The Genetics Of Gene Expression In Human Pancreatic Tissues
Mingfeng ZHANG, National Cancer Institute, United States
ZHU B. 1
, XIAO W. 2
, HOSKINS J. 1
, COLLINS I. 1
, JIA J. 1
, MATTERS G. 3
, POWELL J. 4
, KURTZ R. 5
, CHANOCK S. 1
, SMITH J. 6
, OLSON S. 5
, CHATTERJEE N. 7
, PETERSON G. 8
, SHI J. 1
, AMUNDADOTTIR L. 1
1 Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, Bethesda, Maryland 20892, USA.
2 Division of Bioinformatics and Biostatistics, National Center for Toxicological Research, FDA, Jefferson, Arkansas 72079, USA
3 Division of Computational Bioscience, Center for Information Technology, National Institutes of Health, Bioinformatics and Molecular Analysis Section, Bethesda, Maryland 20892, USA
4 Department of Biochemistry and Molecular Biology, The Pennsylvania State University College of Medicine, Hershey, PA 17033, USA.
5 Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York 10065, USA.
6 Division of Gastroenterology & Hepatology, Georgetown University Hospital, Washington, DC 20007, USA
7 Department of Biostatistics, Johns Hopkins Bloomburg School of Public Health, Baltimore, Maryland 21205, USA
8 Division of Epidemiology, Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota 55905, USA
Purpose: To understand the influence noncoding germline variation exerts on the regulation of gene expression in pancreatic tissues and to explain the underlying molecular mechanisms of pancreatic cancer risk loci.
Methods: We sequenced RNA from 95 histologically normal pancreatic tissue samples and tested associations between ~6million germline variants and the expression of ~17,000 genes. We included 115 tumor derived pancreatic tissue samples from the Cancer Genome Atlas (TCGA) pancreatic cancer dataset for comparison. Allele specific expression (ASE) was tested for heterogeneous loci at coding regions within each individual after taking into account the expected allele imbalance.
Results: We identified 484 cis-eQTL genes in histologically normal tissues and 237 cis-eQTL genes in tumor tissues (FDR<0.1). We observed significant enrichment of eQTLs within noncoding regulatory regions, which was more prominent in pancreatic cancer cell lines as compared to other tissues in the ENCODE data. The eQTLs were enriched in promoters for normal (3.8-fold) and tumor (4.6-fold) tissue eQTLs as compared to non-eQTL SNPs. Enrichment of eQTLs was seen in genic regions and for 5’UTRs, this was more prominent in the tumor tissue (21-fold) as compared to the normal tissues (10-fold). A common pancreatic cancer risk locus on 9q34.2 in the ABO gene demonstrated a regulatory effect on ABO expression in both normal (P=5.8x10-8) and tumor (P=8.3x10-5) tissues, with replication in the GTEx pancreatic samples (n=149, P=3.0x10-12). The ASE analyses further identified 1,286 genes with ASE in normal tissues after adjusting for multiple testing (P<1x105), including 36.8% of the cis-eQTL genes.
Conclusions: We have identified eQTLs representing potential functional regulatory variance in the pancreas and generated a dataset for further studies on regulation of gene expression in pancreatic tissues, which not only elucidates gene regulatory mechanisms in the pancreas, but also aids in interpreting pancreatic cancer GWAS findings.
Funding source: NCI/NIH intramural funding.