Recruitment of FBXO22 for targeted degradation of NSD2.

Targeted protein degradation (TPD) is an emerging therapeutic strategy that would benefit from new chemical entities with which to recruit a wider variety of ubiquitin E3 ligases to target proteins for proteasomal degradation. Here we describe a TPD strategy involving the recruitment of FBXO22 to induce degradation of the histone methyltransferase and oncogene NSD2. UNC8732 facilitates FBXO22-mediated degradation of NSD2 in acute lymphoblastic leukemia cells harboring the NSD2 gain-of-function mutation p.E1099K, resulting in growth suppression, apoptosis and reversal of drug resistance. The primary amine of UNC8732 is metabolized to an aldehyde species, which engages C326 of FBXO22 to recruit the SCFFBXO22 Cullin complex. We further demonstrate that a previously reported alkyl amine-containing degrader targeting XIAP is similarly dependent on SCFFBXO22. Overall, we present a potent NSD2 degrader for the exploration of NSD2 disease phenotypes and a new FBXO22-recruitment strategy for TPD.

MYC Protein Interactome Profiling Reveals Functionally Distinct Regions that Cooperate to Drive Tumorigenesis.

Transforming members of the MYC family (MYC, MYCL1, and MYCN) encode transcription factors containing six highly conserved regions, termed MYC homology boxes (MBs). By conducting proteomic profiling of the MB interactomes, we demonstrate that half of the MYC interactors require one or more MBs for binding. Comprehensive phenotypic analyses reveal that two MBs, MB0 and MBII, are universally required for transformation. MBII mediates interactions with acetyltransferase-containing complexes, enabling histone acetylation, and is essential for MYC-dependent tumor initiation. By contrast, MB0 mediates interactions with transcription elongation factors via direct binding to the general transcription factor TFIIF. MB0 is dispensable for tumor initiation but is a major accelerator of tumor growth. Notably, the full transforming activity of MYC can be restored by co-expression of the non-transforming MB0 and MBII deletion proteins, indicating that these two regions confer separate molecular functions, both of which are required for oncogenic MYC activity.

SLC3A2 N-glycosylation and Golgi remodeling regulate SLC7A amino acid exchangers and stress mitigation.

Proteostasis requires oxidative metabolism (ATP) and mitigation of the associated damage by glutathione, in an increasingly dysfunctional relationship with aging. SLC3A2 (4F2hc, CD98) plays a role as a disulfide-linked adaptor to the SLC7A5 and SLC7A11 exchangers which import essential amino acids and cystine while exporting Gln and Glu, respectively. The positions of N-glycosylation sites on SLC3A2 have evolved with the emergence of primates, presumably in synchrony with metabolism. Herein, we report that each of the four sites in SLC3A2 has distinct profiles of Golgi-modified N-glycans. N-glycans at the primate-derived site N381 stabilized SLC3A2 in the galectin-3 lattice against coated-pit endocytosis, while N365, the site nearest the membrane promoted glycolipid-galectin-3 (GL-Lect)-driven endocytosis. Our results indicate that surface retention and endocytosis are precisely balanced by the number, position, and remodeling of N-glycans on SLC3A2. Furthermore, proteomics and functional assays revealed an N-glycan-dependent clustering of the SLC3A2∗SLC7A5 heterodimer with amino-acid/Na+ symporters (SLC1A4, SLC1A5) that balances branched-chain amino acids and Gln levels, at the expense of ATP to maintain the Na+/K+ gradient. In replete conditions, SLC3A2 interactions require Golgi-modified N-glycans at N365D and N381D, whereas reducing N-glycosylation in the endoplasmic reticulum by fluvastatin treatment promoted the recruitment of CD44 and transporters needed to mitigate stress. Thus, SLC3A2 N-glycosylation and Golgi remodeling of the N-glycans have distinct roles in amino acids import for growth, maintenance, and metabolic stresses.

The MYC oncoprotein directly interacts with its chromatin cofactor PNUTS to recruit PP1 phosphatase.

Despite MYC dysregulation in most human cancers, strategies to target this potent oncogenic driver remain an urgent unmet need. Recent evidence shows the PP1 phosphatase and its regulatory subunit PNUTS control MYC phosphorylation, chromatin occupancy, and stability, however the molecular basis remains unclear. Here we demonstrate that MYC interacts directly with PNUTS through the MYC homology Box 0 (MB0), a highly conserved region recently shown to be important for MYC oncogenic activity. By NMR we identified a distinct peptide motif within MB0 that interacts with PNUTS residues 1-148, a functional unit, here termed PNUTS amino-terminal domain (PAD). Using NMR spectroscopy we determined the solution structure of PAD, and characterised its MYC-binding patch. Point mutations of residues at the MYC-PNUTS interface significantly weaken their interaction both in vitro and in vivo, leading to elevated MYC phosphorylation. These data demonstrate that the MB0 region of MYC directly interacts with the PAD of PNUTS, which provides new insight into the control mechanisms of MYC as a regulator of gene transcription and a pervasive cancer driver.

Rapid 3D phenotypic analysis of neurons and organoids using data-driven cell segmentation-free machine learning.

Phenotypic profiling of large three-dimensional microscopy data sets has not been widely adopted due to the challenges posed by cell segmentation and feature selection. The computational demands of automated processing further limit analysis of hard-to-segment images such as of neurons and organoids. Here we describe a comprehensive shallow-learning framework for automated quantitative phenotyping of three-dimensional (3D) image data using unsupervised data-driven voxel-based feature learning, which enables computationally facile classification, clustering and advanced data visualization. We demonstrate the analysis potential on complex 3D images by investigating the phenotypic alterations of: neurons in response to apoptosis-inducing treatments and morphogenesis for oncogene-expressing human mammary gland acinar organoids. Our novel implementation of image analysis algorithms called Phindr3D allowed rapid implementation of data-driven voxel-based feature learning into 3D high content analysis (HCA) operations and constitutes a major practical advance as the computed assignments represent the biology while preserving the heterogeneity of the underlying data. Phindr3D is provided as Matlab code and as a stand-alone program (https://github.com/DWALab/Phindr3D).

Quantitative Prostate MRI Analysis Following Fluvastatin Therapy for Localized Prostate Cancer - A Pilot Study.

PURPOSE: To assess the role of multi-parametric MRI (mpMRI) in assessment of tumor response to fluvastatin administered prior to radical prostatectomy. METHODS: Men with MRI-visible, clinically significant prostate cancer and due to be treated with radical prostatectomy were prospectively enrolled. mpMRI was performed at baseline and following 6-7 week of neoadjuvant oral statin therapy (40 mg fluvastatin, twice daily), prior to prostatectomy. MRI assessment included tumor size, T2 relaxation time, ADC value, K-trans (volume transfer constant), Kep (reflux constant), and Ve (fractional volume) parameters at the 2 time points. Initial prostate needle biopsy cores, prior to starting oral statin therapy, corresponding to site of tumor on radical prostatectomy specimens were selected for analysis. The effect of fluvastatin on tumor proliferation (marker Ki67) and on tumor cell apoptosis (marker cleaved Caspase-3, CC3) were analyzed and correlated with MRI findings. RESULTS: Nine men with paired MRI studies were included in the study. Binary histopathological data was available for 6 of the participants. No significant change in tumor size (P = 0.898), T2 relaxation time (P = 0.213), ADC value (P = 0.455), K-trans (P = 0.613), Kep (P = 0.547) or Ve (P = 0.883) between the time of biopsy and prostatectomy were observed. No significant change in tumor proliferation (%Ki67-positive cells, P = 0.766) was observed by immunohistochemistry analysis. However, there was a significant increase in tumor cell apoptosis (%CC3-positive cells, P = 0.047). CONCLUSION: mpMRI techniques may not be sufficiently sensitive to detect the types (or magnitude) of tumor cell changes observed following 6-7 weeks of fluvastatin therapy for prostate cancer.

Image-Based Analysis of Protein Stability.

Short half-life proteins regulate many essential processes, including cell cycle, transcription, and apoptosis. However, few well-characterized protein-turnover pathways have been identified because traditional methods to measure protein half-life are time and labor intensive. To overcome this barrier, we developed a protein stability probe and high-content screening pipeline for novel regulators of short half-life proteins using automated image analysis. Our pilot probe consists of the short half-life protein c-MYC (MYC) fused to Venus fluorescent protein (MYC-Venus). This probe enables protein half-life to be scored as a function of fluorescence intensity and distribution. Rapid turnover prevents maximal fluorescence of the probe due to the relatively longer maturation time of the fluorescent protein. Cells expressing the MYC-Venus probe were analyzed using a pipeline in which automated confocal microscopy and image analyses were used to score MYC-Venus stability by two strategies: assaying the percentage of cells with Venus fluorescence above background, and phenotypic comparative analysis. To evaluate this high-content screening pipeline and our probe, a kinase inhibitor library was screened by confocal microscopy to identify known and novel kinases that regulate MYC stability. Compounds identified were shown to increase the half-life of both MYC-Venus and endogenous MYC, validating the probe and pipeline. Fusion of another short half-life protein, myeloid cell leukemia 1 (MCL1), with Venus also demonstrated an increase in percent Venus-positive cells after treatment with inhibitors known to stabilize MCL1. Together, the results validate the use of our automated microscopy and image analysis pipeline of stability probe-expressing cells to rapidly and quantitatively identify regulators of short half-life proteins. © 2019 The Authors. Cytometry Part A published by Wiley Periodicals, Inc. on behalf of International Society for Advancement of Cytometry.

AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress.

Mitochondrial respiration is a crucial component of cellular metabolism that can become dysregulated in cancer. Compared with normal hematopoietic cells, acute myeloid leukemia (AML) cells and patient samples have higher mitochondrial mass, without a concomitant increase in respiratory chain complex activity. Hence these cells have a lower spare reserve capacity in the respiratory chain and are more susceptible to oxidative stress. We therefore tested the effects of increasing the electron flux through the respiratory chain as a strategy to induce oxidative stress and cell death preferentially in AML cells. Treatment with the fatty acid palmitate induced oxidative stress and cell death in AML cells, and it suppressed tumor burden in leukemic cell lines and primary patient sample xenografts in the absence of overt toxicity to normal cells and organs. These data highlight a unique metabolic vulnerability in AML, and identify a new therapeutic strategy that targets abnormal oxidative metabolism in this malignancy.

Myc and its interactors take shape.

The Myc oncoprotein is a key contributor to the development of many human cancers. As such, understanding its molecular activities and biological functions has been a field of active research since its discovery more than three decades ago. Genome-wide studies have revealed Myc to be a global regulator of gene expression. The identification of its DNA-binding partner protein, Max, launched an area of extensive research into both the protein-protein interactions and protein structure of Myc. In this review, we highlight key insights with respect to Myc interactors and protein structure that contribute to the understanding of Myc's roles in transcriptional regulation and cancer. Structural analyses of Myc show many critical regions with transient structures that mediate protein interactions and biological functions. Interactors, such as Max, TRRAP, and PTEF-b, provide mechanistic insight into Myc's transcriptional activities, while others, such as ubiquitin ligases, regulate the Myc protein itself. It is appreciated that Myc possesses a large interactome, yet the functional relevance of many interactors remains unknown. Here, we discuss future research trends that embrace advances in genome-wide and proteome-wide approaches to systematically elucidate mechanisms of Myc action. This article is part of a Special Issue entitled: Myc proteins in cell biology and pathology.

Involvement of Toso in activation of monocytes, macrophages, and granulocytes.

Rapid activation of immune responses is necessary for antibacterial defense, but excessive immune activation can result in life-threatening septic shock. Understanding how these processes are balanced may provide novel therapeutic potential in treating inflammatory disease. Fc receptors are crucial for innate immune activation. However, the role of the putative Fc receptor for IgM, known as Toso/Faim3, has to this point been unclear. In this study, we generated Toso-deficient mice and used them to uncover a critical regulatory function of Toso in innate immune activation. Development of innate immune cells was intact in the absence of Toso, but Toso-deficient neutrophils exhibited more reactive oxygen species production and reduced phagocytosis of pathogens compared with controls. Cytokine production was also decreased in Toso(-/-) mice compared with WT animals, rendering them resistant to septic shock induced by lipopolysaccharide. However, Toso(-/-) mice also displayed limited cytokine production after infection with the bacterium Listeria monocytogenes that was correlated with elevated presence of Listeria throughout the body. Accordingly, Toso(-/-) mice succumbed to infections of L. monocytogenes, whereas WT mice successfully eliminated the infection. Taken together, our data reveal Toso to be a unique regulator of innate immune responses during bacterial infection and septic shock.

Identifying molecular features that distinguish fluvastatin-sensitive breast tumor cells.

Statins, routinely used to treat hypercholesterolemia, selectively induce apoptosis in some tumor cells by inhibiting the mevalonate pathway. Recent clinical studies suggest that a subset of breast tumors is particularly susceptible to lipophilic statins, such as fluvastatin. To quickly advance statins as effective anticancer agents for breast cancer treatment, it is critical to identify the molecular features defining this sensitive subset. We have therefore characterized fluvastatin sensitivity by MTT assay in a panel of 19 breast cell lines that reflect the molecular diversity of breast cancer, and have evaluated the association of sensitivity with several clinicopathological and molecular features. A wide range of fluvastatin sensitivity was observed across breast tumor cell lines, with fluvastatin triggering cell death in a subset of sensitive cell lines. Fluvastatin sensitivity was associated with an estrogen receptor alpha (ERα)-negative, basal-like tumor subtype, features that can be scored with routine and/or strong preclinical diagnostics. To ascertain additional candidate sensitivity-associated molecular features, we mined publicly available gene expression datasets, identifying genes encoding regulators of mevalonate production, non-sterol lipid homeostasis, and global cellular metabolism, including the oncogene MYC. Further exploration of this data allowed us to generate a 10-gene mRNA abundance signature predictive of fluvastatin sensitivity, which showed preliminary validation in an independent set of breast tumor cell lines. Here, we have therefore identified several candidate predictors of sensitivity to fluvastatin treatment in breast cancer, which warrant further preclinical and clinical evaluation.

The role of ligand density and size in mediating quantum dot nuclear transport.

Studying the effects of the physicochemical properties of nanomaterials on cellular uptake, toxicity, and exocytosis can provide the foundation for designing safer and more effective nanoparticles for clinical applications. However, an understanding of the effects of these properties on subcellular transport, accumulation, and distribution remains limited. The present study investigates the effects of surface density and particle size of semiconductor quantum dots on cellular uptake as well as nuclear transport kinetics, retention, and accumulation. The current work illustrates that cellular uptake and nuclear accumulation of nanoparticles depend on surface density of the nuclear localization signal (NLS) peptides with nuclear transport reaching a plateau at 20% surface NLS density in as little as 30 min. These intracellular nanoparticles have no effects on cell viability up to 72 h post treatment. These findings will set a foundation for engineering more sophisticated nanoparticle systems for imaging and manipulating genetic targets in the nucleus.

Transient structure and dynamics in the disordered c-Myc transactivation domain affect Bin1 binding.

The crucial role of Myc as an oncoprotein and as a key regulator of cell growth makes it essential to understand the molecular basis of Myc function. The N-terminal region of c-Myc coordinates a wealth of protein interactions involved in transformation, differentiation and apoptosis. We have characterized in detail the intrinsically disordered properties of Myc-1-88, where hierarchical phosphorylation of S62 and T58 regulates activation and destruction of the Myc protein. By nuclear magnetic resonance (NMR) chemical shift analysis, relaxation measurements and NOE analysis, we show that although Myc occupies a very heterogeneous conformational space, we find transiently structured regions in residues 22-33 and in the Myc homology box I (MBI; residues 45-65); both these regions are conserved in other members of the Myc family. Binding of Bin1 to Myc-1-88 as assayed by NMR and surface plasmon resonance (SPR) revealed primary binding to the S62 region in a dynamically disordered and multivalent complex, accompanied by population shifts leading to altered intramolecular conformational dynamics. These findings expand the increasingly recognized concept of intrinsically disordered regions mediating transient interactions to Myc, a key transcriptional regulator of major medical importance, and have important implications for further understanding its multifaceted role in gene regulation.

Sterol-like drugs potentiate statin-triggered prostate cancer cell death by inhibiting SREBP2 nuclear translocation.

There is an urgent need to provide immediate and effective options for the treatment of prostate cancer (PCa) to prevent progression to lethal castration-resistant PCa (CRPC). The mevalonate (MVA) pathway is dysregulated in PCa, and statin drugs commonly prescribed for hypercholesterolemia, effectively target this pathway. Statins exhibit anti-PCa activity, however the resulting intracellular depletion of cholesterol triggers a feedback loop that restores MVA pathway activity, thus diminishing statin efficacy and contributing to resistance. To identify drugs that block this feedback response and enhance the pro-apoptotic activity of statins, we performed a high-content image-based screen of a 1508 drug library, enriched for FDA-approved compounds. Two of the validated hits, Galeterone (GAL) and Quinestrol, share the cholesterol-related tetracyclic structure, which is also evident in the FDA-approved CRPC drug Abiraterone (ABI). Molecular modeling revealed that GAL, Quinestrol and ABI not only share structural similarity with 25-hydroxy-cholesterol (25HC) but were also predicted to bind similarly to a known protein-binding site of 25HC. This suggested GAL, Quinestrol and ABI are sterol-mimetics and thereby inhibit the statin-induced feedback response. Cell-based assays demonstrated that these agents inhibit nuclear translocation of sterol-regulatory element binding protein 2 (SREBP2) and the transcription of MVA genes. Sensitivity was independent of androgen status and the Fluva-GAL combination significantly impeded CRPC tumor xenograft growth. By identifying cholesterol-mimetic drugs that inhibit SREBP2 activation upon statin treatment, we provide a potent "one-two punch" against CRPC progression and pave the way for innovative therapeutic strategies to combat additional diseases whose etiology is associated with SREBP2 dysregulation.

Dysregulation of the mevalonate pathway promotes transformation.

The importance of cancer metabolism has been appreciated for many years, but the intricacies of how metabolic pathways interconnect with oncogenic signaling are not fully understood. With a clear understanding of how metabolism contributes to tumorigenesis, we will be better able to integrate the targeting of these fundamental biochemical pathways into patient care. The mevalonate (MVA) pathway, paced by its rate-limiting enzyme, hydroxymethylglutaryl coenzyme A reductase (HMGCR), is required for the generation of several fundamental end-products including cholesterol and isoprenoids. Despite years of extensive research from the perspective of cardiovascular disease, the contribution of a dysregulated MVA pathway to human cancer remains largely unexplored. We address this issue directly by showing that dysregulation of the MVA pathway, achieved by ectopic expression of either full-length HMGCR or its novel splice variant, promotes transformation. Ectopic HMGCR accentuates growth of transformed and nontransformed cells under anchorage-independent conditions or as xenografts in immunocompromised mice and, importantly, cooperates with RAS to drive the transformation of primary mouse embryonic fibroblasts cells. We further explore whether the MVA pathway may play a role in the etiology of human cancers and show that high mRNA levels of HMGCR and additional MVA pathway genes correlate with poor prognosis in a meta-analysis of six microarray datasets of primary breast cancer. Taken together, our results suggest that HMGCR is a candidate metabolic oncogene and provide a molecular rationale for further exploring the statin family of HMGCR inhibitors as anticancer agents.

Topoisomerase 1 Inhibition in MYC-Driven Cancer Promotes Aberrant R-Loop Accumulation to Induce Synthetic Lethality.

MYC is a central regulator of gene transcription and is frequently dysregulated in human cancers. As targeting MYC directly is challenging, an alternative strategy is to identify specific proteins or processes required for MYC to function as a potent cancer driver that can be targeted to result in synthetic lethality. To identify potential targets in MYC-driven cancers, we performed a genome-wide CRISPR knockout screen using an isogenic pair of breast cancer cell lines in which MYC dysregulation is the switch from benign to transformed tumor growth. Proteins that regulate R-loops were identified as a potential class of synthetic lethal targets. Dysregulated MYC elevated global transcription and coincident R-loop accumulation. Topoisomerase 1 (TOP1), a regulator of R-loops by DNA topology, was validated to be a vulnerability in cells with high MYC activity. Genetic knockdown of TOP1 in MYC-transformed cells resulted in reduced colony formation compared with control cells, demonstrating synthetic lethality. Overexpression of RNaseH1, a riboendonuclease that specifically degrades R-loops, rescued the reduction in clonogenicity induced by TOP1 deficiency, demonstrating that this vulnerability is driven by aberrant R-loop accumulation. Genetic and pharmacologic TOP1 inhibition selectively reduced the fitness of MYC-transformed tumors in vivo. Finally, drug response to TOP1 inhibitors (i.e., topotecan) significantly correlated with MYC levels and activity across panels of breast cancer cell lines and patient-derived organoids. Together, these results highlight TOP1 as a promising target for MYC-driven cancers. SIGNIFICANCE: CRISPR screening reveals topoisomerase 1 as an immediately actionable vulnerability in cancers harboring MYC as a driver oncoprotein that can be targeted with clinically approved inhibitors.

Exploiting the mevalonate pathway to distinguish statin-sensitive multiple myeloma.

Statin inhibitors, used to control hypercholesterolemia, trigger apoptosis of hematologic tumor cells and therefore have immediate potential as anticancer agents. Evaluations of statins in acute myelogenous leukemia and multiple myeloma have shown that statin efficacy is mixed, with only a subset of tumor cells being highly responsive. Our goal was to distinguish molecular features of statin-sensitive and -insensitive myeloma cells and gain insight into potential predictive markers. We show that dysregulation of the mevalonate pathway is a key determinant of sensitivity to statin-induced apoptosis in multiple myeloma. In sensitive cells, the classic feedback response to statin exposure is lost. This results in deficient up-regulation of 2 isoforms of hydroxymethylglutaryl coenzyme A reductase: the rate-limiting enzyme of the mevalonate pathway and hydroxymethylglutaryl coenzyme A synthase 1. To ascertain the clinical utility of these findings, we demonstrate that a subset of primary myeloma cells is sensitive to statins and that monitoring dysregulation of the mevalonate pathway may distinguish these cancers. We also show statins are highly effective and well tolerated in an orthotopic model of myeloma using cells harboring this dysregulation. This determinant of sensitivity further provides molecular rationale for the significant therapeutic index of statins on these tumor cells.

Prognostic gene signatures for non-small-cell lung cancer.

Resectable non-small-cell lung cancer (NSCLC) patients have poor prognosis, with 30-50% relapsing within 5 years. Current staging criteria do not fully capture the complexity of this disease. Survival could be improved by identification of those early-stage patients who are most likely to benefit from adjuvant therapy. Molecular classification by using mRNA expression profiles has led to multiple, poorly overlapping signatures. We hypothesized that differing statistical methodologies contribute to this lack of overlap. To test this hypothesis, we analyzed our previously published quantitative RT-PCR dataset with a semisupervised method. A 6-gene signature was identified and validated in 4 independent public microarray datasets that represent a range of tumor histologies and stages. This result demonstrated that at least 2 prognostic signatures can be derived from this single dataset. We next estimated the total number of prognostic signatures in this dataset with a 10-million-signature permutation study. Our 6-gene signature was among the top 0.02% of signatures with maximum verifiability, reaffirming its efficacy. Importantly, this analysis identified 1,789 unique signatures, implying that our dataset contains >500,000 verifiable prognostic signatures for NSCLC. This result appears to rationalize the observed lack of overlap among reported NSCLC prognostic signatures.

Statins and prostate cancer-hype or hope? The biological perspective.

Growing evidence suggests that men prescribed a statin for cholesterol control have a lower risk of advanced prostate cancer (PCa) and improved treatment outcomes; however, the mechanism by which statins elicit their anti-neoplastic effects is not well understood and is likely multifaceted. Statins are potent and specific inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR), the rate-limiting enzyme of the mevalonate (MVA) metabolic pathway. This two-part series is a review of the observational and experimental data on statins as anti-cancer agents in PCa. In this article, we describe the functional role that deregulated MVA metabolism plays in PCa progression and summarize the biological evidence and rationale for targeting the MVA pathway, with statins and other agents, for the treatment of PCa.

Bimodal Gene Expression in Patients with Cancer Provides Interpretable Biomarkers for Drug Sensitivity.

Identifying biomarkers predictive of cancer cell response to drug treatment constitutes one of the main challenges in precision oncology. Recent large-scale cancer pharmacogenomic studies have opened new avenues of research to develop predictive biomarkers by profiling thousands of human cancer cell lines at the molecular level and screening them with hundreds of approved drugs and experimental chemical compounds. Many studies have leveraged these data to build predictive models of response using various statistical and machine learning methods. However, a common pitfall to these methods is the lack of interpretability as to how they make predictions, hindering the clinical translation of these models. To alleviate this issue, we used the recent logic modeling approach to develop a new machine learning pipeline that explores the space of bimodally expressed genes in multiple large in vitro pharmacogenomic studies and builds multivariate, nonlinear, yet interpretable logic-based models predictive of drug response. The performance of this approach was showcased in a compendium of the three largest in vitro pharmacogenomic datasets to build robust and interpretable models for 101 drugs that span 17 drug classes with high validation rates in independent datasets. These results along with in vivo and clinical validation support a better translation of gene expression biomarkers between model systems using bimodal gene expression. SIGNIFICANCE: A new machine learning pipeline exploits the bimodality of gene expression to provide a reliable set of candidate predictive biomarkers with a high potential for clinical translatability.