An exome-wide rare variant analysis of korean men identifies three novel genes predisposing to prostate cancer

ABSTRACT Since prostate cancer is highly heritable, common variants associated with prostate cancer have been studied in various populations, including those in Korea. However, rare and

low-frequency variants have a significant influence on the heritability of the disease. The contributions of rare variants to prostate cancer susceptibility have not yet been systematically

evaluated in a Korean population. In this work, we present a large-scale exome-wide rare variant analysis of 7,258 individuals (985 cases with prostate cancer and 6,273 controls). In total,

19 rare variant loci spanning 7 genes contributed to an association with prostate cancer susceptibility. In addition to replicating previously known susceptibility genes (e.g., _CDYL2_,

_MST1R_, _GPER1_, and _PARD3B_), 3 novel genes were identified (_FDR q_ < 0.05), including the non-coding RNAs _ENTPD3-AS1_, _LOC102724438_, and protein-coding gene _SPATA3_.

Additionally, 6 pathways were identified based on identified variants and genes, including estrogen signaling pathway, signaling by MST1, IL-15 production, MSP-RON signaling pathway, and

IL-12 signaling and production in macrophages, which are known to be associated with prostate cancer. In summary, we report novel genes and rare variants that potentially play a role in

prostate cancer susceptibility in the Korean population. These observations demonstrated a path towards one of the fundamental goals of precision medicine, which is to identify biomarkers

for a subset of the population with a greater risk of disease than others. SIMILAR CONTENT BEING VIEWED BY OTHERS ASSESSING THE CONTRIBUTION OF RARE PROTEIN-CODING GERMLINE VARIANTS TO

PROSTATE CANCER RISK AND SEVERITY IN 37,184 CASES Article Open access 19 February 2025 TRANS-ANCESTRY GENOME-WIDE ASSOCIATION META-ANALYSIS OF PROSTATE CANCER IDENTIFIES NEW SUSCEPTIBILITY

LOCI AND INFORMS GENETIC RISK PREDICTION Article 04 January 2021 USING GENE AND GENE-SET ASSOCIATION TESTS TO IDENTIFY LETHAL PROSTATE CANCER GENES Article 17 August 2024 INTRODUCTION

Prostate cancer is a common malignancy of a gland in the male reproductive system. It is the fifth leading cancer diagnosed and the seventh leading cause of cancer deaths in Korean men1. The

overall mortality rate in Korean men due to cancer was 188.7 per 100,000 and 6.6 per 100,000 for prostate cancer in 20142. Additionally, the prevalence, incidence, and mortality of prostate

cancer in Korean men have increased significantly in the past few years1. Heritable genetic factors contribute to the susceptibility of various cancers and the genetic attribution of the

incidence of prostate cancer is more than any other cancer type3. Twin studies have shown that genetic factors contribute to 42% of the incidence of prostate cancer3. Furthermore, first

degree relatives are known to have two to three-fold increased risk of developing prostate cancer4. These observations indicate that germline variants contribute to prostate cancer. However,

the identification of germline factors involved in prostate cancer has been limited in scope, and it is still a subject of ongoing research. In recent years, there has been a push to

discover common and rare variants associated with prostate cancer among different ethnicities, as growing evidence suggests that germline factors associated with prostate cancer

susceptibility may differ among different ethnicities5. Prostate cancer is the second leading cause of cancer mortality in American men6. The prevalence and mortality rates of prostate

cancer differ across European, African, and Asian ethnic groups6. In particular, the incidence rates are particularly high in African American men and substantially low in Japanese and

mainland Chinese population6. However, the incidence is higher among immigrant Japanese in the United States as compared Japanese living in Japan, but it is still about half that of American

European population6. Thus, it is evident that germline factors associated with prostate cancer differ among ethnicities; therefore, in this study, we aimed to find the germline variants

specific to Korean population that are associated with prostate cancer. With the advent of high throughput genotyping technologies in the past few years, it has become easier to sequence

thousands of samples. Many genome-wide association studies (GWAS) have been conducted to identify common variants associated with prostate cancer7,8,9. However, European populations are

better characterized relative to Asian populations8. For instance, more than 100 loci have been identified in GWAS studies using European cohorts7,8, while studies that have performed GWAS

using participants of Chinese and Japanese ethnicities have identified only 12 significant loci associated with prostate cancer8. Another GWAS study using common variants identified 5

significant loci associated with prostate cancer in a Korean population9. However, the common variants discovered to date explain only a small portion of heritability of prostate cancer,

thus leaving the majority of genetic risk unexplained10. Most association studies on prostate cancer to date have focused mainly on common variants. However, a proportion of the missing

heritability in prostate cancer could be further explained through low-frequency and rare variants. Rare variants play a key role in the contribution to heritability among different cancer

types10,11. Many exome-wide studies have shown rare variants associated with susceptibility genes in colorectal cancer12, breast cancer13, prostate cancer14, and endometrial cancer15.

However, very few rare variant studies have been conducted in Asian populations and even fewer in the Korean population. The variants discovered using common and rare variants are rarely

replicated among European, Chinese, Japanese, and Korean population studies7,8,9,16. Thus, it is important to conduct separate association studies on ethnic populations to improve our

understanding of the heritability of prostate cancer among population subsets. To this end, we analyzed whole-exome array data of germline samples from Korean individuals with and without

prostate cancer. Rare variants were collapsed into genes and pathways across the genome and tested for association with prostate cancer. The significant variants in the gene or pathway were

further annotated and evaluated for their association with prostate cancer. RESULTS STUDY DESIGN AND QUALITY CONTROL Case population comprising 1,008 patients and a control group of 6,438

patients were obtained from the Korean Association Resource (KARE) study, which is a part of the Korean Genome and Epidemiology Study (KoGES)17. A summary of patient demographics for cases

and controls is described in Table 1. All patients were men and of Korean ethnicity. The average ages of case and control patients were 67.43 and 54.39 years, respectively. The average body

mass indices (BMI) of cases and controls were similar. The quality control (QC) was performed to filter out bad samples and markers. All samples with marker call rate < 99%, sample call

rate < 99%, and samples which were closely related (based on identity by descent (IBD) cutoff of 0.125) were dropped from the further analysis. After QC, 7,258 samples included 985 cases

and 6,273 controls. The detailed steps involved in quality control are shown in Supplementary Figs 1 and 2. Using 71,270 variants that passed QC filters, rare variants with minor allele

frequency (MAF) < 0.05 were binned into genes and association tests were performed. Further, the genomic inflation rate was calculated. A high genomic inflation rate usually indicates

population substructure in data and spurious associations in the results. We found the genomic inflation factor _λ_1000 to be 1.19618 for an equivalent study of 1,000 cases and 1,000

controls (Supplementary Fig. 3). Finally, 7,258 samples with 71,270 variants that passed QC were used for subsequent statistical analysis, as described in the Methods section. IDENTIFICATION

OF SIGNIFICANT GENES ASSOCIATED WITH PROSTATE CANCER The gene-based rare variant analysis was performed to identify associations between rare variants and prostate cancer. After binning

variants into 5,830 gene bins, seven genes were identified to be associated with prostate cancer (FDR < 0.05 (Fig. 1 and Table 2)). In total, 19 rare variant loci spanning across 7 genes

contributed to an association with prostate cancer susceptibility (Supplementary Table 1). Three genes, including _MST1R_, _GPER1_, and _PARD3B_, have been previously implicated in prostate

cancer19,20,21,22,23 and _CDYL2_ in breast cancer24. Supplementary Fig. 4 shows regional plots for the genes with variants and exon span. Further, the distribution of MAF for the significant

variants in this study population; Northeast Asians reference panel (NARD)25, which consists of the population from Korea (N = 850), Mongolia (N = 386), Japan, China, and Hong Kong; and

gnomeAD26 are shown in Supplementary Table 2. The samples that had more than one rare variant among significant locus are shown in Supplementary Fig. 5. Further, the rare association tests

were rerun by removing one variant at a time from the gene/bin to elucidate the significance of that particular variant in the gene/bin (Supplementary Table 1). A decrease in the

significance of the gene (increase in p-value) represents a significant contribution of the variant while maintaining a significant association suggests that the contribution of the variant

is insignificant. In case of _MST1R_ and _LOC102724438_, when the stop gained variant, rs200626206 was removed and the gene was insignificant (_p__rm_ > 0.05), indicating that it is the

most significant variant in the gene/bin, as most of the signal detected in the genes is attributed to rs200626206 loci (Supplementary Table 1). Nonsense mutations often produce

nonfunctional protein products due to premature termination of translation, thus the significant contribution of rs200626206 loci to the signal was expected. The variants rs181756759 and

rs201829385 in genes _GPER1_ and _PARD3B_, respectively, have insignificant _p_rm and are predicted to be ‘probably damaging’ with Polyphen score > 0.95 (Supplementary Table 1). Thus,

there is a high possibility that these variants result in partial or complete loss of protein function27,28. VARIANT ANNOTATION To characterize the clinical significance, effect of variants

on the protein, and implications in human inherited diseases, the variants within significantly associated genes were annotated using ClinVar and Variant Effect Predictor (VEP). None of the

significant variants binned were found in ClinVar. VEP annotated the variants concerning their effect on the coding region (Supplementary Table 1). VEP also annotates variants by their

potential influence on protein sequence (e.g., high, moderate, and low). High impact variants have a disruptive effect on proteins, such as protein truncation and loss of function. Moderate

effect variants are non-disruptive but can change protein sequence while low effect variants are unlikely to change protein behavior. Here, VEP categorized 16 variants as a moderate effect,

2 variants as high effect, and 1 variant as a modifier (Supplementary Table 1). Therefore, most rare variants within the coding regions of genes associated with prostate cancer possibly

influence protein function. IDENTIFICATION OF SIGNIFICANT PATHWAYS ASSOCIATED WITH PROSTATE CANCER Pathway analysis was performed using variants and genes discovered by gene-based rare

variant association test. The biological pathways were derived using Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity), and 4 canonical pathways were

identified, including IL-15 Production, MSP-RON Signaling Pathway, Sperm Motility and IL-12 Signaling, and Production in Macrophages (Table 3). IL-15 Production, MSP-RON Signaling. Further,

pathway enrichment test was also run using EnrichR, which identified estrogen signaling pathway, signaling by MST1 pathway, and four Gene Ontology (GO) molecular function ontologies29,30.

DISCUSSION In this work, an exome-wide rare variant analysis study was performed, and 19 novel low-frequency variants in 7 genes were identified to be associated with prostate cancer in

Korean men. Since common variants alone do not completely explain the heritability of prostate cancer, integrative analyses of rare variants across the genome can provide us with a new

understanding of prostate cancer heritability. Though common variants have been studied previously in the Korean population, this is the first exome-wide study of rare variants associated

with prostate cancer in the Korean population. The genes and variants discovered in this study can potentially help in early diagnostic and understanding of carcinogenesis in prostate

cancer. The rare variants were binned into genes and association tests were performed across the genes. Binning of rare variants increases statistical power to detect rare variant

associations31 and helps interpret the effect of rare variants on the prognosis and progression of prostate cancer. Several genes identified in this study were previously implicated with

prostate cancer and other cancers. _In-vivo_ studies have shown an association of _MST1R_, _GPER1_, and _PARD3B_ with prostate cancer19,20,21,22,23. Another gene we found, _CDYL2_, has

common variants that are implicated in breast cancer24. Since three of the genes discovered in this study have been previously validated to be associated with prostate cancer, it would

appear that our analysis predicted true associations. Angiogenesis, cell survival, and cell proliferation are hallmarks of cancer32. One of the genes that we found was significantly

associated with prostate cancer, _MST1R_ (Macrophage Stimulating 1 Receptor) or _RON_, is overexpressed in prostate cancer and various other cancers33. _RON_ is known to be overexpressed in

breast cancer and bladder cancer and is associated with poor clinical outcome33. _In vivo_ study using genetically engineered mouse model has shown that the _RON_ receptor plays a functional

role in prostate tumor and that deficient _Ron_ receptor signaling is associated with smaller tumor size34. Another study on stromal cells of the prostate tumor using mouse model showed

that loss of _Ron_ in tumor-associated macrophages inhibits cancer cell growth19. Another gene significantly associated with prostate cancer, _GPER1_ (G Protein-Coupled Estrogen Receptor 1)

or _GPR30_, is known to regulate cell growth by non-genomic signaling of estrogen35. _GPER1_ is also known to stimulate cell proliferation in breast, endometrial, ovarian, and thyroid cancer

cells by rapid but transient activation of Erk1/235. Besides, in the case of prostate cancer, _GPER1_ is known to control cancer cell growth through _GPER1_ mediated pathways35. The

presence of alternate allele in rs11544331, one of the rare variant loci binned in _GPER1_, is known to result in the expression of P16L variant of _GPER1_36. The substitution of proline

with leucine at position 16 of the _GPER_ protein sequence blocks _GPER_ from being glycosylated and causes it to localize to the nucleus, although typically it should localize outside of

nucleus36. The _P16L_ in the nucleus may also regulate transcription of cancer-relevant genes and migration of carcinoma cells36. Of the genes identified in this study, _PARD3B_

(Partitioning defective 3 homolog) plays an essential role in asymmetric cell division, polarized growth, and maintenance of cell-polarity37. Mutational inactivation of its homolog gene

_PARD3_ is known to cause carcinogenesis in prostate cancer38. _In vivo_ studies have shown that downregulation of Par3 in breast cancer induces cell invasion and metastasis by decreasing

cell-cell cohesion in a Tiam1/Rac-GTP pathway-dependent manner23. Higher expression of _PARD3B_ is associated with colorectal cancer malignancy and poor survival, as _PARD3B_ inhibits

Lkb1/AMPK signaling pathway and its knockout induces apoptosis and reduces proliferation, supporting its role in colorectal cancer cell survival39,40. Additionally, a previous genome-wide

association study found rs2335704, which resides in _PARD3B_, to be associated with tuberculosis41. Another gene, _CDYL2_ (Chromodomain Y Like 2 or Prostate Cancer Candidate Protein 1), is

involved in catalytic activity, protein binding, and methylated histone binding42. Genome-wide studies using common variants have identified loci in _CDYL2_ associated with breast cancer43.

_ENTPD3-AS1_ is a long non-coding RNA that we found was significantly associated with prostate cancer. A locus (rs193921050) in _ENTPD3-AS1_ has been reported for ‘Malignant tumor of

prostate’ in ClinVar with uncertain clinical significance and review status of 0/444. The mutation on the locus was discovered in somatic tissue but was not found to be mutated at a

significantly higher rate relative to the background mutation rate45. In addition to the genes, the estrogen signaling pathway was significantly enriched using WikiPathways and KEGG pathway.

Evidence suggests that prostate carcinogenesis and progression involves local estrogen signaling mechanisms46,47. Further, signaling by MST1 pathway was enriched using the Reactome pathway

database. An _in vitro_ study showed that MST1 suppressed prostate cancer growth48. Moreover, MST1 is the key kinase component of the Hippo-YAP pathway, which restricts prostate cancer

progression by interacting with multiple molecular pathways49. The Ingenuity pathway analysis revealed three significant pathways. IL-15 production pathway, one of the significant pathways,

is known to be associated with prostate cancer. The expression of IL-15 is known to decrease the migration, invasion, and angiogenesis but increase tumor volume by increasing lipid

deposition and inflammation in prostate cancer50. IL-15 also alters the expression of genes involved in cell death and immune response50. Vaccinations using IL-15 are effective in

up-regulating immune responses, reducing invasion, and improving survival51. Another pathway, MSP-RON signaling pathway, has been previously known to be associated with many cancer types,

including prostate cancer, and has been extensively studied _in vivo_ and _in vitro_. The MSP-RON signaling generates oncogenic variants and activates downstream pathways, resulting in

tumorigenesis, proliferation, angiogenesis, invasion, and resistance to chemotherapy52. Loss of RON in myeloid cells has been shown to reduce prostate cancer growth in mice models52. IL-12

signaling is anti-carcinogenic, and IL-12 deficiency in mice is known to induce the development of spontaneous tumors and promote their growth53. Further, the GO molecular function ontology

enrichment indicated four significantly enriched molecular function ontologies, including Mitogen-activated protein kinase (MAPK) activity, Transmembrane receptor protein tyrosine activity,

Transmembrane receptor protein kinase activity, and Phosphatidylinositol binding. All the ontologies discovered are known to be associated with prostate cancer33,54,55. Even though some of

the genes we found have already been implicated, we found 19 novel variants and 3 novel genes that are associated with prostate cancer. Moreover, all the pathways that were found to be

associated have been well studied and have been found to play a key role in cancer. All the variants discovered were missense and stop gained, except one variant in an intron, as categorized

by VEP. Variation in amino acid sequence could potentially affect stability, conformational dynamics, drug response, and other protein properties that could result in a pathological

condition and increased susceptibility to disease56. Many variants were also filtered out when re-clustering and filtering were performed using CHARGE criteria, a more relaxed approach shown

in Park _et al_., which could be used to preserve more variants while maintaining the genotyping accuracy of common and rare variants57. Further studies would be required to validate our

findings, as variants discovered in this study were not discovered in the European population studies58. This is probably due to different genetic factors affecting prostate cancer

susceptibility among different ethnic groups5. Additional exploration of the molecular mechanisms is required to understand the exact role of the variants in prostate cancer. Further studies

are also required to elucidate the role of lifestyle/environment, especially dietary factors, in the Korean population, as they have been previously shown to be associated with prostate

cancer4. In conclusion, we found novel genes and rare variants that are associated with prostate cancer in the Korean population, revealing potential biomarkers for prostate cancer that are

unique to Korean ethnicity. They could also help us explain the missing heritability in prostate cancer, which could be applied in the field of precision medicine. METHODS SAMPLES AND DATA

SET Between November 2003 and July 2013, we prospectively recruited 1,008 prostate cancer patients from a single tertiary hospital, Seoul National University Bundang Hospital, and conducted

a case-control study that included 6,438 age-matched controls from the Korean Association Resource (KARE) study, which was a part of the Korean Genome and Epidemiology Study (KoGES)17. Blood

specimens were prospectively collected throughout the course of the study from all of the prostate cancer patients. The automatic firing mechanism was used to perform transrectal

ultrasound-guided multi-core (≥12) biopsies bilaterally near the base, mid-gland, and apex, with at least six biopsies per side. A total of 12 baseline biopsy cores were taken from all of

the men, and additional biopsies of suspicious lesions were obtained if needed. Further, 820 patients among the study population who had prostate cancer were treated with radical

prostatectomy (RP) in the same hospital. The genotyping was done using the blood samples collected. EXOME CHIP The Illumina HumanExome BeadChip 12v1-1 system provides 242,901 variants

selected over 12,000 individual human exome and whole-genome sequences representing diverse populations and ethnicities. The chip focuses on protein-altering variants. A more detailed

explanation is available at http://genome.sph.umich.edu/wiki/Exome_Chip_Design. GENOTYPING AND QUALITY CONTROL The datasets were generated using Illumina HumanExome BeadChip 12v1-1. We used

Illumina’s GenTrain version 2.0 clustering algorithm with the GenomeStudio software (V2011.1) for genotype calling. The genotype calling for the exome chip was performed following the best

practices defined in Grove _et al_.59. We performed manual re-clustering and visual inspection using CHARGE clustering method59 to improve the accuracy of variant calling (Supplementary Fig.

 6). A separate study on the quality of variants showed that re-clustering using CHARGE criteria on KoGES dataset with more relaxed cutoffs has 99.9% concordance rate for rare variants with

whole-exome sequencing data, which indicates that rare variant calls are robust57. Quality control filters were applied to both case and control datasets separately. Since the number of

variants differed in the datasets, only common variants between cases and controls were selected for the analysis. The datasets were merged after QC, and filters were again applied to the

merged dataset. As a part of the quality control, sample call rates, marker call rates, and sample relatedness were checked60. The palindromic SNPs and SNPs with indels were removed. The

samples and markers with a call rate of less than 99% were removed. Identity by descent (IBD) was calculated using plink, and IBD threshold of 0.125 was used to remove related samples. The

detailed quality control steps are shown in Supplementary Figs 1 and 2. The final merged dataset had 7,258 samples with 985 cases and 6,273 controls and 71,270 variants. The dataset was

checked for batch effects. Since datasets from different sources were merged, population stratification in the data could have occurred60. Principal Component Analysis was performed using

SMARTPCA61 on the dataset after the LD pruning using plink option ‘-indep-pairwise 50 5 0.2’ and removing all SNPs with MAF < 0.05. PCA was performed to check case and control sample

clusters (Supplementary Fig. 7). PCA was also performed by projecting onto 1,000 genomes data. The case and controls clustered together around the South Asian population, as shown in

Supplementary Fig. 8. Further, quantile-quantile (Q-Q) plot was drawn using SKAT-O p-values to check for inflation (Supplementary Fig. 3). RARE VARIANT GENE-BASED ASSOCIATION TEST The rare

variant analysis was performed using BioBin (https://ritchielab.psu.edu/software/biobindownload), a tool that can be used to perform rare variant burden tests31. BioBin bins all variants

into gene bins and variants outside genes into intergenic region bins. Subsequently, SKAT-O was used to test for statistical significance of associations62. SKAT-O increases statistical

power by optimally combining burden and dispersion (SKAT) tests and adaptively applying them62. Since rare variants are statistically underpowered for the association test, binning of rare

variants by biologically informed units, such as gene or pathway, increases statistical power to detect rare variant associations by increasing the composite allele frequency and reducing

the degrees of freedom31. BioBin is configured by default to bin all variants with minor allele frequency (MAF) below 5%. Library of Knowledge Integration (LOKI) is a database of genomic

locations of SNPs and genes as well as known relationships among genes and proteins, such as interaction pairs, pathways, and ontological categories integrated from various disparate data

sources31. LOKI provides prior knowledge to BioBin31. All variants with MAF < 0.05% were removed. Variants with MAF > 0.05% in case population or control population were included, and

only genes with at least 2 variants were tested. Age and first 5 principal components were incorporated as covariates to adjust for age and population stratification. The first 5 principal

components that defined maximum variance were selected, as shown in Supplementary Fig. 9. The weight-loci argument was used to add Madsen & Browning weights to each locus63. BioBin

creates bin based on gene regions when the bin-regions argument is set using gene information from LOKI. The false discovery rate (FDR) correction was applied to adjust for multiple testing.

Any FDR adjusted _q-value_ < 0.05 was considered significant. The rare association tests were run again by removing one variant at a time from the significant bins to ascertain the

significance of the variant. The higher the SKAT-O _p-value_ (_p_rm), the more significant is the contribution of the variant in the bin. VARIANT ANNOTATION ClinVar is a public archive that

connects human variation to phenotypes, clinical significance, relationship to human health, and other supporting data obtained through submissions by various groups44. These are aggregated

to reflect both consensus and conflicting assertions44. Variant effect predictor (VEP) provides information about the variants’ location, gene/transcript affected by variants, types of

mutation (i.e., stop gained, missense, stop-lost, and frameshift), and protein change scores, which indicate possible partial/complete loss of function of the protein due to amino acid

substitution. All the variants in significantly associated genes were annotated using ClinVar and VEP. PATHWAY ENRICHMENT ANALYSIS The significant variants and genes were used for pathway

enrichment analysis using Ingenuity pathway analysis and Enrichr. Enrichr is a web-based enrichment analysis tool which contains pathway and ontology libraries from various sources29,30. The

WikiPathways (2019), KEGG (2019), Reactome (2016), and GO molecular function (2018) libraries were used as part of Enrichr. ETHICS STATEMENT This study was approved by our institutional

review board (Seoul National University Bundang Hospital Institutional review board; IRB number, B-1312/232-302) and followed the rules stated in the Declaration of Helsinki. All

participants provided written informed consent. CHANGE HISTORY * _ 07 FEBRUARY 2020 An amendment to this paper has been published and can be accessed via a link at the top of the paper. _

REFERENCES * Han, H. H. _et al_. Epidemiology of prostate cancer in South Korea. _Prostate Int_ 3, 99–102, https://doi.org/10.1016/j.prnil.2015.06.003 (2015). Article  PubMed  PubMed Central

  Google Scholar  * Shin, H.-Y. _et al_. Cause-of-death statistics in the Republic of Korea, 2014. _J Korean Med Assoc_ 59, 221–232 (2016). Article  ADS  Google Scholar  * Lichtenstein, P.

_et al_. Environmental and Heritable Factors in the Causation of Cancer — Analyses of Cohorts of Twins from Sweden, Denmark, and Finland. _New England Journal of Medicine_ 343, 78–85,

https://doi.org/10.1056/NEJM200007133430201 (2000). Article  CAS  Google Scholar  * Grönberg, H. Prostate cancer epidemiology. _The Lancet_ 361, 859–864,

https://doi.org/10.1016/S0140-6736(03)12713-4 (2003). Article  Google Scholar  * Nakagawa, H. Prostate cancer genomics by high-throughput technologies: genome-wide association study and

sequencing analysis. _Endocr Relat Cancer_ 20, R171–R181, https://doi.org/10.1530/ERC-13-0113 (2013). Article  CAS  PubMed  Google Scholar  * Judd Moul, T. Z., Andrew A & Joseph L.

_Prostate Cancer_ (http://www.cancernetwork.com/cancer-management/prostate-cancer, 2015). * Al Olama, A. A. _et al_. A meta-analysis of 87,040 individuals identifies 23 new susceptibility

loci for prostate cancer. _Nat Genet_ 46, 1103–1109, https://doi.org/10.1038/ng.3094 (2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Wang, M. _et al_. Large-scale association

analysis in Asians identifies new susceptibility loci for prostate cancer. _Nature_ _Communications_ 6, https://doi.org/10.1038/ncomms9469 (2015). * Oh, J. J. _et al_. Exome-based

genome-wide association study and risk assessment using genetic risk score to prostate cancer in the Korean population. _Oncotarget_ 5, https://doi.org/10.18632/oncotarget.16540 (2017). *

Mancuso, N. _et al_. The contribution of rare variation to prostate cancer heritability. _Nat Genet_ 48, 30–35, https://doi.org/10.1038/ng.3446 (2016). Article  CAS  PubMed  Google Scholar 

* Wang, Y. _et al_. Rare variants of large effect in BRCA2 and CHEK2 affect risk of lung cancer. _Nat Genet_ 46, 736–741, https://doi.org/10.1038/ng.3002 (2014). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Gylfe, A. E. _et al_. Eleven Candidate Susceptibility Genes for Common Familial Colorectal Cancer. _PLOS Genetics_ 9, e1003876,

https://doi.org/10.1371/journal.pgen.1003876 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Thompson, E. R. _et al_. Exome Sequencing Identifies Rare Deleterious Mutations

in DNA Repair Genes FANCC and BLM as Potential Breast Cancer Susceptibility Alleles. _PLOS Genetics_ 8, https://doi.org/10.1371/journal.pgen.1002894 (2012). Article  CAS  PubMed  PubMed

Central  Google Scholar  * FitzGerald, L. M. _et al_. Germline Missense Variants in the BTNL2 Gene Are Associated with Prostate Cancer Susceptibility. _Cancer Epidemiol Biomarkers Prev_ 22,

1520–1528, https://doi.org/10.1158/1055-9965.EPI-13-0345 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Chen, M. M. _et al_. Exome-Wide Association Study of Endometrial

Cancer in a Multiethnic Population. _Plos One_ 9, e97045, https://doi.org/10.1371/journal.pone.0097045 (2014). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Ren, X.-Q. _et

al_. Variants on 8q24 and prostate cancer risk in Chinese population: a meta-analysis. _Int J Clin Exp Med_ 8, 8561–8570 (2015). CAS  PubMed  PubMed Central  Google Scholar  * Kim, Y. &

Han, B.-G. Cohort Profile: The Korean Genome and Epidemiology Study (KoGES) Consortium. _Int J Epidemiol_, https://doi.org/10.1093/ije/dyv316. Article  PubMed Central  Google Scholar  *

Freedman, M. L. _et al_. Assessing the impact of population stratification on genetic association studies. _Nat Genet_ 36, 388–393, https://doi.org/10.1038/ng1333 (2004). Article  CAS 

PubMed  Google Scholar  * Gurusamy, D. _et al_. Myeloid-Specific Expression of Ron Receptor Kinase Promotes Prostate Tumor Growth. _Cancer Res_ 73, 1752–1763,

https://doi.org/10.1158/0008-5472.CAN-12-2474 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Lee, H.-R., Hwang, K.-A. & Choi, K.-C. The estrogen receptor signaling

pathway activated by phthalates is linked with transforming growth factor-β in the progression of LNCaP prostate cancer models. _International Journal of Oncology_ 45, 595–602 (2014).

Article  CAS  PubMed  Google Scholar  * Nelles, J. L., Hu, W.-Y. & Prins, G. S. Estrogen action and prostate cancer. _Expert Rev Endocrinol Metab_ 6, 437–451,

https://doi.org/10.1586/eem.11.20 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar  * Zhang, K. _et al_. Shp2 promotes metastasis of prostate cancer by attenuating the

PAR3/PAR6/aPKC polarity protein complex and enhancing epithelial-to-mesenchymal transition. _Oncogene_ 35, 1271, https://doi.org/10.1038/onc.2015.184,

https://www.nature.com/articles/onc2015184#supplementary-information (2015). Article  PubMed  Google Scholar  * Xue, B., Krishnamurthy, K., Allred, D. C. & Muthuswamy, S. K. Loss of Par3

promotes breast cancer metastasis by compromising cell-cell cohesion. _Nat Cell Biol_ 15, 189–200, https://doi.org/10.1038/ncb2663 (2013). Article  CAS  PubMed  Google Scholar  *

Ghoussaini, M., Pharoah, P. D. P. & Easton, D. F. Inherited Genetic Susceptibility to Breast Cancer: The Beginning of the End or the End of the Beginning? _The American Journal of

Pathology_ 183, 1038–1051, https://doi.org/10.1016/j.ajpath.2013.07.003 (2013). Article  CAS  PubMed  Google Scholar  * Yoo, S.-K. _et al_. Whole-genome reference panel of 1,781 Northeast

Asians improves imputation accuracy of rare and low-frequency variants. _bioRxiv_, 600353, https://doi.org/10.1101/600353 (2019). * Karczewski, K. J. _et al_. Variation across 141,456 human

exomes and genomes reveals the spectrum of loss-of-function intolerance across human protein-coding genes. _bioRxiv_, 531210, https://doi.org/10.1101/531210 (2019). * Adzhubei, I. A. _et

al_. A method and server for predicting damaging missense mutations. _Nat Methods_ 7, 248–249, https://doi.org/10.1038/nmeth0410-248 (2010). Article  CAS  PubMed  PubMed Central  Google

Scholar  * Ramensky, V., Bork, P. & Sunyaev, S. Human non-synonymous SNPs: server and survey. _Nucleic Acids Res_ 30, 3894–3900, https://doi.org/10.1093/nar/gkf493 (2002). Article  CAS 

PubMed  PubMed Central  Google Scholar  * Chen, E. Y. _et al_. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. _BMC Bioinformatics_ 14, 128–128,

https://doi.org/10.1186/1471-2105-14-128 (2013). Article  PubMed  PubMed Central  Google Scholar  * Kuleshov, M. V. _et al_. Enrichr: a comprehensive gene set enrichment analysis web server

2016 update. _Nucleic Acids Res_ 44, W90–W97, https://doi.org/10.1093/nar/gkw377 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Moore, C. B., Wallace, J. R., Frase, A. T.,

Pendergrass, S. A. & Ritchie, M. D. BioBin: a bioinformatics tool for automating the binning of rare variants using publicly available biological knowledge. _BMC Med Genomics_ 6, S6,

https://doi.org/10.1186/1755-8794-6-S2-S6 (2013). Article  PubMed  PubMed Central  Google Scholar  * Hanahan, D. & Weinberg, R. A. Hallmarks of Cancer: The Next Generation. _Cell_ 144,

646–674, https://doi.org/10.1016/j.cell.2011.02.013 (2011). Article  CAS  PubMed  Google Scholar  * Leonis, M. A., Thobe, M. N. & Waltz, S. E. Ron-receptor tyrosine kinase in

tumorigenesis and metastasis. _Future Oncol_ 3, 441–448, https://doi.org/10.2217/14796694.3.4.441 (2007). Article  CAS  PubMed  Google Scholar  * Thobe, M. N. _et al_. The RON receptor

promotes prostate tumor growth in tramp mouse model. _Oncogene_ 30, 4990–4998, https://doi.org/10.1038/onc.2011.205 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar  * Chan, Q.

K. Y. _et al_. Activation of GPR30 inhibits the growth of prostate cancer cells through sustained activation of Erk1/2, c-jun/c-fos-dependent upregulation of p21, and induction of G(2)

cell-cycle arrest. _Cell Death Differ_ 17, 1511–1523, https://doi.org/10.1038/cdd.2010.20 (2010). Article  CAS  PubMed  Google Scholar  * Pupo, M. _et al_. A genetic polymorphism repurposes

the G-protein coupled and membrane-associated estrogen receptor GPER to a transcription factor-like molecule promoting paracrine signaling between stroma and breast carcinoma cells.

_Oncotarget_ 8, 46728–46744, https://doi.org/10.18632/oncotarget.18156 (2017). Article  PubMed  PubMed Central  Google Scholar  * UniProt: the universal protein knowledgebase. _Nucleic Acids

Res_ 45, D158–D169, https://doi.org/10.1093/nar/gkw1099 (2017). Article  CAS  Google Scholar  * Kunnev, D., Ivanov, I. & Ionov, Y. Par-3 partitioning defective 3 homolog (C. elegans)

and androgen-induced prostate proliferative shutoff associated protein genes are mutationally inactivated in prostate cancer cells. _BMC Cancer_ 9, 318,

https://doi.org/10.1186/1471-2407-9-318 (2009). Article  CAS  PubMed  PubMed Central  Google Scholar  * Li, T., Liu, X., Jiang, Q., Lei, X. & Liu, D. High expression of partitioning

defective 3–like protein is associated with malignancy in colorectal cancer. _Tumor Biology_ 39, 1010428317698393, https://doi.org/10.1177/1010428317698393 (2017). Article  CAS  PubMed 

Google Scholar  * Li, T., Liu, D., Lei, X. & Jiang, Q. Par3L enhances colorectal cancer cell survival by inhibiting Lkb1/AMPK signaling pathway. _Biochemical and Biophysical Research

Communications_ 482, 1037–1041, https://doi.org/10.1016/j.bbrc.2016.11.154 (2017). Article  CAS  PubMed  Google Scholar  * Thye, T. _et al_. Genome-wide association analyses identifies a

susceptibility locus for tuberculosis on chromosome 18q11.2. _Nat Genet_ 42, 739–741, https://doi.org/10.1038/ng.639 (2010). Article  CAS  PubMed  PubMed Central  Google Scholar  *

Ashburner, M. _et al_. Gene Ontology: tool for the unification of biology. _Nat Genet_ 25, 25–29 (2000). Article  CAS  PubMed  PubMed Central  Google Scholar  * Michailidou, K. _et al_.

Large-scale genotyping identifies 41 new loci associated with breast cancer risk. _Nat Genet_ 45, 353–361e3612, https://doi.org/10.1038/ng.2563 (2013). Article  CAS  PubMed  PubMed Central 

Google Scholar  * Landrum, M. J. _et al_. ClinVar: public archive of relationships among sequence variation and human phenotype. _Nucleic Acids Res_ 42, D980–D985,

https://doi.org/10.1093/nar/gkt1113 (2014). Article  CAS  PubMed  Google Scholar  * Lindberg, J. _et al_. The Mitochondrial and Autosomal Mutation Landscapes of Prostate Cancer. _European

Urology_ 63, 702–708, https://doi.org/10.1016/j.eururo.2012.11.053 (2013). Article  CAS  PubMed  Google Scholar  * Bonkhoff, H. Estrogen receptor signaling in prostate cancer: Implications

for carcinogenesis and tumor progression. _The Prostate_ 78, 2–10, https://doi.org/10.1002/pros.23446 (2018). Article  CAS  PubMed  Google Scholar  * Misawa, A. & Inoue, S.

Estrogen-Related Receptors in Breast Cancer and Prostate Cancer. _Front Endocrinol_ (_Lausanne_) 6, https://doi.org/10.3389/fendo.2015.00083 (2015). * Cinar, B. _et al_. MST1 is a

multifunctional caspase-independent inhibitor of androgenic signaling. _Cancer Res_ 71, 4303–4313, https://doi.org/10.1158/0008-5472.CAN-10-4532 (2011). Article  CAS  PubMed  PubMed Central

  Google Scholar  * Ready, D. _et al_. Mapping the STK4/Hippo signaling network in prostate cancer cell. _Plos One_ 12, e0184590–e0184590, https://doi.org/10.1371/journal.pone.0184590

(2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Rohena-Rivera, K. _et al_. IL-15 regulates migration, invasion, angiogenesis and genes associated with lipid metabolism and

inflammation in prostate cancer. _Plos One_ 12, e0172786–e0172786, https://doi.org/10.1371/journal.pone.0172786 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Morris, J. C.

_et al_. Vaccination with tumor cells expressing IL-15 and IL-15Rα inhibits murine breast and prostate cancer. _Gene Ther_ 21, 393–401, https://doi.org/10.1038/gt.2014.10 (2014). Article 

CAS  PubMed  PubMed Central  Google Scholar  * Yao, H.-P., Zhou, Y.-Q., Zhang, R. & Wang, M.-H. MSP–RON signalling in cancer: pathogenesis and therapeutic potential. _Nat Rev Cancer_ 13,

466, https://doi.org/10.1038/nrc3545 (2013). Article  CAS  PubMed  Google Scholar  * Tugues, S. _et al_. New insights into IL-12-mediated tumor suppression. _Cell Death Differ_ 22, 237–246,

https://doi.org/10.1038/cdd.2014.134 (2015). Article  CAS  PubMed  Google Scholar  * Huang, Q., Shen, H. M. & Ong, C. N. Emodin inhibits tumor cell migration through suppression of the

phosphatidylinositol 3-kinase-Cdc42/Rac1 pathway. _CMLS, Cell. Mol. Life Sci._ 62, 1167–1175, https://doi.org/10.1007/s00018-005-5050-2 (2005). Article  CAS  PubMed  Google Scholar  * Zhang,

Z. _et al_. Phosphatidylinositol 3-kinase β and δ isoforms play key roles in metastasis of prostate cancer DU145 cells. _The FASEB Journal_ 32, 5967–5975,

https://doi.org/10.1096/fj.201800183R (2018). Article  CAS  PubMed  Google Scholar  * Stefl, S., Nishi, H., Petukh, M., Panchenko, A. R. & Alexov, E. Molecular mechanisms of

disease-causing missense mutations. _J Mol Biol_ 425, 3919–3936, https://doi.org/10.1016/j.jmb.2013.07.014 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Park, T.-J. _et

al_. Practical Calling Approach for Exome Array-Based Genome-Wide Association Studies in Korean Population. _Int J Genomics_ 2015, 421715–421715, https://doi.org/10.1155/2015/421715 (2015).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Gudmundsson, J. _et al_. A study based on whole-genome sequencing yields a rare variant at 8q24 associated with prostate cancer. _Nat

Genet_ 44, 1326–1329, https://doi.org/10.1038/ng.2437 (2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Grove, M. L. _et al_. Best Practices and Joint Calling of the HumanExome

BeadChip: The CHARGE Consortium. _Plos One_ 8, https://doi.org/10.1371/journal.pone.0068095 (2013). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Turner, S. _et al_. Quality

Control Procedures for Genome Wide Association Studies. _Curr Protoc Hum Genet_ CHAPTER, Unit1.19, https://doi.org/10.1002/0471142905.hg0119s68 (2011). Article  Google Scholar  * Patterson,

N., Price, A. L. & Reich, D. Population Structure and Eigenanalysis. _PLOS Genetics_ 2, e190, https://doi.org/10.1371/journal.pgen.0020190 (2006). Article  CAS  PubMed  PubMed Central 

Google Scholar  * Lee, S. _et al_. Optimal unified approach for rare-variant association testing with application to small-sample case-control whole-exome sequencing studies. _Am J Hum

Genet_ 91, 224–237, https://doi.org/10.1016/j.ajhg.2012.06.007 (2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Madsen, B. E. & Browning, S. R. A Groupwise Association

Test for Rare Mutations Using a Weighted Sum Statistic. _PLOS Genetics_ 5, https://doi.org/10.1371/journal.pgen.1000384 (2009). Article  PubMed  PubMed Central  Google Scholar  Download

references ACKNOWLEDGEMENTS This work was supported by NLM R01 NL012535. This project is also funded, in part, under a grant with the Pennsylvania Department of Health (#SAP 4100070267). The

Department specifically disclaims responsibility for any analyses, interpretations or conclusions. This work was also supported by 1) Korean Urologic Oncology Society research fund

(KUOS17-1), 2) SNUBH Research fund (No. 13-2015-009 & 02-2017-009) and 3) Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of

Education (2017R1D1A1A09000743). AUTHOR INFORMATION AUTHORS AND AFFILIATIONS * Department of Urology, Seoul National University College of Medicine, Seoul National University Bundang

Hospital, Seongnam, Korea Jong Jin Oh, Hakmin Lee, Sung Kyu Hong, Sang Eun Lee & Seok-Soo Byun * Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine,

University of Pennsylvania, Philadelphia, PA, USA Manu Shivakumar & Dokyoon Kim * Department of Genetics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Jason Miller & Shefali Verma * Department of Biomedical Informatics, University of Utah, University of Utah School of Medicine, Salt Lake City, UT, USA Younghee Lee * Complex Diseases

and Genome Epidemiology Laboratory, Department of Public Health, Graduate School of Public Health, Seoul National University, Seoul, Korea Soo Ji Lee & Joohon Sung * Institute for

Biomedical Informatics, University of Pennsylvania, Philadelphia, PA, USA Dokyoon Kim Authors * Jong Jin Oh View author publications You can also search for this author inPubMed Google

Scholar * Manu Shivakumar View author publications You can also search for this author inPubMed Google Scholar * Jason Miller View author publications You can also search for this author

inPubMed Google Scholar * Shefali Verma View author publications You can also search for this author inPubMed Google Scholar * Hakmin Lee View author publications You can also search for

this author inPubMed Google Scholar * Sung Kyu Hong View author publications You can also search for this author inPubMed Google Scholar * Sang Eun Lee View author publications You can also

search for this author inPubMed Google Scholar * Younghee Lee View author publications You can also search for this author inPubMed Google Scholar * Soo Ji Lee View author publications You

can also search for this author inPubMed Google Scholar * Joohon Sung View author publications You can also search for this author inPubMed Google Scholar * Dokyoon Kim View author

publications You can also search for this author inPubMed Google Scholar * Seok-Soo Byun View author publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS

J.J.O., M.S., D.K. and S.S.B. conceived the project. J.J.O., H.L., S.K.H., S.E.L. and S.S.B. curated data. J.J.O., M.S., J.M., S.V., Y.L., S.J.L., J.S. performed the formal analysis along

with visualization. D.K. and S.S.B. provided resources. J.J.O., M.S., J.M., D.K. and S.S.B. supervised all aspects of this work. The original draft was written by J.J.O. and M.S. All authors

reviewed and edited the manuscript. Jong Jin Oh and Manu Shivakumar contributed equally. CORRESPONDING AUTHORS Correspondence to Dokyoon Kim or Seok-Soo Byun. ETHICS DECLARATIONS COMPETING

INTERESTS The authors declare no competing interests. ADDITIONAL INFORMATION PUBLISHER’S NOTE Springer Nature remains neutral with regard to jurisdictional claims in published maps and

institutional affiliations. SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFO RIGHTS AND PERMISSIONS OPEN ACCESS This article is licensed under a Creative Commons Attribution 4.0 International

License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source,

provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons

license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by

statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

http://creativecommons.org/licenses/by/4.0/. Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Oh, J.J., Shivakumar, M., Miller, J. _et al._ An exome-wide rare variant analysis

of Korean men identifies three novel genes predisposing to prostate cancer. _Sci Rep_ 9, 17173 (2019). https://doi.org/10.1038/s41598-019-53445-2 Download citation * Received: 15 November

2018 * Accepted: 25 October 2019 * Published: 20 November 2019 * DOI: https://doi.org/10.1038/s41598-019-53445-2 SHARE THIS ARTICLE Anyone you share the following link with will be able to

read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing

initiative