Detection of BCR-ABL Mutation by High Resolution Melting Analysis Is Affected by the Sequence Content, Position and Type of Nucleotide Change and DNA Dye Together with Salt Composition of PCR Mixes.

Abstract 1565

Poster Board I-588

Mutations in the ABL kinase domain of the BCR-ABL gene are one of the important mechanisms of resistance to tyrosin kinase inhibitor imatinib, now used as a standard first-line treatment of newly diagnosed CML patients. Various methods with different sensitivity are used to detect BCR-ABL mutations including High-resolution melting analysis (HRM) which allows to screen DNA samples for single-base changes, insertions/deletions, or other unknown mutations based on their dissociation behavior as they transit (melt) from double stranded (dsDNA) to single stranded DNA (ssDNA) form with increasing temperature.

To validate previously reported HRM analysis (Poláková KM et al., 2008, Leuk Res 32, 1236–1243) as a routine screening test for detection of mutations in ABL kinase domain we used LCGreen dye combined with Fast Start Taq polymerase and a PCR mix from Roche Applied Science. Four primer pairs divide the ABL kinase domain into four amplicons (HRM1–HRM4); HRM1 (221bp) covers amino acids P216-E275, HRM2 (225bp) amino acids F283-M343, HRM3 (239bp) amino acids A350-K415 and HRM4 (241bp) amino acids L429-T495. Sixteen different mutations at different ratios (previously identified and quantified by direct sequencing) each in at least four samples were used in duplicates for HRM analysis.

HRM results were concordant with those obtained by direct sequencing for 9/16 mutations. For 3/16 mutations we were not able to detect in samples with 100% mutant template content. This issue could be improved by addition of exogenous wild-type DNA after PCR to form heteroduplexes. One mutation in HRM1 and three mutations in HRM3 regions were not detected reliably and presumably PCR reaction conditions need to be changed. Our results proved that HRM analysis is more sensitive than direct sequencing since mutations were detected by HRM earlier in 6 of 9 tested cases.

As a next step we assessed the HRM curves of templates amplified with 4 different PCR mixes combined with several DNA dyes. We tested 6 dyes (SYBR Green, SYTO-9, SYTO-13, EvaGreen, LCGreen, ResoLight) and 4 PCR mixes differing in their salt content (24 combinations in total) to evaluate their use in HRM analysis with respect to mutation detection sensitivity and reliability. For this test, HRM1 was applied to detect E255K mutation in samples containing 100%, 50%, 15%, 5% and 0% of the mutant BCR-ABL. HRM analysis was performed immediately after PCR, and normalized melting curves and difference graphs were calculated.

Mutant templates representing 15%, 50% and 100% of mutation were distinguished using all combinations of dyes and mixes. Enhanced variations of HRM scoring values in individual experiments of 100% mutant templates, reflected by higher SD values, were noticed in some PCR mix-DNA dye combinations. HRM analysis of samples containing 5% of mutated templates showed that various HRM dyes differed significantly in their ability to detect mutants. LCGreen, SYTO-9 and SYTO-13 in all mixes significantly detected 5% mutant templates (P<0.05 or P<0.01). HRM performed with SYBR Green did not reliably detect 5% mutant templates in none of mixes tested whereas other dyes differed in their capabilities to detect this mixture depending on PCR mix used.

In conclusion, our data indicate that the sequence content of HRM1-4, type of mutation, composition of PCR mix as well as DNA-binding dye affect HRM performance indicating careful choice of DNA with PCR mix and primer design for successful HRM assay reliably detecting mutations in BCR-ABL kinase domain. We showed that HRM is a helpful tool for sequencing data confirmation and earlier detection since observed by sequencing.

Supported by MZOUHKT2005.