Lipophilic salts of poorly soluble compounds to enable high-dose lipidic SEDDS formulations in drug discovery

European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

Contents lists available at ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics journal homepage: www.elsevier.com/locate/ejpb

Research paper

Lipophilic salts of poorly soluble compounds to enable high-dose lipidic SEDDS formulations in drug discovery Michael Morgen a,⇑, Ajay Saxena b, Xue-Qing Chen c, Warren Miller a, Richard Nkansah a, Aaron Goodwin a, Jon Cape a, Roy Haskell d, Ching Su c, Olafur Gudmundsson c, Michael Hageman c, Anoop Kumar e, Gajendra Singh Chowan b, Abhijith Rao e, Vinay K. Holenarsipur e a

Bend Research Inc., a division of Capsugel, 64550 Research Road, Bend, OR 97703, USA Biopharmaceutics, Biocon Bristol-Myers Squibb R&D Centre (BBRC), Syngene International Ltd., Biocon Park, Plot 2 & 3, Bommasandra IV Phase, Bangalore 560099, India c Discovery Pharmaceutics, Bristol-Myers Squibb USA, Bristol-Myers Squibb Pharmaceutical Research Institute, Route 206, Province Line Road P.O. Box 4000, Princeton, NJ 08543, USA d Discovery Pharmaceutics, Bristol-Myers Squibb Pharmaceutical Research Institute, Bristol-Myers Squibb USA, 5 Research Pkwy, Wallingford, CT 06492, USA e Pharmaceutical Candidate Optimization, Biocon Bristol-Myers Squibb R&D Centre (BBRC), Syngene International Ltd., Biocon Park, Plot 2 & 3, Bommasandra IV Phase, Bangalore 560099, India b

a r t i c l e

i n f o

Article history: Received 31 December 2016 Revised 16 March 2017 Accepted in revised form 20 April 2017 Available online 21 April 2017 Chemical compounds studied in this article: Atazanavir (PubChem CID: 198904-31-3) Keywords: Lipophilic salt Atazanavir 2-naphthalene sulfonic acid Dioctyl sulfosuccinic acid SEDDS formulation development Emulsion

a b s t r a c t Self-emulsifying drug delivery systems (SEDDS) have been used to solubilize poorly water-soluble drugs to improve exposure in high-dose pharmacokinetic (PK) and toxicokinetic (TK) studies. However, the absorbable dose is often limited by drug solubility in the lipidic SEDDS vehicle. This study focuses on increasing solubility and drug loading of ionizable drugs in SEDDS vehicles using lipophilic counterions to prepare lipophilic salts of drugs. SEDDS formulations of two lipophilic salts—atazanavir-2-naphthalene sulfonic acid (ATV-2-NSA) and atazanavir-dioctyl sulfosuccinic acid (ATV-Doc)—were characterized and their performance compared to atazanavir (ATV) free base formulated as an aqueous crystalline suspension, an organic solution, and a SEDDS suspension, using in vitro, in vivo, and in silico methods. ATV-2-NSA exhibited
6-fold increased solubility in a SEDDS vehicle, allowing emulsion dosing at 12 mg/mL. In rat PK studies at 60 mg/kg, the ATV-2-NSA SEDDS emulsion had comparable exposure to the free-base solution, but with less variability, and had better exposure at high dose than aqueous suspensions of ATV free base. Trends in dose-dependent exposure for various formulations were consistent with GastroPlusTM modeling. Results suggest use of lipophilic salts is a valuable approach for delivering poorly soluble compounds at high doses in Discovery. Ó 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Driven by the complexity of pharmacological targets, pharmaceutical pipelines are increasingly filled with poorly watersoluble compounds. The factors contributing to this trend include the addition of lipophilic residues to increase ligand-receptor affinity, the broadening of the chemical space using combinatorial chemistry, and the use of high-throughput screening [1,2]. Highly potent compounds may be discovered through these means, but their poor water solubility can lead to numerous challenges during Abbreviations: 2-NSA, 2-naphthalene sulfonic acid; ATV, atazanavir; ATF-2-NSA, atazanavir-2-naphthalene sulfonic acid; ATV-Doc, atazanavir-dioctyl sulfosuccinic acid; Doc, dioctyl sulfosuccinic acid. ⇑ Corresponding author. E-mail address: [email protected] (M. Morgen).

development—particularly, reduced systemic exposure after oral administration [3]. Poorly water-soluble molecules present a particular formulation challenge for high-dose studies—both in Discovery and in later stages of development. In Discovery, high-dose studies are invaluable in aiding early identification of toxicity—the leading cause of attrition at all stages of drug development [4]. High doses are also required in pharmacokinetic (PK) studies, especially toxicokinetic (TK) studies, where high-exposure multiples are required to evaluate safety margins over the entire range of efficacious exposure. Advances in the pharmaceutical sciences have produced numerous approaches to address the challenges presented by poorly water-soluble compounds, which are designated Class II and IV compounds in the Biopharmaceutics Classification System (BCS). These strategies include reducing particle size to increase

http://dx.doi.org/10.1016/j.ejpb.2017.04.021 0939-6411/Ó 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

the surface area and dissolution rate [5]; solubilization in cosolvents [6]; complexation [7,8]; formulation as solid amorphous dispersions [9]; formation of salt forms [10]; and the use of lipidic systems for lipophilic drugs [11]. Each approach has its limitations. For example, reducing particle size has little impact on solubilitylimited exposure, solubilization in cosolvents is restricted to a limited range of pharmaceutically acceptable excipients, and the binding constant of inclusion complexes is often insufficient to solubilize the required dose using acceptable amounts of complexing agents. Solid amorphous dispersions offer improved solubilization, but are not as facile to prepare as liquid lipid-based formulations. The formation of salt forms and the use of lipidic systems for lipophilic drugs offers particular promise in addressing the need for stable, high-dose formulations. Salt formation using small, often inorganic, counterions is a well-established technique to increase the dissolution rate and kinetic solubility of drug compounds [12–17]. Although these high-energy salts can increase the oral bioavailability of poorly soluble ionizable compounds, this approach can be limited by the disproportionation of salt to the free-drug form, with concomitant crystallization to a low-energy crystal. To prevent this, many investigators have proposed modifying the physiochemical properties of poorly water-soluble drug compounds using lipophilic counterions to form lipophilic salts (sometimes referred to as ‘‘ionic liquids”) as an alternative to high-energy salt forms, to enable more effective formulation, delivery, and oral absorption [18–23]. Lipid-based formulations have numerous desirable features; they are easy to manufacture and can increase stability, enhance absorption, and improve bioavailability [24]. However, the drug solubility in the lipid often limits the maximum dose deliverable using lipid vehicles. It would be advantageous to extend the applicability of lipid formulations to a broader range of compounds, particularly those with lower log P values (e.g., between 2 and 5) and higher melting-point (Tm) values, such as atazanavir, the model drug used in the work described here. Lipophilic ion-drug pairs can markedly increase solubility in lipidic phases through a combination of two mechanisms: (1) the lipophilicity of the counterion, which can help pull the drug into a lipid phase; and (2) the reduced crystalline forces (and Tm) of the ion-drug pair, which tends to increase solubility in all solvents. The increased lipid solubility of such salts can facilitate development of stable, high-concentration formulations for high-dose pharmacokinetic or toxicokinetic studies, which in turn can result in higher exposures than those achievable with formulations of other salts or free forms of the drug. Reports have detailed the use of lipophilic counterions to form ion-drug pairs that enable higher drug loadings in solid lipid nanoparticles [25–27] or into oil-in-water emulsion droplets [28]. More recently, Shadid et al. showed a 2.5-fold increase in oral exposure in rats for sulfasalazine when formulated as an ‘‘ionic liquid” [23]. In addition, Sahbaz et al. demonstrated improved oral exposure in rats for three poorly soluble compounds by formulating them with a variety of lipophilic counterions [22]. The purpose of the work described here is to demonstrate the feasibility of using lipophilic ion-drug pairs (i.e., salts) in lipid vehicles to enable dosing of poorly water-soluble drugs at high concentrations for preclinical studies. Demonstrating that this promising approach is suitable for high-dose testing, particularly in a Discovery setting, would be valuable, since solubilizing high doses of such drugs is difficult in lipid vehicles using free-acid/free-base drug forms or typical high-energy salt forms. This study describes the use of lipophilic counterions to prepare lipophilic salt forms of a weakly basic drug—atazanavir (ATV)—to increase drug solubility in a self-emulsifying drug-delivery system (SEDDS) vehicle (above that achievable with the free base and typ-

213

ical inorganic salts of the drug) to increase the deliverable dose, enabling high-dose testing. The synthesis, characterization, formulation development, and in vivo PK profiles of the lipophilic salts atazanavir 2-naphthalene sulfonic acid (ATV-2-NSA) and atazanavir-dioctyl sulfosuccinic (‘‘docusate”) acid (ATV-Doc) are described. ATV, which is an antiretroviral drug of the protease inhibitor class [29–32], was chosen as a model compound due to its low aqueous solubility (
1.5 lg/mL at pH 6.5) and weakly basic pKa (
4.6). 2-naphthalene sulfonic acid (2-NSA) and dioctyl sulfosuccinic acid (Doc) were chosen as counteracids because of their low pKa values [16], which result in complete protonation of ATV at intestinal pH, as well as for their pharmaceutical precedence [33,34] and low toxicity [35]. The SEDDS formulation chosen is typical of a Type II formulation in the lipid formulation classification system (LFCS) described in the literature [36,37], but was not optimized for ATV.

2. Materials and methods 2.1. Materials All reagents and solvents were used as received. ATV was obtained from Bristol-Myers Squibb (New York, New York). 2NSA was purchased from TCI Fine Chemicals (Portland, Oregon). Docusate sodium and polysorbate 80 (TweenÒ 80) were purchased from Spectrum Chemical (Gardena, California). Tetrahydrofuran (THF, anhydrous) and hydrochloric acid (HCl, 1.25 M) in methanol were obtained from Aldrich Chemical (Milwaukee, Wisconsin). MiglyolÒ 812, a mixture of medium-chain triglycerides, was a kind gift from Cremer Oleo (Hamburg, Germany). Glyceryl monocaprate (CapmulÒ MCM C10) was obtained from Abitec Corporation (Columbus, Ohio). Dimethylacetamide (DMAC), polyethylene glycol 400 (PEG 400), and d6 dimethyl sulfoxide (d6 DMSO, 99.8% deuterated) were purchased from Sigma Aldrich (Belgium and USA). Hydroxypropyl b cyclodextrin (KleptoseÒ HPB) was sourced from Roquette (France). Citric acid monohydrate and sodium hydroxide (NaOH) were purchased from Merck (India). Highperformance liquid chromatography (HPLC)-grade acetonitrile (ACN) and methanol were obtained from Honeywell, Burdick, and Jackson (Muskegon, Michigan). Trifluoroacetic acid (TFA) was purchased from Thermo Scientific (Rockford, Illinois). Phosphate butter solution (PBS) consisted of 82.1 mM NaCl, 20 mM Na2HPO47H2O, and 46.7 mM KH2PO4, adjusted to pH 6.5 with 30% NaOH, and had an osmolarity adjusted to 290 mOsm/kg with 1:20.4 KCl:NaCl. Methylcellulose (MethocelTM A4M) was a kind gift from The Dow Chemical Company. 2.2. Methods 2.2.1. Preparation of amorphous ATV free base Amorphous ATV free base was prepared by spray-drying ATV free base from solution on a custom mini spray dryer designed and built at Bend Research, a division of Capsugel [9]. ATV free base was dissolved in methanol at a concentration of 18 mg/mL and spray-dried at an inlet temperature of 65 °C and a nitrogen flow rate of 30 standard liters per min. 2.2.2. Preparation of ATV salts The ATV salts were prepared quantitatively by stoichiometric addition of strong acids to the ATV free base. The 2-NSA and Doc salts were both isolated in the amorphous form. The 2-NSA salt was also separately isolated as crystalline material. The Doc salt was not easily isolated in the crystalline form, so only the amorphous form was characterized.

214

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

2.2.3. Preparation of ATV-2-NSA salt Amorphous ATV-2-NSA salt was prepared by dissolving 2.27 g of ATV free base in 50 mL of methanol. To this solution, 0.74 g of 2-NSA was added with stirring. The resulting solution was rotoevaporated to dryness and placed under vacuum overnight, yielding an off-white solid. Crystalline ATV-2-NSA salt was prepared by dissolving 1.22 g of ATV free base in 50 mL of methanol. To this solution, 0.42 g of 2NSA in 10 mL of methanol was added with stirring. The resulting solution was rotoevaporated to dryness. The solid was dissolved in 30 mL of methyl acetate with mild heating, and then cooled to ambient temperature. The mixture was allowed to sit at ambient temperature while being monitored for the appearance of crystals over time. After 3 days, there being no further apparent increase in precipitate, the white crystalline solid was isolated by filtration, rinsed with 50 mL of methyl acetate, and then rinsed with 30 mL of diethyl ether and dried. 2.2.4. Preparation of ATV-Doc salt Amorphous ATV-Doc salt was prepared by converting docusate sodium to docusic acid followed by reaction with ATV free base. Docusate sodium (1.380 g) was dissolved in 25 mL of anhydrous THF. Methanolic HCl (2.005 g, 1.25 M) was added with stirring to produce a white precipitate of sodium chloride. The mixture was chilled to 0 °C and then filtered through a 0.45-lm filter, resulting in a clear solution of docusic acid. The docusic acid was added with stirring to a solution of ATV free base, 2.04 g in 75 mL of methanol, giving a clear solution that was rotoevaporated to dryness and further dried under vacuum to give an amorphous solid. Unsuccessful attempts were made to isolate crystalline ATVDoc salt. ATV-DOC (0.445 g) was dissolved in methyl acetate (1.096 g) and placed in a freezer overnight. As this resulted in no precipitation of crystals, diethyl ether (10.34 g), was then added to the solution, which remained clear at room temperature. This solution was then placed in a freezer overnight, with solids being observed the next day. The sample was kept in the freezer for a total of 4 days. Solids were harvested and analyzed by PXRD to confirm the amorphous nature of the solid. 2.2.5. Measurement of ATV content of salts Solid samples of ATV-2-NSA and ATV-Doc salts were dissolved in methanol in triplicate, and the concentration of the active compound (i.e. ATV free base equivalent) was measured using the HPLC method described below. 2.2.6. Preparation of SEDDS vehicle The SEDDS vehicle was prepared by combining Miglyol 812 (35% w/w), Capmul MCM C10 (15% w/w), and Tween 80 (50% w/ w) and heating to 80 °C, followed by stirring and cooling to room temperature. 2.2.7. Measurement of solubility To determine the solubility of ATV free base, ATV-2-NSA, and ATV-Doc in the SEDDS vehicle, SEDDS vehicles containing excess solids of each drug form were stirred for six days at ambient temperature. Aliquots (900 lL) were taken from each suspension in duplicate and centrifuged using a Beckman OptimaTM Max ultracentrifuge (Beckman-Coulter Inc., Danvers, Massachusetts) at 470,000g for 20 min at 23 °C. The supernatants were sampled and diluted 1:10 in methanol and analyzed by HPLC. The amorphous solubility of ATV in PBS at pH 6.5 was determined by adding a concentrated DMSO solution (5 mg/mL) of ATV free base into PBS, and using ultraviolet (UV) probes to measure the maximum concentration reached before light scattering occurred due to phase separation of drug from solution [38–41].

2.2.8. Evaluation of SEDDS and SEDDS emulsion stability To determine the physical stability of SEDDS formulations, solutions of ATV free base, ATV-2-NSA, and ATV-Doc were prepared in the SEDDS vehicles at up to 100 mg/mL and observed visually for precipitation. The clear SEDDS formulations were diluted 1:4 by weight with 0.01 N HCl to prepare SEDDS emulsions. The resulting emulsions were then placed onto a nutating mixer at 24 rpm at room temperature. Aliquots (500 lL) were taken immediately after preparation and at 1- and 2-h time points and transferred to microcentrifuge tubes. The samples were then centrifuged for 1 min at 16,000g. The supernatant was sampled and diluted 1:10 in methanol and analyzed by HPLC. Any precipitate was isolated and characterized by powder X-ray diffraction (PXRD). In addition to the stability of the emulsion with respect to crystallization of the drug, the size and size stability of the emulsion droplet size of SEDDS emulsions were determined by dynamic light scattering (DLS) using a BI-200SM particle size analyzer with a BI-9000AT correlator (Brookhaven Instruments Inc., Holtsville, NY). Measurements were made of SEDDS emulsions at low drug concentration (1– 2 mg/mL) immediately after preparation, and after storage at ambient for 24 h, in order to assess the stability of the lipid droplet size in the emulsion.

2.2.9. Measurement of composition by HPLC The compositions of the isolated ATV salts were determined by measuring the amount of ATV free base equivalent in a given mass of each isolated solid. Drug-containing samples were analyzed by reverse-phase (RP) HPLC using a Hewlett Packard instrument (Series 1100, Hewlett Packard Development Corp., Palo Alto, California) with a C18 column (Agilent ZorbaxÒ SB-C18, 3.5 lm, 75 mm  4.6 mm) using an injection volume of 10 lL, column temperature of 25 °C, and flow rate of 1.5 mL/min with UV detection at 254 nm. ATV had a retention time of approximately 1.8 min using an isocratic method and a mobile phase of 60:40 ACN:0.1% TFA in water.

2.2.10. Evaluation of physical state by PXRD ATV free base, ATV-2-NSA, and ATV-Doc were analyzed by PXRD using a Bruker D8 Advance X-ray diffractometer (Bruker Corp., Billerica, Massachusetts) with a LynxeyeTM XE positionsensitive detector and Cu K-alpha radiation (1.54 Å). Approximately 10–20 mg of powder was placed in a zero-backgroundholder (ZBH) sample cup containing a single-crystal silicon surface and covered with a polyether ether ketone (PEEK) dome. The sample cup was rotated at 30 rpm in the horizontal plane, while the goniometer was stepped through a 2h angle of 4° to 40° with a 0.04° step size at 1 s/step. The primary beam optics included a Goebel mirror, 1-mm dispersive slit, and a 2.5° soller slit. The diffracted beam optics included a 5.84 mm antiscatter slit and a 2.5° soller slit.

2.2.11. Evaluation of thermal characteristics by modulated differential scanning calorimetry (mDSC) The glass-transition temperatures (Tg) of the amorphous spraydried ATV free base, ATV-2-NSA, and ATV-Doc isolated by rotoevaporation were determined using a TA Q-1000 mDSC instrument (TA Instruments, New Castle, Delaware). Samples were prepared as 5-mm compacts in hermetic TzeroTM pans (TA Instruments) with open lids, and were equilibrated at room temperature at <5% relative humidity (RH) overnight. Samples were sealed and then run in modulated mode from 20 °C to 200 °C at 2.5 °C/min with ±1.5 °C/ min modulation. Tg values were determined using TA Instrument software and quantified as the half-height of the thermal transition in the reversing heat-flow curve.

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

2.2.12. Assessment of salt formation by nuclear magnetic resonance (NMR) ATV free base, ATV-2-NSA, and ATV-Doc were dissolved at 15– 40 mg/mL in d6 DMSO. d1H NMR spectra were recorded on a Varian 600-MHz Fourier transform (FT) instrument with 64 scans. Chemical shifts were reported in parts per million (ppm). 2.2.13. Preparation of SEDDS formulations for PK studies The SEDDS formulations were prepared by first dissolving amorphous ATV free base (at 10 and 30 mg/mL); amorphous ATV-2-NSA (at 10, 30, and 60 mg/mL); and amorphous ATV-Doc (at 10 and 30 mg/mL) separately in the SEDDS vehicle. SEDDS emulsions were then prepared by diluting the SEDDS formulations 1:4 in 0.01 N HCl before administration. The SEDDS formulations were diluted in acidic media instead of in media of higher pH in order to limit the concentration of ATV free base in the aqueous phase, in which it has a propensity to crystallize. The SEDDS formulations were emulsified prior to dosing in order to ensure accurate transfer of the dose, as the high viscosity of the (unemulsified) SEDDS solution can make it difficult to transfer the formulations from the dosing device to the animal during in vivo studies, leading to variability in exposure. Solution formulations of crystalline ATV free base were prepared (at 10 and 30 mg/mL) using 10% v/v DMAC, 40% v/v PEG 400, and 15% v/v Kleptose HPB, and 35% v/v citrate buffer (citric acid monohydrate, pH 4.0, 50 mM). ATV free base was not soluble at 60 mg/mL in this vehicle, so the highest dose was prepared using amorphous ATV free base in 100% PEG 400. ATV free base was also dosed as an aqueous suspension prepared (at 2 and 20 mg/mL, respectively, for 10 and 100 mg/kg doses) in 0.5% w/w Methocel A4 M, 0.1% w/w Tween 80, and 99.4% w/w water. 2.3. Animals Male Sprague-Dawley rats weighing 280 ± 40 g were obtained from the Syngene International Ltd. in-house breeding facility (Bangalore, India). All animal experiments were conducted in the animal research Syngene facility (Bangalore, India), which is registered by the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA) and accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), after approval by the Institutional Animal Ethics Committee (IAEC). The animals were fed with a standard laboratory rodent diet (Tetragon Chemie Pvt. Ltd., Bangalore, India) and housed at room temperature (22 °C ± 3 °C) and 50% RH ± 20% RH on a 12-h light-dark cycle. All animals were fasted overnight before dosing. Water was provided ad libitum throughout the study. 2.4. PK studies Test Preparation: Cannulas were implanted in the jugular vein of all rats and a recovery period of 24 h was provided after surgery. Cannulated rats (n = 3 per group) were fasted for approximately 16 h before peroral dosing and fed 3 h after dosing. ATV freebase suspension (at 10 and 100 mg/kg); ATV free-base solution (at 10, 30, and 60 mg/kg); ATV free-base SEDDS (at 10 and 30 mg/kg); ATV-2-NSA SEDDS (at 10, 30, and 60 mg/kg); and ATV-Doc SEDDS (at 10 and 30 mg/kg) formulations were administered by oral gavage using a feeding needle (5 mL/kg). Blood samples (150 lL) were collected from the jugular-vein catheter 0.25, 0.5, 1, 3, 5, 7 and 24 h after dosing. The blood sample volume was replaced with an equal volume of 0.9% saline. Blood samples were collected into tubes containing K2EDTA (1.33 mM) and were inverted several times to disperse the anticoagulant. They were then kept chilled on ice until centrifugation, which occurred within

215

30 min of sample collection. Blood was centrifuged at 10,000g for 5 min at 4 °C. Plasma was harvested and transferred into tubes, which were then stored at 70 °C (±10 °C) until the samples were analyzed. Sample preparation for bioanalysis: An ATV calibration standard was prepared in rat plasma by a serial dilution method for the concentration range of 1.22–5000 nM. Aliquots (25 lL) of plasma standards/study samples were quenched with 125 lL of ACN containing an internal standard (200 nM, ritonavir) in a Millipore Solvinert hydrophilic plate and vortexed for 30 s. The samples were then centrifuged at 4000 rpm for 10 min at 4 °C. The eluent (2 lL) was injected onto a mass spectrometry (MS) instrument for analysis. Ultraperformance liquid chromatography (UPLC)-MS/MS method: Ò A Waters Acquity UPLC Integrated System (Milliford, USA) was used to analyze samples on an RP column (Waters BEH C18, 1.7 lm, 2.1  50 mm) maintained at 4 °C ± 2 °C. Samples were analyzed in a shallow gradient (Solvent A: 0.1% formic acid in 10-mM ammonium formate; Solvent B: 0.1% formic acid in ACN) at a flow rate of 0.5 mL/min, with a total run time of 2.5 min. Quantitation was achieved by MS/MS detection in positive-ion multiple reaction monitoring (MRM) mode using an API 4000 MS instrument (Applied Biosystems, Foster City, California, USA) with an electrospray positive ionization (ESI) source at a capillary voltage of 5.5 kV and a source temperature of 500 °C. Quadrupoles Q1 and Q3 were set at unit resolution with a dwell time of 30 ms for each MRM (705.5/335.2). Analytical data were acquired and processed using AnalystÒ software (Version 1.5, Sciex, Framingham, Massachusetts, USA). The accuracy and precision of the ATV calibrations standard was well within the acceptance limit for the bioanalytical method (<±20% at lower limit of quantification [LOQ] and <±15% at other concentrations). PK Data Analysis: The PK parameters for ATV were determined by noncompartmental analysis of plasma concentration-versustime data using KineticaTM software (Version 5.1, Thermo Fisher Scientific Corporation, Philadelphia, Pennsylvania, USA). The peak concentration (Cmax) and time for Cmax (Tmax) were recorded directly from experimental observations. The area under the plasma concentration-time curve (AUC0-t) was calculated using mixed log-linear trapezoidal rule up to the last detectable concentration (Clast) and extrapolated to infinity (AUC0-1) by adding Clast/ kz to AUC0-t. Estimation of the terminal elimination rate constant (kz) was made using a minimum of three time points with quantifiable concentrations. PK parameters were expressed as the mean plus or minus the standard deviation (SD). An analysis of variance (ANOVA) was performed to assess differences between more than two groups, followed by a Tukey multiple comparisons test, wherever appropriate. Comparison between two groups was conducted using the two-tailed t-test using GraphPad PrismÒ software (Version 5.02, (GraphPad Software Inc., San Diego, California, USA). A p-value of less than 0.05 was considered statistically significant. 2.5. In silico PK modeling In silico simulations used to predict drug plasma concentrations from an oral dose in a fasted rat animal model were performed using the GastroPlusTM advanced compartmental and transit (ACAT) model (Version 9.0) by Simulations Plus Inc. Clearance and compartment rate constants were obtained for a three-compartment model, by fitting intravenous (IV) bolus PK data (collected inhouse) for 1- and 5-mg/kg doses. Given the extensive cytochrome P450 (CYP) metabolism of ATV [42–44], we also fit MichaelisMenton kinetic parameters—Vmax (the maximum rate achieved by the enzymatic system, at saturating substrate concentration) and km (the substrate concentration at which the enzymatic

216

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

reaction rate is half of Vmax)—for liver metabolism to the IV bolus data. The PK curves for ATV-2-NSA SEDDS and ATV solution at 10, 30, and 60 mg/kg, and the ATV free-base aqueous suspension administered at 100 mg/kg were used in a global optimization routine to fit Michaelis-Menton kinetic terms for CYP metabolism in the gut and efflux transporters in the gut enterocytes [45]. A pattern search method using the Hooke and Jeeves method was used to match predicted AUC0-t and predicted drug plasma concentration for the average of three sets of in vivo data consisting of the SEDDS and PEG solution formulations dosed at 10, 30, and 60 mg/kg. An overall effective permeability value (Peff) of 0.45 cm/s was estimated from single-pass perfusion studies of rat intestinal segments [45]. A solubility of 1.5 lg/mL at a pH of 6.5 was used for the crystalline free base, and the measured ‘‘amorphous solubility” of 65 lg/mL at pH 6.5 was used for the SEDDS formulation. We assumed this solubility was maintained when the drug was in the un-ionized state. Drug partitioning from the dispersed SEDDS oil phase is expected to be rapid, and thus was taken to be instantaneous in the model. Additionally, it was assumed that the solubilities of the crystalline and amorphous drug forms followed the Henderson-Hasselbalch relationship [46]. 3. Results The purpose of this study was to demonstrate the utility of lipophilic salts to enable dosing of poorly soluble drugs at high concentrations. ATV was used as a model weakly basic drug compound, and 2-NSA and Doc were selected as model lipophilic counterions. The chemical structure and properties of ATV free base, 2-NSA, and Doc are shown in Fig. 1. The solid salt forms of ATV-2-NSA and ATV-Doc were characterized by PXRD, mDSC, and d1H NMR, evaluated for solubility and physical stability in the SEDDS vehicle, and assessed for in vivo performance. 3.1. Characterization of the lipophilic salts Reaction of ATV free base with the two sulfonic acids, 2-NSA and Doc, gave the corresponding salt forms in high yield. The ATV-2-NSA and ATV-Doc salts formed readily, as expected, since the pKa of ATV (4.7) is more than four units higher than the pKa

Fig. 1. Chemical structures and physicochemical properties of ATV free base, 2-NSA, and Doc.

Fig. 2. PXRD results for crystalline ATV free base and amorphous ATV-2-NSA and ATV-Doc.

values of the strongly acidic sulfonic acid groups of 2-NSA and Doc. The ATV-2-NSA and ATV-Doc salts were isolated by rotoevaporation followed by drying under high vacuum, producing solids that were characterized by PXRD. As shown in Fig. 2, diffractograms of the solid powders showed amorphous ‘‘halos” with no evidence of distinct diffraction peaks. The amorphous nature of ATV-2-NSA and ATV-Doc suggested by PXRD analysis was supported by mDSC data, showing clear Tg events in the reversing heat flow of the thermal scans. As shown in Fig. 3 and Table 1, the Tg of amorphous ATV-2-NSA was 15 °C higher than that of the amorphous ATV free base, and the Tg of ATV-Doc was more than 30 °C below that of the amorphous ATV free base. The lower Tg of ATV-Doc relative to the free base may be due to the highly flexible nature of the docusate moiety. The docusate sodium itself has a low Tg (13 °C) (data not shown). Only the ATV free base showed a recrystallization, at
162 °C. The lack of a Tm for the 2-NSA salt in Fig. 3 is consistent with the amorphous nature of this sample. The 2-NSA salt sample that was crystallized from methyl acetate did show a Tm of 172 °C (data not shown). The thermal data and ATV free base content of the salts is summarized in Table 1. As the table shows, the measured free base equivalent content of the ATV-2-NSA and ATV-Doc salt samples was consistent with their theoretical values. d1H NMR spectroscopy was performed to confirm the formation of ATV lipophilic salts. The aromatic region of the d1H NMR spectra of ATV free base, 2-NSA, ATV-2-NSA, and ATV-Doc are shown in Fig. 4, and data on the chemical shifts are shown in Table 2. The spectra were assigned based on the spectrum of 2-phenyl pyridine [47] in conjunction with two-dimensional correlation spectroscopy (COSY), total correlation spectroscopy (TOCSY), and heteronuclear single-quantum correlation (HSAC) spectroscopy of the ATV free base, ATV-2-NSA, and ATV-Doc (data not shown). As shown in the spectra of Fig. 4 and the shift data in Table 2, for ATV-2-NSA and ATV-Doc, the pyridyl protons (A, B, C, and D) shifted significantly downfield (by up to 0.52 ppm) relative to those of ATV free base. The phenyl protons (E and F), which are further removed from the protonatable pyridyl ring, shifted only slightly upfield and downfield, respectively. The benzyl protons (G, HR, and I), which are far removed from the protonatable pyridyl group, showed essentially no shift for the two salt forms relative to their chemical shift for the free-base form. The solubility of ATV free base, ATV-2-NSA and ATV-Doc in the SEDDS vehicle was determined by adding excess solids of each

217

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

Fig. 3. Reversing (top) and nonreversing (bottom) mDSC heat flows for amorphous ATV free base, ATV-2-NSA, and ATV-Doc.

Table 1 Thermal properties, ATV free base equivalent, and solubility of amorphous ATV free base, ATV-2-NSA, and ATV-Doc.

a b c d

Physical form

Tm (°C)

Tg (°C)

Theoretical ATV content (mgA/g)a

Measured ATV content (mgA/g)b

ATV solubility in SEDDS vehicle (mg/mL)

Form isolated after solubility assessment

ATV free base ATV-2-NSA ATV-Doc

204 172 (amorph.)

105 120 73

1000 772 625

N/Ac 796 ± 3 613 ± 2

0.75 4.30 2.8d

ATV free base ATV-2-NSA ATV free base

mgA = mg active of ATV free base equivalent. Uncertainty shown is the SD of triplicate measurements. N/A = not applicable. ‘‘Solubility” of the amorphous form.

drug form to the SEDDS vehicle and stirring for several days to reach equilibrium. For ATV free base and ATV-2-NSA, the crystalline form was used. For ATV-Doc, the amorphous solid was used, since crystalline ATV-Doc could not be isolated. The results are summarized in Table 1. Crystalline ATV-2-NSA had a 5.7-fold higher solubility than crystalline ATV free base in the SEDDS vehicle. The amorphous solubility ATV-Doc in the SEDDS vehicle was approximately 3.7-fold higher than the crystalline solubility of ATV free base. 3.2. Physical stability in SEDDS vehicle and as emulsions The physical stability of the various drug forms in SEDDS formulations at several concentrations was assessed by dissolving the amorphous forms of ATV free base (at 10, 30, and 60 mg/ mL); ATV-2-NSA (at 30 and 60 mg/mL); and ATV-Doc (at 10, 30, and 60 mg/mL) in the SEDDS vehicle. The resulting supersaturated solutions were monitored for precipitation. As anticipated, the precipitation rate increased as the concentration of drug in the SEDDS vehicle increased. ATV free base at 10 and 30 mg/mL stayed in solution for up to 6 and 1.5 h, respectively, whereas at 60 mg/mL, precipitation was observed within 0.25 to 0.5 h. Due to the rapid precipitation, no further tests were conducted with ATV free base at 60 mg/mL. ATV-Doc had better stability in the SEDDS vehicle. At 10, 30, and 60 mg/mL, ATV-Doc solution formulations did not precipitate for up to 6, 3.5, and 1 h, respectively. Of the three drug

forms, ATV-2-NSA had the best stability in the SEDDS vehicle. No precipitation was observed for up to 30 h, even at 60 mg/mL. To determine if changes in physical form occurred when the salts were added to the SEDDS vehicle, PXRD was used to compare the crystalline precipitates isolated from the ATV-2-NSA and ATVDoc SEDDS formulations to those for two controls: crystalline ATV2-NSA (prepared by crystallization from methyl acetate) and ATV free base. As the results in Fig. 5 show, the diffraction patterns for crystalline solids obtained from the ATV-2-NSA and ATV-Doc SEDDS vehicle match those of the crystalline ATV-2-NSA and ATV free-base forms, respectively. This confirms that ATV-2-NSA is a more stable salt in the SEDDS vehicle than ATV-Doc, which disproportionates to a larger extent, precipitating as the ATV free base. SEDDS emulsions were diluted to make dosing easier and to ensure accurate transfer of the dose during in vivo testing. Two potential diluents—0.01 N HCl and water—were tested. A significant improvement in the emulsion stability was observed when the ATV-NSA SEDDS vehicle was diluted with 0.01 N HCl (5% precipitation) versus water (33% precipitation). Based on these results, 1:4 dilution with 0.01 N HCl was used for subsequent tests. SEDDS emulsion formulations were tested for stability to ensure minimal precipitation of drug occurred between preparation of the emulsion and dosing. Table 3 summarizes the stability results of SEDDS emulsions of different formulations as a function of concentration. In this study, the potency of ATV in the supernatant of the SEDDS emulsion after centrifugation was measured

218

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

Fig. 4. d1H NMR spectra of 2-NSA, ATV-2-NSA, ATV-Doc, and ATV free base in d6 DMSO, showing that aromatic protons of the pyridyl ring and, to a lesser extent the phenyl ring, are shifted by protonation of the pyridyl ring by the counteracids.

Table 2 Chemical shifts of aromatic protons for ATV free base, ATV-2-NSA, and ATV-Doc by d1H NMR analysis. Proton position

Proton designation

ATV (ppm, multiplet)

ATV-2-NSA (ppm, multiplet)

ATV-Doc (ppm, multiplet)

Pyr Pyr Pyr Pyr Ph Ph Bz Bz

A B C D E F G, H I

8.62 7.29 7.83 7.88 7.94 7.39 7.16 7.10

8.78 7.75 8.35 8.19 7.91 7.53 7.16 7.10

8.68 7.50 8.06 8.02 7.92 7.46 7.16 7.10

(d) (t) (t) (d) (d) (d) (m) (m)

immediately after preparation (i.e., initial), after 1 h, and after 2 h, and the results were compared to the total potency. As the table shows, at 10 mg/mL, the ATV free-base SEDDS emulsion concentration was stable over 2 h, but significant ATV precipitation (77%) was observed at 2 h for the 30 mg/mL ATV free-base SEDDS emulsion. After dilution, the 30 mg/mL ATV-Doc SEDDS emulsion was more stable than the ATV free-base emulsion at the same concentration, but 1:4 dilution of the 60 mg/mL ATV-Doc SEDDS solution with 0.01 N HCl resulted in 82% precipitation of the ATV by 2 h. When diluted similarly, the 30 and 60 mg/mL ATV-2-NSA SEDDS solutions resulted in emulsions that were fairly stable, with only 5% to 15% loss of ATV due to precipitation. Similar dilutions of 100 mg/mL ATV-2-NSA resulted in more than half the drug precipitating from the resulting emulsion within 2 h. The emulsion dro-

(d) (t) (t) (d) (d) (d) (m) (m)

(d) (t) (d) (d) (d) (m) (m)

plet size for SEDDS emulsions, as determined by effective diameter from DLS, was typically 200–300 nm, and did not change appreciably over the course of 24 h after preparation. 3.3. Initial PK studies SEDDS formulations showing acceptable in vitro stability were selected for in vivo assessment in rats: ATV free base (at 10 and 30 mg/mL), ATV-2-NSA (at 10, 30, and 60 mg/mL), and ATV-Doc (at 10 and 30 mg/mL). Two control formulations—an aqueous suspension containing crystalline ATV free base and an organic solution of ATV free base—were also tested (at 10 and 100 mg/kg, and 10, 30, and 60 mg/kg, respectively for the suspension and solution). The ATV free-base solution was used as a positive control to

219

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

Fig. 5. PXRD results for (a) ATV free base, (b) ATV-2-NSA crystallized from methyl acetate, (c) precipitate from ATV-2-NSA dissolved in SEDDS vehicle (100 mg/mL), and (d) precipitate from ATV-Doc dissolved in SEDDS vehicle (120 mg/mL).

Table 3 Kinetic stability of aqueous emulsions prepared from ATV free base, ATV-2-NSA, and ATV-Doc SEDDS formulations after dilution.

a

Drug form

SEDDS formulation concentration (mg/mL)

Total ATV concentration of SEDDS emulsiona (mg/mL)

Initial

t=1 h

t=2h

ATV free base

10 30

2.0 6.0

1.8 1.9

1.7 1.4

1.7 1.4

85 23

ATV-2-NSA

30 60 100

6.0 12.0 20.0

6.0 12.3 9.5

5.8 11.6 9.1

5.7 10.1 9.0

95 84 45

ATV-Doc

30 60

6.0 12.0

6.3 3.2

5.9 2.6

4.8 2.2

80 18

ATV concentration in the supernatant after centrifugation (mg/mL)

Unprecipitated ATV At 2 h (%)

After 1:4 dilution with HCl.

show the maximum exposure achievable when the dissolved-drug concentration was maximized, whereas the ATV free-base crystalline suspension was used as a second control to show the exposure resulting from the very low dissolved-drug levels supplied by the low-energy crystalline drug form. Exposure increased with increasing dose for all three SEDDS formulations and for the ATV free-base solution formulations. Higher-

than-dose-proportional exposure was observed for SEDDS formulations containing ATV free base and ATV-Doc at 10 and 30 mg/ kg. This is shown by the data in Table 4 and Fig. 6. However, after dose normalization at 10 mg/kg, the difference between the 10 mg/ kg and 30 mg/kg dose groups for these two formulations was not statistically significant, due to high variability observed at 30 mg/ kg and the small number of animals in each group (n = 3). Due to

Table 4 PK parameters (mean ± SD) of ATV following oral administration of ATV SEDDS emulsions, ATV free-base crystalline suspensions, and ATV free-base solutions. Formulation

Dose (mg/kg)

Cmax (lg/mL)

Tmax (h)

AUC0-1 (lg/[mL h])

ATV free-base SEDDS

10 30 10 100 10 30 10 30 60 10a 30a 60b

0.6 (0.1) 1.6 (0.6) 0.4 (0.1) 0.5 (0.3) 0.7 (0.1) 2.3 (0.4) 0.6 (0.1) 2.1 (1.1) 11.4 (1.1) 0.9 (0.2) 3.6 (1.3) 6.7 (4.1)

0.8 1.0 1.0 0.9 1.7 2.3 1.0 2.2 3.0 0.5 1.0 3.0

1.4 (0.0) 6.8 (2.4) 0.8 (0.0) 1.6 (1.1) 2.0 (0.8) 7.3 (2.0) 1.7 (0.4) 8.1 (4.4) 45.0 (5.6) 1.6 (0.2) 14.7 (3.2) 33.7 (18.2)

ATV free-base crystalline suspension (control) ATV-Doc SEDDS ATV-2-NSA SEDDS

ATV free-base solution (control)

(0.4) (0) (0) (0.2) (1.1) (1.2) (0) (1.4) (0) (0) (0) (0)

Conditions: Dosed in rats with jugular-vein cannulas (n = 3). a The ATV free-base solutions at 10 and 30 mg/kg (i.e., 2 mg/mL and 6 mg/mL) were prepared using 10% v/v DMAC, 40% v/v PEG 400, 15% v/v KleptoseÒ HPB, and 35% v/v citrate buffer (pH 4.0, 50 mM). b The ATV free-base solution formulation at 60 mg/kg (i.e., 12 mg/mL) was prepared in 100% v/v PEG 400.

220

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

4. Discussion

Fig. 6. Plasma concentration versus time profile of ATV following oral administration in rats (n = 3) from an ATV free-base solution and an ATV-2-NSA SEDDS formulation at 60 mg/kg.

Fig. 7. ATV exposure in rats for ATV free-base crystalline suspension, free base solution, and ATV-2-NSA SEDDS at different doses. The solid curves show the simulated AUC (GastroPlusTM) as a function of dose.

limited stability of the ATV free-base SEDDS formulation and the ATV-Doc SEDDS formulation at 60 mg/mL (as shown by the data in Table 3), neither of these formulations was evaluated at this concentration. Like the ATV free-base SEDDS and ATV-Doc SEDDS formulations, the ATV-2-NSA SEDDS formulation and ATV free-base solution exhibited higher-than-dose-proportional exposure between 10 mg/kg and 60 mg/kg. The interanimal variability for the ATV2-NSA SEDDS formulation at 60 mg/kg was substantially lower (a coefficient of variation [CV] of 13% for AUC0-1) than for the ATV free-base solution (a CV of 54% for AUC0-1), as shown by the data in Table 4 and Fig. 6. Exposure from the ATV free-base crystalline suspension was low at both doses tested, likely due to solubilitylimited absorption. The dose responses (i.e., exposure) for the ATV-2-NSA SEDDS formulations and the two control formulations are shown in Fig. 7. Also plotted are the simulated dose responses predicted by GastroPlusTM simulations (note that since the simulated curves for the ATV-2-NSA SEDDS and the ATV free-base solution overlay, only the former is plotted). As shown, the simulated curves for the ATV2-NSA SEDDS formulation, the ATV free-base solution, and the ATV free-base suspension overlay the data well. In particular, the model predicts the nonlinear increase in exposure seen in the PK data for the ATV-2-NSA SEDDS and ATV free-base solution formulations, and the low, flat exposure for the ATV free-base suspension.

A number of technologies are available for increasing the dissolution rate or kinetic solubility of compounds with poor aqueous solubility to increase oral exposure in toxicology studies. However, most of these techniques have limitations in formulating for highdose studies—both in early discovery and in later stages of preclinical and clinical drug development. Lipid-based formulations, in particular, are typically most applicable to highly lipophilic compounds (log P > 5). Because of the desirable features of lipidbased formulations, including ease of preparation, it would be advantageous to extend the applicability of this approach to compounds with lower log P values. In this study, 2-NSA and Doc lipophilic counterions were used to prepare lipophilic ion-drug pairs to markedly increase solubility in lipid solvents and facilitate development of stable, highconcentration formulations for high-dose studies in discovery. The physical properties of ATV-2-NSA and ATV-Doc are significantly different from those of the free-base form or the common salt forms of ATV. For example, inorganic salt forms of weakly basic drug compounds typically have Tm values greater than or equal to that of the free base. This is true for ATV, where bisulfate and chloride salts melt at approximately 213 °C and 204 °C, respectively – much higher than the 2-NSA salt (172 °C). ATV-2-NSA’s lower Tm relative to the Tm of the free base, chloride, and bisulfate forms suggests that ATV-2-NSA has a lower crystal lattice energy. The 2-NSA may interfere with the ATV intra- or intermolecular bonding. In addition, the steric hindrance imposed by the bulky naphthalene group of the 2-NSA anion likely results in a crystal structure that is less tightly packed. The crystalline form of ATV-Doc was not easily isolated, likely due to the flexible nature of the docusate moiety, which hinders short- and long-range ordering of the salt. Nevertheless, consistent with the discussion above for the 2-NSA crystalline form, one would also expect a crystalline Doc salt to have a low Tm. The low Tms of the lipophilic salts and their resultant reduced tendency to crystallize increase their solubility in lipid vehicles and enhance emulsion stability, allowing effective formulation at high concentrations to enable high doses. The two lipophilic salts described here have increased solubility in the SEDDS vehicle. ATV-2-NSA has a 5.7-fold higher solubility than ATV free base in the SEDDS vehicle. The increased lipid solubility of ATV-2-NSA versus ATV free base was evident in the physical stability of supersaturated solutions in SEDDS vehicle prepared by dissolving the respective amorphous forms into the lipid. ATV2-NSA prepared at 60 mg/mL, or a supersaturation ratio (c/S) of approximately 12.6, did not undergo any crystallization within 30 h, whereas amorphous ATV free base and ATV-Doc both initially dissolved to 60 mg/mL in the SEDDS vehicle, but crystallized to the free base within 0.5 h and 1.5 h, respectively. This conversion to the free base may suggest ATV-Doc undergoes a larger degree of salt dissociation in the vehicle than does ATV-2-NSA. As discussed in Section 3, the ATV-2-NSA SEDDS emulsion, formed by diluting the SEDDS solution 1:4 in 0.01 N HCl, had better physical stability than either the free-base or ATV-Doc SEDDS formulations (as shown by the data in Table 3). In addition, the physical stability of the SEDDS formulation diluted in HCl was better than that of the SEDDS formulation diluted in water. The improved stability of the ATV-2-NSA SEDDS formulation versus the ATV freebase SEDDS formulation in HCl, and the improved stability of ATV2-NSA SEDDS diluted in HCl versus in water are both likely due to the tendency of the free base to crystallize in an aqueous environment. The use of the ATV-2-NSA salt, particularly in HCl, minimizes the concentration, and therefore the crystallization rate, of the free base.

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

The improved physical stability of the ATV-2-NSA SEDDS emulsion relative to the ATV free-base formulation suggests that the 2NSA salt form can facilitate oral dosing at higher drug concentrations than the free-base form. Here the increased solubility of the 2-NSA salt allowed effective dosing at six times higher concentrations than the ATV free base. That is, the 2-NSA could be dissolved at 60 mg/mL in the lipid vehicle versus 10 mg/mL for the free base, enabling dosing of the SEDDS formulations at 60 mg/kg for the 2NSA versus 10 mg/kg for the free base. Differences in SEDDS solubility and emulsion stability for the ATV-2-NSA, ATV-Doc and ATV free-base formulations did not manifest in exposure differences in vivo at the two lower doses tested (i.e., 10 and 30 mg/kg). Likely, all three formulations sustained dissolved drug concentrations long enough in vivo to allow good absorption of the compound at these lower doses, as evident by the similar AUC values. The exposure for the aqueous suspension of free-base crystals was similar to that of the SEDDS formulations at low dose, suggesting that at these doses, the crystalline solubility is sufficient to provide good absorption. The high permeability of ATV results in rapid absorption to mitigate the limitation of low dissolved drug levels [48]. At 60 mg/kg, the exposure for the ATV-2-NSA SEDDS formulation was similar to that of the ATV free-base solution, and higher than that of the ATV free-base suspension in SEDDS, suggesting that the ATV-2-NSA SEDDS formulation provided maximum dissolved drug levels in the GI tract to drive absorption. Interanimal variability in exposure was higher for the ATV free-base solution formulation than for the ATV-2-NSA SEDDS formulation, possibly due to precipitation of the solution formulation in vivo. This highlights the fact that even when high drug concentrations in cosolvents can be achieved, the risk of in vivo precipitation, which can lead to a reduced or variable exposure, can be significant. Of course, cosolvent intolerability is another common drawback of such solutions, which can limit their use in early preclinical studies. The ATV-crystalline aqueous suspension formulation, which is a surrogate for a clinical solid (tablet or capsule) dosage form, gave a 50% lower AUC at 30 mg/kg compared to SEDDS formulations, likely due to lower solubility and dissolution rate of the crystalline form of ATV. Even at 100 mg/kg, no increase in exposure was observed for the crystalline suspension, limiting its utility for toxicology studies. In this study, ATV exposure increased more than dose proportionally, as the dose increased from 10 to 60 mg/kg in ATV solution and ATV-2-NSA SEDDS formulations (Fig. 7). For solution formulations, 3- and 6-fold increases in dose resulted in 9- and 20-fold increases in AUC0-1, respectively, as the data in Table 4 show. The observed nonlinear PK may have resulted from multiple factors, including saturation of efflux and uptake transporters [45] and enzymatic metabolism [49]. ATV is a known substrate for efflux transporters, such as P-glycoprotein (P-gp) and multidrug resistance-associated proteins (MRPs) [45,50]. Kis et al. showed that in the presence of a P-gp inhibitor, the ATV efflux ratio was reduced to 1.5 from 11.7 [45]. Moreover, when the ATV concentration was increased substantially (to 100 lM from 1 lM), the efflux ratio was reduced to 1, suggesting a saturable process. ATV is a substrate and moderate inhibitor of CYP3A4 and a competitive inhibitor of UGT1A1 [51]. Therefore, along with efflux saturation, saturation of enzymatic metabolism at higher doses may have also contributed to an increase in exposure that was higher than dose proportional [49]. In silico modeling was done in an attempt to understand the dose-dependent exposure seen in the rat PK studies. Results from in silico modeling suggests that the absorption of crystalline ATV is solubility-limited. In particular, the model predicts the flat exposure of the aqueous crystalline suspension as a

221

function of dose from 10 to 100 mg/kg. If absorption was primarily limited by dissolution rate, increasing the crystalline surface area by increasing the dose would be expected to increase the exposure across this range. The increased exposure with dose for the SEDDS and solution formulations is therefore likely due to the higher dissolved drug concentrations for these formulations relative to crystalline solubility. The simulations likewise support the hypothesis that metabolism and efflux contribute to the nonlinear PK, as the metabolism and efflux are overcome as dose increases, leading to a superlinear response with dose for the solution and ATV-2-NSA SEDDS formulations. The increase in exposure with dose is predicted to roll over eventually due to (amorphous) solubility-limited absorption from the ATV-2-SEDDS and ATV free-base solution formulations. Based on the modeling, this roll-off is not expected to occur until above the doses tested in this study. Experimentally, it was shown that the ATV tends to crystallize from the SEDDS formulation before reaching the dose concentrations at which absorption is limited by the amorphous solubility. The present ATV-2-SEDDS formulation precipitates rapidly at the 20 mg/mL dose concentration that would be needed to achieve 100 mg/mL (assuming the typical 5 mg/kg dose volume used here). Because of this, the roll-off in exposure for the current (unoptimized) SEDDS formulation occurs between 12 mg/mL and 20 mg/mL (60 and 100 mg/kg at 5 mL/kg dose volume) due to the limited physical stability of the supersaturated formulations. Increasing the dose volume and/or further optimizing the SEDDS formulation would increase the dose at which the roll-off in exposure occurs. The interpretation of the dose response for various formulations was aided by simulations that demonstrated the impact of metabolism and efflux, particularly at low dose. The results shown here demonstrate that the use of lipophilic salts can enable high doses and dose concentrations of ATV in lipid formulations. Lipophilic salts can be formed for a variety of weakly acidic or weakly basic drug compounds, typically using a counter ion having a pKa at least several units higher (lower) than the pKa of the corresponding acidic (basic) drug molecule, to ensure robust association. The use of lipophilic salts in lipid vehicles is especially attractive for compounds that have moderate to low lipophilicities and/or compounds with high lipophilicities and high Tm. For these compounds, the effective lipophilicity of the compound can be increased, and/or the Tm decreased, both of which raise the solubility of the drug compound in lipophilic formulation vehicles, and in the use environment following dispersion and digestion of the vehicle. Lipophilic salts may find utility in other formulation approaches as well, including in solid amorphous dispersions, in which they can increase the solid state stability by increasing drug solubility in a polymeric dispersion matrix, for example. Based on results here, it is expected that for weakly basic drugs, lipophilic salts incorporated into a liquid lipid-based formulation or into a solid dispersion may limit dissolved drug concentrations in the low pH gastric environment, and therefore the resulting supersaturation and concomitant likelihood of crystallization in the intestine. 5. Conclusion We have shown that increasing the lipophilicity of an ionizable compound by preparing a lipophilic salt is a viable approach for increasing its lipid solubility and improving the stability of the associated SEDDS formulation. In this study, lipophilic salts of ATV with 2-NSA and Doc were prepared as model systems for studying the utility of lipophilic salts of weakly basic active compounds with poor aqueous solubility. The physical properties of the two salts—including Tg, Tm, and solubility—were shown to be significantly different from those of the ATV free base. Most

222

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

significantly, the solubility of crystalline ATV-2-NSA in a SEDDS formulation was determined to be 5- to 6-fold higher than that of the crystalline free base. The ATV-2-NSA SEDDS formulation also had better physical stability than corresponding SEDDS ATV freebase or ATV-Doc formulations. The plasma exposure in rats provided by ATV-2-NSA at a given dose was equivalent to that provided by an ATV free-base SEDDS formulation. However, ATV-2-NSA can be dosed at a higher level using this formulation approach, due to its higher solubility in the lipid SEDDS vehicle. Compared to the ATV free-base solution formulation, the ATV-2-NSA salt at high dose had comparable exposure while reducing interanimal variability in exposure. While these in vivo results are preliminary, with limited animal numbers and an unoptimized vehicle, they suggest this approach shows promise in enabling high-dose testing of low-solubility drug compounds both in early Discovery and in later stages of preclinical and clinical drug development. The comparison of the SEDDS formulation of the ATV-2-NSA lipophilic salt with a suspension of ATV free base in SEDDS or in aqueous vehicles, and with the solution formulation of the free base, shows the potential advantage of using lipophilic salts paired with lipid-based formulations for delivering high doses of low-solubility compounds. In particular, the increased solubility of the lipophilic salt in the lipid vehicle allows higher effective dosing for TK and dose-escalation studies relative to free-base suspensions in either aqueous or lipid vehicles. In addition, lipophilic salts in lipid vehicles provide high dissolved-drug levels that resist precipitation without the use of cosolvents that can lead to uncontrolled precipitation and to adverse effects that interfere with toxicology readouts. The work described here represents a systematic approach using in vitro, in vivo, and in silico assessment that can be used to guide rational selection of lipophilic-salt SEDDS formulations with resulting utility in the drug-development process. In addition, this study provides a framework for using in vitro, in vivo, and in silico assessment to understand and select formulations for toxicology studies. Conflict of interest The authors declare that they have no conflict of interest. Acknowledgments The authors would like to acknowledge the contributions of Sheelendra Pratap Singh, Priyanka Ghule, and Sandhya Mandlekar for the work described in this manuscript. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] C.A. Lipinski, Drug-like properties and the causes of poor solubility and poor permeability, J. Pharm. Tox. Methods 44 (2000) 235–249, http://dx.doi.org/ 10.1016/s1056-8719(00)00107-6. [2] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings, Adv. Drug Del. Rev. 64 (2012), http://dx. doi.org/10.1016/s0196-409x(96)00423-1, 644–17. [3] C.J.H. Porter, N.L. Trevaskis, W.N. Charman, Lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs, Nat. Rev. Drug Disc. 6 (2007) 231–248, http://dx.doi.org/10.1038/nrd2197. [4] J.A. Kramer, J.E. Sagarz, D.L. Morris, The application of discovery toxicology and pathology toward the design of safer pharmaceutical lead candidates, Nat. Rev. Drug Disc. 6 (2007) 636–649, http://dx.doi.org/10.1038/nrd2378. [5] J.C. Chaumeil, Micronization: a method of improving the bioavailability of poorly soluble drugs, Methods Find. Exp. Clin. Pharmacol. 20 (3) (1998) 211. [6] Y. Miyako, N. Khalef, K. Matsuzaki, R. Pinal, Solubility enhancement of hydrophobic compounds by cosolvents: role of solute hydrophobicity on the solubilization effect, Int. J. Pharm. 393 (1–2) (2010) 48–54, http://dx.doi.org/ 10.1016/j.ijpharm.2010.03.059.

[7] T. Loftsson, M.E. Brewster, Pharmaceutical applications of cyclodextrins. 1. Drug solubilization and stabilization, J. Pharm. Sci. 85 (10) (1996) 1017–1025, http://dx.doi.org/10.1021/js950534b. [8] R.A. Rajewski, V.J. Stella, Pharmaceutical applications of cyclodextrins. 2. In vivo drug delivery, J. Pharm. Sci. 85 (11) (1996) 1142–1169, http://dx.doi.org/ 10.1021/js960075u. [9] W.B. Curatolo, J.A.S. Nightingale, S.M. Herbig, Utility of hydroxypropyl methylcellulose acetate succinate (HPMCAS) for Initiation and maintenance of drug supersaturation in the GI milieu, Pharm. Res. 26 (2009) 1419–1431, http://dx.doi.org/10.1007/s11095-009-9852-z. [10] S. Agharkar, S. Lindenbaum, T. Higuchi, Enhancement of solubility of drug salts by hydrophilic counterions: properties of organic salts of an antimalarial drug, J. Pharm. Sci. 65 (5) (1976) 747–749. [11] A.J. Humberstone, W.N. Charman, Lipid-based vehicles for the oral delivery of poorly water soluble drugs, Adv. Drug Del. Rev. 25 (1997) 103–128, http://dx. doi.org/10.1016/s0169-409x(96)00494-2. [12] A.T.M. Serajuddin, Salt formation to improve drug solubility, Adv. Drug Del. Rev. 59 (2007) 603–616, http://dx.doi.org/10.1016/j2007.05.010. [13] S.M. Berge, L.D. Bighley, D.C. Monkhouse, Pharmaceutical salts, J. Pharm. Sci. 66 (1) (1977) 1–19, http://dx.doi.org/10.1002/jps.2600660104. [14] P.L. Gould, Salt selection for basic drugs, Int. J. Pharm. 33 (1–3) (1986) 201– 217, http://dx.doi.org/10.1016/0378-5173(86)90055-4. [15] L.D. Bighley, S.M. Berge, D.C. Monkhouse, Salt forms of drug and absorption, in: J. Swarbrick, J.C. Boylan (Eds.), Encyclopedia of Pharmaceutical Technology, vol. 13, Marcel Dekker Inc, New York, 1996, pp. 453–499. [16] Anon, in: P.H. Stahl, C.G. Wermuth (Eds.), Handbook of Pharmaceutical Salts; Properties, Selection and Use, VHCA and Wiley-VCH, Zurich, Switzerland, 2011, p. 363. [17] G.S. Paulekuhn, J.B. Dressman, C. Saal, Trends in active pharmaceutical ingredient salt selection based on analysis of the orange book database, J. Med. Chem. 50 (26) (2007) 6665–6672, http://dx.doi.org/10.1021/jm701032y. [18] W.L. Hough, R.D. Rogers, Ionic liquids then and now: from solvents to materials to active pharmaceutical ingredients, Bull. Chem. Soc. Japan 80 (12) (2007) 2262–2269, http://dx.doi.org/10.1246/bcsj.80.2262. [19] W.L. Hough, M. Smiglak, H. Rodríguez, R.P. Swatloski, S.K. Spear, D.T. Daly, R.D. Rogers, The third evolution of ionic liquids: active pharmaceutical ingredients, New J. Chem. 31 (8) (2007) 1429, http://dx.doi.org/10.1039/b706677p. [20] J. Stoimenovski, D.R. MacFarlane, K. Bica, R.D. Rogers, Crystalline vs. ionic liquid salt forms of active pharmaceutical ingredients: a position paper, Pharm. Res. 27 (4) (2010) 521–526, http://dx.doi.org/10.1007/s11095-0090030-0. [21] H.D. Williams, Y. Sahbaz, L. Ford, T.-H. Nguyen, P.J. Scammells, C.J.H. Porter, Ionic liquids provide unique opportunities for oral drug delivery: structure optimisation and in vivo evidence of utility, Chem. Comm. 50 (2014) 1688– 1690, http://dx.doi.org/10.1039/c3cc48650h. [22] Y. Sahbaz, H.D. Williams, T.-H. Nguyen, J. Saunders, L. Ford, S.A. Charman, P.J. Scammells, C.J.H. Porter, Transformation of poorly water-soluble drugs into lipophilic ionic liquids enhances oral drug exposure from the lipid based formulations, Mol. Pharm. 12 (6) (2015) 1980–1991, http://dx.doi.org/ 10.1021/mp500790t. [23] M. Shadid, G. Gurau, J.L. Shamshina, B.-C. Chuang, S. Hailu, E. Guan, S.K. Chowdhury, J.-T. Wu, S.A.A. Rizvi, R.J. Griffin, R.D. Rogers, Sulfasalazine in ionic liquid form with improved solubility and exposure, Med. Chem. Comm. 6 (2015) 1837–1841, http://dx.doi.org/10.1039/c5md00290g. [24] C.J.H. Porter, H.D. Williams, N.L. Trevaskis, Recent advances in lipid-based formulation technology, Pharm. Res. 30 (12) (2013) 2971–2975, http://dx.doi. org/10.1007/s11095-013-1229-7. [25] G.P. Zara, R. Cavalli, A. Fundarò, A. Bargoni, O. Caputo, M.R. Gasco, PK of doxorubicin incorporated in solid lipid nanospheres (SLN), Pharmacol. Res. 40 (3) (1999) 281–286, http://dx.doi.org/10.1006/phrs.1999.0509. [26] L. Serpe, M.G. Catalano, R. Cavalli, E. Ugazio, O. Bosco, R. Canaparo, Cytotoxicity of anticancer drugs incorporated in solid lipid nanoparticles on ht-29 colorectal cancer cell line, Eur. J. Pharm. Biopharm. 58 (3) (2004) 673–680, http://dx.doi.org/10.1016/j.ejpb.2004.03.026. [27] R.S. Kalhapure, C. Mocktar, D.R. Sikwal, S.J. Sonawane, M.K. Kathiravan, A. Skelton, T. Govender, Ion pairing with linoleic acid simultaneously enhances encapsulation efficiency and antibacterial activity of vancomycin in solid lipid nanoparticles, Colloid Surf. B 117 (2014) 303–311, http://dx.doi.org/10.1016/ j.colsurfb.2014.02.045. [28] T. Zhang, Y. Zheng, Q. Peng, X. Cao, T. Gong, Z. Zhang, A novel submicron emulsion system loaded with vincristine-oleic acid ion-pair complex with improved anticancer effect: in vitro and in vivo studies, Int. J. Nanomed. 8 (2013) 1185–1196, http://dx.doi.org/10.2147/ijn.s41775. [29] P.J. Piliero, Atazanavir: a novel HIV-1 protease inhibitor, Expert Opin. Investig. Drugs 11 (9) (2002) 1295–1301, http://dx.doi.org/10.1517/13543784.11.9.1295. [30] D.R. Goldsmith, C.M. Perry, Atazanavir. Drugs 63 (16) (2003) 1679–1693. [31] B.L. Musial, J.K. Chojnacki, C.I. Coleman, Atazanavir: a new protease inhibitor to treat HIV infection, Am. J. Health – Syst. Pharm. 61 (13) (2004) 1365–1374. [32] K.F. Croom, S. Dhillon, S.J. Keam, Atazanavir—a review of its use in the management of HIV-1 infection, Drugs 69 (8) (2009) 1107–1140. [33] G. Davies, Changing the salt, changing the drug, Pharm. J. 266 (7138) (2001) 322–323. [34] United States Pharmacopeia: Monograph for Docusate Sodium, . [35] G.E. Woody, A.T. McLellan, C.P. O’Brien, F.S. Tennant, J. Mintz, Lack of toxicity of high dose propoxyphene napsylate when used for maintenance treatment of

M. Morgen et al. / European Journal of Pharmaceutics and Biopharmaceutics 117 (2017) 212–223

[36]

[37]

[38]

[39]

[40]

[41]

[42]

addiction, Clin. Toxicol. 16 (4) (1980) 473–478, http://dx.doi.org/10.3109/ 15563658008989976. K. Mohsin, M. Long, C. Pouton, Design of lipid-based formulations for oral administration of poorly water-soluble drugs: precipitation of drug after dispersion of formulations in aqueous solution, J. Pharm. Sci. 98 (10) (2009) 3582–3595, http://dx.doi.org/10.1002/jps.21659. C. Porter, C. Pouton, J. Cuine, W. Charman, Enhancing intestinal drug solubilisation using lipid-based delivery systems, Adv. Drug Del. Rev. 60 (2008) 673–692, http://dx.doi.org/10.1016/j.addr.2007.10.014. S.B. Murdande, M.J. Pikal, R.M. Shanker, R.H. Bogner, Solubility advantage of amorphous pharmaceuticals: 1. A thermodynamic analysis, J. Pharm. Sci. 99 (3) (2010) 1254–1265, http://dx.doi.org/10.1002/jps.21903. S.B. Murdande, M.J. Pikal, R.M. Shanker, R.H. Bogner, Solubility advantage of amorphous pharmaceuticals: II. Application of quantitative thermodynamic relationships for prediction of solubility enhancement in structurally diverse insoluble pharmaceuticals, Pharm. Res. 27 (2010) 2704–2714, http://dx.doi. org/10.1007/s11095-010-0269-5. L. Almeida e Sousa, S.M. Reutzel-Edens, G.A. Stephenson, L.S. Taylor, Assessment of the amorphous ‘‘solubility” of a group of diverse drugs using new experimental and theoretical approaches, Mol. Pharm. 12 (2) (2015) 484– 495, http://dx.doi.org/10.1021/mp500571m. G.A. Ileybare, L.S. Taylor, Liquid–liquid phase separation in highly supersaturated aqueous solutions of poorly water-soluble drugs: implications for solubility enhancing formulations, Cryst. Growth Des. 13 (2013) 1497–1509, http://dx.doi.org/10.1021/cg301679h. Bristol-Myers Squibb Company, Clinical Pharmacology and Biopharmaceutics Review, Application No. 21-567, 2002. Available from and .

223

[43] P.L. Anderson, C.L. Aquilante, E.M. Gardner, J. Predhomme, P. McDaneld, L.R. Bushman, J.H. Zheng, M. Ray, S. MaWhinney, Atazanavir pharmacokinetics in genetically determined CYP3A5 expressors versus non-expressors, J. Antimicrob. Chemother. 64 (2009) 1071–1079, http://dx.doi.org/10.3109/ 15563658008989976. [44] M.F. Wempe, P.L. Anderson, Atazanavir metabolism according to CYP3A5 status: an in vitro-in vivo assessment, Drug Metab. Dispos. 39 (2011) 522–527, http://dx.doi.org/10.1124/dmd.110.036178. [45] O. Kis, J.A. Zastre, M.T. Hoque, S.L. Walmsley, R. Bendayan, Role of drug efflux and uptake transporters in atazanavir intestinal permeability and drug-drug interactions, Pharm. Res. 30 (4) (2013) 1050–1064, http://dx.doi.org/10.1007/ s11095-012-0942-y. [46] K.A. Hasselbalch, Calculation of blood pH based on the free and bound carbonic acid, and oxygen binding of blood as function of pH, Biochem. Z. 78 (1916) 112–144. [47] SBDS Database, . [48] R. Meergans, R. Stefan, Dry processing of atazanavir, U.S. Patent Publication No. US20130224294A1, 2013, . [49] M. Berlin, A. Ruff, F. Kesisoglou, W. Xu, M.H. Wang, J.B. Dressman, Advances and challenges in PBPK modeling – analysis of factors contributing to the oral absorption of atazanavir, a poorly soluble weak base, Eur. J. Pharm. Biopharm. 93 (2015) 267–280, http://dx.doi.org/10.1016/j.ejpb.2015.03.031. [50] M.T. Huisman, J.W. Smit, K.M. Crommentuyn, N. Zelcer, H.R. Wiltshire, J.H. Beijnen, A.H. Schinkel, Multidrug resistance protein 2 (MRP2) transports HIV protease inhibitors, and transport can be enhanced by other drugs, AIDS 16 (17) (2002) 2295–2301. [51] Bristol-Myers Squibb, Atazanavir Briefing Document No. 232262, May 2003, .