Vinodh N. Rajapakse

Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse

To evaluate the role of oncogenic RAS mutations in pancreatic tumorigenesis, we directed endogenous expression of KRAS(G12D) to progenitor cells of the mouse pancreas. We find that physiological levels of Kras(G12D) induce ductal lesions that recapitulate the full spectrum of human pancreatic intraepithelial neoplasias (PanINs), putative precursors to invasive pancreatic cancer. The PanINs are highly proliferative, show evidence of histological progression, and activate signaling pathways normally quiescent in ductal epithelium, suggesting potential therapeutic and chemopreventive targets for the cognate human condition. At low frequency, these lesions also progress spontaneously to invasive and metastatic adenocarcinomas, establishing PanINs as definitive precursors to the invasive disease. Finally, mice with PanINs have an identifiable serum proteomic signature, suggesting a means of detecting the preinvasive state in patients.

The histone variant H2A.X is a regulator of the epithelial–mesenchymal transition

The epithelial–mesenchymal transition (EMT), considered essential for metastatic cancer, has been a focus of much research, but important questions remain. Here, we show that silencing or removing H2A.X, a histone H2A variant involved in cellular DNA repair and robust growth, induces mesenchymal-like characteristics including activation of EMT transcription factors, Slug and ZEB1, in HCT116 human colon cancer cells. Ectopic H2A.X re-expression partially reverses these changes, as does silencing Slug and ZEB1. In an experimental metastasis model, the HCT116 parental and H2A.X-null cells exhibit a similar metastatic behaviour, but the cells with re-expressed H2A.X are substantially more metastatic. We surmise that H2A.X re-expression leads to partial EMT reversal and increases robustness in the HCT116 cells, permitting them to both form tumours and to metastasize. In a human adenocarcinoma panel, H2A.X levels correlate inversely with Slug and ZEB1 levels. Together, these results point to H2A.X as a regulator of EMT.

SLFN11 Is a Transcriptional Target of EWS-FLI1 and a Determinant of Drug Response in Ewing Sarcoma.

Purpose: SLFN11 was identified as a critical determinant of response to DNA-targeted therapies by analyzing gene expression and drug sensitivity of NCI-60 and CCLE datasets. However, how SLFN11 is regulated in cancer cells remained unknown. Ewing sarcoma, which is characterized by the chimeric transcription factor EWS-FLI1, has notably high SLFN11 expression, leading us to investigate whether EWS-FLI1 drives SLFN11 expression and the role of SLFN11 in the drug response of Ewing sarcoma cells. Experimental Design: Binding sites of EWS-FLI1 on the SLFN11 promoter were analyzed by chromatin immunoprecipitation sequencing and promoter-luciferase reporter analyses. The relationship between SLFN11 and EWS-FLI1 were further examined in EWS-FLI1-knockdown or -overexpressing cells and in clinical tumor samples. Results: EWS-FLI1 binds near the transcription start site of SLFN11 promoter and acts as a positive regulator of SLFN11 expression in Ewing sarcoma cells. EWS-FLI1–mediated SLFN11 expression is responsible for high sensitivity of Ewing sarcoma to camptothecin and combinations of PARP inhibitors with temozolomide. Importantly, Ewing sarcoma patients with higher SLFN11 expression showed better tumor-free survival rate. The correlated expression between SLFN11 and FLI1 extends to leukemia, pediatric, colon, breast, and prostate cancers. In addition, expression of other ETS members correlates with SLFN11 in NCI-60 and CCLE datasets, and molecular experiments demonstrate that ETS1 acts as a positive regulator for SLFN11 expression in breast cancer cells. Conclusions: Our results imply the emerging relevance of SLFN11 as an ETS transcription factor response gene and for therapeutic response to topoisomerase I inhibitors and temozolomide–PARP inhibitor combinations in ETS-activated cancers. Clin Cancer Res; 21(18); 4184–93. ©2015 AACR. See related commentary by Kovar, p. 4033

Alterations of DNA repair genes in the NCI-60 cell lines and their predictive value for anticancer drug activity.

Loss of function of DNA repair (DNAR) genes is associated with genomic instability and cancer predisposition; it also makes cancer cells reliant on a reduced set of DNAR pathways to resist DNA-targeted therapy, which remains the core of the anticancer armamentarium. Because the landscape of DNAR defects across numerous types of cancers and its relation with drug activity have not been systematically examined, we took advantage of the unique drug and genomic databases of the US National Cancer Institute cancer cell lines (the NCI-60) to characterize 260 DNAR genes with respect to deleterious mutations and expression down-regulation; 169 genes exhibited a total of 549 function-affecting alterations, with 39 of them scoring as putative knockouts across 31 cell lines. Those mutations were compared to tumor samples from 12 studies of The Cancer Genome Atlas (TCGA) and The Cancer Cell Line Encyclopedia (CCLE). Based on this compendium of alterations, we determined which DNAR genomic alterations predicted drug response for 20,195 compounds present in the NCI-60 drug database. Among 242 DNA damaging agents, 202 showed associations with at least one DNAR genomic signature. In addition to SLFN11, the Fanconi anemia-scaffolding gene SLX4 (FANCP/BTBD12) stood out among the genes most significantly related with DNA synthesis and topoisomerase inhibitors. Depletion and complementation experiments validated the causal relationship between SLX4 defects and sensitivity to raltitrexed and cytarabine in addition to camptothecin. Therefore, we propose new rational uses for existing anticancer drugs based on a comprehensive analysis of DNAR genomic parameters.

Using drug response data to identify molecular effectors, and molecular "omic" data to identify candidate drugs in cancer

The current convergence of molecular and pharmacological data provides unprecedented opportunities to gain insights into the relationships between the two types of data. Multiple forms of large-scale molecular data, including but not limited to gene and microRNA transcript expression, DNA somatic and germline variations from next-generation DNA and RNA sequencing, and DNA copy number from array comparative genomic hybridization are all potentially informative when one attempts to recognize the panoply of potentially influential events both for cancer progression and therapeutic outcome. Concurrently, there has also been a substantial expansion of the pharmacological data being accrued in a systematic fashion. For cancer cell lines, the National Cancer Institute cell line panel (NCI-60), the Cancer Cell Line Encyclopedia (CCLE), and the collaborative Genomics of Drug Sensitivity in Cancer (GDSC) databases all provide subsets of these forms of data. For the patient-derived data, The Cancer Genome Atlas (TCGA) provides analogous forms of genomic information along with treatment histories. Integration of these data in turn relies on the fields of statistics and statistical learning. Multiple algorithmic approaches may be chosen, depending on the data being considered, and the nature of the question being asked. Combining these algorithms with prior biological knowledge, the results of molecular biological studies, and the consideration of genes as pathways or functional groups provides both the challenge and the potential of the field. The ultimate goal is to provide a paradigm shift in the way that drugs are selected to provide a more targeted and efficacious outcome for the patient.

rcellminer: exploring molecular profiles and drug response of the NCI-60 cell lines in R.

PURPOSE The rcellminer R package provides a wide range of functionality to help R users access and explore molecular profiling and drug response data for the NCI-60. The package enables flexible programmatic access to CellMiners unparalleled breadth of NCI-60 data, including gene and protein expression, copy number, whole exome mutations, as well as activity data for ∼21K compounds, with information on their structure, mechanism of action and repeat screens. Functions are available to easily visualize compound structures, activity patterns and molecular feature profiles. Additionally, embedded R Shiny applications allow interactive data exploration. AVAILABILITY AND IMPLEMENTATION rcellminer is compatible with R 3.2 and above on Windows, Mac OS X and Linux. The package, documentation, tutorials and Shiny-based applications are available through Bioconductor (http://www.bioconductor.org/packages/rcellminer); ongoing updates will occur according to the Bioconductor release schedule with new CellMiner data. The package is free and open-source (LGPL 3). CONTACT lunaa@cbio.mskcc.org or vinodh.rajapakse@nih.gov.

Cytidine Deaminase Deficiency Reveals New Therapeutic Opportunities against Cancer

Purpose: One of the main challenges in cancer therapy is the identification of molecular mechanisms mediating resistance or sensitivity to treatment. Cytidine deaminase (CDA) was reported to be downregulated in cells derived from patients with Bloom syndrome, a genetic disease associated with a strong predisposition to a wide range of cancers. The purpose of this study was to determine whether CDA deficiency could be associated with tumors from the general population and could constitute a predictive marker of susceptibility to antitumor drugs. Experimental Design: We analyzed CDA expression in silico, in large datasets for cancer cell lines and tumors and in various cancer cell lines and primary tumor tissues using IHC, PDXs, qRT-PCR, and Western blotting. We also studied the mechanism underlying CDA silencing and searched for molecules that might target specifically CDA-deficient tumor cells using in silico analysis coupled to classical cellular experimental approaches. Results: We found that CDA expression is downregulated in about 60% of cancer cells and tissues. We demonstrate that DNA methylation is a prevalent mechanism of CDA silencing in tumors. Finally, we show that CDA-deficient tumor cells can be specifically targeted with epigenetic treatments and with the anticancer drug aminoflavone. Conclusions: CDA expression status identifies new subgroups of cancers, and CDA deficiency appears to be a novel and relevant predictive marker of susceptibility to antitumor drugs, opening up new possibilities for treating cancer. Clin Cancer Res; 23(8); 2116–26. ©2016 AACR.

Temozolomide in the Era of Precision Medicine

In the January 1, 2017, issue of Cancer Research, Nagel and colleagues demonstrate the value of assays that determine the DNA repair capacity of cancers in predicting response to temozolomide. Using a fluorescence-based multiplex flow cytometric host cell reactivation assay that provides simultaneous readout of DNA repair capacity across multiple pathways, they show that the multivariate drug response models derived from cell line data were applicable to patient-derived xenograft models of glioblastoma. In this commentary, we first outline the mechanism of activity and current clinical application of temozolomide, which, until now, has been largely limited to glioblastoma. Given the challenges of clinical application of functional assays, we argue that functional readouts be approximated by genomic signatures. In this context, a combination of MGMT activity and mismatch repair (MMR) status of the tumor are important parameters that determine sensitivity to temozolomide. More reliable methods are needed to determine MGMT activity as DNA methylation, the current standard, does not accurately reflect the expression of MGMT. Also, genomics for MMR are warranted. Furthermore, based on patterns of MGMT expression across different solid tumors, we make a case for revisiting temozolomide use in a broader spectrum of cancers based on our current understanding of its molecular basis of activity. Cancer Res; 77(4); 823-6. ©2017 AACR.

The NCI-60 Methylome and Its Integration into CellMiner

A unique resource for systems pharmacology and genomic studies is the NCI-60 cancer cell line panel, which provides data for the largest publicly available library of compounds with cytotoxic activity (∼21,000 compounds), including 108 FDA-approved and 70 clinical trial drugs as well as genomic data, including whole-exome sequencing, gene and miRNA transcripts, DNA copy number, and protein levels. Here, we provide the first readily usable genome-wide DNA methylation database for the NCI-60, including 485,577 probes from the Infinium HumanMethylation450k BeadChip array, which yielded DNA methylation signatures for 17,559 genes integrated into our open access CellMiner version 2.0 (https://discover.nci.nih.gov/cellminer). Among new insights, transcript versus DNA methylation correlations revealed the epithelial/mesenchymal gene functional category as being influenced most heavily by methylation. DNA methylation and copy number integration with transcript levels yielded an assessment of their relative influence for 15,798 genes, including tumor suppressor, mitochondrial, and mismatch repair genes. Four forms of molecular data were combined, providing rationale for microsatellite instability for 8 of the 9 cell lines in which it occurred. Individual cell line analyses showed global methylome patterns with overall methylation levels ranging from 17% to 84%. A six-gene model, including PARP1, EP300, KDM5C, SMARCB1, and UHRF1 matched this pattern. In addition, promoter methylation of two translationally relevant genes, Schlafen 11 (SLFN11) and methylguanine methyltransferase (MGMT), served as indicators of therapeutic resistance or susceptibility, respectively. Overall, our database provides a resource of pharmacologic data that can reinforce known therapeutic strategies and identify novel drugs and drug targets across multiple cancer types. Cancer Res; 77(3); 601-12. ©2016 AACR.

Overcoming resistance to DNA targeted agents by epigenetic activation of Schlafen 11 (SLFN11) expression with class I histone deacetylase inhibitors

Purpose: Schlafen 11 (SLFN11), a putative DNA/RNA helicase is a dominant genomic determinant of response to DNA-damaging agents and is frequently not expressed in cancer cells. Whether histone deacetylase (HDAC) inhibitors can be used to release SLFN11 and sensitize SLFN11-inactivated cancers to DNA-targeted agents is tested here. Experimental Design: SLFN11 expression was examined in The Cancer Genome Atlas (TCGA), in cancer cell line databases and in patients treated with romidepsin. Isogenic cells overexpressing or genetically inactivated for SLFN11 were used to investigate the effect of HDAC inhibitors on SLFN11 expression and sensitivity to DNA-damaging agents. Results: SLFN11 expression is suppressed in a broad fraction of common cancers and cancer cell lines. In cancer cells not expressing SLFN11, transfection of SLFN11 sensitized the cells to camptothecin, topotecan, hydroxyurea, and cisplatin but not to paclitaxel. SLFN11 mRNA and protein levels were strongly induced by class I (romidepsin, entinostat), but not class II (roclinostat) HDAC inhibitors in a broad panel of cancer cells. SLFN11 expression was also enhanced in peripheral blood mononuclear cells of patients with circulating cutaneous T-cell lymphoma treated with romidepsin. Consistent with the epigenetic regulation of SLFN11, camptothecin and class I HDAC inhibitors were synergistic in many of the cell lines tested. Conclusions: This study reports the prevalent epigenetic regulation of SLFN11 and the dominant stimulatory effect of HDAC inhibitors on SLFN11 expression. Our results provide a rationale for combining class I HDAC inhibitors and DNA-damaging agents to overcome epigenetic inactivation of SLFN11-mediated resistance to DNA-targeted agents. Clin Cancer Res; 24(8); 1944–53. ©2018 AACR.