Early detection represents one of the most promising approaches to improve lung cancer survival. To date, no reliable circulating biomarkers of lung cancer have been identified that allow early diagnosis. With the advent of gene targeting technologies, new genetically engineered mouse models of lung cancer have been created that closely recapitulate the pathobiology of the human disease. Given the variability, complexity, and dynamic range of human plasma, we hypothesized that the use of appropriate animal models might provide a unique entry point for the discovery of early lung cancer biomarkers. To test this hypothesis, we obtained plasma from three groups of mice, a control group with no lung tumors, a second group with lung tumors induced by expression of oncogenic K-ras (K-rasG12D), and a third group with lung tumors driven by both oncogenic K-ras and mutant p53. Plasma samples were first depleted of highly abundant proteins using an Agilent MARS M3 column. The depleted samples were then reduced, alkylated, and fractionated at the protein level by size exclusion chromatography (SEC). SEC fractions were digested with trypsin and sub-fractionated by gradient strong cation exchange (SCX). Collected fractions were analyzed by LC-MS/MS using a linear ion trap-Fourier transform ion cyclotron resonance tandem mass spectrometer. The resulting LC-MS/MS data were analyzed using SpectrumMill. All analyses were performed as complete replicates. We identified close to 500 proteins with a high degree of confidence. Of these, approximately one-quarter were increased at least two fold in one or both of the lung cancer-bearing groups compared to cancer-free controls. These candidate biomarkers include the expected acute phase reactants as well as a variety of coagulation cascade components. Potential biomarkers of interest include proteins involved in extracellular maintenance and degradation, lung-specific proteins, and tumor endothelial markers. To validate candidate biomarkers of lung cancer, we have performed immunoblotting of individual mouse plasma samples using commercially available antibodies directed against potential biomarkers. We are also performing quantitative analysis of candidate biomarkers using labeled peptides and multiple reaction monitoring. These preliminary studies highlight the potential utility of combining mouse models with innovative technologies to facilitate the discovery of clinically relevant biomarkers of lung cancer. These studies may ultimately lead to the identification of one or a panel of markers for formal evaluation in clinical trials.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]