Single Crystal and Powder Diffraction Characterization of Three Polymorphic Forms of Acitretin

Single Crystal and Powder Diffraction Characterization of Three Polymorphic Forms of Acitretin

Single Crystal and Powder Diffraction Characterization of Three Polymorphic Forms of Acitretin LUCIANA MALPEZZI,1 GRATO ANGELO MAGNONE,2 NORBERTO MASC...

578KB Sizes 0 Downloads 37 Views

Single Crystal and Powder Diffraction Characterization of Three Polymorphic Forms of Acitretin LUCIANA MALPEZZI,1 GRATO ANGELO MAGNONE,2 NORBERTO MASCIOCCHI,3 ANGELO SIRONI4 1

Dipartimento di Chimica, Materiali ed Ingegneria Chimica ‘‘G.Natta’’, Politecnico di Milano, via Mancinelli, 20131, Milano, Italy 2

SOLMAG S.p.A., Garbagnate Milanese, Italy

3

Dipartimento di Scienze Chimiche e Ambientali, Universita` dell’Insubria, via Valleggio 11, 22100 Como, Italy

4

Dipartimento di Chimica Strutturale e Stereochimica Inorganica, Universita` di Milano, via Venezian 21, 20133 Milano, Italy

Received 22 October 2004; revised 6 January 2005; accepted 10 January 2005 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20330

ABSTRACT: Acitretin [all-trans-9-(4-methoxy-2,3,6-trimethylphenyl)-3,7-dimethyl2,4,6,8-nonatetraenoic acid or 3-methoxy-2-methyl-17-nor-1,2,3,4-tetradehydroretinoic acid], a widely marketed oral synthetic retinoid, introduced for clinical use as effective therapy against psoriasis, was found to crystallize in three polymorphic modifications (hereafter, I, II, and III), the crystal structures of which have been determined by singlecrystal diffractometry (form I) or X-ray powder diffraction methods (form II and III) from conventional laboratory data only. In these latter cases, real space techniques (simulated annealing and whole-profile pattern matching) have been employed. Polymorph I crystallizes in space group P21, Z ¼ 8, with unit cell parameters a ¼ 7.894(1), b ¼ 58.454(6), ˚ , b ¼ 102.04(1)8, and V ¼ 3682.9(8) A ˚ 3. Polymorph II crystallizes in space c ¼ 8.161(1) A ˚, group P21/n, Z ¼ 4, with unit cell parameters a ¼ 13.999(2), b ¼ 10.714(1), c ¼ 12.465(2) A 3 ˚ b ¼ 98.76(5)8, and V ¼ 1847.9(3) A . Polymorph III crystallizes in space group P21/c, Z ¼ 4, ˚ , b ¼ 100.41(7)8, and with unit cell parameters a ¼ 3.0751(4), b ¼ 4.0487(4), c ¼ 14.956(2) A ˚ 3. Polymorph I, found to be identical with that deposited in the European V ¼ 1831.3(4) A Pharmacopeia, shows four crystallographically independent Acitretin molecules, arranged in pairs through conventional hydrogen-bonded carboxylic dimers; also in form II, carboxylic dimers are observed, located on crystallographic inversion centres, while in form III, a catameric arrangement of the carboxylic residues, winding up about the rather short monoclinic axis, generates one-dimensional chains of hydrogen-bonded Acitretin molecules. Thermal analysis showed that form I can be quantitatively transformed into form II by moderate heating near 2008C, under vacuum. These results show that ab initio structural studies from conventional laboratory X-ray powder diffraction (XRPD) data are fully providing the opportunity to investigate the structural aspects of moderately complex substances also in the absence of single crystals, disclosing the crystal chemistry of a few polymorphs of pharmaceutically relevant species. ß 2005 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 94:1067–1078, 2005

Keywords: (DSC)

polymorphism; X-ray powder diffraction; crystal structure; calorimetry

INTRODUCTION Correspondence to: Norberto Masciocchi (Telephone: þ39031-326227; Fax: þ39-031-326230; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 94, 1067–1078 (2005) ß 2005 Wiley-Liss, Inc. and the American Pharmacists Association

Acitretin, an aromatic analogue of retinoic acid, is used in the systematic treatment and prevention of a range of skin diseases such as psoriases and other diskeratotic dermatoses.1

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1067

1068

MALPEZZI ET AL.

This species, that is, all-trans-9-(4-methoxy2,3,6-trimethylphenyl)-3,7-dimethyl-2,4,6,8-nonatetraenoic acid, has been established to have clinical advantage with respect to the previously currently used Etretinate, another member of the class of oral synthetic retinoids.2 Despite of the extensive and successful clinical usage of this class of compounds in the treatment of skin diseases, the actual mechanism of their pharmacological action has not yet been clearly identified. While different methods of preparation of Acitretin are reported in the literature since 1976,3 it was not until very recently4 that Acitretin was found to exist in three different crystalline, unsolvated (thus strictly polymorphic) phases. In that study, some different methods of crystallization are described: the three different crystalline forms are then selectively prepared using tailor-made crystallization conditions (and solvents) used in the precipitation process. In the solid state, Acitretin typically appears as a yellow polycrystalline powder. The existence of different crystalline modifications of Acitretin had been initially detected by conventional thermal analysis, and subsequently verified by X-ray powder diffraction (XRPD) measurements. At that stage, a detailed molecular structure analysis was not performed, owing to the difficulty to obtain suitable crystals for conventional X-ray diffraction analysis. Polymorphism5 arises from the possibility of a substance to crystallize in more than one spatial arrangement of the molecules within the crystal lattice. Each different crystalline form, owing to its different molecular packing, exhibits different intermolecular interactions (thermodynamic properties) between nearest neighbors, next nearest neighbors, and so on throughout the whole crystal. The relative stability of the various crystalline modifications depends on the free energy associated with the crystal packing, the most stable having the lowest free energy. To understand the formation of the different polymorphs and the relationships among them, it is therefore crucial to have an accurate knowledge of the structural aspects of the polymorphic entities. Polymorphism, mostly of the conformational type, has been found to affect a large number of pharmaceutical ingredients and has a large impact on the pharmaceutical product development (see, e.g.,6). Since different polymorphic forms of the same drug substance, owing to the differences in the energetic content of the different molecular JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

arrangements within the crystal, may show significant differences in the physicochemical and biopharmaceutical properties, the characterization and the possibility of a rapid identification of each polymorph are essential requirements during the formulation, manufacturing, and processing of a new drug substance. Accordingly, the investigations on polymorphism also appear in many patents and articles on the major pharmaceutical and crystallographic journals. In the following, we complement our very preliminary study on the crystallization products of Acitretin from a variety of solvents, and fully characterize, from a structural point of view, the crystal and molecular structures of three polymorphs. While, after many efforts, within a sample of Acitretin Form I, we (occasionally) discovered a (very) small single crystal, suitable for conventional structural analysis on a CCD-equipped 3circles diffractometer, the structures of two of the three polymorphs (forms II and III), not affording single-crystals of suitable quality or size, were solved directly from conventional powder diffraction data. Indeed, in the recent years, a considerable progress has been made in the techniques for solving the structure of a complex organic crystalline substance7 by using only the XRPD, as beautifully shown by, among others, Dinnebier, Shankland, and David, using synchrotron radiation;8 however, as we demonstrate with our study, interesting and satisfactory results can be gathered even from conventional laboratory data. Similarly, a number of organic crystal structures with limited flexibility have been solved in the last few years by laboratory XRPD, and were reviewed recently.9

EXPERIMENTAL SECTION Materials Several batches of Acitretin, of different origins and crystallized from different solvents and in different thermal and pressure conditions, were obtained as described in reference 4. The samples were examined by X-ray powder diffraction to discriminate among the different polymorphs. The many batches could eventually be classified in the following three classes (or ‘‘samples’’): (a) the first sample, Acitretin reference standard, was obtained from the European Standard Pharmacopoeia and is hereafter named Acitretin form I. Most of the crystallizations aimed to precipitate

THREE POLYMORPHIC FORMS OF ACITRETIN

single crystals, suitable for conventional diffraction analysis, failed. Eventually, some very thin crystals were obtained from tetrahydrofurane solution, maintained for about 2 h at refluxing temperature (ca. 678C), and later allowed to slowly cool to room temperature; (b) a second group of samples were produced by SOLMAG (Garbagnate Milanese, Milano, Italy) and crystallized from one of the following solvents: toluene, methylisobutylketone, ethyl acetate, isopropanol. Hereafter they are named Acitretin form II. All attempts to prepare single crystals of suitable quality failed; (c) a third group of samples were also produced by SOLMAG and were crystallized from one of the following solvents: hexane, methanol, or a mixture of dichloromethane/hexane. Hereafter they are named Acitretin form III. Also in this case, all attempts to prepare single crystals of suitable quality failed. Thermal Analysis DSC measurements (on a Perkin-Elmer DSC-7 instrument) were carried out in order to investigate and compare the thermal behavior of the different modifications. The samples (each of about 2 mg), enclosed in aluminium pans having small holes and accurately weighed, were heated over the temperature range of 50–2408C, at a scan rate of 108C/min, under a nitrogen purge. A few thermograms of Acitretin form I were recorded also at different scan rates: 5, 20, and 308C/min (and commented below). The apparatus was calibrated with (99.9% pure) indium.

1069

Indexing, of the low angle diffraction peaks suggested, for all three species, primitive monoclinic cells of approximate dimensions, form I: ˚ , b ¼ 102.18 [SVDa ¼ 7.86, b ¼ 58.22, c ¼ 8.11 A TOPAS, GOF(21) ¼ 16]; form II: a ¼ 14.00, b ¼ ˚ , b ¼ 98.78 [TREOR,10 M(20)11 ¼ 10.71, c ¼ 12.47 A 12 29; F(20) ¼ 71 (0.007, 43)]; form III; a ¼ 31.64, ˚ , b ¼ 107.28 [SVD-TOPAS,13 b ¼ 4.04, c ¼ 14.90 A ˚ GOF(29) ¼ 25], later converted to a ¼ 30.71 A and b ¼ 100.48 (b and c remaining unchanged) by conventional cell reduction programs. Systematic absences indicated P21/n and P21/c as the probable space group for forms II and III, later confirmed by successful solution and refinement, while, owing to the complexity of the unit cell of form I, its space group assignment remained dubious (apart from the obvious presence of the 21 axis). Structure solutions of forms II and III were performed (with TOPAS13) using a molecular model of the entire Acitretin molecule, obtained from analogous systems retrieved from the CSD database, with torsional freedom about the six C-O and C-C bonds highlighted in Chart I, and free location and orientation. Within a few hours of computational time on a 1500 MHz PC running MS Windows XP, the approximate models were found.14

Chart I. X-Ray Powder Diffraction Analysis The gently ground yellow powders were cautiously deposited in the hollow of an aluminium holder equipped with a quartz monocrystal zero background plate (supplied by The Gem Dugout, State College, PA). Diffraction data (Cu Ka, ˚ ) were collected on a y:y vertical scan l ¼ 1.5418 A Bruker AXS D8 diffractometer, equipped with parallel (Soller) slits, a secondary beam curved graphite monochromator, a Na(Tl)I scintillation detector, and pulse height amplifier discrimination. The generator was operated at 40 kV and 40 mA. Slits used: divergence 0.58, antiscatter 0.58, and receiving 0.2 mm. Nominal resolution for the present set-up is 0.088 2y (FWHM–Ka1) for the LaB6 peak at ca. 21.38 (2y). Long scans were performed with 2 < 2y < 728, t ¼ 75 s/step and D2y ¼ 0.028.

For form I, parallel structure solution processes were started in the different possible space groups, with Z0 ¼ 2 (Z0 ¼ number of crystallographically independent molecules), or even Z0 ¼ 4 in P21, with the same strategy but with much more increased complexity. No satisfactory solutions could be found, even after about 2 weeks of uninterrupted CPU usage. In retrospect, we attribute this failure to the presence of uneliminable preferred orientation effects, difficulties of spanning the whole parameter space in (what was later found to be) the correct space group (P21), as well as in the pseudosymmetric arrangement of the four molecules in this structure eventually found by conventional single-crystal analysis (vide infra).15 The bulk from which the crystal was chosen only contains Acitretin form I, as confirmed by a Rietveld refinement, performed by imposing the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1070

MALPEZZI ET AL.

Figure 1. Final Rietveld refinement plot (2 < 2y < 728) for Acitretin form I with coordinates taken from the single crystal structure. Vertical scale, counts. The insert shows the high angle region magnified 5.

structural model derived from the single-crystal analysis [Figure 1; Rwp ¼ 0.224 (0.148); Rwp ¼ 0.166 (0.114); RBragg ¼ 0.155 (0.076); the values in parentheses refer to a narrower dataset (10<2y<608) with three low-angle peaks omitted, as evidenced by the shaded area (i.e., those most affected by instrumental aberrations). Refined preferred orientation coefficient for the [010] pole g16 ¼ 0.473(2).] The final refinements (forms II and III) were performed by the Rietveld method with the aid of TOPAS.13 In the final cycles, the peak shapes were carefully described by the fundamental parameters approach.17 The background level was modeled by a polynomial function, while systematic errors were corrected with the aid of a sampledisplacement angular shift, finite thickness transparency effects, anisotropic line broadening, and a preferred orientation correction along the [10-1] and [100] poles, respectively. A (single) isotropic atomic displacement parameter was refined. Scattering factors were taken from the internal library of TOPAS. Final Rp, Rwp, and RBragg agreement factors, together with details of the data collections and analyses for all crystal phases, can be found in Table 1. Figure 2 shows the final Rietveld refinement plots for forms II and III. A list of experimental d/I values for the three forms of Acitretin has been deposited with the JCPDS Powder Diffraction File (ICDD, Newton Square, PA). Single Crystal X-Ray Analysis A very small single crystal of Acitretin form I, of about 0.20  0.15  0.03 mm, was mounted on a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

glass fiber and transferred to a BRUKER AXS APEX diffractometer. The resulting cell parameters matched those found from our powder diffraction analysis. Data collection was performed by the o-scan technique, collecting 2400 frames with Do ¼ 0.38; no absorption correction was applied. The structure was solved, in P21, by direct methods (SIR9718), and refined anisotropically on F2 (program SHELX-9719). Hydrogen atoms were included using a riding method. Crystal data and other refinement details can be found in Table 1. Note that agreement factors are less than ideal, but clearly reflect the size and quality of the sample in our hands.

RESULTS AND DISCUSSION Thermal Behavior The DSC thermograms of the three polymorphic forms of Acitretin (measured at the scan rate of 108/min) are shown in Figure 3. Both forms II and III show a melting temperature at ca. 225.58C and a melting enthalpy of about 140–150 J/g. On the other hand, form I possesses a more complex thermal behavior, as evidenced directly by the reported thermograms. Indeed, before melting, another thermal event is observed in the 190– 2108C range (see Figure 3). In particular, the first endothermal event in the DSC thermograms of form I (acquired with temperature gradients of 208 and 308C/min) clearly indicated the presence of a solid-solid transition below 2008C coherent with an enantiotropic system; however, at lower

THREE POLYMORPHIC FORMS OF ACITRETIN

1071

Table 1. Crystal Data and Details of Structure Refinements (From Powders or Single-Crystals)* Acitretin Form I

Acitretin Form II

Acitretin Form III

Powder Diffraction

Powder Diffraction

Species Method

Single-Crystal Diffraction

Formula Fw T(K) ˚) l (A Cryst syst Space group Cell constants ˚ a, A ˚ b, A ˚ c, A

C21H26O3 326.43 298(2) 0.71073 Monoclinic P21

C21H26O3 326.43 298(2) 1.5418 Monoclinic P21/n

C21H26O3 326.43 298(2) 1.5418 Monoclinic P21/c

7.894(1) 58.454(6) 8.161(1) 102.04(1) 3682.9(1) 8 1.177 0.08 1408 0.20  0.15  0.03 3.00–20.00 7  h  7 56  k  56 7  l  7 20112/6886 [R(int) ¼ 0.064] None Full-matrix Least-squares on F2 6886/1/445 1.061 0.084, 0.254 n.a. ˚3 0.38, 0.34 e/A

13.999(2) 10.714(1) 12.465(2) 98.76(5) 1847.9(3) 4 1.173 0.61 704 20  8  0.3 (powders) 4.00–36.00

30.751(4) 4.0487(4) 14.956(2) 100.41(7) 1831.3(4) 4 1.184 0.62 704 20  8  0.3 (powders) 2.50–36.00

n.a.

n.a.

876 Finite thickness layer Full-matrix Least-squares on yi 3201/0/44 n.a. n.a. 0.128, 0.097, 0.055 n.a.

852 Finite thickness layer Full-matrix Least-squares on yi 3351/0/39 n.a. n.a. 0.141, 0.099, 0.047 n.a.

b, deg ˚ 3) V(A Z r(calc) (Mg/m3) m (mm1) F(000) Sample size (mm3) y range (deg) hkl collected No. of reflns (coll./indep.) Abs. Corr Refinement method No. of data/restraints/params S(F2) R1, wR2a Rwp, Rp, RBraggb ˚ 3) Max/min Dr (e/A

*Crystallographic data (excluding structure factors) for the three polymorphic structures of Acitretin, reported in this study, have been deposited with the Cambridge Crystallographic Data Centre supplementary publication no. CCDC 253439–253441. Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (Fax: (þ44)1223-336-033; E-mail: [email protected]). a R1 ¼ kFo jjFc k=jFo for reflections with I2ðIÞ; wR2 ¼ ½½wðFo2  Fc2 Þ2 =½wðFo2 Þ2 1=2  for all reflections; w1 ¼ 2 ðF 2 Þ þ 2 2 P¼ ðaPÞ2 þ bP, where ð2F cþ Fo Þ=3 and a and b are constants set by the program.

b

Rp ¼ i yi;o  yi;c =i yi;o ; Rwp ¼ ½i wi ðyi;o  yi;c Þ2 =i wi ðyi;o Þ2 1=2 ; RBragg ¼ n In;o jjIn;c =i In;o , where yi,o and yi,c are the observed and calculated profile intensities, respectively, while In;o and In;c the observed and calculated structure factors. The summations run over i data points or n independent reflections. Statistical weights wi are normally taken as 1/yi,o.

scanning rates (down to 58C/min) a more complex trace can be appreciated, which we attribute to premelting effects (enhanced chain mobility) followed by recrystallization into form II (as suggested by the presence of a weak exothermic peak). Ex situ XRPD analysis on a sample of form I heated above 2008C, performed after cooling down to room temperature the heated powders, showed the presence of a polycrystalline species containing uniquely form II crystals (even if the m.p. of this batch is, apparently, a few degrees lower than that observed for pristine powders

of form II). Thus, these findings firmly suggest another mode of preparation of Acitretin form II, that is, by the quantitative conversion of form I upon controlled thermal treatment at about 2008C.

Molecular Structure of Acitretin Acitretin molecules contain a polymethylated benzene ring, bearing, para to each other, a methoxo group, and an all-trans tetraalkenyl chain (with some methyl substituents), terminated by a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1072

MALPEZZI ET AL.

Figure 2. Final Rietveld refinement plots for Acitretin form II (top) and form III (bottom), with difference plot and peak markers.

carboxylic group. Thus, each molecule consists of a weakly polar moiety with an extremely hydrophilic head (see Chart I, where a partial labeling scheme is also shown). Since no solvent molecules are present, the three different crystal phases are true polymorphs. Crystals of form I contain four crystallographically independent molecules, hereafter labeled as molecules A–D; the geometrical parameters relevant for the following discussion are synoptically collected in Table 2, together with those of the molecules of forms II and III. A schematic drawing of the molecules C and D is shown in Figure 4. For the sake of simplicity, molecules A and B, which show intramolecular features similar (but not equal!) to molecules D and C, respectively, are not drawn here. As expected, in all four molecules, the methoxy residue is bent away from the neighboring methyl group, because of obvious steric hindrance (torsional COMe angles in the 11, þ88 range). Steric effects are also at work in driving the phenylpolyalkenyl dihedral angle (CCC) away from coplaJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

narity (which would optimize the amount of p conjugation), by as much as 38.48 (molecule B). A further peculiar feature of all these molecules (within form I) is the syn disposition of the MeO- and of the branching (vinylic) methyl substituents, which, apparently, suffer of a small, but significant, stereochemical (or even electronic) differentiation of the two sides of the aromatic ring, that is, those bearing one, or two, methyl substituents, respectively. Crystals of Acitretin forms II and III belong to the same monoclinic space group (although proposed in different settings), but possess markedly different cell parameters (see Table 1); both phases contain a unique crystallographically independent molecule, hereafter labeled molecules E and F, respectively: their molecular structures are shown in Figure 5. Consistently with the description presented above, also in form III, the syn-disposition of the two major ring-subtituents is observed, even if with a slightly different torsional angle (ca. 608, see Table 2). On the contrary, molecule E adopts the anti-disposition, as witnessed by the

THREE POLYMORPHIC FORMS OF ACITRETIN

1073

Figure 3. (Top) DSC thermograms (108C/min) of Acitretin forms II and III; (bottom) DSC thermograms of Acitretin form I measured at two different heating rates: left, at 108C/min; right, at 208C/min.

atoms syn to the methyl substituents, up to 358), while in form I, smaller dihedral angles are observed (see Table 2).

much higher CCC (ca. 1658) value. This structural feature is also manifested by the orange-yellow color of form II, while forms I and III are light yellow. Another interesting structural parameter, the relevance of which will be later evident when discussing the supramolecular features of these species, is the CR-COOH torsional angle, which defines the twist of the carboxylic acid moiety with respect to the polyalkenyl chain. In forms II and III, they reach the largest values (for the oxygen

Crystal Structures of Acitretin Forms I, II, and III As shown in Figure 6, in form I, two slightly different carboxylic dimers are present, with O_€O ˚ contacts, involved in hydrogen bonds, 2.61–2.66 A apart. The fully stretched molecules adopt, in the AB dimer, an only nearly centrosymmetric

Table 2. Synoptic Collection of the Relevant Conformational Parameters for the Six Crystallographically Independent Acitretin Molecules Found in the Three Crystal Modifications Polymorphic form Molecule C5-C4-C7-C8, 8 C2-C1-O1-C20, 8 C13-C14-C15-Osyn, 8 C13-C14-C15-Oanti, 8 ˚ OH_€O, A OH_€O topology

CCC COme CR-COOH COMe

Form Ia

Form I

Form I

Form I

Form IIa

Form IIIa

A

B

C

D

E

F

29.8 7.9 14.5 167.2 2.67 Dimer

38.4 4.8 10.5 168.9 2.67 Dimer

36.5 8.4 2.4 174.4 2.65 Dimer

32.2 2.2 8.5 176.8 2.65 Dimer

165.3 10.9 23.0 158.8 2.68 Dimer

60.8 5.8 35.4 146.4 2.79 Catamer

a Forms II and III are centrosymmetric; thus molecules with opposite torsional angles are also present in the unit cell; at variance, for the acentric Form I (which occurs as enantiomorphic crystals), ‘inverted’ molecules are present in different crystals.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1074

MALPEZZI ET AL.

Figure 4. Structure of molecules C (top) and D (bottom) present in Acitretrin form I.

arrangement (see torsional angles in Table 2), thus structurally differentiating the two sides of the dimer. A differentiating effect of similar magnitude is also present in the CD dimer. In addition, the relative orientation of the nearly coplanar (CO2)2 six-membered rings (generated by the mutual coupling of two hydrogenbonded RCOOH moieties) in the crystal is not related by any symmetry element. Being the two dimers crystallographically unrelated, and due to the slightly different conformations within each dimer, we can easily explain why, although similar

Figure 5. Structure of the Acitretin molecules present in forms II (top) and III (bottom). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

and packed in space in a pseudosymmetrical mode (see Figure 7), all four molecules are crystallographically independent, that is, a Z0 ¼ 4 crystal is observed. As depicted in Figure 7, the very elongated P21 unit cell hosts the dimers arranged in parallel slabs, normal to b, separated by hydrophobic ˚ cell parameter is, therefore, interactions. The 58 A due to the presence of (symmetrically related) ca. ˚ thick (i.e., the molecular length of each dimer) 30 A slabs in the cell. ˚ ) are Hydrogen-bound dimers (O_€O ¼ 2.68 A present in form II, which, at variance from those discussed above, are truly centrosymmetric. The different conformation of the dimer can easily be appreciated by observing Figure 6, where the central core is equioriented with dimers AB and CD, and the different twist about the CCC torsional angle (cited above) is observed. The crystal packing of this crystalline form (see Figure 8) does not clearly show the presence of specific supramolecular arrangements (chains, layers, or slabs). The nearly isotropic unit cell, which contains such elongated dimers, is indeed a manifestation of poorly segragated molecules, packing in space, through normal van der Waals contacts, apparently disregarding their very anisotropic shape. Noteworthy, the phase transformation from form I to form II achieved by moderate heating equalizes the four crystallographically independent molecules of form I, through a reorientation process involving the syn-anti transformation of the C4-C7 torsional angle (possibly by coherent rotation of the polyalkenyl chains within each dimer). Note also that the centric monoclinic form II (with molecules in anti conformation), and not the ‘elusive’ centric monoclinic B21/d crystal structure (a non-standard setting of P21/c, obtained by averaging the orientations of the synAcitretin molecules of form I in P21, see Ref. 15) is formed just below 2008C, thanks to the aforementioned relevant conformational change. Crystals of phase III, which possess a rather ˚ , contain short unit cell axis (b), of about 4.04 A stacked molecules arranged as infinite [_€OCOH_€ OCOH_€OCOH]n chains, with hydrogen-bonded ˚ . This spatial arrangeO_€O interactions of 2.79 A ment (shown in Figure 6), obtained by winding up the catameric chains about the short twofold axis, therefore, does not contain dimeric units, and is favored by the slighly different molecular conformation (of the syn type) present in the molecules of Acitretin form III. Each polymeric chain is polar (see Figure 6), but, giving the presence of inversion

THREE POLYMORPHIC FORMS OF ACITRETIN

1075

Figure 6. Top to bottom: (a) hydrogen bonded dimer (molecules A and B; form I); (b) hydrogen bonded dimer (molecules C and D; form I); (c) centrosymmetric hydrogen bonded dimer (form II); (d) hydrogen bonded catamer, winding up the 21 axis (form III). The correct locations of the acidic H atoms could not be determined experimentally, and, when drawn, are a guide to the eye for evidencing the supramolecular features discussed in the text. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1076

MALPEZZI ET AL.

Figure 7. Crystal packing of Acitretin molecules in form I, viewed down [001]. ˚ thick slabs described in the Horizontal axis, b. The highlighted portion refers to the 30 A text, where the presence of centrosymmetric hydrogen-bonded dimers can be appreciated.

centers in space group P21/c, is surrounded by chains of opposite polarity in a crystal packing that, to some extent, shows segregated slabs aligned normally to a (which, accordingly, is about ˚ long, see Figure 8). 30 A

Form III is also the densest of the three phases, and, as such, can be interpreted as the most thermodynamically stable (at low temperature). Thermal treatment, up to the melting point, did not show any polymorphic transition, while

Figure 8. Crystal packing of Acitretin molecules in form II (a) and form III (b), viewed down [010]. In both cases, horizontal axis, a. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

THREE POLYMORPHIC FORMS OF ACITRETIN

1077

Figure 8. (Continued )

form I was found to convert easily (at about 2008C) into form II. Whether kinetic effects are at work in this transformation (on in the lack of it), we do not know. Certainly, the similar dimeric nature of the constituents in forms I and II, despite their different conformation, may allow the simple reorientation of (partially flexible) dimers, without any bond breaking in their hydrogen-bonded core. Accessibility to the polymeric arrangement found in form III, thus, may be hindered by a much higher activation energy.

CONCLUSIONS It would be unfortunate if the stereochemistry of important molecular functional materials (organics, inorganics, and even low-dimensional organometallic polymers20) would remain unknown because of the lack of suitable well-diffracting specimens. Aiming to overcome these difficulties, the results presented above clearly show that XRPD, coupled to a conventional powder diffractometer, is indeed fully providing the opportunity to investigate the structural aspects of pharmaceutically active substances also in the absence of single crystals, if the complexity of the model sought is not too high. Accordingly, today XRPD methods can be used not only for quali- and quantitative phase analysis or material characterization (probing strain, stress or texture), but (as

recently reviewed by us21–22) also for crystal structure determination.

ACKNOWLEDGMENTS The Fondazione Provinciale Comasca is acknowledged for funding.

REFERENCES 1. (a) Gollnick HPM, Du¨mmler U. 1997. Retinoids. Clin Dermatol 15:799–810. (b) Mayer H, Bollag W, Haenni R, Ruegg R. 1978. Retinoids, a new class of compounds with prophylactic and therapeutic activities in oncology and dermatology. Experientia 34:1105–1119; (c) Ward A, Brodgen RN, Hell RC, Speight TM, Avery GS. 1983. Etretinate. A review of its pharmacological properties and therapeutic efficacy in psoriasis and other skin disorders. Drugs 26:9; (d) Saurat JH. 1999. Retinoids and psoriasis: Novel issues in retinoid pharmacology and implications for psoriasis treatment. J Am Acad Dermatol 41:S2–S6 and references therein. 2. Viegard UW, Chou RC. 1998. Pharmacokinetics of acitretin and etretinate. J Am Acad Dermatol 39: S25–S33. 3. Andriamialisoa Z, Valla A, Cartier A, Labia R. 2002. A New Stereoselective Synthesis of Acitretin (Soriatane1, Neotigason1). Helv Chim Acta 85: 2926–2929; (b) Patent WO03/007871 and references herein; (c) US 4.105.681. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

1078

MALPEZZI ET AL.

4. Malpezzi L, Brenna E, Fuganti C, Masciocchi N, Pellegatta C, Ventimiglia G. 2004. Polymorphism in acitretin. Pharm Ind 66:1551–1557. 5. Bernstein J. 2002. Polymorphism in molecular crystals, Oxford (UK): Oxford University Press. 6. Brittain HG. 1999. Polymorphism in pharmaceutical solids. New York: Marcel Dekker, Inc. 7. See for example: Hammond RB, Jones MJ, Murphy SA, Roberts KJ, Smith EDL, Klapper H, Kutzke H, Docherty R, Cherryman J, Roberts RJ, Fagan PG. 2000. Determining the crystal structures of organic solids using X-ray powder diffraction together with molecular and solid state modeling techniques. Mol Cryst Liq Cryst 356:389–405, and references therein. 8. Dinnebier R, Sieger P, Nar H, Shankland K, David WIF. 2000. Structural characterization of three crystalline modifications of telmisartan by single crystal and high-resolution X-ray powder diffraction. J Pharm Sci 89:1465–1479. 9. Chernyshev VV. 2004. X-ray powder diffraction of organics. In: Masciocchi N, editor. Powder diffraction of molecular functional materials, Vol. 31. Chester, England: IUCr Commission of Powder Diffraction. pp 5–15. 10. Werner PE, Eriksson L, Westdahl M. 1985. TREOR, a semi-exhaustive trial-and-error powder indexing program for all symmetries. J Appl Crystallogr 18: 367–370. 11. De Wolff PM. 1968. A simplified criterion for the reliability of a powder pattern indexing. J Appl Crystallogr 1:108–113. 12. Smith GS, Snyder RL. 1979. FN: A criterion for rating powder diffraction patterns and evaluating the reliability of powder-pattern indexing. J Appl Crystallogr 12:60–65. 13. TOPAS-R V3.1. 2004. Bruker AXS, Karlsruhe (Germany). 14. Inter alia, these complete structural analyses (including indexing, structure solution and Riet-

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 5, MAY 2005

15.

16.

17.

18.

19.

20.

21.

22.

veld refinement) demonstrate that all materials investigated in this study (Acitretin forms I, II and III) were prepared as pure monophasic species without any detectable contamination. The molecular arrangement is nearly, but not exactly, B-centered, in the non-conventional B21/d (i.e. P21/c) space group; in the latter description, Z0 ¼ 1 and only one disordered molecule is present. That the space group is indeed P21 (with Z0 ¼ 4) is manifested by weak, but clearly visible, h0l reflections (h þ l6¼2n) both in the single-crystal and powder diffraction datasets. Dollase WA. 1987. Correction of intensities for preferred orientation in powder diffractometry: Application of the March model. J Appl Crystallogr 19: 267–272. Cheary RW, Coelho AA. 1992. A fundamental parameters approach to X-ray line-profile fitting. J Appl Crystallogr 25:109–121. Altomare A, Burla MC, Camalli M, Cascarano GL, Giacovazzo C, Guagliardi A, Moliterni AGG, Polidori G, Spagna R. 1999, SIR-97: A new tool for crystal structure determination and refinement J Appl Crystallogr 32:115–119. Sheldrick GM, SHELX-97. 1997. Program for crystal structure determination. University of Go¨ttingen, Go¨ttingen (Germany). Masciocchi N, Sironi A. 2004. X-ray powder diffraction of low-dimensional organmetallic systems. In: Masciocchi N, editor. Powder diffraction of molecular functional materials, Vol. 31. Chester, England: IUCr Commission of Powder Diffraction. pp 18–22. Masciocchi N, Sironi A. 2005. Structural powder diffraction characterisation of organometallic species: The role of complementary information. Comptes Rendus (in press). Masciocchi N, Galli S, Sironi A. 2005. X-ray powder diffraction characterization of polymeric metal diazolates. Comm Inorg Chem (in press).