International Journal of Pharmaceutics 460 (2014) 83–91
Contents lists available at ScienceDirect
International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm
Four new polymorphic forms of suplatast tosilate Keiko Nagai a,e,∗ , Takanori Ushio b , Hidenori Miura c , Takashi Nakamura d , Kunikazu Moribe e , Keiji Yamamoto e a
Taiho Pharmaceutical Co. Ltd., Ebisuno, Hiraishi, Kawauchi-cho, Tokushima 771-0194, Japan Ace Japan Co. Ltd., Higashine-koh, Higashine, Yamagata 999-3701, Japan Taiho Pharmaceutical Co. Ltd., Kamikawa-machi, Kodama-gun, Saitama 367-0241, Japan d RIKEN, Wako, Saitama 351-0198, Japan e Graduate School of Pharmaceutical Sciences, Chiba University, Inohana, Chuo-ku, Chiba 260-8675, Japan b
c
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 20 October 2013 Accepted 29 October 2013 Available online 6 November 2013 Keywords: Suplatast tosilate Polymorphs Seeding analogous compounds Physicochemical properties Solubility Rietveld refinement
a b s t r a c t We found four new polymorphic forms (␥-, -, -, and -forms) of suplatast tosilate (ST) by recrystallization and seeding with ST-analogous compounds; three polymorphic forms (␣-, -, and ␦-forms) of ST have been previously reported. The physicochemical properties of these new forms were investigated using infrared (IR) spectroscopy, solid-state nuclear magnetic resonance (NMR) spectroscopy, differential scanning calorimetry, and powder X-ray diffractometry. The presence of hydrogen bonds in the new forms was assessed from the IR and solid-state NMR spectra. The crystal structures of the - and forms were determined from their powder X-ray diffraction data using the direct space approach and the Monte Carlo method, followed by Rietveld refinement. The structures determined for the - and -forms supported the presence of hydrogen bonds between the ST molecules, as the IR and solid-state NMR spectra indicated. The thermodynamic characteristics of the seven polymorphic forms were evaluated by determining the solubility of each form. The ␣-form was the most insoluble in 2-propanol at 35 ◦ C, and was thus concluded to be the most stable form. The -form was the most soluble, and a polymorphic transition from the - to the ␣-form was observed during solubility testing. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Polymorphism has long been a subject of great concern in the pharmaceutical industry because different solid pharmaceutical forms can impart different physical and chemical properties that can occasionally affect the bioavailability, dissolution rate, stability, and manufacturability of solid dosage forms (Brittain, 2012; Grunenberg et al., 1996; Jones, 1997; Kawakami, 2007; Thirunahari et al., 2010; Yamamoto et al., 2011). The characterization of solid pharmaceuticals is therefore very important to ensure that pharmaceuticals with standardized qualities are supplied to the consumer. There have been many studies in which solid pharmaceuticals have been physicochemically characterized, and techniques such as differential scanning calorimetry (DSC), infrared (IR) spectroscopy, microscopy, powder X-ray diffractometry (PXRD), Raman spectroscopy, and solid-state nuclear magnetic resonance (NMR) spectroscopy have been used (Ikeda et al., 2010;
∗ Corresponding author at: Taiho Pharmaceutical Co. Ltd., Ebisuno, Hiraishi, Kawauchi-cho, Tokushima 771-0194, Japan. Tel.: +81 88 665 3571; fax: +81 88 665 7225. E-mail address: [email protected] (K. Nagai). 0378-5173/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2013.10.049
Tozuka et al., 2002). Most polymorphs can be identified using these analytical techniques, but some polymorphs are still difficult to identify. Polymorphism is defined as the occurrence of different arrangements and/or conformations of molecules within the crystal lattice. Elucidating the arrangement and/or conformation of the molecules is often crucial to understanding the solid-state chemistry of active pharmaceutical ingredients. The crystal structure is one of the most crucial pieces of information relating to polymorphism; however, it is not always possible to determine the crystal structure of a metastable crystalline form because single crystals that are suitable for X-ray crystallographic analysis may be unavailable. Such difficulties have recently been overcome by advances in the procedures available for solving structures from powder X-ray diffraction data measured using laboratory X-ray sources (Engel et al., 1999; Harris et al., 1994, 2001; Neumann, 2003; Rietveld, 1969; Stephenson, 2000; Young and Ed, 1993), although there still are some limitations. Suplatast tosilate (ST) [(±)-[2-[4-(3-ethoxy-2-hydroxypropoxy) phenylcarbamoyl]ethyl]dimethylsulfonium p-toluenesulfonate] is an excellent IgE antibody production suppressor, and is a useful therapeutic agent for various allergic diseases. ST is a glycerol
84
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
derivative with one chiral carbon atom, which has been developed as a racemate. It has been reported that changing the crystallization conditions, such as the solvent(s) and temperature, pseudo-seeding the supersaturated solution with a crystal with the desired structure (following an epitaxy protocol), or adding impurities or additives to inhibit the nucleation and/or crystal growth of an undesired polymorph can be used to give a desired polymorph (Maruyama and Ooshima, 2000; Blagden et al., 1998). Using this technique, polymorphs of ST and ST analogous compounds were also prepared (Takahashi et al., 1998a, 1998b, 2002; Tamura et al., 1997, 1998, 2001a, 2001b, 2011). Three polymorphic forms (␣-, -, and ␦-forms) of ST were previously reported, and their crystal structures were determined (Miura et al., 2003; Takahashi et al., 2001; Ushio et al., 1996a, 1996b, 1996c, 2002; Ushio and Yamamoto, 1994). We have found four new polymorphic forms (␥-, -, -, and -forms) by recrystallizing and seeding with ST-analogous compounds. We investigated the physicochemical properties of the new polymorphic forms of ST by DSC, IR and solid-state NMR spectroscopies, and PXRD. We also determined the solubilities of the new polymorphic forms, and compared them with the solubilities of the three crystals that were previously characterized. We determined the crystal structures of the - and -forms from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement.
variable torsional angles are shown in Fig. 1. All other chemicals used were of special reagent grade. 2.2. Preparation of ˛-, ˇ-, and ı-form ST crystals The ␣-, -, and ␦-forms of ST were prepared following the previously reported procedures (Miura et al., 2003; Tamura et al., 1997; Ushio et al., 1996c). 2.3. Preparation of -form ST crystals Suplatast tosilate (5 g) was dissolved in chloroform (20 mL) and the solvent was evaporated under reduced pressure. Crystals prepared in this way were designated as the ␥-form. 2.4. Preparation of -form ST crystals Suplatast tosilate (1 g) was dissolved with heating in 2propanol (4 mL). After cooling the solution to room temperature, (±)-[2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]trimethylammonium p-toluenesulfonate (2) (approximately 5 mg) was added to act as pseudo-seed crystals. The wet crystals obtained from the solution were added to another solution of ST (5 g) in 2-propanol (20 mL) to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form.
2. Materials and methods 2.5. Preparation of -form ST crystals 2.1. Materials Suplatast tosilate [(±)-[2-[4-(3-ethoxy-2-hydroxypropoxy) phenylcarbamoyl]ethyl] dimethylsulfonium p-toluenesulfonate] (1a), which has been developed as a racemate, was provided by Taiho Pharmaceutical Co. Ltd. (Tokyo, Japan). Analogous compounds of ST, (±)-[2-[4-(3-propoxy-2-hydroxypropoxy) p-toluenesulfonate phenylcarbamoyl]ethyl]dimethylsulfonium (1b) and (±)-[2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]trimethylammonium p-toluenesulfonate (2), were provided by the Graduate School of Human and Environmental Studies, Kyoto University (Japan), and Taiho Pharmaceutical Co. Ltd., respectively. The chemical structures of ST, the ST-analogous compounds and the structure formula of (±)-1a with the six
⋅
Suplatast tosilate (1 g) was dissolved with heating in 2-propanol (4 mL). After cooling the solution to room temperature, (±)-[2-[4(3-propoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]dimethylsulfonium p-toluenesulfonate (1b) (approximately 2 mg) was added to act as pseudo-seed crystals. The wet crystals obtained from the solution were added to another solution of ST (5 g) in 2-propanol (20 mL) to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form. 2.6. Preparation of -form ST crystals Suplatast tosilate (5 g) was dissolved with heating in a mixture of acetone (30 mL) and water (0.75 mL). After cooling the solution to 5 ◦ C, the ␣-form of ST crystals (1a) (approximately 5 mg) was added to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form. 2.7. Powder X-ray diffraction measurements
⋅
Powder X-ray diffraction patterns of each polymorphic form of ST were determined on a Spectris X’Pert instrument (Spectris Co., Ltd., Tokyo, Japan) with Cu K␣ radiation at 40 kV and 30 mA, passed through a nickel filter. The analysis was performed at a continuous scanning rate of 2◦ /min over a 2 angular range of 5–35◦ . The diffractograms obtained were analyzed using Spectris diffraction software. 2.8. Differential scanning calorimetry
Fig. 1. The chemical structures of suplatast tosilate (1a), the suplatast tosilate analogous compounds (1b and 2), and the structure formula of (±)-1a with the six variable torsional angles.
Differential scanning calorimetry analysis was performed using a Rigaku 8230 instrument (Rigaku, Tokyo, Japan). A sample (5 mg) was weighed in an open aluminum pan and subjected to a 30–130 ◦ C thermal scan at a heating rate of 4 ◦ C/min with a 100 mL/min dry nitrogen purge.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
2.9. Infrared spectroscopy An infrared spectrum of each polymorphic form was recorded using a JASCO FT/IR-420 instrument (JASCO International Co., Ltd., Tokyo, Japan) and the potassium bromide disk method. 2.10.
13 C
solid-state nuclear magnetic resonance spectroscopy
13 C
solid-state NMR spectra of the ␣-, -, and ␥-forms of ST were measured using a Chemagnetics CMX Infinity 400 instrument (Chemagnetics, Fort Collins, CO, USA) operated at 100.3 MHz with a cross-polarization/magic angle spinning (CP/MAS) probe. CP/MAS with two-pulse phase modulated (TPPM) decoupling was used, and the sample was placed in a 4 mm outer diameter sample tube that was rotated at 12 kHz. The contact time and the repetition time were 5.0 ms, and 5.0 s, respectively. The spectral width was 40 kHz, and 3000 scans were recorded. 13 C solid-state NMR spectra of the ␦-, -, -, and -forms of ST were measured using a Chemagnetics CMX Infinity 400 instrument (Chemagnetics, Fort Collins, CO, USA), operated at 100.4 MHz, using ramped CP/MAS with TPPM decoupling. The sample was placed in a 5 mm outer diameter sample tube that was rotated at 10 kHz. The contact time and the repetition time were 3.0 ms, and 5.0 s, respectively. The spectral width was 40 kHz, and 400 scans were recorded. 2.11. Solubilities An excess amount of a polymorph sample was added to 2propanol (2 mL). The sample was vigorously shaken for 30 s every 5 min for 30 min, and then allowed to stand for 4 h at 35.0 ◦ C. The saturated solution was filtered through a 0.45 m membrane filter, diluted with methanol, and analyzed by HPLC. HPLC analyses were carried out using an octadecylsilanized silica gel column (Shinwa chemical, ULTRON TDP, 4.6 mm × 150 mm), a mixture of 2.5 mmol/L sodium 1-hexanesulfonate solution, acetonitrile, and acetic acid (300:100:3) as the mobile phase at flow rate of 0.9 mL/min, and a UV–vis spectrophotometer (250 nm) as detector. Powder X-ray diffraction was used to determine whether any of the polymorphic forms had been transformed into another during the solubility measurement process. The solubilities were determined in duplicate. 2.12. Solving the crystal structures from the powder X-ray diffraction data The Reflex Plus module in the Materials Studio software (Accelrys, Inc, USA) was used for all calculations. The PXRD patterns of the - and -form crystals of ST were indexed with the X-cell program using 30 reflections (2 < 32◦ ). The cell and function profile parameters obtained were refined using the Pawley method. The space group was determined from within the space group candidates consistent with the systematic absences by a trial-and-error method. Once the initial model molecular conformations of the - and -form crystals of ST were assigned, the structures were solved using the Monte Carlo/parallel tempering method in the PowderSolve software. Rietveld refinement was then performed, which involved the following processes: (i) the pseudo Voight function was used to simulate the peak shape; (ii) the background was determined by linear interpolation using 20 terms; (iii) the Finger–Cox–Jephcoat method was used for asymmetric refinement; and (iv) the March–Dollase method was applied to correct preferred orientation effects. Global isotropic factors were used to refine the temperature factors. The molecular conformation in each of the motion groups was refined by varying the torsional angles
85
around all of the single bonds. The Monte Carlo calculations and subsequent Rietveld refinements were repeated.
3. Results and discussion 3.1. Polymorphic behavior We obtained four new forms of ST by recrystallizing ST and seeding with ST-analogous compounds, and investigated their physicochemical properties. Visual observations of all seven crystalline forms are shown in Fig. 2. The crystalline forms had different morphologies. The ␣-Form crystals were massive and solid. The -, ␦-, and -Form crystals were relatively large but fragile. The ␥- and -Form crystals were fine, and the -form crystals were needles. The PXRD patterns of the seven crystalline ST forms are shown in Fig. 3. We confirmed that the ␣-, -, and ␦-form PXRD patterns coincided with the patterns that were previously reported (Miura et al., 2003; Takahashi et al., 2001; Ushio et al., 1996c, 2002). However, the PXRD patterns of the four new crystals obtained in this study were completely different from those of the other three forms, and were also different from each other; each form showed a high degree of crystallinity and distinct diffraction angles and intensities. These results strongly suggested that the four new crystal forms could be polymorphic forms of each other. Adding a small amount of an ST analogous compound to act as seed crystals may have induced three of the new ST polymorphs (␦-, -, and -forms) to crystallize. It is noteworthy that additives, which made up less than 0.5% of the recrystallization solution in all of the cases, caused new polymorphic forms to crystallize. Kuroda et al. found a new polymorphic form of 6mercaptopurine, which was six to seven times more soluble than the other known crystalline forms, by adding an analogous compound as a seed (Kuroda et al., 1979). Kitamura and Ishizu investigated the growth rate of l-glutamic acid in the presence of lphenylalanine, confirming that the polymorphic (␣- and -) forms had different growth rates (Kitamura and Ishizu, 1998). Addadi et al. successfully resolved d- and l-asparagine conglomerations by adding small amounts of the pure enantiomers to saturated solutions of d- and l-asparagine mixtures (Addadi et al., 1962). It is thought that in each of these mechanisms, the analogous compound adheres to a crystal surface and inhibits crystal growth. In this study, it is considered that this mechanism also effects the formation of new polymorphic forms of ST. Evaporating the chloroform from a supersaturated solution of ST was an effective way of crystallizing the ␥-form, but the other polymorphic forms were prepared by recrystallization. The DSC curves of the seven polymorphic forms of ST are shown in Fig. 4. Each curve showed an endothermic peak due to fusion between 75 and 87 ◦ C. The highest melting points were 86.7 ◦ C for the ␣-form and 86.5 ◦ C for the ␥-form, showing that these polymorphic forms had spatially stable molecular arrangements. In contrast, low melting points were observed for the ␦-, -, and -forms, indicating that these forms were metastable. Furthermore, the polymorphic forms prepared by pseudo-seeding with ST-analogous compounds (␦-, -, and -forms) had broad endothermic peaks caused by meltingpoint depression. The highest enthalpy of fusion was 39.9 kJ/mol for the -form, and the lowest was 23.2 kJ/mol for the -form. The IR spectra are shown in Fig. 5. For the known crystal structures (␣-, -, and ␦-forms), the carbonyl stretching band was found at 1664 cm−1 for the ␣-form (in which the carbonyl group was hydrogen bonded to a OH group), whereas this band was shifted to 1678 and 1684 cm−1 for the - and ␦-forms (in which the carbonyl groups did not interact with a proton donor group).
86
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
Fig. 2. Microscopy photographs of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
It is well known that different hydrogen bonding modes may cause polymorphism in many compounds, such as barbiturate derivatives, sulfonamides, and oxalic acid (Mesley and Clements, 1967; Yang and Guillory, 1972; Pauling, 1960). The occurrence of hydrogen bonds in the four new polymorphic forms of ST was assessed by comparing their IR spectra. The carbonyl group stretching band of the -form was found at 1658 cm−1 , indicating the presence of hydrogen bonds with a proton donor group. However,
the carbonyl group stretching band was found at 1682 cm−1 for the -form, suggesting that there were no hydrogen bonds present. The ␥- and -forms showed carbonyl stretching bands at 1675 and 1672 cm−1 , respectively. It was not appropriate to assess hydrogen bond formation in the polymorphs from these data. In general, OH and NH stretching bands that are not hydrogen bonded are found at 3600–3700 and 3400–3500 cm−1 , respectively, as intense and sharp peaks. The polymorphic forms of ST other than
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
87
Fig. 4. Differential scanning calorimetry curves of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
Fig. 3. Powder X-ray diffraction patterns of suplatast tosilate polymorphs: (a) ␣form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
the -form showed broad bands between 3000 and 3500 cm−1 , revealing that several types of hydrogen bonds were present. The -form had a sharp peak at 3495 cm−1 , suggesting that most of the NH groups were not hydrogen bonded. Since the - and -forms of ST had broad bands around 3500 cm−1 , we accessed that sharp peaks overlapped in the broad bands around 3500 cm−1 , suggesting that some of the NH groups were not hydrogen bonded. The 13 C solid-state NMR results are shown in Fig. 6. The spectra of all seven polymorphic forms of ST were readily distinguishable by the 13 C solid-state NMR. The resonances of methyl carbons in the dimethylsulfonium group observed around 20 –30 ppm were different from each other, and the presence of hydrogen bonds between molecules in the new polymorphic forms of ST
was assessed by comparing the NMR spectra with the spectra of the polymorphic forms that were previously characterized. The chemical shift of the carbonyl carbon was observed at 170.0 ppm for the ␣-form of ST, in which the carbonyl group is hydrogen bonded to an OH group, and at 167.9 and 166.7 ppm for the and ␦-forms, respectively. In both of the latter forms, the carbonyl group did not interact with a proton donor group. The chemical shift for the -form carbonyl was 167.1 ppm, suggesting that the carbonyl group was not hydrogen bonded. However, the carbonyl chemical shift for the -form was observed at 170.2 ppm; therefore, the carbonyl group in this form was assumed to be hydrogen bonded. The carbonyl chemical shifts for the ␥- and -forms were 169.1 and 169.9 ppm, respectively, and these intermediate chemical shift positions could not be used to assess whether the carbonyl groups were hydrogen bonded or not in these crystals. The results of physicochemical properties of the ST polymorphs are summarized in Table 1.
Fig. 5. Infrared spectra of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
88
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
Table 1 Summary of physicochemical characteristic for suplatast tosilate polymorphs. States
Preparation method
DSC
IR
NMR
Melting point (◦ C)
Enthalpy of fusion (kJ/mol)
Stretching vibration of carbonyl band (cm−1 )
Chemical shift of carbonyl group (ppm)
Crystal structure
␣-form
Known
Recrystallization
86.7
35.2
1664
170.0
Already determined Ref: Takahashi et al. (2001)
-form
Known
Recrystallization
80.5
34.8
1678
167.9
Already determined Ref: Ushio et al. (1996c)
␥-form
New
Evaporation
86.5
35.0
1675
169.1
–
␦-form
Known
Seeding an analogous compound
74.8
32.5
1684
166.7
Already determined Ref: Miura et al. (2003)
-form
New
Seeding an analogous compound
75.2
29.2
1682
167.1
Rietveld refinement
-form
New
Seeding an analogous compound
78.1
23.2
1658
170.2
–
-form
New
Recrystallization
83.0
39.6
1672
169.9
Rietveld refinement
into the ␣-form (although most of the crystals remained in the original ␦-form). The -form was mostly transformed into the ␣-form during the solubility measurements, and the ␥-form was completely transformed into the ␣-form. Even though the -form was found to be extremely soluble in the initial dissolution stage, the solubility measurement for this form was not completed because it rapidly transformed into the ␣-form. The solubility of the crystal forms at 35 ◦ C followed the order ␣ <  < ␥ < < ␦ < , with the ␣-form being the most stable. The apparent solubility of the -form was significantly high nearly six times that of the ␣-form. Since the ␥-form completely transformed to the ␣-form, it was the second most unstable crystal after the -form. 3.3. Determination of the crystal structure from the powder X-ray diffraction data
Fig. 6. 13 C solid-state nuclear magnetic resonance spectra of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
3.2. Solubility studies
The crystal structures of the ␣-, -, and ␦-forms were previously reported (Miura et al., 2003; Ushio et al., 1996c). We determined the crystal structures of the - and -forms from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement (Table 3, and Figs. 7 and 8). After two cycles of Monte Carlo calculations (with 266 million trial structures per cycle) and the subsequent Rietveld refinement,
The thermodynamic characteristics of the seven polymorphic forms were evaluated by determining the solubility of each form in 2-propanol; the results are shown in Table 2. Once the solubility measurements had been completed, the remaining crystals were collected and subjected to PXRD measurements. The ␣-, -, and forms did not transform into other polymorphic forms during the solubility measurements, whereas the ␦-form partly transformed Table 2 Solubilities of suplatast tosilate polymorphs in 2-propanol (35 ◦ C).
␣-form -form ␥-form ␦-form -form -form -form a
Solubility (mg/mL)
Polymorphic form after solubility determination
20.95 33.63 35.87 42.22 134.28 73.62a 40.17
␣  ␣ ␣ ␣
-form was rapidly transformed to the ␣-form while measuring solubility.
Fig. 7. Molecular arrangement in -form of suplatast tosilate.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
89
Table 3 Crystal data for suplatast tosilate polymorphs. Cell parameters
␣-form
-form
␦-form
-form
-form
Chemical formula Formula weight Sample type Crystal system Space group a (Å) b (Å) c (Å) ˛ (deg) ˇ (deg) (deg) V (Å) Z Rpa , Rwpb Ref
C23 H33 NS2 O7 499.65 Single crystal Triclinic P-1 14.885 15.673 12.435 100.17 104.65 108.73 2550.6 4 0.063; 0.081 Takahashi et al. (2001)
C23 H33 NS2 O7 499.65 Single crystal Triclinic P1 10.748 15.616 8.234 100.50 108.47 85.51 1288.0 2 0.044; 0.047 Ushio et al. (1996c)
C23 H33 NS2 O7 499.65 Powder Triclinic P-1 9.895 15.030 8.909 98.24 90.28 108.15 1242.0 2 0.147; 0.190 Miura et al. (2003)
C23 H33 NS2 O7 499.65 Powder Monoclinic P21/c 21.380 12.044 10.756
C23 H33 NS2 O7 499.65 Powder Triclinic P-1 16.664 10.357 7.619 78.18 97.47 101.59 1255.4 2 0.167; 0.226 –
a
x Rp =
|cY sim (2i)−I exp (2i)−Y back (2i)|
b
y Rwp =
|I exp (2i)|
wi(I exp (2i))2
2503.5 4 0.146; 0.190 –
.
wi(cY sim (2i)−I exp (2i)−Y back (2i))
115.32
2
1/2 .
Fig. 9. Difference plot for the Rietveld simulated and experimental powder X-ray diffraction patterns for the -form of suplatast tosilate.
Fig. 8. Molecular arrangement in -form of suplatast tosilate.
the weighted- profile R value (Rwp ) converged at below 23%. The experimental and calculated diffraction patterns and the corresponding difference profiles are shown in Figs. 9 and 10. The -form crystal was found to be in a monoclinic system, and it was very different from the polymorphic forms that were previously reported. This crystal was composed of equal amounts of the two ST enantiomers (P21/c, Z = 4) and had a center of symmetry like the ␣- and ␦-forms. The R and S ST enantiomers in the -form were hydrogen bonded between the hydroxyl groups and ethoxy oxygen atoms (with an ˚ as in the ␦-form. However, unlike the O· · ·O distance of 2.873 A), other ST crystals, the -form was found to have a unique spatial
Fig. 10. Difference plot for the Rietveld simulated and experimental powder X-ray diffraction patterns for the -form of suplatast tosilate.
90
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
molecular arrangement, in which the amide NH hydrogen was not hydrogen bonded to the sulfonate ion oxygen, which was consistent with the IR result. The -form crystal structure was similar to that of the ␣-form, consisting of a racemic mixture of R and S enantiomers in the unit cell (Z = 2) of the centrosymmetric space group P-1. The R and S enantiomers were found to be hydrogen bonded between some of a hydroxyl group and carbonyl oxygen (with an O· · ·O distance of ˚ similarly to the ␣-form but unlike the - and ␦-forms. 3.036 A), Concurrently, some of the hydroxyl groups in the -form crystal structure were also found to be hydrogen bonded to sulfonate ˚ as in the group oxygen atoms (with an O· · ·O distance of 2.547 A), form. From these conflicting results, we estimated that some of the hydroxyl groups should therefore be disordered, and free carbonyl groups did seem to exist. The - and -form crystal structures supported the conclusions of intermolecular hydrogen bonding drawn from the IR and 13 C solid-state NMR spectra. The determination of crystal structures of the ␥- and forms were also examined, however, they transformed into the ␣-form during the measurements and they are still being studied. 4. Conclusions We identified four new polymorphic forms (␥-, -, -, and forms) of ST by recrystallization and seeding with ST-analogous compounds. The physicochemical properties of the new ST polymorphs were investigated using DSC, IR, and 13 C solid-state NMR spectroscopies, and PXRD The PXRD patterns for the new crystal forms were completely different from the patterns for the three previously characterized polymorphs, as well as being different from each other. The highest melting points observed from DSC measurements were 86.7 ◦ C for the ␣-form and 86.5 ◦ C for the ␥-form, showing that both these forms were relatively stable. Relatively low melting points were found for the ␦-, -, and -forms (which were prepared by pseudo-seeding with ST-analogous compounds); these forms had broad endothermic peaks associated with melting-point depression. We assessed the likelihood of the presence of hydrogen bonds in the four new polymorphic forms of ST from their IR and 13 C solidstate NMR spectra. The ST carbonyl group in the -form did not appear to be hydrogen bonded, in contrast to the -form carbonyl group. The crystal structures of the - and -forms were determined from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement. These crystal structures supported the conclusions drawn from the IR and solid-state NMR spectra, that there were hydrogen bonds between the molecules in these crystal forms. Acknowledgments The authors would like to thank Dr. R. Tamura and Dr. H. Tsue of the Graduate School of Human and Environmental Studies, Kyoto University, for conformation analyses, and Dr. H. Takahashi of the Graduate School of Human and Environmental Studies, Kyoto University, for performing NMR measurements. References Addadi, L., Berkovich-Yellin, Z., Domb, N., Gati, E., Lahav, M., Leiserowits, L., 1962. Resolution of comglomerates by stereoselective habit modifications. Nature 296, 21–26.
Blagden, N., Davey, R.J., Rowe, R., Roberts, R., 1998. Disappearing polymorphs and the role of reaction by-products: the case of sulphathiazole. Int. J. Pharm. 172, 169–177. Brittain, H.G., 2012. Polymorphism and solvatomorphism 2010. J. Pharm. Sci. 101, 464. Engel, G.E., Wilke, S., König, O., Harris, K.D.M., Leusen, F.J.J., 1999. Powder Solve–A complete package for crystal structure solution from powder diffraction patterns. J. Appl. Crystallogr. 32, 1169–1179. Grunenberg, A., Henck, J.-O., Siesler, H.W., 1996. Theoretical derivation and practical application of energy/temperature diagrams as an instrument in preformulation studies of polymorphic drug substances. Int. J. Pharm. 129, 147–158. Harris, K.D.M., Tremayne, M., Kariuki, B.M., 2001. Contemporary advances in the use of powder X-ray diffraction for structure determination. Angew. Chem. Int. Ed. Engl. 40, 1626–1651. Harris, K.D.M., Tremayne, M., Lightfoot, P., Bruce, P.G., 1994. Crystal structure determination from powder diffraction data by Monte Carlo methods. J. Am. Chem. Soc. 116, 3543–3547. Ikeda, Y., Ishihara, Y., Moriwaki, T., Kato, E., Terada, K., 2010. A novel analytical method for pharmaceutical polymorphs by terahertz spectroscopy and the optimization of crystal form at the discovery stage. Chem. Pharm. Bull. 58, 76–81. Jones, W., 1997. Organic Molecular Solids. Properties and Applications, 1st ed. CRC Press, Boca Ratón, FL, U.S.A. Kawakami, K., 2007. Reversibility of enantiotropically related polymorphic transformations from a practical viewpoint: thermal analysis of kinetically reversible/irreversible polymorphic transformations. J. Pharm. Sci. 96, 982–989. Kitamura, M., Ishizu, T., 1998. Kinetic effect of l-phenylalanine on growth process of l-glutamic acid polymorph. J. Cryst. Growth 212, 225–235. Kuroda, K., Yokoyama, T., Uneda, T., Kira, Y., Konishi, A., Kuroda, T., 1979. Studies on drug nonequivalence VIII. Solubilities of polymorphs and hydrate of mercaptoprine. Yakugaku Zasshi 99, 745–751. Maruyama, S., Ooshima, H., 2000. Crystallization behavior of taltirelin polymorphs in a mixture of water and methanol. J. Cryst. Growth 212, 239–245. Mesley, R.J., Clements, R.L., 1967. Infrared identification of barbiturates with particular reference to the occurrence of polymorphism. J. Pharm. Sci. 20, 341–347. Miura, H., Ushio, T., Nagai, K., Fujimoto, D., Lepp, Z., Takahashi, H., Tamura, R., 2003. Crystallization of a desired metastable polymorph by pseudoseeding crystal structure solution from its powder X-ray diffraction data, and confirmation of polymorphic transition. Cryst. Growth Des. 3, 959–965. Neumann, M.A., 2003. X-Cell: a novel indexing algorithm for routine tasks and difficult cases. J. Appl. Crystallogr. 36, 356–365. Pauling, L., 1960. The Nature of the Chemical Bond, 3rd ed. Cornell University Press, New York, U.S.A., pp. 479–480. Rietveld, H.M., 1969. A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2, 65–71. Stephenson, G.A., 2000. Structure determination from conventional powder diffraction data: application to hydrates, hydrochloride salts, and metastable polymorphs. J. Pharm.Sci. 89, 958–966. Takahashi, H., Tamura, R., Fujimoto, D., Lepp, Z., Kobayashi, K., Ushio, T., 2002. Preferential enrichment: full crystallographic analysis of the unusual phenomenon in the mixed crystals’ version. Chirality 14, 541–547. Takahashi, H., Tamura, R., Lepp, Z., Kobayashi, K., Ushio, T., 2001. Preferential enrichment: an essential crystal structure. Enantiomer 6, 57–66. Takahashi, H., Tamura, R., Ushio, T.z, Nakajima, Y., Hirotsu, K., 1998a. Ideal enantiomeric resolution (preferential enrichment) by recrystallization of a racemic compound. V: relationship between preferential enrichment and crystal structures. Chirality 10, 705–710. Takahashi, H., Tamura, R., Ushio, T., Nakai, Y., Hirotsu, K., Toda, F., 1998b. Preferential enrichment: mode of polymorphic transformation of a mixed crystal into a racemic compound crystal. Mol. Cryst. Liq. Cryst. 313, 211–216. Tamura, R., Iwama, S., Gonnade, R.G., 2011. Control of polymorphic transition inducing preferential enrichment. Cryst. Eng. Commun. 13, 5269–5280. Tamura, R., Takahashi, H., Hirotsu, K., Nakajima, Y., Ushio, T., 2001a. Preferential enrichment and crystal structure. Mol. Cryst. Liq. Cryst. 356, 185–194. Tamura, R., Takahashi, H., Hirotsu, K., Nakajima, Y., Ushio, T., Toda, F., 1998. Unusual disordered crystal structure of a racemate exhibiting a novel enantiomeric resolution: preferential enrichment. Angew. Chem. Int. Ed. 37, 2876–2878. Tamura, R., Takahashi, H., Miura, H., Lepp, Z., Nakajima, Y., Hirotsu, K., Ushio, T., 2001b. Comparison of crystal structures of new racemic chiral compounds showing and not showing the phenomenon of preferential enrichment. Supramol. Chem. 13, 71–78. Tamura, R., Ushio, T., Takahashi, H., Nakamura, K., Azuma, N., Toda, F., Endo, K., 1997. Ideal enantiomeric resolution by recrystallization of a racemic compound (part 4): relationship between enantiomeric resolution phenomenon and crystal properties. Chirality 9, 220–224. Thirunahari, S., Atipamula, S., Chow, P.S., Tan, R.B.H., 2010. Conformational polymorphism of Tolbutamide: a structural, spectroscopic, and thermodynamic characterization of Burger’s forms I–IV. J. Pharm. Sci. 99, 2975–2990. Tozuka, Y., Ito, A., Seki, H., Oguchi, T., Yamamoto, K., 2002. Characterization and quantitation of clarithromycin polymorphs by powder X-ray diffractometry and solid-state NMR spectroscopy. Chem. Pharm. Bull. 50, 1128–1130. Ushio, T., Endo, K., Takahashi, H., Yamamoto, K., 1996a. Physicochemical properties of suplatast tosilate racemate and enantiomers. Yakugaku Zasshi 116, 866–875. Ushio, T., Miura, H., Nagai, K., 2002. Suplatast tosilate crystals. Patent US7955607B2.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91 Ushio, T., Tamura, R., Takahashi, H., Yamamoto, K., 1996b. Unusual enantiomeric resolution phenomenon observed upon recrystallization of a racemic compound. Angew. Chem. Int. Ed. Engl. 35, 2372–2374. Ushio, T., Tamura, R., Azuma, N., Nakamura, K., Toda, F., Kobayashi, K., 1996c. Polymorphism of compound existing as both racemic compound crystals and mixed crystals of enantiomers. Mol. Cryst. Liq. Cryst. 276, 245–252. Ushio, T., Yamamoto, K., 1994. High-performance liquid chromatography of enantiomers of {2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl}dimethylsulfonium p-toluenesulfonate (suplatast tosilate) on a cellulose
91
tris-3,5-dimethylphenylcarbamate column. J. Chromatogr. A 684, 235–242. Yamamoto, K., Limwikrant, W., Moribe, K., 2011. Analysis of molecular interactions in solid dosage forms; Challenge to molecular pharmaceutics. Chem. Pharm. Bull. 59, 147–154. Yang, S.S., Guillory, J.K., 1972. Polymorphism in sulfonamides. J. Pharm. Sci. 61, 26–40. Young, Ed, R.A., 1993. The Rietveld Method (IUCr Monographies of Crystallography 5). Oxford University Press, Oxford.
Contents lists available at ScienceDirect
International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm
Four new polymorphic forms of suplatast tosilate Keiko Nagai a,e,∗ , Takanori Ushio b , Hidenori Miura c , Takashi Nakamura d , Kunikazu Moribe e , Keiji Yamamoto e a
Taiho Pharmaceutical Co. Ltd., Ebisuno, Hiraishi, Kawauchi-cho, Tokushima 771-0194, Japan Ace Japan Co. Ltd., Higashine-koh, Higashine, Yamagata 999-3701, Japan Taiho Pharmaceutical Co. Ltd., Kamikawa-machi, Kodama-gun, Saitama 367-0241, Japan d RIKEN, Wako, Saitama 351-0198, Japan e Graduate School of Pharmaceutical Sciences, Chiba University, Inohana, Chuo-ku, Chiba 260-8675, Japan b
c
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 20 October 2013 Accepted 29 October 2013 Available online 6 November 2013 Keywords: Suplatast tosilate Polymorphs Seeding analogous compounds Physicochemical properties Solubility Rietveld refinement
a b s t r a c t We found four new polymorphic forms (␥-, -, -, and -forms) of suplatast tosilate (ST) by recrystallization and seeding with ST-analogous compounds; three polymorphic forms (␣-, -, and ␦-forms) of ST have been previously reported. The physicochemical properties of these new forms were investigated using infrared (IR) spectroscopy, solid-state nuclear magnetic resonance (NMR) spectroscopy, differential scanning calorimetry, and powder X-ray diffractometry. The presence of hydrogen bonds in the new forms was assessed from the IR and solid-state NMR spectra. The crystal structures of the - and forms were determined from their powder X-ray diffraction data using the direct space approach and the Monte Carlo method, followed by Rietveld refinement. The structures determined for the - and -forms supported the presence of hydrogen bonds between the ST molecules, as the IR and solid-state NMR spectra indicated. The thermodynamic characteristics of the seven polymorphic forms were evaluated by determining the solubility of each form. The ␣-form was the most insoluble in 2-propanol at 35 ◦ C, and was thus concluded to be the most stable form. The -form was the most soluble, and a polymorphic transition from the - to the ␣-form was observed during solubility testing. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Polymorphism has long been a subject of great concern in the pharmaceutical industry because different solid pharmaceutical forms can impart different physical and chemical properties that can occasionally affect the bioavailability, dissolution rate, stability, and manufacturability of solid dosage forms (Brittain, 2012; Grunenberg et al., 1996; Jones, 1997; Kawakami, 2007; Thirunahari et al., 2010; Yamamoto et al., 2011). The characterization of solid pharmaceuticals is therefore very important to ensure that pharmaceuticals with standardized qualities are supplied to the consumer. There have been many studies in which solid pharmaceuticals have been physicochemically characterized, and techniques such as differential scanning calorimetry (DSC), infrared (IR) spectroscopy, microscopy, powder X-ray diffractometry (PXRD), Raman spectroscopy, and solid-state nuclear magnetic resonance (NMR) spectroscopy have been used (Ikeda et al., 2010;
∗ Corresponding author at: Taiho Pharmaceutical Co. Ltd., Ebisuno, Hiraishi, Kawauchi-cho, Tokushima 771-0194, Japan. Tel.: +81 88 665 3571; fax: +81 88 665 7225. E-mail address: [email protected] (K. Nagai). 0378-5173/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2013.10.049
Tozuka et al., 2002). Most polymorphs can be identified using these analytical techniques, but some polymorphs are still difficult to identify. Polymorphism is defined as the occurrence of different arrangements and/or conformations of molecules within the crystal lattice. Elucidating the arrangement and/or conformation of the molecules is often crucial to understanding the solid-state chemistry of active pharmaceutical ingredients. The crystal structure is one of the most crucial pieces of information relating to polymorphism; however, it is not always possible to determine the crystal structure of a metastable crystalline form because single crystals that are suitable for X-ray crystallographic analysis may be unavailable. Such difficulties have recently been overcome by advances in the procedures available for solving structures from powder X-ray diffraction data measured using laboratory X-ray sources (Engel et al., 1999; Harris et al., 1994, 2001; Neumann, 2003; Rietveld, 1969; Stephenson, 2000; Young and Ed, 1993), although there still are some limitations. Suplatast tosilate (ST) [(±)-[2-[4-(3-ethoxy-2-hydroxypropoxy) phenylcarbamoyl]ethyl]dimethylsulfonium p-toluenesulfonate] is an excellent IgE antibody production suppressor, and is a useful therapeutic agent for various allergic diseases. ST is a glycerol
84
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
derivative with one chiral carbon atom, which has been developed as a racemate. It has been reported that changing the crystallization conditions, such as the solvent(s) and temperature, pseudo-seeding the supersaturated solution with a crystal with the desired structure (following an epitaxy protocol), or adding impurities or additives to inhibit the nucleation and/or crystal growth of an undesired polymorph can be used to give a desired polymorph (Maruyama and Ooshima, 2000; Blagden et al., 1998). Using this technique, polymorphs of ST and ST analogous compounds were also prepared (Takahashi et al., 1998a, 1998b, 2002; Tamura et al., 1997, 1998, 2001a, 2001b, 2011). Three polymorphic forms (␣-, -, and ␦-forms) of ST were previously reported, and their crystal structures were determined (Miura et al., 2003; Takahashi et al., 2001; Ushio et al., 1996a, 1996b, 1996c, 2002; Ushio and Yamamoto, 1994). We have found four new polymorphic forms (␥-, -, -, and -forms) by recrystallizing and seeding with ST-analogous compounds. We investigated the physicochemical properties of the new polymorphic forms of ST by DSC, IR and solid-state NMR spectroscopies, and PXRD. We also determined the solubilities of the new polymorphic forms, and compared them with the solubilities of the three crystals that were previously characterized. We determined the crystal structures of the - and -forms from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement.
variable torsional angles are shown in Fig. 1. All other chemicals used were of special reagent grade. 2.2. Preparation of ˛-, ˇ-, and ı-form ST crystals The ␣-, -, and ␦-forms of ST were prepared following the previously reported procedures (Miura et al., 2003; Tamura et al., 1997; Ushio et al., 1996c). 2.3. Preparation of -form ST crystals Suplatast tosilate (5 g) was dissolved in chloroform (20 mL) and the solvent was evaporated under reduced pressure. Crystals prepared in this way were designated as the ␥-form. 2.4. Preparation of -form ST crystals Suplatast tosilate (1 g) was dissolved with heating in 2propanol (4 mL). After cooling the solution to room temperature, (±)-[2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]trimethylammonium p-toluenesulfonate (2) (approximately 5 mg) was added to act as pseudo-seed crystals. The wet crystals obtained from the solution were added to another solution of ST (5 g) in 2-propanol (20 mL) to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form.
2. Materials and methods 2.5. Preparation of -form ST crystals 2.1. Materials Suplatast tosilate [(±)-[2-[4-(3-ethoxy-2-hydroxypropoxy) phenylcarbamoyl]ethyl] dimethylsulfonium p-toluenesulfonate] (1a), which has been developed as a racemate, was provided by Taiho Pharmaceutical Co. Ltd. (Tokyo, Japan). Analogous compounds of ST, (±)-[2-[4-(3-propoxy-2-hydroxypropoxy) p-toluenesulfonate phenylcarbamoyl]ethyl]dimethylsulfonium (1b) and (±)-[2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]trimethylammonium p-toluenesulfonate (2), were provided by the Graduate School of Human and Environmental Studies, Kyoto University (Japan), and Taiho Pharmaceutical Co. Ltd., respectively. The chemical structures of ST, the ST-analogous compounds and the structure formula of (±)-1a with the six
⋅
Suplatast tosilate (1 g) was dissolved with heating in 2-propanol (4 mL). After cooling the solution to room temperature, (±)-[2-[4(3-propoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl]dimethylsulfonium p-toluenesulfonate (1b) (approximately 2 mg) was added to act as pseudo-seed crystals. The wet crystals obtained from the solution were added to another solution of ST (5 g) in 2-propanol (20 mL) to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form. 2.6. Preparation of -form ST crystals Suplatast tosilate (5 g) was dissolved with heating in a mixture of acetone (30 mL) and water (0.75 mL). After cooling the solution to 5 ◦ C, the ␣-form of ST crystals (1a) (approximately 5 mg) was added to act as seed crystals. The crystals that formed were filtered and dried, and classed as the -form. 2.7. Powder X-ray diffraction measurements
⋅
Powder X-ray diffraction patterns of each polymorphic form of ST were determined on a Spectris X’Pert instrument (Spectris Co., Ltd., Tokyo, Japan) with Cu K␣ radiation at 40 kV and 30 mA, passed through a nickel filter. The analysis was performed at a continuous scanning rate of 2◦ /min over a 2 angular range of 5–35◦ . The diffractograms obtained were analyzed using Spectris diffraction software. 2.8. Differential scanning calorimetry
Fig. 1. The chemical structures of suplatast tosilate (1a), the suplatast tosilate analogous compounds (1b and 2), and the structure formula of (±)-1a with the six variable torsional angles.
Differential scanning calorimetry analysis was performed using a Rigaku 8230 instrument (Rigaku, Tokyo, Japan). A sample (5 mg) was weighed in an open aluminum pan and subjected to a 30–130 ◦ C thermal scan at a heating rate of 4 ◦ C/min with a 100 mL/min dry nitrogen purge.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
2.9. Infrared spectroscopy An infrared spectrum of each polymorphic form was recorded using a JASCO FT/IR-420 instrument (JASCO International Co., Ltd., Tokyo, Japan) and the potassium bromide disk method. 2.10.
13 C
solid-state nuclear magnetic resonance spectroscopy
13 C
solid-state NMR spectra of the ␣-, -, and ␥-forms of ST were measured using a Chemagnetics CMX Infinity 400 instrument (Chemagnetics, Fort Collins, CO, USA) operated at 100.3 MHz with a cross-polarization/magic angle spinning (CP/MAS) probe. CP/MAS with two-pulse phase modulated (TPPM) decoupling was used, and the sample was placed in a 4 mm outer diameter sample tube that was rotated at 12 kHz. The contact time and the repetition time were 5.0 ms, and 5.0 s, respectively. The spectral width was 40 kHz, and 3000 scans were recorded. 13 C solid-state NMR spectra of the ␦-, -, -, and -forms of ST were measured using a Chemagnetics CMX Infinity 400 instrument (Chemagnetics, Fort Collins, CO, USA), operated at 100.4 MHz, using ramped CP/MAS with TPPM decoupling. The sample was placed in a 5 mm outer diameter sample tube that was rotated at 10 kHz. The contact time and the repetition time were 3.0 ms, and 5.0 s, respectively. The spectral width was 40 kHz, and 400 scans were recorded. 2.11. Solubilities An excess amount of a polymorph sample was added to 2propanol (2 mL). The sample was vigorously shaken for 30 s every 5 min for 30 min, and then allowed to stand for 4 h at 35.0 ◦ C. The saturated solution was filtered through a 0.45 m membrane filter, diluted with methanol, and analyzed by HPLC. HPLC analyses were carried out using an octadecylsilanized silica gel column (Shinwa chemical, ULTRON TDP, 4.6 mm × 150 mm), a mixture of 2.5 mmol/L sodium 1-hexanesulfonate solution, acetonitrile, and acetic acid (300:100:3) as the mobile phase at flow rate of 0.9 mL/min, and a UV–vis spectrophotometer (250 nm) as detector. Powder X-ray diffraction was used to determine whether any of the polymorphic forms had been transformed into another during the solubility measurement process. The solubilities were determined in duplicate. 2.12. Solving the crystal structures from the powder X-ray diffraction data The Reflex Plus module in the Materials Studio software (Accelrys, Inc, USA) was used for all calculations. The PXRD patterns of the - and -form crystals of ST were indexed with the X-cell program using 30 reflections (2 < 32◦ ). The cell and function profile parameters obtained were refined using the Pawley method. The space group was determined from within the space group candidates consistent with the systematic absences by a trial-and-error method. Once the initial model molecular conformations of the - and -form crystals of ST were assigned, the structures were solved using the Monte Carlo/parallel tempering method in the PowderSolve software. Rietveld refinement was then performed, which involved the following processes: (i) the pseudo Voight function was used to simulate the peak shape; (ii) the background was determined by linear interpolation using 20 terms; (iii) the Finger–Cox–Jephcoat method was used for asymmetric refinement; and (iv) the March–Dollase method was applied to correct preferred orientation effects. Global isotropic factors were used to refine the temperature factors. The molecular conformation in each of the motion groups was refined by varying the torsional angles
85
around all of the single bonds. The Monte Carlo calculations and subsequent Rietveld refinements were repeated.
3. Results and discussion 3.1. Polymorphic behavior We obtained four new forms of ST by recrystallizing ST and seeding with ST-analogous compounds, and investigated their physicochemical properties. Visual observations of all seven crystalline forms are shown in Fig. 2. The crystalline forms had different morphologies. The ␣-Form crystals were massive and solid. The -, ␦-, and -Form crystals were relatively large but fragile. The ␥- and -Form crystals were fine, and the -form crystals were needles. The PXRD patterns of the seven crystalline ST forms are shown in Fig. 3. We confirmed that the ␣-, -, and ␦-form PXRD patterns coincided with the patterns that were previously reported (Miura et al., 2003; Takahashi et al., 2001; Ushio et al., 1996c, 2002). However, the PXRD patterns of the four new crystals obtained in this study were completely different from those of the other three forms, and were also different from each other; each form showed a high degree of crystallinity and distinct diffraction angles and intensities. These results strongly suggested that the four new crystal forms could be polymorphic forms of each other. Adding a small amount of an ST analogous compound to act as seed crystals may have induced three of the new ST polymorphs (␦-, -, and -forms) to crystallize. It is noteworthy that additives, which made up less than 0.5% of the recrystallization solution in all of the cases, caused new polymorphic forms to crystallize. Kuroda et al. found a new polymorphic form of 6mercaptopurine, which was six to seven times more soluble than the other known crystalline forms, by adding an analogous compound as a seed (Kuroda et al., 1979). Kitamura and Ishizu investigated the growth rate of l-glutamic acid in the presence of lphenylalanine, confirming that the polymorphic (␣- and -) forms had different growth rates (Kitamura and Ishizu, 1998). Addadi et al. successfully resolved d- and l-asparagine conglomerations by adding small amounts of the pure enantiomers to saturated solutions of d- and l-asparagine mixtures (Addadi et al., 1962). It is thought that in each of these mechanisms, the analogous compound adheres to a crystal surface and inhibits crystal growth. In this study, it is considered that this mechanism also effects the formation of new polymorphic forms of ST. Evaporating the chloroform from a supersaturated solution of ST was an effective way of crystallizing the ␥-form, but the other polymorphic forms were prepared by recrystallization. The DSC curves of the seven polymorphic forms of ST are shown in Fig. 4. Each curve showed an endothermic peak due to fusion between 75 and 87 ◦ C. The highest melting points were 86.7 ◦ C for the ␣-form and 86.5 ◦ C for the ␥-form, showing that these polymorphic forms had spatially stable molecular arrangements. In contrast, low melting points were observed for the ␦-, -, and -forms, indicating that these forms were metastable. Furthermore, the polymorphic forms prepared by pseudo-seeding with ST-analogous compounds (␦-, -, and -forms) had broad endothermic peaks caused by meltingpoint depression. The highest enthalpy of fusion was 39.9 kJ/mol for the -form, and the lowest was 23.2 kJ/mol for the -form. The IR spectra are shown in Fig. 5. For the known crystal structures (␣-, -, and ␦-forms), the carbonyl stretching band was found at 1664 cm−1 for the ␣-form (in which the carbonyl group was hydrogen bonded to a OH group), whereas this band was shifted to 1678 and 1684 cm−1 for the - and ␦-forms (in which the carbonyl groups did not interact with a proton donor group).
86
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
Fig. 2. Microscopy photographs of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
It is well known that different hydrogen bonding modes may cause polymorphism in many compounds, such as barbiturate derivatives, sulfonamides, and oxalic acid (Mesley and Clements, 1967; Yang and Guillory, 1972; Pauling, 1960). The occurrence of hydrogen bonds in the four new polymorphic forms of ST was assessed by comparing their IR spectra. The carbonyl group stretching band of the -form was found at 1658 cm−1 , indicating the presence of hydrogen bonds with a proton donor group. However,
the carbonyl group stretching band was found at 1682 cm−1 for the -form, suggesting that there were no hydrogen bonds present. The ␥- and -forms showed carbonyl stretching bands at 1675 and 1672 cm−1 , respectively. It was not appropriate to assess hydrogen bond formation in the polymorphs from these data. In general, OH and NH stretching bands that are not hydrogen bonded are found at 3600–3700 and 3400–3500 cm−1 , respectively, as intense and sharp peaks. The polymorphic forms of ST other than
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
87
Fig. 4. Differential scanning calorimetry curves of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
Fig. 3. Powder X-ray diffraction patterns of suplatast tosilate polymorphs: (a) ␣form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
the -form showed broad bands between 3000 and 3500 cm−1 , revealing that several types of hydrogen bonds were present. The -form had a sharp peak at 3495 cm−1 , suggesting that most of the NH groups were not hydrogen bonded. Since the - and -forms of ST had broad bands around 3500 cm−1 , we accessed that sharp peaks overlapped in the broad bands around 3500 cm−1 , suggesting that some of the NH groups were not hydrogen bonded. The 13 C solid-state NMR results are shown in Fig. 6. The spectra of all seven polymorphic forms of ST were readily distinguishable by the 13 C solid-state NMR. The resonances of methyl carbons in the dimethylsulfonium group observed around 20 –30 ppm were different from each other, and the presence of hydrogen bonds between molecules in the new polymorphic forms of ST
was assessed by comparing the NMR spectra with the spectra of the polymorphic forms that were previously characterized. The chemical shift of the carbonyl carbon was observed at 170.0 ppm for the ␣-form of ST, in which the carbonyl group is hydrogen bonded to an OH group, and at 167.9 and 166.7 ppm for the and ␦-forms, respectively. In both of the latter forms, the carbonyl group did not interact with a proton donor group. The chemical shift for the -form carbonyl was 167.1 ppm, suggesting that the carbonyl group was not hydrogen bonded. However, the carbonyl chemical shift for the -form was observed at 170.2 ppm; therefore, the carbonyl group in this form was assumed to be hydrogen bonded. The carbonyl chemical shifts for the ␥- and -forms were 169.1 and 169.9 ppm, respectively, and these intermediate chemical shift positions could not be used to assess whether the carbonyl groups were hydrogen bonded or not in these crystals. The results of physicochemical properties of the ST polymorphs are summarized in Table 1.
Fig. 5. Infrared spectra of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
88
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
Table 1 Summary of physicochemical characteristic for suplatast tosilate polymorphs. States
Preparation method
DSC
IR
NMR
Melting point (◦ C)
Enthalpy of fusion (kJ/mol)
Stretching vibration of carbonyl band (cm−1 )
Chemical shift of carbonyl group (ppm)
Crystal structure
␣-form
Known
Recrystallization
86.7
35.2
1664
170.0
Already determined Ref: Takahashi et al. (2001)
-form
Known
Recrystallization
80.5
34.8
1678
167.9
Already determined Ref: Ushio et al. (1996c)
␥-form
New
Evaporation
86.5
35.0
1675
169.1
–
␦-form
Known
Seeding an analogous compound
74.8
32.5
1684
166.7
Already determined Ref: Miura et al. (2003)
-form
New
Seeding an analogous compound
75.2
29.2
1682
167.1
Rietveld refinement
-form
New
Seeding an analogous compound
78.1
23.2
1658
170.2
–
-form
New
Recrystallization
83.0
39.6
1672
169.9
Rietveld refinement
into the ␣-form (although most of the crystals remained in the original ␦-form). The -form was mostly transformed into the ␣-form during the solubility measurements, and the ␥-form was completely transformed into the ␣-form. Even though the -form was found to be extremely soluble in the initial dissolution stage, the solubility measurement for this form was not completed because it rapidly transformed into the ␣-form. The solubility of the crystal forms at 35 ◦ C followed the order ␣ <  < ␥ < < ␦ < , with the ␣-form being the most stable. The apparent solubility of the -form was significantly high nearly six times that of the ␣-form. Since the ␥-form completely transformed to the ␣-form, it was the second most unstable crystal after the -form. 3.3. Determination of the crystal structure from the powder X-ray diffraction data
Fig. 6. 13 C solid-state nuclear magnetic resonance spectra of suplatast tosilate polymorphs: (a) ␣-form, (b) -form, (c) ␥-form, (d) ␦-form, (e) -form, (f) -form, and (g) -form.
3.2. Solubility studies
The crystal structures of the ␣-, -, and ␦-forms were previously reported (Miura et al., 2003; Ushio et al., 1996c). We determined the crystal structures of the - and -forms from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement (Table 3, and Figs. 7 and 8). After two cycles of Monte Carlo calculations (with 266 million trial structures per cycle) and the subsequent Rietveld refinement,
The thermodynamic characteristics of the seven polymorphic forms were evaluated by determining the solubility of each form in 2-propanol; the results are shown in Table 2. Once the solubility measurements had been completed, the remaining crystals were collected and subjected to PXRD measurements. The ␣-, -, and forms did not transform into other polymorphic forms during the solubility measurements, whereas the ␦-form partly transformed Table 2 Solubilities of suplatast tosilate polymorphs in 2-propanol (35 ◦ C).
␣-form -form ␥-form ␦-form -form -form -form a
Solubility (mg/mL)
Polymorphic form after solubility determination
20.95 33.63 35.87 42.22 134.28 73.62a 40.17
␣  ␣ ␣ ␣
-form was rapidly transformed to the ␣-form while measuring solubility.
Fig. 7. Molecular arrangement in -form of suplatast tosilate.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
89
Table 3 Crystal data for suplatast tosilate polymorphs. Cell parameters
␣-form
-form
␦-form
-form
-form
Chemical formula Formula weight Sample type Crystal system Space group a (Å) b (Å) c (Å) ˛ (deg) ˇ (deg) (deg) V (Å) Z Rpa , Rwpb Ref
C23 H33 NS2 O7 499.65 Single crystal Triclinic P-1 14.885 15.673 12.435 100.17 104.65 108.73 2550.6 4 0.063; 0.081 Takahashi et al. (2001)
C23 H33 NS2 O7 499.65 Single crystal Triclinic P1 10.748 15.616 8.234 100.50 108.47 85.51 1288.0 2 0.044; 0.047 Ushio et al. (1996c)
C23 H33 NS2 O7 499.65 Powder Triclinic P-1 9.895 15.030 8.909 98.24 90.28 108.15 1242.0 2 0.147; 0.190 Miura et al. (2003)
C23 H33 NS2 O7 499.65 Powder Monoclinic P21/c 21.380 12.044 10.756
C23 H33 NS2 O7 499.65 Powder Triclinic P-1 16.664 10.357 7.619 78.18 97.47 101.59 1255.4 2 0.167; 0.226 –
a
x Rp =
|cY sim (2i)−I exp (2i)−Y back (2i)|
b
y Rwp =
|I exp (2i)|
wi(I exp (2i))2
2503.5 4 0.146; 0.190 –
.
wi(cY sim (2i)−I exp (2i)−Y back (2i))
115.32
2
1/2 .
Fig. 9. Difference plot for the Rietveld simulated and experimental powder X-ray diffraction patterns for the -form of suplatast tosilate.
Fig. 8. Molecular arrangement in -form of suplatast tosilate.
the weighted- profile R value (Rwp ) converged at below 23%. The experimental and calculated diffraction patterns and the corresponding difference profiles are shown in Figs. 9 and 10. The -form crystal was found to be in a monoclinic system, and it was very different from the polymorphic forms that were previously reported. This crystal was composed of equal amounts of the two ST enantiomers (P21/c, Z = 4) and had a center of symmetry like the ␣- and ␦-forms. The R and S ST enantiomers in the -form were hydrogen bonded between the hydroxyl groups and ethoxy oxygen atoms (with an ˚ as in the ␦-form. However, unlike the O· · ·O distance of 2.873 A), other ST crystals, the -form was found to have a unique spatial
Fig. 10. Difference plot for the Rietveld simulated and experimental powder X-ray diffraction patterns for the -form of suplatast tosilate.
90
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91
molecular arrangement, in which the amide NH hydrogen was not hydrogen bonded to the sulfonate ion oxygen, which was consistent with the IR result. The -form crystal structure was similar to that of the ␣-form, consisting of a racemic mixture of R and S enantiomers in the unit cell (Z = 2) of the centrosymmetric space group P-1. The R and S enantiomers were found to be hydrogen bonded between some of a hydroxyl group and carbonyl oxygen (with an O· · ·O distance of ˚ similarly to the ␣-form but unlike the - and ␦-forms. 3.036 A), Concurrently, some of the hydroxyl groups in the -form crystal structure were also found to be hydrogen bonded to sulfonate ˚ as in the group oxygen atoms (with an O· · ·O distance of 2.547 A), form. From these conflicting results, we estimated that some of the hydroxyl groups should therefore be disordered, and free carbonyl groups did seem to exist. The - and -form crystal structures supported the conclusions of intermolecular hydrogen bonding drawn from the IR and 13 C solid-state NMR spectra. The determination of crystal structures of the ␥- and forms were also examined, however, they transformed into the ␣-form during the measurements and they are still being studied. 4. Conclusions We identified four new polymorphic forms (␥-, -, -, and forms) of ST by recrystallization and seeding with ST-analogous compounds. The physicochemical properties of the new ST polymorphs were investigated using DSC, IR, and 13 C solid-state NMR spectroscopies, and PXRD The PXRD patterns for the new crystal forms were completely different from the patterns for the three previously characterized polymorphs, as well as being different from each other. The highest melting points observed from DSC measurements were 86.7 ◦ C for the ␣-form and 86.5 ◦ C for the ␥-form, showing that both these forms were relatively stable. Relatively low melting points were found for the ␦-, -, and -forms (which were prepared by pseudo-seeding with ST-analogous compounds); these forms had broad endothermic peaks associated with melting-point depression. We assessed the likelihood of the presence of hydrogen bonds in the four new polymorphic forms of ST from their IR and 13 C solidstate NMR spectra. The ST carbonyl group in the -form did not appear to be hydrogen bonded, in contrast to the -form carbonyl group. The crystal structures of the - and -forms were determined from their PXRD data using the direct space approach in conjunction with the Monte Carlo method, followed by Rietveld refinement. These crystal structures supported the conclusions drawn from the IR and solid-state NMR spectra, that there were hydrogen bonds between the molecules in these crystal forms. Acknowledgments The authors would like to thank Dr. R. Tamura and Dr. H. Tsue of the Graduate School of Human and Environmental Studies, Kyoto University, for conformation analyses, and Dr. H. Takahashi of the Graduate School of Human and Environmental Studies, Kyoto University, for performing NMR measurements. References Addadi, L., Berkovich-Yellin, Z., Domb, N., Gati, E., Lahav, M., Leiserowits, L., 1962. Resolution of comglomerates by stereoselective habit modifications. Nature 296, 21–26.
Blagden, N., Davey, R.J., Rowe, R., Roberts, R., 1998. Disappearing polymorphs and the role of reaction by-products: the case of sulphathiazole. Int. J. Pharm. 172, 169–177. Brittain, H.G., 2012. Polymorphism and solvatomorphism 2010. J. Pharm. Sci. 101, 464. Engel, G.E., Wilke, S., König, O., Harris, K.D.M., Leusen, F.J.J., 1999. Powder Solve–A complete package for crystal structure solution from powder diffraction patterns. J. Appl. Crystallogr. 32, 1169–1179. Grunenberg, A., Henck, J.-O., Siesler, H.W., 1996. Theoretical derivation and practical application of energy/temperature diagrams as an instrument in preformulation studies of polymorphic drug substances. Int. J. Pharm. 129, 147–158. Harris, K.D.M., Tremayne, M., Kariuki, B.M., 2001. Contemporary advances in the use of powder X-ray diffraction for structure determination. Angew. Chem. Int. Ed. Engl. 40, 1626–1651. Harris, K.D.M., Tremayne, M., Lightfoot, P., Bruce, P.G., 1994. Crystal structure determination from powder diffraction data by Monte Carlo methods. J. Am. Chem. Soc. 116, 3543–3547. Ikeda, Y., Ishihara, Y., Moriwaki, T., Kato, E., Terada, K., 2010. A novel analytical method for pharmaceutical polymorphs by terahertz spectroscopy and the optimization of crystal form at the discovery stage. Chem. Pharm. Bull. 58, 76–81. Jones, W., 1997. Organic Molecular Solids. Properties and Applications, 1st ed. CRC Press, Boca Ratón, FL, U.S.A. Kawakami, K., 2007. Reversibility of enantiotropically related polymorphic transformations from a practical viewpoint: thermal analysis of kinetically reversible/irreversible polymorphic transformations. J. Pharm. Sci. 96, 982–989. Kitamura, M., Ishizu, T., 1998. Kinetic effect of l-phenylalanine on growth process of l-glutamic acid polymorph. J. Cryst. Growth 212, 225–235. Kuroda, K., Yokoyama, T., Uneda, T., Kira, Y., Konishi, A., Kuroda, T., 1979. Studies on drug nonequivalence VIII. Solubilities of polymorphs and hydrate of mercaptoprine. Yakugaku Zasshi 99, 745–751. Maruyama, S., Ooshima, H., 2000. Crystallization behavior of taltirelin polymorphs in a mixture of water and methanol. J. Cryst. Growth 212, 239–245. Mesley, R.J., Clements, R.L., 1967. Infrared identification of barbiturates with particular reference to the occurrence of polymorphism. J. Pharm. Sci. 20, 341–347. Miura, H., Ushio, T., Nagai, K., Fujimoto, D., Lepp, Z., Takahashi, H., Tamura, R., 2003. Crystallization of a desired metastable polymorph by pseudoseeding crystal structure solution from its powder X-ray diffraction data, and confirmation of polymorphic transition. Cryst. Growth Des. 3, 959–965. Neumann, M.A., 2003. X-Cell: a novel indexing algorithm for routine tasks and difficult cases. J. Appl. Crystallogr. 36, 356–365. Pauling, L., 1960. The Nature of the Chemical Bond, 3rd ed. Cornell University Press, New York, U.S.A., pp. 479–480. Rietveld, H.M., 1969. A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2, 65–71. Stephenson, G.A., 2000. Structure determination from conventional powder diffraction data: application to hydrates, hydrochloride salts, and metastable polymorphs. J. Pharm.Sci. 89, 958–966. Takahashi, H., Tamura, R., Fujimoto, D., Lepp, Z., Kobayashi, K., Ushio, T., 2002. Preferential enrichment: full crystallographic analysis of the unusual phenomenon in the mixed crystals’ version. Chirality 14, 541–547. Takahashi, H., Tamura, R., Lepp, Z., Kobayashi, K., Ushio, T., 2001. Preferential enrichment: an essential crystal structure. Enantiomer 6, 57–66. Takahashi, H., Tamura, R., Ushio, T.z, Nakajima, Y., Hirotsu, K., 1998a. Ideal enantiomeric resolution (preferential enrichment) by recrystallization of a racemic compound. V: relationship between preferential enrichment and crystal structures. Chirality 10, 705–710. Takahashi, H., Tamura, R., Ushio, T., Nakai, Y., Hirotsu, K., Toda, F., 1998b. Preferential enrichment: mode of polymorphic transformation of a mixed crystal into a racemic compound crystal. Mol. Cryst. Liq. Cryst. 313, 211–216. Tamura, R., Iwama, S., Gonnade, R.G., 2011. Control of polymorphic transition inducing preferential enrichment. Cryst. Eng. Commun. 13, 5269–5280. Tamura, R., Takahashi, H., Hirotsu, K., Nakajima, Y., Ushio, T., 2001a. Preferential enrichment and crystal structure. Mol. Cryst. Liq. Cryst. 356, 185–194. Tamura, R., Takahashi, H., Hirotsu, K., Nakajima, Y., Ushio, T., Toda, F., 1998. Unusual disordered crystal structure of a racemate exhibiting a novel enantiomeric resolution: preferential enrichment. Angew. Chem. Int. Ed. 37, 2876–2878. Tamura, R., Takahashi, H., Miura, H., Lepp, Z., Nakajima, Y., Hirotsu, K., Ushio, T., 2001b. Comparison of crystal structures of new racemic chiral compounds showing and not showing the phenomenon of preferential enrichment. Supramol. Chem. 13, 71–78. Tamura, R., Ushio, T., Takahashi, H., Nakamura, K., Azuma, N., Toda, F., Endo, K., 1997. Ideal enantiomeric resolution by recrystallization of a racemic compound (part 4): relationship between enantiomeric resolution phenomenon and crystal properties. Chirality 9, 220–224. Thirunahari, S., Atipamula, S., Chow, P.S., Tan, R.B.H., 2010. Conformational polymorphism of Tolbutamide: a structural, spectroscopic, and thermodynamic characterization of Burger’s forms I–IV. J. Pharm. Sci. 99, 2975–2990. Tozuka, Y., Ito, A., Seki, H., Oguchi, T., Yamamoto, K., 2002. Characterization and quantitation of clarithromycin polymorphs by powder X-ray diffractometry and solid-state NMR spectroscopy. Chem. Pharm. Bull. 50, 1128–1130. Ushio, T., Endo, K., Takahashi, H., Yamamoto, K., 1996a. Physicochemical properties of suplatast tosilate racemate and enantiomers. Yakugaku Zasshi 116, 866–875. Ushio, T., Miura, H., Nagai, K., 2002. Suplatast tosilate crystals. Patent US7955607B2.
K. Nagai et al. / International Journal of Pharmaceutics 460 (2014) 83–91 Ushio, T., Tamura, R., Takahashi, H., Yamamoto, K., 1996b. Unusual enantiomeric resolution phenomenon observed upon recrystallization of a racemic compound. Angew. Chem. Int. Ed. Engl. 35, 2372–2374. Ushio, T., Tamura, R., Azuma, N., Nakamura, K., Toda, F., Kobayashi, K., 1996c. Polymorphism of compound existing as both racemic compound crystals and mixed crystals of enantiomers. Mol. Cryst. Liq. Cryst. 276, 245–252. Ushio, T., Yamamoto, K., 1994. High-performance liquid chromatography of enantiomers of {2-[4-(3-ethoxy-2-hydroxypropoxy)phenylcarbamoyl]ethyl}dimethylsulfonium p-toluenesulfonate (suplatast tosilate) on a cellulose
91
tris-3,5-dimethylphenylcarbamate column. J. Chromatogr. A 684, 235–242. Yamamoto, K., Limwikrant, W., Moribe, K., 2011. Analysis of molecular interactions in solid dosage forms; Challenge to molecular pharmaceutics. Chem. Pharm. Bull. 59, 147–154. Yang, S.S., Guillory, J.K., 1972. Polymorphism in sulfonamides. J. Pharm. Sci. 61, 26–40. Young, Ed, R.A., 1993. The Rietveld Method (IUCr Monographies of Crystallography 5). Oxford University Press, Oxford.
![[Mutagenicity tests of suplatast tosilate (IPD-1151T)].](/assets/img/hold.png)
