Drug Response Information Research Articles

DrugFormer: Graph-Enhanced Language Model to Predict Drug Sensitivity.

Drug resistance poses a crucial challenge in healthcare, with response rates to chemotherapy and targeted therapy remaining low. Individual patient's resistance is exacerbated by the intricate heterogeneity of tumor cells, presenting significant obstacles to effective treatment. To address this challenge, DrugFormer, a novel graph-augmented large language model designed to predict drug resistance at single-cell level is proposed. DrugFormer integrates both serialized gene tokens and gene-based knowledge graphs for the accurate predictions of drug response. After training on comprehensive single-cell data with drug response information, DrugFormer model presents outperformance, with higher F1, precision, and recall in predicting drug response. Based on the scRNA-seq data from refractory multiple myeloma (MM) and acute myeloid leukemia (AML) patients, DrugFormer demonstrates high efficacy in identifying resistant cells and uncovering underlying molecular mechanisms. Through pseudotime trajectory analysisunique drug-resistant cellular states associated with poor patient outcomes are revealed. Furthermore, DrugFormer identifies potential therapeutic targets, such as COX8A, for overcoming drug resistance across different cancer types. In conclusion, DrugFormer represents a significant advancement in the field of drug resistance prediction, offering a powerful tool for unraveling the heterogeneity of cellular response to drugs and guiding personalized treatment strategies.

D3EGFR: a webserver for deep learning-guided drug sensitivity prediction and drug response information retrieval for EGFR mutation-driven lung cancer.

As key oncogenic drivers in non-small-cell lung cancer (NSCLC), various mutations in the epidermal growth factor receptor (EGFR) with variable drug sensitivities have been a major obstacle for precision medicine. To achieve clinical-level drug recommendations, a platform for clinical patient case retrieval and reliable drug sensitivity prediction is highly expected. Therefore, we built a database, D3EGFRdb, with the clinicopathologic characteristics and drug responses of 1339 patients with EGFR mutations via literature mining. On the basis of D3EGFRdb, we developed a deep learning-based prediction model, D3EGFRAI, for drug sensitivity prediction of new EGFR mutation-driven NSCLC. Model validations of D3EGFRAI showed a prediction accuracy of 0.81 and 0.85 for patients from D3EGFRdb and our hospitals, respectively. Furthermore, mutation scanning of the crucial residues inside drug-binding pockets, which may occur in the future, was performed to explore their drug sensitivity changes. D3EGFR is the first platform to achieve clinical-level drug response prediction of all approved small molecule drugs for EGFR mutation-driven lung cancer and is freely accessible at https://www.d3pharma.com/D3EGFR/index.php.

Genomic and transcriptomic analyses identify a prognostic gene signature and predict response to therapy in pleural and peritoneal mesothelioma

Malignant mesothelioma is an aggressive cancer with limited treatment options and poor prognosis. A better understanding of mesothelioma genomics and transcriptomics could advance therapies. Here, we present a mesothelioma cohort of 122 patients along with their germline and tumor whole-exome and tumor RNA sequencing data as well as phenotypic and drug response information. We identify a 48-gene prognostic signature that is highly predictive of mesothelioma patient survival, including CCNB1, the expression of which is highly predictive of patient survival on its own. In addition, we analyze the transcriptomics data to study the tumor immune microenvironment and identify synthetic-lethality-based signatures predictive of response to therapy. This germline and somatic whole-exome sequencing as well as transcriptomics data from the same patient are a valuable resource to address important biological questions, including prognostic biomarkers and determinants of treatment response in mesothelioma.

Trends and Potential of Machine Learning and Deep Learning in Drug Study at Single-Cell Level.

Cancer treatments always face challenging problems, particularly drug resistance due to tumor cell heterogeneity. The existing datasets include the relationship between gene expression and drug sensitivities; however, the majority are based on tissue-level studies. Study drugs at the single-cell level are perspective to overcome minimal residual disease caused by subclonal resistant cancer cells retained after initial curative therapy. Fortunately, machine learning techniques can help us understand how different types of cells respond to different cancer drugs from the perspective of single-cell gene expression. Good modeling using single-cell data and drug response information will not only improve machine learning for cell-drug outcome prediction but also facilitate the discovery of drugs for specific cancer subgroups and specific cancer treatments. In this paper, we review machine learning and deep learning approaches in drug research. By analyzing the application of these methods on cancer cell lines and single-cell data and comparing the technical gap between single-cell sequencing data analysis and single-cell drug sensitivity analysis, we hope to explore the trends and potential of drug research at the single-cell data level and provide more inspiration for drug research at the single-cell level. We anticipate that this review will stimulate the innovative use of machine learning methods to address new challenges in precision medicine more broadly.

RAMP: response-aware multi-task learning with contrastive regularization for cancer drug response prediction.

The accurate prediction of cancer drug sensitivity according to the multiomics profiles of individual patients is crucial for precision cancer medicine. However, the development of prediction models has been challenged by the complex crosstalk of input features and the resistance-dominant drug response information contained in public databases. In this study, we propose a novel multidrug response prediction framework, response-aware multitask prediction (RAMP), via a Bayesian neural network and restrict it by soft-supervised contrastive regularization. To utilize network embedding vectors as representation learning features for heterogeneous networks, we harness response-aware negative sampling, which applies cell line-drug response information to the training of network embeddings. RAMP overcomes the prediction accuracy limitation induced by the imbalance of trained response data based on the comprehensive selection and utilization of drug response features. When trained on the Genomics of Drug Sensitivity in Cancer dataset, RAMP achieved an area under the receiver operating characteristic curve > 89%, an area under the precision-recall curve > 59% and an $\textrm{F}_1$ score > 52% and outperformed previously developed methods on both balanced and imbalanced datasets. Furthermore, RAMP predicted many missing drug responses that were not included in the public databases. Our results showed that RAMP will be suitable for the high-throughput prediction of cancer drug sensitivity and will be useful for guiding cancer drug selection processes. The Python implementation for RAMP is available at https://github.com/hvcl/RAMP.

Predicting cancer drug response using parallel heterogeneous graph convolutional networks with neighborhood interactions.

Due to cancer heterogeneity, the therapeutic effect may not be the same when a cohort of patients of the same cancer type receive the same treatment. The anticancer drug response prediction may help develop personalized therapy regimens to increase survival and reduce patients' expenses. Recently, graph neural network-based methods have aroused widespread interest and achieved impressive results on the drug response prediction task. However, most of them apply graph convolution to process cell line-drug bipartite graphs while ignoring the intrinsic differences between cell lines and drug nodes. Moreover, most of these methods aggregate node-wise neighbor features but fail to consider the element-wise interaction between cell lines and drugs. This work proposes a neighborhood interaction (NI)-based heterogeneous graph convolution network method, namely NIHGCN, for anticancer drug response prediction in an end-to-end way. Firstly, it constructs a heterogeneous network consisting of drugs, cell lines and the known drug response information. Cell line gene expression and drug molecular fingerprints are linearly transformed and input as node attributes into an interaction model. The interaction module consists of a parallel graph convolution network layer and a NI layer, which aggregates node-level features from their neighbors through graph convolution operation and considers the element-level of interactions with their neighbors in the NI layer. Finally, the drug response predictions are made by calculating the linear correlation coefficients of feature representations of cell lines and drugs. We have conducted extensive experiments to assess the effectiveness of our model on Cancer Drug Sensitivity Data (GDSC) and Cancer Cell Line Encyclopedia (CCLE) datasets. It has achieved the best performance compared with the state-of-the-art algorithms, especially in predicting drug responses for new cell lines, new drugs and targeted drugs. Furthermore, our model that was well trained on the GDSC dataset can be successfully applied to predict samples of PDX and TCGA, which verified the transferability of our model from cell line in vitro to the datasets in vivo. The source code can be obtained from https://github.com/weiba/NIHGCN. Supplementary data are available at Bioinformatics online.

Abstract 6015: Comprehensive evaluation of three-dimensional (3D) culture systems as compound screening platforms with primary cancer cell lines

Abstract There are a number of challenges associated with personalized medicines. For example, many models revolve around the question of whether or not the in vitro models capture the relevant in vivo pathobiology and drug response information that may impact a patient's response to a therapy. Most in vitro oncology drug screening techniques are based on two-dimensional (2D) cell culture, which, to date, has failed to accurately imitate the physiological microenvironments due to the inherent limitations, such as the lack of interactions between intracellular and extracellular environments, or modifications in three-dimensional cell morphology, polarity, and division. These disadvantages prompted the development of models which would better simulate the microenvironment of tumor cells. Three-dimensional (3D) culture systems are becoming increasingly popular due to their ability to mimic natural cell characteristics and tissue-like structures compared with traditional 2D monolayer cell culture. Currently, multicellular spheroid and organoid cell culture are two popular 3D culture methods to evaluate drug activity. Spheroid cell culture utilizes ultra-low attachment plate to allow cells grow in spheroid format. Organoid cell culture is accomplished by growing cells in extracellular matrix scaffolds (such as Matrigel). In this study, we attempted to compare different drug screening platforms, including patient-derived xenograft (PDX) and two 3D cell cultures (spheroid and organoid). Compound sensitivity, genomics, pathology and expression of clinical marker were profiled. The purpose of this study is to identify an optimal model as a drug testing platform and a bridge between in vitro studies and animal experiments for identifying novel therapeutic targets. Citation Format: Chao Zhang, Qinghua Wu, Lei Jin, Deyin Chen, Xuzhen Tang, Hui Qi, Wenjie Gao, Qunsheng Ji. Comprehensive evaluation of three-dimensional (3D) culture systems as compound screening platforms with primary cancer cell lines [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 6015.

CTR-DB, an omnibus for patient-derived gene expression signatures correlated with cancer drug response.

To date, only some cancer patients can benefit from chemotherapy and targeted therapy. Drug resistance continues to be a major and challenging problem facing current cancer research. Rapidly accumulated patient-derived clinical transcriptomic data with cancer drug response bring opportunities for exploring molecular determinants of drug response, but meanwhile pose challenges for data management, integration, and reuse. Here we present the Cancer Treatment Response gene signature DataBase (CTR-DB, http://ctrdb.ncpsb.org.cn/), a unique database for basic and clinical researchers to access, integrate, and reuse clinical transcriptomes with cancer drug response. CTR-DB has collected and uniformly reprocessed 83 patient-derived pre-treatment transcriptomic source datasets with manually curated cancer drug response information, involving 28 histological cancer types, 123 drugs, and 5139 patient samples. These data are browsable, searchable, and downloadable. Moreover, CTR-DB supports single-dataset exploration (including differential gene expression, receiver operating characteristic curve, functional enrichment, sensitizing drug search, and tumor microenvironment analyses), and multiple-dataset combination and comparison, as well as biomarker validation function, which provide insights into the drug resistance mechanism, predictive biomarker discovery and validation, drug combination, and resistance mechanism heterogeneity.

Improving the enrollment of women and racially/ethnically diverse populations in cardiovascular clinical trials: An ASPC practice statement

Cardiovascular disease (CVD) remains the leading cause of death for both women and men worldwide. In the United States (U.S.), there are significant disparities in cardiovascular risk factors and CVD outcomes among racial and ethnic minority populations, some of whom have the highest U.S. CVD incidence and mortality. Despite this, women and racial/ethnic minority populations remain underrepresented in cardiovascular clinical trials, relative to their disease burden and population percentage. The lack of diverse participants in trials is not only a moral and ethical issue, but a scientific concern, as it can limit application of future therapies. Providing comprehensive demographic data by sex and race/ethnicity and increasing representation of diverse participants into clinical trials are essential in assessing accurate drug response, safety and efficacy information. Additionally, diversifying investigators and clinical trial staff may assist with connecting to the language, customs, and beliefs of study populations and increase recruitment of participants from diverse backgrounds. In this review, a working group for the American Society for Preventive Cardiology (ASPC) reviewed the literature regarding the inclusion of women and individuals of diverse backgrounds into cardiovascular clinical trials, focusing on prevention, and provided recommendations of best practices for improving enrollment to be more representative of the U.S. society into trials.

Establishment and characterization of immortalized human breast cancer cell lines from breast cancer patient-derived xenografts (PDX)

The application of patient-derived xenografts (PDX) in drug screening and testing is a costly and time-consuming endeavor. While cell lines permit extensive mechanistic studies, many human breast cancer cell lines lack patient characteristics and clinical treatment information. Establishing cell lines that retain patient’s genetic and drug response information would enable greater drug screening and mechanistic studies. Therefore, we utilized breast cancer PDX from the Mayo Breast Cancer Genome Guided Therapy Study (BEAUTY) to establish two immortalized, genomically unique breast cancer cell lines. Through extensive genetic and therapeutic testing, the cell lines were found to retain the same clinical subtype, major somatic alterations, and drug response phenotypes as their corresponding PDX and patient tumor. Our findings demonstrate PDX can be utilized to develop immortalized breast cancer cell lines and provide a valuable tool for understanding the molecular mechanism of drug resistance and exploring novel treatment strategies.

Facilitating Drug Discovery in Breast Cancer by Virtually Screening Patients Using In Vitro Drug Response Modeling.

Simple SummaryWhile patient datasets such as The Cancer Genome Atlas (TCGA) often contain a plethora of “-omics” data, the corresponding drug response information are limited and not suited for novel drug discovery. By integrating in vitro high throughput drug screening data and patient tumor molecular information, we created a virtual drug screening pipeline that enables drug discovery with simultaneous biomarker identification for a patient population. Using triple-negative breast cancer (TNBC) as our population of interest, we demonstrated the pipeline from lead identification, to biomarker discovery, to in vitro and in vivo validation of the compound AZD-1775.(1) Background: Drug imputation methods often aim to translate in vitro drug response to in vivo drug efficacy predictions. While commonly used in retrospective analyses, our aim is to investigate the use of drug prediction methods for the generation of novel drug discovery hypotheses. Triple-negative breast cancer (TNBC) is a severe clinical challenge in need of new therapies. (2) Methods: We used an established machine learning approach to build models of drug response based on cell line transcriptome data, which we then applied to patient tumor data to obtain predicted sensitivity scores for hundreds of drugs in over 1000 breast cancer patients. We then examined the relationships between predicted drug response and patient clinical features. (3) Results: Our analysis recapitulated several suspected vulnerabilities in TNBC and identified a number of compounds-of-interest. AZD-1775, a Wee1 inhibitor, was predicted to have preferential activity in TNBC (p < 2.2 × 10−16) and its efficacy was highly associated with TP53 mutations (p = 1.2 × 10−46). We validated these findings using independent cell line screening data and pathway analysis. Additionally, co-administration of AZD-1775 with standard-of-care paclitaxel was able to inhibit tumor growth (p < 0.05) and increase survival (p < 0.01) in a xenograft mouse model of TNBC. (4) Conclusions: Overall, this study provides a framework to turn any cancer transcriptomic dataset into a dataset for drug discovery. Using this framework, one can quickly generate meaningful drug discovery hypotheses for a cancer population of interest.

MCAM abnormal expression and clinical outcome associations are highly cancer dependent as revealed through pan-cancer analysis.

MCAM (CD146) is a cell surface adhesion molecule that has been reported to promote cancer development, progression and metastasis and is considered as a potential tumor biomarker and therapeutic target. However, inconsistent reports exist, and its clinical value is yet to be confirmed. Here we took advantage of several large genomic data collections (Genotype-Tissue Expression, The Cancer Genome Atlas and Cancer Cell Line Encyclopedia) and comprehensively analyzed MCAM expression in thousands of normal and cancer samples and cell lines along with their clinical phenotypes and drug response information. Our results show that MCAM is very highly expressed in large vessel tissues while majority of tissues have low or minimal expression. Its expression is dramatically increased in a few tumors but significantly decreased in most other tumors relative to their pairing normal tissues. Increased MCAM expression is associated with a higher tumor stage and worse patient survival for some less common tumors but not for major ones. Higher MCAM expression in primary tumors may be complicated by tumor-associated or normal stromal blood vessels yet its significance may differ from the one from cancer cells. MCAM expression is weakly associated with the response to a few small molecular drugs and the association with targeted anti-BRAF agents suggests its involvement in that pathway which warrants further investigation.

IPCT: Integrated Pharmacogenomic Platform of Human Cancer Cell Lines and Tissues

(1) Motivation: The exponential increase in multilayered data, including omics, pathways, chemicals, and experimental models, requires innovative strategies to identify new linkages between drug response information and omics features. Despite the availability of databases such as the Cancer Cell Line Encyclopedia (CCLE), the Cancer Therapeutics Response Portal (CTRP), and The Cancer Genome Atlas (TCGA), it is still challenging for biologists to explore the relationship between drug response and underlying genomic features due to the heterogeneity of the data. In light of this, the Integrated Pharmacogenomic Database of Cancer Cell Lines and Tissues (IPCT) has been developed as a user-friendly way to identify new linkages between drug responses and genomic features, as these findings can lead not only to new biological discoveries but also to new clinical trials. (2) Results: The IPCT allows biologists to compare the genomic features of sensitive cell lines or small molecules with the genomic features of tumor tissues by integrating the CTRP and CCLE databases with the REACTOME, cBioPortal, and Expression Atlas databases. The input consists of a list of small molecules, cell lines, or genes, and the output is a graph containing data entities connected with the queried input. Users can apply filters to the databases, pathways, and genes as well as select computed sensitivity values and mutation frequency scores to generate a relevant graph. Different objects are differentiated based on the background color of the nodes. Moreover, when multiple small molecules, cell lines, or genes are input, users can see their shared connections to explore the data entities common between them. Finally, users can view the resulting graphs in the online interface or download them in multiple image or graph formats. (3) Availability and Implementation: The IPCT is available as a web application with an integrated MySQL database. The web application was developed using Java and deployed on the Tomcat server. The user interface was developed using HTML5, JQuery v.3.1.0, and the Cytoscape Graph API v.1.0.4. The IPCT web and the source code are available in Sample Availability section.

Investigation of model stacking for drug sensitivity prediction

BackgroundA significant problem in precision medicine is the prediction of drug sensitivity for individual cancer cell lines. Predictive models such as Random Forests have shown promising performance while predicting from individual genomic features such as gene expressions. However, accessibility of various other forms of data types including information on multiple tested drugs necessitates the examination of designing predictive models incorporating the various data types.ResultsWe explore the predictive performance of model stacking and the effect of stacking on the predictive bias and squared error. In addition we discuss the analytical underpinnings supporting the advantages of stacking in reducing squared error and inherent bias of random forests in prediction of outliers. The framework is tested on a setup including gene expression, drug target, physical properties and drug response information for a set of drugs and cell lines.ConclusionThe performance of individual and stacked models are compared. We note that stacking models built on two heterogeneous datasets provide superior performance to stacking different models built on the same dataset. It is also noted that stacking provides a noticeable reduction in the bias of our predictors when the dominant eigenvalue of the principle axis of variation in the residuals is significantly higher than the remaining eigenvalues.

Abstract 3139: In vivo functional analyses of cancer gene variants for cancer driver identification and drug discovery

Abstract Precision oncology requires identifying and understanding of cancer genome changes in a patient tumor tissue and finding the best cancer therapy targeting the changes. Although many cancer gene targets have been validated so far, next-generation genomic profile analyses have uncovered much more cancer gene variants with unconfirmed functions. Developing methods to functionally evaluate mutations/variants and understand their roles in cancer development and drug responses, such as drug resistance or synthetic lethality, will be critical in cancer treatment decision support. In addition, in some clinical cases, multiple treatment choices such as multiple drug combinations exist. Developing cancer models which can test multiple treatments will provide direct comparison of those therapies and select the best options. At GenenDesign, we have performed drug tests on mouse “avatars”, which are also known as Patient-Derived Xenograft (PDX) models. They are personalized cancer models derived from patient tumor samples with cancer mutation profiles and drug responses very similar to the corresponding cancer patients. Drug screenings were carried out in avatars by testing chemotherapies or targeted drugs against specific cancer gene mutations and variants. Selected drugs or drug combinations from avatar studies have been applied to corresponding patients with highly consistent treatment outcome. From genomic profile analysis of our near 1500 PDX tumor models in cancer types such as lung, colorectal, gastric, liver, and esophageal, we are able to identify a large number of cancer gene mutations/variants, gene fusions, as well as gene copy number and RNA expression changes in major cancer signal pathways such as EGFR, Her2, c-Met/ALK, Ras/Raf, FGFRs, PI3K/Akt, Wnt, Notch, DNA repair, cell cycle regulation, angiogenesis. Many of these gene aberrations are potential drug targets and have been functionally tested in PDX models with approved drugs or clinical drug candidates. The mutation/variant information and drug response information generated from PDX models have been organized into our Precision Cancer Information Lab database. Patient tumor DNA test results have been used for searching genetically matched PDX models in our database. Once matched PDX models are identified, the available drug response information can be used as evidence for clinical treatment decision. In addition, the matched PDX models can also been used to test more treatment options such as different combinations and new clinical drug candidates. Citation Format: Jingjing Jiang, Tengfei Yu, Ying Yan, Wei Du, Tingting Tan, Xuqin Yang, Jiali Gu, Ling Qiu, Xin K. Ye, Zhenyu Gu. In vivo functional analyses of cancer gene variants for cancer driver identification and drug discovery [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3139. doi:10.1158/1538-7445.AM2017-3139

Abstract 755: Supporting precision cancer treatment decision with functional evaluation of cancer gene mutations and variants

Abstract Precision oncology requires identifying and understanding of cancer genome changes in a patient tumor tissue and finding the best cancer therapy targeting the changes. Although many cancer gene targets have been validated so far, next-generation genomic profile analyses have uncovered much more cancer gene variants with unconfirmed functions. Developing methods to functionally evaluate mutations/variants and understand their roles in cancer development and drug responses, such as drug resistance or synthetic lethality, will be critical in cancer treatment decision support. In addition, in some clinical cases, multiple treatment choices such as multiple drug combinations exist. Developing cancer models which can test multiple treatments will provide direct comparison of those therapies and select the best options. At GenenDesign, we have performed drug tests on mouse “avatars”, which are also known as Patient-Derived Xenograft (PDX) models. They are personalized cancer models derived from patient tumor samples with cancer mutation profiles and drug responses very similar to the corresponding cancer patients. Drug screenings were carried out in avatars by testing chemotherapies or targeted drugs against specific cancer gene mutations and variants. Selected drugs or drug combinations from avatar studies have been applied to corresponding patients with highly consistent treatment outcome. From genomic profile analysis of our near 1500 PDX tumor models in cancer types such as lung, colorectal, gastric, liver, and esophageal, we are able to identify a large number of cancer gene mutations/variants, gene fusions, as well as gene copy number and RNA expression changes in major cancer signal pathways such as EGFR, Her2, c-Met/ALK, Ras/Raf, FGFRs, PI3K/Akt, Wnt, Notch, DNA repair, cell cycle regulation, angiogenesis. Many of these gene aberrations are potential drug targets and have been functionally tested in PDX models with approved drugs or clinical drug candidates. The mutation/variant information and drug response information generated from PDX models have been organized into our Precision Cancer Information Lab database. Patient tumor DNA test results have been used for searching genetically matched PDX models in our database. Once matched PDX models are identified, the available drug response information can be used as evidence for clinical treatment decision. In addition, the matched PDX models can also been used to test more treatment options such as different combinations and new clinical drug candidates. Citation Format: Jingjing Jiang, Zhongguang Luo, Jia Wei, Guanglei Zhuang, Song Yi, Ying Yan, Tengfei Yu, Wei Du, Tingting Tan, Ling Qiu, Jiali Gu, Xin K. Ye, Jie Liu, Zhenyu Gu. Supporting precision cancer treatment decision with functional evaluation of cancer gene mutations and variants [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 755. doi:10.1158/1538-7445.AM2017-755

Abstract 2590: Analysis of APOBEC3A and APOBEC3B mutational signatures using next-generation sequencing data from cancer cell lines

Abstract The APOBEC (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like) gene family of cytidine deaminases includes evolutionarily conserved genes that play important roles in DNA repair and mRNA editing. Activity of at least two APOBEC family members, APOBEC3A and APOBEC3B, can lead to kataegis, a mutagenic process in cancer cells that generates clusters of closely spaced, single strand specific C-&amp;gt;T DNA substitutions. APOBEC mutagenesis has a characteristic signature, most commonly represented by the 5’-Tp(C-&amp;gt;T)pW-3’ sequence motif, with additional substitutions also reported. This hypermutation signature and high mRNA expression of APOBEC3A and APOBEC3B have been associated with several cancer types. Most previous studies of APOBEC signatures have examined tumor sequence data from clinical samples, for which limited or no information about drug response was available. We investigated the presence of the mutational signature and mRNA expression patterns of the APOBEC3A and APOBEC3B genes in extensively characterized cell lines, in order to identify those cell lines that carry mutations generated by kataegis, with the aim of establishing associations between the APOBEC mutational signature, individual cancer types, and the patterns of sensitivity to antitumor agents. For this purpose, we analyzed whole exome sequencing (WES) data and mRNA expression of the APOBEC3A and APOBEC3B genes in two resources with extensive drug response data: the NCI-60 cell line panel, which includes 59 human cancer cell lines representing 9 cancer types and drug response information for thousands of anticancer agents, and the Cancer Cell Line Encyclopedia (CCLE), which provides WES, whole genome, and RNA-seq sequence information on hundreds of cancer cell lines and drug response data to over 200 agents. We analyzed WES data of 325 CCLE cell lines and 59 NCI-60 cell lines, with variants identified using GATK pipeline and Varscan2 software. The variants in each cell line were filtered to remove common polymorphisms in dbSNP and 1000 Genome Project databases. We searched the discovered variants for the presence of APOBEC signatures, 5’-Tp(C-&amp;gt;AGT)pN-3’, 5’-Tp(C-&amp;gt;AGT)pD-3’, and 5’-Tp(C-&amp;gt;AGT)pW-3’ in closely spaced (1000 and 10,000 bp) windows that appeared on the same DNA strand. We will discuss the use of optimal filters for detecting APOBEC mutational signatures and will present the analyses of associations between APOBEC signatures, mutational load of the tumor cell lines, APOBEC gene expression, and chemosensitivity to treatment. These results contribute to additional characterization of available cell lines by providing information about specific mutational signatures in different categories of cancer. Our findings may assist with identifying antitumor agents that would be appropriate for treatment of cancer cells with specific signature patterns generated by APOBEC mutagenesis. Citation Format: Suleyman Vural, Julia Krushkal, Richard Simon. Analysis of APOBEC3A and APOBEC3B mutational signatures using next-generation sequencing data from cancer cell lines [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2590. doi:10.1158/1538-7445.AM2017-2590

Analysis of Chemopredictive Assay for Targeting Cancer Stem Cells in Glioblastoma Patients

Introduction: The prognosis of glioblastoma (GBM) treated with standard-of-care maximal surgical resection and concurrent adjuvant temozolomide (TMZ)/radiotherapy remains very poor (less than 15 months). GBMs have been found to contain a small population of cancer stem cells (CSCs) that contribute to tumor propagation, maintenance, and treatment resistance. The highly invasive nature of high-grade gliomas and their inherent resistance to therapy lead to very high rates of recurrence. For these reasons, not all patients with similar diagnoses respond to the same chemotherapy, schedule, or dose. Administration of ineffective anticancer therapy is not only costly but more importantly burdens the patient with unnecessary toxicity and selects for the development of resistant cancer cell clones. We have developed a drug response assay (ChemoID) that identifies the most effective chemotherapy against CSCs and bulk of tumor cells from of a panel of potential treatments, offering great promise for individualized cancer management. Providing the treating physician with drug response information on a panel of approved drugs will aid in personalized therapy selections of the most effective chemotherapy for individual patients, thereby improving outcomes. A prospective study was conducted evaluating the use of the ChemoID drug response assay in GBM patients treated with standard of care. Methods: Forty-one GBM patients (mean age 54 years, 59% male), all eligible for a surgical biopsy, were enrolled in an Institutional Review Board–approved protocol, and fresh tissue samples were collected for drug sensitivity testing. Patients were all treated with standard-of-care TMZ plus radiation with or without maximal surgery, depending on the status of the disease. Patients were prospectively monitored for tumor response, time to recurrence, progression-free survival (PFS), and overall survival (OS). Odds ratio (OR) associations of 12-month recurrence, PFS, and OS outcomes were estimated for CSC, bulk tumor, and combined assay responses for the standard-of-care TMZ treatment; sensitivities/specificities, areas under the curve (AUCs), and risk reclassification components were examined. Results: Median follow-up was 8 months (range 3-49 months). For every 5% increase in in vitro CSC cell kill by TMZ, 12-month patient response (nonrecurrence of cancer) increased two-fold, OR=2.2 (P=.016). Similar but somewhat less supported associations with the bulk tumor test were seen, OR=2.75 (P=.07) for each 5% bulk tumor cell kill by TMZ. Combining CSC and bulk tumor assay results in a single model yielded a statistically supported CSC association, OR=2.36 (P=.036), but a much attenuated remaining bulk tumor association, OR=1.46 (P=.472). AUCs and [sensitivity/specificity] at optimal outpoints (>40% CSC cell kill and >55% bulk tumor cell kill) were AUC=0.989 [sensitivity=100/specificity=97], 0.972 [100/89], and 0.989 [100/97] for the CSC only, bulk tumor only, and combined models, respectively. Risk categorization of patients was improved by 11% when using the CSC test in conjunction with the bulk test (risk reclassification nonevent net reclassification improvement [NRI] and overall NRI=0.111, P=.030). Median recurrence time was 20 months for patients with a positive (>40% cell kill) CSC test versus only 3 months for those with a negative CSC test, whereas median recurrence time was 13 months versus 4 months for patients with a positive (>55% cell kill) bulk test versus negative. Similar favorable results for the CSC test were observed for PFS and OS outcomes. Panel results across 14 potential other treatments indicated that 34/41 (83%) potentially more optimal alternative therapies may have been chosen using CSC results, whereas 27/41 (66%) alternative therapies may have been chosen using bulk tumor results. Conclusions: The ChemoID CSC drug response assay has the potential to increase the accuracy of bulk tumor assays to help guide individualized chemotherapy choices. GBM cancer recurrence may occur quickly if the CSC test has a low in vitro cell kill rate even if the bulk tumor test cell kill rate is high.

Abstract 395: Patient stratification and drug combination strategy based on drug response and genomic information from PDX clinical trials (PCTs)

Abstract Cancer is a heterogeneous disease with various molecular lesions and drug response profiles within the same tumor type. Patient stratification in clinical trials based on molecular features has contributed to recent success of several targeted cancer drugs on molecular aberrations such as BCR-ABL translocation in CML, Her2 amplification in breast and gastric cancers, EGFR mutations and ALK fusions in lung cancer and BRAF mutation in melanoma. However, in most cancers, the molecular features which can be used for patient stratification are not as simple as a single genetic aberration. Multiple drug resistance mechanisms caused by various mutations in cancer signaling pathways can also increase the uncertainty of clinical outcomes. To increase the chance of success in human clinical studies, patient-derived xenograft (PDX) clinical trials (PCTs) have increasingly been used for predictive biomarker validation, resistance mechanism investigation and combination therapy selection. PDX tumor models have been demonstrated to have high correlations with human patients in tumor pathology, molecular characteristics and drug responses. Large scale PCTs have also shown consistency in results when compared to related human clinical trials. At GenenDesign, we have established over 1000 PDX tumor models and more than 100 resistance models against various cancer drugs. Many of these PDX models have been characterized at RNA/Exome sequence, gene expression, gene copy number and hot-spot mutation levels. We carried out our PDX clinical trials by testing multiple approved drugs and clinical drug candidates such as targeted inhibitors against FGFRs, c-Met/ALK, HER2, EGFR, cell cycle regulators, Ras/Raf pathway, PI3K/Akt pathway, as well as chemotherapy drugs in biomarker-driven multi-drug multi-arm expanded PDX clinical trials. So far, we have accumulated more than 3000 efficacy data sets and associated PD samples. Analysis of drug response and associated genomic information from PDX clinical trials yielded rich information for predictive biomarker identification and validation. At the same time, many potential resistance mechanisms were also revealed. These information can make human clinical trial better prepared, more efficient and focused. More importantly, testing of a targeted drug with multiple chemotherapies in the same models can also provide guidance on future combination selection strategy. Citation Format: Jingjing Jiang, Tengfei Yu, Ying Yan, Wei Du, Tingting Tan, Xuqin Yang, Jiali Gu, Xin K. Ye, Zhenyu Gu. Patient stratification and drug combination strategy based on drug response and genomic information from PDX clinical trials (PCTs). [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 395.

Abstract 3982: Prediction and selection of cancer drug treatments using personalized tumor models or models with matching genomic profiles

Abstract The main purpose of precision medicine is to find the right drug for the right patient at the right time. Current progress in cancer drug target identification and development of new targeted drugs has presented many successful examples. However, for majority of patients, finding treatments based on the unique mutations in their own tumor cells are still hard. New approaches for precision medicine are in great needs and thus being investigated. A mouse “avatar”, also known as Patient-Derived Xenograft (PDX) model, is a personalized cancer model derived from a patient tumor sample and used to test different drugs for the patient. “GIFTS” (or Genomic Information Fitting based Therapeutics Selection) is a method with which a patient's cancer genomic information is compared to PDX model genomic profiles to find the best genomic fit and related therapeutic options in GenenDesign Drug Response Database and Genomic Information database. So far we have successfully derived more than 1000 PDX models from patient tumor tissues of various cancer types including lung, gastric, liver, esophageal, colorectal, pancreatic and many other cancers. In some cases, both patients and their avatars were tested with same cancer drugs including targeted drugs and chemotherapies. By comparing drug responses in mouse avatars with patient clinical results, we found high correlations between them in both sensitivity and unresponsiveness. We have also developed a bioinformatics algorithm to analyze PDX model genomic and drug response information. Both drug sensitivity and resistance biomarkers have been used in matching cancer patient genomic profiles to those of PDX models in GenenDesign database as a part of our GIFTS method. Our preliminary results show that there are high degree of similarities in drug response profiles between patients and their matched PDX models. Avatar and GIFTS methods can provide predictions of therapeutic effectiveness on both targeted drugs and chemotherapies. Avatar is a drug screening based method, while GIFTS is a genomic profile matching based method. Both methods are especially useful for patients, whose tumors could not be treated with known targeted drugs. Citation Format: Jingjing Jiang, Ying Yan, Zhongguang Luo, Jia He, Tengfei Yu, Wei Du, Xuqin Yang, Jiali Gu, Xin K. Ye, Guanglei Zhuang, Jie Liu, Zhenyu Gu. Prediction and selection of cancer drug treatments using personalized tumor models or models with matching genomic profiles. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3982.