Acute myeloid leukemia (AML) is the most common hematopoietic tumor and mainly affects older adults (1). In 2020, worldwide statistics indicated that AML was the thirteenth most common type of cancer in terms of incidence and had the tenth highest mortality rate (2). Men are more susceptible to AML than women (6.4 vs. 4.3%) (2), and the overall survival rate is ~50% in young patients (18–50 years) but only 10% in adults >65 years old (3).

AML is characterized by accelerated and autonomous growth of immature myeloid cells, leading to the dysregulation of hematological components (4–6). Some mutations are relevant for cell survival and proliferation, such as tyrosine-protein kinase (c-Kit) and FMS-like tyrosine kinase 3 (Flt3) mutations, which are directly related to AML progression (7,8). Notably, 25–35% of patients with AML have mutations that result in the upregulation of expression of the FLT3 receptor, which is related to high malignancy and poor prognosis (9).

FLT3 is part of the class III receptor tyrosine kinase (RTK) family. It is typically expressed in early myeloid and lymphoid CD34⁺ precursors, and its expression gradually decreases during cell differentiation (10). The FLT3 receptor consists of five domains: An immunoglobulin-like extracellular domain, which is the site of ligand binding and FLT3 activation (which is activated in synergy with IL-3 and granulocyte-macrophage colony-stimulating factor), a transmembrane domain, a juxtamembrane domain (JMD) and two cytoplasmic tyrosine kinase-like domains (TKDs; TKD1 and TKD2), which constitute the catalytic domain of FLT3 and are critical for receptor activation (10–12). Upon ligand binding, FLT3 activity is mediated by the JMD through phosphorylation of Tyr589 and Tyr591 amino acid residues. This phosphorylation promotes conformational changes that free the activation loop (A-loop) present in TKD1 and the phenylalanine residue of the aspartic acid-phenylalanine-glycine motif (DFG motif) (13).

The DFG motif, which is part of the ATP-binding site, is highly conserved and essential for the function of RTKs such as FLT3 (14). Depending on the type of ligand, the DFG motif can change between its inactive DFG motif (DFG-out) or active DFG motif (DFG-in) form (13,15). The DFG-out conformation occurs when Asp flips the ATP binding site outward and Phe moves out of the hydrophobic site (14,16,17). Conversely, the DFG-in conformation occurs when Phe is packed in a hydrophobic pocket and Asp moves toward the ATP binding site to interact with Mg²⁺ (18). FLT3/DFG-in activation stimulates proliferation and survival signaling pathways via Janus kinase (JAK)/STAT5, AKT/ERK and PI3K (10–13,19).

In patients with AML, common activating mutations in FLT3 include the Asp835Tyr (D835Y) point mutation in TKD1 and the internal tandem duplication (ITD) in JMD or TKD1 (3,20). ITD mutations in JMD and TKD1 affect 5–16% of pediatric patients (<10 years old) and 25–35% of adult patients (3,20). Patients with ITD-FLT3 mutations exhibit poor prognosis associated with the upregulation of this receptor, which leads to high proliferation and survival of leukemic cells, resistance to treatment, increased risk of relapse and decreased survival (3,20). Therefore, the ITD-FLT3 receptor requires special consideration as a drug target (21).

The ITD insertion (3–400 base pairs) occurs between exons 14 and 15 of the Flt3 gene, corresponding to the JMD or TKD1 domains of FLT3. At the protein level, the new amino acids are commonly inserted within a sequence of one to seven amino acids [F(Y)DFREYE/YDLK], corresponding to codons 591–597 of the original open reading frame (3,8,22). The length of ITD insertion changes the size of FLT3 and may influence its activity and the effects of its inhibitors (3,8,22).

Drugs that inhibit tyrosine kinase activity are widely used as AML treatments, alone or in combination with chemotherapy (23,24). FLT3 inhibitors have improved the overall survival of patients with leukemia. Due to their efficacy and being associated with fewer adverse effects compared with chemotherapy, they represent a major improvement in AML treatment. However, first-generation tyrosine kinase inhibitors (TKIs; including midostaurin and sorafenib) lack specificity for FLT3 as they bind to several RTKs (C-KIT, platelet-derived growth factor receptor, RAS, RAF, JAK2 and VEGFR), contributing to strong side effects, such as diarrhea, fatigue, nausea, cardiovascular problems and myelosuppression (25). Additionally, some patients with AML may develop resistance to TKIs (24,26,27). Conversely, second-generation TKIs, including quizartinib and gilteritinib, have a more specific effect, both for wild-type (WT) and ITD-mutated forms of the FLT3 receptor. Therefore, they can be used in lower concentrations than first-generation TKIs, resulting in lower toxicity and fewer side effects (11,23,24). TKIs can also be classified based on their selective binding to the DFG-out or DFG-in conformations of the FLT3 receptor. Type I inhibitors (midostaurin and gilteritinib) bind preferably to the ATP pocket of the DFG-in conformation (active motif), while type II inhibitors (sorafenib and quizartinib) bind preferably to the hydrophobic region of FLT3 adjacent to the ATP pocket of the DFG-out conformation (inactive motif) (23).

Over the past 15 years, the efficiency of several TKIs has been evaluated in patients with AML. Midostaurin, gilteritinib and quizartinib have been approved by the Food and Drug Administration (FDA) for AML treatment (24). However, little is known about the effects of ITD mutations on the molecular interactions and affinity forces of these inhibitors. Therefore, it is crucial to carefully investigate these parameters and their impact on the inhibitory activity of TKIs in AML cells.

Bioinformatics provides valuable tools to assess molecular interactions and predict the theoretical potential of specific inhibitors for a particular target. Examples of these tools include homology modeling of protein structure (HMP) and molecular docking. HMP helps approximate the atomic structure of a specific protein based on the known structure of a related protein available from databases such as the Protein Data Base (PDB) (28–30). HMP is achieved using two phylogenetically close protein structures with a high percentage of similarity, one serving as a template and the other as a target (30). Once a good quality model of the protein of interest is obtained, molecular docking facilitates the prediction of the molecular interactions between a ligand and a protein, calculating their affinity force through polar (H-bonds) and non-polar bonds (31).

Several AML cell lines expressing WT or ITD-mutated FLT3 receptors have been established for experimental evaluation of inhibitors. These are useful for obtaining an in vitro model to study the disease and validate the predictions of in silico approaches. Previous studies have used WT-FLT3 homozygous cell lines (HL60, THP1, OCI-AML3 and MUTZ-2 cell lines), while other studies have used WT/ITD-FLT3 heterozygous cell lines (MOLM-13, MOLM-14 and PL21 cell lines) or the MV4-11 and MULT-11 ITD-FLT3 cell lines with loss of heterozygosity (LOH) of FLT3 (15,32–35).

The present study was designed to assess the molecular interactions and affinity forces of four TKIs (sorafenib, midostaurin, gilteritinib and quizartinib) with the WT-FLT3 and ITD-FLT3 receptors in their DFG-out and DFG-in conformations in silico. Furthermore, the present study evaluated the effect of the second-generation inhibitors gilteritinib and quizartinib on MV4-11 (ITD-FLT3) and HL60 (WT-FLT3) AML cell lines, and peripheral blood mononuclear cells (PBMCs; FLT3-negative) from a healthy volunteer in vitro.

The upregulation of FLT3 expression leads to the excessive proliferation and death evasion of leukemic cells, as well as the rapid progression of the disease. FLT3 participates in cell proliferation, motility and differentiation. Its activation is induced by autophosphorylation of Tyr589, Tyr591 and Tyr599, which promotes conformational changes in the conserved DFG motif of this receptor and makes the ATP-binding site available (75,76). Additionally, ITD insertions in the JMD of FLT3 are present in 69.5% of patients with AML and are associated with higher cell proliferation and malignancy (13,77,78). Therefore, WT-FLT3 and ITD-FLT3 are essential targets for AML therapy.

The conformation of the FLT3 receptor is dynamic in biological systems, alternating between DFG-out (inactive) and DFG-in (active) forms depending on cellular processes and signaling (79). Thus, it was essential to obtain the structural models of the DFG-out and DFG-in forms of WT-FLT3 and ITD-FLT3 receptors. This contributes to understanding the impact of the ITD in JMD on the conformational changes of FLT3, and the molecular interactions, affinity and effectiveness of TKIs.

The first homology model of WT-FLT3/DFG-in was constructed by Ke et al (13), who used the FLT3/DFG-out (1RJB) and the colony stimulating factor 1 (3LCD) (DFG-in) structures. This model helped define the molecular interaction parameters of type I TKIs and allowed the identification of novel inhibitors through structure-based virtual screening (13). Furthermore, Mashkani et al (15) described two potential models of WT-FLT3/DFG-in based on the structures of c-Kit (1PKG) and Abelson kinases (2GQG), and their molecular interactions with adenosine diphosphate and a type I TKI (dasatinib). Lee et al (80) also constructed a model of WT-FLT3/DFG-in using dual specificity tyrosine phosphorylation regulated kinase 1A (4NCT) to study the effect of the G697R point mutation on the molecular interactions and drug responses to sorafenib and midostaurin. The WT-FLT3/DFG-out structure obtained in the present study complements these previous models and the ones reported by Griffith et al (38) (1RJB) and Thomas (6IL3) (39).

Based on this structural model, WT-FLT3/DFG-in was then constructed using the DFG-in portion of the crystalized c-Kit structure (1PGK) as described in previous works employing c-Kit (42) and other RTKs as templates (13,15,42). Despite the relevance of the ITD-FLT3 mutation in AML progression, few structural models have been reported so far. Most of them focus on elucidating the effect of ITD length on conformational changes of the receptor (37,81). Therefore, to the best of our knowledge, the present study was the first to report the ITD-FLT3 structural models in DFG-out and DFG-in conformations.

To evaluate the quality scores of the four models, seven programs (ERRAT, Profile 3D Verify, QMEAN, Molprobity, ProSA, ProQ and PROCHECK) were used, providing a set of parameters indicating their high structural quality. This strategy demonstrated that they could be used to assess the molecular interactions and affinity scores of sorafenib, midostaurin, gilteritinib and quizartinib.

The molecular interactions of TKIs with the FLT3 receptor are determined by the biochemical characteristics of each inhibitor and the conformation of the receptor to which they bind (76,82). First-generation TKIs were initially designed to target RTKs, but they are not specific for FLT3. For example, sorafenib and midostaurin can efficiently treat solid tumors, such as hepatocarcinoma, advanced renal carcinoma and non-small cell lung cancer, through the inhibition of Raf serine/threonine kinase (83) and protein kinase C (84), respectively. However, the lack of specificity of these inhibitors leads to strong cytotoxic effects at therapeutic doses, limiting their therapeutic use (77,85). Conversely, gilteritinib and quizartinib (second-generation inhibitors) were explicitly designed against FLT3 to treat patients with AML bearing ITD mutations or the D835 mutation in the TKD (86–88). The molecular docking presented in the present study confirmed that gilteritinib and quizartinib (second-generation TKIs) have higher affinities for the WT-FLT3/DFG-out and WT-FLT3/DFG-in models compared with sorafenib and midostaurin (first-generation TKIs).

The molecular docking of WT-FLT3 and the use of PyMOL2 and Discovery Studio visualizer programs indicated that Leu616 and Lys614 [from the phosphate-binding loop (P-loop)], Tyr693, Cys695, and Tyr696 [from the amino terminal lobe (N-lobe)] were the main residues participating in the interactions of WT-FLT3/DFG-out with its inhibitors, whereas Asp698 and Cys694 (from the N-lobe) were involved in the interaction with WT-FLT3/DFG-in. These data partially coincide (Leu616, Tyr693, Cys694, Cys695 and Tyr696) with those of Pandurang et al (89), who also reported the participation of Gly617, Val624, Ala642, Phe830 and Gly697 during the formation of inhibitor/WT-FLT3 complexes. The present study also indicated the participation of other residues, including Arg764, Asp788, Asn790 (from the DFG-motif), Arg704 (from the N-lobe) for WT-FLT3/DFG-out, and Asn701 (from the N-lobe), Asn836 (from the A-loop), Lys802 and Val801 (from the catalytic loop) for WT-FLT3/DFG-in. These sites are associated with the backbone and hinge region (JMD-TKD) of FLT3 and are crucial for the inhibitory activity of the drugs (13,85). Therefore, the ITD insertion in JMD, known to affect the conformation of this region of FLT3, may modify the molecular interactions with its inhibitors and modulate their activity.

Regarding the ITD-FLT3 models, it should be noted that the amino acid position (number) increased downstream of the ITD insertion site (9 aa long). The dockings showed that Asn710 (WT, Asn701), Asp707 (WT, Asp698), Arg773 (WT, Arg764), Cys703 (WT, Cys694), Asn799 (WT, Asn790), Asp797 (WT, Asp788), Val810 (WT, Val801) and Lys811 (WT, Lys802) were involved for both ITD-FLT3 and WT-FLT3 receptors (common amino acids). Most amino acids participating in the formation of the inhibitor/ITD-FLT3 complexes belonged to the hinge region, P-loop, A-loop and N-lobe of the receptor, as observed for WT-FLT3. The ITD insertion in FLT3 also led to the participation of new amino acids from the interacting region of the receptor. In particular, the sorafenib/ITD-FLT3/DFG-out complex involved Arg792 (WT, Arg783), the gilteritinib/ITD-FLT3/DFG-out complex involved Ser796 (WT, Ser787), the quizartinib/ITD-FLT3/DFG-out complex involved Cys623 (WT, Cys614), and the midostaurin/ITD-FLT3/DFG-in complex involved Arg773, Asn774 and Asp787 (WT, Arg764, Asn765 and Asp778, respectively). The different residues forming these complexes may change the affinity of the inhibitors.

The ITD mutation in FLT3 modified the affinity forces of the TKIs included in the present study. This observation supports previous reports by Zorn et al (36) and Friedman (76), who showed that the ITD mutation increased receptor flexibility and facilitated the interaction between ATP or TKIs (drugs) with their binding site, leading to constitutive activation (through ATP binding) or modulation of the activity of the receptors (through TKI binding). Sorafenib and midostaurin exhibited increased affinity towards ITD-FLT3/DFG-out compared with WT-FLT3/DFG-out. Conversely, sorafenib, midostaurin and quizartinib exhibited no marked changes in their affinities towards ITD-FLT3/DFG-in and WT-FLT3/DFG-in. Furthermore, gilteritinib exhibited decreased affinity towards ITD-FLT3 compared with WT-FLT3 (both DFG-out and DFG-in). Therefore, this inhibitor was more sensitive to the effect of the ITD mutation than the rest. Despite this in silico prediction, Perl et al (90) and Smith et al (91) reported that gilteritinib was associated with a good survival outcome in patients with refractory AML, 70% of which presented the ITD mutation.

Regarding the DFG motif, changes in its conformation may modulate the affinity of its inhibitors and their efficacy in AML treatment. Type I TKIs (midostaurin and gilteritinib) preferentially bind to the ATP-binding pocket of the DFG-in conformation, which is highly conserved between RTKs. By contrast, type II TKIs (sorafenib and quizartinib) preferentially bind to a less conserved region adjacent to the ATP binding pocket of the DFG-out conformation (23,92). The present data on type I TKIs showed that gilteritinib exhibited improved affinity compared with midostaurin for WT-FLT3 (DFG-out and DFG-in). These data agree with those of Egbuna et al (93) and Bultum et al (94), who reported high affinity of gilteritinib for WT-FLT3/DFG-out. Mashkani et al (15) reported that midostaurin had a similar affinity for WT-FLT3, in DFG-out and DFG-in conformations. The present results on type II TKIs also showed that sorafenib and quizartinib had high affinity for the four FLT3 models, which coincides with reports by Egbuna et al (93) for sorafenib and Mirza et al (95) for quizartinib using a WT-FLT3/DFG-out structural model of the receptor.

The differences in the molecular interactions and affinities of the inhibitors may be related to their biochemical characteristics. Both type I and type II inhibitors included in the present study are heterocyclic aromatic compounds; however, they have different characteristic structural groups; the type I inhibitors are indoles and the type II inhibitors are benzimidazoles. This may favor the binding of type I TKIs (WT-FLT3/DFG-in) to the ATP-binding site and the binding of type II TKIs (WT-FLT3/DFG-out) adjacent to the ATP-binding site (96). The changes in affinity forces may also be due to the differences in polar and non-polar bonds, which modulate the stability of the ligand/protein complexes (80).

Gilteritinib (type I TKIs) exhibited low affinity for the active ITD-FLT3/DFG-in conformation compared with sorafenib and quizartinib (type II TKIs) in both conformations (DFG-out and DFG-in), while midostaurin showed an increase in affinity for DFG-out. This difference may be related to the impact of the ITD insertion on the availability of the binding site of FLT3 (97,98).

Although both inhibitors (type I and type II) suppress FLT3, Wodicka et al (99) and Ke et al (13) reported that type I TKIs bind more strongly to the WT-FLT3/DFG-in model. Contrarily, Vijayan et al (18) reported that type II TKIs have an improved selectivity for the inactive conformations of FLT3 (DFG-out) because they are specific for each RTK.

In summary, the HMP and molecular docking tools applied revealed that the four models were suitable for studying the molecular interactions of TKIs with FLT3. They also highlighted the affinity of the second-generation drugs gilteritinib and quizartinib for WT and mutated FLT3 (DFG-out and DFG-in). To corroborate the accuracy of these predictions, their effect was evaluated in an in vitro cellular model of AML.

Several AML cell lines are available to evaluate the cytotoxicity and specificity of TKIs, including the HL60 (WT-FLT3) and MV4-11 (ITD-FLT3-LOH) homozygote cell lines (32). It has been reported that LOH of FLT3 in MV4-11 leukemia cells and patients with AML leads to the constitutive phosphorylation of FLT3, the activation of its related pathways and high malignancy (15). Commonly, ITD-FLT3-LOH is present in 35% of patients with newly diagnosed AML but in 70–80% of refractory patients (32,100); thus, it is particularly important to characterize inhibitors of this receptor.

The presence of the ITD insertion (~30 bp) in the FLT3 sequence of MV4-11 cells was validated by end-point PCR, confirming other reports (32,43,67). In addition, the expression of FLT3 in the MV4-11 and HL60 cells was confirmed by flow cytometry, which also showed that MV4-11 cells expressed more FLT3 receptors on their surface than HL60 cells. Quentmeier et al (32) also reported high expression levels of ITD-FLT3 in MV4-11 cells, which enhanced their proliferation compared with that of MULT-11 or MOLM-13 heterozygote cells, which express WT-FLT3 and ITD-FLT3.

The upregulation of WT-FLT3 expression is commonly observed in patients with newly diagnosed AML who respond satisfactorily to conventional therapy (induction treatments with daunorubicin and cytarabine, and consolidation treatment by bone marrow transplant) (101). However, if the patient (new or refractory) is positive for ITD-FLT3 (homozygous or heterozygous), the only viable alternative is TKI therapy (11). At present, the only FDA-approved TKIs for the treatment of AML are midostaurin (Rydapt^®; 2017), gilteritinib (Xosapta^®; 2018) and quizartinib (Vanflyta^®; 2023) (25,68,90,102). Midostaurin induces adverse effects and presents high toxicity (103). Additionally, this molecule may induce drug resistance, leading to a progressive loss of effectiveness and the need to increase the therapeutic dose during the illness (76). Therefore, it is mainly used in chemotherapy as a complement to cytostatic drugs (daunorubicin and cytarabine) in newly diagnosed patients (86). Sorafenib has not yet been approved for AML because of its strong side effects (104). Gilteritinib and quizartinib are considered competitive inhibitors of WT- and ITD-FLT3 because of their transitory binding to the receptor at the ATP-binding site (type I) or adjacent to it (type II) (76,96). Although both molecules were explicitly designed against WT- and ITD-FLT3, quizartinib has a more potent inhibitory activity (22 mg/day) than gilteritinib (120 mg/day) (68,105,106), Gilteritinib is better tolerated by patients than quizartinib (76). Thus, at present, quizartinib is approved to be used in combination with chemotherapy in newly diagnosed patients, and gilteritinib is recommended for monotherapy in patients with refractory or relapsed AML (68,106).

In the MTT assay, both compounds strongly affected AML cell viability, with IC₅₀ values of 7.99 nM (gilteritinib) and 4.76 nM (quizartinib) against MV4-11 cells, and 57.54 nM (gilteritinib) and 38.75 nM (quizartinib) against HL60 cells. These IC₅₀ values were consistent with previous reports in MV4-11 cells. Kawase et al (34) reported an IC₅₀ of 3.8 nM for gilteritinib and 0.7 nM for quizartinib; Wang and Baron (103) reported an IC₅₀ of 3.34 nM for gilteritinib and 1.07 nM for quizartinib; Hu et al (69) reported an IC₅₀ of 3.02 nM for gilteritinib; and Reiter et al (44) reported an IC₅₀ of 1.8 nM for quizartinib. Furthermore, Mori et al (25) reported that, in Ba/F3 cells (murine interleukin-3-dependent pro-B cell line) transfected with the ITD mutation, gilteritinib (IC₅₀, 1.8 nM) and quizartinib (IC₅₀, 0.46 nM) showed a similar effect to that observed in the present study. Kawase et al (34) also reported an IC₅₀ of 15.5 nM for gilteritinib and 1.6 nM for quizartinib in MOLM-13 cells expressing the WT and ITD FLT3 receptor. These data highlight the strong in vitro effect of gilteritinib (type I TKI) and quizartinib (type II TKI) against leukemic cells and support the clinical evidence showing that they are currently the best TKIs for the treatment of patients with refractory AML.

The discrepancy between the high cytotoxic effect of gilteritinib against MV4-11 cells that express ITD-FLT3 and the prediction using molecular docking of a low affinity for ITD-FLT3 may be related to the Cys703 residue of ITD-FLT3 (WT, Cys694). Ke et al (13), Mashkani et al (15) and Pandurang et al (89) pointed out the importance of this amino acid for the formation of the gilteritinib/FLT3 complexes because it helps maintain the inhibitory effect. This may also explain why gilteritinib is useful (as monotherapy) in AML in refractory patients that present a high percentage of ITD-positive cells (68,90,100,106). Furthermore, the IC₅₀ values revealed that the MV4-11 cells (ITD-FLT3) were more sensitive to both inhibitors than the HL60 cells (WT-FLT3), reflecting the fact that these drugs were designed to inhibit ITD-FLT3.

Cell death by apoptosis and inhibition of cell proliferation are among the main mechanisms that may explain the effect of gilteritinib and quizartinib on MV4-11 and HL60 cells through inhibition of the FLT3 receptor (69,104). Thus, apoptosis was examined by flow cytometry using annexin V (a marker of phosphatidylserine implicated in the induction of apoptosis) and PI (a DNA marker related to loss of membrane integrity) labeling, and cell proliferation was examined using CFSE labeling (a marker of cell generation).

Under the present experimental conditions, both gilteritinib (7.99 nM) and quizartinib (4.76 nM) slightly induced early and late apoptosis in MV4-11 cells after 48 h of exposure, and significantly induced necrosis compared with the vehicle control. These results differ from those of Hu et al (69), who reported a higher percentage (37.2%) of annexin V-positive MV4-11 cells exposed to 2.5 nM gilteritinib for 48 h. However, the present results align with those of other authors, such as Ueno et al (107), who observed that doses of 3–30 nM only induced 20% apoptosis. Similarly, Qiao et al (108) tested gilteritinib concentrations of 12.5, 25 and 50 nM, and observed that the highest concentration promoted apoptosis in only 25% of MV4-11 cells after 24 h. Quizartinib was recently evaluated by Qiu et al (109), who showed that 48 h of exposure to a high concentration of this molecule (50 nM) induced apoptosis in 65.8% of MV4-11 cells. In addition, the present study demonstrated that 48 h of exposure to gilteritinib or quizartinib did not induce apoptosis in HL60 cells or PBMCs. These data agreed with Qiao et al (108) and Darici et al (110) for THP1 cells, HL60 cells and PBMCs.

The present study also demonstrated that the 48-h exposure to the inhibitors significantly decreased MV4-11 cell proliferation from 80.68% (vehicle control) to 20.36% (gilteritinib) and 24.48% (quizartinib). However, this effect was not observed in HL60 cells which showed a similar percentage of proliferation after exposure to gilteritinib or quizartinib compared with the vehicle control, suggesting that one of the mechanisms responsible for the activity of these drugs is through their specific and competitive binding to ITD-FLT3 in MV4-11 cells, which inhibits its activity and delays the generation of new cells. This observation confirmed previous studies with other methodologies (Cell Counting Kit-8 and MTT assays) reporting that both inhibitors (used at nM ranges) have a higher specificity for ITD-FLT3 than for WT-FLT3 (70,111,112).

The viability of PBMCs was not affected following 48 h of exposure to gilteritinib or quizartinib. This result agrees with reports by Qiao et al (108) and Darici et al (110) on PBMCs exposed to high concentrations of gilteritinib (8,000 nM) and quizartinib (500 nM). These cells exhibited no sign of apoptosis or impact on their proliferation. These results are supported by the fact that the main target of both inhibitors is the FLT3 protein. This observation is of importance to avoid unwanted toxic effects in non-tumor blood cells, which are crucial in the functioning and defense of the organism. AML treatments have progressed in the last decade, fostering the development of precision TKIs targeting WT and ITD-mutated FLT3 receptors (76,85,89). The present in silico study showed a strong affinity of second-generation TKIs for WT-FLT3 (DFG-out and DFG-in) compared with the first-generation drugs. Additionally, the ITD mutation in FLT3 impacted the binding of the TKIs tested, and the effects were different for each inhibitor. The high affinities of gilteritinib (for WT-FLT3/DFG-out and WT-FLT3/DFG-in) and quizartinib (for all four FLT3 models) predicted in silico were tested in vitro by evaluating their IC₅₀ in MV4-11 (ITD-FLT3) and HL60 (WT-FLT3) cell lines. The MV4-11 cells showed a slight induction of apoptosis and were more sensitive to both inhibitors than the HL60 cells. A significant decrease in proliferation in MV4-11 cells and a milder effect on HL60 cells were observed. No cytotoxic effect against non-tumoral blood cells (PBMCs from a healthy volunteer with null FLT3 expression) was observed, suggesting a specific effect of these TKIs against the FLT3 receptor expressed in leukemic cells.

As aforementioned, in the early stages of AML, patients express low levels of WT-FLT3 (such as HL60 cells). By contrast, 70–80% of patients with refractory or advanced-stage AML present the ITD-FLT3 mutation due to LOH (such as MV4-11 cells), which favors leukemic cell proliferation because of the constitutive expression of FLT3 (12,100). Thus, the rationale for using MV4-11 cells that possess the homozygous ITD-FLT3 mutation and HL60 cells that only express WT-FLT3 was to compare the effect of gilteritinib and quizartinib in a progressive cellular model of AML. Based on the present data and to further explore the contribution of FLT3 in leukemic cell proliferation and the inhibitory mechanism of TKIs, it would be interesting to complement the present in vitro model with more AML cell lines that express WT-FLT3 (THP1 cells) and WT/ITD-FLT3 (MOLM-13 and PL21 cells), and with null expression of FLT3 (U937 and HEL cells) (32,63). Additionally, more work is needed to include an MV4-11 cell line with equal expression of WT-FLT3 and ITD-FLT3 receptor, as well as an FLT3-free control system, which could be achieved by transfection and knockout of expression of FLT3 using interfering RNA (short hairpin RNA or small interfering RNA).

The present study provides a valuable starting point for further research. The construction of the structural models of WT-FLT3 and ITD-FLT3 (DFG-out and DFG-in), the descriptions of the molecular interactions and the calculations of the affinity forces of TKIs are relevant as they contribute to the knowledge of their mechanisms of action and provide high-quality data to search for novel, more specific and potent FLT3 inhibitory molecules of synthetic or natural origins, which are urgently needed for patients with AML. In addition, the present study provides the basis of an in vitro progressive model of AML to test their efficiency.