Successful treatment of relapsed FLT3-mutated donor cell-derived MDS/AML with FLT3 inhibitor gilteritinib

Abstract

Donor cell-derived myelodysplastic syndrome/acute myeloid leukemia is a rare but serious complication of allogeneic hematopoietic stem cell transplantation, and its optimal treatment has not been established. Here, we report a case of a 44-year-old woman who was diagnosed with donor cell-derived myelodysplastic syndromes with excess blasts carrying RAD21 and KMT2D mutations after six-year clinical remission of her initial acute myeloid leukemia treated with bone marrow transplantation from an unrelated male donor. The disease subsequently transformed to acute myeloid leukemia harboring a newly acquired FLT3 internal tandem duplication mutation. Azacitidine and venetoclax with cytarabine were both ineffective, but gilteritinib monotherapy rapidly reduced blasts and achieved near complete remission. The patient then underwent haploidentical stem cell transplantation from her son, followed by gilteritinib maintenance therapy. She achieved complete remission with clearance of donor cell-derived mutations and has remained in remission for more than one year. To our knowledge, this is the first case report of a relapsed FLT3-mutated donor cell-derived acute myeloid leukemia who was successfully treated with a FLT3 inhibitor, gilteritinib, which highlights the potential effectiveness of gilteritinib not only as a salvage therapy but also as an effective bridging and maintenance therapy in relapsed FLT3-mutated donor cell-derived acute myeloid leukemia.

Introduction

Donor cell-derived myelodysplastic syndromes/acute myeloid leukemia (DC-MDS/AML) is a rare but serious complication of allogeneic hematopoietic stem cell transplantation (allo-HSCT). Among 2390 allo-HSCT recipients engrafted with either bone marrow (BM), peripheral blood (PB) or umbilical cord blood, the incidence of DC-MDS/AML within a median follow-up of 7.1 years was 0.53–0.56% [1]. More recently, a group of the European Society for Blood and Marrow Transplantation reported that the estimated prevalence of DC-MDS/AML was 80.5/100,000 transplants [2]. Overall survival (OS) was poor with 29 out of 38 patients were dead at a median of 11 (range 0–91) months after DC-MDS/AML diagnosis [2]. These studies clarified the rarity of DC-MDS/AML among allo-HSCT recipients, nonetheless, it has a significantly negative impact on HSCT outcome and its standard treatment has yet to be established.

FMS-like tyrosine kinase 3 (FLT3) is a receptor tyrosine kinase which regulates hematopoietic stem/progenitor cell proliferation and differentiation [3]. FLT3-activating mutations, such as internal tandem duplications within the juxtamembrane region (FLT3-ITD) or missense mutations in the tyrosine kinase domain (FLT3-TKD), occur in 25–30% of AML patients [4, 5]. Although FLT3-ITD mutations are adverse prognostic factors in relapsed or refractory (r/r) AML, recent studies have shown that FLT3 inhibitors, including gilteritinib and quizartinib, can significantly improve OS, complete remission (CR) rates and bridging rates to allo-HSCT in r/r FLT3-mutated AML compared to conventional chemotherapy [5–7]. In addition, several studies have demonstrated clinical benefits of post-transplant maintenance therapy with FLT3 inhibitors, including sorafenib, quizartinib or gilteritinib, in r/r FLT3-mutated AML [8–13]. However, whether FLT3 inhibitors are similarly effective to the treatment of r/r FLT3-mutated DC-MDS/AML remains to be elucidated.

Here, we report a case of relapsed FLT3-mutated DC-MDS/AML who achieved long-term molecular CR after gilteritinib monotherapy followed by haploidentical (haplo-) HSCT and post-transplant gilteritinib maintenance therapy.

Case report

The case was a 38-year-old woman who was referred to her previous physician in April, year X-7 due to leukocytosis (white blood cell, WBC, 175 × 10³/µL, blast 89.0%), anemia (hemoglobin, Hb, 9.5 g/dL) and thrombocytopenia (platelet, plt, 7.1 × 10⁴/µL) (Fig. 1). BM smear presented with an expansion of myeloperoxidase (MPO)-positive myeloblasts (79.8%) and she was diagnosed with AML (French-American-British classification: AML-M2, World Health Organization (WHO) classification: AML-not otherwise specified (NOS), FLT3-ITD mutation: negative, karyotype: 47, XX, +r1 [18/20], 47, idem, i(7)(q10) [2/20]). The patient was decided to be treated based on Japan Adult Leukemia Study Group (JALSG) AML201 protocol. Induction chemotherapy and three courses of consolidation chemotherapies led to the first hematological CR (Fig. 1). At the first CR, she received bone marrow transplantation (BMT) on October 4th, year X-7, from an unrelated male donor (human leukocyte antigen, HLA, one allele mismatch, HLA-A allele mismatch) using busulfan (BU, 3.2 mg/kg) and cyclophosphamide (CY, 120 mg/kg) as a preconditioning regimen (Fig. 1). The patient achieved CR after the first allo-HSCT and maintained CR thereafter for more than 6 years.

Fig. 1.

Clinical course of the patient. Changes in PB blast (%, red line) of the patient. The upper panels show chemotherapy regimens (black) with dosing periods (double-headed arrows). The two large black arrows in the graph show timepoints of the first and second HSCTs. The lower pie charts represent percentages of blasts in BM at the indicated time points (black portion). AZA; azacitidine, VEN; venetoclax, Ara-C; cytarabine

On December 19th, year X-1, approximately 6 years after the first allo-HSCT when the patient was 44 years old, slight anemia (Hb 11.3 g/dL), thrombocytopenia (Plt 8.0 × 10⁴/µL) and increased blasts in PB (2.0%) were pointed out at her routine follow-up (Fig. 1). BM examination performed on December 21st, year X-1, presented with an increase of MPO-positive myeloblasts (7.8%), megakaryocytic dysplasia (micromegakaryocyte and multinucleation, Fig. 2A) and granulocytic dysplasia (decreased granules and nuclear hyposegmentation, Fig. 2B). Karyotype was 46, XY male donor type in all 20 cells analyzed in metaphase. Therefore, the patient was diagnosed with DC-MDS (international consensus classification (ICC): MDS with excess blasts, WHO classification: MDS with increased blasts 1 (MDS-IB1)) and was categorized into intermediate risk based on revised international prognostic scoring system (IPSS-R). Targeted deep sequencing (TDS) of the BM sample detected pathogenic mutations in genes encoding a cohesion factor RAD21 (RAD21 c.1684 C > T, p.Gln562Ter, variant allele frequency (VAF) 35.4%) and a histone modifier KMT2D (KMT2D c.1311G > T, p.Glu437Asp, VAF 45.7%) which are both recurrently mutated in MDS and AML [14, 15] (Fig. 3). The patient was treated with three courses of azacitidine (AZA) as a bridging therapy prior to the second allo-HSCT (Fig. 1). However, the percentage of PB blasts gradually increased and subsequent BM exam performed in May, year X, revealed increased BM blasts to 11% (Fig. 1). Therefore, she was diagnosed with transformation to MDS/AML (ICC). Karyotype was 46, XY, t(11;19)(q23.2;p13.1) in all 20 cells analyzed in metaphase. TDS of the BM sample revealed persistent RAD21 mutation (RAD21 c.1684 C > T, p.Gln562Ter, VAF 41.1%) and KMT2D mutation (KMT2D c.1311G > T, p.Glu437Asp, VAF 40.7%) at this time point (Fig. 3). The patient was switched to venetoclax (VEN) + Ara-C, though it was discontinued due to grade 4 neutropenia (Fig. 1). Unfortunately, PB blasts rapidly increased and BM exam on June 25th, year X, revealed blast percentage of 87.4% (Fig. 1). At this point, genetic testing demonstrated that the patient was positive for FLT3-ITD mutation. Treatment with gilteritinib 120 mg/day was promptly started on June 27th, year X (Fig. 1). The patient achieved near CR (BM blasts 5.7%) after 3 weeks of gilteritinib monotherapy (Fig. 1). She received haploidentical related peripheral blood stem cell transplantation (haplo-rPBSCT) on August 14th, year X, from her son using BU, fludarabine (FLU) and total body irradiation (TBI) as a preconditioning regimen and post-transplant cyclophosphamide (PTCY) as a prophylaxis of graft-versus-host disease (GVHD) (Fig. 1). The patient achieved marrow CR with donor chimerism of 99% on day 27. BM karyotype was restored to 46, XY male normal karyotype in all 20 cells analyzed in metaphase. TDS of the BM sample derived on day 27 showed complete clearance of both RAD21 and KMT2D mutations (Fig. 3). Gilteritinib 40 mg/day was started on day 28, as a post-transplant maintenance therapy (Fig. 1). The patient was discharged on day 43 and has remained CR for more than a year under gilteritinib maintenance therapy.

Fig. 2.

Bone marrow pathology and dysplasia. Representative images of BM smears with BM dysplasias. (A) Images of megakaryocytic dysplasia (left panel; micromegakaryocyte, right panel; multinucleation). (B) Images of granulocytic dysplasias and myeloid blasts (large arrows; Pseudo-Pelger-Huët anomaly, arrow heads; decreased granules, small arrows; myeloid blasts)

Fig. 3.

VAF of RAD21 and KMT2D mutations. Bar graph showing the VAFs of RAD21 (red bar graph) and KMT2D (blue bar graph) mutations detected in the patient on 12/21/X − 1, 05/21/X, and 09/10/X. The patient was diagnosed with DC-MDS (MDS-EB1) on 12/21/X − 1, when high VAFs of both mutations were observed. AZA treatment was initiated on 2/27/X, and administered again on 3/26/X and 4/23/X. Although high VAFs of RAD21 and KMT2D mutations persisted on 05/21/X even after AZA treatment, both mutations became undetectable on 09/10/X after gilteritinib bridging therapy and haplo-rPBSCT

Discussion

Here, we report a case of relapsed FLT3-mutated DC-MDS/AML in which gilteritinib was effective in terms of bridging therapy to haplo-rPBSCT and post-transplant maintenance therapy. This case suggests the effectiveness of gilteritinib not only against r/r FLT3-mutated primary MDS/AML but also against r/r FLT3-mutated DC-MDS/AML cases.

Previous studies have shown risk factors of DC-MDS/AML. Engel et al. identified three factors on multivariate analysis as significantly associated with an increased risk for DC-MDS/AML: use of granulocyte-colony stimulating factor (G-CSF) within the first 100 days after transplantation, in vivo T-cell depletion by anti-thymocyte globulin or alemtuzumab and multiple allografts [2]. Our case started to receive G-CSF on day 1 in the first unrelated BMT to promote neutrophil recovery in the HLA-A allele mismatch setting. As G-CSF administration after allo-HSCT before engraftment is not uncommon to facilitate neutrophil recovery and reduce infections, caution against DC-MDS/AML should be paid in allo-HSCT recipients who received G-CSF after transplantation [2].

Recent studies have identified genetic abnormalities potentially associated with DC-MDS/AML. Williams et al. reported that over 20% of donor-derived leukemias carry chromosome 7 abnormalities, mostly monosomy 7 [16]. KMT2A rearrangement has been identified in approximately 3% of DC-MDS/AML which was also found in our case, suggesting that KMT2A rearrangement is a recurrent cytogenetic event in DC-MDS/AML [16]. In addition, recent advances in whole exome/genome sequencing have detected candidate genetic mutations related to DC-MDS/AML, including mutations in CEBPA, GATA2, JAK2, RUNX1, DDX41, EZH2, IDH1/2, DNMT3A, ASXL1, XPD, XRCC3, and CHEK1 [16]. FLT3-ITD, detected in our case, is not in this gene list, nonetheless, there are previous case reports of FLT3-mutated DC-AML. Hahn et al. reported two sibling cases of FLT3-mutated AML arising from a single pre-leukemic DNMT3A-mutant clone [17]. They have shown that DNMT3A^R882H positive clonal hematopoiesis (CH) BM was accidentally transplanted from an apparently healthy 53-year-old elder brother to his 49-year-old younger brother [17]. Consequently, the donor developed normal karyotype AML with mutations in DNMT3A, NPM1 and FLT3 fourteen months after BMT, whereas the recipient developed normal karyotype DC-AML with mutations in DNMT3A, NPM1 and SMC1A [17]. Herold et al. reported another sibling pair of FLT3-mutated AML arising from a single pre-leukemic DNMT3A/TET2-double mutant clone [18]. Again, accidental transplantation of DNMT3A/TET2-double mutant CH BM triggered abnormal karyotype (47, XX, + 8) AML with mutations in DNMT3A, IDH1, PHF6, CBL and SMC1A in the female donor approximately seven years after BMT, whereas the male recipient developed normal female karyotype (46, XX) DC-AML with mutations in DNMT3A, TET2, NPM1 and FLT3-ITD around the same time period [18]. Chonabayashi et al. also reported a successful case of second cord blood transplantation in a patient with cord blood DC-AML harboring FLT3-ITD and NPM1 mutations [19]. These studies indicate that, similar to adult primary AML, FLT3-ITD functions as a driver event in leukemic transformation from precedent CH and commonly co-segregates with DNMT3A and NPM1 mutations in DC-AML as well [17–20]. As FLT3-ITD and KMT2A rearrangement were both acquired at the time of transformation to MDS/AML in our case, FLT3-ITD may have driven her MDS progression in our DC-MDS/AML case as previously reported [17–19]. In contrast, our DC-MDS/AML case is distinct from the reported DC-AML cases in the way that FLT3-ITD did not co-occur with DNMT3A or NPM1 mutations but rather co-segregated with RAD21 and KMT2D mutations. Although RAD21 and KMT2D mutations identified in our case have not been reported in existing cases of DC-MDS/AML, both of these genes are included in proposed genes for donor screening prior to transplantation for donors with personal or family history of hematologic and solid cancers, which corroborates that RAD21 and KMT2D mutations are previously unreported candidate genetic mutations related to DC-MDS/AML [16]. This notion is further supported by recent studies showing that lymphoma/myeloma patients undergoing autologous stem cell transplantation with CH mutations in DNA repair genes (RAD21, TP53, PPM1D) had a significant inferior OS with increased rates of therapy-related myeloid neoplasms (t-MN) compared to patients without CH mutations in these genes [21, 22]. Of note, among CH mutations in these DNA repair genes, RAD21 mutations have been reported to be correlated with patients’ younger age [23]. Although KMT2D is recurrently mutated as a lymphoid lineage CH gene in the context of CH and therefore the role of this gene mutation as a putative pre-leukemic event is elusive in this case, these observations indicate that the precedent RAD21-mutant clone may have contributed to the development of DC-MDS/AML in relatively young age in this case [23, 24]. Importantly, lymphoma patients with CH mutations in DNA repair genes (RAD21, TP53, PPM1D) showed 13% cumulative incidence of t-MN only at 3 years or later since PB stem cell harvest [21]. Moreover, Rad21 knock-down in murine BM cells in vivo led to enhanced HSC fitness and myeloid skewing without leukemic transformation [25]. These findings clearly underscore a critical role of RAD21-mutated clones as putative pre-leukemic origins, nonetheless, RAD21 mutation alone is insufficient to drive leukemic transformation, and additional genetic events may be required for disease progression. This may explain the long latency before the onset of DC-MDS/AML in our case where FLT3-ITD may have driven leukemic transformation in her final stage. However, this remains a matter of speculation in this case as the BM sample before the first BMT is lacking which precluded TDS and therefore the presence of ancestral RAD21-mutated clones cannot be verified.

Several studies have shown clinical efficacy of FLT3 inhibitors for r/r FLT3-mutated primary AML as a salvage chemotherapy, a bridging therapy prior to allo-HSCT or a post-transplant maintenance therapy [5–13]. However, whether FLT3 inhibitors are also beneficial for the treatment of r/r FLT3-mutated DC-MDS/AML are elusive. Here, we reported a case of r/r FLT3-mutated DC-MDS/AML who was successfully treated with a FLT3 inhibitor, gilteritinib, as a bridging therapy to haplo-rPBSCT, achieved and maintained molecular CR with haplo-rPBSCT and post-transplant gilteritinib maintenance therapy. To our knowledge, this is the first case report of a r/r FLT3-mutated DC-MDS/AML who achieved clinical response to sequential treatment with gilteritinib as a salvage, bridging and maintenance therapies. Our case suggests that a FLT3 inhibitor, gilteritinib, is a promising therapeutic modality as a salvage, bridging and maintenance therapy with manageable toxicity not only for r/r FLT3-mutated primary AML but also for DC-MDS/AML. As mutated FLT3 targeted therapy by gilteritinib has been shown to be more effective and less toxic for r/r FLT3-mutated AML than conventional salvage chemotherapy, it is vital to confirm FLT3 mutation status in r/r DC-MDS/AML cases, start treatment with gilteritinib and verify treatment response in FLT3-mutated cases [5].

Author contributions

M.K. and H.K. wrote the manuscript; M.K. and H.K. developed figures; H.K., A.I., A.A., A.M., C.Y., K.H., M.S., M.T., T.S., T.O., H.T., T.M., T.T., M.H., K.M. and H.N. carried out all the laboratory tests and were involved in patient management. All authors reviewed the manuscript.

Funding

This work was supported in part by the Japanese Society for the Promotion of Science (JSPS) KAKENHI (grant number JP24K11565 (H.K.)), Research Grant for the Promotion of Advanced Medicine in Yokohama City University (H.K.), and JSPS KAKENHI (grant number JP23H02939 (H.N.)).

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Consent to publish

The participant has consented to the submission of the case report to the journal and written informed consent for publication was obtained from the patient.

Ethics declaration

This work was conducted in accordance with the Declaration of Helsinki.