(M1230-04-22) Food Effect Risk Assessment in Preformulation Stage Using Material Sparing µFLUX Methodology

Purpose: The intake of food and meal type can strongly impact the bioavailability of orally administered drugs and can consequently impact drug efficacy and safety. A positive food effect (FE) may compromise safety if exposure exceeds maximum tolerable dose while a negative food effect may compromise efficacy of the drugs by decreasing its absorption. Therefore, assessing food effect on the absorption of a drug is critical to optimize the safety and efficacy of the final product. During the early stages of drug development, when only a small amount of drug substance is available, scientists might not be aware of the mechanism of food effect, and the limiting steps in oral absorption of a new drug candidate. The aim of this study was to investigate the suitability of a small volume dissolution-permeation set up combined with in-situ UV concentration monitoring, to predict food effect early in preformulation stage using unformulated drug substance

Methods: 11 model compounds (BCS class I-IV) with diverse physicochemical properties and published human clinical data on food effects were selected. In vitro dissolution-permeation studies were carried out using µFLUX apparatus (Pion Inc., Billerica MA, USA). Measurements were performed in side-by-side donor -acceptor pairs (Figure 1) filled with 16-20 mL of simulated intestinal fluid in the fasted (FaSSIF) and fed state (FeSSIF) in the donor side and matching volume of Acceptor Sink Buffer (ASB pH 7.4) in the acceptor compartments. The compartments were separated by a filter support membrane which at the beginning of assays was impregnated with solution of lecithin in dodecane (GIT Lipid, Pion Inc.) forming the lipophilic barrier. The compositions of the lipophilic barrier and the receiver solution were identical to the Double-Sink™ PAMPA model described elsewhere (ref 1). Unformulated drugs were added as a powder into the donor compartments and concentration was monitored 3-6 hours in both chambers of µFLUX pair usingin-situ monitoring technique implemented with Rainbow fiber optic  UV-Vis spectrometer (Rainbow R2D8, Pion Inc).  Direct spectra or second derivative analysis and path length from 2-20 mm were used to further improve concentration analysis. The temperature in both compartments was maintained at 37 ˚C and the media were stirred at 100 RPM

Results: in-situ concentration time profiles from donor and acceptor chambers were collected for all compounds. Figure 2 shows representative concentration profiles in donor (left) and acceptor (right) chambers obtained during µFlux experiment in FeSSIF media for celecoxib. The flux across the membrane representing the amount of drug substance crossing a unit area of the membrane per unit time, was calculated from the concentration-time profiles in the receiver compartments using AuPro software in a range that capture the initial rate as the drug enter the receiver compartment. Clinical data on food effect were collected from literature and the change in area under the plasma concentration time curve with the change in the maximum concentration-time curve from fed state compared to fasted state AUC (fed/fa) and Cmax (fed/fa) were calculated. Among the 11 selected compounds, 4 compounds showed positive FE, 3 compounds showed no FE and 4 compounds showed negative FE. Solubility of model compounds in FaSSIF and FeSSIF were collected from the literature and the change in solubility from fed to fasted state was calculated and compared to AUC (fed/fa) and Cmax (fed/fa). Solubility (fed/fa) was able to predict FE for 5 of the compounds. Solubility ratio was able to predict FE for all the compounds with positive FE and for 1 compound with negative FE. Change in flux, measured in the µFLUX, from the fed to fasted state was calculated and compared with AUC and Cmax ratios. The flux ratio was able to predict food effect for 9 of the 11 compounds. Flux ratio was able to predict FE for all compounds with positive FE, with no FE and half of the compounds with negative FE

Conclusion: This study shows that dissolution-permeation setup under biorelevant conditions could be used in preformulation to estimate food effect of new drug candidates. While the solubility-based approach was able to predict positive FE for low solubility compounds, the flux-based approach was able to predict positive, no FE and some negative FE of compounds with passive absorption mechanism. For drugs with active absorption (by transporter) or compounds undergoing intestinal metabolism, flux-based approach with lipid-coated membrane may not reliably predict the food effect. Overall, the flux-based approach using a small amount of unformulated API was more predictive than the solubility-based approach and may provide good direction to design a clinical plan and adjustment of the drug dose in fed state

References: Ref 1: Avdeef, A., 2003. Absorption and Drug Development. Wiley–Interscience, New York, pp. 116–246

Figure 1 represents a schematic of the µFlux apparatus (Pion Inc., Billerica MA, USA) used in the study



Figure 2. Celecoxib: in situ concentration profiles in donor (left) and acceptor(right) chambers obtained during µFlux experiment in FeSSIF medium