The presence of internal tandem duplications (ITD) in the FLT3 receptor tyrosine kinase gene have long been known to confer a poor prognosis in patients with acute myeloid leukemia. Now, specific structural features of the ITDs are also suggested to alter patient outcome, including sensitivity to targeted therapies, prompting their evaluation in therapeutic algorithms.

See related article by Schwartz et al., p. 573

In this issue of Clinical Cancer Research, Schwartz and colleagues show that structural features of FLT3-internal tandem duplications (ITD) mutations, a highly recurrent mutation in acute myeloid leukemia (AML), influence patient responses and overall outcome following both standard cytotoxic chemotherapy and targeted FLT3 inhibitor therapy (1).

AML is a genetically heterogeneous disease and several subtypes, characterized by different biological and prognostic features, have been described on the basis of their mutational patterns and the presence of gross chromosomal alterations (2). Mutations affecting genes involved in signaling pathways are among the most frequently found in AML. In particular, mutations in the type III receptor tyrosine kinase FLT3 are present in about 30% of patients with AML. FLT3 mutations are mostly secondary to an ITD (FLT3ITD) of the juxta membrane domain which abrogates its negative regulatory function, resulting in constitutive activation of the receptor's tyrosine kinase activity, and predict for an increased relapse rate following standard therapies and a poor prognosis (3).

FLT3ITD mutations activate survival and proliferative signaling pathways, including phosphatidyl-inositol3-kinase (PI3K)/AKT, MAPK/ERK, and STAT5 thus explaining the hyperproliferative and aggressive phenotype of AML carrying these mutations. Quantification of variant allele frequency in patient DNA samples, genetic assessment at diagnosis, and relapse and studies in animal models have respectively shown that FLT3ITD occur later in leukemia evolution and are not sufficient to produce an AML phenotype without collaborating mutations. Nevertheless, preclinical studies demonstrate that AML cells are oncogenically addicted to FLT3ITD thus suggesting its role as a therapeutic target (4). This has been also confirmed by the results of several recent phase II and III clinical trials that have demonstrated a survival benefit for patients treated with FLT3 tyrosine kinase inhibitors (TKI; ref. 3). However, despite our improved understanding of the role played by FLT3ITD mutations in AML and the rational design of targeted inhibitors of their tyrosine kinase activity, the overall outcome of patients with AML carrying FLT3ITD mutations remains poor and somewhat variable, with average overall response rates of less than 50% even following treatment with targeted FLT3 TKI. Several mechanisms have been shown to alter responses to FLT3 TKI including mutations in the FLT3 tyrosine kinase domain, cellular adaptive responses, and disease features inherent to individual patients, such as FLT3ITD mutational burden and the presence of cooccuring mutations (2–4).

The article from Schwartz and colleagues provides a novel tool to further classify patients' response to both standard cytotoxic and novel FLT3 TKI therapies based on the ITD structural features of FLT3ITD mutations. The authors developed a novel, publicly available, bioinformatic algorithm (HeatITup) which is able to detect the presence of FLT3ITD mutations from DNA next-generation sequencing data from patients with AML at diagnosis with better accuracy than other commonly used algorithms, validating their findings using both sequencing data from their own independent patient cohort and The Cancer Genome Atlas (TCGA) AML dataset. Moreover, they show that HeatITup is capable of defining in fine details of the structural features of ITD mutations in each patient. These include not only length of the duplicated section but also nucleotide composition including spacer regions and exogenous sequence within the insertion, as well as potentially relevant point mutations within the ITDs. Using HeatITup, the authors then divide FLT3ITD mutations into “typical” or “atypical” subtypes, based respectively on the absence or the presence of nucleotides exogenous to the wild-type FLT3 locus within the mutated region. Interestingly, they demonstrate that typical and atypical mutations tend to cluster within each patient that is patient harboring multiple ITD clones tend to have mostly either typical or atypical clones. This suggests that these two types of mutations might represent different biological entities and/or represent the consequences of different mutational exposures or processes. Partially supporting this notion, they then show that patient carrying atypical FLT3ITD mutations have lower survival rates when treated with both standard cytotoxic chemotherapy and FLT3 TKI therapy, within both their patient cohort and validated in the TCGA cohort.

Although FLT3ITD allele burden, reflective of the size of the FLT3-mutated clone and or uniparental disomy for chromosome 13q, has been implicated in the prognosis of patients treated with standard chemotherapy, the role of duplication length is less well understood as there have been conflicting reports regarding a link between overall survival and ITD length (5). The work by Schwartz and colleagues adds to our knowledge on the role of structural features of FLT3ITD mutations by showing that, beside length, other structural features of the ITDs may have an impact on patient outcome and that this might also hold true in the era of novel targeted therapy with FLT3 TKI. These findings will need to be validated in other independent datasets and the software the authors have developed will provide an accessible computational algorithm to enable these studies. If confirmed in larger studies, these results will also support the concept that characterization of FLT3ITD complexity, beside their detection, should be used to inform clinical decision-making and in the therapeutic algorithm for patients with AML carrying FLT3ITD mutations.

Some of the observations from the author's work, if further confirmed, also raise interesting biological questions that, if confirmed, should be specifically addressed in experimental models of FLT3ITD leukemia. In particular, it would be interesting to understand how the two classes of ITD mutations confer a different response/sensitivity to treatment. Important questions could include; does the type of ITD alter the native activity of mutated FLT3 and are typical and atypical ITDs different in the intensity of their signaling and FLT3 activity? Alternatively, as the two classes predict response to both standard cytotoxic and FLT3 TKI therapy, is the differential sensitivity to therapy independent of FLT3 activity and reflect other processes such as increased/altered mutagenesis? Moreover, given that typical and atypical mutations tend to cluster within each patient, do these two types of mutations represent the consequences of different mutational exposures or processes? Interestingly, the authors' data show that the pattern and number of cooccuring mutations is not different between the two classes of ITDs within their cohort, which seems to suggest similar rates of mutation. Nevertheless, further validation in bigger cohorts and a more in depth analysis of their biology might help to clarify whether patients carrying each subtype of mutations show differential response to DNA damage and highlight the mechanisms behind their almost complete mutual exclusivity.

In conclusion, the study by Schwartz and colleagues represents another incremental step towards implementing precision medicine for patients with AML carrying FLT3ITD mutations. Their HeatITup algorithm will be a valuable, easy-to-use, and publicly available tool to stratify patients according to the structural features of their ITD mutations (Fig. 1). It is also conceivable that this algorithm could be used to generically detect and classify ITD mutations reported in various genes for other subtypes of leukemia and solid cancers. Future studies will help to further validate these findings, clarify the features modulating the response to FLT3 TKI and standard chemotherapy in patients with AML, and understand their underlying biological mechanisms.

Figure 1.

Sequencing data from AML patients' DNA can be analyzed using the HeatITup algorithm to detect FLT3ITD mutations and classify them, based on the ITD structural features, into typical or atypical ITD variants. This classification can then be used to inform patient prognosis and design a more effective therapeutic algorithm for patients with AML carrying FLT3ITD mutations and may suggest intrinsic differences in biology between typical and atypical patients.

Figure 1.

Sequencing data from AML patients' DNA can be analyzed using the HeatITup algorithm to detect FLT3ITD mutations and classify them, based on the ITD structural features, into typical or atypical ITD variants. This classification can then be used to inform patient prognosis and design a more effective therapeutic algorithm for patients with AML carrying FLT3ITD mutations and may suggest intrinsic differences in biology between typical and atypical patients.

Close modal

P. Gallipoli reports receiving speakers bureau honoraria from and is a consultant/advisory board member for Bristol-Myers Squibb. B.J.P. Huntly reports receiving speakers bureau honoraria from Novartis, Pfizer, and Bristol-Myers Squibb. No other potential conflicts of interest were disclosed.

Conception and design: P. Gallipoli, B.J.P. Huntly

Writing, review, and/or revision of the manuscript: P. Gallipoli, B.J.P. Huntly

Study supervision: B.J.P. Huntly

P. Gallipoli is supported by the Wellcome Trust (109967/Z/15/Z) and an ASH Global Research Award. B.J.P. Huntly is supported by the European Research Council (COMAL: 647685) and Cancer Research UK (A25508).

1.
Schwartz
GW
,
Manning
B
,
Zhou
Y
,
Velu
P
,
Bigdeli
A
,
Astles
R
, et al
Classes of ITD predict outcomes in AML patients treated with FLT3 inhibitors
.
Clin Cancer Res
2019
;
25
:
573
83
.
2.
Papaemmanuil
E
,
Dohner
H
,
Campbell
PJ
. 
Genomic classification in acute myeloid leukemia
.
N Engl J Med
2016
;
375
:
900
1
.
3.
Leick
MB
,
Levis
MJ
. 
The future of targeting FLT3 activation in AML
.
Curr Hematol Malig Rep
2017
;
12
:
153
67
.
4.
Smith
CC
,
Wang
Q
,
Chin
CS
,
Salerno
S
,
Damon
LE
,
Levis
MJ
, et al
Validation of ITD mutations in FLT3 as a therapeutic target in human acute myeloid leukaemia
.
Nature
2012
;
485
:
260
3
.
5.
Gale
RE
,
Green
C
,
Allen
C
,
Mead
AJ
,
Burnett
AK
,
Hills
RK
, et al
The impact of FLT3 internal tandem duplication mutant level, number, size, and interaction with NPM1 mutations in a large cohort of young adult patients with acute myeloid leukemia
.
Blood
2008
;
111
:
2776
84
.