Gilteritinib: potent targeting of FLT3 mutations in AML

Abstract

Since the discovery of FMS-like tyrosine kinase-3 (FLT3)-activating mutations as genetic drivers in acute myeloid leukemia (AML), investigators have tried to develop tyrosine kinase inhibitors that could effectively target FLT3 and alter the disease trajectory. Giltertinib (formerly known as ASP2215) is a novel compound that entered the field late, but moved through the developmental process with remarkable speed. In many ways, this drug's rapid development was facilitated by the large body of knowledge gained over the years from efforts to develop other FLT3 inhibitors. Single-agent gilteritinib, a potent and selective oral FLT3 inhibitor, improved the survival of patients with relapsed or refractory FLT3-mutated AML compared with standard chemotherapy. This continues to validate the approach of targeting FLT3 itself and establishes a new backbone for testing combination regimens. This review will frame the preclinical and clinical development of gilteritinib in the context of the lessons learned from its predecessors.

Figure 1.

Hematopoiesis and FLT3 expression. The green zone surrounds the cell types that express FLT3. CLP, common lymphoid progenitor; CMP, common myeloid progenitor; GMP, granulocyte-macrophage progenitor; LT-HSC, long-term hematopoietic stem cell; MEP, megakaryocyte-erythroid progenitor; MPP, multipotent progenitor; NK, natural killer; ST-HSC, short-term hematopoietic stem cell. Professional illustration by Somersault18:24.

Figure 2.

FLT3 ITDs. This diagram illustrates how FLT3 mutations can be defined as “juxtamembrane insertion” and “kinase domain insertion.” For the large majority of FLT3-ITD mutations, the duplicated sequence includes the codon for residue arginine 595 (R595). The resultant amino acid sequence inserted typically ranges from 3 to 42 residues. When the insertion is short, it consists only of juxtamembrane sequence, whereas longer duplications include residues from the kinase 1 domain. Therefore, in the case of these longer insertions, the actual duplicated sequence begins within that kinase domain and are sometimes referred to as a kinase domain insertion. Professional illustration by Somersault18:24.

Figure 3.

PIA assay for FLT3. This assay serves as a validated surrogate for in vivo FLT3 inhibition in patients treated with FLT3 inhibitors. Whole blood is collected at a trough time point from patients treated with an FLT3 inhibitor, preferably when the drug is at steady-state. An FLT3-ITD–expressing cell line, such as Molm14, is then incubated in the plasma, and the phosphorylation status of FLT3 is analyzed by immunoblotting. The results are normalized to control or pretreatment plasma. The immunoblots can be quantitated by densitometry. The blots are from a patient treated with 80 mg/d or 60 mg/d of lestaurtinib, a first-generation FLT3 inhibitor. Professional illustration by Somersault18:24.

Figure 4.

Type I vs type II FLT3 inhibitors. Gilteritinib is a type I inhibitor. As such, it is more or less an ATP mimetic, and its binding is relatively less influenced by the conformation of the activation loop. “DFG” refers to 3 highly conserved amino acid residues (aspartate-phenylalanine-glycine) at the start of the activation loop. A type II inhibitor fits into a hydrophobic groove adjacent to the ATP binding site, and, in the case of quizartinib, actually fits in between phenyl rings of phenylalanine 691 (F691, the “gatekeeper”) and phenylalanine 830 (F830). Professional illustration by Somersault18:24.

Figure 5.

Molecular structure of gilteritinib.

Figure 6.

KINOMEscan assay profile for gilteritinib at 100 nM. This commercially available assay of a kinase inhibitor’s relative potency and selectivity is widely displayed but often misunderstood. This profile displays gilteritinib’s affinity for the different kinases in the human genome at 100 nM. A concentration of 100 nM gilteritinib is a crude approximation of steady-state levels achieved in plasma by patients receiving the standard dose of 120 mg/d. The KINOMEscan assay measures the ability of a test compound, such as gilteritinib, at specific concentrations (100 nM in this case) to displace a reference ligand from the active site of a kinase. Therefore, it is not an in vitro kinase assay but rather a drug-binding assay. The reference ligand is a proprietary compound designed to bind in highly promiscuous fashion to kinase active sites (earlier developmental versions of this assay used staurosporine). Kinases, derived from a library of 468 genes, are individually expressed in bacteriophages for use in the assay. In the case of transmembrane receptor kinases, just the cytoplasmic domain is expressed. The library includes the vast majority of wild-type human kinases, as well as key mutant kinases, such as FLT3-ITD and FLT3-TKD variants, including the F691L gatekeeper mutation. Each small gray circle on a branch represents an individual kinase. A red circle over any point of a branch indicates that gilteritinib competes with the reference compound for binding. “Percent control” refers to the efficacy of the competition. The larger the circle, the more effectively it displaces the reference ligand. A value of “0%” (the largest red circle) indicates the highest-affinity binding, meaning no reference ligand binds to the kinase. Note that these are not dissociation constant (Kd) values, although this assay can be used to calculate such values. Images generated using the TREEspot Software Tool and reprinted with permission from KINOMEscan, a division of DiscoveRx Corporation. © 2010 DiscoveRx Corporation.