Research ArticleCancer

Preventing chemotherapy-induced myelosuppression by repurposing the FLT3 inhibitor quizartinib

See allHide authors and affiliations

Science Translational Medicine  09 Aug 2017:
Vol. 9, Issue 402, eaam8060
DOI: 10.1126/scitranslmed.aam8060

Rock-a-bye bone marrow

Although chemotherapy saves the lives of many cancer patients, it is a difficult treatment that induces many major side effects, with one of the most common being myelosuppression (depletion of bone marrow cells). The consequences of myelosuppression include anemia, thrombocytopenia, and neutropenia, all of which can cause severe complications and delay subsequent courses of chemotherapy. Taylor et al. discovered that quizartinib, a tyrosine kinase inhibitor, can decrease the risk of myelosuppression during cancer treatment by transiently suppressing the proliferation of bone marrow progenitor cells. In contrast, cancer cells continue to proliferate during treatment, making them a target for chemotherapy even while the bone marrow is protected, as the authors demonstrated in mice with leukemia.

Abstract

We describe an approach to inhibit chemotherapy-induced myelosuppression. We found that short-term exposure of mice to the FLT3 inhibitor quizartinib induced the transient quiescence of multipotent progenitors (MPPs). This property of quizartinib conferred marked protection to MPPs in mice receiving fluorouracil or gemcitabine. The protection resulted in the rapid recovery of bone marrow and blood cellularity, thus preventing otherwise lethal myelosuppression. A treatment strategy involving quizartinib priming that protected wild-type bone marrow progenitors, but not leukemic cells, from fluorouracil provided a more effective treatment than conventional induction therapy in mouse models of acute myeloid leukemia. This strategy has the potential to be extended for use in other cancers where FLT3 inhibition does not adversely affect the effectiveness of chemotherapy. Thus, the addition of quizartinib to cancer treatment regimens could markedly improve cancer patient survival and quality of life.

INTRODUCTION

Cytotoxic chemotherapy for the treatment of cancer causes a range of side effects that adversely affect patient health and quality of life. One such side effect is myelosuppression, where chemotherapy massively depletes bone marrow progenitor cells resulting in anemia, neutropenia, and thrombocytopenia (13). Patients suffering from myelosuppression can experience complications such as fatigue, dizziness, bruising, hemorrhage, and potentially fatal opportunistic infections. Consequently, drug dosage and frequency may be limited to abrogate these complications, in turn, compromising the effectiveness of the treatment (4).

Current management of chemotherapy-induced myelosuppression is costly and of limited effectiveness. The most common treatments are blood transfusion, growth factor injections, and, in some cases, where complete myeloablation is inevitable, bone marrow transplantation. Blood transfusions provide rapid improvements to the peripheral blood profile but are ineffective in reversing the most serious complication of myelosuppression, neutropenia (2). Growth factor injections promote blood cell differentiation and are used with regimens that cause profound myelosuppression. Granulocyte colony-stimulating factor (G-CSF) is used to promote myeloid cell generation (57), whereas thrombopoietin receptor agonists, such as romiplostim, are under development to promote platelet recovery (8). However, to be effective, growth factors require a sufficiently large pool of hematopoietic progenitors to exert their therapeutic effects. Therefore, survival of these cells is essential for successful recovery of bone marrow cellularity. The scarcity of options available for treating chemotherapy-induced myelosuppression therefore warrants the investigation of additional therapeutic approaches.

The hematopoietic cells that are most affected by chemotherapy are the rapidly cycling multipotent and lineage-committed progenitors. Therefore, finding ways to protect these cells from the cytotoxic effects of chemotherapy could circumvent the development of myelosuppression. Multipotent progenitors (MPPs) have recently been identified in mice as the main drivers of lifelong blood cell production (9), so their depletion by chemotherapy has profound consequences for all blood lineages. MPPs are found within the lineage-negative, Sca-1+, c-Kit+ (LSK) population in the bone marrow, and the expression of the FLT3 receptor tyrosine kinase distinguishes them from other LSK cells. In recent years, we have been studying the enhanced activity of FLT3 in MPPs from c-Cbl RING finger mutant mice and the effects of treatment with the FLT3 inhibitor quizartinib (AC220) (1012). An unexpected finding from our studies of dosing c-Cbl RING finger mutant mice with quizartinib was the rapid but transient induction of quiescence in MPPs (11). Here, we show that quizartinib similarly induces the quiescence of wild-type (WT) MPPs and that this effect provides marked protection when mice are treated with chemotherapy. Thus, by using quizartinib to preferentially protect hematopoietic progenitors from chemotherapy, without affecting the ability of cytotoxic drugs to eliminate FLT3-independent tumors, we show that the repurposing of quizartinib has the potential to provide a safe and effective addition to cancer treatment.

RESULTS

Quizartinib dosing induces a rapid and transient quiescence of MPPs in C57BL/6 mice

To determine whether dosing with quizartinib induces quiescence in MPPs from WT mice, in a manner similar to MPPs from c-Cbl RING finger mutant mice (11), we dosed C57BL/6 (B6) mice once daily with quizartinib (10 mg/kg) or vehicle over 1 to 16 days. We found that after 1 or 2 days of dosing with quizartinib, there was a marked reduction in the proliferative activity of MPPs with a significant decrease in Ki-67+ cells (P < 0.01 and P < 0.0001, respectively) and a marked reduction in cell size (as measured by forward light scatter), compared to vehicle-dosed mice (Fig. 1, A to C). Most notable was the decrease in the proportion of large Ki-67+ MPPs (Fig. 1C). Consistent with these findings were the marked reductions in the proportions of MPPs in S + G2-M phases of the cell cycle after 1 or 2 days of quizartinib dosing (Fig. 1D).

Fig. 1. Quizartinib induces a rapid and transient quiescence in MPPs.

(A) B6 mice were dosed once daily over a 16-day time course with either vehicle (VEH) or quizartinib (QUIZ; 10 mg/kg) and analyzed 24 hours after the final dose. FLT3+ LSK cells (MPPs) were analyzed by flow cytometry for the expression of Ki-67 and cell size [forward light scatter (FSC)], and the percentages of cells within each quadrant are shown. Cells in the top right quadrant are large Ki-67+ cells (highly proliferative cells), and the cells in the lower left quadrant are small Ki-67 cells (G0 quiescent cells). (B) Percentages of Ki-67+ MPPs and (C) large Ki-67+ MPPs at each time point. (D) Cell cycle analysis showing the proportion of MPPs in the S + G2-M phases after 1 or 2 days of vehicle or quizartinib dosing. (E) Amount of FLT3 ligand in the serum over the 16-day time course. Data are from three to four mice at each time point for each group, and the graphs are expressed as mean ± SEM. All statistics were calculated using Student’s t tests; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

After 4 days of daily dosing, the proportion of Ki-67+ MPPs from quizartinib-dosed mice remained low, but a recovery of large Ki-67+ MPPs became evident (Fig. 1C) and was associated with a rise in the amount of FLT3 ligand (Fig. 1E). By day 8, the proliferative activity of MPPs from quizartinib-dosed mice was restored (Fig. 1, A and B), and the magnitude of this recovery was evident by the higher proportion of large Ki-67+ MPPs from quizartinib-dosed mice compared to vehicle-dosed mice (Fig. 1C). This response at day 8 correlated with high amounts of FLT3 ligand in the serum (Fig. 1E), suggesting the presence of a recovery feedback loop.

Quizartinib priming protects hematopoietic progenitors and stem cells from fluorouracil-induced death

The transient induction of quiescence of MPPs in quizartinib-dosed mice raised the possibility that short-term quizartinib dosing may provide an approach for protecting these cells from chemotherapy-induced death. To test this, we investigated fluorouracil (5-FU), a pyrimidine antagonist that primarily induces cell death during the S phase of cell growth and is used in the treatment of many human cancers. We first examined the extent to which 5-FU depletes hematopoietic stem cells (HSCs) and progenitors after a single intraperitoneal injection of 150 mg/kg. This is a clinically relevant dose using body surface area normalization (13). Two days after 5-FU administration, there was a massive depletion of LSK cells and committed progenitors within the lineage-negative, Sca-1, c-Kit+ (LK) population (fig. S1, A to D). Analysis of the HSC and MPP populations within the LSK gate revealed marked losses by day 2 after 5-FU exposure (fig. S1, E to G). MPPs showed the greatest proportional loss, and long-term HSCs (LT-HSCs) showed the least loss (fig. S1H), a result consistent with MPPs being highly proliferative and the LT-HSC population containing a greater proportion of quiescent cells. Therefore, a single injection of 5-FU (150 mg/kg) provided a clinically relevant model of myelosuppression.

We first determined whether quizartinib can protect MPPs from 5-FU by dosing 10-week-old B6 mice with vehicle or quizartinib (10 mg/kg), followed 24 hours later by a second dose of vehicle or quizartinib that coincided with an injection of 5-FU (150 mg/kg) (Fig. 2A). For this experiment, we restricted quizartinib to no more than two doses because dosing for longer reduces the numbers of hematopoietic progenitors (fig. S2, A to E), and this would be counterproductive to our aim of protecting these cells. The results from analyzing bone marrow 2 days after 5-FU administration showed that quizartinib provided marked protection to MPPs compared to vehicle-treated mice (Fig. 2A). Two doses of quizartinib 24 hours apart did not provide greater protection against 5-FU than a single dose given at the same time as 5-FU (Fig. 2A, compare hatched red bar with red bar). It was also evident that protection was lost when 5-FU was administered 24 hours after quizartinib (Fig. 2A, hatched white bar). This was unexpected because after 1 day, there is a significant decrease in the percentage of Ki-67–positive MPPs (Fig. 1, A and B; P < 0.01) and cycling MPPs (Fig. 1D; P < 0.05), which would be predicted to make them resistant to 5-FU. A possible explanation is that the cell cycle induction machinery may precede detectable increases in Ki-67 protein or DNA. Thus, although Ki-67 and DNA are useful indicators of cell cycle status, they may not precisely correlate with cells that have just entered the cell cycle (14). In summary, these results show that a single dose of quizartinib given at the same time as 5-FU can provide four- to fivefold protection to MPPs compared to vehicle-primed mice (Fig. 2A).

Fig. 2. A single priming dose of quizartinib protects hematopoietic progenitors from 5-FU–induced death.

(A) Numbers of MPPs from B6 mice dosed twice as indicated, with either vehicle or quizartinib (10 mg/kg) 24 hours apart, and an intraperitoneal injection of 5-FU (150 mg/kg) that was given at the time of the second dose. Bone marrow cells were analyzed by flow cytometry 2 days later to identify MPP numbers. (B) Experimental scheme of a 24-hour dosing time course before the mice were injected with 5-FU. Quizartinib or vehicle was administered to B6 mice at 0, 2, 6, 12, 18, or 24 hours before receiving an intraperitoneal (ip) injection of 5-FU. Bone marrow was examined 2 days later by flow cytometry. Fold change in the numbers of MPPs (C), ST-HSCs (D), LT-HSCs (E), and LK cells (F) from quizartinib-primed mice over a 24-hour time course compared to vehicle + 5-FU–treated mice, as represented by dashed line. (G) Experimental scheme for B6 mice dosed with vehicle or quizartinib 6 hours before an intraperitoneal injection of 5-FU. The bone marrow was examined at 2, 4, 6, 8, or 10 days after 5-FU treatment. (H) Representative flow cytometry profiles showing the lineage-negative bone marrow cells from either vehicle- or quizartinib-primed mice at the indicated time points. Gates shown identify the LK (left) and LSK (right) populations. Numbers of (I) LK cells, (J) LSK cells, (K) MPPs, (L) ST-HSCs, (M) LT-HSCs, and (N) total nucleated bone marrow (BM) cells over the time course. The blue shading shows the normal range of cell numbers from two femurs and two tibias from untreated age-matched B6 mice. Data are from three to six mice at each time point for each group, and the graphs are expressed as mean ± SEM. All statistics were calculated using Student’s t tests; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Optimization of quizartinib dosing protects HSCs and progenitor cells from 5-FU–induced cytotoxicity

To further characterize the effectiveness of quizartinib’s protection of MPPs and other bone marrow cells from 5-FU–induced death, we investigated single priming doses at time points between 0 and 24 hours (Fig. 2B). We found that quizartinib provided the highest level of protection to MPPs, short-term HSCs (ST-HSCs), LT-HSCs, and LK cells when it was administered between 2 and 12 hours before 5-FU (Fig. 2, C to F, and fig. S3A). The population that showed the greatest protection was MPPs, with ~10-fold more cells in the bone marrow after 2, 6, or 12 hours of quizartinib priming compared to vehicle-primed mice (Fig. 2C). Cells in the ST-HSC and LK gates also showed protection (Fig. 2, D and F), a finding that was not surprising because cells within these populations express a range of FLT3 levels (15, 16). However, the protection conferred to LT-HSCs (Fig. 2E) was not expected because they do not express FLT3 (17).

To further characterize quizartinib’s protection of HSCs and progenitor cells from 5-FU, we examined higher priming doses of 30 and 60 mg/kg. A dose of 30 mg/kg equates to 90 mg/m2, a dose well tolerated in leukemia patients (18, 19). We found that a priming dose of quizartinib (30 mg/kg) 6 hours before 5-FU provided slightly greater protection to HSCs and progenitor populations than 10 mg/kg (fig. S3, B to F), with a significant difference occurring within the ST-HSC population (fig. S3E; P < 0.05). In contrast, protection was not evident when the dose was increased to 60 mg/kg, possibly because this higher dose is myelosuppressive. In most subsequent experiments, mice were primed with quizartinib (30 mg/kg) before 5-FU treatment.

The protection of LT-HSCs, and possibly MPPs and LK cells, may partially involve quizartinib’s inhibitory activity toward c-Kit, albeit 10- to 20-fold less compared to FLT3 in cellular assays (20, 21). However, we previously found that dasatinib, a potent c-Kit inhibitor that also targets multiple tyrosine kinases including platelet-derived growth factor (PDGF) receptors and Src family kinases (22), does not induce quiescence of LT-HSCs or LSK or LK populations (23). The possible involvement of c-Kit was further examined by dosing mice with imatinib, an inhibitor of c-Kit, at a concentration of 100 nM (24), which also inhibits the PDGF and CSF-1 receptors (25, 26). Mice were dosed with imatinib (100 mg/kg) twice daily for 2 or 4 days before analysis. This approximates a dose of 500 mg twice daily for humans, which is slightly higher than the 800-mg daily dose used to treat patients with advanced gastrointestinal stromal tumors (27). We found that imatinib did not induce the quiescence of either LK or LSK cells (fig. S4, A and B). Furthermore, a priming dose of imatinib (100 mg/kg) either 6 or 12 hours before 5-FU administration did not protect LK or LSK cells from 5-FU–induced death (fig. S4C). Therefore, c-Kit inhibition in the absence of FLT3 inhibition is insufficient to induce quiescence or protect HSCs and progenitor cells from 5-FU.

FLT3 inhibitor crenolanib does not induce quiescence nor protect MPPs from 5-FU

Crenolanib is a potent next-generation FLT3 inhibitor that was originally developed as an inhibitor of PDGF receptors alpha and beta (28, 29). To test for quiescence induction in MPPs, we injected mice twice with vehicle or crenolanib (15 mg/kg) (30), with a 12-hour interval, and quantified Ki-67 the following day by flow cytometry. Rather than inducing quiescence, crenolanib caused a significant increase in large Ki-67+ MPPs compared to those from vehicle-treated mice (fig. S5, A and B; P < 0.01). A similar effect was observed when mice were treated for 2 days with two crenolanib injections per day (fig. S5, C and D; P < 0.0001), although the effectiveness of crenolanib treatment was evident by the reduced proportion of FLT3+ LSK cells (fig. S5E). The induction of proliferation of MPPs in crenolanib-treated mice was surprising, and the mechanism unknown; however, it was not due to an induction of FLT3 ligand (fig. S5F). Consistent with the proliferative activity induced by crenolanib, it did not protect LK, LSK, or MPP cells from 5-FU, which is in marked contrast to quizartinib (fig. S5, G and H). These findings demonstrate that crenolanib does not offer the same promise as quizartinib as a protector of the hematopoietic system against chemotherapy-induced myelosuppression.

Quizartinib priming enables a rapid recovery of bone marrow cellularity after 5-FU

To further investigate the protective ability of quizartinib priming, we examined the bone marrow from mice at time points 2 to 10 days after an injection with 5-FU (150 mg/kg) (Fig. 2G). The results showed that the protection conferred by quizartinib was most profound by day 6, with very large increases in LK and LSK cells compared to vehicle-primed mice (Fig. 2, H to J). Vehicle-primed mice showed little evidence of recovery of these populations at day 6. Within the LSK population, the recovery in quizartinib-primed mice was evident for MPPs, ST-HSCs, and LT-HSCs (Fig. 2, K to M). This resulted in the restoration of total bone marrow cell numbers by day 8, whereas the bone marrow from vehicle-primed mice remained markedly depleted (Fig. 2N). The normal range of cell numbers from two femurs and two tibias from untreated aged-matched B6 mice is depicted by blue shading to show the extent of the depletion and the kinetics of recovery within each population. The data on day 10 suggest that vehicle-primed mice had a greater requirement for a proliferative response within the LT-HSC population to contribute to bone marrow recovery (Fig. 2M).

Quizartinib priming protects the hematopoietic system from multiple rounds of 5-FU

Chemotherapy routinely involves multiple rounds of treatments, and therefore, we examined whether quizartinib protects mice from multiple rounds of 5-FU. We injected B6 mice with 5-FU (150 mg/kg) once every 7 or 10 days, 12 hours after priming with vehicle or quizartinib (10 mg/kg) (Fig. 3A). These cycles of 5-FU are potent regimens that cause bone marrow failure and lethality (31). We found that quizartinib priming enhanced survival of mice receiving 5-FU every 7 days, with a median survival of 17 days for vehicle-primed mice compared to 27 days for quizartinib-primed mice (Fig. 3B). Quizartinib priming before each 10-day cycle of 5-FU further enhanced survival, with only one of five mice dying over a period of 15 cycles (Fig. 3C). In contrast, all vehicle-primed mice succumbed to myelosuppression by the sixth cycle (Fig. 3C).

Fig. 3. Quizartinib priming before multiple 5-FU treatments enhances survival and blood recovery and prevents weight loss.

(A) Experimental scheme. B6 mice were primed with vehicle or quizartinib (10 mg/kg) 6 or 12 hours before an intraperitoneal injection of 5-FU (150 mg/kg). This process was repeated every 7 or 10 days, with mouse survival, blood counts, and weights monitored. (B and C) Kaplan-Meier survival plots of B6 mice primed with vehicle or quizartinib 12 hours before a 5-FU injection. The treatment cycles were repeated every 7 days (B) or 10 days (C), and the arrows indicate each treatment. Another cohort of B6 mice was treated with three 10-day cycles of 5-FU and primed with vehicle or quizartinib 6 hours before 5-FU treatment. The mice were bled every 5 days after the last injection to track numbers of (D) WBCs, (E) lymphocytes, (F) neutrophils, (G) RBCs, and (H) platelets. (I) The weights of mice over the time course were also recorded. One vehicle-primed mouse succumbed 8 days after the third 5-FU injection. Data are from five mice in each treatment group, with blood counts expressed as mean ± SEM. All statistics were calculated using Student’s t tests; *P < 0.05 and **P < 0.01.

To determine why quizartinib was less effective in protecting mice from 7-day cycles of 5-FU compared to 10-day cycles, we examined whether a single dose of quizartinib could induce quiescence in LSK cells from mice that had received 5-FU 7 or 10 days earlier. Ki-67 and cell cycle analysis revealed that quizartinib was unable to induce quiescence 7 days after 5-FU administration (fig. S6, A and B) and that this correlated with high amounts of FLT3 ligand (fig. S6C). This suggested that quizartinib could not overcome the greater proliferative activity induced by FLT3 ligand, and possibly other growth factors, at this time. In contrast, sensitivity to quizartinib-induced quiescence was evident 10 days after 5-FU (fig. S6, A and B), and this correlated with a return to normal amounts of FLT3 ligand (fig. S6C). Thus, a minimum 10-day cycle was deemed necessary for further investigations with serial treatments of quizartinib priming plus 5-FU.

Peripheral blood and body weights were also analyzed when mice received three 10-day cycles of 5-FU 12 hours after a priming dose of vehicle or quizartinib (10 mg/kg) (Fig. 3, D to I). Mice were bled every 5 days after the third cycle, and the measurement of total white blood cells (WBCs), lymphocytes, neutrophils, red blood cells (RBCs), and platelets showed that quizartinib-primed mice recovered more rapidly from this regimen (Fig. 3, D to H). The extent of the recovery was most notable 11 days after the third 5-FU injection, where all blood parameters, except for RBCs (Fig. 3G), had returned to pretreatment levels (Fig. 3, D to F and H). In contrast, vehicle-primed mice were yet to show a recovery of their blood cell numbers (Fig. 3, D to H). Furthermore, mice primed with quizartinib retained their body weight, whereas all vehicle-primed mice showed significant weight loss (Fig. 3I; P < 0.05 and P < 0.01), with one mouse succumbing to myelosuppression 8 days after the third injection.

Quizartinib priming protects against gemcitabine-induced myelosuppression

To investigate the possibility that quizartinib’s protective properties could be used with another chemotherapeutic compound, we investigated gemcitabine (Fig. 4A). Analysis of total bone marrow cells and lineage-negative cells 1 and 3 days after the third gemcitabine injection showed that quizartinib provided significant protection (Fig. 4, B and C; P < 0.01 on day 1 and P < 0.0001 on day 3). Mice primed with quizartinib had significantly higher numbers of LK cells (Fig. 4D; P < 0.001) and LSK cells (Fig. 4E; P < 0.0001 on day 1 and P < 0.001 on day 3) compared to vehicle-primed mice. Furthermore, by day 3, the LSK cells from quizartinib-primed mice had largely returned to a normal profile, with most of the population expressing high levels of c-Kit (Fig. 4F). For comparison, an LSK profile from an untreated mouse can be seen in fig. S1A. In contrast, at day 3, the vehicle-primed mice showed a high proportion of c-Kitlo cells within the expanded LSK gate used for this study (Fig. 4F). Further analysis of the LSK population showed that the protective effect was seen across MPPs, ST-HSCs, and LT-HSCs (Fig. 4, G to I). These findings demonstrate that quizartinib is effective at inhibiting myelosuppression induced by gemcitabine.

Fig. 4. Quizartinib priming protects hematopoietic progenitors and stem cells from gemcitabine-induced death.

(A) Experimental scheme. B6 mice were primed with vehicle or quizartinib at 0, 6, and 12 hours, and intraperitoneal injections of gemcitabine were administered at 6, 12, and 24 hours. Bone marrow was analyzed 1 and 3 days after the final gemcitabine injection. (B) Total nucleated bone marrow cell numbers and (C) lineage-negative bone marrow cell numbers at 1 and 3 days. Numbers of (D) LK cells and (E) LSK cells from mice at 1 and 3 days after the final gemcitabine injection. (F) Representative flow cytometry profiles from vehicle or quizartinib plus gemcitabine-treated mice at 1 or 3 days after treatment, displaying the expression of c-Kit and Sca-1 on lineage-negative bone marrow cells. Displayed are the gates that identify LK and LSK populations. Numbers of (G) MPPs, (H) ST-HSCs, and (I) LT-HSCs from mice at 1 and 3 days after the final gemcitabine injection. Data are from four mice at each time point for each group, and the graphs are expressed as mean ± SEM. All statistics were calculated using Student’s t tests; *P < 0 .05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Quizartinib preferentially protects WT MPPs from 5-FU cytotoxicity in WT/FLT3-ITD(F692L) chimeric mice

FLT3–internal tandem duplication (ITD) mutations are found in about 25% of human acute myeloid leukemias (AMLs), and knock-in mice with an ITD mutation develop a myeloproliferative disease characterized by an expansion of highly proliferative MPPs (32). These mice carry a germline F692L mutation in FLT3 that confers resistance to quizartinib (33). The origin of this mutation is not known, but it corresponds to F691L mutations found in FLT3-ITD+ AML patients who develop resistance to quizartinib (34). The presence of this mutation provided an in vivo model to test whether quizartinib preferentially protects WT MPPs from 5-FU toxicity, whereas the leukemic cells continue to cycle and remain susceptible. Resistance to quizartinib was confirmed by dosing FLT3-ITD(F692L) mice with quizartinib or vehicle over 2 days. No significant induction of quiescence was evident in the mutant MPPs (Fig. 5A). This contrasts with the effect of quizartinib on MPPs from WT B6 mice (Fig. 1B). The effect of quizartinib on FLT3-ITD(F692L) MPPs was further tested in an 18-hour time course of vehicle or quizartinib priming followed by 5-FU treatment, as described for B6 mice (Fig. 2B). Examination of bone marrow cells by flow cytometry 2 days after 5-FU treatment showed that quizartinib was unable to protect FLT3-ITD(F692L) MPPs from 5-FU cytotoxicity (Fig. 5B). The changes in the numbers of surviving MPPs from mice primed with quizartinib were normalized to vehicle-primed mice (shown by dashed line) at each time point (Fig. 5B). This lack of protection is in marked contrast to MPPs from WT B6 mice, where quizartinib priming provided marked protection from 5-FU cytotoxicity (Fig. 2C).

Fig. 5. Hematopoietic progenitors from FLT3-ITD(F692L) mutant mice dosed with quizartinib are not induced into quiescence or protected from 5-FU cytotoxicity.

(A) Percentage of Ki-67+ MPPs from FLT3-ITD(F692L)–transplanted mice dosed daily for 1 or 2 days with vehicle or quizartinib (10 mg/kg). n = 5 per group. (B) Numbers of surviving MPP cells from FLT3-ITD(F692L)–transplanted mice primed with quizartinib (10 mg/kg) at 0, 2, 6, 12, or 18 hours before an injection of 5-FU normalized to vehicle-primed mice at each time point, as shown by the dashed line. n = 5 per time point. (C) Representative flow cytometry profiles of CD11b+ cells from the blood of WT:FLT3-ITD(F692L) chimeric mice, showing the percentages of CD45.1 (WT) versus CD45.2 [FLT3-ITD(F692L)] cells. The mice were bled at the indicated times after the first treatment, and the arrows represent the administration of quizartinib (10 mg/kg; 12-hour priming) + 5-FU (150 mg/kg) at days 0 and 10. (D) Proportions of WT CD11b+ blood cells over the time course (n = 5), where the arrows represent the two treatment cycles. PreB, prebleed. (E) Kaplan-Meier survival curves of WT:FLT3-ITD(F692L) chimeric mice primed with vehicle or quizartinib before 5-FU administration (n = 10 and 5, respectively).

Chimeric mice were then generated by repopulating lethally irradiated B6.CD45.1 mice with a mix of WT (B6.CD45.1) bone marrow and FLT3-ITD(F692L) (CD45.2) spleen cells. This model mimics the coexistence of normal and leukemic hematopoiesis in patients. Spleen cells from FLT3-ITD mice were used because they are a richer source of LT-HSCs than bone marrow (35). Eleven weeks after transplantation, the chimeric mice were bled, and the proportions of CD11b+ cells expressing CD45.1 or CD45.2 were determined. CD11b+ myeloid cells provide a measure of new blood production because their turnover is rapid (36, 37) and therefore provide an indication of the genotype of the active hematopoietic progenitors. Before treatment, the FLT3-ITD(F692L) cells markedly outcompeted WT cells (Fig. 5, C and D), with some mice having as little as 1% contribution from WT CD11b+ cells in their blood. The mice were then subjected to two rounds of quizartinib priming plus 5-FU 10 days apart, and blood was analyzed 7 days after each treatment. After the two rounds, there was a striking reversal, whereby WT CD11b+ cells were now markedly in excess compared to FLT3-ITD(F692L) cells (Fig. 5, C and D). Furthermore, the predominance of WT CD11b+ cells was still evident 31 days after the treatment commenced. However, at 60 and 105 days after the first treatment, the FLT3-ITD(F692L) cells reemerged, indicating that further rounds of treatment would be necessary to again suppress the inherent dominance of these cells (Fig. 5, C and D). The findings from this model provide strong evidence that quizartinib can preferentially protect WT hematopoiesis from chemotherapy.

The analysis of recipient irradiated WT/FLT3-ITD(F692L) chimeric mice also involved a cohort dosed with vehicle 12 hours before 5-FU administration. All mice in this cohort died within 11 days of the second treatment, whereas all quizartinib-primed mice remained active and healthy, providing further evidence of the marked protection provided by quizartinib (Fig. 5E). Prior irradiation markedly affects bone marrow recovery after 5-FU treatment (38), and this is illustrated by comparing Fig. 5E with Fig. 3C, where nonirradiated B6 mice receiving multiple 10-day cycles of vehicle and 5-FU showed a median survival of 46 days.

A combination of quizartinib priming and 5-FU is an effective treatment for FLT3-ITD(F692L)/NPM1c AML

The results from treating chimeric WT/FLT3-ITD(F692L) mice by a combination of quizartinib priming and 5-FU suggested that this approach may also provide an effective treatment for more aggressive leukemias. We therefore crossed FLT3-ITD(F692L) mice with nucleophosmin knock-in mutant mice. Mutations in the nucleophosmin gene (NPM1) are among the most frequent genetic lesions in human AML, and mice with a humanized NPM1c conditional knock-in mutation develop a delayed-onset AML in about one-third of cases (39). However, when they are crossed with FLT3-ITD mice, the double-mutant mice rapidly develop an aggressive AML (40, 41).

We first tested whether quizartinib affected the proliferative activity of MPPs from FLT3-ITD(F692L)/NPM1c mice. Mice were dosed once daily for 2 days with vehicle or quizartinib (30 mg/kg), and MPPs were analyzed for Ki-67 expression 24 hours after the second dose. Consistent with findings in the FLT3-ITD(F692L) mice, no induction of quiescence was evident in MPPs from the double-mutant mice (Fig. 6A). To test quizartinib as a monotherapy for FLT3-ITD(F692L)/NPM1c–driven AML, we transplanted spleen cells from a leukemic FLT3-ITD(F692L)/NPM1c mouse into nonirradiated B6 syngeneic recipients. Unlike irradiated recipient mice, this model has the advantage that it does not eliminate WT hematopoietic progenitors and stem cells. Treatment commenced 8 days after transplantation and consisted of daily dosing with vehicle or quizartinib (10 mg/kg). The results showed that quizartinib-dosed mice succumbed faster than vehicle-dosed mice, indicating that quizartinib, as a monotherapy, is an ineffective treatment for this leukemia (Fig. 6B). This result provided further evidence of the inherent resistance of FLT3-ITD(F691L)/NPM1c AML cells to quizartinib. The reason why quizartinib-dosed mice succumbed faster than vehicle-dosed mice may be the suppression of WT hematopoiesis, which could adversely affect the ability of the mice to combat leukemia progression. In humans, quizartinib-induced myelosuppression is a major dose-limiting toxicity (42)

Fig. 6. Quizartinib priming in combination with 5-FU provides an effective treatment for FLT3-ITD(F692L)/NPM1c–driven AML.

(A) Percentage of Ki-67+ MPPs from naturally bred double-mutant FLT3-ITD(F692L)/NPM1c mice dosed for 2 days with vehicle or quizartinib (30 mg/kg). Bone marrow was analyzed 24 hours after the second dose. n = 3 for each group. (B) Kaplan-Meier survival curves of B6 mice transplanted with FLT3-ITD(F692L)/NPM1c AML cells and dosed daily with either vehicle or quizartinib (10 mg/kg). (C) A second cohort of FLT3-ITD(F692L)/NPM1c–transplanted mice was established to investigate the effectiveness of the quizartinib priming plus 5-FU protocol versus induction chemotherapy of Ara-C and Dox. Displayed are pretreatment percentages of CD45.2+ CD11b+ blood cells and total WBC counts for each test group (n = 5). (D) Kaplan-Meier survival curves of FLT3-ITD(F692L)/NPM1c–transplanted mice from the three test groups. (E) WBC counts and (F) spleen weights from untreated and Ara-C + Dox groups at the time of death. (G) Percentages of CD45.2+ CD11b+ WBCs and (H) total WBC counts from individual mice (M1 to M5) treated with quizartinib + 5-FU over the 176-day time course. (I) Flow cytometry profiles displaying CD45.1+ (WT) versus CD45.2+ [FLT3-ITD(F692L)/NPM1c] LSK cells from the bone marrow of the four surviving quizartinib + 5-FU–treated mice culled at day 176.

To determine whether the combination of quizartinib priming and 5-FU can provide an effective treatment, we transplanted 15 nonirradiated B6.CD45.1 mice with 3 × 105 spleen cells each from a FLT3-ITD(F691L)/NPM1c mouse that succumbed to AML at 6 weeks of age. The mice were bled 15 days after transplantation, and all mice showed evidence of CD45.2+ AML cells in the periphery; however, the WBC counts were not yet markedly elevated (Fig. 6C). Treatment commenced on day 16 and consisted of (i) no treatment, (ii) 10-day cycles with a priming dose of quizartinib (30 mg/kg) 6 hours before 5-FU (150 mg/kg) in each cycle, or (iii) conventional induction therapy with cytarabine (Ara-C) + doxorubicin (Dox). The induction therapy involved intraperitoneal delivery of Ara-C at 50 mg/kg and intravenous delivery of Dox at 1.5 mg/kg for 3 days, followed by Ara-C alone at 50 mg/kg on days 4 and 5 (5 + 3 therapy) (43).

All mice in the untreated group succumbed to AML within 30 days of transplantation (Fig. 6D), with very high WBC counts and enlarged spleens (Fig. 6, E and F). The Ara-C + Dox cohorts were treated twice, on days 16 to 21 and 36 to 41, and all mice succumbed by day 56 after transplantation (Fig. 6D), with very high WBC counts and enlarged spleens (Fig. 6, E and F). In contrast, four of the five mice treated with nine cycles of quizartinib priming and 5-FU remained healthy for 176 days after transplantation and 80 days after cessation of treatment (Fig. 6D). Analysis of the blood revealed no detectable CD45.2+ AML cells when the mice were last bled on day 160 (Fig. 6G). Furthermore, all four mice (M1 to M4) showed WBC counts within the normal range of 5 × 109 to 14 × 109 cells/liter (Fig. 6H). The mouse that died 121 days after transplantation (M5) appears to have developed resistance to 5-FU because the disease returned while the treatment was ongoing (Fig. 6G). When the four surviving mice were culled 176 days after transplantation, no AML cells were detected in the bone marrow, and none were detected in the LSK population, which is likely to harbor leukemic stem cells (Fig. 6I).

A combination of quizartinib priming and 5-FU provides an effective treatment for NPM1c/NrasG12D AML

To examine an additional model of AML, where the leukemic cells were unlikely to be sensitive to quizartinib, we crossed NPM1c mutant mice with NrasG12D mutant mice (44). Cohorts were established by transplanting spleen cells from NPM1c/NrasG12D leukemic mice into nonirradiated B6.CD45.1 recipients. We first examined whether NPM1c/NrasG12D AML cells are affected by quizartinib. The emergence of AML cells in the peripheral blood was confirmed 18 days after transplantation, and the mice were then dosed once daily for 2 days with vehicle or quizartinib (30 mg/kg). Bone marrow from the chimeric mice was analyzed by flow cytometry 24 hours after the second dose to determine Ki-67 expression in the NPM1c/NrasG12D AML and WT LSK populations. Gating on the CD45.2+ AML cells showed that quizartinib did not reduce the proportion of Ki-67+ cells, indicating that these cells continued to cycle and would therefore not be expected to be protected from chemotherapy (Fig. 7A). In contrast, WT LSK (CD45.1+) cells in the quizartinib-dosed mice showed a significant reduction in the proportion of Ki-67+ cells (Fig. 7A; P < 0.001). The gating strategies and representative flow cytometry profiles are shown in fig. S7A.

Fig. 7. Quizartinib priming in combination with 5-FU provides an effective treatment for NPM1c/NrasG12D-driven AML.

(A) Percentages of Ki-67+ cells within the CD45.2+ (AML) and CD45.1+ (WT) LSK populations from NPM1c/NrasG12D-transplanted mice dosed for 2 days with vehicle or quizartinib (30 mg/kg). n = 5 for each group. ***P < 0.001. (B) A cohort of NPM1c/NrasG12D-transplanted mice was established to investigate the effectiveness of the quizartinib priming plus 5-FU protocol versus standard induction chemotherapy of Ara-C + Dox. Displayed are pretreatment percentages of CD45.2+ CD11b+ blood cells and WBC counts for each test group (n = 5). (C) Kaplan-Meier survival curves of NPM1c/NrasG12D-transplanted mice from the three test groups. (D) WBC counts and (E) spleen weights from untreated and Ara-C + Dox groups at the time of death. (F) Percentages of CD45.2+ CD11b+ WBCs and (G) WBC counts from individual mice (M1 to M5) treated with quizartinib + 5-FU over the time course. (H) A second cohort of NPM1c/NrasG12D-transplanted B6.CD45.1 mice was established to repeat the examination of quizartinib priming plus 5-FU versus induction chemotherapy. Displayed are pretreatment percentages of CD45.2+ CD11b+ blood cells and total WBC counts for each test group (n = 5). (I) Kaplan-Meier survival curves of NPM1c/NrasG12D-transplanted mice from the three test groups. (J) WBC counts and (K) spleen weights from untreated and Ara-C + Dox groups at the time of death. (L) Percentages of CD45.2+ CD11b+ WBCs and (M) WBC counts from individual mice (M1 to M5) treated with quizartinib and 5-FU over the time course.

Two cohorts of mice transplanted with spleen cells from NPM1c/NrasG12D mice were established to test the relative effectiveness of treating the mice with quizartinib priming plus 5-FU or induction chemotherapy. The first cohort was bled 12 days after transplantation, and all mice showed detectable CD45.2+ AML cells; however, WBC numbers were not markedly elevated at this time (Fig. 7B). The mice were divided into three groups that were (i) left untreated, (ii) treated with 10-day cycles consisting of a priming dose of quizartinib (30 mg/kg) 6 hours before 5-FU (150 mg/kg), or (iii) treated with Ara-C + Dox induction therapy, as described above. Treatment commenced 15 days after transplantation, and all five untreated mice succumbed by 32 days (Fig. 7C). Induction therapy provided minimal benefit, with all mice succumbing by 35 days after transplantation (Fig. 7C). At the time of death, both of these groups had high WBC counts and enlarged spleens (Fig. 7, D and E), and the blood films showed a high proportion of undifferentiated blasts (fig. S7B).

Mice treated with quizartinib priming and 5-FU exhibited a profound improvement in survival (Fig. 7C). These mice initially received four 10-day cycles, starting on days 15, 25, 35, and 45 after transplantation (Fig. 7C). Blood analysis on day 53 showed minimal or undetectable numbers of CD45.2+ AML cells (Fig. 7F and fig. S8A), and WBC counts were within, or slightly lower than, the normal range of 5 × 109/liter to 10 × 109/liter for B6 mice (Fig. 7G). About 1 month after the cessation of treatment (day 81), four of the five mice displayed detectable CD45.2+ AML cells in the blood (0.93, 49, 0.22, and 0.25% of CD11b+ cells expressing CD45.2 for M1, M3, M4, and M5, respectively) (Fig. 7F and fig. S8). The following day, treatment was recommenced in an attempt to curb resurgence of the disease. Four additional cycles of treatment reduced the AML cells to undetectable levels in all five mice (Fig. 7F and fig. S8A). As before, AML cells again became detectable in the blood 1 month after the cessation of treatment (day 146), but this time, the mice were left untreated to monitor disease progression (Fig. 7F and fig. S8A). One mouse (M3) was found dead at day 196 and could not be analyzed, and another (M2) was culled at day 197 because of weight loss due to feeding difficulties but displayed no evidence of AML (Fig. 7, F and G). The three surviving mice remained active and healthy, although they had showed high proportions of CD45.2+ myeloid cells in their blood since day 183 (Fig. 7F). The mice were culled at day 214 to conclude the experiment, and two (M1 and M4) had increased WBC counts from day 197 (Fig. 7G), indicating disease progression, albeit at a slow rate. One mouse (M5) showed a decrease in WBC counts over this time (Fig. 7G). Furthermore, blood films from these three mice revealed very few undifferentiated blasts (fig. S8B) compared to the untreated or induction therapy–treated groups (fig. S7B).

A second cohort of NPM1c/N-RasG12D-transplanted mice was established, and when the mice were bled 14 days after transplantation, AML cells were evident in all 15 mice (Fig. 7H). The percentages of AML cells in all but one mouse were markedly higher than in the first cohort, as were the WBC counts (Fig. 7H). With the disease well established, treatment commenced the following day. The untreated mice again succumbed rapidly, with all mice dying by day 29 after transplantation (Fig. 7I). The Ara-C + Dox induction therapy again provided minimal benefit, with all mice succumbing by 31 days after transplantation (Fig. 7I). At death, both groups had markedly increased WBC counts and enlarged spleens (Fig. 7, J and K).

Consistent with the first cohort, 10-day cycles of quizartinib priming in combination with 5-FU proved to be a very effective treatment, leading to an increase in survival (Fig. 7I). When the mice were bled 33 days after transplantation, all five showed a marked reduction in the proportion of AML cells, with three showing undetectable numbers (Fig. 7L and fig. S8C). Similarly, the WBC counts were markedly reduced after two 10-day treatment cycles (Fig. 7M). Because the disease returned after four cycles of treatment in the first experiment examining NPM1c/NrasG12D-transplanted mice, we continued the treatment for six cycles. Thereafter, the mice were monitored for AML cells and WBC counts. The mice remained disease-free when bled at 78 days after transplantation; however, a small amount of AML cells in the blood was evident by day 91 (fig. S8C). From this point, the disease progressed rapidly, and when the mice were bled on day 113, all five showed greater than 96% CD45.2+ myeloid cells (Fig. 7, L and M, and fig. S8C). One mouse required culling on this day, and the remaining four were culled over days 118 to 120 (Fig. 7I). Blood films from these mice displayed fewer blasts compared to the untreated or induction therapy–treated groups (compare figs. S8D and S7B); however, the reduction in severity of this disease was not as profound as in the first NPM1c/NrasG12D cohort (compare fig. S8, B and D). This could be due to the additional rounds of treatment that the first cohort received, and thus, it is likely that these mice would have survived longer with further treatment.

DISCUSSION

Small-molecule kinase inhibitors, such as quizartinib, were developed to specifically target tumor cells with constitutively activated kinases; however, the effects of these inhibitors upon WT cells are often less well characterized. Here, we reveal that inhibition of FLT3 with quizartinib provides an effective approach to protect WT hematopoietic progenitor and stem cells from chemotherapy-induced death, which, in turn, prevents the development of lethal myelosuppression. Exploitation of this phenomenon in a clinical setting has the potential to improve the health and quality of life of cancer patients receiving chemotherapy, as well as greatly reduce hospitalization time and health care costs.

Myelosuppression is one of the predominant dose-limiting side effects of chemotherapy, causing marked patient morbidity and often impeding optimal cytotoxic drug usage. Results from our study support the concept that protection of the hematopoietic system is possible through induction of quiescence in progenitors and stem cells, a finding recently described through administration of an inhibitor of cyclin-dependent kinases 4 and 6 (CDK4/6) known as G1T28 (45, 46). The ability of G1T28 to induce quiescence holds promise for inhibiting myelosuppression in a subset of patients with tumors that are functionally independent of CDK4/6.

A major advantage of quizartinib’s protection of the hematopoietic system is that it alleviates multilineage myelosuppression, and thus, it is more broadly effective than current measures, such as the commonly implemented strategy of postchemotherapy G-CSF treatment to combat neutropenia (47, 48). A recent phase 1 trial described quizartinib as a safe single-dose agent with no serious adverse events (49). Therefore, investigations to examine quizartinib priming as a strategy for combating myelosuppression in humans are likely to be feasible and safe. Furthermore, the extended expression profile of FLT3 in humans, which is expressed on long-term repopulating HSCs, MPPs, multilymphoid progenitors, common myeloid progenitors, and granulocyte and monocyte progenitors, suggests that quizartinib priming may be an effective approach for protecting human bone marrow from chemotherapy-induced myelosuppresion (50, 51). However, the first steps toward translating from mouse models will require trials that identify a quizartinib dose, or doses, that achieve protection of the bone marrow but do not affect the ability of chemotherapy to kill tumor cells.

The concept of cellular protection mediated by small-molecule inhibitors that induce transient quiescence could also be extended, with the possibility of protecting other cell types that are eliminated by chemotherapy, such as intestinal epithelial cells. The potential for small-molecule inhibitors to mediate a selective and transient protection of proliferating nontumor cells could markedly enhance their current clinical use.

The inability of crenolanib, another potent FLT3 inhibitor, to induce quiescence of MPPs and provide protection from 5-FU–induced death was unexpected. It is possible that crenolanib’s potent inhibition of PDGF receptors, which are expressed on bone marrow stromal cells, could be a factor. Furthermore, crenolanib could be affecting different subpopulations of FLT3+ cells because it functions as a type I kinase inhibitor, which binds to an active kinase confirmation, whereas quizartinib, being a type II inhibitor, binds to an inactive kinase conformation (21, 52, 53). In addition, c-Kit inhibition is unlikely to be a factor mediating the contrasting effects because both show similar inhibitory effects in medium and plasma (52, 54).

Our study of quizartinib priming of mice treated with 5-FU found this to be a highly effective treatment for two mouse models of AML where quizartinib did not affect the proliferative activity of the AML cells. These models indicated that quizartinib priming should be beneficial for treating tumors where FLT3 signaling is not a driver. However, it is possible that this approach could be detrimental for treating disorders such as FLT3-ITD+ AMLs. This prediction is supported by an in vitro study that found that pretreatment of FLT3-ITD mutant MV4-11 AML cells with the FLT3 inhibitor CEP-701 reduced the number of cycling cells, and as a consequence, the cytotoxic effects of cell cycle–dependent chemotherapeutic agents were less pronounced (55).

The role of 5-FU in this study is also a key point because it was the potent cytotoxic effect of 5-FU against leukemic cells that helped facilitate the benefits observed in the quizartinib-primed treatment groups. The effectiveness of bolus 10-day cycle injections of 5-FU as an antileukemic agent was surprising, given that this compound is not routinely used to treat leukemia. However, the effectiveness of 5-FU therapy was undeniable, with most mice displaying an absence of detectable AML cells in the blood. This contrasted with conventional induction therapy of Ara-C and Dox, which produced minimal improvements in survival. Thus, the promising 5-FU treatment results encourage investigation into the use of 5-FU in combination with quizartinib priming for the treatment of AML patients without FLT3 mutations.

The protection of the hematopoietic system afforded by quizartinib priming has the potential to improve the treatment of many cancers beyond the non–FLT3-dependent AMLs described here. It is likely that this approach will be more applicable for treating solid tumors. This is because most, if not all, solid tumors are functionally independent of FLT3 and therefore should not be induced into quiescence by quizartinib. The restricted cellular expression of FLT3 indicates that quizartinib priming could potentially be used for most tumor types. To further expand the utility of this strategy, it will also be important to determine whether other chemotherapeutic compounds, in addition to 5-FU and gemcitabine, can be used in combination with quizartinib to protect the hematopoietic system. It is likely that quizartinib’s protection will be most effective with cell cycle–specific chemotherapeutic agents, such as the antimetabolites described here, and less applicable for drugs that are cytotoxic for all phases of the cell cycle.

In addition to its activity against FLT3, quizartinib is also a potent inhibitor of the drug transporter ABCG2 (adenosine 5′-triphosphate–binding cassette subfamily G member 2; breast cancer–resistant protein) (56). This property is pertinent for 5-FU because it is an ABCG2 substrate (57), raising the possibility that quizartinib dosing may result in higher 5-FU concentrations in hematopoietic progenitors and stem cells. However, despite this, quizartinib provides marked protection to these cells. This highlights the importance of 5-FU’s cell cycle specificity because higher intracellular concentrations of 5-FU would be ineffective when the cells are quiescent. Furthermore, quizartinib’s inhibition of ABCG2 may enhance the efficacy of 5-FU toward cycling tumor cells, thereby providing an additional benefit of quizartinib priming.

In summary, the ability of quizartinib to inhibit 5-FU– and gemcitabine-induced myelosuppression may allow for more intensive chemotherapy of resistant tumors and also allow for elderly patients, who are more susceptible to myelosuppression, to receive optimal doses. The protection of the hematopoietic system may also aid the recovery of immune cell function. Recent evidence indicates that chemotherapy plays an important part in antitumor immune responses (58, 59), and therefore, quizartinib’s protection of the hematopoietic system may enhance the ability of the host to respond more rapidly and effectively to tumor antigens. Furthermore, given the growing importance of immunotherapies for cancer treatment, the restoration of the hematopoietic system by quizartinib priming may prove to be a beneficial addition when combining chemotherapy with immunomodulating therapies.

MATERIALS AND METHODS

Study design

This study was designed to determine the benefits of priming mice with the FLT3 inhibitor quizartinib before their exposure to chemotherapies that induce myelosuppression. To determine the optimal quizartinib doses and time points that induced quiescence of hematopoietic progenitors and therefore may provide protection to these cells against the cytotoxic effects of 5-FU and gemcitabine, we used 8- to 10-week-old male B6 mice. The differential effects on bone marrow and blood cells from mice primed with either vehicle or quizartinib before chemotherapy were assessed by flow cytometry using well-characterized antibodies that identify progenitor and stem cell populations in addition to Ki-67 antibodies for determining their proliferative activity.

From these initial results, we assessed whether the quizartinib priming protocol in combination with chemotherapy could provide an effective anticancer treatment. For this, we used two models of AML. It was essential that the leukemic cells from these AML models could be transplanted into nonirradiated congenic C57BL/6;CD45.1 mice to ensure that recipient hematopoiesis was maintained and therefore could be protected by quizartinib. The leukemic cells to establish these models were from FLT3-ITD(F692L)/NPM1c and NPM1c/NrasG12D mutant mice, the details of which are described in Supplementary Materials and Methods. The dosing schedule for the treatments was based on the optimization studies (Figs. 1 to 3), and a standard induction chemotherapy protocol was based on published literature (43). Mice were grouped after initial peripheral blood analysis to establish an equivalent and comparable level of disease across treatments. Survival endpoints were reached if the mice displayed excessive disease or treatment burden (according to Association for Assessment and Accreditation of Laboratory Animal Care regulations). It was determined that between 3 and 10 mice at each time point was a sufficient sample size given the precise scope of the study and the robust effect elicited by the treatment. The study was not blinded. Additional information regarding treatment timing, sample sizes, flow cytometry and Hemavet analyses, and tumor endpoints is provided in Results, figure legends, and Supplementary Materials and Methods.

Statistical analyses

To assess significance, we used unpaired two-sided Student’s t tests (GraphPad Prism 5 software). P values less than 0.05 were considered statistically significant. All statistical data are presented as means ± SE. Kaplan-Meier plots were generated using GraphPad Prism 5 software.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/9/402/eaam8060/DC1

Materials and Methods

Fig. S1. Hematopoietic progenitors and stem cells are markedly depleted after a single intraperitoneal injection of 5-FU.

Fig. S2. Daily dosing with quizartinib for 4 days markedly reduces the number of hematopoietic progenitors.

Fig. S3. Optimizing quizartinib dosing protects HSCs and progenitor cells from 5-FU–induced cytotoxicity.

Fig. S4. Imatinib does not induce quiescence of LK or LSK cells nor provide protection from 5-FU cytotoxicity.

Fig. S5. Crenolanib does not induce quiescence of MPPs nor provide protection from 5-FU cytotoxicity.

Fig. S6. A quizartinib dose 10 days, but not 7 days, after 5-FU is effective in inducing quiescence in LSK cells.

Fig. S7. Quizartinib does not induce quiescence of NPM1c/NrasG12D AML cells.

Fig. S8. Analysis of blood from NPM1c/NrasG12D mice shows the development of AML in mice treated with 10-day cycles of quizartinib priming followed by 5-FU.

REFERENCES AND NOTES

  1. Acknowledgments: We thank the support of the staff from the Sir Charles Gairdner Hospital Cancer Centre for mouse irradiation and the Sir Charles Gairdner Hospital Pharmacy for the provision of chemotherapeutic drugs. We also thank the Experimental Immunology Group at the Lions Eye Institute for sharing their flow cytometry facility. Funding: This work was supported by the National Health and Medical Research Council Project (grants 572516, 1057762, and 1101318), the Medical and Health Research Infrastructure Fund, and Ph.D. scholarships (to S.J.T.) from the L.T. Thean Memorial Medical Research Scholarship and the Cancer Council Western Australia. G.S.V. and O.M.D. are funded by a Wellcome Trust Senior Fellowship in Clinical Science (WT095663MA), and work in G.S.V.’s laboratory is also funded by the Wellcome Trust Sanger Institute (WT098051). Author contributions: S.J.T., J.M.D., and W.Y.L. conceived, designed, and performed the experiments. E.J.D. and C.A.R. contributed to the experiments. S.A.D. genotyped and maintained our mouse colonies. O.M.D. and G.S.V. provided the NPM1flox-cA;Mx1-Cre+ mutant mice and AML cells from NPM1c/NRasG12D mice and contributed to the editing of the manuscript. C.S.G. facilitated the provision of mice, analyzed the blood films, and contributed to the editing of the manuscript. S.J.T. and W.Y.L. wrote the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article