Gynecologic Oncology 133 (2014) 346–352

Contents lists available at ScienceDirect

Gynecologic Oncology

j ourna l homepage: www.e lsev ie r .com/ locate /ygyno

Assessing the efficacy of targeting the phosphatidylinositol 3-kinase/AKT/mTOR signaling pathway in endometrial cancer

Leslie S. Bradford a,b,e,1,2, Alejandro Rauh-Hain a,b,e,1, Rachel M. Clark a,b,e, Jolijn W. Groeneweg a,e, Ling Zhang a,Darrell Borger d,e, Lawrence R. Zukerberg c,e, Whitfield B. Growdon a,b,e, Rosemary Foster a,b,e, Bo R. Rueda a,b,e,⁎a Vincent Center for Reproductive Biology, Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital, Boston, MA 02114, USAb Gynecologic Oncology Division, Vincent Department of Obstetrics & Gynecology, Massachusetts General Hospital, Boston, MA 02114, USAc Department of Pathology, Massachusetts General Hospital, Boston, MA 02114, USAd Division of Hematology/Oncology, Department of Medicine, Massachusetts General Hospital, Boston, MA, USAe Harvard Medical School, Boston, MA 02115, USA

H I G H L I G H T S

• We treated mice bearing primary human endometrial cancer xenografts with single agent PI3K inhibitor NVP BKM-120.• Response to inhibition of PI3K was unaffected by the presence of an activating mutation in PIK3CA.• Combination of PI3K inhibition with conventional paclitaxel and carboplatin abrogated the consistent acquired resistance pattern observed with single agenttherapy.

⁎ Corresponding author at: Massachusetts GeneralReproductive Biology, Vincent Department of ObstetricsFruit Street, Boston, MA 02114, USA. Fax: +1 617 726 05

E-mail address: brueda@partners.org (B.R. Rueda).1 These authors contributed equally.2 Present address: Division of Gynecologic Oncology, 1

01605, USA.

http://dx.doi.org/10.1016/j.ygyno.2014.02.0220090-8258/© 2014 Elsevier Inc. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 11 December 2013Accepted 13 February 2014Available online 19 February 2014

Keywords:Endometrial cancerPI3K inhibitorPI3K/AKT/mTOR pathway

Objective. Alterations in the PI3K pathway are prevalent in endometrial cancer due to PIK3CA mutationand loss of PTEN. We investigated the anti-tumor activity of the PI3K inhibitor NVP BKM-120 (BKM) as asingle agent and in combination with standard cytotoxic chemotherapy in a human primary endometrialxenograft model.

Methods. NOD/SCID mice bearing xenografts of primary human tumors with and without PIK3CA genemutations were divided into two and four arm cohorts with equivalent tumor volumes. BKM was adminis-tered alone and in combination with paclitaxel and carboplatin (P/C) and endometrial xenograft tumor vol-umes were assessed. Tumors from the BKM, P/C, P/C + BKM and vehicle treated mice were processed for

determination of PI3K/AKT/mTOR pathway activation.

Results. In both single agent experiments, BKM resulted in significant tumor growth suppressionstarting at days 5–10 compared to the linear growth observed in vehicle treated tumors (p b 0.04in all experiments). Tumor resurgence manifested between days 14 and 25 (p b 0.03). When BKMwas combined with P/C, this resistance pattern failed to develop in three separate xenograft lines(p b 0.05). Synergistic tumor growth suppression (p b 0.05) of only one xenograft tumor with nodetected PIK3CA mutation was observed. Acute treatment with BKM led to a decrease in pAKTlevels.

Conclusion. Independent of PIK3CA gene mutation, BKM mediated inhibition of the PI3K/AKT/mTORpathway in endometrial tumors precludes tumor growth in a primary xenograft model. While a patternof resistance emerges, this effect appears to be mitigated by the addition of conventional cytotoxicchemotherapy.

© 2014 Elsevier Inc. All rights reserved.

Hospital, Vincent Center forand Gynecology, THR 901, 55

61.

19 Belmont St., Worcester, MA

Introduction

Endometrial cancer is the most common gynecologic malignancy inthe United States. In 2013, it is estimated that 47,130 American womenwill be diagnosed with endometrial carcinoma and 8010 women willnot survive their disease [1]. While the majority of these women will becuredof their disease [2], a subset (15–25%)whopresentwith carcinomas

347L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

of various histologic types and grades [3] are at increased risk for develop-ing metastatic disease, disease recurrence and chemotherapy resistance[3,4]. Unfortunately, conventional cytotoxic and radiation therapies arenot effective for all of these women leading to high mortality within thissubset of endometrial cancer patients [4,5]. These therapeutic limitationshighlight the distinct need for additional scientific investigation to under-stand the driving molecular pathways that may be amenable to targetedtherapy.

Correlative scientific investigations have shown that the phos-phatidylinositol 3-kinase (PI3K) pathway is one of the most altered on-cogenic pathways in cancer and that endometrial cancers have one ofthe highest rates of oncogenic activation in bothhigh and low risk endo-metrial cancer subtypes [6,7]. This signaling pathway regulates cell pro-liferation and differentiation.While three classes of PI3K enzymes havebeen described, class IA PI3Ks have beenmost associated with promot-ing carcinogenesis [6]. Class IA PI3K is activated by receptor tyrosinekinases and G-protein-coupled receptors and transfer phosphategroups to the inositol ring of phosphatidylinositol 4,5 bi-phosphate(PIP2) to produce the signaling molecule phosphatidylinositol 3,4,5tri-phosphate (PIP3). This process is negatively regulated by the phos-phatase and tensin homolog (PTEN) [6]. The underlying mechanismsfor PI3K pathway activation in cancer include receptor tyrosine kinaseactivation or amplification, gain of function mutation, deletion, silenc-ing of negative regulators of the PI3K pathway and activation or ampli-fication of downstream kinasemediators [8]. In endometrial cancer, themost commonmechanisms of PI3K pathway activation are loss of PTENtumor suppressor function and activating mutations in the catalyticPI3K subunit p110α encoded by the PIK3CA gene [9]. Alterations inPTEN and PIK3CA have been identified in 50–80% and 25–40% of endo-metrial cancers respectively across the spectrum of grade and histology[7,9].

Given the high prevalence of PI3K/AKT/mTOR pathway activa-tion [9], targeted strategies inhibiting this cascade could hold tre-mendous potential to benefit patients with endometrial cancer. Inaddition to genomic data supporting activation of the PI3K path-way in endometrial cancer [9], pre-clinical in vitro studies havedemonstrated that PI3K pathway inhibition can impair the prolifer-ation of endometrial carcinoma cells [10]. Collectively, these datahave provided rationale for the multiple phase I and II clinical trialsinvolving endometrial cancer patients that have tested agents thattarget the PI3K pathway [11]. Phase II trials of rapalogs that inhibitthe mammalian target of rapamycin (mTOR), a downstream medi-ator of the PI3K pathway, revealed both objective responses as wellas clinically significant disease stabilization supporting the impor-tant role of PI3K mediated signaling in driving endometrialcarcinoma [12,13]. In addition to the rapalogs, several otherclasses of PI3K pathway inhibitors including direct PI3K inhibitors,PI3K/mTOR dual inhibitors, and AKT inhibitors are in developmentfor treating endometrial cancer [6,14]. Early reports from phase Iclinical trial suggest that responses to single agent blockade havean approximately 30% prevalence, occur with or without gain offunction mutations in PIK3CA, and manifest limited clinical durabil-ity that will give way to the inevitable development of resistance[11,15].

The PI3K pathway is an attractive target for therapeutic interven-tion [7]. It is currently unclear how gain of function mutations inPIK3CA and/or inactivation of PTEN in endometrial tumors affect sen-sitivity to these inhibitors. In this investigation, we utilized patientderived xenografts (PDXs) from women surgically treated at our in-stitution to explore the pattern of response to pan-P13K inhibitionwith the agent NVP BKM-120 as both a single agent, and in concertwith conventional paclitaxel and carboplatin chemotherapy (P/C).We report that although all xenograft tumors manifest an initialtumorstatic response independent of PIK3CA mutation status, a con-sistent resistance pattern emerges that can be blunted by the addi-tion of cytotoxic chemotherapy.

Methods

Tumor collection and propagation in vivo

Excess samples of human endometrial tumorwere obtained throughan IRB-approved centralized banking infrastructure at our institution.Written informed consentwas received from all participants. All animalstudies were approved by our institution's Sub-Committee on Researchand Animal Care andwere in compliancewith the humane treatment ofresearch animals.

Single cell suspensions from enzymatically processed solid tumortissue were depleted of hematologic components, as previously de-scribed [16]. Cells were suspended in PBS:MatrigelH (1:1) and injectedsubcutaneously into 6–8week old female NOD/SCIDmice (Jackson Lab-oratory, Bar Harbor, ME). All animal experiments were conducted instrict accordance with institutional guidelines and the recommenda-tions in the Guide for the Care and Use of Laboratory Animals of the Na-tional Institutes of Health. Animals were euthanized when they hadexcessive tumor burden, excessive weight loss, or became moribund.

For continued propagation in mice, the xenograft tumors were ex-cised and enzymatically processed to a single cell suspension. The sus-pension was depleted of mouse H-2Kd+ cells using MACS beads(Miltenyi Biotech, Bergisch Gladbach, Germany) and tumor-derivedcells were re-injected subcutaneously into new recipient NOD/SCIDmice or propagated in cell culture. Every primary endometrial tumorutilized in this study underwent three to four passages in vivo with his-tologic features preserved over the serial transplantation process. For allof the in vivo experiments, the histology and grade of both the primaryendometrial tumor sample and the subsequent xenograft tumors wereevaluated by a pathologist.

Genotyping of endometrial tumors

An adapted version of the Applied Biosystems (ABI) Prism SNaPshotmultiplex system designed to screen for 209 well-characterized muta-tions in 22 oncogenes and tumor suppressor genes was used to geno-type 3 mm core samples from the primary tumor as previouslydescribed [17]. The xenografts derived from the analyzed primary tu-mors were also analyzed on the SNaPshot platform to confirm thatthey retained the original genotype.

Immunohistochemistry

PTEN status in all tumors was assessed by immunohistochemistry.Antigen retrievalwas carried out in citrate target retrieval buffer follow-ed by treatmentwith 3% hydrogen peroxide inmethanol and incubationin Protein Block (DAKO, Glostrup, Denmark). Sections were incubatedwith anti-human PTEN clone 6H2.1 antibody (Cascade Bioscience, Win-chester, MA) diluted at 1:100 in phosphate buffered saline (PBS) over-night at 4 °C followed by incubation with biotinylated IgG secondaryantibody (Vector Laboratories, Burlingame, CA) for 60min at room tem-perature. For visualization, slideswere treatedwith a Vectastain ABC re-agent prior to applying 3,39-diaminobenzidine chromagen (DAB) for2 min. Sections were then counterstained with hematoxylin,dehydrated and mounted [18].

Treatment with single agent NVP BKM-120 in vivo

In two independent experiments,mice bearing xenografts of either aPIK3CA mutant or PIK3CA wild type endometrioid endometrial tumorwere randomly divided into two cohorts with equivalent tumor vol-umes (250–600mm3). One cohort (n=6)was treatedwith the PI3K in-hibitor NVP BKM-120 (30 mg/kg) as a daily oral gavage and the secondcohort (n=6)was treatedwith vehicle only. The effect of treatment ontumor growth was assessed by regular monitoring of tumor volume.

Table 1The histological andmolecular characteristics of the eight prospectively collected primaryendometrial cancer specimens analyzed in these studies.

ENCA Histology Stage Grade Mutation PTEN

1 Endometrioid IB 2 PIK3CA (H1047L) Negative2 Carcinosarcoma IA 3 PIK3CA (R88Q) Positive4 Endometrioid IA 2 PIK3CA (R88Q) Negative5 Endometrioid IB 2 KRAS (G12D) Negative6 Endometrioid IIIC 2 None detected Negative7 Uterine serous carcinoma IVA 3 None detected Negative8 Carcinosarcoma IIIC 3 None detected Negative

348 L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

Potential toxicity was assessed by monitoring mouse weight over thecourse of treatment.

Generation of cell lines from xenograft tumors

ENCA1 (H1047L PIK3CAmutation) xenograft tumorswere harvestedfrom NVP BKM-120 and vehicle treated mice after the development ofresistance to NVP BKM-120. A subsection of each tumor was mincedinto b1 mm3 pieces, suspended in 10 ml DMEM:F12 medium supple-mented with insulin (5 μg/ml), bFGF (10 ng/ml), EGF (20 ng/ml) andBSA (0.4%), and incubated at 37 °C 95% CO2 and 5% O2.

Re injection of NVP BKM-120 resistant tumor cells

To test the stability of the NVP BKM-120 resistance phenotype, cellsderived from the non-resistant and resistant xenograft derived lineswere injected subcutaneously into NOD/SCID mice (1 × 106 cells permouse). Once tumors formed the mice were evenly distributed to twodifferent groups and treated with 30 mg/kg NVP BKM-120 daily as de-scribed. Tumor volume and mouse weight were measured twiceweekly.

Treatment of mice with NVP BKM-120, paclitaxel and carboplatin (P/C)in vivo

In three separate experiments, mice harboring xenografts derivedfrom a primary human endometrial tumor (ENCA2, ENCA4 or ENCA5)were randomly distributed into four cohorts (5–7 mice/cohort) withequivalent tumor volumes (250–600 mm3) and treated with either ve-hicle alone, NVP BKM-120 alone (30 mg/kg, daily, by oral gavage),carboplatin (50 mg/kg) and paclitaxel (15 mg/kg) alone once weeklyby intraperitoneal injection (IP), or the combination of NVP BKM-120and paclitaxel and carboplatin. Treatment effect on tumor growth andpotential toxicity were assessed as described.

Immunoblotting

Total protein lysates prepared from post-treatment xenograft sam-ples were electrophoresed on 10% Bis-Tris gels (Nupage Novex,Invitrogen) and transferred to polyvinylidene difluoride (PVDF) mem-branes (Millipore, Bedford). The membranes were incubated overnightat 4 °C in TBSTM (TBS, 0.1% Tween20 and 5% non-fat drymilk) with therelevant primary antibodies against AKT, phospho-AKT (Ser473),phospho-Akt (Thr308), PRAS40, phospho-PRAS40, p44/42MAPK(ERK1/2) and phospho-p44/42MAPK (ERK1/2) (Cell Signaling). Theblots were incubated with goat anti-rabbit horseradish peroxidase(HRP) at room temperature and developed with ECL-plus detection re-agent (GEHealthcare). Themembraneswere stripped using 1Mglycineand then re-probed with an antibody directed against either β-actin orGAPDH as loading controls.

Statistical analysis

Non-parametric statistical analysis using Wilcoxan rank-sum testsfor unpaired and sign-rank tests for paired data on tumor volumesand weights, as well as mouse weights was performed. A p value ofb0.05 was considered statistically significant. STATA (College Station,TX) v10 software was used for all statistical analysis.

Results

Summary of endometrial tumors used for xenograft experiments

This investigation utilized primary human endometrial carcinomatissue from eight patients whose characteristics are summarized inTable 1. A range of histology, stage, grade, PIK3CA mutation status and

PTEN status was observed in this cohort. All tumors were grade 2 or 3and comprised endometrioid, serous (USC) and carcinosarcoma histo-logical subtypes. Immunohistochemical analyses determined that onlyENCA2 (PIK3CA R88Q mutation) stained positively for PTEN (Supple-mental Fig. 1). The remaining seven tumors showed loss of PTEN ex-pression. None of the analyzed tumors had mutations at the PTENhotspots detected by the SNaPshot platform (data not shown).

Effects of acute treatment with NVP BKM-120 on PI3K/AKT/mTOR andMAPK signaling pathways

To confirm that we were targeting the PI3K/AKT/mTOR cell-signaling pathway, we assessed the acute effect of PI3K pathway inhibi-tion on downstream signaling events. Mice harboring xenograft tumorswere treatedwith either vehicle or single agent NVP BKM-120 adminis-tered at 30 mg/kg for 6 h and the phosphorylation status of AKT andPRAS40 was determined (Fig. 1). Regardless of PIK3CAmutation status,NVP BKM-120 treatment reduced relative levels of pAKT (Thr308) andpPRAS-40 in the treated xenograft tumors. In contrast, no change innon-phosphorylated AKT or PRAS-40 in response to treatment was ob-served (Fig. 1A and B). We also assessed the levels of phosphorylatedand non-phosphorylated ERK as PI3K can also signal through theMAPK pathway. As seen in Fig. 1C, we observed a change in the relativelevels of phosphorylated and non-phosphorylated ERK in only one ana-lyzed xenograft tumor suggesting that this pathwaywas not as affectedin the acute setting by inhibition of the PI3K/AKT/mTOR signalingpathway.

Effect of single agent NVP BKM-120 therapy on endometrial xenografttumors

Daily treatment of mice with NVP BKM-120 at 30 mg/kg resulted insignificant xenograft tumor growth suppression starting at days 5–10compared to the linear growth observed in vehicle treated tumors(p b 0.04). Following this initial tumorstatic response, tumor resurgencemanifested between days 14 and 25 (p b 0.03) in both endometrioidtumor single agent experiments (Fig. 2). The pattern of response and re-surgence was independent of PIK3CA mutation status. NVP BKM-120therapy was not associated with any change in mouse weights duringthe course of treatment.

Since chronic therapy resulted in tumor resurgence, we analyzedsignaling events downstream of PI3K inhibition to understand the mo-lecular effects associated with acquired resistance. To accomplish this,tumors were harvested from mice 6 h after their last treatment whenthere was evidence of resistance. In contrast to what we observed fol-lowing acute treatment, immunoblotting demonstrated no differencein the relative levels of phosphorylated AKT (Thr308), phosphorylatedPRAS40 or phosphorylated ERK after chronic treatment with NVPBKM-120 (Supplemental Fig. 2).

We derived a cell line from anENCA1 (PIK3CAH1047Lmutation) xe-nograft tumor that acquired resistance to NVP BKM-120 over the in vivotreatment period and from an ENCA1 vehicle treated tumor. Mice bear-ing xenograft tumors generated following re-injection of these NVP

Fig. 1. NVP BKM-120 suppresses PI3K/AKT/mTOR signaling in endometrial cancer tumorxenografts. Mice bearing endometrial cancer tumor xenografts from samples ENCA4, 5,7 and 8 were randomized into two cohorts with matched tumor volumes. Tumors weretreatedwith either vehicle or NVP BKM-120 and harvested 6 h after treatment. Protein ly-sates from the xenografts were evaluated by Western blot to determine total and phos-phorylated protein levels. NVP BKM-120 treatment led to on target decreases inphosphorylated isoforms of AKT and PRAS-40. Only one xenograft tumordemonstrated al-tered phosphoERK expression.

Fig. 2. NVP BKM-120 alone impedes endometrial cancer tumor growth. Mice bearing endometumes and treatedwith either vehicle orNVPBKM-120 (30 mg/kg) by oral gavage. Tumor volumat the beginning of the experiment. Error bars represent the standard error of the mean (S.E.M

349L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

BKM-120 resistant cells did not demonstrate tumorstasis in response toNVP BKM-120 treatment in vivo (Fig. 3).

Effects of NVP BKM-120 in combination with paclitaxel and carboplatin onendometrial xenograft tumors

Single agent NVP BKM-120 treatment manifested ultimate loss oftumorstasis in our primary human xenograft model and it was unclearif the addition of P/C could overcome this resistance pattern. The threeexperiments depicted in Fig. 4 demonstrate that the addition of P/C toNVP BKM-120 therapy significantly abrogated the observed resistancepattern of three separate xenograft tumors (p b 0.05). NVP BKM-120and P/C therapy exhibited synergistic tumor growth suppression(p b 0.05) in ENCA5 (PIK3CA wild type). In contrast, treatment withNVP BKM-120 + P/C or P/C alone resulted in similar tumorstasis inmice harboring ENCA4 (PIK3CA R88Q mutation) or ENCA2 (PIK3CAR88Q mutation) derived xenograft tumors.

Discussion

The PI3K signaling pathway contributes to the pathology of a signif-icant subset of endometrial cancers across histology and grade [7,11,14].Targeting the PI3K pathway is likely to hold significant potential for in-novating the standard of care for women with metastatic or recurrentendometrial cancer, and early phase clinical trials support that re-sponses are prevalent [11]. We utilized PDXs with known PTEN andPIK3CA mutation status to better understand the tumor response pat-terns following treatment with a pan-PI3K inhibitor, NVP BKM-120. Inrecent years, numerous reports have confirmed that endometrial can-cers across grade and histology harbor a complex molecular landscapecharacterized by numerous and often synchronous alterations in manyoncogenic pathways [9,19]. Specifically, endometrial tumors commonlypresent with therapeutically relevant mutations or alterations inPIK3CA, PTEN and/or KRAS making strategies that target the PI3K/AKT/mTOR and MAPK pathways attractive [14]. Pre-clinical and clinical ex-periences have suggested that inhibiting one point within this complexnetwork will commonly lead to inevitable resistance secondary to theintricate crosstalk these two pathways exhibit [20–22]. These data con-firm that single agent inhibition of PI3Kwith a pan inhibitor does initial-ly manifest an in vivo response. A pattern of resistance, however,consistently emerged and appeared to be durable as it was maintainedin subsequent in vitro and in vivo experiments. Interestingly, theacute on target disruption of phosphorylation of AKT and PRAS40 ob-served with pan-PI3K inhibition was lost after chronic therapy suggest-ing a restoration of oncogenic signaling. The behavior manifested inthese PDXmodels echoes early reports fromphase I clinical trials testingpipeline PI3K inhibitors that demonstrate high response rates with lim-ited progression free survival due to tumor resurgence [11]. While theexact mechanism of resistance is unclear in endometrial cancer, data

rial cancer tumor xenografts were randomized into two cohorts with matched tumor vol-eswere assessed regularly. Change in tumor volume is shown relative to the tumor volume.). Left panel: ENCA1, PIK3CA H1047L mutation; right panel: ENCA6, wild type PIK3CA.

Fig. 3. Established NVP BKM-120 resistance persists in subsequent in vivo passages. Cellscultured from an ENCA1 xenograft tumor that acquired resistance to NVP BKM-120were injected into NOD/SCID mice to form xenograft tumors. The mice were randomizedinto two cohorts (n = 5) and received either NVP BKM-120 or vehicle. As shown, xeno-grafts derived from the NVP BKM-120 resistant cells failed to manifest any response toPI3K inhibition in vivo. Percent change in tumor volume represents the change in tumorvolume relative to the tumor volume at the beginning of the experiment. Error bars repre-sent the standard error of the mean (S.E.M.).

350 L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

in other disease sites would strongly suggest alternate signalingthrough related pathways [20,21,23].

The success of targeted therapy depends upon the development ofbiomarkers that can identify which patients are most likely to respond[24,25]. Based upon in vitro data, investigators have suggested thatgain of function mutations in PIK3CA may associate with heightenedsensitivity to PI3K inhibitors [26,27]. In our primary human xenograftmodel, the presence of gain of function mutations in PIK3CA did not in-fluence the initial tumorstatic response or pattern of subsequent resis-tance to PI3K inhibition. Early phase clinical trials testing PI3Kinhibitors in solid tumors including endometrial cancer have alsoshown that PIK3CA mutations do not appear to associate with survivalor response [28,29], though one report suggested that the H1047RPIK3CAmutationwas associatedwith response [15]. Themajority of en-dometrial tumors in this investigation had evidence of PTEN loss, andthis may in part account for the responses to pan-PI3K inhibition ob-served independent of PIK3CA mutation status. In endometrial cancer,where alterations in PTEN function have been reported in up to 91% oftumors [9], the lack of negative PI3K regulation may result in a baselineactivation of the PI3K/AKT/mTOR pathway that may associate with re-sponse to PI3K inhibition even in the absence of PIK3CA gain of function

Fig. 4.NVPBKM-120 in combinationwith paclitaxel and carboplatin (P/C) abrogates thepatterncancer xenograft tumors were randomized into four cohorts (n= 4–6 animals per treatment cohicle), NVPBKM-120-alone, paclitaxel and carboplatin (P/C) alone orNVPBKM-120+P/C as indthe tumor volume at the beginning of the experiment. Error bars represent the standard error ofmutation; right panel: ENCA2, PIK3CA R88Qmutation. In all experiments, the addition of P/C bluBKM-120 blunted the development of resistance observed in the NVP BKM-120 alone arm in agraft lines (ENCA5).

mutations. Another intriguing possibility is that the array of PIK3CAdriver mutations are cancer site specific and responses may be mediat-ed by a more complex network of factors of which PIK3CA mutation isjust one [30].While modeling the genomic heterogeneity of endometri-al cancer is a formidable challenge, clinical response data is anticipatedfrom several actively enrolling phase II trials testing single agent NVPBKM-120, amongst other PI3K pathway inhibitors (NCT01420081,NCT01226316, NCT01550380, NCT01307631).

In this endometrial cancer setting marked by simultaneous activat-ing mutations in oncogenes and loss of tumor suppressor function,both pan and α-selective PI3K inhibitors are being investigated. Our ra-tionale for selecting a pan PI3K inhibitor was based upon in vitro expe-rience that suggested that inhibition of p110β is required to reverse theloss of PTEN negative regulation of PI3K [31]. While this interaction iscontroversial, recent in vitro data testing various PI3K inhibitors in aset of 24 endometrioid endometrial cancer cell lines observed that panPI3K inhibition led to more pronounced anti-proliferative effects whencompared to p110α- or p110β selective inhibitors in PTEN mutantcells [26]. Interestingly, a recent investigation of cell lines from serouscarcinomas of the uterus demonstrated that in those tumors withPIK3CA mutations, α-selective inhibition produced dose dependentanti-proliferative effects [27]. Currently, it is not known whether ornot a pan or α-selective PI3K inhibitor will be more potent in endome-trial cancer or if PTEN status can serve as a biomarker to better predictsensitivity. Our data confirm that NVP BKM-120 induced significant in-hibition of downstream PI3K pathway signaling suggesting on target ef-fects, but after the development of resistance, this signaling wasrestored. It remains to be elucidated if anα-selective PI3K inhibitor pro-duces a different pattern of sensitivity or resistance in vivo.

Many investigators have hypothesized that tumor resistance to sin-gle agent targeted therapies is an inevitable event resulting frominteracting pathways that can compensate and circumvent isolatedpoints of inhibition [32]. Combination therapies that impede both fun-damental oncogenic pathways and the primary escape pathwayswould therefore hold the promise of synergistic tumor inhibition. Inthis investigation, the observed pattern of NVP BKM-120 resistancewas abrogated when NVP BKM-120 therapy was combined with con-ventional P/C chemotherapy. While synergistic activity was observedin only one of the three primary human xenograft tumors analyzed,the lack of characteristic tumor resurgence in all three suggests thatusing DNA damaging agents with a pan-inhibitor induces complemen-tary anti-tumor efficacy. One explanation for this effective combinationtherapy was recently explored using breast cancer cells treated with

of resistance seenwith isolated PI3Kpathway inhibition.Mice bearing humanendometrialhort) with matched tumor volumes. Themice were treatedwith either vehicle alone (ve-icated. Percent change in tumor volume represents the change in tumor volume relative tothemean (S.E.M.). Left panel: ENCA5, PIK3CAwild type; center panel: ENCA4, PIK3CAR88Qnted the development of characteristic single agent resistance. The addition of P/C to NVP

ll experiments, and demonstrated synergistic anti-tumor activity in one of the three xeno-

351L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

NVP BKM-120 [33,34]. Juvekar et al. demonstrated the PI3K inhibitionwith NVP BKM-120 induced defective homologous DNA repair (HR)that led to a dose dependent increase in overall poly ADP-ribose po-lymerase (PARP) activity. As a result of these alterations, less effi-cient base excision repair mechanisms were activated as has beenobserved in BRCA deficient cells [34]. In the setting of defectiveDNA repair, conventional cytotoxics have been shown to manifestheightened sensitivity [35]. NVP BKM-120 induction of a BRCA-likephenotype has yet to be validated in endometrial cancer, but this ex-perience offers a plausible explanation for both the more durable ef-ficacy in all models and the observed synergy with conventional P/Cchemotherapy we observed in only one of our models that meritsfurther investigation.

The emerging role of PTEN as a “custodian” of genomic stability mayalso play a role in the activity observed in this investigation [36]. Recentinvestigations have observed that PTEN loss of function led to pro-nounced defects in HR and upregulated PARP compared to tumorswith functional PTEN [37,38]. Given the high prevalence of PTEN inacti-vation in endometrial cancer, this loss could not only work to activatethe PI3K pathway, but could also promote genomic instability makingthese tumors more vulnerable to dual therapy with NVP BKM-120 andchemotherapy. Importantly, we utilized a primary carcinosarcoma(ENCA2, PIK3CA R88Q mutation) in one of the four arm experimentswith NVP BKM-120 and P/C because this PDX line exhibited intactPTEN expression.While a pattern of abrogated resistancewas observed,the degree of tumor control with P/C or P/C + NVP BKM-120 was lessthan that observed in the other two primary endometrial xenografts.These pilot data suggest that PTEN statusmay be an important factor as-sociated with response to combined PI3K inhibition and cytotoxictherapy.

Collectively, these consistent observations support that PI3K inhibi-tion with the pan class IA inhibitor NVP BKM-120 induced primaryhuman xenograft tumor growth suppressionwith the ultimate develop-ment of resurgent tumor growth over time. This pattern of responsewas observed in xenografts independent of the presence of a PIK3CAgene mutation and resistance was maintained in vivo. The addition ofP/C to NVP BKM-120 resulted in a loss of this resistance pattern, andin one line, NVP BKM-120 + P/C showed superior anti-tumor activitywhen compared to P/C alone. Because of highly prevalent alterationsin the PI3K pathway in endometrial tumors, targeted PI3K inhibitionhas the potential for improving the therapeutic options for womenwith metastatic or recurrent disease. Future investigations examin-ing the mechanisms of resistance to PI3K inhibition will be crucialto optimizing combination therapy and selecting those endometrialcancer patients most likely to gain benefit from targeted PI3Kinhibition.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ygyno.2014.02.022.

Conflict of interest statement

The authors declare no conflict of interest with regard to the work presented in this man-uscript. The NVP BKM-120 was supplied by the Novartis Oncology.

Acknowledgments

Funding for this project was provided by the Advanced Medical Re-search Fund (BRR) and Vincent Memorial funds (BRR). We appreciatethe efforts of Minji Kim for conducting protein analysis. In addition,we would like to acknowledge the contributions of the clinical researchcoordinators (Celeste DiGloria and Virginia Byron) who were responsi-ble for educating and consenting our gynecologic patients. Jolijn W.Groeneweg, MD received financial support via a scholarship awardedby the VSB Foundation. Finally, we want to give special thanks to allthe patients who have consented to allow their malignant tumors tobe part of our research efforts.

References

[1] Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin2012;62:10–29.

[2] Robbe EJ, Pijnenborg JM. Survival and recurrent disease after postoperative radio-therapy for early endometrial cancer: systematic review and meta-analysis. BJOG2008;115:926–7 [author reply 7].

[3] Rose PG. Endometrial carcinoma. N Engl J Med 1996;335:640–9.[4] Humber CE, Tierney JF, Symonds RP, CollingwoodM, Kirwan J, Williams C, et al. Che-

motherapy for advanced, recurrent or metastatic endometrial cancer: a systematicreview of Cochrane collaboration. Ann Oncol 2007;18:409–20.

[5] Hamilton CA, CheungMK, Osann K, Chen L, Teng NN, Longacre TA, et al. Uterine pap-illary serous and clear cell carcinomas predict for poorer survival compared to grade3 endometrioid corpus cancers. Br J Cancer 2006;94:642–6.

[6] Courtney KD, Corcoran RB, Engelman JA. The PI3K pathway as drug target in humancancer. J Clin Oncol 2010;28:1075–83.

[7] Dedes KJ, Wetterskog D, Ashworth A, Kaye SB, Reis-Filho JS. Emerging therapeutictargets in endometrial cancer. Nat Rev Clin Oncol 2011;8:261–71.

[8] Yuan TL, Cantley LC. PI3K pathway alterations in cancer: variations on a theme. On-cogene 2008;27:5497–510.

[9] Kandoth C, Schultz N, Cherniack AD, Akbani R, Liu Y, Shen H, et al. Integrated geno-mic characterization of endometrial carcinoma. Nature 2013;497:67–73.

[10] Shoji K, Oda K, Kashiyama T, Ikeda Y, Nakagawa S, Sone K, et al. Genotype-dependent efficacy of a dual PI3K/mTOR inhibitor, NVP-BEZ235, and an mTOR in-hibitor, RAD001, in endometrial carcinomas. PLoS One 2012;7:e37431.

[11] Janku F, Wheler JJ, Naing A, Stepanek VM, Falchook GS, Fu S, et al. PIK3CAmutationsin advanced cancers: characteristics and outcomes. Oncotarget 2012;3:1566–75.

[12] Oza AM, Elit L, Tsao MS, Kamel-Reid S, Biagi J, Provencher DM, et al. Phase II study oftemsirolimus in women with recurrent or metastatic endometrial cancer: a trial ofthe NCIC Clinical Trials Group. J Clin Oncol 2011;29:3278–85.

[13] Slomovitz BM, Lu KH, Johnston T, Coleman RL, Munsell M, Broaddus RR,et al. A phase 2 study of the oral mammalian target of rapamycin inhib-itor, everolimus, in patients with recurrent endometrial carcinoma. Cancer2010;116:5415–9.

[14] Slomovitz BM, Coleman RL. The PI3K/AKT/mTOR pathway as a therapeutic target inendometrial cancer. Clin Cancer Res 2012;18:5856–64.

[15] Janku F, Wheler JJ, Naing A, Falchook GS, Hong DS, Stepanek V, et al. PIK3CA muta-tion H1047R is associated with response to PI3K/AKT/mTOR signaling pathway in-hibitors in early phase clinical trials. Cancer Res 2013;73:276–84.

[16] Curley MD, Therrien VA, Cummings CL, Sergent PA, Koulouris CR, Friel AM, et al.CD133 expression defines a tumor initiating cell population in primary human ovar-ian cancer. Stem Cells 2009;27:2875–83.

[17] Dias-Santagata D, Akhavanfard S, David SS, Vernovsky K, Kuhlmann G, Boisvert SL,et al. Rapid targeted mutational analysis of human tumours: a clinical platform toguide personalized cancer medicine. EMBO Mol Med 2010;2:146–58.

[18] Garg K, Broaddus RR, Soslow RA, Urbauer DL, Levine DA, Djordjevic B. Pathologicscoring of PTEN immunohistochemistry in endometrial carcinoma is highly repro-ducible. Int J Gynecol Pathol 2012;31:48–56.

[19] Oda K, Stokoe D, Taketani Y, McCormick F. High frequency of coexistent mu-tations of PIK3CA and PTEN genes in endometrial carcinoma. Cancer Res2005;65:10669–73.

[20] Carracedo A, Pandolfi PP. The PTEN-PI3K pathway: of feedbacks and cross-talks. On-cogene 2008;27:5527–41.

[21] O'Reilly KE, Rojo F, She QB, Solit D, Mills GB, Smith D, et al. mTOR inhibition inducesupstream receptor tyrosine kinase signaling and activates Akt. Cancer Res2006;66:1500–8.

[22] Elkabets M, Vora S, Juric D, Morse N, Mino-Kenudson M, Muranen T, et al. mTORC1inhibition is required for sensitivity to PI3K p110alpha inhibitors in PIK3CA-mutantbreast cancer. Sci Transl Med 2013;5:196ra99.

[23] Mendoza MC, Er EE, Blenis J. The Ras-ERK and PI3K-mTOR pathways: cross-talk andcompensation. Trends Biochem Sci 2011;36:320–8.

[24] Salvesen HB, Haldorsen IS, Trovik J. Markers for individualised therapy in endome-trial carcinoma. Lancet Oncol 2012;13:e353–61.

[25] Juric D, Baselga J. Tumor genetic testing for patient selection in phase I clinical trials:the case of PI3K inhibitors. J Clin Oncol 2012;30:765–6.

[26] Weigelt B, Warne PH, Lambros MB, Reis-Filho JS, Downward J. PI3K pathwaydependencies in endometrioid endometrial cancer cell lines. Clin Cancer Res2013;19:3533–44.

[27] English DP, Bellone S, Cocco E, Bortolomai I, Pecorelli S, Lopez S, et al. OncogenicPIK3CA genemutations and HER2/neu gene amplifications determine the sensitivityof uterine serous carcinoma cell lines to GDC-0980, a selective inhibitor of Class I PI3kinase and mTOR kinase (TORC1/2). Am J Obstet Gynecol 2013;209:465.e1–9.

[28] Tredan O, Treilleux I, Wang Q, Gane N, Pissaloux D, Bonnin N, et al. Predicting evero-limus treatment efficacy in patients with advanced endometrial carcinoma: aGINECO group study. Target Oncol 2012;8:243–51.

[29] Bendell JC, Rodon J, Burris HA, de Jonge M, Verweij J, Birle D, et al. Phase I, dose-escalation study of BKM120, an oral pan-Class I PI3K inhibitor, in patients with ad-vanced solid tumors. J Clin Oncol 2012;30:282–90.

[30] Ashworth A, Lord CJ, Reis-Filho JS. Genetic interactions in cancer progression andtreatment. Cell 2011;145:30–8.

[31] Wee S, Wiederschain D, Maira SM, Loo A, Miller C, deBeaumont R, et al. PTEN-deficient cancers depend on PIK3CB. Proc Natl Acad Sci U S A 2008;105:13057–62.

[32] McCubrey JA, Steelman LS, Chappell WH, Abrams SL, Franklin RA, Montalto G, et al.Ras/Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR cascade inhibitors: how mutations canresult in therapy resistance and how to overcome resistance. Oncotarget2012;3:1068–111.

352 L.S. Bradford et al. / Gynecologic Oncology 133 (2014) 346–352

[33] Ibrahim YH, Garcia-Garcia C, Serra V, He L, Torres-Lockhart K, Prat A, et al. PI3K in-hibition impairs BRCA1/2 expression and sensitizes BRCA-proficient triple-negativebreast cancer to PARP inhibition. Cancer Discov 2012;2:1036–47.

[34] Juvekar A, Burga LN, Hu H, Lunsford EP, Ibrahim YH, Balmana J, et al. Combining aPI3K inhibitor with a PARP inhibitor provides an effective therapy for BRCA1-related breast cancer. Cancer Discov 2012;2:1048–63.

[35] Clark CC, Weitzel JN, O'Connor TR. Enhancement of synthetic lethality via combina-tions of ABT-888, a PARP inhibitor, and carboplatin in vitro and in vivo using BRCA1and BRCA2 isogenic models. Mol Cancer Ther 2012;11:1948–58.

[36] Shen WH, Balajee AS, Wang J, Wu H, Eng C, Pandolfi PP, et al. Essential role for nu-clear PTEN in maintaining chromosomal integrity. Cell 2007;128:157–70.

[37] Dedes KJ, Wetterskog D, Mendes-Pereira AM, Natrajan R, Lambros MB, Geyer FC,et al. PTEN deficiency in endometrioid endometrial adenocarcinomas predicts sensi-tivity to PARP inhibitors. Sci Transl Med 2010;2:53ra75.

[38] Forster MD, Dedes KJ, Sandhu S, Frentzas S, Kristeleit R, Ashworth A, et al. Treatmentwith olaparib in a patient with PTEN-deficient endometrioid endometrial cancer.Nat Rev Clin Oncol 2011;8:302–6.