Date Published: April 7, 2009
Publisher: Public Library of Science
Author(s): H. C. Harsha, Kumaran Kandasamy, Prathibha Ranganathan, Sandhya Rani, Subhashri Ramabadran, Sashikanth Gollapudi, Lavanya Balakrishnan, Sutopa B. Dwivedi, Deepthi Telikicherla, Lakshmi Dhevi N. Selvan, Renu Goel, Suresh Mathivanan, Arivusudar Marimuthu, Manoj Kashyap, Robert F. Vizza, Robert J. Mayer, James A. DeCaprio, Sudhir Srivastava, Samir M. Hanash, Ralph H. Hruban, Akhilesh Pandey
Abstract: Akhilesh Pandey and colleagues describe a compendium of potential biomarkers that can be systematically validated by the pancreatic cancer community.
Partial Text: With a large number of studies in the published literature (a keyword search for “pancreatic cancer” fetches ∼50,000 published articles), our approach (Figure 1) was to first identify the relevant publications and datasets (e.g., microarray data submitted to repositories such as GEO, ArrayExpress, and Oncomine –) that might contain information on overexpression of mRNAs or proteins in pancreatic cancer. The type of pancreatic cancer and the cell type where overexpression was observed were also annotated. After generating this list of molecules, specific searches were carried out to identify the presence of these molecules in body fluids or on the plasma membrane. Finally, queries were carried out to determine the status of these molecules in chronic pancreatitis, which is an important consideration in the differential diagnosis of pancreatic cancer.
Our efforts at creating this compendium represent the first step in tackling biomarkers for pancreatic cancer in a global and systematic fashion. In fact, it is already being used by a consortium of investigators who are developing antibodies against the 60 most promising targets in PDACs as part of a new initiative funded by the Lustgarten Foundation for Pancreatic Cancer Research. Our compendium also included data on other, less common subtypes of pancreatic cancer (Table G in Table S1 provides a partial list of molecules). The entire list of molecules overexpressed in pancreatic cancers that are included in the compendium is provided in Table S2. It must be pointed out that 74% of the molecules in this compendium are based solely on mRNA evidence. As is inherent to mRNA-based methods, especially DNA microarrays, the data often require subsequent validation by other methods. Further, several high-throughput studies carried out to identify genes that are differentially expressed in pancreatic cancer have used tissues that are not microdissected to separate cancer from stroma. Thus, it is unclear in many instances if the observed difference in the expression of a particular gene originates in the stroma or in cancer cells. This further underscores the importance of validating these observations using alternative methods by targeted studies.
Given the explosion of data from multiple platforms, the information must be integrated before a systems view of cancers can emerge. In this regard, we have used pancreatic cancer as an example to create a resource that should serve as a model for other cancers. Discovering a single biomarker that would be both sensitive and specific for cancer of a given organ might be more difficult than discovering a panel of biomarkers. Identifying components to be used in such a panel would require systematic cataloging and testing of the most promising candidates that are available. To this end, we have carried out a systematic curation of the literature that took approximately over 7,000 person hours. This was an international effort and was possible because of concerted efforts of trained scientists at the Institute of Bioinformatics, where the majority of the curation work was carried out, working closely with several scientists in the United States. Such database efforts are crucial for systems biology approaches to human diseases because the data are not available in a single location and often not accessible to those without any bioinformatics experience. We feel that a multipronged approach to cancer will require not only continued discovery efforts, but also resources that maximize what we already know about cancers. Our future goal is to develop a Web-based searchable database of all molecular alterations in pancreatic cancer—from the genome to the proteome—that will help initiate a systems medicine approach to cancer.