Pharmacological Inhibition of the Protein Kinase MRK/ZAK Radiosensitizes Medulloblastoma Daniel Markowitz1, Caitlin Powell1, Nhan L

Published OnlineFirst May 20, 2016; DOI: 10.1158/1535-7163.MCT-15-0849

Small Molecule Therapeutics Molecular Cancer Therapeutics Pharmacological Inhibition of the Protein Kinase MRK/ZAK Radiosensitizes Medulloblastoma Daniel Markowitz1, Caitlin Powell1, Nhan L. Tran2, Michael E. Berens2, Timothy C. Ryken3, Magimairajan Vanan4, Lisa Rosen5, Mingzhu He6, Shan Sun6, Marc Symons1, Yousef Al-Abed6, and Rosamaria Ruggieri1

Abstract

Medulloblastoma is a cerebellar tumor and the most com- loblastomacellsaswellasUI226patient–derived primary cells, mon pediatric brain malignancy. Radiotherapy is part of the whereas it does not affect the response to radiation of normal standard care for this tumor, but its effectiveness is accompa- brain cells. M443 also inhibits radiation-induced activation nied by significant neurocognitive sequelae due to the delete- of both p38 and Chk2, two proteins that act downstream of rious effects of radiation on the developing brain. We have MRK and are involved in DNA damage–induced cell-cycle previously shown that the protein kinase MRK/ZAK protects arrest.Importantly,inananimal model of medulloblastoma tumor cells from radiation-induced cell death by regulating that employs orthotopic implantation of primary patient– cell-cycle arrest after ionizing radiation. Here, we show that derived UI226 cells in nude mice, M443 in combination with siRNA-mediated MRK depletion sensitizes medulloblastoma radiation achieved a synergistic increase in survival. We primary cells to radiation. We have, therefore, designed and hypothesize that combining radiotherapy with M443 will allow tested a specific small molecule inhibitor of MRK, M443, which us to lower the radiation dose while maintaining therapeutic binds to MRK in an irreversible fashion and inhibits its activity. efficacy, thereby minimizing radiation-induced side effects. We found that M443 strongly radiosensitizes UW228 medul- Mol Cancer Ther; 15(8); 1799–808. 2016 AACR.

Introduction tion, together with high-dose chemotherapy, has improved 5- year survival rates in patients (2). However, it also causes Radiotherapy is commonly used as part of cancer treat- serious deleterious effects that include neurocognitive deficits, ment, and it is important in the management of 40% of developmental problems, and secondary malignancies in the patients who are cured of cancer (1). The therapeutic benefits majority of survivors (3–6). Identifying new approaches that of radiotherapy are, however, accompanied by late toxicity would allow reduction of the total radiation dose in these that severely affects quality of life, especially in pediatric treatments without compromising therapeutic efficacy is crit- patients. Specifically, in the treatment of medulloblastoma, ical for improving tumor management and quality of life. the most common malignant pediatric brain tumor, radia- Radiosensitizers may offer this opportunity by increasing the efficacy of radiotherapy while reducing effects on the normal developing brain. 1Center of Oncology and Cell Biology, Feinstein Institute, Manhasset, The protein kinase MRK (mixed lineage kinase–related kinase), 2 New York. Translational Genomics Research Institute, Phoenix, Ari- also known as ZAK (sterile alpha motif and leucine zipper con- zona. 3Department of Neurosurgery, Kansas University Medical Cen- ter, Kansas City, Kansas. 4Section of Pediatric Hematology/Oncology/ taining kinase AZK; referred to as MRK in this study; ref. 7), BMT, University of Manitoba, Winnipeg, Canada. 5Biostatistic Unit, belongs to the family of MAPK kinase kinase (MAPKKK) and has 6 Feinstein Institute, Manhasset, New York. Center for Molecular Inno- close homology to the MLK proteins (7–10). MRK has two splice vation, Feinstein Institute, Manhasset, New York. variants, MRKa and MRKb, which share some functions but differ Note: Supplementary data for this article are available at Molecular Cancer in other regards. In particular, while MRKa overexpression pro- Therapeutics Online (http://mct.aacrjournals.org/). motes tumor growth (11), MRKb has been shown to have an D. Markowitz and C. Powell are co-first authors of this article. important role in the response to DNA damage and in the S and M. Symons, Y. Al-Abed, and R. Ruggieri are co-senior authors of this article. G2–M checkpoints (12). At the mRNA level, the two MRK splice b The following authors, Yousef Al- Abed, Marc Symons and Rosamaria Ruggieri, forms are ubiquitously expressed (7, 8), although the MRK have submitted a US patent application: Treatment of solid tumors by inhibiting mRNA is more abundant than the MRKa mRNA. Recently, MRK/ZAK (#50425/507) 2015. This application, however, does not alter the however, the MRKa mRNA has been shown to be abnormally authors' adherence to all the Molecular Cancer Therapeutics policies on sharing spliced in gastric tumors as well as in colorectal, bladder, and data and materials. breast cancers with consequent overexpression of the respective Corresponding Authors: Rosamaria Ruggieri, The Feinstein Institute for Medical protein (13). MRKb is activated by stress (9, 14–16), including Research, 350 Community Drive, Manhasset, NY 11030. Phone: 516-562-3410; ionizing radiation (IR; refs. 7, 12, 17) upon which it contributes to Fax: 519-562-1022; E-mail: [email protected]; and Marc Symons, activation of Chk2 and p38 proteins, which leads to IR-induced [email protected] cell-cycle arrest. MRK silencing caused both failure to arrest cell- doi: 10.1158/1535-7163.MCT-15-0849 cycle progression in response to IR and increased killing by 2016 American Association for Cancer Research. radiation (12).

www.aacrjournals.org 1799

Markowitz et al.

In this study, we examine the role of MRK as a radiosensitiza- Human primary tumor cells and medulloblastoma xenograft tion target in medulloblastoma. In addition, we describe a novel model specific small molecule that we have designed to inhibit MRK The primary UI226 medulloblastoma cells are patient- activity and show that it works as an effective radiosensitizer in an derived xenografts that were established by Dr. Timothy Ryken, orthotopic xenograft model of medulloblastoma. University of Kansas Medical Center (Kansas City, KS). UI226 were largely propagated as flankculturesinnudemiceand Materials and Methods cultured in StemPro media (Life Technologies) for less than 3 weeks before intracranial injections. Cell lines and transfections Allanimalworkwascarriedout in accordance with NIH The human medulloblastoma cell line UW228 (18) was kindly guidelines and was approved by the Institutional Animal Care provided in 2007 by Dr. James Rutka, Hospital for Sick Children, and Use Committee of the Feinstein Institute in Manhasset, NY. Toronto, Canada, and it was authenticated by PCR. Cells were Medulloblastoma cells (5.0 105 UI226 in 5 mLofStemPro cultured in DMEM containing 10% FBS, 1 mmol/L glutamine and medium) were injected over 5 minutes into the cerebellum of 1% penicillin/streptomycin. UMDI cells are a derivative of the 4-week-old athymic female mice (Taconic), using a stereotactic osteosarcoma cell line U2-OS, described in Korkina and collea- frame and the following coordinates: 2 mm to the right and gues (19). In UMDI cells, expressing plasmid pC4-Fv1E-MRK, 1 mm posterior to the lambda, using a sterile dental drill. The MRK can be activated by induced homodimerization with the Hamilton 10-mL syringe loaded with cells was lowered 3 mm drug AP20187 (Clontech). The UMDI cells are cultured in McCoy under the surface of the brain and then lifted for 0.5 mm to media supplemented with 10% FBS, 0.6 mg/mL puromycin, create a pocket for the cells. Three weeks after tumor cell 50 IU/mL penicillin, 50 mg/mL streptomycin, and 2 mmol/L implantation and approximately 2 weeks before the animals L-glutamine. The K-562 and 4T1 cell lines were purchased from became moribund, treatment was started by implantation of an ATCC and cultured according to distributor's protocol. The human osmotic pump (Alzet, Durect) filled with either 0.05 mg/mL neural stem cells (NSC) were purchased from Invitrogen (Thermo solution of M443 or vehicle (0.01% DMSO in PBS) and Fisher Scientific) and cultured in StemPro media (Life Technolo- implanted in a subcutaneous pocket on the dorsal flank of the gies), as directed by the manufacturer. Human astrocytes were animal. A catheter with attached cannula delivered the drug purchased from Gibco (Life Technologies) and cultured in Gibco intracranially and directly into the tumor over a period of 2 astrocyte medium. weeks, at a steady rate of 0.25 mL/hour. Irradiation of the mice The MRK and control luciferase siRNAs sequences were previ- head was initiated 2 days after pump implantation and con- ously described (12) and transfected at a concentration of ducted over 2 days (two fractions of 3 Gy each). It was 10 nmol/L using Dharmafect #1 (Dharmacon) as recommended performed using a biologic irradiator (RS2000, Rad Source by the manufacturer. Technologies), whereas the rest of the body was shielded. A B Figure 1. MRK downregulation radiosensitizes medulloblastoma cells and inhibits downstream signals. UW228 cells were transfected with luciferase control (Luc) or MRK siRNAs, exposed to the indicated doses of radiation and cell survival was determined by the MTT assay (A) or by the clonogenic assay (B) as described in Materials and Methods. The Western blot analysis shows a typical level of MRK silencing. Data are the average SEM of three

Cell survival (% of control) independent experiments. P values for all treatment points were <0.005. Colony number (% of control) number Colony C, forty-eight hours after transfection with the respective siRNAs, cells were exposed to the indicated doses of radiation. Cell lysates were processed C for Western blotting with the indicated siRNAs: Luciferase MRK antibodies. Histograms show quantiﬁcation of phospho-Chk2 signals Time post IR (h): 01240124 from 3 independent experiments SE. MRK Data represent Chk2 speciﬁc activity, that is, p-Chk2/Chk2 ration. As the P-Chk2 values of p-Chk2 in the absence of radiation are very low and in some Chk2 experiments undetectable by Western (% of 1 hr activity)

Chk2-specific activity blot analysis, the Chk2 speciﬁc activity P-p38 data at the 1-hour time point for the Time (h): luciferase siRNA control cells were Tubulin siRNAs: Luciferase MRK chosen to normalize the other data.

1800 Mol Cancer Ther; 15(8) August 2016 Molecular Cancer Therapeutics

Downloaded from mct.aacrjournals.org on October 2, 2021. © 2016 American Association for Cancer Research. tteidctdcnetain(otm ntesrmfe eimadte ujce oMKatvto ihtehmdmrzrA217(P o hour 1 for (AP) AP20187 homodimerizer the domain. with site activation active MRK kinase to conserved subjected the then antibodies. of and indicated part the medium with serum-free blotting Western the for in harvesting before (bottom) concentration indicated the at oprsno ioii gen.M4 silsrtdatrfrigacvln odwt ytie2.Tenltnbsrcuewsetatdfo -b crystal c-Abl from extracted was structure nilotinib hidden. The are 22. hydrogens cysteine Nonpolar with gray. bond in covalent shown a are forming residues after MRK illustrated (PDB:3CS9). is structure M443 (green). nilotinib of comparison www.aacrjournals.org eino h niio fMK M443. MRK, of inhibitor the of Design 2. Figure D, ibnrpeetto ftesrcueo R oooymdli ope ihcmon 43(magenta). M443 compound with complex in model homology MRK of structure the of representation ribbon Downloaded from F D A Tubulin - P-MRK - Tubulin - P-MRK - N N N N H DMSO-EtOH

O DMSO-AP Published OnlineFirstMay20,2016;DOI:10.1158/1535-7163.MCT-15-0849 mct.aacrjournals.org H N DMSO-AP Nilotinib - AP N

N 40 nmol/L Nilotinib - AP A,

F Sorafenib - AP MIclsta xrs ooieial R eeperae o orwt 0.5 with hour 1 for pretreated were MRK homodimerizable express that cells UMDI F F 20 nmol/L Sorafenib - AP Motesanib - AP 700 nmol/L Imatinib - AP

nqeCs2 nMRK. in 22 Cys unique , Dasatinib - AP 100 nmol/L Dasatinib - AP Imatinib - AP on October 2,2021. ©2016 American Association forCancer Research. 100 nmol/L Motesanib - AP DHP2 - AP 20 nmol/L DHP2 - AP B, C B G mn cdsqec lgmn fteidctdpoen kontreso ioii)across nilotinib) of targets (known proteins indicated the of alignment sequence acid amino E C, mn cdsqec lgmn fteidctdpoen MKhomologs). (MRK proteins indicated the of alignment sequence acid amino pO31_A13-6IEVEEV FDVLEK sp|O43318_TAK1|36-66 LDLQWV sp|Q13043_MST|30-62 LRLEEV sp|Q12852_MUK|126-155 LQLEEI sp|Q16584_MLK3|117-147 LTLEEI sp|Q02779_MLK2|98-128 sp|P80192_MLK1|144-174 WEIVGE ITFEDL sp|P35590_TIE1|839-873 LSFGKT sp|O94804_LOK|36-68 LHLGRV sp|P10721_cKIT|589-626 LEFGKV sp|P35916_FLT4|845-882 LKLGKS sp|P36888_FLT|610-647 LTFKEK sp|P17948_FLT1|827-864 LRFKEK sp|Q16832_DDR2|563-611 sp|Q08345_DDR1|610-658 LRLEVK LKLGKP sp|P35968_VGFR2|834-871 VMLSTR sp|P12931_SRC|270-301 sp|P04049_RAF1|349-378 LVLGRT sp|P09619_PGFRB|600-637 LVLGRV ITVGQR sp|P16234_PGFAR|593-630 LQFFEN ITMKHK sp|P15056_BRAF|457-486 sp|P00519_ABL|242-274 sp|Q9NYL2_MRK|16-48 NO N F, N aisniiaino eulbatm yIhbto fMRK of Inhibition by Medulloblastoma of Radiosensitization ioii structure. nilotinib N H O H N N N F G, F F E, 43structure. M443 ls-pve ftebnigst fM4 in M443 of site binding the of view close-up m o acrTe;1()Ags 2016 August 15(8) Ther; Cancer Mol o/ fteidctdihbtr tp or (top) inhibitors indicated the of mol/L V L - I I I I L L L L L L L L L I L L I L C * GR GE GS GI GV GI GE GD GA GY GS GR GE GE GR GQ GS GS GS GS G G G G G G G G G G G G G G G G G G G G G G G G AFG SYG AQG GFG GFG GFG NFG AFG AFG AFG AFG AFG QFG QFG AFG CFG SFG AFG AFG SFG QYG SFG SVYKAI- AVFLGR- KVYRGS- KVYRAL- KVYRAF- QVIRAM- KVYKAK- KVVEAT KVVEAS KVMNAT KVVQAS EVHLCE EVHLCE QVIEAD EVWM--- TVYKGK- QVVEAT KVVEGT TVYKGK- EVYE--- SVYRAK- A A A A V V A A A 1801 Published OnlineFirst May 20, 2016; DOI: 10.1158/1535-7163.MCT-15-0849

Markowitz et al.

Western blot analysis and antibodies backbone of cAbl template (rms deviation from 0.37–0.43 Å). The Cells were processed for Western blotting as previously phi/psi backbone angles and side chain rotamers were evaluated described (19). The MRK mAb, 4–23, was previously described to discard structures with high-energy, disallowed bonds and the (7). The MRK phospho-specific mAb, A12, was generated against best model was selected for the following studies. the following di-phospho MRK peptide: FHNHpTTHMpS- The MRK homology model obtained from the previous step LVGTFP. The b-tubulin antibody was from Millipore; the Chk2, was subjected in Maestro (v. 10.2, Schrodinger,€ LLC, 2015). M443 P-Chk2, c-Abl, and P-cAbl antibodies were from Cell Signaling structure was constructed in 2D Sketcher and subsequently pre- Technology. The Hu-HLA antibody was from Abcam. The pared by LigPrep module (version 3.4, Schrodinger,€ LLC, 2015) in P-MPM2 antibody was purchased from Millipore. Maestro. The CovDock docking program (20) within the Schrodinger€ suite was employed for M443 and MRK covalent Cell viability and colony formation assay docking. Cysteine 22 on MRK was targeted as the reactive residue Two days after transfection with siRNAs, medulloblastoma to undergo Michael addition with compound M443. The ranking fi cells were seeded in a 96-well plate, and the following day, they of poses by af nity score calculates the average of the pre-reacted were exposed to different doses of radiation. Cell viability was and post-reacted Glide score for the given pose (20). The top fi fi determined 72 hours after radiation treatment by MTT (Sigma) ranking CovDock af nity pose was chosen as the nal pose of absorbance at 595 nm. For cells that were treated with M443, 6 M443. hours after treatment with different concentrations of the drug or vehicle control, they were exposed to radiation and processed for Immunofluorescence the MTT assay 72 hours later as above. UW228 cells were seeded at 70,000 cells/well on coverslips in a Clonogenic survival was determined using a colony formation 24-well plate. The following day, cells were treated with M443 or assay. Five hundred cells were seeded in 6-cm dishes 2 days after vehicle and 3 hours later exposed to 6 Gy of radiation. Cells were siRNA transfections. The following day, cells were exposed to the subsequently fixed at different time points and processed for different doses of radiation using a biologic irradiator (RS2000, immunofluorescence as previously described (21) with the P- Rad Source Technologies) and cultured for 7 days with media MPM2 antibody (Millipore) and counterstained with DAPI. changes every other day. Subsequently, cells were fixed in 3.7% Images were collected using a Zeiss axiovert 200M inverted formaldehyde in PBS and stained with 0.2% (w/v) sulforhoda- microscope, equipped with a 20 objective (0.3 NA) and a Zeiss mine B (SRB) dye in 1% acetic acid for 20 minutes. The dishes AxiocamMR camera running on Axiovision software. were washed with 1% glacial acetic acid and allowed to dry. Colonies containing more than 50 cells were counted, and the Statistical analysis results were used to calculate the surviving fractions. For the Two-way ANOVA was used to compare the mean of each treatment with M443 or vehicle control, 24 hours after seeding, outcome between groups of interest and either radiation dose or cells were exposed to the drug for 6 hours and subsequently to time. The interaction term between group and dose/time was also radiation. Colonies were fixed and processed as described above. examined. If a significant difference between means was found, The dose enhancement factor (DEF) was calculated from the then multiple comparisons between pairs were computed. The dose–response curves, as the ratio between the dose reducing Tukey–Kramer method was used to adjust for multiple compar- survival to 10% for radiation alone and that for radiation of MRK isons. Results were considered significant at a significance level of siRNA–transfected or M443-treated cells. P < 0.05. Analyses were conducted using SAS version 9.4 (SAS Institute, Inc.). The Kaplan–Meier estimate and a log-rank (Man- M443 synthesis tel–Cox) test were used to generate survival curves with the M443 was synthesized in 7 steps with a total yield of 3%. GraphPad Prism 6.0 software. To determine whether there was Methyl 3-amino-4-methylbenzoate (Aldrich CAS :18595-18-1) a synergistic effect between M443 and radiation treatment (IR), was reacted with cyanamide to obtain the guanidine M439 at the null hypothesis of an M443 IR interaction was tested in a 90% yield; the reaction of compound A (CAS: 858643-92-2) Cox model applied to a 2 2 factorial design [M443 (no/yes) IR with N,N-dimethylformamide dimethyl acetal yielded M422. (no/yes)]. Results were considered significant at a significance By refluxing M422 with M439 in absolute ethanol, M440 was level of P < 0.05. Analyses were conducted using SAS version 9.3 obtained with a yield of 75%. After saponification of ethyl ester (SAS Institute, Inc.). with NaOH in hot aqueous ethanol, M441 was obtained and coupled with compound C (CAS-641571-11-1) employing Results DECP (diethyl cyanophosphate) to provide the amide M442; Downregulation of MRK radiosensitizes medulloblastoma after deprotection with acid, coupling with acid chlorides cells completed the synthesis of M443. We have previously shown that MRK is activated by IR and controls the cellular response to DNA damage in osteosarcoma MRK homology model and M443 covalent docking and colorectal cancer cells (7, 12). To test whether MRK has a The MRK sequence was obtained from UniProt (ID: Q9NYL2). similar role in medulloblastoma cells, we asked whether down- A homology model was produced for MRK by alignment with the regulation of MRK expression can sensitize these cells to killing cAbl crystal structure (PDB: 3CS9) followed by three-dimensional by radiation. UW228 medulloblastoma cells were transfected model building and energy minimization using MOE (Molecular with control (luciferase) or specific MRK siRNAs and subsequent- Operating Environment, 2014.0901; Chemical Computing ly tested for viability after exposure to different doses of IR with Group Inc.). A database of 10 distinct protein models was gen- the MTT and clonogenic assays. Figure 1A shows that MRK erated and showed good overall alignment with the a-carbon depletion decreased cell viability after IR by 33% of control at

1802 Mol Cancer Ther; 15(8) August 2016 Molecular Cancer Therapeutics

Radiosensitization of Medulloblastoma by Inhibition of MRK

A Time:

Wash: – + – + – + P-MRK -

Figure 3. M443 inhibits MRK in an irreversible and Tubulin - speciﬁc fashion. A, time course of MRK inhibition by M443. UMDI cells were pretreated with 500 nmol/L M443 (M) or treated with DMSO for the indicated time. Then, cells were washed 3 times with fresh medium (þ), medium was DMSO Nilotinib M443 refreshed after 15 minutes twice, and B Wash: – ––++ AP was added for 30 minutes before harvesting and processing cells for P-MRK - Western blotting. B, UMDI cells were treated as in A with 100 nmol/L nilotinib or 500 nmol/L M443 for 6 hours. MRK - Histogram shows quantiﬁcation of data from three independent experiments, expressed as mean SD. C, K-562 cells were treated with 500 nmol/L M443,

Tubulin - P-MRK (% of control) nilotinib, or vehicle control for 6 hours and cell lysates were processed for Western blotting with the indicated antibodies. GAPDH was used as a loading control. The blot is a C representative of three independent Control M443 Nilotinib Wash: ––++ – experiments. DMSO M443 P-BCR-Abl - Nilotinib

P-MRK -

GAPDH

3 Gy. Similarly, the clonogenic assay showed a significant decrease nant form of MRK, which can be activated by homodimerization in survival with a dose enhancement factor (DEF) of 1.6 at 10% with the drug AP20187. Figure 2A shows that preincubation of viability (Fig. 1B). In both assays, downregulation of MRK expres- these cells with the different drugs led to various levels of MRK sion alone did not significantly affect viability by more than 5% inhibition. Among the inhibitors, we selected nilotinib, which (R. Ruggieri, unpublished observations). Analysis of the sig- had been previously shown by Manley and colleagues to bind naling elements known to be activated downstream of MRK with very high affinity to MRK (24).This study also included a (12) demonstrated a significant attenuation of IR-induced model of the MRK active site bound to nilotinib, based on the Chk2 phosphorylation at 1 and 2 hours after IR and effective corresponding structure of nilotinib bound to its original target c- reduction in p38 phosphorylation (Fig. 1C). The lack of Abl. We noticed that the pyridine ring in nilotinib was located in complete Chk2 inhibition following MRK downregulation close proximity to cysteine 22 in MRK, suggesting the possibility could be explained by the fact that Chk2 is known to be of designing a nilotinib derivative that can covalently interact with activated by multiple upstream factors that include DNA- MRK. Importantly, c-Abl has a leucine at the corresponding dependent protein kinase and Polo-like kinase-3 (22). Thus, position. In addition, none of the known off-targets of nilotinib MRK influences the DNA damage response in medulloblasto- (Fig. 2B) and none of the MRK family members or related kinases ma cells, and MRK downregulation sensitizes medulloblasto- (Fig. 2C) have a cysteine at this position. Thus, in an effort to ma cells to radiation. increase the specificity of nilotinib for inhibiting MRK, we decided to derivatize nilotinib by modifying the pyridine ring to a piper- Design of an irreversible inhibitor of MRK idine ring with the addition of an ab unsaturated amide to create To identify MRK-specific small molecule inhibitors, we selected compound M443 (see Materials and Methods, Fig. 2D–G and a number of known kinase inhibitors that had been tested against Supplementary Fig. S1). the kinome and found to bind to MRK (23). We assessed their We anticipated that M443 would bind covalently to MRK via ability to inhibit MRK kinase activity using a cell-based assay that cysteine 22 and irreversibly inhibit MRK activity (Fig. 2D and we previously described (19). Briefly, we made use of the M28 E). To examine whether M443 indeed inhibits MRK in an osteosarcoma cell derivative (UMDI) that expresses a recombi- irreversible manner, we pretreated UMDI cells with M443 for

www.aacrjournals.org Mol Cancer Ther; 15(8) August 2016 1803

Markowitz et al.

different times. After extensive washes, we treated cells with the A homodimerizing AP compound to activate MRK. We ﬁrst noted Time after IR (min): 0 15 30 60 120 that the extent of MRK inhibition by M443 was time-dependent P-MRK - and not reversed by washing out the inhibitor (Fig. 3A). As Tubulin - MRK inhibition was optimal after 6 hours of exposure to M443 (Fig. 3A), we chose this time point to compare M443 with nilotinib. Figure 3B shows that in contrast to M443, inhibition UW228 UI226 by nilotinib could be totally reversed after washing out the M443: – + – + drug. These results provide strong evidence that M443 is an IR: – + – + – + – + irreversible inhibitor of MRK. P-MRK -

MRK - Specificity of M443 To assess the specificity of M443, we compared the degree to P-Chk2 - which nilotinib and M443 inhibit the activity of BCR-Abl, the CHK2 - original target of nilotinib. To this end, we used the chronic myelogenous leukemia cell line K-562, which expresses the con- P-p38 - stitutively active BCR-Abl protein. Figure 3C shows that while both drugs inhibit MRK activity, nilotinib and M443 clearly Tubulin - diverge in their effect on BCR-Abl: nilotinib efficiently inhibits BCR-Abl whereas M443 does not. Thus, the nilotinib derivative M443 selectively inhibits MRK. B Control M443

M443 effectively inhibits MRK downstream signals in primary medulloblastoma cells and prevents cell-cycle arrest induced by IR MPM2 To examine whether M443 inhibits signaling downstream of MRK, we tested both the UW228 cell line as well as the UI226 primary patient–derived medulloblastoma cells for their response to IR in the presence and absence of M443. Figure 4A shows that MRK activation by IR is maximal at 30 minutes after exposure to

DAPI radiation. Therefore, for subsequent analysis, this time point was used. In both cell cultures, the IR-stimulated activation of MRK, Chk2, and p38 was greatly inhibited by 500 nmol/L M443. As Chk2 activity is important for the S and G2–M checkpoints, we also examined the effect of M443 on cell-cycle arrest after IR. To this end, we measured the mitotic index in UW228 cells after treatment with M443 and exposure to IR. Cells were seeded on

Bright field coverslips, pretreated with 500 nmol/L M443 or vehicle, exposed to 6 Gy of IR, fixed at different times after IR and processed for immunofluorescence with the MPM2 phospho-specific antibody that specifically stains mitotic cells. Figure 4B shows that in contrast to control cells, the M443-treated cells failed to arrest after IR and maintained a similar mitotic index as the nonirradi- Overlay ated cells. Thus, inhibition of MRK leads to inhibition of Chk2 and failure to arrest in the cell cycle in response to IR-induced DNA damage. 5

4 Figure 4. M443 inhibits MRK downstream signaling and prevents radiation-induced cell- cycle arrest. A, top, time course of MRK activation by IR. UW228 cells were 3 exposed to 3 Gy of IR, harvested at the indicated time points, and processed Control for Western blotting with the p-MRK antibody. Tubulin was used as loading 2 M443 control. Bottom, UW228 cells and primary UI226 cells were pretreated with 500 nmol/L M443 for 6 hours and then they were exposed to 3 Gy of radiation, P < 0.0001 and, 30 minutes later, they were harvested and processed for Western blotting 1 with the indicated antibodies. B, UW228 cells, seeded on coverslips, were P < 0.0001 pretreated with 500 nmol/L M443 for 3 hours, exposed to 6 Gy of radiation, and (% of total) MPM2-positive cells P < 0.0001 0 then processed for immunofluorescence with the phospho-specific MPM2 antibody as described in Materials and Methods. Mitotic cells from 10 fields per Time (h): 0 1 2 3 4 5 condition were calculated from 3 independent experiments and plotted in the (post IR) -1 graph below. Bars, 50 mm.

1804 Mol Cancer Ther; 15(8) August 2016 Molecular Cancer Therapeutics

Radiosensitization of Medulloblastoma by Inhibition of MRK

M443 radiosensitizes medulloblastoma cells A As a consequence of failed cell-cycle arrest in M443-treated 100 medulloblastoma cells, we would expect to see increased sensi- 0 nmol/L tivity to IR. To conﬁrm this, we used both UI226 and UW228 P < 0.0001 250 nmol/L 500 nmol/L cells. As the UI226 medulloblastoma primary cells do not form P < 0.0001 colonies when plated sparsely, we tested colony formation by P < 0.0001 the UW228 cell line and used the MTT assay for the UI226 10 cells. Figure 5 shows that M443 effectively radiosensitizes both P < 0.0001 cell cultures to radiation. The clonogenic assay in UW228 cells showed a DEF at 10% viability of 1.6, which is the same observed upon silencing of MRK in these cells (Fig. 1B). A similar response P < 0.0001 was observed with the parental compound, nilotinib (Fig. 5B). 1

The MTT assay showed that M443 in conjunction with IR caused a Colony number (% of control) 48% reduction in viability of the primary medulloblastoma IR (Gy): 0 2 4 6 cells in comparison to IR alone. Interestingly, these cells responded maximally already to 125 nmol/L of M443, the lowest dose used, indicating that the IC50 for the radiosensitization effect B is lower than 125 nmol/L. Similar results were obtained with a 0 nmol/L second medulloblastoma primary line (R. Ruggieri, unpublished 100 500 nmol/L 1,000 nmol/L observations). Importantly, M443 did not show any toxicity in P < 0.0006 the absence of radiation and failed to radiosensitize normal neuronal stem cells and astrocytes (Supplementary Fig. S2). Thus, P < 0.017 P < 0.0001 these data indicate that M443 can be used to radiosensitize medulloblastoma tumors without compromising the surround- 10 P < 0.004 ing parenchyma. To test the extent to which M443 radiosensitizes tumors other than medulloblastoma, we tested breast cancer 4T1 cells for their response to the drug in conjunction with IR. Supplementary Figure S3 shows that M443 can radiosensitize breast cancer cells 1 as well. Colony number (% of control) IR (Gy): 0 2 4 6 M443 synergizes with IR to extend survival in an orthotopic medulloblastoma xenograft model To examine whether M443 can act as a radiosensitizer in a C 0 nmol/L mouse model of medulloblastoma, we used the primary UI226 100 125 nmol/L cells to generate an orthotopic medulloblastoma tumor in 250 nmol/L nude mice. Previous work had established that approximately 80 500 nmol/L 3 weeks after tumor cell implantation, mice become symp- tomatic within a relatively narrow time frame, just over a week, 60 indicating similar tumor growth rates among the animals. Two P < 0.0001 weeks before mice became moribund, they were randomized 40 into four groups: control, treated with M443 alone, treated P < 0.0001 with 6 Gy of IR (2 3 Gy over 2 days), and with the 20 combination of M443 and IR. M443 was delivered directly 0 into the tumor via an osmotic pump, and 2 days later, mice Cell viability (% of control) were exposed to IR. IR (Gy): 0 3 6 Control mice survived with a median of 32 days after tumor cell implantation. Treatment with M443 alone added 5.5 days to this Figure 5. M443 radiosensitizes medulloblastoma cells in vitro. UW228 cells were survival, whereas the chosen low dose of radiation did not fi pretreated with the indicated doses of M443 (A) or nilotinib (B), exposed to the signi cantly increase survival (median of 1 day additional survival indicated doses of radiation, and then processed for the clonogenic assay as over that of control). In contrast, the combination of M443 and IR described in Materials and Methods. C, UI226 primary medulloblastoma cells extended survival with a median of 16 days longer than control. were pretreated with the indicated doses of M443, exposed to various doses of Statistical analysis of the data demonstrated a synergistic effect of radiation, and processed for the MTT assay 72 hours later. the two combined treatments (Fig. 6A). Supporting the safety data shown in Supplementary Fig. S2, treatment with M443 did not affect the animal weight, as the weight loss observed was observed harvested and their cerebella were divided into four sections that in all groups just a few days before the animals became moribund were then examined by Western blotting with phospho-specific (Supplementary Fig. S4). and total MRK antibodies. The tumor-containing sections of the To confirm in vivo target inhibition by M443, we tested MRK cerebellum were identified by the human anti-HLA antibody, levels and activity in brains from treated and untreated mice 18 because despite performing the tumor cell injections in a standard hours after exposure to the drug. For this assay, the brains were fashion, it is difficult to predict the invasion and growth path of

www.aacrjournals.org Mol Cancer Ther; 15(8) August 2016 1805

Markowitz et al.

A IR treatment * * toma and designed a small molecule inhibitor M443 that irre- Drug treatment versibly blocks its activity. Importantly, administration of M443 in vivo effectively synergizes with radiation in a mouse model of 100 medulloblastoma, where it increases the survival time by 50%. Control This result provides a proof of principle that MRK is a target for M443 radiosensitization in medulloblastoma. IR We have developed an irreversible inhibitor of MRK, M443, 50 IR+M443 based on the close proximity of the M443 parental compound nilotinib to cysteine 22 in the model structure of the MRK ATP- binding domain (24). The unique presence of this cysteine at this position in MRK, compared with other nilotinib targets Percent survival and other MRK-related proteins, suggested that an irreversible 0 inhibitor of MRK could provide high selectivity toward MRK, 0204060which is likely to result in reduced off-target effects of a Time (days after tumor cell implantation) treatment based on this molecule. Our initial investigation on the effects of M443 on BRC-Abl, the original target of nilotinib, indicates that M443 selectively inhibits MRK, but not BCR-Abl. B Although a wider specificity screen is needed, this result sug- Treatment: Vehicle M443 gests good selectivity for this compound. RT RB LT LB RT RB LT LB cells The past reservations about irreversible inhibitors that were P-MRK - based on safety concerns recently have been mitigated by several clinical successes. In fact, irreversible kinase inhibitors for EGF MRK - receptor (EGFR), Bruton tyrosine kinase (BTK), and MEK1 are being tested in the clinic as anticancer agents (25). Importantly, hu-HLA - this type of inhibitors is expected to have extended pharmaco- dynamic duration of inhibition, which could allow for lower Tubulin - circulating doses with less potential side effects. Although the toxicity profile of M443 needs to be assessed after systemic in vivo Figure 6. administration , it is interesting to note that we did not M443 synergizes with radiation to prolong survival in a murine model of observe any apparent toxicity in tumor cells treated with M443 medulloblastoma. A, Kaplan–Meier survival curves demonstrating synergism (R. Ruggieri, unpublished observations), nor in animals based on between M443 and radiation treatment. n ¼ 10–12 mice/group. Time refers to lack of weight loss following drug administration. In addition, days after intracranial tumor cell injections. The bar and asterisks above the no effect was observed on normal brain cells (Supplementary Kaplan–Meier curves indicate the respective treatment periods (2 weeks of Fig. S2). We believe that this characteristic of the MRK inhibitor is continues exposure to the drug via the osmotic pump and 2 doses of IR indicated by the asterisks). B, in vivo target inhibition and specificity of M443. Eighteen particularly important in the context of pediatric brain tumors for hours after pump implantation, brains were harvested and each cerebellum was which safer therapeutic approaches are needed. Future character- sectioned in four parts: right top (RT), right bottom (RB), left top (LT), and left ization of the pharmacokinetic properties of M443 will establish bottom (LB). Tissue extracts were processed for Western blotting with the the extent to which this compound crosses the blood–brain indicated antibodies. The hu-HLA antibody identifies the tumor-containing barrier. fractions. Analysis of the tumor tissue after convection delivery revealed a clear in vivo inhibition of MRK activity by M443. The phospho- the tumor. Figure 6B shows that the tumor-containing fraction specific MRK antibody should be a very valuable tool to identify had elevated levels of both total and active MRK (lane RB in the patients that might have elevated activation levels of MRK and vehicle-treated brain). In contrast, the tumor-containing fraction therefore may be more responsive to a treatment that includes from the M443-treated brain showed total loss of MRK activity. MRK inhibition. Interestingly, the MRK protein levels were also reduced in the M443 inhibits MRK downstream pathways that include IR- treated fractions. This result is consistent with our previous induced activation of Chk2 and the consequent cell-cycle arrest. observation that sustained inhibition of MRK leads to its insta- The fact that these effects of MRK silencing were replicated by bility (R. Ruggieri, unpublished observation). Interestingly, the M443 supports the specificity of these phenotypes and strength- fractions containing normal brain, which in the control brain ens a role for the MRK-b isoform in the DNA damage response, showed some level of MRK protein, in the treated brain also had previously observed in other cell types (12). Both Chk2 and p38 lost MRK, suggesting that diffusion of M443 across the whole have been shown to control cell-cycle arrest after DNA damage by cerebellum inhibited normal MRK as well. regulating the activity of the CDC25 family of proteins and thus Taken together, our study demonstrates that MRK is a suitable the activity of the CDK proteins (26). The role of p38a in the DNA target for radiosensitization of medulloblastoma tumors and damage response has, however, been demonstrated only in likely additional types of tumors. response to UV-induced damage (27). The activation of p38a by IR in medulloblastoma is clearly dependent on MRK activity and it may point to either a role for p38a in IR-induced DNA damage, as Discussion suggested for the response to UV, or to additional roles that In this study, we have identified the protein kinase MRK/ZAK as MRK may have in response to IR. As we have shown previously, a promising novel target for radiosensitization in medulloblas- IR-induced MRK activation is dependent on Nbs1 and partially on

1806 Mol Cancer Ther; 15(8) August 2016 Molecular Cancer Therapeutics

Radiosensitization of Medulloblastoma by Inhibition of MRK

ATM (17), two proteins involved in the DNA damage response ing MRK/ZAK (2015). No potential conflicts of interest were disclosed by the and in cell-cycle checkpoint regulation (28, 29). In turn, MRK other authors. activates Chk2, a checkpoint protein that, together with Chk1, is considered a good target for radiosensitization in cancer and for Authors' Contributions which inhibitors are actively being evaluated in the clinic (30, 31). Conception and design: D. Markowitz, M.E. Berens, S. Sun, M. Symons, Y. Al-Abed, R. Ruggieri Thus, M443 represents an additional novel approach to target the Development of methodology: N.L. Tran, M. He, S. Sun, Y. Al-Abed, R. Ruggieri DNA damage response in cancer. Acquisition of data (provided animals, acquired and managed patients, As the two splice forms of MRK share the same kinase provided facilities, etc.): D. Markowitz, C. Powell, T.C. Ryken, M. Vanan, domain, M443 is expected also to inhibit MRKa.Thus,because R. Ruggieri MRKa is aberrantly expressed in cancers like gastric, colorectal, Analysis and interpretation of data (e.g., statistical analysis, biostatistics, bladder, and breast tumors, where it plays a critical role in computational analysis): D. Markowitz, C. Powell, M. Vanan, L. Rosen, S. Sun, M. Symons, R. Ruggieri tumor cell proliferation, M443 also may have therapeutic Writing, review, and/or revision of the manuscript: N.L. Tran, T.C. Ryken, effects in these cancers (13). Interestingly, we found that in M. Vanan, L. Rosen, S. Sun, M. Symons, R. Ruggieri addition to medulloblastoma cells, breast cancer cells are also Administrative, technical, or material support (i.e., reporting or organizing sensitized to radiation following MRK inhibition by M443 data, constructing databases): C. Powell, M. Vanan, M. Symons (Supplementary Fig. S2). In addition, we found that siRNA- Study supervision: M. Symons, R. Ruggieri mediated MRK silencing radiosensitizes several types of cancer Other (synthesized the MRK inhibitor/M443): M. He Other (design and synthesis of the pharmacological inhibitor M443): cells including osteosarcoma and glioblastoma cells (ref. 12 Y. Al-Abed and data not shown). Thus, inhibition of MRK may be a useful approach to radiosensitize different types of solid tumors. Acknowledgments Future studies will investigate this possibility. The authors thank James Rutka (The Hospital for Sick Children) for kindly In summary, we have developed a specific and irreversible providing the UW228 cells and Amanda Chan (The Feinstein Institute) for her inhibitor of MRK that effectively radiosensitizes medulloblasto- help with microscopy data acquisition. ma tumors and potentially other cancers as well. Future development of this small molecule may provide a new therapeutic Grant Support approach for radiosensitization in cancer. This study was supported by the following fund sources: Swim Across America (SAA) Foundation and Project to Cure (PTC) Foundation (to M. Symons), the Bradley Foundation (to R. Ruggieri), and The Ben & Catherine Disclosure of Potential Conflicts of Interest Ivy Foundation (to M.E. Berens). R. Ruggieri has ownership interest as co-inventor on patent application The costs of publication of this article were defrayed in part by the payment of advertisement for treatment of solid tumors by inhibiting MRK/ZAK (2015). Y. Al-Abed page charges. This article must therefore be hereby marked in ownership interest as co-inventor on patent application for treatment of accordance with 18 U.S.C. Section 1734 solely to indicate this fact. solid tumors by inhibiting MRK/ZAK (2015) and is president and co- founder of Fatimo Innovation LLC. M. Symons has ownership interest as Received November 9, 2015; revised May 5, 2016; accepted May 10, 2016; co-inventor on patent application for treatment of solid tumors by inhibit- published OnlineFirst May 20, 2016.

References 1. Bentzen SM, Heeren G, Cottier B, Slotman B, Glimelius B, Lievens Y, et al. 9.GotohI,AdachiM,NishidaE.Identiﬁcation and characterization Towards evidence-based guidelines for radiotherapy infrastructure and of a novel MAP kinase kinase kinase, MLTK. J Biol Chem 2001; stafﬁng needs in Europe: the ESTRO QUARTS project. Radiother Oncol 276:4276–86. 2005;75:355–65. 10. Liu TC, Huang CJ, Chu YC, Wei CC, Chou CC, Chou MY, et al. Cloning and 2. Packer RJ, Vezina G. Management of and prognosis with medulloblastoma: expression of ZAK, a mixed lineage kinase-like protein containing a therapy at a crossroads. Arch Neurol 2008;65:1419–24. leucine-zipper and a sterile-alpha motif. Biochem Biophys Res Commun 3. Christopherson KM, Rotondo RL, Bradley JA, Pincus DW, Wynn TT, Fort J A, 2000;274:811–6. et al. Late toxicity following craniospinal radiation for early-stage medul- 11. Cho YY, Bode AM, Mizuno H, Choi BY, Choi HS, Dong Z. A novel role for loblastoma. Acta Oncol 2014;53:471–80. mixed-lineage kinase-like mitogen-activated protein triple kinase alpha in 4. Fossati P, Ricardi U, Orecchia R. Pediatric medulloblastoma: toxicity of neoplastic cell transformation and tumor development. Cancer Res current treatment and potential role of proton therapy. Cancer Treat Rev 2004;64:3855–64. 2009;35:79–96. 12. Tosti E, Waldbaum L, Warshaw G, Gross EA, Ruggieri R. The stress 5. Gajjar A, Chintagumpala M, Ashley D, Kellie S, Kun LE, Merchant TE, et al. kinase MRK contributes to regulation of DNA damage checkpoints Risk-adapted craniospinal radiotherapy followed by high-dose chemo- through a p38gamma-independent pathway. J Biol Chem 2004;279: therapy and stem-cell rescue in children with newly diagnosed medullo- 47652–60. blastoma (St Jude Medulloblastoma-96): long-term results from a pro- 13. Liu J, McCleland M, Stawiski EW, Gnad F, Mayba O, Haverty PM, et al. spective, multicentre trial. Lancet Oncol 2006;7:813–20. Integrated exome and transcriptome sequencing reveals ZAK isoform usage 6. Moxon-Emre I, Bouffet E, Taylor MD, Laperriere N, Scantlebury N, Law N, in gastric cancer. Nat Commun 2014;5:3830. et al. Impact of craniospinal dose, boost volume, and neurologic compli- 14. Mao X, Bravo IG, Cheng H, Alonso A. Multiple independent kinase cations on intellectual outcome in patients with medulloblastoma. J Clin cascades are targeted by hyperosmotic stress but only one activates stress Oncol 2014;32:1760–8. kinase p38. Exp Cell Res 2004;292:304–11. 7. Gross EA, Callow MG, Waldbaum L, Thomas S, Ruggieri R. MRK, a mixed 15. Takahashi M, Gotoh Y, Isagawa T, Nishimura T, Goyama E, Kim HS, et al. lineage kinase-related molecule that plays a role in gamma-radiation- Regulation of a mitogen-activated protein kinase kinase kinase, MLTK by induced cell cycle arrest. J Biol Chem 2002;277:13873–82. PKN. J Biochem (Tokyo) 2003;133:181–7. 8. Bloem LJ, Pickard TR, Acton S, Donoghue M, Beavis RC, Knierman MD, 16. Wang X, Mader MM, Toth JE, Yu X, Jin N, Campbell RM, et al. Complete et al. et al.Tissue distribution and functional expression of a cDNA encoding inhibition of anisomycin and UV radiation but not cytokine induced JNK a novel mixed lineage kinase. J Mol Cell Cardiol 2001;33:1739–50. and p38 activation by an aryl-substituted dihydropyrrolopyrazole

www.aacrjournals.org Mol Cancer Ther; 15(8) August 2016 1807

Markowitz et al.

quinoline and mixed lineage kinase 7 small interfering RNA. J Biol Chem 24. Manley PW, Drueckes P, Fendrich G, Furet P, Liebetanz J, Martiny- 2005;280:19298–305. Baron G, et al. Extended kinase proﬁle and properties of the 17. VananI,DongZ,TostiE,WarshawG,SymonsM,RuggieriR.Roleofa protein kinase inhibitor nilotinib. Biochim Biophys Acta 2010;1804: DNA damage checkpoint pathway in ionizing radiation-induced glio- 445–53. blastoma cell migration and invasion. Cell Mol Neurobiol 2012; 25. Liu Q, Sabnis Y, Zhao Z, Zhang T, Buhrlage SJ, Jones LH, et al. Developing 32:1199–208. irreversible inhibitors of the protein kinase cysteinome. Chem Biol 18. Keles GE, Berger MS, Srinivasan J, Kolstoe DD, Bobola MS, Silber JR. 2013;20:146–59. Establishment and characterization of four human medulloblastoma- 26. Reinhardt HC, Yaffe MB. Kinases that control the cell cycle in response derived cell lines. Oncol Res 1995;7:493–503. to DNA damage: Chk1, Chk2, and MK2. Curr Opin Cell Biol 2009; 19. Korkina O, Dong Z, Marullo A, Warshaw G, Symons M, Ruggieri R. The 21:245–55. MLK Related Kinase (MRK) is a novel RhoC effector that mediates Lysopho- 27. Bulavin DV, Higashimoto Y, Popoff IJ, Gaarde WA, Basrur V, Potapova O, sphatidic Acid (LPA)-stimulated tumor cell invasion. J Biol Chem et al. Initiation of a G2/M checkpoint after ultraviolet radiation requires 2013;288:5364–73. p38 kinase. Nature 2001;411:102–7. 20. Zhu K, Borrelli KW, Greenwood JR, Day T, Abel R, Farid RS, et al. Docking 28. Kuroda S, Urata Y, Fujiwara T. Ataxia-telangiectasia mutated and the covalent inhibitors: a parameter free approach to pose prediction and Mre11-Rad50-NBS1 complex: promising targets for radiosensitization. scoring. J Chem Inf Model 2014;54:1932–40. Acta Med Okayama 2012;66:83–92. 21. Chuang YY, Tran NL, Rusk N, Nakada M, Berens ME, Symons M. Role of 29. Shiloh Y. ATM: expanding roles as a chief guardian of genome stability. Exp synaptojanin 2 in glioma cell migration and invasion. Cancer Res Cell Res 2014;329:154–61. 2004;64:8271–5. 30. Matthews TP, Jones AM, Collins I. Structure-based design, discovery and 22. Zannini L, Delia D, Buscemi G. CHK2 kinase in the DNA damage response development of checkpoint kinase inhibitors as potential anticancer ther- and beyond. J Mol Cell Biol 2014;6:442–57. apies. Expert Opin Drug Discov 2013;8:621–40. 23. Bergh J. Quo vadis with targeted drugs in the 21st century? J Clin Oncol 31. Weber A, Ryan AJ. ATM and ATR as therapeutic targets in cancer. Pharmacol 2009;27:2–5. Ther 2014;149:124–38.

1808 Mol Cancer Ther; 15(8) August 2016 Molecular Cancer Therapeutics

Pharmacological Inhibition of the Protein Kinase MRK/ZAK Radiosensitizes Medulloblastoma

Daniel Markowitz, Caitlin Powell, Nhan L. Tran, et al.

Mol Cancer Ther 2016;15:1799-1808. Published OnlineFirst May 20, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-15-0849

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2016/05/20/1535-7163.MCT-15-0849.DC1

Cited articles This article cites 31 articles, 9 of which you can access for free at: http://mct.aacrjournals.org/content/15/8/1799.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/15/8/1799.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/15/8/1799. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.