P124

Abstract

Epigenetic biomarkers including aberrantly methylated DNA (amDNA) and non-coding RNA such as microRNAs (miRNAs) and long non-coding RNAs (lncRNAs) in biological fluids represent markers with the highest potential for cancer diagnostics. Tumor-specific amDNA is more easily and efficiently detected in blood plasma/serum in presence of large excess of wild type DNA compared with mutated DNA. miRNAs are much more stable than coding mRNA and are present in blood in detectable amounts. However still only a few cell-free DNA (cfDNA) -based diagnostics detecting epigenetic markers have come to the market. The main limitations of aberrantly methylated cfDNA (am-cfDNA) markers are age- and tissue-specific DNA methylation, low concentration of tumor-specific DNA along with presence of any conceivable methylated DNA pattern in circulation (Korshunova, 2009). Potential miRNA markers selection demands large scale primary analysis and overcoming some analytical complication like efficiency and uniformity of miRNA isolation and normalization at verification step. To overcome limitations of am-cfDNA use we utilized two approaches: Biased comprising selection of markers valid for tumor tissues, elaboration of the demands for inclusion/exclusion of cfDNA samples, development of PCR with absolute specificity and sensitivity of 1–10 molecules per sample, increasing of cfDNA concentration in the analytical sample. Epigenomics GA successfully used this approach for the EpiColon test development. Unbiased comprising search for potential markers in cfDNA/cfRNA using modern methods including MPS and subsequent bioinformatics data analysis, formulation of the requirements to analytical systems based on MPS data, development and validation of the diagnostics. Using first approach we selected 4 markers from 5 studied which proposed diagnostic value for lung cancer. Analytical sensitivity of methyl-specific TagMan PCR for these 4 genes and a control gene are 2, 3, 4, 10 and 10 copies of methylated DNA in 1000-fold excess of unmethylated DNA, correspondingly. Diagnostic sensitivity of lung cancer detection in the training group (30 healthy, 30 lung cancer patients, training cohort) for 4 genes are 13%, 46%, 73%, 30%, specificity – 100%, 97%, 100%, 97%, respectively. Inclusion/exclusion criteria are based on a control gene concentration in blood plasma. 2 markers in combination provide 89%sensitivity of lung cancer diagnostics and 97% specificity, 3 markers in combination demonstrate sensitivity and specificity of 92% and 97%. Unbiased approach included locus-specific MPS using MiSeq platform of 2 aberrantly methylated genes amplified with methyl-independent primers (barcoded individually) after bisulfite conversion of cfDNA isolated from blood of 18 healthy donors (HD), 20 prostate cancer (PC) and 17 benign prostatic hyperplasia patients (BPH). Analysis of individual methylation patterns and their representation demonstrates that methylation of GSTP1 gene-located 4 CpG-pairs in cfDNA from healthy donors differs compared with PC patients. Logistic regression-based model of 4 CpG-pairs methylation demonstrates 95% specificity and 100% sensitivity of PC diagnostics. Looking for miRNA markers 20 primary lung cancer patients (14 – squamous cell carcinoma, 6 – adenocarcinoma) and 10 healthy donors were enrolled in the study (training cohort). Circulating miRNAs were isolated by one-step phenol-free protocol (Zaporozhchenko et al., 2015). Concentrations of 179 miRNAs were measured in circulating miRNA pool by miRCURY LNA microRNA PCR Serum/Plasma Panel (Exiqon Ltd, Denmark). 18 miRNAs were differentially expressed between groups of healthy donors and patients with squamous cell carcinoma. 2 miRNAs provide 85% and 79% diagnostic sensitivity with 100% specificity. Verification using qRT-TaqMan PCR on a testing cohort including 75 lung cancer patients (squamous cell carcinoma and adenocarcinoma) and 50 healthy donors demonstrates 66% sensitivity and 89% specificity of the best selected miRNA. In conclusion, wide scale modern methods demonstrate high efficacy for epigenetic cfNA cancer markers development, whereas conventional approaches based on the current experience have not yet exhausted their potential.