Current Applications of Powder X-Ray Diffraction in Drug Discovery and Development

• University of Cape Town

Introduction

Physicochemical characterization of solid materials generated in any program of drug discovery and development is an essential requirement aimed at ensuring reproducibility in the preparation of a given phase and on-going monitoring of its integrity during scale-up, processing, manufacture, formulation and storage [1, 2]. Among the numerous techniques available for solid-state characterization, powder X-ray diffraction (PXRD) plays a pivotal role due to its non-destructive nature and its ability to produce a unique pattern for any given crystalline phase. Such a pattern (Figure 1), represented as a plot of diffracted X-ray intensity (absolute or relative) versus the angular parameter 2θ, is viewed as a ‘fingerprint’ of the phase in question and a significant feature is that, unlike the three-dimensional, fully resolved X-ray diffraction pattern obtained from a single crystal of the same material, the PXRD pattern of a polycrystalline sample is a one-dimensional record of diffracted intensity as a function of diffraction angle.

Figure 1. An experimental PXRD pattern of an inclusion compound containing a steroidal drug

Detailed treatments of the theory and practice of the PXRD technique are readily accessible in the literature and it is not the intention here to repeat them. However, it may be useful to highlight a few salient points in anticipation of what is to follow. As for single crystal XRD, production of the powder pattern is governed by the Bragg Law, λ = 2d_hkl sinθ, in which λ is the X-ray wavelength, d_hkl the interplanar spacing of the set of crystal planes with indices hkl, and θ the Bragg angle. Recording a PXRD pattern that is truly representative of a specific solid phase requires a sample composed of minute crystallites (typical average particle size ~100 μm) in random orientation. Failure to meet the last condition will yield a pattern with erroneous peak intensities due to the ‘preferred orientation’ of microcrystals within the sample arising from their anisotropic shape, resulting in a statistically biased exposure of the various crystallographic planes to the incident X-ray beam. A second noteworthy point is that every atom in the unit cell of the crystalline phase makes a contribution to the intensity at every 2θ-value. This implies that in order to deduce that two samples are structurally identical, every peak in the PXRD pattern of one sample should have a counterpart in the PXRD pattern of the other at the same diffraction angle and with the same relative intensity.

A consequence of the drastic condensation of the diffraction record (from e.g. several thousand resolved reflections in single crystal XRD compared with several hundred intensity peaks in PXRD) is that physically distinct sets of planes (i.e. planes with different indices hkl) having similar interplanar distances d_hkl will produce peaks at similar angular positions, i.e. there will be extensive overlap of peaks in the PXRD pattern. This is evident from the Bragg Law (above) which shows that the angular position of a peak depends only on d_hkl when monochromatic X-rays of constant wavelength are employed. Thus, what appears as a single peak (e.g. that at 2θ ~17° in Figure 1) actually represents the resultant of numerous contributions from different sets of planes with similar d_hkl values. This is a serious drawback of PXRD in the context of detailed structural elucidation of materials available only as microcrystalline powders, as will be discussed later. However, since the overall PXRD pattern is certainly sufficiently definitive to serve as a characteristic ‘fingerprint’ of the material in question, the most common application of PXRD patterns is precisely their use in such identification. As alluded to above, identifying a specific crystalline phase by comparing its experimental PXRD pattern with that of a standard (or ‘reference’) pattern requires, in principle, a match between peak intensities at each 2θ-value. The permitted USP tolerance on diffraction angle for establishing identity is ± 0.2° in 2θ for the ten most intense reflections. Permitted tolerances on peak intensities have been discussed elsewhere [3].

Finally, the importance of the relationship between the three-dimensional, fully resolved X-ray diffraction pattern of a single crystal and the PXRD pattern of the same material cannot be understated. Thus, if the single crystal structure of a material has been determined by X-ray diffraction from a full set of resolved reflection intensity data, it is a trivial matter to use the parameters of the refined structural model to compute the simulated (‘idealized’) PXRD pattern, which would then serve as the best representative pattern for that phase.

During recent years, significant technological advances have led to the development of powder diffractometers with highly stable, intense X-ray sources and ultrasensitive detectors, facilitating rapid and accurate recording of PXRD patterns. This technique has thus become an indispensable analytical tool with widespread applications in both drug discovery and drug development. In addition, PXRD patterns of new drug candidates, polymorphs and other potential products are required for patent registrations [4]. Several important illustrations of the applications of the PXRD technique follow.

Quantitative Phase Analysis

Since each crystalline phase presents a unique PXRD pattern, a physical mixture of two or more crystalline phases is the sum of the individual patterns, their respective intensity profile contributions being weighted by the mass fraction of each phase present. This is the basis of traditional quantitative analysis of mixtures of crystalline phases, one of the best-known applications of the PXRD technique. In particular, the PXRD patterns of two or more polymorphic forms of a given drug will generally differ significantly and their mixtures can thus be quantified. This is of major importance in the control of polymorphic purity when an API occurs in different crystal forms having different physical properties (e.g. dissolution rates). An illustrative study relates to polymorphs of the artificial sweetener neotame, where PXRD was employed to quantify mixtures of crystalline forms A and G [5]. The PXRD technique was used in a supporting role to confirm the accuracy of composition assays of mixtures of two polymorphs of aprepitant (a substance P antagonist for chemotherapy-induced emesis) obtained using attenuated total reflectance Fourier transform infrared spectroscopy (ATRFT-IR) as the primary analytical tool [6]. If the PXRD patterns of the individual components in a mixture are available, or can be computed from single crystal data, the PXRD pattern of an arbitrary mixture can then be simulated by summing the component patterns in various ratios and refining these by least-squares to obtain the best match of the simulated PXRD pattern with that of an experimental multiphase sample. The use of the Rietveld refinement method (‘whole-pattern’ analysis) to quantify multiphase systems has led to significant improvements in the accuracy of quantitative analysis of mixtures as well as estimation of amorphous content [7].

Multi-component Compounds

The PXRD technique is used extensively in the identification of multi-component compounds and increasingly in the context of the characterization of new pharmaceutically relevant phases that might display advantageous properties, such as enhanced solubility of the active pharmaceutical ingredient (API). This is illustrated below with respect to two representative types of multi-component systems in which one component is an API. The first concerns an interaction compound (a co-crystal [8]) formed between two components, as exemplified by the API nevirapine and a GRAS (‘generally regarded as safe’) co-former, maleic acid [9, 10]. As shown in Figure 2 (top), the two pure components have distinctly different PXRD patterns, as expected. Following co-grinding of an equimolar mixture of the two in the presence of chloroform for 10 minutes, the PXRD pattern of the product (NVPMLE) was recorded; this pattern differs significantly from those of the components, indicating probable formation of a binary compound, a salt or a co-crystal being the most likely products.

Figure 2. The PXRD patterns of the API nevirapine, the co-former maleic acid and the product (NVPMLE) obtained when they are coground (top); experimental and computed PXRD patterns for the 1:1 co-crystal nevirapine-maleic acid (bottom). The structure of the co-crystal is shown on the right.

Subsequent isolation of monocrystals of the same product via coprecipitation and their structural elucidation by single crystal XRD revealed that this phase was indeed a co-crystal, no proton transfer between the two components having taken place [8]. Furthermore, in this co-crystal, the API and maleic acid molecules are linked by strong, complementary hydrogen bonds between the amide group of the API and the carboxylic acid group of the co-former. The computersimulated PXRD pattern based on the single crystal XRD structure compared favorably with that of the experimental pattern (Figure 2, bottom), proving that the same co-crystal phase is obtained by both the co-grinding and co-precipitation methods.

Cyclodextrin inclusion complexes of drugs [11] represent a second type of multi-component compound for which the PXRD approach is an invaluable characterization technique. These crystalline complexes comprise a cyclodextrin (a cyclic oligosaccharide host), one or more hydrophobic guest molecules and water molecules. In such complexes, the entire guest molecule or a guest residue is included within the hydrophobic cavity of the macrocyclic host molecule, being complexed by non-covalent interactions. As for drug polymorphs, solvates, cocrystals and other pharmaceutically relevant solid phases, the complete solid-state characterization of cyclodextrin complexes requires the application of several complementary techniques, thermal analysis, FTIR, Raman and/or solid-state NMR spectroscopy, and XRD methods being the most common. It is interesting to note, however, that for cyclodextrin complexes in particular, one can generally identify a newly formed complex and infer a considerable amount of structural detail using PXRD alone. This is based on the fact that there is a high probability that an inclusion complex formed between e.g. the host β-cyclodextrin and a new API will be isostructural with an already structurally documented β-cyclodextrin inclusion complex containing a different API as guest. ‘Isostructurality’ in this context means that the three-dimensional structural arrangement of the host molecules in both inclusion complex crystals is identical; the only difference therefore being the nature and precise location of the respective guest molecules (and the possible locations of water molecules). Since the cyclodextrin host molecule typically has a mass which is several times that of either of the included guest components, its crystalline structural arrangement is the dominant X-ray scatterer and the net result is that the PXRD patterns of the two inclusion complexes are very similar. A systematic examination of available cyclodextrin inclusion complexes yields a series of isostructural families, each of which has a distinct crystal packing arrangement and a corresponding distinct ‘reference’ PXRD pattern. Compilation of a library of such patterns enables a new inclusion complex to be identified rapidly and definitively from a single PXRD trace [12]. Figure 3 illustrates the unequivocal identification of a co-precipitation product of γ-cyclodextrin and cimetidine as a genuine inclusion complex [13].

Figure 3. Use of a reference PXRD pattern to establish complex formation between cimetidine and γ-cyclodextrin, the common host packing arrangement in this series of inclusion complexes is shown on the right

The experimental PXRD pattern closely resembles the reference PXRD pattern for a known series of γ-CD inclusion complexes crystallizing in the tetragonal space group P4212 with a = b ~ 23.8, c ~23.2 Å. The close similarity of the PXRD traces thus not only distinguishes the experimental sample as a genuine inclusion complex (as opposed to e.g. a physical mixture), but enables one to deduce the crystal system, space group, approximate unit cell parameters for the new complex and the complete structural arrangement of the host molecules in the complex crystal as well (see Figure 3). This approach to gleaning maximum structural information from a single PXRD pattern can obviously be applied to series of isostructural complexes containing other host compounds as drug carrier molecules (e.g. cucurbiturils, calixarenes).

Variable Temperature PXRD (VT-PXRD)

Heating solid APIs, or multi-component compounds containing APIs, generally induces reactions and phase transitions such as loss of solvent of crystallization (for solvated materials), conversion of one polymorphic form to another, transformation of amorphous phases to crystalline phases, fusion and thermal decomposition. While in many cases these processes can be suitably characterized using a combination of thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), variable-temperature PXRD is an invaluable supporting technique yielding different diffraction patterns as new phases appear during programmed heating or cooling of the material in question. Figure 4 shows a series of PXRD patterns recorded during the heating of an API. A distinct change in the pattern of the starting material is seen to occur between 65-70 °C; the new crystalline phase that appears at the higher temperature persists until fusion occurs at 188 °C. The use of in situ VT-PXRD for monitoring a wide variety of solid-state reactions and phase transformations during various stages of drug processing has become a popular technique. The combined use of this method and thermal analysis to investigate pharmaceutically relevant materials was reviewed earlier in this journal [14].

Figure 4. Variable temperature PXRD monitoring of a polymorphic phase change followed by fusion of the high-temperature form.

Structural Elucidation Directly from PXRD Data

Single crystal X-ray structure analysis currently retains its position as the optimum method for determining the detailed structures of crystalline materials. For pharmaceutically relevant materials that are available only in the form of powders, this method is not applicable. However, in such instances structural elucidation using PXRD data may be successful. This type of analysis is rapidly becoming a routine procedure, certainly in the case of small molecules, and the technique is currently enjoying an explosive growth, benefiting from significant advances in computational methods, progress in crystal structure prediction, instrumental improvements and the use of synchrotron X-rays. A considerable number of successful structure determinations of drugs have been carried out to date using PXRD data. Examples include a polymorph of hydrochorothiazide (Form II) [15], amodiaquinium dichloride dihydrate (from laboratory PXRD data) [16] and ampicillin trihydrate (from synchrotron PXRD data) [17]. Full details of the procedures involved in structure determination from PXRD data can be found in authoritative articles [18, 19]. Essentially, if the integrated intensities of individual reflections can be extracted from the severely overlapping peaks in the highly compressed PXRD pattern, then the wellestablished methods of single crystal X-ray analysis can be used to solve the crystal structure. However, this deconvolution process is non-trivial and its complexity increases with the complexity of the structure to be analyzed. More recent developments include the direct-space strategy [20], which involves the generation of a large number of trial structural models and their assessment by comparison of their calculated PXRD patterns with the experimental PXRD pattern of the compound in question. Comparison of calculated and observed integrated intensities can conveniently be replaced by point-by-point intensity comparison of the entire PXRD profiles. Following identification of the best structural model (as indicated by appropriate residual indices), refinement by the Rietveld method is performed to yield the final structural model. One serious current drawback of the method is that due to the low X-ray scattering power of hydrogen atoms, they cannot generally be imaged using PXRD data and are usually added to the structural model in idealized positions which may not be correct. Thus, ambiguities of H atom location on specific molecules may arise in situations where tautomerism may be involved or where the question of whether intramolecular or intermolecular proton transfer might have occurred is crucial (e.g. in molecular drugs, amino acids, salts and co-crystals). In the case of the long-acting angiotensin-converting enzyme inhibitor lisinopril (as the dihydrate), it was necessary to employ the single crystal XRD method to locate the hydrogen atoms unequivocally to establish the double zwitterionic nature of the API [21].

Successful ab initio determination of the structure of the 5-residue peptide acetyl-YEQGL-amide from PXRD data was recently reported [22], illustrating the potential of the method for its applicability in structural biology. The authors of this study indicate that the ultimate goal is the ab initio structural elucidation of small proteins using PXRD data. Success in this area would clearly have very significant positive implications for drug discovery and development.

Concluding Remarks

The powder X-ray diffraction technique is used ubiquitously in drug discovery and development and a comprehensive account would require a lengthy exposition. The examples in this article illustrating the application of the PXRD method in the pharmaceutical context were intended to convey the versatility of the technique in the types of challenges addressed, ranging from technically fairly modest but essential quantification of multiphase powders, to the daunting challenge of complete structural elucidation of complex molecular and crystal structures of biologically relevant compounds.

Acknowledgements

The author is grateful to the University of Cape Town and the National Research Foundation (Pretoria) for financial support.

Author Biography

Dr. Mino Caira is Professor of Physical Chemistry at the University of Cape Town and Director of the Science Faculty’s Centre for Supramolecular Chemistry Research. His expertise is in the area of solid-state chemistry of drug polymorphs and novel multi-component systems containing APIs, especially solvates, inclusion complexes and cocrystals.

References

1. R. Hilfiker, “Polymorphism in the Pharmaceutical Industry“, Wiley-VCH 2006, Weinheim.
2. S. R. Byrn, R. R. Pfeiffer and J. G. Stowell, “Solid-state Chemistry of Drugs”, SSCI, Inc., 1999, West Lafayette, Indiana, USA.
3. H. G. Brittain, “X-ray Diffraction III: Pharmaceutical Applications of X-ray Powder Diffraction”, Spectroscopy, 2001, 16, 14.
4. https://www.americanlaboratory.com/914-Application-Notes/36153-Increasing-Application-of-X-Ray-Powder-Diffraction-in-the-Pharmaceutical-Industry/: accessed 08 September 2013.
5. Z. Dong, E. J. Munson, S, A, Schroeder, I. Prakash, D. J. W. Grant. “Neotame anhydrate polymorphs. II: Quantitation and relative physical stability”, Pharm. Res., 2002, 19(9), 1259.
6. R. Helmy, G. X. Zhou, Y. W. Chen, L. Crocker, T. Wang, R. M. Wenslow, Jr., A. Vailaya, “Characterization and Quantitation of Aprepitant Drug Substance Polymorphs by Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy”, Anal. Chem. 2003, 75, 605.
7. D. L. Bish, S. A. Howard, “Quantitative phase analysis using the Rietveld method”, J. Appl. Cryst., 1998, 21, 86.
8. N. Schultheiss and A. Newman, “Pharmaceutical Cocrystals and Their Physicochemical Properties”, Cryst. Growth Des., 2009, 9, 2950.
9. M. R. Caira, S. A. Bourne, H. Samsodien, E. Engel, W. Liebenberg, N. Stieger and M. Aucamp. “Co-crystals of the antiretroviral nevirapine: crystal structures, thermal analysis and dissolution behaviour”. CrystEngComm, 2012, 14(7), 2541.
10. E. Engel. “Supramolecular modification of selected antiretroviral drugs”, MSc dissertation, University of Cape Town, 2011.
11. “Cyclodextrins and Their Complexes”, H. Dodziuk ed., Wiley-VCH 2006, Weinheim.
12. M. R. Caira. “On the isostructurality of cyclodextrin inclusion complexes and its practical utility”. Rev. Roum. Chim., 2001, 46(4), 371.
13. D. R. Dodds, “Physicochemical Study of Inclusion of Drug Molecules in Cyclodextrins”, PhD thesis, University of Cape Town, 1999.
14. S. X. Yin, R. P. Scaringe, M. F. Malley, J. Z. Gougoutas, “In-situ variable temperature powder X-ray diffraction and thermal analysis -applications in the pharmaceutical industry”. American Pharmaceutical Review, 2005, 8, 56-58, 60, 62, 67.
15. A. Florence, A. Johnston, P. Fernandes, K. Shankland, H.N.E. Stevens,S. Osmundsen, A. B. Mullen. “Powder study of hydrochlorothiazide form II”. Acta Crystallogr., 2005, E61:o2798.
16. A. Llinàs, L. Fábián, J. C. Burley, J. van de Streek, J. M. Goodman. 2006. “Amodiaquinium dichloride dihydrate from laboratory powder diffraction data”, Acta Crystallogr., E62:o4196.
17. J. C. Burley, J. van de Streek, P. W. Stephens. P. W. “Ampicillin trihydrate from synchrotron powder diffraction data”. Acta Crystallogr., 2006, E62:o797.
18. W. I. F. David and K. Shankland, “Structure determination from powder diffraction data”. Acta Crystallogr., Section A: Foundations of Crystallography, 2008, A64, 52.
19. K. D. M. Harris, “ Structure solution from powder X-ray diffraction data by genetic algorithm techniques, applied to organic materials generated as polycrystalline products from solid state processes”. Mater. Manuf. Process, 2009, 24, 293.
20. K. D. M. Harris, E. Y. Cheung, “How to determine structures when single crystals cannot be grown: opportunities for structure determination of molecular materials using powder diffraction data”, Chem. Soc. Rev., 2004, 33, 526.
21. M. Sorrenti, L. Catenacci, D. L. Cruickshank, M. R. Caira, “Lisinopril dihydrate: single-crystal X-ray structure and physicochemical characterization of derived solid forms”. J. Pharm. Sci., 2013, 102, 3596.
22. K. Fujii, M. T. Young, K. D.M. Harris, “Exploiting powder X-ray diffraction for direct structure determination in structural biology: The P2X4 receptor trafficking motif YEQGL”, J. Struct. Biol. 2011, 174, 461–467.