High Frequency of Pik3r1and PIK3R2 Mutations in Endometrial Cancer

Published OnlineFirst June 7, 2011; DOI: 10.1158/2159-8290.CD-11-0039 Cancer Discovery; Published OnlineFirst on June 29, 2011 as doi:10.1158/2159-8290.CD-11-0039

RESEARCH ARTICLE

high Frequency of PIK3R1 and PIK3R2 Mutations in endometrial cancer elucidates a Novel Mechanism for regulation of PteN Protein stability

Lydia W.T. Cheung1,* , Bryan T. Hennessy 6, 7,*, J1 i e, Shuangxing L i Yu1 , Andrea P. Myers 8 ,, 9 Bojana Djordjevic11 , Y i l i n g1 , L Katherine u Stemke-Hale1 , Mary D. Dyer1 , Fan Zhang1. Z h e n1 l i, n2, Lewis J u C. Cantley9, 10, Steven E. Scherer5 , Han Liang2 , Karen H. Lu3 , Russell R. Broaddus4 , and Gordon B. Mills1

Downloaded from cancerdiscovery.aacrjournals.orgCopyright 2011 American Association on September for 29, Cancer 2021. © Research.2011 American Association for Cancer Research. Published OnlineFirst June 7, 2011; DOI: 10.1158/2159-8290.CD-11-0039

The BATTLE Trial: Personalizing Therapy for Lung Cancer research article

ABSTRACT We demonstrate that phosphatidylinositol 3-kinase (PI3K) pathway aberrations occur in .80% of endometrioid endometrial cancers, with coordinate mutations of multiple PI3K pathway members being more common than predicted by chance. PIK3R1 (p85α) mutations occur at a higher rate in endometrial cancer than in any other tumor lineage, and PIK3R2 (p85β), not previously demonstrated to be a cancer gene, is also frequently mutated. The dominant activation event in the PI3K pathway appears to be PTEN protein loss. However, in tumors with retained PTEN protein, PI3K pathway mutations phenocopy PTEN loss, resulting in pathway activation. KRAS mutations are common in endometrioid tumors activating independent events from PI3K pathway aberrations. Multiple PIK3R1 and PIK3R2 mutations demonstrate gain of function, including disruption of a novel mechanism of pathway regulation wherein p85α dimers bind and stabilize PTEN. Taken together, the PI3K pathway represents a critical driver of endometrial cancer pathogenesis and a novel therapeutic target.

siGNiFicaNce: Our data indicate that the PI3K pathway is targeted in the vast majority of endometrioid endometrial cancers leading to PI3K pathway activation. Frequent oncogenic mutations in PIK3R1 and PIK3R2 provide evidence for their role in endometrial cancer pathophysiology with patient-specific mutations revealing a novel mechanism by which p85α regulates the PI3K pathway through stabi- lizing PTEN. Cancer Discovery; 1(2); OF1–OF16. ©2011 AACR.

INTRODUCTION aberrations at multiple nodes ( 1 ). Inactivation of PTEN, the Endometrial cancer (EC) is the most prevalent gyneco- major negative regulator of the PI3K pathway, has been re- logic malignancy and the fourth most common cancer ported in phenotypically normal endometrium and is sufficient among women in Western countries. Mortality for localized to drive tumorigenesis in mice, suggesting that pathway activa- low-grade, low-stage endometrioid endometrial carcinomas tion is an early event in EC ( 2 , 3 ). Loss of PTEN, which can be (EEC) is low. However, treatment options for patients with caused by gene mutation, promoter methylation, and protein metastatic or recurrent EEC or nonendometrioid endome- degradation, is present in 20% of endometrial hyperplasia, 35% trial carcinomas (NEEC) are limited and outcomes are ex- to 50% of EEC, and 10% of NEEC ( 4–6 ). Activating PIK3CA tremely poor. Thus there is an urgent need to develop novel mutations are present in 30% of EEC and 15% of NEEC and effective targeted therapies. are frequently coexistent with PTEN aberrations ( 7 , 8 ). Somatic The phosphatidylinositol 3-kinase (PI3K) pathway that sig- mutations in AKT1 occur in 2% of EEC ( 9 ). Frequent mutations nals downstream from receptor tyrosine kinases (RTK) is fre- in fibroblast growth factor receptor 2 (FGFR2) in EEC (12%) quently activated in many cancer lineages, including EC, by also point to the importance of RTK signaling in the etiology of this disease ( 10 ). In addition, other molecular features of EC include gene mutations in pathways that interact with the PI3K Authors' Affiliations: Departments of 1 Systems Biology, 2 Bioinformatics pathway, including CTNNB1 and TP53 ( 11 , 12 ). and Computational Biology, 3 Gynecologic Oncology, and 4 Pathology, The The PI3K pathway can interact bidirectionally with the University of Texas MD Anderson Cancer Center; 5 Department of Molecular Ras/mitogen-activated protein kinase (MAPK) pathway, sug- and Human Genetics, Human Genome Sequencing Center, Baylor College of gesting that the two pathways might cooperate to determine Medicine, Houston, Texas; 6 Department of Medical Oncology, Beaumont Hospital; 7 Royal College of Surgeons of Ireland, Dublin, Ireland; 8 Division of functional outcomes ( 13 ). Activating KRAS mutations are Women's Cancers, Department of Medical Oncology, Dana-Farber Cancer found in approximately 20% of EEC ( 7 , 14 ). However, the Institute; 9 Division of Signal Transduction, Department of Medicine, Beth significance of the crosstalk between these pathways in EC 10 Israel Deaconess Medical Center; Department of Systems Biology, remains to be explored. Harvard Medical School, Boston, Massachusetts; and 11 Department of Pathology and Laboratory Medicine, University of Ottawa, Ottawa, Ontario, The frequent deregulation of RAS and PI3K signaling Canada in EC offers attractive candidates for targeted therapy. *These authors contributed equally to this article. Indeed, compounds that target these pathways are cur- Note: Supplementary data for this article are available at Cancer Discovery rently in preclinical and clinical development for EC (15 , Online (http://www.cancerdiscovery.aacrjournals.org). 16 ). Achievement of optimal therapeutic benefit requires Corresponding Author : Lydia W.T. Cheung, Department of Systems Biology, identification of patients likely to benefit combined with The University of Texas MD Anderson Cancer Center, 7435 Fannin Street, rational combinatorial therapy, such as co-targeting of the Suite 2SCR3.2211, Houston, TX 77054-1942. Phone: 713-563-0431; Fax: PI3K and RAS pathways. In this study, we first performed 713-563-4235; E-mail: [email protected] a comprehensive mutational analysis of candidate genes doi: 10.1158/2159-8290.CD-11-0039 in 243 well-characterized endometrial tumors. Frequent ©2011 American Association for Cancer Research. anomalies were found in multiple members of the PI3K

JULY 2011 CANCER DISCOVERY | OF2

research article Cheung et al. pathway and KRAS, and the effects of these mutations on RESULTS downstream signaling were systematically investigated. Overview of Nonsynonymous Mutations in ec Second, we determined whether aberrations in the PI3K and MAPK pathways dictate sensitivity toward drugs target- In terms of nonsynonymous somatic mutations in 243 ing these pathways. Third, we show for the first time that endometrial tumors (see Methods), PTEN was the most fre- PIK3R1 and PIK3R2, the genes encoding p85α and p85β, are quently mutated gene (108 tumors, 44%) followed by PIK3CA frequently mutated in EC and novel activating mutations (97 tumors, 40%) (see Fig. 1A and Supplementary Table S1). were identified. Finally, we describe a mechanism defined by Moreover, 24 (10%), 39 (16%), and 35 (14%) samples had one of the PIK3R1 gain of function mutations, which desta- FGFR2, CTNNB1, and TP53 mutations, respectively. PIK3R1 bilizes PTEN through disruption of p85α homodimeriza- mutations were detected at a higher frequency (48 tumors, tion, offering evidence of the importance of PTEN and p85 20%) than in any other cancer lineage (17, 18). Moreover, interactions in human cancer. we found PIK3R2 mutations, which had not been previously

All endometrial cancer tumors (n=243)

Endometrioid grade 1 and grade 2 (n=132) Endometrioid grade 3 (n=29) Endometrioid grade 1 and grade 2 (n=132) Endometrioid grade 3 (n=29) c Endometrioid grade 1 andD grade 2 (n=132) Endometrioid grade 3 (n=29)

Mixed endometrial (n=60) Malignant mixed mullerian tumors (n=18) Mixed endometrial (n=60) Malignant mixed mullerian tumors (n=18) Figure 1. Mutation diagrams: a, the full set of endometrial tumor samples (n = 243); B, endometrioid grade 1 and grade 2 (n = 132; left), endometrioid grade 3 (n = 29; right); c, mixed endometrial (n = 60); and D, MMMT (n = 18). Each column represents a tumor and each row corresponds to aMixed single gene. endometrial (n=60) Malignant mixed mullerian tumors (n=18)

OF3 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article reported at a significant frequency in any tumor lineage, in mutations, and significantly more frequent TP53 mutations 12 (5%) endometrial tumors, establishing PIK3R2 as a can- compared to EEC (P = 0.001). cer-associated gene. A mass spectroscopy–based analysis (MassARRAY, Sequenom) revealed that 43 (18%) and 2 (1%) PIK3CA andPIK3R1 Mutations Frequently coexist of the tumors carried mutations in KRAS and AKT1 , respec- with PTEN heterozygous Mutation tively, whereas BRAF, AKT2 , and AKT3 hotspot mutations Of particular interest, in contrast to other tumor lineages were not detected (0%). The majority of the mutations, in- ( 19 , 20 ), mutations in PIK3CA , PIK3R1 , PIK3R2 , KRAS , and cluding those in PTEN , were heterozygous. PTEN were not mutually exclusive ( Fig. 1 ). To examine the Expression data from reverse-phase protein array (RPPA) patterns and frequencies of these co-mutations in detail, we were used to impute PTEN levels where tumor slides for im- combined data from EEC grades 1, 2, 3, and mixed carcino- munohistochemistry (IHC) analysis were unavailable (53 cases) mas to increase statistical power (n = 221, Table 1). KRAS or the staining was heterogeneous (35 cases). Notably, where mutations frequently coexisted with PTEN (7%), PIK3CA PTEN expression data were available from both IHC and (8%), or PIK3R1/PIK3R2 (5%). The co-mutations occurred at RPPA, they were concordant in 177 of 190 cases assessed (93%) frequencies expected based on the frequency of each indepen- (Supplementary Fig. S1), suggesting that RPPA allows reliable dent mutation (Supplementary Table S2). characterization of PTEN protein levels. Absence of PTEN pro- Our data showed that only 10 of 106 (9%) PTEN mutations tein was observed in 119 of 243 (49%) tumors. are homozygous with all exhibiting complete loss of PTEN The mutation spectrum was similar for tumors defined protein (Table 1). Five (50%) of these tumors harbored PTEN as EEC and mixed endometrioid and serous carcinomas mutation alone while 5 (50%) had a concomitant PI3K path- ( Fig. 1B–D and Supplementary Table S1). In contrast, ma- way mutation. In contrast, the majority of the heterozygous lignant mixed mullerian tumors (MMMT) displayed mark- PTEN-mutant tumors (92%) had coincident mutations in edly fewer mutations in the PI3K pathway, no CTNNB1 other pathway genes. Strikingly, heterozygous PTEN mutation

Table 1. summary of mutation patterns in the Pi3K pathway and KRAS in 221 endometrial cancers

PTEN protein Percentage MMR deficienta Positive Negative all wild type 26.24% 15.38% 10.86% 8/43 (18.60%) PteN2 /2 2.26% 0.00% 2.26% 0/5 (0%) PteN1 /2 6.79% 3.17% 3.62% 2/10 (20%) PIK3CA 8.60% 3.17% 5.43% 3/16 (18.75%) PIK3R1/PIK3R2 4.07% 1.36% 2.71% 2/8 (25%) PIK3CA 1 PIK3R1/PIK3R2 0.90% 0.45% 0.45% 0/2 (0%) PTEN 2/2 1 PIK3CA 1.36% 0.00% 1.36% 1/3 (33.33%) PTEN 1/2 1 PIK3CA 14.93% 9.05% 5.88% 9/30 (30%) PTEN 2/2 1 PIK3R1/PIK3R2 0.45% 0.00% 0.45% 0/1 (0%) PTEN 1/2 1 PIK3R1/PIK3R2 7.24% 4.98% 2.26% 3/15 (20%) PTEN 2/2 1 PIK3CA 1 PIK3R1/PIK3R2 0.45% 0.00% 0.45% 0/1 (0%) PTEN 1/2 1 PIK3CA 1 PIK3R1/PIK3R1 7.69% 3.17% 4.52% 5/17 (29.41%) KRAS 4.98% 1.81% 3.17% 1/9 (11.11%) K R A S1 PTEN 1/2 1.81% 0.90% 0.90% 2/4 (50%) KRAS 1 PIK3CA 3.62% 1.81% 1.81% 1/8 (12.5%) K R A S1 PIK3R1/PIK3R2 2.71% 0.90% 1.81% 1/5 (20%) KRAS 1 PTEN 1/2 1 PIK3CA 3.17% 1.36% 1.81% 3/6 (50%) K R A S1 PTEN 1/2 1 PIK3R1/PIK3R2 0.90% 0.45% 0.45% 1/2 (50%) K R A S1 PTEN 1/2 1 PIK3CA 1 PIK3R1/PIK3R1 0.90% 0.45% 0.45% 1/1 (100%) AKT 0.90% 0.90% 0.00% 0/2 (0%) Abbreviations: MMR, mismatch repair; PTEN +/−, PTEN heterozygous mutation; PTEN −/−, PTEN homozygous mutation. a 188 endometrial tumor slides were available and were stained for the MMR proteins

JULY 2011 CANCER DISCOVERY | OF4

research article Cheung et al. frequently co-occurred with PIK3CA (59/96; 61%) or PIK3R1/ determined by RPPA. Based on the apparent differential PIK3R2 (37/96; 39%) mutations (Table 1). The frequencies of effects of PI3K pathway and KRAS mutations postulated these co-mutations were higher than expected due to chance, from the co-mutations, RPPA was assessed in four tumor suggesting mutual inclusivity, in contrast to co-mutations be- groups: (1) wild-type (WT) in both PI3K pathway and KRAS, tween homozygous PTEN and PIK3CA or PIK3R1/PIK3R2 that (2) PI3K pathway aberrations without KRAS mutation, (3) were present at expected frequencies (Supplementary Table S2). PI3K pathway aberrations with KRAS mutation, and (4) Moreover, 7 of 15 (47%) tumors carrying only heterozygous KRAS mutation only. Levels of phosphorylated AKT (pAKT) PTEN mutation showed retained PTEN expression, compared at Thr308 and Ser473, but not total AKT, were markedly higher to tumors with concomitant mutations in PIK3CA (20/33; in tumors with PI3K pathway aberrations (Supplementary 61%) or PIK3R1/PIK3R2 (11/16; 69%). PTEN protein loss was Fig. S3A). KRAS mutation was not associated with changes in significantly less frequent in tumors with co-mutations in ad- pAKT. Instead, KRAS mutation was associated with increased ditional pathway members (P = 0.04). This observation led phosphorylation of mitogen-activated protein kinase kinase us to hypothesize that PI3KCA, PIK3R1, or PIK3R2 mutations (MEK)1/2 at Ser217/221, ERK1/2 at Thr202/Tyr204, and p38 might be co-selected with heterozygous PTEN mutations to MAPK at Thr180/Tyr182 without significant change in their re- compensate for incomplete loss of PTEN protein. spective total proteins (Supplementary Fig. S3B). Activation of the MAPK pathway was not seen in PI3K pathway mu- Deficient Mismatch repair May contribute to the tant tumors, suggesting that distinct downstream signaling high Frequency of co-mutations events are activated in PI3K pathway- and KRAS-mutant EC. Defects in the DNA mismatch repair (MMR) pathway PTEN protein loss, regardless of whether PTEN was mu- are common in EC, leading to genomic instability (21). tant or WT, was associated with a consistent increase in Consistent with previous studies, 43 of 188 (23%) tumors were pAKT indicative of pathway activation (Supplementary MMR-deficient as indicated by loss on IHC of MLH1, MSH2, Fig. S3C, left). The effect of PTEN loss on pathway activation or MSH6 (Table 1). Intriguingly, we found MMR deficiency to was dominant because PIK3CA, PIK3R1, or PIK3R2 mutation be significantly more frequent in tumors carrying 2 or 3 PI3K in addition to PTEN loss did not result in a further increase pathway mutations than tumors with 0 or 1 mutation (29% in pAKT (Supplementary Fig. S3C, left). PTEN heterozygous and 32% versus 17% and 19%, respectively; P < 0.05). This mutation alone where PTEN protein was retained was not finding suggested that MMR deficiency might partly contrib- associated with increased pAKT compared to tumors bearing ute to the high frequency of co-mutations in the PI3K path- WT PTEN (Supplementary Fig. S3C, right). way in EC. However, the mutations in PI3K pathway members Strikingly, in tumors where PTEN protein was retained, were not classic MMR-related aberrations because they did including tumors with PTEN heterozygous mutation, mu- not occur in regions of nucleotide repeats. tation in other PI3K pathway members, including PIK3CA, PIK3R1, or PIK3R2, resulted in marked increases in pAKT Distribution of PTEN, PIK3CA, PIK3R1, and PIK3R2 levels comparable to that observed in tumors where PTEN is Mutations lost (Fig. 3A and 3B). Furthermore, similar changes in stathmin, caveolin 1, and insulin growth factor binding protein Approximately 40% and 60% of PTEN mutations in the 2 (IGFBP2) were observed in tumors with either PTEN loss endometrial tumor set occur in the C2 and phosphatase do- or PIK3CA, PIK3R1, or PIK3R2 mutations (Fig. 3C). Thus it mains, respectively (Supplementary Fig. S2). Of 114 PIK3CA appears that PIK3CA, PIK3R1, or PIK3R2 mutation results mutations, 20% were in the adaptor-binding domain (ABD), in a phenocopy of PTEN loss in terms of pathway activity. 17% in the C2 domain, 28% in the helical domain, and 24% These overlapping effects on cellular signaling were also ob- were in the kinase domain (Fig. 2A). This distribution is dis- served in unsupervised hierarchical clustering of the RPPA tinct from that observed in other cancers in which the vast data, as tumors with PTEN loss intermingled with tumors majority of mutations cluster within the helical and kinase carrying PIK3CA, PIK3R1, or PIK3R2 mutations (Fig. 3A). No domains (22, 23). All substitutions in PIK3R1 were somatic significant difference in the pAKT level among tumors car- except one known germline single-nucleotide polymorphism rying PIK3CA, PIK3R1, or PIK3R2 mutations or among tu- (SNP) (M326I). A significant proportion of substitutions mors carrying mutations in different domains of PIK3CA or (12/29; 41%) and indels (37/42; 88%) were located within the PIK3R1 (Supplementary Fig. S3D) was detected, suggesting iSH2 domain (Fig. 2B), supporting this region as a muta- similar effects on cellular function in EC. Together, these tional hotspot (18). We also detected mutations in all other data support the functional consequence of PIK3CA, PIK3R1, PIK3R1 domains, albeit at lower frequencies. Unlike PIK3R1, or PIK3R2 mutations on PI3K pathway activation, which is all PIK3R2 mutations were substitutions with no apparent manifest selectively when PTEN is heterozygously mutated hotspot region (Fig. 2C). Two alterations observed at the and PTEN protein is retained. same amino acid, A727T and A727V, were confirmed to be Phosphorylation of MEK1/2, ERK1/2, and p38 MAPK germline SNPs. was higher in KRAS-mutant tumors (Fig. 3D). Strikingly, as with the lack of effect of KRAS mutations on PI3K pathway Pi3K Mutations Functionally Mimic PteN loss signaling, there was no interaction between the effects of The functional impact of PI3K anomalies and the inter- PTEN loss, PIK3CA, PIK3R1, or PIK3R2 mutations and the action between KRAS and PI3K pathway aberration in EC effects of KRAS mutation on the MAPK pathway (Fig. 3A). are not well characterized. Therefore, the consequences of Together, the signaling data suggest that EC can be ro - these genetic aberrations on downstream signaling were bustly classified as either (1) WT PI3K and PTEN-retained,

OF5 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article

a PIK3CA R38S (2) E39K (1) R88Q (18) R93W (2) L339I (1) V344M (3) D350G (2) R357Q (2) G364R (1) E365K (2) C378R (1) I406V (1) C407Y (1) C420R (3) E453K (1) P471L (1) E542K/V (5) E545A/K/Q (20) Q546R/H/K (7) H665Q (1) H701P (1) R818C (1) K884R (1) F909L (1) R916C (1) D926N (1) M1043V (3) H1047L/R (19) H1065L (1) R38S (2) E39K (1) R88Q (18) R93W (2) L339I (1) V344M (3) D350G (2) R357Q (2) G364R (1) E365K (2) C378R (1) I406V (1) C407Y (1) C420R (3) E453K (1) P471L (1) E542K/V (5) E545A/K/Q (20) Q546R/H/K (7) H665Q (1) H701P (1) R818C (1) K884R (1) F909L (1) R916C (1) D926N (1) M1043V (3) H1047L/R (19) H1065L (1) R38S (2) E39K (1) R88Q (18) R93W (2) L339I (1) V344M (3) D350G (2) R357Q (2) G364R (1) E365K (2) C378R (1) I406V (1) C407Y (1) C420R (3) E453K (1) P471L (1) E542K/V (5) E545A/K/Q (20) Q546R/H/K (7) H665Q (1) H701P (1) R818C (1) K884R (1) F909L (1) R916C (1) D926N (1) M1043V (3) H1047L/R (19) H1065L (1) E110K (3) K111N (1) G118D (6) V146I (1) Y165H (1) D300Y (1) E110K (3) K111N (1) G118D (6) V146I (1) Y165H (1) D300Y (1) E110K (3) K111N (1) G118D (6) V146I (1) Y165H (1) D300Y (1)

ABD (19%) RBD (0%) C2 (18%) HELICAL (28%) KINASE (25%)

1 108 191 291 330 480 525 696 1068 K944fs(1) K944fs(1) K944fs(1) E453deI (2) E453deI (2) E453deI (2)

B PIK3R1 L30F (1) F69L (1) I338V (1) R348* (5) G353R (1) L449I (1) E458* (1) R461* (1) E468* (1) R503W (2) I521M (1) D548Q (1) R557* (1) K567E (2) S608* (1) R639P (1) R642* (1) L30F (1) F69L (1) I338V (1) R348* (5) G353R (1) L449I (1) E458* (1) R461* (1) E468* (1) R503W (2) I521M (1) D548Q (1) R557* (1) K567E (2) S608* (1) R639P (1) R642* (1) L30F (1) F69L (1) I338V (1) R348* (5) G353R (1) L449I (1) E458* (1) R461* (1) E468* (1) R503W (2) I521M (1) D548Q (1) R557* (1) K567E (2) S608* (1) R639P (1) R642* (1) E217K (1) I177N (1) E160* (1) E217K (1) I177N (1) E160* (1) E217K (1) I177N (1) E160* (1) I82F (1) R724* (1) M326I (85) E297K (1) I82F (1) R724* (1) M326I (85) E297K (1) I82F (1) M326I (85) R724* (1) E297K (1) SH3 (3%) Rho-GAP (6%) nSH2 (15%) iSH2 (66%) cSH2 (6%)

6 78 126 278 331 414 226 407 l571fs (1) l571fs (1) l571fs (1) E596fs (1) R503fs (1) R574fs (1) R577fs (1) Y657fs (1) K575fs (1) K142fs (1) E596fs (1) R503fs (1) R574fs (1) R577fs (1) Y657fs (1) K575fs (1) K142fs (1) N673fs (1) D605fs (1) D578fs (2) Q415fs (1) E596fs (1) R503fs (1) R574fs (1) R577fs (1) Y657fs (1) K575fs (1) K142fs (1) N673fs (1) D605fs (1) D578fs (2) Q415fs (1) M582fs (1) N673fs (1) D605fs (1) D578fs (2) Q415fs (1) M582fs (1) I566lnsl (1) M582fs (1) I566lnsl (1) L449del (1) E439del (2) R574del (1) K382del (1) L449del (1) I566lnsl (1) E439del (2) R574del (1) K382del (1) H450del (2) N453del (1) L449del (1) E439del (2) R574del (1) K382del (1) H450del (2) N453del (1) H450del (2) N453del (1) EY451del (1) EY451del (1) HE450del (1) LI380deIF (1) RD577del (1) EY451del (1) HE450del (1) LI380deIF (1) RD577del (1) HE450del (1) LI380deIF (1) RD577del (1) TRD576del (1) Q552lnsKQ (1) TRD576del (1) RDQ577del (1) Q552lnsKQ (1) TRD576del (1) RDQ577del (1) Q552lnsKQ (1) RDQ577del (1) EEYT467delS (1) EEYT467delS (1) EEYT467delS (1) RLYEE465delK (1) RLYEE465delK (1) RLYEE465delK (1) LTFSSVVE396del (1) LTFSSVVE396del (1) LTFSSVVE396del (1) IEAVGKKLHEY442del (2) IEAVGKKLHEY442del (2) IEAVGKKLHEY442del (2) EAVGKKLHEY443delD (2) EAVGKKLHEY443delD (2) EAVGKKLHEY443delD (2) P568lnsALIQLRKTRDQYLMKP(1) P568lnsALIQLRKTRDQYLMKP(1) P568lnsALIQLRKTRDQYLMKP(1)

c PIK3R2 A36E(1) A415T (1) R35Q(1) S273C (1) G373R (1) K376N (1) N403I (1) A533V (1) N561D (1) A171V (1) A36E(1) A415T (1) R35Q(1) S273C (1) G373R (1) K376N (1) N403I (1) A533V (1) N561D (1) A171V (1) A36E(1) A415T (1) R35Q(1) S273C (1) G373R (1) K376N (1) N403I (1) A533V (1) N561D (1) A171V (1) A727V/T (4/22) P323S (1) F118L(1) A727V/T (4/22) P323S (1) F118L(1) A727V/T (4/22) P323S (1) F118L(1) SH3 (17%) Rho-GAP (17%) nSH2 (25%) iSH2 (25%) cSH2 (0%)

7 79 122 292 328 411 620 702

Figure 2. Distribution of nonsynonymous mutations in a, PIK3CA, B, PIK3R1, and c, PIK3R2. Substitutions (top) and indels (bottom) are graphed with the amino acid indicated and the number of independent observations within parentheses. The percentage of mutations identified in each domain is shown in the box. p110α is characterized by 5 domains: ABD, Ras-binding domain (RBD), C2 domain, helical domain, and kinase catalytic domain. p85 consists of SH3, Rho-GAP, and two SH2 domains (nSH2 and cSH2) that flank an intervening domain (iSH2). *, nonsense mutation; bold, mutations confirmed somatic; red, germline SNPs. or (2) PTEN-lost or PTEN-retained with PIK3CA, PIK3R1, Mutation status Predicts sensitivity to mtOr and or PIK3R2 mutated. The independent effects of KRAS mu- MeK inhibitors tations in each of the subsets suggest that there are 4 To determine if mutation status can predict sensitiv- independent groups based on PI3K pathway status and ity to PI3K pathway or MEK inhibitors, we obtained 16 EC KRAS mutation. cell lines and characterized mutation and signaling status

JULY 2011 CANCER DISCOVERY | OF6

research article Cheung et al.

a B ANOVA P=0.001 1 2 ANOVA P=0.001 2 ANOVA P=0.001 1 PI3K WT, PTEN retained, KRAS WT 2 1 1 0 AktpS47 3 AktpT30 8 Caveolin 1 PTEN pERK1/2 p-MEK1/ 2 p-p38 MAPK IGFBP2 Stathmin PI3K WT, PTEN retained, KRAS WT PI3K WT, PTEN retained, KRAS Mut 1 0 AktpS47 3 AktpT30 8 Caveolin 1 PTEN pERK1/2 p-MEK1/ 2 p-p38 MAPK IGFBP2 Stathmin PI3K WT, PTEN retained, KRAS WT 1 308 PI3K WT, PTEN retained, KRAS Mut 473 0 AktpS47 3 AktpT30 8 Caveolin 1 PTEN pERK1/2 p-MEK1/ 2 p-p38 MAPK IGFBP2 Stathmin PI3K WT, PTEN loss, KRAS WT PI3K WT, PTEN retained, KRAS Mut 0 308 473 –1 PI3K WT,PI3K PTEN WT, loss, PTEN KRAS loss, WT KRAS Mut 0 308 PI3K WT, PTEN loss, KRAS WT 473 PI3K WT, PTEN loss, KRAS Mut 0 –1 Akt pS PI3K WT,Mut, PTEN PTEN loss, retained, KRAS KRASMut WT Akt pT –1 –1 Akt pS PI3K Mut,PI3K PTEN Mut, retained, PTEN retained, KRAS WT KRAS Mut Akt pT –1 –2 Akt pS PI3K Mut, PTEN retained, KRAS WT Akt pT PI3K Mut, PTEN retained, KRAS Mut –1 –2 PI3K Mut, PTEN retained, KRAS Mut –2 –2 –2 –2 –3 –3 –3 KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut PI3K PTENWT retainedPTEN loss PI3K MutPTEN retainedPI3K PTENWT retainedPTEN loss PI3K MutPTEN retained PTEN retainedPI3K WT PTENPTEN loss retainedPI3K MutPTEN retainedPI3K WT PTENPTEN loss retainedPI3K Mut 1.0 ANOVA P=0.02 ANOVA P=0.04 ANOVA P=0018 PTEN retained4 PTEN retained PTEN retained PTEN retained 1.0 ANOVA P=0018 ANOVA P=0.02 2 ANOVA P=0.04 0.51.0 4 ANOVA P=0.02 ANOVA P=0.04 c ANOVA P=0018 34 2 0.5 2 0.00.5 3 23 ANOVA P=0.001 0.0 2 1 –0.50.0 2 1 12 1

–0.5 IFGBP2 Stathmin PI3K WT, PTEN retained, KRAS WT Caveolin 1 –1.0–0.5 1 1 0 IFGBP2 AktpS47 3 AktpT30 8 Caveolin 1 PTEN pERK1/2 p-MEK1/ 2 p-p38 MAPK IGFBP2 Stathmin 0 Stathmin PI3K WT, PTEN retained, KRAS Mut Caveolin 1 01

–1.0 IFGBP2 Stathmin –1.0 0 Caveolin 1 0 308 –1.5 473 PI3K WT, PTEN loss, KRAS WT –100 0 –1.5 –1 PI3K WT, PTEN–2.0–1.5 loss, KRAS Mut –1 –1 –2–1 –1 Akt pS PI3K Mut,–2.0 PTEN retained, KRAS WT Akt pT –1 –2.0 –2 –1 PI3K Mut, PTEN retained, KRAS Mut –2 –2 KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS–2 WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRASPI3K–3 Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut PI3K PTENWT retainedPTEN loss PI3K MutPTEN retained PI3K PTENWT retainedPTEN loss PI3K MutPTEN retainedPI3K PTENWT retainedPTEN loss PI3K MutPTEN retained PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PTEN retained 1.0 PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTENKRAS retained WTKRASANOVA MutKRAS P=0.037 WTKRASPTEN MutKRAS retained WTKRAS Mut PTENKRAS retained WTKRAS MutKRAS WTKRASPTEN MutKRAS retained WTKRAS Mut 1.0 1.0 PI3KANOVA WT3 P=0.037PTEN loss PI3K Mut PI3K WT1 PTEN loss PI3K Mut 0.5 ANOVA P=0.037 4 PTEN3 retained PTEN retained PTEN 182 retained PTEN retained 0.5 1 20 3 /Y 4 2 182 1 1.00.5 0 ANOVA P=0.04 4 /Y ANOVA P=0.02 182 D 0.0 ANOVA P=001820 0 2 4 /Y 217/221 18 0 20 /Y 0.0 /Y 2 0 2 202 0

217/221 1 0.5 18 0.0 /Y 3 0 217/221 18 –0.5 202 1 –1 –0.5 0.0 202 01 2 –1 –1.0–0.5 0 –11

–1.0 MEK1/2 pS –0.5 –10 –2 ERK1/2 pT 1 MEK1/2 pS IFGBP2 –1.0 –2 p38 MAPK pT Stathmin –1 Caveolin 1

–1.5 ERK1/2 pT MEK1/2 pS –1.0 –2

–1 p38 MAPK pT 0 –1.5 ANOVA P=0.05 ERK1/2 pT –20 –3 ANOVA P=0.05 –1.5 p38 MAPK pT –1.5 ANOVA P=0.05 –2 –3 ANOVA P=0.05 ANOVA P=0.05 –2–1 –3 ANOVA P=0.05 –2.0 –1 KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS–2 WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRASPI3K Mut WTKRAS WTKRASPTEN MutlossKRAS WTKRASPI3K Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut PI3K PTENWT retainedPTEN loss PI3K MutPTEN retained PI3K PTENWT retainedPTEN loss PI3K MutPTEN retainedPI3K PTENWT retainedPTEN loss PI3K MutPTEN retained PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PTEN retainedKRAS WTKRAS MutKRASPTEN WTKRAS retained MutKRAS WTKRASPTEN Mut retainedKRAS WTKRAS MutKRASPTEN WTKRAS retained MutKRAS WTKRASPTEN Mut retainedKRAS WTKRAS MutKRASPTEN WTKRAS retained MutKRAS WTKRAS Mut PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained 1.0 ANOVA P=0.037 3 0.5 1 4 182 20 2 /Y

Figure 3. PI3K mutation in tumors where PTEN protein is retained mimics the effect of PTEN loss on downstream signaling0 regardless of KRAS 0.0 /Y 0 217/221 status. a, heat map of unsupervised cluster analysis of proteins and samples by RPPA. Proteins are listed across the top18 of the heat map and 202 1 mutational status of the tumor is represented at the–0.5 right. Red, higher expression; green, low relative to the other samples.–1 Expression levels of B, phosphorylated AKT at Thr 308 (left) and Ser473 (right); c, stathmin (left), caveolin 1 (center),0 insulin-like growth factor binding protein 2 (IGFBP2; 217/221 –1.0 202 204 180 182 right); D, phosphorylated MEK1/2 at Ser (left),MEK1/2 pS ERK1/2 at Thr /Tyr (center), and–1 p38 MAPK at Thr /Tyr (right)–2 were logarithmically ERK1/2 pT

converted, normalized by mean, and presented on the–1.5 Y axis. The boxes represent the distribution of individual values fromp38 MAPK pT the lower 25th percentile to upper 75th percentile; solid line in the middle, median value;ANOVA P=0.05 lower and upper–2 whiskers, 5th and 95th percentiles.–3 WT, wild-type;ANOVA Mut, P=0.05 mutated. KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut KRAS WTKRAS MutKRAS WTKRAS MutKRAS WTKRAS Mut PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PI3K WT PTEN loss PI3K Mut PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained PTEN retained (Supplementary Table S3). Six of seven (86%) PTEN hetero- needed to achieve 50% growth inhibition) could not be ob- zygous mutations co-occurred with PIK3CA or PIK3R1 or tained. In contrast, cells with PI3K pathway mutations PIK3R2 mutations. As in patient samples, KRAS mutation without KRAS mutations were markedly more sensitive to was not mutually exclusive with PI3K pathway mutations. rapamycin. Notably, higher GI50 values were obtained in cell Loss of PTEN protein was observed in 5 cell lines. More lines with concomitant KRAS and PI3K pathway mutations importantly, these 5 cell lines had relatively high pAKT and than with PI3K pathway mutations alone. There was no obvi- low caveolin 1 levels as predicted from the patient RPPA ous correlation between mutational status and responsive- data (Supplementary Fig. S4A). Furthermore, Western blot ness to GDC-0941, a selective class I PI3K inhibitor that is (WB) and RPPA showed that PIK3CA, PIK3R1, or PIK3R2 currently in clinical trials, under the culture conditions as- mutations in cells with retained PTEN produced a pheno- sessed (Supplementary Fig. S5B). copy of PTEN loss in terms of activation and expression of All cell lines harboring KRAS mutations had concurrent the proteins altered in the patient samples (Supplementary PI3K pathway mutations and were relatively sensitive to MEK Fig. S4). inhibition (Supplementary Fig. S5C). However, although Cell lines with WT PI3K pathway members were resistant cell lines with aberrations in the PI3K pathway tended to to the mTOR inhibitor rapamycin (Supplementary Fig. S5A). be resistant to MEK inhibition, some of the cell lines that Three cell lines, two of which carry both KRAS and PIK3CA lacked KRAS mutation were sensitive to MEK inhibition. mutations, were very resistant and a GI50 value (concentration Thus, KRAS mutation is likely a sufficient but not obligatory

OF7 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article sensitivity marker for MEK inhibitors, which may be useful in in PTEN levels, with diminished pAKT in parallel (Fig. 4B, patients with KRAS-mutant EC. left). The effects of WT p85α on PTEN were confirmed with 3 independent constructs. WT p85β did not alter PTEN or identification of activating Mutations in PIK3R1 pAKT levels (Fig. 4B, right), suggesting that the two p85 iso- and PIK3R2 forms display different functional consequences in EC. These The majority of PIK3R1 mutations found in EC are novel observations were confirmed with multiple cell lines includ- and frequent PIK3R2 mutation has not been reported pre- ing EFE184, Ba/F3, and HEK293FT (data not shown). viously in any tumor lineage. We therefore investigated These changes in PTEN expression are unlikely to have the functional significance of a subset of these mutations been caused by transcriptional regulation, because PTEN by assessing their ability to induce interleukin-3 (IL-3)– mRNA levels were not altered (data not shown). Rather, independent survival of Ba/F3 cells, an assay that has been cycloheximide-based chase studies demonstrated that WT used previously to characterize mutations in PI3K pathway p85α significantly extended the half-life of PTEN protein members (18). Three PIK3R1 indels identified in our dataset (Fig. 5A, left). In contrast, PTEN protein stability in cells (E439del, R574fs, and T576del) were previously character- transfected with E160* was substantially reduced. Because ized and R574fs and T576del have been shown to be onco- PTEN protein can be regulated via the ubiquitin-depen- genic (24). Concordantly, we found that these two mutants dent proteasome pathway (25), we treated cells transfected were sufficient to convert Ba/F3 cells to IL-3 independence with WT p85α or E160* with the proteasome inhibitor (Fig. 4A, left). In addition, two truncation mutations—E160* MG132. PTEN protein levels were modestly increased by in the Rho-GAP domain that retains the src-homology 3 proteasome inhibition (Fig. 5A, center). Strikingly, E160* (SH3) and N terminal proline-rich domain and, with less effi- enhanced PTEN ubiquitination, whereas WT p85α de- ciency, R348* in the nSH2 domain that retains the SH3, Rho- creased the abundance of PTEN-ubiquitin conjugates (Fig. GAP, and two proline-rich domains—and the point mutation 5A, right). These results demonstrated that the ubiquitin R503W in iSH2 conferred IL-3–independent growth to Ba/F3 degradative pathway likely contributes to the ability of cells compared with WT p85α or LacZ control. The germline p85α to regulate PTEN protein. SNP, M326I, had no effect on Ba/F3 survival. It is worth not- Recent studies have demonstrated the ability of the SH3 ing that an F268I mutation that was seen in the initial muta- and RhoGAP domains of p85α to bind PTEN (26). As in- tion screen but not confirmed in orthologous mutational dicated in Fig. 5B (left), WT and the other p85α mutants analysis had no effect on Ba/F3 survival, further confirming directly interact with PTEN, but not E160*, which lacks the the validity of the assay. RhoGap domain but retains the BH3 and N terminal proline- We report for the first time herein activating mutations rich domain. These observations were confirmed by recipro- in PIK3R2 including A171V in the Rho-GAP domain and cal IP (data not shown). Intriguingly, the binding between N561D in the iSH2 domain (Fig. 4A, right). The germline PTEN and WT p85α was decreased by transfection of in- SNP, V727T, had no impact on Ba/F3 survival. WT and creasing amounts of E160* (Fig. 5B, center). The increase in E160* PIK3R1 were assessed in parallel, demonstrating that PTEN protein expression and stability induced by WT p85α PIK3R2 may have greater activity than PIK3R1 in increasing was also inhibited by co-expression of E160* (Fig. 5B, right). viability of Ba/F3 cells (Fig. 4A, right, gray bars). These data suggest that E160* competitively inhibits interactions between PTEN and WT p85α, and this disruption PIK3R1 or PIK3R2 activating Mutations increase is associated with PTEN destabilization. Additionally, con- aKt Phosphorylation sistent with previous studies (27), we confirmed that WT We sought to understand the mechanism underlying the ac- p85α forms homodimers and this homodimer formation was tivity of mutant PIK3R1 and PIK3R2 by examining their effects decreased in the presence of E160* (Fig. 5C, left and cen- on downstream signaling. The PIK3R1 or PIK3R2 mutants that ter). In parallel, we observed that E160* is able to interact potentiated Ba/F3 cell proliferation increased AKT phosphor- with WT p85α (Fig. 5C, right). These data suggested that ho- ylation in the HEC1A EC cell line that contains a G1049R modimerization of WT p85α is inhibited by E160* likely by PIK3CA and a G12D KRAS mutation (Fig. 4B; Supplementary formation of WT p85α–E160* dimers. This process is associ- Fig. S6), whereas mutants that had no effect on Ba/F3 sur- ated with diminished binding between WT p85α and PTEN. vival did not increase pAKT. Indeed, pAKT levels significantly Furthermore, overexpression of p110α disrupted the interac- correlated with IL-3-independent survival (P = 0.001). There tion between WT p85α and E160*, formation of WT p85α was no observable change in the expression of p110α in cells homodimer and binding of PTEN to p85α (Fig. 5D), sug- transfected with any of the mutants. Because p85 binds and gesting that binding of p110α and PTEN to p85α is mutually stabilizes p110, this suggests that p85α is in excess of p110α exclusive and that p110α does not bind to the PTEN–p85α in HEC1A cells. Immunoprecipitation (IP) experiments dem- homodimer complex. onstrated that the majority of the mutants retained the ability to bind p110 , except E160* and R348*, which lack the iSH2 α DISCUSSION domain implicated in p110α binding (Fig. 4C). Characterization of mutations in cancer cells provides in- the e160* p85a Mutant Destabilizes PteN sights into tumorigenesis and reveals candidates for targeted Strikingly, the E160* p85α mutant induced a marked de- therapeutics. Most previous studies of genetic aberrations in crease in PTEN protein levels with a concordant increase in EC are small with analysis of a limited number of candidate pAKT. In contrast, expression of WT p85α led to an increase genes. The functional relevance of these alterations in EC is

JULY 2011 CANCER DISCOVERY | OF8

research article Cheung et al. a p85a p85b b p856p85aa p852.5p85b * * 5 6 6 2.52.52 * * * * * * 5 4 * 5 1.52 * * 2 * * 4 34 * * * * 1.51.51 2 3 3 * * * * 0.51 1 Relative cell viability 1 2 2 Relative cell viability 0 0.50.50 Relative cell viability 1 Relative cell viability 1 Relative cell viability Relative cell viability

0 0 0 0

p85a p85b B p85p85aa p85p85bb

95kDa b 72kDa p85 95kDa 55kDa95kDa b 95kDa 72kDa p8p85 5b p85a 72kDa 72kDa 36kDa 55kDa55kDa p85a 95kDa55kDa95kDa a 28kDa p85p85a 72kDa36kDa 36kDa17kDa 72kDa 36kDa p8p85a5a 55kDa28kDa55kDa 28kDa28kDa 17kDa p85b 36kDa 17kDa17kDa 36kDa p110a 28kDa28kDa p85p110p85bab 17kDa17kDa PTENa PTENa p110p110a p110p110a 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 2.3 0.2 1.0 2.3 0.3 1.4 1.2 1.2 1.1 1.3 1.2 PTENPTEN PTENbPTEN-actin b-actin 1.01.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 2.3 2.3 0.2 0.2 1.01.0 2.3 2.3 0.3 0.3 1.4 1.4 1.2 1.2 1.2 1.2 1.1 1.1 1.3 1.3 1.2 1.2 308 308 Aktb-actinb-actinpT Aktb-actib-actipTn n 1.0 0.2 1.9 1.0 1.0 1.0 1.9 1.8 1.0 1.0 1.0 1.0 2.2 1.0 1.0 1.0 2.2 1.0 0.3 2.4 308 308308 Akt pT 473038 AktAktpTpT AktAktpSpT AktpS473 1.0 0.2 1.9 1.0 1.0 1.0 1.9 1.8 1.0 1.01.0 1.0 1.0 1.0 1.0 2.2 2.2 1.0 1.0 1.0 1.0 1.0 1.0 2.2 2.2 1.0 1.0 0.3 0.3 2.4 2.4 1.01.0 0.30.2 2.31.9 1.11.0 1.31.0 1.01.0 1.91.9 2.11.8 1.21.0 1.0 1.0 1.0 2.3 1.0 1.1 1.0 2.2 1.0 0.3 2.4 473 Akt pS 473 473473 Akt pSAkt AkAktptpSSAkt 1.01.0 0.3 0.3 2.3 2.3 1.1 1.1 1.3 1.3 1.0 1.0 1.9 1.9 2.1 2.1 1.2 1.2 1.01.0 1.0 1.0 1.0 1.0 2.3 2.3 1.0 1.0 1.1 1.1 1.0 1.0 2.2 2.2 1.0 1.0 0.3 0.3 2.4 2.4 AktAkt AktAkt p85a p85b c p85p85aa p85p85bb

IP:p85a IP:p85IP:p85b WB: p110a WB: p110a IP:p85IP:p85a a IP:p85IP:p8IP:p85IP:p85b5b 95kDa IP:p85b WB:WB:p110p110a a WB:WB:p110p110a a IP:p85a 72kDa WB:p85b IP:p85IP:p85b b WB:p85a 95kDa55kD95kDaa IP:p85a 72kD72kDa a WB:p8WB:p85b5b IP:p85a 36kDa WB:p85a a 55kD55kDa a WB:p85 28kDa 36kD17kD36kDa aa 28kD28kDa a 17kD17kDa a

Figure 4. Ba/F3 cells were transfected with a, WT p85α (p85α WT, left) or patient mutants, or WT p85β or patient mutants (right two gray bars, p85α WT or E160*). Cells were cultured without IL-3 for 4 weeks and harvested for viability assay. LacZ served as the control. The means (±SD) of triplicate samples of 3 independent experiments are shown. *, P < 0.05, compared with WT. The plasmids were transiently transfected into HEC1A cells. Whole-cell lysates were collected at 72 hours post-transfection for B , WB or c, IP with anti-p85α and then subjected to WB with anti-p110α . Numerical values below each lane of the immunoblots represent quantification of the relative protein level by densitometry (normalized to β-actin or AKT).

also obscure, hampering the development and implementa- (endometrioid/mucinous) (7, 11, 12), the EEC in our dataset tion of pathway-targeted therapy. Therefore, we performed are characterized by high frequencies of aberrations in the an integrated analysis to determine the effects of a broad PI3K pathway, KRAS, and CTNNB1, whereas TP53 mutations set of candidate mutations on downstream signaling in a are more frequent in grade 3 EEC. The mixed endometrioid/ large sample set. In agreement with the model of type 1 EC serous histotype is genetically similar to EEC, suggesting that

OF9 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article

aLacZ LacLacZ LacZ Z WT WTWTWT E160* E160*E160*E160* LacZWLacZWLacLacTEZWZWTE160*TETE160* 160*160* -Ub -Ub -U+Ubb +Ub +Ub 0 2 80 1220 0 82 24 2128 0 8 12 24 2 12 24 80 24 1220 0 82 24 2 12 8 0 8 12 24 212 24 80 24 122 0 08 2 24 212 8 h 8 1224 12 h 24 24 h h -+-+ -+-+-+ -+ -+-+-+ -+ -+-+ LacZ LacZ WT WT E160* E160* LacZWLacZWTETE160* 160* -Ub -Ub +Ub +Ub CHX CHX CHX LacZ WT E160* LacZWTE160* -Ub +Ub CHX WT 0 2 8 0 12 2 Lac 248 Z 12 0 224 8 0 12 2 248 12 0 224 8 0 12 2E 160* 248 12h 24 h MG132 MG132-+MG132MG132 -+ Lac-+ZW -+-+TE -+160* -Ub +Ub 0 2 8 12 24 0 2 8 12 24 0 2 8 12 24 h -+ -+ -+ 95kDa 95kDa 9595kDkDaa CHX 0 2 8 12 24 0 2 8 12 24 0 2 8 12 24 h PTEN PTEN PTEN PTEN PTENCHXPTENPTEN PTEN -+ -+ -+ CHX MG132 MG132 72kDa 72kDa 72kDa CHX MG132 95kDa 95kD72akDa 95kDa l l PTEN l PTEN MG132PTEN PTEN l β-actin β-actinPTEN120 β-actin120 120 β-actin β-actinPTENβ-actin IP: PTENIP: PTENIP: PTEN 55kDa 55kD95akD55kDa a ve ve β-actinve 120 β-actin IP: PTEN 72kDa 72kD55akDa ve PTEN 72kDa l l PTEN l le le le

100 100 100 l le l l 100 3.5 3.5 3.5 72kDa ve ve ve

l 3.5 * * * β-actin β120-actin80 12080 80 β-actin β-actinve * IP: PTEN 55kDa 55kDa ve le le le IP: PTEN β-actin ve 120 80 3 β-actin3 3 WB: UbiquitinWB: UbiquitinWB: Ubiquitin l 36kDa 36kD55akD36kDa a le IP: PTEN ve WB: Ubiquitin l l 3

le 36kDa β100-actin le 100 β 60 12060 60 l -actin* * * le 55kDa PTEN PTEN PTEN IP: PTEN ve 100 2.5 2.5 2.5

60 3.5 3.5

ve * ve PTEN 2.5 * *

* * * l 3.5 le 80 80 * * * ve *

100 le 40 40 LacZ40 LacZ LacZ * 2 le 2 2 WB: UbiquitinWB: Ubiquitin ve ve ve 80 * 3 3 * * * 36kDa 36kDa 40 LacZ le PTEN PTEN PTEN 3.5 2

ve WB: Ubiquitin ve ti ti ti

3 * * 36kDa WT WT WT PTEN

60 60ti 1.5 1.5 * 1.5 80 * * * *

PTEN 20 20 20 PTEN 2.5 le la la la WT 2.5 WB: Ubiquitin

60 20 * 31.5 * IP: PTENIP: PTENIP: PTEN 36kDa PTEN ve ve ve

la E160* E160* E160* * * 2.5 IP: PTEN

* ve ti ti 400 60400LacZ 0 LacZE160* * * 12 ti 1 2 1 * ve Re Re Re ve PTEN * 2.5 ti 1 * * PTEN

40 LacZ PTEN la la 0 la 2 ve Re ti ti

* * WB: PTENWB: PTENWB: PTEN WT WT 0.5 0.PTEN 5 0.5

02020281812281224 224 4 la ti * 1.5 20 4020 LacZ * * 1.5

la 2 WB: PTEN la WT 0.5 ve 028122* 4 1.5 IP: PTEN IP: PTEN Re Re 20 Re * ve ve PTEN

la E160* E160*

ti 0 0 0 IP: PTEN ve Re WT ti 0 0 E160* * 1 ti 1.15 0

Re 20 Re la 0 ti 1 − IP:− PTEN − + − + − + +− + + + + + + + + la la Re E160* 0.5 ve 0.5 WB: PTENWB: PTEN 0202818122224 4 la − − + + + + 0 ti 0.15 WT −WT WB: + PTENWT− − + − − + − Re 0281224

Re WT Re B 0 la 0 − + − 0281224 Re 0.05 WB: PTEN E160* E160*− −E160* +− −+ + + + + + + Re 0 WT WTE160* − − + + + + − WT + −− + − PTEN 1.0PTEN 0.21.0PTEN 2.4 −0.2 1.0 1.6 +− 2.4 0.21.2 −+ 1.6 0.2.4 +9 1.2 1.6 +0. 91.2 + 0.9 WT WT WT WT 1.0 0.2 2.4 1.6 1.2 0.9 WT E160* E160*PTEN − + − 95kDa 95kDa95kDa E160* E160*− E160*− − E160* 72kDa 72kDa95k72kDaDa E160* − 1.0E160* 0.2 2.4 1.6 1.2 0.9 β-actinPTEN β-actinPTENβ-actin1.0 0.2 2.4 1.6 1.2 0.9 72kDa WT WT PTEN 1.0 0.2 2.4 1.6 1.2 0.9 55kDa 55kDa55kDa WT β-actinWT1.0 0.2WT 2.4WT 1.6WT 1.2 + E WT 0.160*9 + WTE160* + E160* 95kDa 95k55kDaDa E160* E160* PTEN WT 95kDa 1.0−E160* 1.01.0 − 0.6 1.01.0 0. WT 20.6 1.0 0. 0.62 0.2 WT + E160* IP: PTENIP: PTENIP: PTEN 72k36kDa 36k72kDaDa36kDaIP: PTENIP: PTENIP: PTEN− 0 2 0 8 212 0 8 24 212 0 8 24 122 0 8 24 212 0 8 24 2 12 h 8 24 12 h 24 h IP: PTEN 95k72k36kDaDa E160* −1.0 1.0 0.6 0.2 β-actin β-actin 0 2 8 12 24 0 2 8 12 24 h 55k28kDa 28k72k55kDaDa28kDa IP: PTEN CHX CHXβ-actinCHXWT WT WT + E160*WT + E160* 17kDa 17k55kDa28kDa17kDaDaWB: p85αWB: p85WB:α p85α β-actinCHX WT WT + E160* WB: p85αWB: p85WB:α p85α 1.0WB: p85 1.0α 1.0 0.6 1.0 0. 2 0.6 0.2 IP: PTEN IP: PTEN 36kDa 36k55k17kDaDa IP: PTEN IP: PTEN 1.0 1.0 0.6 0.2 0 2 80 12 2 WT 248 12 0 242 80 WT12 2 + 824 E 160*12 h 24 h IP:WB: PTEN p85α 36kDa IP: PTENIP:IP: PTEN PTENIP: PTEN PTEN 2.0PTEN 1.82.0PTEN 1.9 1.8 0 2.0 1.9 2 1.8 2.0 82.0 1.90.7 12 2.0 0.72.0 24 0.7 0.52.0 0.70 0.30.7 20.5 0.1 0.7 80.3 0.5 12 0.1 0.3 24 0.1 h IP: PTEN 28kDa 28kDa 1.0 1.0 0.6 0.2 CHX CHX 02.0 2 1.8 8 1.9 12 2.0 24 2.0 0 0.7 2 0.7 8 0.512 0.3 24 0.1 h 36k28kDa WB: p85α WB:IP: PTENp85IP: PTENα CHXPTEN 17kDa 17kDa WB: p85α WB: p85α WB: p85α 28k17kDa WB: PTENWB: PTENWB: PTEN CHX WB: p85α WB:WB: p85 PTENα IP: PTENIP: PTENIP: PTEN 17kDa IP: PTEN IP: PTEN β-actinPTEN β-actin2.0PTEN 1.8β-actin 2.01.9 1.8 2.0 1.9 2.0 2.00.7 2.00.7 0.70.5 0.70.3 0.50.1 0.3 0.1 WB: p85α IP: PTEN PTEN 2.0 1.8 1.9 2.0 2.0 0.7 0.7 0.5 0.3 0.1 IP: PTEN β-actin IP: PTEN PTEN 2.0 1.8 1.9 2.0 2.0 0.7 0.7 0.5 0.3 0.1 WB: PTENWB: PTEN WB: PTENWB: PTENWB: PTEN WB: PTEN IP: IgGIP: IgGIP:HA IgG HA HA β-actin β-actin IP: PTEN IP:WB: PTEN PTEN WB: PTEN HA-p85HA-α p85HA-α p85αβ IP: PTEN + IP −+: Ig G ++− + +− HA + + -actin IP: PTEN E160* E160*E160* HA-p85αβ-actin WB: PTENWB: PTEN + + + −+ − + −+ + + − + + + −+ −+ − c WB: PTENIP: IgG IP: IgG+HA + HA − + HA-p85HA-α +p85 FlHA-ag-α +p85 Flag-α +p85 Flag-α p85α E160* + − HA-p85αHA-p85αHA-p85IP: IgαG HA IP: IgGIP: IgIPGHA-: HAIgp85GαHA-HAp85αHA WB: PTEN + − + + − + + 95k Da + 95kDa95kDa HA-p85α + Flag-p85α HA-p85α HA-p85IP:α Ig+G − HA + + − − − IP: IgG HA Flag-p85αFlag-p85Flag-p85α α 72kDa 72kDa95k72kDaDa E160* E+160* − + −HA-p85α + + + − +− + −+ + − E160* + − 95kDa 95kDa 95kDa Flag-p85α 72kDaE160*HA-E160*p85αHA-E160* + Flp85ag-αp85 + Flαag-p85α HA-p85α HA-p85α + + − 55k Da + 55kDa55kDa HA-p85α + Flag-p85α IP: IgE160*GIP: Ig+G HA − HA 95kDa HA-p85α 95kDa 95kDa E160* IP: IgG HA 72kDa 72kDa 72kDa 95k55kDaDaIP: HA IP:− HA IP:HA- HA−p85α + Flag-p85α IP: IgG HA Flag-p85α Flag-p85HA-p85αα 72kDa 72kDa − 36kDa 36k95k72kDaDa36kDa IP: HA 5595kDkDaa55kD95akD7255akDkDaa Flag-p85α E160* E160* − 95kDa WB: FlagWB:Flag-p85 FlagWB:α Flag 55k28kDaDa 28k55k72kDa36kDa28kWB:DaDa FlagWB: FlagE160*WB: Flag 55kDa α α α 36kDa 36kD95akD36akDa 17kDa 17k55kDaDa17kIP:Da HA IP:E160* HA IP: HA WB: p85 WB: p85WB: p85 72kDa 72kDa WB: Flag 28kDaIP: HA IP:WB: HA Flag 28kDa 28kD72akD28akDa 55k17kDaDa IP: HAIP: HA WB: p85α 17kDa 17kDa3617kDkDaa 36kDa 36kDaWB: HA WB: HA WB: HA 55kDa 7255kD28akDa 36kDa IP: HA 55kD17akDa WB: Flag WB: Flag 28kDa 28kDaWB: Flag WB: Flag E160* E160* E160* WB: Flag 17kDa 17k36k28kDa IP: HA WB:IP: FlagWB: HA HA WB: p85α WB: p85α 36kDa 3655kDa 17kDa IP: HA WB: p85α 28kDa 2836kDkDaa WB: Flag 28kDa WB: Flag E160* 28kDa WB: HA WB: HA WB: HA IP: HA WB: p85α 17kDa 3617kDa 17kDaWB: HA WB: HA 17kDa WB: HA 28kDa WB: HA E160*WB: HA WB:E160* HA WB: HA 17kDa Flag-p85Flag-α +p85 FlE160*ag-α +p85 E160*α + E160* WB: HA E160* WB: HA E160* WB: HA WB: HA Flag-p85α + E160* : HA WB:− HA − − HA-p85HA-α +p85 FlHA-ag-α +p85 Flag-α +p85 Flag-α WB:p85 HAα WB HA-p85HA-α p85HA-α p85α WB: HA α − 95kDa 95kDa95kDa WB: HA p110α p110Flag-α p110p85Flα ag-+ Ep85160*α + E160* HA-p85α + Flag-p85α Flag-p85α + E160*72kDa 72kDa95k72kDaDa − − − WB: HA HA-p85α Flag-p85α + E160* − − − D − p110α − 55kDa 55kDa72k55kDaDa − p110αHA-p110p85α αHA-p110 + αFlp85ag-−αp85 + Flαag-p85α 95kDa 95k55kDaDa HA-p85α + Flag-p85α p110α p110α HA-p110p85α −αHA-p85α − 36kDa 36k95kDaDa36kDa HA-p85α p110α p110α IP: HA IP: HAp110IP:HA-α HAp85α + Flag-p85α IP: Flag IP:p110 FlagαIP: Flag 72kDa 95k72kDa − − p110α HA-p85α 72k36kDaDa − − − p110α 55k28kDa 28k72k55kDaDa28kDa IP: HA IP: PTENIP: PTENIP: PTEN− IP: Flag 55kDa p110α p110α − 55k28kDaWB:Da FlagWB: Flagp110WB:α Flag p110α p110α − WB: p85αWB: p85WB:α p85α 36kDa 36kDa IP: HA IP: HA IP: HA WB: p85αWB: p85p110IP:WB:α PTENα p85α 17kDa 17k36kDaDa17kIP:Da HA IP:p110WB: HAα Flag p110α IP: Flag IP: Flag IP: HA IP:WB: Flag p85α 36kDa IP: HA IP: PTENIP: PTENWB:IP: p85 PTENα E160* E160* E160* 28kDa 28k17kDaWB:Da HA WB:IP: HA HAWB: HA IP: PTEN IP: PTEN IP: Flag 28kDa IP: PTEN WB: Flag WB: Flag WB: PTENWB: PTENIP:WB: PTEN PTEN WB: p85α WB: p85E160*α 28kDa IP: HA WB:IP: FlagWB: HA HA WB: p85α WB:IP: PTENp85α WB: p85α 17kDa 17kDa IP: HA WB: p85α 17kDa WB: Flag WB: p85α IP: HA IP: PTEN WB:IP:WB: PTENp85 PTENα E160* E160* 17kDaWB: HA WB: HA IP: PTEN E160* WB: HA WB: FlagWB: FlagWB: Flag IP: PTEN E160* WB: HA WB: PTENWB: PTEN WB: PTEN WB: Flag WB: PTEN

WB: Flag WB: Flag WB: Flag WB: Flag

Figure 5. a, HEC1A cells transfected with LacZ or p85α WT or E160* were treated with cycloheximide (CHX; left) for indicated duration or with MG132 (center) for 24 hours. Cells were then harvested for WB. PTEN levels were normalized to β-actin by densitometry (below). *, P < 0.05. Right, HEC1A cells were transfected with LacZ or WT or E160* in the absence or presence of ubiquitin (Ub) for 72 hours. Whole-cell lysates were collected for IP with anti-PTEN and then subjected to WB with anti-ubiquitin. PTEN protein levels were normalized prior to IP by using proportionally different amounts of lysates. B, left, HEC1A cells were transfected with LacZ or WT or its mutants for 72 hours and were collected for IP with PTEN and WB with anti-p85α. HEC1A cells were co-transfected with WT or increasing amount of E160*. Cell lysates were collected for IP (center) or WB (right top). Right bottom, transfected HEC1A cells were treated with CHX for the indicated time points and harvested for WB. c, cells were transfected with HA-tagged p85α (HA-p85α) and/or Flag-tagged p85α (Flag-p85α) in the absence (left) or presence (center) of increasing amounts of E160*. IP was performed with anti-HA and WB with anti-Flag. Right, cells were transfected with HA-p85 α and increasing amounts of E160*. IP was performed with anti-HA and WB with anti-p85α . D, left, cells were transfected with WT p110α , Flag-p85α, and E160*. IP was performed with anti-Flag and WB with anti-p85α. Center, cells were transfected with WT p110α , Flag-p85α, and HA-p85α. IP was performed with anti-HA and WB with anti-Flag. Right, cells were transfected with WT p110α and HA-p85α, IP was performed with PTEN and WB with p85α . Numerical values below each lane of the immunoblots represent quantification of the relative protein level by densitometry.

JULY 2011 CANCER DISCOVERY | OF10

research article Cheung et al. the endometrioid component is dominant at least in terms additional mechanisms contribute to PTEN loss. Further, of mutational status. MMMT is poorly characterized to date. PTEN loss was also seen in a significant fraction of WT PTEN In addition to the frequent TP53 mutations previously re- tumors. PTEN protein levels and function are extensively reg- ported (28), we noted a low rate of aberrations in the PI3K ulated by multiple genomic and epigenetic mechanisms in pathway, KRAS, and CTNNB1 in MMMT, clearly suggesting addition to mRNA and protein modification, emphasizing that MMMT has more in common mutationally with type the importance of this tumor suppressor (25). Thus multiple 2 EC (serous and clear cell), in which TP53 mutation is the mechanisms are likely to converge on regulation of PTEN most common genetic alteration and is considered an early protein levels and function in EC. event (29). EC exhibits mutations throughout the coding region Based on frequent aberrations in PIK3CA and PTEN, we of PIK3CA with the exception of the RAS binding domain explored whether mutations in other members of the PI3K compatible with interactions with RAS being critical to the pathway are present in EC. Strikingly, the mutation rate of function of p110 (33). Consistent with a recent study, we PIK3R1 (20%) was markedly higher in EC than that previ- observed a high frequency of mutations within the ABD ously reported for any other lineage, demonstrating selec- domain in PIK3CA, compared to other tumor lineages (34). tive targeting in EC (17, 18). In the COSMIC database (30), Interestingly, 87% of mutations within the ABD domain PIK3R1 mutations were detected in 36 of 1693 (2%) tumors. were associated with PTEN heterozygous mutation, suggest- PIK3R2 mutation had only been reported in 1 of 108 (0.9%) ing that ABD mutations may selectively interact with PTEN colon cancers (18) and in 3 of 817 (0.4%) tumors (30). Our heterozygous mutations. ABD domain mutations have been studies demonstrated PIK3R2 mutations in 5% of EC, with suggested to have lower transforming activity and to activate several mutations being demonstrated to exhibit gain of AKT to a lesser degree than catalytic and helical domain mu- function, establishing PIK3R2 as a novel cancer gene. tations (31, 34). It is possible that co-mutation of the ABD Previous reports suggested that KRAS and PIK3CA muta- domain with PTEN increases information flow through the tions are mutually exclusive in EC (7, 14). However, these PI3K pathway. Although complete PTEN loss has been re- studies only sequenced exons 9 (helical) and 20 (catalytic do- ported to elicit senescence in prostate cancer models (35), it main) of PIK3CA in small sample sets. The large sample size does not appear to do so in EC, as several tumors did exhibit in this study, the characterization of multiple PI3K pathway complete PTEN loss. Furthermore, although PTEN loss is members, and our observation that PIK3CA mutations com- often a late event in prostate tumorigenesis, PTEN aberra- monly occur outside of exons 9 and 20 likely accounts for the tions occur early during the pathogenesis of EC. discrepancy, supporting our observation that KRAS and PI3K The roles of p85 in PI3K pathway activation and tumori- pathway members are coordinately mutated in EC. The con- genesis are complex. Levels of p85α are decreased in a num- current mutations in PI3K pathway members and KRAS in ber of tumor lineages (36). Haploinsufficiency of PIK3R1 can cell lines support the concept that the mutations occur in a result in PI3K pathway activation, whereas homozygous de- single cell population rather than in independent subclones. pletion inhibits the pathway (37). On the other hand, PIK3R1 This may reflect, in part, a lack of functional redundancy knockout mice demonstrated increased PI3K pathway activ- among PI3K pathway and KRAS mutations in EC that is ity and decreased PTEN protein levels, compatible with our supported by RPPA demonstrating distinct downstream sig- data on the effect of WT p85α on PTEN levels (36). If in naling consequences associated with these mutations. This excess of p110, p85 could compete with the p85–p110 com- activation of different pathways could thus cooperate for ef- plex for binding to phosphorylated insulin receptor substrate ficient transformation. (IRS) or to other phosphotyrosine-containing proteins, sug- Co-mutations in different components of the PI3K path- gesting that free p85 could negatively modulate PI3K sig- way might also cooperate for efficient transformation (31). naling (37). A direct and possibly bidirectional interaction Further, we have demonstrated that PTEN protein loss and between the N-terminal SH3-Rho-GAP domain of p85 and PIK3CA mutations have markedly different functional ef- PTEN has been demonstrated (26, 38, 39). We showed herein fects on PI3K pathway activation in human breast cancer that this p85α-PTEN interaction is associated with increased (22). Co-mutations in PI3K pathway members in EC occur at PTEN stability through decreased ubiquitination. The trun- frequencies significantly higher than predicted, in contrast to cated gain-of-function mutant E160*, which does not bind most cancers (19, 20). PTEN protein loss, regardless of PTEN either p110α or PTEN, binds WT p85α and decreases its mutational status, resulted in PI3K pathway activation. binding to PTEN, resulting in PTEN destabilization through PIK3CA, PIK3R1, or PIK3R2 mutation was more common in ubiquitination. cells where PTEN protein was retained, and these mutations Overall, the data are most compatible with a model phenocopy the functional effects of PTEN loss on down- wherein free p85α forms a homodimer that is able to bind stream signaling. MMR deficiency, which is an early event PTEN (Fig. 6). This model is supported by evidence that (1) in the pathogenesis of EC (21), might contribute to these overexpression of p85α did not increase p110α levels, indi- co-mutations. However, the mutations in the PI3K pathway cating that p85α is not limiting in the cells; (2) p110α is not members were not typical of MMR aberrations. part of the complex because p110α overexpression inhib- The majority of PTEN mutations in EC are heterozygous. ited p85α-PTEN interaction, suggesting a mutually exclusive PTEN haploinsufficiency has been reported to contribute binding; (3) E160* binds to free p85α because overexpres- to development of prostate cancer (32). However, despite sion of p110α competed with E160* for binding to p85α; PTEN mutations being heterozygous, almost half of EC dem- and (4) p85α self-dimerized and this homodimerization was onstrated complete loss of PTEN protein, suggesting that inhibited by overexpression of p110α and E160*. The SH3

OF11 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article a With wild-type p85a 3 SH KINA SE PR

C2 Ubiquitinated 3 3 AP PTEN SH SH hoG CAL R e Phosph atas PR PR LI Ub PR AP AP HE Ub H2 hoG hoG nS R R

C2 PR PR H2 H2 H2 iS nS nS PTEN RBD H2 H2 H2 cS Proteasome iS iS ABD WT p85 H2 H2

p110 cS cS

PTEN degradation

p85a homodimer-PTEN complex

B In the presence of p85a mutants

Ubiquitinated PTEN 3 3 C2 C2 3 3 SH SH SH SH Ub Ub PR PR PR PR Phos Phos AP AP AP hoG hoG hoG E160* pha pha R R R t ase t ase PR PR PR H2 H2 nS nS R348* PTEN PTEN H2 H2 iS

iS Proteasome H2 H2 cS cS WT p85a WT p85a

PTEN degradation

Figure 6. Schematic working model of p85α-mediated PTEN stabilization. a, in cells with WT p85α, in addition to p110α-bound p85α, excess p85α forms homodimers via intermolecular interactions between SH3 and proline-rich motif and between Rho-GAP domains (see Discussion). The homodimer likely undergoes conformation changes that allow binding to PTEN and stabilizes PTEN through inhibition of ubiquitination. B , the E160* truncated mutant binds WT p85α but inhibits PTEN binding, resulting in PTEN degradation. However, the R348* mutant retains the ability to bind PTEN after dimerization, either alone or with WT p85α, thereby inhibiting PTEN degradation. PR, proline-rich motif. and first proline-rich motif have been shown to be suffi- homodimer formation (27). The ability of E160* to bind cient to mediate homodimerization; furthermore, the trun- with WT p85α suggests that SH3 and the first proline-rich cated isoform p55γ, which lacks the N-terminal domains, domains constitute a major homodimerization interface. failed to form dimers (27). Intermolecular interaction be- However, the failure of the E160*–p85α heterodimer to bind tween Rho-GAP domains might also contribute to the p85α and stabilize PTEN whereas the R348*–p85α heterodimer

JULY 2011 CANCER DISCOVERY | OF12

research article Cheung et al.

(or possibly R348* homodimers) binds and stabilizes PTEN, premature to predict the pathophysiologic relevance of the suggests that the Rho-GAP domain likely plays a role in the types of PIK3R1 mutations on tumor initiation and pro- conformation change in p85α required for binding to PTEN. gression. Possibly the different mutations will contribute How the homodimer enhances PTEN stability warrants to sensitivity to a specific type of PI3K pathway inhibitor. further investigation, but it is tempting to speculate that Therefore, the effects of each class of PIK3R1 mutations on the homodimer may provide a combinatorial binding site tumor initiation and progression warrant further study. for PTEN or may facilitate recruitment of other molecules WT p85β demonstrated greater ability to induce Ba/F3 to the complex to stabilize PTEN. Phosphorylation of the survival than WT p85α. Unlike p85α, p85β did not stabilize C-terminal tail of PTEN inhibits its proteosome-mediated PTEN protein in spite of its ability to bind PTEN (unpub- degradation (40). p85 and unphosphorylated PTEN are part lished data). It is possible that the two isoforms mediate dif- of a high molecular weight complex with formation of the ferent functions (45). Another plausible explanation is that complex associated with diminished AKT activation (38). the effects of p85α on PTEN are dominant over p85β when In this scenario, it is possible that phosphorylated PTEN both isoforms are expressed. Indeed, we found that expres- adopts a closed conformation that results in lower affinity sion of exogenous p85α inhibited binding of p85β to PTEN for the complex (41). The association of PTEN with p85α (unpublished observation), suggesting that PTEN interacts is readily observed in EEC in the absence of growth factor preferentially with p85α. This finding is in contrast to previ- addition, in contrast to previous studies (26). This can be ous studies in other cell lines where p85β appears to be the potentially due to the frequent presence of other activating preferred partner of PTEN (39) and may represent the unique mutations in EEC (Fig. 1). In contrast to studies in other context of EEC cells. Of note, p85α and p85β share only systems (38), we were unable to find evidence for a trimeric 30% protein homology in the Rho-GAP domain that con- (or tetrameric) complex of PTEN, p85α, and p110α by cross- tributes to PTEN binding, in contrast to 80% in the SH2 do- immunoprecipitation in EEC (not presented). Further, the mains. Although p85β possesses similar proline-rich motifs ability of exogenous p110α to decrease p85α homodimers as p85α, whether p85β homodimerizes and/or bind PTEN as and p85α–PTEN heterodimers suggests that in EEC, the efficiently as p85α is unclear and may be context dependent. binding of PTEN and p110α to p85α may be mutually ex- Therefore, due to the complex effects of p85α and p85β on clusive, with p85α monomers interacting with p110α and the PI3K pathway, tumor development is likely to be tissue dimers with PTEN. The spontaneous E160* mutation from context dependent and determined by relative levels and ac- EEC will provide a powerful tool to elucidate the mecha- tivities of p110, p85, PTEN and phosphotyrosine residues. nisms underlying these processes. Germline variants of PIK3R1 (M326I) and PIK3R2 (V727T) Other activating p85α mutations found in EC likely eluci- were found in our samples, with homozygous mutations oc- date additional regulatory mechanisms as they do not appar- curring in frequencies fitting Hardy-Weinberg equilibrium. ently alter PTEN protein levels but do increase pAKT levels. Although M326I has been reported to exhibit increased bind- Three of five activating mutations identified in the Ba/F3 as- ing to IRS-1, it had no impact on PI3K activity and signaling say reside in the iSH2 domain, which binds the C2 domain of events (46). Herein, we found that M326I and V727T had no p110 inhibiting the catalytic activity of p110. These activat- effect on Ba/F3 survival or AKT phosphorylation. RPPA analy- ing mutations appear to disrupt iSH2–C2 contact, releasing sis also showed no difference in downstream signaling between the inhibitory effect on p110 while retaining the ability to patients with or without these variants (data not shown). stabilize p110 (24, 42). One of the novel activating mutations The prevalence of abnormalities in members of the PI3K in EC (R503W) is part of an exposed surface composed of pathway and KRAS implicates these pathways as critical driv- basic residues that may bind to negatively charged membrane ers of pathogenesis of EC and thus as exciting targets for phosphatidylinositide substrates compatible with an inde- cancer treatment. Encouragingly, we found that mutations pendent mechanism (43). in the PI3K pathway and KRAS predict in vitro sensitivity to The recurrent R348* activating mutation lacks intact SH2 rapamycin and MEK inhibitor but, like other studies, not to and iSH2 domains and thus lacks ability to bind p110. It GDC-0941 (47). Due to feedback loops in the PI3K pathway, demonstrated lower activity than E160* in multiple repeat rational targeted therapeutic combinations may be required Ba/F3 assays. The portion of p85α remaining in the R348* for optimal outcomes. Our data also suggest that EEC cells mutant is similar to the minimal domain defined to bind with both PI3K pathway and KRAS mutations are relatively PTEN (26), a property we confirmed for R348*, which may more resistant to PI3K pathway inhibitors than those with contribute to its activity. Because the R348* mutation is het- PI3K pathway aberrations alone, as previously observed (48, erozygous in each of the 5 tumors, p110 might be stabilized 49). Whether coordinate targeting of the PI3K and RAS path- by intact p85α expressed from the remaining PIK3R1 allele. ways will demonstrate higher activity with acceptable toxicity Indeed, an artificial p85α mutant unable to bind p110 has in EC xenografts is under investigation. Further, we found been demonstrated to increase kinase activity of a p110α that EC patients with PI3K pathway aberration appear to mutant (44). Interestingly, heterozygous mutations in PTEN have a significantly lower recurrence risk (P = 0.003), sug- and/or PIK3CA were concurrent in tumors harboring the gesting that PI3K pathway mutation status might also pre- R348* mutation, suggesting that the activity of this mutant dict tumor recurrence. is most clearly manifest in cooperation with other mutants in Currently, the choice of therapy for individual patients the PI3K pathway. with EC is largely empirical. Our data, which include compre- Thus, there appear to be multiple mechanisms by which hensive characterization of genetic alterations coupled with aberrations in PIK3R1 can contribute to tumorigenesis. It is mechanistic studies, suggest that manipulation of the PI3K

OF13 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article and RAS pathways could provide a promising molecular- Ba/F3 cell viability assay targeted therapy in EC. The genomic and protein signatures Lentiviruses carrying WT or mutated PIK3R1 or PIK3R2 were gen- we identified should affect the stratification of patients in tri- erated by cotransfection of HEK293FT cells with the constructs and als and hence accelerate the fulfillment of personalized mo- ViralPower Lentiviral Expression System. Ba/F3 cells were incubated lecular medicine. with the viral supernatant supplemented with IL-3. At 48 hours postinfection, the cells were washed three times with PBS and re- 4 MethODs suspended in medium without IL-3. Cells (2 × 10 ) were plated in 96-well plate and cultured for 4 weeks. CellTiter-Blue (Promega) was Patient Samples used to assess cell viability. Endometrial tumor samples (26 EEC grade 1, 106 EEC grade 2, 29 EEC grade 3, 60 mixed endometrioid and serous, 18 MMMT, 4 se- statistical analysis rous) were obtained under institutional review board–approved pro- To test the significance of pairwise comparisons between aberra- tocols from 243 patients who were diagnosed at the MD Anderson tions, we applied Tukey's Honestly Significant Difference (HSD) test Cancer Center (Houston, Texas) from 1998 to 2009. Specimens were to a 1-way ANOVA model. The analysis was performed with the R reviewed by a pathologist (R.R. Broaddus) for ≥80% tumor content, software (http://www.r-project.org/). Difference was considered sig- histologic verification, and classification by grade and stage. nificant when P < 0.05. Additional experimental procedures are available in the Cell Cultures and Plasmids Supplementary Data. For descriptions of the cell lines used, see Supplementary Data. Disclosure of Potential conflicts of interest DNA fingerprinting (STR analysis) was performed for the characterization of cell lines. G.B. Mills received commercial research grants from AstraZeneca, Full-length open reading frames of PIK3R1 and PIK3R2 GlaxoSmithKline, Celgene, Exelixis, Roche, and Wyeth/Pfizer, and were cloned into pLenti-7.3 Dest lentiviral expression system served as consultant to AstraZeneca, Celgene, Enzon, Novartis, and (Invitrogen). Mutations were generated by site-directed mutagene- Wyeth/Pfizer. The other authors disclosed no potential conflicts of sis using QuikChange II XL Site-Directed Mutagenesis Kit (Agilent interest. Technologies). Four PIK3R1 mutants were kindly provided by Dr. The Editor-in-Chief of Cancer Discovery is an author of this article. Lynda Chin (Harvard Medical School, Boston, Massachusetts). All In keeping with the AACR's Editorial Policy, the paper was peer re- constructs were confirmed by Sanger sequencing. Transfection was viewed and a member of the AACR's Publications Committee ren- performed using Lipofectamine 2000 (Invitrogen). dered the decision concerning acceptability. acknowledgments Mutational analysis We thank Drs. Lynda Chin and Peter Vogt for providing p85α DNA preparation plasmids; and CCSG DNA analysis facility, RPPA core, and character- Genomic DNA from frozen tumor resections or endometrial tu- ized cell line core (funded by NCI CA16672) in the MD Anderson mor cell lines was extracted using QIAamp DNA Mini Kit (Qiagen). Cancer Center. We wish to acknowledge the contributions of the Normal DNA was extracted from peripheral blood leukocytes using Baylor College of Medicine Human Genome Sequencing Center and QIAamp Blood kit (Qiagen). in particular, Dr. Richard Gibbs, director, Donna Muzny, Dr. David Wheeler, Lora Lewis, Robert Ruth, Kyle Chang, Humeira Akbar, and SNP analysis Shannon Gross. A mass spectroscopy–based approach was used to detect SNPs Grant support (MassARRAY, Sequenom) as described previously (22). We designed high-throughput assays for somatic hotspot mutations in KRAS, This work was supported by a Stand Up to Cancer grant to L.C. PIK3CA, CTNNB1, BRAF and AKT. Mutations in the panel are listed Cantley, A.P. Myers, and G.B. Mills; a Uterine Cancer SPORE grant in Supplementary Data. (NIH/NCI P50 CA098258) to R.R. Broaddus, K.H. Lu, and G.B. Mills; and a Career Development Award from the Conquer Cancer Whole-gene resequencing Foundation of the American Society of Clinical Oncology to B.T. Hennessy. Resequencing of PTEN, PIK3CA, PIK3R1, PIK3R2, FGFR2, TP53, and CTNNB1 was performed at the Human Genome Sequencing Received February 26, 2011; revised April 21, 2011; accepted May Center at the Baylor College of Medicine (Houston, Texas). The last 12, 2011; published OnlineFirst June 7, 2011. exon of PTEN failed to amplify. All other sequencing covered all exons and intron–exon boundaries. Ninety-two percent of PIK3CA mutations and 89% of CTNNB1 mutations detected by Sanger sequencing were confirmed by reFereNces MassARRAY (Sequenom), indicating a high accuracy of base-calling. 1. Hennessy BT, Smith DL, Ram PT, Lu Y, Mills GB. Exploiting the PI3K/AKT pathway for cancer drug discovery. Nat Rev Drug Discov Mutations verification 2005;4:988–1004. PI3KR1 and PIK3R2 mutations detected by resequencing were con- 2. Daikoku T, Hirota Y, Tranguch S, Joshi AR, DeMayo FJ, Lydon JP, et al. Conditional loss of uterine Pten unfailingly and rapidly induces firmed by independent PCR-based sequencing in the MD Anderson endometrial cancer in mice. Cancer Res 2008;68:5619–27. Cancer Center Sequencing Core. Variants were deemed somatic if 3. Mutter GL, Ince TA, Baak JP, Kust GA, Zhou XP, Eng C. Molecular they were absent in matched normal DNA. identification of latent precancers in histologically normal endometrium. Cancer Res 2001;61:4311–4. Reverse-phase protein array 4. Sun H, Enomoto T, Fujita M, Wada H, Yoshino K, Ozaki K, et al. High-throughput RPPA for 127 proteins (antibody list provided in Mutational analysis of the PTEN gene in endometrial carcinoma and Supplementary Data) was performed as described previously (22, 50). hyperplasia. Am J Clin Pathol 2001;115:32–8.

JULY 2011 CANCER DISCOVERY | OF14

research article Cheung et al.

5. Kanamori Y, Kigawa J, Itamochi H, Shimada M, Takahashi M, p85 subunit of phosphatidylinositol 3-kinase. Proc Natl Acad Sci Kamazawa S, et al. Correlation between loss of PTEN expression U S A 2010;107:5471–6. and Akt phosphorylation in endometrial carcinoma. Clin Cancer Res 27. Harpur AG, Layton MJ, Das P, Bottomley MJ, Panayotou G, 2001;7:892–5. Driscoll PC, et al. Intermolecular interactions of the p85alpha 6. Quddus MR, Ologun BA, Sung CJ, Steinhoff MM, Lawrence WD. regulatory subunit of phosphatidylinositol 3-kinase. J Biol Chem Utility of PTEN expression of endometrial "surface epithelial 1999;274:12323–32. changes" and underlying atypical endometrial hyperplasia. Int J 28. Taylor NP, Zighelboim I, Huettner PC, Powell MA, Gibb RK, Rader Gynecol Pathol 2009;28:471–6. JS, et al. DNA mismatch repair and TP53 defects are early events in 7. Velasco A, Bussaglia E, Pallares J, Dolcet X, Llobet D, Encinas M, uterine carcinosarcoma tumorigenesis. Mod Pathol 2006;19:1333–8. et al. PIK3CA gene mutations in endometrial carcinoma: correlation 29. Jarboe EA, Pizer ES, Miron A, Monte N, Mutter GL, Crum CP. with PTEN and K-RAS alterations. Hum Pathol 2006;37:1465–72. Evidence for a latent precursor (p53 signature) that may pre- 8. Hayes MP, Douglas W, Ellenson LH. Molecular alterations of cede serous endometrial intraepithelial carcinoma. Mod Pathol EGFR and PIK3CA in uterine serous carcinoma. Gynecol Oncol 2009;22:345–50. 2009;113:370–3. 30. Forbes SA, Bindal N, Bamford S, Cole C, Kok CY, Beare D, 9. Shoji K, Oda K, Nakagawa S, Hosokawa S, Nagae G, Uehara Y, et al. COSMIC: mining complete cancer genomes in the Catalogue of et al. The oncogenic mutation in the pleckstrin homology domain of Somatic Mutations in Cancer. Nucleic Acids Res 2011;39: D945–50. AKT1 in endometrial carcinomas. Br J Cancer 2009;101:145–8. 31. Oda K, Okada J, Timmerman L, Rodriguez-Viciana P, Stokoe D, 10. Dutt A, Salvesen HB, Chen TH, Ramos AH, Onofrio RC, Hatton C, Shoji K, et al. PIK3CA cooperates with other phosphatidylinositol et al. Drug-sensitive FGFR2 mutations in endometrial carcinoma. 3'-kinase pathway mutations to effect oncogenic transformation. Proc Natl Acad Sci U S A 2008;105:8713–7. Cancer Res 2008;68:8127–36. 11. Moreno-Bueno G, Hardisson D, Sanchez C, Sarrio D, Cassia R, 32. Kwabi-Addo B, Giri D, Schmidt K, Podsypanina K, Parsons R, Garcia-Rostan G, et al. Abnormalities of the APC/beta-catenin path- Greenberg N, et al. Haploinsufficiency of the Pten tumor suppressor way in endometrial cancer. Oncogene 2002;21:7981–90. gene promotes prostate cancer progression. Proc Natl Acad Sci U S A 12. Lax SF, Kendall B, Tashiro H, Slebos RJ, Hedrick L. The frequency of 2001;98:11563–8. p53, K-ras mutations, and microsatellite instability differs in uterine 33. Orme MH, Alrubaie S, Bradley GL, Walker CD, Leevers SJ. Input endometrioid and serous carcinoma: evidence of distinct molecular from Ras is required for maximal PI(3)K signalling in Drosophila. genetic pathways. Cancer 2000;88:814–24. Nat Cell Biol 2006;8:1298–302. 13. Gupta S, Ramjaun AR, Haiko P, Wang Y, Warne PH, Nicke B, et al. 34. Rudd ML, Price JC, Fogoros S, Godwin AK, Sgroi DC, Merino M, et al. Binding of ras to phosphoinositide 3-kinase p110alpha is required A unique spectrum of somatic PIK3CA (p110a) mutations within for ras-driven tumorigenesis in mice. Cell 2007;129:957–68. primary endometrial carcinomas. Clin Cancer Res 2011;17:1331–40. 14. Kang S, Seo SS, Chang HJ, Yoo CW, Park SY, Dong SM. Mutual 35. Alimonti A, Nardella C, Chen Z, Clohessy JG, Carracedo A, Trotman exclusiveness between PIK3CA and KRAS mutations in endometrial LC, et al. A novel type of cellular senescence that can be enhanced in carcinoma. Int J Gynecol Cancer 2008;18:1339–43. mouse models and human tumor xenografts to suppress prostate 15. Slomovitz BM, Lu KH, Johnston T, Coleman RL, Munsell M, tumorigenesis. J Clin Invest 2010;120:681–93. Broaddus RR, et al. A phase 2 study of the oral mammalian target of 36. Taniguchi CM, Winnay J, Kondo T, Bronson RT, Guimaraes AR, rapamycin inhibitor, everolimus, in patients with recurrent endome- Aleman JO, et al. The phosphoinositide 3-kinase regulatory subunit trial carcinoma. Cancer 2010;116:5415–9. p85alpha can exert tumor suppressor properties through negative 16. Courtney KD, Corcoran RB, Engelman JA. The PI3K pathway as drug regulation of growth factor signaling. Cancer Res 2010;70:5305–15. target in human cancer. J Clin Oncol 2010;28:1075–83. 37. Ueki K, Fruman DA, Brachmann SM, Tseng YH, Cantley LC, Kahn 17. Cancer Genome Atlas Research Network. Comprehensive genomic CR. Molecular balance between the regulatory and catalytic subunits characterization defines human glioblastoma genes and core path- of phosphoinositide 3-kinase regulates cell signaling and survival. ways. Nature 2008;455:1061–8. Mol Cell Biol 2002;22:965–77. 18. Jaiswal BS, Janakiraman V, Kljavin NM, Chaudhuri S, Stern HM, 38. Rabinovsky R, Pochanard P, McNear C, Brachmann SM, Duke- Wang W, et al. Somatic mutations in p85alpha promote tumorigen- Cohan JS, Garraway LA, et al. p85 Associates with unphosphory- esis through class IA PI3K activation. Cancer Cell 2009;16:463–74. lated PTEN and the PTEN-associated complex. Mol Cell Biol 19. Broderick DK, Di C, Parrett TJ, Samuels YR, Cummins JM, 2009;29:5377–88. McLendon RE, et al. Mutations of PIK3CA in anaplastic oligo- 39. He J, de la Monte S, Wands JR. The p85beta regulatory subunit of dendrogliomas, high-grade astrocytomas, and medulloblastomas. PI3K serves as a substrate for PTEN protein phosphatase activity Cancer Res 2004;64:5048–50. during insulin mediated signaling. Biochem Biophys Res Commun 20. Murugan AK, Hong NT, Fukui Y, Munirajan AK, Tsuchida N. 2010;397:513–9. Oncogenic mutations of the PIK3CA gene in head and neck squa- 40. Vazquez F, Ramaswamy S, Nakamura N, Sellers WR. Phosphorylation mous cell carcinomas. Int J Oncol 2008;32:101–11. of the PTEN tail regulates protein stability and function. Mol Cell 21. Karamurzin Y, Rutgers JK. DNA mismatch repair deficiency in endo- Biol 2000;20:5010–8. metrial carcinoma. Int J Gynecol Pathol 2009;28:239–55. 41. Vazquez F, Grossman SR, Takahashi Y, Rokas MV, Nakamura N, 22. Stemke-Hale K, Gonzalez-Angulo AM, Lluch A, Neve RM, Kuo WL, Sellers WR. Phosphorylation of the PTEN tail acts as an inhibitory Davies M, et al. An integrative genomic and proteomic analysis of switch by preventing its recruitment into a protein complex. J Biol PIK3CA, PTEN, and AKT mutations in breast cancer. Cancer Res Chem 2001;276:48627–30. 2008;68:6084–91. 42. Wu H, Shekar SC, Flinn RJ, El-Sibai M, Jaiswal BS, Sen KI, et al. 23. Samuels Y, Wang Z, Bardelli A, Silliman N, Ptak J, Szabo S, et al. Regulation of class IA PI 3-kinases: C2 domain-iSH2 domain con- High frequency of mutations of the PIK3CA gene in human cancers. tacts inhibit p85/p110alpha and are disrupted in oncogenic p85 mu- Science 2004;304:554. tants. Proc Natl Acad Sci U S A 2009;106:20258–63. 24. Sun M, Hillmann P, Hofmann BT, Hart JR, Vogt PK. Cancer-derived 43. Huang CH, Mandelker D, Schmidt-Kittler O, Samuels Y, Velculescu mutations in the regulatory subunit p85alpha of phosphoinositide VE, Kinzler KW, et al. The structure of a human p110alpha/p85alpha 3-kinase function through the catalytic subunit p110alpha. Proc complex elucidates the effects of oncogenic PI3Kalpha mutations. Natl Acad Sci U S A 2010;107:15547–52. Science 2007;318:1744–8. 25. Salmena L, Carracedo A, Pandolfi PP. Tenets of PTEN tumor sup- 44. Hsu JH, Shi Y, Frost P, Yan H, Hoang B, Sharma S, et al. Interleukin-6 pression. Cell 2008;133:403–14. activates phosphoinositol-3' kinase in multiple myeloma tumor cells 26. Chagpar RB, Links PH, Pastor MC, Furber LA, Hawrysh AD, by signaling through RAS-dependent and, separately, through p85- Chamberlain MD, et al. Direct positive regulation of PTEN by the dependent pathways. Oncogene 2004;23:3368–75.

OF15 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

PI3K Pathway Mutations in Endometrial Cancer research article

45. Ueki K, Fruman DA, Yballe CM, Fasshauer M, Klein J, Asano T, et al. 48. Ihle NT, Lemos R Jr Wipf P, Yacoub A, Mitchell C, Siwak D, et al. Positive and negative roles of p85 alpha and p85 beta regulatory sub- Mutations in the phosphatidylinositol-3-kinase pathway predict for units of phosphoinositide 3-kinase in insulin signaling. J Biol Chem antitumor activity of the inhibitor PX-866 whereas oncogenic Ras is 2003;278:48453–66. a dominant predictor for resistance. Cancer Res 2009;69:143–50. 46. Almind K, Delahaye L, Hansen T, Van Obberghen E, Pedersen O, 49. Engelman JA, Chen L, Tan X, Crosby K, Guimaraes AR, Upadhyay Kahn CR. Characterization of the Met326Ile variant of phos- R, et al. Effective use of PI3K and MEK inhibitors to treat mutant phatidylinositol 3-kinase p85alpha. Proc Natl Acad Sci U S A Kras G12D and PIK3CA H1047R murine lung cancers. Nat Med 2002;99:2124–8. 2008;14:1351–6. 47. O'Brien C, Wallin JJ, Sampath D, GuhaThakurta D, Savage H, 50. Zhang L, Wei Q, Mao L, Liu W, Mills GB, Coombes K. Serial dilution Punnoose EA, et al. Predictive biomarkers of sensitivity to the phos- curve: a new method for analysis of reverse phase protein array data. phatidylinositol 3' kinase inhibitor GDC-0941 in breast cancer pre- Bioinformatics 2009;25:650–4. clinical models. Clin Cancer Res 2010;16:3670–83.

JULY 2011 CANCER DISCOVERY | OF16

Correction: High Frequency of PIK3R1 and PIK3R2 Mutations in Endometrial Cancer Elucidates a Novel Mechanism for Regulation of PTEN Protein Stability

In this article (Cancer Discovery 2011;1:170–85), which appeared in the July 2011 issue of Cancer Discovery (1), the labels along the vertical axes in Figure 5 were incorrectly formatted. The amended ﬁ gure appears on the next page. The publisher regrets this error.

750 | CANCER DISCOVERYAUGUST 2012 www.aacrjournals.org A LacZWT E160* LacZ WT E160* -Ub +Ub CHX 0 2 8 12 24 0 2 8 12 24 0 2 8 12 24 h MG132 -+ -+ -+

PTEN PTEN 95kDa IP: PTEN β-actin β-actin 72kDa

120 WB: Ubiquitin 55kDa 100 3.5 * 80 3 36kDa 60 2.5 * IP: PTEN * 40 LacZ * 2 * WB: PTEN 20 WT * 1.5 E160* 0 1 Relative PTEN level PTEN Relative 0281224 0.5

Relative PTEN level PTEN Relative 0

WT − − + + + + B E160* − + − PTEN E160* WT 95kDa 1.0 0.2 2.4 1.6 1.2 0.9 72kDa − IP: PTEN β-actin 55kDa IP: PTEN WB: p85α 36kDa 1.0 1.0 0.6 0.2 WT WT + E160* WB: p85α 28kDa IP: PTEN 17kDa CHX 0 2 8 12 24 0 2 8 12 24 h WB: PTEN IP: PTEN PTEN WB: PTEN 2.0 1.8 1.9 2.0 2.0 0.7 0.7 0.5 0.3 0.1 β-actin

IP: IgG HA C HA-p85α HA-p85α + − + + Flag-p85α + + − + E160* + − HA-p85α + Flag-p85α IP: IgG HA 95kDa E160* 72kDa − IP: HA 95kDa 55kDa WB: Flag WB: Flag WB: p85α 72kDa 36kDa IP: HA 55kDa 28kDa WB: HA 36kDa 17kDa 28kDa E160* 17kDa WB: HA WB: HA

D Flag-p85α + E160* p110α − HA-p85α + Flag-p85α 95kDa HA-p85α α 72kDa p110 − p110α 55kDa − IP: Flag IP: HA IP: PTEN WB: p85α 36kDa WB: Flag WB: p85α 28kDa IP: HA IP: PTEN WB: HA WB: PTEN E160* 17kDa

WB: Flag

Figure 5. A, HEC1A cells transfected with LacZ or p85α WT or E160* were treated with cycloheximide (CHX; left) for indicated duration or with MG132 (center) for 24 hours. Cells were then harvested for WB. PTEN levels were normalized to β-actin by densitometry (below). *, P < 0.05. Right, HEC1A cells were transfected with LacZ or WT or E160* in the absence or presence of ubiquitin (Ub) for 72 hours. Whole-cell lysates were collected for IP with anti-PTEN and then subjected to WB with anti-ubiquitin. PTEN protein levels were normalized prior to IP by using proportionally different amounts of lysates. B, left, HEC1A cells were transfected with LacZ or WT or its mutants for 72 hours and were collected for IP with PTEN and WB with anti-p85α. HEC1A cells were co-transfected with WT or increasing amount of E160*. Cell lysates were collected for IP (center) or WB (right top). Right bottom, transfected HEC1A cells were treated with CHX for the indicated time points and harvested for WB. C, cells were transfected with HA-tagged p85α (HA-p85α) and/or Flag-tagged p85α (Flag-p85α) in the absence (left) or presence (center) of increasing amounts of E160*. IP was performed with anti-HA and WB with anti-Flag. Right, cells were transfected with HA-p85α and increasing amounts of E160*. IP was performed with anti-HA and WB with anti-p85α. D, left, cells were transfected with WT p110α, Flag-p85α, and E160*. IP was performed with anti-Flag and WB with anti-p85α. Center, cells were transfected with WT p110α, Flag-p85α, and HA-p85α. IP was performed with anti-HA and WB with anti-Flag. Right, cells were transfected with WT p110α and HA-p85α, IP was performed with PTEN and WB with p85α. Numerical values below each lane of the immunoblots represent quantiﬁ cation of the relative protein level by densitometry.

REFERENCE 1. Cheung LW, Hennessy BT, Li J, Yu S, Myers AP, Djordjevic B, et al. High frequency of PIK3R1 and PIK3R2 mutations in endometrial cancer elucidates a novel mechanism for regulation of PTEN protein stability. Cancer Discovery 2011;1:170–85.

AUGUST 2012CANCER DISCOVERY | 751 Published OnlineFirst June 7, 2011; DOI: 10.1158/2159-8290.CD-11-0039

High Frequency of PIK3R1 and PIK3R2 Mutations in Endometrial Cancer Elucidates a Novel Mechanism for Regulation of PTEN Protein Stability

Lydia W.T. Cheung, Bryan T. Hennessy, Jie Li, et al.

Cancer Discovery Published OnlineFirst June 7, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-11-0039

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2011/08/09/2159-8290.CD-11-0039.DC2 http://cancerdiscovery.aacrjournals.org/content/suppl/2011/05/23/2159-8290.CD-11-0039.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2011/06/23/2159-8290.CD-11-0039. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.