Drug Response Profiles Research Articles

From Proteomic Mapping to Invasion-Metastasis-Cascade Systemic Biomarkering and Targeted Drugging of Mutant BRAF-Dependent Human Cutaneous Melanomagenesis.

Simple SummaryDespite the recent advances in human malignancy therapy, metastasis and chemoresistance remain the principal causes of cancer-derived deaths. Given the fatal forms of cutaneous metastatic melanoma, we herein employed primary (WM115) and metastatic (WM266-4) melanoma cells, both obtained from the same patient, to identify novel biomarkers and therapeutic agents. Through state-of-the-art technologies including deep proteome landscaping, immunofluorescence phenotyping, and drug toxicity screening, we were able to describe new molecular programs, oncogenic drivers, and drug regimens, controlling the invasion-metastasis cascade during BRAFV600D-dependent melanomagenesis. It proved that proteomic navigation could foster the development of systemic biomarkering and targeted drugging for successful treatment of advanced disease.Melanoma is classified among the most notoriously aggressive human cancers. Despite the recent progress, due to its propensity for metastasis and resistance to therapy, novel biomarkers and oncogenic molecular drivers need to be promptly identified for metastatic melanoma. Hence, by employing nano liquid chromatography-tandem mass spectrometry deep proteomics technology, advanced bioinformatics algorithms, immunofluorescence, western blotting, wound healing protocols, molecular modeling programs, and MTT assays, we comparatively examined the respective proteomic contents of WM115 primary (n = 3955 proteins) and WM266-4 metastatic (n = 6681 proteins) melanoma cells. It proved that WM115 and WM266-4 cells have engaged hybrid epithelial-to-mesenchymal transition/mesenchymal-to-epithelial transition states, with TGF-β controlling their motility in vitro. They are characterized by different signatures of SOX-dependent neural crest-like stemness and distinct architectures of the cytoskeleton network. Multiple signaling pathways have already been activated from the primary melanoma stage, whereas HIF1α, the major hypoxia-inducible factor, can be exclusively observed in metastatic melanoma cells. Invasion-metastasis cascade-specific sub-routines of activated Caspase-3-triggered apoptosis and LC3B-II-dependent constitutive autophagy were also unveiled. Importantly, WM115 and WM266-4 cells exhibited diverse drug response profiles, with epirubicin holding considerable promise as a beneficial drug for metastatic melanoma clinical management. It is the proteome navigation that enables systemic biomarkering and targeted drugging to open new therapeutic windows for advanced disease.

Abstract PO-052: Exploring patient derived xenografts based pharmacogenomic data for precision oncology

Abstract Preclinical cancer models play a vital role in oncology research and precision medicine. Patient-derived tumor xenografts (PDXs) are used as reliable preclinical models for studying tumor biology and for testing anti-cancer therapies that are tailored according to genomic characteristics of tumors. Several academic groups, research institutes, and commercial organizations are generating and distributing PDX models. However the distributed nature of PDX model generation and lack of central repository make it challenging to find PDX models with specific characteristics. Furthermore this also hinders meta-analysis (across datasets) of PDX pharmacogenomic data. International consortia and catalogs of PDX models such as PDXNet, EurOPDX and PDXFinder are being developed to standardize PDX associated metadata and facilitate material sharing. Recently we have developed Xenograft Visualization & Analysis (Xeva), an open-source software package in R programming language. Xeva allows PDX growth curve visualization, different response metrics computation and biomarker discovery. Extending to this we have developed XevaDB, a database of PDX drug response and genomic profiles. XevaDB is the first resource to allow concurrent visualization of drug response and associated molecular data such as mutation and copy number alterations. Furthermore XevaDB enables exploration of the tumor growth curve of a PDX model, along with corresponding control. XevaDB contains PDXs from >600 individual patients, spanning across nine different tissue types and >70 drugs. Using XevaDB, we have performed meta-analysis of PDX pharmacogenomic data and have identified 90 pathways significantly associated with response to 53 drugs (FDR < 5%). Our results show that activity of the EGFR signaling pathway is significantly associated with erlotinib response in lung cancer PDXs. We have also found that in PDXs, response to binimetinib is associated with the MAP kinase activation pathway. XevaDB provides a comprehensive resource to search and explore PDX pharmacogenomic data. By combining drug response with genomic data of PDXs, XevaDB allows researchers to quickly find the model of interest and access the data to answer their biological questions. As PDXs based pharmacogenomic datasets continue to expand, XevaDB will facilitate easy access and analysis of this valuable data by the scientific community. Citation Format: Arvind Singh Mer, Benjamin Haibe-Kains. Exploring patient derived xenografts based pharmacogenomic data for precision oncology [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-052.

Evaluating and comparing immunostaining and computational methods for spatial profiling of drug response in patient-derived explants

AbstractPatient-derived explants (PDEs) represent the direct culture of fragments of freshly-resected tumour tissue under conditions that retain the original architecture of the tumour. PDEs have advantages over other preclinical cancer models as platforms for predicting patient-relevant drug responses in that they preserve the tumour microenvironment and tumour heterogeneity. At endpoint, PDEs may either be processed for generation of histological sections or homogenised and processed for ʻomic' evaluation of biomarker expression. A significant advantage of spatial profiling is the ability to co-register drug responses with tumour pathology, tumour heterogeneity and changes in the tumour microenvironment. Spatial profiling of PDEs relies on the utilisation of robust immunostaining approaches for validated biomarkers and incorporation of appropriate image analysis methods to quantitatively and qualitatively monitor changes in biomarker expression in response to anti-cancer drugs. Automation of immunostaining and image analysis would provide a significant advantage for the drug discovery pipeline and therefore, here, we have sought to optimise digital pathology approaches. We compare three image analysis software platforms (QuPath, ImmunoRatio and VisioPharm) for evaluating Ki67 as a marker for proliferation, cleaved PARP (cPARP) as a marker for apoptosis and pan-cytokeratin (CK) as a marker for tumour areas and find that all three generate comparable data to the views of a histomorphometrist. We also show that Virtual Double Staining of sequential sections by immunohistochemistry results in imperfect section alignment such that CK-stained tumour areas are over-estimated. Finally, we demonstrate that multi-immunofluorescence combined with digital image analysis is a superior method for monitoring multiple biomarkers simultaneously in tumour and stromal areas in PDEs.This paper reports the optimisation of digital pathology approaches for evaluating drug responses in patient-derived explants at endpoint. The authors compare three image analysis software platforms and find that all three generate comparable data to that of a histomorphometrist. They also demonstrate the power of multi-immunofluorescence staining for monitoring multiple biomarkers simultaneously compared to virtual double staining of sequential sections by immunohistochemistry.

Abstract PS17-52: A preclinical platform of breast cancer PDX-derived cell lines as a tool for pharmacological screening and functional studies

Abstract Despite considerable progress in understanding the biology and genetics of breast cancer progression, the development of effective therapies need physiological and predictive preclinical models. In this context, breast cancer patient-derived xenograft (PDX) models has become a standard tool as they reproduce accurately the behavior of tumor of origin, in term of histological and molecular phenotype and response to chemotherapy. Although PDXs in vivo models are indispensable for preclinical studies, they suffer from some limitations due to study costs related to tumor maintenance on mice, variable engraftment rate, growth delay and limited throughput for large-scale drug screening.To address this problem and propose a time and cost effective preclinical screening tool, we developed a panel of breast cancer PDX-derived low-passage 2D cell lines as a convenient in vitro pre-screening platform to profile compound activity.30 different breast cancer PDX models including TNBC, HER2+ and ER+ were tested for their capacity to generate cell lines maintaining the characteristics of the parental PDX tumor and usable for in vitro assays.Today, we succeeded with a series of 14 PDX models.Tumor cells isolated from PDX tumor tissue were cultured under different media and matrix conditions, allowing at least 5 passages in culture. A Short Tandem Repeat (STR) comparison profile was done with the parental PDX before performing a master bank. We succeeded in establishing a panel of 14 PDX-derived cellular models (14/30 = 46% success rate).We performed short term 2D cytotoxicity assays and long term colony assays to compare cell lines in vitro drug sensitivity with their parental PDX in vivo drug response and overall, the results show that this panel reproduced the drug response profile of the original PDXs with chemotherapies, PARP inhibitors, an ADC (T-DM1) therapies.Moreover, cellular models engrafted back onto mice showed in vivo response to chemotherapies similar to that of the parental PDX confirming the identical behavior of cell line / PDX couples.As the use of cellular models is still considered as a standard for early preclinical test to evaluate drug response before moving to in vivo assays, our breast cancer PDX-derived cell line platform appeared to be a robust and relevant tool. Furthermore, since the main concern when using in vitro models is the representativeness of the results obtained when transposed to in vivo models, the similarities between cell lines and parental PDX should maximize success of further in vivo preclinical drug development. Citation Format: Stefano Cairo, Olivier Deas, Sophie Banis, Kathleen Flosseau, Enora Le Ven, Jean-Gabriel Judde. A preclinical platform of breast cancer PDX-derived cell lines as a tool for pharmacological screening and functional studies [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS17-52.

Abstract PS10-06: Comparative analysis of anti-proliferative effects and gene profiling of lapatinib, neratinib, and tucatinib

Abstract Introduction: Human epidermal growth factor 2 (HER2/ERBB2) is frequently amplified or mutated across various cancer types. The tyrosine kinase inhibitors (TKIs) lapatinib, neratinib, and tucatinib are FDA-approved for the treatment of HER2-positive breast cancer. All three TKIs bind and inhibit the kinase domain of HER2 but differ both in the mechanism of binding and in specificity for other HER family members. Direct comparisons to differentiate the pre-clinical efficacy of the three TKIs have been limited to small-scale studies and novel biomarkers of response to further define appropriate patient populations are required. Methods: In this study, the anti-proliferative effects of the three TKIs were compared using a 115-cancer cell line panel, including 12 breast cancer cell lines and 22 cell lines harbouring point mutations or amplifications of EGFR, HER2, or HER3. Hierarchical clustering analysis was carried out to compare the IC50 “fingerprint” of the three TKIs to 168 other anti-cancer agents. Novel markers of TKI sensitivity and resistance were identified through cross-analysis of each drug response profile with mutation, copy number variation, and gene expression data. Results: All three TKIs were effective against HER2-positive breast cancer models; neratinib showed the most potent activity, followed by tucatinib and lapatinib respectively (Table 1). Neratinib displayed the greatest anti-proliferative activity in HER2-mutant and EGFR-mutant cell lines. Clustering analysis revealed that the anti-proliferative profile of tucatinib was most similar to trastuzumab, while neratinib and lapatinib were most like other HER family inhibitors. Mutation and gene expression analysis identified potential markers of response for each TKI. High expression of four genes (HER2, VTCN1, CDK12, and RAC1) correlated with response to all three TKIs. DNA damage repair genes were significantly associated with resistance to the HER2-targeted TKIs. BRCA2 mutation was correlated with neratinib and tucatinib response, and high gene expression of ATM, BRCA2, and BRCA1 were all associated with neratinib resistance. Conclusions: Neratinib was the most effective HER2-targeted TKI against HER2-amplified, -mutant, and EGFR-mutant cell lines. This analysis revealed possible mechanisms that may be exploited using combinatorial strategies involving CDK inhibitors, immunotherapies, and targeting DNA repair pathways. Table: IC50 values for neratinib, lapatinib, and tucatinib in the HER2+ breast cancer cell linesIC50 values (nM)Cell linesNeratinibLapatinibTucatinibAU-56520294125BT-4745926229HCC195413814262122MDA-MB-453306228445928SKBR3715222 Citation Format: Neil T Conlon, Jeffrey J Kooijman, Suzanne JC van Gerwen, Winfried R Mulder, Guido JR Zaman, Irmina Diala, Lisa D Eli, Alshad S Lalani, John Crown, Denis Collins. Comparative analysis of anti-proliferative effects and gene profiling of lapatinib, neratinib, and tucatinib [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS10-06.

Comparative analysis of drug response and gene profiling of HER2-targeted tyrosine kinase inhibitors

BackgroundHuman epidermal growth factor 2 (HER2/ERBB2) is frequently amplified/mutated in cancer. The tyrosine kinase inhibitors (TKIs) lapatinib, neratinib, and tucatinib are FDA-approved for the treatment of HER2-positive breast cancer. Direct comparisons of the preclinical efficacy of the TKIs have been limited to small-scale studies. Novel biomarkers are required to define beneficial patient populations.MethodsIn this study, the anti-proliferative effects of the three TKIs were directly compared using a 115 cancer cell line panel. Novel TKI response/resistance markers were identified through cross-analysis of drug response profiles with mutation, gene copy number and expression data.ResultsAll three TKIs were effective against HER2-amplified breast cancer models; neratinib showing the most potent activity, followed by tucatinib then lapatinib. Neratinib displayed the greatest activity in HER2-mutant and EGFR-mutant cells. High expression of HER2, VTCN1, CDK12, and RAC1 correlated with response to all three TKIs. DNA damage repair genes were associated with TKI resistance. BRCA2 mutations were correlated with neratinib and tucatinib response, and high expression of ATM, BRCA2, and BRCA1 were associated with neratinib resistance.ConclusionsNeratinib was the most effective HER2-targeted TKI against HER2-amplified, -mutant, and EGFR-mutant cell lines. This analysis revealed novel resistance mechanisms that may be exploited using combinatorial strategies.

Genetic characterization of B-cell prolymphocytic leukemia: a prognostic model involving MYC and TP53

AbstractB-cell prolymphocytic leukemia (B-PLL) is a rare hematological disorder whose underlying oncogenic mechanisms are poorly understood. Our cytogenetic and molecular assessments of 34 patients with B-PLL revealed several disease-specific features and potential therapeutic targets. The karyotype was complex (≥3 abnormalities) in 73% of the patients and highly complex (≥5 abnormalities) in 45%. The most frequent chromosomal aberrations were translocations involving MYC [t(MYC)] (62%), deletion (del)17p (38%), trisomy (tri)18 (30%), del13q (29%), tri3 (24%), tri12 (24%), and del8p (23%). Twenty-six (76%) of the 34 patients exhibited an MYC aberration, resulting from mutually exclusive translocations or gains. Whole-exome sequencing revealed frequent mutations in TP53, MYD88, BCOR, MYC, SF3B1, SETD2, CHD2, CXCR4, and BCLAF1. The majority of B-PLL used the IGHV3 or IGHV4 subgroups (89%) and displayed significantly mutated IGHV genes (79%). We identified 3 distinct cytogenetic risk groups: low risk (no MYC aberration), intermediate risk (MYC aberration but no del17p), and high risk (MYC aberration and del17p) (P = .0006). In vitro drug response profiling revealed that the combination of a B-cell receptor or BCL2 inhibitor with OTX015 (a bromodomain and extra-terminal motif inhibitor targeting MYC) was associated with significantly lower viability of B-PLL cells harboring a t(MYC). We concluded that cytogenetic analysis is a useful diagnostic and prognostic tool in B-PLL. Targeting MYC may be a useful treatment option in this disease.

Creating In Vitro Three-Dimensional Tumor Models: A Guide for the Biofabrication of a Primary Osteosarcoma Model.

Osteosarcoma (OS) is a highly aggressive primary bone tumor. The mainstay for its treatment is multiagent chemotherapy and surgical resection, with a 50-70% 5-year survival rate. Despite the huge effort made by clinicians and researchers in the past 30 years, limited progress has been made to improve patient outcomes. As novel therapeutic approaches for OS become available, such as monoclonal antibodies, small molecules, and immunotherapies, the need for OS preclinical model development becomes equally pressing. Three-dimensional (3D) OS models represent an alternative system to study this tumor: In contrast to two-dimensional monolayers, 3D matrices can recapitulate key elements of the tumor microenvironment (TME), such as the cellular interaction with the bone mineralized matrix. The advancement of tissue engineering and biofabrication techniques enables the incorporation of specific TME aspects into 3D models, to investigate the contribution of individual components to tumor progression and enhance understanding of basic OS biology. The use of biomaterials that mimic the extracellular matrix could also facilitate the testing of drugs targeting the TME itself, allowing a larger range of therapeutics to be tested, while averting the ethical implications and high cost associated with in vivo preclinical models. This review aims at serving as a practical guide by delineating the OS TME ("what it is like") and, in turn, propose various biofabrication strategies to create a 3D model ("how to recreate it"), to improve the in vitro representation of the OS tumor and ultimately generate more accurate drug response profiles.

Quantitative Structure\u2013Mutation\u2013Activity Relationship Tests (QSMART) model for protein kinase inhibitor response prediction

BackgroundProtein kinases are a large family of druggable proteins that are genomically and proteomically altered in many human cancers. Kinase-targeted drugs are emerging as promising avenues for personalized medicine because of the differential response shown by altered kinases to drug treatment in patients and cell-based assays. However, an incomplete understanding of the relationships connecting genome, proteome and drug sensitivity profiles present a major bottleneck in targeting kinases for personalized medicine.ResultsIn this study, we propose a multi-component Quantitative Structure–Mutation–Activity Relationship Tests (QSMART) model and neural networks framework for providing explainable models of protein kinase inhibition and drug response (hbox {IC}_{50}) profiles in cell lines. Using non-small cell lung cancer as a case study, we show that interaction terms that capture associations between drugs, pathways, and mutant kinases quantitatively contribute to the response of two EGFR inhibitors (afatinib and lapatinib). In particular, protein–protein interactions associated with the JNK apoptotic pathway, associations between lung development and axon extension, and interaction terms connecting drug substructures and the volume/charge of mutant residues at specific structural locations contribute significantly to the observed hbox {IC}_{50} values in cell-based assays.ConclusionsBy integrating multi-omics data in the QSMART model, we not only predict drug responses in cancer cell lines with high accuracy but also identify features and explainable interaction terms contributing to the accuracy. Although we have tested our multi-component explainable framework on protein kinase inhibitors, it can be extended across the proteome to investigate the complex relationships connecting genotypes and drug sensitivity profiles.

INNV-16. CLINICAL APPLICABILITY OF INDIVIDUALIZED DRUG RESPONSE PROFILING UTILIZING EX-VIVO TISSUE-DERIVED 3D CELL CULTURE ASSAYS IN HIGH-GRADE GLIOMA: A SINGLE INSTITUTION CASE SERIES USING 3D-PREDICT RESULTS

Abstract Recurrent high-grade glioma is a challenging disease process, without consensus on effective second-line therapy options. Individualized, patient-specific, biologically-based data is desirable in driving therapeutic decision-making. Patients with recurrent high-grade glioma and planned surgical re-resection at our institution were prospectively enrolled into the 3D-PREDICT study. Tissue was collected at the time of surgery for ex vivo 3D cell culture assays comprising a panel of agents commonly used for high-grade glioma, including chemotherapies and targeted therapies used in other solid cancers. In all cases, therapeutic agent selection was guided by the neuro-oncologist’s clinical judgement, factoring the patient’s age, performance status, comorbidities, toxicities/side effect profile of potential agents, and drug accessibility, plus ex-vivo drug response RESULTS: We present 3 cases in which the selection of agents was influenced by the tissue-derived 3D cell culture results; treatment led to clinical response observed in terms of progression free survival, quality of life, and pharmacologic tolerability. In Case 1, a patient with recurrent anaplastic astrocytoma was treated with a BRAF inhibitor for 12 months with excellent tolerability and no radiographic progression. Case 2 demonstrates the use of combination bevacizumab and irinotecan after disease progression subsequent to standard treatment. This patient had local radiographic control for 7 months, tolerating the regimen well. In Case 3, an individual with recurrent glioblastoma was treated with combination carboplatin and etoposide based on assay response prediction to both agents; treatment has been tolerated well with radiographic stability at 6 months while maintaining good performance status. This case series represents our institutional experience of utilizing patient-specific, ex-vivo tissue-derived cell drug response profiling to guide choice of therapy for recurrent high-grade glioma patients. Using individualized, tumor-specific drug sensitivity data to guide these decisions is representative of the ongoing paradigm shift into the realm of individualized medicine to improve outcomes in cancer patients.

In Vitro Drug Response Profiling in BCP- and T-ALL Primary Samples Adds a Robust Functional Layer Enabling Optimized Guidance of Individualized Therapy in Relapsed and Refractory Pediatric Acute Leukemia Patients

In Vitro Drug Response Profiling in BCP- and T-ALL Primary Samples Adds a Robust Functional Layer Enabling Optimized Guidance of Individualized Therapy in Relapsed and Refractory Pediatric Acute Leukemia Patients

Metabolic Drug Survey Highlights Cancer Cell Dependencies and Vulnerabilities

Here we assembled a novel metabolic drug library covering 243 compounds allowing to systematically identify metabolic dependencies in high-throughput phenotypic screens. The CeMM Library of Metabolic Drugs (CLIMET) was compiled in a stepwise fashion, starting from 8000 candidate compounds, after a survey of public drug-target databases, and ending with 243 highly-curated compounds including extensive crosschecking for approval status, structural information, compound's potency and selectivity for the intended target, pathway/target redundancy, and commercial availability. To assess the potency of the compounds in CLIMET, we screened the full collection against a panel of 15 diverse myeloid leukemia cell lines. Each compound was tested for its effect on cell growth and survival in a 10,000-fold concentration range (1nM to 10uM) enabling the generation of dose response curves and calculation of area under the curve values (AUC) for each drug. We, further, functionally grouped the cell lines and drugs based on metabolic drug efficacy patterns and associated them with distinct genomic and metabolic attributes. Analysis of the metabolic drug response profiles revealed that 77 compounds (32%) affected cell viability with the top effective compounds targeting nucleotide metabolism, oxidative stress, and the PI3K/mTOR pathway. Unsupervised hierarchical clustering of the drug sensitivity profiles stratified the cell lines in 5 functional taxonomic groups, with the activity of 19 compounds significantly contributing to the cell line grouping (e.g. PF-02545920, GW 4064, mTOR inhibitors, daporinad). Comparison of the oxygen consumption rate and extracellular acidification rate showed that the examined cell lines have analogous baseline metabolic phenotypes, suggesting that the mitochondrial function of the cells as assessed by Seahorse analysis did not significantly influence the clustering. Genotype to phenotype associations were identified between FLT3mutations and sensitivity to 5-FU, lestaurtinib, and PF-02545920. Moreover, RAS mutations negatively correlated to mTOR and mitochondrial respiration inhibitor sensitivity, whereas TP53 mutations conferred a resistance phenotype to PI3K pathway inhibitors and antineoplastic agents. Selective sensitivities were detected to the lactate transporter (SLC16A1) inhibitor AZD3965, the PI3K inhibitor pictilisib, and the fatty acid synthase inhibitor GSK2194069, which could be explained by varied gene expression in sensitive cell lines and target/process dependency. CLIMET allows for identification of metabolic susceptibilities, grouping of cancer cells based on metabolic dependencies, as well as understanding of context-dependent mechanism of action of drugs. Functional drug testing may provide a rapid and robust approach to identify metabolic vulnerabilities, responding patients, and prioritize compounds for clinical evaluation as illustrated with our study. Disclosures Staber: Janssen: Consultancy, Honoraria; AbbVie: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Astra Zeneca: Consultancy, Honoraria; Roche: Consultancy, Honoraria, Research Funding; Celgene/ BMS: Consultancy, Honoraria; Gilead: Consultancy, Honoraria; msd: Consultancy, Honoraria.

Proteogenomic Subtyping of Chronic Lymphocytic Leukemia Identifies a Novel Poor Outcome Subgroup with a Distinct Drug Response Profile

Proteogenomic Subtyping of Chronic Lymphocytic Leukemia Identifies a Novel Poor Outcome Subgroup with a Distinct Drug Response Profile

Identifying Drug Sensitivity Subnetworks with NETPHIX

SummaryPhenotypic heterogeneity in cancer is often caused by different patterns of genetic alterations. Understanding such phenotype-genotype relationships is fundamental for the advance of personalized medicine. We develop a computational method, named NETPHIX (NETwork-to-PHenotype association with eXclusivity) to identify subnetworks of genes whose genetic alterations are associated with drug response or other continuous cancer phenotypes. Leveraging interaction information among genes and properties of cancer mutations such as mutual exclusivity, we formulate the problem as an integer linear program and solve it optimally to obtain a subnetwork of associated genes. Applied to a large-scale drug screening dataset, NETPHIX uncovered gene modules significantly associated with drug responses. Utilizing interaction information, NETPHIX modules are functionally coherent and can thus provide important insights into drug action. In addition, we show that modules identified by NETPHIX together with their association patterns can be leveraged to suggest drug combinations.

Tissue dynamics spectroscopic imaging: functional imaging of heterogeneous cancer tissue.

.Significance: Tumor heterogeneity poses a challenge for the chemotherapeutic treatment of cancer. Tissue dynamics spectroscopy captures dynamic contrast and can capture the response of living tissue to applied therapeutics, but the current analysis averages over the complicated spatial response of living biopsy samples.Aim: To develop tissue dynamics spectroscopic imaging (TDSI) to map the heterogeneous spatial response of tumor tissue to anticancer drugs.Approach: TDSI is applied to tumor spheroids grown from cell lines and to ex vivo living esophageal biopsy samples. Doppler fluctuation spectroscopy is performed on a voxel basis to extract spatial maps of biodynamic biomarkers. Functional images and bivariate spatial maps are produced using a bivariate color merge to represent the spatial distribution of pairs of signed drug-response biodynamic biomarkers.Results: We have mapped the spatial variability of drug responses within biopsies and have tracked sample-to-sample variability. Sample heterogeneity observed in the biodynamic maps is associated with histological heterogeneity observed using inverted selective-plane illumination microscopy.Conclusion: We have demonstrated the utility of TDSI as a functional imaging method to measure tumor heterogeneity and its potential for use in drug-response profiling.

Abstract 3911: High-throughput screening of patient-derived HCC organoids for drug discovery

Abstract Current in vitro models of primary liver cancer insufficiently recapitulates the tumor heterogeneity and pathophysiology of the original patient's tumor. This significantly hinders translational studies and often leads to the high rate of clinical trial attrition and poor patient response. Recent technological advances in organoid culture methods have been reported to overcome these limitations. Our group has successfully established 3 patient-derived HCC tumor organoids that closely recapitulate the patient's pathophysiology. Conventional short-term monolayer culture of the same 3 tumors as well as 3 non-tumoral adjacent liver and 3 normal liver organoids were established and compared with 5 HCC cell lines from ATCC for their drug response profiles. Subsequent drug screening was performed using a panel of 100 drugs, of which 70 are FDA-approved for either HCC or other cancers, as well as 30 pipeline compounds targeting frequently mutated pathways in HCC. We found that our screening pipeline could achieve robust z' scores and reproducible data points with technical and biological replicates. Principal component analysis of Total Area under curve (AUC) values show distinct and progressive changes between Normal Liver, Non-tumoral adjacent liver, and Tumor organoids. Upon comparison with HCC cell lines Hep3B, HepG2, Huh7, SNU398 and PLC5, we found HCC patient-derived short term cultures, particularly organoids have more diverse drug sensitivity profiles and may be a better indicator of patient heterogeneity. Interestingly, FDA-approved drugs for HCC such as Regorafenib show high sensitivity and tumor specificity to ATCC cell lines (ATCC vs Normal liver P=0.00017***), but this specificity was reduced and lost in patient-derived 2D cultures (P=0.045*) and organoids (P=0.895), respectively. This may explain the observed hepatotoxicity observed in HCC patients and warrant further investigation of more targeted ways of treating HCC. Citation Format: Alissa Michelle Go Wong, Winnie Cheung, Maggie Wong, Aikha Wong, Helena Lam, Zhe Zhang, Matthew Man, Hui Chen, David Close, Paul Johnston, Anthony Chan, Paul Lai, Joseph Kwong, Nathalie Wong. High-throughput screening of patient-derived HCC organoids for drug discovery [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3911.

Abstract 3929: Patient-derived pancreatic cancer organoids as in-vitro cancer models for personalizing therapy

Abstract Pancreatic ductal adenocarcinoma (PDAC) is the most aggressive solid tumor. Unfortunately, patients with this disease have not benefited from the advances in precision medicine seen in other malignancies. The lack of actionable mutations as well as predictive biomarkers has hampered progress in this area. Combination therapy remains the cornerstone of systemic treatment, with FOLFIRINOX being the most effective therapeutic regimen in PDAC. However, it is clear that only subset of patients are responsive, and the toxicities of this 4-drug combination are intolerable to many patients particularly in the adjuvant setting. There is a clear need for more personalized rational drug combinations that are optimized to maximize efficacy while minimizing toxicity. We have established a protocol for the generation of PDAC organoids from endoscopic ultrasound biopsies for personalizing systemic therapy based on the principle of phenotype driven oncology. Organoids generated are analyzed for KRAS and p53 mutations to confirm their tumor origin. By applying a phenotypic optimization platform (QPOP) on these organoids, we are able to identify optimized drug combinations from a pool of both FDA-approved and preclinical drugs. As expected, FOLFIRINOX did not always rank high by treatment efficacy suggesting that alternative treatment may be more beneficial for these patients. In addition, we identified that the efficacy of FOLFIRINOX was driven by two-drug synergistic combinations which varied between patients. This can potentially advise modifications to this regimen when toxicity arises. From a discovery perspective, our platform has the capability of deriving novel drug combinations. The combination of gemcitabine and Abraxane with a WEE1-inhibitor was found to be synergistic in selected PDAC patients. Furthermore, patients with similar drug response profiles can be studied together to identify predictive markers of sensitivity. To demonstrate that this platform can indeed be applied to prospectively recommend treatment strategies for patients with progressive disease on first-line chemotherapy, we are establishing a clinical trial to personalize second-line therapy, with the endpoint of overall response rate. Overall, by combining patient-derived organoids with rational optimized drug combinations, we aim to personalize therapy for patients with PDAC to improve overall survival. Citation Format: Claire Alexandra Zhen Chew, Li Min Tay, Lissa Hooi, Edward Kai-Hua Chow, Glenn K. Bonney, Shing Leng Chan. Patient-derived pancreatic cancer organoids as in-vitro cancer models for personalizing therapy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3929.

Dissecting intratumour heterogeneity of nodal B-cell lymphomas at the transcriptional, genetic and drug-response levels.

Tumour heterogeneity encompasses both the malignant cells and their microenvironment. While heterogeneity between individual patients is known to affect the efficacy of cancer therapy, most personalized treatment approaches do not account for intratumour heterogeneity. We addressed this issue by studying the heterogeneity of nodal B-cell lymphomas by single-cell RNA-sequencing and transcriptome-informed flow cytometry. We identified transcriptionally distinct malignant subpopulations and compared their drug-response and genomic profiles. Malignant subpopulations from the same patient responded strikingly differently to anti-cancer drugs ex vivo, which recapitulated subpopulation-specific drug sensitivity during in vivo treatment. Infiltrating T cells represented the majority of non-malignant cells, whose gene-expression signatures were similar across all donors, whereas the frequencies of T-cell subsets varied significantly between the donors. Our data provide insights into the heterogeneity of nodal B-cell lymphomas and highlight the relevance of intratumour heterogeneity for personalized cancer therapy.

Abstract GS2-02: Acquired activating mutations in RTKs confer endocrine resistance in ER+ metastatic breast cancer through ER-reprogramming, MAPK signaling, and an induced stem-like cell state

While recent studies have begun characterizing the metastatic breast cancer (MBC) genomics, our understanding of mechanisms of acquired endocrine-resistance, their induced cell-state, and their altered drug-response profile, remains lacking. We collected biopsies from patients with MBC with detailed clinicopathologic features. To date we profiled 520 exomes, and 291 transcriptomes, with 126 patients having multiple biopsy exomes. Curated endocrine-relapse set include 60 patients with pre-treatment and post-relapse exomes. Characterization of candidate mechanisms of resistance (MOR) included 909 RNA-seq profiles of T47D cells with introduced MOR under various drugs. In the acquired endocrine resistance cohort, as expected, we found frequent ESR1 acquired mutations (13 pt, 22%). Additionally, we identified acquired activating SNVs and amplifications in oncogenic receptor tyrosine kinases (RTKs) in 18/60 (30%), including EGF family - HER2 (n=7), ERRB3 (R525Q), and EGFR (S116F), and FGF family - FGFR1 (n=5), FGFR2 (n=4), and FGF3 amplicon (n=4). RNA-seq of T47D cells overexpressing HER2 activating mutations revealed a distinct cell-state (HER2-ACT). Similarly, FGFR activation revealed a district FGFR-ACT state. The transcriptional signatures of these HER2 and FGFR states were remarkably similar (odd-ratio= 91, p= 2.64E-94). To characterize the cell-state common to HER and FGFR, we defined RTK-ACT with 358 overlapping marker genes (see table). Canonical (estradiol) ER signaling is slightly elevated in RTK-ACT, however this state is strongly associated with growth-factor driven ER signaling, suggesting reprogramming of ER from AF2, to AF1 signaling. Consistent with this, RTK-ACT had significantly higher MAPK activation. Additionally, RTK-ACT Induced stronger similarity to Basal-state, enrichment in motility/migration, mesenchymal, and stem-like features, with top genes including CDH3, MBP7, and S100, ETV, DUSP, SPRY families (see table). To study RTK-driven state in-vivo, we analyzed RNA-seq from our MBC biopsies, and compared tumors with activating RTK mutations (n=38) with WT (n=118), and inferred activated RTK in-vivo (RTK.ACT.iv). Our in-vitro and in-vivo states show significant overlap (OR=4.71, p=9.42E-20). Furthermore, characterization of RTK.ACT.iv, recapitulated all the cell-state features observed in RTK.ACT - including higher growth-factors ER signaling, MAPK, and basal-like state (see table). We further studied the viability and transcription of these RTKs under various drugs (12 treatments), including fulvestrant, palbociclib, and specific tyrosine kinase inhibitors (TKIs) - Neratinib (pan-HER inhibitor) and FIIN-3 (FGFR-i). We found that HER2-ACT and FGFR-ACT signatures remain robust when treated with fulvestrant, palbociclib, and their combinations, as compared to TKIs suppression (see table). in-line with our viability results - suggesting intrinsic resistance to CDK4/6i and sensitivity to TKIs. This study demonstrated that activating RTKs constitute of prevalent modality of acquired resistance to endocrine therapies, inducing a distinct state with clinical implications - suggesting the potential benefit of combination therapies with specific TKIs over CDK4/6i. The common MAPK activity and our preliminary results - suggests the potential of convergence-node targeting strategy with added MEK or SHP2 inhibition. Citation Format: Ofir Cohen, Pingping Mao, Utthara Nayar, Jorge E Buendia-Bue, Dewey Kim, Esha Jain, Karla Helvie, Daniel Abravanel, Kailey J Kowalski, Christian Kapstad, Samuel Freeman, Victor Adalsteinsson, Seth A Wander, Adrienne G Waks, Gad Getz, Aviv Regev, Eric Winer P Winer, Nancy U Nancy U. Lin, Nikhil Wagle. Acquired activating mutations in RTKs confer endocrine resistance in ER+ metastatic breast cancer through ER-reprogramming, MAPK signaling, and an induced stem-like cell state [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr GS2-02.

Tissue-guided LASSO for prediction of clinical drug response using preclinical samples.

Prediction of clinical drug response (CDR) of cancer patients, based on their clinical and molecular profiles obtained prior to administration of the drug, can play a significant role in individualized medicine. Machine learning models have the potential to address this issue but training them requires data from a large number of patients treated with each drug, limiting their feasibility. While large databases of drug response and molecular profiles of preclinical in-vitro cancer cell lines (CCLs) exist for many drugs, it is unclear whether preclinical samples can be used to predict CDR of real patients. We designed a systematic approach to evaluate how well different algorithms, trained on gene expression and drug response of CCLs, can predict CDR of patients. Using data from two large databases, we evaluated various linear and non-linear algorithms, some of which utilized information on gene interactions. Then, we developed a new algorithm called TG-LASSO that explicitly integrates information on samples’ tissue of origin with gene expression profiles to improve prediction performance. Our results showed that regularized regression methods provide better prediction performance. However, including the network information or common methods of including information on the tissue of origin did not improve the results. On the other hand, TG-LASSO improved the predictions and distinguished resistant and sensitive patients for 7 out of 13 drugs. Additionally, TG-LASSO identified genes associated with the drug response, including known targets and pathways involved in the drugs’ mechanism of action. Moreover, genes identified by TG-LASSO for multiple drugs in a tissue were associated with patient survival. In summary, our analysis suggests that preclinical samples can be used to predict CDR of patients and identify biomarkers of drug sensitivity and survival.