Finding Genes for Human Prostate Cancer

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

Prostate Cancer Studies Prostate cancer (PCa) is the most common solid tumor in (humans) men and is a major cause of cancer-related morbidity and mortality, yet its etiology and molecular underpinnings are unresolved (Eeles et al., 2014). Last year, more than 161,360 men were diagnosed with PCa, and 26,730 were predicted to die of the disease (Siegel et al., 2017). We have worked for over 20 years to identify genes and markers associated with disease risk, progression and outcomes. Much of our work has been in consortia, or with long time collaborator Dr. Janet Stanford at the Fred Hutchinson Cancer Research Center. Risk of Aggressive and Lethal Prostate Cancer Biomarkers to identify men who at high risk of metastatic disease are needed for prostate cancer. We have worked with the Stanford group to continue to identify germline variants associated with prostate cancer specific mortality. In a study by FitzGerald et al., (FitzGerald et al., 2018). we analyzed 12,082 PC cases and found variants in IL4, MGMT and AKT1 are associated PC mortality, providing evidence that genetic background plays a role in modulating tumor aggressiveness. We also interrogated the tumor transcriptome to identify genes with predictive utility. Primary tumor samples from 383 patients were used as a discovery set with validation in an independent cohort of 78 patients. All patients were followed for 5 years after radical prostatectomy to ascertain outcomes. We showed that 23 differently expressed transcripts were validated and have predictive value for metastatic-lethal prostate cancer in men initially treated for localized disease. Transcripts represent genes with functions related to tumor aggressiveness, including cell cycle/proliferation, the immune/inflammatory pathway, and steroid hormone signal transduction (Rubicz et al., 2017). Epigenetic Studies To continue our epigenetic studies, we examined PTEN lost in tumors (Geybels et al., 2017). We studied 471 patients with PTEN loss in tumors and found that loss was significantly associated with a higher risk of recurrence. Hazard ratios for hemi- and homozygous loss were 1.39 (95% CI: 0.73-2.64) and 2.84 (95% CI: 1.30-6.19), respectively. Epigenome-wide methylation profiling identified 4,208 differentially methylated CpGs (FDR Q-value < 0.01) in tumors with any versus no PTEN loss. Tumor methylation data were used to build a methylation signature of PTEN loss in our cohort, which was confirmed in The Cancer Genome Atlas (TCGA), and included CpGs in ATP11A, GDNF, JAK1, JAM3, and VAPA. Risk in Hereditary PC Families In this study, we focused on the important but elusive class of low-frequency, moderately penetrant variants by performing disease model-based variant filtering of whole exome sequence data from 75 hereditary PCa families (Karyadi et al., 2017). In a study led by Ostrander lab staff scientist Danielle Karyadi, we analyzed 341 candidate risk variants and identified nine variants significantly associated with increased prostate cancer (PCa) risk in a population-based, case-control study of 2,495 men. In an independent nested case-control study of 7,121 men, we observed a risk association for TANGO2 p.Ser17Ter, and the established HOXB13 p.Gly84Glu variant. Meta-analysis combining the case-control studies identified two additional variants suggestively associated with risk: OR5H14 p.Met59Val and CHAD p.Ala342Asp. The TANGO2 and HOXB13 variants co-occurred in cases more often than expected by chance and never in controls. Finally, TANGO2 p.Ser17Ter was associated with aggressive disease in both case-control studies separately. Using WGS and SNP chip analysis, our analyses identified three new PCa susceptibility. Prostate Tumor Analyses Because Gleason score (GS) is one of the best predictors of PCa aggressiveness we developed a gene expression signature of GS to enhance the prediction of PCa outcomes (Jhun et al., 2017). Elastic net was used to construct a gene expression signature by contrasting GS 8-10 vs. 6 tumors in the TCGA dataset. The constructed signature was then evaluated for its ability to predict recurrence and metastatic-lethal (ML) progression in a Fred Hutchinson Cancer Research Center (FH) patient cohort (N=408; NRecurrence=109; NMLprogression=27). The expression signature included transcripts representing 49 genes and a 25% increase in the signature was associated with a hazard ratio (HR) of 1.51 (P=2.710-5) for recurrence. Compared to a model with age at diagnosis, pathological stage and GS, the gene expression signature improved the Area Under the Curve (AUC) for recurrence (3%) and ML progression (6%). This gene expression signature based on GS may improve the prediction of recurrence as well as ML progression in PCa patients after radical prostatectomy. In collaboration with the Center for Inherited Disease Research, we optimized a technique for generating exome-enriched sequencing libraries using DNA extracted from formalin-fixed paraffin-embedded (FFPE) samples. The workflow can start with as little as 50 ng of DNA. The process includes an assessment of the quality and quantity of the DNA, use of a DNA repair enzyme that fixes the most common damage resulting from formalin fixation and a bead-based clean-up to minimize sample loss. Afterwards, targeted selection for exons is performed followed by Illumina next-generation short-read sequencing. This study reports a technique which is successful at generating whole exome sequencing from archived FFPE samples (Marosy et al., 2017). Methods Development Due to low power for single-variant testing, aggregative methods are commonly used to combine observed rare variation within a single gene. Causal variation may also aggregate across multiple genes within relevant biomolecular pathways. Kernel-machine regression and adaptive testing methods for aggregative rare-variant association testing are powerful approaches for pathway-level analysis, but are computationally intensive. We were thus part of a large group that outlined a statistical strategy for aggregative rare-variant analysis using component gene-level linear kernel score test summary statistics as well as derive simple estimators of the effective number of tests for family-wise error rate control (Larson et al., 2017). We then conduct extensive simulation studies to characterize the behavior of our approach relative to direct application of kernel and adaptive methods under a variety of conditions. We also apply our method to two case-control studies, respectively, evaluating rare variation in hereditary prostate cancer and now provide open-source R code for public use to facilitate easy application of our methods to existing rare-variant analysis results. Consortia We continued our participation in large consortia publishing two high profile papers. In the first, we identified 63 new prostate cancer risk genes in an association study of 140,000 men (Schumacher et al., 2018) in a paper published in Nature Communication. In a second paper, published in Nature Genetics, we did fine mapping of prostate cancer susceptibility loci in a large meta analysis to identify specific disease associated variants (Dadaev et al., 2018). Our future projects will focus on finding genes that increase tumor aggressiveness and metastasis.