Integration of pan-cancer cell line and single-cell transcriptomic profiles enables inference of therapeutic vulnerabilities in heterogeneous tumors

Single-cell RNA-sequencing (scRNA-seq) greatly advanced the understanding of intratumoral heterogeneity by identifying distinct cancer cell subpopulations. However, translating biological differences into treatment strategies is challenging due to a lack of tools to facilitate efficient drug discovery that tackles heterogeneous tumors. Developing such approaches requires accurate prediction of drug response at the single-cell level to offer therapeutic options to specific cell subpopulations. Here, we developed a transparent computational framework (nicknamed scIDUC) to predict therapeutic efficacies on an individual-cell basis by integrating single-cell transcriptomic profiles with large, data-rich pan-cancer cell line screening datasets. This method achieved high accuracy in separating cells into their correct cellular drug response statuses. In three distinct prospective tests covering different diseases (rhabdomyosarcoma, pancreatic ductal adenocarcinoma, and castration-resistant prostate cancer), the predicted results using scIDUC were accurate and mirrored biological expectations. In the first two tests, the framework identified drugs for cell subpopulations that were resistant to standard-of-care (SOC) therapies due to intrinsic resistance or tumor microenvironmental effects, and the results showed high consistency with experimental findings from the original studies. In the third test using newly generated SOC therapy resistant cell lines, scIDUC identified efficacious drugs for the resistant line, and the predictions were validated with in vitro experiments. Together, this study demonstrates the potential of scIDUC to quickly translate scRNA-seq data into drug responses for individual cells, displaying the potential as a tool to improve treatment of heterogenous tumors.

A Web Application for Predicting Drug Combination Efficacy Using Monotherapy Data and IDACombo

We recently reported a computational method (IDACombo) designed to predict the efficacy of cancer drug combinations using monotherapy response data and the assumptions of independent drug action. Given the strong agreement between IDACombo predictions and measured drug combination efficacy in vitro and in clinical trials, we believe IDACombo can be of immediate use to researchers who are working to develop novel drug combinations. While we previously released our method as an R package, we have now created an R Shiny application to allow researchers without programming experience to easily utilize this method. The app provides a graphical interface which enables users to easily generate efficacy predictions with IDACombo using provided data from several high-throughput cell line screens or using custom, user-provided data.

Simplicity: Web-Based Visualization and Analysis of High-Throughput Cancer Cell Line Screens

High-throughput drug screens are a powerful tool for cancer drug development. However, the results of such screens are often made available only as raw data, which is intractable for researchers without informatics skills, or as highly processed summary statistics, which can lack essential information for translating screening results into clinically meaningful discoveries. To improve the usability of these datasets, we developed Simplicity, a robust and user-friendly web interface for visualizing, exploring, and summarizing raw and processed data from high- throughput drug screens. Importantly, Simplicity allows for easy recalculation of summary statistics at user-defined drug concentrations. This allows Simplicity's outputs to be used with methods that rely on statistics being calculated at clinically relevant doses. Simplicity can be freely accessed at https://oncotherapyinformatics.org/simplicity/.

Inferring therapeutic vulnerability within tumors through integration of pan-cancer cell line and single-cell transcriptomic profiles

Single-cell RNA sequencing greatly advanced our understanding of intratumoral heterogeneity through identifying tumor subpopulations with distinct biologies. However, translating biological differences into treatment strategies is challenging, as we still lack tools to facilitate efficient drug discovery that tackles heterogeneous tumors. One key component of such approaches tackles accurate prediction of drug response at the single-cell level to offer therapeutic options to specific cell subpopulations. Here, we present a transparent computational framework (nicknamed scIDUC) to predict therapeutic efficacies on an individual-cell basis by integrating single-cell transcriptomic profiles with large, data-rich pan-cancer cell line screening datasets. Our method achieves high accuracy, with predicted sensitivities easily able to separate cells into their true cellular drug resistance status as measured by effect size (Cohen's d > 1.0). More importantly, we examine our method's utility with three distinct prospective tests covering different diseases (rhabdomyosarcoma, pancreatic ductal adenocarcinoma, and castration-resistant prostate cancer), and in each our predicted results are accurate and mirrored biological expectations. In the first two, we identified drugs for cell subpopulations that are resistant to standard-of-care (SOC) therapies due to intrinsic resistance or effects of tumor microenvironments. Our results showed high consistency with experimental findings from the original studies. In the third test, we generated SOC therapy resistant cell lines, used scIDUC to identify efficacious drugs for the resistant line, and validated the predictions with in-vitro experiments. Together, scIDUC quickly translates scRNA-seq data into drug response for individual cells, displaying the potential as a first-line tool for nuanced and heterogeneity-aware drug discovery.

Simplicity: web-based visualization and analysis of high-throughput cancer cell line screens

High-throughput drug screens are a powerful tool for cancer drug development. However, the results of such screens are often made available only as raw data, which is intractable for researchers without informatic skills, or as highly processed summary statistics, which can lack essential information for translating screening results into clinically meaningful discoveries. To improve the usability of these datasets, we developed Simplicity, a robust and user-friendly web interface for visualizing, exploring, and summarizing raw and processed data from high-throughput drug screens. Importantly, Simplicity allows for easy recalculation of summary statistics at user-defined drug concentrations. This allows Simplicity's outputs to be used with methods that rely on statistics being calculated at clinically relevant doses. Simplicity can be freely accessed at https://oncotherapyinformatics.org/simplicity/.

Cooperativity between H3.3K27M and PDGFRA poses multiple therapeutic vulnerabilities in human iPSC-derived diffuse midline glioma avatars

Diffuse midline glioma (DMG) is a leading cause of brain tumor death in children. In addition to hallmark H3.3K27M mutations, significant subsets also harbor alterations of other genes, such as TP53 and PDGFRA. Despite the prevalence of H3.3K27M, the results of clinical trials in DMG have been mixed, possibly due to the lack of models recapitulating its genetic heterogeneity. To address this gap, we developed human iPSC-derived tumor models harboring TP53R248Q with or without heterozygous H3.3K27M and/or PDGFRAD842V overexpression. The combination of H3.3K27M and PDGFRAD842V resulted in more proliferative tumors when gene-edited neural progenitor (NP) cells were implanted into mouse brains compared to NP with either mutation alone. Transcriptomic comparison of tumors and their NP cells of origin identified conserved JAK/STAT pathway activation across genotypes as characteristic of malignant transformation. Conversely, integrated genome-wide epigenomic and transcriptomic analyses, as well as rational pharmacologic inhibition, revealed targetable vulnerabilities unique to the TP53R248Q; H3.3K27M; PDGFRAD842V tumors and related to their aggressive growth phenotype. These include AREG-mediated cell cycle control, altered metabolism, and vulnerability to combination ONC201/trametinib treatment. Taken together, these data suggest that cooperation between H3.3K27M and PDGFRA influences tumor biology, underscoring the need for better molecular stratification in DMG clinical trials.

TB-1 ADDITIONAL GENETIC ALTERATIONS DIFFERENTIALLY ALTER THE TRANSCRIPTOMIC LANDSCAPE OF H3 K27M-MUTANT DIFFUSE MIDLINE GLIOMA

Histone H3 K27M mutation is a hallmark mutation for H3 K27M-mutant diffuse midline glioma (DMG), but targeting this mutation has yet to achieve a significant survival benefit in clinical trials. Recent analyses revealed alterations in several genes, such as NF1 and PDGFRA, are observed in substantial subpopulations of H3 K27M-mutant DMG patients in addition to H3 mutation and recurrent TP53 mutations, indicating patient-to-patient tumor heterogeneity and the potential necessity of tailored target therapy for the treatment of this disease. Here, using our human induced pluripotent stem cells (iPSC)-derived glioma avatar platform, we designed DMG models by introducing TP53R248Q with or without heterozygous H3 K27M mutation in combination with further genetic modifications of either NF1 knockout or PDGFRAD842V overexpression to recapitulate DMG subpopulations. Mice with TP53R248Q; H3F3AK27M (QM) tumors survived significantly longer than those harboring QM;NF1-/- (QMN) tumors and QM; PDGFRAD842Voe (QMP) tumors. RNA-sequencing of those induced DMG (iDMG) neurospheres revealed altered patterns of upregulation of MAPK pathway genes both in QMN and QMP-iDMG neurospheres compared to their H3 wildtype counterparts with the same combinations of genetic alterations, suggesting that those additional mutations modifies the oncogenic signaling associated with H3 K27M mutation. Further, differential expression analysis comparing QMN and QMP-iDMG neurospheres revealed 405 differentially expressed genes. Gene set enrichment analysis showed upregulation of transcriptional programs related to mesenchymal signature in QMN-iDMG neurospheres and proneural signature in QMP-iDMG neurosphere as expected. These data show that NF1 deletion and PDGFRAD842V overexpression significantly alter gene expression in H3 K27M-mutant iDMG tumors, potentially opening up a new therapeutic avenue in these devastating tumors with patient-to-patient heterogeneity. Further work using these models will shed light on the development of tailored therapy based on detailed genetic information on each patient sample, such as combining targeted kinase inhibition with HDAC inhibitors that have shown promise in the clinic.

Abstract LB067: NF1 deletion potentiates tumorigenesis and activates expression of cancer-related kinases in an iPSC-based model of H3.3K27M diffuse intrinsic pontine glioma

Diffuse intrinsic pontine glioma (DIPG) is a subset of high-grade glioma that occurs predominantly in children and has no cure. Up to 80% of DIPG harbor a heterozygous point mutation that results in a lysine 27 to methionine substitution in histone variant H3.3 (H3.3K27M). However, H3.3K27M alone is insufficient for tumorigenesis in existing DIPG models, suggesting that it interacts with co-occurring oncogenic mutations. NF1 deletion co-occurs with TP53 mutation and H3.3K27M in DIPG, but the impact of these mutations individually and together are understudied. To address this gap, we designed a model of DIPG based on human induced pluripotent stem cells (iPSC) edited via CRISPR to express either wild-type or deleted NF1 in conjunction with heterozygous H3.3K27M and a loss of function mutation in TP53. Edited iPSC were chemically differentiated into neural progenitor cells (iNPC), which upon implantation into the brainstems of immunodeficient mice formed diffusely invasive tumors that were histologically consistent with high-grade glioma. Mice with TP53 mut;H3.3K27M tumors survived significantly (p<0.05) longer than those harboring TP53 mut;NF1 -/-;H3.3K27M tumors. In vitro proliferation assays of neurospheres generated from these tumors (termed iDIPG) confirmed this result, suggesting that NF1 deletion synergizes with the other two mutations to accelerate tumor growth. To determine which transcriptional pathways could be involved in this faster tumorigenesis, we performed RNA-sequencing TP53 mut;H3.3K27M (TK) and TP53 mut;NF1 -/-;H3.3WT (TNK) iDIPG neurospheres as well as their iNPC precursors. iDIPG and iNPC clustered by NF1 status. Differential expression analysis comparing TNK and TK iDIPG neurospheres revealed 7226 differentially expressed genes (p<0.05). Gene set enrichment analysis showed upregulation of transcriptional programs related to cell cycle and kinase signaling cascades in TNK iDIPG neurospheres; therefore, we focused our analysis on differentially expressed kinases to determine whether NF1 deletion significantly impacts expression of the kinome. We found 141 kinases upregulated in TNK iDIPG neurospheres, including several members of the MAP kinase family and EGFR, which is amplified or activated in a large number of high-grade gliomas and therefore represents an attractive therapeutic target. Taken together, these data show that NF1 deletion is associated with a significant alteration of kinase expression in H3.3K27M iDIPG, potentially opening up a new therapeutic avenue in these devastating tumors. Further work using this model will focus on screening for kinases necessary for TNK iDIPG neurospheres survival in culture and investigating synergy between targeted kinase inhibition and HDAC inhibitors, which have shown promise in H3.3K27M DIPG.

Citation Format: Kasey Skinner, Tomoyuki Koga, Shunichiro Miki, Robert F. Gruener, R. Stephanie Huang, Frank Furnari, Ryan Miller. NF1 deletion potentiates tumorigenesis and activates expression of cancer-related kinases in an iPSC-based model of H3.3K27M diffuse intrinsic pontine glioma [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 LB067.

oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data

Cell line drug screening datasets can be utilized for a range of different drug discovery applications from drug biomarker discovery to building translational models of drug response. Previously, we described three separate methodologies to (1) correct for general levels of drug sensitivity to enable drug-specific biomarker discovery, (2) predict clinical drug response in patients and (3) associate these predictions with clinical features to perform in vivo drug biomarker discovery. Here, we unite and update these methodologies into one R package (oncoPredict) to facilitate the development and adoption of these tools. This new OncoPredict R package can be applied to various in vitro and in vivo contexts for drug and biomarker discovery.

Abstract 994: Investigating inhibition of nuclear export in breast cancer cells

XPO1 (Exportin aka CRM1) is a mediator of nuclear export that is known to be dysregulated in cancers. Selective inhibitors of nuclear export (SINEs) have shown some promise in cancer treatment, but the mechanism is poorly understood in cancer specific subtypes. Here, we investigated XPO1 inhibition in breast cancer by treating a panel of breast cancer cell lines with a SINE (Selinexor) and saw a bimodal response, with most triple-negative cell lines showing sensitivity and some ER+ cell lines showing resistance to the therapy. Changes in the expression of BIRC5 (Survivin) have been previously implicated in XPO1 inhibitor response. Using RT-qPCR experiments we were able to show that after 24 hours of selinexor treatment, the expression of BIRC5 does change in two sensitive cell lines MDA-MB-231 and MCF7, but is not necessarily changed at earlier time points. To more holistically explore the mechanism of XPO1 inhibition induced cell death, we performed RNA-Seq and differential gene expression analysis on these two cell lines at 2, 8, and 24 hours post treatment with Selinexor (compared to vehicle). Our results indicate that several breast cancer and cell cycle associated genes show expression changes as early as 2 hours post-treatment and showed-consistent changes across the time-points. Overall, our analysis sheds light on the mechanism by which XPO1 inhibition induces cell death in breast cancer, which may help elucidate potential biomarkers as well as to nominate rationale combinations to explore with XPO1 inhibition.

Citation Format: Robert F. Gruener, Geoffrey L. Greene, R. Stephanie Huang. Investigating inhibition of nuclear export in breast cancer cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 994.

HGG-12. HUMAN IPSC-DERIVED H3.3K27M NEUROSPHERES: A NOVEL MODEL FOR INVESTIGATING DIPG PATHOGENESIS AND DRUG RESPONSE

Diffuse intrinsic pontine glioma (DIPG) is a subset of high-grade glioma that occurs predominantly in children and has no cure. Up to 80% of DIPG harbor a heterozygous point mutation that results in a lysine 27 to methionine substitution in histone variant H3.3 (H3.3K27M). Existing DIPG models have provided insight into the role of H3.3K27M but have limitations: genetically engineered murine models often rely on overexpression of the mutant histone to form tumors; patient-derived xenografts (PDX) are more genetically faithful but preclude examination of the effect of individual mutations on pathogenesis. To address these shortcomings and better recapitulate the genetics of human tumors, we designed a novel DIPG model based on human induced pluripotent stem cells (iPSC) edited via CRISPR to express heterozygous H3.3K27M. Edited iPSC were chemically differentiated into neural progenitor cells, which upon implantation into the brainstems of immunodeficient mice formed diffusely invasive tumors that were histologically consistent with high-grade glioma. Further, neurospheres cultured from primary tumors formed secondary tumors upon reimplantation with more diffuse invasion, suggesting in vivo evolution. To validate this model’s relevance to DIPG transcriptionally, we performed RNA-sequencing on a cohort of primary and secondary tumor neurospheres (termed primary and secondary iDIPG) and compared them to published RNA-seq data from pediatric PDX and patient tumor samples. Hierarchical clustering and principal component analysis on differentially expressed genes (P<0.05) showed that H3.3K27M iDIPG cluster with H3.3K27M PDX and patient tumors. Further, ssGSEA showed that H3.3K27M iDIPG are enriched for astrocytic and mesenchymal signature genes, a defining feature of H3.3K27M DIPG. Finally, we found that primary H3.3K27M iDIPG neurospheres are sensitive to panobinostat, an HDAC inhibitor shown to be effective against H3.3K27M DIPG cells in vitro. Overall, these data suggest that H3.3K27M iDIPG are a promising tool for investigating DIPG biology and new therapeutic strategies.

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

(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.

Abstract 15: Imputing a more targeted therapy for triple-negative breast cancers

Triple-negative breast cancer (TNBC) is a clinically defined subset of breast cancers that are inherently difficult to treat since, unlike the other breast cancer subtypes, they are defined by their lack of a distinct molecular target. Previously our lab developed a computational framework to impute drug response in patients that accurately mirrored the observed patient response. In short, this methodology involves building unsupervised linear regression models between cell line baseline gene expression values and the respective drug sensitivity metric (EC50 or AUC), which allows us to generate imputed sensitivity scores for cell line or patient tumor samples for which we have gene expression data. For this study, we applied this methodology to prospectively identify agents that are predicted to be more effective specifically in triple-negative breast cancer. We began by expanding the drug imputation training dataset to the Broad’s Cell Therapeutics Response Portal, which contains 545 drugs and over 1,000 cancer cell lines (CCLs) and imputed sensitivity in The Cancer Genome Atlas’s (TCGA’s) breast cancer cohort. Our imputation data indicated that TNBC tumors were expected to be comparatively sensitive to cell cycle pathway inhibitors. Among these compounds, the Wee1 inhibitor MK1775 (aka adavosertib and AZD1775) was identified for having the most significant predicted effect in TNBC patients compared to the receptor-positive patients. We were able to validate the accuracy of the MK1775 imputation model using traditional cross-validation approaches as well as in independent cell line dataset and showed that the measured and our predicted response to Wee1 inhibition strongly correlated. Additionally, we were able to uses our predicted MK1775 response scores to recapitulate the significant association between MK1775 response and p53 mutation status. We then looked to validate the efficacy of MK1775’s efficacy in TNBC cell line and xenograft experiments. We observed strong inhibition of cancer growth by MK1775 alone but also identified a combination effect between MK1775 and the TNBC standard-of-care agent paclitaxel as assayed by both in vitro viability experiments as well as in vivo tumor growth assays. We believe these results indicate that this methodology could generally be applied as a hypothesis-generating tool for identifying targeted agents for other disease subtypes as well as supporting the use of MK1775, especially in combination with paclitaxel, for TNBC.

Citation Format: Robert F. Gruener, Geoffrey L. Greene, R. Stephanie Huang. Imputing a more targeted therapy for triple-negative breast cancers [abstract]. In: Proceedings of the AACR Special Conference on Advancing Precision Medicine Drug Development: Incorporation of Real-World Data and Other Novel Strategies; Jan 9-12, 2020; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(12_Suppl_1):Abstract nr 15.

More than fishing for a cure: The promises and pitfalls of high throughput cancer cell line screens

High-throughput screens in cancer cell lines (CCLs) have been used for decades to help researchers identify compounds with the potential to improve the treatment of cancer and, more recently, to identify genomic susceptibilities in cancer via genome-wide shRNA and CRISPR/Cas9 screens. Additionally, rich genomic and transcriptomic data of these CCLs has allowed researchers to pair this screening data with biological features, enabling efforts to identify biomarkers of treatment response and gene dependencies. In this paper, we review the major CCL screening efforts and the large datasets these screens have made available. We also assess the CCL screens collectively and include a resource with harmonized CCL and compound identifiers to facilitate comparisons across screens. The CCLs in these screens were found to represent a wide range of cancer types, with a strong correlation between the representation of a cancer type and its associated mortality. Patient ages and gender distributions of CCLs were generally as expected, with some notable exceptions of female underrepresentation in certain disease types. Also, ethnicity information, while largely incomplete, suggests that African American and Hispanic patients may be severely underrepresented in these screens. Nearly all genes were targeted in the genetic perturbations screens, but the compounds used for the drug screens target less than half of known cancer drivers, likely reflecting known limitations in our drug design capabilities. Finally, we discuss recent developments in the field and the promise they hold for enabling future screens to overcome previous limitations and lead to new breakthroughs in cancer treatment.

Discovering novel pharmacogenomic biomarkers by imputing drug response in cancer patients from large genomics studies

Obtaining accurate drug response data in large cohorts of cancer patients is very challenging; thus, most cancer pharmacogenomics discovery is conducted in preclinical studies, typically using cell lines and mouse models. However, these platforms suffer from serious limitations, including small sample sizes. Here, we have developed a novel computational method that allows us to impute drug response in very large clinical cancer genomics data sets, such as The Cancer Genome Atlas (TCGA). The approach works by creating statistical models relating gene expression to drug response in large panels of cancer cell lines and applying these models to tumor gene expression data in the clinical data sets (e.g., TCGA). This yields an imputed drug response for every drug in each patient. These imputed drug response data are then associated with somatic genetic variants measured in the clinical cohort, such as copy number changes or mutations in protein coding genes. These analyses recapitulated drug associations for known clinically actionable somatic genetic alterations and identified new predictive biomarkers for existing drugs.