Hidden Leukemic Stem Cells That Fuel Relapse in FLT3-Mutated AML Finally Exposed

Acute myeloid leukemia (AML) carrying the FLT3-ITD mutation has long been one of the most stubborn enemies in hematology. Even after the arrival of potent FLT3 inhibitors such as gilteritinib, a substantial share of patients relapse, and once the disease returns it is often far harder to control. A new study published in Cell Death Discovery by Serena Travaglini, Maria Teresa Voso and colleagues at the University of Rome Tor Vergata, together with collaborators across Italy, offers a detailed explanation of why: a rare population of leukemic progenitor cells survives frontline therapy, hides from standard monitoring, and acts as a reservoir from which the leukemia re-emerges. The work combines flow cytometry, mouse xenograft experiments, transcriptomics and drug testing into one of the most complete portraits yet of these treatment-resistant cells.

The team focused on a population they had previously identified, marked by the co-expression of the surface proteins CD34, CD123, CD25 and CD99. In their earlier work, these CD34/CD123/CD25/CD99-positive leukemic progenitor cells (LPCs) were seen to persist in patients after induction chemotherapy and to herald disease recurrence in FLT3-ITD-mutated AML. What remained unclear was whether these cells were merely passive survivors or genuine leukemia-initiating cells with stem-like capacity. The new study answers that question directly, providing what the authors describe as the first functional evidence that FLT3-ITD-mutated LPCs maintain stem cell potential in vivo.

The clinical part of the investigation began with a diagnostic twist that could change practice. Using multiparametric flow cytometry on prospective patient samples, the researchers showed that when the CD34/CD123/CD25/CD99-positive fraction exceeds 7.4 percent of the relevant cell compartment, the presence of an FLT3-ITD mutation can be predicted with 88.2 percent sensitivity and, remarkably, 100 percent specificity. In other words, an immunophenotypic signature visible on a standard flow cytometry panel can act as a surrogate flag for the mutation, and the same signature can be tracked over time as a form of measurable residual disease (MRD) monitoring that does not depend on molecular assays alone.

To test whether these cells truly behave like leukemia stem cells, the team purified LPCs by cell sorting and transplanted them into conditioned NSG mice, an immunodeficient strain widely used for human hematopoietic engraftment studies. The sorted LPCs engrafted efficiently and reinitiated leukemia in the recipient animals. Critically, when the resulting leukemic cells were harvested and transplanted into successive cohorts of mice, the LPC compartment expanded with each sequential passage. This serial transplantation behavior, with the progenitor pool growing rather than exhausting, is exactly what would be expected from a self-renewing disease reservoir capable of sustaining relapse.

With the functional identity of the LPCs established, the researchers turned to molecular profiling to understand what makes them tick. They performed transcriptomic analysis on 13 paired samples of LPCs and leukemic blasts from the same patients, allowing a direct comparison between the progenitor population and the bulk tumor cells. The most striking finding was an upregulation of genes encoding protein kinases in the LPCs, with the lymphocyte-specific protein tyrosine kinase, known as LCK, standing out as the dominant signal. LCK is classically associated with T-cell receptor signaling, so its prominence in myeloid leukemic progenitors was unexpected and immediately suggested a druggable vulnerability.

The LCK finding was not a one-off artifact of a small cohort. The team went on to confirm elevated LCK expression in 28 additional FLT3-ITD-mutated AML samples, consolidating the observation across an independent validation set. This reproducibility matters, because candidate therapeutic targets in AML have a long history of failing to survive independent validation. Here, the kinase overexpression pattern appeared consistently in the very cells that seed relapse, not just in the bulk blast population that most genomic studies interrogate.

On the therapeutic side, the researchers examined how the FLT3 inhibitor gilteritinib, already approved for relapsed or refractory FLT3-mutated AML, affects these progenitor cells. Gilteritinib reduced proliferation and dampened the FLT3-STAT5 signaling pathway, a canonical downstream axis of the mutant receptor, in both FLT3-ITD-mutated blasts and LPCs. That the drug reaches its canonical target even within the stem-like compartment is encouraging, but the experiments also revealed why single-agent inhibition is not enough: the LPCs retained survival capacity, pointing to parallel signaling routes that keep the cells alive when FLT3 itself is blocked.

That parallel route appears to run through LCK. When the investigators added LCK inhibition on top of gilteritinib, the effects were more than additive in key respects. Dual blockade further attenuated FLT3/STAT5 activity, disrupted the aberrant subcellular localization of the FLT3 receptor, and downregulated ABC transporter-associated multidrug resistance, the family of efflux pumps that chemotherapy and targeted agents alike must overcome. The combination of gilteritinib with LCK blockade enhanced the overall anti-leukemic activity, consistent with the idea that FLT3 and LCK cooperate to sustain FLT3-ITD-mutated leukemic progenitors and that hitting both pathways simultaneously collapses their support network.

The mechanistic detail about FLT3 localization deserves particular attention. The biosynthetic trafficking of FLT3-ITD has been implicated in the mutant receptor’s ability to transduce autonomous growth signals from intracellular compartments, and the finding that LCK inhibition corrects aberrant FLT3 localization suggests that LCK may help position the mutant receptor where it can signal most effectively. Meanwhile, the suppression of ABC transporter expression links the LCK axis to one of the classic mechanisms of chemoresistance in AML, offering a potential explanation for why LPCs survive induction chemotherapy in the first place.

Taken together, the study delivers three practical takeaways. First, a defined immunophenotypic signature above a 7.4 percent threshold provides a clinically actionable tool for identifying FLT3-ITD-mutated disease and monitoring residual leukemic progenitors by flow cytometry. Second, the in vivo engraftment and serial passage data cement LPCs as the functional reservoir of relapse, making them the logical target for eradication strategies aimed at long-term remission. Third, the LCK pathway emerges as a novel molecular vulnerability that can be attacked in combination with existing FLT3 inhibitors, a strategy the authors argue could selectively eliminate residual FLT3-ITD-mutated LPCs and improve long-term outcomes in a disease where relapsed and refractory cases remain a pressing unmet medical need. As gilteritinib-resistant relapses continue to challenge clinicians, the prospect of pairing FLT3 blockade with LCK-directed agents gives the field a concrete, mechanistically grounded next step.

Subject of Research: Functional and molecular characterization of FLT3-ITD-mutated leukemic progenitor cells in acute myeloid leukemia

Article Title: Characterization of FLT3-ITD-mutated leukemic stem cells: functional and clinical insights

Article References: Travaglini, S., Ottone, T., Antonelli, S., Serra, M., Silvestrini, G., Masciarelli, S., Sniegocka, M., Noguera, N. I., De Bardi, M., Lomi, M., Consalvo, M. A. I., Gurnari, C., Piazza, R., Divona, M., Marchesi, F., Fazi, F., Mattei, M., Buccisano, F., Battistini, L., … Voso, M. T. (2026). Characterization of FLT3-ITD-mutated leukemic stem cells: functional and clinical insights. Cell Death Discovery. https://doi.org/10.1038/s41420-026-03319-0

Image Credits: AI Generated

DOI: 10.1038/s41420-026-03319-0

Keywords: acute myeloid leukemia, FLT3-ITD, leukemic stem cells, leukemic progenitor cells, gilteritinib, LCK, STAT5 signaling, measurable residual disease, flow cytometry, NSG mice, drug resistance, relapse