The FMS-like tyrosine kinase 3 (FLT3) is a transmembrane receptor tyrosine kinase, which is expressed by hematopoietic stem cells and stimulates the development of myeloid and lymphoid progenitor cells (1). Mutations of the FLT3 gene are identified in ~30% of patients with acute myeloid leukemia (AML). About 20%–25% of AML patients show internal tandem duplications (FLT3-ITD), affecting the juxtamembrane domain (JMD) and/or the tyrosine kinase domain-1 (TKD1) of FLT3 (2). In addition, mutations in exon 20, coding for the TKD2 region (most frequently affecting codons D835 and I836), can be detected in 7%–10% of AML patients (FLT3-TKD) (1, 2).

Both FLT3-ITD and FLT3-TKD mutations constitutively activate the FLT3 kinase, inducing proliferation of leukemic populations, although differences in the signaling induced have been reported between these two mutations (1). While the clinical significance of FLT3-TKD mutations is uncertain, the presence of FLT3-ITD mutations in AML patients confers a poor prognosis with an increased risk of relapse and shorter overall survival (2, 3). Consequently, the FLT3-ITD mutational status is included in the current risk classification of the ELN-2017 recommendations and is recognized as target for specific TK inhibitors (2, 4, 5). However, several factors modify the prognostic impact of FLT3-ITD mutations, such as the allelic ratio and the presence of a concomitant NPM1 mutation (4).

In addition to typical ITD mutations, previous reports indicated the presence of rare non-ITD mutations (small deletions and point mutations) in the FLT3 JMD of AML patients (6–16). However, due to the low frequency of these alterations, there are so far only casuistic reports available, with limited information on molecular and clinical associations. To investigate the prevalence and prognostic impact of non-ITD mutations in the FLT3 JMD of adult patients with AML, we analyzed a large cohort of 1,539 patients with newly diagnosed AML to correlate these mutations with clinical characteristics, co-mutations, and outcome.

In the present study, we analyzed 1,539 adult AML patients for the prevalence and prognostic impact of non-ITD mutations in the JMD/TKD1 region of FLT3. We confirm that JMD point mutations and deletions are rare, but recurrent alterations in patients with AML at a frequency in the range of previous estimates (8, 9). Compared to AML, higher rates (~4-fold) of JMD non-ITD mutations were previously detected in patients with acute lymphoblastic leukemia (ALL) (8). Among non-ITD mutations in our cohort, the point mutations Y572C, V592G (9), L576Q (10), G583S (8), Y591H (11), and V592A (12, 13) as well as the deletion EY598_599del (14) were previously recognized as gain-of-function mutations that result in constitutive kinase activation and stimulate AML growth through aberrant STAT5 signaling (9, 12, 14). In addition, in one patient we identified a novel and likely damaging (PolyPhen score = 1) point mutation at Y599N (30% VAF), not previously reported in the literature. However, functional data indicate that any alteration of the JMD sequence interferes with the kinase auto-inhibition, similar to the effect of FLT3-ITD mutations (9, 11, 15). Accordingly, in vitro studies illustrated the response of AML cells with non-ITD JMD mutations to pharmacologic FLT3 inhibition (9, 10, 12, 14–16), which supports the idea to use any of the approved TKIs in patients with this type of mutation (5). In line with this, we demonstrate high similarities between patients with FLT3-ITD and non-ITD mutations, with respect to major laboratory parameters (i.e., an increased percentage of blast cells in PB and BM) and the rate of co-mutated driver variants, which is consistent with the mutational landscape and clinical phenotype typically observed for patients with FLT3-ITD mutations (1–3, 21). For example, a high rate of concomitant NPM1 and DNMT3A mutations but mutual exclusivity with downstream effectors such as NRAS in FLT3 mutant patients has been described in detail (21). Similar to our data, partial tandem duplications (PTD) of the histone methyltransferase KMT2A (MLL) occur in 3%–5% of AML cases and are typically enriched in patients with FLT3-ITD mutations (as compared to FLT3-wt) (21, 22). In addition, we show a not previously reported high association of KMT2A-PTD mutations with FLT3 JMD point mutations, which adds on initial observations on the genomic landscape of KMT2A-PTD-mutated AML (22). More importantly, although the presence of FLT3 non-ITD mutations alone was not an independent prognostic factor, patients with dual non-ITD and KMT2A-PTD mutations showed a trend for inferior outcome. While KMT2A-PTD mutations are typically associated with poor prognosis, isolated KMT2A-PTD mutations are not sufficient to establish leukemic transformation of hematopoietic cells, highlighting the importance of cooperating genetic lesions, such as FLT3 mutations in KMT2A-rearranged leukemias (22–24). Likewise, functional data indicate the transcriptional regulation of FLT3 by KMT2A via aberrant MEIS1 gene expression (25). More recently, the combined inhibition of menin-MLL (MLL1, KMT2A) and FLT3 demonstrated a synergistic therapeutic opportunity in these leukemia subtypes (26). However, mutations of genes associated with RAS signaling (i.e., FLT3) may be frequently lost or emerge during AML progression, with relevance for post-remission strategies (27). In this regard, future investigations are needed to evaluate the impact of FLT3 non-ITD mutations for relapse development. In conclusion, we show that FLT3 non-ITD mutations are rare but recurrent alterations in AML and associated with similar clinical features like FLT3-ITD variants. Interestingly, our data point at a functional interaction of FLT3 non-ITD mutations with KMT2A-PTD and 11q-abnormalities as cooperating genetic events, which might be indicative of a distinct pathway of genesis in this subset of AML.

The datasets presented in this article are not readily available because patients did not consent to have data uploaded to a public database. Requests to access the datasets should be directed to the corresponding author on reasonable request: christian.thiede@uniklinikum-dresden.de.

The studies involving human participants were reviewed and approved by Technical University Dresden (EK98032010). The patients/participants provided their written informed consent to participate in this study.

Conception of the work: CT, SS. Sample/Data Collection: All authors. Aquisition/Analysis of Data: CT, SS. Bioinformatic Analysis: SS. Statistical Analysis: MK, SZ. Interpretation of Data: CT, SS. Drafted the manuscript: SS. Administrative support: GE, MB. All authors have read and approved the manuscript being submitted.

CT is CEO and co-owner of AgenDix GmbH, a company performing molecular diagnostics.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.