Article Cite This: Mol. Pharmaceutics 2018, 15, 1052−1061
Drug-Excipient Interactions: Effect on Molecular Mobility and Physical Stability of Ketoconazole−Organic Acid Coamorphous Systems Michelle H. Fung,† Marla DeVault,‡ Keith T. Kuwata,‡ and Raj Suryanarayanan*,† †
Department of Pharmaceutics, University of Minnesota, Minneapolis, Minnesota 55455, United States Department of Chemistry, Macalester College, Saint Paul, Minnesota 55105, United States
‡
Downloaded via KAOHSIUNG MEDICAL UNIV on August 6, 2018 at 00:11:28 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: The use of excipients other than polymers for enhancing the physical stability of amorphous active pharmaceutical ingredients (APIs) has largely been unexplored. We investigated several organic acids (oxalic, tartaric, citric, and succinic acid) for the purpose of stabilizing a weakly basic API, ketoconazole (KTZ), in the amorphous state. Coamorphous systems with each acid, in 1:1 KTZ−acid molar ratio, were prepared by spray drying. The interaction of KTZ with each acid was investigated by FT-IR, solid-state NMR, and quantum chemical calculations. Each acid exhibited ionic and/or hydrogen-bonding interactions with KTZ, and quantum chemical calculations provided a measure of the strength of this interaction. The αrelaxation times, a measure of molecular mobility, were determined by dielectric spectroscopy, and their crystallization propensity by variable temperature X-ray powder diffractometry. Crystallization was observed only in two systems, KTZ−oxalic salt and KTZ−succinic as a cocrystal. An increase in the strength of KTZ−acid interaction translated to a decrease in molecular mobility. When the two systems prepared with structurally similar dicarboxylic acids (succinic and oxalic acid) were compared, the physical stability enhancement of KTZ−oxalic coamorphous system could be attributed to its lower mobility. However, the exceptional stability of KTZ−tartaric and KTZ−citric could not be explained by mobility alone, indicating that structural factors may also contribute to stabilization. The interaction between KTZ and acid may alter the system sufficiently so that the crystallization propensity of the KTZ−acid complex (salt or cocrystal) becomes relevant. We conclude that small molecule excipients have the potential to improve the physical stability of amorphous APIs. KEYWORDS: ketoconazole, oxalic acid, citric acid, tartaric acid, succinic acid, coamorphous, molecular mobility, crystallization, dielectric spectroscopy, amorphous
■
INTRODUCTION Amorphous active pharmaceutical ingredients (APIs) are characterized by poor physical stability leading to crystallization. A popular stabilization technique is to formulate amorphous solid dispersions (ASDs) by incorporating the API in a polymeric matrix. Polymers with high glass transition temperatures (Tg) can stabilize an amorphous API with a low Tg (for example, < RT) by antiplasticization of the system. As a result, the Tg of the ASD will be higher than that of the drug. A high polymer content can amplify this effect, yielding ASDs which can be in a deeply glassy state at room temperature. The Tg value, however, is not a reliable predictor of physical stability of amorphous pharmaceuticals. For example, though the Tg values of the acetaminophen ASDs prepared with polyvinylprrolidone (PVP) and poly(acrylic acid) (PAA) were virtually identical, the latter was significantly more stable.1 The interactions between drug and polymer provided an avenue for amorphous phase stabilization. For example, ASDs of ketoconazole (KTZ), a weakly basic drug, were formulated with each poly(acrylic acid) (PAA), poly(2-hydroxyethyl © 2018 American Chemical Society
methacrylate) (PHEMA), and polyvinylpyrrolidone (PVP). The strong ionic interaction between KTZ and PAA translated to a pronounced reduction in molecular mobility as well as inhibition in drug crystallization.2,3 Thus, PAA, even at a low concentration of 4% w/w, prevented drug crystallization for time scales of pharmaceutical interest. An ASD is thermodynamically stable if the API content is below the equilibrium solubility of crystalline drug in the polymer.4 Unfortunately, many crystalline APIs are characterized by low solubility in polymers (at the temperature of interest). On the other hand, the ASDs can be kinetically stabilized when the drug concentration is below the drug− polymer miscibility limit. However, under such circumstances, the risk of drug crystallization is not eliminated. As a result, while polymers have been extensively used to fabricate ASDs, Received: Revised: Accepted: Published: 1052
October 25, 2017 December 11, 2017 January 8, 2018 January 8, 2018 DOI: 10.1021/acs.molpharmaceut.7b00932 Mol. Pharmaceutics 2018, 15, 1052−1061
Article
Molecular Pharmaceutics
each of the following acids: oxalic, tartaric, citric, and succinic acid. A solution composed of KTZ and each acid in methanol (total solid content of 5% w/v) was introduced at a feed rate of 3.5 mL/min. The inlet temperature was maintained at 55 °C, which resulted in an outlet temperature of ∼35 °C. The spray dried samples were placed under vacuum for 20 h (RT) to remove the residual solvent. Variable-Temperature and Isothermal Powder X-ray Diffractometry (PXRD). Amorphous KTZ and coamorphous systems were evaluated using an X-ray diffactometer (D8 ADVANCE, Bruker AXS, WI, USA) equipped with a variable temperature stage (TTK 450; Anton Paar, Graz-Straβgang, Austria) and a Si strip one-dimensional detector (LynxEye). The measurements were performed in 10 °C increments from 45 to 205 °C. The heating rate was 10 °C/min, and the sample was maintained under isothermal conditions during each XRD scan. Data were collected, using Cu Kα radiation (1.54 Å, 40 kV × 40 mA), in the range of 5−35° 2θ with a step size of 0.02° and a dwell time of 0.5 s. Isothermal crystallization studies were conducted by subjecting the coamorphous systems to XRD after storing the amorphous samples in desiccators containing anhydrous calcium sulfate (∼0% RH) at 75 °C for 1 week. Differential Scanning Calorimetry (DSC). A differential scanning calorimeter (TA Instruments Q2000, DE, USA) equipped with a refrigerated cooling unit was used. The instrument was calibrated with indium. The samples, in hermetically crimped (Tzero) pans, were heated at 10 °C/min under nitrogen purge (50 mL/min). While the glass transition (Tg) and melting (Tm) temperatures reported are the midpoints of the transition, the crystallization temperature (Tc) was the onset of the transition. Thermogravimetric Analysis (TGA). The samples were heated at 10 °C/min under nitrogen purge (50 mL/min) from RT to 200 °C (TGA Q50, TA Instruments, DE, USA). The residual solvent content was determined based on the % weight loss at 100 °C during heating. Fourier Transform Infrared Spectroscopy (FT-IR). FTIR spectra were obtained at room temperature using Bruker Vertex 80 spectrometer (MA, USA). At a resolution of 4 cm−1, 64 scans were obtained across the spectral range of 3500800 cm−1. Amorphous citric, oxalic, tartaric, and succinic acids were prepared by melt-quenching. Each crystalline acid was heated to ∼10 degrees above its melting point and quench cooled in liquid nitrogen. The melt-quenched samples were immersed in liquid nitrogen until IR measurements. Dielectric Spectroscopy (DES). Molecular mobility was measured using a dielectric spectrometer (Novocontrol AlphaAK high performance frequency analyzer, Novocontrol Technologies, Germany) equipped with a temperature controller (Novocool Cryosystem) at temperatures ranging from 25 to 120 °C in steps of 5 °C in the frequency range of 10−2 to 106 Hz. The experiments were carried out as a function of frequency at a fixed temperature. Solid-State NMR (ssNMR). 13C spectra were acquired under magic angle spinning using cross-polarization (CP/ MAS) using 4 mm double resonance probe (Bruker ADVANCE III-500 spectrometer). Approximately 300 mg of samples were used for each run. Computational Studies. The stabilization energy of each KTZ−acid complex was determined by ONIOM (our own nlayered integrated molecular orbital and molecular mechanics) calculations using Gaussian 09.13 ONIOM is a computational approach that utilizes several levels of model chemistry.14 A
the polymer concentration tends to be high, limiting its use for low dose APIs. Processing can be a challenge at high polymer concentrations. In the hot melt extrusion process, the desirable flow properties are often realized at temperatures substantially higher than the glass transition temperature of the polymer.5 High processing temperature may lead to chemical degradations of the API and polymer. Some of the disadvantages with polymers can be overcome through the use of low molecular weight additives.6,7 Using this approach, stabilization may be accomplished at a relatively low additive concentration. Coamorphous drug-amino acid mixtures are emerging as a promising stabilization strategy.8 As with drug and polymer, the additive−API interactions provide an avenue to prepare a stable amorphous product. Spray drying of indomethacin−arginine mixture resulted in amorphous salt formation. 9 When formulated into tablets, the salt was physically stable and yielded a substantial “solubility” enhancement with respect to crystalline indomethacin. While API−excipient interactions play an important role in the stabilization of coamorphous mixtures, there are several examples where a reduction in molecular mobility translated to physical stability enhancement. For example, in binary mixtures of indapamid and ezetimibe, increasing concentrations of the former led to a progressive reduction in molecular mobility of the system, resulting in improved physical stability.10 This suggested that similar to ASDs, molecular mobility of coamorphous systems may be a predictor of physical stability. Interaction between weakly basic APIs and acidic excipients (including organic acids) provides an avenue to prepare stable amorphous phases. The strength of drug−excipient interaction is expected to influence the ASD stability. The current study aims to investigate the impact of several organic acids on the molecular mobility and physical stability of coamorphous systems prepared with a weakly basic API, ketoconazole (KTZ). Four organic acids (oxalic, tartaric, citric, and succinic acids) with different strengths (pKa values) were chosen. Based on the pKa difference between KTZ and each of the acids, the strongest interaction would either be ionic or H-bonding.11 Using FT-IR spectroscopy, specific interactions in coamorphous systems will be identified. The strength of interactions will be assessed using solid-state NMR spectroscopy. Quantum chemical calculations will provide a theoretical estimate of drug−acid interactions. We hypothesize that the extent of reduction in molecular mobility brought about by the addition of the acids will be influenced by the strength of interaction between the drug and each acid. Finally, the crystallization propensity of these systems will be evaluated by powder X-ray diffractometry and differential scanning calorimetry. This comprehensive approach will enable a mechanistic understanding of the stabilizing effect of excipients.
■
EXPERIMENTAL SECTION Materials. Ketoconazole (KTZ), the model drug, was a gift from Laborate Pharmaceuticals (Haryana, India). It is a weakly basic compound that is practically insoluble in water.12 There are no reported polymorphic forms of KTZ. The organic acids used, oxalic acid, anhydrous citric acid, D, L-tartaric acid, and succinic acid, were purchased from Sigma-Aldrich (Missouri, USA). Sample Preparation. Ketoconazole (KTZ) coamorphous systems with 1:1 drug-acid molar ratio were prepared by spray drying (BUCHI mini spray dryer B-290, Delaware, USA) with 1053
DOI: 10.1021/acs.molpharmaceut.7b00932 Mol. Pharmaceutics 2018, 15, 1052−1061
Article
Molecular Pharmaceutics prototypical example of ONIOM is geometric optimization of enzymes, where the active site of the enzyme is computed using the highest level model chemistry while the area around the active site is calculated using lower level model chemistries.15 In this study, three levels of model chemistry were used. The atoms surrounding the sites of specific interactions (acid−base or hydrogen-bonding interaction) were optimized using DFT with model chemistry B3LYP/6-31+G(d,p) (Supporting Information; Figure S2). The remaining atoms in each acid molecule utilized a lower level model chemistry (B3LYP/631G). The atoms in KTZ that were not involved in specific interactions were optimized with a semiempirical method (PM6). For each calculation, only the electronic energy of the system was taken into account. The input geometry of ketoconazole was based on the crystal structure obtained from the CCDC database.16 A self-consistent solvent reaction field was applied to each calculation using water as a solvent. In order to test the validity of this approach, optimizations were done for complexes of either oxalic or tartaric acid with ammonia using a single model chemistry, B3LYP/6-31+G(d,p), and with ONIOM layers mimicking those used for the optimization of the ketoconazole complexes with the same acids. The results from these calculations are summarized in the Supporting Information (Table S3). When compared with single model chemistry, the ONIOM method revealed a higher stabilization energy for both acids. In addition, when ONIOM was applied, there was a dampening of the acidity of oxalic acid that was disproportionately lower than that of tartaric acid.
Table 1. Acids Used for Preparing KTZ Coamorphous Systemsa pKa1, acid
ΔpKa = pKa1, KTZ − pKa1,acid
1.3 2.9 3.1 4.2
5.3 3.5 3.4 2.3
oxalic acid tartaric acid citric acid succinic acid a
The pKa of the most acidic site in each acid (pKa1) and its difference from the pKa of the most basic site in KTZ are provided.
Since the pKa values are reflections of the equilibrium behavior in aqueous solution,17 it is difficult to predict the extent of proton transfer in coamorphous solid dispersion based on pKa alone. Baseline Characterization. Spray dried KTZ was X-ray amorphous. KTZ spray dried with each oxalic, tartaric, citric, and succinic acid also exhibited a broad halo over the angular range of 5 to 35° 2θ (Supporting Information; Figure S3). The residual solvent content, determined by TGA, was 3, it is generally expected that there will be proton transfer from the acid to the base, resulting in salt formation.17 When the ΔpKa is
