Purpose: Often, enabled formulations are needed to achieve desired exposure of BCS Class II compounds in preclinical safety studies. Amorphous solid dispersions are used to improve bioavailability by creating supersaturated drug concentrations upon dissolution, they can reach a higher concentration of unbound drug than the crystalline form, and often generates a large number of drug-rich colloids species. Conventional dissolution test using static one chamber vessel can be misleading in selecting formulation. The purpose of this study was to evaluate an in-vitro screening method with an absorption phase during drug dissolution for rank-ordering bioavailability enhancement formulations tested in rats.
Methods: μFlux apparatus (Pion Inc.) was used to monitor the drug concentration in the donor and acceptor. Dissolution medium at pH 4 (mimicking rat stomach pH) and FaSSIF were used in the donor compartment and Acceptor Sink Buffer (ASB pH 7.4) in the acceptor, separated by a lipophilic membrane. Concentrations were monitored using in situ fiber optic probe. HPLC was performed on manually pulled samples from donor compartment to determine the total drug concentration and isolated smaller species (centrifuged + filtered). A weak base, BCS class II drug BI XYZ with solubility of 22 µg/mL (pH 4 and above) was used for this study. Dose formulations of jet-milled API (JM) or spray-dried amorphous dispersion (ASD) of BI XYZ/HPMCAS (1:2 weight ratio) were dispersed in donor compartment at drug load equivalent to 30 mg/kg in rats. Formulations tested were JM in 1% HPMC/pH2, ASD in 1% HPC/pH2 and ASD in natrosol. Second derivative spectral analysis between 286-290 nm was used. Zero Intercept Method (ZIM) previously described was used to recognize spectroscopic changes and assess BI XYZ amorphous solubility.
Results: •Initial formulation screening showed that drug fraction dissolved in dose formulations for JM-1% HPMC/pH2, ASD-natrosol, ASD-1% HPC/pH2 were 33%, 1-0.5% and 73-33% respectively. PLM and XRPD showed that API crystallizes in natrosol formulation within 2 minutes and this formulation was not evaluated further for flux.
•2 Step-dissolution in pH4/6.5 shows a significant increase in drug concentration: 280 µg/mL corresponding to 10-fold crystalline API solubility, with maintenance of supersaturation for 45 minutes for ASD-1% HPC/pH2 while JM-1% HPMC/pH2 concentration reached only 90 to 50 µg/mL at 45 minutes.
•In FaSSIF media, dissolution profiles for JM-1% HPMC/pH2 are comparable to profiles in pH4. However, ASD-1% HPC/pH2 reached supersaturation to 10-fold its crystalline solubility but rapidly precipitated within 10 minutes and the concentration dropped to 80 µg/mL at 1 hour. PLM microscopy of ASD-1% HPC/pH2 dispersed into FaSSIF, taken at 5 minutes, shows the presence of birefringent particles of 1-3 µm in size indicating API crystallization.
•HPLC analysis of the supernatant and smaller species of ASD-1% HPC/pH2 in FaSSIF in the donor shows that more than 80% of the drug concentration is contributed by smaller species (< 0.2 µm) suggesting no substantial amount of drug-rich larger polymer colloids aggregates present in the system.
•Flux data in FaSSIF correlates with in-vivo rat data. Flux ratio of ASD-1% HPC/pH2 over JM-1% HPMC/pH2 is 1.3 and corresponds to the AUC and Cmax ratio in rat species 1.3 and 1.5 respectively.
•ZIM data analysis combined with turbidity measurements also supports no large colloidal aggregate formation during dissolution. Amorphous solubility measured using ZIM was found to be 50 µg/mL.
Conclusion: For this weakly basic, fast crystallizer drug, conventional in vitro dissolution test in pH4 overpredicted formulation performance of the ASD-1% HPC/pH2. Although API solubility is comparable in pH 4 and pH 6.5 (FaSSIF), the rate of HPMCAS dissolving (ionization) is different and could influence the API crystallization. Choice of lower in vitro pH media relevant to rat species can be misleading in predicting absorption. In this case, flux data in FaSSIF were able to predict in-vivo formulation performance.