Towards virtual bioequivalence studies for oral dosage forms containing poorly water-soluble drugs: a physiologically based biopharmaceutics modeling (PBBM) approach

Advancing Virtual Bioequivalence for Orally Administered Drug Products: Methodology, Real-World Applications and Future Outlook

Bioequivalence studies are pivotal in generic drug development wherein therapeutic equivalence is provided with an innovator product. However, bioequivalence studies represent significant complexities due to the interplay of multiple factors related to drug, formulation, physiology, and pharmacokinetics. Approaches such as physiologically based biopharmaceutics modeling (PBBM) can enable virtual bioequivalence (VBE) assessment through appropriately developed and validated models. Such models are now being extensively used for bioequivalence risk assessment, internal decision-making, and the evaluation of drug and formulation factors related to bioequivalence. Depiction of the above-mentioned factors through the incorporation of variability and development of a virtual population for bioequivalence assessment is of paramount importance in utilizing such models. In this manuscript, we have portrayed our current understanding of VBE. A detailed explanation was provided with respect to study designs, in vivo variability, and the impact of physiological, drug, and formulation factors on the development of the population for VBE. Furthermore, strategies are suggested to incorporate variability in GastroPlus with an emphasis on intra-subject and inter-occasion variability. Two industrial case studies pertaining to immediate and modified release formulation were portrayed wherein VBE was utilized for decision-making and regulatory justification. Finally, regulatory understanding in the area of VBE, along with future perspectives, was detailed.

Physiologically Based Biopharmaceutics Modeling for Gefapixant IR Formulation Development and Defining the Bioequivalence Dissolution Safe Space

Gefapixant is a weakly basic drug which has been formulated as an immediate release tablet for oral administration. A physiologically based biopharmaceutics model (PBBM) was developed based on gefapixant physicochemical properties and clinical pharmacokinetics to aid formulation selection, bioequivalence safe space assessment and dissolution specification settings. In vitro dissolution profiles of different free base and citrate salt formulations were used as an input to the model. The model was validated against the results of independent studies, which included a bioequivalence and a relative bioavailability study, as well as a human ADME study, all meeting acceptance criteria of prediction errors ≤ 20% for both Cmax and AUC.  PBBM was also applied to evaluate gastric pH-mediated drug-drug-interaction potential with co-administration of a proton pump inhibitor (PPI), omeprazole. Model results showed good agreement with clinical data in which omeprazole lowered gefapixant exposure for the free base formulation but did not significantly alter gefapixant pharmacokinetics for the citrate based commercial drug product. An extended virtual dissolution bioequivalence safe space was established.  Gefapixant drug product batches are anticipated to be bioequivalent with the clinical reference batch when their dissolution is > 80% in 60 minutes. PBBM established a wide dissolution bioequivalence space as part of assuring product quality.

Mechanistic Modeling of In Vitro Biopharmaceutic Data for a Weak Acid Drug: A Pathway Towards Deriving Fundamental Parameters for Physiologically Based Biopharmaceutic Modeling

Mechanistic modeling of in vitro experiments using metabolic enzyme systems enables the extrapolation of metabolic clearance for in vitro-in vivo predictions. This is particularly important for successful clearance predictions using physiologically based pharmacokinetic (PBPK) modeling. The concept of mechanistic modeling can also be extended to biopharmaceutics, where in vitro data is used to predict the in vivo pharmacokinetic profile of the drug. This approach further allows for the identification of parameters that are critical for oral drug absorption in vivo. However, the routine use of this analysis approach has been hindered by the lack of an integrated analysis workflow. The objective of this tutorial is to (1) review processes and parameters contributing to oral drug absorption in increasing levels of complexity, (2) outline a general physiologically based biopharmaceutic modeling workflow for weak acids, and (3) illustrate the outlined concepts via an ibuprofen (i.e., a weak, poorly soluble acid) case example in order to provide practical guidance on how to integrate biopharmaceutic and physiological data to better understand oral drug absorption. In the future, we plan to explore the usefulness of this tutorial/roadmap to inform the development of PBPK models for BCS 2 weak bases, by expanding the stepwise modeling approach to accommodate more intricate scenarios, including the presence of diprotic basic compounds and acidifying agents within the formulation.

Importance of Considering Fed-State Gastrointestinal Physiology in Predicting the Reabsorption of Enterohepatic Circulation of Drugs

PURPOSE

The purpose of this study was to develop a simulation model for the pharmacokinetics (PK) of drugs undergoing enterohepatic circulation (EHC) with consideration to the environment in the gastrointestinal tract in the fed state in humans. The investigation particularly focused on the necessity of compensating for the permeability rate constant in the reabsorption process in consideration of drug entrapment in bile micelles.

METHODS

Meloxicam and ezetimibe were used as model drugs. The extent of the entrapment of drugs inside bile micelles was evaluated using the solubility ratio of Fed State Simulated Intestinal Fluid version 2 (FeSSIF-V2) to Fasted State Simulated Intestinal Fluid version 2 (FaSSIF-V2). Prediction accuracy was evaluated using the Mean Absolute Percentage Error (MAPE) value, calculated from the observed and predicted oral PK profiles.

RESULTS

The solubilization of ezetimibe by bile micelles was clearly observed while that of meloxicam was not. Assuming that only drugs in the free fraction of micelles permeate through the intestinal membrane, PK simulation for ezetimibe was performed in both scenarios with and without compensation by the permeation rate constant. The MAPE value of Zetia® tablet, containing ezetimibe, was lower with compensation than without compensation. By contrast, Mobic® tablet, containing meloxicam, showed a relatively low MAPE value even without compensation.

CONCLUSION

For drugs which undergo EHC and can be solubilized by bile micelles, compensating for the permeation rate constant in the reabsorption process based on the free fraction ratio appears an important factor in increasing the accuracy of PK profile prediction.

In Silico Modeling Approaches Coupled with In Vitro Characterization in Predicting In Vivo Performance of Drug Delivery System Formulations

Optimization of the in vivo performance of dosage forms in humans is essential in developing not only conventional formulations but also drug delivery system (DDS) formulations. Although animal experiments are still useful for these formulations, in silico approaches have become increasingly important for DDS formulations with regard to species-specific differences in physiology that can affect the in vivo performance of dosage forms between animals and humans. Furthermore, it is also important to couple in vitro characterizations with in silico models to predict in vivo performance in humans precisely. In this review article, I summarized in vitro-in silico approaches to predicting the in vivo performance of oral DDS formulations (amorphous solid dispersions, lipid-based formulations, nanosized formulations, cyclodextrins-based formulations, sustained release products, enteric coat products, and orally disintegrating tablets) and parenteral DDS formulations (cyclodextrins-based formulations, liposomes, and inhaled formulations).

Preparation of pazopanib-fumarate disodium glycyrrhizinate nanocrystalline micelles by liquid-assisted ball milling

Pazopanib (PZ) is a multikinase inhibitor, which is mainly used in the treatment of soft tissue sarcoma and advanced renal cancer. However, because of its water insolubility, oral bioavailability is poor. At the same time, photo lability and high dose oral administration lead to severe hepatotoxicity, which is limited in clinical application. In this paper, the novel pazopanib-fumarate disodium glycyrrhizinate nanocrystalline micelles are successfully prepared by liquid-assisted ball milling. The prepared cocrystals and nanocrystalline micelle structures are systematically characterized by X-ray diffraction (XRD), differential scanning calorimetry (DSC) and Fourier Transform Infrared Spectrometer (FTIR) analysis. In vitro solubility and dissolution experiments show that the solubility and dissolution of nanocrystalline micelles are significantly improved under different simulated physiological conditions. The accelerated stabilization experiments show that the nanocrystalline micelles have good physical and chemical stability and showed excellent stability in water (Zeta potential was 62.39 mV). In addition, the in vivo bioavailability of nanocrystalline micelles is 3 times higher than that of PZ, and the therapeutic threshold (> 20 μg/mL) is up to 30 h. This new strategy provides a feasible solution to the undesirable properties of PZ.

Integration of artificial neural network and physiologically based biopharmaceutic models in the development of sustained-release formulations

Model-informed drug development is an important area recognized by regulatory authorities and is gaining increasing interest from the generic drug industry. Physiologically based biopharmaceutics modeling (PBBM) is a valuable tool to support drug development and bioequivalence assessments. This study aimed to utilize an artificial neural network (ANN) with a multilayer perceptron (MLP) model to develop a sustained-release matrix tablet of metformin HCl 500 mg, and to test the likelihood of the prototype formulation being bioequivalent to Glucophage® XR, using PBBM modeling and virtual bioequivalence (vBE). The ANN with MLP model was used to simultaneously optimize 735 formulations to determine the optimal formulation for Glucophage® XR release. The optimized formulation was evaluated and compared to Glucophage® XR using PBBM modeling and vBE. The optimized formulation consisted of 228 mg of hydroxypropyl methylcellulose (HPMC) and 151 mg of PVP, and exhibited an observed release rate of 42% at 1 h, 47% at 2 h, 55% at 4 h, and 58% at 8 h. The PBBM modeling was effective in assessing the bioequivalence of two formulations of metformin, and the vBE evaluation demonstrated the utility and relevance of translational modeling for bioequivalence assessments. The study demonstrated the effectiveness of using PBBM modeling and model-informed drug development methodologies, such as ANN and MLP, to optimize drug formulations and evaluate bioequivalence. These tools can be utilized by the generic drug industry to support drug development and biopharmaceutics assessments.

Application of a novel PhysioCell apparatus for biopredictive dissolution tests of oral immediate release formulations - a case study workflow for in vitro-in vivo predictions

Biorelevant dissolution tests of oral solid dosage forms open the gate to valid in vitro-in vivo predictions (IVIVP). A recently developed apparatus, PhysioCell, allows mimicking the fluid flow and pressure waves occurring in the human fasted stomach. In this work, we used the PhysioCell to perform IVIVP for vortioxetine immediate-release (IR) tablets: the originator (Brintellix) and generic product candidates (VORTIO). The dissolved drug was monitored in the gastric (StressCell) and intestinal (Collection Vessel) compartments that contained biorelevant media. Simulated intermittent gastric stress at 15 min and 'housekeeping wave' at 30 min increased the dissolution of Brintellix formulations only. A mechanistic model that best described the observations involved the first-order tablet disintegration with a stress-induced enhancement for Brintellix, dissolution of solid particles in the StressCell, and drug transfer to the Collection Vessel. Then, a semi-mechanistic pharmacokinetic model with dissolution parameters as inputs simulated vortioxetine plasma concentrations in healthy volunteers after single and multiple dosing of Brintellix. Despite different dissolution characteristics, VORTIO provided similar concentration profiles to the originator. In conclusion, PhysioCell dissolution tests, combined with semi-mechanistic IVIVP, can be successfully used to develop IR dosage forms exhibiting gastric stress-related effects.

Conjunction of semi-mechanistic in vitro-in vivo modeling and population pharmacokinetics as a tool for virtual bioequivalence analysis - a case study for a BCS class II drug

Virtual bioequivalence trial (VBE) simulations based on (semi)mechanistic in vitro-in vivo (IVIV) modeling have gained a huge interest in the pharmaceutical industry. Sophisticated commercially available software allows modeling variable drug fates in the gastrointestinal tract (GIT). Surprisingly, the between-subject and inter-occasion variability (IOV) of the distribution volumes and clearances are ignored or simplified, despite substantially contributing to varied plasma drug concentrations. The paper describes a novel approach for IVIV-based VBE by using population pharmacokinetics (popPK). The data from two bioequivalence trials with a poorly soluble BCS class II drug were analyzed retrospectively. In the first trial, the test drug product (biobatch 1) did not meet the bioequivalence criteria, but after a reformulation, the second trial succeeded (biobatch 2). The popPK model was developed in the Monolix software (Lixoft SAS, Simulation Plus) based on the originator's plasma concentrations. The modified Noyes-Whitney model was fitted to the results of discriminative biorelevant dissolution tests of the two biobatches and seven other reformulations. Then, the IVIV model was constructed by joining the popPK model with fixed drug disposition parameters, the drug dissolution model, and mechanistic approximation of the GIT transit. It was used to simulate the drug concentrations at different IOV levels of the primary pharmacokinetic parameters and perform the VBE. Estimated VBE success rates for both biobatches well reflected the outcomes of the bioequivalence trials. The predicted 90% confidence intervals for the area under the time-concentration curves were comparable with the observed values, and the 10% IOV allowed the closest approximation to the clinical results. Simulations confirmed that a significantly lower maximum drug concentration for biobatch 1 was responsible for the first clinical trial's failure. In conclusion, the proposed workflow might aid formulation screening in generic drug development.

Dynamic Changes in Gastrointestinal Fluid Characteristics after Food Ingestion Are Important for Quantitatively Predicting the In Vivo Performance of Oral Solid Dosage Forms in Humans in the Fed State

The aim of this study was to develop a simulation model to predict the in vivo performance of solid oral dosage forms in humans in the fed state. We focused on investigating the effect of dynamic changes in gastrointestinal (GI) fluid characteristics in the fed state on the in vivo performance of solid dosage forms. We used six solid dosage forms containing weak base drugs as model formulations, two with positive food effects in humans, two with negative food effects, and two which are not affected by food ingestion. These model drug formulations were used to perform biorelevant dissolution tests in the stomach and small intestine under both prandial states. The in vitro properties of the drug products obtained from these tests were then coupled with in silico models (fasted or fed) to predict food effects in humans. We successfully incorporated the dynamic changes in GI fluid characteristics and their effects on the in vivo dissolution of drugs into the prediction model for the fed state. This newly designed physiologically based biopharmaceutics modeling approach provided the precise and quantitative prediction of food effects (i.e., changes in C_max and AUC after food ingestion) in humans while considering the dynamic changes in fluid characteristics in the fed state.

Model-Informed drug development of gastroretentive release systems for sildenafil citrate

Gastroretentive drug delivery systems (GRDDS) are modified-release dosage forms designed to prolong their residence time in the upper gastrointestinal tract, where some drugs are preferentially absorbed, and increase the drug bioavailability. This work aimed the development of a novel GRDDS containing 60 mg of sildenafil citrate, and the evaluation of the feasibility of the proposed formulation for use in the treatment of pulmonary arterial hypertension (PAH), for once a day administration, by using in silico pharmacokinetic (PK) modeling and simulations using GastroPlus^TM. The Model-Informed Drug Development (MIDD) approach was used in formulation design and pharmacokinetic exposure prospecting. A 2² factorial design with a central point was used for optimization of the formulation, which was produced by direct compression and characterized by some tests, including buoyancy test, assay, impurities, and in vitro dissolution. A compartmental PK model was built using the GatroPlus^TM software for virtual bioequivalence of the proposed formulations in comparison with the defined target release profile provided by an immediate release (IR) tablet formulation containing 20 mg of sildenafil administered three times a day (TID). The results of the factorial design showed a direct correlation between the dissolution rate and the amount of hydroxypropyl methyl cellulose (HPMC) in the formulations. By comparing the PK parameters predicted by the virtual bioequivalence, the formulations F1, F2, F3 and F5 failed on bioequivalence. The F4 showed bioequivalence to the reference and was considered the viable formulation to substitute the IR. Thus, GRDDS could be a promising alternative for controlling the release of drugs with a pH-dependent solubility and narrow absorption window, specifically in the gastric environment, and an interesting way to reduce dose frequency and increase the drug bioavailability. The MIDD approach increases the level of information about the pharmaceutical product and guide the drug development for more assertive ways.

Quantitative assessment of disintegration rate is important for predicting the oral absorption of solid dosage forms containing poorly soluble weak base drugs

This study aimed to develop a novel in silico modeling and simulation that considers the disintegration rate in the stomach to predict the in vivo performance of oral solid dosage forms with slow disintegration rates containing poorly soluble weak base drugs. Oxatomide and manidipine hydrochloride were used as model drugs. First, the in vitro disintegration rate and dissolution rate were determined in biorelevant media that simulate the gastrointestinal fluids in fasted humans using a USP apparatus II paddle dissolution tester. Next, the oral absorption of the dosage forms was predicted using the novel simulation model coupled with not only the dissolution rate but also the estimated disintegration rate. As the in vitro disintegration time was 45 min or longer for both drugs in Fasted State Simulated Gastric Fluid, the disintegration rate of these dosage forms was considered slow as immediate release (IR) tablets. While the predicted and observed pharmacokinetic profiles of both drugs were comparable using the new model, the conventional model, which did not consider the disintegration step, underestimated the oral absorption of both drugs. Thus, our novel simulation model coupled with the disintegration rate estimated from in vitro tests is promising for predicting the in vivo performance of oral solid dosage forms with slow disintegration rates containing poorly soluble weak base drugs.