A Genome-Wide Association Study of Prostate Cancer Susceptibility Using Occupational and Environmental Factors as Confounding Factors

Background In addition to genetic predisposition, occupational and environmental factors are important for the risk of prostate cancer. We investigated the effect of single nucleotide polymorphisms (SNPs) on the development of prostate cancer in Japan, including occupational and industrial history as confounding factors in addition to age, smoking, and alcohol drinking. Methods We enrolled 210 prostate cancer patients and 504 male control patients. We conducted four genome-wide association study (GWAS) patterns for prostate cancer development. In the association test, logistic regression models incorporated age, smoking history, alcohol consumption history, and each pattern of industrial/occupational classification. Results No SNPs satisfying the genome-wide significance level of 5×10-8 were detected in GWAS. SNPs with a suggestive association level of 1×10-6 were found near the long intergenic non-protein coding RNA 1824 (LINC01824) and tripartite motif family like 2 (TRIML2) genes in the GWAS using occupational history as a confounder and near the ribosomal protein S2 pseudogene 25 (RPS2P25) gene in the GWAS using industrial history as a confounder. No SNPs that met the suggestive association level were observed in the GWAS that did not include occupational and industrial history. Conclusion By adding occupational and industrial history to the confounding factors, there were SNPs detected in the GWAS for prostate cancer development. The consideration of occupational and industrial history may increase the usefulness of GWAS.


Introduction
There were approximately 288,300 new cases of prostate cancer and approximately 34,700 deaths from prostate cancer in the United States in 2023.Since 2014, the incidence rate has increased by 3% per year overall and by about 5% per year for advanced-stage prostate cancer [1].In Japan, 92,021 males were diagnosed with prostate cancer in 2018, and there were 12,759 deaths from prostate cancer in 2020 [2].
In addition to genetic predisposition, occupational and environmental factors are important for the risk of prostate cancer.Smoking increases the risk of death from prostate cancer [3], but the overall association with prostate cancer incidence remains unclear [4].Alcohol drinking is a risk factor for the development of prostate cancer [5][6][7].Regarding occupation, being a white-collar worker [8] or a professional in a whitecollar industry [9] was reported to be risk factors for prostate cancer.In a meta-analysis, exposure to pesticides and chromium and shift work were risk factors for prostate cancer [10].
In this study, we investigated the effect of single nucleotide polymorphisms (SNPs) on the development of prostate cancer in Japan, including occupational and industrial history as confounding factors in addition to age, smoking, and alcohol drinking.As occupational history is a risk factor for prostate cancer, we added it to the confounding factors in order to detect the genetic contribution more clearly.
A genome-wide association study (GWAS) adopted in this study is a comprehensive genome-wide polymorphism search in order to detect SNPs that are susceptible to a disease or condition.Analyzing the genomes of prostate cancer patients, 1,534 associations were identified, leading to 74 articles as shown in the GWAS Catalog [11].

Materials And Methods
The genome purification was performed using ethylenediaminetetraacetic acid (EDTA)-containing blood (10 mL) collected from 210 prostate cancer patients and 504 male control (non-prostate cancer) patients at the Japan Organization of Occupational Health and Safety, Kanto Rosai Hospital.No special inclusion and exclusion criteria were established for male control patients.
We obtained occupational and environmental data from the Inpatient Clinico-Occupational Database of Rosai Hospital Group (ICOD-R) [12] including occupational history and information on smoking and alcohol drinking, with the provision by the Japan Organization of Occupational Health and Safety.Detailed coding of occupational and industrial history can be found elsewhere [13,14].We obtained other clinical data from electronic medical records.There were missing values because of the omission or lack of description by patients.

Zaitsu classification of industry/occupation
Zaitsu et al. [9] created a new taxonomy (tentatively named the Zaitsu classification) that combines the classifications of industrial and occupational history to create 12 different categories [15,16].

Clinical and environmental factors
Categorical variables were analyzed using Fisher's exact test between two and multiple groups, and age was analyzed using an unpaired t-test.

Genotyping
The genotyping of samples, quality control of samples, quality control of genotypes, and SNP imputation were previously described [16].

GWAS
We conducted four GWAS patterns for prostate cancer development by logistic linear models with PLINK 2.0, using SNP dosage obtained by SNP imputation with a minor allele frequency (MAF) of >0.01.In the association test, age, alcohol drinking history (yes/no), smoking history (the Brinkman index, ordered category with 0-3 levels), and each pattern of industrial/occupational classifications were added to logistic regression models.We tested four industrial and occupational classifications: (i) one variable with 20 levels for industrial classification divisions, (ii) one variable with 12 levels for occupational classification major groups, (iii) the Zaitsu classification, and (iv) GWAS without occupational and industrial history.The genome-wide significance level at p=5×10 -8 and suggestive association level at p=1×10 -6 were used in this study.

Clinical and environmental factors
The prostate cancer patients were older than the control patients (Table 1

TABLE 1: Clinical and environmental factors
Ages were analyzed using the unpaired t-test, whereas the Brinkman index and alcohol drinking were analyzed using Fisher's exact test between two or multiple groups In this study, a high Brinkman index was classified into four stages, and alcohol drinking history was not different between the prostate cancer cases and controls (Table 1).The distributions of the divisions of industrial classification (Table 2), occupational classification major groups (Table 3), and categories in the Zaitsu classification (Table 4) were not significantly different between prostate cancer cases and controls as a whole.Individually, prostate cancer was less frequent in industrial classification division l (scientific research, professional, and technical services) as shown in Table 2. Prostate cancer was significantly less common (Table 3) in the occupational classification major group f (security workers).

TABLE 4: Distribution of groups in the Zaitsu classification
The P-value analyzed among multiple groups using Fisher's exact test was 0.1564

Results of GWAS
No SNPs satisfying the genome-wide significance level of 5×10 -8 were detected in GWAS.SNPs with a suggestive association level of 1×10 -6 were found near the long intergenic non-protein coding RNA 1824 (LINC01824) and tripartite motif family like 2 (TRIML2) genes in the GWAS using occupational history as a confounder (Figures 1-3) and near the ribosomal protein S2 pseudogene 25 ( RPS2P25) gene in the GWAS using industrial history as a confounder (Figures 1, 4).No SNPs that met the suggestive association level were observed in the GWAS with the Zaitsu classification as a confounding factor or in the GWAS that did not include occupational or industrial history (Figure 1).SNPs that satisfied the suggestive association level of 1×10 -6 in GWAS are summarized in Table 5. 2024     The added industrial/occupational factor was one variable with 12 levels for occupational classification major groups TRIML2: tripartite motif family like 2

FIGURE 4: Regional plot of the RPS2P25 region (produced by
LocusZoom) The added industrial/occupational factor was one variable with 20 levels for industrial classification divisions RPS2P25: ribosomal protein S2 pseudogene 25

Discussion
Kanto Rosai Hospital is located in Kawasaki City, which is a heavy industrial area traditionally neighboring the Tokyo metropolitan area.Accordingly, there were a limited number of workers engaged in the primary sector of industry (farming, logging, fishing, and forestry) and the mining industry in this study.
This study primarily employed a relatively broad classification of industries/occupations.Thus, the aim of the present study was not to explore the connection between SNPs and particular environmentally exposed substances but rather to analyze the influence of wider industrial/occupational environmental factors, for example, stress stimulation and work environment, as confounding elements in prostate cancer development.Under these conditions, several SNPs were detected near genes (LINC01824, TRIML2, and RPS2P25) by GWAS that may be associated with the development of prostate cancer.
The association between these three genes and prostate cancer has not been previously reported, even in the GWAS Catalog.LINC01824 is an RNA gene located on 2p25.2 and is affiliated with the long non-coding RNA (lncRNA) class.LINC01824 is a lncRNA that is strongly correlated with transforming growth factor beta 1 (TGF-β1) expression in triple-negative breast cancer tissue [17].As TGF-β1 signaling is involved in the tumorigenesis of prostate cancer [18], LINC01824 can also be involved in prostate cancer development.
The TRIML2 gene, located on 4q35.2, encodes a member of the tripartite motif family of proteins.This protein may be regulated by tumor suppressor p53 and may regulate p53 through the enhancement of p53 SUMOylation [19].p53 is deeply involved in various aspects of prostate cancer [20].RPS2P25 is a pseudogene located on 5q14. 3 [21].For these genes, future clinical and biological studies will be necessary.
Utilizing ICOD-R occupational classification major groups, a relationship between occupations that harbor high physical activity and the reduction of cancer risk was demonstrated [12].By comparing the categories included in the manufacturing industry division of ICOD-R, it was shown that ureter cancer incidence in workers engaged in electronics is higher than that in workers in food manufacturing [22].Therefore, it is justified to add the industrial/occupational classification to the confounding factors of GWAS for the examination of the development of cancer.It may emphasize genetic contribution by reducing the contribution of environmental factors.In fact, GWAS performed without occupation and industrial history did not find any SNPs that met the suggestive association level; however, when they were added as confounders, several SNPs were detected in this study.
Several malignant tumor diseases other than prostate cancer were included in the control group.The rationale for the inclusion of many other cancers in the control group of GWAS for prostate cancer may be controversial.Pathways shared by various malignancies are less likely to be detected; however, it can be more effective for pathways specific to prostate cancer to appear.
The limitations of this study are that, compared to the GWAS conducted in recent years by a huge number of cases, the number of cases in this study is by far the smallest.Therefore, it is assumed that only limited results will be obtained from the GWAS itself.However, the genome-wide polymorphism information that accompanies detailed occupational and industrial history is scarce and has an advantage there.

Conclusions
By adding occupational and industrial history to the confounding factors, SNPs were detected in the GWAS for prostate cancer development.Our findings suggest that including occupational and industrial history increases the usefulness of GWAS.

FIGURE 1 :
FIGURE 1: Manhattan plots of the GWAS of prostate cancer cases Industrial/occupational factors added in GWAS: (A) GWAS without industrial/occupational factors, (B) one variable with 20 levels for industrial classification divisions, (C) one variable with 12 levels for occupational classification major groups, and (D) the Zaitsu classification GWAS: genome-wide association study

FIGURE 2 :
FIGURE 2: Regional plot of the LINC01824 region (produced by LocusZoom, University of Michigan, Ann Arbor, MI)The added industrial/occupational factor was one variable with 12 levels for occupational classification major groups LINC01824: long intergenic non-protein coding RNA 1824

FIGURE 3 :
FIGURE 3: Regional plot of the TRIML2 region (produced by LocusZoom) ). Malignant tumors other than prostate cancer, mainly urothelial cancer, were observed in 28.6% of the prostate cancer group and 70.8% of the control group.