BA16 - Identifying biomarkers for pancreatic cancer

Investigator Names and Contact Information

Sunil Hingorani (srh@fhcrc.org)

Introduction/Intent

It is axiomatic in clinical oncology that early detection of cancers increases the likelihood of cure. This is especially true of pancreatic ductal adenocarcinoma (PDA), or pancreas cancer, a disease whose onset and progression are effectively obscured by anatomic location, vague and non-localizing symptoms – when they occur at all – and a complete dearth of noninvasive detection and diagnostic modalities. With an annual incidence and mortality of almost 40,000 cases, PDA is now the fourth leading cause of cancer-related death in the United States (1). The 5-year overall survival for patients with PDA is less than 3% (reviewed in (2)). The relatively advanced stage of disease in the vast majority of patients at presentation together with an inherent resistance to all current forms of chemical and radiotherapies conspire to make PDA an essentially uniformly fatal disease.

The successful development of early detection biomarkers requires four essential elements: 1) the identification of appropriate patient populations for the investigation and validation of testing strategies; 2) biopsy or resection of the organ of interest to definitively establish a histopathologic diagnosis and stage of disease; 3) biospecimens (e.g. plasma, serum) obtained prior to treatment, and in the ideal circumstance, prior to clinical presentation; and 4) a discovery platform with sufficient sensitivity to probe the femtomolar concentration range of the circulating proteome where most useful early detection markers are hypothesized to reside. To date, cancer of the pancreas has defied all of these requirements. In addition, biopsy of the pancreas carries a very real risk of pancreatitis with the potential for prolonged hospitalization and even death. Thus, the identification of a circulating biomarker with sufficient sensitivity and specificity to detect PDA early enough to resect would immediately and dramatically alter the management of these patients. The longitudinal design, cohort size and number of PDA cases of the WHI study make it a unique and an ideal resource for addressing the extraordinary challenge of developing methods to detect pancreas cancer early enough to intervene meaningfully.

We propose a nested case-control study consisting of 215 pancreas cancer cases diagnosed among women enrolled in WHI Observational Study (OS) through August 15, 2007, and 215 individually matched cancer-free controls, designed to identify biomarkers for the early detection of infiltrating pancreatic ductal adenocarcinoma (PDA). Plasma samples from cases and controls will be interrogated using a novel, high-density microarray platform developed by our laboratories. The platform encompasses over 18,000 antibodies as binding agents that have been specifically selected for utility in cancer early detection (discussed below). The current array contains approximately 1300 commercially-available, full-length antibodies together with a recombinant antibody sub-library composed of single-chain variable domains (scFv). Sub-libraries of scFv are selected from 2 master libraries that contain between 108-1010 unique sequences to bind epitomes present in pooled plasma from cancer patients that are not present in cancer-free control plasma. The nanomolar to femtomolar affinity constants of these binding agents afford several advantages over other proteomic methods: 1) ability to interrogate the plasma proteome with a dynamic range far greater than possible with mass spectrometry methods and using comparatively minute sample volumes; 2) ability to discover biomarkers resulting from differential post-translational modifications of otherwise identical peptide backbones; and 3) ability to discover biomarkers that may not be readily immunogenic in animals. We have already demonstrated the high specificity and sensitivity of these arrays to detect known and potential novel biomarkers for breast and ovarian cancer (3) and Preliminary Studies).

These studies will be further informed by findings from the highly faithful animal models of PDA developed in our laboratories that recapitulate the full spectrum of preinvasive, invasive and metastatic stages of human PDA. Scrutinizing plasma from these sophisticated mouse models of pancreatic cancer with our microarray platform has identified a circumscribed list of strong candidate markers which was then used to augment newly fabricated antibody arrays to probe patient samples. We have additionally performed gene expression and proteomics analyses on fractionated cells isolated at discrete stages of tumor progression, and these data further informed the enrichment of the primary antibody arrays. Thus, our methodologies for directed microarray analyses permit a highly cost effective approach to interrogate the plasma proteome utilizing very small sample volumes (100-200 μl) to obtain reliable comparative data on thousands of proteins simultaneously. This combination of technologies, expertise and richly annotated biospecimen repository provides an unprecedented opportunity to penetrate a previously intractable clinical problem with profound implications for decreasing morbidity and mortality from this uniformly devastating disease. Our proposal encompasses the following focused aims:

AIM 1: To identify biomarkers for the early detection of invasive pancreas cancer.

AIM 2: To determine whether integrating analyses of known and suspected risk factors for pancreas cancer will enhance the performance and potential clinical utility of the early detection biomarkers identified in Aim 1.

References

1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J Clin 2007;57(1):43-66.
2. Hruban RH. Tumors of the pancreas. In: Hruban RH, Pitman, M.B. and Klimstra, D.S., editor. Atlas of tumor pathology. Washington, DC: Armed Forces Institute of Pathology; 2007.
3. Loch CM, Ramirez AB, Liu Y, et al. Use of High Density Antibody Arrays to Validate and Discover Cancer Serum Biomarkers. Molecular Oncology 2007;1(3):313-20.