Genome-wide copy number analysis reveals candidate gene loci that confer susceptibility to high-grade prostate cancer

Highlights

• CNVs can be used as a means for PCa prognostics.

• Approximately 70% of the identified regions were CN gains while another 30% were CN losses.

• Three out of 5 putative CNVs identified in PCa were rare (1q21.3, 15q15, and 7p12.1).

• A novel CNV region at 12q23.1 found to be unique to patients with PCa was identified.

• A deletion event important in Wnt Signaling was observed at region 8p11.21.

Abstract

Background

Two key issues in prostate cancer (PCa) that demand attention currently are the need for a more precise and minimally invasive screening test owing to the inaccuracy of prostate-specific antigen and differential diagnosis to distinguish advanced vs. indolent cancers. This continues to pose a tremendous challenge in diagnosis and prognosis of PCa and could potentially lead to overdiagnosis and overtreatment complications. Copy number variations (CNVs) in the human genome have been linked to various carcinomas including PCa. Detection of these variants may improve clinical treatment as well as an understanding of the pathobiology underlying this complex disease.

Methods

To this end, we undertook a pilot genome-wide CNV analysis approach in 36 subjects (18 patients with high-grade PCa and 18 controls that were matched by age and ethnicity) in search of more accurate biomarkers that could potentially explain susceptibility toward high-grade PCa. We conducted this study using the array comparative genomic hybridization technique. Array results were validated in 92 independent samples (46 high-grade PCa, 23 benign prostatic hyperplasia, and 23 healthy controls) using polymerase chain reaction–based copy number counting method.

Results

A total of 314 CNV regions were found to be unique to PCa subjects in this cohort (P<0.05). A log₂ ratio-based copy number analysis revealed 5 putative rare or novel CNV loci or both associated with susceptibility to PCa. The CNV gain regions were 1q21.3, 15q15, 7p12.1, and a novel CNV in PCa 12q23.1, harboring ARNT, THBS1, SLC5A8, and DDC genes that are crucial in the p53 and cancer pathways. A CNV loss and deletion event was observed at 8p11.21, which contains the SFRP1 gene from the Wnt signaling pathway. Cross-comparison analysis with genes associated to PCa revealed significant CNVs involved in biological processes that elicit cancer pathogenesis via cytokine production and endothelial cell proliferation.

Conclusion

In conclusion, we postulated that the CNVs identified in this study could provide an insight into the development of advanced PCa.

Introduction

Prostate cancer (PCa) is a slow growing, noncutaneous cancer that continues to be the most common malignancy and the second leading cause of cancer-related deaths among men in developed countries [1]. Cancer registry in Malaysia ranks PCa as the fourth most common cancer among men [2]. However, the PCa prevalence in Asians has been reported to be 20 times lower than in Western countries [3]. Although substantial research on PCa has been conducted, the exact etiology and pathogenesis of PCa have yet to be fully elucidated. To date, only prostate-specific antigen (PSA) test in conjunction with digital rectal examination has been widely used in clinical practice to identify patients who require transrectal ultrasound biopsy for detection of PCa [4]. Nonetheless, PSA test comes with well-known limitations. Although PSA test contributes toward reducing PCa-related death by 20%, overdiagnosis has been a huge concern, as it is often not possible to differentiate a benign condition from that of a cancerous one [5]. Therefore, histopathology remains the gold standard for PCa diagnosis.

Cancer genetics can offer ways to identify predictive markers for cancers, and such efforts have been deemed successful. This is proven by mounting evidence from studies conducted in hereditary breast and ovarian cancers regarding the BRCA1 and BRCA2 genes [6], the KRAS, p53, and epidermal growth factor receptor (EGFR) genes being notable biomarkers in colorectal, esophageal, liver, and pancreatic cancers [7], respectively; hypermethylation of the DAP-kinase in bladder cancer [8]; and mutation or polymorphisms in SRD5A2 that contribute toward androgen metabolism in PCa [9], [10]. A deletion copy number variation (CNV) at 20p13 is associated with aggressiveness of PCa in a genome-wide study [11]. CNVs have also been characterized in several other PCa-related studies [12], [13], [14]. Nonetheless, findings of these studies have been inconclusive due to ethnicity, lifestyle, and environment differences. Unlike familial breast cancer, in which high penetrance of the BRCA1 mutation results in 80% lifetime risk of developing breast cancer among individuals with the mutation, PCa on the contrary, is a complex disease with multiple risk factors involved and reduced penetrance, hence providing challenges in predicting clinical outcomes.

In this study, we compared patients with PCa with high-grade (Gleason 8–10) cancer, the stage in which cells are moderately to poorly differentiated and which would be a more suitable group to assess the prognosis of PCa [15] to benign prostatic hyperplasia (BPH). PCa of lower stages were excluded from the study owing to possibilities of coexistence with BPH because lower-grade PCa often resembles foci of BPH [16]. Here, we described a pilot study designed to detect somatic CNV loci in Chinese patients within the Malaysian population, with high-grade PCa. Although the literature is mostly focused on the Western population owing to its high disease prevalence, relatively little is known about this condition in Asians. In addition, the Malaysian Chinese differ in the population structure from that of the Chinese Han Beijing, commonly reported in the HapMap [17].