Sodium bisulfite analysis of the methylation status of DNA from small portions of paraffin slides.

DNA methylation at CpG dinucleotides has been intensely studied to elucidate its roles in normal and aberrant cellular processes, including genomic imprinting, modulation of transcription, X-inactivation, gene silencing, and carcinogenesis. A number of techniques for the detection of methylated cytosines within a DNA sequence have been developed, and these have been refined and improved over time. DNA methylation analysis involving the conversion of genomic DNA by the sodium bisulfite technique, coupled with cloning and sequencing, is a powerful method for ascertaining the methylation status of individual cytosine residues within a continuous stretch of DNA (2–4). In this method, unmethylated cytosines are converted to uracil residues via a series of chemical treatments, including incubation with sodium bisulfite (8–10), while methylated cytosines remain unmodified. A segment of treated DNA can then be amplified by PCR, resulting in the representation of unmethylated cytosines by thymine residues, and the PCR product can be cloned and sequenced. By comparing the original and bisulfite-converted sequences, unmethylated cytosines can be identified by their position in the thymine lane, in contrast with methylated cytosines, which maintain their normal positions as cytosines. Aside from the ability to determine the methylation status of every cytosine residue within the PCR product, the sodium bisulfite method has the further advantage of requiring smaller quantities of starting DNA than many other methods. More recent modifications of the protocol have further reduced the amount of starting material needed by embedding DNA or cells in agarose to retain the genetic material through a series of washes and to keep DNA in a denatured state. This allows for full reactivity of sodium bisulfite (7). For many applications, particularly those investigating the connection between aberrant DNA methylation and cancer, archival, formalin-fixed, paraffin-embedded specimens would be a desirable starting material. However, the heterogeneity and close proximity of normal tissue inherent in many tumors on a microscopic scale, coupled with the difficulties associated with extracting sufficient quantities of DNA from small volumes of paraffin to perform methylation analysis, have rendered these archival specimens of limited use. Here, we present a simple protocol for performing sodium bisulfite sequencing on pieces of paraffinembedded tissue sections with areas as small as 1 mm2. Intact glass microscope slides containing 10-μm sections of formalinfixed, paraffin embedded human liver, a 10-μm section of normal human breast tissue, or a 5-μm section of breast tumor tissue were deparaffinized through 3 × 5-min washes in xylene, followed by rehydration in descending alcohols: 100%, 2 × 2 min; 95%, 2 × 2 min; 70%, 30 s. Slides were then soaked in PBS. The desired portions of each slide as described in Table 1 were then carefully scraped away with razor blades and sterile pipet tips and put into 600-μL tubes. Small quantities of lysis buffer (20–50 μL; 0.5% Tween 20 in TE) and proteinase K (10 mg/mL) were added to tubes as shown in Table 1. Mineral oil was added to each tube; all tubes were then centrifuged briefly to ensure that tissue and proteinase K were contained in the aqueous phase. Tubes were incubated for three days at 65oC, with the addition of proteinase K every 24 h as described in Table 1. Following the fourth and final addition of proteinase K, samples were incubated for 1 h at 65oC and then heated to 95oC for 5 min. For each bead, 9 μL molten 2% low-melt agarose were heated to 80oC and mixed with 1 μL digested tissue solution. A 10-μL aliquot of this mixture was then carefully dropped into 500 μL ice-cold mineral oil and either stored at 4oC or used directly in the following steps. Sodium bisulfite conversion of agarose-embedded genomic DNA was performed as described (7), beginning with a series of washes in TE. All washes were performed in 500 μL solution. To determine the efficacy of the technique in converting DNA from small amounts of paraffin embedded tissue, a 305-bp portion of the estrogen receptor-α gene (GenBank accession no. X03635) CpG island was amplified by PCR with semi-nested primers; the need for nested or semi-nested primers is gene and primer-pair dependent and was determined empirically. All PCRs were hot started with 1 μL of the reverse primer and carried out under the following conditions: 5 cycles at 94oC for 20 s, 55oC for 30 s, 72oC for 45 s, followed by 30 cycles at 94oC for 25 s, 55oC for 40 s, 72oC for 25 s, and a final extension step at 72oC for 16 min. In the first round of amplification, outside forward primer DR73 (5′-TGATTTTTTATAAAGTATTTGGGATGG-3′) and reverse primer DR208 (5′-ATACAATAACATCAACAAACTCAAAAACAC-3′) were used to amplify directly from the agarose/DNA slurry in a 50Benchmarks