DNA damage accumulation and repair defects in FLT3‐ITD acute myeloid leukemia: Implications for clonal evolution and disease progression

Abstract Acute myeloid leukemia is a group of hematological diseases that have a high mortality rate. During the development of this pathology, hematopoietic cells acquire chromosomal rearrangements and multiple genetic mutations, including FLT3‐ITD. FLT3‐ITD is a marker associated with a poor clinical prognosis and involves the activation of pathways such as PI3K/AKT, MAPK/ERK, and JAK/STAT that favor the survival and proliferation of leukemic cells. In addition, FLT3‐ITD leads to overproduction of reactive oxygen species and defective DNA damage repair, both implicated in the appearance of new mutations and leukemic clones. Thus, the purpose of this review is to illustrate the molecular mechanisms through which FLT3‐ITD generates genetic instability and how it facilitates clonal evolution with the generation of more resistant and aggressive cells. Likewise, this article discusses the feasibility of combined therapies with FLT3 inhibitors and inhibitors of DNA repair pathways.

exhibits five extracellular immunoglobulins-like domains, a singlepass transmembrane domain, a juxtamembrane domain (JM), and two intracellular tyrosine kinase domains (TKDs) linked by an insertion kinase domain ( Figure 1A). 9 FLT3 signaling involves the binding of this receptor to its ligand (FL) through contacts between the third immunoglobulin-like domain of FLT3 and the N-terminal segment of FL. 10 This causes a series of rapid conformational changes that lead to the protein acquiring a catalytically active form which initially promotes autophosphorylation and homodimerization events. 11 Subsequently, FLT3 exposes binding sites and phosphorylates signal transduction effector molecules involved in multiple pathways such as PI3K/AKT, MAPK/ERK, JAK/STAT, and PLC, among others ( Figure 1B). 12 It should be noted that, given its importance, FLT3 signaling is highly regulated by different mechanisms including a narrow range of FLT3-expressing tissues and cells, the need for conformational changes for its activation, and the rapid internalization and degradation of the FLT3-FL complex, among others. 9,11,13 Despite all this, there are mutations that make FLT3 constitutively active and contribute to the development of diseases such as AML. Interestingly, although mutations in FLT3 may be driver mutations in AML, it is known that they cannot induce disease development on their own and require other helper mutations during leukemogenesis, including chromosomal alterations such as t(8; 21) (q22; q22); RUNX1-RUNX1T1, inv(16) (p13.1q22); CBFB-MYH11 and t(15; 17) (q22; q12); PML-RARA, as well as point mutations in genes such as NPM1 and DNMT3A with which frequent coexistence has been found in patients with AML with normal karyotypes. 14 The most common form of mutation affecting FLT3 is an internal tandem duplication (ITD) in exons 14 and 15 that affects the juxtamembrane region, while the second common type of mutation is a missense point mutation in exon 20 known as FLT3-TKD because affects the second kinase domain. 15,16 Several studies have shown that FLT3-ITD mutations have a greater clinical impact than FLT3-TKD mutations, being associated with a poor clinical prognosis. 16,17 FLT3-ITD can activate the same pathways as wild-type FLT3, albeit in an aberrant manner. Also, because it is retained longer in the endoplasmic reticulum (ER)-Golgi network, it can interact with other proteins and also activate other pathways. 18 As a product of this, FLT3-ITD acts to initiate a cycle of genomic instability through increased production of reactive oxygen species (ROS) and inefficient DNA damage repair 19,20 which, in turn, creates an environment conducive to the appearance of new mutations and clonal evolution. 21 In this way, the purpose of this review is to discuss how FLT3-ITD favors the generation of DNA damage and regulates repair mechanisms in such a way that the appearance of new clones that are more resistant to treatment and more aggressive is facilitated.
Likewise, this article analyzes the feasibility of therapies that combine inhibitors of FLT3 signaling and inhibitors of DNA repair pathways.

| FLT3-ITD PROMOTES DNA DAMAGE
Every day, DNA integrity and stability are challenged by exogenous physical, chemical, or biological agents, as well as endogenous processes, including DNA replication errors, spontaneous hydrolytic reactions, and reactive oxygen species (ROS). 22 23 ROS are normally present at low and constant levels, playing an important role in cell signaling and homeostasis, but at high concentrations, ROS react rapidly with proteins, lipids, carbohydrates, and nucleic acids, causing irreversible functional alterations or even its complete destruction. 24 FLT3-ITD, by overactivating the STAT5 and PI3K/AKT pathways, maintains and sometimes increases the expression of p22phox and NOX proteins, thus favoring the production of ROS that diffuse into the nucleus and cause DNA damage, especially double-strand breaks and mismatches ( Figure 2). 19,20,25,26 Furthermore, the presence of FLT3-ITD at the plasma membrane is required to maintain NOX protein levels and prevent proteosomal degradation of p22phox triggered by GSK3-β signaling. 23,27 Overall, NOX proteins associate with p22phox to costabilize in the membrane, and then RAC1 binds and exchanges GDP for GTP to initiate the conversion of oxygen to superoxide. 28 In this regard, it has also been reported that RAC1-GTP binding to p-STAT5 is increased in FLT3-ITD cells and, consequently, a greater amount of RAC1-GTP is recruited to NADPH oxidase complexes. 19 In particular, the nuclear membrane-bound NOX4D isoform is overexpressed in patients and cell lines expressing FLT3-ITD but is almost absent in their wild-type FLT3-expressing counterparts. It stands out that this protein generates ROS that favor the survival of leukemia cells. 25 Due to a feedback mechanism, a vicious circle is created between FLT3 signaling and ROS production. In this way, ROS create an oxidative environment that enhances wild-type FLT3 and FLT3-ITD signaling (presumably by oxidation of certain cysteines such as residue 790) which, in turn, leads to increased ROS production. 29 On the other hand, FLT3-ITD expressing cells have high expression of succinate-CoA ligases and high mitochondrial electron transport chain (ETC) complex II activity. Thus, a high respiratory activity is induced with increased production of mitochondrial ROS associated with the Krebs cycle and oxidative phosphorylation. 30 Although there is no related report, there is a possibility that FLT3-ITD, by increasing global transcription, may also cause DNA damage.
In this sense, FLT3-ITD could increase the number of transcriptionreplication conflicts, activate responses to replicative stress and cause an accumulation of R loops, among other things. 31 All this demonstrates that cells that express FLT3-ITD present greater DNA damage than those that do not present the mutation, this being mainly caused by ROS, although other mechanisms may be associated.

| FLT3-ITD REGULATES DNA REPAIR PATHWAYS
Due to the important consequences that DNA damage can cause, cells have developed sophisticated strategies to repair it as quickly and precisely as possible, which are collectively known as DNA repair mechanisms and include several pathways (Table 1) whose functions include detecting DNA damage, signaling its presence, and promoting its repair. 32 Because cancer must initially acquire permanent genomic mutations, it is, by definition, a disease due to inadequate DNA repair. In this manner, many cancers may be constantly acquiring new mutations. 33 For this reason, it is interesting that the Cancer Genome Atlas (TCGA) reported that AML cells have a low average mutation frequency (approximately 1 mutation per megabase), these being mainly C > T transitions and affecting genes such as PIK3CA, IDH1, IDH2, NPM1, FLT3, RUNX1, MIR142, DNMT3A, TET2, GATA3 and MAP3K1 where on average two non-synonymous mutations occur. 34 This would indicate that DNA repair mechanisms function relatively well in AML, although several reports indicate that mutability is increased when FLT3-ITD is expressed. 35 FLT3-ITD has been reported to increase interchromosomal homologous recombination F I G U R E 2 FLT3-ITD promotes the formation of ROS. FLT3-ITD, by activating PI3K/AKT and STAT5 signaling, promotes the transcription of NOX proteins, one of the main producers of ROS. In particular, NOX4D bound to the nuclear membrane produces ROS that help leukemia cells survive and damage DNA, mainly by producing double-strand breaks events and these were also positively correlated with intracellular ROS levels. 36 Similarly, patients with FLT3-ITD positive AML have been reported to have a high frequency of rare structural chromosomal abnormalities at relapse. 37 Figure 3A). Notably, the resulting alterations are mainly deletions in regions of sequence microhomology. 39 It should be noted that, like the classical NHEJ (c-NHEJ) pathway, alt-NHEJ can act at any phase of the cell cycle and in areas with high chromatin condensation. 42 In contrast, HR repair occurs more efficiently in FLT3-ITD cells than in wild-type FLT3 cells ( Figure 3B).
Remarkably, the HR pathway can only be used during mitosis when chromosomes can line up and in areas without high chromatin condensation. 42

| FLT3-ITD FACILITATES CLONAL EVOLUTION
Cancers evolve through an iterative process of clonal expansion, genetic diversification, and clonal selection that results in a predominance of only those clones that are best adapted to the conditions of the surrounding microenvironment (constantly changing). 50 There are two main models of clonal evolution, linear and branching ( Figure 4A). The linear model involves the gradual acquisition of individual mutations, with the cell at the final step carrying all the mutations that arose during evolutionary history and outcompeting earlier clones that carry only some of the mutations. 51 For its part, the branching model implies the eradication of the dominant clone, followed by the growth of a subclone. 52 With little evolutionary pressure, it is likely to follow a linear evolution, but if the pressure changes profoundly (e.g., initiation of chemotherapy, stromal modification, major changes in growth factors, cytokines, or hormones), it is more likely that follow a branching evolution. 53 Explaining this in more detail and in the context of FLT3-ITD, the genetic instability that is generated by constant DNA damage and defects in its repair allows the appearance of a diverse collection of cells that harbor different molecular signatures (linear evolution) which then be selected in the surrounding microenvironment (branching evolution) ( Figure 4B). 52,54 Thus, a clonal evolution can be clearly observed in patients studied at the time of diagnosis, during treatment and when they suffer relapses. 35,55 For example, one study mentioned that in patients without FLT3-ITD more than half of the mutations were maintained from diagnosis to relapse, but in patients with FLT3-ITD this did not occur and some mutations disappeared while others appeared. 55 This would suggest that FLT3-ITD promotes branching evolution. On the other hand, in FLT3-ITD patients treated with midostaurin, a FLT3 inhibitor, the number of mutations that appeared at relapse was lower than in patients who did not receive the drug. 35,55 Furthermore, FLT3-ITD persists in a considerable proportion of patients who relapse, which reinforces the conception of this as a driver mutation, although the number of ITD clones in the same patient decreases. 55 This would suggest that linear evolution also occurs in FLT3-ITD patients and thus indicate that, whether linear or branched, cell populations are constantly undergoing clonal evolution to better adapt to their environment.
Increased homologous interchromosomal recombination enables loss of heterozygosity (LOH) during relapse by deleting the allele encoding wild-type FLT3, and this is associated with a more aggressive phenotype. 36 Remarkably, the selection of these clones is accelerated by the administration of chemotherapeutic drugs, providing a powerful source of artificial selection, and resulting in a selective pressure that allows only variant cells that resist treatment to proliferate and cause massive death of cells that do not do it. 50,56,57 In this sense, the appearance of clones that present additional mutations in inhibitor midostaurin has been approved together with conventional induction therapy, as well as gilteritinib, which is a more potent and specific FLT3 inhibitor, as monotherapy for relapsed/refractory FLT3-ITD positive AML. 67  One of the areas to investigate and that would provide important information, which is why it is discussed in this review, is clonal evolution. Clonal evolution is complicated and complex to study, which is reflected in the few studies that currently exist in this area.
In this context, although the different investigations can be oriented in many directions, we will highlight only three here: The first category would be to analyze how clones evolve and which mechanism of evolution is preferred (linear or branched), for which it would be necessary to identify the different molecular signatures and how they behave at different points of the disease, such as diagnosis, at various points in treatment and during relapse. Also, it would be interesting to know the role of other types of mutations in FLT3 with respect to DNA damage and its repair and to compare them with those of FLT3-ITD. To do this, tools such as single-cell sequencing would be of great help, along with artificial intelligence and machine learning systems that can extrapolate these results to other patients and make predictions to provide more personalized treatment.
The second category is to elucidate the mechanisms by which clones are selected, which could mainly consider the context of the cellular microenvironment in the bone marrow, as well as the effect of chemotherapeutic drugs and/or radiation. To this end, threedimensional cell culture systems in bone marrow tissue, gene editing (CRISPR-Cas9, TALENs, ZFNs) and high-throughput animal models could provide important insights. Previous RNA-seq, ChIP-seq, microarray, epigenomic, proteomic and metabolomic data in this context could also be collected, homogenized and analyzed together.
The third category involves, based on clonal evolution, selecting the appropriate combinations of drugs and/or therapies. Although AML therapy protocols are well established (7 + 3 therapy), it would be good to know if they can be customized (e.g., by adding combined therapies with FLT3 inhibitors and inhibitors of some DNA repair pathways). Identifying whether it is possible to control clonal evolution with drugs would also be an interesting option.
Clearly, all of these categories take time and a lot of work and effort, but as we learn more about the evolution of AML disease, more patients will have better survival and disease-free survival.

AUTHOR CONTRIBUTIONS
Francisco Alejandro Lagunas-Rangel conceptualized the idea, looked up the references and wrote the article.

ACKNOWLEDGMENTS
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

CONFLICT OF INTEREST
The author declares no conflict of interest.

DATA AVAILABILITY STATEMENT
Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

ETHICS STATEMENT
This is a review manuscript based on previously published articles so ethical approval is not required.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/hon.3076.