Genetically Defined Subsets of Human Pancreatic Cancer Show Unique in Vitro Chemosensitivity

Published OnlineFirst July 2, 2012; DOI: 10.1158/1078-0432.CCR-12-0827

Clinical Cancer Predictive Biomarkers and Personalized Medicine Research

Genetically Deﬁned Subsets of Human Pancreatic Cancer Show Unique In Vitro Chemosensitivity

Yunfeng Cui1, Jacqueline A. Brosnan1, Amanda L. Blackford2, Surojit Sur2,3, Ralph H. Hruban1,2, Kenneth W. Kinzler2,3, Bert Vogelstein2,3, Anirban Maitra1,2, Luis A. Diaz Jr2,3, Christine A. Iacobuzio-Donahue1,2, and James R. Eshleman1,2

Abstract Purpose: Pancreatic cancer is the fourth cause of death from cancer in the western world. Majority of patients present with advanced unresectable disease responding poorly to most chemotherapeutic agents. Chemotherapy for pancreatic cancer might be improved by adjusting it to individual genetic profiles. We attempt to identify genetic predictors of chemosensitivity to broad classes of anticancer drugs. Experimental Design: Using a panel of genetically defined human pancreatic cancer cell lines, we tested gemcitabine (antimetabolite), docetaxel (antimicrotubule), mitomycin C (MMC; alkylating), irinotecan (topoisomerase I inhibitor), cisplatin (crosslinking), KU0058948 (Parp1 inhibitor), triptolide (terpenoid drug), and artemisinin (control). Results: All pancreatic cancer cell lines were sensitive to triptolide and docetaxel. Most pancreatic cancer cells were also sensitive to gemcitabine and MMC. The vast majority of pancreatic cancer cell lines were insensitive to cisplatin, irinotecan, and a Parp1 inhibitor. However, individual cell lines were often sensitive to these compounds in unique ways. We found that DPC4/SMAD4 inactivation sensitized pancreatic cancer cells to cisplatin and irinotecan by 2- to 4-fold, but they were modestly less sensitive to gemcitabine. Pancreatic cancer cells were all sensitive to triptolide and 18% were sensitive to the Parp1 inhibitor. P16/ CDKN2A-inactivated pancreatic cancer cells were 3- to 4-fold less sensitive to gemcitabine and MMC. Conclusions: Chemosensitivity of pancreatic cancer cells correlated with some specific genetic profiles. These results support the hypothesis that genetic subsets of pancreatic cancer exist, and these genetic backgrounds may permit one to personalize the chemotherapy of pancreatic cancer in the future. Further work will need to confirm these responses and determine their magnitude in vivo. Clin Cancer Res; 18(23); 6519–30. 2012 AACR.

Introduction 5% because of its advanced stage at diagnosis, aggressive Pancreatic cancer is the fourth leading cause of cancer- growth, and resistance to most anticancer drugs (2). Despite related death in the United States and leads to an estimated this, systemic chemotherapy is generally used to treat 227,000 deaths per year worldwide (1). Five-year survival patients with pancreatic cancer. Gemcitabine is a nucleoside rates for pancreatic cancer are estimated to be less than analog and historically the standard first-line choice for the treatment of advanced pancreatic cancer. The median overall survival of single-agent gemcitabine is approximately 6 months with a response rate of 10% in patients Authors' Affiliations: 1Department of Pathology, Sol Goldman Pancreatic Cancer Research Center; 2Department of Oncology; and 3Ludwig Center with metastatic disease (3). Recently, the combination of for Cancer Genetics and Howard Hughes Medical Institute, Johns Hopkins biweekly administered fluorouracil, leucovorin, irinotecan, Medical Institutions, Baltimore, Maryland and oxaliplatin (FOLFIRINOX) was studied in a phase III Note: Supplementary data for this article are available at Clinical Cancer trial yielding a response rate of 31.6% and a median overall Research Online (http://clincancerres.aacrjournals.org/). survival of 11.1 months (4). De Jesus-Acosta and colleagues Current address for Y. Cui: Department of Surgery, Tianjin Nankai Hospital, also showed a substantial survival benefit with gemcitabine, Nankai Clinical School of Medicine, Tianjin Medical University, Tianjin docetaxel, and capecitabine (GTX) chemotherapy in pati- 300100, China. ents with metastatic and locally advanced pancreatic cancer, Corresponding Author: James R. Eshleman, The Sol Goldman Pancreatic and achieved a median survival of 11.3 and 25.0 months, Cancer Research Center, Department of Pathology and Oncology, The Johns Hopkins Medical Institutions, Suite 344, David Koch Cancer respectively (5). Research Building-II, 1550 Orleans Street, Baltimore 21231, MD. Phone: Recent studies in human pancreatic cancer have revealed 410-955-3511; Fax: 410-614-0671; E-mail: [email protected] complex genetic alterations in its pathogenesis and progres- doi: 10.1158/1078-0432.CCR-12-0827 sion. Jones and colleagues (6) sequenced 20,661 protein- 2012 American Association for Cancer Research. coding genes in 24 pancreatic cancers and found alterations

www.aacrjournals.org 6519

Cui et al.

bolite drug (gemcitabine), an antimicrotubule drug (doc- Translational Relevance etaxel), an alkylator (MMC; mitomycin C), a topoisomerase The costs of next generation sequencing are rapidly I (TopI) inhibitor (irinotecan), a cross-linking agent (cis- decreasing, and the US$ 1,000 genome is near. This "new platin), a Parp inhibitor (KU0058948), a terpenoid (trip- tool" may permit more rational and effective use of tolide), and a negative control drug (artemisinin). We chemotherapeutic drugs. We studied the chemosensitiv- tested this panel of drugs for cytotoxicity using a panel of ity of a broad panel of genetically defined pancreatic genetically defined human pancreatic cancer cells. We cancers to 7 chemotherapeutic drugs with distinct also verified some initial findings using DPC4/SMAD4 and mechanisms of cytotoxicity. We found 7 different TP53 isogenic pairs of cancer cells. The preliminary data gene–drug correlations in vitro. Additional work will be show that genetic subsets of pancreatic cancer correlate needed to confirm and determine the magnitude of these with in vitro chemosensitivity, and these genetic back- effects in vivo. These findings provide an initial frame- grounds may be useful for personalizing chemotherapy of work for personalized chemotherapy of this nearly lethal patients with pancreatic cancer in the future. malignancy.

Materials and Methods Human pancreatic cancer cell lines, human pancreatic in KRAS (>95%), p16/CDKN2A (95%), TP53 (50–75%), ductal epithelial cell line, human DPC4/SMAD4 and DPC4/SMAD4 (55%), the 4 high-frequency pan- isogenic pancreatic cancer cell lines, human TP53 creatic cancer driver genes (6–8). Recent studies reported isogenic colon cancer cell lines, and cell culture that inactivation of DPC4/SMAD4 was associated with Thirty-four human pancreatic cancer cell lines used poorer prognosis in patients with surgically resected pan- in this study were isolated from the tumor samples of creatic cancer and a widely metastatic phenotype in unre- patients in Departments of Pathology and Oncology, sected patients (9, 10). Chemoresistance of tumor cells Johns Hopkins School of Medicine (Baltimore, MD; can be enhanced by mutations in oncogenes, loss of tumor Supplementary Table S1). Among them, 19 cell lines were suppressors, or dysregulation of genes involved in DNA used for the Discovery Screen and 10 cell lines were used repair, cell cycle, cell proliferation, signal transduction, for the Prevalence Screen of the original pancreatic can- angiogenesis, or apoptosis (11, 12). cer sequencing study (6). The sequences of 23,219 tran- Using genetic alterations to predict chemotherapeutic scripts representing 20,661 protein-coding genes were response may prove to be one of the most important con- determined, and somatic mutations identified in the cepts in the therapy of human cancer. The field of targeted discovery and prevalence screen, homozygous deletions, therapy is rapidly expanding with more than 800 new and SAGE gene expression data have been published in agents developed, 140 of which are under clinical evalua- detail elsewhere (6). Twelve of the patients contributing tion. For example, Her2/neu is closely associated with samples to this study had a family history of the disease, aggressive clinical behavior and poor outcome (13), and defined as having at least 2 first-degree relatives affected trastuzumab was the first Her2-targeted therapy approved with pancreatic cancer. The use of human tissue was by the FDA in 1998 for the treatment of metastatic Her2- approved by the local Institutional Review Board (John mutated patients (14). Parp1 is a critical enzyme involved Hopkins University). All human pancreatic cancer cell in cell proliferation and repair of single-stranded DNA lines were cultured in minimum essential medium breaks and multiple PARP1 inhibitors have been tested (Gibco) supplemented with 10% FBS (Gibco), 100 U/mL clinically with encouraging results, particularly in triple penicillin, and 100 mg/mL streptomycin. negative BRCA1/2-mutated breast cancer (15). Extra- An HPV transfected normal human pancreatic ductal polation of this finding to other BRCA1/2 defective can- epithelial cell line (HPDE), generously provided by cers may not necessarily and directly follow (16). The Dr Ming-Sound Tsao, Toronto, Ontario, Canada. was also cytogenetic lesion of chronic myelogenous leukemia is analyzed. Apart from slightly aberrant expression of p53, the Philadelphia chromosome, which arises from a bal- molecular profiling of this cell line has shown that anced reciprocal translocation between the long arms of expression of other proto-oncogenes and tumor suppres- chromosomes 9 and 22 (17). This translocation produces sor genes are normal (23). The HPDE cell line was grown a constitutively active tyrosine kinase (BCR-ABL; ref. 18) in keratinocyte serum-free media (Invitrogen) supple- for which imatinib mesylate, a first-generation tyrosine mented with bovine pituitary extract and recombinant kinase inhibitor (TKI), is a potent inhibitor (19). However, EGF. although numerous studies have addressed chemosensiti- Pa01C and Pa02C parental cell lines were both generated vity testing in a variety of malignancies, yet there are few from liver metastases of patients with pancreatic cancer, studies in pancreatic cancer (20–22). both of which had homozygous deletions of DPC4. Pa01C We wished to test the hypothesis that the genetic muta- and Pa02C cells were stably transfected with either a cDNA tions in pancreatic cancer might predict chemotherapeutic expression plasmid to overexpress DPC4 (pcDNA3.1– response. For these initial studies, we selected a panel of DPC4) or an empty plasmid (pcDNA3.1Mock) using broadly acting anticancer agents including, an antimeta- the Attractene transfection agent (Qiagen). The Pa01C and

6520 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetic Subsets and Chemosensitivity of Pancreatic Cancer

Pa02C derivative isogenic cells were both analyzed as single mmol/L. Positive drug effects and isogenic cells drugs were clones. All of the genetically complemented clones express tested using final concentrations: 0 nmol/L, 1 nmol/L, 3 DPC4, whereas the control cells do not (data not shown). nmol/L, 10 nmol/L, 30 nmol/L, 100 nmol/L, 300 nmol/L, TGF-b pathway activities were tested for these isogenic pairs 1 mmol/L, 3 mmol/L, and 10 mmol/L in 3 independent using a previously reported luciferase assay (24). These cell experiments (18 replicate wells totally per dose). lines were cultured in Dulbecco’s modified Eagle medium Gene set enrichment analysis for gene expression data (Gibco) supplemented with 10% FBS (Gibco), 100 U/mL (Supplementary Materials and Methods). penicillin, 100 mg/mL streptomycin, and 0.35 mg/mL G418 sulfate (Mediatech). Supplementary Table S2 summarizes Statistical analysis the DPC4/SMAD4 isogenic cell lines. IC50 values were calculated by fitting sigmoid dose– The parental colorectal cancer cell lines HCT116, SW48, response curves with GraphPad Prism 5.0 (GraphPad Soft- and RKO were from American Type Culture Collection. All ware, Inc.). IC50 distributions were compared by genetic the isogenic cell pairs for TP53 were created with genetic status (mutations, deletions, or both) and family history knockout/knockin method, as was previously reported (25). status with Wilcoxon rank sum tests. A drug response was These celllines were culturedinMcCoy’s 5A medium (Gibco) defined as an IC50 value below the median or below the supplemented with 10% FBS (Gibco), 100 U/mL penicillin, highest IC50 that did not exceed 10,000 (this happened and 100 mg/mL streptomycin. Supplementary Table S3 for drugs where more than 50% of cell lines had IC50 shows the characteristics of TP53 isogenic cell lines. values of 10,000 or more). The relationship between All cells were maintained at 37C in a humidified atmo- response and genetic status was summarized with an OR sphere with 5% CO2. from a logistic regression model and Fisher exact test. When comparing the IC50 values between different drugs or to fold Preparation of anticancer drugs change in pathway activity, Spearman rank-based correla- Triptolide, docetaxel, MMC, cisplatin, irinotecan, and tion was used. IC50 values between isogenic pairs were artemisinin were purchased from Sigma. Gemcitabine was compared with a paired Student t test. Chemosensitivity obtained from Net Qem (Research Triangle Park). Parp1 correlation analysis of clustered cell lines was examined with inhibitor (KU0058948) was synthesized in Dr Vogelstein’s heatmaps and a network map, which is a visual depiction of lab. Artemisinin (an antimalarial drug) was selected as a how cell lines cluster based on whether the Spearman negative control drug because it possessed no known anti- correlation of the within-cell line standardized IC50 values cancer activity. The 8 drugs were dissolved in dimethyl between a pair of cell lines was 0.65 or higher. Statistical sulfoxide (DMSO; triptolide, docetaxel, cisplatin, irinote- analyses were conducted using GraphPad Prism 5.0 and R can, and artemisinin) or PBS (MMC and gemcitabine), (version 2.13.1). Because of the exploratory nature of the stock solutions of 10 mmol/L were made and they were study, P values are not corrected for multiple comparisons stored at 80C. and are included for descriptive purposes only.

In vitro cell growth assay for drug screening Cells were plated in 96-well plates at a density of 3,000 Results cells per well using standard culture media. Edge wells of Substantial variation of pancreatic cancer cell lines to the 96-well plate were filled with 1 PBS. After overnight broad classes of chemotherapeutic agents growth, media were removed and 200 mL of medium with We screened 29 genetically defined pancreatic cancer drug was added into each well. Drugs were serially diluted cell lines with the 7 broad classes of anticancer drugs listed to desired final concentrations with medium containing above. Pancreatic cancer cells were exposed to increasing 10% serum and each dose had 6 replicate wells. Cells concentrations of the drugs in 6 replicate wells and IC50 incubated in the medium with the vehicle dimethyl values calculated (Fig. 1 and Supplementary Tables S4 sulfoxide (DMSO) or PBS served as controls. Plates were and S5). Most pancreatic cancer cell lines were sensitive incubated with drugs for 72 hours. Media were aspirated to gemcitabine, triptolide, docetaxel, and MMC. The major- and each well was washed with PBS 3 times to remove ity of pancreatic cancer cell lines were insensitive to cis- cellular debris. One hundred microliters of ddH2Owas platin, irinotecan, Parp1 inhibitor, and the negative control addedtoeachwell,andplateswereincubatedatroom drug. Among the generally sensitive drugs, some cell lines temperaturefor1hourtolysethecells.Onehundred were notably less sensitive (Fig. 2A and D). For example, microliters of deionized H2O containing 0.15% of SYBR with gemcitabine, the vast majority of cell lines were green I solution (Molecular Probes; Catalog No. S7567) sensitive with IC50s in the single digit nanomolar range, wereaddedtowellsandmixed10times.Fluorescence whereas 3 cell lines (Pa09C, Pa08C, and Pa43C) were was measured by using BMG FluoStar Galaxy (BMG significantly less sensitive (IC50s: 40–60 nmol/L). Similarly, Labtechnologies; excitation at 480 nm and measurement for MMC, most of cell lines were sensitive with IC50s at 520 nm; ref. 25). under 50 nmol/L, whereas 5 cell lines (Pa18C, Pa07C, Human pancreatic cancer cell lines and HPDE were Pa02C, Pa113C, and Pa19C) were substantially resistant. tested initially at the following doses: 0 nmol/L, 1 nmol/L, Surprisingly, among the generally ineffective drugs (Fig. 1), 10 nmol/L, 100 nmol/L, 1 mmol/L, 10 mmol/L, and 100 individual cell lines were uniquely sensitive (Fig. 3). For

www.aacrjournals.org Clin Cancer Res; 18(23) December 1, 2012 6521

Cui et al.

SMAD4, TGFbR2,orTGFbR3) were 2.2-fold more sensitive to cisplatin (P ¼ 0.02; Supplementary Table S6). In logistic regression analysis, comparisons of DPC4 deleted or SMADPath gene deleted or SMADPath gene defective (mutated or deleted) pancreatic cancer cells with wild- type cells treated with cisplatin confirmed this finding (ORs: 11.93, 15.07, and 7.56; P: 0.01, 0.001, and 0.05, respectively; Supplementary Table S7). One concern about this study is that of multiple comparisons. Accordingly, to confirm these findings, we con- structed isogenic cell lines from Pa01C and Pa02C with or without homozygous DPC4 deletions. With cisplatin, we found that IC50 values of DPC4-deficient cell lines for cisplatin were 1.8- and 1.2-fold lower than for DPC4-containing cell lines (Fig. 5A and Supplementary Table S8). This was less significant in another isogenic cell line pair derived from Pa02C, but this cell line showed lower TGF- b signaling even when DPC4 was present (Supplementary Fig. S3A). Because all the pairs of isogenic cells had differ- Figure 1. Chemosensitivity of pancreatic cancer cell lines to broad classes ent levels of TGF-b pathway signaling (Supplementary of the IC values) of of chemotherapeutic agents. Drug responses (log10 50 Fig. S3A), we attempted to correlate the cisplatin IC50 values 34 pancreatic cancer cell lines tested with 7 chemotherapeutic drugs with TGF-b pathway activities in these cells. Our results representing different mechanisms of action and the control drug (artemisinin). Each cell line result is represented by a single circle. confirmed that IC50 values of cisplatin were positively cor- b r ¼ Horizontal lines indicate the median IC50 value for each drug for all cell related with TGF- activity ( 0.86, 95% CI: 0.39 to 0.97, lines. The 4 drugs on the left are generally sensitive drugs and the 4 drugs P ¼ 0.007), although this relationship is largely driven on the right are generally inactive. by the 2 cell lines with highest TGFbR activity (Supple- mentary Fig. S3B and Supplementary Table S9). The anal- example, with cisplatin, most cell lines were quite insensi- ysis of expression data showed that cytotoxicity of cisplatin tive, whereas a few (Pa18C, Pa21C, and Pa228C) had was closely positively associated with MAP3K12 (an acti- relatively lower IC50 values (100–200 nmol/L). Similarly, vator of the SAPK/JNK pathway), MAPK11 (most closely for irinotecan, 3 cell lines (Pa21C, Pa37C, and Pa227C) related to p38 MAP kinase), and MAP4K2 (involved in showed IC50 values less than 200 nmol/L, and for the Parp1 activation of MAP3K1/MEKK1; Supplementary Table S10). inhibitor, 6 cell lines (Pa03C, Pa01C, Pa11C, Pa104C, Pancreatic cancer cells with DPC4 defects (mutations or Pa221C, and Pa227C) had IC50s less than 200 nmol/L. In deletion combined) were 4.5-fold more sensitive to irino- summary, among the sensitive drugs, outlier pancreatic tecan than those with wild-type DPC4 (P ¼ 0.04; Fig. 4A and cancer cell lines were more resistant, whereas among the Supplementary Table S6). In logistic regression analysis, resistant drugs, unique cell lines were more sensitive. comparison of DPC4-defective cells to wild-type cells treated with irinotecan produced an OR of 8.64 and a P value of DPC4/SMAD4 inactivation sensitized pancreatic cancer 0.02, and when only DPC4-deleted cells were compared cells to cisplatin and irinotecan but reduced their with wild-type cells, a P value of 0.02 was obtained (Sup- sensitivity to gemcitabine plementary Table S7). To verify these findings, we tested the We attempted to explain the variation in drug response pairs of DPC4 isogenic cell lines with irinotecan. We found noted above by correlating it with the genetic status of IC50 values of DPC4-deficient cell lines were 1.2- and 1.7- these lines. We checked all the somatic mutations and fold lower than those of DPC4-complemented cell lines homozygous deletions which existed in the 19 pancreatic (Fig. 5B and Supplementary Table S8). Correlation analysis cancer cell lines included in the discovery screen and the for IC50 values and TGF-b pathway activities was not sig- 10 pancreatic cancer cell lines included in the prevalence nificant (P ¼ 0.067; Supplementary Table S9). screen (Supplementary Table S1; ref. 6). We attempted to We tested 2 pairs of DPC4 isogenic cells for possible correlate drug responses to genotype (Fig. 4 and Supple- synergy between cisplatin and irinotecan at a constant ratio mentary Fig. S1 and S2), considering each gene as either of 1:1 with simultaneous treatment. The combination indi- mutated only (mut), deleted only (del), mutated or deleted ces (CI) were calculated using CompuSyn (26). Simulta- combined (md), or wild-type (wt). We initially focused on neous treatment with the 2 drugs resulted in antagonistic DPC4/SMAD4 (Supplementary Tables S6 and S7). interactions with DPC4-defective cells and synergistic inter- Pancreatic cancer cells with DPC4 homozygous dele- actions in DPC4-complemented clones at ID50, but these tions were modestly (2-fold) more sensitive to cisplatin cell lines all showed antagonistic or additive interactions at compared with those with wild-type genotype (P ¼ 0.04; ID75 (Supplementary Table S11). Supplementary Fig. S1A and Table S6). Cells with the Despite the lack of correlation between DPC4 status and deletion of any of 4 SMAD pathway genes (SMAD3, gemcitabine response in the panel of pancreatic cancer cell

6522 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetic Subsets and Chemosensitivity of Pancreatic Cancer

Figure 2. Reponses of individual pancreatic cancer cell lines to the 4 sensitive drugs (note Y-axis scale, 0–120 nmol/L). Cytotoxic effects of gemcitabine (A), triptolide (B), docetaxel (C), and MMC (D) on pancreatic cancer cells. Gemcitabine (A) showed generally good responses, although unique cell lines showed more than 100-fold less sensitivities ("" indicates outlier cell lines). Triptolide (B) and docetaxel (C) showed nearly universal toxicities at doses in the single-digit nanomolar concentration. MMC (D) showed more variation in response, and unique cell lines () showed more than 40-fold less sensitivity.

lines (Supplementary Fig. S1C and S1E), we tested the DPC4 activities (P ¼ 0.015; Supplementary Fig. S3C and Sup- isogenic cell lines with gemcitabine and found that DPC4- plementary Table S11). defective cells were about 2-fold less sensitive than wild-type cells (P ¼ 0.058 and 0.033; Fig. 5C and Sup- Chemotherapeutic effects of triptolide and a Parp1 plementary Table S8). We made the correlation ana- inhibitor were related to TP53 inactivation lysis for IC50 values and TGF-b pathway activities in these Pancreatic cancer cells with TP53 defects were 2.3-fold isogenic cells. Our results conﬁrmed that gemcitabine more sensitive to triptolide than wild-type TP53 cells, P ¼ IC50s were negatively correlated with TGF-b pathway 0.04 (Fig. 4B and Supplementary Table S6). We conﬁrmed

www.aacrjournals.org Clin Cancer Res; 18(23) December 1, 2012 6523

Cui et al.

Figure 3. Reponses of individual pancreatic cancer cell lines to the 4 insensitive drugs (note Y-axis scale, 0–10,000 nmol/L). Cytotoxic effects of cisplatin (A), irinotecan (B), and Parp1 inhibitor (C) on pancreatic cancer cell viability showed variable drug responses, and unique cell lines showed more than 200-fold variation in their sensitivities to these drugs ("" indicates uniquely sensitive cell lines). Artemisinin (D) showed the nearly invariant resistance at doses in greater than 10,000 nanomolar, although unique cell lines () were sensitive.

these results in logistic regression analysis, P ¼ 0.02 (Sup- TP53 alleles, P ¼ 0.04 (Fig. 4C and Supplementary plementary Table S7), and conﬁrmed a modest effect Table S6). In this analysis, only 3 cell lines were wild-type, (1.5- to 2.7-fold) using 3 pairs of TP53 isogenic colon and the response of TP53-defective cell lines was clearly cancer cells (1.4- to 2.7-fold; P < 0.05; Fig. 5D and Supple- bimodal with one highly resistant population (IC50s: > mentary Table S12). 10,000 nmol/L), and a second variably sensitive popula- Pancreatic cancer cells with defective TP53 were 4.3-fold tion (IC50s: 20–4,000 nmol/L). Logistic regression analysis more sensitive to Parp1 inhibitor than those with wild-type conﬁrmed that a subset of TP53-defective cancer cells

6524 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetic Subsets and Chemosensitivity of Pancreatic Cancer

Figure 4. Correlations of drug response to genotype. Cytotoxic effect of irinotecan (A) was related to DPC4/SMAD4 inactivation, those of triptolide (B) and Parp1 inhibitor (C) were related to TP53 inactivation, and that of MMC (D) was related to P16 inactivation. Wilcoxon rank sum tests were used to compare differences in the median IC50 by mutation status. "md" indicates either mutation or deletion and the "wt" indicates wild-type. For TP53. md group in (C), solid circles indicate TP53-mutated pancreatic cancer cell lines, and triangles indicate TP53- deleted pancreatic cancer cell lines.

were more sensitive to the Parp1 inhibitor (an OR of 11.93 although this difference did not achieve statistical signi- for TP53 mutation vs. wild-type, P ¼ 0.02, and P value of ficance (P ¼ 0.15; Supplementary Fig. S2D). Pancreatic 0.01 for TP53 mutation or deletion vs. wild-type; Supple- cancer cells with wild-type P16 alleles were 3.8-fold more mentary Table S7). We investigated possible explanations sensitive to MMC than those with either P16 mutation or for this effect (Supplementary Tables S13 and S14), and deletion (P ¼ 0.04; Fig. 4D and Supplementary Table S5). identified a possible association between sensitivity and We confirmed this in logistic regression analysis (an OR wild-type P16 status, although this was not statistically of 0.19 for P16 defects vs. wild type, P ¼ 0.05; Supplemen- significant. We attempted to confirm these findings using tary Table S7). We attempted to explain the variation in 3 pairs of TP53 isogenic colon cancer cell lines (Supple- drug response of the P16-defective cancer cells, but were mentary Fig. S3D and Supplementary Table S12); however, unable to identify cooperating mutations in other genes the opposite effect was seen, possibly because the different (Supplementary Table S15). genetic and cellular backgrounds of the pancreatic cancer cells and the colon cancer cell lines. Other findings in the anticancer drug screening We selected the antimalarial drug, artemisinin, as a neg- P16/CDKN2A inactivation made human pancreatic ative control drug, and the majority (88.2%) of cancer cells cancer cells less sensitive to gemcitabine and MMC showed no response, as expected. To our surprise, 3 pan- Pancreatic cancer cells with P16 homozygous deletions creatic cancer cell lines showed sensitive responses with showed approximately 3.1-fold higher IC50 values for gem- IC50s less than 300 nmol/L (Fig. 3D). We examined the citabine than those with wild-type P16 (P ¼ 0.04; Supple- SAGE expression data in attempt to explain these results. mentary Fig. S2A and SupplementaryTable S6), although We noted significant the lack of expression of P16, p15, this was less significant when mutations and deletions were and a series of mitogen-activated protein kinase genes in considered together (P ¼ 0.11; Supplementary Fig. S2C). the sensitive lines (Supplementary Table S16). Because wild-type alleles for DPC4 and P16 were each Other drug response–genotype associations, drug associated with gemcitabine sensitivity, we also compared response–gene expression profile correlations and DPC4/ pancreatic cancer cells having both wild-type P16 and SMAD4, TP53,orP16/CDKN2A status gene expression pro- DPC4 with those having both P16 and DPC4 defects, file correlations are presented in Supplementary Results and found cells with wild-type DPC4 and P16 were 4-fold (Supplementary Table S6, S7, S10, S16–S21 and Fig. S4 more sensitive to gemcitabine than DPC4 and P16 defects, and S5). With regard to gemcitabine sensitivity, we were

www.aacrjournals.org Clin Cancer Res; 18(23) December 1, 2012 6525

Cui et al.

Figure 5. Reponses of isogenic cancer cell lines to anticancer drugs. Two pairs of isogenic pancreatic cancer cell lines for DPC4/SMAD4 were tested with cisplatin (A), irinotecan (B), and gemcitabine (C). Results for Pa01C isogenic cells are shown, as they show higher TGF-b functional complementation. Three pairs of isogenic colon cancer cell lines for TP53 showed differences in response for triptolide (D). Differences of IC50 value for isogenic pairs were compared with paired t test analysis and signiﬁcant P values indicated. IC50s were calculated for each experiment using 6 replicates per dose, and 3 separate experiments were used to calculate the mean IC50 for each drug and cell line combination (18 replicate wells totally per dose).

unable to correlate it with quasi-mesenchymal subtype (27) identify genes responsible for the clustering were not or ZEB1/2 expression (data not shown). successful.

Cluster analysis for anticancer drug sensitivities We also analyzed the chemosensitivity without consi- Discussion deration of the genetics using cluster analysis. Using an Personalized chemotherapy based on the pattern of evolutionary cluster analysis, it seems that the drug pat- genetic alterations and gene expression in a tumor can, in tern is dominant over the cell line clustering. As expected, theory, identify drugs to which the tumor will be sensitive as sensitive drugs clustered on the right branch, and insen- well as those unlikely to provide a therapeutic response. sitive drugs clustered on the left branch (Fig. 6A). We Comprehensive germline and somatic changes in patients found that cisplatin with irinotecan and gemcitabine and their cancers can now be determined at a reasonable with Parp1 inhibitors had similar drug response pro- cost in 2012. However, the bioinformatics to translate files. Some pancreatic cancer cell lines clustered together these data to chemoprediction is years behind. Our study forming several groups with similar drug response pro- attempts to fill this gap. In our study, we screened a set of files for 8 drugs (Fig. 6B). For example, Pa01C, Pa03C, genetically defined pancreatic cancers to determine and Pa21C in a triangle network showed similar drug whether there were correlations between genetic and tran- responses and similar different gene expression profiles scriptomic profiles and drug sensitivities. We found that that distinguished them from other pancreatic cancer somatic mutations or homozygous deletions of DPC4/ cell lines (Supplementary Table S22). Other attempts to SMAD4, TP53, and P16/CDKN2A genes correlated with

6526 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetic Subsets and Chemosensitivity of Pancreatic Cancer

Figure 6. Chemosensitivity correlation analysis of pancreatic cancer cell lines using evolutionary trees and cluster analysis. Cluster analysis was used to analyze chemosensitivity correlation of pancreatic cancer cell lines, and a threshold of 0.65 (Spearman correlation) indicates a positive correlation. A, clustered heatmap; B, network map for pancreatic cancer cell lines. Different drugs clustered because of similar drug response profiles for pancreatic cancer cell lines (A). In B, the dots represent individual cell lines. Cell lines are connected by a line to every other sample with which it was correlated by at least 0.65. Related cancer cells are showed by "boxes" (thin arrows) with "Xs" in the center for 4 cell lines and triangles between 3 cell lines (thick arrows). The line thickness indicates the strength of the correlation. Individual points mean that sample does not correlate with anyone else. in vitro drug responses, although the differential effects were CDKN1B and upregulation of CDCA4 and CDC2L6 with not pronounced. Additional studies will be necessary to DPC4 inactivation in our panel of pancreatic cancer cells determine whether these correlations can be reproduced in (Supplementary Table S19 and Supplementary Figs. S5A preclinical animal models and in patients. and S35B). In a recent study, Bornstein and colleagues We identified an association between DPC4/SMAD4 (32) reported increased genomic instability in both normal defects and sensitivity to cisplatin and irinotecan in the head and neck tissues and squamous cell carcinoma of initial panel of pancreatic cancer cell lines, and confirmed the head and neck (HNSCC) from HN-mice with selec- these effects using isogenic cell lines. The observation that tive inactivation of DPC4/SMAD4 in head and neck cisplatin cytotoxicity was correlated with alteration of epithelium, which correlated with downregulated expres- MAPKs indicates that the cytotoxic mechanism of cisplatin sion and function of genes encoding proteins in the Fanconi in pancreatic cancer cell may partly be linked to noncanon- anemia/BRCA DNA repair pathway. Under inactivation of ical TGF-b pathways. The activation of MAPK pathways by DPC4, cell-cycle checkpoint arrest is reduced, thereby TGF-b depends upon the cell type, often in a DPC4-inde- increasing genomic instability and resulting in apoptosis, pendent manner (28). For irinotecan, we observed that which may contribute to the sensitivity of DNA damaging DPC4 inactivation sensitized pancreatic cancer cells to this drugs in pancreatic cancer cells. In our study, we also drug. DNA TopI is essential in metazoans for the relaxation found that wild-type DPC4 was a sensitive factor for gem- of DNA supercoiling by forming Top1 cleavage complexes citabine. Pancreatic cancers are represented by distinct (Top1cc; ref. 29), and Top1 inhibitors can stabilize Top1cc genetic subtypes with significantly different patterns of and Top1-associated DNA damage (30). TGF-b signaling failure (9). So it may be valuable to determine DPC4 status influences cellular proliferation through inhibition of at initial diagnosis and stratify patients into personalized G1–S transition by inducing expression of cyclin kinase chemotherapy. inhibitors such as P15, P21, and P27, although TGF-b We found substantial variation in Parp1 inhibitor sensi- signaling represses c-Myc expression (31). Consistent with tivity among pancreatic cancer cell lines, where sensi- these prior studies, we observed downregulation of P15/ tivity correlated with inactivated TP53. All TP53 wild-type

www.aacrjournals.org Clin Cancer Res; 18(23) December 1, 2012 6527

Cui et al.

pancreatic cancer cells were highly resistant to the Parp1 was found that elevated expressions of HSF1 and HSPA9 inhibitor, although there was considerable variation among were related to the triptolide sensitivities in pancreatic TP53-inactivated cells. This variation was not correlated cancer cell lines (Supplementary Table S16). with the type of p53 mutation or mutation status at a second In addition, we showed that pancreatic cancer cells with gene (Supplementary Table S13). Nor could we explain P16 inactivation were less sensitive to gemcitabine and these results through differences in expression (Supplemen- MMC. These findings are consistent with those of Halloran, tary Table S14). TP53 downregulates homologous recom- who reported that gemcitabine combined with adenoviral- bination (HR) either directly (33) or by interacting and/or mediated reintroduction of P16 greatly enhanced cytotox- regulating HR-related proteins, such as Rad51 (34). In icity in Panc-1 pancreatic adenocarcinoma cells (49). P16/ this sense, TP53 inactivation results in increased exacerbat- CDKN2A is a tumor suppressor and plays a pivotal role ed spontaneous or damage-induced HR (35, 36). TP53- in regulating the G1–S cell-cycle checkpoint (50). Gemci- deficient mice exhibit high HR frequency at different tabine and MMC can both cause DNA damage by incor- developmental stages (37) and TP53 mutations correlate poration of nucleoside analogs or production of crosslinks with elevated recombination rates (38, 39). Moreover, and wild-type P16 may facilitate the process for cycle TP53 preferentially suppresses both spontaneous and checkpoint arrest and apoptosis in pancreatic cancer cells. DSB-induced nonconservative intrachromosomal HR Cell-cycle checkpoints are important in cellular response between imperfect homologous sequences (ICHR), suscep- to DNA damage by arresting the cell cycle to provide time tible of generating gross rearrangements such as deletions, for repair and by inducing transcription of genes that duplications, or inversions (40). In our study, it was observ- facilitate repair (51). This is consistent with a clinical trial ed that RAD51 and other HR genes were upregulated in that suggested that the ability of cells to progress into the TP53-inactivated pancreatic cancer cells, which indicated a S-phase after gemcitabine and radiation treatment was high HR frequency in TP53-deficient pancreatic cancer the key factor in radiosensitization (52). A recent study (Supplementary Table S20 and Fig. S5C). The correct bal- indicated that the polymorphic genotypes of p16 were ance between desirable and undesirable repair is attained associated with significantly shorter time to tumor pro- by the finely tuned HR regulation of TP53, as evidenced by gression and poorer response to therapy (53). Our analysis the fact that both excessive and defective HR are associated of expression data also showed that cytotoxicity of gemci- with cancer predisposition and resistance to therapeutic tabine was closely associated with the expression of genes drugs (41). Therefore, TP53 inactivation may result in regulating the cell cycle, apoptosis, autophagy, DNA repair, an imbalance between desirable and undesirable repair tumor suppressor gene, and anticancer drug metabolism because of the lack of proper regulation by TP53 in HR. pathway (Supplementary Table S16). So, we hypothesize Some studies report that cells lacking PARP1 also show a that mutations and deletions in cell-cycle genes such as hyper-recombinogenic phenotype (an increased frequency P16 may modulate the response to treatment with gemci- of HR; ref. 42–44), which has also been observed in mice tabine or MMC-based chemotherapy. using the pun assay (45). We propose that TP53 inactivation We also analyzed the chemosensitivity data by ignoring may have a synthetic lethality with Parp1 inhibitor in the genetics. For active drugs, we found uniquely insensitive pancreatic cancer cells because TP53 controls the fidelity lines, and among inactive drugs, including even the control of Rad51-dependent HR and represses aberrant processing drug, we found unique sensitive lines. By using cluster of replication forks after stalling at unrepaired DNA analysis, we found that some anticancer drugs have similar lesions. In our study, we also found that the sensitivity of drug response profiles for pancreatic cancer cell lines, and Parp1 inhibitor was related to expression profiles of some some pancreatic cancer cell lines clustered together forming DNA damage and repair pathway genes (ERCC3, RAD17, several groups with similar drug response profiles. Consis- SUMO1, MUTYH, CRY1, HSP90B1, CDC37, RXRA, USP5) tent with the findings of Collisson and colleagues (27), in (Supplementary Tables S10, S14, and S16). this provides evidence that functional subsets of pancreatic Triptolide covalently binds to human XPB (also known cancer exist, however, despite our best effort, we were as ERCC3), a subunit of the transcription factor TFIIH, unable to correlate clusters in Fig. 6B with specific genetic and inhibits its DNA-dependent ATPase activity, which defects. Drugs that closely cluster may indicate that they leads to the inhibition of RNA polymerase II–mediated target a common pathway or downstream effector. transcription and likely nucleotide excision repair (46). In In vitro studies, such as ours, have limitations. Several our study, we found that the expression level of ERCC3 was hypotheses have been suggested to explain why purified upregulated in triptolide-sensitive pancreatic cancer cells malignant cells may exhibit sensitivities in vitro that do not and triptolide was more sensitive to pancreatic cancer cells extrapolate to in vivo tumors. Malignant cells may possess with TP53 inactivation (Supplementary Table S16 and biologic properties that are lost in tissue culture. Alterna- Fig. S4A). This indicates that ERCC3 may exert a synthetic tively, the resistance to systemic therapies in patients with lethality with TP53. Recent reports also indicate that trip- pancreatic cancer might also be partly explained by ineffi- tolide causes pancreatic cancer cell death in vitro and in vivo cient drug delivery to malignant cells, because stromal by inducing apoptosis via inhibition of HSP70 that can microenvironment may create a hypovascular penetration inhibit apoptosis in cancer cells by attenuating cytosolic barrier that impairs effective chemotherapeutic delivery calcium and stabilizing lysosomes (47, 48). In our study, it (54). Furthermore, the degree of genetic heterogeneity of

6528 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetic Subsets and Chemosensitivity of Pancreatic Cancer

a tumor is also likely to be an important determinant of Analysis and interpretation of data (e.g., statistical analysis, biosta- tistics, computational analysis): Y. Cui, A. L. Blackford, S. Sur, R.H. therapeutic outcome (55), the reconstruction of tumor Hruban, K.W. Kinzler, A. Maitra, L.A. Diaz, C.A. Iacobuzio-Donahue, J.R. clonal architectures, and the identification of early driver Eshleman mutations may provide more robust biomarkers and ther- Writing, review, and/or revision of the manuscript: Y. Cui, J.A. Brosnan, A.L. Blackford, R.H. Hruban, B. Vogelstein, A. Maitra, L.A. Diaz, J.R. apeutic approaches (56). Eshleman In conclusion, we screened broad classes of anticancer Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): Y. Cui, J.A. Brosnan, C.A. Iacobuzio- drugs for their antiproliferative properties in a panel of Donahue, J.R. Eshleman human pancreatic cancer cell lines. These results suggest Study supervision: J.R. Eshleman that genetic subsets of pancreatic cancer exist, and that many are significantly correlated with chemosensitivity to specific Acknowledgments agents. These genetic subsets may be valuable for person- The authors thank Scott E. Kern, Christopher L. Wolfgang, and Leslie alizing pancreatic cancer chemotherapy in the future. Cope, for helpful discussion, and Ming-Sound Tsao for generously providing the HPDE cell line in this project. Disclosure of Potential Conflicts of Interest B. Vogelstein is a consultant/advisory board member in Inostics, Mor- Grant Support photek, Exact Sciences, and PGDx. L.A. Diaz has ownership interest (includ- These studies were supported by PanCan (to J.R. Eshleman), CA130938 ing patents) in Personal Genome Diagnostics and is a consultant/advisory (to J.R. Eshleman), CA140599 (to C.A. Iacobuzio-Donahue), Lustgarten (to board member in Inostics. J.R. Eshleman and R.H. Hruban disclose a patent B. Vogelstein), the Sol Goldman Pancreatic Cancer Research Center, and the licensed to Myriad Genetics No potential conflicts of interest were disclosed Michael Rolfe Foundation for Pancreatic Cancer Research (to A. Maitra, C.A. by other authors. Iacobuzio-Donahue, J.R. Eshleman). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked Authors' Contributions advertisement Conception and design: Y. Cui, A. Maitra, C.A. Iacobuzio-Donahue, J. R. in accordance with 18 U.S.C. Section 1734 solely to indicate Eshleman this fact. Development of methodology: Y. Cui, S. Sur, A. Maitra Acquisition of data (provided animals, acquired and managed patients, Received March 13, 2012; revised May 8, 2012; accepted June 16, 2012; provided facilities, etc.): Y. Cui, J.A. Brosnan, S. Sur, B. Vogelstein published OnlineFirst July 2, 2012.

References 1. Raimondi S, Maisonneuve P, Lowenfels AB. Epidemiology of pancre- 11. Lee W, Lockhart AC, Kim RB, Rothenberg ML. Cancer pharmacoge- atic cancer: an overview. Nat Rev Gastroenterol Hepatol 2009;6: nomics: powerful tools in cancer chemotherapy and drug develop- 699–708. ment. Oncologist 2005;10:104–11. 2. Shaib YH, Davila JA, El-Serag HB. The epidemiology of pancreatic 12. Kleeff J, Michalski C, Friess H, Buchler MW. Pancreatic cancer: from cancer in the United States: changes below the surface. Aliment bench to 5-year survival. Pancreas 2006;33:111–8. Pharmacol Ther 2006;24:87–94. 13. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. 3. Burris HA III, Moore MJ, Andersen J, Green MR, Rothenberg ML, Human breast cancer: correlation of relapse and survival with ampli- Modiano MR, et al. Improvements in survival and clinical bene- fication of the HER-2/neu oncogene. Science 1987;235:177–82. fit with gemcitabine as first-line therapy for patients with 14. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, advanced pancreas cancer: a randomized trial. J Clin Oncol 1997; et al. Use of chemotherapy plus a monoclonal antibody against HER2 15:2403–13. for metastatic breast cancer that overexpresses HER2. N Engl J Med 4. Conroy T, Desseigne F, Ychou M, Bouche O, Guimbaud R, Becouarn 2001;344:783–92. Y, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic 15. Fong PC, Boss DS, Yap TA, Tutt A, Wu P, Mergui-Roelvink M, et al. cancer. N Engl J Med 2011;364:1817–25. Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA muta- 5. De Jesus-Acosta A, Oliver GR, Blackford A, Kinsman K, Flores EI, tion carriers. N Engl J Med 2009;361:123–34. Wilfong LS, et al. A multicenter analysis of GTX chemotherapy in 16. Ledermann J, Harter P, Gourley C, Friedlander M, Vergote I, Rustin G, patients with locally advanced and metastatic pancreatic adenocar- et al. Olaparib maintenance therapy in platinum-sensitive relapsed cinoma. Cancer Chemother Pharmacol 2012;69:415–24. ovarian cancer. N Engl J Med 2012;366:1382–92. 6. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al. 17. Heisterkamp N, Stephenson JR, Groffen J, Hansen PF, de Klein A, Core signaling pathways in human pancreatic cancers revealed by Bartram CR, et al. Localization of the c-ab1 oncogene adjacent to a global genomic analyses. Science 2008;321:1801–6. translocation break point in chronic myelocytic leukaemia. Nature 7. Caldas C, Hahn SA, da Costa LT, Redston MS, Schutte M, Seymour 1983;306:239–42. AB, et al. Frequent somatic mutations and homozygous deletions of 18. Daley GQ, Van Etten RA, Baltimore D. Induction of chronic myelog- the p16 (MTS1) gene in pancreatic adenocarcinoma. Nat Genet enous leukemia in mice by the P210bcr/abl gene of the Philadelphia 1994;8:27–32. chromosome. Science 1990;247:824–30. 8. Iacobuzio-Donahue CA, Song J, Parmiagiani G, Yeo CJ, Hruban RH, 19. Buchdunger E, Zimmermann J, Mett H, Meyer T, Muller M, Druker BJ, Kern SE. Missense mutations of MADH4: characterization of the et al. Inhibition of the Abl protein-tyrosine kinase in vitro and in vivo by a mutational hot spot and functional consequences in human tumors. 2-phenylaminopyrimidine derivative. Cancer Res 1996;56:100–4. Clin Cancer Res 2004;10:1597–604. 20. Dan S, Tsunoda T, Kitahara O, Yanagawa R, Zembutsu H, Katagiri T, 9. Iacobuzio-Donahue CA, Fu B, Yachida S, Luo M, Abe H, Henderson et al. An integrated database of chemosensitivity to 55 anticancer CM, et al. DPC4 gene status of the primary carcinoma correlates with drugs and gene expression profiles of 39 human cancer cell lines. patterns of failure in patients with pancreatic cancer. J Clin Oncol Cancer Res 2002;62:1139–47. 2009;27:1806–13. 21. Zembutsu H, Ohnishi Y, Tsunoda T, Furukawa Y, Katagiri T, Ueyama Y, 10. Blackford A, Serrano OK, Wolfgang CL, Parmigiani G, Jones S, Zhang et al. Genome-wide cDNA microarray screening to correlate gene X, et al. SMAD4 gene mutations are associated with poor prognosis in expression profiles with sensitivity of 85 human cancer xenografts to pancreatic cancer. Clin Cancer Res 2009;15:4674–9. anticancer drugs. Cancer Res 2002;62:518–27.

www.aacrjournals.org Clin Cancer Res; 18(23) December 1, 2012 6529

Cui et al.

22. Staunton JE, Slonim DK, Coller HA, Tamayo P, Angelo MJ, Park J, et al. 39. Akyuz N, Boehden GS, Susse S, Rimek A, Preuss U, Scheidtmann KH, Chemosensitivity prediction by transcriptional profiling. Proc Natl et al. DNA substrate dependence of p53-mediated regulation of Acad Sci U S A 2001;98:10787–92. double-strand break repair. Mol Cell Biol 2002;22:6306–17. 23. Furukawa T, Duguid WP, Rosenberg L, Viallet J, Galloway DA, Tsao 40. Dudenhoffer C, Rohaly G, Will K, Deppert W, Wiesmuller L. Specific MS. Long-term culture and immortalization of epithelial cells from mismatch recognition in heteroduplex intermediates by p53 suggests normal adult human pancreatic ducts transfected by the E6E7 gene aroleinfidelity control of homologous recombination. Mol Cell Biol of human papilloma virus 16. Am J Pathol 1996;148:1763–70. 1998;18:5332–42. 24. Sohn TA, Su GH, Ryu B, Yeo CJ, Kern SE. High-throughput drug 41. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human screening of the DPC4 tumor-suppressor pathway in human pancre- cancers. Nature 1998;396:643–9. atic cancer cells. Ann Surg 2001;233:696–703. 42. Godon C, Cordelieres FP, Biard D, Giocanti N, Megnin-Chanet F, Hall 25. Sur S, Pagliarini R, Bunz F, Rago C, Diaz LA Jr, Kinzler KW, et al. A panel J, et al. PARP inhibition versus PARP-1 silencing: different outcomes in of isogenic human cancer cells suggests a therapeutic approach for terms of single-strand break repair and radiation susceptibility. Nucleic cancers with inactivated p53. Proc Natl Acad Sci U S A 2009;106: Acids Res 2008;36:4454–64. 3964–9. 43. Schultz N, Lopez E, Saleh-Gohari N, Helleday T. Poly(ADP-ribose) 26. Chou TC. Theoretical basis, experimental design, and computerized polymerase (PARP-1) has a controlling role in homologous recombi- simulation of synergism and antagonism in drug combination studies. nation. Nucleic Acids Res 2003;31:4959–64. Pharmacol Rev 2006;58:621–81. 44. Waldman AS, Waldman BC. Stimulation of intrachromosomal homol- 27. Collisson EA, Sadanandam A, Olson P, Gibb WJ, Truitt M, Gu S, et al. ogous recombination in mammalian cells by an inhibitor of poly(ADP- Subtypes of pancreatic ductal adenocarcinoma and their differing ribosylation). Nucleic Acids Res 1991;19:5943–7. responses to therapy. Nat Med 2011;17:500–3. 45. Claybon A, Karia B, Bruce C, Bishop AJ. PARP1 suppresses homol- 28. de Caestecker MP, Piek E, Roberts AB. Role of transforming growth ogous recombination events in mice in vivo. Nucleic Acids Res 2010; factor-beta signaling in cancer. J Natl Cancer Inst 2000;92:1388–402. 38:7538–45. 29. Pommier Y. DNA topoisomerase I inhibitors: chemistry, biology, and 46. Titov DV, Gilman B, He QL, Bhat S, Low WK, Dang Y, et al. XPB, a interfacial inhibition. Chem Rev 2009;109:2894–902. subunit of TFIIH, is a target of the natural product triptolide. Nat Chem 30. Pommier Y, Cushman M. The indenoisoquinoline noncamptothecin Biol 2011;7:182–8. topoisomerase I inhibitors: update and perspectives. Mol Cancer Ther 47. Dudeja V, Mujumdar N, Phillips P, Chugh R, Borja-Cacho D, Dawra RK, 2009;8:1008–14. et al. Heat shock protein 70 inhibits apoptosis in cancer cells through 31. Peng B, Fleming JB, Breslin T, Grau AM, Fojioka S, Abbruzzese JL, simultaneous and independent mechanisms. Gastroenterology et al. Suppression of tumorigenesis and induction of p15(ink4b) by 2009;136:1772–82. Smad4/DPC4 in human pancreatic cancer cells. Clin Cancer Res 48. Phillips PA, Dudeja V, McCarroll JA, Borja-Cacho D, Dawra RK, Grizzle 2002;8:3628–38. WE, et al. Triptolide induces pancreatic cancer cell death via inhibition 32. Bornstein S, White R, Malkoski S, Oka M, Han G, Cleaver T, et al. of heat shock protein 70. Cancer Res 2007;67:9407–16. Smad4 loss in mice causes spontaneous head and neck cancer with 49. Halloran CM, Ghaneh P, Shore S, Greenhalf W, Zumstein L, Wilson increased genomic instability and inflammation. J Clin Invest 2009; D, et al. 5-Fluorouracil or gemcitabine combined with adenoviral- 119:3408–19. mediated reintroduction of p16INK4A greatly enhanced cytotoxicity 33. Janz C, Susse S, Wiesmuller L. p53 and recombination intermediates: in Panc-1 pancreatic adenocarcinoma cells. J Gene Med 2004;6: role of tetramerization at DNA junctions in complex formation and 514–25. exonucleolytic degradation. Oncogene 2002;21:2130–40. 50. Macaluso M, Montanari M, Cinti C, Giordano A. Modulation of cell 34. Linke SP, Sengupta S, Khabie N, Jeffries BA, Buchhop S, Miska S, et al. cycle components by epigenetic and genetic events. Semin Oncol p53 interacts with hRAD51 and hRAD54, and directly modulates 2005;32:452–7. homologous recombination. Cancer Res 2003;63:2596–605. 51. Elledge SJ. Cell cycle checkpoints: preventing an identity crisis. 35. Mekeel KL, Tang W, Kachnic LA, Luo CM, DeFrank JS, Powell SN. Science 1996;274:1664–72. Inactivation of p53 results in high rates of homologous recombination. 52. Ostruszka LJ, Shewach DS. The role of cell cycle progression in Oncogene 1997;14:1847–57. radiosensitization by 20,20-difluoro-20-deoxycytidine. Cancer Res 36. Saintigny Y, Rouillard D, Chaput B, Soussi T, Lopez BS. Mutant p53 2000;60:6080–8. proteins stimulate spontaneous and radiation-induced intrachromo- 53. Chen J, Li D, Killary AM, Sen S, Amos CI, Evans DB, et al. Polymorph- somal homologous recombination independently of the alteration of isms of p16, p27, p73, and MDM2 modulate response and survival of the transactivation activity and of the G1 checkpoint. Oncogene pancreatic cancer patients treated with preoperative chemoradiation. 1999;18:3553–63. Ann Surg Oncol 2009;16:431–9. 37. Bishop AJ, Hollander MC, Kosaras B, Sidman RL, Fornace AJ Jr, 54. Neesse A, Michl P, Frese KK, Feig C, Cook N, Jacobetz MA, et al. Schiestl RH. Atm-, p53-, and Gadd45a-deficient mice show an Stromal biology and therapy in pancreatic cancer. Gut 2011;60: increased frequency of homologous recombination at different stages 861–8. during development. Cancer Res 2003;63:5335–43. 55. Marusyk A, Almendro V, Polyak K. Intra-tumour heterogeneity: a 38. Bertrand P, Rouillard D, Boulet A, Levalois C, Soussi T, Lopez BS. looking glass for cancer? Nat Rev Cancer 2012;12:323–4. Increase of spontaneous intrachromosomal homologous recombina- 56. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos tion in mammalian cells expressing a mutant p53 protein. Oncogene E, et al. Intratumor heterogeneity and branched evolution revealed by 1997;14:1117–22. multiregion sequencing. N Engl J Med 2012;366:883–92.

6530 Clin Cancer Res; 18(23) December 1, 2012 Clinical Cancer Research

Genetically Defined Subsets of Human Pancreatic Cancer Show Unique In Vitro Chemosensitivity

Yunfeng Cui, Jacqueline A. Brosnan, Amanda L. Blackford, et al.

Clin Cancer Res 2012;18:6519-6530. Published OnlineFirst July 2, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-12-0827

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2012/07/02/1078-0432.CCR-12-0827.DC1

Cited articles This article cites 56 articles, 24 of which you can access for free at: http://clincancerres.aacrjournals.org/content/18/23/6519.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/18/23/6519.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://clincancerres.aacrjournals.org/content/18/23/6519. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.