Introduction

Hypermethylation of CpG islands in promoter regions of numerous genes involved in cell growth, differentiation, and DNA repair processes is now recognized as an important and early event in carcinogenesis (1,2,3,4,5,6,7). Methylation of the cytosine residue of CpG dinucleotides epigenetically silences gene expression through two main mechanisms: (i) by triggering a series of events leading to a closed chromatin structure that inhibits transcription and (ii) by altering the DNA binding and specific action of transcriptional regulatory proteins via methylation of specific cytosines located within their binding sites (8,9).

Given the importance of cytosine methylation in several key biological processes, such as genomic imprinting, chromosome X inactivation, and carcinogenesis, there has been great interest in developing methods to detect and quantify methylation. One major breakthrough central to the majority of these methylation detection methods has been the use of sodium bisulfite (NaBi) to treat genomic DNA, a process that converts all nonmethylated cytosines to uracil (later replicated as thymidine during PCR), while leaving methylated cytosines intact (10,11). The ability to predict sequences of methylated, M, or unmethylated, U, DNA after NaBi conversion facilitated the subsequent development of a PCR-based method called methylation-specific PCR (MSP) (12). In MSP, the specificity of methylation detection hinges on primer design, whereby two sets of primers (an M and a U primer set) are used to characterize methylation for each gene.

Because standard MSP is a qualitative reaction that renders either positive or negative results, it is not informative regarding relative amounts of methylation; a critical problem that is magnified by the presence of low levels of DNA methylation in some normal tissues (13,14,15). To address this concern, a number of quantitative MSP techniques have been developed in recent years to evaluate either relative amounts of DNA promoter hypermethylation between normal and affected tissues or methylation of a specific CpG dinucleotide. These include realtime applications such as quantitative MSP (also known as MethyLight) (16), HeavyMethyl (17), MethylQuant (18), and quantitative multiplex MSP (QM-MSP) (15). To our knowledge, all of these methods currently use two reaction wells per sample to characterize methylation of a single gene. Specifically, in MethyLight, the M reaction is carried out in one well, while a reference gene such as actin is amplified in the second well (16). Similarly, the gel-based MSP methods of Herman et al. (12) and the real-time QM-MSP method of Fackler et al. (15) use one well for primers specific to the unmethylated U gene fraction and a second well for primers specific to the methylated M gene fraction.

MSP presents unique challenges that distinguish it from traditional PCR, including successful modification of genomic DNA template via NaBi treatment prior to MSP and the need to characterize both the methylated and unmethylated fractions of a particular target DNA; or in the case of some quantitative MSP techniques, to measure the methylated fraction of a gene relative to an unmethylated reference gene such as actin. Moreover, the use of multiple fluorophores in any real-time quantitative MSP presents additional concerns above and apart from traditional reverse transcription PCR (RT-PCR) that are related to quantification of the U + M DNA targets in real-time and the added competition between U + M probes and primers for similar targets. Finally, the special requirements for U + M real-time primers and probes with complementary melting temperatures (T_ms) are more difficult to achieve, because all unmethylated cytosines are converted to uracil during NaBi treatment. All of these factors contribute to the complexity of real-time MSP.

While several quantitative MSP techniques offer between 10- to 100-fold increases in sensitivity over standard gel-based MSP, they are more costly, and the real-time reaction is further compromised by singleplexing of U and M primer/probe sets, which limits the number of samples that can be analyzed per plate. As quantitative MSP techniques hold promise for translation into the clinical setting, improved efficiency and affordability is needed for their application as diagnostic tests.

We now report an improvement in the real-time step of QM-MSP, which utilizes two primer/probe sets labeled with fluorophores of FAM™ and VIC®, respectively, to co-amplify two methylation-specific DNA targets in the same well. In this paper, we compare the realtime step of the traditional one-color QM-MSP method with three possible protocols of our improved method: (A) a two-gene reaction wherein U₁+U₂ (unmethylated primer/probe sets for genes 1 + 2) and M₁+M₂ (methylated primer/probe sets for genes 1 + 2) are co-amplified in one well, respectively; (B) a two-gene reaction where U₁+M₂ (an unmethylated primer/probe set for gene 1 and a methylated primer/probe set for gene 2) and M₂+U₁ (a methylated primer/probe set for gene 2 and an unmethylated primer/probe set for gene 1) are co-amplified in the same well; and (C) a single-gene reaction where the U₁ and M₁ primers and probes for the same gene are co-amplified in one well; hereafter referred to as protocols A, B, and C, respectively. While co-amplification with two or more fluorophores has been performed for standard RT-PCR (19,20,21), to our knowledge, this is the first report for real-time MSP applications that either two genes or U + M primer/probe sets for the same gene have been simultaneously quantified in a single well. This two-color modification can be applied not only to QM-MSP, but to any real-time MSP experiment.