DNA Methylation Biomarkers of Primary Liver Cancer Risk

NIH Small Grant Program (R03, submission October 2014) and Showalter Research Trust (internal funding, pending) will be used to perform genome-wide DNA methylation analysis and validation of the top changes by pyrosequencing in 36 cases and 36 controls. These preliminary data will be used for NIH Research Project Grant Program application (ROl) that will be aimed at accomplishing the entire project.

Epigenetic modifications, particularly DNA methylation, have attracted a significant amount of attention for the prevention and treatment of different illnesses with cancer at the forefront, mainly due to the inherent reversibility of epigenetic states. Aberrations in DNA methylation patterns have been linked to cancer development and progression in many studies in the last decades. Hypermethylation of tumor suppressor genes linked to transcriptional silencing, global DNA demethylation associated with genome rearrangements and instabi lity, and recently reported promoter hypomethylation linked to activation of oncogenes and prometastatic genes are hallmarks of nearly all types of cancer. DNA hypermethylation of tumor suppressor genes has been shown to have diagnostic potential in several cancers whereas DNA hypomethylation of specific gene promoters has recently been shown by our group as potential biomarker of hepatocellular carcinoma (HCC), one of the most common liver cancer types with rising morbidity and mortality rates. As late onset of clinical symptoms in HCC accounts for late diagnosis and poor prognosis and early detection of HCC increases cure rate from 5% to 80%, identifying reliable biomarkers of risk prediction is of high interest. Such markers would have a high application in clinics only if detectable by minimally invasive tests like blood test. There is previous evidence for the feasibility of this approach in other cancers. A MAGEA gene expression test that provides information on multiple MAGEA genes in a single reaction detected circulating tumor cells in blood of melanoma, breast and colorectal patients. Brennan et al. discovered methylation in ATM intragenic loci in DNA from white blood cells as a potential marker of breast cancer risk.

Circulating methylated SEPTIN 9 DNA and IFFOl DNA in plasma were found to correlate with the occurrence of colorectal cancer and ovarian cancer, respectively. Such biomarkers would improve individual risk prediction and enable selection of individuals for prevention or early detection interventions.

We intend to identify DNA methylation biomarkers of HCC in blood using samples from case-control studies in prospective cohorts. All cohorts are welcome to participate if they have eligible samples, namely blood samples collected prior to diagnosis. Around 300-500 cases will be matched with controls on age, recruitment center, and blood collection date. We will perform genome-wide analysis of DNA methylation profile in 100 cases and 100 controls using Illumina Infinium 450K microarray platform and derive DNA methylation signatures (differential methylation) associated with liver cancer risk. We will use Genome Studio Data Analysis Illumina software for bioinformatics analyses of data in order to calculate methylation level at every CpG site (average beta), differential methylation level between cases and controls (delta beta) and statistical significance (differential score) . As a white blood cell mixture will be used as a source of DNA in our proposal, we will apply existing data and statistical methodologies (Houseman's algorithm-negative R) in R software environment for statistical computing to account for the possibility that changes in cell type proportions could be registered as DNA methylation differences. The top differences between cases and controls will be validated by pyrosequencing in the entire set of samples and biomarkers of risk prediction will be established. The functional significance and the predictive value of combined biomarkers will be tested using molecular and bioinformatics methods.

Findings of the proposed research project will be of great relevance to liver cancer prevention and epidemiology. We will identify epigenetic biomarkers that can predict the risk of cancer development opening the door to prevention and improving disease management.

The goal of the project is to establish DNA methylation biomarkers of primary liver cancer risk in blood samples. Good risk predictors can identify a group of the population in which a majority of cases will occur and the benefits of prevention or early detection will outweigh the harms.

A. Delineate genome -wide DNA methylation profile and derive DNA methylation signatures associated with liver cancer risk. We will use Illumina Infinium-Human Methylation -450K Beadchip microarray for analysis of genome-wide DNA methylation profile. Using bioinformatics tools, we will then determine differences in DNA methylation between controls and case samples collected prior to diagnosis with primary liver cancer.
B. Validate the top differences in DNA methylation signatures for biomarkers of risk prediction (changes prior to diagnosis)using pyrosequencing.

Genome-wide DNA methylation profile will be assessed using Illumina Infinium 450K microarrays and analyzed with Genome Studio Illumina software and R software environment. Established biomarkers will be validated using pyrosequencing following bisulfite DNA conversion; the results will be analyzed with PyroMark Q24/Q96 Analysis Software.

Our goal is to elucidate DNA methylation in 300-500 blood samples from cases and controls collected prior to diagnosis. Taking into account 5-15% survival rate in liver cancer, the Cox regression model indicates testing approximately 300-500 samples for high predictive value of biomarkers. Including different cohorts from different geographical regions will ensure a proper number of samples and will establish robust DNA methylation biomarkers for primary liver cancer risk .

Per-cohort minimum of 10 cases (blood samples collected prior to diagnosis)

Clinical characteristics including type of the treatment that has been applied and the response to the treatment.

None

Data on when diagnosis was made relative to blood collection date. Stage of tumor at the time of diagnosis.

Yes

None