Pancreatic cancer is one of the deadliest malignancies with less than 10% survival rate within 5-years of diagnosis. The disease etiology is believed to be multifactorial including both genetic and environmental factors, and the malignant tumor manifests unique molecular phenotypes resulting from several somatic mutations. Based on these understandings, we hypothesize that gene-environment interactions induce somatic mutations that lead to malignant transformation of cells in the exocrine pancreas. The transformation alters cell transcriptomes which change cell metabolisms. Based on the hypothesis, we speculate that distinct metabolomes, which are detectable in the circulation, are associated with specific transcriptomes which are connected to somatic mutations in the tumor. We further think that the metabolomic changes involve specific microbial metabolites implicated in the disease and that some of the metabolites are useful for early detection of pancreatic cancer. To test our hypothesis, we propose a nested case-control study with 3 specific aims.
Aim 1. To identify metabolomic signatures representing specific mutation phenotypes of pancreatic cancer for better understanding of the disease etiology. To achieve this aim, we will use high-throughput methods to analyze the metabolomes of 1,448 pancreatic cancer patients and 2,896 matched controls selected from 5 prospective cohorts, including the Multiethnic Cohort (MEC), the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO), the Shanghai Men and Women Health Cohorts (SMHS/SWHS) and the Southern Community Cohort Study (SCCS). The metabolomic profiles will be interrogated in pre-diagnostic plasma with bioinformatics tools and GWAS data to identify metabolomic signatures implicated in specific mutation phenotypes.
Aim 2. To examine the role of microbiome in pancreatic cancer by analyzing microbial metabolites in pre-diagnostic plasma and assessing their associations with mutation phenotypes and risk factors of pancreatic cancer. Using our unique reference library and protocols established for identifying microbial metabolites, we will interrogate the metabolomes of pancreatic cancer and compare the profiles with healthy controls to find microbial metabolites that are associated with pancreatic cancer and its risk factors. We will also examine the relationship between microbial metabolites and mutation phenotypes to assess the involvement of microbiota in pancreatic cancer etiology.
Aim 3. To test and validate specific metabolites that are useful for early detection of pancreatic cancer in different racial and ethnic groups. To achieve this aim, we will interrogate the pre-diagnostic metabolomes of pancreatic cancer to identify distinct metabolites for the disease. The differential metabolites will be cross-validated in each cohort and tested for sensitivity and specificity in racial/ethnic diverse populations while adjusting for covariates and confounders. We will also evaluate the time dependence of biomarkers prior to diagnosis and test their specificity in comparison to other types of cancer.
The proposed study will provide new insights into the relationship of mutation, transcription and metabolism in pancreatic cancer, which will help to understand the disease mechanisms and offer opportunities to develop novel methods for risk assessment and early detection, both of which are important for reducing the burden of pancreatic cancer.

Herbert Yu (University of Hawaii Cancer Center)
Wei Jia (University of Hawaii Cancer Center)
Loic Le Marchand (University of Hawaii Cancer Center)
Lynne Wilkens (University of Hawaii Cancer Center)
Lang Wu (University of Hawaii Cancer Center)
Xiao-ou Shu (Vanderbilt University School of Medicine)
Wendy Setiawa (University of Southern California School of Medicine)