Potentiation of Carboplatin-Mediated DNADamage by the Mdm2 Modulator Nutlin-3a in aHumanized Orthotopic Breast-to-Lung MetastaticModelEva Tonsing-Carter1, Barbara J. Bailey2,3, M. Reza Saadatzadeh3,4, Jixin Ding3,4,Haiyan Wang3, Anthony L. Sinn2, Kacie M. Peterman2, Tiaishia K. Spragins2,Jayne M. Silver2, Alyssa A. Sprouse3, Taxiarchis M. Georgiadis5, T. Zachary Gunter5,Eric C. Long5, Robert E. Minto5, Christophe C. Marchal3, Christopher N. Batuello3,Ahmad R. Safa1, Helmut Hanenberg3,6,7, Paul R. Territo8, George E. Sandusky9,Lindsey D. Mayo3,10, Christine M. Eischen11, Harlan E. Shannon3, and Karen E. Pollok1,2,3
Abstract
Triple-negative breast cancers (TNBC) are typically resistant totreatment, and strategies that build upon frontline therapy areneeded. Targeting themurine doubleminute 2 (Mdm2) protein isan attractive approach, as Mdm2 levels are elevated in manytherapy-refractive breast cancers. The Mdm2 protein–proteininteraction inhibitor Nutlin-3a blocks the binding of Mdm2 tokey signaling molecules such as p53 and p73a and can result inactivation of cell death signaling pathways. In the present study,the therapeutic potential of carboplatin and Nutlin-3a to treatTNBC was investigated, as carboplatin is under evaluation inclinical trials for TNBC. In mutant p53 TMD231 TNBC cells,carboplatin and Nutlin-3a led to increased Mdm2 and wasstrongly synergistic in promoting cell death in vitro. Furthermore,sensitivity of TNBC cells to combination treatment was depen-dent on p73a. Following combination treatment, gH2AX
increased and Mdm2 localized to a larger degree to chromatincompared with single-agent treatment, consistent with previousobservations that Mdm2 binds to the Mre11/Rad50/Nbs1complex associated with DNA and inhibits the DNA damageresponse. In vivo efficacy studies were conducted in theTMD231 orthotopic mammary fat pad model in NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice. Using an intermittent dosingschedule of combined carboplatin and Nutlin-3a, there was asignificant reduction in primary tumor growth and lung metas-tases compared with vehicle and single-agent treatments. Inaddition, there was minimal toxicity to the bone marrow andnormal tissues. These studies demonstrate that Mdm2 holdspromise as a therapeutic target in combination with conventionaltherapy and may lead to new clinical therapies for TNBC. MolCancer Ther; 14(12); 2850–63. �2015 AACR.
IntroductionThe development of efficacious therapies for triple-negative
breast cancer (TNBC) has been challenging due to its aggressivenature and lack of hormone receptors that can be therapeuticallytargeted (1, 2). Platinumagents, such as cisplatin and carboplatin,form DNA–platinum adducts resulting in intra- and interstrand
DNA crosslinks leading to increased double-strand breaks andcell death (3). Clinical studies have indicated that platinum-based therapy can provide enhanced efficacy in TNBC (4, 5).Furthermore, clinical trials are currently evaluating the use ofcombination therapies that include carboplatin to specificallytreat TNBCs with metastases (NCT01881230, NCT00691379,
1Department of Pharmacology and Toxicology, Indiana UniversitySchool of Medicine, Indianapolis, Indiana. 2In Vivo Therapeutics Core,Indiana University Melvin and Bren Simon Cancer Center, IndianaUniversity School of Medicine, Indianapolis, Indiana. 3Department ofPediatrics, Herman B Wells Center for Pediatric Research, IndianaUniversity School of Medicine, Indianapolis, Indiana. 4GoodmanCampbell Brain and Spine, Department of Neurosurgery, Indiana Uni-versity School of Medicine, Indianapolis, Indiana. 5Department ofChemistry and Chemical Biology, Indiana University-Purdue Univer-sity Indianapolis, Indianapolis, Indiana. 6Department of Medical &Molecular Genetics, Indiana University School of Medicine, Indiana-polis, Indiana. 7Department of Otorhinolaryngology and Head/NeckSurgery, Heinrich Heine University, D€usseldorf, Germany. 8IndianaInstitute for Biomedical Sciences Imaging, Department of Radiology,Indiana University School of Medicine, Indianapolis, Indiana. 9Depart-ment of Pathology, Indiana University School of Medicine, Indianapo-lis, Indiana. 10Department of Biochemistry and Molecular Biology,Indiana University School of Medicine, Indianapolis, Indiana. 11Depart-
ment of Pathology, Microbiology and Immunology,Vanderbilt Univer-sity Medical Center, Nashville, Tennessee.
Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).
Current address for T.Z. Gunter: Eli Lilly and Company, Product Design andDevelopability, Indianapolis, Indiana; and current address for H. Hanenberg:Department of Pediatrics III, University Children's Hospital Essen, UniversityDuisburg-Essen, Essen, Germany.
Corresponding Author: Karen E. Pollok, Indiana University School of Medicine,Cancer Research Institute R4 302, 1044 W. Walnut St, Indianapolis, IN 46202.Phone: 317-274-8891; Fax: 317-274-8046; E-mail: [email protected]
and NCT01281150; www.clinicaltrials.gov). While platinumagents show promise in the clinic, new combination treatmentmodalities are needed that optimize tumor cell kill withoutenhancing platinum-mediated toxicity to normal tissues, therebyincreasing the therapeutic window.
Targeting the mouse double minute 2 (Mdm2) signaling axisholds promise as a novel approach to combination therapy in thetreatment of TNBC. Mdm2 is a multifaceted protein involved innumerous aspects of cell growth, survival, and invasion (6).Mdm2 is an E3 ubiquitin ligase that targets p53 for degradationby the proteasome (7) and is expressed in a number of tumortypes. There is mounting evidence demonstrating the critical roleMdm2 plays in cell growth regulation and cancer (6). In breastcancers, studies have shown that MDM2 gene amplificationsrange from 10% to 60% with some studies indicating a worseprognosis with overexpression of Mdm2 or gene amplification(6, 8). Moreover, evidence suggests that high Mdm2 levels cor-relate with a higher incidence of metastasis in vivo (6, 8, 9).
Nutlin-3a is a small molecule that binds to the N-terminalhydrophobic pocket of Mdm2 and blocks protein–protein inter-actions (PPI) between Mdm2 and p53 (10). Nutlin-3a alsoinhibits the binding of the p73 isoform p73a, as well as E2F1and hypoxia-inducible factor 1a (Hif-1a), to the N-terminalhydrophobic pocket of Mdm2 (11–13). It is estimated thatp53 is mutated in approximately 50% of all cancers with 60%of TNBCs bearing mutations in p53 (14, 15); in contrast, p73 israrely mutated in cancers (14, 16). p73 is a member of the p53family of tumor suppressors and has similar transactivationfunctions relating to the induction of proapoptotic genes inresponse to cellular stress (17, 18). Lau and colleagues showedthat when cells were treated with Nutlin-3a, the binding of p73ato Mdm2 was inhibited, leading to p73a-mediated induction ofproapoptotic downstream targets and increased apoptosis in cellslacking wild-type p53 (12). The use of Nutlin-3a to inhibit PPIsbetweenMdm2and binding partners, including p53, p73a, E2F1,and Hif-1a (10–13), may lead to a multitargeted approach totreating cancer, especially when coupled with clinically relevantDNA-damaging drugs such as carboplatin.
The purpose of the present study was to investigate thetherapeutic potential of modulating Mdm2 function in thecontext of carboplatin-mediated DNA damage using an opti-mized human mutant p53 TNBC orthotopic xenograft model.In vitro, carboplatin and Nutlin-3a were strongly synergistic inincreasing cell death in TNBC cells with a mutant p53 back-ground. Studies using siRNA indicated that drug sensitivity,as well as Mdm2 protein levels, were dependent on p73a.Following combination treatment in our model system, therewas increased Mdm2 localized to chromatin compared withsingle-agent treatment. In vivo efficacy experiments, conductedin the TMD231 orthotopic mammary fat pad model in NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice, demonstrated a statisti-cally significant reduction in primary tumor volume as well aslung metastases with significantly increased probability ofsurvival compared with vehicle and single-agent treatments.In regards to normal tissue toxicity, body weights were onlytransiently affected and recovered to normal levels after treat-ment. Treatment-mediated decreases in hematopoietic functionwere similar in mice treated with carboplatin alone or combi-nation Nutlin-3a/carboplatin, and these decreases did not leadto bone marrow aplasia. The present study demonstrates thatMdm2 is a valid therapeutic target in mutant p53 TNBC and
that Mdm2 PPI inhibitors offer new avenues for exploringnovel combination therapies for treatment of TNBC.
Materials and MethodsCells and cell culture
MDA-MB-231 (HTB-26), MDA-MB-468 (HTB-132), MCF10A(CRL-10317), and MCF-7 (HTB-22) were purchased from ATCCin 2010 (MCF-7), 2012 (MDA-MB-468), and 2014 (MDA-MB-231 and MCF10A). Primary human fibroblast cells were kindlyprovided by Kathrin Scheckenbach in 2014. TMD231 cellswere obtained from Dr. Harikrishna Nakshatri in 2009 and area derivative of theMDA-MB-231 cell line (19). TMD231 cells weretransduced with E2-Crimson lentiviral vector, p2CL7CR2wo(TMD231-CR), for in vivo imaging as described (20). Uponreceipt, cell stocks were cryopreserved at low passage. Authenti-cation of molecular profiles was verified by STR analysis (IDEXXBioResearch), and cells tested negative for mycoplasma. TMD231cells were cultured in MEM-a (Gibco) supplemented with 10%FBS (Atlanta Biologicals) and 1%HEPES (Invitrogen). MDA-MB-231, MDA-MB-468, and primary human fibroblast cells werecultured in DMEM (Gibco) supplemented with 10% FBS (AtlantaBiologicals). MCF10A cells were cultured inMedium171 (Gibco)supplemented with 1% MEGS (Invitrogen) and 0.1% choleratoxin (Sigma). All cells were cultured at 37�C with 5% CO2.
CompoundsNutlin-3a was synthesized at the IUPUI Chemical Synthesis
and Organic Drug Lead Development Core and confirmedthrough HPLC-MS analysis. Carboplatin was purchased fromSigma. For in vitro studies, carboplatin was dissolved in H2O,whereasNutlin-3awas dissolved inDMSO. Final concentration ofDMSO was less than 0.2%. For in vivo studies, carboplatin wasdissolved in PBS, whereas Nutlin-3a was suspended in 0.5%methylcellulose (Sigma) and 0.05% Tween-80 (Sigma).
Methylene blue proliferation assayThe methylene blue proliferation assay was derived fromOliver
and colleagues (21). Cells were treatedwith varying concentrationsofNutlin-3a, carboplatin,or combinationdose ratios for3 to5daysin triplicate or sextuplicate. IC50 valueswere calculated according tothe linearization method of Chou and Talalay (22) and were usedto construct isobolograms as previously described (23).
Clonogenic proliferation assayTMD231 cells were plated at low cell density and treated with
Nutlin-3a, carboplatin, or a 1:1 combination and colonies enu-merated at day 14.
Cell counting proliferation assayCells were treated with Nutlin-3a, carboplatin, 1:1 Nutlin-3a/
carboplatin, or appropriate vehicle control. Each experiment wasconducted in triplicate. The number of live cells was determinedin the presence of trypan blue dye via hemocytometer.
In vitro analysis of activated caspase-3/7, gH2AX, Mdm2, andcleaved PARP
For in vitro studies, activated caspase-3/7 was measured usingApoTox-Glo Triplex Assay (Promega) as per manufacturer'sinstructions. Formeasurement of gH2AX foci, TMD231 cells wereseeded on chamber slides (Lab-Teck Brand Products) and treated
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2851
the next day. Cells were fixed with 2% paraformaldehyde, stainedwith a fluorescein isothiocyanate (FITC)–conjugated phospho-histone H2AX (Ser139) primary antibody (1:200 dilution, CellSignaling), incubated with 40,6-diamidino-2-phenylindole(DAPI), and analyzed as described in the Supplementary Materi-als and Methods. gH2AX, Mdm2, and cleaved PARP were mea-sured using MILLIPLEX MAP gH2AX/b-Tubulin, Mdm2/b-Tubu-lin, or PARP/GAPDH Magnetic Beads (EMD Millipore) as permanufacturer's instructions. Mean fluorescence intensities (MFI)per mg protein were normalized to GAPDH or b-tubulin MFI.
Annexin V and 7-AAD apoptosis assayTMD231 cells were treated for 96 hours with the IC50 concen-
trations of carboplatin or Nutlin-3a alone, or in 1:0.3, 1:1, or 1:3dose ratios of Nutlin-3a:carboplatin in triplicate. Cells werecollected and stained using Annexin V-FITC (BD Biosciences)and 7-AAD (BD Biosciences) per manufacturer's instructions andanalyzed by flow cytometry.
Western blot analysisCells were lysedwith RIPA buffer containing 1Complete-EDTA
freemini tablet (Roche), and 1%phosphatase inhibitor 3 (Sigma;refs. 24, 25). Protein was quantified using the DC Protein Assay(Bio-Rad) per manufacturer's instructions. Western blot densi-tometry was determined using ImageJ software (http://imagej.nih.gov/), and proteins of interest were normalized to loadingcontrol and expressed relative to untreated, vehicle, or siControlcells.
AntibodiesFor in vitro and in vivo Western blot analyses, Mdm2 (90-kDa
band) antibody cocktail included SMP14 (sc-965, Santa Cruz),2A9 (OP155T, Calbiochem), and 4B11 (OP143, Calbiochem).Antibodies for PUMA (21 kDa, #4976, Cell Signaling), p21 (21-kDa, DCS60, Cell Signaling), p53, (DO-1, sc-126, Santa Cruz),MdmX (55 kDa, ab154324, Abcam), gH2AX (17 kDa, #2577, CellSignaling),a-tubulin (50 kDa, clone B-5-1-2, Sigma-Aldrich), andGAPDH(37kDa, 14C10, #2118,Cell Signaling)were alsoused. Inthe TMD231 cells, the predominant form of p73 detected was thealpha isoform (�80 kDa, A300-126A, Bethyl Laboratories).Although the p73 polyclonal antibody A300-126A recognizesmultiple isoforms of p73, it does not recognize DNp73. For thechromatin association assay, antibodies included Mdm2 (3G9,Millipore), H2AX (A300-082A, Bethyl), and b-actin (AC-15,Sigma) as previously described (25).
Transient knockdown of p73a with siRNAON-TARGETplus non-targeting control and SMARTpool ON-
TARGETplus p73 siRNAs were purchased from Dharmacon (GEHealthcare). See Supplementary Materials and Methods for p73siRNA constructs. Lipofectamine RNAiMAX (Life Technologies)was used for transfection. For Western blot and proliferationassays, TMD231 cells were transfected with either nontargetingcontrol siRNAorp73 siRNA.Onday1 after transfection, cellswereseeded into 6-well plates (Western) or 96-well plates (prolifera-tion assays), treated with Nutlin-3a or carboplatin, alone or incombination for 72 hours, and IC50 values determined. ForWestern blot experiments, cells were treated with Nutlin-3a orcarboplatin, alone or in combination for 24 hours.
Chromatin association assayTMD231 cells were treated with Nutlin-3a, carboplatin, 1:1
combination, or vehicle control for 6 hours. Cells were harvested,and soluble and chromatin-bound proteins separated with CSKbuffer as described (26). The chromatin fraction was extractedwith RIPA buffer containing protease inhibitors (24, 25). Whole-cell lysates were prepared as previously reported (27).
AnimalsAll studieswere carried out in accordancewith, and approval of,
the Institutional Animal Care and Use Committee of IndianaUniversity School of Medicine and the Guide for the Care andUse of Laboratory Animals. Female NOD/scid and NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice 6 to 8 weeks of age wereobtained from the In Vivo Therapeutics Core of the IndianaUniversity SimonCancer Center. Animals weremaintained underpathogen-free conditions and maintained on Irradiated Global2018 (Uniprim 4,100 ppm; TD.06596, Harlan Laboratories USA)food pellets with ad libitum access to autoclaved, acidified tapwater under a 12-hour light–dark cycle at 22�C to 24�C.
Animal strain comparisonsNOD/scid and NSG mice were implanted with 1 � 106
TMD231 cells into the mammary fat pad. Tumor volumes werecalculated according to the formula (a2 � b)/2, where a was theshorter and b was the longer of the two dimensions followingcaliper measurements. Mice were sacrificed throughout the studyto better understand longitudinal metastasis formation in thelungs via hematoxylin and eosin (H&E) staining.
Orthotopic xenograft studiesIn the first study, NSG mice were implanted with 1 � 106
TMD231 or TMD231-CR cells in themammary fat pad.Mice wereoptically imaged on day 7 and block-randomized to treatmentgroups basedon tumorfluorescence intensity: vehicle (Veh; PBSþmethylcellulose/Tween80), 25 mg/kg carboplatin intraperitone-ally (Carb; in a.m.), 200 mg/kg Nutlin-3a per os (Nut; in p.m.), or25mg/kg carboplatinþ 200mg/kgNutlin-3a (Combo), andweredosed three timesweekly for 2weeks. Bodyweights were collectedweekly and primary tumors were measured via caliper biweekly.The endpoint for the study was when the primary tumors reached�1000 mm3. At the time of sacrifice, the primary tumors andlungs were collected for Ki67 and H&E staining.
In the second study, animals were implanted and block-ran-domized into treatment groups: vehicle (Veh; PBS þ methylcel-lulose/Tween80), 20 mg/kg carboplatin (Carb), 200 mg/kgNutlin-3a (Nut), and 20 mg/kg carboplatin þ 200 mg/kgNutlin-3a combination (Combo; n ¼ 12 per group). Mice weredosed twiceweekly for 4weeks. Bodyweights andprimary tumorswere measured as described above. To determine drug effects onbone marrow, four mice from each treatment group were sacri-ficed at 5 days following the completion of treatment and totalbone marrow cell counts determined. At the time of sacrifice, theprimary tumors and lungs were collected and fixed for H&Eanalysis. Femurs were collected to determine bone marrow cellu-larity. The remaining mice from each treatment group (n ¼ 8 pergroup) were used to examine the probability of survival fortumors to reach �800 mm3. Total bone marrow cell counts werealso determined at the endpoint (n ¼ 3 per group).
Tonsing-Carter et al.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2852
In the third study, mice were implanted as described above anddosed with vehicle (Veh; PBS þ methylcellulose/Tween80), 20mg/kg carboplatin (Carb; in a.m.), 200 mg/kg Nutlin-3a (Nut; inp.m.), or 20 mg/kg carboplatin þ 200 mg/kg Nutlin-3a in com-bination (Combo) twice weekly for 4 weeks. Endpoint was whenprimary tumors reached �1000 mm3. Following necropsy, pri-mary tumors, lungs, liver, spleen, and femur bones were collectedand evaluated with H&E staining. Bone marrow cellularity, com-plete blood counts (CBC), and progenitor assays were completed.
Tissue processing and stainingTissues were fixed in 10% neutral-buffered formalin at 4�C for
24 hours followed by tissue processing and then embedded inparaffin. Five-micrometer sections were cut and stained for rou-tine H&E or Ki67 (DAKO).
Whole slide digital imagingThe Aperio ScanScope CS system whole slide digital imaging
system was used for imaging (Leica Biosystems). All slides wereimaged at 20�. Images were captured from the whole slides andstored in the Spectrum software system. See Supplementarysection for details of analysis.
In vivo imagingAnesthesia was induced with 4% to 5% isoflurane (balance
medical oxygen) andmaintained with 1% to 2% isoflurane. Micewere imaged using an Optix MX3 (ART Technologies), whereexcitation and emission of E2-Crimsonwas carried out at 635 and650 nm, respectively.
Bone marrow cellularityFemurs were collected, crushed, and cells were filtered through
70-mm filters. Red blood cells were depleted with RBC Lysis Buffer(Qiagen). Total cell counts were determined using a BeckmanCoulter Counter.
Colony-forming unit progenitor assayBone marrow cells were plated in triplicate at 2� 104 per plate
inMethoCult GFM3434 (StemCell Technologies) andprogenitorcolonies counted as described in Cai and colleagues (28).
Total CBCsTumor-bearing mice were treated with vehicle control (Veh),
carboplatin (Carb), Nutlin-3a (Nut), or carboplatin þ Nutlin-3acombination (Combo). After a 2- to 4-week recovery period, analiquot of peripheral blood was analyzed via Hemavet for redblood cells, platelets, and white blood cells.
Statistical analysisData were analyzed by one-, two-way, or repeated measures
ANOVAor the Student t test, as appropriate, using SigmaPlot 11.2(Systat Software, Inc.). To normalize the variability, data werelogarithmically transformed for the purposes of statistical analysisfor the MILLIPLEX data, Ki67 and H&E stain quantification, andCBCs data. Differences among individual pairs of means weredetermined by the Holm–Sidak post-hoc test. IC50 were valuesdetermined using CalcuSyn (BioSoft). Kaplan–Meier survivalplots were generated using SigmaPlot, and changes in survivalanalyzed by the log-rank test. Datawere considered significant atP< 0.05. Data are presented asmean� 1SDunless noted otherwise.
ResultsIn vitro evaluation of carboplatin and Nutlin-3a on growth ofmutant p53 TNBC cells
The effects of single or combination Nutlin-3a/carboplatintreatment was evaluated in a panel of TNBC cell lines usingmethylene blue proliferation assays. Both carboplatin andNutlin-3a produced dose-related decreases in cell proliferationin MDA-MB-231, TMD231, and MDA-MB-468 cells (Fig. 1A–C).Furthermore, the IC50 values for both drugs were greatly reducedin combination treatments compared with single-agent treat-ments (Supplementary Table S1). In the isobologram analyses,the isoboles of different dose ratios of carboplatin and Nutlin-3afell well below the line of additivity indicating a synergistic effectof combination treatments (Fig. 1E–G). In addition, combinationindices were <1 for all Nutlin-3a:carboplatin ratios tested, againindicating synergism between Nutlin-3a and carboplatin (Sup-plementary Fig. S1). As expected,MCF-7 (wt-p53) cells weremoresensitive to single-agent Nutlin-3a (Fig. 1D), and isobologramanalysis indicated that Nutlin-3a was synergistic with carboplatin(Fig. 1H). In addition, combination indices <1 were observed inMCF-7 cells (Supplementary Fig. S1).
Effects of combination treatment on cell death in TNBC cellsand normal cells in vitro
On the basis of proliferation assay results, the effect of com-bination treatment on clonogenicity was evaluated in TMD231cells. Colony formation was inhibited in a concentration-depen-dent manner following Nutlin-3a, carboplatin, and 1:1 combi-nation treatment (Fig. 2A–C). Growth was inhibited by about50% at concentrations of 2.5 mmol/L carboplatin or 30 mmol/LNutlin-3a alone (Fig. 2A and B). In contrast, following combina-tion treatment, clonogenic growth was inhibited �50% by 1.5mmol/L of each compound, consistent with synergism (Fig. 2C).Tounderstandhowcell proliferationwas inhibited, kinetic experi-ments were conducted using TMD231 cells to quantify the num-ber of viable cells throughout a5-day treatment period.We electedto use a 1:1 dose ratio of Nutlin-3a:carboplatin for this study, asthis ratio exhibited the largest synergism in the isobologramanalysis (Fig. 1F). On the basis of analysis of carboplatin plasmalevels in clinical trials formetastatic disease, a concentration of 15mmol/L carboplatin was selected for the in vitro studies (29).Consistent with our cell proliferation data, 15 mmol/L Nutlin-3a did not markedly affect cell viability, whereas 15 mmol/L ofcarboplatin significantly reduced cell number, over time (Fig. 2D).For the combination-treated cells, there was a significant furtherreduction in the number of cells relative to carboplatin alone (Fig.2D, inset).
The sensitivity of MCF10A mammary epithelial and normalhuman fibroblast cells to single agent and combination treatmentwas compared with that of TMD231 cells. Cells were exposed to15 mmol/L of carboplatin, Nutlin-3a, or a 1:1 ratio of carboplatinand Nutlin-3a for 5 days. As expected, wild-type p53 normalhuman cells were significantly more sensitive to Nutlin-3a thanthe mtp53 TMD231 cells (Fig. 2E). Interestingly, the TMD231cells were significantly more sensitive to carboplatin than thenormal human cells (Fig. 2E), consistent with previous reportsthat mtp53 cancer cells can have inefficient nucleotide excisionrepair (30), rendering them more susceptible to platinum-medi-ated DNA damage. As observed in Fig. 2D, TMD231 cells weresignificantly more sensitive to the combination treatment than
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2853
Figure 1.Nutlin-3a sensitizes TNBC cells to carboplatin-mediated DNA damage in vitro. A–D, dose-related inhibition of growth by carboplatin, Nutlin-3a, and the 1:1combination in MDA-MB-231, TMD231, and MDA-MB-468 TNBC cells and MCF-7 breast cancer cells following 5-day proliferation assays. E–H, IC50 values ofcarboplatin and Nutlin-3a were determined at the indicated dose ratios and analyzed by isobolograms. Individual isobole points that lie below the diagonal line ofadditivity indicate synergism, points on the line indicate additivity, and points above the line indicate antagonism. Each point represents the mean of threeexperiments. Vertical and horizontal lines indicate �1 SD and are absent when less than the size of the point.
Tonsing-Carter et al.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2854
Figure 2.Potentiation of carboplatin-mediated DNA damage by Nutlin-3a increases cell death and differentially affects normal cells. A–C, dose-dependent reductionin colony-forming units (CFU) by Nutlin-3a (A), carboplatin (B), and in the 1:1 combination (C). ��� , P < 0.001 versus V (vehicle), Holm–Sidakpost-hoc test, n ¼ 3. D, time course of the number of viable TMD231 cells following treatment with vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/Lcarboplatin (Carb), or 1:1 combination (Combo). ��� , P < 0.001 carboplatin versus vehicle, Holm–Sidak post-hoc test, n ¼ 3. Inset, number of viablecells following carboplatin and combination treatment. � , P < 0.05 versus carboplatin alone, Student t test, n ¼ 3. Vertical lines indicate �1 SDand are absent when less than the size of the point. E, effects of 15 mmol/L of Nutlin-3a, carboplatin, or 1:1 combination in TMD231, MCF10A mammaryepithelial, and human fibroblast cells on the number of viable cells, expressed as percentage of vehicle control, following 5 days of treatment.The y-axis was plotted on a log scale to better illustrate the comparisons. †††, P < 0.001 versus TMD231 treated with 15 mmol/L Nutlin-3a; zzz, P < 0.001versus MCF10A and fibroblasts treated with 15 mmol/L carboplatin; $$$, P < 0.001 versus MCF10A and fibroblasts treated with combination;��� , P < 0.001 versus TMD231 treated with 15 mmol/L Nutlin-3a and carboplatin alone, Holm–Sidak post-hoc test, n ¼ 3. F, activated caspase-3/7 foldinduction in TMD231 cells treated with Vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or the combination (1:1 combo) for3 days. ��� , P < 0.001 versus Veh, Nut, and Carb, Holm–Sidak post-hoc test, n ¼ 3. G, cleaved PARP fold induction in TMD231 cells treated with Vehicle(Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or the combination (1:1 combo) for 3 days. �� , P < 0.01 versus Veh, Student t test,n ¼ 3, �SEM. H, total apoptosis/necrosis in TMD231 cells treated with Nutlin-3a (N), carboplatin (C), or the combination (NC) at the indicated dosesand dose ratios. �� , P < 0.01 versus Veh, N, and C, Holm–Sidak post-hoc test, n ¼ 3.
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2855
single agent alone. Moreover, TMD231 cells were significantlymore sensitive than the MCF10A epithelial and human fibroblastcells to 15 mmol/L combination treatment (Fig. 2E). Although invitro toxicity data do not necessarily predict toxicity in vivo, thesedata indicate that careful selection of compound doses andscheduling will be crucial to avoid unacceptable levels of normaltissue toxicity in vivo.
We next examined the effects of the combination treatment onactivated caspase-3/7 following dual Nutlin-3a and carboplatintreatment in TMD231 cells. The 1:1 combination of Nutlin-3a:carboplatin led to a significant increase in activated caspase-3/7compared with single drug treatments after 3 days of treatment(Fig. 2F). In addition, cleaved PARP was significantly increased incombination-treated TMD231 cells following 3-day treatment(Fig. 2G). Further, based on our observation that cell number isreduced by 4 days posttreatment, apoptosis was also examinedin TMD231 cells on day 4 (Fig. 2D, inset). The IC50 values fromthree different dose ratios [1:1 (0.8:0.8 mmol/L), 1:0.3 (3.75:1.25mmol/L), and 1:3 (0.7:2.1 mmol/L) Nutlin-3a:carboplatin], ascalculated from the TMD231 proliferation assays (Fig. 1 andSupplementary Table S1), were used in the apoptosis assays.Total apoptosis/necrosis was significantly increased in all com-bination treatments (50%–75%) compared with either drugalone (20%–30%; Fig. 2H), indicating that cell death pathwaysare activated in TNBC cells following combination treatment invitro. Normal fibroblast cells were relatively resistant to Nutlin-3aand carboplatin both as single agents and in combination withtotal apoptosis/necrosis <5% (Supplementary Fig. S2) indicatingspecificity of the combination treatment inducing apoptosis inTNBC cells compared with normal cells.
Target modulation and mechanism of action followingcombination treatment
Wehypothesized, on the basis of previous studies (12), that theeffects of Nutlin-3a in amutant p53 background were most likelyattributable to p73a-mediated signaling. To gain an understand-ing of molecular mechanisms that contribute to cell death inTNBC cells, the effect of treatment on levels of p73a and down-stream targets Mdm2, p21, and PUMA was evaluated in TMD231cells (Fig. 3A). TMD231 cells were transiently transfected withnontargeting control or p73 siRNAs followed by treatment withsingle agent or combination for 24 hours. In the siControl RNATMD231 cells, Nutlin-3a alone and combination Nutlin-3a/car-boplatin produced increases in Mdm2 when compared withvehicle-treated cells (Fig. 3A, left). p73a was constitutivelyexpressed in TMD231 cells, and no large changes in p73a levelswere evident following treatment. PUMA levels were higher incombination-treated TMD231 cells compared with vehicle andsingle-agent–treated cells, and increases in p21 were observedfollowing carboplatin or combination treatment. Efficient knock-downof p73awasobtained (Fig. 3A, right), and the effect of p73aknockdown following treatment was evaluated. In contrast tosiControl cells treated with Nutlin-3a or combination, Mdm2protein levels were lower in siRNAp73-transfected cells, indicat-ing that p73a plays a role in regulating Mdm2 levels (Fig. 3A). Incomparison to siControl cells, p21 levels were lower in carbopla-tin- and combination-treated siRNAp73 cells, and PUMA levelswere lower in combination-treated siRNAp73 cells, indicatingthat p21 and PUMA induction is dependent to some extent onp73a (Fig. 3A). The levels of the Mdm2-binding partner Mdmxwere also evaluated; no significant changes in protein levels in
both siControl and siRNAp73-transfected cells were observedfollowing treatment (Fig. 3A).
Sensitivity of TMD231 cells to carboplatin and Nutlin-3a wasdependent on p73a protein levels. When cellular sensitivity wasexamined by methylene blue proliferation assays, there was asignificant increase in IC50 values for carboplatin alone andcarboplatin plus Nutlin-3a at a 1:1 dose ratio in siRNAp73-transfected TMD231 cells compared with siControl-transfectedcells (Fig. 3B).
Cell-cycle analysis indicated that the TMD231 cell line hadbothdiploid and aneuploid subpopulations and that carboplatin-treated TMD231 cells accumulated in S and G2–M in both sub-populations (Supplementary Fig. S3), which is consistent withprevious reports that platinum agents lead to S and G2–M accu-mulation (31). Nutlin-3a alone did not induce cell-cycle arrest inmutant p53 TMD231 cells (Supplementary Fig. S3). In addition,TMD231 cells treated with the 1:1 Nutlin-3a:carboplatin combi-nation arrested in S and G2–M and also exhibited an increasedsub-G1 apoptotic population (Supplementary Fig. S3).
Mdm2 can modulate the ability of cells to sense DNA damagethrough its direct binding to Nbs1 of the MRN DNA repaircomplex (24). Therefore, we next determined whether increasedlevels of total Mdm2 correlated with increased levels of Mdm2associatedwith the chromatin fraction. TMD231 cellswere treatedwith 15 mmol/L Nutlin-3a, 15 mmol/L carboplatin, or 1:1 com-bination for 6 hours, and the levels of Mdm2 in whole-cell lysatesand chromatin fraction lysates were determined by Western blot.FollowingNutlin-3a or combination treatment,whole-cell lysatesshowed increased levels of Mdm2 (Fig. 3C). In chromatin frac-tions isolated from the different treatment groups, association ofMdm2 with the chromatin was increased in combination-treatedcells compared with the untreated or single-agent–treated cells(Fig. 3D). These findings are consistent with the interpretationthat the increased localization of Mdm2 to chromatin leads toinhibition of theMRN complex (24) andmay decrease the abilityof cells to sense carboplatin-mediated DNA damage leading toincreased cell death. Time course studies indicated that gH2AXfoci were significantly elevated in combination-treated versussingle-agent and vehicle-treated cells by 48 hours after treatmentusing confocalmicroscopy as a highly sensitivemeasure of gH2AXfoci formation (Fig. 3E). Analysis of gH2AX via MILLIPLEXindicated significant increases in gH2AX by 72 hours after treat-ment (Fig. 3F). Significant increases in Mdm2 were evident inTMD231 cells treated with Nutlin-3a and combination at all timepoints analyzed (Fig. 3G).
It is possible that blockade of theN-terminal p53-binding site ofMdm2 by Nutlin-3a could result in stabilization of mutant p53(32). However, in TMD231 cells, there was no increase in mtp53steady-state protein levels upon treatment with combinationNutlin-3a/carboplatin compared with vehicle (Fig. 3A). Cyclohex-imide experiments indicated that mtp53 was highly stable, and nodifferences in mtp53 stability were observed in vehicle- versuscombination-treated TMD231 cells (Supplementary Fig. S4).Therefore,Nutlin-3adidnothave aneffect onmtp53protein levels.
Development of a fluorescence imaging model for earlydetection of orthotopic TMD231 primary mammary fat padtumors
To optimize an in vivo animalmodel to study human TNBC, weinitially compared the growth kinetics and metastasis to the lungin Nod/Scid versus NSG mice. In contrast to Nod/Scid mice,
Tonsing-Carter et al.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2856
TMD231 tumors implanted in NSGmice grewmore consistently,at a faster rate, and lungmetastaseswere detected earlier followingimplant (Fig. 4A and B). The NSG strain was thus selected forsubsequent studies. In NSG mice, TMD231 fat pad tumors werepalpable by day 7 after implantation, but the tumors were toosmall to be accurately measured by calipers until approximatelyday 14 or later, a time point at which tumor burden rapidly
increases at both the primary and lung metastatic sites. To opti-mize the therapeutic window in the model, mice were random-ized to treatment groups on the basis of primary tumor size at day7 using fluorescence imaging.
TMD231-CR transduced cells expressed high levels of E2-Crim-son (Supplementary Fig. S5A and S5B), and fluorescence intensitywas linearly related to the number of cells implanted in the
Figure 3.Effects of single agent and combination treatment onMdm2anddependency of treatment-mediated cell death onp73a. A, TMD231 cellswere transiently transfectedwith siControl or sip73 siRNA and treated with Vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or 1:1 combination (Combo) for24 hours and levels of Mdm2, MdmX, p73a, p53, p21, PUMA, and GAPDH determined by Western blot. Densitometry for each protein was quantified byImageJ and shown below each protein relative to siControl Vehicle-treated cells. B, TMD231 cells transfected with sip73 or siControl (siCon) were treatedwith 15mmol/LNutlin-3a (Nut), 15mmol/L carboplatin (Carb), or Nutlin-3a/carboplatin (Combo), for 3 days and IC50 values for inhibiting cell growthwere determinedafter methylene blue staining. � , P < 0.05, vs siControl, Student' t test, n ¼ 5 independent experiments. C and D, cells were not treated (Untx) or treatedwith Vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or 1:1 combination (Combo) for 6 hours. Mdm2 levels of whole-cell lysates(WCL; C) and the chromatin fraction (Chromatin; D) were determined by Western blotting. b-Actin (WCL) and H2AX (chromatin) served as loading controls.Densitometry is shown above the blots. Data are representative of 2 independent experiments. E, gH2AX foci formation treated with Vehicle (Veh), 7.5 mmol/LNutlin-3a (Nut), 7.5 mmol/L carboplatin (Carb), or 1:1 combination (Combo). �� , P < 0.01 versus Vehicle; ��� , P < 0.001 versus Vehicle; ††, P < 0.01 versus Nut;†††, P < 0.001 versus Nut, #, P < 0.05 versus Carb; ###, P < 0.001 versus Carb at same time point, Holm–Sidak post-hoc test, n ¼ 5 fields per group. F, gH2AXlevels normalized to b-tubulin following Vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or 1:1 combination (Combo) in TMD231cells. �� , P < 0.01 versus Vehicle; ��� , P < 0.001 versus Vehicle; ###, P < 0.001 versus Nut; @, P < 0.01 versus Carb at the same time point, Holm–Sidak post-hoc test,n ¼ 3 independent repeats. G, Mdm2 levels normalized to b-tubulin following Vehicle (Veh), 15 mmol/L Nutlin-3a (Nut), 15 mmol/L carboplatin (Carb), or 1:1combination (Combo) in TMD231 cells. �� , P < 0.01 versus Vehicle; ��� , P < 0.001 versus Vehicle; @@@, P < 0.001 versus Carb; and ††, P < 0.01 versus Nut at the sametime point, Holm–Sidak post-hoc test, n ¼ 3 independent repeats.
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2857
mammary fat pad (Supplementary Fig. S5C and S5D). Further-more, sensitivity to Nutlin-3a and carboplatin in an in vitroproliferation assay, alone or in combination was similar to thatobserved in nontransduced TDM231 parental cells (Fig. 1B,Supplementary Table S1 and Supplementary Fig. S5E). Fluores-cence intensity from tumors was detected in themammary fat padas early as 7 days after implantation in nearly 100% of the mice,which is a time point before detection of metastatic foci in thelungs (Fig. 4A and B) and was used to block-randomize mice intotreatment groups at day 7 after implantation (Supplementary Fig.S5F and S5G). We also evaluated whether fluorescence imagingcould be used to longitudinally monitor the appearance ofmetastatic foci in the lungs. However, the photon emission fromtumor foci in the lungs could not be detected, most likely due tosecondary inner-filtering and light scattering by tissue surround-ing the foci. Therefore, H&E staining of lung tissue was used toquantify metastatic foci at the end of each study. Using thisoptimized model, we performed dose finding studies for carbo-
platin to identify a dose and schedule to combine with Nutlin-3a.Carboplatin produced dose-related decreases in TMD231-CRgrowth and increased probability of survival using an endpointof primary tumor volume �1,000 mm3 (Fig. 4C and D). Themedian survival was 41� 3.6 days for vehicle, 44� 4.3 days for 1mg/kg, 48� 1.9 days for 3mg/kg, and 58� 0 days for 30mg/kg ofcarboplatin (Fig. 4D). On the basis of these data, we elected toadminister carboplatin at 20–25 mg/kg in the combinationefficacy studies.
In vivo effects of combination Nutlin-3a/carboplatin treatmenton growth of primary tumor and metastatic foci in the lung
Combination treatment significantly inhibited primary tumorgrowth compared with vehicle and single-agent treatments(Fig. 5A). Body weights were transiently reduced by carboplatinand the combination of Nutlin-3a/carboplatin; however, bythe end of the study, all body weights were within a normal range(Fig. 5B). In addition, there were no visual signs of dehydration,
Figure 4.Metastatic breast-to-lung orthotopic model optimization and carboplatin dose-finding studies. Female Nod/Scid or NSG mice were implanted with 1� 106 TMD231cells in the mammary fat pad. A, longitudinal tumor growth in both mouse strains measured with calipers (�SEM). Inset, number of lung metastatic fociin mice sacrificed at the indicated time points as determined by H&E staining (�SEM). B, representative microscopic images of H&E-stained lung sections:normal mouse lung (i), days 18 (ii), 26 (iii), and 32 (iv) postimplantation of TMD231 cells. C and D, NSG mice were implanted with 1 � 106 TMD231-CR cells anddosed with Vehicle (Veh), 1, 3, or 30 mg/kg carboplatin intraperitoneally 3� per week for 2 weeks. C, dose-related decreases in tumor volume produced bycarboplatin. ��� , P < 0.001 versus Vehicle, Holm–Sidak post-hoc test, n ¼ 8–9 per group, �SEM. D, dose-related increases in survival produced by carboplatin.��� , P < 0.001 versus Vehicle; $$, P < 0.01 versus 3 mg/kg, Holm–Sidak post-hoc test, n ¼ 8–9 per group.
Tonsing-Carter et al.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2858
diarrhea, or hemorrhaging of the gastrointestinal tract. Further-more, primary tumors excised at the termination of the studyweighed significantly less from mice treated with combinationtherapy compared with vehicle and single-agent treatments (Fig.5C). In addition, there was a significant reduction in Ki67 stainingin primary tumors of combination-treated mice as measured bywhole slide digital imaging (Fig. 5D).
At termination of the study, the metastatic burden in the lungswas compared among the treatment groups (Fig. 5E and F). Incontrast to mice treated with vehicle or single agent, the combi-nation treatment significantly decreased the metastatic burden asmeasured by H&E staining and whole slide digital imaginganalysis (Fig. 5F). From the present data, it is not clear whetherthe combination treatment kills metastatic cells at the primarytumor site, inhibits the metastatic process, and/or leads to tumorcell death after the TMD231 cells metastasize to the lungs. In vitroinvasion assays, however, indicated that the combination treat-ment did not significantly affect cell invasion when comparedwith vehicle- or single-agent treatments (Supplementary Fig. S6).
To evaluate potential pharmacodynamic biomarkers oftherapeutic response in vivo, NSG mice were implanted with
TMD231-CR cells and treated for 3 consecutive days with vehicle,carboplatin, Nutlin-3a, or Nutlin-3a/carboplatin combination.The tumors were excised 2 hours after the last dose and tumorlysates examined by Western blotting (Supplementary Fig. S7Aand S7B). In PD study 1, tumors were lysed in 1% SDS toefficiently isolate nuclear-localized proteins (Supplementary Fig.S7A). Basal levels of p73a did not significantly change in any ofthe groups. While Mdm2 levels were overall higher in tumorsfrom the Nutlin-3a–treated group versus vehicle, it was notsignificant (P ¼ 0.08). In comparison with the carboplatin-alone–treated group, Mdm2 levels were significantly higher inthe Nutlin3a-alone and combination-treated groups (P < 0.05).Furthermore, p21 levels significantly increased in tumors frommice treated with the combination compared with vehicle (P <0.05). In PD study 2, the tumor sampleswere lysed in RIPA buffer,as we have found that this extraction method is optimal fordetection of the phosphoprotein gH2AX in tumor tissue (Sup-plementary Fig. S7B). Analysis of Western blot analyses viadensitometry indicated that while gH2AX increased in the sin-gle-agent– and combination-treated tumors compared with vehi-cle-treatedmice, statistical significancewas not reached (P>0.05);
Figure 5.Carboplatin in combinationwith Nutlin-3a significantly decreases primarytumor growth and lung metastases.A, longitudinal tumor growth of miceadministered Nutlin-3a (Nut),carboplatin (Carb), or the combination(Combo; 3� weekly for 2 weeks).��� , P < 0.001 versus all other groups,Holm–Sidak post-hoc test, n ¼ 7–9 pergroup, �SEM. B, body weightsfollowing drug treatment (�SEM).C, primary tumor weight at studytermination. ��� , P < 0.001 versus allother groups, Holm–Sidak post-hoctest, n ¼ 7–9 per group, �SEM. D, Ki67staining positivity of primary tumorsusing whole slide digital imaging.� , P < 0.05 versus all groups, Holm–
Sidak post-hoc test. E, representativeH&E-stained lung sections(magnification, 20�). F, metastaticburden as measured by the proportionof the lungs positive for H&E staining atstudy termination using whole slidedigital imaging. �� , P < 0.01 versus allgroups, Holm–Sidak post-hoc test,n ¼ 5 per group.
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2859
mutant p53 levels remained unchanged across the treatmentgroups (Supplementary Fig. S7B). In addition, the levels ofactivated caspase-3 and cleaved PARP were evaluated in tumorlysates derived fromPD study 2. Although activated caspase-3wasdetectable in50%of the vehicle-treated tumors, activated caspase-3 was detected in 100% of the tumors treated with Nutlin-3a,carboplatin, or combination (Supplementary Fig. S7C). Similarresults were obtained in tumor samples probed for cleaved PARPwith 100% of tumors having measureable levels of cleaved PARPfollowing combination treatment, whereas only 50% to 66% oftumors contained measureable levels of cleaved PARP followingvehicle, Nutlin-3a, or carboplatin treatment (SupplementaryFig. S7D).
Effects of combination Nutlin-3a/carboplatin treatment onsurvival and normal tissue toxicity in vivo
To confirm therapeutic effects on tumor growth and evaluateeffects of treatment on hematopoiesis, a sensitive indicator ofnormal tissue toxicity, a second orthotopic study was conducted.Similar to the previous study (Fig. 5), the combination treatmentled to significantly reduced tumor volumes when compared withvehicle and single-agent treatments (Fig. 6A) with only a transientdecrease in body weights during therapy (Fig. 6B). Significantreductions in metastatic burden following combination treat-ment were also observed (Fig. 6C). Survival was evaluated usingthe day post implant at which the primary tumor volume reached�800 mm3 as the survival endpoint. In contrast to single-agenttherapy, the Nutlin-3a/carboplatin combination significantlyincreased the probability of survival (Fig. 6D). Median survivalwas 40� 1.5 days for vehicle, 40� 8.1 for Nutlin-3a, 47� 2.6 forcarboplatin, and 54.3 � 0.9 days for the combination (Fig. 6D).Bone marrow cellularity was determined at 5 days and �14 daysafter treatment. At 5 days after treatment, total bone marrow cellcounts (Fig. 6E, left) were significantly reduced in combination-treated mice relative to the other groups. However, by 14 to 28days after treatment (Recovery), bone marrow cellularity hadrecovered to control levels (Fig. 6E, right).
Normal tissue toxicity was further examined in a third study.Total bone marrow cell counts were analyzed following a 2-weekrecovery period after treatment in all groups, and the frequency ofhematopoietic progenitor cells in thebonemarrowwas evaluated.While CBCs and hematopoietic progenitors were significantlydecreased in samples from mice treated with carboplatin aloneor the combination compared with vehicle or Nutlin-3a alone,there were no differences between carboplatin alone and combi-nation groups (Fig. 6F–I). Furthermore, no indications of normaltissue toxicity were noted at the tissue level upon analysis of H&Estaining of liver, spleen, and femur bone marrow smears, evalu-ated by a board-certified pathologist (G.E. Sandusky; personalcommunication).
DiscussionThe objective of the present study was to evaluate the outcome
of modulating Mdm2 function in the context of current chemo-therapeutic approaches for TNBC as well as to assess Mdm2 as atherapeutic target. TNBCs are highly refractive to therapy and thedevelopment of multitargeted treatments that exhibit an accept-able toxicity profile are needed. Approximately 60% of basalTNBCs bear mutations in p53 (15) and more than 90% ofmutations in p53 occur in the DNA-binding domain. The
MDA-MB-231 (mtp53 R280K), the MDA-MB-231 derivativeTMD231 (mtp53 R280K), and the MDA-MB-468 (mtp53R273H) cells used in our study all have missense mutationswithin the DNA-binding domain of p53 (33). The mtp53R280K mutation in the MDA-MB-231 and the TMD231 cells hasbeen reported to have gain-of-function properties (34); however,its exact role is still not well-understood. Gain-of-functionmutantp53may antagonize other tumor-suppressing capabilities of cells.In some cell types, mtp53 can sequester p73, which leads to ablockade of p73-mediated downstream signaling (35). Xu andcolleagues demonstrated that mutant p53 (mtp53; R282W andR110P) exhibit a misfolded/aggregated state which leads toincreased aggregation of mtp53 with p73 causing inhibition ofp73 function (35). It is not known whether the gain-of-functionability ofmutant p53 in theMDA-MB-231 and TMD231 (R280K)cells plays a similar role in coaggregation with p73 and will beaddressed in future studies. However, a sufficient pool of "free"p73must exist, for exposure of TNBC cells in vitro to combinationNutlin-3a/carboplatin treatment led to a synergistic inhibition ofcell proliferation, which was dependent, at least in part, on p73a-mediated signaling (Fig. 3B).
Previous studies have demonstrated that mutant p53 cancercells canharbor nucleotide excision repair defects, rendering themhighly sensitive to platinum agents (30). Furthermore, ongoingclinical trials for metastatic TNBC are including platinum agentsin their treatment regimens (4, 36–38). In an orthotopic human-ized breast-to-lung model, treatment with Nutlin-3a/carboplatinsignificantly inhibited primary tumor growth, andmetastatic fociin the lung were fewer in number and smaller in size relative totreatment with vehicle or single-agent therapy. Pharmacokineticstudies have demonstrated efficient delivery of Nutlin-3a to thelungs, which was comparable to plasma levels (39). In addition,carboplatin is routinely used to treat non–small cell lung cancer,providing further rationale for the use of this combination ther-apy to treat metastatic foci in the lung (40, 41). The p53 familymember p73a has functions similar to p53 (17, 18, 42, 43) but isregulated differently by Mdm2. While p53 and p73a can bothinteract with Mdm2 in the same N-terminal hydrophobic pocket(12), Mdm2 does not ubiquitinate p73a (44–46). Rather, Mdm2sequesters p73a thereby preventing downstream p73a-mediatedsignaling (45). Blockade of the binding of Mdm2 to p73a byNutlin-3a increases the pool of p73a available to transcriptionallyactivate proapoptotic gene expression and promote apoptosis inmutant p53 and p53-null cells (12). TMD231 cells transfectedwith p73 siRNAweremore resistant to carboplatin andNutlin-3a/carboplatin compared with siRNA control–transfected cells. Inaddition, decreases in p73a protein levels via siRNA knockdowncorrelated with decreased Mdm2 protein in TMD231 cells. Thesedata are consistent with previous reports indicating that p73a candirectly bind to the Mdm2 promoter and increase Mdm2 geneexpression (12, 47).
It has been previously demonstrated that Mdm2 binds directlyto Nbs1 of the Mre11/Rad50/Nbs1 (MRN) complex, colocalizeswith Nbs1 to DNA damage foci, and inhibits the DNA damageresponse (24, 25). While our studies do not demonstrate a directinteraction of chromatin-associated Mdm2 with the MRN com-plex, increases in Mdm2-associated chromatin following Nutlin-3a or combination Nutlin-3a/carboplatin are consistent withstudies of Eischen and colleagues. They demonstrated thatNutlin-3a–induced increased levels of Mdm2 directly inhibitDNA damage response signaling and delayed DNA break repair
Tonsing-Carter et al.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2860
Figure 6.Carboplatin in combination with Nutlin-3a significantly decreases primary tumor growth and increases survival with minimal normal tissue toxicity. A, primarytumor volumes over time in Vehicle (Veh), 200 mg/kg Nutlin-3a (Nut), 20 mg/kg carboplatin (Carb), and Nutlin-3a/carboplatin (Combo)-treated mice.Treatment was administered 2 times weekly for 4 weeks. ��� , P < 0.001 versus Vehicle; $$, P < 0.01 versus Nutlin-3a and carboplatin, n ¼ 12 per group atinitiation of study, �SEM. Note: bone marrow was analyzed in some mice at 5 days after treatment (n ¼ 4 per group); the remaining mice (n ¼ 8 pergroup) were monitored until the survival endpoint (�800 mm3) was met, at which time bone marrow was analyzed. B, body weights following drugtreatment (�SEM). C, metastatic burden as measured by H&E followed by whole slide digital imaging. � , P < 0.05 versus Veh, Holm–Sidak post-hoc test,n ¼ 4 per group. D, probability of survival. ��� , P < 0.001 versus all other groups, n ¼ 8 per group. E, total bone marrow cells per femur determined either5 days after the completion of treatment, (left, �� , P < 0.01, Veh vs. Combo, n ¼ 4 per group), or, after a period of recovery and when a mouse met the survivalendpoint for tumor size [right, Veh vs. Combo, n.s. (nonsignificant) P > 0.05]. Recovery phase varied depending on treatment group (Veh and Nut: 14 days,Carb: 18 days, and Combo: 28 days). F–I, normal tissue toxicity analysis of (F) total number of colony-forming units (CFU), (G) WBCs, (H) platelets (PLT),and (I) RBC. � , P < 0.05 versus Vehicle; �� , P < 0.01 versus Vehicle, n.s. P > 0.05 carb versus combo, Holm–Sidak post-hoc test, n ¼ 7–8 per group.
Modulation of Mdm2 in the Context of DNA Damage
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2861
in p53-nullMEFs and inovarian cancer cell lines that lackedp53orthat contained mutant p53 (48). Combination treatment anddosing strategies that reduce the amount of required chemother-apeutics are important clinically to decrease normal tissue toxicityas well as prevent the emergence of secondary malignanciescaused by therapies that damage DNA. In MCF10A epithelialand human fibroblast cells, the combination treatment decreasedcell growth in vitro compared with carboplatin alone but did notdecrease cell growth beyond that of Nutlin-3a alone. In addition,the TMD231 cells were more sensitive to the combination treat-ment than these other cell types. Jiang and colleagues previouslyreported that Nutlin-3 could protect renal cells from cisplatin-based therapy (49). To what extent Nutlin-3a provides chemo-protective effects in the present model will require further study.
The dosing schedule used in the present breast-to-lung ortho-topic model support the use of intermittent dosing schedules forMdm2 inhibitors, particularly in the context of cytotoxic therapy.Bone marrow toxicity is one of the principal side effects of che-motherapeutic drugs and was monitored closely in the present invivo studies.While therewas adecrease inhematopoieticprogenitorcells in the bone marrow of carboplatin- and Nutlin-3a/carbopla-tin–treated mice, the inclusion of Nutlin-3a did not lead to furtherincreases in carboplatin-mediated toxicity. Analysis of normaltissues (liver, spleen, and bone marrow smears) by H&E indicatedthere were no obvious effects of the combination therapy on tissueintegrity. Thus, combination Nutlin-3a/carboplatin treatmentshows promise of efficacy with minimal effects on normal tissuetoxicity compared with carboplatin-alone.
The optimized breast-to-lung orthotopic model described herecan be used to test novel combination platinum-based regimensand further increase our understanding of how to therapeuticallypotentiate DNA damage in TNBC at primary and metastaticsites. The present studies demonstrate the promise of Mdm2 asa therapeutic target in combination with current therapeuticapproaches for TNBC and may lead to new clinical therapies forTNBC and metastases.
Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.
Authors' ContributionsConception anddesign: E. Tonsing-Carter,M.R. Saadatzadeh, J. Ding,H.Wang,A.L. Sinn, A.R. Safa, K.E. PollokDevelopment of methodology: E. Tonsing-Carter, B.J. Bailey, M.R. Saadatza-deh, J. Ding, H. Wang, A.L. Sinn, K. Peterman, T.M. Georgiadis, E.C. Long,C.C. Marchal, A.R. Safa, H. Hanenberg, K.E. PollokAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): E. Tonsing-Carter, B.J. Bailey, M.R. Saadatzadeh,J. Ding, A.L. Sinn, K. Peterman, T.K. Spragins, A.A. Sprouse, T.Z. Gunter,R.E. Minto, P.R. Territo, C.M. Eischen, H.E. Shannon, K.E. PollokAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): E. Tonsing-Carter, A.A. Sprouse, T.Z. Gunter, C.N.Batuello, A.R. Safa, P.R. Territo, G.E. Sandusky, H.E. Shannon, K.E. PollokWriting, review, and/or revision of the manuscript: E. Tonsing-Carter,B.J. Bailey, M.R. Saadatzadeh, J. Ding, A.A. Sprouse, T.Z. Gunter, E.C. Long,C.N. Batuello, A.R. Safa, H. Hanenberg, P.R. Territo, G.E. Sandusky, L.D. Mayo,C.M. Eischen, H.E. Shannon, K.E. PollokAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): E. Tonsing-Carter, B.J. Bailey,M.R. Saadatzadeh,J. Ding, H.Wang, K. Peterman, T.K. Spragins, J.M. Silver, A.A. Sprouse, K.E. PollokStudy supervision: E. Tonsing-Carter, P.R. Territo, C.M. Eischen, K.E. PollokOther (synthesized compound, assisted with dosing formulation):T.M. GeorgiadisOther (development of novel reagents): H. Hanenberg
AcknowledgmentsThe authors thank Courtney Hemenway (In Vivo Therapeutics Core), Mal-
gorzata M. Kamocka (Indiana Center for Biological Microscopy), Maria PiaArrate (Eischen Lab), Jenny Then, and Sydney Ross (Translational ResearchIntegrated Biology Laboratory) for their expert assistance.
Grant SupportResearch in this publication was supported by the NCI of the NIH under
award number R01CA138798 (E. Tonsing-Carter, B.J. Bailey, H. Wang, L.D.Mayo, andK.E. Pollok); R01CA181204 (C.M.Eischen); theDeVault Fellowship-IUSCC Cancer Biology Training Program (E. Tonsing-Carter); pilot fundingthrough the IUSCC Breast Cancer Program, the Vera Bradley Foundation,ITRAC, and the Indiana Clinical and Translational Sciences Institute (K.E.Pollok).
The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ReceivedMarch 20, 2015; revised September 3, 2015; accepted September 30,2015; published OnlineFirst October 22, 2015.
References1. Schnitt SJ. Classification and prognosis of invasive breast cancer: from
morphology tomolecular taxonomy.Mod Pathol 2010;23 Suppl 2:S60–4.2. Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, et al.
Triple-negative breast cancer: clinical features and patterns of recurrence.Clin Cancer Res 2007;13:4429–34.
3. Fuertes MA, Castilla J, Alonso C, Perez JM. Cisplatin biochemical mech-anism of action: from cytotoxicity to induction of cell death throughinterconnections between apoptotic and necrotic pathways. Curr MedChem 2003;10:257–66.
4. Villarreal-Garza C, Khalaf D, Bouganim N, Clemons M, Pena-Curiel O,Baez-Revueltas B, et al. Platinum-based chemotherapy in triple-negativeadvanced breast cancer. Breast Cancer Res Treat 2014;146:567–72.
5. Hurley J, Reis IM, Rodgers SE, Gomez-FernandezC,Wright J, Leone JP, et al.The use of neoadjuvant platinum-based chemotherapy in locally advancedbreast cancer that is triple negative: retrospective analysis of 144 patients.Breast Cancer Res Treat 2013;138:783–94.
6. Rayburn E, Zhang R, He J, Wang H. MDM2 and human malignancies:expression, clinical pathology, prognostic markers, and implications forchemotherapy. Curr Cancer Drug Targets 2005;5:27–41.
7. Haupt Y,Maya R, Kazaz A, OrenM.Mdm2 promotes the rapid degradationof p53. Nature 1997;387:296–9.
8. Araki S, Eitel JA, Batuello CN, Bijangi-Vishehsaraei K, Xie XJ, Daniel-pour D, et al. TGF-beta1-induced expression of human Mdm2 corre-lates with late-stage metastatic breast cancer. J Clin Invest 2010;120:290–302.
9. Chen X, Qiu J, Yang D, Lu J, Yan C, Zha X, et al. MDM2 promotes invasionand metastasis in invasive ductal breast carcinoma by inducing matrixmetalloproteinase-9. PLoS One 2013;8:e78794.
10. Vassilev LT, Vu BT, Graves B, Carvajal D, Podlaski F, Filipovic Z, et al. In vivoactivation of the p53 pathway by small-molecule antagonists of MDM2.Science 2004;303:844–8.
11. Ambrosini G, Sambol EB, Carvajal D, Vassilev LT, Singer S, Schwartz GK.Mouse double minute antagonist Nutlin-3a enhances chemotherapy-induced apoptosis in cancer cells with mutant p53 by activating E2F1.Oncogene 2007;26:3473–81.
12. Lau LM, Nugent JK, Zhao X, IrwinMS. HDM2 antagonist Nutlin-3 disruptsp73-HDM2 binding and enhances p73 function. Oncogene 2008;27:997–1003.
13. LaRusch GA, Jackson MW, Dunbar JD, Warren RS, Donner DB, Mayo LD.Nutlin3 blocks vascular endothelial growth factor induction by preventingthe interaction between hypoxia inducible factor 1alpha and Hdm2.Cancer Res 2007;67:450–4.
Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2862
14. Soussi T, Beroud C. Assessing TP53 status in human tumours to evaluateclinical outcome. Nat Rev Cancer 2001;1:233–40.
15. Shah SP, Roth A, Goya R, Oloumi A, Ha G, Zhao Y, et al. The clonal andmutational evolution spectrum of primary triple-negative breast cancers.Nature 2012;486:395–9.
16. Dotsch V, Bernassola F, CoutandinD, Candi E,MelinoG. p63 and p73, theancestors of p53. Cold Spring Harb Perspect Biol 2010;2:a004887.
17. Jost CA, Marin MC, Kaelin WG Jr. p73 is a simian [correction of human]p53-related protein that can induce apoptosis. Nature 1997;389:191–4.
18. Melino G, Bernassola F, Ranalli M, Yee K, ZongWX, Corazzari M, et al. p73Induces apoptosis via PUMA transactivation and Bax mitochondrial trans-location. J Biol Chem 2004;279:8076–83.
19. Helbig G, Christopherson KW II, Bhat-Nakshatri P, Kumar S, Kishimoto H,Miller KD, et al. NF-kappaB promotes breast cancer cell migration andmetastasis by inducing the expression of the chemokine receptor CXCR4. JBiol Chem 2003;278:21631–8.
20. Wiek C, Schmidt EM, Roellecke K, Freund M, Nakano M, Kelly EJ, et al.Identification of amino acid determinants in CYP4B1 for optimal catalyticprocessing of 4-ipomeanol. Biochem J 2015;465:103–14.
21. Oliver MH, Harrison NK, Bishop JE, Cole PJ, Laurent GJ. A rapid andconvenient assay for counting cells cultured in microwell plates: applica-tion for assessment of growth factors. J Cell Sci 1989;92:513–8.
22. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: thecombined effects of multiple drugs or enzyme inhibitors. Adv EnzymeRegul 1984;22:27–55.
23. Tallarida RJ. Quantitative methods for assessing drug synergism. GenesCancer 2011;2:1003–8.
24. Alt JR, Bouska A, Fernandez MR, Cerny RL, Xiao H, Eischen CM. Mdm2binds to Nbs1 at sites of DNA damage and regulates double strand breakrepair. J Biol Chem 2005;280:18771–81.
25. Bouska A, Lushnikova T, Plaza S, Eischen CM. Mdm2 promotes geneticinstability and transformation independent of p53. Mol Cell Biol2008;28:4862–74.
26. BradenWA, Lenihan JM, Lan Z, Luce KS, ZagorskiW, Bosco E, et al. Distinctaction of the retinoblastoma pathway on the DNA replication machinerydefines specific roles for cyclin-dependent kinase complexes in prereplica-tion complex assembly and S-phase progression. Mol Cell Biol 2006;26:7667–81.
27. Alt JR, Greiner TC, Cleveland JL, Eischen CM. Mdm2 haplo-insufficiencyprofoundly inhibits Myc-induced lymphomagenesis. EMBO J 2003;22:1442–50.
28. Cai S,WangH, Bailey B, Ernstberger A, Juliar BE, Sinn AL, et al. Humanizedbonemarrowmousemodel as a preclinical tool to assess therapy-mediatedhematotoxicity. Clin Cancer Res 2011;17:2195–206.
29. Shea TC, Flaherty M, Elias A, Eder JP, Antman K, Begg C, et al. A phase Iclinical and pharmacokinetic study of carboplatin and autologous bonemarrow support. J Clin Oncol 1989;7:651–61.
30. Sengupta S, Harris CC. p53: traffic cop at the crossroads of DNA repair andrecombination. Nat Rev Mol Cell Biol 2005;6:44–55.
31. Smith ML. Mdm2 sensitizes MCF7 breast cancer cells to cisplatin orcarboplatin. Breast Cancer Res Treat 1999;58:99–105.
32. Terzian T, Suh YA, Iwakuma T, Post SM, Neumann M, Lang GA, et al. Theinherent instability ofmutant p53 is alleviatedbyMdm2or p16INK4a loss.Genes Dev 2008;22:1337–44.
33. JunkDJ, Vrba L,Watts GS,OshiroMM,Martinez JD, Futscher BW.Differentmutant/wild-type p53 combinations cause a spectrum of increased inva-
sive potential in nonmalignant immortalized humanmammary epithelialcells. Neoplasia 2008;10:450–61.
34. Hui L, ZhengY, YanY, Bargonetti J, FosterDA.Mutant p53 inMDA-MB-231breast cancer cells is stabilized by elevated phospholipase D activity andcontributes to survival signals generated by phospholipase D. Oncogene2006;25:7305–10.
36. O'Shaughnessy J, Osborne C, Pippen JE, Yoffe M, Patt D, Rocha C, et al.Iniparib plus chemotherapy in metastatic triple-negative breast cancer.N Engl J Med 2011;364:205–14.
37. O'Shaughnessy J, Schwartzberg L, Danso MA, Miller KD, Rugo HS, Neu-bauerM, et al. Phase III study of iniparib plus gemcitabine and carboplatinversus gemcitabine and carboplatin in patients with metastatic triple-negative breast cancer. J Clin Oncol 2014;32:3840–7.
38. Carey LA, Rugo HS, Marcom PK, Mayer EL, Esteva FJ, Ma CX, et al. TBCRC001: randomized phase II study of cetuximab in combination with carbo-platin in stage IV triple-negative breast cancer. J Clin Oncol 2012;30:2615–23.
39. Zhang F, Tagen M, Throm S, Mallari J, Miller L, Guy RK, et al. Whole-body physiologically based pharmacokinetic model for nutlin-3a inmice after intravenous and oral administration. Drug Metab Dispos2011;39:15–21.
40. Okamoto H, Watanabe K, Kunikane H, Yokoyama A, Kudoh S, AsakawaT, et al. Randomised phase III trial of carboplatin plus etoposide vs splitdoses of cisplatin plus etoposide in elderly or poor-risk patients withextensive disease small-cell lung cancer: JCOG 9702. Br J Cancer2007;97:162–9.
41. Zukin M, Barrios CH, Pereira JR, Ribeiro Rde A, Beato CA, do NascimentoYN, et al. Randomized phase III trial of single-agent pemetrexed versuscarboplatin and pemetrexed in patients with advanced non-small-cell lungcancer and Eastern CooperativeOncology Group performance status of 2. JClin Oncol 2013;31:2849–53.
42. Zawacka-Pankau J, Kostecka A, Sznarkowska A,HedstromE, Kawiak A. p73tumor suppressor protein: a close relative of p53 not only in structure butalso in anti-cancer approach? Cell Cycle 2010;9:720–8.
43. Zhu J, Jiang J, Zhou W, Chen X. The potential tumor suppressor p73differentially regulates cellular p53 target genes. Cancer Res 1998;58:5061–5.
44. Dobbelstein M, Wienzek S, Konig C, Roth J. Inactivation of the p53-homologue p73 by the mdm2-oncoprotein. Oncogene 1999;18:2101–6.
45. Zeng X, Chen L, Jost CA,Maya R, Keller D,Wang X, et al. MDM2 suppressesp73 function without promoting p73 degradation. Mol Cell Biol1999;19:3257–66.
46. Balint E, Bates S, Vousden KH. Mdm2 binds p73 alpha without targetingdegradation. Oncogene 1999;18:3923–9.
47. Wang XQ, Ongkeko WM, Lau AW, Leung KM, Poon RY. A possible role ofp73 on the modulation of p53 level through MDM2. Cancer Res2001;61:1598–603.
48. Carrillo AM, Hicks M, Khabele D, Eischen CM. Pharmacologically increas-ingMdm2 InhibitsDNA repair and cooperates with genotoxic agents to killp53-inactivated ovarian cancer cells. Mol Cancer Res 2015;13:1197–205.
49. JiangM, PablaN,MurphyRF, Yang T, Yin XM,Degenhardt K, et al. Nutlin-3protects kidney cells during cisplatin therapy by suppressing Bax/Bakactivation. J Biol Chem 2007;282:2636–45.
www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2863
2015;14:2850-2863. Published OnlineFirst October 22, 2015.Mol Cancer Ther Eva Tonsing-Carter, Barbara J. Bailey, M. Reza Saadatzadeh, et al. Metastatic ModelModulator Nutlin-3a in a Humanized Orthotopic Breast-to-Lung Potentiation of Carboplatin-Mediated DNA Damage by the Mdm2
Updated version
10.1158/1535-7163.MCT-15-0237doi:
Access the most recent version of this article at:
