Epidemiological Study Contributing to Understanding of Cancer Etiology and Development of Preventive Measures

▼ Environmental Epidemiology

In addition to examining factors related to lifestyles such as smoking, alcohol consumption, and diet, we vigorously investigate an association between various environmental factors exposed in daily life and the risk of cancer incidence. While some chemicals pose carcinogenic risks due to occupational exposure, the impact of chemicals potentially exposed in everyday life on cancer incidence among the Japanese population requires evaluation. From this perspective, we have advanced investigations in the following areas 1) to 3). Additionally, our research and findings to date, including specific findings detailed in 1) to 3), are also described in the review article provided below.

Exposure to environmental chemicals and cancer risk: epidemiological evidence from Japanese studies

Moreover, recent advancements in analytical techniques utilizing next-generation sequencers have enabled the assessment of microbiota, with a growing focus on its impact on diseases. As shown in the epidemiological study in 4), we have initiated the collection of fecal specimens and analyses targeting gut microbiota.

1) Organochlorines
Organochlorines are suspected to possess endocrine disruption. We investigate the influence of these substances in daily living environments on the risk of hormone-related cancers. We have reported the following findings:

1 Plasma organochlorine levels and subsequent risk of breast cancer among Japanese women: a nested case-control study
2 Serum organochlorines and breast cancer risk in Japanese women: a case-control study 
3 Plasma organochlorines and subsequent risk of prostate cancer in Japanese men: a nested case-control study 

2) Perfluoroalkyl substances
Perfluoroalkyl substances are utilized in various applications, including fire extinguishing agents, water repellents, stain repellents, and surface treatment agents for materials such as paper, textiles, and cookware. Due to their stable structures, these compounds resist decomposition in the environment and exhibit high bioaccumulative potential, raising concerns about possible health impacts on humans. The International Agency for Research on Cancer (IARC) currently classifies perfluorooctanoic acid (PFOA) as Group 1 (carcinogenic to humans) and perfluorooctanesulfonic acid (PFOS) as Group 2B (possibly carcinogenic to humans), emphasizing the need for evidence from epidemiological studies. In our division, we have reported findings such as the following:

1 Serum perfluoroalkyl substances and breast cancer risk in Japanese women: A case-control study 

3) Dietary carcinogens
Among chemical substances generated during food processing and cooking, some are suspected carcinogens. One example is heterocyclic amines produced when cooking meat or fish at high temperatures. IARC categorizes them as Group 2B (possibly carcinogenic to humans), supported by sufficient evidence from animal experiments. However, evidence from epidemiological studies for humans is currently insufficient. Therefore, we developed a method to assess the intake of heterocyclic amines from food frequency questionnaire (FFQ) and evaluated the validity of the method. Further, we applied it to epidemiological research and investigated an association between heterocyclic amines and cancer risk, as summarized in the following review article. 

1 Dietary heterocyclic aromatic amine intake and cancer risk: epidemiological evidence from Japanese studies 

Next, we address acrylamide, a chemical used as a raw material in substances such as paper strength enhancers and adhesives. It is classified as Group 2A (probably carcinogenic to humans) by IARC. Recently, it has been pointed out that, since acrylamide is formed by chemical reactions during the processing and cooking of foods containing the nutrients asparagine and reducing sugars at temperatures exceeding 120℃, it is present in food. Hence, there has been a growing focus on a potential association between the intake of acrylamide from diets and cancer risk. In our division, we have reported findings, as follows: 

2 Validity of a Self-administered Food Frequency Questionnaire for the Estimation of Acrylamide Intake in the Japanese Population: The JPHC FFQ Validation Study 
3 Dietary acrylamide intake and risk of breast cancer: The Japan Public Health Center-based Prospective Study 
4 Dietary acrylamide intake and the risk of endometrial or ovarian cancers in Japanese women 

4) Gut Microbiota
Basic research has reported the presence of Escherichia coli producing a toxin called colibactin, which has genotoxic properties in the intestines, suggesting its potential involvement in colorectal carcinogenesis. Thus, we conducted epidemiological studies focusing on colibactin-producing E. coli and reported findings, as shown in the following: 

1^† Association of Escherichia coli containing polyketide synthase in the gut microbiota with colorectal neoplasia in Japan

^†The study was a collaboration with University of Shizuoka and National Institutes of Biomedical Innovation.

▼ Molecular Epidemiology

Sporadic cancer is a multifactorial disease wherein various environmental factors and genetic elements intricately interact, leading to carcinogenesis. Environmental factors have been shown to play a substantial role as a cause while the contribution of genetic factors is comparatively modest. To explore genetic factors that determine the risk of cancer incidence, genome-wide association studies (GWAS) have been vigorously conducted. Our involvement in international collaborative studies has contributed to the identification of novel genetic region associated with the risk of cancers such as breast and colorectal cancers. The representative studies are listed below:

- GWAS on colorectal cancer in collaboration with Professor Loic Le Marchand at the University of Hawaii
Trans-ethnic genome-wide association study of colorectal cancer identifies a new susceptibility locus in VTI1A

- Genome-wide association study on breast cancer conducted by the Breast Cancer Association Consortium (BCAC)
Association analysis identifies 65 new breast cancer risk loci 

With the accumulation of insights from GWAS, the knowledge is actively utilized in analytical studies. Particularly from the standpoint of epidemiological and preventive studies, key investigations include 1) investigation into gene-environment interactions, 2) Mendelian randomization study with genomic information, exemplified by genetic polymorphisms, as instrumental variables, and 3) application to predictive models for estimating absolute risk. We present our findings based on these investigations below. Additionally, the development of analytical techniques by next-generation sequencers also has enabled the evaluation of rare mutations and abnormalities of genome structure, advancing the elucidation of genetic factors significantly contributing to carcinogenesis. In our division, studies with genome analysis using next-generation sequencers were initiated, as outlined in 4).


1) Investigation into Gene-Environment Interactions
Cancer is considered to arise from multiple factors, necessitating the evaluation of simultaneous effects of various factors, namely interactions. Understanding gene-environment interactions, particularly an association between environmental factors and cancer risks observed exclusively in individuals with specific genetic factors, can provide evidence for prevention based on genetic factors (i.e., personalized prevention). We have reported the following findings contributing to this area of investigation. 

1^‡ Genetic polymorphisms of ADH1B, ADH1C and ALDH2, alcohol consumption, and the risk of gastric cancer: the Japan Public Health Center-based prospective study 
2^‡ CYP1A1, GSTM1 and GSTT1 genetic polymorphisms and gastric cancer risk among Japanese: A nested case-control study within a large-scale population-based prospective study 
3 Vitamin D Receptor Gene Polymorphism and the Risk of Colorectal Cancer: A Nested Case-Control Study 
4 Alcohol consumption, genetic variants in the alcohol- and folate metabolic pathways and colorectal cancer risk: the JPHC Study 
5 Fat mass and obesity-associated gene polymorphisms, pre-diagnostic plasma adipokine levels and the risk of colorectal cancer: The Japan Public Health Center-based Prospective Study
6 p53 Arg72Pro polymorphism, adiposity status, and cancer risk: Two case-cohorts within a Japanese prospective study  

^‡These studies were a collaboration with Aichi Cancer Center.


2) Mendelian randomization study
Mendelian randomization analysis utilizes an instrumental variable method with genomic information to analytically control the impact of unknown and unobserved confounders. Regarding genetic polymorphisms employed as instrumental variables, alleles are randomly allocated according to Mendel's law. Therefore, an equal distribution of background factors is assumed between groups with and without risk alleles. Moreover, causal effect can be inferred under the following assumptions: 1) genetic polymorphisms are associated with an exposure, 2) genetic polymorphisms exclusively influence an outcome via an exposure, and 3) there is no confounders between genetic polymorphisms and an outcome. In controlling confounders, a crucial issue in epidemiological studies, traditional observational studies such as cohort studies have primarily relied on known and observed factors. Namely, controlling unknown and unobserved confounders has been a challenge. In recent years, with the accumulation of insights from GWAS across various traits, Mendelian randomization analysis has been widely conducted. We have reported our findings in line with this methodology, as shown below:

1 Diabetes and cancer risk: A Mendelian randomization study 
2^‡‡ Body mass index and colorectal cancer risk: A Mendelian randomization study
3^‡‡ Blood Lipids and the Risk of Colorectal Cancer: Mendelian Randomization Analyses in the Japanese Consortium of Genetic Epidemiology Studies 
4^‡‡ Association of 25-hydroxyvitamin D with risk of overall and colorectal cancer among Japanese using a Mendelian randomization approach 
5^‡‡ Investigating the association between glycaemic traits and colorectal cancer in the Japanese population using Mendelian randomisation

^‡‡These studies were a collaboration in Japanese Consortium of Genetic Epidemiology studies.


3) Development of Risk Prediction Models
A risk prediction model is a statistical tool designed to estimate the absolute risk of disease based on an individual's presence of risk factors. Understanding one's absolute risk is anticipated to encourage behavioral changes such as lifestyle improvements. In the JPHC Study, we have developed risk prediction models based on information gathered from questionnaires and made them accessible on the web, enabling individuals to easily grasp their absolute risks (linked to an external Japanese website for NCC; https://epi.ncc.go.jp/riskcheck/). To enhance the predictive performance of these risk prediction models based on information derived from questionnaires, we have incorporated risk scores utilizing genetic polymorphisms revealed through GWAS. Additionally, we have also constructed models that consider gene-environment interactions. Our division has reported the following findings in this field:

1 Inclusion of a Genetic Risk Score into a Validated Risk Prediction Model for Colorectal Cancer in Japanese Men Improves Performance
2 Inclusion of a gene-environment interaction between alcohol consumption and the aldehyde dehydrogenase 2 genotype in a risk prediction model for upper aerodigestive tract cancer in Japanese men
3 Development and validation of genome-wide polygenic risk scores for predicting breast cancer incidence in Japanese females: a population-based case-cohort study 
 

4) Studies in Genome Analysis Utilizing Next-Generation Sequencers
Utilizing next-generation sequencers for analysis, we conduct studies targeting participants with informed consent in the JPHC-NEXT Study to explore impactful rare mutations and structural abnormalities in the genome. As an initial step, we initiated whole genome sequencing for cancer cases in their 40s, where the influence of their heredity is deemed relatively strong. Future plans involve focusing on specific cancer sites or adding age groups to further develop this project. Moreover, we examine the applicability of the detection of somatic cell mutations derived from tumors to diagnostic purposes by collecting cell-free DNA (cfDNA) in the bloodstream and analyzing it with next-generation sequencers. If the techniques established in this project prove applicable to liquid biopsy, it is anticipated that early detection of cancer with minimal invasiveness may become achievable.

▼ Metabolic Epidemiology

In traditional epidemiological studies, the assessment of lifestyles such as diet has primarily been conducted through questionnaire surveys. However, it is recognized that levels of certain dietary and nutritional factors estimated through questionnaires may not necessarily reflect exposure levels within the body. Additionally, metabolites within the body may also exhibit carcinogenic or preventive effects. Moreover, established risk factors such as obesity and diabetes are assumedly involved in carcinogenesis via adipocytokines, hormones, chronic inflammation, blood glucose, insulin, and other substances within the body. Investigation into elements reflected in blood concentrations as a result of exposure to such factors, exemplified by nutrients, cytokines, hormones, and metabolites, can lead to understanding the mechanisms of carcinogenesis and obtaining evidence for the evaluation of the causal relationship between risk factors and cancer.


1) Nested Case-Control Study on Pancreatic Cancer (Based on a Baseline Survey in the JPHC Study)
We conducted metabolomic analysis to identify novel risk factors for pancreatic cancer. In the targeted metabolomic analysis, we focused on branched-chain amino acids in relation to the risk of diabetes and metabolites reported as potential diagnostic markers for pancreatic cancer. This approach yielded the following findings:

1^§ Increased Levels of Branched-Chain Amino Acid Associated With Increased Risk of Pancreatic Cancer in a Prospective Case-Control Study of a Large Cohort
2^§ Metabolome analysis for pancreatic cancer risk in nested case-control study: Japan Public Health Center-based prospective Study 

^§These studies were a collaboration with Kobe University


2) Case-Cohort Study on Overall Cancer (Based on a Baseline Survey in the JPHC Study)
We conducted case-cohort studies on overall cancer, establishing a sub-cohort of approximately 4,500 individuals randomly selected from the approximately 34,000 participants in the baseline survey of the JPHC Study. These studies utilized the data of 3,750 cancer cases identified by follow-up surveys until the end of 2009. By analyzing blood biomarkers reflecting carcinogenic mechanisms relatively common across various sites in the case-cohort studies, we reported findings to date, as listed below:

1 Plasma 25-hydroxyvitamin D concentration and subsequent risk of total and site specific cancers in Japanese population: large case-cohort study within Japan Public Health Center-based Prospective Study cohort
2 Plasma C-peptide and glycated albumin and subsequent risk of cancer: From a large prospective case-cohort study in Japan
3 Association between C-reactive protein and risk of overall and 18 site-specific cancers in a Japanese case-cohort
4 Association of Plasma Iron Status with Subsequent Risk of Total and Site-Specific Cancer: A Large Case-Cohort Study within JPHC Study
5 Plasma Albumin, Bilirubin, and Uric Acid and the Subsequent Risk of Cancer: A Case-Cohort Study in the Japan Public Health Center-based Prospective Study


3) Elucidating the Mediating Effects of Blood Biomarkers
Nutrients, cytokines, hormones, and metabolites reflected in blood concentrations are assumed to either directly participate in carcinogenic mechanisms or reflect the consequences of exposure to certain risk factors. For example, obesity influences adipocytokines, insulin, chronic inflammation, and sex hormones, and in relation to cancer risk, it is possible to clarify the significance of blood biomarkers by decomposing the total effect into the direct effects of obesity and the indirect effects mediated by these blood biomarkers, respectively. We have conducted studies to elucidate the mediating effects of blood biomarkers below.

1 Importance of circulating leptin and adiponectin in the causal pathways between obesity and the development of colorectal cancer in Japanese men

▼ Molecular Pathological Epidemiology

In various types of cancers, molecular subtypes based on genetic mutations, mRNA levels, and protein expression levels in tumors are being proposed. Investigating risk factors of each molecular subtype of cancer by this novel classification method, we aim to examine how these factors influence carcinogenesis via various molecular abnormalities. Additionally, for factors that lack sufficient evidence of association with cancer, an association can be revealed between these factors and specific molecular subtypes of cancer. In the JPHC Study, tumor tissues from gastric and colorectal cancers identified through the follow-up surveys were collected from participants in selected areas. Currently, we measure genetic mutations and expression level of proteins in these tumors, initiating analyses focusing on molecular subtypes. This type of study represents the first attempt in Asia to utilize information and samples from a prospective cohort study.


1) Study Utilizing Tumor Tissues from Colorectal Cancer
Among participants in the JPHC Study, approximately 20,000 individuals residing in Yokote, Akita Prefecture and the central part of Okinawa Prefecture were followed up from 2000 to the end of 2014, and approximately 650 cases of colorectal cancer were identified. Tumor tissues were successfully collected from around 560 cases through collaboration with major hospitals in these areas. Utilizing the collected tumor tissues, genetic mutations and protein expression were examined by targeted sequencing analysis and immunohistochemistry, respectively. To date, this study has reported the following findings:

1 Smoking and risk of colorectal cancer according to KRAS and BRAF mutation status in a Japanese prospective Study
2 Dietary vitamin D intake and risk of colorectal cancer according to vitamin D receptor expression in tumors and their surrounding stroma
3 Height, body mass index, physical activity, and risk of colorectal cancer in relation to expression of insulin receptor: The Japan Public Health Center-based Prospective Study