Comparison of Different Techniques for the Detection of FLT3-Mutations in Exons 14/15 and 20 in Patients with Acute Myeloid Leukemia.

Mutations of the FLT3 receptor tyrosine kinase are among the most important molecular abnormalities in patients with acute myeloid leukaemia (AML). Two types of mutations are predominant, internal tandem duplication mutations located in exons 14/15 and point mutations in exon 20, coding for the activation loop (codons 835/836) of the second tyrosine kinase domain (TKD). Numerous assays have been described for the detection of these abnormalities, including standard agarose gel electrophoresis, high resolution fragment analysis (Genescan), restriction fragment length polymorphism (RFLP) analysis and denaturing high performance liquid chromatography (DHPLC). Limited data exist on the capacity of these methods to detect FLT3 mutations. We compared agarose gel electrophoresis (AGE), Genescan analysis (GS) and DHPLC for the detection of FLT3 mutations in exon 14/15 and RFLP and DHPLC for the detection of mutations in exon 20 in a total of 1071 primary AML samples from patients treated in the AML96 protocol of the Deutsche Studieninitiative Leukämie (DSIL). The sensitivity of the ITD-detection was compared using dilutions of the FLT3-ITD pos. cell line MV4-11 in normal cells. Clinical samples which were found positive only in DHPLC were cloned and sequenced.

Results: The dilution experiments indicated the highest sensitivity for the GS-method (0.1%) whereas the other techniques were less sensitive (AGE: 6.4%; DHPLC: 3.2%). Altogether 210/1071 (19.6%) cases scored positive in all three techniques for mutations in exon 14/15, 35 (3.3%) samples where positive only in the GS, 19 (1.8%) in GS and DHPLC and 14 (1.3%) in DHPLC only. Sequence analysis confirmed the presence of point mutations/small deletions within the juxtamembrane domain in 12/14 of those samples positive in DHPLC only. Mutational screening for the TKD domain revealed 64 positive cases with both techniques (6%), 14 (1.3%) cases were only positive in the DHPLC, 1 (0.1%) case was found by RFLP only. Interestingly, the majority of those samples positive with the DHPLC showed mutations outside the hot-spot region (codons 835/836), namely mutations in codons 837, 839, 841 and 842.

Conclusions: Taken together these data indicate that more advanced techniques (GS; DHPLC) might be able to detect more mutations in the FLT3 gene than previously reported. Although AGE appears to detect most clinically relevant FLT3-ITD mutations, GS analysis might be more appropriate to screen for this lesion. DHPLC screening for ITD mutations has a lower sensitivity but is able to detect point mutations, the clinical relevance of which is currently unknown. In contrast DHPLC might be an alternative for the assessment of mutations within the TKD domain, in this cohort, about 20% more TKD mutations were found using this method. This higher sensitivity might be especially relevant in the context of novel targeted therapies. In addition, the data presented here are the first comprehensive report on the prevalence of unusual FLT3-mutations in a large cohort of patients.