Protein Tyrosine Phosphatase PTPN13 Negatively Regulates Her2/Erbb2 Malignant Signaling

Oncogene (2008) 27, 2525–2531 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Protein tyrosine phosphatase PTPN13 negatively regulates Her2/ErbB2 malignant signaling

J-H Zhu1, R Chen1,WYi1, GT Cantin2, C Fearns1, Y Yang3, JR Yates III2 and J-D Lee1

1Department of Immunology, The Scripps Research Institute, La Jolla, CA, USA; 2Department of Cell Biology, The Scripps Research Institute, La Jolla, CA, USA and 3Johnson and Johnson Pharmaceutical Research and Development, San Diego, CA, USA

Deregulated Her2/ErbB2 receptor tyrosine kinase drives domain, consequently activating many mitogenic/onco- tumorigenesis and tumor progression in a variety of human genic intracellular signaling cascades such as PI(3)K/ tissues. Her2 transmits oncogenic signals through phos- AKT and MAP kinases, and inducing cellular processes phorylation of its cytosolic domain. To study innate cellular such as cell migration, invasion, differentiation and mechanisms for containing Her2 oncogenic phosphoryla- proliferation (Hackel et al., 1999; Gensler et al., 2004). tion, a siRNA phosphatase library was screened for cellular Overexpression of Her2 has been found in a variety of phosphatase(s) that enhance phosphorylation in the signal- solid tumors such as bladder, breast and ovarian cancers ing motif of Her2 after knockdown. We found that (Menard et al., 2004), and numerous reports from silencing protein tyrosine phosphatase PTPN13 signifi- in vitro and mouse models have demonstrated that cantly augmented growth factor-induced phosphorylation continued activation of Her2 has a causal role in of the Her2 signaling domain and promoted the invasive- tumorigenesis and leads to tumor progression (Slamon ness of Her2-deregulated cancer cells. In addition, we et al., 1987; Moody et al., 2002; Menard et al., 2004). discovered that growth factor-induced phosphorylation of Phosphorylation of tyrosine residues is usually a critical PTPN13 was essential for the dephosphorylation of Her2 step for regulating activities of certain cellular proteins and suggesting a negative feedback mechanism induced by subsequent activation/deactivation of their downstream growth factor to inhibit cellular Her2 activity through signaling events. This type of modification is generally PTPN13. Importantly, we showed that PTPN13 mutations controlled by the balance between the opposing activities previously reported in human tumors significantly reduced of protein kinases and tyrosine phosphatases (PTPs). In the phosphatase activity of PTPN13, and consequently recent years, much attention has been given to the elevated the oncogenic potential of Her2 and the invasive- regulation by PTPs and their physiological functions ness of Her2-overexpressing human cancer cells. Taken (Gil-Henn and Elson, 2003). PTPs are broadly divided together, these results suggest that cellular PTPN13 into single transmembrane receptor-like forms and non- inhibits Her2 activity by dephosphorylating the signal receptor forms. Recent evidences have shown that PTPs domain of Her2 and plays a role in attenuating invasiveness can function as tumor suppressors but some PTPs such as and metastasis of Her2 overactive tumors. SHP2 can be oncogenic (Ostman et al., 2006; Tonks, Oncogene (2008) 27, 2525–2531; doi:10.1038/sj.onc.1210922; 2006). Activating mutations in SHP2 have been identified published online 5 November 2007 in various tumors, particularly in leukemias (Tartaglia et al., 2003; Tartaglia and Gelb, 2005; Ostman et al., 2006). Keywords: PTPN13; Her2; dephosphorylation; invasion Through screening a phosphatase siRNA library, we and negative feedback found protein tyrosine phosphatase PTPN13 (also known as FAP-1, PTP-BAS and PTPL1) was involved in the regulation of phosphorylation of the Her2 signaling domain. PTPN13, a nonreceptor PTP, is the largest of known PTPs (Alonso et al., 2004). PTPN13 has been Introduction previously suggested to have tumor-suppressive activity (Wang et al., 2004; Ostman et al., 2006; Yeh et al., 2006) Her2/ErbB2 together with Her1/ErbB1, Her3/ErbB3 and mutations in PTPN13 have been reported in colon and Her4/ErbB4 are members of epidermal growth and hepatocellular cancers (Wang et al., 2004; Yeh et al., factor (EGF) receptor family of receptor tyrosine 2006). In this study, we demonstrated a role of PTPN13 in kinases. Activation of Her2 triggers autophosphoryla- regulating Her2 activity. tion of specific tyrosine residues within its cytoplasmic Results Correspondence: Professor J-D Lee, Department of Immunology, The Scripps Research Institute, IMM-12, 10550 N. Torrey Pines Road., La PTPN13 knockdown by siRNA increases the Jolla, CA 92037, USA. E-mail: [email protected] phosphorylation of Her2 signaling domain Received 19 June 2007; revised 8October 2007; accepted 11 October A phosphatase siRNA library was used to screen for 2007; published online 5 November 2007 phosphatase(s) that affected Her2 phosphorylation. PTPN13 negatively regulates Her2 activity J-H Zhu et al 2526 Among 193 phosphatases (40 protein tyrosine phospha- expression of wild-type PTPN13 inhibited Her2 phos- tases) screened by siRNA knockdown, knockdown of 7 phorylation after EGF stimulation, while the mutant protein tyrosine phosphatases (PTPs) significantly in- PTPN13D2359A promoted it (Figure 2b). This modulation creased phospho-Her2 levels after growth factor stimu- of the phosphorylation of Her2 raised the question lation (Table 1). We further confirmed these data by whether PTPN13 can form a complex with Her2. To increasing expression of these PTPs by transfection of address this question, an in vitro binding experiment was expression plasmids and found that only increased performed using either wild-type PTPN13 or mutant expression of PTPN13 and PTPN14 can reduce the Her2 phosphorylation induced by growth factor (No data is available about increased expression of PTPRD a NT Mock siBMK1 siPTPN13 since expression plasmid of this gene is toxic to cells). We further analysed the phosphatase activity of both P-Her2 PTPN13 and PTPN14 and detected phosphatase activity only in PTPN13 but not in PTPN14 (Table 1). Thus, Her2 we focused our effort on PTPN13. PTPN13 knockdown by siRNA significantly in- BMK1 creased phosphorylation of the Her2 signaling domain after growth factor stimulation (Figure 1a and Supple- PTPN13 mentary Figure S1B). As controls, mock transfection and siRNA-mediated big mitogen-activated kinase 1 GADPH (BMK1) knockdown had no effect on Her2 phosphorylation. The effect of PTPN13 knockdown on Her2 b Mock siPTPN13 phosphorylation was validated by two individual 0 5 15 30 60 0 5 15 30 60 (min) siRNAs against PTPN13 (Supplementary Figure S2). Time course experiments showed that EGF-induced P-Her2 Her2 phosphorylation was enhanced by PTPN13 knockdown at all time points tested (Figure 1b), but Her2 the augmentation of Her2 phosphorylation was more significant at the earlier time points (5 and 15 min) P-Akt indicating that PTPN13 has a considerable role in

inhibiting the early (5 and 15 min) rather than late P-Stat3 (60 min) growth factor-induced Her2 signaling. Accord- ingly, activation of downstream Akt and Stat3 was elevated by PTPN13 knockdown, particularly at the P-Erk1/2 earlier time points (5 and 15 min). A slight increase in Erk1/2 activation at the earlier times was also observed GADPH in the PTPN13-deﬁcient cells (Figure 1b). Figure 1 PTPN13 knockdown by siRNA increases the phosphorylation of Her2 signaling domain. (a) HeLa cells were transfected Increased expression of cellular PTPN13 inhibits Her2 with siRNA against BMK1 (siBMK1) or against PTPN13 phosphorylation, and Her2 forms a stable complex with a (siPTPN13). Mock-transfected or siBMK1-transfected HeLa cells phosphatase inactive form of PTPN13 were used as controls. Forty-eight hours post transfection, the cells À1 We constructed a catalytically inactive mutant of PTPN13 were serum starved overnight followed by 10 ng ml EGF stimulation for 15 min. The levels of phospho-Her2 (P-Her2), by mutating aspartic acid 2359 in the phosphatase domain Her2, BMK1, PTPN13 and GADPH in these cells were analysed to alanine. As predicted, PTPN13D2359A had little phos- by western blot using speciﬁc antibodies (described in Materials phatase activity compared with the wild-type PTPN13 and methods) as indicated. (b) HeLa cells were mock transfected or transfected with siPTPN13 and starved as in (a). These cells were (Figure 2a). We next tested whether expression of wild- À1 D2359A stimulated with 10 ng ml EGF for 0, 5, 15, 30 and 60 min as type PTPN13 or inactive PTPN13 had any effect on indicated and the levels of P-Her2, Her2, phospho-Akt, phospho- Her2 phosphorylation. Compared to the mock-trans- Stat3, phospho-Erk1/2 and GADPH in these cells were examined fected and empty vector-transfected controls, increased by western blot.

Table 1 Effect of seven protein tyrosine phosphatases on Her2 phosphorylation PTPN3 PTPN13 PTPN14 PTPRB PTPRD PTPRE PTPRN2

siRNA knockdown mm mmmm m Overexpression À kkÀ NA ÀÀ Phosphatase activity + + ND + NA + ND

Abbreviations: NA, data not available due to toxic plasmid; ND, activity not detectable. HeLa cells were stimulated 15 min with 10 ng mlÀ1 EGF. Phosphatase activity was assayed as described in Materials and methods. m, increased Her2 phosphorylation compared to control; k, decreased Her2 phosphorylation compared to control; À, no change of Her2 phosphorylation compared to control; +, activity detectable.

Oncogene PTPN13 negatively regulates Her2 activity J-H Zhu et al 2527 PTPN13D2359A. Indeed, by immunoprecipitation, we phosphorylation (Figure 3b). Following EGF stimula- showed that catalytically inactive PTPN13D2359A,formed tion, expression of PTPN13T1336A or PTPN13T1336A/S1339A a much more stable complex with Her2 protein than wild- led to elevated Her2 phosphorylation while expression type PTPN13 did (Figure 2c), suggesting that the inactive of PTPN13S1339A has an effect similar to that of wild-type form of PTPN13 not only loses its phosphatase activity PTPN13 (Figure 3c). Phosphatase activity assay showed but also may protect the dephosphorylation of the Her2 that PTPN13T1336A and PTPN13T1336A/S1339A retained only signaling domain by stably interacting with Her2 thereby around 35 and 25% of its wild-type activity, respec- preventing the access of other cellular phosphatase(s) to tively, while PTPN13S1339A retained about 80% activity the Her2 signaling domain. (Figure 3d). These data indicate that growth factor- induced phosphorylation of PTPN13 is important for its The growth factor-mediated phosphorylation of PTPN13 phosphatase activity and for deactivating Her2. is involved in the dephosphorylation of Her2 by PTPN13 Through functional phosphoproteomic analysis, we PTPN13 mutations in human cancer promote Her2 proﬁled phosphorylation state of cellular proteins before phosphorylation and after growth factor stimulation (manuscript in Mutations in PTPN13 have been reported in clinical human preparation). We discovered that EGF induced the tumors (Wang et al., 2004). To examine their impact on phosphorylation of cellular PTPN13 at both threonine Her2 phosphorylation in cancer cells, we generated expres- 1336 and serine 1339 (Figure 3a). These two sites in sion plasmids for two of these cancer mutants, PTPN13 were then individually or together mutated to PTPN13E2455D and PTPN13Y2260X (Figure 4a). Expression alanine (PTPN13T1336A, PTPN13S1339A and PTPN13T1336A/ of either PTPN13E2455D or PTPN13Y2260X led to increased S1339A) to prevent their phosphorylation after growth phosphorylation of cellular Her2 after growth factor factor stimulation and to study whether phosphoryla- induction (Figure 4b). Phosphatase activity assays showed tion of these two sites of PTPN13 has any effect on Her2 that PTPN13E2455D retained around 50% of wild-type a b a Mock EV WT DA 100 P-Her2 75 50 Her2 K.T*FSSSPPKPGDIFEVELAK.N K.TFSS*SPPKPGDIFEVELAK.N 25 activity (%) 0 Flag b T1336A S1339A Relative phosphatase EV WT DA KIND FERM PDZ1 PDZ2 PDZ3 PDZ4 PDZ5 PTP domain Flag GADPH c d Mock EV WT TA SA TSA c EV MDC1 WT DA 100 P-Her2 Her2 75 50 Her2

Flag activity (%) 25 Flag 0 Relative phosphatase Figure 2 Increased expression of cellular PTPN13 inhibits Her2 EV WT TA SA TSA phosphorylation, and Her2 forms a stable complex with a GAPDH Flag phosphatase inactive form of PTPN13. (a) HEK293 cells were transfected with empty vector (EV), expression plasmids encoding Figure 3 The growth factor-mediated phosphorylation of Flag-tagged wild-type PTPN13 (WT) or its catalytically inactive PTPN13 is involved in the dephosphorylation of Her2 by PTPN13. mutant PTPN13D2359A (DA). Forty-eight hours post transfection, (a) The mass spectrometry diagram of phosphorylation of cells were lysed and the phosphatase activities of WT and DA were PTPN13T1336 and PTPN13S1339 after EGF stimulation. The lower assessed as described in Materials and methods. Relative phos- (indicated by —4) and the upper lines (indicated by -) in the phatase activity in these cells was normalized using WT transfected graphs show the phosphopeptides (sequences indicated at the cells whose value was taken as 100%. (b) HEK293 cells were bottom of graphs) detected by mass spectrometry before and after transfected with expression plasmids encoding WT or DA. Mock- EGF stimulation, respectively. (b) The linear diagram of function transfected cells and EV transfected cells were used as controls. domains and the positions of T1336A and S1339A mutation sites in Forty-eight hours post transfection, cells were serum starved PTPN13. (c) HEK293 cells were transfected with expression overnight. The cells were then treated with EGF (10 ng mlÀ1) for plasmids encoding wild-type PTPN13 (WT), PTPN13T1336A (TA), 10 min and the levels of phospho-Her2 (P-Her2), Her2, the PTPN13S1339A (SA) or PTPN13T1336A/S1339A (TSA). Forty-eight hours expression of Flag-tagged proteins and GAPDH in these cells post transfection, the cells were serum starved and then treated were analysed by western blot. (c) HEK293 cells were cotransfected with EGF (10 ng mlÀ1) for 10 min before harvesting. The levels of with an expression plasmid encoding Her2 along with plasmid phospho-Her2 (P-Her2), Her2, the expression of Flag-tagged encoding WT, DA or Flag-tagged MDC1 (mediator of DNA proteins and GAPDH in these cells were examined by western damage checkpoint 1) as a control. Forty-eight hours post blot. (d) HEK293 cells were transfected with empty vector (EV) or transfection, cells were serum starved overnight followed by EGF expression plamids encoding WT, TA, SA or TSA. 48h post (10 ng mlÀ1) stimulation for 10 min. Cell lysates were then prepared transfection, cells were collected and phosphatase activities of EV, and immunoprecipitation were performed with anti-Flag antibody. WT, TA, SA and TSA were assessed as described in Materials and The protein levels of Her2 in these immunocomplexes were methods. Relative phosphatase activity in these cells was normal- detected by western blot. ized using WT transfected cells whose value was taken as 100%.

Oncogene PTPN13 negatively regulates Her2 activity J-H Zhu et al 2528 PTPN13 activity and PTPN13Y2260X lost almost all its knockdown SKOV3 cells (SKOV3-shN13), Her2 phos- phosphatase activity (Figure 4c). We also found that cancer phorylation induced by growth factor was increased mutant PTPN13Y2260X, but not PTPN13E2455D, stably asso- compared to parental SKOV3 and SKOV3 infected with ciated with cellular Her2 protein (Figure 4d). These results virus carrying empty vector (SKOV3-shEV) (Figure 5a). suggested that the dysfunctional PTPN13 cancer mutants There was no difference in cell proliferation and migration can upregulate Her2 malignant signaling by decreasing their rate between SKOV3-shN13 and SKOV3-shEV cells (data phosphatase activity and/or by stably interacting with the not shown). However, the invasive potential of SKOV3- Her2 signaling domain and consequently blocking the shN13 was increased by about two fold compared with that accessibility of other functional cellular phosphatases to of parental SKOV3 and SKOV3-shEV (Figures 5b and c). Her2. We further tested whether PTPN13 mutations in human cancer have a similar effect in promoting tumor cell invasiveness. We found that SKOV3 cells expressing PTPN13 knockdown or PTPN13 cancer mutants promote PTPN13 cancer mutants, Y2260X or E2455D, promoted the invasiveness of Her2-overexpressing cancer cells the SKOV3 invasion by about 50 and 25% compared to Her2 overexpression is detected in about 30% of human mock- or empty vector-transfected SKOV3 cells breast and ovarian cancers and is a marker of poor (Figure 5d). The increase of invasion in SKOV3-shN13 prognosis for cancer patients (Hayes and Thor, 2002). As cells is mostly derived from upregulated Her2 signaling PTPN13 inhibited Her2 signaling, we tested whether since Her2 knockdown in SKOV3-shN13 cells by siRNA PTPN13 has a role in regulating malignant properties of Her2-overexpressing tumor cells. PTPN13 expression was silenced in a Her2-overexpressing ovarian carcinoma cell line, SKOV3, by infecting it with recombinant lentivirus a Parental shEV shPTPN13 b carrying shRNA against PTPN13 (shPTPN13). In PTPN13 P-Her2 2.5 2 Her2 1.5

Y2260X E2455D 1 a PTPN13 0.5 KIND FERM PDZ1 PDZ2 PDZ3 PDZ4 PDZ5 PTP domain

Relative cell invasion 0 GADPH Parental shEV shPTPN13 b Mock EV WT YX ED c

P-Her2 100 c Parental d 75 2 Her2 50 1.5

activity (%) 25 Flag 1 0 Relative phosphatase 0.5 GAPDH EV WT YX ED shEV Flag 0 Relative cell invasion Mock EV YX ED d EV WT YX ED Flag Her2 shPTPN13 P-Her2

Flag Her2 Figure 4 PTPN13 mutations in human cancer promote Her2 phosphorylation. (a) A linear diagram showing the locations of cancer mutations Y2260X (X, stop codon) and E2455D in Figure 5 PTPN13 knockdown or PTPN13 cancer mutants PTPN13. (b) HEK293 cells were transfected with empty vector promote the invasiveness of Her2-overexpressing cancer cells. (a) (EV) or expression plasmids encoding wild-type PTPN13 (WT), Stable SKOV3 cell line carrying empty vector (shEV) or vector PTPN13E2455D (ED), or PTPN13Y2260X (YX). Forty-eight hours post expressing shRNA against PTPN13 (shPTPN13) was stimulated by transfection, these cells were serum starved overnight. Then the 10 ng mlÀ1 EGF for 15 min. Parental cell line or cell line carrying cells were stimulated with EGF (10 ng mlÀ1) for 10 min and the levels shEV were used as controls. These cells were collected and the of phospho-Her2 (P-Her2), Her2, Flag-tagged proteins and protein levels of phospho-Her2 (P-Her2), Her2, PTPN13 and GAPDH in these cells were examined by western blot. (c) GAPDH were detected by western blot as indicated. (b) Invasive HEK293 cells were transfected with EV or expression plasmids potential of parental, shEV and shPTPN13 cells were analysed as encoding WT, ED or YX. Forty-eight hours post transfection, described in Materials and methods. The relative cell invasion in phosphatase activity was assessed as described in Materials and these cells was normalized using parental cells whose value was methods. Relative phosphatase activity in these cells was normal- taken as 1. (c) Representative images of invaded cells of parental, ized using WT transfected cells whose value was taken as 100%. shEV and shPTPN13 as described in (b). (d) SKOV3 cells were (d) HEK293 cells were cotransfected with expression plasmids transfected with empty vector (EV) or with expression plasmids encoding Her2 along with plasmids encoding WT, ED or YX. encoding PTPN13E2455D (ED) or PTPN13Y2260X (YX). Forty-eight Forty-eight hours post transfection, these cells were serum starved hours post transfection, the invasive capacity of the cells was overnight. Subsequently, the cells were stimulated by EGF analysed and the relative cell invasion in these cells was normalized (10 ng mlÀ1) for 10 min, cell lysates were prepared and used for using mock-transfected cells whose value was taken as 1. The level immunoprecipitation with anti-Flag antibody. The protein levels of of P-Her2 and Her2 was detected after 48h transfection and Her2 in the immunocomplexes were detected by western blot as overnight starvation followed by stimulation with 10 ng mlÀ1 EGF indicated. for 15 min.

Oncogene PTPN13 negatively regulates Her2 activity J-H Zhu et al 2529 reduced their invasive potential to the level of control 2007; Julien et al., 2007). Herein, we described a third SKOV3-shEV cells (Figure 6). We also tested the effect of phosphatase, PTPN13, which blocked Her2 activity by PTPN13 knockdown on Her2-overexpressing breast cancer dephosphorylating the Her2 cytoplasmic domain cell line BT474. Similar to HeLa and SKOV3, PTPN13 (Figure 7). We found knockdown of PTPN13 by siRNA knockdown lead to increased Her2 phosphorylation and significantly increased Her2 phosphorylation, while in- cell invasiveness of BT474 (Supplementary Figure S3). As creased expression of PTPN13 inhibited Her2 activity. tumor cell invasiveness is directly linked to tumor Similar to BDP1 and PTP1B, PTPN13 also belongs to the metastasis, these data imply that PTPN13 may have a role family of nonreceptor phosphatases. However, besides the in inhibiting the invasion and metastasis of Her2-over- phosphatase catalytic domain, PTPN13 also contains one expressing tumor cells (Figure 7). FERM domain and five PDZ domains, which may contribute to protein or substrate binding specificity (Erdmann, 2003; Tonks, 2006). Intriguingly, we found that growth factor induced the phosphorylation of Discussion PTPN13 at two sites, which are important for phosphatase activity of PTPN13 and consequent inhibition of Her2 Overactive Her2 is found in nearly 30% of mammary and activity. These results indicate the existence of a growth ovarian cancers and is associated with a poor clinical factor-induced negative feedback mechanism of Her2 prognosis (Hayes and Thor, 2002). Consequently, since signaling through phosphatase PTPN13 (Figure 7). Her2 is an important molecular target for cancer therapy Both decreased expression and mutations of PTPN13 (Kim et al., 2004), understanding the regulatory mechan- have been observed in human tumors (Inazawa et al., isms that control the malignant activity of Her2 in cells 1996; Wang et al., 2004; Yeh et al., 2006). It has been should help in devising new and more effective approaches reported that PTPN13 is located on chromosome to treat Her2-dependent cancers. Her2 oncogenic signaling 4q21.3, a region frequently deleted in liver and ovarian is through kinase–mediated phosphorylation of its cyto- cancers (Inazawa et al., 1996). Lower expression levels solic signaling domain. Thus, dephosphorylation of the of PTPN13 contribute to higher proliferation rate of Her2 signaling domain by phosphatase should confine the certain cancer cell lines (Yeh et al., 2006), while no magnitude and duration of Her2 oncogenic signaling after biochemical or cellular phenotype has been attributed to its activation in cells. To date, two nonreceptor phospha- the mutations of PTPN13 found in human tumors tases (BDP1/PTPN18(Gensler et al., 2004) and PTP1B/ (Wang et al., 2004). Herein, we demonstrated that PTPN1 (Bentires-Alj and Neel, 2007; Julien et al., 2007)) expression of two of the cancer mutants of PTPN13, have been shown to be involved or implicated in regulation PTPN13E2455D and PTPN13Y2260X, individually led to of Her2 activity and its downstream signaling. BDP1/ enhancement of Her2 activity. In addition, similar to the PTPN18inhibits Her2 signaling by inhibiting Her2 catalytically inactive mutant PTPN13D2359A described in phosphorylation (Gensler et al., 2004). Conversely, PTP1B Figure 2, the cancer mutant with little phosphatase positively regulates Her2 signaling since PTP1B deficiency activity, PTPN13Y2260X was also able to bind Her2 postponed Her2-induced mammary tumorigenesis and protein with much higher persistent affinity than wild- protected from lung metastasis (Bentires-Alj and Neel, type PTPN13. This suggests that nonfunctional PTPN13 mutation(s) in cancer may promote Her2- dependent malignancy not only by their inactivity but 2 also by binding to Her2 and preventing the entry of 1.6 other phosphatase(s) or molecules that downmodulate

1.2

0.8 Growth Factor Growth Factor 0.4 Relative cell invasion 0 shEV shEV shN13 shN13 Her1 Her1 + siHer2 + siHer2 Her3 Her2 Her3 Her2 Her4 Her4 Her2 P P

PTPN13 PTPN13 P Cancer P GAPDH P P Mutation PTPN13 PTPN13 Y2260X Figure 6 Increased invasion caused by PTPN13 knockdown is primarily due to Her2 signaling. Stable SKOV3 cell line carrying empty vector (shEV) or vector expressing shRNA against PTPN13 Invasion (shN13) was mock-transfected or transfected with siRNA against Invasion Her2 (siHer2) as indicated. After 3 days, the invasive potential of mock or Her2 siRNA transfected cells were analysed as described Metastasis Metastasis in Materials and methods. The relative cell invasion in these cells was normalized using mock-transfected shEV cells whose value was Figure 7 Schematic illustration of mechanisms of action of taken as 1. PTPN13 on Her2 malignant activity.

Oncogene PTPN13 negatively regulates Her2 activity J-H Zhu et al 2530 the activity of Her2. Importantly, expression of these Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS), two cancer mutants or knockdown of cellular PTPN13 1% penicillin/streptomycin and 1% L-glutamine (Invitrogen, increased cell invasiveness of Her2-overactive tumor Carlsbad, CA, USA). cells. As cell invasiveness is closely connected to the metastatic potential of tumor, these data suggest that siRNA library screen PTPN13 negatively regulates Her2 activity and conse- The phosphatase siRNA library and transfection reagent were quently may inhibit metastasis of the Her2-overexpres- purchased from Dharmacon (Lafayette, CO, USA). The sing tumor (Figure 7). Therefore, cellular PTPN13 likely siRNA against each gene was a pool composed of four individual siRNA duplexes. siRNA transfection into HeLa possesses tumor suppressive activity, which has also cells was conducted following the manufacturer’s protocol. been suggested by reports from other labs (Bompard Cells were transfected with siRNA for 2 days and then serum et al., 2002; Ostman et al., 2006; Yeh et al., 2006). starved for 1 day before harvest. After stimulated for 15 min Her2/ErbB2 belongs to the family of ErbB/EGF with 10 ng mlÀ1 EGF, cells were lysed in RIPA buffer (9.1 mM receptors, which includes three other members, Her1/ dibasic sodium phosphate, 1.7 mM monobasic sodium phos- EGFR/ErbB1, Her3/ErbB3 and Her4/ErbB4. Upon phate, 150 mM NaCl, pH 7.4, 1% Triton X-100, 0.5% sodium growth factor stimulation, these receptors homodimer- deoxycholate, 0.1% SDS, 100 mgmlÀ1 PMSF and 1 mM sodium ize or heterodimerize leading to autophosphorylation orthovanadate). Western blot was performed to screen for and subsequent downstream effector activation (Engel- changes in Her2 phosphorylation. man and Cantley, 2006). Due to the similar architecture among the ErbB receptors, a single phosphatase may be DNA constructs and transfection able to modulate more than one of these receptors. The lentiviral shRNA plasmid against PTPN13 was purchased from Open Biosystems (Huntsville, AL, USA). The SKOV3 cell However, to date, no phosphatase has been shown to lines carrying PTPN13 shRNA or the empty vector were directly dephosphorylate both Her1 and Her2. PTP1B established according to the manufacturer’s protocol. The Flag- was reported to regulate Her1 signaling since cells tagged PTPN13 expression plasmid was a generous gift from Dr deficient in PTP1B exhibited increased and sustained Carl-Henrik Heldin (Ludwig Institute for Cancer Research, phosphorylation of Her1 after growth factor treatment Uppsala University, Sweden). PTPN13 mutations were con- (Haj et al., 2003). This phosphatase was also involved in structed using Quikchange II XL site-directed mutagenesis kit regulation of Her2 signaling, likely by functioning on (Stratagene, La Jolla, CA, USA). The Flag-tagged MDC1 downstream pathways (Bentires-Alj and Neel, 2007; (mediator of DNA damage checkpoint 1) expression plasmid Julien et al., 2007). BDP1/PTPN18can only regulate was provided by Dr Maoyi Lai (The Scripps Research Institute, one ErbB receptor as increased expression of this La Jolla, CA, USA). HEK293 cells were transfected using a calcium phosphate method modified from the protocol provided phosphatase-inhibited growth factor induced activation by Open Biosystems. On the other hand, SKOV3 cells were of Her2 but not that of closely related Her1 (Gensler transfected using GenJet DNA in vitro transfection reagent et al., 2004). Interestingly, similar to BDP1, PTPN13 (SignaGen Laboratories, Gaithersburg, MD, USA) following knockdown led to minimal, if any, change of Her1 the manufacturer’s recommended protocol. phosphorylation level after growth factor treatment (Supplementary Figure S1A). Antibodies, western blot and immunoprecipitation In summary, our data showed that PTPN13 nega- Anti-Flag antibody was purchased from Stratagene. Anti- tively regulated Her2 activity and inhibited invasiveness phospho-Her2 (Tyr1248), anti-Her2, anti-phospho-Akt of Her2-overexpressing tumor cells. In addition, we (Ser473) and anti-Stat3 (Tyr705) antibodies were from Upstate discovered that growth factor induced the phosphoryla- (Lake Placid, NY, USA). Anti-phospho-Erk1/2 (Thr202/ tion of PTPN13, which was important for the ability of Tyr204) antibody was from Cell Signaling (Boston, MA, PTPN13 to inhibit the Her2 malignant signaling USA). Anti-GAPDH antibody was from Calbiochem (San indicating a growth factor-induced negative feedback Diego, CA, USA). Anti-BMK1 antibody was previously described (Kato et al., 1997). Anti-PTPN13 antibody, anti- mechanism of Her2 activity through PTPN13. PTPN13 rabbit and anti-mouse secondary antibodies were purchased mutations in human tumors escalated Her2 activity and from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). the invasive potential of Her2-deregulated cancer cells For western blot, cells were lysed in RIPA buffer as described suggesting that cellular PTPN13 has a role in containing above. For immunoprecipitation, cells were lysed in 50 mM the Her2 oncogenic signaling and inhibiting the invasive Hepes, 150 mM NaCl, 1 mM EDTA, 10% glycerin, 1% Triton and metastatic potential of tumors bearing overactive X-100, 100 mgmlÀ1 PMSF and protease inhibitor cocktail Her2. Finally, the identification of PTPN13 as a direct (Roche, Indianapolis, IN, USA). Samples were loaded onto 6 regulator of Her2 activity provides an important clue for or 8% sodium dodecyl sulfate–polyacrylamide gel electro- designing novel approaches to slow down or stop tumor phoresis gel and examined by western blot. invasiveness and metastasis of Her2-dependent cancers. Phosphoproteomic analysis Phosphoproteomic analysis of the EGF pathway was carried out as detailed elsewhere (Cantin GT, Yi W, Lu B, Park SK, Materials and methods Xu T, Lee JD and Yates JR; manuscript in preparation). Briefly, HeLa cells were grown separately in heavy and light Cell culture SILAC D-MEM media (Invitrogen), and one culture was The human cell lines HeLa, HEK293 and SKOV3 were treated with EGF. Protein lysates were prepared from both obtained from American Type Culture Collection (Manassas, cultures, mixed 1:1 and subjected to two rounds of immobilized VA, USA). All cells were cultured in Dulbecco’s modified metal affinity chromatography (IMAC). Peptides from the final

Oncogene PTPN13 negatively regulates Her2 activity J-H Zhu et al 2531 IMAC eluate were subjected to MudPIT (Washburn et al., Phosphatase activity assay 2001) and relative quantification analysis (Cantin et al., 2006). The method of phosphatase activity assay was modified from Xu et al. (2005) and the protocol of Upstate. Briefly, Flag-tagged Invasion assay protein was immunoprecipitated using anti-Flag antibody. À1 Cell invasion was analysed as described by Hauck et al. (2002) Bovine serum albumin (5 ml, 5mgml ) was added to the with slight modification. Briefly, Matrigel (Sigma, Saint Louis, immunocomplex and total volume was brought to 80 mlwith MI, USA) was thawed at 4 1C overnight, and diluted to around phosphatase assay buffer (25 mM HEPES, pH 7.2, 50 mM NaCl, 3mgmlÀ1 with cold DMEM. Millicell plate inserts (12 mm 5mM dithiothreitol, 2.5 mM EDTA). Thereafter, 120 mlof À1 diameter with 8.0 mM pores, Millipore, Billerica, MA, USA) p-nitrophenyl phosphate (4 mg ml ) freshly prepared in the assay were coated with 100 ml diluted Matrigel and incubated at buffer was added into the mixture to start the reaction. After 37 1C 4–5 h for gelling. Cells were harvested by Trypsin/ 15 min at 37 1C, the reaction was stopped by adding 20 ml of 13% EDTA, washed twice with DMEM, and resuspended in K2HPO4.Then60ml of the mixture was transferred to a 96-well DMEM at a density of 106 per ml. Lower chambers were plate, and hydrolysis of p-nitrophenyl phosphate was measured filled with 600 ml warm DMEM containing 10% FBS as the spectrometrically at 405 nm. Activity was expressed as a relative attractant, and the Matrigel-coated upper chambers were filled value of wild-type PTPN13 activity, which was set to 100%. with 100 ml of the cell suspension. After 24 h incubation at Mean values were obtained from three individual experiments. 37 1C, cells were fixed for 30 min in 5% glutaraldehyde prepared in phosphate-buffer saline at room temperature, Acknowledgements and stained 30 min in 0.1% crystal violet and 2% ethanol. The numbers of invasive cells were measured by counting three This work was supported by funds from National Cancer Institute fields per chamber. Mean values were obtained from at least (CA079871 and CA114059), California Tobacco-Related Disease three separate experiments. (15RT-0104) and Department of Defense BCRP (BC031105).

References

Alonso A, Sasin J, Bottini N, Friedberg I, Friedberg I, Osterman A Kato Y, Kravchenko VV, Tapping RI, Han J, Ulevitch RJ, Lee JD. et al. (2004). Protein tyrosine phosphatases in the human genome. (1997). BMK1/ERK5 regulates serum-induced early gene expression Cell 117: 699–711. through transcription factor MEF2C. EMBO J 16: 7054–7066. Bentires-Alj M, Neel BG. (2007). Protein-tyrosine phosphatase 1B is required Kim SW, Chao TH, Xiang R, Lo JF, Campbell MJ, Fearns C et al. (2004). for HER2/Neu-induced breast cancer. Cancer Res 67: 2420–2424. Tid1, the human homologue of a Drosophila tumor suppressor, reduces Bompard G, Puech C, Prebois C, Vignon F, Freiss G. (2002). Protein- the malignant activity of ErbB-2 in carcinoma cells. Cancer Res 64: tyrosine phosphatase PTPL1/FAP-1 triggers apoptosis in human 7732–7739. breast cancer cells. J Biol Chem 277: 47861–47869. Menard S, Casalini P, Campiglio M, Pupa SM, Tagliabue E. (2004). Cantin GT, Venable JD, Cociorva D, Yates III JR. (2006). Role of HER2/neu in tumor progression and therapy. Cell Mol Life Quantitative phosphoproteomic analysis of the tumor necrosis Sci 61: 2965–2978. factor pathway. J Proteome Res 5: 127–134. Moody SE, Sarkisian CJ, Hahn KT, Gunther EJ, Pickup S, Dugan Engelman JA, Cantley LC. (2006). The role of the ErbB family KD et al. (2002). Conditional activation of Neu in the mammary members in non-small cell lung cancers sensitive to epidermal growth epithelium of transgenic mice results in reversible pulmonary factor receptor kinase inhibitors. Clin Cancer Res 12: 4372s–4376s. metastasis. Cancer Cell 2: 451–461. Erdmann KS. (2003). The protein tyrosine phosphatase PTP-Basophil/ Ostman A, Hellberg C, Bohmer FD. (2006). Protein-tyrosine Basophil-like. Interacting proteins and molecular functions. Eur J phosphatases and cancer. Nat Rev Cancer 6: 307–320. Biochem 270: 4789–4798. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Gensler M, Buschbeck M, Ullrich A. (2004). Negative regulation of (1987). Human breast cancer: correlation of relapse and survival with HER2 signaling by the PEST-type protein-tyrosine phosphatase amplification of the HER-2/neu oncogene. Science 235: 177–182. BDP1. J Biol Chem 279: 12110–12116. Tartaglia M, Gelb BD. (2005). Germ-line and somatic PTPN11 Gil-Henn H, Elson A. (2003). Tyrosine phosphatase-epsilon activates mutations in human disease. Eur J Med Genet 48: 81–96. Src and supports the transformed phenotype of Neu-induced Tartaglia M, Niemeyer CM, Fragale A, Song X, Buechner J, Jung A mammary tumor cells. J Biol Chem 278: 15579–15586. et al. (2003). Somatic mutations in PTPN11 in juvenile myelomo- Hackel PO, Zwick E, Prenzel N, Ullrich A. (1999). Epidermal growth nocytic leukemia, myelodysplastic syndromes and acute myeloid factor receptors: critical mediators of multiple receptor pathways. leukemia. Nat Genet 34: 148–150. Curr Opin Cell Biol 11: 184–189. Tonks NK. (2006). Protein tyrosine phosphatases: from genes, to Haj FG, Markova B, Klaman LD, Bohmer FD, Neel BG. (2003). function, to disease. Nat Rev Mol Cell Biol 7: 833–846. Regulation of receptor tyrosine kinase signaling by protein tyrosine Wang Z, Shen D, Parsons DW, Bardelli A, Sager J, Szabo S et al. phosphatase-1B. J Biol Chem 278: 739–744. (2004). Mutational analysis of the tyrosine phosphatome in color- Hauck CR, Hsia DA, Puente XS, Cheresh DA, Schlaepfer DD. (2002). ectal cancers. Science 304: 1164–1166. FRNK blocks v-Src-stimulated invasion and experimental metastases Washburn MP, Wolters D, Yates III JR. (2001). Large-scale analysis without effects on cell motility or growth. EMBO J 21: 6289–6302. of the yeast proteome by multidimensional protein identification Hayes DF, Thor AD. (2002). c-erbB-2 in breast cancer: development technology. Nat Biotechnol 19: 242–247. of a clinically useful marker. Semin Oncol 29: 231–245. Xu Y, Tan LJ, Grachtchouk V, Voorhees JJ, Fisher GJ. (2005). Inazawa J, Ariyama T, Abe T, Druck T, Ohta M, Huebner K et al. (1996). Receptor-type protein-tyrosine phosphatase-kappa regulates PTPN13, a fas-associated protein tyrosine phosphatase, is located on the epidermal growth factor receptor function. J Biol Chem 280: long arm of chromosome 4 at band q21.3. Genomics 31: 240–242. 42694–42700. Julien SG, Dube N, Read M, Penney J, Paquet M, Han Y et al. (2007). Yeh SH, Wu DC, Tsai CY, Kuo TJ, Yu WC, Chang YS et al. (2006). Protein tyrosine phosphatase 1B deficiency or inhibition delays Genetic characterization of fas-associated phosphatase-1 as a ErbB2-induced mammary tumorigenesis and protects from lung putative tumor suppressor gene on chromosome 4q21.3 in hepato- metastasis. Nat Genet 39: 338–346. cellular carcinoma. Clin Cancer Res 12: 1097–1108.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene