Discovery of novel cell cycle regulatory and signal transduction modules driving Retinoblastoma using a correlative multi-omics approach

Abstract

Purpose: Applying a multi-omics methodology to molecular analysis of primary intraocular retinoblastoma tumor samples covering global transcriptomic and metabolomics assays for elucidating functional pathways driving this cancer.

Methods: All samples were collected after informed written consent and approval of Institutional Ethics Committee. We obtained total RNA from tumors of 9 patients (5 male, 4 female; age range 4-30 months) who underwent enucleation of the affected eyes. Additional tumor, aqueous humor, vitreous humor and tear samples were also collected from the same patients for metabolomics analyses. Control retina, vitreous and aqueous humor was extracted from pediatric donor eyes. Global gene expression and miRNA microarrays were performed simultaneously with the tumor RNA, followed by pathway analysis. Metabolites were extracted using monophasic solvent extraction of aqueous and vitreous humor samples followed by analysis on Accurate Mass QTOF mass spectrometer on reverse phase C18 and HILIC columns.

Results: Differential expression analysis carried out using moderated t-test with Benjamini Hochberg multiple testing correction revealed 1404 significantly regulated genes (p≤0.005 and fold change≥10) as compared to normal pediatric retina. Analysis of miRNA arrays revealed 18 previously unreported, deregulated miRNAs (p≤0.005 and fold change≥10). Using these data sets in concert, pathway analysis of the miRNAs and their possible target differentially expressed genes revealed novel networks. The results show that cell cycle, mTOR, PI3-AKT, retinal function and Rap1 signaling modules are most significantly deregulated. We further validated 20 genes selected from these deregulated pathways by quantitative PCR. IHC staining of 33 independent patient tumor cohort further validated E2F and Rap1 up regulation. Analysis of the differentially expressed metabolites in vitreous and aqueous humor revealed enrichment of several additional pathways.

Conclusions: The study illustrates an integrated method of discovering new biological insights that are made accessible by correlating data from different large scale techniques in the same patient sample, thereby reducing intra-cohort bias. In particular, E2F, Rap1, CDK and cyclin dependent signaling pathways were up regulated while retinal function and visual cycle related pathways were down regulated in Rb patients.