Dasatinib does not exacerbate dexamethasone-induced osteonecrosis in murine models of acute lymphoblastic leukemia therapy

Abstract

Introduction:

There are clinical reports that the incorporation of dasatinib may increase the frequency of osteonecrosis in acute lymphoblastic leukemia (ALL) treatment regimens. No rigorous testing of this hypothesis is available to guide clinicians.

Methods:

We tested whether oral dasatinib increased the frequency of dexamethasone-induced osteonecrosis in a murine model and tested its effects on dexamethasone’s antileukemic efficacy in a murine BCR-ABL⁺ model of ALL.

Results:

Dasatinib did not change the frequency of osteonecrosis (p = 0.99) nor of arteriopathy (p = 0.36) in dexamethasone-treated mice when given at dosages that achieved clinically relevant steady-state dasatinib plasma concentrations of 53.1 ng/mL (95% CI 43.5–57.3 ng/mL). These dasatinib exposures were not associated with increased dexamethasone plasma exposure in non-leukemia bearing mice.These same dosages were not associated with any decrement in antileukemic efficacy of dexamethasone in a responsive BCR-ABL⁺ model of ALL.

Conclusions:

Based on the results of our preclinical murine studieswe conclude that dasatinib is unlikely to increase the osteonecrotic effects of dexamethasone in ALL regimens.

Introduction:

Dasatinib is a second-generation tyrosine kinase inhibitor (TKI) with activity against both ABL and SRC family kinases. Recent studies show dasatinib is superior to the first generation TKI imatinib for treatment of BCR-ABL1-fusion and BCR-ABL1-like acute lymphoblastic leukemia (ALL) due to its increased potency¹,efficacy against CNS leukemia², activity against imatinib-resistant BCR-ABL1 (except T315I mutation), and improved survival in a randomized clinical trial for children with newly diagnosed BCR-ABL1-fusion ALL³.

As dasatinib has entered wider use, it is critical to understand how it may contribute to the short- and long-term toxicities of ALL therapy. In addition to myelosuppression, dasatinib has also been reported to be associated with the development of osteonecrosis in patients with chronic myeloid leukemia (CML)⁴. The addition of TKI to multiagent chemotherapy was also associated with a high rate of osteonecrosis in patients withALL⁵. In a preliminary report of 270 patients on AIEOP-BFM ALL 2009 protocol, 75% (3/4) of BCR-ABL1+ patients on TKI therapy developed osteonecrosis compared to an overall frequency of just 16%⁵.

Moreover, dasatinib is a substrate for CYP3A metabolism⁶, as is dexamethasone, and dexamethasone can also induce CYP3A as well as common transporters⁷; thus, there is the possibility of pharmacokinetically-based drug interactions between dasatinib and dexamethasone.

We evaluated the impact of oral dasatinib given at clinically relevant concentrations on the frequency of dexamethasone-induced osteonecrosis and on antileukemic efficacy inmurine models.

Methods:

Treatment

All experiments were approved by the Institutional Animal Care and Use Committee of St. Jude Children’s Research Hospital (SJCRH; Memphis, TN): protocol numbers 423–100428 and 465–100549.

Dexamethasone (dexamethasonesodium phosphate solution; Fresenius Kabi, Lake Zurich, IL) was given in drinking water, herein as “dexamethasone water” (2 mg/L in osteonecrosis model and 4 mg/L in BCR-ABL⁺ ALL model)^8–10.All mice received prophylactic antibiotics in drinking water (“base-water”): continuous tetracycline (1 g/L; Sigma-Aldrich, St. Louis, MO) and intermittent sulfamethoxazole/trimethoprim oral suspension (600/120 mg/L for 3.5 days/week; Aurobindo Pharma, USA, Inc., Dayton, NJ).^8–12

The folic acid deficient diet (0.2 ppm folic acid; TestDiet, Richmond, IN) was the “base-diet”⁸. Dasatinib (dasatinib, free base, >99%; LC Laboratories, Woburn, MA) was added to the base-diet for a final concentration of 0.0172% dasatinib (w/w) (“dasatinib-supplemented diet”; TestDiet, Richmond, IN), which equates to ~40 mg/kg dasatinib daily¹³. This dose was chosen to test the upper-tolerable exposure to dasatinib while still being within clinically relevant exposure¹⁴. Mouse feed was changed 1–2 times/week; diets were stored at 4 C for up to three months.

Mice were randomized to one of four treatment groups: control (base-water and base-diet); dasatinib-only (base-water and dasatinib-supplemented diet); dexamethasone-only (dexamethasone-water and base-diet); or dexamethasone + dasatinib (dexamethasone-water and dasatinib-supplemented diet).

Osteonecrosis model

The protocol was modified from that previously described^{8,9,11,12,15,16}. At postnatal days 26 to 28, male Balb/cJ mice (bred in house) were stratifiedby body weightand randomized to 6-weeks of treatment.Osteonecrosis and epiphyseal arteriopathy were evaluated and scored in both distal femurs by a board-certified veterinary pathologist (L.J.J.) blinded to treatment assignment, as described^11,15.Mice that succumbed to therapy-induced toxicity prior to completion of 6-weeks of therapy were not evaluated for osteonecrosis or arteriopathy. Legs in which the epiphyseal arteriole was not fully visualized were inevaluable for arteriopathy.

BCR-ABL⁺ ALL model

The BCR-ABL⁺ (p185⁺,Arf^−/−) luciferase-positive cell line (BCR-ABL⁺) was generated as previously described^10,17–19.Eight-week-old female C57Bl/6J, Arf-wildtype mice (Jackson Laboratory, ME) received intravenous injections of 2,000 BCR-ABL⁺ cells. Bioluminescent imaging was performed weekly¹⁰, to monitor leukemic burden. At day 3, mice were stratified by luminescent signal and body weight to treatment (n=10/treatment group). Treatment ended at day 28, after which mice received base-diet/base-water (control-group conditions) until a humane disease related endpoint or the end of study at day 59.Disease related endpoints included ventral luminescence near saturation (pre-determined ventral-luminescencethreshold of 1×10¹⁰) or clinical signs of compromise from leukemia (≥ 20% weight loss, hind limb paralysis, ruffled fur,respiratory distress, poor mobility). One dasatinib-only mouse was censored due to death from isoflurane.

Dexamethasone and Dasatinib Pharmacokinetics

Steady-state dexamethasone and dasatinib were measured from plasma taken by cardiac puncture at the end of treatment in the osteonecrosis model and by retroorbital blood collection at day 28 in the BCR-ABL⁺ model. Mice were fasted overnight prior to collection. Samples were frozen at −80°Cuntil analysis. Dexamethasonewas measured by HPLC-MS, with a linear reportable range of 5.1–1275 nM, as previously described¹⁰. Dasatinib was measured by LC-MS/MS using a novel method. Quantitation of analytes was carried out with a Waters ACQUITY separation system (Milford, MA) and TQD triple-quadrupole system (Beverly, MA). Separation was achieved on a Waters ACQUITY BEHC₁₈ column (1.7 μm, 50 × 2.1 mm) using a column heater operating at 40°C with a Waters ACQUITY in-line filter. Autosampler temperature was maintained at 15 °± 5 °C. The gradient mobile phase was composed of 0.1% formic acid in acetonitrile (B) and 10mM ammonium acetate in 0.5% formic acid water at 0.7 mL/min and the separation was completed within 5 minutes. The instrument was equipped with an electrospray interface, and was controlled by Masslynx 4.1 software (Micromass, UK). The analysis was performed in MRM mode: m/z 488.16>401.11 for dasatinib; m/z 444.17>275.11 for N-deshyhydroxydasatinib;m/z504.17>387.12 for dasatinib N-oxide; and m/z 502.12>401.13 for dasatinib carboxylic acid. Dasatinib_d8 (m/z 496.23>405.93), dasatinib_N_deshyhydroxy_d8 (m/z 452.18>283.18) and dasatinib N-oxide_d8 (512.31>389.93) were used as internal standards. The MS/MS conditions were capillary voltage: 0.6kV; source temperature: 150°C; desolvation temperature: 450°C; cone gas flow: 10 L/h; desolvation gas flow: 900 L/h.Calibration curves of dasatinib and its metabolites were created by plotting the peak area ratios of analyte to the internal standard against the analyte concentrations in the spiked plasma. The back-calculated concentration for each standard was to be less than 15% of the nominal concentration except the LOQ which was to be less than 20%. A 10 μL aliquot of standard or murine plasma was spiked into a polypropylene microcentrifuge tube and 60 μL of internal standards solution at 15 ng/mL. The tube was vortex-mixed for 60 seconds, followed by centrifugation at 10,000 rpm for 8 minutes at 4°C. The supernatant was transferred to an autosampler vial and 3 μL was injected for analysis.The within- and between-run precisions for dasatinib and its metabolites were within 6.9% and accuracies were all within 5.3%.

Statistical Analysis

The frequency of osteonecrosis/arteriopathy was compared using chi-square test. Survival comparisons utilized log-rank testing. For continuous variables, the two-tailed Mann-Whitney test (twogroups) or the Kruskal-Wallis test with Dunn’s correction for multiple comparisons (three or more groups) were utilized. All continuous statistics are expressed as median with 95% confidence interval (CI). Statistical analyses and visual representation utilized the R program²⁰and GraphPad²¹. A significance threshold of p<0.05 was used for all analyses.The osteonecrosis experiment was poweredto detect an increase in osteonecrosis from 20% in the control arm to 50% in the experimental arm with 40 evaluable mice per arm, a two-sided alpha of 0.05, and an 80% power.Anti-leukemic efficacy experiments were powered to detect a hazard ratio of 4.35 with a two-sided alpha of 0.05 and an 80% power using 10 mice per arm.

Results:

Osteonecrosis

Mice were stratified by body weight and randomized to one of four treatment groups(Figure 1A): control (n=9); dasatinib-only (n=18); dexamethasone-only (n=45); dexamethasone + dasatinib (n=48).Consistent with previous observations^8,9,11,12,16, mice treated with dexamethasone (dexamethasone-only and dexamethasone + dasatinib treatment groups) gained less weight over time versus those not treated with dexamethasone (control and dasatinib-only treatment groups) (Figure 1B). Mice treated with dexamethasone + dasatinib gained less weight than dexamethasone-only mice (p=0.006). There was no difference between weight in control versus dasatinib-only mice at any time point (Figure 1B).

Figure 1.

Osteonecrosis experiments in BalbC mice. (A) Treatment groups. (B) Median weight of mice by treatment groups. (C) Survival of BalbC mice.control (n=9); dasatinib-only (n=18); dexamethasone-only (n=45; n=44 evaluable at 6-weeks); dexamethasone + dasatinib (n=48; n=43 evaluable at 6-weeks).

There was no difference in overall survival between treatment groups (p=0.16; Figure 1C). All mice not treated with dexamethasone (control and dasatinib-only groups) survived the entire six-weeks of treatment (Figure 1C). There was a non-significant increase in treatment related toxic death with dexamethasone + dasatinib (5/48) than dexamethasone-only (1/45) (p=0.11).

Consistent with previous studies^8,9,11,12,16, only micetreated with dexamethasone developed osteonecrosis (Figure 2). The frequency of osteonecrosis was not affected by the addition of dasatinib: 22.7% in dexamethasone-only versus 23.3% in dexamethasone + dasatinib (p>0.99; Figure 2A). Additionally, arteriopathy, a precursor of osteonecrosis, was not affected by the addition of dasatinib to dexamethasone treatment (p=0.36; Figure 2B).

Figure 2.

Percent of mice developing osteonecrosis or arteriopathy in those treated with dexamethasone only (N=44; DEX) versus dexamethasone plus dasatinib (N=43; DEX + Dasatinib).

Dexamethasone increased dasatinib exposure. The median concentration of dasatinibwas 18.8 ng/mL (95% CI 13.9–24.9 ng/mL) in dasatinib-only mice versus 53.1 ng/mL (95% CI 43.5–57.3 ng/mL) in dexamethasone + dasatinib mice (p<0.0001, Figure 3A).There was a non-significant decrease in dexamethasone plasma levels in mice treated with combination dexamethasone + dasatinib vs dexamethasone alone,(median 21.9 nM, 95% CI 18.0–27.3 nM versusmedian 27.5 nM, 95% CI 21.3–32.9 nM(p=0.059, Figure3B).

Figure 3.

Pharmacokinetics of dasatinib and dexamethasone after 6-weeks of treatment in Balbc mice. (A) Steady state concentrations of dasatinib by treatment group (n=18 Dasatinib Only; n= 43 DEX + Dasatinib) (B) Steady state concentrations of dexamethasone by treatment group (n= 44 Dexamethasone Only; n= 43 DEX + Dasatinib). Only mice that were treated with the drug being measured were evaluated for said drug.

Dasatinib in BCR-ABL leukemia

To test whether the dasatinib provided in the diet was effective against ALL, we utilized an established BCR-ABL⁺ model of murine ALL¹⁰(Figure 4A).Dasatinib, whether dasatinib-only or dexamethasone + dasatinib groups, suppressed leukemia, as evidenced by suppression of ventral luminescence from marked ALL (Figure 4B) and by longer duration of survival (Figure 4C). All control mice succumbed to leukemia by day 25 of treatment (Figure 4C). At day 28, mice receiving dasatinib-only or dexamethasone + dasatinib were taken off treatment and monitored for disease relapse. Following an additional 30 days of observation, all mice originally treated with dasatinib diet were alive (Figure 4C) and had the lowest observed luminescent signal (Figure 4B).

Figure 4.

Antileukemic efficacy of dexamethasone with or without dasatinib in a BCR/ABL+ model of murine ALL in C57Bl/6J mice. (A) Treatment groups and experimental schema. (B) Ventral luminescence over time for marked ALL cells by treatment group. (C) Probability of survival in ALL-bearing miceby treatment group. N=10/treatment group.

Mice treated with dasatinib only had prolonged survival compared to mice treated with dexamethasone only (p=0.02) but not compared to the dexamethasone + dasatinib group (p=0.08; Figure 4C).

As observed in the osteonecrosis BalbC model, BCR-ABL⁺ mice treated with dexamethasone + dasatinib had higher exposure to dasatinib compared to dasatinib-only mice: median 27.2 nM (95% CI 19.6–31.1 nM) versus 12.5 nM (95% CI 8.7–13.5nM), respectively (p<0.001; Figure 5A).In contrast to the osteonecrosis mice, dexamethasone + dasatinib-treated BCR-ABL+ mice had higher dexamethasone exposure at day 28 compared to mice treated with dexamethasone alone: median 4.4 nM (95% CI 3.1–5.2 nM) versus 2.3 nM (95% CI 1.5–5.0 nM), respectively (p<0.01; Figure 5B).

Figure 5.

Plasma steady state concentrations of dasatinib (A) and dexamethasone (B) by treatment group among C57Bl/6J ALL-bearing mice at day 28 of treatment. N=10/treatment group. Only mice that were treated with the drug being measured were evaluated for said drug.

Discussion:

Dasatinib, at exposures that were highly effective against ALL, did not increase dexamethasoneassociated osteonecrosis. Although there are a few worrisome clinical reports describing osteonecrosis in patients treated with dasatinib^4,5,these reports are not convincing in ALL. In the report of ALL patients, small patient numbers preclude attribution of observed osteonecrosis to either dasatinib or the therapeutic backbone which included drugs (glucocorticoids, asparaginase) known to cause osteonecrosis⁵. Although cases of osteonecrosis were also reported among CML patients, the cause of the osteonecrosis was not clearly related to dasatinib. Given that there are now several large studies reporting the addition of dasatinib to ALL regimens^3,22that did not reportincreased rates of osteonecrosis, there is not strong clinical evidence of an effect of dasatinib on glucocorticoid-induced osteonecrosis.

Although there are theoretical concerns of a drug interaction between dexamethasone and dasatinib, there have yet to be convincing clinical pharmacokinetic data demonstrating a drug interaction. Predicting dexamethasone-based drug interactions has proven difficult, probably because dexamethasone is both a competitive inhibitor and an inducer of CYP3A²³. Indeed, in these murine studies, the BalbC mice showed a non-significant trend toward decreased concentrations of dexamethasone in the presence of dasatinib. In contrast, dexamethasone concentrations were higher in the presence of dasatinib in the C57Bl/6J mice. The reasons for these conflicting results are not clearbut could be in part due to differences in genetics between BalbC versus C57Bl/6J mouse strains and/or the fundamental differences in the models (non-leukemia versus leukemia-bearing mice). In both murine models, the addition of dexamethasone increased exposure to dasatinib, suggesting that inhibition of dasatinib metabolism or excretion seemed to predominate over any inductive effect dexamethasone may have had on drug clearance mechanisms.

It should be noted that these experiments used dasatinib added to mouse chow, instead of oral gavage. Given the likelihood that dasatinib will be used chronically for months or even years to treat leukemias, this mode of administration will be very convenient for future experiments. Given estimated chow consumption of ~4.4 g/day, we estimate this dasatinib dose at ~40 mg/kg/day, which is slightly higher than the dasatinib doses reported by others of 20 mg/kg/day by oral gavage^17,24,25. Because the plasma concentrations were similar to what has been observed in the clinic²⁶ and in other preclinical models²⁷, and because this regimen was well tolerated and demonstrated excellent ALL efficacy, the dosages used in our model were reasonable and sufficient to test whether, at therapeutic exposures, dasatinibincreases the osteonecrotic effect of dexamethasone.While the continuous exposure to dasatinib via food administration in our model may result in a flatter PK profile than the peak-and-trough profile seen with daily oral administration, this is unlikely to alter our conclusions. Specifically, hepatic drug interactions are more likely when both cytochrome substrates are present simultaneously, suggesting that our model is unlikely to underestimate drug-drug interactions. Additionally, data from clinical²⁸ and preclinical^8,10 trials suggests that continuous drug exposure is more likely than discontinuous exposure to induce osteonecrosis. This suggests that our model is unlikely to underestimate the potential osteonecrotic effect of dasatinib, Based on the results of our preclinicalmurine studies, we conclude that it is unlikely that dasatinib increases the risk of dexamethasone-induced osteonecrosis in ALL regimens.

Table of abbreviations

TKI

tyrosine kinase inhibitor

ALL

acute lymphoblastic leukemia

CML

chronic myeloid leukemia

CNS

central nervous system

SJCRH

St. Jude Children’s Research Hospital

HPLC-MS

high performance liquid chromatography and mass spectrometry