Toward an Understanding of the Propensity for Crystalline Hydrate Formation by Molecular Compounds. Part 2

The propensity of molecular organic compounds to form stoichiometric or nonstoichiometric crystalline hydrates remains a challenging aspect of crystal engineering and is of practical relevance to fields such as pharmaceutical science. In this work, we address the propensity for hydrate formation of a library of eight compounds comprised of 5- and 6-membered N-heterocyclic aromatics classified into three subgroups: linear dipyridyls, substituted Schiff bases, and tripodal molecules. Each molecular compound studied possesses strong hydrogen bond acceptors and is devoid of strong hydrogen bond donors. Four methods were used to screen for hydrate propensity using the anhydrate forms of the molecular compounds in our library: water slurry under ambient conditions, exposure to humidity, aqueous solvent drop grinding (SDG), and dynamic water vapor sorption (DVS). In addition, crystallization from mixed solvents was studied. Water slurry, aqueous SDG, and exposure to humidity were found to be the most effective methods for hydrate screening. Our study also involved a structural analysis using the Cambridge Structural Database, electrostatic potential (ESP) maps, full interaction maps (FIMs), and crystal packing motifs. The hydrate propensity of each compound studied was compared to a compound of the same type known to form a hydrate through a previous study of ours. Out of the eight newly studied compounds (herein numbered 4–11), three Schiff bases were observed to form hydrates. Three crystal structures (two hydrates and one anhydrate) were determined. Compound 6 crystallized as an isolated site hydrate in the monoclinic space group P21/a, while 7 and 10 crystallized in the monoclinic space group P21/c as a channel tetrahydrate and an anhydrate, respectively. Whereas we did not find any direct correlation between the number of H–bond acceptors and either hydrate propensity or the stoichiometry of the resulting hydrates, analysis of FIMs suggested that hydrates tend to form when the corresponding anhydrate structure does not facilitate intermolecular interactions.


Scheme 1.
The library of N-heterocyclic compounds investigated herein for their propensity to form hydrates. REFCODEs for structures reported in the CSD of anhydrate (Anh) and hydrate (H2O) forms, and previously unreported structures (New) are listed.

Hydrates in the context of pharmaceutical industry
In orally delivered drug products, active drug substances are almost always used as crystalline solids because of the inherent stability and purity of crystal forms, reproducibility of properties and the scaleability of crystallization processes. There is strong motivation for pharmaceutical scientists to study the range of crystal forms that exist for drug compounds, as the physicochemical properties of a drug substance can be affected by its solid form. [1][2][3][4][5] Routine crystallization screening can be used to identify those crystal form(s), typically polymorphs, solvates or hydrates, with properties suitable for use in a drug product. These screening experiments, which can be conducted robotically, also facilitate setting of the parameters needed to control the formation of a specific crystal form during manufacturing. 6 The latter issue is especially important in the context of hydrates, as the small size of water molecules and its multi-faceted H-bonding capability can favor hydrate formation. Further, the ubiquity of water vapor means that hydrates can form spontaneously. 2,[7][8][9][10] Formation of a hydrate is not necessarily a problem as a hydrate can ultimately be more suited for use in a drug product than an anhydrate. 11 Indeed, the presence of water in a drug substance does not raise any serious regulatory concerns as there are no toxicology risks associated with its presence in the resulting drug product. Hydrate formation can, however, impact intermolecular interactions and alter the physicochemical properties (e.g. solubility, 12 hygroscopicity, 13 stability 14 and bioavailability 11 ) of a drug substance. 2,15 In addition, a change in the hydration state of a drug substance may impact mechanical behavior during tableting or grinding and thereby affect product performance. 4,15,16 Nevertheless, a hydrated solid form does not limit the use of a drug substance in a marketed drug product. 17 Indeed, identification of hydrated crystal forms is a step in drug development. 15,18 Overall, hydrates are often used in solid drug products, either as the active ingredient or an inactive excipient. 15,19,20 A survey of the literature has revealed that drug products based upon hydrates include creatine phosphate sodium, 21 morphine sulphate, 22 azithromycin, 23 erythromycin, 24 amoxicillin, 25 fosamax, 25 lipitor, 25 protonix, 25 darunavir, 26 lisinopril, 27 cefaclor, 28 ampicillin, 29 cephalexin, 30 cefadroxil, 31 theophylline, 32,33 nitrofurantoin 34 and paroxetine hydrochloride. 35 However, there are also cases where formation of a solvate or hydrate renders the substance less suitable for use in a drug product as exemplified by paracetamol, cimetidine or naproxen, the active ingredients of the respective over-the-counter (OTC) drug products Tylenol (Panadol), Tagamet and Aleve. Mometasone furoate, pazopanib and seratraline hydrochloride, present in prescription medicines, Elocon, Votrient, and Zoloft, respectively, also have concerns related to hydrate formation. 25,36 The process used to manufacture a formulated drug product can also be impacted by polymorpism or pseudopolymorphism in its excipients. 15 Hydrate formation by excipients such as lactose, glucose, magnesium stearate or calcium phosphate are among examples studied in the literature. 15,19 Control over hydrate and solvate formation in molecular compounds is therefore of particular interest to pharmaceutical science, where one-third of drug substances are thought to form crystalline hydrates. 18,36 Indeed, 31.9% of entries in the European Pharmacopeia (1991) are hydrates and 11.2% are solvates. 37 Similar statistics were reported in 1999 for organic compounds in general, with hydrates more prelevant (33%) than solvates (10%). 38 A study by Rodriguez-Spong et al. based on a survey of the Cambridge Structural Database (CSD) also indicated that hydrate formation for organic compounds occurs more frequently than solvate formation with organic solvents. 39

Synthesis of Compounds 1-11
1,4-Bis(4-pyridyl)benzene (1). Compound 1 was synthesised following a previously published method by our group by 2-fold Pd0-catalyzed Suzuki coupling of 4-pyridinylboronic acid with 1,4dibromobenzene. 40 A 250 mL oven-dried two-necked round bottom flask was cooled under N2 atmosphere and charged with 1,4-dibromobenzene (1.61 g, 6.85 mmol), 4-pyridinylboronic acid (2.52 g, 20.5 mmol), Pd(PPh3)4 (0.39 g, 0.42 mmol), powdered NaOH (1.10 g, 27.4 mmol), 30 mL of toluene, 20 mL of EtOH and 10 mL of distilled water. The resultant reaction mixture was refluxed at 110 o C. The contents dissolved completely to give clear yellow coloration over a period of 1.5 h. Heating was continued at reflux under N2 atmosphere for 2 d. The change in the color of reaction mixture from yellow to dark brown indicated completion of the reaction, which was further verified by TLC. Subsequently, the reaction mixture was cooled and extracted with CHCl3 and washed with brine solution. The organic phase was dried over anhyd Na2SO4 and concentrated in vacuo. The pure product was isolated by Silica gel column chromatography using CHCl3/pet. ether (40%) mixture as an eluent to afford 1 as a white solid in 96% yield.
Bis(Pyridin-4-ylmethylene)benzene-1,4-diamine (2). Compound 2 was prepared following a previously published method by our group. 41 p-phenylenediamine (0.11 g, 1.0 mmol) and 4pyridinecarboxaldehyde (0.21 g, 2.0 mmol) were ground until a free-flowing yellow powder was obtained. Upon addition of ~100 μL of MeOH the resultant yellow paste was further ground (ca. 10 min) until a free-flowing deep yellow powder was obtained as pure product in 95.4% yield.

General aspect
Slurry experiments of 4-11 were performed under ambient conditions. 50 mg of each compound was slurried in pure water up to 7 days. The volume of solvent used to suspend the sample was one-third of the volume required to dissolve it completely. Powder X-ray diffraction (PXRD) and thermogravimetric analyses (TGA) were used to determine the hydrate formation. To examine the relative stability of the isolated hydrates, mixtures of hydrate and anhydrate forms in 1:1 w/w ratio were slurried in mixed solvent system of EtOH and H2O in varied EtOH:H2O vol. ratios: 1:1, 5:1 and 1:5.
3,6-Di(pyridin-4-yl)-1,2,4,5-tetrazine (4). 50 mg of 4 was slurried in water at room temperature in a sealed glass vial. Aliquots of sample were removed after 7 days in order to record the PXRD and TGA patterns ( Figures S2 and S3). (5). 50 mg of 5 was slurried for 24 h in water at room temperature in a sealed glass vial. Competitive slurry was conducted using 50 mg of the mixture of hydrate and anhydrate of 5 in a 1:1 w/w ratio, and 1:1, 5:1 and 1:5 EtOH:H2O solvent system. Aliquots of sample were removed after 24 h in order to record the PXRD and TGA patterns ( Figures S4 and S5). (6). 50 mg of 6 was slurried for 24 h in water at room temperature in a sealed glass vial. Competitive slurry was conducted using 50 mg of the mixture of hydrate and anhydrate of 6 in a 1:1 w/w ratio, and 1:1, 5:1 and 1:5 EtOH:H2O solvent system. Aliquots of sample were removed after 24 h in order to record the PXRD and TGA patterns ( Figure S6 and S7). (7). 50 mg of 7 was slurried for 24 h in water at room temperature in a sealed glass vial. Competitive slurry was conducted using 50 mg of the mixture of hydrate and anhydrate of 7 in a 1:1 w/w ratio, and 1:1, 5:1 and 1:5 EtOH:H2O solvent system. Aliquots of sample were removed after 24 h in order to record the PXRD and TGA patterns ( Figures S8 and S9). N,N'-Bis(4-pyridylmethylene)naphthalene-1,5-diamine (8). 50 mg of 8 was slurried in water at room temperature in a sealed glass vial. Aliquots of sample were removed after 7 days in order to record the PXRD and TGA patterns (Figures S10 and S11). N,N'-(1,4-phenylenebis(methan-1-yl-1-ylidene))bis(4-(1H-imidazol-1-yl)aniline) (9). 50 mg of 9 was slurried in water at room temperature in a sealed glass vial. Aliquots of sample were removed after 7 days in order to record the PXRD and TGA patterns (Figures S12 and S13).