SUMMARY Immune checkpoint inhibitors (ICIs) produce durable responses in some melanoma patients, but many patients derive no clinical benefit, and the molecular underpinnings of such resistance remain elusive. Here, we leveraged single-cell RNA sequencing (scRNA-seq) from 33 melanoma tumors and computational analyses to interrogate malignant cell states that promote immune evasion. We identified a resistance program expressed by malignant cells that is associated with T cell exclusion and immune evasion. The program is expressed prior to immunotherapy, characterizes cold niches in situ, and predicts clinical responses to anti-PD-1 therapy in an independent cohort of 112 melanoma patients. CDK4/6-inhibition represses this program in individual malignant cells, induces senescence, and reduces melanoma tumor outgrowth in mouse models in vivo when given in combination with immunotherapy. Our study provides a high-resolution landscape of ICI-resistant cell states, identifies clinically predictive signatures, and suggests new therapeutic strategies to overcome immunotherapy resistance.

Comprehensive genomic characterization of prostate cancer has identified recurrent alterations in androgen signaling, DNA repair, and PI3K among others. However, larger and uniform genomic analysis may reveal additional recurrently mutated genes at lower frequencies. Here we aggregate and uniformly analyze exome sequencing data from 1013 prostate cancers. We identify and validate a new class of E26 transformation-specific (ETS) fusion negative tumors defined by mutations in epigenetic regulators, as well as alterations in pathways not previously implicated in prostate cancer, such as the spliceosome pathway. We find that the incidence of significantly mutated genes (SMGs) follows a long-tail distribution, with many genes mutated in less than 3% of cases. We identify a total of 97 SMGs, including 70 not previously implicated in prostate cancer, such as the ubiquitin ligase CUL3 and the transcription factor SPEN. Finally, comparing primary and metastatic prostate cancer reveals a set of genomic markers that may inform risk stratification.

Tumor mutational burden correlates with response to immune checkpoint blockade in multiple solid tumors, although in microsatellite-stable tumors this association is of uncertain clinical utility. Here we uniformly analyzed whole-exome sequencing (WES) of 249 tumors and matched normal tissue from patients with clinically annotated outcomes to immune checkpoint therapy, including radiographic response, across multiple cancer types to examine additional tumor genomic features that contribute to selective response. Our analyses identified genomic correlates of response beyond mutational burden, including somatic events in individual driver genes, certain global mutational signatures, and specific HLA-restricted neoantigens. However, these features were often interrelated, highlighting the complexity of identifying genetic driver events that generate an immunoresponsive tumor environment. This study lays a path forward in analyzing large clinical cohorts in an integrated and multifaceted manner to enhance the ability to discover clinically meaningful predictive features of response to immune checkpoint blockade.

Cancer genomes contain large numbers of somatic mutations, but few of these mutations drive tumor development. Current approaches identify driver genes based on mutational recurrence, or they approximate the functional consequences of nonsynonymous mutations using bioinformatic scores. While passenger mutations are enriched in characteristic nucleotide contexts, driver mutations occur in functional positions, which are not necessarily surrounded by a particular nucleotide context. We observed that mutations in contexts that deviate from the characteristic contexts around passenger mutations provide a signal in favor of driver genes. We therefore developed a method that combines this feature with the signals traditionally used for driver gene identification. We applied our method to whole-exome sequencing data from 11,873 tumor-normal pairs and identified 460 driver genes that clustered into 21 cancer-related pathways. Our study provides a resource of driver genes across 28 tumor types with additional driver genes identified based on mutations in unusual nucleotide contexts.

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.

BACKGROUND. Understanding the integrated immunogenomic landscape of advanced prostate cancer (APC) could impact stratified treatment selection.METHODS. Defective mismatch repair (dMMR) status was determined by either loss of mismatch repair protein expression on IHC or microsatellite instability (MSI) by PCR in 127 APC biopsies from 124 patients (Royal Marsden [RMH] cohort); MSI by targeted panel next-generation sequencing (MSINGS) was then evaluated in the same cohort and in 254 APC samples from the Stand Up To Cancer/Prostate Cancer Foundation (SU2C/PCF). Whole exome sequencing (WES) data from this latter cohort were analyzed for pathogenic MMR gene variants, mutational load, and mutational signatures. Transcriptomic data, available for 168 samples, was also performed.RESULTS. Overall, 8.1% of patients in the RMH cohort had some evidence of dMMR, which associated with decreased overall survival. Higher MSINGS scores associated with dMMR, and these APCs were enriched for higher T cell infiltration and PD-L1 protein expression. Exome MSINGS scores strongly correlated with targeted panel MSINGS scores (r = 0.73, P < 0.0001), and higher MSINGS scores associated with dMMR mutational signatures in APC exomes. dMMR mutational signatures also associated with MMR gene mutations and increased immune cell, immune checkpoint, and T cell–associated transcripts. APC with dMMR mutational signatures overexpressed a variety of immune transcripts, including CD200R1, BTLA, PD-L1, PD-L2, ADORA2A, PIK3CG, and TIGIT.CONCLUSION. These data could impact immune target selection, combination therapeutic strategy selection, and selection of predictive biomarkers for immunotherapy in APC.FUNDING. We acknowledge funding support from Movember, Prostate Cancer UK, The Prostate Cancer Foundation, SU2C, and Cancer Research UK.

BackgroundThe diversity of clinical tumor profiling approaches (small panels to whole exomes with matched or unmatched germline analysis) may engender uncertainty about their benefits and liabilities, particularly in light of reported germline false positives in tumor-only profiling and use of global mutational and/or neoantigen data. The goal of this study was to determine the impact of genomic analysis strategies on error rates and data interpretation across contexts and ancestries.MethodsWe modeled common tumor profiling modalities—large (n = 300 genes), medium (n = 48 genes), and small (n = 15 genes) panels—using clinical whole exomes (WES) from 157 patients with lung or colon adenocarcinoma. We created a tumor-only analysis algorithm to assess germline false positive rates, the impact of patient ancestry on tumor-only results, and neoantigen detection.ResultsAfter optimizing a germline filtering strategy, the germline false positive rate with tumor-only large panel sequencing was 14 % (144/1012 variants). For patients whose tumor-only results underwent molecular pathologist review (n = 91), 50/54 (93 %) false positives were correctly interpreted as uncertain variants. Increased germline false positives were observed in tumor-only sequencing of non-European compared with European ancestry patients (p < 0.001; Fisher’s exact) when basic germline filtering approaches were used; however, the ExAC database (60,706 germline exomes) mitigated this disparity (p = 0.53). Matched and unmatched large panel mutational load correlated with WES mutational load (r2 = 0.99 and 0.93, respectively; p < 0.001). Neoantigen load also correlated (r2 = 0.80; p < 0.001), though WES identified a broader spectrum of neoantigens. Small panels did not predict mutational or neoantigen load.ConclusionsLarge tumor-only targeted panels are sufficient for most somatic variant identification and mutational load prediction if paired with expanded germline analysis strategies and molecular pathologist review. Paired germline sequencing reduced overall false positive mutation calls and WES provided the most neoantigens. Without patient-matched germline data, large germline databases are needed to minimize false positive mutation calling and mitigate ethnic disparities.Electronic supplementary materialThe online version of this article (doi:10.1186/s13073-016-0333-9) contains supplementary material, which is available to authorized users.

Discussion | Both cancer and PD occur primarily at advanced ages and can be debilitating. Consequently, medical surveillance for cancer could be influenced by a prior PD diagnosis, which may bias the observed PD-cancer associations. Moreover, patterns of medical surveillance may differ by health care system (eg, US Medicare and Taiwan National Health Insurance Research Database). It is therefore important that in evaluating PD-cancer associations investigators adjust for medical surveillance and include negative controls.Although we cannot rule out small risks attributable to sample size, we did not find a positive association between PD and subsequent cancer in Asian Americans in SEER-Medicare data. Although according to SEER-Medicare data Asian Americans are ethnically more diverse than the Taiwanese population, our null findings do not support the hypothesis that ethnicity is a major contributor to the elevated risk in the study by Lin et al. 4 The elevated risks observed in the study by Lin et al 4 could reflect various factors, including possibly more cancer surveillance after PD.