Prediction of whole-cell transcriptional response with machine learning.

MOTIVATION: Applications in synthetic and systems biology can benefit from measuring whole-cell response to biochemical perturbations. Execution of experiments to cover all possible combinations of perturbations is infeasible. In this paper, we present the host response model (HRM), a machine learning approach that maps response of single perturbations to transcriptional response of the combination of perturbations. RESULTS: The HRM combines high-throughput sequencing with machine learning to infer links between experimental context, prior knowledge of cell regulatory networks, and RNASeq data to predict a gene's dysregulation. We find that the HRM can predict the directionality of dysregulation to a combination of inducers with an accuracy of >90% using data from single inducers. We further find that the use of prior, known cell regulatory networks doubles the predictive performance of the HRM (an R2 from 0.3 to 0.65). The model was validated in two organisms, Escherichia coli and Bacillus subtilis, using new experiments conducted after training. Finally, while the HRM is trained with gene expression data, the direct prediction of differential expression makes it possible to also conduct enrichment analyses using its predictions. We show that the HRM can accurately classify >95% of the pathway regulations. The HRM reduces the number of RNASeq experiments needed as responses can be tested in silico prior to the experiment. AVAILABILITY AND IMPLEMENTATION: The HRM software and tutorial are available at https://github.com/sd2e/CDM and the configurable differential expression analysis tools and tutorials are available at https://github.com/SD2E/omics_tools. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Transcriptional profiling identifies an androgen receptor activity-low, stemness program associated with enzalutamide resistance.

The androgen receptor (AR) antagonist enzalutamide is one of the principal treatments for men with castration-resistant prostate cancer (CRPC). However, not all patients respond, and resistance mechanisms are largely unknown. We hypothesized that genomic and transcriptional features from metastatic CRPC biopsies prior to treatment would be predictive of de novo treatment resistance. To this end, we conducted a phase II trial of enzalutamide treatment (160 mg/d) in 36 men with metastatic CRPC. Thirty-four patients were evaluable for the primary end point of a prostate-specific antigen (PSA)50 response (PSA decline ≥50% at 12 wk vs. baseline). Nine patients were classified as nonresponders (PSA decline <50%), and 25 patients were classified as responders (PSA decline ≥50%). Failure to achieve a PSA50 was associated with shorter progression-free survival, time on treatment, and overall survival, demonstrating PSA50's utility. Targeted DNA-sequencing was performed on 26 of 36 biopsies, and RNA-sequencing was performed on 25 of 36 biopsies that contained sufficient material. Using computational methods, we measured AR transcriptional function and performed gene set enrichment analysis (GSEA) to identify pathways whose activity state correlated with de novo resistance. TP53 gene alterations were more common in nonresponders, although this did not reach statistical significance (P = 0.055). AR gene alterations and AR expression were similar between groups. Importantly, however, transcriptional measurements demonstrated that specific gene sets-including those linked to low AR transcriptional activity and a stemness program-were activated in nonresponders. Our results suggest that patients whose tumors harbor this program should be considered for clinical trials testing rational agents to overcome de novo enzalutamide resistance.

LSD1 activates a lethal prostate cancer gene network independently of its demethylase function.

Medical castration that interferes with androgen receptor (AR) function is the principal treatment for advanced prostate cancer. However, clinical progression is universal, and tumors with AR-independent resistance mechanisms appear to be increasing in frequency. Consequently, there is an urgent need to develop new treatments targeting molecular pathways enriched in lethal prostate cancer. Lysine-specific demethylase 1 (LSD1) is a histone demethylase and an important regulator of gene expression. Here, we show that LSD1 promotes the survival of prostate cancer cells, including those that are castration-resistant, independently of its demethylase function and of the AR. Importantly, this effect is explained in part by activation of a lethal prostate cancer gene network in collaboration with LSD1's binding protein, ZNF217. Finally, that a small-molecule LSD1 inhibitor-SP-2509-blocks important demethylase-independent functions and suppresses castration-resistant prostate cancer cell viability demonstrates the potential of LSD1 inhibition in this disease.

Robustness and reproducibility of simple and complex synthetic logic circuit designs using a DBTL loop.

Computational tools addressing various components of design-build-test-learn (DBTL) loops for the construction of synthetic genetic networks exist but do not generally cover the entire DBTL loop. This manuscript introduces an end-to-end sequence of tools that together form a DBTL loop called Design Assemble Round Trip (DART). DART provides rational selection and refinement of genetic parts to construct and test a circuit. Computational support for experimental process, metadata management, standardized data collection and reproducible data analysis is provided via the previously published Round Trip (RT) test-learn loop. The primary focus of this work is on the Design Assemble (DA) part of the tool chain, which improves on previous techniques by screening up to thousands of network topologies for robust performance using a novel robustness score derived from dynamical behavior based on circuit topology only. In addition, novel experimental support software is introduced for the assembly of genetic circuits. A complete design-through-analysis sequence is presented using several OR and NOR circuit designs, with and without structural redundancy, that are implemented in budding yeast. The execution of DART tested the predictions of the design tools, specifically with regard to robust and reproducible performance under different experimental conditions. The data analysis depended on a novel application of machine learning techniques to segment bimodal flow cytometry distributions. Evidence is presented that, in some cases, a more complex build may impart more robustness and reproducibility across experimental conditions. Graphical Abstract.

A toolkit for enhanced reproducibility of RNASeq analysis for synthetic biologists.

Sequencing technologies, in particular RNASeq, have become critical tools in the design, build, test and learn cycle of synthetic biology. They provide a better understanding of synthetic designs, and they help identify ways to improve and select designs. While these data are beneficial to design, their collection and analysis is a complex, multistep process that has implications on both discovery and reproducibility of experiments. Additionally, tool parameters, experimental metadata, normalization of data and standardization of file formats present challenges that are computationally intensive. This calls for high-throughput pipelines expressly designed to handle the combinatorial and longitudinal nature of synthetic biology. In this paper, we present a pipeline to maximize the analytical reproducibility of RNASeq for synthetic biologists. We also explore the impact of reproducibility on the validation of machine learning models. We present the design of a pipeline that combines traditional RNASeq data processing tools with structured metadata tracking to allow for the exploration of the combinatorial design in a high-throughput and reproducible manner. We then demonstrate utility via two different experiments: a control comparison experiment and a machine learning model experiment. The first experiment compares datasets collected from identical biological controls across multiple days for two different organisms. It shows that a reproducible experimental protocol for one organism does not guarantee reproducibility in another. The second experiment quantifies the differences in experimental runs from multiple perspectives. It shows that the lack of reproducibility from these different perspectives can place an upper bound on the validation of machine learning models trained on RNASeq data. Graphical Abstract.

BET Bromodomain Inhibition Blocks an AR-Repressed, E2F1-Activated Treatment-Emergent Neuroendocrine Prostate Cancer Lineage Plasticity Program.

PURPOSE: Lineage plasticity in prostate cancer-most commonly exemplified by loss of androgen receptor (AR) signaling and a switch from a luminal to alternate differentiation program-is now recognized as a treatment resistance mechanism. Lineage plasticity is a spectrum, but neuroendocrine prostate cancer (NEPC) is the most virulent example. Currently, there are limited treatments for NEPC. Moreover, the incidence of treatment-emergent NEPC (t-NEPC) is increasing in the era of novel AR inhibitors. In contradistinction to de novo NEPC, t-NEPC tumors often express the AR, but AR's functional role in t-NEPC is unknown. Furthermore, targetable factors that promote t-NEPC lineage plasticity are also unclear. EXPERIMENTAL DESIGN: Using an integrative systems biology approach, we investigated enzalutamide-resistant t-NEPC cell lines and their parental, enzalutamide-sensitive adenocarcinoma cell lines. The AR is still expressed in these t-NEPC cells, enabling us to determine the role of the AR and other key factors in regulating t-NEPC lineage plasticity. RESULTS: AR inhibition accentuates lineage plasticity in t-NEPC cells-an effect not observed in parental, enzalutamide-sensitive adenocarcinoma cells. Induction of an AR-repressed, lineage plasticity program is dependent on activation of the transcription factor E2F1 in concert with the BET bromodomain chromatin reader BRD4. BET inhibition (BETi) blocks this E2F1/BRD4-regulated program and decreases growth of t-NEPC tumor models and a subset of t-NEPC patient tumors with high activity of this program in a BETi clinical trial. CONCLUSIONS: E2F1 and BRD4 are critical for activating an AR-repressed, t-NEPC lineage plasticity program. BETi is a promising approach to block this program.

Alternative splicing of LSD1+8a in neuroendocrine prostate cancer is mediated by SRRM4.

Neuroendocrine prostate cancer (NEPC) is the most virulent form of prostate cancer. Importantly, our recent work examining metastatic biopsy samples demonstrates NEPC is increasing in frequency. In contrast to prostate adenocarcinomas that express a luminal gene expression program, NEPC tumors express a neuronal gene expression program. Despite this distinction, the diagnosis of NEPC is often challenging, demonstrating an urgent need to identify new biomarkers and therapeutic targets. Our prior work demonstrated that the histone demethylase LSD1 (KDM1A) is important for survival of prostate adenocarcinomas, but little was known about LSD1's role in NEPC. Recently, a neural-specific transcript variant of LSD1-LSD1+8a-was discovered and demonstrated to activate neuronal gene expression in neural cells. The splicing factor SRRM4 was previously shown to promote LSD1+8a splicing in neuronal cells, and SRRM4 promotes NEPC differentiation and cell survival. Therefore, we sought to determine if LSD1+8a might play a role in NEPC and whether LSD1+8a splicing was linked to SRRM4. To investigate a potential role for LSD1+8a in NEPC, we examined a panel of prostate adenocarcinoma and NEPC patient-derived xenografts and metastatic biopsies. LSD1+8a was expressed exclusively in NEPC samples and correlated significantly with elevated expression of SRRM4. Using SRRM4-overexpressing cell lines, we determined that SRRM4 mediates alternative splicing of LSD1+8a. Finally, using gain of function studies, we confirmed that LSD1+8a and SRRM4 co-regulate target genes distinct from canonical LSD1. Our findings suggest further study of the interplay between SRRM4 and LSD1+8a and mechanisms by which LSD1+8a regulates gene expression in NEPC is warranted.

Copy Number Loss of 17q22 Is Associated with Enzalutamide Resistance and Poor Prognosis in Metastatic Castration-Resistant Prostate Cancer.

PURPOSE: The purpose of this study was to measure genomic changes that emerge with enzalutamide treatment using analyses of whole-genome sequencing and RNA sequencing. EXPERIMENTAL DESIGN: One hundred and one tumors from men with metastatic castration-resistant prostate cancer (mCRPC) who had not been treated with enzalutamide (n = 64) or who had enzalutamide-resistant mCRPC (n = 37) underwent whole genome sequencing. Ninety-nine of these tumors also underwent RNA sequencing. We analyzed the genomes and transcriptomes of these mCRPC tumors. RESULTS: Copy number loss was more common than gain in enzalutamide-resistant tumors. Specially, we identified 124 protein-coding genes that were more commonly lost in enzalutamide-resistant samples. These 124 genes included eight putative tumor suppressors located at nine distinct genomic regions. We demonstrated that focal deletion of the 17q22 locus that includes RNF43 and SRSF1 was not present in any patient with enzalutamide-naïve mCRPC but was present in 16% (6/37) of patients with enzalutamide-resistant mCRPC. 17q22 loss was associated with lower RNF43 and SRSF1 expression and poor overall survival from time of biopsy [median overall survival of 19.3 months in 17q22 intact vs. 8.9 months in 17q22 loss, HR, 3.44 95% confidence interval (CI), 1.338-8.867, log-rank P = 0.006]. Finally, 17q22 loss was linked with activation of several targetable factors, including CDK1/2, Akt, and PLK1, demonstrating the potential therapeutic relevance of 17q22 loss in mCRPC. CONCLUSIONS: Copy number loss is common in enzalutamide-resistant tumors. Focal deletion of chromosome 17q22 defines a previously unappreciated molecular subset of enzalutamide-resistant mCRPC associated with poor clinical outcome.

Maintenance of MYC expression promotes de novo resistance to BET bromodomain inhibition in castration-resistant prostate cancer.

The BET bromodomain protein BRD4 is a chromatin reader that regulates transcription, including in cancer. In prostate cancer, specifically, the anti-tumor activity of BET bromodomain inhibition has been principally linked to suppression of androgen receptor (AR) function. MYC is a well-described BRD4 target gene in multiple cancer types, and prior work demonstrates that MYC plays an important role in promoting prostate cancer cell survival. Importantly, several BET bromodomain clinical trials are ongoing, including in prostate cancer. However, there is limited information about pharmacodynamic markers of response or mediators of de novo resistance. Using a panel of prostate cancer cell lines, we demonstrated that MYC suppression-rather than AR suppression-is a key determinant of BET bromodomain inhibitor sensitivity. Importantly, we determined that BRD4 was dispensable for MYC expression in the most resistant cell lines and that MYC RNAi + BET bromodomain inhibition led to additive anti-tumor activity in the most resistant cell lines. Our findings demonstrate that MYC suppression is an important pharmacodynamic marker of BET bromodomain inhibitor response and suggest that targeting MYC may be a promising therapeutic strategy to overcome de novo BET bromodomain inhibitor resistance in prostate cancer.

BET bromodomain inhibition blocks the function of a critical AR-independent master regulator network in lethal prostate cancer.

BET bromodomain inhibitors block prostate cancer cell growth at least in part through c-Myc and androgen receptor (AR) suppression. However, little is known about other transcriptional regulators whose suppression contributes to BET bromodomain inhibitor anti-tumor activity. Moreover, the anti-tumor activity of BET bromodomain inhibition in AR-independent castration-resistant prostate cancers (CRPC), whose frequency is increasing, is also unknown. Herein, we demonstrate that BET bromodomain inhibition blocks growth of a diverse set of CRPC cell models, including those that are AR-independent or in which c-Myc is not suppressed. To identify transcriptional regulators whose suppression accounts for these effects, we treated multiple CRPC cell lines with the BET bromodomain inhibitor JQ1 and then performed RNA-sequencing followed by Master Regulator computational analysis. This approach identified several previously unappreciated transcriptional regulators that are highly expressed in CRPC and whose suppression, via both transcriptional or post-translational mechanisms, contributes to the anti-tumor activity of BET bromodomain inhibitors.

Integrative molecular network analysis identifies emergent enzalutamide resistance mechanisms in prostate cancer.

Recent work demonstrates that castration-resistant prostate cancer (CRPC) tumors harbor countless genomic aberrations that control many hallmarks of cancer. While some specific mutations in CRPC may be actionable, many others are not. We hypothesized that genomic aberrations in cancer may operate in concert to promote drug resistance and tumor progression, and that organization of these genomic aberrations into therapeutically targetable pathways may improve our ability to treat CRPC. To identify the molecular underpinnings of enzalutamide-resistant CRPC, we performed transcriptional and copy number profiling studies using paired enzalutamide-sensitive and resistant LNCaP prostate cancer cell lines. Gene networks associated with enzalutamide resistance were revealed by performing an integrative genomic analysis with the PAthway Representation and Analysis by Direct Reference on Graphical Models (PARADIGM) tool. Amongst the pathways enriched in the enzalutamide-resistant cells were those associated with MEK, EGFR, RAS, and NFKB. Functional validation studies of 64 genes identified 10 candidate genes whose suppression led to greater effects on cell viability in enzalutamide-resistant cells as compared to sensitive parental cells. Examination of a patient cohort demonstrated that several of our functionally-validated gene hits are deregulated in metastatic CRPC tumor samples, suggesting that they may be clinically relevant therapeutic targets for patients with enzalutamide-resistant CRPC. Altogether, our approach demonstrates the potential of integrative genomic analyses to clarify determinants of drug resistance and rational co-targeting strategies to overcome resistance.

Cellular androgen content influences enzalutamide agonism of F877L mutant androgen receptor.

Prostate cancer is the most commonly diagnosed and second-most lethal cancer among men in the United States. The vast majority of prostate cancer deaths are due to castration-resistant prostate cancer (CRPC) - the lethal form of the disease that has progressed despite therapies that interfere with activation of androgen receptor (AR) signaling. One emergent resistance mechanism to medical castration is synthesis of intratumoral androgens that activate the AR. This insight led to the development of the AR antagonist enzalutamide. However, resistance to enzalutamide invariably develops, and disease progression is nearly universal. One mechanism of resistance to enzalutamide is an F877L mutation in the AR ligand-binding domain that can convert enzalutamide to an agonist of AR activity. However, mechanisms that contribute to the agonist switch had not been fully clarified, and there were no therapies to block AR F877L. Using cell line models of castration-resistant prostate cancer (CRPC), we determined that cellular androgen content influences enzalutamide agonism of mutant F877L AR. Further, enzalutamide treatment of AR F877L-expressing cell lines recapitulated the effects of androgen activation of F877L AR or wild-type AR. Because the BET bromodomain inhibitor JQ-1 was previously shown to block androgen activation of wild-type AR, we tested JQ-1 in AR F877L-expressing CRPC models. We determined that JQ-1 suppressed androgen or enzalutamide activation of mutant F877L AR and suppressed growth of mutant F877L AR CRPC tumors in vivo, demonstrating a new strategy to treat tumors harboring this mutation.