J. D. Rice1, H. Roberts1, M. R. Eichenberger1, J. Pan2, S. Rai2, S. Galandiuk1 1University Of Louisville,Hiram C. Polk, Jr., MD Department Of Surgery,Louisville, KY, USA 2University Of Louisville,2Department Of Bioinformatics And Biostatistics,Louisville, KY, USA
Introduction:
Plasma miRNAs have been shown to be promising biomarkers for cancer. A large problem is the lack of standardization in current data acquisition and reporting. Part of this problem lies in the numerous different housekeeping genes (HKG) used in real time PCR data acquisition. With multiple accepted HKG, data comparison and validation between different investigators becomes difficult. Due to issues of data reproducibility and standardization, it is crucial that a plasma HKG 1) be expressed in all samples, 2) have medium to high levels of expression and 3) have consistent levels of expression.
Methods:
We screened for 380 miRNAs using microfluidic array technology (Applied Biosystems), in a discovery cohort of 20 colorectal cancer (CRC) patients, 10 patients each with colorectal adnoma (CRAd), breast cancer (BC), lung cancer (LC), pancreatic cancer (PC) and controls. We investigated Let-7a, Let-7d, Let-7g, miR-16, RNU6, RNU48, miR-191, miR-223, miR-484 and miR-520d-5p as potential HKG’s based both upon our previous data and reports in the literature. The mean ∆CT and standard deviation were calculated for each HKG. Only microRNAs with >90% expression were included in this statistical analysis.
Results:
RNU6, miR-520d-5p, miR-16, miR-191, miR-223, miR-484 were expressed in all samples. Let-7a, Let-7d, Let-7g and RNU48 were not expressed in 26%, 7%, 10% and 8% of samples respectively. U6 and miR-520d-5p had the most consistent ∆CT and lowest standard deviation.
Conclusion:
Let-7a, Let-7d, Let-7g and RNU48 were not expressed in all samples and were therefore not considered reliable HKG’s. Investigating the remaining HKG with 100% expression in all samples, RNU6 and miR-520d-5p are excellent HKG candidates for studies of plasma miRNA due to consistent ∆CT in all samples and a very narrow standard deviation.
