Determination of Drug Binding Kinetics to Fat in Biorelevant Gastric Fed State Media

Purpose The objective of the study was to develop a method to predict the rate of drug binding to fat for drugs with different physicochemical properties, when biorelevant fed state gastric media are used in in vitro dissolution tests. Investigation of a possible correlation between the drug binding rate to the lipid part of the fed state gastric media with the physicochemical properties of the model compounds and the presence of positive or negative food effect observed in vivo was also performed. Methods For each drug studied, 30 mL of concentrations equal to their aqueous solubility in Fed State Simulated Gastric Fluid [prepared with 0% fat milk (FeSSGF 0%) with pH adjusted to 5] were placed outside dialysis tubing cellulose membrane of average flat width 25 mm and MWCO 12- 14 KDa (Sigma, UK) in a 50 mL centrifuge tube. Anhydrous milk fat, commercially purchased, after being heated at 37 °C, was placed inside the membrane (8 cm length) in quantities equivalent to 5%, 8%, 15%, 20%, 25% and 33% w/v fat concentrations in FeSSGF for nifedipine (10 μg/ mL) and to 5%, 15% and 25% w/v for danazol (1 μg/ mL), atorvastatin calcium (2.6 μg/ mL), ketoconazole (2.7 μg/ mL) and propafenone HCl (200 μg/ mL). Anhydrous milk fat has a similar fatty acid profile with milk and in this study, it was used as a surrogate of the lipid part of FeSSGF. The experiments were performed in a 37 °C incubator room under constant stirring (300 rpm). 0.5 mL samples were taken at 0, 0.5, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 24 (nifedipine) and 48 hours (all other drugs). Proteins were precipitated after addition of 2 parts of methanol at 1 part of FeSSGF and the samples were vortexed (30 sec), centrifuged (8000 rpm, 15 min, 4 °C), filtered and analysed with HPLC-UV. All experiments were performed in triplicate. The drug binding rate for all drugs was expressed as a decrease in drug concentration from the donor compartment and in case of nifedipine, the rate of binding (k) was also expressed as a function of % fat content (SigmaPlot®). Results Drug binding to fat was proven to be an extremely slow process. Using nifedipine as a starting model compound, a predictive model was developed so as to access the relationship between medium fat content and drug binding process and predict the binding rate to the fat, even at low fat concentrations. The binding process followed a first order kinetics pattern in all drugs during the selected sampling period except for propafenone HCl, the binding rate of which was too low to be determined. The study with nifedipine revealed that the drug binding rate was proportional to the % fat content with the binding rate vs % drug bound relationship following a hyperbolic model [k = 0.1195 × % fat content/ (10.4086 + % fat content)]. Using the equation derived from the predictive model developed, the k value of nifedipine’s binding to 1.8% fat content of FeSSGF was 0.018 h-1. With the exception of nifedipine, an amount equivalent to less than 15% of the initial drug concentration was bound to the fat in the first 8 hours, even at a 25% w/v fat concentration medium. The k values at 25% w/v fat, were 0.0881, 0.0165, 0.0136, 0.0123 and 0.0013 h-1 for nifedipine, danazol, ketoconazole, atorvastatin calcium and propafenone HCl respectively. Compared to nifedipine, all other drugs tested had higher log P values but were bound to fat multiple times more slowly. Interestingly, drug binding rate was not correlated to logP, logD, aqueous solubility or food effect. Even though propafenone HCl, which was bound to the fat much slower than the other drugs, had significantly lower log D (pH= 5) value (0.68), no clear trend was observed between log D and the rate of binding. Conclusion Drug binding to fat is a very slow process which followed a first order kinetics pattern for all model drugs chosen. Binding rate was proportional to the fat content in the medium, following a hyperbolic model. The binding rate did not seem to have a direct correlation either with drug lipophilicity or with food effect observed in in vivo studies.