Preparation and Evaluation of an Elusive Polymorph of Mefenamic Acid by High Pressure Techniques

High pressure crystallization technique has been successfully used to prepare an elusive form II of a non-steroidal anti-inflammatory drug, Mefenamic acid. Single crystal of form II was grown at 0.3 GPa from an 4:1 methanol/ ethanol mixture as a solvent using Diamond Anvil Cell. Comparison of crystal structures show that the efficient packing of MA molecules in Form II results from the structural flexibility of MA acid molecules. Compression studies performed on the single crystal of form I resulted in 14 % decrease of unit cell volume up to 2.5 GPa. No phase transition was observed up to this pressure. A reconstructive phase transition is required to induce conformational changes in the structure, which is confirmed by crystallization at high pressure results.


Introduction
The phenomenon polymorphism is solid state is the ability of a substance to crystallize with more than one crystal forms; called as polymorphs.The polymorphism of organic compounds is of fundamental interest to experimental and theoretical chemists, and is also of crucial importance to industry for a wide range of materials that includes pharmaceuticals, pigments, explosives (energetic materials), food products and proteins [1,2].In pharmaceuticals, polymorphism of the active pharmaceutical ingredient (API) is of great importance.Difference in the three dimensional arrangements and/or conformations of molecules within the unit cell of polymorphs can have profound effects on the physical and chemical properties of the API such as melting point, sublimation temperature, heat capacity, solubility, dissolution and bio-availability [2].Many processing properties are also dependent on the polymorphic behaviour of the active pharmaceutical agents (API) [1].Polymorphic inter-conversion during manufacturing process (such as compression, milling and drying) and storage of drug compounds is also of crucial importance, where changes in conditions e.g.temperature, relative humidity and pressure may results in the formation of new forms.Patent rights are also become an issue, as polymorphs of a same compound can be patented separately.The cases of Ritonavir (an HIV drug by Abbot Laboratories) and ranitidine hydrochloride (Zantac, the largest selling anti-ulcer drug from GlaxoSmithKline), provide examples of the importance of polymorphism in processing and patent litigation, where emergence of new polymorphs results in huge financial loss to the companies [3,4].For all the above reasons pharmaceutical companies deploy a great effort and resources for the identification of all the possible polymorphs in the early stage of development.
In recent years considerable effort has been put into developing high-throughput, automated approaches to polymorph screening where key crystallization variables such as solvent and antisolvent selection, evaporation and cooling rate followed by the characterization by spectrosopic, diffraction and thermodynamical techniques are employed [5].The main aim is to develop methods that are capable of producing and identifying large numbers of samples in a relatively short period of time.Even after these exhaustive tests, a new polymorph may remain undetected for many years, or a sample of a new polymorph may be obtained once but never again, the phenomenon of so-called "disappearing polymorphs" [6,7].Polymorphism remains one of the major challenges for both the industry and academic sectors.The development of complementary methods with better control on the crystallization conditions, and more importantly the understanding of factors that result in different packing arrangements of organic molecule in the solid state would therefore be highly desirable [8].
There is a need to explore alternative crystallisation conditions e.g.use of pressure as additional variable in conjunction with conventional strategies.Polymorph screening is normally performed at ambient conditions of pressure.Pressure can change the relative thermodynamic stability of polymorphs and also can have effect on the conformation of molecules as they pack within a crystal structure.It can act as an additional variable for the identification of new polymorphs.High-pressure has already been extensively applied in polymorphic studies of many simple organic compounds such as amines, ketones, carboxylic acids, alcohol and amino acids [9][10][11][12][13][14][15], where direct compression has successfully been used to generate new polymorphs.However sometime only compression is not enough for these compounds where kinetic barrier of molecular rearrangement is high and cannot be achieved by simple compression.To overcome this problem a technique has been developed which involves recrystallization of compound from solutions under elevated pressure (i.e.pressure induced crystallization).This has successfully been used to generate new polymorphs of simple compounds e.g.new polymorphs of parabanic acid [8], urea [16], acetamide and paracetamol [8,17].
Mefenamic acid (2-[(2, 3-dimethyphenyl)-amino] benzoic acid), is a well known fenamate derivative showing potent analgesic effect [18] and is marketed as Ponstel®.It belongs to the group of medicines called non-steroidal anti-inflammatory drugs (NSAIDs), which are among the most frequently used medicinal drugs.MA is known to exist as three polymorphic forms, denoted as forms I, II and III.These polymorphs show different solubility and stability behaviour [19,20].Form I is the most stable form under ambient conditions, but form II is the stable form above 160 o C, form III is the least stable form at ambient conditions and converts back to form I immediately [21].The crystal structures of form I, II and II was reported in 1976, 2008 and 2013, respectively [22].NMR spectroscopy and theoretical studies have shown that form I has some conformational similarities with related fenamate compounds e.g.flufenamic acid, niflumic acid and maclofenamic acid [23,24].
MA molecule shows structural flexibility that may results in more conformational polymorphs using high pressure conditions.Compression behaviour of the two polymorphs has been studied in connection with the mechanical properties of the finished tablets, but very little attention has been paid to possible polymorphic transitions of the active ingredient under mechanical treatment [25].
All polymorph screening of MA has been performed at ambient pressure [26].For all the above 3 of 13 reasons the aims of the current investigation were therefore to explore whether high-pressure methods could be used to obtain the elusive form II, or other new forms.

Materials and Methods
Mefenamic acid was obtained from the Sigma Aldrich, UK.Single crystals were obtained by slow cooling of a hot saturated solution (ca.1M) in THF.A small, colourless, rectangular shaped crystal was taken directly from the crystallised sample and was confirmed as being form I by indexing the diffraction data from single crystal X-ray diffraction.Crystals from the same solution were used in all of the single crystal compression experiments.Analytical grade solvents were used for the recrystallisation experiments.
High-pressure experiments were carried out using Merrill-Bassett diamond-anvil cell (DAC) (half-opening angle 40 o ) [27], equipped with 600 μm culet diamonds and a pre-indented tungsten gasket of thickness 250 μm and a 300 μm diameter hole.For high-pressure crystallisation experiments, diamond-anvil cells using tungsten carbide backing disc [28] were used.

Compression Studies
For compression studies a single crystal of suitable size was loaded into DAC along with a small piece of ruby in order to allow the determination of the pressure by laser fluorescence method [29] using a 632.8 nm excitation line from a He-Ne laser.The fluorescence was detected by a Jobin-Yvon LabRam 300.A methanol-ethanol (4:1) mixture was used as the pressure-transmitting medium [30].

Crystallization at High Pressure
For high pressure crystallization, a ca.1M solution of mefenamic acid was prepared in ethanol.
The solution was then loaded along with a few crystallites of form I at 293(2) K into a DAC.The pressure was increased to a minimum in order to form a sealed system and the cell was then gently heated to dissolve all of the solid material.Pressure was then applied by tightening the screws of the DAC to induce precipitation.Crystallisation was achieved at ca. 0.6 GPa.The pressure was then reduced to 0.3 GPa in order to facilitate growth of a single crystal.The temperature was then cycled near 353 K in order to dissolve all but one of the crystallites.On slow cooling to 298 K a single crystal grew.

Single Crystal X-ray Diffraction
Diffraction data were collected on a Bruker APEX II CCD diffractometer at 293(2) K using Mo-Kα radiation (λ = 0.71073).Data indexing was performed using CELL_NOW [31].Data processing was performed according to the procedure described by Dawson et al. (2004) [32].Integration of data sets and global cell refinement were carried out using the program SAINT [33], in which 'dynamic masks' were used in order to prevent integration of areas of the detector shaded by the body of the DAC.An analytical correction for the absorption by the DAC component was then applied using SHADE [34] and an absorption correction for the crystal was applied using SADABS [35].Known coordinates were taken from the literature [22].Refinement was then performed using CRYSTALS [36].All non-hydrogen atoms were refined isotropically and hydrogen atoms were placed in calculated positions.A completeness of only ca.40 % was obtained for this dataset and this led to a rather high R-factor of 9.5 %.Nevertheless, this was sufficient to identify the main structural feature of form II and make a comparison between the polymorphs.Full structural refinements details and Crystallographic data in CIF format is attached as supplementary information.

Direct compression of a single crystal of form I
The first data-set was collected at ambient pressure.Indexing of the reflections confirmed the known triclinic form I of MA.Diffraction data were then collected in incremental steps of pressure up to 2.5 GPa.Lattice parameters at pressures up to 2.5 GPa are shown in Table 1.On raising the pressure to 3.0 GPa the crystal disintegrated.All data-sets up to 2.50 GPa were processed according to the procedure described above.As a result of the limitations caused by shading from the steel body of the diamond-anvil cell, highpressure data-sets are frequently incomplete compared with data-sets recorded at ambient pressure.
These factors combined to make structural refinement particularly challenging and in this case resulted in poor structural refinement with high R-factors.Therefore, only lattice parameters are reported here.Figure 2 shows the variation of unit cell volume with increasing pressure.There was a ca.14 % decrease in the unit cell volume over the studied pressure range.Within the limits of experimental errors, the volume decreases smoothly over this pressure range, indicating that there is no phase transition associated with direct compression of the single crystal.The observation that at 3.0 GPa the crystal disintegrated to give a polycrystalline powder, suggests reconstructive phase transition at this pressure.These observations prompted a study by recrystallisation from solution at high pressure.

Crystallization from solution at high pressure
Crystal structure obtained from the single crystal grown at 0.3 GPa gave a triclinic unit cell with lattice parameters shown in Table 1.Crystallographic data obtained from the high pressure crystallization experiments is compared with already published forms I, II and III (Table 1).All three polymorphs (I−III) crystallize in the triclinic space group (P-1) with a molecule in the asymmetric unit; such an observation is relatively rare, wherein all the polymorphic forms consist of Z′ = 1.

Description of crystal structure
The molecular structure of MA with numbering scheme used in this work is shown in Figure 1.
The conformation of the molecule in the crystal structure can be described by three torsional angles (see Figure 1): θ1 (O1-C7-C6-C1) i.e. twisting of the carboxylic group with respect to C6-C7 axis; θ2 (C1-N1-C8-C13) i.e. twisting of the second phenyl group with respect to the N1-C8 axis; and θ3 (C6-C1-N1-C8) i.e. twisting of bridging group relative to the C1-N2 axis.These torsional angles are shown by arrows in Figure 1 and are calculated by using the program MERCURY [37].The crystal packing diagram of form I, II and II (high pressure form) are compared in Figure 3.
Although the polymorphs exhibit significant variation in the unit cell dimensions, the MA molecules pack as dimer units through an intermolecular hydrogen bond involving the carboxylic acid group (O2-H…O1) with an O2…O1.The length of this hydrogen bond is comparable in all the three forms with the value of ~1.68 Å.The phenyl ring bearing the carboxyl group is coplanar with the carboxyland bridging amino-groups.This coplanar conformation is stabilised by the resonance interaction and internal hydrogen bond between the amino-and carboxyl-groups [38].The distance between the atoms (N1…O1) involved in this interaction in form I, II and III 1.88, 2.01 and 2.09 Å, respectively.
The description of the hydrogen-bond pattern according to the graph-set notation at the first level graph set is N1 = S(6)R2 2 (8) (The graph-set notations are assigned by the method described by Bernstein et al. [39], and is obtained from PLUTO [40] and MERCURY [37]).Where S and R motifs represents internal and external hydrogen bonds, respectively.

Decompression studies
On progressive decompression from 0.3 GPa at 293 K to ambient pressure, no colour change or destruction of the crystal was observed.This tiny crystal was then removed from gasket, placed on the goniometer and was successfully indexed as form II.However, owing to poor quality and the damage done during unloading from the gasket it was not possible to obtain sufficiently good data for full structure refinements.The crystal was also used as a seed in order to attempt the growth of single crystals of form II at ambient pressure, but this resulted in the production of form I instead.
This suggests that presence of form I seeds in solution may dominate the crystallisation process.
In order to demonstrate reproducibility, repeated crystallisation at high pressure always resulted in form II.These results are of great significance as they demonstrate that form II can be prepared reproducibly at high pressure and recovered successfully to ambient pressure.Furthermore, unlike in the study by Lee et al., [24] no additives were required.This may suggest that the form II is the more thermodynamically stable form at high pressure, mirroring the behaviour of paracetamol [41].

Discussion
Mefenamic acid (MA) is an example of polymorphic system where the stable form (form I) can be obtained with relative ease, and its crystal structure is known from beginning (1976).
However, metastable elusive forms (form II and III) are very difficult to obtain.Our attention was drawn to this compound by the observation that form II is slightly denser than form I and has better packing of molecules within the unit cell.Hence our aim was to explore effect of pressure on the polymorphic transformation of MA and to explore weather form II could be prepared reproducibly either high pressure techniques.
There are two known methods to obtain the metastable form II; (i) heating of form I above the transition temperature (ii) cooling of a supersaturated solution in N,N-Dimethylformamide or chloroform.The thermal method resulted in the formation of a polycrystalline mass, and solution methods resulted in the conversion to form I on cooling.A recently reported method [18] involves crystallization of MA in the presence of structurally similar additive, flufenamic acid (FFA).They have described that MA molecules in form I have similar conformation as compared to FFA form I crystal structure.When MA crystallized in the presence of FFA, FFA blocked the nucleation process of MA form I and resulted in production of form II.This allowed crystal structure of form II to be determined for first time.This procedure is tedious, involves a number of controlled and lengthy processes (takes more than 20 days) and often results in failure.This is undesirable from both a scientific plus regulatory point of view, especially as crystal structures of as many polymorphic forms as possible, are required quickly and at an early stage of development.
Our initial attempts to obtain form II by recrystallisation from solution, using variety of solvent and at ambient pressure conditions were always resulted in the stable form I (see supplementary material).Different level of supersaturation was also applied but the results were same as stable form I was produces every time.Optical observation of recrystallisation process showed that on crystallisation the solution turned yellowish green first before white precipitates came out, which is indication of crystallization of form II.This may be explained by the fact that initially nucleated metastable polymorph (form II) dissolved rapidly into the surrounding phase and again nucleated as the stable form I. All these observations suggested that MA is an example of concomitant polymorphism where two or more crystal forms that nucleate and grow simultaneously under identical conditions but ends with the conversion of metastable form to most stable form.
Occasionally, in case of concomitant polymorphs metastable form is not detected during the conventional screening process as it remains in contact with solution and may convert to the more stable form via a solution-mediated transformation process by means of dissolution and recrystallisation [42,43].There was a need to explore alternative crystallisation conditions e.g.use of pressure as additional variable in conjunction with conventional strategies.
Compression behaviour of MA has been studied in connection with mechanical and dissolution properties [44], but very little attention has been paid to polymorphic transition under the conditions of high pressure.Previous studies have shown that the difference between polymorphs of MA comes from the torsional flexibility (figure 2) of molecules where rotation of one phenyl ring to other (torsional angle θ2), at the polymorphic transition from form I to form II. Direct compression of either single crystal or powder sample is one way to exploit this torsional flexibility.However our result on the single crystal compression studies shows that no phase transition has been observed by compressing single crystal of form I up to 1.62 GPa.Pressure was then raised to 3.8 GPa.At this pressure crystal disintegrated, suggesting a destructive phase transition.This is confirmed by the recrystallization from solution at high pressure, where form II was generated at 0.3 GPa pressure.
Further, a nice single crystal of form II was produce by this procedure, which was previously very difficult to achieve.
High pressure encourages the efficient packing of molecules to normally give a denser structure.
In the case of MA, conformation of MA molecule changed from form I to form II with the applied pressure in order to achieve better packing and the torsional angle θ2 has been changed from -119.9 o to -85.1 o respectively.This conformational change results in weakening of internal hydrogen bond as length of this contact increases from 1.88 to 2.01 Å.This is depicted by IR spectra of two forms (figure 5) where NH stretching frequency has changed from 3311cm -1 to 3500 cm -1 as a result of transformation from form I to II.

Conclusions
Single crystal of form II was obtained by high pressure crystallization technique at 0.3 GPa.The example of mefenamic acid has demonstrated that pressure can be used to induce polymorphism in

Figure 1 :
Figure 1: variation of unit cell volume of form I with pressure.

PreprintsFigure 3 :CFigure 4 :
Figure 3: (A) Crystal packing diagram of form I viewed down the c-axis (B) Dimer unit of MA molecule in form I (C) Crystal packing diagram of form II (D) dimer unit of MA molecule in form II (E) Crystal packing diagram of form III viewed down the c-axis (F) Dimer unit in form III.

Table 1 :
Lattice parameters for form I with increasing pressure up to 2.50 GPa.

Table 2 :
Crystallographic information of various polymorphs of MA compared with high pressure structure.

Table 3 :
Comparison of torsional angles between forms I and II of MA.