Ostrander, Elaine A.

AbstractBackgroundMolecular studies have tried to address the unmet need for prognostic biomarkers in prostate cancer (PCa). Some gene expression tests improve upon clinical factors for prediction of outcomes, but additional tools for accurate prediction of tumor aggressiveness are needed.MethodsBased on a previously published panel of 23 gene transcripts that distinguished patients with metastatic progression, we constructed a prediction model using independent training and testing datasets. Using the validated messenger RNAs and Gleason score (GS), we performed model selection in the training set to define a final locked model to classify patients who developed metastatic‐lethal events from those who remained recurrence‐free. In an independent testing dataset, we compared our locked model to established clinical prognostic factors and utilized Kaplan‐Meier curves and receiver operating characteristic analyses to evaluate the model's performance.ResultsThirteen of 23 previously identified gene transcripts that stratified patients with aggressive PCa were validated in the training dataset. These biomarkers plus GS were used to develop a four‐gene (CST2, FBLN1, TNFRSF19, and ZNF704) transcript (4GT) score that was significantly higher in patients who progressed to metastatic‐lethal events compared to those without recurrence in the testing dataset (P = 5.7 × 10−11). The 4GT score provided higher prediction accuracy (area under the ROC curve [AUC] = 0.76; 95% confidence interval [CI] = 0.69‐0.83; partial area under the ROC curve [pAUC] = 0.008) than GS alone (AUC = 0.63; 95% CI = 0.56‐0.70; pAUC = 0.002), and it improved risk stratification in subgroups defined by a combination of clinicopathological features (ie, Cancer of the Prostate Risk Assessment‐Surgery).ConclusionOur validated 4GT score has prognostic value for metastatic‐lethal progression in men treated for localized PCa and warrants further evaluation for its clinical utility.

ABSTRACTNext‐generation sequencing technologies have afforded unprecedented characterization of low‐frequency and rare genetic variation. 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 have been demonstrated to be powerful approaches for pathway‐level analysis, although these methods tend to be computationally intensive at high‐variant dimensionality and require access to complete data. An additional analytical issue in scans of large pathway definition sets is multiple testing correction. Gene set definitions may exhibit substantial genic overlap, and the impact of the resultant correlation in test statistics on Type I error rate control for large agnostic gene set scans has not been fully explored. Herein, we first outline 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. 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 schizophrenia. Finally, we provide open‐source R code for public use to facilitate easy application of our methods to existing rare‐variant analysis results.

ABSTRACTRare variants (RVs) have been shown to be significant contributors to complex disease risk. By definition, these variants have very low minor allele frequencies and traditional single‐marker methods for statistical analysis are underpowered for typical sequencing study sample sizes. Multimarker burden‐type approaches attempt to identify aggregation of RVs across case‐control status by analyzing relatively small partitions of the genome, such as genes. However, it is generally the case that the aggregative measure would be a mixture of causal and neutral variants, and these omnibus tests do not directly provide any indication of which RVs may be driving a given association. Recently, Bayesian variable selection approaches have been proposed to identify RV associations from a large set of RVs under consideration. Although these approaches have been shown to be powerful at detecting associations at the RV level, there are often computational limitations on the total quantity of RVs under consideration and compromises are necessary for large‐scale application. Here, we propose a computationally efficient alternative formulation of this method using a probit regression approach specifically capable of simultaneously analyzing hundreds to thousands of RVs. We evaluate our approach to detect causal variation on simulated data and examine sensitivity and specificity in instances of high RV dimensionality as well as apply it to pathway‐level RV analysis results from a prostate cancer (PC) risk case‐control sequencing study. Finally, we discuss potential extensions and future directions of this work.

Abstract African American Hereditary Prostate Cancer Study (AAHPC) was developed as a national collaboration to explore the role of genetics in the causation of hereditary prostate cancer (HPC) in African American (AA) men. AAHPC is in partnership with the International Consortium for Prostate Cancer Genetics (ICPCG), which conducts collaborative studies of HPC genetics in multiplex families. As part of an ICPCG sequencing study of 539 affected individuals from 366 HPC pedigrees, we performed whole exome sequencing in 21 ICPCG AA families, of which there were 14 AAHPC affected men from 11 pedigrees. The combined ICPCG AA cohort consisted of N=26 affected members. Post-variant calling quality control (QC) was implemented using Golden Helix SVS 8 software with filters set for removal of variants with Read Depth < 10, Quality Score < 20, Quality Score: Read Depth Ratio < 0.5, Call Rate < 0.75. Variants were additionally filtered by MAF based on the NHLBI ESP650051-V2 exomes variant frequencies for the AA population using a MAF threshold of 5%. Following QC, 176/299 SNVs and 20/39 INDELs remained for further analysis. In these analyses, we focused on 13 known cancer genes (MSR1, AR, BRCA1, BRCA2, BTNL2, EPHB2, CDH1, RNASEL, ELAC2, HOXB13, CHEK2, TP53 and NBN). Three sequenced families had > 1 affected members sequenced (2 or 3 per family) and the remaining 18 families had one member sequenced. Under the dominant model, our preliminary results show that no rare variants in the 13 candidate genes were found in 3/3 affecteds in two families. Rare SNVs in seven candidate genes (AR, CDH1, ELAC2, HOXB13, RNASEL, BRCA2 and EPHB2) were found in 2/3 affecteds for two families and 2/2 affecteds in one family. Several of the remaining 18 affected men (1 sequenced per pedigree) shared the same rare SNV in these candidate genes. For INDELs, rare variants in three candidate genes were found in pedigrees with ≥ 2 affecteds. Several of the remaining 18 affected men (one sequenced per pedigree) shared the same rare INDEL. Additional QC is underway to validate these variants and bioinformatic analyses are being used to predict effects of the variants in an effort to unravel the complex genetic heterogeneity of HPC in AA. Citation Format: Cheryl D. Cropp, Shannon K. McDonnell, Sumit Middha, Melissa DeRycke, Danielle M. Karyadi, Daniel Schaid, Stephen N. Thibodeau, William B. Isaacs, Elaine A. Ostrander, Janet Stanford, Kathleen A. Cooney, Joan E. Bailey-Wilson, John D. Carpten. Rare variant discovery in known cancer genes from whole-exome sequencingof African American hereditary prostate cancer families. [abstract]. In: Proceedings of the Eighth AACR Conference on The Science of Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; Nov 13-16, 2015; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2016;25(3 Suppl):Abstract nr B40.

Abstract Prostate cancer (PCa) will impact one in six American men and many low-risk men endure therapy, while other men die of PCa. Optimal treatment selection depends on distinguishing indolent tumors from those that will cause prostate cancer specific mortality (PCSM). To date, sequencing of prostate tumor DNA has highlighted the fact that prostate tumors develop differently than other cancers. Instead of recurrent protein-altering mutations, many prostate tumors harbor “closed chain” structural rearrangements that may mediate tumor evolution. However, we do not yet understand which somatic mutations are involved in aggressive disease and PCSM. To address this crucial question, we are using whole genome sequencing to pinpoint somatic genetic features in tumor/normal paired samples from four PCSM and six non-PCSM patients with treatment-naïve, Gleason 8+ tumors. Tumor and normal DNA were sequenced to a mean coverage of 95.3X and 48.3X, respectively. Tumor DNA purity ranged from 49%-79% and tumor ploidy ranged from 1.96 to 2.22. To define the genomic landscape of aggressive PCa, we will compare our findings with published studies and genotype high-priority candidates in at least 50 additional low- and high-grade tumors. We will also compare somatic mutations in PCSM versus non-PCSM cases and use network analysis to identify pathways that may be involved in progression and metastasis. Finally, we will characterize the clonal evolution of PCa tumors using deep digital sequencing to distinguish mutations that were present in the original tumor clone from subclonal mutations that arose after oncogenesis. In this way, we aim to identify potential oncogenic drivers and candidate mediators of progression and metastasis. Ultimately, we aim to 1) Uncover somatic mutations that are specific to aggressive tumors 2) Determine whether PCSM tumors have unique genetic features, and 3) Understand the role of clonal evolution in oncogenesis and progression. As more tumors are sequenced, this approach may yield valuable markers for diagnosis, prognosis and treatment selection. Note: This abstract was not presented at the meeting. Citation Format: Brennan J. Decker, Danielle M. Karyadi, Eric Karlins, Brian W. Davis, Lori S. Tillmans, Stephen N. Thibodeau, Elaine A. Ostrander. Whole genome sequencing of high-grade treatment-naïve prostate tumors. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1536. doi:10.1158/1538-7445.AM2014-1536

Abstract Background: Genome-wide association studies have identified multiple genetic variants associated with prostate cancer risk which explain a substantial proportion of familial relative risk. These variants can be used to stratify individuals by their risk of prostate cancer. Methods: We genotyped 25 prostate cancer susceptibility loci in 40,414 individuals and derived a polygenic risk score (PRS). We estimated empirical odds ratios (OR) for prostate cancer associated with different risk strata defined by PRS and derived age-specific absolute risks of developing prostate cancer by PRS stratum and family history. Results: The prostate cancer risk for men in the top 1% of the PRS distribution was 30.6 (95% CI, 16.4–57.3) fold compared with men in the bottom 1%, and 4.2 (95% CI, 3.2–5.5) fold compared with the median risk. The absolute risk of prostate cancer by age of 85 years was 65.8% for a man with family history in the top 1% of the PRS distribution, compared with 3.7% for a man in the bottom 1%. The PRS was only weakly correlated with serum PSA level (correlation = 0.09). Conclusions: Risk profiling can identify men at substantially increased or reduced risk of prostate cancer. The effect size, measured by OR per unit PRS, was higher in men at younger ages and in men with family history of prostate cancer. Incorporating additional newly identified loci into a PRS should improve the predictive value of risk profiles. Impact: We demonstrate that the risk profiling based on SNPs can identify men at substantially increased or reduced risk that could have useful implications for targeted prevention and screening programs. Cancer Epidemiol Biomarkers Prev; 24(7); 1121–9. ©2015 AACR.

Abstract A recent genome-wide association study found that genetic variants on chromosomes 3, 6, 7, 10, 11, 19 and X were associated with prostate cancer risk. We evaluated the most significant single-nucleotide polymorphisms (SNP) in these loci using a worldwide consortium of 13 groups (PRACTICAL). Blood DNA from 7,370 prostate cancer cases and 5,742 male controls was analyzed by genotyping assays. Odds ratios (OR) associated with each genotype were estimated using unconditional logistic regression. Six of the seven SNPs showed clear evidence of association with prostate cancer (P = 0.0007-P = 10−17). For each of these six SNPs, the estimated per-allele OR was similar to those previously reported and ranged from 1.12 to 1.29. One SNP on 3p12 (rs2660753) showed a weaker association than previously reported [per-allele OR, 1.08 (95% confidence interval, 1.00-1.16; P = 0.06) versus 1.18 (95% confidence interval, 1.06-1.31)]. The combined risks associated with each pair of SNPs were consistent with a multiplicative risk model. Under this model, and in combination with previously reported SNPs on 8q and 17q, these loci explain 16% of the familial risk of the disease, and men in the top 10% of the risk distribution have a 2.1-fold increased risk relative to general population rates. This study provides strong confirmation of these susceptibility loci in multiple populations and shows that they make an important contribution to prostate cancer risk prediction. (Cancer Epidemiol Biomarkers Prev 2008;17(8):2052–61)

Abstract Background: Family-based linkage studies, association studies, and studies of tumors have highlighted human chromosome 8q as a genomic region of interest for prostate cancer susceptibility loci. Recently, a locus at 8q24, characterized by both a single nucleotide polymorphism (SNP) and a microsatellite marker, was shown to be associated with prostate cancer risk in Icelandic, Swedish, and U.S. samples. Although the data were provocative, the U.S. samples were not population based, which precludes assessment of the potential contribution of this locus to prostate cancer incidence in the United States. Methods: We analyzed both markers in a population-based, case-control study of middle-aged men from King County, Washington. Results: Overall, there was a significant positive association between the A allele of the SNP rs1447295 and prostate cancer risk [odds ratio, 1.4; 95% confidence interval (95% CI), 1.1-2.0] but no significant association with the microsatellite DG8S737. However, significant associations were observed for both markers in men with high Gleason scores. Adjusting for age, first-degree family history of prostate cancer, and prostate cancer screening history, the adjusted odds ratios were 1.4 (95% CI, 1.1-1.8) for the A allele of the SNP and 1.9 (95% CI, 1.2-2.8) for the −10 allele of the microsatellite. Conclusions: These data suggest that the locus on chromosome 8q24 harbors a genetic variant associated with prostate cancer and that the microsatellite marker is a stronger risk factor for aggressive prostate cancers defined by poorly differentiated tumor morphology. (Cancer Epidemiol Biomarkers Prev 2007;16(4):809–14)

Abstract Prostate cancer is a leading and increasingly prevalent cause of cancer death in men. Whereas family history of disease is one of the strongest prostate cancer risk factors and suggests a hereditary component, the predisposing genetic factors remain unknown. We first showed that KLF6 is a tumor suppressor somatically inactivated in prostate cancer and since then, its functional loss has been further established in prostate cancer cell lines and other human cancers. Wild-type KLF6, but not patient-derived mutants, suppresses cell growth through p53-independent transactivation of p21. Here we show that a germline KLF6 single nucleotide polymorphism, confirmed in a tri-institutional study of 3,411 men, is significantly associated with an increased relative risk of prostate cancer in men, regardless of family history of disease. This prostate cancer–associated allele generates a novel functional SRp40 DNA binding site and increases transcription of three alternatively spliced KLF6 isoforms. The KLF6 variant proteins KLF6-SV1 and KLF6-SV2 are mislocalized to the cytoplasm, antagonize wtKLF6 function, leading to decreased p21 expression and increased cell growth, and are up-regulated in tumor versus normal prostatic tissue. Thus, these results are the first to identify a novel mechanism of self-encoded tumor suppressor gene inactivation and link a relatively common single nucleotide polymorphism to both regulation of alternative splicing and an increased risk in a major human cancer.