Box 1. Methods for detecting CpG island methylation.

Methods for the analysis of CpG-island methylation are available both genome-wide and at the single gene level. Restriction landmark genomic scanning (RLGS) is performed by digesting genomic DNA with a methylation-sensitive restriction enzyme, end labelling of the resulting DNA fragments and subsequent digestion with two different restriction enzymes and 2-dimensional gel electrophoresis (Costello et al, 2000). Comparison of signal intensities between tumour and normal DNA after autoradiography allows estimation of the number of aberrantly methylated CpG islands in tumours, and individual aberrantly methylated CpG islands can be identified by sequencing. Differential methylation hybridisation (DMH) is an alternative means of examining genome-wide methylation patterns that uses restriction digestion of genomic DNA and ligation to linkers (Huang et al, 1999), followed by digestion with a methylation-sensitive restriction enzyme such as BstUI, PCR amplification and hybridisation to CpG-rich DNA sequences (representing putative CpG islands). Comparison to hybridisation signals obtained from undigested linker-ligated DNA allowed the identification of aberrantly methylated CpG islands. Methylation sensitive-representational difference analysis (MS-RDA) uses genomic tester and driver DNA samples digested with the methylation-sensitive restriction enzyme HpaII (Ushijima et al, 1997). Sequences that are specific for the tester amplicon are subsequently enriched by repeated cycles of subtractive hybridisations.

Several methods for the analysis of the methylation status of individual CpG islands utilise bisulphate treatment of DNA, which has been described in detail (Grunau et al, 2001; Warnecke et al, 2002). Bisulphite treatment of DNA converts unmethylated cytosines into uracil but does not affect methylated cytosines. A difference in methylation is thus converted into a difference in sequence. A widely used method for analysing the methylation status of specific sequences is methylation-specific PCR (MSP) (Herman et al, 1996). Methylation-specific PCR is performed using primers specific for either unmethylated or methylated sequences, thereby allowing the detection of the respective methylation state. Among the advantages of MSP are its easy detection due to its gain-of-signal character and its high sensitivity, allowing the detection of as little as 0.1% methylation in a DNA sample (Herman et al, 1996). The MethyLight technique also involves bisulphite modification. Fluorescence-based PCR is then performed with primers that either overlap CpG methylation sites or that do not overlap any CpG dinucleotides. Sequence discrimination can occur either at the level of the PCR amplification process or at the level of the probe hybridisation process or both (Eads et al, 2000). Combined restriction analysis (COBRA) uses primers that amplify the template irrespective of its methylation state (Xiong and Laird, 1997). The PCR product should therefore be heterogeneous and reflect the various methylation states present in the template. Discrimination of methylation states is achieved by restriction digest using a restriction site whose presence after bisulphite modification depends on the methylation state of the DNA. Combined restriction analysis allows the quantification of the methylation, but its disadvantage is that the methylation of one CpG site is not necessarily representative for the other CpG sites in the analysed sequence. The highest accuracy of methylation density in a region of DNA is achieved by bisulphite sequencing. As in COBRA, the modified DNA is amplified irrespective of its methylation state, but subsequently the amplicon is subcloned and sequenced. This not only allows detection of methylation with a single-nucleotide resolution but also gives information about the distribution of methylated cytosines within individual DNA molecules. The disadvantage is that bisulphite sequencing is relatively labour-intensive.