Vitamin D levels do not cause vitamin-drug interactions with dexamethasone or dasatinib in mice

Vitamin D3 (VD3) induces intestinal CYP3A that metabolizes orally administered anti-leukemic chemotherapeutic substrates dexamethasone (DEX) and dasatinib potentially causing a vitamin-drug interaction. To determine the impact of VD3 status on systemic exposure and efficacy of these chemotherapeutic agents, we used VD3 sufficient and deficient mice and performed pharmacokinetic and anti-leukemic efficacy studies. Female C57BL/6J and hCYP3A4 transgenic VD3 deficient mice had significantly lower duodenal (but not hepatic) mouse Cyp3a11 and hCYP3A4 expression compared to VD3 sufficient mice, while duodenal expression of Mdr1a, Bcrp and Mrp4 were significantly higher in deficient mice. When the effect of VD3 status on DEX systemic exposure was compared following a discontinuous oral DEX regimen, similar to that used to treat pediatric acute lymphoblastic leukemia patients, male VD3 deficient mice had significantly higher mean plasma DEX levels (31.7 nM) compared to sufficient mice (12.43 nM) at days 3.5 but not at any later timepoints. Following a single oral gavage of DEX, there was a statistically, but not practically, significant decrease in DEX systemic exposure in VD3 deficient vs. sufficient mice. While VD3 status had no effect on oral dasatinib’s area under the plasma drug concentration-time curve, VD3 deficient male mice had significantly higher dasatinib plasma levels at t = 0.25 hr. Dexamethasone was unable to reverse the poorer survival of VD3 sufficient vs. deficient mice to BCR-ABL leukemia. In conclusion, although VD3 levels significantly altered intestinal mouse Cyp3a in female mice, DEX plasma exposure was only transiently different for orally administered DEX and dasatinib in male mice. Likewise, the small effect size of VD3 deficiency on single oral dose DEX clearance suggests that the clinical significance of VD3 levels on DEX systemic exposure are likely to be limited.


Dasatinib VitD DDI
FOR P-PKSR APPROVED USE AND DISTRIBUTION 1.0 OBJECTIVES § To evaluate the plasma pharmacokinetics (PK) of dasatinib after two separate single oral gavage doses of solution, separated by a washout period, in normal C57BL/6 mice § To assess the effects of Vitamin D dietary status (VD3 Suf vs VD3 Def), sex (M vs F) and age (8wk vs 10wk) on the plasma PK of oral dasatinib using a nonlinear mixed effects (NLME) PK modeling approach 2.0 MATERIALS AND METHODS

In Vivo Pharmacokinetic (PK) Study
The total plasma PK of dasatinib in C57BL/6 mice (The Jackson Laboratory) was assessed after single oral gavages of 10 mg/kg of dasatinib free base equivalents. Dasatinib (St. Jude Compound Management, SJ000518976-9, LC Labs, D-3307, Lot BDS-106, purity >99%) was dissolved in 80 mMol/L sodium citrate buffer pH 3.1 for a final dasatinib concentration of 1 mg/mL for a 10 mL/kg gavage volume. Mice were grouped by Vitamin D dietary sufficiency status and sex, and studied at 8 and 10 weeks of age. Survival saphenous and facial vein bleeding of mice was conducted using IACUC-approved methods at 0. 125, 0.25, 0.5, 1, 2, 4, 8, 16 and 24 hr. post-dose, with 3 mice per time point. Each mouse was sampled 3 times on the 2 separate occasions of dasatinib dosing. Blood was collected into a Sarstedt Minivette ® POCT 50 µl K3 EDTA capillary device, dispensed into a microtube and vortexed to mix the anticoagulant. The tubes were then immediately centrifuged to plasma, and stored on dry ice for the remainder of the study. At the end of the in vivo procedures, plasma samples were transferred from dry ice and placed at -80 °C until analysis.

Bioanalysis
Total plasma dasatinib concentrations were assessed using a sensitive and specific liquid chromatography, tandem mass spectrometry assay. Dasatinib (St. Jude Compound Management, SJ000518976-9, Purity >99%) stock solutions were prepared in acetonitrile and used to spike matrix calibrators and quality controls. Protein precipitation was performed using a 1: 4 ratio of plasma to 15 ng/mL erlotinib HCl (St. Jude Compound Management, S00004053, Inventory ID: SJCH0042996, Purity >95%) in methanol as an internal standard. A 3 µL aliquot of the extracted supernatant was injected onto a Shimadzu LC-20ADXR high performance liquid chromatography system via a LEAP CTC PAL autosampler. The LC separation was performed using a Phenomenex Kinetex 2.6 µm EVO C18 (100 Å, 50 x 2.1 mm) maintained at 50 °C with gradient elution at a flow rate of 0.5 mL/min. The binary mobile phase consisted of water: acetonitrile: 200 mM ammonium acetate in H2O pH = 6.0 (9:1:1 v/v) in reservoir A and acetonitrile: H2O: 200 mM ammonium acetate in H2O pH = 6.0 (9:1:1 v/v) in reservoir B. The mobile phase gradient began with a linear increase to 100% B in 4.0 minutes. The column was then rinsed for 2.0 minutes at 100% B and then equilibrated at the initial conditions for two minutes for a total run time of eight minutes. Under these conditions, the analyte and IS eluted at 1.81 and 2.21 minutes, respectively.
Analyte and IS were detected with tandem mass spectrometry using a SCIEX API 5500 Q-TRAP in the positive ESI mode with monitoring of the following mass transitions: dasatinib 488.20 -> 401.00, erlotinib HCl 394.20 -> 250.10.
The experimental bioanalytical runs were all found to be acceptable for the purpose of a singlicate non-GLP, preclinical PK assessment. A linear model (1/X 2 weighting) fit the calibrators across the 1.00 to 500 ng/mL range, with a correlation coefficient (R) of ≥0.9958. The lower limit of quantitation (LLOQ), defined as a peak area signal-to-noise ratio of 5 or greater verses a matrix blank with IS, was 1.00 ng/mL. The intra-run precision and accuracy was < 11.6% CV and 90.6% to 107%, respectively.

Pharmacokinetic Analysis
Plasma concentration-time (Ct) data in ng/mL for dasatinib were grouped by individual mouse, Vitamin D status, sex, and age, and were analyzed using nonlinear mixed effect (NLME) modeling as implemented in Monolix version 2018R1 (Lixoft SAS, Antony, France). Briefly, parameters and the Fisher Information Matrix (FIM) were estimated using the stochastic approximation expectation maximization (SAEM) algorithm, and the final log-likelihood estimated with importance sampling, all using the default Monolix initial settings, except that 1000 iterations were permitted for estimation of FIM using stochastic approximation. A variety of models were fit to the dasatinib Ct data, parameterized using apparent clearances, volumes of distribution, and absorption rate constant as needed. These models were assessed for goodness of fit using the -2 log likelihood (-2LL) value, Akaike and Bayesian Information Criterion (AIC, BIC), visual predictive checks, plots of model individual and population predicted vs. observed data, residual plots, and the standard error of parameter estimates. A log-normal inter-individual and inter-occasion parameter distribution was assumed on selected supported parameters, with both onand off-diagonal elements of parameter covariance matrices tested. Additive and/or proportional error models were tested and implemented as supported. Beal's M3 method was used to handle any data that were below the LLOQ or above the upper limit of the assay range [1]. The grouping levels were tested as categorical covariates on supported PK parameters, primarily the apparent oral clearance (Cl). A covariate effect was considered significant if its addition reduced the -2LL by at least 3.84 units (P < 0.05, based on the χ 2 test for the difference in the -2LL between two hierarchical models that differ by 1 degree of freedom). Additionally, Wald test P values were outputted for each covariate effect by the Monolix software.

RESULTS AND DISCUSSION
The oral dasatinib plasma PK in the studied mice was best described using a linear, two-compartment, zero order absorption model, with proportional residual error. Inter-individual variability upon apparent oral systemic clearance (Cl) and apparent volume of distribution of the central compartment (V1) was supported, improved the model fit and performance, as did the off-diagonal correlation between these two parameters. This was defined as the Base model. Inter-occasion variability on either Cl or V1 was not supported, resulted in model instability (failure of FIM convergence via SAEM) and a poorer model fit. This prevented formal testing age as a covariate, but implies that dasatinib PK is not significantly different between 8 and 10 week old mice.
Vitamin D sufficiency and sex were then tested alone and combined as covariates on Cl, and were also found to be insignificant at the predefined p<0.05 threshold level. Overall, the Base model without any covariates performed the best, and suggested no influence of Vitamin D status, sex, or age on oral dasatinib PK in mice. Table 3.1 presents the results from the Base and Final models and   is adequately converted in vivo. Despite the statistical significance, the magnitude of the effect was small -ranging from 4.9% to 17.2%. The authors conceded that while the clinical significance of this effect is unclear, it's likely to be limited given the small effect size. It is unclear what the inter-and intra-assay precision and accuracy values were for these TDM assays, but it's assumed they were at least within standard values of ≤15%. Through the sheer number of samples, the investigators achieved adequate power to detect such small, and likely clinically irrelevant, effects. For reference, these differences are within the FDA guidance for therapeutic bioequivalence by plasma AUCs or clearance values, i.e. ± 20%.
The current mouse study was primarily designed to evaluate the effect of Vitamin D status on apparent oral Cl of dasatinib, and to explore any effect of sex. Eighteen mice total, balanced for sex, were allocated to each parallel Vitamin D status group. The population Cl estimate was precise, with a RSE of 6.66%, and the inter-individual variance was well estimated (20% RSE). Considering the log-normally distributed variability in Cl, and assuming only an effect of Vitamin D without confounding sex effect, 44 mice in each parallel Vitamin D group would be required to determine a 17% difference in Cl, with an alpha = 0.1 and beta = 0.20 [3]. Such small effect sizes are very difficult to determine in preclinical mouse studies, and are of questionable relevance.

REFERENCES
1. Beal SL. Ways to fit a PK model with some data below the quantification limit.