Abstract 4807: Identification of pre-invasive pancreatic cancer serum biomarkers using genetically engineered mouse models and mass spectrometry techniques

• Published in: Cancer research (Chicago, Ill.) Ill.), 2012-04, Vol.72 (8_Supplement), p.4807-4807
• Main Authors: Ludwig, Michael R., Kojima, Kyoko, Grizzle, William, Mobley, James A., Klug, Christopher A.
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2012-4807

Background: Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer death in the United States. Surgery is the only curative treatment approach if the tumor is diagnosed early at an organ confined stage. Therefore, serum biomarkers specific for early stage pancreatic neoplasia would be of tremendous value in the treatment of this disease. Genetically engineered mouse models (GEM) of pancreatic neoplasia have been generated that closely recapitulate the molecular and biological properties of human PDAC. Here we take advantage of these GEMs and a high-throughput proteomics approaches to “dig deep” into the serum proteome of these mice to identify representative biomarkers of early stage pancreatic neoplasia. Methods: Serum was collected from 9 Pdx1-Cre (control) and 9 Pdx1-Cre;LSL-RasG12D (pre-invasive PDAC model) mice at 4 months of age. Serum was then pooled into 3 sets of 3 samples for each arm. The top 7 most abundant proteins were depleted from each pooled sample prior to labeling with a tandem mass tag (TMT) sixplex isobaric labeling system and long-column MudPIT analysis. Data were searched in SEQUEST, and quantified in ProteoIQ. Candidate proteins were accepted only with 97-99% confidence, and at least 2 unique TMT tagged peptides. Only biomarkers with >2 fold change and passing 2 non-parametric statistical tests were further evaluated. The same samples were also analyzed with a 1D MALDI-TOF MSMS approach to survey the low-mass serum proteome. In addition, a multi-dimensional MALDI-TOF approach was carried out on whole protein serum samples using spectral counting to quantify proteins instead of TMT labeling. Results/Discussion: Long-column MudPIT analysis resulted in identification of 3001 proteins with high confidence. Seventy of these proteins passed fold change and statistical filters, with 25 of these having known biological connections to pancreatic disease or function. A number of these proteins are currently being validated by immune assays. Multi-dimensional MALDI-TOF analysis resulted in identification of 8 differentially abundant proteins. Two of these proteins have been validated by Western blot. Taken together, these mass spectrometry techniques allow for a relatively in-depth survey of the serum proteome with accurate quantitation of potentially novel early stage pancreatic neoplasia biomarkers. Ultimately, the biomarkers identified in this study will be validated in human studies with the expectation that high ris