Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2019 Jan 14;16(3):1245–1254. doi: 10.1021/acs.molpharmaceut.8b01200

Use of PBPK Modeling To Evaluate the Performance of DissolvIt, a Biorelevant Dissolution Assay for Orally Inhaled Drug Products

Mireille Hassoun †, Maria Malmlöf ‡,§, Otto Scheibelhofer ∥, Abhinav Kumar †, Sukhi Bansal †, Ewa Selg ‡, Mattias Nowenwik ‡, Per Gerde ‡,§, Snezana Radivojev ∥, Amrit Paudel ∥,⊥, Sumit Arora ∥, Ben Forbes †,*
PMCID: PMC6503535  PMID: 30640475

Abstract

graphic file with name mp-2018-01200v_0008.jpg

The dissolution of inhaled drug particles in the lungs is a challenge to model using biorelevant methods in terms of (i) collecting a respirable emitted aerosol fraction and dose, (ii) presenting this to a small volume of medium that is representative of lung lining fluid, and (iii) measuring the low concentrations of drug released. We report developments in methodology for each of these steps and utilize mechanistic in silico modeling to evaluate the in vitro dissolution profiles in the context of plasma concentration–time profiles. The PreciseInhale aerosol delivery system was used to deliver Flixotide aerosol particles to DissolvIt apparatus for measurement of dissolution. Different media were used in the DissolvIt chamber to investigate their effect on dissolution profiles, these were (i) 1.5% poly(ethylene oxide) with 0.4% l-alphaphosphatidyl choline, (ii) Survanta, and (iii) a synthetic simulated lung lining fluid (SLF) based on human lung fluid composition. For fluticasone proprionate (FP) quantification, solid phase extraction was used for sample preparation with LC–MS/MS analysis to provide an assay that was fit for purpose with a limit of quantification for FP of 312 pg/mL. FP concentration–time profiles in the flow-past perfusate were similar irrespective of the medium used in the DissolvIt chamber (∼0.04–0.07%/min), but these were significantly lower than transfer of drug from air-to-perfusate in isolated perfused lungs (0.12%/min). This difference was attributed to the DissolvIt system representing slower dissolution in the central region of the lungs (which feature nonsink conditions) compared to the peripheral regions that are represented in the isolated lung preparation. Pharmacokinetic parameters (Cmax, Tmax, and AUC0-∞) were estimated from the profiles for dissolution in the different lung fluid simulants and were predicted by the simulation within 2-fold of the values reported for inhaled FP (1000 μg dose) administered via Flixotide Evohaler 250 μg strength inhaler in man. In conclusion, we report methods for performing biorelevant dissolution studies for orally inhaled products and illustrate how they can provide inputs parameters for physiologically based pharmacokinetic (PBPK) modeling of inhaled medicines.

Keywords: Flixotide Evohaler, fluticasone, PreciseInhale, isolated perfused lungs, simulated lung fluid, Survanta

1. Introduction

In vitro dissolution testing is well established for enteral solid dosage forms for quality control purposes, for comparing products under drug classification frameworks, and for predicting drug pharmacokinetics in vivo.1−4 The therapeutic effect of an inhaled particulate aerosol is only realized after drug release into solution; thus, investigating the dissolution of solid particle aerosol dosage forms has attracted interest.5−8 Dissolution testing for orally inhaled products (OIP) is currently a “hot topic” with research groups adapting a panoply of adaptations of pharmacopoeial apparatus for aerosol collection and dissolution to function as in vitro tests for discerning the quality attributes of inhaled medicines. The latest developments in oral biopharmaceutics demonstrate convincingly that biorelevant methods are important if dissolution testing is to be used as an in vivo predictive tool and realize its full potential in a regulatory context and to predict clinically relevant performance.3,4

The complexity of biorelevant dissolution for inhaled products derives from the need to capture representative aerosol particles in a dispersed manner that reflects their deposition in the lungs, to present the particles to low volumes of lung fluid-like dissolution medium, and to measure reliably the low mass of drug delivered by aerosol medicines. Of the systems reported to date,5−11 none accommodates all these features. The disparate OIP dissolution methods that have been studied tend to be nonintegrated and utilize large volumes of dissolution medium, which precludes the use of a dissolution medium that represents human lung lining fluid.12,13 For some studies of poorly soluble drugs, the medium has been supplemented by addition of protein or phospholipid components, e.g., surfactants such as DPPC6,14 or lung surfactant preparations such as Survanta.15 However, biorelevant media are either expensive or difficult to prepare, and often represent only the surfactant component of distal respiratory tract lining fluid, with the highly abundant proteins absent.

Recently, an integrated apparatus has been developed by Inhalation Sciences for depositing aerosols to a flow past dissolution cell,16 comprising the PreciseInhale and DissolvIt systems, respectively. The PreciseInhale can deliver carefully controlled doses of aerosols from powder inhalers or pressurized metered dose inhalers to the DissolvIt system, in which particle dissolution can be followed by simultaneous observation of aerosol particles using microscopy and measurement of dissolved drug transferred to a flow-past perfusate. Although DissolvIt addresses various limitation of dissolution systems used for OIP, the dissolution vessel contains 5.7 μL of a poly(ethylene oxide) (PEO) gel as the dissolution matrix rather than a biorelevant medium. Due to the novelty of the system, there is little reported data on the performance of the system in predicting dissolution.16,17

To study clinically relevant scenarios, dissolution studies to date have focused on the dissolution of poorly soluble inhaled drugs, in particular fluticasone proprionate (FP).10,11,18 Delivery of FP to the DissolvIt with different biorelevant media in the chamber permits comparison to FP dissolution–absorption profiles in other systems, e.g., isolated perfused lungs (IPL). To perform these experiments requires accurate quantification of submicromolar concentrations of FP using a sensitive assay and an efficient extraction method.19,20 Liquid-chromatography with tandem mass spectrometric detection (LC–MS/MS) provides selective and sensitive analysis of glucocorticoids in biological fluids.21−23 However, poor repeatability using reported methods21−23 required development of a new solid phase extraction (SPE) method, which was reliable and quick and required minimal sample preparation and solvent use.

The value of in vitro systems is in providing decision-making data, e.g., dissolution measurements for predicting and modeling impacts on drug pharmacokinetics in the early stages of the drug development process. Such data can expedite drug development and prevent unexpected toxico-kinetics and ultimately avoid costly end-stage failures.24 Reliable predictive models for pharmacokinetics depend on selecting appropriate mathematical approaches, and more current studies tend to utilize in silico techniques.25−27 For modeling dissolution, Backman et al. have described how mechanistic models may aid in obtaining a better understanding of dissolution, which can be used to predict systemic exposure (AUC) and hence its influence on drug therapeutic effect.28 For this study, a mechanistic model was developed to evaluate the dissolution data derived from the biorelevant approach using the DissolvIt system.

In summary, the aim of the present study was to develop a biorelevant dissolution method by utilizing simulated lung fluid in the DissolvIt system. To measure the dissolution of FP, a LC–MS/MS method was validated for measurement of low drug concentrations. The effect of dissolution medium on FP aerosol particle dissolution was investigated using three different media: (i) 1.5% poly(ethylene oxide) + 0.4% l-alphaphosphatidyl choline, (ii) Survanta , and (iii) a synthetic simulated lung lining fluid (SLF), synthesized based on human lung fluid composition.29,30 Finally, an in silico model based on the method of Boger et al.31 was adapted to explore the impact of the dissolution rates derived on pharmacokinetics.

2. Experimental Section

2.1. Materials

Flixotide 50 μg Evohaler (GSK), poly(ethylene oxide) (PEO), and l-alphaphosphatidyl choline were supplied by Sigma-Aldrich Limited (Dorset, UK), whereas Survanta was obtained from Abbvie Ltd. (Berkshire, UK). The chemicals required for the production of SLF and the preparation of SLF were carried out according to a recently published method.30 For solid phase extraction validation, the chemicals included were micronized FP (USP grade, purity 98%) supplied by LGM Pharma Inc. (Boca Raton, USA), pentadeuterated FP (FP-d5; USP grade, purity 97%) by Insight Biotechnology Limited (Wembley, UK), and rabbit serum, purchased from Sigma-Aldrich Company Limited (Dorset, UK). Chemicals needed for the extraction procedure were zinc sulfate powder, supplied by VWR International Limited (Lutterworth, UK), HPLC-gradient grade acetonitrile, 35% v/v ammonium hydroxide solution, and Analytical-Reagent grade dichloromethane, which were all purchased from Fischer Chemical (Loughborough, UK). The materials required for aerosolization, deposition, and dissolution of FP were provided by Inhalation Sciences, Sweden. For FP dissolution in rat IPL, female CD IGS (Sprague–Dawley) rats were obtained from Charles River (Sulzfeld, Germany), and the necessary equipment was provided by Inhalation Sciences, Sweden.

2.2. Preparation of Calibration Curve and Validation of Assay

Primary stock solutions of FP and FP-d5 were prepared by adding 1 mg of FP or FP-d5 into a 10 mL volumetric flask and filled to the volume with pure acetonitrile, producing 100 μg/mL solutions, and stored at −20 °C. A 1 μg/mL FP working solution was prepared by the appropriate dilution of the stock with pure acetonitrile. The calibration standards (156, 313, 625, 1250, 2500, 5000, and 10,000 pg/mL) were prepared from serial dilution of the working solution with pure acetonitrile. Method validation was conducted in terms of linearity, precision (intraday and interday), accuracy, limit of detection, and limit of quantification. Linearity was evaluated by plotting a calibration curve of mean peak area ratio of FP/FP-d5 (n = 9) against the concentrations of seven standards, using a weighted (1/x) linear regression model. The coefficient of variation (%CV) was calculated across three calibration sets prepared on the same day for intraday precision. For interday precision, another three fresh series of calibration standards prepared on days 2 and 3 were analyzed. Accuracy of the data was also evaluated across nine determinants of each standard, ensuring it was within 15% of each standard concentration. The limit of detection (LOD) and limit of quantification (LOQ) were calculated based on eqs 1 and 2, respectively.19

2.2. 1
2.2. 2

where SD is the standard deviation of the y estimate (peak area ratio) and slope is the gradient of the line.

2.3. Deposition and Dissolution of FP Aerosol in the DissolvIt System

The aerosolization of Flixotide was carried out by connecting the Flixotide pMDI canister to the US Pharmacopeia Induction Port No. 1 (standardized simulation of the throat) of the PreciseInhale aerosol system from Inhalation Sciences (Stockholm, Sweden) (Figure 1). The aerosol particles were deposited on nine circular microscope glass coverslips, 13 mm in diameter, and the dissolution of the deposited particles was investigated by interfacing the particles with the dissolution medium in the DissolvIt dissolution system from Inhalation Sciences (Stockholm, Sweden),16 thermostated to 37 °C. Prewarmed dissolution medium, 5.7 μL of PEO, Survanta, or SLF, was applied to the polycarbonate membrane (pore size 0.03 μm) of each DissolvIt dissolution chamber, with the perfusate buffer streaming on the other side. The flow past perfusate consisted of 0.1 M phosphate buffer containing 4% w/v albumin solution, mixed using a magnetic stirrer. The perfusate was degassed using helium to remove excess bubbles and streamed at a flow rate of 0.4 mL/min over a period of 4 h with samples collected by an automated fraction collector at 0, 3, 6, 9, 12, 15, 20, 25, 30, 40, 50, 60, 120, and 240 min.

Figure 1.

Figure 1

Open in a new tab

Schematic diagram of (a) fluticasone propionate aerosolization and particle deposition and (b) the dissolution system.

2.4. Dissolution of FP Aerosol in Rat Isolated Perfused Lungs

Female rats with body weight 279 ± 20 g were euthanized with phenobarbital sodium (100 mg/kg, i.p.), and their whole lungs were maintained ex vivo as described in other reports.32,33 The lungs were placed in the artificial thoracic chamber. They were ventilated with room air at 75 breaths/min by creating an alternating negative pressure (−0.2 to −0.8 kPa)3 inside the chamber, using an Ugo Basile model 7025 animal respirator (Varese, Italy), with a stroke volume of 6 mL, superimposed on a constant vacuum source connected to the chamber. The tracheal air flow velocity and pressure inside the chamber were measured with a heated Hans Rudolph 8430 series pneumotachograph (Kansas City, USA) at 0–3 L/min and a differential pressure transducer from EMKA Technologies (Paris, France), respectively. The physiological lung-function variables: tidal volume (Vt), dynamic lung compliance (Cdyn),34 and lung conductance (Gaw), which is inversely proportional to lung resistance (RL),34 were calculated from each breath in real time and logged by a data acquisition system using the EMKA Technologies software IOX v. 6.1a. The lungs were perfused via the pulmonary artery in a single-pass mode, at a constant hydrostatic pressure of approximately 12 cm H2O, and the perfusate reservoir was continually overflowing into a recirculation drain pipe, in order to keep a constant liquid pressure head. Throughout the experiments, the perfusate flow rate after the passage through the lungs (Qperf) was measured gravimetrically using a custom-made fraction collector with a balance. The perfusion medium consisted of Krebs–Henseleit buffer, 5.5 mM glucose, 12.6 mM HEPES, and 4% w/v bovine serum albumin. The temperature of the perfusate and the artificial thoracic chamber were maintained at 37 °C. The lungs were left to stabilize for 30 min prior to aerosol exposures, and only the lung preparations with stable baseline values for Vt, Cdyn, Gaw, and Qperf during at least a 15 min period were used. The measured values were Vt = 1.8 ± 0.2 mL, Cdyn = 6.6 ± 1.0 mL/kPa; Gaw = 279 ± 20 mL/s/kPa, and Qperf = 32 ± 2 mL/min (n = 6). Administration of Flixotide aerosol to the IPL was carried out using the PreciseInhale system as described above, where the aerosol was delivered to the lungs by the active dosing system, and the system automatically terminated the exposure when the inhaled target dose was reached. The perfusate was sampled using an automatic fraction collector over a 2 h period from the start of the aerosol exposure with sampling intervals of 4.5, 6, 7.5, 9, 12, 15, 30, 60, and 120 min. After the end of the perfusion period, the lungs and trachea were harvested for analysis of the amount of FP retained in the tissues after the perfusion period to enable mass balance calculations. The experiments were approved by a local ethical review board in Stockholm.

2.5. Sample Extraction

Samples were prepared for analysis following a new solid phase extraction method. Each sample, 325 μL, was loaded into a deep-well sample plate from Thermo-Scientific (Surrey, UK) followed by 50 μL of internal standard (0.1 μg/mL FP-D5). Zinc sulfate 0.1 M, 300 μL, followed by 75 μL of 10% ammonium hydroxide were added and mixed using a multichannel pipet. The SPE plate was placed on an orbital shaker for 30 min followed by centrifugation at 3700 rpm for 5 min. The samples were then transferred to a preconditioned Evolute Express ABN 10 mg SPE 96-well plate by Biotage (Uppsala, Sweden) and washed by applying low vacuum with 200 μL HPLC-grade water followed by 200 μL of 25% v/v methanol in water. The analytes were eluted twice with 200 μL of pure acetonitrile, once with 100 μL of dichloromethane then vacuum centrifuged to dryness. Samples were reconstituted with 30 μL of 55% v/v acetonitrile in water and sonicated rapidly for 10 min. Finally, an aliquot of the sample (20 μL) was injected into the LC–MS/MS system.

2.6. FP Quantification Using LC–MS/MS

Quantification of FP was carried out by Waters Xevo TQ tandem quadrupole mass spectrometer by Waters (Elstree, UK) equipped with an ESI interface, coupled with a Waters Acquity Ultra High Performance LC system (UPLC), equipped with a binary solvent delivery system. Chromatographic separations were carried out on a Waters Acquity UPLC BEH C18 column 130 Å, 1.7 μm, 2.1 × 50 mm. The mobile phase was a mix of mobile phase A and mobile phase B, which were 0.1% ammonium hydroxide in water and 1:1 v/v acetonitrile in water, respectively. The flow rate of the mobile phase was 0.2 mL/min with a 2 min gradient from 50% to 95% B. Argon was used as the collision gas and the collision energy was set at 12 V. The LC–MS/MS operations were controlled by the computer software, MassLynx 4.1, and analyte quantification was performed with multiple reaction monitoring using the following transitions: m/z 501.4 > 313.1 for FP and m/z 506.4 > 313.1 for FP-d5.

2.7. Data Analysis

For the validation process, peak integrations and data analysis were performed using the MassLynx 4.1 computer software. The relationship between peak area ratio and FP concentration (pg/mL) was calculated using the LINEST function in Microsoft Excel. Data was expressed as the mean ± standard deviation of replicate determinations, where n ≥ 3. For the DissolvIt system, the FP transferred to the perfusate was expressed as a percent of the deposited amount on the glass slide. For statistical analysis, one-way ANOVA was applied to the data followed by Tukey post-hoc analysis, using the IBM SPSS version 22 software. Data was identified as statistically significant when p ≤ 0.05.

2.8. Mechanistic Modeling

2.8.1. Simulation of Plasma Concentration–Time Profiles of Fluticasone

A mechanistic physiologically based pharmacokinetic (PBPK) model for predicting the fate of inhaled FP (as illustrated in Figure 2) was developed using Java (version 1.8.0_111, Oracle, Redwood City, US). The integration of the system of ordinary differential equations was performed via the 8(5,3) Dormand–Prince integrator35 as realized in the Apache Commons Math library version 3.6.1 from Apache Software Foundation (Forest Hill, US). The model was adapted from that published by Boger et al.31 Briefly, the model was based on the respiratory physiology divided into three compartments; extra-thoracic, tracheobronchial (central lung), and alveolar (peripheral lung) region (Figure 2). The particles deposited in the extra thoracic region were swallowed and transferred to gut, where they were subjected to systemic absorption, based on their bioavailable fraction (F). Particles deposited in the central and peripheral lung regions were modeled for their dissolution in epithelial lung lining fluid, using input from the in vitro dissolution experiments in DissolvIt system. The in vitro data were fitted to a Weibull function to extract the shape and time scale parameters that were then used to model the dissolution of particles in the model. FP permeation in lung tissues and mucociliary clearance of particles deposited in the central lung were modeled as described by Boger et al.31 The central and peripheral lung areas were perfused by the bronchial blood flow (Q_central lung) and entire cardiac output (Q_cardiac output), respectively. Perfusion-rate limited distribution was assumed to apply for all tissues. System-specific input parameters for central lung, peripheral lung, blood flows, and volume of the tissue compartments are provided as Supporting Information (Tables S1 and S2).

Figure 2.

Figure 2

Open in a new tab

Schematic diagram representing the whole body physiologically based pharmacokinetic (PBPK) model.

For regional lung deposition modeling, the particle size distribution of the tested formulations was determined using next generation impactor (NGI), resulting in a discrete distribution of seven particle sizes with corresponding mass fraction deposited (f0, ..., f6). Multiple-Path Particle Dosimetry model MPPD V2.11 2009 from Applied Research Associates Inc. (Albuquerque, US) was used to calculate the regional deposition of particles from the tested formulations. A breathing pattern with 2 s inspiration, 1 s expiration, 10 s breath hold, and a tidal volume of 625 mL was used.36 The Yeh-Shum 5-lobe lung model was chosen for the calculations of regional deposition fraction.37 The drug and formulation specific parameters for FP inhaled in the model are provided as Supporting Information (Table S3).

2.8.2. Sensitivity Analysis of Dissolution Kinetics

A sensitivity analysis of the pharmacokinetic parameters to the in vitro dissolution kinetics of FP was performed using the mechanistic PBPK model (described in section 2.8.1). Hypothetical in vitro dissolution profiles of FP were created by means of numerical approximation with maximum cumulative percent dissolved fixed to mimic the cumulative percent of FP in SLF. The numerical approximations were selected in order to probe three different possible in vitro dissolution scenarios: a profile where release greatly exceeded that observed experimentally in SLF (case 1) and two profiles that are similar to SLF but initially more rapid (case 2) or slower (case 3). The data was fitted to a Weibull function to extract the shape (b) and time scale (a) parameters of these profiles (Table 1). The Weibull eq (eq 3) was applied to describe the hypothetical dissolution curves and used as an input to the PBPK model. It describes the accumulated fraction of the drug (m) in solution at time t. The location parameter (Ti) is the lag time before the onset of the dissolution and, in all investigated cases, was zero.

2.8.2. 3
Table 1. Fitted Weibull Shape Factor (b) Together with Pharmacokinetic Data of FP Following Its Dissolution in SLF and Artificial Dissolution Profiles (Cases 1–3)a.
parameter SLF* case 1** case 2** case 3**
Weibull shape parameter 1.5285 ± 0.08 3.0204 1.1508 1.8716
Cmax (μg/L) 0.74 ± 0.05 4.61 1.44 0.53
Tmax (h) 3.01 ± 0.58 0.50 0.75 6.00
AUC0-∞ (μg/L·h) 6.46 ± 0.08 6.92 6.87 6.04
Open in a new tab
a

*n = 3; **n = 1.

3. Results

3.1. Extraction and Quantification of Fluticasone Propionate Using LC–MS/MS

As published methods for FP analysis21−23 proved difficult to replicate with adequate reproducibility and sensitivity, a new SPE method for sample preparation was developed for use with LC–MS/MS for the assay of FP in biorelevant media. The methodology was easy to perform, and the relationship between the mean peak area ratio of FP/FP-d5 and the concentration of FP in the samples was linear (R2 value = 0.999) with interday and intraday precision (CV) being <20% (in according to ICH guidelines), except for 156 pg/mL. The accuracy for all FP standard concentrations was within 85–115% (Figure 3). The LOD and LOQ were 106 and 312 pg/mL, respectively. Since the FP concentrations in all dissolution experiments fell within the upper range of the assay, the method was fit for purpose.

Figure 3.

Figure 3

Open in a new tab

Validation of the solid phase extraction and LC–MS/MS assay of fluticasone propionate (FP): (a) linearity of the mean peak area ratio vs concentration; (b) FP concentration, precision, and accuracy. Data expressed as mean ± SD (n = 9).

3.2. Dissolution of FP in DissolvIt and IPL

The penetration of FP, manifested as perfusate concentration, was higher at all time points when the dissolution medium was PEO or Survanta with lipid content lower than that of SLF (Figure 4), in good agreement with the theoretical models. However, overall the influence of medium on FP dissolution was limited since the difference in the FP perfusate concentration values were not statistically significant (one-way ANOVA, p > 0.05) between dissolution in any of three lung fluids at most time points, except the difference in FP concentration for PEO and SLF at 20 min. The FP concentration–time profile in perfusate was also similar between PEO and Survanta, both reaching a Cmax at approximately 20 min. The cumulative percent of FP transferred into the perfusate over time in the DissolvIt system showed similar profiles in each dissolution medium reflecting the ranking observed in the perfusate concentrations, whereas administration to the rat IPL resulted in concentrations of FP and cumulative % of FP in the perfusate that were significantly higher at nearly all time points (Figure 5).

Figure 4.

Figure 4

Open in a new tab

(a) Protein and lipid concentration in poly(ethylene oxide) in phosphate buffer solution (PEO), simulated lung lining fluid (SLF), and Survanta and (b) concentration of FP in the perfusate over time following dissolution in PEO, SLF, and Survanta normalized to mass deposited on the glass coverslips. **Difference in FP concentration in PEO and SLF is statistically significant (one-way ANOVA, p < 0.05). Data expressed as mean ± SD (n = 3).

Figure 5.

Figure 5

Open in a new tab

(a) Concentration of FP in the perfusate over time following dissolution in poly(ethylene oxide) in buffer solution (PEO), simulated lung lining fluid (SLF), Survanta, and rat isolated perfused lung (IPL). *Difference in FP concentration in IPL and SLF is statistically significant (one-way ANOVA, p < 0.05). **Difference in FP concentration in IPL and the remaining three lung fluids, PEO, SLF, and Survanta is statistically significant (one-way ANOVA, p < 0.05). (b) Cumulative % of FP transferred into the perfusate over time, following its dissolution in PEO, SLF, Survanta, and IPL. Data expressed as mean ± SD (n = 3).

3.3. In Silico Modeling of FP Dissolution

Pharmacokinetic parameters (Cmax, Tmax, and AUC0-∞), calculated from the simulated plasma concentration time profiles for the different lung fluid simulants, predicted within two-folds the observed pharmacokinetic parameters of inhaled FP (1000 μg dose) administered via Flixotide Evohaler 250 μg strength inhaler38 (Figure 6). No significant difference was found between the clinically observed and simulated pharmacokinetic parameters when in vitro dissolution input from PEO and Survanta was used in the developed PBPK model. However, differences (p > 0.05) in Cmax and AUC0-∞ compared to the clinical data were found when the slower in vitro dissolution of FP in SLF was modeled. The AUC0-∞ predicted by the model for all three media were slightly underestimated owing to the underestimation of terminal time points of plasma concentration–time profile of inhaled FP suggesting that FP is retained for longer in the airways, which if incorporated into the model would improve the simulation.

Figure 6.

Figure 6

Open in a new tab

In silico modeling. (a) Simulated plasma concentration of FP over time, following its dissolution in poly(ethylene oxide) in buffer solution (PEO), simulated lung lining fluid (SLF), and Survanta. (b) Pharmacokinetic data of FP absorbed in plasma from healthy volunteers, after inhalation of FP pMDI (in vivo) and of FP absorbed in perfusate, following its dissolution in PEO, SLF, and Survanta. Data expressed as mean ± SD (n = 3 or 9).

To understand the sensitivity of the predicted PK parameters toward the dissolution profiles of FP, different hypothetical dissolution profiles were created (Figure 7). In the cases where the dissolution–time curves differed from the SLF profile only in terms of faster or slower initial rate (cases two and three), a similar shape parameter described the exponential curves (b ≈ 1). Fitting of an immediate release type hypothetical dissolution profile (case one) resulted in a value describing a sigmoidal curve (b ≫ 1). Calculated values of AUC for the cases were similar to the values generated for SLF, which reflect the fixing of the cumulative percentage of dissolved FP to 9.34% in 4 h. Differences were observed in terms of Cmax and Tmax with profiles when drug dissolution was faster/slower than in vitro dissolution profile of FP in SLF. Dissolution profiles mimicking the faster dissolution rates (case one and case two) predicted higher values of Cmax (6- and 2-fold), and lower values of Tmax (6- and 4-fold) compared to the values observed in SLF.

Figure 7.

Figure 7

Open in a new tab

Sensitivity testing using numerical approximation to derive three dissolution profiles that vary from the experimental observations for dissolution of fluticasone in SLF (observed): a profile where release greatly exceeded that observed experimentally in SLF (case 1) and two profiles that are similar to dissolution SLF but initially more rapid (case 2) or slower (case 3).

4. Discussion

The use of different dissolution media in the DissolvIt dissolution assay was investigated. A PEO-based medium is used as the “standard” solvent for the DissolvIt system and possesses a lipid content of 4 mg/mL, which was lower than that of SLF (5.4 mg/mL; Figure 4a). Survanta is a lung surfactant extract concentrate and was diluted (1:5 with water) to normalize the lipid concentration to that of PEO. PEO has no biological relevance beyond providing a viscosity that could be regarded as analogous to that provided by respiratory mucus in the airways.39 The slower appearance of FP in the perfusate when using SLF compared to PEO or Survanta may reflect slower dissolution or greater retention of FP as a result of the drug preferentially residing or becoming trapped within the more abundant lipid/lamellar structures in SLF, which also contains cholesterol. Cholesterol can form tight nanodomain complexes with DPPC, stabilizing DPPC in lipid structures in which FP can be solubilized and retained.40

Appearance of a low-soluble inhalant in perfusate or plasma is a serial process of dissolution in lung lining fluid followed by diffusion through the air-to-blood barrier. The second step is controlled by barrier thickness and lipid content and distribution within the barrier. While the mathematics of transport in such two-phase heterogeneous barriers was established decades ago,41,42 the concept was later investigated for lipophilic toxicants in the airway lining layer.43 By adding a small amount of surfactant to an aqueous model of the airway lining layer, the penetration of lipophilic benzo(a)pyrene through the experimental barrier was greatly reduced.44 Thus, a higher content of disperse lipids SLF would be expected to reduce penetration of lipophilic drugs.

Although the simulations in this study were based entirely on human parameters, including the ratio of central/peripheral aerosol deposition, the ex vivo rat IPL model was used as a comparator for experimentally determined dissolution–permeation profiles. The PreciseInhale system provides the advantage of a common delivery platform that can be used to deliver accurate dose and identical respirable aerosol fractions from the pMDI to the in vitro dissolution apparatus and ex vivo model. The concentration of FP and cumulative proportion of FP in the perfusate was significantly higher at nearly all time points following administration to the rat IPL compared to DissolvIt. The higher rate of absorptive clearance was attributed to the IPL possessing a comparatively rapid peripheral (alveolar) dissolution–permeation component in addition to slower central (airway) dissolution–permeation. In contrast, the DissolvIt system is hypothesized to model better the dissolution and absorptive clearance mechanisms in the central airways. In the central regions of the lungs, nonsink conditions may be expected as the dose is distributed over a smaller area compared to the alveolar region, and dissolved FP molecules are required to diffuse across the 5–20 μm pseudostratified epithelium, compared to 1–2 μm in the alveoli of the lungs, to reach the perfusate.17 The DissolvIt system possesses an effective dissolution area of 0.95 cm2, and the penetration distance is approximately 60 μm. Despite being an ex vivo nonhuman model, the IPL is an adaptable tool for teasing out the contributions of dissolution and permeation in different regions of the lungs to drug absorption and local exposure.

As FP exhibits dissolution rate-limited absorption from the lungs of humans,31,45 modeling was carried out to understand the sensitivity of simulated plasma concentration–time profiles of inhaled FP to dissolution profiles. When faster dissolution rates compared to the values observed in SLF were modeled (Figure 7), the higher predicted higher values of Cmax and lower values of Tmax were obtained as a result of higher drug concentration in solution during the early stages of the dissolution process. Where the initial rate of in vitro dissolution was lower than that in SLF, a lower Cmax and higher Tmax value were predicted. This showed clearly the ability of the developed PBPK model to respond to the differences in the in vitro dissolution profiles and translate the differences to the respective PK parameters despite the rapid peripheral dissolution and absorption implied by the IPL studies being unaccounted. These results illustrate how dissolution profiles can have significant impact on the PK parameters of a poorly soluble inhaled drug and demonstrate the application of biorelevant in vitro assays together with PBPK modeling.

5. Conclusion

We report the development of experimental methods for performing biorelevant dissolution studies for orally inhaled products illustrated by a study into the impact of the dissolution of FP, an archetypal poorly soluble inhaled drug, on plasma pharmacokinetics when the drug was delivered using Flixotide. The in silico model was able to translate the in vitro data for FP dissolution in the lungs into impacts on physiologically relevant simulated plasma concentration–time profiles. This approach can lead to enhanced understanding regarding how dissolution processes of inhaled poorly soluble drugs may influence absorptive clearance from the lungs.

Acknowledgments

M.H. was supported by a BBSRC-CASE studentship (BB/K012762/1) hosted at King’s College London.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.molpharmaceut.8b01200.

  • System-specific input parameters for humans; system-specific input parameters for the central lung and peripheral lung in humans; drug and formulation specific input parameters for fluticasone propionate; data obtained from FP absorption and concentration profile in the perfusate (PDF)

Author Present Address

∇ Certara UK, Simcyp Division, Level-2 Acero, 1 Concourse Way, Sheffield S1 2BJ, U.K.

Author Contributions

# Denotes equal contribution to authorship.

Author Contributions

○ Denotes joint senior authors.

The authors declare no competing financial interest.

Supplementary Material

mp8b01200_si_001.pdf (294.1KB, pdf)

References

  1. US FDA CDER . Guidance for Industry: Dissolution Testing of Immediate Release Solid Oral Dosage Forms. http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm070237.pdf.
  2. Cardot J. M.; Davit B. M. In vitro-In Vivo Correlations: Tricks and Traps. AAPS J. 2012, 14 (3), 491–499. 10.1208/s12248-012-9359-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Grady H.; Elder D.; Webster G. K.; Mao Y.; Lin Y.; Flanagan T.; Mann J.; Blanchard A.; Cohen M. J.; Lin J.; Kesisoglou F.; Hermans A.; Abend A.; Zhang L.; Curran D. Industry’s View on Using Quality Control, Biorelevant, and Clinically Relevant Dissolution Tests for Pharmaceutical Development, Registration, and Commercialization. J. Pharm. Sci. 2018, 107, 34–41. 10.1016/j.xphs.2017.10.019. [DOI] [PubMed] [Google Scholar]
  4. Lennernas H.; Lindahl A.; Peer A. V.; Oliier C.; Flanagan T.; Lionberger R.; Nordmark A.; Yamashita S.; Yu L.; Amidon G. L.; Fischer V.; Sjögren E.; Zane P.; McAllister M.; Abrahamsson B. Vivo Predictive Dissolution (IPD) and Biopharmaceutical Modelling and Simulation: Future Use of Modern Approaches and Methodologies in a Regulatory Context. Mol. Pharmaceutics 2017, 14 (4), 1307–1314. 10.1021/acs.molpharmaceut.6b00824. [DOI] [PubMed] [Google Scholar]
  5. Arora D.; Shah K. A.; Halquist M. S.; Sakagami M. In vitro aqueous fluid-capacity-limited dissolution testing of respirable aerosol drug particles generated from inhaler products. Pharm. Res. 2010, 27, 786–795. 10.1007/s11095-010-0070-5. [DOI] [PubMed] [Google Scholar]
  6. Son Y. J.; Horng M.; Copley M.; McConville J. T. Optimization of an in vitro dissolution test method for inhalation formulations. Dissolution Technol. 2010, 17, 6. 10.14227/DT170210P6. [DOI] [Google Scholar]
  7. May S.; Jensen B.; Wolkenhauer M.; Schneider M.; Lehr C. M. Dissolution techniques for in vitro testing of dry powders for inhalation. Pharm. Res. 2012, 29, 2157–2166. 10.1007/s11095-012-0744-2. [DOI] [PubMed] [Google Scholar]
  8. Rohrschneider M.; Bhagwat S.; Krampe R.; Michler V.; Breitkreutz J.; Hochhaus G. Evaluation of the transwell system for characterisation of dissolution behaviour of inhalation drugs: effects of membrane and surfactant. Mol. Pharmaceutics 2015, 12, 2618. 10.1021/acs.molpharmaceut.5b00221. [DOI] [PubMed] [Google Scholar]
  9. Riley T.; Christopher D.; Arp J.; Casazza A.; Colombani A.; Cooper A.; Dey M.; Maas J.; Mitchell J.; Reiners M.; Sigari N.; Tougas T.; Lyapustina S. Challenges with developing in vitro dissolution tests for orally inhaled products (OIPs). AAPS PharmSciTech 2012, 13 (3), 978–989. 10.1208/s12249-012-9822-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Franek F.; Fransson R.; Thörn H.; Backman P.; Andersson P. U.; Tehler U. Ranking in vitro dissolution of inhaled micronized drug powders including a candidate drug with two different particle sizes. Mol. Pharmaceutics 2018, 15, 5319–5326. 10.1021/acs.molpharmaceut.8b00796. [DOI] [PubMed] [Google Scholar]
  11. Bhagwat S.; Schilling U.; Chen M. J.; Wei X.; Delvadia R.; Absar M.; Saluja B.; Hochhaus G. Predicting Pulmonary Pharmacokinetics from In Vitro Properties of Dry Powder Inhalers. Pharm. Res. 2017, 34, 2541–2556. 10.1007/s11095-017-2235-y. [DOI] [PubMed] [Google Scholar]
  12. Hatch G. Comparative biochemistry of airway lining fluid. Treatise on pulmonary toxicology 1992, 1, 617–632. [Google Scholar]
  13. Meyer K. C.; Sharma A.; Brown R.; Weatherly M.; Moya F. R.; Lewandoski J.; Zimmerman J. J. Function and composition of pulmonary surfactant and surfactant-derived fatty acid profiles are altered in young adults with cystic fibrosis. Chest 2000, 118 (1), 164–174. 10.1378/chest.118.1.164. [DOI] [PubMed] [Google Scholar]
  14. Marques M. R. C.; Loebenberg R.; Almukainzi M. Simulated biological fluids with possible application in dissolution testing. Dissolution Technol. 2011, 18, 15–28. 10.14227/DT180311P15. [DOI] [Google Scholar]
  15. Davies N. M.; Feddah M. R. A novel method for assessing dissolution of aerosol inhaler products. Int. J. Pharm. 2003, 255, 175–187. 10.1016/S0378-5173(03)00091-7. [DOI] [PubMed] [Google Scholar]
  16. Gerde P.; Malmlöf M.; Havsborn L.; Sjöberg C.; Ewing P.; Eirefelt S.; Ekelund K. DissolvIt: An in vitro method for simulating the dissolution and absorption of inhaled dry powder drugs in the lungs. Assay Drug Dev. Technol. 2017, 15 (2), 77. 10.1089/adt.2017.779. [DOI] [PubMed] [Google Scholar]
  17. Börjel M.; Selg E.; Gerde P.. In Vitro- Ex Vivo Correlation of Fluticasone Propionate Pharmacokinetic Profiles; DDL: Edinburgh, 2015. [Google Scholar]
  18. Hastedt J.; Bäckman P.; Clark A.; Doub W.; Hickey A.; Hochhaus G.; Kuehl P.; Lehr C.; Mauser P.; McConville J.; Niven R.; Sakagimi M.; Weers J. Scope and relevance of a pulmonary biopharmaceutical classification system AAPS/FDA/USP Workshop March 16–17th, 2015 in Baltimore, MD. AAPS open 2016, 2, 1. 10.1186/s41120-015-0002-x. [DOI] [Google Scholar]
  19. US FDA CDER . Guidance for Industry: Bioanalytical Method Validation. http://www.fda.gov/downloads/Drugs/Guidances/ucm070107.pdf.
  20. Lombardi C. Solid phase extraction. Chemistry in New Zealand 2015, 88–90. [Google Scholar]
  21. Krishnaswami S.; Mollmann H.; Derendorf H.; Hochhaus G. A sensitive LC-MS/MS method for the quantification of fluticasone propionate in human plasma. J. Pharm. Biomed. Anal. 2000, 20, 123–129. 10.1016/S0731-7085(99)00246-0. [DOI] [PubMed] [Google Scholar]
  22. Matuszewski B. K.; Constanzer M. L.; Chavez-Eng C. M. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal. Chem. 2003, 75, 3019–3030. 10.1021/ac020361s. [DOI] [PubMed] [Google Scholar]
  23. Li Y. N.; Tattam B. N.; Brown K. F.; Seale J. P. A sensitive method for the quantification of fluticasone propionate in human plasma by high-performance liquid chromatography/atmospheric pressure chemical ionisation mass spectrometry. J. Pharm. Biomed. Anal. 1997, 16 (3), 447–452. 10.1016/S0731-7085(97)00073-3. [DOI] [PubMed] [Google Scholar]
  24. Di L.; Goosen T. C.; Lai Y.; Steyn S. J.; Varma M. V.; Obach S.; Feng B. A Perspective on the Prediction of Drug Pharmacokinetics and Disposition in Drug Research and Development. Drug Metab. Dispos. 2013, 41 (12), 1975–1993. 10.1124/dmd.113.054031. [DOI] [PubMed] [Google Scholar]
  25. Frohlich E.; Mercuri A.; Wu S.; Salar-Behzadi S. Measurements of Deposition, Lung Surface Area and Lung Fluid for Simulation of Inhaled Compounds. Front. Pharmacol. 2016, 7, 181. 10.3389/fphar.2016.00181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Gaohua L.; Wedagedera J.; Small B. G.; Almond L.; Romero K.; Hermann D.; Hanna D.; Jamei M.; Gardner I. Development of a Multicompartment Permeability-Limited Lung PBPK Model and Its Application in Predicting Pulmonary Pharmacokinetics of Antituberculosis Drugs. CPT: Pharmacometrics Syst. Pharmacol. 2015, 4 (10), 605–613. 10.1002/psp4.12034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Chen A.; Yarmush M. L.; Maguire T. Physiologically Based Pharmacokinetic Models: Integration of In Silico Approaches with Micro Cell Culture Analogues. Curr. Drug Metab. 2014, 13 (6), 863–880. 10.2174/138920012800840419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Bäckman P.; Adelman H.; Petersson G.; Jones C. B. Advances in inhaled technologies: understanding the therapeutic challenge, predicting clinical performance, and designing the optimal inhaled product. Clin. Pharmacol. Ther. 2014, 95 (5), 509–20. 10.1038/clpt.2014.27. [DOI] [PubMed] [Google Scholar]
  29. Kumar A.; Terakosolphan W.; Hassoun M.; Vandera K.; Novicky A.; Harvey R.; Royall P.; Bicer E. M.; Eriksson J.; Edwards K.; Hollanders K.; Valkenborg D.; Nelissen I.; Hassall D.; Mudway I. S.; Forbes B. A biocompatible synthetic lung fluid based on human respiratory tract lining fluid composition. Pharm. Res. 2017, 34 (12), 2454–2465. 10.1007/s11095-017-2169-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hassoun M.; Royall P. G.; Harvey R. D.; Forbes B. Design and development of a biorelevant simulated human lung fluid. J. Drug Delivery Sci. Technol. 2018, 47, 485–491. 10.1016/j.jddst.2018.08.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Boger E.; Evans N.; Chappell M.; Lundqvist A.; Ewing P.; Wigenborg A.; Friden M. Systems Pharmacology Approach for Prediction of Pulmonary and Systemic Pharmacokinetics and Receptor Occupancy of Inhaled Drugs. CPT: Pharmacometrics Syst. Pharmacol. 2016, 5 (4), 201–10. 10.1002/psp4.12074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kröll F.; Karlsson J. A.; Nilsson E.; Persson C. G.; Ryrfeldt A. Lung mechanics of the guinea-pig isolated perfused lung. Acta Physiol. Scand. 1986, 128, 1–8. 10.1111/j.1748-1716.1986.tb07942.x. [DOI] [PubMed] [Google Scholar]
  33. Sundström E.; Låstbom L.; Ryrfeldt Å.; Dahlén S. E. Interactions among three classes of mediators explain antigen-induced bronchoconstriction in the isolated perfused and ventilated guinea pig lung. J. Pharmacol. Exp. Ther. 2003, 307, 408–418. 10.1124/jpet.103.053546. [DOI] [PubMed] [Google Scholar]
  34. Uhlig S.; Wollin L. An improved setup for the isolated perfused rat lung. J. Pharmacol. Toxicol. Methods 1994, 31 (2), 85–94. 10.1016/1056-8719(94)90047-7. [DOI] [PubMed] [Google Scholar]
  35. Hairer E.; Norsett S. P.; Wanner G.. Solving Ordinary Differential Equations I, 2nd ed.; Springer-Verlag: Berlin, 1993. [Google Scholar]
  36. Ibrahim M.; Verma R.; Garcia-Contreras L. Inhalation drug delivery devices: technology update. Med. Devices: Evidence Res. 2015, 12 (8), 131–139. 10.2147/MDER.S48888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yeh H. C.; Schum G. M. Models of Human-Lung Airways and Their Application to Inhaled Particle Deposition. Bull. Math. Biophys. 1980, 42, 461–480. 10.1007/BF02460796. [DOI] [PubMed] [Google Scholar]
  38. Mackie A. E.; et al. Systemic Exposure to Fluticasone Propionate Administered via Metered-Dose Inhaler Containing Chlorofluorocarbon or Hydrofluoroalkane Propellant. Clin. Pharmacokinet. 2000, 39 (1), 17–22. 10.2165/00003088-200039001-00003. [DOI] [PubMed] [Google Scholar]
  39. Shah S.; Fung K.; Brim S.; Rubin B. K. An in vitro evaluation of the effectiveness of endotracheal suction cathetars. Chest 2005, 128 (5), 3699–3704. 10.1016/S0012-3692(15)52954-1. [DOI] [PubMed] [Google Scholar]
  40. Kim K.; Choi S. Q.; Zell Z. A.; Squires T. M.; Zasadzinski J. A. Effect of cholesterol nanodomains on monolayer morphology and dynamics. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, E3054–E3060. 10.1073/pnas.1303304110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Higuchi W. I.; Higuchi T. Theoretical analysis of diffusional movement through heterogeneous barriers. J. Am. Pharm. Assoc., Sci. Ed. 1960, 49 (9), 598–606. 10.1002/jps.3030490910. [DOI] [Google Scholar]
  42. Ash R.; Barrer R. M.; Petropoulos J. H. Diffusion in heterogeneous media: properties of the laminated slab. Br. J. Appl. Phys. 1963, 14, 854–862. 10.1088/0508-3443/14/12/307. [DOI] [Google Scholar]
  43. Gerde P.; Scholander P. A mathematical model of the penetration of polycyclic aromatic hydrocarbons through the bronchial lining layer. Environ. Res. 1987, 44, 321–334. 10.1016/S0013-9351(87)80241-4. [DOI] [PubMed] [Google Scholar]
  44. Gerde P.; Scholander P. An experimental study on the penetration of polycyclic aromatic hydrocarbons through a model of the bronchial lining layer. Environ. Res. 1989, 48, 287–295. 10.1016/S0013-9351(89)80041-6. [DOI] [PubMed] [Google Scholar]
  45. Edsbacker; et al. Airway Selectivity: An Update of Pharmacokinetic Factors Affecting Local and Systemic Disposition of Inhaled Steroids. Basic Clin. Pharmacol. Toxicol. 2006, 98, 523–536. 10.1111/j.1742-7843.2006.pto_355.x. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

mp8b01200_si_001.pdf (294.1KB, pdf)

Articles from Molecular Pharmaceutics are provided here courtesy of American Chemical Society

RESOURCES