Molecular diagnosis of human prostate cancer (CaP) by RRT-qPCR determination of gene expression signature

INTRODUCTION & OBJECTIVES: We previously found that human CaPprogression is accompanied by differential expression of a panel of 8 informativegenes by Northern blot analysis. Determination of expression of these genes in theTRAMP mice model of CaP progression provided reliable diagnostic (benign Vscancer) and prognostic prediction to green tea extract (GTCs) administration (GTCssensitiveVs GTCs-resistant tumor, Scaltriti et al., Carcinogenesis 2006). Here wechallenged the potential of the same 8-genes signature, detected by implementedRT-qPCR, in tumor detection/classification of human CaP.MATERIAL & METHODS: RT-qPCR was performed on 77 tissue specimensobtained from radical prostatectomy, immediately excised after surgery and quicklyfrozen. Histopatological examination classified 40 benign tissues and 37 CaPs. ForRNA extraction we used RNA-Fast reagent. RNA preps were quantified and resolvedon agarose gel to check quality and integrity. Retro-trascription (RT) was performedwith Superscript II reverse trascriptase. On each cDNA sample, the 8-genes signature(ODC, OAZ, SSAT, AdoMetDC, CLU, H3, Gas1, GAPDH) plus two housekeepergenes (PGK1 and HMBS) were determined in duplicate with optimised Sybr-Greenbased master mix. Data were analysed by Quadratic Discriminant Analysis (QDA).The CT-normalized database included 154 gene signature determinations. For CaPsamples we also collected clinical information, such as age and PSA value at diagnosis,and final Gleason score. QDA was used for discrimination of benign Vs CaP tissue,age65y Vs>65y, PSA10 Vs>10, Gleason score<7 Vs7.RESULTS: QDA analysis revealed that the concordance between our gene signatureclassification and final pathological evaluation (benign Vs CaP) was 83%. Alsofinal Gleason score correct classification was 83%. Gene signature classification wassuccessful in 91% of patients with regard of age, and in 74% with regard to PSA.CONCLUSIONS: RT-qPCR gene profiling, based on our 8-genes signature, appearto be an appropriate means for molecular diagnosis of presence/absence of CaP. Inaddition, it might became a new tool of potential prognostic value for discriminationof indolent Vs aggressive CaP because, in consideration that this approach was alreadysuccessful in predicting the biological behaviour of CaP in previous experimentsperformed in TRAMP mice, a possible explanation for misclassified cases could beindolent/aggressive behaviour. To this purpose, the follow-up of patients will continueto check this hypothesis.