Pharmacogenomic analysis of patient-derived tumor cells in gynecologic cancers
Gynecologic malignancy is one of the leading causes of mortality in female adults worldwide. Comprehensive genomic analysis has revealed a list of molecular aberrations that are essential to tumorigenesis, progression, and metastasis of gynecologic tumors. However, targeting such alterations has frequently led to treatment failures due to underlying genomic complexity and simultaneous activation of various tumor cell survival pathway molecules. A compilation of molecular characterization of tumors with pharmacological drug response is the next step toward clinical application of patient-tailored treatment regimens.
Toward this goal, we establish a library of 139 gynecologic tumors including epithelial ovarian cancers (EOCs), cervical, endometrial tumors, and uterine sarcomas that are genomically and/or pharmacologically annotated and explore dynamic pharmacogenomic associations against 37 molecularly targeted drugs. We discover lineage-specific drug sensitivities based on subcategorization of gynecologic tumors and identify TP53 mutation as a molecular determinant that elicits therapeutic response to poly (ADP-Ribose) polymerase (PARP) inhibitor. We further identify transcriptome expression of inhibitor of DNA biding 2 (ID2) as a potential predictive biomarker for treatment response to olaparib.
Together, our results demonstrate the potential utility of rapid drug screening combined with genomic profiling for precision treatment of gynecologic cancers.
KeywordsGynecologic malignancy Pharmacogenomic analysis PARP inhibitor TP53 mutations ID2
A fundamental principle of precision oncology is that molecular profiling of the tumor enables identification of appropriate therapeutic choice for individual patients [1, 2, 3, 4, 5, 6, 7, 8]. However, predicting successful therapies on the sole basis of computational approach remains challenging [9, 10, 11]. Large-scale pharmacogenomic analyses using conventional cancer cell-line models have shown significant conceptual advances in discovering alternative therapeutic options for subsets of cancer patients [12, 13, 14, 15, 16, 17, 18]. However, molecular and pharmacological discrepancies between patient tumors and long-term cultured cancer cell-lines discourage clinical application of current gene-drug atlas. We have previously established a pharmacogenomic landscape of patient-derived tumor cell (PDC) models to reveal unprecedented insights into dynamic gene-drug associations and demonstrated its clinical feasibility . To further interrogate the dynamics of pharmacogenomic interactions at single tumor-lineage resolution, we generated a collection of gynecologic tumors, including cervical, endometrial/uterine, and epithelial ovarian cancers (EOCs), and explored potential gene-drug associations against 37 molecularly targeted agents.
Currently, there are over 100,000 newly diagnosed cases and approximately 32,000 mortalities from gynecologic cancers in the US. Gynecologic tumors can be categorized into 5 distinct subgroups: ovarian, endometrial/uterine, cervical, vulvar, and vaginal tumors based on geographical locations. The current standard treatment consists of aggressive surgical intervention followed by platinum–taxane chemotherapy. Despite such intensive treatment modalities, approximately 25% of the patients invariably undergo tumor relapse within 6 months from the initial treatment and there is no alternative therapeutic avenue that is readily available. Although large-scale genomic characterizations of ovarian, uterine, and cervical cancers have been profiled by The Cancer Genome Atlas (TCGA) Research Network [20, 21, 22, 23], clinical application potential of molecular targeted therapy remains obscure. Toward this goal, we have established a library of short-term cultured PDC models and performed comprehensive analyses of pharmacogenomic interactions to identify potential molecular determinants that could guide personalized treatment in gynecologic tumors.
Establishment of patient-derived gynecologic tumor cell library
Tumor cell isolates were cultured under serum-free conditions for generally 2 to 4 weeks with epidermal growth factor (EGF) and basic fibroblast growth factor (FGF) supplements for enrichment of tumor initiating cell (TIC) populations (Additional file 2: Table S1) [29, 30]. Afterwards, PDCs were subjected to systematic drug sensitivity screening against 37 anti-cancer agents, targeting major oncogenic pathways including receptor tyrosine kinase (RTK), histone deacetylase (HDAC), and poly (ADP-ribose) polymerase (Additional file 3: Table S2 and Additional file 1: Figure S3) . Drug sensitivities were determined using the area under curve (AUC) of the dose response curve (Fig. 1a) [19, 31, 32, 33]. A number of PDCs were further subjected to targeted exome sequencing and/or WTS to investigate whether the major gynecologic cancer-driver genes were retained from the parental tumors to PDCs. Consistent with previous observation, major drug-target genetic aberrations, including TP53, ERBB3, EGFR, and BRAF, were highly conserved in PDCs (Fig. 1c, Additional file 1: Figure S4). Moreover, transcriptome analysis of the parental tumors with matched PDCs demonstrated a strong positive correlation (Fig. 1c). To assess tumorigenicity of PDCs in vivo, we established patient-derived xenograft (PDX) models and evaluated their histological features . Notably, PDX models recapitulated the original morphologic and pathologic characteristics of their parental tumors in situ (Additional file 1: Figure S5). Collectively, these results suggest that our gynecologic PDCs manifest molecular characteristics of the parental tumors and can be employed as surrogates for comprehensive pharmacogenomic analyses.
Therapeutic landscape of gynecologic cancers reveals tumor type-specific drug sensitivity
Next, we established a pharmacological landscape of 66 PDCs that were derived from cervical (n = 6), endometrial (n = 10), uterine (n = 8, including leiomyoma), and EOCs (n = 42) using 37 molecularly targeted drugs. A total of 2442 drug-PDC combinations were analyzed and plotted (Additional file 4: Table S3 and Additional file 1: Figure S6). Distribution of drug sensitivities varied widely, portraying the heterogeneous nature of gynecologic PDCs. A subset of drugs, including afatinib, dacomitinib, neratinib (EGFR), AZD2014 (mTOR), panobinostat (HDAC), and trametinib (MEK), showed exceptionally high anti-tumor activities across all gynecologic tumors. In contrast, several agents, such as cabozantinib (VEGFR, MET, RET), ABT-888 (PARP), dabrafenib (BRAF), imatinib (Bcr/Abl), and sunitinib (PDGFR), demonstrated relatively minimal anti-cancer activities .
Pharmacogenomic interactions in EOCs
Identification of genomic biomarkers for drug sensitivity in gynecologic cancers
PARP inhibition demonstrated potent therapeutic efficacies in patients who were diagnosed with either metastatic breast or advanced ovarian cancers with germline BRCA1/2 mutations [43, 44]. Although statistically not significant, BRCA1/2 mutations were enriched in olaparib-sensitive PDCs (represented with Z-score < 0) (Fig. 4b). Interestingly, TP53 mutation also portrayed profound anti-tumor activity towards olaparib, regardless of histopathological subtype. Receiver operating characteristic (ROC) analysis revealed that genomic alterations of BRCA1/2 demonstrated positive correlations with olaparib sensitivity (AUC of ROC > 0.5; Fig. 4c). Furthermore, sole TP53 mutation or combination of TP53 with BRCA1/2 mutations revealed enhanced predictability to olaparib treatment (AUC = 0.9074 and 0.9 for TP53 and TP53 with BRCA1/2, respectively; Fig. 4c). To functionally validate our findings, we established cancer cell-line models that stably express various dominant-negative mutant forms of TP53 (R273H, R249S, and R175H)  in an OVISE ovarian cancer cell-line (TP53 wild-type). As suspected, cytotoxic activities of olaparib were significantly enhanced on all TP53 mutants (log (IC50) values for TP53 wild-type, R273H, R249S, and R175H mutants were 2.155 (95% CI 2.055 to 2.259), 1.421 (95% CI 1.336 to 1.506), 1.269 (95% CI 1.177 to 1.362), and 1.408 (95% CI 1.296 to 1.520) μM, respectively) (Fig. 4d).
Transcriptomic correlates of olaparib response in EOCs
The success of precision oncology depends on identification of effective drugs tailored to individual patients based on molecular profiling of the tumor. Comprehensive analyses of cancer genome have revealed a landscape of key genetic ablations that constitute complex process of tumorigenesis [1, 8, 46]. A large-scale compilation of pharmacological drug response with molecular characterization of cancer cell-line models has provided a reference point for evaluating potential genomic markers of drug sensitivity [12, 13, 16, 17]. Moreover, we have previously established a landscape of pharmacogenomic interactions using a library of short-term cultured PDCs to explore dynamic gene-drug associations and presented its clinical feasibility . As an extension, we generated a collection of 139 gynecologic tumors from patients with cervical, endometrial/uterine, or ovarian cancers. Through integrative genomic, transcriptomic, and pharmacological analyses, we have provided several new therapeutic insights for gynecologic malignancies (Fig. 1a).
We evaluated lineage-specific drug sensitivity in gynecologic tumors and discovered that EOCs demonstrated enhanced sensitivities to multiple EGFR inhibitors, while ECs were particularly sensitive to everolimus, an mTOR inhibitor. A number of clinical observations further advocated our results. Despite the small number of patients and lack of randomization, addition of erlotinib (EGFR) to platinum or paclitaxel provided favorable clinical outcomes for EOC patients, compared to platinum or paclitaxel treatment alone . Moreover, several clinical investigations demonstrated potential therapeutic benefits of mTOR targeted therapy in EC patients [36, 48, 49, 50]. Consistent with previous observations, we confirmed that serous and clear cell EOCs can be distinguished by evidently distinct genomic compositions (Fig. 3a), highlighting the need for molecular-based therapeutic approaches. Interestingly, our drug screening results, coupled with in vivo validations, propose clinical utility of cediranib (VEGFR) for HGSC patients, while PAM inhibitors could be more beneficial for those with clear cell carcinomas. Recent phase II and III clinical trials with cediranib also revealed that cediranib plus olaparib treatment resulted in a significant improvement in progression-free survival of HGSC patients compared to olaparib alone [51, 52]. As in vivo PDX results were only representative cases, further functional validations in a larger cohort are warranted to explore potential clinical applications of VEGFR and PAM inhibitors in serous and clear cell type tumors, respectively.
We also identified genomic correlates of drug sensitivity and resistance to olaparib. Approximately 13–18% of the HGSCs are attributable to BRCA1 or 2 germline mutations, and PARP inhibition therapy has been a successful approach for these patients. However, the need for discovering an alternative molecular biomarker to better predict the clinical efficacy against PARP inhibition treatment has been increasingly recognized due to global acquisition of olaparib resistance. Notably, we demonstrated that addition of TP53 mutation could be a significant molecular determinant for predicting potential clinical response to PARP inhibition therapy. The tumor suppressor protein p53 provides an essential role in governing cell cycle arrest or apoptosis upon DNA damage . However, the underlying molecular cascades that determine p53 protein stability and its activation are not fully understood. Accumulation of PARP1 is an early event where a single strand DNA break is generated and initiates base excision repair (BER) pathway . PARP-1 interacts with p53 to modulate DNA damage , and genotoxic drugs promote accumulation and activation of p53 in parp-1-deficient cells. Furthermore, p53 deficiency enhanced pharmacological sensitivity towards PARP inhibition therapy in mantle cell lymphoma . Consistent with such findings, our results suggest the prospect of targeting p53-deficient tumors with PARP inhibitors regardless of histopathological characteristics. Prevalence of TP53 mutation in HGSCs could contribute to the clinical success of PARP inhibitors [39, 42, 44].
Lastly, we identified SRC activation to be associated with therapeutic resistance to olaparib. The SRC family of non-receptor tyrosine kinases regulates essential cellular programs, including cellular proliferation, differentiation, migration, survival, and angiogenesis . A substantial number of studies have postulated that activation of SRC pathway contributes to inherent resistance to chemotherapy and inhibition of SRC pathway could potentially circumvent such mechanism [57, 58]. Moreover, transcriptional expressions of ID family genes were identified as molecular correlates of olaparib sensitivity. ID proteins are members of the large family of the helix-loop-helix (HLH) transcription factors. During development, ID proteins govern cell cycle and differentiation programs by modulating various cell cycle regulators . From the perspective of tumorigenesis, upregulation of ID protein is mediated by a group of proto-oncogenes, including Myc, Ras, and (EWS)-Ets, and prevents activation of various tumor suppressor genes , making it a promising therapeutic target .
Here, we generated an additional cohort at single tumor-lineage resolution, specializing in gynecologic malignancies. We performed systematic analyses of tumor genome and transcriptome to identify molecular determinants that dictate drug sensitivity to various molecular targeted drugs that are currently being used or under development. Our work provides an extension to current pharmacogenomic database in identifying predictive biomarkers and combinational approach to overcome cellular-intrinsic resistance to particular drug classes, including PARP inhibitors.
Gynecologic cancer specimens and their derivative cells
After receiving informed consent, tumor specimens and clinical records were obtained from patients undergoing surgery at Samsung Medical Center (SMC) in accordance with its Institutional Review Board. Surgical samples measuring ~ 5 × 5 × 5 mm3 were snap frozen using liquid nitrogen for genomic analysis. Portions of the surgical samples were enzymatically dissociated using Liberase TM (Roche) and cultured in DMEM/F12 media with l-glutamine (Thermofisher), N2 and B27 supplements (0.5× each; Thermofisher), human recombinant basic fibroblast growth factor (bFGF), and epidermal growth factor (EGF; 20 ng/ml each; R&D Systems).
Orthotopic xenograft animal models and drug treatment
Female BALB/c nude mice were purchased from ORIENT BIO (Sungnam, Korea). This study was performed in accordance with relevant guidelines and regulations. This study was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Samsung Biomedical Research Institute (SBRI). SBRI is an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC International, protocol no. H-A9-003)-accredited facility and abides by the Institute of Laboratory Animal Resources (ILAR) guidelines. To generate PDX models, patient tumor specimens were cut into small pieces (below 2–3 mm), implanted into the subrenal capsule of the left mouse kidney, and propagated by serial transplantation . After 1–2 weeks, the mice (n = 10 mice per group) were treated with either 1% polysorbate 80 or 0.5% methylcellulose or cediranib (6 mg/kg, qd, p.o.) or AZD8835 (25 mg/kg, bid, p.o.). Mice were monitored daily for tumor development and postoperative complications and were sacrificed between day 35 and 40 or if mice seemed moribund. Total body weight and tumor weight of each mouse were recorded. Tumors were fixed in formalin and embedded in paraffin or snap frozen in the OCT compound (Sakura Finetek, Japan, Tokyo, Japan) in liquid nitrogen.
Isolation of genomic DNA and quality control
Genomic DNA was extracted from fresh tissue specimens using the QIAamp DNA mini kit (Qiagen, Valencia, CA, USA) or from FFPE tissues using either the Promega Maxwell 16 CSC DNA FFPE kit or the QIAamp DNA FFPE Tissue kit according to the manufacturer’s manual. The purity, amount, and median size of the extracted DNA were measured by the Nanodrop 8000 UV-Vis spectrometer (Thermo Scientific Inc., Wilmington, DE, USA), Qubit 2.0 fluorometer (Life Technologies Inc., Grand Island, NY, USA), and 2200 TapeStation Instrument (Agilent Technologies, Santa Clara, CA, USA). In addition, ΔCt values were determined using real-time PCR (Agilent Technologies) with Mx3005p instrument (Agilent Technologies, USA), FFPE QC kit (Illumina, cat no. WG-321-1001), and Brilliant Ultra-Fast SYBR Green qPCR (Agilent Technologies, cat no. 600882). If DNA meets the quality criteria such as (i) purity to absorption ratio (260 nm/280 nm) > 1.8, 260 nm/230 nm > 1.8; (ii) total amount > 250 ng; (iii) degradation to ΔCt value < 2.0; or DNA median size > 0.35 kb, it is proceeded onto the sequencing step.
Panel design and sequencing
Samples were profiled on CancerSCAN, a targeted sequencing platform designed at Samsung Medical Center. This customized platform offers flexibility to include target genes that are curated from the literature or requested by researchers and clinicians. To obtain cancer panel sequencing data, CancerSCAN probes were designed to enrich the exons of 80 target genes (Additional file 9: Table S8). Genomic DNA was sheared using the Covaris S220 (Covaris, Woburn, MA) to construct a sequencing library using the SureSelect XT Reagent Kit, HSQ (Agilent Technologies) on target genes. A paired-end sequencing library was purified and amplified with a barcode tag, and the library quality and quantity were determined. Sequencing was carried out using the 100-bp paired-end mode of the TruSeq Rapid PE Cluster kit and TruSeq Rapid SBS kit on a HiSeq 2500 sequencing platform (Illumina, San Diego, CA, USA).
Bulk RNA sequencing
RNA-seq libraries were prepared using the Illumina TrueSeq RNA Sample Prep kit. Sequenced reads were mapped onto hg19 using the Burrows-Wheeler Aligner (BWA). The initial alignment BAM files were sorted and summarized into BED files using SAMtools and bedTools. The BED files were used to calculate values of RPKM (reads per kilobase of transcript per million reads) for each gene, using DEGseq package.
PDCs were cultured in serum-free medium, dissociated into single cells, and seeded in 384-well plates at a density of 500 cells per well in duplicate or triplicate for each treatment. The drug panel consisted of 37 anti-cancer agents targeting oncogenic signals. All drug libraries were purchased from Selleckchem. PDCs were treated with drugs in a four-fold and seven-point serial dilution series from 20 to 4.88 nM using a Janus Automated Workstation (PerkinElmer, Waltham, MA, USA). After 7 days of incubation at 37 °C in a 5% CO2 humidified incubator, cell viability was analyzed using an adenosine triphosphate (ATP) monitoring system based on firefly luciferase (ATPLite™ 1step, PerkinElmer). Viable cells were estimated using an EnVision Multilabel Reader (PerkinElmer). The controls, dimethyl sulfoxide (DMSO) vehicle, were used to calculate relative cell viability for each plate and to normalize the data on a per-plate basis. Dose response curve (DRC) fitting was performed using GraphPad Prism 5 (GraphPad) and evaluated by measuring the area under curve (AUC). In brief, each plate was normalized to the mean of the seven conditions on the plate with a DMSO control. After normalization, best-fit lines and the resulting IC50values were calculated using GraphPad: [log(inhibitor) vs. response − variable slope (four parameters)]. Y = Bottom + (Top − Bottom)/(1 + 10^((logIC50 − X) × HillSlope)). The AUC of each curve was determined using GraphPad Prism, ignoring regions defined by fewer than two peaks. Non-convergence or ambiguous curves are excluded in every analysis.
Pharmacogenomic interactions on major genomic alterations
For gene-drug associations, a list of major cancer-driver alterations, including single nucleotide variations, small insertions, and deletions, was considered as a predictive genomic biomarker to evaluate drug response interactions. For each drug candidate, drug sensitivity data (AUCs) were analyzed by comparing tumors with the selected genomic alteration to those without using the Wilcoxon rank-sum test. Samples with unknown status of a given alteration were excluded from the analysis. To evaluate lineage-specific drug sensitivities in gynecologic tumors, drug sensitivity data were analyzed by comparing tumors from each pathologic entity to the rest using the Wilcoxon rank-sum test. For transcriptome analysis, tumors were classified as “sensitive” (Z-score < − 0.5) or “resistant” (Z-score > 0.5) based on drug sensitivity data and subjected to Gene Set Enrichment Analysis (GSEA).
Plasmid preparation and stable cell establishment
Lentiviral vectors encoding TP53 mutants (R175H, R273H, or R249S) in pLenti6/V5 plasmid were purchased from Addgene (cat no. 22936, 22934, and 22935, respectively). Lentivirus was prepared by transfecting plasmids into the 293T cells using pMD2G, psPAX2 (Addgene), and Lipofectamine 2000 (Thermofisher). After the initial transfection, supernatants were collected after 48 and 72 h and filtered through 0.45 μM filters (Milipore). To generate stable TP53 mutant-expressing cell lines, lentivirus particles were incubated with ovarian clear cell carcinoma, OVISE, and treated with polybrene (8 μg/ml, Sigma) for 48 h and blasticidin (5 μg/ml, Sigma) selection was performed for 2 weeks.
Immunohistochemical staining was performed on formalin-fixed, paraffin-embedded, 4–5-μm-thick tissue sections, using the Bond-maxTM automated immunostainer (Leica Biosystems, Melbourne, Australia) and BondTM Polymer Refines Detection Kit (Vision Biosystems, Melbourne, Australia). Mouse monoclonal anti-ID2 antibody (1:100; NBP2-66898, Novus Biologicals, USA) was used. Briefly, antigen retrieval was carried out at 97 °C for 20 min in ER1 buffer. After blocking the endogenous peroxidase activity with 3% hydrogen peroxidase for 10 min, primary antibody was incubated for 60 min at room temperature. To verify antibody specificity, anti-mouse IgG (AI-2000; Vector Laboratories, Burlingame, CA, USA) was used as a control. The degree of immunostaining of ID2 was evaluated according to staining proportion of positively stained cancer cell nucleus and the staining intensity, as previously described . Briefly, the areas of stained cancer cells were scored as follows: the percentage of positive cells (0–100%) and intensity scaled from 0 to 2 (null = 0, weak to moderate = 1, strong = 2).
All statistical analyses were conducted by either Wilcoxon’s rank-sum test (two-sided), Pearson’s correlation coefficient test, or Fisher’s exact test (two-sided). Survival curves were estimated with the Kaplan-Meier method. All statistical analyses were conducted and obtained using the R software (https://www.r-project.org).
We thank the Samsung Medical Center BioBank for providing the biospecimens that were used in this study.
The review history is available as Additional file 10.
JKS, JRH, and YJC are the co-first authors. JKS, JRH, and YJC performed the majority of the experiments and analyses. JWL, DHN, and JKL provided the concept of the study. JYR and JJC participated in several experiments. JJC, SYJ, JK, MSK, ESP, YYL, CHC, TJK, BGK, and DSB provided the surgical samples and performed the clinical interpretation. JSB and WYP generated the genome and transcriptome data. JKS, HJC, and HK processed and analyzed the sequencing data. YL, NGH, JWO, YJS, JYK, and YJS generated the patient-derived cell library and conducted the drug response analysis. TL, HSK, SYS, HDH, and HJA performed the pathological analysis of biospecimens. JKS, JKL, and JWL wrote the manuscript and organized the figures and tables. AKS and RR reviewed and edited the manuscript. JWL, DHN, and JKL designed and supervised the entire project. All authors read and approved the final manuscript.
This research was supported in part by a grant from the Korea Health Technology R&D project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (HI14C3418 and HI18C1953); the National R&D Program for Cancer Control, Ministry for Health, Welfare and Family Affairs, Republic of Korea (1520100); and the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (NRF-MSIP-2016R1A5A2945889, NRF-MEST-2016R1A2B3006644, NRF-2017R1A2B4003434, and NRF-2018R1C1B3001648).
Ethics approval and consent to participate
The study was approved by the local committee on the use of human samples for experimental studies of the Samsung Medical Center (SMC), Seoul, Republic of Korea (IRB file #201004004). Written informed consents were provided by the participants prior to enrollment. All experimental methods abided by the Helsinki Declaration.
Consent for publication
Do-Hyun Nam is the CEO of AimedBio Inc. and owns shares of AimedBio Inc. which owns IPs for Avatascan. The other authors declare that they have no competing interests.
- 4.Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, Fleming T, Eiermann W, Wolter J, Pegram M, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med. 2001;344:783–92.PubMedCrossRefPubMedCentralGoogle Scholar
- 6.O'Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, Cervantes F, Cornelissen JJ, Fischer T, Hochhaus A, Hughes T, et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med. 2003;348:994–1004.PubMedCrossRefPubMedCentralGoogle Scholar
- 9.Rubio-Perez C, Tamborero D, Schroeder MP, Antolin AA, Deu-Pons J, Perez-Llamas C, Mestres J, Gonzalez-Perez A, Lopez-Bigas N. In silico prescription of anticancer drugs to cohorts of 28 tumor types reveals targeting opportunities. Cancer Cell. 2015;27:382–96.PubMedCrossRefPubMedCentralGoogle Scholar
- 21.Cancer Genome Atlas Research N, Albert Einstein College of M, Analytical Biological S, Barretos Cancer H, Baylor College of M, Beckman Research Institute of City of H, Buck Institute for Research on A, Canada's Michael Smith Genome Sciences C, Harvard Medical S, Helen FGCC, et al. Integrated genomic and molecular characterization of cervical cancer. Nature. 2017;543:378–84.CrossRefGoogle Scholar
- 29.House CD, Hernandez L, Annunziata CM. In vitro enrichment of ovarian cancer tumor-initiating cells. J Vis Exp. 2015;96:e52446.Google Scholar
- 32.Jang IS, Neto EC, Guinney J, Friend SH, Margolin AA. Systematic assessment of analytical methods for drug sensitivity prediction from cancer cell line data. Pac Symp Biocomput. 2014:63–74.Google Scholar
- 35.Gao W, Wang M, Wang L, Lu H, Wu S, Dai B, Ou Z, Zhang L, Heymach JV, Gold KA, et al. Selective antitumor activity of ibrutinib in EGFR-mutant non-small cell lung cancer cells. J Natl Cancer Inst. 2014;106(9).Google Scholar
- 38.Cole AJ, Dwight T, Gill AJ, Dickson KA, Zhu Y, Clarkson A, Gard GB, Maidens J, Valmadre S, Clifton-Bligh R, Marsh DJ. Assessing mutant p53 in primary high-grade serous ovarian cancer using immunohistochemistry and massively parallel sequencing. Sci Rep. 2016;6:26191.PubMedPubMedCentralCrossRefGoogle Scholar
- 41.Chandler RL, Damrauer JS, Raab JR, Schisler JC, Wilkerson MD, Didion JP, Starmer J, Serber D, Yee D, Xiong J, et al. Coexistent ARID1A-PIK3CA mutations promote ovarian clear-cell tumorigenesis through pro-tumorigenic inflammatory cytokine signalling. Nat Commun. 2015;6:6118.PubMedPubMedCentralCrossRefGoogle Scholar
- 47.Blank SV, Christos P, Curtin JP, Goldman N, Runowicz CD, Sparano JA, Liebes L, Chen HX, Muggia FM. Erlotinib added to carboplatin and paclitaxel as first-line treatment of ovarian cancer: a phase II study based on surgical reassessment. Gynecol Oncol. 2010;119:451–6.PubMedPubMedCentralCrossRefGoogle Scholar
- 48.Slomovitz BM, Lu KH, Johnston T, Coleman RL, Munsell M, Broaddus RR, Walker C, Ramondetta LM, Burke TW, Gershenson DM, Wolf J. A phase 2 study of the oral mammalian target of rapamycin inhibitor, everolimus, in patients with recurrent endometrial carcinoma. Cancer. 2010;116:5415–9.PubMedPubMedCentralCrossRefGoogle Scholar
- 51.Liu JF, Barry WT, Birrer M, Lee JM, Buckanovich RJ, Fleming GF, Rimel B, Buss MK, Nattam S, Hurteau J, et al. Combination cediranib and olaparib versus olaparib alone for women with recurrent platinum-sensitive ovarian cancer: a randomised phase 2 study. Lancet Oncol. 2014;15:1207–14.PubMedPubMedCentralCrossRefGoogle Scholar
- 52.Ledermann JA, Embleton AC, Raja F, Perren TJ, Jayson GC, Rustin GJS, Kaye SB, Hirte H, Eisenhauer E, Vaughan M, et al. Cediranib in patients with relapsed platinum-sensitive ovarian cancer (ICON6): a randomised, double-blind, placebo-controlled phase 3 trial. Lancet. 2016;387:1066–74.PubMedCrossRefPubMedCentralGoogle Scholar
- 63.Sa JK, Hwang JR, Cho YJ, Ryu JY, Choi JJ, Jeong SY, Kim J, Kim MS, Paik ES, Lee YY, Choi CH, Kim TJ, Kim BG, Bae DS, Lee Y, Her NG, Shin YJ, Cho HJ, Kim JY, Seo YJ, Koo H, Oh JW, Lee T, Kim HS, Song SY, Bae JS, Park WY, Han HD, Ahn HJ, Sood AK, Rabadan R, Lee JK, Nam DH, Lee JW. Gynecologic cancer genome. European Genome-phenome Archive. 2019.https://www.ebi.ac.uk/ega/studies/EGAS00001003556. Accessed 17 Nov 2019.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.