, the majority does not breast cancer cells (Yakes et al., 2002), suggesting that respond and/or eventually escape Herceptin action (1) inhibition of PI3K and Akt might be an obligate step (Vogel et al., 2002), supporting de novo as well as for the cell cycle arrest and/or apoptosis mediated by acquired mechanisms of therapeutic resistance. Similar HER2 inhibitors, and (2) abnormal hyperactivation of to EGFR inhibitors, the antitumor effect of HER2 Akt can counteract the cellular effects of HER2 inhibitors would require the subversion of key post- inhibitors. We speculate that the same should apply to receptor signaling pathways and cell cycle/apoptosis EGFR inhibitors as well. One mechanism for aberrant regulatory molecules that mediate the transforming activation of Akt is loss of PTEN, a phosphatase that effects of HER2. These postreceptor pathways are negatively regulates PI3K and Akt by dephosphorylat- shared with and receive simultaneous inputs from ing PI3, 4, 5P3 (Cantley and Neel, 1999; Simpson and heterologous receptor networks (Carpenter, 1999; Parsons, 2001). Surveys of several tumors for PTEN Gschwind et al., 2001). Indeed, overexpression of the gene deletion or mutations have shown PTEN loss in a IGF-I receptor has been reported to abrogate the wide spectrum of human cancers (Ali et al., 1999; antitumor effect of Herceptin as well as EGFR tyrosine Vivanco and Sawyers, 2002). Although mutations of kinase inhibitors against human tumor cells (Lu et al., PTEN occur in o5% of breast cancers, some reports 2001; Chakravarti et al., 2002). In these studies, suggest severe loss of PTEN protein in breast tumors simultaneous blockade of either the IGF-I receptor or and frequent hemizygous deletions of the PTEN gene PI3K restored sensitivity to the HER2 and EGFR (Teng et al., 1997; Perren et al., 1999). Therefore, we inhibitor, respectively. Taken together, these data have examined the effects of the EGFR tyrosine kinase

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2820 inhibitor ZD1839 on EGFR-overexpressing tumor cells with EGFR inhibitors in patients with advanced that lack PTEN function. metastatic cancers (Baselga et al., 2002; Ranson et al., Submicromolar concentrations of ZD1839 markedly 2002), it is possible that those responses are limited to inhibited the growth of A431 but not PTEN-null MDA- tumors without amplification of PI3K-Akt signaling and 468 cells. Of note, at concentrations that did not inhibit in which inhibition of the EGFR was sufficient to inhibit MDA-468 cell growth, ZD1839 induced loss of EGFR this postreceptor signaling pathway. phosphorylation, dissociation of EGFR from p85a,and inhibition of MAPK activity. Potentially, owing to the low level of HER3 in these cells, we were unable to Materials and methods detect HER3 protein in EGFR precipitates from MDA- 468 cells, suggesting the possibility that (1) EGFR- Cell lines and kinase inhibitors associated HER3 levels are too low to be detected by A431 and MDA-468 cancer cell lines were purchased from the immunoblot, and/or (2) in these cells, an unidentified American Type Culture Collection (Manassas, VA, USA) and adaptor molecule couples the EGFR to p85a. These maintained in Improved Minimal Essential Zinc Option possibilities will require additional investigation. De- Medium (IMEM, Life Technologies, Inc., Rockville, MD, spite its ability to inhibit EGFR phosphorylation in USA) supplemented with 10% fetal calf serum (FCS), 100 U/ MDA-468 cells, ZD1839 treatment did not inhibit P-Akt ml penicillin, and 100 mg/ml streptomycin. All NSCLC lines levels. This result suggests that (1) in MDA-468 cells, were provided by Freddy Kaye (NCI/Naval Medical Oncol- PI3K and Akt activities are the result of at least two ogy, Bethesda, MD, USA) and maintained in DMEM/10% inputs, EGFR autocrine signals and loss of PTEN, and FCS (Life Technologies, Inc.) supplemented with penicillin and streptomycin (Brognard et al., 2001). LY294002 was (2) the lack of inhibition of PI3K and Akt contributes to purchased from BIOMOL (Plymouth Meeting, PA, USA) and the resistance to ZD1839. reconstituted in DMSO. ZD1839 (Iressa) was provided by To test these possibilities, we stably reconstituted Steven Averbuch (AstraZeneca Pharmaceuticals, Wilmington, PTEN in MDA-468 cells. Re-expression of PTEN DE, USA). OSI-774 (Tarceva) was a gift from Mark resulted in a reduction in basal PI3K activity and P- Sliwkowski (Genentech, Inc., South San Francisco, CA, Akt levels but did not change total Akt and EGFR USA). Both ZD1839 and OSI-774 were reconstituted in protein levels. However, it restored the ability of DMSO and stored as 10-mm stock solutions. C225 (Cetux- ZD1839 to inhibit active Akt as well as to induce imab) was provided by Dan Hicklin (Imclone Systems, Inc., nuclear localization of FKHRL1. Simultaneous with New York, NY, USA). these biochemical responses, treatment of ZD1839 resulted in marked induction of apoptosis and a modest Cell proliferation and apoptosis assays delay in cell cycle progression in PTEN-expressing but Cells were seeded in full medium in 12-well plates at a density not PTEN-null MDA-468 cells. These results imply that of 3 Â 104 cells/well in triplicate; ZD1839 or LY294002 were re-expression of PTEN in EGFR-overexpressing cells added the next day. Medium and inhibitors were replenished restores the receptor dependence of PI3K and Akt every other day until cells were harvested by trypsinization and activation. This difference in sensitivity to EGFR counted with a Zeiss Coulter Counter (Beckman Coulter, Miami, FL, USA). To measure apoptosis, cells were seeded in inhibitors was also seen with the EGFR antibody 5 C225 and the small-molecule tyrosine kinase inhibitor full growth medium in 60-mm dishes at the density of 2 Â 10 cells/dish. The following day, the medium with or without OSI-774. Furthermore, the differential sensitivity to ZD1839 or LY294002 was changed. Both adherent and ZD1839 observed between A431 and MDA-468 cells floating cells were harvested and pooled 72 h later and and between MDA-468/vector and MDA-468 cells was subjected to TUNEL assay utilizing the Apo-BrdU assay kit also observed between PTEN-positive H1355 and (Phoenix Flow Systems, San Diego, CA, USA) according to the PTEN-mutant H157 NSCLC cells. manufacturer’s protocol. Detection and quantitation of FITC- There are several implications that can be derived positive (apoptotic) cells were done with a FACS/Calibur Flow from these data. For example, EGFR-overexpressing Cytometer (Becton Dickinson, Mansfield, MA, USA). tumors with concomitant amplification(s) of PI3K-Akt signaling may benefit from the combined blockade of Immunoblot analysis and immunoprecipitation the EGFR and Akt kinases. In addition to loss of PTEN After washes with phosphate-buffered saline (PBS) on ice, cell function, the possible mechanisms that will result in a monolayers were lysed in a buffer containing 20 mm Tris, pH high threshold of PI3K and Akt activity will include 7.4, 150 mm NaCl, 1% Nonidet P-40, 20 mm NaF, 1 mm overexpression of PI3K isoforms (Shayesteh et al., 1999; Na3VO4,1mm PMSF, 2 mg/ml aprotinin, and 2 mg/ml Fry, 2001), Akt isoforms (Bellacosa et al., 1995; leupeptin. Equal amount of protein in whole-cell lysates (as Nakatani et al., 1999; Sun et al., 2001a, b), the measured by Bradford) were separated by 8% (for EGFR, cytoplasmic tyrosine kinases like Src and BRK (re- HER3, p85, and P-Tyr) or 12% (for Akt and MAPK) SDS– viewed in Fry, 2001), and heterologous receptor net- PAGE and transferred to nitrocellulose membranes. Immuno- works like the IGF-I receptor (Yu and Rohan, 2000). blot analyses were performed as described previously (Moulder et al., 2001; Yakes et al., 2002) using the following Whether subjects bearing EGFR-expressing cancers antibodies: polyclonal EGFR Ab-12 (NeoMarkers, Freemont, with aberrant PI3K-Akt activity are less responsive to CA, USA); polyclonal HER3 C-17 and monoclonal PTEN EGFR inhibitors can be tested prospectively in current A2B1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA); p85 clinical trials with EGFR inhibitors. Secondly, as and monoclonal P-Tyr (Upstate Biotechnology, Lake Placid, clinical responses have been reported in recent trials NY, USA); S473 P-Akt, total Akt, and total MAPK (New

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2821 England BioLabs, Beverly, MA, USA); T202/Y204 P-MAPK FKHRL1 transfections and immunofluorescence (Promega Corp., Madison, WI, USA). Horseradish perox- An expression vector encoding WTHA-tagged FKHRL1 was idase-linked anti-rabbit or anti-mouse secondary antibodies provided by Anne Brunet (Harvard Medical School, Boston, were used as described (Yakes et al., 2002) and immunor- MA, USA). Cells were grown in 60-mm dishes (106 cells/dish) eactive bands visualized by enhanced chemiluminescence and then transfected with 10 mg HA-FKHRL1 for 16 h using (ECL; Roche Molecular Biochemicals, Indianapolis, IN, FUGENE 6 (Roche). Cells were transferred to coverslips on USA). For immunoprecipitation, 300 mg of protein from 12-well plates, incubated with 1 mm ZD1839 for 24 h, and then whole-cell lysates were incubated overnight at 41C with 3 mg subjected to immunostaining with anti-HA fluorescein (1 : 500; of EGFR Ab-12 (Neomarkers); protein A-Sepharose (Sigma, Roche) as described (Shin et al., 2001). For detection of nuclei, St Louis, MO, USA) was then added for 2 h while rocking. The cells were incubated with 1 mg/ml Hoechst 33258 (Sigma) after precipitates were washed four times with ice-cold PBS, incubation with secondary antibody. Coverslips were mounted resuspended in Â 6 Laemmli buffer, and resolved by SDS– on glass slides with AquaPolyMount (Polysciences, Warring- PAGE followed by Western blot analysis. ton, PA, USA) and examined by laser scanning confocal microscopy (Carl Zeiss LSM410). Retroviral infection WT-PTEN (Lu et al., 1999) was subcloned into the pBMN- Transcriptional reporter assays IRES-EGFP retroviral vector (Grignani et al., 1998) (provided Cells were seeded at a density of 105 cells/well in 12-well plates. by Gary Nolan, Stanford University). 293 Phoenix amphi- On the following day, the cells were transfected with 0.5 mg/ tropic retrovirus-packaging cells (ATCC), expressing gag-pol well of FHRE-Luciferase (provided by Anne Brunet) and and envelope proteins, were plated in a 60-mm dish at the 0.005 mg pCMV-R1 (Promega) for 16 h with the use of density of 2.5 Â 106 cells/dish. After transfection of 3 mg FUGENE 6. Transfected cells were treated or not treated pBMN-IRES-EGFP (vector control) or pBMN-PTEN-IRES- with 1 mm ZD1839 for 24 h. Firefly and Renilla reniformis EGFP for 24 h using FUGENE 6 (Roche Molecular Bio- luciferase activities in cell lysates were determined with the chemicals), virus was obtained by allowing the producing cells Dual Luciferase Assay System (Promega) in a Monolight 2001 to reach confluence, removing the medium, and replacing it luminometer (Analytical Luminescence Laboratory) as de- with half-volume of fresh serum-free medium. The conditioned scribed previously (Shin et al., 2001). Firefly luciferase activity medium was collected 48–72 h later and passed through a 45- was normalized to Renilla reniformis luciferase activity. mm filter. MDA-468 cells (5 Â 105) were plated on a 100-mm dish in IMEM/10% FCS. The following day, the medium was Cell cycle analysis removed and replaced with 3 ml of a cocktail containing the retroviral supernatant (at an MOI of 80 plaque-forming units/ Cells were harvested by trypsinization, fixed in ethanol, and cell), 4 mg/ml polybrene, and IMEM/10% FCS. After 6 h, the labeled with 50 mg/ml propidium iodide (Sigma) containing medium was changed with fresh medium and the cells collected 125 U/ml protease-free RNAse (Calbiochem) as described 48 h later. GFP-positive cells were sorted by flow cytometry previously (Busse et al., 2000). Cells were filtered through a and expanded in full medium. PTEN expression was confirmed 95-mm pore size nylon mesh (Small Parts, Inc., Miami Lakes, by immunoblot analysis. FL, USA). A total of 15 000 labeled nuclei were analysed in a FACS/Calibur Flow Cytometer (Becton Dickinson). Phosphatidylinositol-3 kinase in vivo assay Flow cytometry of cell surface EGFR Approximately 2 Â 105 MDA-468/vector and MDA-468/ EGFR levels in H157 and H1355 cells were estimated by flow PTEN cells/well were plated into six-well plates. The following cytometry using the 528 EGFR monoclonal antibody (Santa day, cells were labeled for 16 h with 100 mCi/ml of 32P- Cruz Biotechnology) and a nonspecific IgG (Sigma) follow- orthophosphate (sp. act. 10 mCi/ml; Amersham, Piscataway, 2a ing a previously described protocol (Moulder, 2001). NJ, USA) on phosphate-free DMEM with 1 mm ZD1839 or 20 mm LY294002. Lipids were then extracted from the labeled cells and 32P-labeled phospholipids analysed by thin-layer chromatography (TLC) as described previously [Chan, 2002 Acknowledgements #533]. The intensity of the 32P-IP3, 4, 5P3 band was This work was supported in part by NIH Grant R01 CA80195 quantitated using Image Gauge (version 4.0, Fuji) and (to CLA) and Vanderbilt-Ingram Comprehensive Cancer expressed as a percentage relative to untreated cells. Center support Grant CA68485.

References

Albanell J, Codony-Servat J, Rojo F, Del Campo JM, Sauleda S, Bellacosa A, de Feo D, Godwin AK, Bell DW, Cheng JQ, Anido J, Raspall G, Giralt J, Rosello J, Nicholson RI, Altomare DA, Wan M, Dubeau L, Scambia G, Masciullo V Mendelsohn J and Baselga J. (2001). Cancer Res., 61, et al. (1995). Int. J. Cancer, 64, 280–285. 6500–6510. Brognard J, Clark AS, Ni Y and Dennis PA. (2001). Cancer Ali IU, Schriml LM and Dean M. (1999). J. Natl. Cancer Inst., Res., 61, 3986–3997. 91, 1922–1932. Brunet A, Bonni A, Zigmond MJ, Lin MZ, Juo P, Hu LS, Arteaga CL. (2001). J. Clin. Oncol., 19, 32S–40S. Anderson MJ, Arden KC, Blenis J and Greenberg ME. Baselga J, Rischin D, Ranson M, Calvert H, Raymond E, (1999). Cell, 96, 857–868. Kieback DG, Kaye SB, Gianni L, Harris A, Bjork T, Busse D, Doughty RS, Ramsey TT, Russell WE, Price JO, Averbuch SD, Feyereislova A, Swaisland H, Rojo F, & Flanagan WM, Shawver LK and Arteaga CL. (2000). J. Albanell J. (2002). J. Clin. Oncol., 20, 4292–4302. Biol. Chem., 275, 6987–6995.

Oncogene RH Loss of PTEN/MMAC1/TEP R Bianco et al 2822 Cantley LC and Neel BG. (1999). Proc. Natl. Acad. Sci. USA, Nelson JM and Fry DW. (2001). J. Biol. Chem., 276, 14842– 96, 4240–4245. 14847. Carpenter G. (1999). J. Cell Biol., 146, 697–702. Olayioye MA, Neve RM, Lane HA and Hynes NE. (2000). Chakravarti A, Loefﬂer JS and Dyson NJ. (2002). Cancer Res., EMBO J., 19, 3159–3167. 62, 200–207. Perren A, Weng LP, Boag AH, Ziebold U, Thakore K, Dahia Chan TO, Rittenhouse SE and Tsichlis PN. (1999). Annu. Rev. PL, Komminoth P, Lees JA, Mulligan LM, Mutter GL and Biochem., 68, 965–1014. Eng C. (1999). Am. J. Pathol., 155, 1253–1260. Datta SR, Brunet A and Greenberg ME. (1999). Genes Dev., Prenzel N, Fischer OM, Streit S, Hart S and Ullrich A. (2001). 13, 2905–2927. Endocr. Relat. Cancer, 8, 11–31. D’Cruz CM, Gunther EJ, Boxer RB, Hartman JL, Sintasath Ranson M, Hammond LA, Ferry D, Kris M, Tullo A, Murray L, Moody SE, Cox JD, Ha SI, Belka GK, Golant A, Cardiff PI, Miller V, Averbuch S, Ochs J, Morris C, Feyereislova A, RD and Chodosh LA. (2001). Nat. Med., 7, 235–239. Swaisland H and Rowinsky EK. (2002). J. Clin. Oncol., 20, Ennis BW, Valverius EM, Bates SE, Lippman ME, Bellot F, 2240–2250. Kris R, Schlessinger J, Masui H, Goldenberg A and Rodrigues GA, Falasca M, Zhang Z, Ong SH and Schlessinger Mendelsohn J et al. (1989). Mol. Endocrinol., 3, 1830–1838. J. (2000). Mol. Cell Biol., 20, 1448–1459. Fan Z, Shang BY, Lu Y, Chou JL and Mendelsohn J. (1997). Shayesteh L, Lu Y, Kuo WL, Baldocchi R, Godfrey T, Collins Clin. Cancer Res., 3, 1943–1948. C, Pinkel D, Powell B, Mills GB and Gray JW. (1999). Nat. Fedi P, Pierce JH, di Fiore PP and Kraus MH. (1994). Mol. Genet., 21, 99–102. Cell Biol., 14, 492–500. Shin I, Bakin AV, Rodeck U, Brunet A and Arteaga CL. Forgacs E, Biesterveld EJ, Sekido Y, Fong K, Muneer S, (2001). Mol. Biol. Cell, 12, 3328–3339. Wistuba II, Milchgrub S, Brezinschek R, Virmani A, Gazdar Shin I, Yakes PM, Rojo R, Shin NY, Bakin AV, Baselga J and AF and Minna JD. (1998). Oncogene, 17, 1557–1565. Arteaga CL. (2002). Nat. Med., 8, 1145–1152. Fry MJ. (2001). Breast Cancer Res., 3, 304–312. Simpson L and Parsons R. (2001). Exp. Cell Res., 264, 29–41. Goldstein NI, Prewett M, Zuklys K, Rockwell P and Sun H, Lesche R, Li DM, Liliental J, Zhang H, Gao J, Mendelsohn J. (1995). Clin. Cancer Res., 1, 1311–1318. Gavrilova N, Mueller B, Liu X and Wu H. (1999). Proc. Grignani F, Kinsella T, Mencarelli A, Valtieri M, Riganelli D, Natl. Acad. Sci. USA, 96, 6199–6204. Lanfrancone L, Peschle C, Nolan GP and Pelicci PG. (1998). Sun M, Paciga JE, Feldman RI, Yuan Z, Coppola D, Lu YY, Cancer Res., 58, 14–19. Shelley SA, Nicosia SV and Cheng JQ. (2001a). Cancer Res., Gschwind A, Zwick E, Prenzel N, Leserer M and Ullrich A. 61, 5985–5991. (2001). Oncogene, 20, 1594–1600. Sun M, Wang G, Paciga JE, Feldman RI, Yuan ZQ, Ma XL, Hermanto U, Zong CS and Wang LH. (2001). Oncogene, 20, Shelley SA, Jove R, Tsichlis PN, Nicosia SV and Cheng JQ. 7551–7562. (2001b). Am. J. Pathol., 159, 431–437. Kohno T, Takahashi M, Manda R and Yokota J. (1998). Teng DH, Hu R, Lin H, Davis T, Iliev D, Frye C, Swedlund B, Genes Chromosomes Cancer, 22, 152–156. Hansen KL, Vinson VL, Gumpper KL, Ellis L, El-Naggar Kops GJ and Burgering BM. (1999). J. Mol. Med., 77, 656– A, Frazier M, Jasser S, Langford LA, Lee J, Mills GB, 665. Pershouse MA, Pollack RE, Tornos C, Troncoso P, Yung Lane HA, Beuvink I, Motoyama AB, Daly JM, Neve RM and WK, Fujii G, Berson A and Steck PA et al. (1997). Cancer Hynes NE. (2000). Mol. Cell Biol., 20, 3210–3223. Res., 57, 5221–5225. Lenferink AE, Busse D, Flanagan WM, Yakes FM and Van de Vijver MJ, Kumar R and Mendelsohn J. (1991). J. Arteaga CL. (2001). Cancer Res., 61, 6583–6591. Biol. Chem., 266, 7503–7508. Lenferink AE, Simpson JF, Shawver LK, Coffey RJ, Forbes Vivanco I and Sawyers CL. (2002). Nat. Rev. Cancer, 2, 489– JTand Arteaga CL. (2000). Proc. Natl. Acad. Sci. USA, 97, 501. 9609–9614. Vlahos CJ, Matter WF, Hui KY and Brown RF. (1994). J. Li DM and Sun H. (1998). Proc. Natl. Acad. Sci. USA, 95, Biol. Chem., 269, 5241–5248. 15406–15411. Vogel CL, Cobleigh MA, Tripathy D, Gutheil JC, Harris LN, Lu Y, Lin YZ, LaPushin R, Cuevas B, Fang X, Yu SX, Davies Fehrenbacher L, Slamon DJ, Murphy M, Novotny WF, MA, Khan H, Furui T, Mao M, Zinner R, Hung MC, Steck Burchmore M, Shak S, Stewart SJ and Press M. (2002). J. P, Siminovitch K and Mills GB. (1999). Oncogene, 18, 7034– Clin. Oncol., 20, 719–726. 7045. Wakeling AE, Guy SP, Woodburn JR, Ashton SE, Curry BJ, Lu Y, Zi X, Zhao Y, Mascarenhas D and Pollak M. (2001). J. Barker AJ and Gibson KH. (2002). Cancer Res., 62, 5749– Natl. Cancer Inst., 93, 1852–1857. 5754. Moasser MM, Basso A, Averbuch SD and Rosen N. (2001). Yakes FM, Chinratanalab W, Ritter CA, King W, Seelig S Cancer Res., 61, 7184–7188. and Arteaga CL. (2002). Cancer Res., 62, 4132–4141. Moulder SL, Yakes FM, Muthuswamy SK, Bianco R, Yarden Y and Sliwkowski MX. (2001). Nat. Rev. Mol. Cell Simpson JF and Arteaga CL. (2001). Cancer Res., 61, Biol., 2, 127–137. 8887–8895. Yu H and Rohan T. (2000). J. Natl. Cancer. Inst., 92, 1472– Moyer JD, Barbacci EG, Iwata KK, Arnold L, Boman B, 1489. Cunningham A, DiOrio C, Doty J, Morin MJ, Moyer MP, Zhou BP, Liao Y, Xia W, Spohn B, Lee MH and Hung MC. Neveu M, Pollack VA, Pustilnik LR, Reynolds MM, Sloan (2001a). Nat. Cell Biol., 3, 245–252. D, Theleman A and Miller P. (1997). Cancer Res., 57, 4838– Zhou BP, Liao Y, Xia W, Zou Y, Spohn B and Hung MC. 4848. (2001b). Nat. Cell Biol., 3, 973–982. Nakatani K, Thompson DA, Barthel A, Sakaue H, Liu W, Weigel RJ and Roth RA. (1999). J. Biol. Chem., 274, 21528– 21532.

Oncogene