1. Introduction
Crystallization is a critical process in the manufacturing of small-molecule pharmaceuticals. Drug molecules are regularly arranged in crystal lattices through weak but numerous intermolecular interactions such as van der Waals forces, hydrogen bonding, and electrostatic interactions [
1]. Crystallization is commonly used in separation and purification processes. It involves the solidification of single molecules from solutions containing other compounds and depends on differences in their solubilities [
2,
3]. In manufacturing, scale-up is a vital factor for crystallization in which nucleation and crystal growth should be controlled [
2]. The stability of the crystalline form of a drug molecule is also important. The quality control of drug products is a stringent requirement for the global manufacturing of pharmaceuticals; therefore, long-term stability and the detection and control of impurities and decomposition products are regulated [
4,
5]. Amorphous forms have favorable solubility and dissolution rates owing to their higher free energies than crystalline forms, but chemical and physical instabilities are often an issue for quality control [
6,
7]. The stability of a drug molecule can be enhanced by crystallization from the amorphous form, owing to the relatively low-energy state of the crystal lattice [
1].
Polymorphs in the crystal form of pharmaceutical drugs are formed during the process of drug development. Most small-molecule drugs have at least two crystal forms (polymorphs) that arise from the formation of different crystal lattices through changes in the arrangement of drug molecules [
8,
9,
10]. It is well known that the polymorphs of a drug have different physicochemical properties. These properties include solubility, the ability to form tablets, flowability, chemical and physical stability due to density, the state of intermolecular interactions, and the proportion of polar functional groups that cover the crystal surface [
8,
9,
10,
11]. Ritonavir is a useful illustration of the importance of pharmaceutical polymorphs. It was developed as an HIV protease inhibitor. However, it was withdrawn from the market owing to the generation of an unexpected polymorph with reduced solubility, which is critical for bioavailability and pharmacological activity [
12]. This incident has significantly affected pharmaceutical research and the drug industry; consequently, there has been considerable focus on ritonavir polymorphs [
13,
14,
15]. A new polymorph of ritonavir, Form III, was recently discovered [
16,
17]. Indomethacin, an ancient drug first marketed in 1965, is commonly used in pharmaceutical research. New polymorphs and their physicochemical properties have been reported in recent decades [
18,
19,
20]. These cases imply that the comprehensive discovery and control of drug polymorphs prior to their release into the market is vitally important.
Multicomponent crystals comprising drug molecules and other molecules are of interest for crystal engineering. Salts are representative multicomponent crystals. The functional groups of the drug molecules ionize and form ionic interactions with organic or inorganic coformers. Approximately 40% of the first-in-class drugs approved by the United States Food and Drug Administration are salts, and numerous counterions are used as coformers [
21]. The cocrystals have also emerged as salts. Non-ionic interactions play a role in the formation of crystal lattices between drug molecules and coformers. This enables the design of multicomponent crystals of drug substances without ionized functional groups [
22]. Both salts and cocrystals can significantly affect physicochemical properties such as solubility and stability [
21,
22,
23]. A multicomponent crystal is considered solvated if the coformer is liquid under ambient conditions. Organic solvents often form solvates with the drug molecules. However, this form should be treated carefully because residual solvents are strictly regulated for safety [
24,
25]. Water is a common and safe pharmaceutical coformer. Drug molecules form hydrates with water, and the number of such hydrates has significantly increased in recent decades [
24]. Drug molecules can become hydrated or dehydrated depending on the ambient conditions [
26]. This is an important consideration in crystal engineering.
Recently, the market for anti-obesity treatments has grown drastically due to the release of glucagon-like peptide 1 receptor (GLP-1R) agonists, which can help reduce body weight and showcases antidiabetic effects [
27]. With increasing investment in the anti-obesity market, monoacylglycerol acyltransferase 2 (MGAT2) is considered a pharmacological target for anti-obesity treatments [
28,
29]. It is highly expressed in the small intestine and catalyzes triacylglycerol synthesis during absorption. The inhibition of MGAT2 can suppress food intake in mice fed a high-fat diet through peripheral vagus nerve signaling and has potential as a novel anti-obesity strategy [
29]. In the present study, we aimed to identify and control the polymorphs of a drug originally designed for MGAT2 inhibition. First, we performed polymorphic screening and classified the crystal forms obtained. We determined the physicochemical properties and transformation mechanisms of the polymorphs.
2. Materials and Methods
2.1. Materials
S-309309 was originally designed as an MGAT2 inhibitor and synthesized by Shionogi & Co., Ltd. (Osaka, Japan) [
30,
31,
32]. Ethanol (EtOH), methanol (MeOH), isopropanol (IPA), and ethyl acetate (AcOEt) were purchased from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan). Purified water was obtained from Otsuka Pharmaceutical Co., Ltd. (Tokyo, Japan). The molecular weights and pKas were determined from the PhysChem Profiler Module using ACD/Percept ver. 14.3.0 (Advanced Chemistry Development, Toronto, ON, Canada).
2.2. Polymorphic Screening
2.2.1. Solvent Evaporation
Approximately 10 mg of S-309309 was placed in an S-1 4 mL glass vial (Nichiden Rika Glass Co., Ltd., Hyogo, Japan), and each solvent (MeOH, EtOH, IPA, AcOEt, and water) was added. Solubilization was confirmed by observation. S-309309 was dissolved in 1 mL of AcOEt at ambient temperature (approximately 20–23 °C) and 1 mL of MeOH at 60 °C. EtOH, IPA, and water were unable to solubilize S-309309. The resulting solutions and suspensions were subjected to solvent evaporation at 30 °C and 1750 rpm under reduced pressure using a Genevac HT-8 Series II evaporation system (Genevac Ltd., Ipswich, UK), and the residual powders were obtained.
2.2.2. Slurry Conversion
Two hundred micrograms of solvent and 10 mg of S-309309 were added to an S-1 4 mL glass vial (Nichiden Rika Glass Co., Ltd., Hyogo, Japan) equipped with a magnetic stirrer. After the cap was tightened, slurry conversion was performed at 300 rpm at ambient temperature using a CPS-300 cool stirrer (Scinics Corporation, Tokyo, Japan). After 8 days of slurry conversion, each suspension was treated by suction filtration using OmniporeÒ and a 0.45 mm polytetrafluoroethylene membrane (Merck KGaA, Darmstadt, Germany). The filtrate on the membrane was obtained.
2.3. X-ray Powder Diffraction
2.3.1. X-ray Powder Diffraction
The crystal forms of the samples were determined by X-ray powder diffraction (XRD) using a D8 Discover (Bruker AXS K.K., Kanagawa, Japan) or SmartLab (Rigaku Corporation, Tokyo, Japan) system. The Cu Kα radiation point source was operated at 40 kV and 40/200 mA. Other procedures were performed as in previous reports [
33,
34]. The data were analyzed using GADDS for XP/2000 ver. 4.1.27 (Bruker AXS K.K., Kanagawa, Japan) or Smart Lab studio II X64 version 4.2.111.0 software (Rigaku Corporation, Tokyo, Japan).
2.3.2. XRD under Controlled Temperature and Humidity
Changes in the crystal forms of the samples under temperature and humidity control were recorded using a Rigaku RINT 2100 Ultima instrument combined with XRD-DSC Thermo Plus II and HIM-1 systems (Rigaku Corporation, Tokyo, Japan). Approximately 2–3 mg of each sample was placed on an XRD/DSC aluminum pan (7 × 7 × 0.3 mm), and its XRD pattern was obtained while heating at a rate of 5 °C/min from room temperature (approximately 21–25 °C) to 80 °C. The effect of humidity was studied on the crystal form analyzed at 25 °C with relative humidity (RH) from 25% to 60% with 5% intervals. The voltage and current were set at 40 kV and 40 mA, respectively. The X-ray data were collected in the range of 5–35° (2-theta) at a scan rate of 60°/min with a scan step of 0.02°. The resulting data were analyzed using XRD-DSC ver. 2. 04 (Rigaku Corporation).
2.4. Thermal Analyses
2.4.1. Differential Scanning Calorimetry
Differential scanning calorimetry (DSC) was performed to determine the heat flow profile of S-309309 using a Discovery Q-1000 system (TA Instruments Japan, Tokyo, Japan) with nitrogen as the purge gas, supplied at a rate of 50 mL/min and calibrated using an indium standard. Each sample (1–3 mg) was weighed in an aluminum pan, which was then sealed. The thermal behavior of each sample was investigated over the temperature range of 0–250 °C at a rate of 10 °C/min. Dehydration behavior and melting point (Tm) were determined using Universal Analysis 2000 ver. 4.7A (TA Instruments, Tokyo, Japan).
2.4.2. Thermogravimetric–Differential Thermal Analysis
Thermogravimetric–differential thermal analysis (TG-DTA) was performed using a STA7200RV instrument (Hitachi High-Tech Science Corporation, Tokyo, Japan). Each sample (1–5 mg) was placed in an aluminum pan, and its change in weight and thermal profile were determined at 10 °C/min from room temperature (21–25 °C) to 300 °C. Data were analyzed using TA7000 standard analysis version 11.2 (Hitachi High-Tech Science Corporation, Tokyo, Japan).
2.5. Raman Spectroscopy
The Raman spectra of S-309309 were obtained using a Raman touch laser microscope (Nanophoton Corporation, Osaka, Japan). Each sample was placed on an aluminum plate, and its Raman spectrum was obtained using the following parameters: excitation wavelength, 671 nm; excitation power, 130.56 mW; ND filter, 56.31%; spectrograph center wavelength, 1400 cm−1; grating, 600 gr/mm; slit width, 50 μm; exposure time, 0.5 s; averaging, 10; gain, high; readout port, low noise; readout speed, 2 MHz; CCD temperature, −70 °C; objective lens, 5×/NA 0.15. The wavenumber was calibrated using the Si spectrum provided by the equipment. The peak positions in the Raman spectra were assigned using a smoothing process based on the fast Fourier transform method. The spectra were analyzed using a Raman Viewer (Nanophoton Corporation, Osaka, Japan).
2.6. Dynamic Vapor Sorption
The dynamic vapor sorption (DVS) profile of the water sorption and desorption of S-309309 was obtained using a DVS Advantage system (East Core Ltd., Tokyo, Japan). Each sample (8–9 mg) was weighed and placed in a pan. Water sorption was gravimetrically measured at 25 °C under various RH conditions. The RH was increased from 0% to 90% at a rate of 0.02%/min in 10% steps and was maintained until the weight change reached a plateau at each RH. The results were analyzed using the DVS Advantage control software ver. 2.1. 0.9 (East Core, Ltd., Tokyo, Japan).
2.7. Single-Crystal XRD and Structural Analysis
Single-crystal XRD data were collected using a Rigaku XtaLAB P200 system with CrysAlisPro 1.171.39.46e software (Rigaku Oxford Diffraction) using thin-layer mirror monochromated Cu-Kα radiation (λ = 1.54184 Å). The hydrate was mounted and examined at 213 K under a dry nitrogen purge and re-examined following transformation to an anhydrate at 298 K. The direct SHELXT method was used for the structural solution of the crystals [
35]. All calculations were performed with the observed reflections [I > 2σ(I)] with CrysAlisPro 1.171.39.46e (Rigaku Oxford Diffraction), except for refinement, which was performed using the SHELXL program [
36]. All nonhydrogen atoms were refined with anisotropic displacement parameters, and hydrogen atoms were placed in idealized positions and refined as rigid atoms with relative isotropic displacement parameters without hydrated water. Packing images and void spaces were prepared using Mercury.
2.8. Stability Test
The stability test was performed under accelerated conditions. The sample was stored in each desiccator with silica gel or saturated solution of sodium chloride (75% RH). The desiccators were stored at 40 °C for 2 months. The purity of the samples before and after storage were measured by using a high-performance liquid chromatography system Prominence (Shimadzu Corporation, Kyoto, Japan). The measurement conditions were as follows: mobile phase, 0.1% formic acid aqueous solution/acetonitrile with 0.6 mL/min; column, YMC-Triart C18 ExRS at 40 °C; injection volume, 3 μL; UV at 266 nm wavelength.
2.9. Dissolution Test
A dissolution test was performed using a μDiss with UV monitoring system (pION Inc., Billerica, MA, USA). Phosphate buffer (pH 6.8) was prepared as a simulated intestinal medium according to the Japanese Pharmacopoeia—18th edition, and 20 mL of the medium was placed in a dissolution vessel. Approximately 5 mg of the sample was weighed into a dissolution vessel, and the dissolution test started with stirring at 300 rpm at 37 °C. The drug concentration was monitored using a UV probe (5 mm) inserted into the donor at 240 nm every 1 min for 120 min.
4. Conclusions
In this study, we investigated the polymorphs of S-309309 and the mechanisms by which the anhydrates and pseudo-polymorphs of the hydrates transformed into each other. Polymorphic screening identified two crystal forms, Form I and Form II. Form I was obtained using various solvents, whereas Form II could only be obtained using AcOEt. Form II was determined to be an AcOEt solvate and demonstrated thermal behavior similar to that of Form I after desolvation. The XRD profile of Form I changed reversibly with increasing temperature, and three patterns were observed. A study in which the RH was varied revealed that Form I formed anhydrate, intermediate hydrate, or hydrate according to the RH within the range of 25–60%, i.e., corresponding to ambient conditions. Single crystals of the hydrate and anhydrate were successfully prepared, and their crystal structures were investigated. The hydrate comprised a 0.25 molar ratio of water and had a crystal structure similar to that of anhydrate. The spaces in the crystal lattice enabled water sorption/desorption via an intermediate state without changing the crystal lattice. The stability and dissolution profiles of the anhydrate and the hydrate did not change significantly. These findings led to the conclusion that the pseudo-polymorphic transformation of S-309309 occurred depending on environmental conditions; however, this is not a critical issue for pharmaceutical use.