Biologically informed deep neural network for prostate cancer discovery.

The determination of molecular features that mediate clinically aggressive phenotypes in prostate cancer remains a major biological and clinical challenge1,2. Recent advances in interpretability of machine learning models as applied to biomedical problems may enable discovery and prediction in clinical cancer genomics3-5. Here we developed P-NET-a biologically informed deep learning model-to stratify patients with prostate cancer by treatment-resistance state and evaluate molecular drivers of treatment resistance for therapeutic targeting through complete model interpretability. We demonstrate that P-NET can predict cancer state using molecular data with a performance that is superior to other modelling approaches. Moreover, the biological interpretability within P-NET revealed established and novel molecularly altered candidates, such as MDM4 and FGFR1, which were implicated in predicting advanced disease and validated in vitro. Broadly, biologically informed fully interpretable neural networks enable preclinical discovery and clinical prediction in prostate cancer and may have general applicability across cancer types.

PIK3CA in cancer: The past 30 years.

Almost thirty years ago, PI3K was discovered as a lipid kinase associated with certain oncoproteins. The first decade of research on PI3K saw the identification, purification and cloning of PI3Kα. The second decade of research was noted for the identification of some of PI3K's activators and effectors. This was accompanied by the discovery that PI3K acts as a retroviral oncogene. The third decade was known for the establishment of the direct involvement of PI3K in cancer, demonstrated by the identification of cancer-specific mutations. Efforts to target PI3K were on the rise from that moment on, accompanied by the first clinical trials for PI3K inhibitor therapies. In the fourth decade of research, PI3K-based cancer drugs will continue to emerge, as will new knowledge regarding other uncovered functions of this protein and pathway.

Systematic Interrogation of Tumor Cell Resistance to Chimeric Antigen Receptor T-cell Therapy in Pancreatic Cancer.

Chimeric antigen receptor (CAR) T-cell therapy can lead to dramatic clinical responses in B-cell malignancies. However, early clinical trials with CAR T-cell therapy in non-B-cell malignancies have been disappointing to date, suggesting that tumor-intrinsic features contribute to resistance. To investigate tumor-intrinsic modes of resistance, we performed genome scale CRISPR-Cas9 screens in mesothelin (MSLN)-expressing pancreatic cancer cells. Co-culture with MSLN-targeting CAR T cells identified both antigen-dependent and antigen-independent modes of resistance. In particular, loss of the majority of the genes involved in the pathway responsible for GPI-anchor biosynthesis and attachment abrogated the ability of CAR T cells to target pancreatic cancer cells, suggesting that disruption of this pathway may permit MSLN CAR T-cell evasion in the clinic. Antigen-independent mediators of CAR T-cell response included members of the death receptor pathway as well as genes that regulate tumor transcriptional responses, including TFAP4 and INTS12. TFAP4-mediated CAR T resistance depended on the NFκB transcription factor p65, indicating that tumor resistance to CAR T-cell therapy likely involves alterations in tumor-intrinsic states. Overall, this study uncovers multiple antigen-dependent and -independent mechanisms of CAR T-cell evasion by pancreatic cancer, paving the way for overcoming resistance in this disease that is notoriously refractory to immunotherapy. SIGNIFICANCE: The identification and validation of key determinants of CAR T-cell response in pancreatic cancer provide insights into the landscape of tumor cell intrinsic resistance mechanisms and into approaches to improve therapeutic efficacy.

NRas activity is regulated by dynamic interactions with nanoscale signaling clusters at the plasma membrane.

NRas is a key mediator of the mitogenic pathway in normal cells and in cancer cells. Its dynamics and nanoscale organization at the plasma membrane (PM) facilitate its signaling. Here, we used two-color photoactivated localization microscopy to resolve the organization of individual NRas and associated signaling proteins in live melanoma cells, with resolution down to ∼20 nm. Upon EGF activation, a fraction of NRas and BRAF (dis)assembled synchronously at the PM in co-clusters. NRas and BRAF clusters associated with GPI-enriched domains, serving as possible nucleation sites for these clusters. NRas and BRAF association in mutual clusters was reduced by the NRas farnesylation inhibitor lonafarnib, yet enhanced by the BRAF inhibitor vemurafenib. Surprisingly, dispersed NRas molecules associated with the periphery of self-clusters of either Grb2 or NF1. Thus, NRas-mediated signaling, which is critical in health and disease, is regulated by dynamic interactions with functional clusters of BRAF or other related proteins at the PM.

CREB5 reprograms FOXA1 nuclear interactions to promote resistance to androgen receptor-targeting therapies.

Metastatic castration-resistant prostate cancers (mCRPCs) are treated with therapies that antagonize the androgen receptor (AR). Nearly all patients develop resistance to AR-targeted therapies (ARTs). Our previous work identified CREB5 as an upregulated target gene in human mCRPC that promoted resistance to all clinically approved ART. The mechanisms by which CREB5 promotes progression of mCRPC or other cancers remains elusive. Integrating ChIP-seq and rapid immunoprecipitation and mass spectroscopy of endogenous proteins, we report that cells overexpressing CREB5 demonstrate extensive reprogramming of nuclear protein-protein interactions in response to the ART agent enzalutamide. Specifically, CREB5 physically interacts with AR, the pioneering actor FOXA1, and other known co-factors of AR and FOXA1 at transcription regulatory elements recently found to be active in mCRPC patients. We identified a subset of CREB5/FOXA1 co-interacting nuclear factors that have critical functions for AR transcription (GRHL2, HOXB13) while others (TBX3, NFIC) regulated cell viability and ART resistance and were amplified or overexpressed in mCRPC. Upon examining the nuclear protein interactions and the impact of CREB5 expression on the mCRPC patient transcriptome, we found that CREB5 was associated with Wnt signaling and epithelial to mesenchymal transitions, implicating these pathways in CREB5/FOXA1-mediated ART resistance. Overall, these observations define the molecular interactions among CREB5, FOXA1, and pathways that promote ART resistance.

MEK Inhibition Reverses Aberrant Signaling in Melanoma Cells through Reorganization of NRas and BRAF in Self Nanoclusters.

Hotspot mutations of the oncogenes BRAF and NRas are the most common genetic alterations in cutaneous melanoma. Still, the nanoscale organization and signal coupling of these proteins remain incompletely understood, particularly upon expression of oncogenic NRas mutants. Here we employed single-molecule localization microscopy to study the nanoscale organization of NRas and BRAF at the plasma membrane (PM) of melanoma cells. NRas and BRAF resided in self-clusters that did not associate well in resting cells. In EGF-activated cells, NRas clusters became more diffused while overall protein levels at the PM increased; thus allowing enhanced association of NRas and BRAF and downstream signaling. In multiple melanoma cell lines, mutant NRas resided in more pronounced self-clusters relative to wild-type (WT) NRas yet associated more with the clustered and more abundant BRAF. In cells resistant to trametinib, a clinical MEK inhibitor (MEKi), a similar coclustering of NRas and BRAF was observed upon EGF activation. Strikingly, treatment of cells expressing mutant NRas with trametinib reversed the effect of mutant NRas expression by restoring the nonoverlapping self-clusters of NRas and BRAF and by reducing their PM levels and elevated pERK levels caused by mutant NRas. Our results indicate a new mechanism for signal regulation of NRas in melanoma through its nanoscale dynamic organization and a new mechanism for MEKi function in melanoma cells carrying NRas mutations but lacking MEK mutations. SIGNIFICANCE: Nanoscale dynamic organization of WT and mutant NRas relative to BRAF serves as a regulatory mechanism for NRas signaling and may be a viable therapeutic target for its sensitivity to MEKi.

Reprogramming of the FOXA1 cistrome in treatment-emergent neuroendocrine prostate cancer.

Lineage plasticity, the ability of a cell to alter its identity, is an increasingly common mechanism of adaptive resistance to targeted therapy in cancer. An archetypal example is the development of neuroendocrine prostate cancer (NEPC) after treatment of prostate adenocarcinoma (PRAD) with inhibitors of androgen signaling. NEPC is an aggressive variant of prostate cancer that aberrantly expresses genes characteristic of neuroendocrine (NE) tissues and no longer depends on androgens. Here, we investigate the epigenomic basis of this resistance mechanism by profiling histone modifications in NEPC and PRAD patient-derived xenografts (PDXs) using chromatin immunoprecipitation and sequencing (ChIP-seq). We identify a vast network of cis-regulatory elements (N~15,000) that are recurrently activated in NEPC. The FOXA1 transcription factor (TF), which pioneers androgen receptor (AR) chromatin binding in the prostate epithelium, is reprogrammed to NE-specific regulatory elements in NEPC. Despite loss of dependence upon AR, NEPC maintains FOXA1 expression and requires FOXA1 for proliferation and expression of NE lineage-defining genes. Ectopic expression of the NE lineage TFs ASCL1 and NKX2-1 in PRAD cells reprograms FOXA1 to bind to NE regulatory elements and induces enhancer activity as evidenced by histone modifications at these sites. Our data establish the importance of FOXA1 in NEPC and provide a principled approach to identifying cancer dependencies through epigenomic profiling.

A genome-wide CRISPR screen identifies FBXO42 involvement in resistance toward MEK inhibition in NRAS-mutant melanoma.

NRAS mutations are the most common alterations among RAS isoforms in cutaneous melanoma, with patients harboring these aggressive tumors having a poor prognosis and low survival rate. The main line of treatment for these patients is MAPK pathway-targeted therapies, such as MEK inhibitors, but, unfortunately, the response to these inhibitors is variable due to tumor resistance. Identifying genetic modifiers involved in resistance toward MEK-targeted therapy may assist in the development of new therapeutic strategies, enhancing treatment response and patient survival. Our whole-genome CRISPR-Cas9 knockout screen identified the target Kelch domain-containing F-Box protein 42 (FBXO42) as a factor involved in NRAS-mutant melanoma-acquired resistance to the MEK1/2 inhibitor trametinib. We further show that FBXO42, an E3 ubiquitin ligase, is involved in the TAK1 signaling pathway, possibly prompting an increase in active P38. In addition, we demonstrate that combining trametinib with the TAK1 inhibitor, takinib, is a far more efficient treatment than trametinib alone in NRAS-mutant melanoma cells. Our findings thus show a new pathway involved in NRAS-mutant melanoma resistance and provide new opportunities for novel therapeutic options.

Author Correction: Combined inhibition of MEK and nuclear ERK translocation has synergistic antitumor activity in melanoma cells.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Correction: CAPN1 is a novel binding partner and regulator of the tumor suppressor NF1 in melanoma.

[This corrects the article DOI: 10.18632/oncotarget.25805.].

CAPN1 is a novel binding partner and regulator of the tumor suppressor NF1 in melanoma.

Neurofibromin 1 (NF1), a tumor suppressor that negatively regulates RAS through its GTPase activity, is highly mutated in various types of sporadic human cancers, including melanoma. However, the binding partners of NF1 and the pathways in which it is involved in melanoma have not been characterized in an in depth manner. Utilizing a mass spectrometry analysis of NF1 binding partners, we revealed Calpain1 (CAPN1), a calcium-dependent neutral cysteine protease, as a novel NF1 binding partner that regulates NF1 degradation in melanoma cells. ShRNA-mediated knockdown of CAPN1 or treatment with a CAPN1 inhibitor stabilizes NF1 protein levels, downregulates AKT signaling and melanoma cell growth. Combination treatment of Calpain inhibitor I with MEKi Trametinib in different melanoma cells is more effective in reducing melanoma cell growth compared to treatment with Trametinib alone, suggesting that this combination may have a therapeutic potential in melanoma. This novel mechanism for regulating NF1 in melanoma provides a molecular basis for targeting CAPN1 in order to stabilize NF1 levels and, in doing so, suppressing Ras activation; this mechanism can be exploited therapeutically in melanoma and other cancers.

Harnessing synthetic lethality to predict the response to cancer treatment.

While synthetic lethality (SL) holds promise in developing effective cancer therapies, SL candidates found via experimental screens often have limited translational value. Here we present a data-driven approach, ISLE (identification of clinically relevant synthetic lethality), that mines TCGA cohort to identify the most likely clinically relevant SL interactions (cSLi) from a given candidate set of lab-screened SLi. We first validate ISLE via a benchmark of large-scale drug response screens and by predicting drug efficacy in mouse xenograft models. We then experimentally test a select set of predicted cSLi via new screening experiments, validating their predicted context-specific sensitivity in hypoxic vs normoxic conditions and demonstrating cSLi's utility in predicting synergistic drug combinations. We show that cSLi can successfully predict patients' drug treatment response and provide patient stratification signatures. ISLE thus complements existing actionable mutation-based methods for precision cancer therapy, offering an opportunity to expand its scope to the whole genome.

Combined inhibition of MEK and nuclear ERK translocation has synergistic antitumor activity in melanoma cells.

Genetic alterations in BRAF, NRAS and NF1 that activate the ERK cascade, account for over 80% of metastatic melanomas. However, ERK cascade inhibitors have been proven beneficial almost exclusively for BRAF mutant melanomas. One of the hallmarks of the ERK cascade is the nuclear translocation of ERK1/2, which is important mainly for the induction of proliferation. This translocation can be inhibited by the NTS-derived peptide (EPE) that blocks the ERK1/2-importin7 interaction, inhibits the nuclear translocation of ERK1/2, and arrests active ERK1/2 in the cytoplasm. In this study, we found that the EPE peptide significantly reduced the viability of not only BRAF, but also several NRAS and NF1 mutant melanomas. Importantly, combination of the EPE peptide and trametinib showed synergy in reducing the viability of some NRAS mutant melanomas, an effect driven by the partial preservation of negative feedback loops. The same combination significantly reduced the viability of other melanoma cells, including those resistant to mono-treatment with EPE peptide and ERK cascade inhibitors. Our study indicates that targeting the nuclear translocation of ERK1/2, in combination with MEK inhibitors can be used for the treatment of different mutant melanomas.

miR-27a and miR-27a* contribute to metastatic properties of osteosarcoma cells.

Osteosarcoma (OS) is the most common primary malignant bone tumor in adolescents and young adults. The essential mechanisms underlying osteosarcomagenesis and progression continue to be obscure. MicroRNAs (miRNAs) have far-reaching effects on the cellular biology of development and cancer. We recently reported that unique miRNA signatures associate with the pathogenesis and progression of OS. Of particular interest, we found that higher expression of miR-27a is associated with clinical metastatic disease. We report here that overexpression of miR-27a/miR-27a*, a microRNA pair derived from a single precursor, promotes pulmonary OS metastases formation. By contrast, sequestering miR-27a/miR-27a* by sponge technology suppressed OS cells invasion and metastases formation. miR-27a/miR-27a* directly repressed CBFA2T3 expression among other target genes. We demonstrated that CBFA2T3 is downregulated in majority of OS samples and its over expression significantly attenuated OS metastatic process mediated by miR-27a/miR-27a* underscoring CBFA2T3 functions as a tumor suppressor in OS. These findings establish that miR-27a/miR-27a* pair plays a significant role in OS metastasis and proposes it as a potential diagnostic and therapeutic target in managing OS metastases.

Recurrent inactivating RASA2 mutations in melanoma.

Analysis of 501 melanoma exomes identified RASA2, encoding a RasGAP, as a tumor-suppressor gene mutated in 5% of melanomas. Recurrent loss-of-function mutations in RASA2 were found to increase RAS activation, melanoma cell growth and migration. RASA2 expression was lost in ≥30% of human melanomas and was associated with reduced patient survival. These findings identify RASA2 inactivation as a melanoma driver and highlight the importance of RasGAPs in cancer.