Use of formalin-fixed paraffin-embedded tumor tissue as a DNA source in molecular epidemiological studies of pediatric CNS tumors.

Formalin-fixed paraffin-embedded tissue (FFPET) samples are a potential source of DNA for molecular epidemiological studies. However, the use of FFPET samples can be restricted by the yield and quality of DNA isolated. The aim of this study was to examine whether FFPET biopsies from pediatric central nervous system tumors were a feasible alternative to archival frozen tissue when characterizing common gene polymorphisms. DNA was isolated from 50 frozen pediatric central nervous system tumor biopsies and matched FFPET samples. Real-time polymerase chain reaction (PCR) was used to quantify DNA and characterize GSTT1, GSTM1, GSTP1, and MTHFR gene polymorphisms. The use of whole-genome amplification (WGA) to increase DNA yields was also investigated. The results showed that DNA isolated from FFPET samples was more fragmented and provided smaller yields than DNA isolated from frozen samples. Attempts to increase the DNA yield from FFPET using WGA were unsuccessful. DNA from FFPET samples was successfully genotyped for the GSTP1 Ile105Val and MTHFR 677 C>T polymorphisms in 98% of samples and was 100% concordant with the results from frozen tissue. However, DNA from FFPET performed poorly in real-time PCR assays for GSTM1 and GSTT1 deletion polymorphisms. Our investigations show that DNA extracted from FFPET is substantially fragmented and not readily amplified using WGA. In addition, careful validation of PCR assays should be carried out due to the variable amplification of fragmented FFPET DNA.