Flux-Based Formulation Development-A Proof of Concept Study.

The work aimed to develop the Absorption Driven Drug Formulation (ADDF) concept, which is a new approach in formulation development to ensure that the drug product meets the expected absorption rate. The concept is built on the solubility-permeability interplay and the rate of supersaturation as the driving force of absorption. This paper presents the first case study using the ADDF concept where not only dissolution and solubility but also permeation of the drug is considered in every step of the formulation development. For that reason, parallel artificial membrane permeability assay (PAMPA) was used for excipient selection, small volume dissolution-permeation apparatus was used for testing amorphous solid dispersions (ASDs), and large volume dissolution-permeation tests were carried out to characterize the final dosage forms. The API-excipient interaction studies on PAMPA indicated differences when different fillers or surfactants were studied. These differences were then confirmed with small volume dissolution-permeation assays where the addition of Tween 80 to the ASDs decreased the flux dramatically. Also, the early indication of sorbitol's advantage over mannitol by PAMPA has been confirmed in the investigation of the final dosage forms by large-scale dissolution-permeation tests. This difference between the fillers was observed in vivo as well. The presented case study demonstrated that the ADDF concept opens a new perspective in generic formulation development using fast and cost-effective flux-based screening methods in order to meet the bioequivalence criteria. Graphical Abstract.

Prediction of Bioequivalence and Food Effect Using Flux- and Solubility-Based Methods.

In this work, two different approaches have been developed to predict the food effect and the bioequivalence of marketed itraconazole (ITRA) formulations. Kinetic solubility and simultaneous dissolution-permeation tests of three (ITRA) formulations (Sporanox capsules and solution and SUBA-ITRA capsules) were carried out in simulated fasted and fed states. Fraction of dose absorbed ratios estimating food effect and bioequivalence were calculated based on these results and were compared to the in vivo study results published by Medicines Agencies. The comparison demonstrated that kinetic solubility and flux values could be used as input parameters for biopharmaceutics modeling and simulations to estimate food effect and bioequivalence. Both prediction methods were able to determine a slightly negative food effect in the case of the Sporanox solution and also a pronounced positive food effect for the Sporanox capsule. Superior bioavailability was predicted when the Sporanox solution was compared to the Sporanox capsule (in agreement with in vivo data).

Using pH Gradient Dissolution with In-Situ Flux Measurement to Evaluate Bioavailability and DDI for Formulated Poorly Soluble Drug Products.

This study described a pH-gradient dissolution method combined with flux measurements as an in vitro tool for assessing the risk of bioavailability reduction due to drug-drug interactions (DDI) caused by acid reducing agents (ARAs). The device incorporates absorption chambers into USP II dissolution vessels, with fiber optic UV-probes monitoring concentration in situ. Dosage forms of Genentech BCS class II drugs, GDC-0810, GDC-0941, and compound A, were tested by starting the dissolution in either pH 1.6 or pH 4.0 media then converting to FaSSIF after 30 min. GDC-0810 showed no significant difference in flux between the two conversion experiments. A supersaturation phase was observed for GDC-0941 in the pH 1.6 experiments after media conversion to FaSSIF; however, it did not appear to occur in the pH 4.0 experiment due to low drug solubility at pH 4.0, resulting in a 95% decrease in flux compared to pH 1.6 experiment. The extent of flux reduction and the total accumulated API mass in the absorption chamber agreed well with the 89% reduction in mean Cmax and the 82% reduction in mean AUC from dog PK study between animals treated with pentagastrin and famotidine. Testing of the compound A optimized formulation tablets showed a 25% reduction in flux and in vitro absorbed amount by changing pH 1.6 to 4.0, correlating well with the AUC decrease in clinical studies. Good correlation between in vitro data and in vivo PK data demonstrated the applicability of the method for formulators to develop drug products mitigating DDI from ARAs.

Effect of Formulation Additives on Drug Transport through Size-Exclusion Membranes.

The aim of this research was to investigate the driving force of membrane transport through size-exclusion membranes and to provide a concentration-based mathematical description of it to evaluate whether it can be an alternative for lipophilic membranes in the formulation development of amorphous solid dispersions. Carvedilol, an antihypertensive drug, was chosen and formulated using solvent-based electrospinning to overcome the poor water solubility of the drug. Vinylpyrrolidone-vinyl acetate copolymer (PVPVA64) and Soluplus were used to create two different amorphous solid dispersions of the API. The load-dependent effect of the additives on dissolution and permeation through regenerated cellulose membrane was observed by a side-by-side diffusion cell, µFLUX. The solubilizing effect of the polymers was studied by carrying out thermodynamic solubility assays. The supersaturation ratio (SSR, defined as the ratio of dissolved amount of the drug to its thermodynamic solubility measured in exactly the same medium) was found to be the driving force of membrane transport in the case of size-exclusion membranes. Although the transport through lipophilic and size-exclusion membranes is mechanistically different, in both cases, the driving force of membrane transport in the presence of polymer additives was found to be the same. This finding may enable the use of size-exclusion membranes as an alternative to lipid membranes in formulation development of amorphous solid dispersions.

Ranking Itraconazole Formulations Based on the Flux through Artificial Lipophilic Membrane.

PURPOSE: The goal of the study was to evaluate a miniaturized dissolution-permeation apparatus (µFLUX™ apparatus) for its ability to benchmark several itraconazole (ITZ) formulations for which in vivo PK data was available in the literature. METHOD: Untreated and micronized powders of ITZ and various enabling formulations of ITZ (commercial Sporanox® solid dispersion, a Soluplus®-based solid dispersion and a nanosuspension) were introduced to the donor compartment of µFLUX™ apparatus. Donor and acceptor chambers were divided from each other by a lipophilic membrane. In addition to the flux evaluations, changes in solid state as a function of time were investigated to gain further insight into the flux changes observed over time for the solid dispersion formulations. RESULTS: Initial flux values from Sporanox®, the nanosuspension and the micronized ITZ showed ratios of 52/4/1 with a decreasing flux from nanosuspension and both solid dispersions after 2.5-3 h. Although the initial flux from the Soluplus® formulation was 2.2 times lower than the one observed for Sporanox®, the decrease in flux observed was milder and became ~ 2 times higher than Sporanox® after approximately 2.5 h. The total amounts of ITZ in the receiver compartment after 240 min showed the same rank order as the rodent AUCs of these formulations reported in literature. CONCLUSIONS: It was demonstrated that in vitro flux measurements using lipophilic artificial membranes could correctly reproduce the rank order of PK results for ITZ formulations. The drop in flux over time for solid dispersions could be backed by experimental indications of crystallization.

The effect of formulation additives on in vitro dissolution-absorption profile and in vivo bioavailability of telmisartan from brand and generic formulations.

In this study, brand and four generic formulations of telmisartan, an antihypertensive drug, were used in in vitro simultaneous dissolution-absorption, investigating the effect of different formulation additives on dissolution and on absorption through an artificial membrane. The in vitro test was found to be sensitive enough to show even small differences between brand and generic formulations caused by the use of different excipients. By only changing the type of filler from sorbitol to mannitol in the formulation, the flux through the membrane was reduced by approximately 10%. Changing the salt forming agent as well resulted in approximately 20% of flux reduction compared to the brand formulation. This significant difference was clearly shown in the published in vivo results as well. The use of additional lactose monohydrate in the formulation also leads to approximately 10% reduction in flux. The results show that by changing excipients, the dissolution of telmisartan was not altered significantly, but the flux through the membrane was found to be significantly changed. These results pointed out the limitations of traditional USP dissolution tests and emphasized the importance of simultaneously measuring dissolution and absorption, which allows the complex effect of formulation excipients on both processes to be measured. Moreover, the in vivo predictive power of the simultaneous dissolution-absorption test was demonstrated by comparing the in vitro fluxes to in vivo bioequivalence study results.

Using Potentiometric Free Drug Sensors to Determine the Free Concentration of Ionizable Drugs in Colloidal Systems.

The present study investigates the use of free drug sensors (FDS) to measure free ionized drug concentrations in colloidal systems, including micellar solutions, emulsions, and lipid formulations during in vitro lipolysis. Diphenhydramine hydrochloride (DPH) and loperamide hydrochloride (LOP) were selected as model drugs. Self-diffusion nuclear magnetic resonance studies were performed and confirmed the entrapment of drugs in micelles in Brij 35 and sodium taurodeoxycholate (TDC)/phosphatidylcholine (PC) micellar solutions. The FDS measurements indicated that with a constant level of drug, the percentage of free DPH and LOP decreased from 84% to 57% and from 51% to 18%, respectively, as the concentration of Brij 35 was increased from 4.7 to 22 mM; and from 99% to 46% and from 100% to 21%, respectively, as the concentration of TDC/PC was increased from 0.49/0.04 to 8.85/0.78 mM. During the in vitro lipolysis of a lipid formulation, free drug concentration decreased with lipolysis time. The percentage of free DPH was higher than for LOP in the same colloidal system because DPH is less lipophilic than LOP. The study showed that FDS can be used to monitor the free drug concentration in colloidal systems with fast response, no sample treatment and simple data analysis.

Investigation and Mathematical Description of the Real Driving Force of Passive Transport of Drug Molecules from Supersaturated Solutions.

The aim of this study was to investigate the impact of formulation excipients and solubilizing additives on dissolution, supersaturation, and membrane transport of an active pharmaceutical ingredient (API). When a poorly water-soluble API is formulated to enhance its dissolution, additives, such as surfactants, polymers, and cyclodextrins, have an effect not only on dissolution profile but also on the measured physicochemical properties (solubility, pKa, permeability) of the drug while the excipient is present, therefore also affecting the driving force of membrane transport. Meloxicam, a nonsteroidal anti-inflammatory drug, was chosen as a poorly water-soluble model drug and formulated in order to enhance its dissolution using solvent-based electrospinning. Three polyvinylpyrrolidone (PVP) derivatives (K30, K90, and VA 64), Soluplus, and (2-hydroxypropyl)-ß-cyclodextrin were used to create five different amorphous solid dispersions of meloxicam. Through experimental design, the various formulation additives that could influence the characteristics of dissolution and permeation through artificial membrane were observed by carrying out a simultaneous dissolution-permeation study with a side-by-side diffusion cell, µFLUX. Although the dissolution profiles of the formulations were found to be very similar, in the case of Soluplus containing formulation the flux was superior, showing that the driving force of membrane transport cannot be simplified to the concentration gradient. Supersaturation gradient, the difference in degree of supersaturation (defined as the ratio of dissolved amount of the drug to its thermodynamic solubility) between the donor and acceptor side, was found to be the driving force of membrane transport. It was mathematically derived from Fick's first law, and experimentally proved to be universal on several meloxicam containing ASDs and DMSO stock solution.

Spectrophotometric pKa determination of ionizable pharmaceuticals: Resolution of molecules with weak pH-dependent spectral shift.

The extent of ionization of a drug molecule at different pH values can be characterized by its pKa (acid dissociation constants). It is an important parameter in pharmaceutical development to rationalize the physiochemical and biopharmaceutical properties of the drug molecule. UV titration for pKa determination is one of the popular methods. The success of this method requires the molecule exhibiting strong pH-dependent spectral shift related to the ionization process. Depending on the proximity between the ionizable group and the chromophore, the spectral shift may not be strong enough to warrant a successful determination. In a previous study, it has been reported that a distance of three σ bonds between the chromophore and the ionizable group was the limit for a precise pKa determination. In this work, a UV titration method for pKa determination, with a particular emphasis on molecules with weak pH-dependent spectral shift is investigated. It has been shown that the pKa values determined from this study are in good agreement with those determined using potentiometric method and literature data (R(2)=0.998). Our methodology revealed that successful pKa determination is feasible even with a separation distance of five σ bonds between the chromophore and the ionizable group.

Physicochemical selectivity of the BBB microenvironment governing passive diffusion--matching with a porcine brain lipid extract artificial membrane permeability model.

PURPOSE: To mimic the physicochemical selectivity of the blood-brain barrier (BBB) and to predict its passive permeability using a PAMPA model based on porcine brain lipid extract (PBLE 10%w/v in alkane). METHODS: Three PAMPA (BD pre-coated and PBLE with 2 different lipid volumes) models were tested with 108 drugs. Abraham solvation descriptors were used to interpret the in vitro-in vivo correlation with 282 in situ brain perfusion measurements, spanning over 5 orders of magnitude. An in combo PAMPA model was developed from combining measured PAMPA permeability with one H-bond descriptor. RESULTS: The in combo PAMPA predicted 93% of the variance of 197 largely efflux-inhibited in situ permeability training set. The model was cross-validated by the "leave-many-out" procedure, with q(2) = 0.92 ± 0.03. The PAMPA models indicated the presence of paramembrane water channels. Only the PBLE-based PAMPA-BBB model with sufficient lipid to fill all the internal pore space of the filter showed a wide dynamic range window, selectivity coefficient near 1, and was suitable for predicting BBB permeability. CONCLUSION: BBB permeability can be predicted by in combo PAMPA. Its speed and substantially lower cost, compared to in vivo measurements, make it an attractive first-pass screening method for BBB passive permeability.

Dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media.

A series of poorly soluble BCS class II compounds with "grease ball" characteristics were assessed for solubility and dissolution rate in biorelevant dissolution media (BDM) with the purpose of investigating which molecular structures gain most in solubility when dissolved under physiologically relevant conditions. The compounds were studied in four media (simulated intestinal fluid in fasted (FaSSIF pH 6.5) and fed state (FeSSIF pH 5.0), and their corresponding blank buffers (FaSSIF(blk) and FeSSIF(blk))) at a temperature of 37 °C. The experimental results were used to analyze which molecular characteristics are of importance for the solubility in BDM and for in silico modeling using multivariate data analysis. It was revealed that a majority of the compounds exhibited a higher dissolution rate and higher solubility in the FaSSIF and FeSSIF than in their corresponding blank buffers. Compounds which were neutral or carried a positive charge were more soluble in FeSSIF than FaSSIF. The acidic compounds displayed clear pH dependency, although the higher concentration of solubilizing agents in FeSSIF than FaSSIF also improved the solubility. Five of the ten compounds were upgraded to BCS class I when dissolved in FaSSIF or FeSSIF, i.e., the maximum dose of these compounds given orally was soluble in 250 mL of these BDMs. Lipophilicity as described by the log D(oct) value was identified as a good predictor of the solubilization ratio (R(2) = 0.74), and computed molecular descriptors were also shown to successfully predict the solubilities in BDM for this data set. To conclude, the physiological solubility of "grease ball" molecules may be largely underestimated in in vitro solubility assays unless BDM is used. Moreover, the results herein indicate that the improvement obtained in BDM may be possible to predict from chemical features alone.

Miniaturization of powder dissolution measurement and estimation of particle size.

The objective was to investigate the applicability and limitations of an approach for estimating particle size from powder dissolution measurement using as little as 50 microg of sample in 1 ml of buffer solutions. The powder dissolution profiles of five sparingly-soluble drugs (hydrochlorothiazide, phenazopyridine hydrochloride, 2-naphthoic acid, indomethacin, and dipyridamole) were evaluated with a novel biexponential spherical particle equation and also the Wang-Flanagan spherical particle non-sink equation. The results were compared to particle sizing based on measured specific surface area by the Brunauer-Emmett-Teller (BET) method, and also based on Coulter counting. With the exception of hydrochlorothiazide, the model compounds indicated some agglomeration in the dissolution media. The dry-state specific surface area was larger than expected from either the Coulter method or the powder-dissolution data, especially for phenazopyridine hydrochloride. The particle radii estimated by the powder dissolution method ranged from 10 to 68 microm, with equilibrium solubilities spanning from 5 microg/ml (dipyridamole) to 911 microg/ml (hydrochlorothiazide). Powder dissolution data collected with the miniaturized apparatus can be used to determine particle size, with estimated values agreeing reasonably with those measured by the Coulter counter method.

Powder dissolution method for estimating rotating disk intrinsic dissolution rates of low solubility drugs.

PURPOSE: The objective was to investigate the applicability and limitations of a novel approach for measuring intrinsic dissolution rates (IDR) of very small quantities of compounds introduced as powders to buffered solutions and comparing these results to disk IDR obtained using the traditional Wood's apparatus. METHODS: The powder dissolution profiles of 13 model drugs were determined at 37 degrees C in USP buffers at pH 1.2, 4.5, and 6.8, stirred at 100 RPM. As little as 0.06 mg of drug were added to 1 mL buffer media. Drug concentration was measured by an in situ fiber optic UV method. The results were converted to rotating disk IDR values by a novel mathematical procedure. RESULTS: The comparison of the powder-based IDR values to those obtained by traditional Wood's apparatus indicated r(2) = 0.97 (n = 26). CONCLUSION: The results demonstrate that using potentially 10,000-fold less drug material does not sacrifice the quality of the measurement, and lends support to an earlier study that the disk IDR measurement may possibly serve as a surrogate for the BCS solubility classification.

Absorption-excipient-pH classification gradient maps: sparingly soluble drugs and the pH partition hypothesis.

This study of sparingly soluble model drugs assesses (a) how pH and the aqueous boundary layer factors may affect in vitro and in vivo absorption, (b) to what extent single excipients (sodium taurocholate, hydroxypropyl-beta-cyclodextrin, KCl, propylene glycol, methylpyrrolidone, and polyethylene glycol 400) can mitigate adverse absorption effects, and (c) how a novel rank-order visualization tool can be applied in high-throughput screening to identify promising single-excipient effects on the absorption potential of test compounds. The products of accurately measured solubility and artificial-membrane permeability (PAMPA) values at pH 5.0, 6.2, and 7.4, fully taking into account factors such as aqueous boundary layer resistance, membrane retention, and the formation of drug dimers and trimers, were used to define a flux function. A "self-organized" data visualization tool based on the flux function was mined for the promising excipient-drug combinations. In excipient-free solutions, most of the compounds studied formed aggregates. The presence of an excipient predominantly lowered permeability, but most often not by the same amount as solubility was elevated. The compounds with absorption potential most helped by excipients were: clotrimazole>griseofulvin>progesterone>dipyridamole>glibenclamide>mefenamic acid>butacaine>astemizole. The HP-beta-CD effect observed for albendazole and glibenclamide appeared to follow Cmax trends in published pharmacokinetics studies. A surprising outcome of the in vitro measurements was that the classical pH Partition Hypothesis can be "inverted" in its monotonicity by sparingly soluble compounds.

Solubility-excipient classification gradient maps.

This study assessed the effect of excipients (sodium taurocholate, 2-hydroxypropyl-f-cyclodextrin, potassium chloride, propylene glycol, 1-methyl-2-pyrrolidone, and polyethylene glycol 400) on the apparent intrinsic solubility properties of eight sparingly soluble drugs (four bases, two neutrals, and two acids): astemizole, butacaine, clotrimazole, dipyridamole, griseofulvin, progesterone, glibenclamide, and mefenemic acid. Over 1,200 UV-based solubility measurements (pH 3-10) were made with a high-throughput instrument. New equations, based on the "shift-in-pKa" method, were derived to interpret the complicated solubility-pH dependence observed, and poorly predicted by the Henderson-Hasselbalch equation. An intrinsic solubility-excipient classification gradient map visualization tool was developed to rank order the compounds and the excipients. In excipient-free solutions, all of the ionizable compounds formed either uncharged or mixed-charge aggregates. Mefenamic acid formed anionic dimers and trimers. Glibenclamide displayed a tendency to form monoanionic dimers. Dipyridamole and butacaine tended to form uncharged aggregates. With strong excipients, the tendency to form aggregates diminished, except in the case of glibenclamide. We conclude that a low-cost, compound-sparing, and reasonably accurate high-throughput assay which can be used in early screening to prioritize candidate molecules by their eventual developability via the excipient route is possible with the aid of the "self-organized" intrinsic solubility-excipient classification gradient maps.

PAMPA--a drug absorption in vitro model. 5. Unstirred water layer in iso-pH mapping assays and pKa(flux)--optimized design (pOD-PAMPA).

Iso-pH mapping unstirred parallel artificial membrane permeability assay (PAMPA) was used to measure the effective permeability, P(e), as a function of pH from 3 to 10, of five weak monoprotic acids (ibuprofen, naproxen, ketoprofen, salicylic acid, benzoic acid), an ampholyte (piroxicam), five monoprotic weak bases (imipramine, verapamil, propranolol, phenazopyridine, metoprolol), and a diprotic weak base (quinine). The intrinsic permeability, P(o), the unstirred water layer (UWL) permeability, P(u), and the apparent pK(a) (pK(a)(flux)) were determined from the pH dependence of logP(e). The underlying permeability-pH equations were derived for multiprotic weak acids, weak bases and ampholytes. The average thickness of the unstirred water layer on each side of the membrane was estimated to be nearly 2000 microm, somewhat larger than that found in Caco-2 permeability assays (unstirred). Since the UWL thickness in the human intestine is believed to be about forty times smaller, it is critical to correct the in vitro permeability data for the effect of the UWL. Without such correction, the in vitro permeability coefficient of lipophilic molecules would be indicative only of the property of water. In single-pH PAMPA (e.g. pH 7.4), the uncertainty of the UWL contribution can be minimized if a specially-selected pH (possibly different from 7.4) were used in the assay. From the analysis of the shapes of the log P(e)-pH plots, a method to improve the selection of the assay pH, called pK(a)(flux)-optimized design (pOD-PAMPA), was described and tested. From an optimally-selected assay pH, it is possible to estimate P(o), as well as the entire membrane permeability-pH profile.