The ability to distinguish between methylated and unmethylated DNA was the first obstacle this technique had to overcome.To create a known methylated sample, unmethylated Lambda DNA was treated with SssI Methylase and then purified. Both treated and untreated DNA, in the amount of 100 ng, were added to assay wells in triplicate and analyzed as indicated above. The untreated sample was comparable to the blank indicating no methylation present. The treated sample had a sixfold increase in signal indicating that methylation occurred due to treatment with the methylase (Figure 1).
The next experiments were designed to prove that the kit would work with DNA purified from a biological system. These tests took advantage of azacytidine, which is a chemical analog of cytidine that cannot be methylated. When added to cell culture media it is incorporated into DNA during replication and thus creates DNA with a decreased global methylation. Azacytidine was added directly to the media of CHO DG44 cells. The cells were harvested after 48 hours. Global methylation was measured on 40 ng of the purified CHO DG44 DNA in triplicate from treated and untreated cells. The treated cells had a threefold decrease in methylation due to the azacytidine treatment (Figure 2).
DNA hypomethylation is of particular interest to cancer researchers. In cancer, hypomethylation is seen in heterochromatic DNA repeats, dispersed retrotransposons, endogenous retroviral elements, and unique sequences such as transcription control sequences. It is thought that these events occur early in the tumorigenesis process, and global hypomethylation could be a potential biomarker.
DNA from matched tissue and tumor samples was evaluated with our method to determine if shifts in global methylation could be detected. DNA was isolated from bronchioalveolar carcinoma and adjacent normal lung tissue from a 72-year-old male. Global DNA methylation was measured for both samples and compared to a fully methylated human genomic DNA control.
As expected, the adjacent normal lung tissue was less methylated than the fully methylated control and the tumor DNA was hypomethylated when compared to normal lung tissue (Figure 3). In a second experiment, DNA was isolated from hepatocellular carcinoma and adjacent normal liver tissue from a 54-year-old male.
Global DNA methylation was measured for both samples and compared to a fully methylated human genomic DNA control. The normal liver DNA was slightly less methylated than the control. Previous publications indicate that hepatocelluar carcinoma is hypomethylated. Our data confirms that tumor liver DNA is hypomethylated in this comparison.