Characterization and Interconversion of Two Crystal Forms of NEt-3IB, a Retinoid X Receptor Agonist

Abstract

Retinoid X receptor (RXR) agonist NEt-3IB (1) is a candidate for the treatment of inflammatory bowel disease (IBD), and we have established a process synthesis of 1 in which the final product is obtained by recrystallization from 70% EtOH. However, we found that there are two crystal forms of 1. Here, to characterize and clarify the relationship between them, we conducted thermogravimetry, powder X-ray diffraction, and single crystal X-ray diffraction. The crystal forms were identified as the monohydrate form I and anhydrate form II. The crystal form I, obtained as a stable form by our established synthesis, was easily dehydrated simply by drying to afford the form II′, which was similar to the crystal form II obtained by recrystallization from anhydrous EtOH. Storage of the form II′ in air regenerated the form I. The molecular conformations of 1 in the crystals of the two forms are similar, and they can be reversibly interconverted. The solubility of the monohydrate form I and anhydrate form II was examined and the former was found to be less soluble than the latter. Thus, form I may be superior to form II for targeting IBD, because of higher delivery to the lower gastrointestinal tract and reduction of systemic side effects associated with lower absorption due to lower water solubility.

Introduction

Different crystal forms of a compound itself are known as polymorphs,^1,2) and often show differences in stability, solubility, and/or bioavailability.^3,4) Therefore, in the process development of active pharmaceutical materials, it is necessary to take into account the possibility of polymorphism.

We have been interested in the clinical potential of NEt-3IB (1, Fig. 1A), which is an agonist of retinoid X receptor (RXR), one of the nuclear receptors.⁵⁾ Bexarotene^6,7) (Targretin^®, Fig. 1A), an RXR agonist clinically used for the treatment of cutaneous T cell lymphoma (CTCL),⁸⁾ has been reported to have therapeutic effects in animal models of colorectal cancer and central nervous system disease,^9,10) but it induces marked triglyceride elevation and hypothyroidism.^11,12) In contrast, 1 causes lower triglyceride elevation than bexarotene.¹³⁾ Positron emission tomography (PET) imaging revealed that 1 accumulates in the liver immediately after oral administration to mice and migrates to the lower gastrointestinal tract via biliary excretion.^14,15) Focusing on the characteristic pharmacokinetics of 1 and the known anti-inflammatory effects of RXR agonists, we were interested in the therapeutic potential of 1 for inflammatory bowel disease (IBD). Indeed, Matsumoto et al. demonstrated a therapeutic effect of 1 in a mouse model of T-cell-mediated colitis.¹⁵⁾ Thus, we are aiming to develop 1 as a new anti-IBD agent. As a first step, we have established a process method to synthesize 1 in large amounts.¹⁶⁾ This method employs only lipophilic ether and EtOH as organic solvents, which can be recovered and reused, and affords 1 with at least 99% purity in 7 steps in a yield of 30% or more.

Fig. 1. (A) Chemical Structures of RXR Agonists NEt-3IB (1) and Bexarotene

(B) Time-dependent weight change of 1 (prepared by recrystallization from 70% EtOH and dried under reduced pressure overnight) during storage in air and effect of subsequent drying under reduced pressure at day 33 and 47.

In this process, the final product 1 was obtained by recrystallization from 70% EtOH, then dried under reduced pressure overnight. However, when the sample was stored in air, an increase in weight was observed, and this weight increase of the product was reversed by drying under reduced pressure (Fig. 1B). In this study, we found that there are two crystal forms of 1. The crystal forms were identified as the monohydrate form I and anhydrate form II. In addition, the solubility of the monohydrate form I and anhydrate form II was examined and the former was found to be less soluble than the latter. For targeting IBD, form I may be superior to form II because of higher delivery to the lower gastrointestinal tract and reduction of systemic side effects associated with lower absorption due to lower water solubility. The details are described below.

Results and Discussion

Elucidation of the Cause of the Weight Increase/Decrease of 1

First, we carried out elemental analysis of samples of 1 with and without drying. The data for the dried sample were in consistent with the expected composition of anhydrous 1, while the undried sample showed a lower percent carbon content, which is consistent with that of a monohydrate (Table 1). These results suggest that undried 1 is a hydrate.

Table 1. Comparison of Elemental Analysis Results with/without Drying

	Elemental composition (%)
	C	H	N
With drying	70.74	7.55	7.84
Without drying	68.34	7.46	7.56
1 anhydrate (calcd.)a)	70.76	7.92	7.86
1 monohydrate (calcd.)a)	67.35	8.08	7.48

a) Calculated with ChemDraw20.0.

Next, we conducted thermogravimetry analysis (TG) of dried and undried samples of 1 (Fig. 2). A weight decrease of 4.9% was observed between 30 and 83 °C only for the undried sample, which suggested hydration of 1. Assuming all 1 is a monohydrated, dehydration causes a weight decrease of about 4.8%. Differential thermal analysis (DTA) confirmed that both samples melted at around 192 °C, in agreement with previous findings.

Fig. 2. Thermogravimetry (TG, Left Axis, Red Line) and Differential Thermal Analysis (DTA, Right Axis, Blue Line) of 1

(A) Undried sample, (B) dried sample.

Next, we performed single crystal X-ray diffraction of 1 obtained by recrystallization from 70% EtOH (Fig. 3, Table 2, Supplementary Movie S1). The triclinic crystal of space group P1̄ contained two molecules of H₂O together with two molecules of 1 per unit cell. We defined this crystal as the form I.

Fig. 3. Single Structure X-Ray Diffraction of a Single Crystal of Form I (Obtained from 70% EtOH)

(A) The crystal structure. (B) Ortep structure of 1 monohydrate.

Table 2. The Statics of Form I

Experimental formula	C21H28N2O3・H2O
Formula weight	374.48
Crystal color	Habit colorless, prism
Crystal dimensions	0.100 × 0.050 × 0.030 mm
Crystal system	Triclinic
Lattice type	Primitive
Lattice parameters	a = 8.521(2) Å
	b = 9.0003(2) Å
	c = 14.884(4) Å
	α = 102.508(8)°
	β = 100.112(8)°
	γ = 106.564(6)°
Lattice volume	1033.5(4) Å3
Space group	P1̄ (#2)
Z value	2

Examination of Recrystallization Conditions

It was suggested that the crystals of 1 obtained from 70% EtOH exist in hydrated form I. Therefore, we investigated recrystallization using anhydrous EtOH, MeOH or isopropanol (iPrOH) as a solvent. The recrystallization yield, melting point and elemental analysis were evaluated and compared (Supplementary Fig. S1, Table 3). The 70% EtOH-recrystallized sample was obtained in quantitative yield, whereas the yields in the case of anhydrous EtOH and the other alcoholic solvents were around 75%. When the concentration of 1 in the filtrate after recrystallization was measured, it was 65 mM in anhydrous EtOH and 1.8 mM in 70% EtOH, raising concerns about its high solubility in absolute alcoholic solvents. The melting points of all the products showed no significant difference. In elemental analysis, a composition consistent with the monohydrate was obtained only for the product recrystallized from 70% EtOH and not dried. Similarly, TG-DTA showed a weight decrease at less than 100 °C only for the product recrystallized from 70% EtOH and undried (Supplementary Fig. S2). Thus, only crystals obtained from 70% EtOH without drying were considered to be in the monohydrate form I.

Table 3. Comparison of Physical Properties of Crystals of 1 Recrystallized under Various Conditions

Sample	Yield	M.p. (°C)	Elemental analysis (%)
			C	H	N
70% EtOH without drying	103% as anhydrate (98% as monohydrate)	194.9–196.2	68.34	7.90	7.47
70% EtOH with drying	—	195.8–196.5	70.74	7.55	7.84
EtOH	79%	195.7–196.3	70.71	7.78	7.86
MeOH	77%	195.7–196.5	70.79	7.89	7.82
iPrOH	73%	195.9–196.6	70.74	7.81	7.85

M.p.; melting point.

Next, powder X-ray diffraction (PXRD) analysis of these samples was performed. All samples showed the same diffraction pattern, except for the sample recrystallized from 70% EtOH without drying (Fig. 4).

Fig. 4. Powder X-Ray Diffraction Patterns of 1 Recrystallized from 70% EtOH without Drying (Blue), 70% EtOH with Drying (Purple), EtOH (Red), MeOH (Pink) and iPrOH (Orange), Respectively

To compare the lattice parameters changes associated with the putative hydration/dehydration, single crystal X-ray diffraction of the anhydrate obtained by recrystallization from 70% EtOH followed by drying should be performed. However, since dehydration made the crystals cloudy, we used a crystal obtained from anhydrous EtOH instead. Therefore, in order to verify the results of PXRD, we carried out X-ray structure analysis of a single crystal obtained from anhydrous EtOH (Fig. 5, Table 4, Supplementary Movie S2). Like the crystal form I (Fig. 3), the crystal of the anhydrate was triclinic with space group P1̄, and the unit cell contained two molecules of 1, with no H₂O. We defined this crystal as the form II. When the PXRD pattern was simulated using VESTA based on the single crystal X-ray diffraction data shown in Figs. 3 and 5, the results were in agreement with the measured data (Supplementary Figure S3). These results support the presence of two crystal forms in 1. In addition, although 1 is dehydrated by efflorescence, the lattice parameters of the form II obtained from anhydrous EtOH is the same as that of the sample recrystallized from 70% EtOH and dried.

Fig. 5. Single Structure X-Ray Diffraction of Form II (Obtained from Anhydrous EtOH)

(A) The crystal structure. (B) Ortep structure of 1 anhydrate.

Table 4. The Statics of Form II

Experimental formula	C21H28N2O3
Formula weight	356.46
Crystal color	Colorless, prism
Crystal dimensions	0.380 × 0.350 × 0.190 mm
Crystal system	Triclinic
Lattice type	Primitive
Lattice parameters	a = 8.5195(10) Å
	b = 10.3418(11) Å
	c = 12.7631(12) Å
	α = 102.168(3)°
	β = 74.987(4)°
	γ = 110.080(6)°
Lattice volume	1010.86(19) Å3
Space group	P1̄ (#2)
Z value	2

Comparison of the Crystal Structures of Two Crystal Forms

The molecular structures of 1 in the crystal forms I and II obtained by single crystal X-ray diffraction appear to be essentially the same (Fig. 6A). However, in the crystal structures (Figs. 3, 5), the length in direction a is similar (form I; 8.521(2) Å, form II; 8.5195(10) Å), but the length in direction b is increased in the form II (form I; 9.003(2) Å, form II; 10.3418(10) Å), and the length in direction c is decreased (form I; 14.884(4) Å, form II; 12.7631(12) Å). Regarding the lattice angles, b is decreased in the form II (form I; 100.112(8)°, form II; 74.987(4)°), while there is no significant change in other angles. Comparison of the single crystal X-ray diffraction data indicates that in the case of the form I, the carboxylic acids form hydrogen bonds via H₂O molecules, and the intermolecular oxygen-oxygen distance is 5.085(4) Å (Fig. 6B). In the form II, the oxygen-oxygen distance of the carboxylic acids of the two molecules in the unit cell is 2.586(2) Å (Fig. 6C). Furthermore, the nitrogen atom on the pyridine of 1 is located 2.844(4) Å from the oxygen atom of H₂O (Fig. 6B), which is consistent with hydrogen bond formation. This may be the reason for the shorter distance in direction b compared with the form II. This suggests that hydration/dehydration can occur reversibly without major change in the molecular conformation in the crystal structures, as shown in the schematic diagram (Fig. 6D). On the other hand, interestingly, even when form II was stored in air for a long period, the crystal form remained unchanged (Supplementary Fig. S4 and Supplementary Table S1). Thus, while form II cannot readily accommodate water molecules, the anhydrate obtained by drying form I (defined as form II′) may not be completely dehydrated; we think the remaining hydrated water serves as a trigger for further hydration and transition to form I.

Fig. 6. (A) Comparison of the Molecular Conformations of 1 Recrystallized from 70% EtOH (Form I, Blue) and from Anhydrous EtOH (Form II, Red)

(B, C) Comparison of the crystal structures of forms I and II. (D) Schematic diagram of changes in the crystal structures.

Solubility of Each Crystalline Form in Water

As mentioned in the introduction, 1 accumulates in the liver after oral administration and is rapidly excreted in bile.^14,15) This enables high migration to the lower gastrointestinal tract and reduces systemic side effects.¹³⁾ Since 1 is being developed as an oral drug, we compared the solubility of each crystal form such as the anhydrate obtained by crystallization from anhydrous EtOH (form II), the anhydrate obtained by crystallization from 70% EtOH followed by drying (form II), and the monohydrate (form I). The form II showed higher solubility than the form I, up to 1.37 µg/mL (Fig. 7). In addition, there was no significant difference between the solubilities of the form II obtained by crystallization from anhydrous EtOH and obtained by crystallization from 70% EtOH followed by drying. This is consistent with the above finding that the crystal structures were not significantly different. 1 is being developed as a new drug for the treatment of IBD, which is associated with inflammation in the lower gastrointestinal tract. Salazosulfapyridine (SASP), a therapeutic agent for ulcerative colitis, is poorly absorbed in the upper gastrointestinal tract due to its low water solubility, and is degraded by the intestinal flora of the lower gastrointestinal tract to form 5-aminosalicylic acid (5-ASA), which has anti-inflammatory properties.¹⁷⁾ In the case of RXR agonists, distribution to the systemic blood may lead to side effects such as hypothyroidism. Thus, the less soluble form of 1 is expected to be preferable in terms of delivery to the lower gastrointestinal tract and prevention of side effects. To predict the solubility of 1 in the stomach and intestinal tract, we compared the water solubility of each crystal form at pH 1.2 and 6.8. The water solubility of 1 was higher in the more acidic environment, probably due to pyridine protonation, though form II still showed significantly higher solubility (Fig. 7B and Supplementary Fig. S5). Examination of the residue by PXRD after water solubility tests of forms II and II′ indicated that the spectrum of form II corresponded to that of a mixture of I and II, and the spectrum of form II′ was similar to that of form I (Supplementary Figure S6). Considering that form II does not become hydrated even during long-term storage, it appears that the dissolution process occurs as the anhydrate (form II) and subsequently the monohydrate (form I) with lower water solubility is precipitated. Thus, the residue was thought to be the monohydrate (form I). Since form I shows significantly lower solubility even in the 2-h water solubility test, form I may be more effective from the viewpoint of delivering 1 to the affected area and reducing systemic side effects.

Fig. 7. Comparison of Time- and pH-Dependence of the Water Solubility of Forms I (Blue), II′ (Purple) and II (Red)

Data shown are mean ± standard deviation (S.D.). (N = 3). Bonferroni test. ****; p < 0.0001, ***; p < 0.001, **; p < 0.01, *; p < 0.1, NS; not significant.

Conclusion

NEt-3IB (1), a candidate for the treatment of inflammatory bowel disease, shows two crystal forms. To clarify the relationship between them, we employed TG-DTA, PXRD, and single crystal X-ray diffraction. These experiments revealed that the crystals obtained by recrystallization from 70% EtOH are the monohydrate, the form I. These crystals can be easily dehydrated to be the form II by drying under reduced pressure, while the form II is transformed to the form I during storage in air. Anhydrate of 1 in the form II gave higher water solubility than the form I. For targeting IBD, form I will be superior to form II that takes into consideration the transferability to the affected site.

Experimental

Preparation of RXR Agonist NEt-3IB (1)

NEt-3IB (1) was synthesized according to the previous report.¹⁶⁾

Weight Change Tracking of NEt-3IB (1)

Dried NEt-3IB (1, 3.56 g, 1.0 mmol) was accurately weighed into a glass bottle, sealed with filter paper, and stored in a desiccator. The weight of the container was measured at 10 a.m., and the weight change of NEt-3IB (1) was measured by subtracting the tare weight. The temperature and humidity inside the desiccator were measured using Humidity/Baro/Temperature data recorder (MHB-382SD, Mother Tool, Japan).

Elemental Analysis

This experiment was performed using a PerkinElmer, Inc. 2400II Elemental Analyzer. The combustion tube and reduction tube temperatures were set at 1253.15 and 923.15 K, respectively. Blank measurements were performed in the presence or absence of oxygen, and the blank signal was stabilized. Each test sample (1.5–3 mg) was accurately weighed and quantitatively converted into H₂O, CO₂, N₂ by combustion decomposition. The contents of C, H, N, etc. were determined using a thermal conductivity detector for each component. Separately, the elemental composition of the test sample was calculated from the calibration curve for each element created using standard samples.

Investigating of Recrystallization Conditions

NEt-3IB (1) was dried overnight at 25 hPa and 40 °C. Then, 1.0 g of 1 was added to a 3-necked 100-mL flask equipped with a stirring bar, a Dimroth condenser and a thermometer under an Ar atmosphere. Solvents were added to the flask and heated to prepare saturated solutions. These solutions were cooled to 30 °C with stirring (400 rpm). The precipitated solid was collected by filtration and dried at normal pressure and room temperature overnight. The samples were weighed, and the recrystallization yields were calculated. In the case of the sample recrystallized from 70% EtOH, after calculation of the recrystallization yield, a part was dried overnight at 25 hPa and 40 °C.

Measuring the Melting Point

The melting point of each sample was measured using a hot stage melting point measuring device (Yanagimoto Seisakusho, Japan). The measurement was performed three times for each sample, and the melting point range was recorded.

Thermogravimetric Analysis and Differential Thermal Analysis

Analysis was carried out using a Simultaneous Thermogravimetric Analyzer (STA7200RV, Hitachi, Japan) at the Industrial Technology Center of Okayama Prefecture. Measurements were performed from 30 to 300 °C at heating rate of 5 °C/min under an N₂ atmosphere.

Powder X-Ray Diffraction Analysis

Powder X-ray diffraction data were collected using a RIGAKU-TTRIII-MTA diffractometer. Finely ground samples were loaded into a 0.7 mm borosilicate glass capillary and mounted on the diffractometer operating in transmission geometry, and equipped with a Johansson monochromator using Cu Kα radiation and Lynxeye detector. An Oxford Cryosystems Cryostream was used to control the temperature of the sample prior to data collection. Data were collected over the angular range 5 ≤ 2θ/θ° ≤ 60 at 2°/min.

Single Crystal X-Ray Diffraction

Single crystal X-ray diffraction was carried out as described in Supplementary Materials. Briefly, the data were collected using graphite-monochromated Mo Kα radiation on a Rigaku R-AXIS RAPID diffractometer. The crystal structures were solved by direct methods. Fourier calculations and least-squares refinements were performed on F². All non-hydrogen atoms were refined anisotropically. All hydrogen atoms were refined using a ring model. The crystallographic data of form I and II are provided in the Supplementary Materials.

Solubility of NEt-3IB (1)

The samples (Form I, II’ and II, about 35.6 mg) were accurately weighed and placed in 15-mL centrifuge tubes. Phosphate-buffered saline (PBS) (about 10 mL, adjusted to pH 1.2, 6.8, and 7.4 with 6M HCl) were added to each tube. The suspensions were vortexed for 30 s, sonicated for 3 min, and shaken in a 37 °C electric water bath. Aliquots of 3 mL were taken after 2, 4 and 24 h and filtered using a syringe filter. The AUC at 250–350 nm in the UV absorption spectrum of the filtrates were measured using a UV-1800 (Shimadzu, Japan). The concentration of 1 in the filtrates was calculated by using a separately prepared calibration curve. Experiments under each condition were carried out three times, and the average value was taken.

Acknowledgments

The authors are grateful to the Division of Instrumental Analysis, Okayama University for the elemental analysis. TG-DTA was supported by Dr. Eiji Fujii (Industrial Technology Center of Okayama Prefecture). Part of the single structure X-ray diffraction was performed by Dr. Hiromi Ota (Advanced Science Research Center, Okayama University).

Funding

This work was supported by JST SPRING, Grant Number JPMJSP2126 (to Y.T.). This study received funding from AIBIOS K.K. The founder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication.

Author Contributions

Y.T. and H.K. conceived and designed the project. Y.T. and S.K. examined each recrystallization condition. Y.T. and H.K. performed thermogravimetric analysis and differential thermal analysis. Y.S. performed X-ray analysis. Y.T., M.F. and K.H. evaluated the solubility. The manuscript was written by Y.T. and H.K.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

Experimental flow chart, thermogravimetry and differential thermal analysis, predicted powder X-ray diffraction, single crystal X-ray diffraction method, supporting movies. The Supplementary Materials are available free of charge on the website.

References

• 1) Bernstein J., “Polymorphism in Molecular Crystals,” Clarendon Press, Oxford, 2002.
• 2) Brittain H. G., “Polymorphism in Pharmaceutical Solids,” Marcel Dekker AG, Basel, Switzerland, 1999.
• 3) Haleblian J. K., J. Pharm. Sci., 64, 1269–1288 (1975).
• 4) Datta S., Grant D. J. W., Nat. Rev. Drug Discov., 3, 42–57 (2004).
• 5) Takamatsu K., Takano A., Yakushiji N., Morohashi K., Morishita K., Matsuura N., Makishima M., Tai A., Sasaki K., Kakuta H., ChemMedChem, 3, 780–787 (2008).
• 6) Pileri A., Delfino C., Grandi V., Pimpinelli N., Immunotherapy, 5, 427–433 (2013).
• 7) Trautinger F., Knobler R., Willemze R., Peris K., Stadler R., Laroche L., D’Incan M., Ranki A., Pimpinelli N., Ortiz-Romero P., Dummer R., Estrach T., Whittaker S., Eur. J. Cancer, 42, 1014–1030 (2006).
• 8) Boehm M. F., Zhang L., Badea B. A., White S. K., Mais D. E., Berger E., Suto C. M., Goldman M. E., Heyman R. A., J. Med. Chem., 37, 2930–2941 (1994).
• 9) Janakiram N. B., Mohammed A., Qian L., Choi C. I., Steele V. E., Rao C. V., Neoplasia, 14, 159–168 (2012).
• 10) McFarland K., Spalding T. A., Hubbard D., Ma J. N., Olsson R., Burstein E. S., ACS Chem. Neurosci., 4, 1430–1438 (2013).
• 11) Sherman S. I., Clin. Lymphoma, 3, 249–252 (2003).
• 12) Standeven A. M., Thacher S. M., Yuan Y. D., Escobar M., Vuligonda V., Beard R. L., Chandraratna R. A., Biochem. Pharmacol., 62, 1501–1509 (2001).
• 13) Kakuta H., Ohsawa F., Yamada S., Makishima M., Tai A., Yasui H., Yoshikawa Y., Biol. Pharm. Bull., 35, 629–633 (2012).
• 14) Kobayashi T., Furusawa Y., Yamada S., Akehi M., Takenaka F., Sasaki T., Akahoshi A., Hanada T., Matsuura E., Hirano H., Tai A., Kakuta H., ACS. Med. Chem. Lett., 6, 334–338 (2015).
• 15) Matsumoto R., Takahashi D., Watanabe M., Nakatani S., Takamura Y., Kurosaki Y., Kakuta H., Hase K., Front. Pharmacol., 12, 715752 (2021).
• 16) Takamura Y., Morishita K. I., Kikuzawa S., Watanabe M., Kakuta H., Chem. Pharm. Bull., 70, 146–154 (2022).
• 17) Singleton J. W., Summers R. W., Kern F. Jr., Becktel J. M., Best W. R., Hansen R. N., Winship D. H., Gastroenterology, 77, 887–897 (1979).