Universal amplification of DNA isolated from small regions of paraffin-embedded, formalin-fixed tissue.

The polymerase chain reaction (PCR) analysis of DNA isolated from formalin-fixed, paraffin-embedded tissue can be difficult because of the inherently poor quality of template DNA available for amplification. As a consequence of the formalin fixation process, DNA is complexed with proteins and is often nicked, giving relatively short fragments. Such samples are often of low DNA concentration and poor quality. A high rate of random failure, the inability to routinely amplify fragments greater than 650 bp and poor storage make paraffin-embedded tissue a less than ideal source of DNA template (2). We wanted to assay specific cells from very small regions of paraffin-embedded tissue sections for loss of heterozygosity (LOH) using microsatellite markers. Various protocols have been written for the successful extraction of DNA from a small number of paraffinembedded cells (2,5), but these protocols were not amenable to genomic screening projects because of the difficulties previously described. Universal genomic amplification is a way to overcome the problems associated with DNA isolated from paraffin-embedded tissue. There have been several protocols reported for the generation of synthetic genomic template, the most notable being Alu PCR (4), degenerate oligonucleotide-primed PCR (DOP-PCR) (8) and primer extension pre-amplification (PEP) (9). The latter has been applied to achieve multiple PCR assays on single cells and has recently been proven useful for extending limited DNA samples (1). The strength of the PEP protocol is the opposite of PCR—it relies on random amplification without any preference for specific sequences. Control genomic DNA was prepared from blood collected in EDTA using standard salting out procedures and quantitated spectrophotometrically (3). DNA was isolated from formalin-fixed, paraffin-embedded tissue sections of unknown age, fixation method or fixation time by the following procedure. Regions of cells were identified from hematoxylin and eosin-stained sections, and the selected cells were scratched from a single corresponding, unstained 10-μm slide section and transferred to 500-μL PCR tubes. The area of the scratched material ranged from about 1 to 10 mm2, depending on the number of cells selected. Samples were heated to 95°C for 10 min under paraffin oil in 20 μL of 1× PCR Buffer II (Perkin-Elmer, Norwalk, CT, USA) with periodic mixing to aid de-waxing of each sample. After cooling briefly, samples were digested with proteinase K (1000 μg/mL) at 65°C for 90 min followed by enzyme inactivation at 85°C for 15 min. This solution (10 μL) was then used as DNA template for universal amplification. Larger samples were purified by chloroform extraction followed by salt/isopropyl alcohol precipitation. DNA pellets were resuspended in 20 μL TE buffer (2 mM TrisHCl, 0.2 mM EDTA, pH 8.0). Universal amplification products were generated using a modification of the PEP protocol (9). Standard PEP reactions were performed under paraffin oil in 60-μL volumes containing 200 pmol of random 12-mer oligonucleotides, 5 U AmpliTaq DNA Polymerase (Perkin-Elmer), 200 μM of each dNTP, 5 mM MgCl2, 1× PCR Buffer II and DNA template. Templates consisted of either purified DNA extracted from paraffin-embedded tissue sections, 10 μL of the proteinase K digest of cells dissected from paraffinembedded tissue sections or a known quantity of the control DNA. Water blanks were included as contamination checks for each amplification series. In