The Senior Investigator moved to the NCI during FY2011 and hired personnel to set up a research laboratory during the first half of FY2012. A Staff Scientist was recruited in November 2011, and this individual initially helped to set up the laboratory;this scientist then began working on this RMS methylation project in March 2012. The staff scientist applied bioinformatic strategies to the methylation data from the first set of RMS tumors to identify genes that are differentially methylated between fusion-positive ARMS and fusion-negative ERMS tumors. Supervised analysis algorithms were used to compare mean methylation levels for each gene in the fusion-negative and fusion-positive tumors, and then to determine those genes whose mean methylation levels were significantly different between the two groups of tumors. As there were greater than 27,000 comparisons, corrections were made for multiple comparisons. Genes were found for which there was both evidence of hypermethylation in ARMS and hypomethylation in ERMS, and other genes were found for which there was evidence of hypomethylation in ARMS and hypermethylation in ERMS. As each methylation measurement applies to a single CpG site, confidence and interest increased when the relationship was found for more than one CpG in the same gene. Our next goal was to validate the differential methylation of selected genes by determining the methylation status of the larger CpG-containing regulatory region in fusion-negative and positive RMS tumors. Based on some trial experiments, it was clear that we needed to optimize methodology in our laboratory to accurately measure gene-specific methylation levels in tumor samples. We initially used RMS tumor cell lines to work out this methodology and then we will subsequently use DNA from RMS specimens in our later analyses. Our goal was to treat genomic DNA with bisulfite, amplify selected CpG containing regions, and then characterize the methylation status of this region by pyrosequencing. We determined that one of several commercially available kits for controlled bisulfite treatment of genomic DNA efficiently converts all cytosines to uracil. In previous studies, using in vitro methylated genomic DNA controls, we also confirmed that the bisulfite treatment from these kits does not modify methylated cytosines. Therefore, we can confidently use this bisulfite procedure to modify the sequence of DNA in a methylation-specific fashion. To amplify the selected regions, there were several considerations in our design of PCR primers and the subsequent PCR reaction. First, the sequence of the region was not the native sequence but the modified sequence after bisulfite treatment. The sequences of interest were the CpG-rich regulatory regions including the CpG's of interest identified in the Illumina array. However, it should be emphasized that we select flanking primers that do not contain CpG dinucleotides and thus do not have a variable sequence dependent on methylation status. Of the available algorithms for selecting primers, we determined that Methyl Primer Express software (Applied Biosystems) was useful in selecting robust primer pairs. In addition to the gene specific sequences in the primer, we also determined the utility of adding tags on the 5'ends of the forward and reverse primers (such as M13 forward and reverse primer sequences). These tags permit a second nested step if the first PCR results in a low yield of product. In addition, a single common biotin-labeled M13 forward primer can be used to isolate the single stranded product from the double stranded PCR products in preparation for the pyrosequencing studies. Once the PCR primers were selected, we looked for optimal reaction conditions, particularly for these high-CpG-containing regions. In particular, we found that addition of BSA and glycerol to the reaction conditions permitted increased specificity by allowing use of higher annealing temperatures during cycling. Based on these various parameters, we have amplified CpG-containing regions for several differentially methylated genes from a series of fusion-negative and positive RMS cells lines, and are now commencing pyrosequencing analysis to determine the C:T ration at cytosines within CpG dinucleotides and thereby determining the methylation status of these sites within the larger CpG-containing region.
