Who's dancing with FLT3?

Mutations of the receptor tyrosine kinase FLT3 are present in approximately 30% of patients with newly diagnosed AML.¹  These mutations come in 2 forms, the first being the internal tandem duplication (FLT3/ITD) mutations, localized to the juxtamembrane region of the receptor, and the second being the point mutations located in the tyrosine kinase domain (FLT3/TKD). The FLT3/ITD mutations, which are the most common type, have clearly been shown to have a strong negative impact on clinical outcome, while the prognostic influence of an FLT3/TKD mutation is still uncertain.

Several lines of laboratory data now suggest that FLT3 mutations by themselves do not cause AML.²  Rather, these mutations represent a single (albeit relatively common) event in a complex dance of mutations eventually leading to transformation. The interaction between the FLT3 mutation and other genetic lesions presumably influences the nature of the leukemia, which is reflected in its clinical behavior. In support of this concept, 2 articles in this issue of Blood provide evidence that the context in which these mutations occur influences their impact on patient prognosis.

In the first report, by Gale and colleagues, leukemia DNA samples from 1425 patients enrolled on the United Kingdom Medical Research Council (MRC) AML 10 and 12 trials were analyzed for both FLT3 and nucleophosmin 1 (NPM1) mutations, and the results were compared with patient outcomes. Such a comparison has been performed by several other groups around the world, but the virtue of this particular MRC study lies in the large sample size and the resultant Kaplan-Meier curves. The findings confirm what others have suggested—namely, that when FLT3/ITD mutations dance without an NPM1 mutation, the prognosis is dismal, and when NPM1 mutations dance without an FLT3 mutation, the prognosis is better. When the 2 mutations dance a duet, the prognosis falls in the middle.

The impact of FLT3/TKD mutations has been more difficult to discern, primarily because of their relatively low frequency in AML (5%-7% of cases). Bacher and colleagues present the results of a truly Herculean task, that of analyzing 3082 consecutive AML DNA samples from the German cooperative group trials for FLT3/ITD, FLT3/TKD, NPM1, NRAS, MLL, and CEBPA mutations. As with the work of Gale and colleagues, the sheer number of samples analyzed distinguishes this study from prior ones. Here, an FLT3/TKD mutation has an unfavorable effect when paired with an MLL mutation or a 15;17 translocation. Yet, when the FLT3/TKD mutation dances with an NPM1 mutation, the effect is favorable. Interestingly, the group from the United Kingdom previously re-ported a favorable influence of FLT3/TKD mutations in patients with core binding factor translocations.³ 

Thus, when FLT3 is engaged in the molecular danse macabre that leads to AML, the fate of the patient appears to depend on which dance partner it chooses.