ORIGINAL ARTICLE

PTPN12 promotes resistance to oxidative stress and supportstumorigenesis by regulating FOXO signalingIS Harris1,2, H Blaser1, J Moreno1,3,4, AE Treloar1, C Gorrini1, M Sasaki1, JM Mason1, CB Knobbe1,5, A Rufini6, M Halle7,AJ Elia1, A Wakeham1, ML Tremblay7, G Melino6, S Done1,2,3,4 and TW Mak1,2

It is well known that protein tyrosine phosphatases (PTPs) that become oxidized due to exposure to reactive oxygen species (ROS)undergo a conformational change and are inactivated. However, whether PTPs can actively regulate ROS levels in order to preventPTP inhibition has yet to be investigated. Here, we demonstrate that PTP non-receptor type 12 (PTPN12) protects cells againstaberrant ROS accumulation and death induced by oxidative stress. Murine embryonic fibroblasts (MEFs) deficient in PTPN12underwent increased ROS-induced apoptosis under conditions of antioxidant depletion. Cells lacking PTPN12 also showeddefective activation of FOXO1/3a, transcription factors required for the upregulation of several antioxidant genes. PTPN12-mediated regulation of ROS appeared to be mediated by phosphoinositide-dependent kinase-1 (PDK1), which was hyperstimulatedin the absence of PTPN12. As tight regulation of ROS to sustain survival is a key feature of cancer cells, we examined PTPN12 levelsin tumors from a cohort of breast cancer patients. Patients whose tumors showed high levels of PTPN12 transcripts had asignificantly poorer prognosis. Analysis of tissues from patients with various breast cancer subtypes revealed that more triple-negative breast cancers, the most aggressive breast cancer subtype, showed high PTPN12 expression than any other subtype.Furthermore, both human breast cancer cells and mouse mammary epithelial tumor cells engineered to lack PTPN12 exhibitedreduced tumorigenic and metastatic potential in vivo that correlated with their elevated ROS levels. The involvement of PTPN12 inthe antioxidant response of breast cancer cells suggests that PTPN12 may represent a novel therapeutic target for this disease.

Oncogene advance online publication, 25 February 2013; doi:10.1038/onc.2013.24

Keywords: oxidative stress; reactive oxygen species; antioxidant; protein tyrosine phosphatase; breast cancer

INTRODUCTIONInefficiencies in the cellular machinery governing mitochondrialmetabolism and protein translation lead to the production of freeradicals known as reactive oxygen species (ROS). These moleculeshave important roles in numerous aspects of cellular homeostasisbecause they oxidize and either activate or inactivate severalclasses of proteins, thereby modifying multiple signaling path-ways. Arguably the best-studied targets of ROS-mediated mod-ification are protein tyrosine phosphatases (PTPs).1 PTPs contain aconserved cysteine residue in their catalytic domain that, whenoxidized, triggers a conformational change in the protein andsubsequent inactivation of its enzymatic function. Thus, almost allPTPs become oxidized under conditions of oxidative stress.2

Although extensive research has been conducted to determinehow ROS inactivate PTPs, very little effort has been dedicated toinvestige the converse: whether PTPs can regulate ROS levels toprotect against PTP oxidation and inactivation.

Although the role of ROS in mammalian cancers has beeninvestigated for decades, it is not yet completely understood. It isclear that ROS can contribute to de novo tumorigenesis by inducingDNA damage and causing mutagenic events,3 but ROS can alsohinder tumorigenesis by inducing cell senescence or cell death.Accordingly, cancer cells exercise intricate control over their ROS

levels, and several studies have demonstrated that deregulating theROS/antioxidant balance in tumor cells leads to their death.4 Keyproteins used by both normal and cancer cells to control ROS arethe forkhead box O (FOXO) transcription factors.5 Severaldownstream effectors of FOXOs can neutralize ROS and thus aredirectly involved in counteracting oxidative stress.6 This reliance onFOXO-mediated ROS regulation has raised the intriguing possibilitythat disruption of a cancer cell’s ROS management machinerymight offer a novel therapeutic approach to cancer treatment.7

In this study, we investigated whether PTP non-receptor type 12(PTPN12; also known as PTP-PEST) is crucial for controlling ROS intumor cells and particularly in breast cancers, the most prevalentmalignancy among women. Our data indicate that PTPN12potentially has an oncogenic function (at least in breast cancers)because it regulates ROS in a manner that promotes tumor cellsurvival.

RESULTSPTPN12 promotes cell survival under conditions of antioxidantdepletionTo investigate whether PTPN12 promotes survival under condi-tions of increased oxidative stress, we utilized immortalized MEFs

1The Campbell Family Institute for Breast Cancer Research at Princess Margaret Hospital, University Health Network, Toronto, Ontario, Canada; 2Department of MedicalBiophysics, University of Toronto, Toronto, Ontario, Canada; 3Laboratory Medicine Program, University Health Network, Toronto, Ontario, Canada; 4Department of LaboratoryMedicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada; 5Department of Neuropathology, Heinrich Heine University, Dusseldorf, Germany; 6Laboratory ofApoptosis in Cancer, MRC Toxicology Unit, University of Leicester, Leicester, UK and 7Department of Biochemistry, Rosalind and Morris Goodman Cancer Centre, McGill University,Montreal, Quebec, Canada. Correspondence: Professor TW Mak, The Campbell Family Institute for Breast Cancer Research at Princess Margaret Hospital, University HealthNetwork, 620 University Avenue, Suite 706, Toronto, Ontario, Canada M5G 2C1.E-mail: tmak@uhnres.utoronto.caReceived 11 September 2012; revised 21 December 2012; accepted 23 December 2012

Oncogene (2013), 1–8& 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13

www.nature.com/onc

derived from PTPN12-null mice.8 To ensure that cells of anidentical genetic background were studied, PTPN12-null MEFswere transfected with either empty vector (to generate cellsreferred to as Ptpn12-KO) or a Ptpn12 expression construct (togenerate cells re-expressing Ptpn12 and referred to as Ptpn12-RE).8

The level of expression of PTPN12 protein in Ptpn12-RE MEFs wascomparable to its expression in MEFs derived from Ptpn12-competent mice (referred to as Ptpn12-WT) (Figure 1a, Supple-mentary Figure S1), validating our approach.

Cells metabolize glucose either via glycolysis to produceadenosine-5’-triphosphate, or via the pentose phosphate pathwayto produce the antioxidant reduced nicotinamide adeninedinucleotide phosphate (NADPH).9 Without NADPH productionthrough the pentose phosphate pathway, cells succumb to ROS-mediated apoptosis. To create an oxidizing environment, wetreated Ptpn12-RE and Ptpn12-KO MEFs with 6-aminonicotinamide(6-AN). 6-AN inhibits glucose-6-phosphate dehydrogenase, thefirst enzyme in the pentose phosphate pathway, and so blocksNADPH production and promotes ROS-mediated cell death.10 Wefound that Ptpn12-KO MEFs treated with 6-AN showed asignificantly higher rate of apoptosis than control cells(Figure 1b). As one function of NADPH is the regeneration ofreduced glutathione (GSH), the most abundant intracellularantioxidant,11 we investigated whether direct depletion of GSHwould also increase the apoptosis of Ptpn12-KO cells. To depleteGSH, we took advantage of the fact that de novo GSH synthesis isdriven by the action of glutamate cysteine ligase on themetabolites glutamate and cysteine. Buthionine sulfoximine(BSO) is a potent inhibitor of glutamate cysteine ligase activityand thus depletes a cell of GSH. We treated Ptpn12-RE andPtpn12-KO MEFs with varying concentrations of BSO andfound that, in the absence of PTPN12, cells were more sensitiveto GSH depletion and underwent increased apoptosis in adose-dependent manner (Figure 1c). These results suggest thatPTPN12 protects cells from oxidative stress caused by deficits inantioxidants such as NADPH and GSH.

L-glutamine (Gln) is essential for cell survival, and a commoneffect of many oncogenes is to increase Gln metabolism in cancercells.12 Gln undergoes glutaminolysis to generate glutamate,11

which can then be processed via glutamate cysteine ligase to formGSH. It has therefore been hypothesized that Gln deprivationleads to diminished antioxidant production and thus susceptibilityto the induction of endogenous oxidative stress. To determine ifPTPN12 counteracts the effects of Gln deprivation, we culturedPtpn12-RE and Ptpn12-KO MEFs in fully supplemented (normal) orGln-free medium. Measurement of intracellular ROS levels inPtpn12-RE and Ptpn12-KO MEFs revealed that, although ROSincreased in both cell lines under Gln-free conditions, Ptpn12-KOcells contained approximately twice as much ROS as controls(Figure 2a). When we monitored cell viability, we observedapproximately three-fold more cell death in Gln-free Ptpn12-KOcultures compared with Gln-free Ptpn12-RE cultures (Figure 2b). Totest if this Gln deprivation-induced death could be avoided bysupplementation with an exogenous antioxidant, we culturedPtpn12-RE and Ptpn12-KO MEFs in Gln-free medium with orwithout the antioxidant N-acetyl cysteine (NAC). Although Ptpn12-RE cells were unaffected by NAC treatment, this antioxidantsignificantly reduced apoptosis in Gln-free Ptpn12-KOMEF cultures(Figure 2b). Next, we examined the proliferation of Ptpn12-RE andPtpn12-KO MEFs cultured without Gln. In normal medium, MEFs ofboth genotypes grew at similar rates, as did Ptpn12-null MEFstransfected with a vector expressing a phosphatase-dead form ofPTPn12 (R237M) (Figure 2c). However, in the absence of Gln, onlyMEFs expressing PTPN12 were able to proliferate robustly, and thisproliferation defect was not rescued by the expression of PTPn12-KO-R237M (Figure 2d). These data indicate that the phosphatasefunction of PTPN12 is required for cell growth in an environmentlacking antioxidants, and imply that the increased apoptosis and

reduced proliferation of Ptpn12-KO cells subjected to Gln with-drawal is directly related to the elevated ROS within these cells.

PTPN12 promotes FOXO activation by inhibiting PDK1 signalingThe results of a previously reported large-scale screen showedthat PTPN12 dephosphorylates PDK1.13 PDK1 modulates oxidativestress, which may explain the sensitivity of Ptpn12-KO cells to ROSaccumulation induced by antioxidant depletion.14 We serum-starved Ptpn12-RE and Ptpn12-KO MEFs overnight and restimu-lated them with 20% FBS over a 2 h time course. Immunoblottingrevealed that more PDK1 was phosphorylated at Tyr9 and Ser241in Ptpn12-KO MEFs than in Ptpn12-RE MEFs; even at time zero(Figure 3a). Although autophosphorylation of Ser241 in theactivation loop of PDK1 is necessary for its activity, tyrosinephosphorylation at Tyr9 has been shown to promote PDK1 proteinstabilization and activity.15,16 Our data suggests that loss ofPTPN12 results in excessive phosphorylation of PDK1 and thusupregulation of PDK1 activity.

Next, we examined the downstream targets of PDK1 known toinfluence the regulation of antioxidants, namely the FOXO1/3atranscription factors. In a resting cell, FOXO proteins aresequestered in a phosphorylated inactive form in the cytoplasm.17

This phosphorylation is mediated by serum/glucocorticoid-regulated kinase 1 (SGK1), which can be directly phosphorylatedand activated by PDK1 independent of membrane recruitment byphosphoinositide-3-kinase.18 As FOXO1/3a dephosphorylation andactivation is necessary for antioxidant enzyme expression,increased PDK1 phosphorylation leading to heightened PDK1

Figure 1. PTPN12 supports cell survival under conditions of NADPHor GSH deprivation. (a) Immunoblot detecting PTPN12 protein inextracts of Ptpn12-KO and Ptpn12-RE MEFs. Actin, loading control. (b)Apoptosis of Ptpn12-RE and Ptpn12-KO MEFs treated for 24 h witheither DMSO (vehicle) or 150 mM 6-AN, followed by staining withAnnexin V/PI. (c) Apoptosis of Ptpn12-RE and Ptpn12-KO MEFstreated for 24 h with the indicated concentrations of BSO, followedby staining with Annexin V/PI. For all Figures, results arerepresentative of at least three trials. Where indicated, data shownare the mean±s.e.m. of triplicate experiments (n.s.¼non-signifi-cant; **Po0.01; ***Po0.001).

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

2

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

activity could result in FOXO1/3a inactivation and decreasedantioxidant activity. Accordingly, we postulated that thehyperphosphorylated PDK1 in Ptpn12-KO MEFs might result indecreased FOXO activity that could account for the increasedsensitivity of these cells to oxidative stress. Indeed, we observed amarked increase in phosphorylated FOXO1/3a in Ptpn12-KO MEFscompared with Ptpn12-RE MEFs (Figure 3a). These data suggestthat, in the absence of PTPN12, PDK1 is ‘hyperactivated’ such thatFOXOs are inactivated, compromising the intracellular antioxidantresponse. To support our hypothesis, we assayed mRNA levels ofan array of FOXO target genes in Ptpn12-KO and PTPn12-KO-R237M MEFs. Among other validated FOXO target genes,H6PD (which controls NADPH production in the endoplasmicreticulum), CD44 (promotes GSH production through cystineuptake) and HO-1 (controls iron-mediated ROS regulation) weredownregulated in the absence of functional PTPN12 (Figure 3b,Supplementary Figure S2).19–21

FOXO3a has been reported to be inactivated under increasingROS levels.22 To confirm that the increased phosphorylation (andsubsequent inactivation) of FOXO1/3a seen in PTPn12-KO MEFs isdue to PTPn12 directly and not simply an indirect effect of highROS levels, we treated cells with the antioxidant NAC. Althoughthe FOXO3a phosphorylation decreased in control cells, PTPn12-KO MEFs still contained a high level of FOXO1/3a phosphorylation(Figure 3c). This result indicates that PTPn12 has a direct effect onpreventing the phosphorylation and inactivation of FOXO1/3a.

Finally, we modulated PDK1 and FOXO1/3a in an effort torescue the increased ROS levels observed in PTPn12-KO MEFs. Wewere unable to observe any changes in ROS levels when PDK1mRNA was reduced (Supplementary Figures S3a and b), poten-tially because any remaining hyperactive PDK1 protein afterknockdown of PTPn12, may still inhibit FOXO1/3a activation.Alternatively, this result suggests that PTPn12 may control ROSlevels in a PDK1-independent fashion. As FOXO1/3a is the

downstream target of PDK1 that ultimately could affect the ROSlevels in PTPn12-KO MEFs, we decided to express a constitutivelyactive FOXO3a:A3 in cells under stressed condition. Under Gln-freeconditions, expression of FOXO3a:A3 was able to reduce the ROSlevels in PTPn12-KO MEFs cells (Figure 3d). We did not see arescue of the cell death in PTPn12-KO MEFs, yet this may be dueto the fact that FOXO3a also induces expression of pro-apoptotictargets, such as Bim23 (Supplementary Figures S4a and b). Overall,these results suggest that deregulation of FOXO signaling mayaccount for the increased sensitivity of PTPN12-deficient cells tooxidative stress.

High levels of PTPN12 mRNA predict poor prognosis in breastcancer patientsTight regulation of oxidative stress within a cell is closely linked toits tumorigenic potential.9 Significantly, previous work has shownthat human breast tumors and numerous commonly used breastcancer cell lines possess a mutated form of PTPN12 that hasincreased phosphatase activity.24 However, whether PTPN12confers a growth advantage on cancer cells under stress has yetto be determined. To shed light on this issue, we examined twopreviously published datasets of mRNA levels in human breastcancer samples25,26 and found that, in both cases, patients withhigh PTPN12 mRNA expression had a reduced probability ofrelapse-free disease (Figures 4a and b). To further explore whetherPTPN12 expression correlated with breast tumorigenesis, we usedimmunohistochemistry (IHC) to examine PTPN12 protein expres-sion in 240 patient-derived primary breast tumors of varioussubtypes (Supplementary Table S1). We found that significantlymore triple-negative breast cancers (TNBC; ER�PR�HER2� )showed high levels of PTPN12 compared with non-TNBC breastcancers (Figures 4c and d). There was no correlation betweenPTPN12 expression and either tumor grade or size (Supplementary

Figure 2. PTPN12 promotes cell survival in a Gln-free environment. (a) Intracellular ROS levels as determined by CM-H2DCFDA staining inPtpn12-RE and Ptpn12-KO MEFs cultured for 48 h in normal or Gln-free medium. Results are expressed as arbitrary units (a.u.)-based meanfluorescent intensity. (b) Apoptosis of Ptpn12-RE and Ptpn12-KO MEFs after 72 h in normal, Gln-free or Gln-free medium supplemented withNAC (1mM), followed by staining with Annexin V/PI. (c, d). Numbers of viable cells in cultures of Ptpn12-RE, Ptpn12-KO and Ptpn12-KO-R237MMEFs grown for 4 days in normal (c) or Gln-free (d) medium as determined by trypan blue exclusion. Data are expressed as the fold change incell number compared to day 1 values (*Po0.05; **Po0.01; ***Po0.001).

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

3

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

Figures S5a and b). Similar results were obtained when the samesamples were subjected to IHC using an alternative anti-PTPN12antibody (Supplementary Table S2). These data suggest that highlevels of PTPN12 allow breast tumor cells to better combatexogenous stress, and imply that PTPN12 may promote and/orsupport breast cancer progression.

Loss of PTPN12 increases ROS levels and inhibits breast cancer cellgrowth in vivoTNBCs show an overlap with tumors of the ‘basal’ breast cancersubtype, which are generally aggressive breast tumors associatedwith very poor patient survival.27 To determine if loss of PTPN12 inbasal-like breast cancer cell lines would decrease theirtumorigenic potential, we used the short hairpin RNA (shRNA)approach to deplete PTPN12 in MDA-MB-231 cells (Figure 5a, left).MDA-MB-231 cells cultured under either normal or Gln-freeconditions contained higher ROS when PTPN12 expression wasreduced (Figure 5b). This increase in ROS levels was not as large asthe change observed between Ptpn12-RE and Ptpn12-KO MEFs(Figure 2a); potentially because the Ptpn12-KO MEFs completelylack PTPN12 protein, whereas shRNA against PTPN12 can neverfully ablate the mRNA and subsequent protein expression. Toinvestigate the impact of PTPN12 depletion on the tumorigenicpotential of MDA-MB-231 cells in vivo, we subcutaneously injectedMDA-MB-231 cells expressing control shRNA or PTPN12 shRNA

into NIH-III nude mice and monitored tumor growth. Strikingly, thetumors formed by MDA-MB-231 cells lacking PTPN12 weresignificantly smaller than tumors derived from control cells(Figure 5c). IHC analysis of these tumors confirmed that PTPN12protein expression was strongly reduced in vivo (data not shown),suggesting that the observed reduction in tumorigenicity was dueto loss of PTPN12.

We bolstered these results by conducting parallel experimentsin which Ptpn12 shRNA was introduced into a spontaneouslyimmortalized mouse mammary tumor cell line (referred to asPyMT cells) derived from MMTV-PyMT mice.28 We injected PyMTcontrol and PyMT Ptpn12 knockdown cells (Figure 5a, right) intothe tail veins of 129P2Ola mice and monitored tumor develop-ment and mouse survival. We found that lung metastasis wasdrastically reduced and overall survival was prolonged in animalsinjected with PyMT cells lacking PTPN12 (Figure 5d). These datareinforce our contention that PTPN12 promotes and/or supportstumorigenesis and metastasis of breast cancer cells in vivo.

To investigate a potential mechanism by which PTPN12 mightpromote tumorigenesis and/or metastasis, we determinedwhether the signaling alterations we observed in Ptpn12-KO MEFsalso occurred in breast cancer cells expressing PTPN12 shRNA.Indeed, like Ptpn12-KO MEFs, PTPN12-depleted MDA-MB-231 cellsshowed elevated phospho-FOXOs (Figure 6a). Taken together, ourdata indicate that cells lacking PTPN12 experience diminishedFOXO signaling, which may be due to dysregulation of the

Figure 3. PTPN12 regulates PDK1 signaling and has downstream effects on FOXOs and their targets. (a) Immunoblot to detect the indicatedtotal and phosphorylated (P) proteins in lysates of Ptpn12-RE and Ptpn12-KO MEFs that were serum-starved overnight and stimulated with 20%FBS for the indicated times. (b) Quantitative RT-PCR analysis of Ptpn12-RE, Ptpn12-KO and Ptpn12-KO-R237M MEFs to detect mRNAs for theindicated FOXO target genes. Data are expressed as mRNA level relative to housekeeping geneRps9. (c) Immunoblot to detect the indicatedtotal and phosphorylated (P) proteins in lysates of Ptpn12-RE and Ptpn12-KO MEFs that were serum-starved overnight or cultured in mediumwith 10% FBS, and treated with or without NAC (1mM). (d) Intracellular ROS levels in Ptpn12-RE and Ptpn12-KO MEFs transiently transfectedwith either vector control or FOXO3a:A3 expression vector and cultured for 48 h in Gln-free medium. Results are expressed as ROS levelsrelative to control Ptpn12-RE MEFs and based on mean fluorescent intensity (*Po0.05; **Po0.001).

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

4

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

PDK1–SGK1 axis. Such a defect could account for the increasedROS levels observed in PTPN12-deficient cancer cells and thedecreased tumorigenicity of these cells in vivo. A model of thishypothetical series of events is shown in Figure 6b.

DISCUSSIONROS are known to modulate PTPs by inducing changes to theirprotein structure that cause their functional inactivation.1 Weinitiated our study in an effort to determine if PTPs canreciprocally regulate ROS in order to prevent this inactivation.Our investigation of the relatively poorly characterized enzymePTPN12 has demonstrated that this PTP can function tocounteract or reduce oxidative stress. When we modulated ROSlevels either directly by inhibiting NADPH or GSH synthesis using6-AN or BSO, or indirectly by Gln deprivation, we found that MEFslacking PTPN12 were unable to cope with the resultingperturbations in ROS and ultimately underwent apoptosis.PTPN12 is thus required for effective ROS regulation.

Recent studies have indicated that cancer cells develop anintricate system of ROS control in order to survive underconditions of oxidative stress.4 Although a lack of NADPHproduction due to loss of glucose-6-phosphate dehydrogenasemight be predicted to inhibit tumorigenesis, women withinherited mutations of glucose-6-phosphate dehydrogenase(mimicking the effect of 6-AN) show no decrease in breastcancer incidence.29 In line with this observation, there is noevidence from explorations of BSO in the clinical setting that BSO

treatment alone is therapeutic for breast cancer patients.30

Similarly, Gln depletion via L-asparaginase treatment is effectiveonly in acute lymphoblastic leukemia patients.9 Considering thesedata, we hypothesize that these therapeutic avenues have failedin part because of oncogenic factors (like PTPN12) that promoteantioxidant mechanisms and thus maintain ROS regulation. Wetherefore believe that targeting such molecules may leave tumorcells vulnerable to antioxidant inhibition, which in turn mayultimately force them to succumb to ROS-induced death.

A previous report by Sun et al.31 claimed that PTPN12 is a tumorsuppressor because PTPN12 expression was often missing intumor samples from patients with TNBCs. However, in our study,we used two different anti-PTPN12 antibodies, including thatemployed by Sun et al.,31 to perform IHC on tumor samples from240 patients with various subtypes of breast cancer. Our results donot support the conclusion that TNBCs have diminished or absentexpression of PTPN12. On the contrary, our broad screen hasdemonstrated that not only is PTPN12 protein present in almostall subtypes of breast tumors, but also that more TNBCs show highexpression than any other breast cancer subtype. We bolsteredthis finding with our analysis of two breast cancer mRNA datasets,which established that high levels of PTPN12 mRNA correlate witha less favorable prognosis. Thus, the results of our study clearlyindicate that elevated PTPN12 expression is beneficial for andsupportive of tumor development, and are consistent with thehypothesis that rapidly proliferating cancer cells developmechanisms that confer a high level of resistance to cellularstress, in particular oxidative stress. Additional in-depth

Figure 4. High levels of PTPN12 mRNA in a breast tumor predict poor patient prognosis. (a, b) PTPN12 mRNA levels in tumors of breast cancerpatients examined in the Perou2 (a) and Chin (b) datasets were plotted and correlated with the fraction of relapse-free patients. (c, d) Sectionsof tumors from 240 patients with TNBC or ‘non-TNBC’ (ERþHER2þ ) breast cancers were subjected to IHC to detect PTPN12 protein. Tumorswere scored as exhibiting either high (total Allred score of 7-8) or low (total Allred score of p6) PTPN12 protein expression. (c) Representativeimages of tumors with low or high PTPN12 protein shown at low magnification (upper panels; scale bar¼ 100 mm) and high magnification(lower panels; scale bar¼ 50 mm). (d) Percentages of patients with TNBC or non-TNBC tumors showing high or low PTPN12 expression.

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

5

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

investigation of the impact of other PTPs on ROS control, and thedelineation of alterations to these PTPs that may be associatedwith tumorigenesis, will no doubt yield many useful insights intohow oxidative stress is overcome by cancer cells.

MATERIALS AND METHODSCell lines, culture conditions and reagentsImmortalized mouse embryonic fibroblasts (MEFs) from PTPN12-null micewere transfected with either empty vector (pcDNA3.1—referred to asPtpn12-KO) or a wild-type mouse full length open reading frame (ORF)Ptpn12 expression construct (referred to as Ptpn12-RE).8 The human breastcancer cell line MDA-MB-231 was obtained from ATCC. A cell line derivedfrom MMTV-PyMT mice (referred to as PyMT cells) was maintained aspreviously described.28 FOXO3a:A3 expression was used as previouslydescribed.32 Unless otherwise indicated, all cell lines were maintained inDulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetalbovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin. Forserum starvation, the cells were cultured overnight in serum-free DMEM.For Gln deprivation, cells were cultured in DMEM without Gln (GIBCO10313, Burlington, ON, Canada). All antibodies were from Cell Signaling,except for anti-PTPN12 (Bethyl Laboratories Inc., Montgomery, TX, USA orSigma HPA007097, Oakville, ON, Canada) and anti-FOXO3A (Millipore,Billerica, MA, USA). All chemical compounds were from Sigma if nototherwise stated.

RNA interferenceMDA-MB-231 or PyMT cells stably expressing control shRNA or shPTPN12were generated by lentiviral transduction, using the pGIPZ system(OpenBiosystems Catalogue no. RHS4531-NM_002835 and Catalogue no.RMM4532-NM_011203, Pittsburgh, PA, USA). Transduced cells wereselected for minimum 7 days in culture in the presence of 1 mg/mlpuromycin. Transient knockdown of PDK1 mRNA was accomplished usingsmall interfering (siRNA) against mouse PDK1 (Dhamarcon, L-040658-00-0020, Pittsburgh, PA, USA) with Lipofectomine 2000 (Invitrogen, Burling-ton, ON, Canada).

Cell proliferationReplicate cultures of cells were grown for 4 days before staining withtrypan blue to determine viability. Viable cells were counted on each day(days 1–4) using a Vi-Cell XR cell viability analyzer (Beckman Coulter,Mississauga, ON, Canada) to establish growth curves.

Apoptosis and ROSCells were cultured in Gln-free medium or treated with NAC, BSO or 6-ANat the concentrations and for the times indicated in the Figures. Apoptoticcells were measured by staining with AnnexinV (BD Biosciences,Mississauga, ON, Canada) plus propidium iodide (PI), followed by flowcytometric measurement of fluorescence using a FACSCalibur (BectonDickinson, Franklin Lakes, NJ, USA). To determine relative ROS levels, cellswere stained with CM-H2DCFDA (Invitrogen) and subjected to FACSanalysis. ROS levels are reported as mean fluorescence intensity expressedin arbitrary units.

Quantitative real-time PCR (qRT–PCR)mRNA was prepared from cell extracts using the NucleoSpin RNA II kit(Macherey-Nagel, Duren, Germany)). mRNA was reverse-transcribed intocDNA using the iScriptcDNA synthesis kit (Bio-Rad, Mississauga, ON,Canada). Real-time PCR analysis was performed using POWER SYBR GreenPCR Master Mix and a 7900HT Fast Real-Time PCR System (AppliedBiosystems, Burlington, ON, Canada). mRNA levels of PDK1 and the FOXOtarget genes BIM, TRAIL, CD44,H6PD and HO-1 were determined relative tothe housekeeping gene RPS9. All experiments were performed intriplicates.

Analysis of breast cancer patient datasetsThe Perou226 and Chin25 datasets were independently analyzed todetermine PTPN12 mRNA expression. Patient survival was examinedusing Cox regression analysis of survival times vs gene expression logratios, followed by the estimation of false discovery rates using theBenjamini and Hochberg method. High expression of PTPN12 mRNA wasassociated with greater risks of tumor recurrence and death. These

Figure 5. PTPN12 suppresses ROS levels and promotes tumorigenesis. (a) Immunoblot to detect PTPN12 protein in extracts of MDA-MB-231 andPyMT cell lines expressing control shRNA or shPTPN12. (b) Quantitation of ROS in MDA-MB-231 cells expressing control shRNA or shPTPN12andgrown in normal or Gln-free medium. ROS levels were determined as for Figure 2a. (c) Total tumor mass in immunodeficient NIH-III nude mice(n¼ 10) subcutaneously injected with MDA-MB-231 cells expressing control shRNA (left flank) or shPTPN12 (right flank). After 84 days, tumorswere excised and weighed. Results are the mean±s.e.m. of total tumor mass per mouse. (d) Kaplan-Meier analysis of the survival of 129P2Olamice injected with PyMT cells expressing control shRNA or shPTPN12 (n¼ 10/group). (*Po0.05).

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

6

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

associations reached formal statistical significance with P values o0.05.Kaplan–Meier curves were generated for each dataset for illustrationpurposes by rank-ordering the log ratios and splitting them into threeequal groups. The curves for the highest and the lowest tertiles weredisplayed on the plots shown in Figures 4a and b.

Immunohistochemistry of human breast tumor tissue arraySix in-house tissue microarrays with a total of 240 unique breast cancercases represented in duplicate, together with appropriate control sections,were subjected to immunostaining with anti-PTPN12 antibody (Bethyl IHC-00260 or Sigma HPA007097) after obtaining UHN research ethics boardapproval. The TMAs contained consecutive breast cancer cases collected ata single institution (UHN) over a calender year (2006). Receptor status (ER/PR/HER2) had been routinely assessed on all cases. Stained tissue sectionswere evaluated by light microscopy and scored blinded by two

independent pathologists (SJD and JM) to produce an Allred score.33

The total Allred score is the sum of the proportion score (percentage ofcells stained: 0%¼ 0; o1%¼ 1; 1–10%¼ 2; 10–33%¼ 3; 34–66%¼ 4;466%¼ 5) and the intensity score (no staining¼ 0; weak¼ 1;moderate¼ 2; strong¼ 3). Values were stratified into two groups, namelyhigh PTPN12 expression (total Allred score of 7-8) or low PTPN12expression (total Allred score of p6).34

XenograftsFor MDA-MB-231 cell analyses, immunodeficient NIH-III nude mice (5weeks old) were subcutaneously injected in both flanks with 3� 106 MDA-MB-231 cells expressing control shRNA (left flank) or shPTPN12 (right flank).Mice were monitored every 4 days for the presence of palpable tumors.After 84 days, the mice were killed and the tumors were excised. Tumorweights were recorded and tissues were photographed and processed forfurther analysis. For PyMT cell analyses, control or shPTPN12 PyMTtumor cells (2.5� 105 in 200ml PBS) were injected into the tail veins of129P2Ola mice. Mouse survival over 200 days was evaluated using Kaplan–Meier methodology. All animal experiments were approved by theAnimal Care and Use Committee of the University Health Network(Toronto, Canada).

StatisticsKaplan–Meier survival plots and statistics were created and analyzedusing GraphPad PRISM-5 software (La Jolla, CA, USA). The two-tailedStudent’s t-test with the s.e.m. was used to evaluate differences in tumorcell proliferation and survival.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSWe thank Dr Anne Brustle and Dr Dirk Brenner for helpful discussions, Dr Han Youand Dr Walbert Bakker for reagents and Dr Mary Saunders for insightful scientificediting. This work was supported by a PGS-D grant from the Natural Sciences andEngineering Council of Canada (ISH) and grant no. 179815 from the CanadianInstitutes of Health Research (TWM).

REFERENCES1 Xu D, Rovira II, Finkel T. Oxidants painting the cysteine chapel: redox regulation of

PTPs. Dev Cell 2002; 2: 251–252.2 Karisch R, Fernandez M, Taylor P, Virtanen C, St-Germain JR, Jin LL et al. Global

proteomic assessment of the classical protein-tyrosine phosphatome and‘redoxome’. Cell 2011; 146: 826–840.

3 Wiseman H, Halliwell B. Damage to DNA by reactive oxygen and nitrogenspecies: role in inflammatory disease and progression to cancer. Biochem J 1996;313: 17–29.

4 Schafer ZT, Grassian AR, Song L, Jiang Z, Gerhart-Hines Z, Irie HY et al. Antioxidantand oncogene rescue of metabolic defects caused by loss of matrix attachment.Nature 2009; 461: 109–113.

5 Storz P. Forkhead homeobox type O transcription factors in the responses tooxidative stress. Antioxid Redox Signal 2011; 14: 593–605.

6 Greer EL, Brunet A. FOXO transcription factors at the interface between longevityand tumor suppression. Oncogene 2005; 24: 7410–7425.

7 Myatt SS, Brosens JJ, Lam EW. Sense and sensitivity: FOXO and ROS in cancerdevelopment and treatment. Antioxid Redox Signal 2011; 14: 675–687.

8 Cote JF, Charest A, Wagner J, Tremblay ML. Combination of gene targeting andsubstrate trapping to identify substrates of protein tyrosine phosphatases usingPTP-PEST as a model. Biochemistry 1998; 37: 13128–13137.

9 Cairns RA, Harris IS, Mak TW. Regulation of cancer cell metabolism. Nat Rev Cancer2011; 11: 85–95.

10 Vaughn AE, Deshmukh M. Glucose metabolism inhibits apoptosis in neuronsand cancer cells by redox inactivation of cytochrome c. Nat Cell Biol 2008; 10:1477–1483.

11 Meister A. Selective modification of glutathione metabolism. Science 1983; 220:472–477.

12 DeBerardinis RJ, Cheng T. Q’s next: the diverse functions of glutamine inmetabolism, cell biology and cancer. Oncogene 2010; 29: 313–324.

Figure 6. PTPN12 regulates ROS levels in tumor cells via FOXOsignaling. (a) Immunoblot to detect the indicated proteins in lysatesof MDA-MB-231 cells expressing control shRNA or shPTPN12 andcultured under normal (þ ) or serum starvation (� ) conditions.(b) Model of alterations to FOXO signaling that might occur in theabsence of PTPN12. Left: PTPN12 promotes FOXO1/3a activation(either directly or through PDK1 inhibition) and subsequentlypromote the expression of FOXO1/3a target genes that combatoxidative stress. The cell is permitted to survive. Right: in theabsence of PTPN12, excessive FOXO1/3a inactivation occurs and afailure to express antioxidant molecules. Oxidative stress cannot besuccessfully combatted and the cell is more likely to die.

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

7

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

13 Barr AJ, Ugochukwu E, Lee WH, King ON, Filippakopoulos P, Alfano I et al.Large-scale structural analysis of the classical human protein tyrosine phospha-tome. Cell 2009; 136: 352–363.

14 Nogueira V, Park Y, Chen CC, Xu PZ, Chen ML, Tonic I et al. Akt determinesreplicative senescence and oxidative or oncogenic premature senescence andsensitizes cells to oxidative apoptosis. Cancer Cell 2008; 14: 458–470.

15 Casamayor A, Morrice NA, Alessi DR. Phosphorylation of Ser-241 is essential forthe activity of 3-phosphoinositide-dependent protein kinase-1: identification offive sites of phosphorylation in vivo. Biochem J 1999; 342(Pt 2): 287–292.

16 Yang KJ, Shin S, Piao L, Shin E, Li Y, Park KA et al. Regulation of 3-phosphoino-sitide-dependent protein kinase-1 (PDK1) by Src involves tyrosine phosphoryla-tion of PDK1 and Src homology 2 domain binding. J Biol Chem 2008; 283:1480–1491.

17 Brunet A, Park J, Tran H, Hu LS, Hemmings BA, Greenberg ME. Protein kinase SGKmediates survival signals by phosphorylating the forkhead transcription factorFKHRL1 (FOXO3a). Mol Cell Biol 2001; 21: 952–965.

18 Park J, Leong ML, Buse P, Maiyar AC, Firestone GL, Hemmings BA. Serum andglucocorticoid-inducible kinase (SGK) is a target of the PI 3-kinase-stimulatedsignaling pathway. EMBO J 1999; 18: 3024–3033.

19 Rogoff D, Black K, McMillan DR, White PC. Contribution of hexose-6-phosphatedehydrogenase to NADPH content and redox environment in the endoplasmicreticulum. Redox Rep 2010; 15: 64–70.

20 Ishimoto T, Nagano O, Yae T, Tamada M, Motohara T, Oshima H et al. CD44 variantregulates redox status in cancer cells by stabilizing the xCT subunit of system xc(-)and thereby promotes tumor growth. Cancer Cell 2011; 19: 387–400.

21 Cheng Z, Guo S, Copps K, Dong X, Kollipara R, Rodgers JT et al. Foxo1 integrates insulinsignaling with mitochondrial function in the liver. Nat Med 2009; 15: 1307–1311.

22 Nemoto S, Finkel T. Redox regulation of forkhead proteins through a p66shc-dependent signaling pathway. Science 2002; 295: 2450–2452.

23 Essafi A, Fernandez de Mattos S, Hassen YA, Soeiro I, Mufti GJ, Thomas NS et al.Direct transcriptional regulation of Bim by FoxO3a mediates STI571-inducedapoptosis in Bcr-Abl-expressing cells. Oncogene 2005; 24: 2317–2329.

24 Streit S, Ruhe JE, Knyazev P, Knyazeva T, Iacobelli S, Peter S et al. PTP-PEST phos-phatase variations in human cancer. Cancer Genet Cytogenet 2006; 170: 48–53.

25 Chin K, DeVries S, Fridlyand J, Spellman PT, Roydasgupta R, Kuo WL et al. Genomicand transcriptional aberrations linked to breast cancer pathophysiologies. CancerCell 2006; 10: 529–541.

26 Hoadley KA, Weigman VJ, Fan C, Sawyer LR, He X, Troester MA et al. EGFRassociated expression profiles vary with breast tumor subtype. BMC Genomics2007; 8: 258.

27 Cheang MC, Voduc D, Bajdik C, Leung S, McKinney S, Chia SK et al. Basal-likebreast cancer defined by five biomarkers has superior prognostic value thantriple-negative phenotype. Clin Cancer Res 2008; 14: 1368–1376.

28 Arpaia E, Blaser H, Quintela-Fandino M, Duncan G, Leong HS, Ablack A et al. Theinteraction between caveolin-1 and Rho-GTPases promotes metastasis by con-trolling the expression of alpha5-integrin and the activation of Src, Ras and Erk.Oncogene 2011; 31: 884–896.

29 Forteleoni G, Argiolas L, Farris A, Ferraris AM, Gaetani GF, Meloni T. G6PD defi-ciency and breast cancer. Tumori 1988; 74: 665–667.

30 Bailey HH, Ripple G, Tutsch KD, Arzoomanian RZ, Alberti D, Feierabend C et al.Phase I study of continuous-infusion L-S,R-buthionine sulfoximine withintravenous melphalan. J Natl Cancer Inst 1997; 89: 1789–1796.

31 Sun T, Aceto N, Meerbrey KL, Kessler JD, Zhou C, Migliaccio I et al. Activation ofmultiple proto-oncogenic tyrosine kinases in breast cancer via loss of the PTPN12phosphatase. Cell 2011; 144: 703–718.

32 Bakker WJ, Harris IS, Mak TW. FOXO3a is activated in response to hypoxic stressand inhibits HIF1-induced apoptosis via regulation of CITED2. Mol Cell 2007; 28:941–953.

33 Allred DC, Harvey JM, Berardo M, Clark GM. Prognostic and predictivefactors in breast cancer by immunohistochemical analysis. Mod Pathol 1998; 11:155–168.

34 Yerushalmi R, Gelmon KA, Leung S, Gao D, Cheang M, Pollak M et al. Insulin-likegrowth factor receptor (IGF-1R) in breast cancer subtypes. Breast Cancer Res Treat2011; 132: 131–142.

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

PTPN12 prevents cell death induced by oxidative stressIS Harris et al

8

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具