The model is based upon CaP progression due to the inactivation of the tumor-suppressor gene PTEN. PTEN is inactivated by loss or mutation in numerous human cancers, resulting in aberrant activation of the PI3-signaling pathway with uncontrolled cell growth, proliferation, and survival. The researchers focused on this pathway with the hypothesis that PTEN loss will produce alterations in surface and secreted proteomes in cells that may be recognized as a unique biomarker signature in the serum.
Using quantitative proteomic screening to detect and quantify N-linked glycoproteins that differ in prostate tissue and sera of prostatic Pten-deficient (Pten knockout, cKO) mice and littermate controls, they sought to focus on a sub-proteome that is enriched for validated serum biomarkers. Purified tissue and serum samples from control and Pten-cKO mice underwent solid-phase extraction of N-glycopeptides (SPEG). A total of 775 glycoproteins were identified and the majority was manually annotated to be secreted, reside on the plasma membrane, or be a part of the intracellular compartments. They found 152 of 658 (23.1%) of the proteins in the prostate tissue to be exclusive in the Pten-cKO mouse and 91 (13.8%) were only in the control mouse tissue. Next they detected Pten-dependent changes in the N-glycosite profiles of prostate tissue and serum by comparing liquid chromatography mass-spec. Quantification of 213 tissue and 105 serum proteins revealed that 68 proteins differed significantly in Pten-cKO mice compared with controls. Only 12 sera proteins were significantly different. Spectral counting confirmed these data and identified another 43 proteins with significant difference between Pten-cKO mice and control tissues. These experiments found differential expression of proteins associated with cell differentiation and the stem cell phenotype.
This identified biomarker list underwent further analysis based upon 3 criteria; Pten-dependency, prostate specificity, and detectability in serum. There were 126 proteins identified that were expected to possibly contain one or more specific candidate biomarker signatures that correspond to PTEN loss in human CaP. To study this further, they used human malignant and benign prostate tissue and serum. PTEN status was determined by calculating the percentage of epithelial cells having lost at least one PTEN gene copy number. They found that 72% of CaP samples had focal loss of PTEN gene copy numbers compared with controls, thus suggesting deletion of one or both alleles of PTEN in at least 20% of cells analyzed. Many of these CaP specimens had PI3K-pathway activation. In an analysis of 57 candidate serum proteins by targeted mass-spec, 36 peptides representing 33 proteins were consistently quantified in up to 105 patients. They used all this data to build predictive models for the discrimination between normal and aberrant PTEN status. Statistical methods selected 20 top-ranked variables for predicting focal loss of PTEN by combining 1 to 5 serum proteins. This resulted in 21,699 different models for analysis. A signature comprised of THBS-1, TIMP-1, CFH, and LRP-1 could correctly predict 78% of cases having aberrant or normal PTEN status with a sensitivity of 79.2% and specificity of 76.7%. This suggests that the reduction in PTEN gene copy number in CaP results in measurable changes of the serum proteome. The signature was independently validated to predict for the PTEN signaling network and status. It also stratified patients based on their Gleason grading.
Cima I, Schiess R, Wild P, Kaelin M, Schüffler P, Lange V, Picotti P, Ossola R, Templeton A, Schubert O, Fuchs T, Leippold T, Wyler S, Zehetner J, Jochum W, Buhmann J, Cerny T, Moch H, Gillessen S, Aebersold R, Krek W
Proc Natl Acad Sci U S A. 2011 Feb 22;108(8):3342-7
10.1073/pnas.1013699108
PubMed Abstract
PMID: 21300890
UroToday.com Prostate Cancer Section
