Loss of MAP3K7 Sensitizes Prostate Cancer Cells to CDK1/2 Inhibition and DNA Damage by Disrupting Homologous Recombination Satoshi Washino1, Leah C

Published OnlineFirst July 12, 2019; DOI: 10.1158/1541-7786.MCR-18-1335

Cancer Genes and Networks Molecular Cancer Research Loss of MAP3K7 Sensitizes Prostate Cancer Cells to CDK1/2 Inhibition and DNA Damage by Disrupting Homologous Recombination Satoshi Washino1, Leah C. Rider1, Lina Romero1, Lauren K. Jillson1, Trisiani Affandi2, Angela M. Ohm2, Elaine T. Lam3, Mary E. Reyland2, James C. Costello1, and Scott D. Cramer1

Abstract

The combined loss of CHD1 and MAP3K7 promotes aggres- without loss of Map3k7. Dinaciclib treatment reduced expres- sive prostate cancer by unknown mechanisms. Because both of sion of homologous recombination (HR) repair genes such as these genes are lost genetically in prostate cancer, they cannot ATM, ATR, BRCA2, and RAD51, blocked BRCA1 phosphory- be directly targeted. We applied an established computational lation, reduced RAD51 foci formation, and increased gH2AX systems pharmacology approach (TRAP) to identify altered foci selectively in prostate cells with suppression of Map3k7, signaling pathways and associated druggable targets. We com- thus inhibiting HR repair of chromosomal double-strand pared gene expression proﬁles of prostate cancer with coloss of breaks. Dinaciclib-induced HR disruption was also observed CHD1 and MAP3K7 with prostate cancer diploid for these in human prostate cells with knockdown of MAP3K7. Cotreat- genes using The Cancer Genome Atlas patient samples. This ment of dinaciclib with DNA-damaging agents or PARP inhib- analysis prioritized druggable target genes that included CDK1 itor resulted in a stronger cytotoxic effect on prostate cells with and CDK2. We validated that inhibitors of these druggable suppression of MAP3K7 compared with those without loss of target genes, including the CDK1/ CDK2 inhibitor dinaciclib, MAP3K7, or to each single agent. had antiproliferative and cytotoxic effects selectively on mouse prostate cells with knockdown of Chd1 and Map3k7. Dinaci- Implications: These ﬁndings demonstrate that loss of clib had stronger effects on prostate cells with suppression of MAP3K7 is a main contributing factor to drug response Map3k7 independent of Chd1 and also compared with cells through disruption of HR in prostate cancer.

Introduction which encodes TGFb-activated kinase 1 (TAK1; ref. 4; P < 0.001; Fisher exact test). Homozygous or heterozygous deletion of Prostate cancer is characterized by large genomic rearrange- MAP3K7 in primary prostate cancer is significantly correlated ments including translocations, copy number losses, and with high Gleason grade tumors (2, 3). We previously demon- gains (1). MAP3K7 is deleted in 30%–40% of primary prostate strated that genomic deletion of CHD1 and MAP3K7 correlates cancer (homozygous and heterozygous deletion combined; with their respective mRNA expression in multiple patient cohorts refs. 2, 3). Loss of the Chromodomain Helicase DNA-binding and that low mRNA expression is correlated with poor disease- protein 1 (CHD1) gene (located on 5q21.1), which occurs in free survival (5). We recently demonstrated that MAP3K7 is a 20%–30% of prostate cancer (4, 5), is associated with an increased prostate tumor suppressor and that coloss with CHD1 promotes frequency of copy number alterations (CNA) and massive geno- aggressive prostate cancer (3, 5). The CHD1 gene encodes a mic rearrangements, termed chromothripsis (4). Not all of the histone-associated protein that is required for DNA double- CNAs associated with loss of CHD1 are random (4). CHD1 loss is strand break repair and CHD1 depletion enhances PARP inhibitor significantly associated with loss of MAP3K7 (located on 6q15.1), therapy (6). MAP3K7 encodes MAP kinase kinase kinase 7 that is immediately downstream of cytokine and stress response path- 1Department of Pharmacology, University of Colorado Anschutz Medical 2 ways. Loss of these two tumor suppressors represents a molecu- Campus, Aurora, Colorado. Department of Craniofacial Biology, University of fi Colorado Anschutz Medical Campus, Aurora, Colorado. 3Department of Internal larly de ned subtype of aggressive prostate cancer with unknown Medicine, Division of Medical Oncology, University of Colorado Anschutz therapeutic sensitivities. Because both of these genes are lost Medical Campus, Aurora, Colorado. genetically in prostate cancer, they cannot be directly targeted. Note: Supplementary data for this article are available at Molecular Cancer Therefore, the pathways that are altered must be characterized to Research Online (http://mcr.aacrjournals.org/). identify therapeutic vulnerabilities. To identify potential signaling pathways altered in prostate J.C. Costello and S.D. Cramer contributed equally to this article. cancer with loss of CHD1 and/or MAP3K7, we applied an estab- Corresponding Author: Scott D. Cramer, University of Colorado, Anschutz lished computational systems pharmacology approach, Tran- Medical Center, 12801 E 17th Ave, L18-6117, Mail Stop 8303, PO Box 6511, Aurora, scriptional Regulatory Association with Pathways (TRAP; ref. 7), CO 80045. Phone: 303-724-6276; Fax: 303-724-3663; E-mail: [email protected] to gene expression data in The Cancer Genome Atlas (TCGA; ref. 8). Through this approach, we identified a number of altered Mol Cancer Res 2019;XX:XX–XX signaling pathways and a set of predicted druggable targets. We doi: 10.1158/1541-7786.MCR-18-1335 investigated several targets in prostate cells with suppression of 2019 American Association for Cancer Research. CHD1 and MAP3K7 expression and identified several promising

www.aacrjournals.org OF1

Washino et al.

drugs, including dinaciclib, a multiple cyclin-dependent kinase luminescence was read using a Modulus Microplate reader. For (CDK) inhibitor. We showed that loss of MAP3K7 was the main WFU3 and LNCaP, cell proliferation was determined by the contributing alteration driving increased sensitivity to these percent conﬂuence of each well using Incucyte Zoom scanning. drugs, and that treatment with a CDK inhibitor paired with either The proportion of cells per treatment group was normalized to PARP inhibitor or DNA-damaging agents resulted in potent cell control wells when indicated. Annexin V- and PI-positive cell death in cells with loss of MAP3K7. counts were normalized to percent conﬂuence in each well.

Cell-cycle analysis and cell sorting Materials and Methods Staining of cells for cell-cycle analysis was performed as Agents described previously (11). Briefly, WFU3 (2–5 105) and All drugs were purchased from Selleckchem except for doce- BHPrE1 cells (1 105) were seeded in a 6-well plate and taxel that was purchased from Sigma-Aldrich. incubated overnight. WFU3 cells were exposed to vehicle, olaparib 10 mmol/L, dinaciclib 60 nmol/L, or the combination of Cell culture these drugs for 24 hours prior to collection. BHPrE1 cells were All cell lines were routinely tested for Mycoplasma contamina- exposed to vehicle, docetaxel 1 nmol/L, dinaciclib 4 nmol/L, or tion by DAPI staining and PCR. LNCaP cells were obtained from the combination of these drugs for 24 hours prior to collection. ATCC and used below passage 50. LNCaP cells were validated by Cells were washed once with PBS, pelleted, and vortexed before spectral karyotyping. Generation and maintenance of WFU3 adding 1 mL Krishan stain solution (11). Cells were vortexed mouse prostate epithelial progenitor/stem cells were described again and incubated overnight at 4C before analysis with a previously (9). Wild-type WFU3 cells generate benign ductal Gallios instrument using CXP software (Beckman Coulter). A structures and represent a clean model to dissect the contributions total of 10,000 events were acquired for each sample and the of specific genetic alterations to therapeutics and signaling (5, 9). percent of cells in G1, S, and G2–M phases was determined with LNCaP cells were maintained as described previously (5). Gen- ModFit LT v4.0.5 software (Verity Software House) eration of shControl, shMAP3K7, shCHD1, and shDouble cell Cell sorting for the cell-cycle phases was performed as lines in WFU3 and LNCaP was described previously (5). BHPrE1 follows: WFU3 cells (0.3–2 106) were seeded in a 15-cm cells, a spontaneously immortalized benign prostate cell line dish and incubated overnight. Cells were exposed to vehicle, or that generates benign ductal structures in vivo, were a generous the combination of dinaciclib and olaparib for 36 hours prior gift from Simon Hayward NorthShore University HealthSystem to collection. Cells were washed once with PBS, pelleted, and Research Institute, Evanston, IL (10). As there is no record of prior vortexed before adding Hoechst33342 (Thermo Fisher Scien- spectral karyotyping, we validated that BHPrE1 cells had a unique tific, #H3570). Cells were incubated in a medium containing karyotype. BHPrE1 cells were maintained and cultured in DMEM/ 60 mg/mL of Hoecht33342 for 45 minutes with rocking in an F12 with 1% insulin–transferrin selenium (Gibco, #41400045), incubator. Around 4 106 cells in each condition were sorted 1% bovine pituitary extract (Hammond Cell Technology, #1078- on the basis of G1,S,andG2–M phases, respectively, using UV NZ), 10 mg/mL EGF (Corning, #354001), 1% penicillin/strepto- laser with MoFlo XDP70 (Beckman Coulter). Collected cells mycin (Thermo Fisher Scientific, #15140122), and 5% FBS were harvested and proteins were isolated. (Gemini Bio-Products, #100-106). Generation of shControl and shMAP3K7 BHPrE1 cell lines was performed as follows. For Immunoblotting knockdown (KD) of MAP3K7, cells were subjected to lentiviral Standard techniques were used for protein isolation, quantifi- transduction and infection with shMAP3K7-puro or shControl- cation, and immunoblotting. The following antibodies were puro lentiviruses. Production of lentivirus for these shRNA plas- used: MAP3K7 (Cell Signaling Technology, #4505), CHD1 mids was described previously (5). Cells were infected with (Bethyl Laboratories, #A301-218A), BRCA2 (Thermo Fisher Sci- lentivirus in growth media containing 8 mg/mL polybrene and entific, #MA5-23942), phospho-BRCA1 (Ser 1524; Thermo Fisher selected with 12 mg/mL puromycin (Sigma, #9620) after 48 hours. Scientific, #PA5-17503), RAD51 (Abcam, #ab88572), anti-phos- Cells were maintained in medium with 12 mg/mL puromycin. pho-Histone H2A.X (Ser139; Millipore, #JBW201), p53 (Santa Cruz Biotechnology, #sc-6243), phospho-ATR (Ser428; Cell Sig- Cell proliferation and cell death assay naling Technology, #2853), ATR (Thermo Fisher Scientific, #PA1- Cells were cultured at 37 C and 5% CO2 and seeded at 1,000 450), ATM (Cell Signaling Technology, #2873), caspase 3 cells (WFU3) and 4,000 cells (BHPrE1 and LNCaP) per well in a (Cell Signaling Technology, #9662), cleaved caspase-3 (Asp175; 96-well plate (Costar). Cells were cultured at 37 C and 5% CO2 in Cell Signaling Technology, #9661), b-actin (Cell Signaling Tech- an IncuCyte Zoom Live Cell Analysis system (Essen BioScience). nology, #3700), a-tubulin (Cell Signaling Technology, #2144), Cells were exposed to drugs for up to 48 hours for WFU3, 36 hours goat anti-rabbit-HRP (Santa Cruz Biotechnology, #sc-2004), and for LNCaP, and 72 hours for BHPrE1 in normal growth medium goat anti-mouse-HRP (Santa Cruz Biotechnology, #sc-2005). with or without 125 ng/mL of Annexin V (CF488 Annexin V, Where appropriate, bands were quantified using ImageJ Biotium #29005 for WFU3, or CF633 Annexin V, and Biotium (NIH, Bethesda, MD). #29008 for LNCaP and BHPrE1) and/or 1.0 mg/mL of propidium iodide (PI; Biotium, #40016), and images were captured at 2-, 3-, Annexin V and PI assay using flow cytometry or 4-hour intervals using a 4 objective. Each experiment was WFU3 cells were seeded at 2–5 105 in a 6-well plate or 1.5 done in triplicate unless otherwise indicated. After monitoring in 106 in a 60-mm dish and incubated overnight. Cells were exposed the Incucyte Zoom, cell proliferation assay for BHPrE1 was to either vehicle or the combination of these drugs for 48 hours performed using CellTiter-Glo Luminescent Cell Viability Assay prior to collection. Cells were harvested and stained using (Promega G7572) following the manufacturer's protocol and Annexin V-APC Detection kit from BioTool (#B32117) according

OF2 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

to the manufacturer's instructions. Flow data were acquired on a gene/canonical pathway pairs in M'. Each value, xi,j, in row i was z- xi;jmi 0 Gallios (Beckman Coulter) and analyzed using Summit v5.1 M xi,j j score normalized ( si ) to create r . Each value, , in row was (Beckman Coulter). xi;jmj M0 z-score normalized ( sj ) to create c. These two matrices were Immunofluorescence assay subsequently combined to create the TRAP matrix, MT, where the Methods for immunofluorescence assay were performed as z-scored value for the interaction between druggable gene, i, 5 j M0 zr described previously (12). WFU3 cells (1–6 10 ) and BHPrE1 and canonical pathway, , from r , i;j, was combined with the cells (3–6 105) were seeded on coverslips in a 12-well plate and z-scored value for the interaction between druggable gene, i, and 0 c incubated overnight. WFU3 cells were exposed to vehicle, ola- canonical pathway, j, from Mc, zi;j, using Stouffer's method (i.e. parib 5 mmol/L, dinaciclib 60 nmol/L, or the combination of these zr þzc i;jpffiffi i;j). We set the combined z-score threshold at 4 to arrive at a drugs for 12 hours for the RAD51 foci assay and 36 hours for the 2 gH2AX foci assay prior to harvest. BHPrE1 cells were exposed to final set of significant relationships that define the TRAP network vehicle, docetaxel 0.5 nmol/L, dinaciclib 4 nmol/L, or the com- (7,846 edges and 1,636 nodes: 691 pathways, 945 druggable bination of these drugs for 24 hours prior to harvest. Immuno- genes; Supplementary Table S1). The TRAP network established reaction was performed using the RAD51 antibody (Santa Cruz the background for which we used to identify druggable targets Technology, #sc-398587, 1:100) overnight at 4C or anti-phos- specific to patients with primary prostate tumor defined by the pho-Histone H2A.X (Ser139; Millipore, #JBW301. 1:500) for loss of CHD1 and MAP3K7. 1 hour at room temperature. Cells were washed three times with To prioritize druggable targets, we first identified patient tumor PBS before incubating with secondary antibody, Alex Fluor594, samples from TCGA (n ¼ 499) that had coloss (homozygous or goat anti-mouse IgG (HþL; Thermo Fisher Scientific, # A11005, heterozygous deletion) of CHD1 and MAP3K7 and patient 1:250 for RAD51, and 1:500 for gH2AX), for 1 hour at room tumors that were diploid at CHD1 and MAP3K7 (8). In addition, temperature. After washing three times, the coverslip was because the coloss of CHD1 and MAP3K7 is mutually exclusive of mounted onto a microscope slide using VECTASHIELD with ETS transcription factor fusions, we only considered tumors that DAPI (Vector Laboratories, #H-1200). Images were acquired with were negative for ETS fusions. In total, we identified 69 patients a Nikon DS-Ri2 confocal laser-scanning microscope with a 40 with loss of CHD1 and MAP3K7 (and were ETS fusion negative) objective lens. For quantitative analysis, software for automatic and 111 patients that were diploid for CHD1 and MAP3K7 (and focus counting was used to count gH2AX and RAD51 foci as were ETS fusion negative). Transcript counts were downloaded described previously (13). More than 100 cells from each group from Gene Expression Omnibus (accession number GSE62944; were chosen at random, and nuclei were counted to determine the ref. 16) and normalized using the calcNormFactors function in percent positive for RAD51 or gH2AX based on a threshold of 5 or the limma R package (17). Genes were ranked according to their greater foci per nucleus. log fold change between the two patient groups (Supplementary Table S2). We ran gene-set enrichment analysis (GSEA) in the TRAP analysis preranked mode on this ranked list of genes using the C2.CP gene TRAP is an algorithm that was first described by Kwong and sets (v6.0; Supplementary Table S3; refs. 18, 19). The significantly colleagues (7) to identify gene-to-pathway transcriptional regu- upregulated pathways (enriched in coloss patients compared with latory relationships using mouse data. Here, we developed a diploid; FDR < 0.05) were mapped to the TRAP network and the separate application of TRAP to evaluate human data. We first first neighbors were selected to define a subnetwork of druggable compiled a human gene expression compendium of 3,531 genes-to-differentially upregulated pathways (Supplementary human gene expression profiles on the Affymetrix HG133plus2.0 Table S4). Using this subnetwork and the full TRAP network, platform. Raw, human gene expression files (.CEL) were down- weighted degree centrality was calculated for both networks loaded from either Gene Expression Omnibus (GEO; https:// according to Opsahl and colleagues (20). To briefly summarize, www.ncbi.nlm.nih.gov/pubmed/23193258) or ArrayExpress the weighted average of the edges connecting each node (weight- (https://academic.oup.com/nar/article/47/D1/D711/5144130). ed degree centrality) in the subnetwork is calculated and then We preferentially selected experiments with multiple samples per compared with the weighted degree centrality in the full TRAP condition and sampled a diversity of treatment conditions, cell network. The difference from the subnetwork to the full network lines, and tissues. Arrays were batch normalized using RMAex- defines the gain in centrality, which is the measure used to press (14) with background adjustment, quantile normalization, prioritize gene targets (Supplementary Table S5). In addition to and median polish summarization. Probesets were mapped to comparing patients with coloss of CHD1 and MAP3K7 to patients HGNC gene symbols using Ensembl (v81) annotations from the diploid at those loci, we performed GSEA analysis on patients that BioMart online resource (https://www.ensembl.org/biomart). had loss of CHD1 alone (n ¼ 14) and loss of MAP3K7 alone (n ¼ Multiple probesets mapping to the same gene were averaged. 53), compared with patients diploid at both loci (n ¼ 111) This final gene expression compendium matrix, M, is the input to following the same procedure described previously. define the TRAP network. Druggable genes were defined using DGIdb (downloaded October 16, 2016; ref. 15); these genes were RNAseq in WFU3 6 selected from M to define M1. Canonical pathways were defined A total of 1.5 10 shControl, shChd1, shMap3k7, or shDou- using the C2.CP gene set collection from MSigDB (v6.0; ref. 15). ble WFU3 cells were plated onto 10-cm dishes in complete growth The average expression of genes in each gene set was calculated medium. Twenty-four hours later, RNA was isolated using 5prime fi using the M - M1 matrix. The new pathway expression matrix M2 Perfect RNA Cell kit (Thermo Fisher Scienti c) according to the was then concatenated with M1 to create M', where the columns manufacturer's protocol. After quality control, RNA samples were are samples and the rows are druggable genes or canonical path- sequenced by Novagene using the Illumina Novoseq 6000. PE150 ways. We calculated mutual information between all druggable reads were generated at an estimated 20 million reads per sample.

www.aacrjournals.org Mol Cancer Res; 2019 OF3

Washino et al.

Raw reads (fastq files) were mapped to the Ensembl v90 tran- selecting the directly connected neighbors, we created a subnet- scriptome of the Mus musculus genome build GRCm38 using work of druggable targets connected to the 65 pathways specificto Bowtie2 with the parameter settings "–sensitive –dpad 0 –gbar patients with loss of CHD1 and MAP3K7 (Fig. 1B; Supplementary 99999999 –mp 1,1 –np 1 –score-min L,0,-0.1 –no-mixed –no- Table S4). By identifying the most central genes in this subnet- discordant"; transcripts were quantified using the rsem-calculate- work, we were able to calculate a prioritization score to each gene expression function in RSEM (21, 22). A summary and details of target (Supplementary Table S5). The top 20 druggable gene sequence report are in Supplementary Tables S6–S8. Genes that targets identified by TRAP are shown in Fig. 1C, along with the had an average count of less than 5 across all replicates were fold changes in gene expression for the associated genes from removed. The log fold change was calculated for each gene using TCGA patient tumors (Supplementary Table S2) and the WFU3 the voom workflow in the limma R package (17). Using the same cells (Supplementary Table S8) that we describe next to evaluate C2.CP gene set collection (v6.0) as was used in the analysis of target hits. TCGA patient samples, we performed GSEA on the shDouble, shChd1, or shMap3k7 compared with WFU3 shControl cells. For In vitro validation of TRAP predictions this analysis, mouse gene symbols were mapped to human gene To model prostate cancer with loss of CHD1 and MAP3K7 symbols using the vertebrate homology mapping supplied by the expression, we used a previously characterized mouse prostate Mouse Genome Informatics resource (http://www.informatics. stem cell line (WFU3; ref. 9) with stably expressing shRNAs jax.org/downloads/reports/index.html). against Chd1 and Map3k7 (5). These cells have been extensively characterized in vitro and in vivo for their tumorigenic pheno- Statistical analyses types (5). We compared WFU3 cells with shRNAs targeting both Statistical analyses were performed with GraphPad Prism V7 Chd1 and Map3k7 (shDouble) to control cells with scrambled (GraphPad). Two-way ANOVA with Sidak multiple comparison shRNAs (shControl). The knockdown status of Map3k7 and Chd1 test, Tukey multiple comparison test, or Dunnett multiple com- is shown in Supplementary Fig. S1A. To assess whether similar parison test was used to analyze two different categorical inde- pathways are modulated in WFU3 cells with suppression of Chd1 pendent variables. IC50 values were determined using the four- and Map3k7 to those of patient samples, we generated RNAseq parameterlogisticmodel and comparison of curves was performed data from WFU3 shDouble and shControl cells (Supplementary using extra sum-of-square F test of logIC50 and hill slope. The Table S7 and S8) and ran GSEA to identify differentially expressed synergism between the two drugs was assessed using CalcuSyn V2 pathways (Supplementary Table S9). Of the 65 differentially (Biosoft). Two-tailed P < 0.05 was considered statistically signif- upregulated pathways identified from TCGA data, 63 were found icant. Data are shown by mean SE unless otherwise indicated. in the mouse results, and 56 of the 63 pathways were also upregulated in the shDouble WFU3 cells compared with WFU3 Data and materials availability: accession codes shControl cells. In total, 18 gene sets were significantly upregu- The RNA-Seq data are deposited in GEO with accession number lated (FDR < 0.05) comparing shDouble to shControl in the GSE132664. WFU3 cells with 12 overlapping the 63 pathways (P ¼ 1.7e–11; Fisher exact test) from the human data (Supplementary Tables S3 and S9). This comparison of human and mouse expression Results profiles supports the WFU3 cells as being a strong model of the Identification of therapeutic targets using TRAP human prostate tumor biology and thus a representative model to Because both CHD1 and MAP3K7 are genetically deleted in assess therapeutic targeting. prostate cancer, they cannot be directly targeted. Therefore, we We initially selected a CDK-1/-2 inhibitor for testing because used a computational systems pharmacology approach (TRAP) to CDK1 and CDK2 are central to the TRAP network and are also identify targetable pathways that are modulated in prostate cancer upregulated in primary patient tumors with coloss of CHD1 and with loss of CHD1 and MAP3K7. Figure 1A provides an overview MAP3K7, as well as in WFU3 shDouble cells (Fig. 1C). In addition, of TRAP (7) with details described in the Materials and Methods we screened several selective inhibitors for target genes among the section. The goal of TRAP is to prioritize druggable genes highly top TRAP hits (Fig. 1C). We also screened olaparib, a PARP connected to pathways that are dysregulated in a given experi- inhibitor, because this drug has been shown to directly affect the ment. A compendium of 3,531 diverse human gene expression responsiveness of prostate cancer cells with loss of CHD1 (6). As profiles was compiled and strength of connections between shown in Fig. 1D and Supplementary Table S10, CDK-1/-2 druggable genes and canonical pathways calculated using mutual inhibitor [dinaciclib (IC50 86.80 vs. 52.85 nmol/L in WFU3 information, where significant edges define the TRAP network shControl vs. shDouble, P < 0.0001)], Chk-1/-2 inhibitors (Supplementary Table S1). This network is the background for [MK8776 (IC50: 13.99 vs. 3.177 mmol/L, P < 0.0001)], and which we compare the following CHD1 and MAP3K7 expression AZD7762 (IC50: 3.629 vs. 0.5672 mmol/L, P ¼ 0.0010)], BIRC5 profile. Because CHD1 and MAP3K7 are mutually exclusive of the inhibitor [YM155 (IC50: 880.6 vs. 118.6 nmol/L, P < 0.0001)], ERG, ETV1, ETV4,orFLI1 transcription factor fusions (5), we docetaxel (IC50 18.84 vs. 4.627 nmol/L, P < 0.0001), TOP2A compared gene expression profiles from TCGA patient tumors inhibitors [doxorubicin (IC50: 193.1 vs. 44.43 nmol/L, P < with coloss of CHD1 and MAP3K7 and were ETS transcription 0.0001) and etoposide (IC50: 0.5596 vs. 0.3712 mmol/L, P ¼ factor fusion negative (n ¼ 69) to tumors diploid for CHD1 and 0.1329], and PARP inhibitor [olaparib (IC50: 31.53 vs. 15.01 MAP3K7 and were also ETS fusion negative (n ¼ 111; Supple- mmol/L, P ¼ 0.0002)] had increased potency on WFU3 shDouble mentary Table S2), then used GSEA to identify dysregulated compared with shControl. As negative controls, we assessed the pathways (Supplementary Table S3; refs. 18, 19). A total of 65 efficacy of cisplatin and temozolomide, both of which were not pathways were significantly differentially upregulated (FDR < predicted in the TRAP analysis and showed no effect on WFU3 0.05). We mapped these pathways to the TRAP network and by shDouble compared with shControl (Supplementary Table S10).

OF4 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

Figure 1. TRAP analysis and validation in WFU3 cells. A, The TRAP network construction workflow is shown. A human gene expression compendium of over 3,500 expression profiles was compiled. Druggable genes were defined using DGIdb and canonical pathways using the C2.CP gene sets from MSigDB. Mutual Information was calculated for all druggable genes/canonical pathway pairs; significant relationships define the edges in the TRAP network, which is the background model that defines how druggable genes are connected to pathways. See Materials and Methods section for more detail. B, TCGA patient tumors with coloss of CHD1 and MAP3K7 were used to identify significantly upregulated pathways. These pathways and their direct druggable gene neighbors were selected from the TRAP network and the most centrally located genes were identified. C, The top 20 druggable target genes are shown. Network centrality from the TRAP analysis, the log fold change between TCGA patients with coloss of CHD1 and MAP3K7 and those diploid at CHD1 and MAP3K7, and the log fold change between WFU3 shDouble and shCtrl. D, Dose–response curves of candidate drugs in cell growth of WFU3 shControl (black) versus shDouble (red). Cells were treated with each drug for 48 hours and cell growth was assessed by % confluence of each well in IncuCyte Zoom. Data are shown by mean and SE (n ¼ 3–5per group). These results are representative of three independent experiments.

www.aacrjournals.org Mol Cancer Res; 2019 OF5

Washino et al.

Loss of MAP3K7, but not CHD1, sensitized the effect of dependent kinase (CDK) inhibitor, and recent studies have dinaciclib through disruption of homologous recombination revealed a critical role for CDKs in regulating homologous We next asked whether loss of either Map3k7 or Chd1 individ- recombination (HR; ref. 23). We tested whether dinaciclib dis- ually was the main driver to increase the sensitivity to dinaciclib rupted HR in WFU3 cells. Dinaciclib decreased BRCA1 phosphor- in WFU3. Cells with knockdown of Chd1 alone did not show ylation and the expression of ataxia-telangiectasia mutated kinase any increase in sensitivity to dinaciclib relative to shControl (ATM), BRCA2, and RAD51 in shMap3k7 and shDouble cells cells (Fig. 2A). In contrast, cells with knockdown of Map3k7 were compared with shControl and shChd1 cells (Fig. 2B; Supplemen- more sensitive to dinaciclib, similar to that of shDouble cells tary Table S10A). The decrease in HR-related proteins was accom- (Fig. 2A). Importantly, knockdown of Chd1 did not alter/reverse panied by induction of gH2AX and p53 in shMap3k7 and the sensitizing effects of knockdown of Map3k7 (shDouble cells shDouble cells but not in shControl and shChd1 cells (Fig. 2B; respond similar to shMap3k7 cells). Dinaciclib is a multiple cyclin Supplementary Table S11A). These results are consistent with the

Figure 2. Loss of MAP3K7 sensitizes prostate cells to dinaciclib through disruption of homologous recombination in WFU3 cells. A, Dose–response curves of dinaciclib in cell growth of WFU3 shControl (black), shChd1 (purple), shMap3k7 (blue), and shDouble (red) are shown. Cells were treated with vehicle or dinaciclib (25–100 nmol/L) for 48 hours. Cell growth was assessed by percent confluence of each well and PI-positive cells were normalized to % confluence in IncuCyte Zoom. Data are shown by mean and SE (n ¼ 3 per group). The comparison of dose–response curve fit between the two cells was performed using extra sum-of-square F test. Two different categorical independent variables were analyzed using two-way ANOVA with Sidak multiple comparison test. , P < 0.05; , P < 0.01; , P < 0.001; #, P < 0.0001 vs. shCtrl. These results are representative of three independent experiments. B, Reduction of ATM, BRCA2, BRCA1 phosphorylation and RAD51, but not ATR, by dinaciclib accompanied by gH2AX and p53 induction in WFU3 cells. Cells were treated with vehicle or dinaciclib (60 or 120 nmol/L) for 24 hours. The results are representative (immunoblot), and mean and SE (bar graphs) of two independent experiments. C, Spearman correlation of normalized enrichment scores of the C2.CP gene sets from MSigDB for TCGA primary prostate tumor with loss of CHD1, MAP3K7, and both, and WFU3 shChd1, shMap3k7, and shDouble as determined by GSEA. The overlap of significant C2.CP gene sets from GSEA (FDR < 0.05) among the TCGA primary prostate tumor with loss of CHD1, MAP3K7, and both (D), and WFU3 shChd1, shMap3k7, and shDouble (E).

OF6 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

stronger effects of dinaciclib on cells with loss of Map3k7 (Fig. 2A). Supplementary Tables S2 and S8), further supporting the The time-dependent decrease of BRCA2 and RAD51 started by 3 WFU3 cells as a reliable model of human disease. We also to 6 hours after dinaciclib treatment in WFU3 shMap3k7 while found a higher frequency of overlaps of upregulated pathways p53 induction started by 3 hours and gH2AX induction was between TCGA coloss and MAP3K7 loss alone, and between observed at 24 hours (Supplementary Fig. S2A and S2B). The WFU3 shDouble and WFU3 shMap3k7 compared with that knockdown of Chd1 alone showed increased sensitivity to doxo- between TCGA coloss and CHD1 loss alone, and between rubicin, etoposide, and olaparib compared with control; howev- WFU3 shDouble and shChd1 (Fig. 2D and E). The expression er, the drugs' sensitizing effects were greater with suppression of patterns show that the shDouble and shMap3k7 are very Map3k7 compared with Chd1 suppression (Supplementary similar, while the shChd1 expression pattern is quite different. Fig. S3). In addition, the knockdown of Map3k7 sensitized WFU3 These data suggest that the expression patterns and drug sen- cells to docetaxel, but the knockdown of Chd1 did not have an sitivities are driven by loss of Map3k7. effect (Supplementary Fig. S3). We next considered whether disruption of HR by dinaciclib We found a strong similarity of drug sensitivity and dinaci- was observed in human prostate cells as well. To this end, clib-induced HR disruption in WFU3 shMap3k7 and shDouble, we used the BHPrE1 nontumorigenic human prostatic epithe- but differences comparing the WFU3 shChd1 cells to the lial cell line. BHPrE1 cells display epithelial intermediate/ shDouble and shMap3k7 cells. To further investigate these transit amplifying cell characteristics and were generated with- differences, we compared gene expression patterns across these out external DNA or viral modiﬁcation (10). The knockdown cells and also across TCGA primary prostate tumor with loss of of MAP3K7 was done using previously validated shRNA target- CHD1, MAP3K7, and the combination. We found a strong ing MAP3K7 (5). We compared the shMAP3K7 cells to cells correlation in altered pathways between TCGA coloss and stably expressing a nontargeting control shRNA (shControl; MAP3K7 loss alone (r ¼ 0.85) and between WFU3 shDouble Supplementary Fig. S1B). Loss of MAP3K7 by itself caused a and shMap3k7 (r ¼ 0.97), whereas altered pathways between slight increase in Annexin V and PI-positive cells in the WFU3 shChd1 and shDouble were quite different (Fig. 2C). We condition without drugs (Supplementary Fig. S1C). Suppres- found strong correlations between TCGA coloss with WFU3 sion of MAP3K7 augmented the antiproliferative effects and shDouble (r ¼ 0.49) and with shMap3k7 (r ¼ 0.49; Fig. 2C; potency of dinaciclib in BHPrE1 (IC50: 6.567 vs. 5.881 nmol/L

Figure 3. Loss of MAP3K7 sensitizes prostate cells to dinaciclib through disruption of homologous recombination in BHPrE1. A, Dose–response curves of dinaciclib in BHPrE1 shControl (black) versus shMAP3K7 (blue). Cells were treated with vehicle or dinaciclib (1.25–40 nmol/L) for 72 hours and cell growth was assessed using CellTiterGlo. Dinaciclib-induced cytotoxicity was assessed by PI assay in IncuCyte Zoom. Cells were treated with dinaciclib (1–16 nmol/L) for 48 hours. Data are shown by mean and SE (n ¼ 3 per group). The results are representative of three independent experiments. B, Dose–response reduction of HR-related proteins and induction of gH2AX in response to dinaciclib, which was more prominent, especially BRCA2 and ATR reduction, in BHPrE1 shMAK3K7 than shControl. Cells were treated with vehicle or dinaciclib (4–12 nmol/L) for 24 hours. The results are representative (immunoblot) and mean SE (bar graph, n ¼ 3 per group) of three independent experiments. Data were analyzed using two-way ANOVA with Dunnett or Sidak multiple comparison test. The comparison of dose–response curve ﬁt between the two cells was performed using extra sum-of-square F test (, P < 0.05; , P < 0.01; , P < 0.001; and #, P < 0.0001 vs. shCtrl).

www.aacrjournals.org Mol Cancer Res; 2019 OF7

Washino et al.

in BHPrE1 shControl vs. shMAP3K7, P < 0.05;Fig.3A).The dinaciclib, and the combination treatment in WFU3 shMap3k7, IC50 of dinaciclib in BHPrE1 shMAP3K7 (5.881 nmol/L) was respectively) while it decreased S-phase (Fig. 5B). The combina- much lower than that in WFU3 shMap3k7 (51.77 nmol/L). tion therapy induced a strong cytotoxic effect in WFU3 with loss of Suppression of MAP3K7 by itself increased the HR-related Map3k7, but not loss of Chd1 (Fig. 5C). The increase of Annexin V- proteins (Fig. 3B; Supplementary Table S11B; Supplementary and PI-positive cells by the combination of dinacilib and olaparib Fig. S4). Dinaciclib decreased phosphorylation of Ataxia Tel- in WFU3 with loss of Map3K7 was confirmed by flow cytometry angiectasia and Rad3-related protein (ATR) and BRCA1. Dina- analysis (% of Annexin V- and PI-positive cells: 14.50 0.9613 vs. ciclib also decreased expression of ATR, ATM, BRCA2, and 77.82 1.507% in WFU3 shControl vs. shMap3k7 in the com- RAD51, and this decrease in ATR, ATM, and BRCA2 was more bination therapy, P < 0.0001; Supplementary Fig. S6A and S6B). prominent in BHPrE1 shMAP3K7 compared with shControl. Dinaciclib did not decrease HR-related proteins in WFU3 These results were accompanied by an increased induction of shControl while the combination of olaparib and dinaciclib gH2AX (Fig. 3B; Supplementary Table S11B). Because p53 was slightly decreased HR-related proteins (Fig. 5D; Supplementary already stabilized in BHPrE1 (10), p53 did not increase in Table S14A). In contrast, in WFU3 shMap3k7, dinaciclib response to dinaciclib. The prominent decreases of HR-related decreased the expression of HR-related proteins and the combi- proteins, with induction of gH2AX in BHPrE1 shMAP3K7 nation therapy caused nearly complete losses of BRCA1 phos- compared with shControl were mostly consistent with WFU3 phorylation, BRCA2, and RAD51, accompanied by the strong cells, with some differences between BHPrE1 and WFU3 induction of gH2AX, p53, and caspase-3 cleavage (Fig. 5D; Sup- cells: for example, dinaciclib-induced RAD51 decrease was plementary Table S14A). The time-dependent decrease of BRCA2 similar in BHPrE1 shControl and shMAP3K7 in contrast to and RAD51 started by 6–12 hours and the increase of p53 and the prominent decrease of RAD51 in WFU3 with knockdown of gH2AX started by 6 and 24 hours, respectively (Fig. 5D; Supple- Map3k7 compared with WFU3 shControl. The decrease of ATR mentary Table S14B). The decrease in BRCA2 and RAD51 in expression by dinaciclib was observed in BHPrE1, but not in response to dinaciclib and olaparib was observed in all of cell- WFU3 cells. cycle phases in WFU3 shMap3k7 (Supplementary Fig. S7), dem- onstrating that the decrease in these proteins was not reflective of a Dinaciclib augmented the effect of DNA-damaging agents in decrease in S-phase cells. prostate cells with loss of MAP3K7 We next used immunofluorescence to detect the formation of In the absence of HR, DNA double-strand breaks persist RAD51 foci following olaparib treatment in all four types of ultimately resulting in cell death unless repaired by error- WFU3 (Fig. 5E). The formation of RAD51 foci in the nucleus prone nonhomologous end joining mechanisms (24–26). We was suppressed by dinaciclib with or without olaparib in WFU3 asked whether dinaciclib-induced disruption of HR sensitizes shMap3k7 and shDouble, while the decrease of RAD51 foci the cells to the effects of DNA-damaging agents including formation was not significant in WFU3 shControl and shChd1 inhibition of PARP. Dinaciclib augmented cytotoxic and anti- (relative decrease in RAD51 foci–positive cells in the combination proliferative effects of doxorubicin, etoposide, docetaxel, and therapy with olaparib: 0.458 and 0.510 in WFU3 shMap3k7 and olaparib in WFU3 shMap3k7, whereas this effect of dinaciclib shDouble vs. 0.700 and 0.896 in WFU3 shControl and shCHD1, on each DNA-damaging agent was less evident in WFU3 respectively). The combination therapy induced perinuclear shControl (Fig. 4A and B; Supplementary Fig. S5; Supplemen- RAD51 foci formation in WFU3 shMap3k7 and shDouble while tary Table S12A and S12B). In BHPrE1, dinaciclib increased the this was rarely observed in WFU3 shControl and shChd1 cytotoxic effects of doxorubicin, etoposide, docetaxel, and (Fig. 5E). Dinacilib also decreased the RAD51 foci formation in olaparib. Dinaciclib's augmentation of the cytotoxic effect of response to etoposide treatment in WFU3 shMap3k7 and shDou- docetaxel, doxorubicin, and olaparib, but not etoposide, was ble while perinuclear RAD51 formation was rarely observed in more prominent in BHPrE1 shMAP3K7 compared with BHPrE1 these combinations (Fig. 5E). gH2AX foci were increased 4- to shControl (Fig. 4C and D; Supplementary Table S12C and 5-fold in WFU3 shMap3k7 and shDouble in the vehicle condition S12D). The combination of dinaciclib with olaparib had less compared with WFU3 shControl or shChd1 (Fig. 5F). WFU3 cells cytotoxicity compared with dinaciclib single agent in BHPrE1 treated with the combination of dinaciclib and olaparib displayed (Fig. 4C and D; Supplementary Tables S12C and S12D), which greater induction of gH2AX compared with those treated with was inconsistent with the WFU3 results in which the combi- each single agent (Fig. 5F). gH2AX foci formation was signifi- nation therapy induced an obvious increase of cytotoxicity cantly more induced by these drugs in WFU3 shMap3k7 and (Fig. 4A and B; Supplementary Table S12A and S12B). shDouble, but not in WFU3 shChd1, compared with WFU3 shControl. Synergistic cytotoxic and antiproliferative effects between We next asked whether the combination of these two drugs dinaciclib and olaparib had similar effects on human prostate cancer cells. To this end, We focused on the combination of dinaciclib and olaparib we used LNCaP, androgen-sensitive human prostate adenocar- because there were obvious synergistic antiproliferative and cyto- cinoma cells. The combination of dinaciclib and olaparib toxic effects between the two drugs in WFU3 cells with loss of caused stronger cytotoxic and antiproliferative effects com- Map3k7 (Figs. 4A and B and 5A; Supplementary Fig. S5D; Sup- pared with each single agent in both LNCaP shControl and plementary Table S13A). In cell-cycle analysis, the combination of shDouble (Supplementary Fig. S8A). Dinaciclib decreased dinaciclib and olaparib increased G2–M in WFU3 shMap3k7 HR-related proteins with the induction of p53 and gH2AX in compared with WFU3 shControl (2.5- vs. 1.8-fold increase com- both LNCaP shControl and shDouble. When dinaciclib was pared with vehicle in WFU3 shMap3k7 vs. shControl) as well as combined with olaparib, it caused more induction of gH2AX each single agent (% of G2–M phase: 20.69 2.023, 24.00 and p53 accompanied by an increase of caspase-3 cleavage 2.103, 28.34 1.037, and 51.05 2.256% in vehicle, olaparib, (Supplementary Fig. S8B).

OF8 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

Figure 4. Dinaciclib augments effects of DNA-damaging agents and olaparib in prostate cells with loss of MAP3K7. Annexin V (A) and PI-positive cells (B) in response to each DNA-damaging agent or olaparib with or without dinaciclib in WFU3 shControl versus shMap3k7. Cells were treated with vehicle or each DNA-damaging agent, doxorubicin 50–400 nmol/L, etoposide 0.63–5.0 mmol/L, docetaxel 1.25–10 nmol/L, or olaparib 1.25–10.0 mmol/L, with or without dinaciclib 60 nmol/L for 48 hours. Annexin V and PI-positive cells were assessed using IncuCyte Zoom and normalized to % conﬂuence. Annexin V (C) and PI-positive cells (D) in BHPrE1 shControl versus shMAP3K7 in response to vehicle or each DNA-damaging agent, doxorubicin (Doxo) 50–200 nmol/L, etoposide (Etop) 1.0–4.0 mmol/L, docetaxel (Doce) 0.25–1 nmol/L or olaparib (Olap) 2.5–10.0 mmol/L with or without dinaciclib (Dina) 4 nmol/L are shown. Cells were treated with these drugs for 48 hours. Annexin V and PI-positive cells were assessed using IncuCyte Zoom and normalized to % conﬂuence. Data were analyzed using two-way ANOVA and F ratio of column factor (each DNA-damaging agent/ olaparib vs. dinaciclib þ each DNA-damaging agent/olaparib) and P values are also shown (Supplementary Table S12). Data are shown by mean and SE (n ¼ 3 per group). These results are representative of three independent experiments.

Combined cytotoxic and antiproliferative effects between compensation of DNA damage by docetaxel, while the increase dinaciclib and docetaxel in BHPrE1 of those proteins by docetaxel was not observed in BHPrE1 Loss of MAP3K7 augmented the antiproliferative and cytotoxic shMAP3K7 (Fig. 6D; Supplementary Table S15). Dinaciclib effects in response to docetaxel and the combination of dinaciclib slightly suppressed HR-related proteins in BHPrE1 shMAP3K7 and docetaxel in BHPrE1 (Fig. 6A and B). This combination while the combination of dinaciclib and docetaxel further sup- therapy induced an additive antiproliferative effect (Supplemen- pressed them (Fig. 6D; Supplementary Table S15). The decreases tary Table S13B) and slightly increased the proportion of cells in S- in ATR, ATM, and BRCA2 by the combination therapy were phase of the cell cycle in both BHPrE1 shControl and shMAP3K7 selectively observed in BHPrE1 shMAP3K7, and were accompa- (1.2- and 1.2-fold increase compared with vehicle in BHPrE1 nied by greater induction of gH2AX and caspase-3 cleavage shControl and shMAP3K7, respectively; Fig. 6C). Docetaxel (Fig. 6D). Similar to what we observed in WFU3 cells (Fig. 5F), increased HR-related proteins in BHPrE1 shControl, likely a gH2AX foci formation was 5-fold more prevalent in BHPrE1

www.aacrjournals.org Mol Cancer Res; 2019 OF9

Washino et al.

Figure 5. The combination therapy of dinaciclib (Dina) and olaparib (Olap) in WFU3. A, Antiproliferative effects in response to vehicle, dinaciclib (60 nmol/L), olaparib (10 mmol/L) or the combination of dinaciclib and olaparib in WFU3 shControl, shChd1, shMap3k7, and shDouble. The combination of dinaciclib and olaparib inhibited the cell growth of WFU3 with suppression of Map3k7 without regard to Chd1 status. Cells were treated with these drugs for 48 hours. , P < 0.0001 compared with WFU3 shControl by two-way ANOVA with Dunnett multiple comparison test. Data are shown by mean and SE (n ¼ 3 per group). The results are representative of three independent experiments. B, The results of cell-cycle analysis in response to vehicle, dinaciclib (60 nmol/L), olaparib (10 mmol/L), or the combination of dinaciclib and olaparib in WFU3 shControl and shMap3k7. Cells were treated with these drugs for 24 hours. Data are shown by mean and SE (n ¼ 3 per group). The results are representative of two independent experiments. C, Time course of Annexin V and PI-positive cells in response to vehicle (black), dinaciclib (60 nmol/L; blue), olaparib (10 mmol/L; green), or the combination (red) in WFU3 shControl, shMap3k7, shChd1, versus shDouble. Cells were treated with these drugs for up to 48 hours. Data are shown by mean and SE (n ¼ 3 per group). The results are representative of three independent experiments. D, HR-related proteins were selectively suppressed in WFU3 shMap3k7 in response to the combination of dinaciclib (60 nmol/L) and olaparib (10 mmol/L), which was accompanied by induction of gH2AX, p53, and cleaved caspase-3. Cells were treated with drugs for 36 hours. Time-dependent reduction of BRCA2 and RAD51 accompanied by increases in p53, gH2AX, and cleaved caspase-3 in response to the combination dinaciclib and olaparib are also shown. The results are representative of at least two independent experiments. E, RAD51 foci were selectively suppressed in WFU3 shMap3k7 and shDouble in response to the dinaciclib (60 nmol/L) with or without olaparib (10 mmol/L)/etoposide (Etop; 10 mmol/L). Cells were treated with drugs for 12 hours. Pictures are representative of merged images of RAD51 foci with DAPI. x, P < 0.1; , P < 0.05 by two-way ANOVA with Tukey multiple comparison test. Data are shown by mean and SE (n ¼ 3 per group). F, gH2AX foci were more observed in WFU3 shMap3k7 and shDouble in vehicle condition (small graph) as well as in response to the dinaciclib (60 nmol/L) and/or olaparib (10 mmol/L; large graph) compared with WFU3 shControl. Cells were treated with drugs for 36 hours. Pictures are representative of images of gH2AX foci, DAPI, and merge. , P < 0.05; , P < 0.01; #, P < 0.0001 versus shControl by two-way ANOVA with Dunnett multiple comparison test. Data are shown by mean and SE (n ¼ 3 per group).

OF10 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

Figure 6. The combination therapy of dinaciclib (Dina) and docetaxel (Doce) in BHPrE1. A, Cell growth assay in response to vehicle, dinaciclib (4 nmol/L), docetaxel (1 nmol/L), or the combination of dinaciclib and docetaxel (Combo) in BHPrE1 shControl versus shMAP3K7 are shown. Cells were treated with these drugs for 72 hours and cell growth was assessed using CellTiterGlo. , P < 0.01; , P < 0.0001 by two-way ANOVA with Sidak multiple comparison test. Data are shown by mean and SE (n ¼ 3 per group). These results are representative of three independent experiments. B, Time course of Annexin V and PI-positive cells in BHPrE1 shControl versus BHPrE1 shMAP3K7 in response to vehicle (black), dinaciclib (4 nmol/L; blue), docetaxel 1 nmol/L (green), or the combination of dinaciclib and docetaxel (red). Cells were treated with these drugs for up to 60 hours. Data are shown by mean and SE (n ¼ 3 per group). These results are representative of three independent experiments. C, Cell-cycle analyses in BHPrE1 shControl and shMAP3K7 in response to vehicle, dinaciclib (4 nmol/L), docetaxel (1 nmol/L), or the combination of dinaciclib and docetaxel. Cells were treated with these drugs for 24 hours. Data are shown by mean and SE (n ¼ 3 per group). These results are representative of two independent experiments. D, Cells were treated with vehicle, dinaciclib (4 nmol/L), docetaxel (1 nmol/L), or the combination of dinaciclib and docetaxel for 48 hours before harvesting. HR-related proteins were suppressed by the combination therapy compared with each single agent, and this was accompanied by induction of gH2AX and cleaved caspase-3. These results are representative (bands) and mean and SE (bar graphs) of three independent experiments. E, gH2AX foci increased in BHP shMAP3K7 in vehicle condition as well as in response to the dinaciclib (4 nmol/L) and docetaxel (1 nmol/L). Cells were treated with drugs for 24 hours. Pictures are representative of merged images of gH2AX foci with DAPI. The results are representative of two independent experiments (n ¼ 3 per each group). , P < 0.001 versus BHPrE1 shControl by two-way ANOVA with Dunnett multiple comparison test. Data are shown by mean and SE (n ¼ 3 per group).

www.aacrjournals.org Mol Cancer Res; 2019 OF11

Washino et al.

shMAP3K7 compared with BHPrE1 shControl in the vehicle of dinaciclib and olaparib (Figs. 2A and 4A and B; Supplementary condition (Fig. 6E) while it was also significantly more induced Fig. S3). Although we observed downregulation of CHD1 expres- in response to dinaciclib and docetaxel in BHPrE1 shMAP3K sion in response to dinaciclib independent of MAP3K7 status, the compared with BHPrE1 shControl. suppression of CHD1 did not affect the dinaciclib-induced disruption of HR and its sensitivity (Fig. 2A and B; Supplementary Table S11A). The mechanisms of downregulation of CHD1 by Discussion dinaciclib remain unknown. We applied an established computational systems pharmacol- MAP3K7 encodes a MAP kinase kinase kinase that is immedi- ogy approach (TRAP) to gene expression profile data in TCGA to ately downstream of cytokine and stress response pathways. identify potential signaling pathways altered in prostate cancer MAP3K7 is lost in 30%–40% of primary prostate tumors (2, 3). with loss of CHD1 and MAP3K7, and we prioritized druggable About half of prostate cancer samples with loss of MAP3K7 have target genes, including CDK1, CDK2, CHEK1, CHEK2, TOP2A, coloss of CHD1 (5). We demonstrated that suppression of and BIRC5 (Fig. 1C). We also validated a subset of the TRAP MAP3K7 sensitized cells to the effects of dinaciclib through the predictions using mouse prostate epithelial/stem cells (WFU3) augmented disruption of HR in both mouse and human prostate and demonstrated that CDK1/2 inhibitor, CHEK1/2 inhibitors, cells (Figs. 2A and B and 3A and B). Suppression of MAP3K7 also TOP2A inhibitors, and BIRC5 inhibitors caused a 1.5- to 7.4-fold affected the antitumor effects of DNA-damaging agents such as decrease in IC50 in WFU3 with cosuppression of Chd1 and doxorubicin, etoposide, docetaxel, and olaparib (Supplementary Map3k7 compared with WFU3 without suppression of these genes Fig. S3). The association of MAP3K7 with HR has not been (Fig. 1D and E). In contrast, several anticancer drugs not predicted previously demonstrated. We hypothesize that prostate cells with to be effective by TRAP analysis showed no effect. loss of MAP3K7 likely have defects in DNA repair pathways Dinaciclib inhibits multiple CDKs, including CDK-1, -2, -5, -9, leading to especially high dependence on HR activities while and -12 activity in in vitro kinase assays with IC50 values of 3, 1, 1, normal cells have intact DNA repair pathways and accumulate 4, and 58 nmol/L, respectively (27, 28). Dinaciclib has been less unrepaired DNA damage. In support of this hypothesis, mainly studied in chronic lymphocytic leukemia, breast cancer, suppression of MAP3K7 by itself in the absence of any drug and multiple myeloma in single agent or combination treat- treatment increased gH2AX foci formation in WFU3 as well as ments (12, 27, 29, 30), but, to the best of our knowledge, has BHPrE1 cells (Figs. 5F and 6E). In addition, prostate cells with not been studied in prostate cancer. Although CDKs are estab- suppression of MAP3K7 with or without suppression of CHD1 lished cell-cycle regulators, multiple lines of evidence suggest a increased HR-related gene expression, including RAD51, CHEK1, highly diverse and complex role for CDKs in the cellular response CHEK2, ATM, ATR, and/or BRCA2 in TCGA data, WFU3, and/or to DNA damage (23). CDKs play crucial roles in both the BHPrE1(Figs. 1C and 3B; Supplementary Fig. S4; Supplementary activation of DNA damage checkpoint signaling and HR- Table S2). Docetaxel increased the HR-related proteins in BHPrE1 dependent DNA repair. (31–34). CDK-1, -2, and -5 directly without loss of MAP3K7, whereas HR-related proteins were regulate ATR and ATM, both of which function as DNA damage already upregulated in BHPrE1 with loss of MAP3K7 with no checkpoints and mediate the initiation of HR (25). CDK1 also additional increase of these proteins upon docetaxel treatment regulates BRCA1 phosphorylation and subsequent recruitment to (Fig. 6D; Supplementary Table S15). Together, our results support sites of DNA damage, which is associated with RAD51 and BRCA2 the hypothesis that loss of MAP3K7 results in repair defects in recruitment, and HR DNA repair (23). Previous studies have prostate cancer. Although we previously reported single loss of shown that dinaciclib disrupts HR-dependent decreases in BRCA1 MAP3K7 is not an independent prognostic contributor (5), find- phosphorylation, RAD51 and BRCA2 (12, 27). We demonstrated ings in this study suggest that loss of MAP3K7, not loss of CHD1,is for the first time that dinaciclib decreased phosphorylation and the main contributor of response to the drugs we tested, which expression of ATR and expression of ATM, while also disrupting suggests MAP3K7 could be used as a predictive marker of treat- the HR DNA repair as shown by decreases in BRCA1 phosphor- ment response. The prevalence of prostate tumors with single loss ylation, BRCA2 and RAD51, as well as decreased RAD51 foci of MAP3K7 is twice as that with dual loss of MAP3K7 and formation in prostate cells (Fig. 2B, 3B, and 5E). Dinaciclib also CHD1 (5) and therefore these findings would open the thera- induced an accumulation of DNA damage as shown by induction peutic window for prostate cancer treatment. of gH2AX and p53 as a single agent (Figs. 2B, 3B, 5F, and 6E). Although we found that dinaciclib alone significantly Taken together, dinaciclib disrupts HR, most likely through CDK decreased cell viability in prostate cells with suppression of inhibition leading to the accumulation of DNA damage in pros- MAP3K7, a greater effect occurs when combined with DNA- tate cells. damaging agents such as doxorubicin, etoposide, and docetaxel, CHD1 is a member of the SNF2-like family of helicase-related or a PARP inhibitor (Fig. 4). Previous studies have also shown that enzymes, and part of a subgroup of CHD proteins that contain dinaciclib enhances the antitumor effects of the PARP inhibitors, tandem chromodomains (35). Recent studies have demonstrated doxorubicin and gemcitabine (12, 27). In WFU3 with suppres- that targeted disruption of CHD1 in human cells leads to a defect sion of Map3k7, the combination of dinaciclib and olaparib in early double-strand break repair via HR, resulting in hyper- strongly disrupted HR compared with each single agent, and this sensitivity to ionizing radiation as well as PARP and PTEN was associated with accumulation of DNA damage and the inhibition (6, 36, 37). Although, in this study, suppression of striking synergistic antiproliferative and cytotoxic effects Chd1 somewhat sensitized WFU3 cells to doxorubicin, etoposide, (Figs. 5A and C–F; Supplementary Tables S13A and S14). WFU3 and olaparib, the drug-sensitizing effect of Chd1 suppression was with suppression of Map3k7 displayed perinuclear RAD51 foci not as strong as that of Map3k7 suppression (Supplementary formation in response to the combination of dinaciclib and Fig. S3). Furthermore, suppression of Chd1 alone did not sensitize olaparib for 12 hours while this was rarely observed in each mouse prostate cells to dinaciclib, docetaxel, or the combination single agent or the combination of dinaciclib with etoposide

OF12 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Disrupts HR and Induces Cell Death

(Fig. 5E). Perinuclear RAD51 formation might reflect RAD51 foci dinaciclib and DNA-damaging agents causes an accumulation of formed with damaged DNA in the cytoplasm of dying cells. DNA damage that results in synergistic or additive antitumor However, in cell death assays, there were few Annexin V and effects on prostate cells with loss of MAP3K7 (Supplementary PI-positive cells at 12 hours of treatment (Fig. 5C), suggesting Fig. S9). another mechanism may be responsible for these observations. The exact mechanisms of perinuclear RAD51 formation by the Disclosure of Potential Conflicts of Interest combination of dinaciclib and olaparib remain unknown. The No potential conflicts of interest were disclosed. combination of dinaciclib and docetaxel strongly disrupted HR compared with each single agent, leading to accumulation of Disclaimer DNA damage and an additive antiproliferative effect in BHPrE1 The contents of this article are the authors' sole responsibility and do not with suppression of MAP3K7 (Fig. 6A and C–E; Supplementary necessarily represent official NIH views. Tables S13B and S15). Importantly, these effects in BHPrE1 were obtained with very low concentration of drugs (docetaxel Authors' Contributions 1 nmol/L and dinaciclib 4 nmol/L), which would open the Conception and design: J.C. Costello, S.D. Cramer therapeutic window for prostate cancer treatment. However, it Development of methodology: S. Washino, L.C. Rider, L. Romero, T. Affandi, MAP3K7 A.M. Ohm, J.C. Costello, S.D. Cramer has yet to be determined how is associated with DNA Acquisition of data (provided animals, acquired and managed patients, damage response including HR. provided facilities, etc.): S. Washino, L.C. Rider, L. Romero, L.K. Jillson, Dinaciclib decreased the expression and phosphorylation of M.E. Reyland, S.D. Cramer HR-related proteins to a similar degree in LNCaP with sup- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, pression of CHD1 and MAP3K7 and those without loss of these computational analysis): S. Washino, L.C. Rider, L. Romero, E.T. Lam, genes. This resulted in similar antiproliferative effects, yet still J.C. Costello, S.D. Cramer Writing, review, and/or revision of the manuscript: S. Washino, L.C. Rider, stronger cytotoxic effects on LNCaP with suppression of these L.K. Jillson, E.T. Lam, M.E. Reyland, J.C. Costello, S.D. Cramer genes (Supplementary Fig. S8A and S8B). Because LNCaP cells Administrative, technical, or material support (i.e., reporting or organizing already have hemizygous deletion of MAP3K7,andthuslower data, constructing databases): S. Washino, L. Romero, E.T. Lam, S.D. Cramer expression of MAP3K7 (2, 38, 39), additional suppression of Study supervision: J.C. Costello, S.D. Cramer MAP3K7 expression might have little effect on augmenting disruption of HR. Accordingly, increased cytotoxicity with Acknowledgments the combination of dinaciclib and olaparib was observed in This project was supported by grants from the NIH (R21-CA187354 and R01- CA199741; to S.D. Cramer), Cancer League of Colorado Grant (to S.D. Cramer LNCaP with or without knockdown of CHD1 and MAP3K7 and J.C. Costello), American Cancer Society (126572-PF-14-223-01-CCE; to L. (Supplementary Fig. S8A and S8B). C. Rider), Boettcher Foundation Webb-Waring grant (to J.C. Costello), and Paul Previous studies also have shown that MAP3K7 functions as a Calabresi Award for Clinical Oncology (5K12CA086913; to E.T. Lam). The critical mediator of the genotoxic stress response and is required University of Colorado Anschutz Medical Campus Cancer Center Cores used in for NF-kB, p38 MAPK, and JNK activation (40, 41). MAP3K7 this study were supported in part by the NIH/NCI Cancer Center Support Grant silencing results in an impaired NF-kB–dependent antiapoptotic P30CA046934. response and increased sensitivity to DNA damage. MAP3K7 was also demonstrated to be a prosurvival mediator in colon cancer The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked cells displaying hyperactive KRAS-dependent Wnt signaling (42). advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Therefore, loss of MAP3K7 could modulate these pathways and this fact. affect sensitivity to anticancer drugs in prostate cells. In conclusion, dinaciclib selectively disrupts HR in prostate Received December 17, 2018; revised May 31, 2019; accepted July 8, 2019; cells with decreased MAP3K7 expression. The combination of published first July 12, 2019.

Reference 1. Baca SC, Prandi D, Lawrence MS, Mosquera JM, Romanel A, Drier Y, 7. Kwong LN, Costello JC, Liu H, Jiang S, Helms TL, Langsdorf AE, et al. et al. Punctuated evolution of prostate cancer genomes. Cell 2013;153: Oncogenic NRAS signaling differentially regulates survival and prolifera- 666–77. tion in melanoma. Nat Med 2012;18:1503–10. 2. Liu W, Chang BL, Cramer S, Koty PP, Li T, Sun J, et al. Deletion of a small 8. Cancer Genome Atlas Research Network. The molecular taxonomy of consensus region at 6q15, including the MAP3K7 gene, is significantly primary prostate cancer. Cell 2015;163:1011–25. associated with high-grade prostate cancers. Clin Cancer Res 2007;13: 9. Barclay WW, Axanova LS, Chen W, Romero L, Maund SL, Soker S, et al. 5028–33. Characterization of adult prostatic progenitor/stem cells exhibiting 3. Wu M, Shi L, Cimic A, Romero L, Sui G, Lees CJ, et al. Suppression of Tak1 self-renewal and multilineage differentiation. Stem Cells 2008;26: promotes prostate tumorigenesis. Cancer Res 2012;72:2833–43. 600–10. 4. Liu W, Lindberg J, Sui G, Luo J, Egevad L, Li T, et al. Identification of novel 10. Jiang M, Strand DW, Fernandez S, He Y, Yi Y, Birbach A, et al. Functional CHD1-associated collaborative alterations of genomic structure and func- remodeling of benign human prostatic tissues in vivo by spontaneously tional assessment of CHD1 in prostate cancer. Oncogene 2012;31: immortalized progenitor and intermediate cells. Stem Cells 2010;28: 3939–48. 344–56. 5. Rodrigues LU, Rider L, Nieto C, Romero L, Karimpour-Fard A, Loda M, et al. 11. Krishan A.Rapid flow cytofluorometric analysis of mammalian cell cycle by Coordinate loss of MAP3K7 and CHD1 promotes aggressive prostate propidium iodide staining. J Cell Biol 1975;66:188–93. cancer. Cancer Res 2015;75:1021–34. 12. Alagpulinsa DA, Ayyadevara S, Yaccoby S, Shmookler Reis RJ. A cyclin- 6. Kari V, Mansour WY, Raul SK, Baumgart SJ, Mund A, Grade M, et al. Loss of dependent kinase inhibitor, dinaciclib, impairs homologous recombina- CHD1 causes DNA repair defects and enhances prostate cancer therapeutic tion and sensitizes multiple myeloma cells to PARP inhibition. Mol Cancer responsiveness. EMBO Rep 2016;17:1609–23. Ther 2016;15:241–50.

www.aacrjournals.org Mol Cancer Res; 2019 OF13

Washino et al.

13.IvashkevichAN,MartinOA,SmithAJ,RedonCE,BonnerWM, 28. Parry D, Guzi T, Shanahan F, Davis N, Prabhavalkar D, Wiswell D, et al. Martin RF, et al. gammaH2AX foci as a measure of DNA damage: Dinaciclib (SCH 727965), a novel and potent cyclin-dependent kinase a computational approach to automatic analysis. Mutat Res 2011; inhibitor. Mol Cancer Ther 2010;9:2344–53. 711:49–60. 29. Ghia P, Scarfo L, Perez S, Pathiraja K, Derosier M, Small K, et al. Efficacy and 14. Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normal- safety of dinaciclib vs ofatumumab in patients with relapsed/refractory ization methods for high density oligonucleotide array data based on chronic lymphocytic leukemia. Blood 2017;129:1876–8. variance and bias. Bioinformatics 2003;19:185–93. 30. Mita MM, Joy AA, Mita A, Sankhala K, Jou YM, Zhang D, et al. Randomized 15. Griffith M, Griffith OL, Coffman AC, Weible JV, McMichael JF, Spies NC, phase II trial of the cyclin-dependent kinase inhibitor dinaciclib (MK- et al. DGIdb: mining the druggable genome. Nat Methods 2013;10: 7965) versus capecitabine in patients with advanced breast cancer. 1209–10. Clin Breast Cancer 2014;14:169–76. 16. Rahman M, Jackson LK, Johnson WE, Li DY, Bild AH, Piccolo SR. 31. Ira G, Pellicioli A, Balijja A, Wang X, Fiorani S, Carotenuto W, et al. DNA Alternative preprocessing of RNA-Sequencing data in The Cancer end resection, homologous recombination and DNA damage checkpoint Genome Atlas leads to improved analysis results. Bioinformatics activation require CDK1. Nature 2004;431:1011–7. 2015;31:3666–72. 32. Jazayeri A, Falck J, Lukas C, Bartek J, Smith GC, Lukas J, et al. ATM- and cell 17. Law CW, Chen Y, Shi W, Smyth GK. voom: precision weights unlock linear cycle-dependent regulation of ATR in response to DNA double-strand model analysis tools for RNA-seq read counts. Genome Biol 2014;15:R29. breaks. Nat Cell Biol 2006;8:37–45. 18. Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, 33. Johnson N, Cai D, Kennedy RD, Pathania S, Arora M, Li YC, et al. Cdk1 et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation participates in BRCA1-dependent S phase checkpoint control in response are coordinately downregulated in human diabetes. Nat Genet 2003;34: to DNA damage. Mol Cell 2009;35:327–39. 267–73. 34. Scully R, Xie A.In my end is my beginning: control of end resection and 19. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, DSBR pathway `choice' by cyclin-dependent kinases. Oncogene 2005;24: et al. Gene set enrichment analysis: a knowledge-based approach for 2871–6. interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 35. Zhou J, Li J, Serafim RB, Ketchum S, Ferreira CG, Liu JC, et al. Human CHD1 2005;102:15545–50. is required for early DNA-damage signaling and is uniquely regulated by its 20. Opsahl T, Agneessens F, Skvoretz J. Node centrality in weighted networks: N terminus. Nucleic Acids Res 2018;46:3891–905. Generalizing degree and shortest paths. Social Networks 2010;32:245–51. 36. ShenoyTR,BoysenG,WangMY,XuQZ,GuoW,KohFM,etal.CHD1 21. Langmead B, Salzberg SL.Fast gapped-read alignment with Bowtie 2. loss sensitizes prostate cancer to DNA damaging therapy by promoting Nat Methods 2012;9:357–9. error-prone double-strand break repair. Ann Oncol 2017;28:1495–507. 22. Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq 37. Tian B, Yang Q, Mao Z. Phosphorylation of ATM by Cdk5 mediates DNA data with or without a reference genome. BMC Bioinformatics 2011;12: damage signalling and regulates neuronal death. Nat Cell Biol 2009;11: 323. 211–8. 23. Johnson N, Shapiro GI. Cyclin-dependent kinases (cdks) and the DNA 38. Klijn C, Durinck S, Stawiski EW, Haverty PM, Jiang Z, Liu H, et al. A damage response: rationale for cdk inhibitor-chemotherapy combinations comprehensive transcriptional portrait of human cancer cell lines. as an anticancer strategy for solid tumors. Expert Opin Ther Targets 2010; Nat Biotechnol 2015;33:306–12. 14:1199–212. 39. Kluth M, Hesse J, Heinl A, Krohn A, Steurer S, Sirma H, et al. Genomic 24. Bryant HE, Schultz N, Thomas HD, Parker KM, Flower D, Lopez E, et al. deletion of MAP3K7 at 6q12-22 is associated with early PSA recurrence in Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP- prostate cancer and absence of TMPRSS2:ERG fusions. Mod Pathol 2013; ribose) polymerase. Nature 2005;434:913–7. 26:975–83. 25. Helleday T.The underlying mechanism for the PARP and BRCA syn- 40. Wang C, Deng L, Hong M, Akkaraju GR, Inoue J, Chen ZJ. TAK1 is a thetic lethality: clearing up the misunderstandings. Mol Oncol 2011;5: ubiquitin-dependent kinase of MKK and IKK. Nature 2001;412: 387–93. 346–51. 26. Patel AG, Sarkaria JN, Kaufmann SH. Nonhomologous end joining 41. Yang Y, Xia F, Hermance N, Mabb A, Simonson S, Morrissey S, et al. A drives poly(ADP-ribose) polymerase (PARP) inhibitor lethality in cytosolic ATM/NEMO/RIP1 complex recruits TAK1 to mediate the NF- homologous recombination-deficient cells. Proc Natl Acad Sci U S A kappaB and p38 mitogen-activated protein kinase (MAPK)/MAPK- 2011;108:3406–11. activated protein 2 responses to DNA damage. Mol Cell Biol 2011;31: 27. Johnson SF, Cruz C, Greifenberg AK, Dust S, Stover DG, Chi D, et al. CDK12 2774–86. inhibition reverses de novo and acquired PARP inhibitor resistance in 42. Singh A, Sweeney MF, Yu M, Burger A, Greninger P, Benes C, et al. TAK1 BRCA wild-type and mutated models of triple-negative breast cancer. inhibition promotes apoptosis in KRAS-dependent colon cancers. Cell Cell Rep 2016;17:2367–81. 2012;148:639–50.

OF14 Mol Cancer Res; 2019 Molecular Cancer Research

Loss of MAP3K7 Sensitizes Prostate Cancer Cells to CDK1/2 Inhibition and DNA Damage by Disrupting Homologous Recombination

Satoshi Washino, Leah C. Rider, Lina Romero, et al.

Mol Cancer Res Published OnlineFirst July 12, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-18-1335

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2019/07/12/1541-7786.MCR-18-1335.DC1

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

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

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/early/2019/08/21/1541-7786.MCR-18-1335. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.