A Proteomics-Based Approach To Identifying Mechanisms Of Cancer-Associated Thrombosis: Potential Role For Immunoglobulins

Venous thromboembolism (VTE) is highly prevalent in cancer and has serious consequences, but underlying mechanisms are incompletely understood. Proteomics is a powerful technique to measure protein expression patterns. We utilized a proteomics-based approach to identify novel potential mechanisms of cancer-associated VTE.

We analyzed the plasma of matched lung or pancreas cancer patients with and without VTE from the Roswell Park Cancer Institute Biorepository. Plasma samples were depleted of albumin and immunoglobulin by applying them simultaneously to Cibacron blue and protein G-Sepharose resins. The remaining plasma proteins were size-fractionated by SDS-PAGE then analyzed by liquid chromatography-mass spectrometry (LC-MS) for each mass range, following in-gel trypsin digestion. Mascot searches were conducted for protein matches with the NCBI database of the human proteome. Search results were uploaded into Scaffold. Data were filtered based on several criteria including two matching peptides, peptide threshold of 95%, protein threshold of 95% and identified in at least 3 samples. The nSC values were calculated by the Scaffold program and statistical significance was determined by a t-test.

The cohort comprised 60 patients with lung (N=30; 15 with VTE) or pancreas (N=30; 15 with VTE) cancer. In pancreas cancer patients, 288 proteins were identified of which 66% were present in both VTE and non-VTE samples. Five proteins were more abundant in the thrombotic samples and all were immunoglobulin-derived. These include IgM Fc (nSC = 4.2 and p= 0.02),immunoglobulin kappa chain variable region (MW 13 kDa) (nSC = 6.0 and p= 0.04), Ig kappa chainVKIII-JK3 (nSC = 9.0 and p= 0.03), immunoglobulin heavy chain variable region, (nSC = 13.3 and p= 0.03), and immunoglobulin kappa light chain variable region(only present in VTE samples, p= 0.002). Two proteins were more abundant in the non-VTE samples including an immunoglobulin kappa light chain variable region of different sequence (MW 8 kDa) (nSC=0, p=0.005) and phospholipase D (nSC=0.1, P=0.006). In lung cancer patients, 392 proteins were identified of which 91% were present in both groups. One protein, IgV kappa light chain, was more abundant in VTE samples (only present in VTE samples, p=0.03) and one protein, tetranectin, was more abundant in non-VTE samples (nSC=0.5, p=0.04). There was no overlap between differentially expressed proteins in pancreas and lung cancer specimens.

We were able to successfully analyze the plasma proteome of a cohort of matched cancer patients with and without VTE. This analysis suggests that immunoglobulin-derived proteins and tetranectin are differentially expressed in cancer patients with and without VTE. Additionally, mechanisms appear to differ based on primary site of cancer. Further studies to investigate the role of these proteins in the pathophysiology of cancer-associated thrombosis and the mechanisms underlying their differential expression are warranted.

No relevant conflicts of interest to declare.