FLT3 in AML: ITD, D835Y, and the type-I vs type-II pivot

FLT3 is the most frequently mutated kinase in acute myeloid leukemia, and the resistance story splits cleanly along a structural fault line: type-I inhibitors hold their potency through the TKD activation-loop mutations, type-II inhibitors do not. The same fault line that explains why quizartinib lost to D835Y in QuANTUM-R is what makes gilteritinib the salvage option after a type-II failure. Worth walking through, because most readers will hit this same decision point in their own program eventually.

Two mutation classes, very different drugs

Roughly a third of newly diagnosed AML patients carry an activating FLT3 mutation. They fall into two structurally distinct buckets:

• FLT3-ITD (internal tandem duplication, ~25% of AML) — an in-frame duplication of 3-400 base pairs in the juxtamembrane domain that destabilizes the autoinhibitory conformation. The kinase gets stuck signaling. Carries an independently poor prognosis when the allelic ratio is high.
• FLT3-TKD (point mutations in the tyrosine-kinase domain, ~7% of AML) — most often at D835 in the activation loop (D835Y, D835V, D835H), occasionally I836del. These also lock the kinase in the active conformation, but through a different structural route.

That structural distinction is what the inhibitor classes hinge on. Type-I inhibitors bind the active (DFG-in) conformation. They’re indifferent to whether the activation loop is held open by ITD or by a D835 substitution — both look like “active” to a type-I binder. Type-II inhibitorsbind the inactive (DFG-out) conformation, accessing an extra hydrophobic pocket only available when the activation loop folds back. D835 mutations destabilize the DFG-out state, so type-II compounds lose grip.

The three approved drugs and where they live on this map

• Midostaurin (Rydapt) — FDA approved April 2017 based on RATIFY (Stone et al., NEJM 2017). Type-I, multi-kinase. Used with 7+3 induction and consolidation for newly diagnosed FLT3-mutant AML, not for relapsed/refractory monotherapy. Improved 4-year overall survival from 44% to 51%.
• Quizartinib (Vanflyta) — FDA approved July 2023 for newly diagnosed FLT3-ITD AML based on QuANTUM-First (Erba et al., Lancet 2023). Type-II, highly selective for FLT3. Median OS 31.9 vs 15.1 months with chemotherapy alone. The label is ITD-specific: D835Y patients were excluded from QuANTUM-First because the in vitro data already showed quizartinib loses potency against TKD mutants.
• Gilteritinib (Xospata) — FDA approved November 2018 for relapsed/refractory FLT3-mutant AML based on ADMIRAL (Perl et al., NEJM 2019). Type-I, selective for FLT3 and AXL. Median OS 9.3 vs 5.6 months on salvage chemotherapy. Active against both ITD and D835Y because type-I doesn’t need the DFG-out pocket.

How resistance actually unfolds in practice

Smith et al. (Nature 2012) showed it before quizartinib was even approved: pick a type-II FLT3 inhibitor, treat ITD-positive AML, and the on-treatment relapse clones come back enriched for D835Y, D835V, and F691L. The patterns recur whenever a type-II drug applies selection pressure:

• D835Y / D835V (activation-loop) — the canonical escape from type-II compounds. The aspartate-to-tyrosine or -valine swap destabilizes DFG-out, the conformational state the drug needs. Sensitivity to type-I drugs is preserved, which is the rationale for sequencing gilteritinib after a type-II failure.
• F691L (gatekeeper) — the FLT3 analog of ABL T315I and EGFR T790M. A bulky leucine in the gatekeeper position sterically clashes with most ATP-pocket binders, including both quizartinib and gilteritinib. Rare in absolute terms but the most common pan-resistance signal when it appears.
• Off-target bypass — RAS pathway activations (NRAS, KRAS, PTPN11), AXL upregulation, BCR-ABL-like signatures. The clones that pick up bypass mutations don’t care which FLT3 inhibitor you switch to.

Where the next generation is heading

The unmet need is a drug that covers ITD, D835Y, and F691L simultaneously. Three threads matter:

• Next-gen type-I inhibitors with F691L coverage.Crenolanib remains in late-stage trials and retains activity against F691L in vitro. Newer chemotypes (FF-10101 / luxeptinib) are designed to make a covalent bond with Cys828 in the ATP pocket, intentionally bypassing the gatekeeper-clash problem.
• Combinations with venetoclax. The BCL-2 inhibitor shifts the FLT3-AML cell to depend more on the apoptotic threshold and less on FLT3 signaling alone. Several trials (e.g. quizartinib plus venetoclax plus decitabine) have shown durable composite remissions in older patients who would not tolerate intensive chemo.
• FLT3 degraders. Early-stage PROTACs targeting FLT3 (e.g. ARV-FLT3-class molecules in preclinical reports) follow the same logic as the KRAS and EGFR degrader programs: a target that gets degraded can’t escape via second-site mutation in the binding pocket.

What this looks like in a docking workflow

The clean signature you want to see when validating a docking setup against FLT3: a type-II benchmark like quizartinib should score a couple kcal/mol worse against D835Y than against ITD, while a type-I like gilteritinib should look roughly flat across the two. That ΔΔ is the structural rationale the clinic ratified. F691L should degrade both, because the gatekeeper bulk reaches into the ATP pocket regardless of DFG state.

Try the docking yourself

Open Studio and pick FLT3 from the target catalog. The mutation chips include D835Y and F691L. Dock quizartinib and gilteritinib side-by-side against wild-type, D835Y, and F691L — you should see quizartinib degrade on D835Y while gilteritinib holds, and both compounds degrade on F691L. The ΔΔ between WT and D835Y is the type-II resistance signal that drove QuANTUM-R’s exclusion criteria before it ever drove a clinical relapse.

Liganx is molecular docking online: free, browser-based, no install. It is a fast way to put molecular docking on the type-I versus type-II FLT3 question and watch quizartinib and gilteritinib diverge across D835Y and F691L.

Primary sources

• Stone RM, et al. Midostaurin plus chemotherapy for acute myeloid leukemia with a FLT3 mutation. NEJM 377, 454-464 (2017). doi:10.1056/NEJMoa1614359
• Perl AE, et al. Gilteritinib or chemotherapy for relapsed or refractory FLT3-mutated AML. NEJM 381, 1728-1740 (2019). doi:10.1056/NEJMoa1902688
• Erba HP, et al. Quizartinib plus chemotherapy in newly diagnosed patients with FLT3-internal-tandem-duplication-positive acute myeloid leukaemia (QuANTUM-First): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet 401, 1571-1583 (2023). doi:10.1016/S0140-6736(23)00464-6
• Smith CC, et al. Validation of ITD mutations in FLT3 as a therapeutic target in human acute myeloid leukaemia. Nature 485, 260-263 (2012). doi:10.1038/nature11016