Towards the Development of Novel Diclofenac Multicomponent Pharmaceutical Solids

: Multicomponent pharmaceutical materials offer new opportunities to address drug physicochemical issues and to obtain improved drug formulation, especially on oral administration drugs. This work reports three new multicomponent pharmaceutical crystals of the non-steroidal anti-inflammatory drug diclofenac and the nucleobases adenine, cytosine, and isocytosine. They have been synthesized by mechanochemical methods and been characterized in-depth in solid-state by powder and single crystal X-ray diffraction, as well as other techniques such as thermal analyses and infrared spectroscopy. Stability and solubility tests were also performed on these materials. This work aimed to evaluate the physicochemical properties of these solid forms, which revealed thermal stability improvement. Dissociation of the new phases was observed in water, though. This fact is consistent with the reported observed layered structures and BFDH


Introduction
Diclofenac (DIC), a phenylacetic derivative non-steroidal anti-inflammatory drug (NSAID), is widely used in human and veterinary practice for the treatment of acute and chronic pain as well as in inflammatory and degenerative rheumatic diseases [1,2]. Diclofenac exerts its action through the inhibition of cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2) enzymes, which inhibit the synthesis of prostaglandins [3]. According to the Biopharmaceutics Classification System (BCS), DIC is a class II drug with low solubility and high permeability [4]. Due to its low solubility (0.9 ± 0.1 μg/mL) [5], achieving its minimum effective concentration requires a higher dosage in the formulation. However, the side effects of DIC have shown dosage-dependency; these include gastrointestinal damage and bleeding, nausea, hepatotoxicity, or renal failure. Moreover, when DIC is administrated orally, its low solubility increases the residence time in the stomach and the contact with the gastric mucosa, increasing the risk of gastric damage [6]. However, poor solubility is a major drawback not only for DIC but also for other Active Pharmaceutical Ingredients (APIs). For that reason, significant efforts have been made by both the industry and academia to develop new methodologies to enhance the physicochemical properties of APIs. Pharmaceutical multicomponent solid forms have gained much interest in the last decade due to their great potential to overcome drug performance limitations [7]. These solid forms are crystalline materials composed of two or more components. At least one must be an API, and the other, called cocrystal former or coformer, must be pharmaceutically acceptable, which means to be recognized as a safe molecule. Both components are in a stoichiometric ratio and interact through non-covalent interactions, mainly hydrogen bonds. These non-covalent interactions guide the organization of the molecules in the crystalline structure and allow the modulation of the physicochemical properties without covalent alterations of the API, whose activity and efficacy remain intact [8]. The literature reports pharmaceutical salts and cocrystals of DIC with amide (isonicotinamide [9]), amine (metformin [10], L-proline [11]), and xanthine (theophylline [12]), as well as pyridine-based coformers [13].
Nucleobases, the main component of nucleic acids, have attracted interest from the crystal engineering point of view because they can establish different hydrogen bond patterns [14]. This ability has been explored previously to form cocrystals and salts through NH···O=C hydrogen bond motifs [15]. Amine-carbonyl synthon has a remarkable key role in the transfer of genetic information and nucleic acid-protein recognition [16]. Moreover, nucleobase-derived drugs exhibit different biological roles, including anti-viral, antibacterial, and antitumoral activities [17,18].
This work reports the synthesis and physicochemical characterization of new multicomponent forms with diclofenac and a nucleobase: adenine, cytosine, and isocytosine (Scheme 1). The single crystal structure of all solid forms is thoroughly described, providing valuable insights into the structural differences that drive their physicochemical properties, mainly stability and solubility. Scheme 1. Chemical formula of diclofenac (DIC), adenine (ADE), cytosine (CYT), and isocytosine (ICT).

Synthesis of Diclofenac Acid Form
Diclofenac acid form (DIC) was obtained from hydrolysis of DICNa. For this purpose, 5 mol of DICNa (1.590 g) were dissolved in 30 mL of ultrapure water (Milli-Q, Millipore, Burlington, MA, USA) at 40 °C. HCl 1 M was added dropwise to the solution until no more diclofenac was precipitating. The product was filtrated and washed three times with cold deionized water and let dry at 35 °C for 24 h. Powder X-ray diffraction (PXRD) was used to confirm the purity of DIC.

Coformer Selection
A search of the Cambridge Structural Database (CSD) [19] was conducted to identify complementary functional groups with the potential for molecular recognition with DIC. A virtual cocrystal screening was performed afterwards using COSMOQuick software [20] (COSMOlogic, Germany, Version 1.4), calculating the excess enthalpy (Hex) of mixing between DIC and selected coformers from an internal library.

General Procedure for Mechanochemical Synthesis
Mechanochemical experiments were carried out via liquid-assisted grinding (LAG) using a Retsh MM200 ball mill (Retsch, Haan, Germany) operating for 30 min at a 25 Hz frequency using methanol as a solvent.
Synthesis of DIC-ADE: A mixture of DIC (74.1 mg, 0.25 mmol) and ADE (33.80 mg, 0.25 mmol) in a 1:1 stoichiometric ratio was placed in a 10 mL stainless-steel jar along with 100 μL of methanol and two stainless-steel balls of 5 mm diameter.
Synthesis of DIC-CYT: A mixture of DIC (74.1 mg, 0.25 mmol) and CYT (22.20 mg, 0.25 mmol) in a 1:1 stoichiometric ratio was placed in a 10 mL stainless-steel jar along with 100 μL of methanol and two stainless-steel balls of 5 mm diameter.
Synthesis of DIC-ICT: A mixture of DIC (74.1 mg, 0.25 mmol) and ICT (22.20 mg, 0.25 mmol) in a 1:1 stoichiometric ratio was placed in a 10 mL stainless-steel jar along with 100 μL of methanol and two stainless-steel balls of 5 mm diameter.

Powder X-ray Diffraction (PXRD)
PXRD data were collected using a Bruker D8 Advance Series II Vαrio diffractometer (Bruker-AXS, Karlsruhe, Germany) equipped with a LYNXEYE detector and Cu-Kα1 radiation (1.5406 Å). Diffraction patterns were collected over a 2θ range of 5-60° using a continuous step size of 0.02° and a total acquisition time of 30 min.

Preparation of Single Crystals
Single crystals were grown from saturated solutions (methanol) of the polycrystalline material obtained from LAG synthesis. Suitable single crystals for X-ray diffraction studies were grown by slow solvent evaporation at room temperature for two days.

Single-Crystal X-ray Diffraction (SCXRD)
Measured crystals were prepared under inert conditions immersed in perfluoropolyether as protecting oil for manipulation. Suitable crystals were mounted on MiTeGen Micromounts™, and these samples were used for data collection. Data for DIC-ADE, DIC-CYT, and DIC-ICT were collected with a Bruker D8 Venture diffractometer with graphite monochromated Cu-Kα radiation (λ = 1.54178 Å). The data were processed with APEX4 suite [21]. The structures were solved by intrinsic phasing using the ShelXT program [22], which revealed the position of all non-hydrogen atoms. These atoms were refined on F 2 by a full-matrix least-squares procedure using an anisotropic displacement parameter [23]. All hydrogen atoms were located in difference Fourier maps and included as fixed contributions riding on attached atoms with isotropic thermal displacement parameters 1.2 or 1.5 times those of the respective atom. The Olex2 software was used as a graphical interface [24]. Intermolecular interactions were calculated using PLATON [25]. Molecular graphics were generated using Mercury [26,27]. The crystallographic data for the reported structures were deposited with the Cambridge Crystallographic Data Center as supplementary publication no. CCDC 2180776-2180778. Additional crystal data are shown in Table 1. Copies of the data can be obtained free of charge at https://www.ccdc.cam.ac.uk/structures/.

Thermal Analysis
Simultaneous thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) measurements were performed using a Mettler Toledo TGA/DSC1 thermal analyzer (Mettler Toledo, Columbus, OH, USA). Samples (3-5 mg) were placed into sealed aluminium pans and heated in a stream of nitrogen (100 mL min -1 ) from 25 to 400 °C at a heating rate of 10 °C min -1 .

Infrared Spectroscopy
Fourier-transform infrared (FT-IR) spectroscopic measurements were performed on a Bruker Tensor 27 FT-IR instrument (Bruker Corporation, Billerica, MA, USA) equipped with a single-reflection diamond crystal platinum ATR unit and OPUS data collection program. The scanning range was from 4000 to 400 cm -1 with a resolution of 4 cm -1 .

Stability Test
Stability in aqueous solution was evaluated through slurry experiments. Excess of powder samples of each phase was added to 1 mL of water and stirred for 24 h in sealed vials. The solids were collected, filtered, and dried for further analysis by PXRD.
Stability at accelerated ageing conditions was also studied: 200 mg of solid was placed in watch glasses and left at 40 °C in 75% relative humidity using a Memmert HPP110 climate chamber (Memmert, Schwabach, Germany). The samples were subjected to the above-accelerated stability conditions for two months. PXRD was used to monitor the stability of the solid forms.

Solubility Test
Solubility studies were performed using the Crystal16 equipment (Technobis Crystallization Systems, Alkmaar, The Netherlands) in water. The equipment comprised four individually controlled reactors, each with a working volume of 1 mL, allowing the measurement of cloud and clear points based on the turbidity of 16 aliquots of 1 mL of solution in parallel and automatically. Each solution was heated at 0.3 °C/min from 20 to 90 °C with a magnetic stirring rate of 700 rpm, held at this temperature for 10 min and then cooled to 20 °C at 0.3 °C/min. The dissolution temperature for each compound was measured using different amounts of solid, and the solubility data of the pure components were fitted to a quadratic equation [29] using the CrystalClear software (Technobis Crystallization Systems, Alkmaar, The Netherlands).

Coformer Selection
Before the experimental trials were conducted, a virtual cocrystal screening was performed to improve the success ratio. A survey on the Cambridge Structural Database (CSD version 5.43, update from June 2022) based on DIC resulted in 70 hits. After excluding the three reported polymorphs [28,30,31] and metal complexes [32][33][34][35][36][37][38][39][40][41], the dataset contained 28 hits corresponding to multicomponent forms (salts, cocrystals, hydrates, and solvates). Only in one hit, the dimer DIC-DIC, observed in the monoclinic DIC polymorphs [28,30], was maintained; meanwhile, the remaining hits exhibited common structural features for DIC salts, COO -···amine and COO -··· ammonium synthons, or cocrystals: COOH···N(pyridine) and COOH···N(imidazole) synthons. According to the above-mentioned, our main prerequisite for the coformer selection was having the above-referred Ngroups and being a safe molecule. From our library of coformers, two groups of molecules fulfil these criteria: amino acids and nucleobases. COSMOQuick software was used to validate our selection. Table 2 shows calculations from a list of candidates to form multicomponent crystals with DIC. The list includes our reported coformers and other coformer molecule involved in the formation of cocrystals/salts reported in the survey. Compounds with negative Hex values show an increased probability of forming cocrystals.

Mechanochemical Synthesis
Liquid-Assisted Grinding is a versatile and efficient methodology widely used to obtain pharmaceutical multicomponent solid forms [42]. A screening through LAG was conducted with DIC and all coformers listed in Table 2. Unfortunately, despite the promising results obtained by the COSMOQuick analysis, only those LAG synthesis with the coformers ADE, CYT, and ICT were successful and achieved new phases whilst the other coformers yielded only physical mixtures of the two components. The product of these reactions was characterized by PXRD and compared with the X-ray powder pattern of the parent components (Table S1, Figure S1). Only experiments using ADE, CYT, and ICT as coformers provided a new PXRD pattern and were used for subsequent characterization (Figure 1). After the screening procedure, the work was focused on the search for finetuning conditions to obtain multicomponent materials of DIC with ADE, CYT, and ICT. Neat grinding experiments resulted in physical mixtures of the components ( Figure S1). Neat grinding approach only led to physical mixtures ( Figure S2). It is reported that this synthesis technique sometimes yields products with low crystallinity, partial reactions, or not even a reaction at all [43]. However, it is well known that adding small amounts of liquids accelerate the reaction, which essentially drove us to the idea of using LAG synthesis instead. LAG experiments were then performed using methanol and different stoichiometries (1:1, 1:2, and 2:1) ( Figure S3). A new common pattern was observed in the three stoichiometries. However, in the 1:2 and 2:1 ratios, there were also peaks corresponding to the coformer and DIC, respectively. Only the 1:1 ratio provided unique different PXRD patterns. Comparing these patterns with those simulated from the crystal structures confirmed the monophasic nature of the bulk solids ( Figure S4).

Thermal Analysis
DSC was used to evaluate the thermal behaviour and to determine the melting point of the new DIC phases. Figure 5 shows the melting point of DIC, as well as the DSC traces of DIC-ADE, DIC-CYT, and DIC-ICT. Each plot shows a well-defined endothermic event that corresponds with the melting point of the material. A single endothermic transition indicates the absence of solvation or hydration phenomena and also demonstrates the stability of the phase until the melting point. Above the melting point, some endothermic events are also observed, corresponding to the degradation of the samples. The multicomponent materials display a melting point that falls in a region between the melting point of DIC (179 °C) and the coformer (ADE: 360 °C; CYT: 320-325 °C; ICT:248-254 °C). This feature has already been described by other researchers [46]. A higher melting point was obtained through multicomponent crystallization, resulting in better thermal stability, probably due to stronger intermolecular interactions between DIC and nucleobases. TGA showed no weight loss until melting, suggesting that the new DIC phases were not hydrated or solvated. The occurrence of mass loss was observed after melting points, which was attributed to the degradation of cocrystals ( Figure S5).

Fourier Transform Infrared (FT-IR) Spectroscopy
Due to its simplicity and reduced consumption of time and samples, FT-IR spectroscopy is a widely used technique for detecting new multicomponent materials [47]. Functional groups exhibit defined bands in the IR spectrum, and intermolecular interactions, such as hydrogen bonds, induce changes in the position of these bands. Hence, the study of the shifts can detect the formation of a cocrystal or a salt and gives information about the groups involved in the interaction [48]. Figure 6 shows the FT-IR spectra of DIC and DIC multicomponent materials. The DIC spectrum has a characteristic band at 3322 cm −1 , ascribed to the stretching mode of -NH. In DIC-ADE, DIC-CYT, and DIC-ICT, this band is shifted to 3326, 3300, and 3298 cm −1 , respectively. Another characteristic band of DIC is the C=O stretching vibration that appears at 1961 cm −1 . This band is shifted to 1671 (DIC-ADE), 1693 (DIC-CYT), and 1678 cm −1 (DIC-ICT). FT-IR data support the information observed in SCXRD, where -COOH and -NH groups from DIC and coformers, respectively, drive the formation of the crystalline structures.

Stability Studies
Thermodynamic stability in an aqueous solution was evaluated by placing an excess of the sample in a vial and stirring in deionized water at 25 °C. After 24 h, the product was filtered, dried at room temperature, and characterized by PXRD. Powder patterns showed high stability for the pure DIC and DIC-ADE cocrystal. However, DIC-CYT and DIC-ICT phases could not remain stable for more than 3 h and 30 min, respectively (Figures 7 and  S7). Multicomponent DIC phases were also stored at accelerating ageing conditions (40 °C, 75% RH). At the determined time, samples were characterized by PXRD to assess their stability. DIC was included in the experiment for better comparison with the new materials. Under these conditions, it was observed that all samples remained stable for two months (Figure 8).

Solubility Studies
As observed in the previous section, only the DIC-ADE cocrystal was thermodynamically stable in a water solution at room temperature. Initial attempts to determine solubility by the shake-flask method [49] were not possible due to the overlap of UV absorption maxima of both the API and coformer ( Figure S8). Evaluation of solubility was then performed by a polythermal method using the Crystal16 equipment. Results showed an improvement in the solubility of DIC-ADE (0.993 mg/mL, Figure 9) compared with the reported solubility of DIC (0.9 μg/mL) [5]. Although the difference in solubility between DIC and the DIC-ADE cocrystal was significant, the amount of solubility improvement is not significant compared with the solubility of the sodium salt (16.18 mg/mL) [50]. The layered structure observed in the DIC-ADE cocrystal directly impacts the solubility improvement of DIC. Different studies have reported the effect of a layer structure consisting of high-solubility molecules on the solubility of multicomponent solid forms [51][52][53]. Although the dimeric DIC structure is disrupted and a better solubility is obtained in DIC-ADE, the intercalated layers composed of low water-soluble ADE molecules do not confer enough solubility improvement themselves in comparison with other multicomponent DIC solids.
A potential risk observed in the use of multicomponent systems is their tendency to experience unexpected dissociation in contact with water or with high relative humidity (RH), which leads to a return to the respective free API and coformer [54,55] and denies the solubility advantage achieved by multicomponent solid formation. To rationalize the dissociation observed for DIC-CYT and DIC-ICT solids, crystal morphologies of the three reported multicomponent DIC forms were computed using the Bravais-Friedel-Donnay-Harker (BFDH) method included in the visualization software package Mercury [27]. As described previously, all the crystal structures consist of alternate layers of type DIC···coformer···DIC··· and these supramolecular arrangements seem responsible for the enhanced properties. Figure 10 shows the predicted morphologies for the reported multicomponent solids. Notably, the facets with the largest surface, following the order: DIC-ICT (57.2%) > DIC-CYT (50%) > DIC-ADE (24.2%), contain hydrogen bond donor and acceptor groups that potentially could interact with water during the dissolution process. Water solubility depends not only on the groups exposed on the surface of a crystal but on other different factors, including density, coformer solubility, or lattice energy [56]. We cannot argue that the high polarity of the crystal surfaces could impact the solubility performance by itself, but it does affect the dissolution of the reported solids as evidenced by the rapid dissociation observed for DIC-CYT and DIC-ICT during the slurry experiments in water. Dissolution is carried out at higher rate in the DIC-CYT and DIC-ICT species. As expected, dissolution in these species is favored by extensive surface exposure. On the other hand, dissolution of the DIC-ADE phase occurs at a slower rate, evidenced by the apparent stability at 24 h.

Conclusions
In conclusion, we have described three new pharmaceutical multicomponent crystals containing DIC and nucleobases (ADE, CYT, and ICT) as coformers. Expected heterosynthons assist formation of the new solid forms, disrupting the robust acid:acid dimmer synthon observed in reported DIC polymorphs. All solids consist of alternated layered structures connected by hydrogen bonds. This supramolecular organization confers good thermal stability, and good stability under accelerated ageing conditions and seems to have an important role in the dissolution properties of the solids. Relevant insights are inferred from the BFDH calculations where CYT and ICT, containing solids, possess a large crystal surface that expose hydrogen donor and acceptor groups, which interact with the water molecules of the bulk solvent.