Validation of a whole-blood RNA prognostic signature in metastatic castration-resistant prostate cancer (mCRPC) patients.

Abstract

36 Background: We developed a whole-blood RNA transcript-based six gene signature (ABL2, SEMA4D, ITGAL, C1QA, TIMP1, CDKN1A), which separates patients with metastatic castration-resistant prostate cancer (mCRPC) into high- and low-risk survival groups (Ross et al. Lancet Oncology, 2012). In this study, the prognostic utility of the signature is validated on two independent platforms: an updated qPCR platform and an RNA-sequence platform in a new cohort of patients. The dynamic changes in the gene expression are examined during disease progression in serial samples from individual patients. Methods: Whole blood RNA was extracted from PAXgene tubes drawn from 63 patients with prostate cancer representing a range of clinical disease states. Serial samples were collected from a subset of patients (n=34). Survival prediction scores were derived independently based on the RNA-transcript levels of the six genes as quantified by a new ViiA 7 qPCR platform and by Illumina RNA-sequence. Results: The six gene signature scores generated on both qPCR and RNA-sequence platforms remain highly prognostic, separating mCRPC patients into high- and low-risk survival groups (logrank test; p-value for qPCR <0.001; p-value for RNA-seq <0.001). In contrast, prostate-specific antigen (PSA) fails to predict survival in the same cohort. There is a weak correlation between PSA and the six gene score (p-value of spearman correlation=0.14), indicating that the six gene score provides unique information regarding disease status. Patients with serial blood draws demonstrated no relationship with PSA changes. Patients with localized and hormone sensitive prostate cancer had significantly lower six gene scores (p-value=0.036). Individual gene analysis revealed that two genes, ABL2 and ITGAL, as the major contributors to the prognostic score. Conclusions: Our results further validate the prognostic significance of a six gene whole blood RNA signature, using an independent cohort and independent platforms.