Acute myeloid leukemia with a normal karyotype (NK-AML) is considered to have an intermediate prognostic risk with 5 year disease free survival (DFS) ranging between 24–42% 12. However, there is marked variability in outcome suggesting significant biological and molecular heterogeneity within this group of AML 3.

In 2005, Falini et al. described a set of common mutations within the final exon of the NPM1 gene in primary NK-AML patients, which alter the N-terminal domain nuclear localisation signal leading to abnormal cytoplasmic accumulation of the NPM1 phosphoprotein 4. While the precise functional effect of the NPM1 mutation is incompletely understood, several groups confirmed that NK-AML patients have a high incidence of NPM1 exon 12 mutations (~24% – 60%) 56789. Mutations in NPM1 are the most frequent genetic change known in patients with NK-AML and a number of studies have shown that NPM1 mutation positive patients have a better prognosis with longer event-free and overall survival (OS) 10.

Schnittger et al. demonstrated that the favourable prognostic implications of NPM1 mutation status are overridden in FLT3-ITD positive cases which have a uniformly poor prognosis 7. These findings demonstrate the need to screen patients for mutations in FLT3-ITD alongside NPM1 10. However, such a molecular screening program can be demanding on the resources of a diagnostic laboratory. Therefore, in this study we assessed the use of high resolution melting (HRM) analysis as a rapid method to screen NK-AML patient samples for the critical molecular changes in NPM1 and FLT3.

In this study, we developed HRM assays allowing rapid assessment of the mutation status of NPM1 and the presence of the FLT3-ITD in the same run. In HRM, the PCR product is subjected to melting in the presence of a dye that only fluoresces when bound to double stranded DNA 11. As melting is sequence dependent, monitoring the precise melting behaviour by observing the change in fluorescence allows the detection of variant sequences. In addition, sequence variants in the DNA such as mutations give rise to heteroduplexes that form earlier melting products allowing ready detection of mutations even at comparatively low concentrations.

Samples from 44 patients with NK-AML were analysed. The median age of the patients was 62 years (range 18–89 years) and 27 (61%) patients were male. Twenty nine (66%) had de novo AML and 15 (34%) had secondary AML. Sixteen patients generated an abnormal melting profile in one of the two tested amplicons, 8 were NPM1 mutation positive only, 4 were NPM1 positive and FLT3-ITD positive and 4 were FLT3-ITD positive only (Figure 1.).

Sequencing confirmed all the HRM detected mutations and did not reveal any further mutations, indicating that HRM was capable of detecting mutations with 100% sensitivity in this cohort.

All the NPM1 mutations detected involved one of two 4 base insertions that altered the tryptophan at amino acid position 288 and the FLT3-ITD ranged from 33–102 bases (Table 1). These mutations were similar to those previously described 41213. All 12 NPM1 mutation positive patients were also positive by immunohistochemistry (IHC) on bone marrow trephine sections, showing typical cytoplasmic localisation (data not shown).

The incidence of NPM1 mutations in the de novo AML cases was 40% (12/29), consistent with the incidence reported in previous studies 56789. Interestingly, 3/15 of the secondary AML cases were NPM1 mutation positive which contrasts with an earlier study, where cytoplasmic localisation of NPM indicative of NPM1 mutations was not seen in 135 secondary AML samples by IHC 4.

A novel point mutation Y572C in exon 14 of FLT3 was also detected. This tyrosine residue within the juxtamembrane domain of FLT3 has been shown to be phosphorylated in vivo 14 and could be included in the newly described class of FLT3 juxtamembrane domain point mutations for which the similar mutation Y591C has been reported 15. This illustrates the power of HRM to detect novel as well known mutations. The use of HRM to screen for FLT3-ITD has been previously reported 16.

HRM is rapidly becoming the most important mutation scanning methodology. It is an in-tube method, meaning that PCR amplification and subsequent analysis are sequentially performed in the one tube or well. This makes it more convenient than other scanning methodologies such as denaturing high-performance liquid chromatography 17. We used a real-time PCR machine with HRM capability rather than a stand-alone HRM instrument. This facilitates quality control as the success of the amplification can be assessed on the same platform as the melting analysis.

HRM has no real disadvantages in mutation scanning except that extra care needs to be taken in designing PCR reactions to avoid primer dimers and non-specific amplification. Secondly, DNA needs to be prepared in a uniform fashion to avoid variation in salt concentration that will affect the melting. In addition, the exact nature of any mutation cannot be determined without sequencing. Nevertheless, performing HRM as an initial screen for potential mutations significantly reduces the volume of samples requiring sequencing with consequent reduction of cost and labour, and improvements to turn around time.

HRM is likely to play a major role in clinical applications as it enables rapid detection of defined and novel molecular changes in clinical samples. In this study, the conditions have been optimised to enable screening of normal karyotype AML patients for both NPM1 and FLT3-ITD in the same run. This has enhanced patient prognostication and clinical decision making regarding therapeutic approaches. The assays are suitable both for individual patient diagnosis and for large scale clinical trials.

Alex Dobrovic has received honoraria from Roche Diagnostics for speaking about HRM.

AYCT wrote the paper and performed the experiments, AD developed the assay with AYCT, co-wrote the paper and revised the paper in accordance with the reviewers' comments, DAW, DC, and JFS initiated the project, provided the specimens and assisted with writing, SJ provided specimens and performed the immunohistochemical analysis. All authors read and approved the final manuscript.

After this manuscript was submitted, another report of NPM1 mutations in secondary AML has appeared 18.