The role of protein tyrosine phosphatase 1B in Ras signaling

Nadia Dube´ *, Alan Cheng*, and Michel L. Tremblay†

McGill Cancer Centre and Department of Biochemistry, McGill University, 3655 Promenade Sir-William-Osler, Room 715, Montreal, QC, Canada H3G 1Y6

Edited by Jack E. Dixon, University of California at San Diego School of Medicine, La Jolla, CA, and approved December 19, 2003 (received for review July 8, 2003) Protein tyrosine phosphatase (PTP) 1B has been implicated as a mediated Erk and Src activation in PTP1B-deficient fibroblasts negative regulator of multiple signaling pathways downstream of are both impaired (12). However, most of our studies were receptor tyrosine kinases. Inhibition of this enzyme was initially performed with cells immortalized with the SV40 large T antigen thought to potentially lead to increased oncogenic signaling and (TAg). Importantly, TAg has been shown to abrogate the tumorigenesis. Surprisingly, we show that platelet-derived growth requirements of Src kinases during PDGF-induced mitogenesis factor-stimulated extracellular-regulated kinase signaling in (13), and PDGF-induced Erk activation in TAg-immortalized PTP1B-deficient cells is not significantly hyperactivated. Moreover, fibroblasts lacking Src kinases is relatively unchanged (14). This these cells exhibit decreased Ras activity and reduced proliferation result suggested to us that there were additional mechanisms in by way of previously uncharacterized pathways. On immortaliza- PTP1B-deficient fibroblasts that were responsible for the dimin- tion, PTP1B-deficient fibroblasts display increased expression of ished Erk activity. Ras GTPase-activating protein (p120RasGAP). Furthermore, we In this study, we show that, although PTP1B-deficient cells demonstrate that p62Dok (downstream of tyrosine kinase) is a exhibit increased PDGFR and AKT phosphorylation, Erk acti- putative substrate of PTP1B and that tyrosine phosphorylation of vation does not occur to the same extent. We show that loss of p62Dok is indeed increased in PTP1B-deficient cells. Consistent PTP1B results in diminished Ras activity and that this event with the decreased Ras activity in cells lacking PTP1B, introduction occurs through increased p120RasGAP (Ras GTPase-activating of constitutively activated Ras restored extracellular-regulated protein) expression and p62Dok (downstream of tyrosine ki- kinase signaling and their proliferative potential to those of WT nase) phosphorylation. Taken together, these results propose cells. These results indicate that loss of PTP1B can lead to decreased how PTP1B can act as a positive regulator of Ras signaling Ras signaling, despite enhanced signaling of other pathways. This downstream of RTKs and may in part explain why PTP1B finding may in part explain the absence of increased tumor inci- knockout mice do not present an increased incidence of tumors. dence in PTP1B-deficient mice. Thus, PTP1B can positively regulate Ras activity by acting on pathways distal to those of receptor Methods tyrosine kinases. Antibodies. Rabbit polyclonal antibodies against PTP1B have been described (3). Additional antibodies were purchased from rotein tyrosine phosphatase (PTP) 1B is the prototype for Cell Signaling Technology (Beverly, MA) (pan and phospho Pthe superfamily of PTPs and has been implicated in multiple anti-AKT, anti-Erk); Upstate Biotechnology (Lake Placid, NY) signaling pathways (1). Of particular interest, gene-targeting (anti-phosphotyrosine 4G10); Santa Cruz Biotechnology (anti- ␤ studies in mice have established PTP1B as a critical physiological PDGFR , anti-Dok1); Transduction Laboratories (Lexington, regulator of metabolism by attenuating insulin, leptin, and KY) (anti-p190RhoGAP, anti-p120RasGAP, anti-H-Ras, and growth hormone signaling (2–6). PTP1B function seems to be anti-Shc); and BioSource International (Camarillo, CA) (anti- dispensable for embryonic development. However, PTP1B- Src). deficient mice exhibit resistance to diabetes and obesity, the two major metabolic diseases in industrialized societies. Not surpris- Cell Culture and Cell Lines. All cell lines were maintained in ingly, PTP1B is a highly regarded target of the pharmaceutical DMEM (Invitrogen) supplemented with 10% FBS (BioSource ͞ ͞ industry in the treatment of these disorders (7). International) and antibiotics (5 mg ml penicillin streptomycin, Because PTP1B is a negative regulator of multiple receptor Invitrogen). All spontaneously immortalized fibroblast cell lines tyrosine kinases (RTKs) (1), the concern is that PTP1B inhibi- derived from PTP1B or T cell PTP (TCPTP) knockout embryos tion may lead to increased oncogenic signaling. Indeed, PTP1B- have been described (9, 15). The SV40 TAg cells were rescued deficient fibroblasts display increased insulin-like growth factor as described (9). I (IGF-I) receptor, epidermal growth factor receptor, and platelet-derived growth factor receptor (PDGFR) tyrosine phos- Generation of the V12Ras Clones. PTP1B WT or knockout spon- phorylation (8, 9). Regardless of this potentially enhanced taneously immortalized fibroblasts were transfected with a lin- oncogenic signaling, PTP1B-deficient mice do not overtly un- earized V12Ras vector and a hygromycin selection vector dergo tumorigenesis. One possibility is that PTP1B may not (pMC1-HygR-pA) at a 10:1 ratio, respectively. Transfected cells ␮ ͞ regulate RTK signaling in all cell types, or that functional were selected in DMEM containing 50–75 g ml hygromycin, redundancy may exist. Alternatively, loss of PTP1B may affect and colonies were picked. Screening of positive clones was done the progression of a tumorigenic event, but not its rate of initiation. Finally, it is also possible that PTP1B may be involved This paper was submitted directly (Track II) to the PNAS office. in the activation of oncogenic pathways downstream of RTKs. Abbreviations: IGF-I, insulin-like growth factor I; MAPK, mitogen-activated protein kinase; We decided to pursue the third alternative based on our Erk, extracellular-regulated kinase; p62Dok, downstream of tyrosine kinase; p120RasGAP, previous studies with PTP1B in the IGF-I receptor pathway. Ras GTPase-activating protein; PDGFR, platelet-derived growth factor receptor; RTK, re- Paradoxically, IGF-I-stimulated extracellular-regulated kinase ceptor tyrosine kinase; TAg, SV40 large T antigen; PTP, protein tyrosine phosphatase; (Erk) phosphorylation in PTP1B-deficient fibroblasts is signif- TCPTP, T cell PTP. icantly diminished. This finding could, in part, be explained by *N.D. and A.C. contributed equally to this work. previous results suggesting that PTP1B is involved in the acti- †To whom correspondence should be addressed. E-mail: [email protected]. vation of Src (10, 11). Indeed, we demonstrated that adhesion- © 2004 by The National Academy of Sciences of the USA

1834–1839 ͉ PNAS ͉ February 17, 2004 ͉ vol. 101 ͉ no. 7 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0304242101 Downloaded by guest on September 27, 2021 by Western blotting by using anti-H-Ras antibody. In each case, at least 10 clones were isolated and further characterized.

Preparation of Cell Lysate and Immunoblotting. Cells were washed in ice-cold PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl͞50 mM Tris⅐HCl, pH 7.5͞1% ͞ ͞ ͞ Nonidet P-40 0.25% sodium deoxycholate 1mMNa3VO4 50 mM NaF) supplemented with Complete EDTA-free protease inhibitor mixture (Roche Molecular Biochemicals). Cell lysates were rotated end-over-end at 4°C for 10 min and cleared by centrifugation at 14,000 ϫ g for 10 min at 4°C. The protein concentration was measured by the Bradford method (Bio-Rad). Protein samples were resolved by 8% SDS͞PAGE and subjected to immunoblotting with the indicated antibodies.

Immunoprecipitation. Cell lysates were incubated with anti- p62Dok antibodies and 20 ␮l of protein G-Agarose (Invitrogen) at 4°C for 2 h. Precipitates were washed in lysis buffer, resus- pended in SDS sample buffer, and resolved by 8% SDS͞PAGE for immunoblot analysis.

Semiquantitative RT-PCR. Total RNA was isolated from cells by using TRIzol (Invitrogen), and first-strand cDNA synthesis was obtained from 1 ␮g of RNA with random hexamers by using SuperScript II reverse transcriptase (Invitrogen) according to the manufacturer’s instructions. The cDNA was then used as a template for PCR with two sets of synthesized primers. Aliquots of 1 ␮l of the reverse transcription reaction were amplified by using 1 unit of AmpliTaq Gold (Applied Biosystems) under the following PCR conditions: 5 min at 95°C, then 30 s at 94°C, 30 s Fig. 1. Differential regulation of PDGF-signaling pathways by PTP1B. (a) at 60°C, and 30 s at 72°C for 30 cycles, 2 min at 72°C) in a 25-␮l Fibroblasts immortalized with SV40 TAg (ϩ͞ϩ PTP1B and Ϫ͞Ϫ PTP1B) were reaction mixture, by using 100 pmol each of the sense and serum-starved and stimulated with 10 ng͞ml PDGF for indicated times or left antisense primers: 5ЈRasGAP, 5Ј-GGGTGTTTACAGAAAT- unstimulated. The lysates were analyzed by Western blotting, and the mem- CAGTTC-3Ј; and 3ЈRasGAP, 5Ј-CTCATTGCTGAGTGT- brane was probed with anti-phosphotyrosine antibodies. (b) The samples Ј were analyzed for Erk and AKT proteins and phosphorylation levels by using TCTCAG-3 . In parallel, a GAPDH PCR was performed to phosphospecific antibodies. control for the RNA input in the RT-PCR: 5ЈGAPDH, 5Ј- AACGACCCCTTCATTGAC-3Ј; and 3ЈGAPDH, 5Ј-TCCAC- GACATACTCAGCAC-3Ј. Reaction products were separated by 1% agarose gel electrophoresis and detected with ethidium Ras-GTP Pull-Down Assay. Cells were lysed in MLB buffer (25 mM ͞ ͞ ͞ bromide staining. Hepes, pH 7.5 150 mM NaCl 1% Nonidet P-40 0.25% sodium ͞ ͞ ͞ ͞ deoxycholate 10% glycerol 25 nM NaF 1mMNa3VO4 10 mM ͞ Subtrate-Trapping Experiments. Details of the PTP1B WT and MgCl2 1 mM EDTA) supplemented with Complete EDTA-free

D181A constructs have been described (16). NIH 3T3 c-Src protease inhibitor mixture. The level of Ras-GTP was deter- BIOCHEMISTRY Y527F cells were transfected by Lipofectamine (Invitrogen) mined by precipitation with a GST fusion protein of the Ras- according to the manufacturer’s instructions. Forty-eight hours binding domain on Raf1, which recognizes only active, GTP- posttransfection, cells were lysed in buffer containing 50 mM bound Ras. Pull-downs were resolved by SDS PAGE and Hepes (pH 7.5), 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, immunoblotted with an anti-Ras antibody to detect precipitated 1% Triton X-100, and 10% glycerol (supplemented with Com- Ras-GTP. plete EDTA-free protease inhibitors). Crude lysates were then cleared by centrifugation at 14,000 ϫ g. GST-tagged proteins Results ␮ were precipitated by using 25 l of Glutathione Sepharose beads Differential Regulation of Downstream Pathways of the PDGFR by (Pharmacia), washed extensively in lysis buffer, and then resus- PTP1B. To investigate the mechanisms by which loss of PTP1B pended in SDS sample buffer. Aliquots were resolved by SDS͞ can attenuate Erk activation, we used the PDGF signaling system PAGE and analyzed by immunoblotting with the indicated antibodies. in TAg immortalized cells (TAg cells), for which Src function is dispensable (13, 14). Previous studies with PTP1B-deficient TAg Soft Agar Assay. PTP1B WT and knockout cells stably expressing cells showed that PDGF-stimulated Erk and AKT activation V12Ras were assessed for anchorage-independent growth by were not dramatically altered (8, 9). We used PDGF at a high ͞ colony formation in soft agar. NIH 3T3 c-Src Y527F cells were concentration (50 ng ml), allowing us to suspect that subtle used as positive control. The cells were plated at 3 ϫ 103 cells per changes by PTP1B may not be detectable in this case. Hence, in ͞ well in a six-well plate in triplicate, by using 0.35% low melting our studies, we used PDGF at much lower levels (10–20 ng ml). point agarose and grown in DMEM with 20% FBS. Media were As expected, stimulation of TAg cells with low levels of PDGF changed every 3 days. Colony number was determined by scoring resulted in increased cellular tyrosine phosphorylation and AKT for those with a size Ͼ0.1 mm in size. Representative colonies activation, which was enhanced in PTP1B-deficient cells (Fig. 1). were photographed in phase contrast from plates at day 10 of the In contrast, PDGF-induced Erk phosphorylation was decreased assay (ϫ10 magnification). Values are reported as the average of in PTP1B knockout cells. Thus, PTP1B seems to differentially three experiments Ϯ SE. regulate signaling pathways diverging from the PDGFR.

Dube´ et al. PNAS ͉ February 17, 2004 ͉ vol. 101 ͉ no. 7 ͉ 1835 Downloaded by guest on September 27, 2021 Fig. 2. Decreased cell growth and Ras activity of SV40 TAg transformed fibroblasts lacking PTP1B. (a and b) Cells were seeded in a 24-well plate at a density of 1 ϫ 104 cells per well in 10% or 1% FBS. At the indicated time points, the cells were trypsinized, and the total cell number per well was determined with a hemacytometer. Values are reported as the average of triplicate Fig. 3. PTP1B-deficient cells display increased p120RasGAP expression. (a) experiments Ϯ SE. (c) PDGFR expression is increased in cells lacking PTP1B. Increased expression of p120RasGAP but not p190RhoGAP in TAg immortal- Serum-starved cells were stimulated with PDGF for indicated times or left ized PTP1B-deficient fibroblasts. (b) Increased expression of RasGAP is also unstimulated. The lysates were analyzed by Western blotting for PDGFR seen in spontaneously immortalized PTP1B-deficient fibroblasts. (c) Stable Ϫ͞Ϫ expression. Equal loading was done by reprobing the membrane with anti- reexpression of PTP1B into PTP1B cells decreases RasGAP levels. R3 and R5 Ϫ͞Ϫ Ϫ͞Ϫ bodies against AKT͞protein kinase B. (d) Serum-starved cells were stimulated are cells that expressed PTP1B, and R9 and R10 are cells mock Ϫ͞Ϫ with 20 ng͞ml PDGF for 10 min or left unstimulated. Lysates were incubated transfected. (d) Expression level of RasGAP is not affected in TCPTP with immobilized GST-Raf1-RBD (Raf-RBD) to precipitate active (GTP-bound) immortalized fibroblasts. (e and f) Increased p120RasGAP mRNA in PTP1B- Ras. Ras-GTP was detected by using anti-H-Ras antibodies. deficient fibroblasts assessed by RT-PCR. GAPDH was used as a loading control, and lane C (no DNA) was used as a negative control.

PTP1B-Deficient Fibroblasts Display Decreased Growth and Ras Activ- ity. Consistent with our data that IGF-I- (9) and PDGF-induced altered. The importance of this finding is underlined by the fact activation of Erk is impaired in PTP1B-deficient cells, we also that p120RasGAP can attenuate Ras activity by promoting the observed that these cells display decreased monolayer growth intrinsic GTPase activity of Ras (19). compared with their WT counterparts (Fig. 2a). Similarly, To exclude the possibility that this phenomenon was due to although both WT and knockout cells were able to proliferate in TAg, we also analyzed p120RasGAP levels in spontaneously 1% serum, PTP1B-deficient cells clearly present a diminished immortalized cells, as well as primary fibroblasts. Similar to TAg capacity to do so (Fig. 2b). Moreover, PTP1B-deficient TAg cells cells, spontaneously immortalized PTP1B knockout cells also exhibit a consistent and reproducible elevation in PDGFR levels possess increased p120RasGAP levels compared with WT con- (Fig. 2c). This result is unlikely to be due to defective receptor trols (Fig. 3b). Interestingly, however, this effect was not seen in internalization and degradation because stimulation with PDGF primary cells, suggesting that an event during immortalization is caused a rapid decrease in receptor levels in both PTP1B WT required for this process (data not shown). To further exclude and knockout cells. the possibility of clonal variation effects, reexpression of myc- It has been established in fibroblasts that expression of the tagged PTP1B into knockout cells was able to restore PDGFR is inversely proportional to Ras activity (17). Further- p120RasGAP expression (Fig. 3c). Finally, to show that the more, Ras activity is necessary for full transformation by TAg increase in p120RasGAP expression was specific for PTP1B (18). By using a Ras-GTP pull down assay, we indeed show that knockout cells, we used TCPTP knockout cells as a control. PTP1B-deficient TAg cells display diminished Ras-GTP levels Although TCPTP knockout cells also display decreased cell (Fig. 2d). In addition, introduction of dominant active V12Ras proliferation (15), p120RasGAP levels are not altered (Fig. 3d). in both PTP1B WT and knockout cells was able to suppress PDGFR levels (data not shown). To determine whether there were increased levels of p120RasGAP mRNA in PTP1B-deficient fibroblasts, we per- Immortalization Increases p120RasGAP Expression in PTP1B-Deficient formed semiquantitative RT-PCR analysis by using the GAPDH Fibroblasts. To gain insight into the potential decrease in Ras gene as a control. As shown in Fig. 3E, this is indeed the case in activity in PTP1B-deficient cells, we first profiled several pro- two independent PTP1B-deficient TAg cell lines compared with teins upstream of the Erk signaling pathway. We found no two WT controls. In Fig. 3f, reexpression of PTP1B into significant alterations in the expression levels of the adapter PTP1B-deficient spontaneously immortalized cells restores Ras- proteins Shc or Grb2, or the kinases Src and Erk (data not GAP mRNA levels. Although the mechanism by which this event shown). Importantly, however, we observed that the levels of occurs is unclear, our results suggest that PTP1B expression is p120RasGAP are elevated in PTP1B-deficient TAg cells (Fig. required to suppress expression of the p120RasGAP during 3a). Yet, the levels of the related protein p190RhoGAP were not immortalization of fibroblasts.

1836 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0304242101 Dube´ et al. Downloaded by guest on September 27, 2021 We next explored the identity of pp60. Using a candidate approach, we revealed that pp60 is the adaptor protein p62Dok. As controls, two other tyrosine-phosphorylated proteins Ϸ60 kDa, Shc and Src, were not found to be precipitated by the PTP1B D181A mutant. To further confirm that PTP1B is a substrate of PTP1B, we analyzed the amounts of tyrosine- phosphorylated p62Dok in the PTP1B WT and knockout cells (Fig. 4b). As expected, p62Dok phosphorylation was increased in PTP1B-deficient TAg cells, and reexpression of myc-tagged PTP1B into these cells decreased p62Dok phosphorylation (Fig. 4c). Previous gene targeting studies have shown that p62Dok is a negative regulator of Erk signaling (23, 24). Furthermore, tyrosine phosphorylation of p62Dok can contribute to its inhib- itory effect on Ras (25). Thus, this result provides another mechanism by which PTP1B may regulate Ras activity.

V12Ras Rescues the Proliferative Potential of PTP1B Knockout Cells. Our results demonstrate that, in immortalized PTP1B-deficient fibroblasts, both p120RasGAP expression and p62Dok phos- phorylation are increased. These events can lead to the suppres- sion of the Erk pathway and lie upstream of Ras activation. Thus, to further confirm our findings, we tested the effect of stably introducing a dominant active Ras mutant (V12Ras) in our PTP1B WT and knockout cells. This mutant is known to inhibit its intrinsic GTPase activity, thus stabilizing the active GTP- bound form of Ras (26). Furthermore, this modification renders V12Ras independent from the modulation by upstream signals, such as RTKs and GAPs. For our experiments, we chose three clones each for PTP1B WT and knockout cells. In this group, clones M (mock) do not express V12Ras, which we treated as our negative controls. Fig. 4. p62Dok is a putative substrate of PTP1B. (a) NIH 3T3 c-Src Y527F cells Expression of V12Ras in the other clones (A ϩ͞ϩ,Bϩ͞ϩ,A were transfected with GST-vector (V), GST-PTP1B WT (WT), and GST-PTP1B Ϫ͞Ϫ,BϪ͞Ϫ) leads to an increase in Ras-GTP levels (Fig. 5a). D181A (DA). Cell lysates were subjected to pull-down with glutathione beads Furthermore, both PTP1B WT and knockout cells were effi- and then analyzed by immunoblotting with the indicated antibodies. TCL, ciently transformed by V12Ras, as judged by cell morphology total cell lysate. (b and c) Increased p62Dok phosphorylation in fibroblasts and growth in soft agar (Fig. 5 b–d). In fact, the PTP1B-deficient lacking PTP1B. Serum-starved cells were stimulated with 20 ng͞ml PDGF for 5 and 15 min or left unstimulated. p62Dok phosphorylation was analyzed by clones even displayed enhanced soft agar growth, probably immunoprecipitating p62Dok, probing with anti-phosphotyrosine antibod- correlating with the slightly higher levels of Ras-GTP in these ies, and then reprobing with anti-Dok1 antibodies. Lane C is antibody and cells (Fig. 5a). Retroviral expression of PTP1B in these clones beads alone as a control. Rescued, PTP1B Ϫ͞Ϫ cells that reexpressed PTP1B. was able to suppress approximately half of the number of colonies (Fig. 5e), likely through its effects in down-regulating AKT activity (9). Thus, V12Ras is able to transform fibroblasts, p62Dok Is a Putative Substrate for PTP1B. Previous studies demon- even in the absence of PTP1B. strated that p120RasGAP is tyrosine phosphorylated in cells BIOCHEMISTRY transformed by protein tyrosine kinases, including Src (20). Discussion Furthermore, tyrosine phosphorylation of p120RasGAP allows Overexpression and͞or activating mutations of at least 30 pro- it to bind to other proteins to contribute to its ability to inhibit tein tyrosine kinases have been linked to malignant transforma- the Ras͞mitogen-activated protein kinase (MAPK) pathway tion and cancer (27). In contrast, much less is known about the (19). To investigate whether p120RasGAP could be a potential role of PTPs in human diseases. PTP1B is the prototypical PTP, substrate of PTP1B, we used a D181A mutant of PTP1B that was and biochemical studies have implicated this enzyme in the previously shown to possess diminished catalytic activity but dephosphorylation of several RTKs (1). We previously demon- retain binding ability, thus producing a ‘‘substrate trapping strated that PTP1B-deficient fibroblasts display enhanced IGF- mutant’’ (21). I-mediated receptor phosphorylation and AKT activation (9). We coexpressed GST alone or GST-tagged PTP1B (WT or Paradoxically, IGF-I-stimulated Erk activation was significantly D181A) into NIH 3T3 cells stably expressing an activated Src impaired. Similarly, we also showed that PTP1B-deficient fibro- mutant (Src Y527F cells) and examined whether we could detect blasts exhibit impaired adhesion-mediated Erk activation (12). tyrosine-phosphorylated proteins in a complex with PTP1B In the present study, we provide evidence to explain how loss of proteins. Pull-downs of lysates were performed with glutathione PTP1B can lead to attenuation of Erk activation by way of beads and then analyzed by immunoblotting with a panel of impaired Ras signaling. antibodies. From Fig. 4a, it is clear that neither GST alone nor Stimulation of our TAg immortalized cells with PDGF re- GST-PTP1B (WT) could appreciably precipitate tyrosine- sulted in enhanced tyrosine phosphorylation of cellular proteins phosphorylated proteins. In contrast, GST-PTP1B (D181A) was in PTP1B knockout cells (Fig. 1a). Similarly, AKT phosphory- found to precipitate two such proteins of Ϸ60 and Ϸ130 kDa lation was also increased (Fig. 1b), but PDGF-stimulated Erk (pp60 and pp130). We were able to confirm the identity of pp130 activation was decreased in PTP1B knockout cells. In contrast, as p130Cas, previously shown to be a potential substrate of Haj et al. (8) reported only minor differences in Erk and AKT PTP1B (22). In contrast, probing with p120RasGAP antibodies activation using independently established PTP1B knockout cell failed to demonstrate binding to PTP1B, suggesting that lines. One possible reason for this discrepancy is that those p120RasGAP is not a substrate of PTP1B (data not shown). studies used high levels of PDGF (50 ng͞ml), which may not

Dube´ et al. PNAS ͉ February 17, 2004 ͉ vol. 101 ͉ no. 7 ͉ 1837 Downloaded by guest on September 27, 2021 Fig. 5. Transformation of PTP1B-deficient fibroblasts by activated Ras. (a) Expression of constitutively activated Ras leads to Ras activation. Cells were plated at 1 ϫ 105 per 10-cm dish and cultured in supplemented DMEM until confluent. The lysates were incubated with immobilized GST-Raf1-RBD (Raf-RBD) to precipitate active (GTP-bound) Ras. Ras-GTP was detected with anti-H-Ras antibodies. (b) Stable cell lines expressing V12Ras were grown near confluence. Representative clones were photographed in phase contrast. (c) Anchorage-independent growth of PTP1B-deficient fibroblasts. Stable cell lines were grown in soft agar for 10 days. Colonies were counted under a microscope, and pictures of representative clones were taken in phase contrast. (d) Summary of the results obtained in c. Values are reported as the average of triplicate Ϯ SE of three independent experiments. NIH 3T3 c-src Y527F were used as positive control. P, parental cell line; M, mock stable cell line (negative control); ϩ͞ϩV12Ras A and ϩ͞ϩV12Ras B, WT clones expressing V12Ras; Ϫ͞ϪV12Ras A and Ϫ͞ϪV12Ras B, PTP1B knockout clones expressing V12Ras. (e) Reexpression of PTP1B in V12Ras PTP1B Ϫ͞Ϫ fibroblasts decreases the colonies’ formation in soft agar. Both V12Ras PTP1B Ϫ͞Ϫ cell lines were infected with a retroviral vector encoding PTP1B and grown in soft agar for 10 days. Colonies were counted as described in d.

allow for detecting subtle differences. Indeed, this situation has attenuate Src activity. Thus, it is intriguing to speculate that been shown for ShcA knockout cells (28). Thus, in our studies, increased p62Dok phosphorylation in PTP1B knockout cells we used lower amounts of PDGF (10 ng͞ml). When we increased may contribute to decreased Src activity that was previously the levels of PDGF to 25 ng͞ml, the differences we observed observed during fibronectin signaling (12). More studies will be were diminished (data not shown). required to verify this hypothesis. Analysis of cell growth demonstrated that PTP1B is required If the impaired Ras signaling in PTP1B-deficient cells was due for efficient proliferation of immortalized TAg fibroblasts (Fig. to p120RasGAP and p62Dok, then V12Ras should be able to 2). Furthermore, we also noticed that PTP1B knockout cells had transform these cells similar to that of WT controls. Indeed this increased levels of the PDGFR (Fig. 3a), which suggested that is the case, and V12Ras-transformed PTP1B-deficient clones Ras activity was lower in these cells (17). Consistent with this actually grow slightly better in soft agar (Fig. 5c). One possibility notion, Ras activity has been demonstrated to be essential for is that these clones possess Ras-GTP levels slightly higher than transformation by TAg in several model systems (18, 29, 30). ͞ their WT counterparts (Fig. 5a). Alternatively, PTP1B-deficient Profiling of key proteins involved in the Ras MAPK pathway cells also display enhanced AKT activity, which may promote a revealed an increased expression of p120RasGAP that was due survival advantage. Nevertheless, our results show that V12Ras to up-regulated transcription͞message stability (Fig. 4). The can transform cells in the absence of PTP1B. significance of this finding is underscored by the ability of Taken together, a model can be put forward to explain the p120RasGAP to negatively regulate Ras activity by promoting the intrinsic GTPase activity of Ras (19). In addition, p62Dok, impaired Erk activation seen in TAg immortalized PTP1B- a binding partner of p120RasGAP, was found to be a potential deficient fibroblasts (Fig. 6). Loss of PTP1B leads to increased substrate of PTP1B (Fig. 5). Previously, an unidentified phos- RTK phosphorylation and enhanced signaling of most down- phorylated 60-kDa protein was shown to be a potential substrate stream pathways. However, loss of PTP1B can also lead to of PTP1B during epidermal growth factor signaling in COS cells cellular alterations that attenuate Erk signaling downstream of (21). It is likely that this protein is also p62Dok. Importantly, RTKs. For example, loss of PTP1B can lead to decreased Src p62Dok is a negative regulator of MAPK signaling (23, 24), and activation by way of increased phosphorylation of the inhibitory tyrosine phosphorylation of p62Dok can contribute to its inhib- site (12). In addition, loss of PTP1B leads to increased expres- itory effect on Ras (25). Collectively, this finding suggested that sion of p120RasGAP by way of an unidentified mechanism. loss of PTP1B in TAg cells leads to impaired Ras signaling. Finally, p62Dok, a potential substrate of PTP1B, is hyperphos- In cells with high levels of Src activity, p62Dok has been shown phorylated in PTP1B-deficient fibroblasts. All these events can to bind Csk, a negative regulator of Src kinases (31). The binding contribute to attenuate Ras activity and thus Erk signaling. is thought to recruit Csk to cytoskeletal compartments to Consistent with this model, introduction of V12Ras into PTP1B-

1838 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0304242101 Dube´ et al. Downloaded by guest on September 27, 2021 If loss of PTP1B leads to impaired Ras signaling, then what is the role of PTP1B during tumorigenesis? Approximately 30% of human cancers harbor activating mutations in the Ras gene (35). Because most of these mutations render Ras resistant to the actions of p120RasGAP, Src, and p62Dok, it is unlikely that loss of PTP1B would affect the Ras activity in this subset of cancer. However, loss of PTP1B does lead to increased IGF-I-induced AKT͞protein kinase B activity (9), and, in this case, inhibition of PTP1B could offer an increased survival advantage to the transformed cells. Breast cancer provides an interesting aspect in that rarely are Fig. 6. Model of impaired Ras signaling in PTP1B-deficient fibroblasts. Loss Ras mutations found (36). In fact, most breast cancer cases are of PTP1B leads to increased RTK phosphorylation and enhanced signaling of associated with increased expression of Src and members of the most downstream pathways. However, the absence of PTP1B can also lead to epidermal growth factor receptor family. Importantly, PTP1B cellular alterations that attenuate MAPK signaling downstream of RTKs. For has been identified as one of the major phosphatases that example, PTP1B deficiency can lead to decreased Src activation by way of activate Src in breast cancer cells (11). Furthermore, increased increased phosphorylation of inhibitory site. In addition, loss of PTP1B leads to expression of PTP1B has also been demonstrated in transformed increased expression of p120RasGAP by way of an unidentified mechanism. Finally, p62Dok, a potential substrate of PTP1B, is hyperphosphorylated in human breast cells (37) and ovarian carcinomas (38). This result PTP1B-deficient fibroblasts. All these events can contribute to attenuate Ras raises the intriguing possibility that PTP1B may positively con- activity and thus MAPK signaling. Consistent with this model, introduction of tribute to the progression of these cancers by way of activation activated Ras into PTP1B null cells can bypass these inhibitory events on Ras of Src. It will be interesting to determine the effects of intro- signaling and action. ducing the PTP1B null background into transgenic models of breast cancer (39). In conclusion, we have identified mechanisms by which deficient cells can bypass these inhibitory events on Ras signaling PTP1B deficiency can actually lead to impaired Ras signaling and action. and proliferation. Our results suggest that decreasing Ras activ- In addition to the plasma membrane, a recent study demon- ity through inhibition of PTP1B could even provide a means to strated that Ras activation and signaling can occur at the treat a subset of cancers. endoplasmic reticulum (ER) (32). PTP1B predominantly local- izes at the ER where it has been suggested to act on internalized We thank Drs. Morag Park, Pankaj Tailor, Christophe Blanchetot, and PDGFRs and epidermal growth factor receptors (33, 34). It will Feng Gu for helpful discussions. N.D. and A.C. are recipients of a be interesting to determine whether loss of PTP1B results in Canadian Institutes of Health Research doctoral research award and studentship, respectively. M.L.T. is a Canadian Institutes of Health impaired Ras signaling globally, or in specific subcellular com- Research Scientist. This work was supported by Cancer Research Society partments. Targeted expression of PTP1B within knockout cells and Canadian Institute of Health Research (MOP-62887) operating may provide further insight into this issue. grants to M.L.T.

1. Ostman, A. & Bohmer, F. D. (2001) Trends Cell Biol. 11, 258–266. 19. Donovan, S., Shannon, K. M. & Bollag, G. (2002) Biochim. Biophys. Acta 1602, 2. Cheng, A., Uetani, N., Simoncic, P. D., Chaubey, V. P., Lee-Loy, A., McGlade, 23–45. C. J., Kennedy, B. P. & Tremblay, M. L. (2002) Dev. Cell 2, 497–503. 20. Ellis, C., Moran, M., McCormick, F. & Pawson, T. (1990) Nature 343, 377–381. 3. Elchebly, M., Payette, P., Michaliszyn, E., Cromlish, W., Collins, S., Loy, A. L., 21. Flint, A. J., Tiganis, T., Barford, D. & Tonks, N. K. (1997) Proc. Natl. Acad. Normandin, D., Cheng, A., Himms-Hagen, J., Chan, C. C., et al. (1999) Science Sci. USA 94, 1680–1685. 283, 1544–1548. 22. Liu, F., Hill, D. E. & Chernoff, J. (1996) J. Biol. Chem. 271, 31290–31295. 4. Gu, F., Dube´, N., Kim, J. W., Cheng, A., Ibarra-Sanchez, M. d. J., Tremblay, 23. Zhao, M., Schmitz, A. A., Qin, Y., Di Cristofano, A., Pandolfi, P. P. & Van

M. L. & Boisclair, Y. R. (2003) Mol. Cell. Biol. 23, 3753–3762. Aelst, L. (2001) J. Exp. Med. 194, 265–274. BIOCHEMISTRY 5. Klaman, L. D., Boss, O., Peroni, O. D., Kim, J. K., Martino, J. L., Zabolotny, 24. Di Cristofano, A., Niki, M., Zhao, M., Karnell, F. G., Clarkson, B., Pear, W. S., J. M., Moghal, N., Lubkin, M., Kim, Y. B., Sharpe, A. H., et al. (2000) Mol. Cell. Van Aelst, L. & Pandolfi, P. P. (2001) J. Exp. Med. 194, 275–284. Biol. 20, 5479–5489. 25. Wick, M. J., Dong, L. Q., Hu, D., Langlais, P. & Liu, F. (2001) J. Biol. Chem. 6. Zabolotny, J. M., Bence-Hanulec, K. K., Stricker-Krongrad, A., Haj, F., Wang, 276, 42843–42850. Y., Minokoshi, Y., Kim, Y. B., Elmquist, J. K., Tartaglia, L. A., Kahn, B. B. & 26. Macaluso, M., Russo, G., Cinti, C., Bazan, V., Gebbia, N. & Russo, A. (2002) Neel, B. G. (2002) Dev. Cell 2, 489–495. J. Cell Physiol. 192, 125–130. 7. Johnson, T. O., Ermolieff, J. & Jirousek, M. R. (2002) Nat. Rev. Drug Discov. 27. Blume-Jensen, P. & Hunter, T. (2001) Nature 411, 355–365. 1, 696–709. 28. Lai, K. M. & Pawson, T. (2000) Genes Dev. 14, 1132–1145. 8. Haj, F. G., Markova, B., Klaman, L. D., Bohmer, F. D. & Neel, B. G. (2003) 29. Thomas, M., Suwa, T., Yang, L., Zhao, L., Hawks, C. L. & Hornsby, P. J. (2002) J. Biol. Chem. 278, 739–744. Neoplasia 4, 493–500. 9. Buckley, D. A., Cheng, A., Kiely, P. A., Tremblay, M. L. & O’Connor, R. (2002) 30. Beachy, T. M., Cole, S. L., Cavender, J. F. & Tevethia, M. J. (2002) J. Virol. 76, Mol. Cell. Biol. 22, 1998–2010. 3145–3157. 10. Arregui, C. O., Balsamo, J. & Lilien, J. (1998) J. Cell Biol. 143, 861–873. 31. Neet, K. & Hunter, T. (1995) Mol. Cell. Biol. 15, 4908–4920. 11. Bjorge, J. D., Pang, A. & Fujita, D. J. (2000) J. Biol. Chem. 275, 41439–41446. 32. Chiu, V. K., Bivona, T., Hach, A., Sajous, J. B., Silletti, J., Wiener, H., Johnson, 12. Cheng, A., Bal, G. S., Kennedy, B. P. & Tremblay, M. L. (2001) J. Biol. Chem. R. L., 2nd, Cox, A. D. & Philips, M. R. (2002) Nat. Cell Biol. 4, 343–350. 276, 25848–25855. 33. Haj, F. G., Verveer, P. J., Squire, A., Neel, B. G. & Bastiaens, P. I. (2002) 13. Broome, M. A. & Courtneidge, S. A. (2000) Oncogene 19, 2867–2869. Science 295, 1708–1711. 14. Klinghoffer, R. A., Sachsenmaier, C., Cooper, J. A. & Soriano, P. (1999) EMBO 34. Frangioni, J. V., Beahm, P. H., Shifrin, V., Jost, C. A. & Neel, B. G. (1992) Cell J. 18, 2459–2471. 68, 545–560. 15. Ibarra-Sanchez, M. J., Wagner, J., Ong, M. T., Lampron, C. & Tremblay, M. L. 35. Bos, J. L. (1989) Cancer Res. 49, 4682–4689. (2001) Oncogene 20, 4728–4739. 36. Andrechek, E. R. & Muller, W. J. (2000) Breast Cancer Res. 2, 211–216. 16. Simoncic, P. D., Lee-Loy, A., Barber, D. L., Tremblay, M. L. & McGlade, C. J. 37. Zhai, Y. F., Beittenmiller, H., Wang, B., Gould, M. N., Oakley, C., Esselman, (2002) Curr. Biol. 12, 446–453. W. J. & Welsch, C. W. (1993) Cancer Res. 53, 2272–2278. 17. Stice, L. L., Vaziri, C. & Faller, D. V. (1999) Front. Biosci. 4, D72–D86. 38. Wiener, J. R., Hurteau, J. A., Kerns, B. J., Whitaker, R. S., Conaway, M. R., 18. Raptis, L., Brownell, H. L., Corbley, M. J., Wood, K. W., Wang, D. & Haliotis, Berchuck, A. & Bast, R. C., Jr. (1994) Am. J. Obstet. Gynecol. 170, 1177–1183. T. (1997) Cell Growth Differ. 8, 891–901. 39. Hutchinson, J. N. & Muller, W. J. (2000) Oncogene 19, 6130–6137.

Dube´ et al. PNAS ͉ February 17, 2004 ͉ vol. 101 ͉ no. 7 ͉ 1839 Downloaded by guest on September 27, 2021