Crystallization of Itraconazole Polymorphs from Melt - Crystal Growth

May 20, 2016 - *(T.W.Y.L.) Address: School of Pharmacy, The Chinese University of Hong Kong, Rm 801H, Lo Kwee-Seong Integrated Biomedical Sciences Bui...
2 downloads 12 Views 2MB Size
Download PDF
Article pubs.acs.org/crystal

Crystallization of Itraconazole Polymorphs from Melt Shuai Zhang, Thomas W. Y. Lee,* and Albert H. L. Chow* School of Pharmacy, The Chinese University of Hong Kong, Shatin, Hong Kong ABSTRACT: The impacts of the nucleation site and temperature on the melt crystallization of itraconazole (ITZ) were investigated. ITZ recrystallized from melt into three distinct spherulite zones between two glass coverslips, as observed under a polarized light microscope. The first two corresponded to Forms I and II which are monotropically related, as determined by solubility and thermal analyses. The third zone differed from the other two in the X-ray powder diffraction pattern, thermal properties, Fourier transform infrared spectrum, and dissolution rate, indicative of a new form (III) which has not been reported previously. Form III readily converted to the more stable Form II with the evolution of heat upon heating above 100 °C or Form I upon dissolution in aqueous ethanol. At 70−150 °C, Form II exhibited the fastest bulk crystal growth, followed successively by Form I and Form III. Analysis of the crystal growth data using the two-dimensional nucleation model afforded good data fitting with the model and revealed an inverse relationship between the bulk crystal growth rates and activation energies of individual polymorphs, reflecting the suitability of this model for describing the ITZ crystallization from melt. The polymorph selectivity of recrystallized ITZ appeared to be dictated by both the nucleation and crystal growth rates of individual polymorphs.

■

INTRODUCTION Poor or limited drug solubility is one of the major challenges in oral drug delivery. It is well recognized that amorphization through molecular dispersion in a hydrophilic polymer (commonly known as amorphous solid dispersion) can increase the aqueous solubilities, dissolution rates, and possibly oral bioavailabilities of poorly water-soluble drugs.1 Despite the widely documented utility of this approach to resolving the solubility-related absorption issue of problematic drugs, the number of amorphous drug products that have successfully made their way to the market is still limited, which can be attributed to their inherently poor stability.2,3 Owing to their high free energy and thermodynamic activity, amorphous solids tend to revert back to their low-energy crystalline counterparts during storage, processing, and/or dissolution,4−6 resulting in a complete loss of their unique advantage in solubility and bioavailability enhancement. A review of the literature showed that most of the studies on amorphous solid dispersions focused on the characterization of drug−polymer miscibility, thermal properties, spectral characteristics, dissolution performance, and physical stability under stressed conditions. Very few of these studies have attempted to elucidate the recrystallization behaviors of dispersed drugs in these systems. In addition, while the literature is replete with information on the solution-phase crystallization of drugs, very little is known about the solid-state crystallization behaviors of amorphous drugs in polymeric matrix,6−8 particularly for drugs capable of crystallizing into different polymorphs with different free energies and thermodynamic stabilities. Typically, upon recrystallization from the amorphous state, the relatively unstable polymorph will form first, as it is relatively close in free energy to the parent phase (Ostwald rule of stages). © XXXX American Chemical Society

However, this unstable (metastable) polymorph will eventually revert to the stable form because of the associated favorable reduction in free energy. By elucidating the mechanisms of polymorph selectivity and establishing the conditions under which this occurs in the solid state, rational and effective strategies can be devised to control the formation of a particular polymorph or to kinetically stabilize the solid drug in its amorphous form. The latter is commonly achieved by dispersing the drug in hydrophilic polymers as amorphous solid dispersions. Polymers can be selected based on their ability to inhibit the crystallization of the initially formed (unstable) polymorph, and the best polymer would be the one that exhibits the strongest crystal growth inhibition and offers the drug maximum physical stability.9,10 Aimed at acquiring a mechanistic understanding of the crystallization behaviors and polymorph selectivity of amorphous drugs, the present study has employed itraconazole (ITZ) as a model compound. ITZ, a synthetic triazole antifungal agent, is a BCS II compound with extremely low water solubility (1 ng/mL at pH 7.4), which may account for the relatively low oral bioavailability (∼55%) of its commercial product Sporanox.11 Moreover, ITZ has a relatively high glass transition temperature (Tg ≈ 59 °C) and tends to form a fairly stable amorphous form through various processing treatments such as melt-quenching and lyophilization.11−13 Thus far, only one polymorph (Form I) has been reported in the literature. U.S. Patent Application Publication No. 2003/0100568 and U.S. Patent No. 7,193,084 document the existence of a second Received: March 2, 2016 Revised: May 15, 2016

A

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

hermetically sealed aluminum pan and scanned from 30 to 180 °C at 3 °C/min under N2 purge at 50 mL/min. The enthalpies of fusion (ΔHf) of the different forms were determined and the corresponding entropies of fusion (ΔSf) were calculated from the relation ΔSf = ΔHf/ Tm, where Tm is the melting point. In addition, the thermal behaviors of Form III were investigated at different heating rates. X-ray Powder Diffraction (XRPD). X-ray powder diffraction (XRPD) data were collected using a Panalytical X-ray diffractometer (Philips X’Pert PRO, The Netherlands) equipped with a Cu radiation source operating at 40 kV and 40 mV. Sample was uniformly packed into an aluminum holder with a 2 mm depth and scanned from 2θ of 3° to 40° at a rate of 4°/min with a step size of 0.017°. Fourier Transform Infrared (FTIR) Spectroscopy. FTIR measurement was conducted using a Bruker Alpha FTIR spectrophotometer (Bruker Corp., MA, USA) in the wavenumber range of 4000 cm−1 to 400 cm−1 in the transmittance mode. One hundred and twenty-eight scans were collected at a resolution of 4 cm−1 for each sample. Solubility Determination. Solubilities of different polymorphs were determined by adding excess drug to 10 mL of 50% v/v aqueous ethanol in glass vials, followed by equilibration with occasional shaking at different temperatures (4 °C, 10 °C, 20 °C, 25 °C, 34 °C, 40 °C) in a temperature-controlled incubator (Lovibond TC 135S, Tintometer GmbH, Dortmund, Germany) for 3 days (i.e., until equilibrium was reached). The solution was then filtered, and the drug content was analyzed by UPLC (Agilent 1200 Infinity LC systems) using a UPLC ZORBAX Eclipse Plus C18 column (2.1 mm × 50 mm; 1.8 μm). The mobile phase consisted of acetonitrile and 0.1% formic acid (60:40 v/ v) eluted at 1 mL/min. Quantification was based on the peak area measurement by UV absorption at λ = 260 nm. After the study, the residual solid in each glass vial was analyzed by DSC and XRPD to check for possible polymorphic transitions during the study. Solubility data were expressed in mole fraction as follows: n2 x2 = n1 + n2

polymorph (Form II).14,15 Although the physical stability of amorphous ITZ formulations has been extensively studied, none of these studies has delved into the polymorphic crystallization mechanism of pure amorphous ITZ, which may be critical for successful development of stable ITZ solid dispersion systems.16 Accordingly, the present work carried the following objectives: (1) to identify the polymorphs of ITZ recrystallized from melt; (2) to determine the melt recrystallization kinetics and thermodynamic relationships of these polymorphs; and (3) to investigate the factors governing the crystal growth of different ITZ polymorphs at the surface and in the bulk of the melt as a function of temperature.

■

MATERIALS AND METHODS

Chemicals and Reagents. Itraconazole ((2R,4S)-rel-1-(butan-2yl)-4-{4-[4-(4-{[(2R,4S)-2-(2,4-dichloro phenyl)-2-(1H-1,2,4-triazol1-ylmethyl)-1,3-dioxolan-4-yl]methoxyphenyl)piperazin-1yl]phenyl}4,5-dihydro-1H-1,2,4-triazol-5-one) with a purity >99.0% was purchased from YICK-VIC (Hong Kong). Ethanol, acetonitrile, and formic acids were of analytical grade (with >99% purity) and purchased from VWR Chemical Co. Ltd. (Hong Kong). Preparation of Amorphous ITZ for Crystal Growth Measurement. About 4−5 mg of crystalline ITZ was melted at 175 °C between two glass coverslips in a temperature-controlled oven, and the sample was then rapidly quenched to room temperature to form an amorphous solid. The solidified sample was confirmed to be crystalfree by means of a polarized light microscope (Axioplan 2 imaging and Axiophot 2 microscope, Carl Zeiss Corp., Germany). For the bulk recrystallization studies, samples were stored together with the attached coverslips at different elevated temperatures to initiate recrystallization, whereas, for the surface recrystallization experiments, the top coverslip was carefully removed to expose a free surface before storing the samples under the same conditions. Crystal growth measurements for both the surface and bulk of amorphous melts basically followed the protocols of Cai et al. with modifications.17 Here bulk crystal growth is defined as the crystal growth around a preformed crystal seed within an amorphous melt, which is sandwiched between two glass coverslips to avoid contact of the adhered melt surface with the surrounding air, while surface crystal growth is defined as the crystal growth around a preformed crystal seed at a melt surface, which is being exposed to open air after removal of the top coverslip. Determination of Crystal Growth Rates of ITZ Polymorphs. Because of the extremely slow crystal growth rates of ITZ polymorphs at room temperature, all growth rate measurements were conducted at elevated temperatures (i.e., 70−140 °C). Measurements at temperatures above 100 °C were performed by hot stage microscopy (HSM); samples were placed on a Linkam THMS 600 hot stage, and growth rate was measured in real time. For the studies conducted at 70−100 °C, samples were maintained at a constant temperature in a temperature-controlled oven in the presence of P2O5 (∼0% RH) and removed at predetermined time points for growth rate measurement under the polarized light microscope. Crystal growth rate at a defined temperature was measured from the advancing speed of crystal front into the melt, and morphology of growing crystal was concurrently examined. A plot of crystal size versus time was found to be linear, and the slope was taken as the growth rate. Each reported growth rate was the average of at least four independent measurements. Differential Scanning Calorimetry (DSC). Thermal analysis was conducted using a PerkinElmer Q6000 differential scanning calorimeter (PerkinElmer, MA, USA). Different forms of ITZ were allowed to recrystallize between two glass coverslips. Once recrystallization was completed, the top coverslip was gently removed, and the materials in different growth zones were carefully scraped off from the surface of the bottom coverslip. An accurately weighed quantity (2−3 mg) of each ITZ polymorph was encapsulated in a

where n1 is the number of moles of solvent, and n2 is the number of moles of solute. Powder Dissolution Studies. All powder samples for dissolution testing were first sieved through an 80 mesh-screen (180 μm), and only particles with sizes smaller than 180 μm were used. An accurately weighed amount (20 ± 0.2 mg) of each ITZ polymorph or ITZ amorphous powder (made by rapid melt quenching) was dispersed in 25 mL of 50% v/v aqueous ethanol with continuous stirring at 100 rpm in a round-bottom flask, equilibrated inside a temperaturecontrolled incubator (Lovibond TC 135S, Tintometer GmbH, Dortmund, Germany) at 25 °C. At appropriate time intervals, a 400 μL aliquot of each solution was withdrawn, filtered through 0.22 μm nylon membrane filters (Sigma-Aldrich Corp., MO, USA), and diluted with a suitable amount of 50% v/v aqueous ethanol prior to UPLC analysis. Each sample withdrawn was replaced by the same volume of fresh dissolution medium. The method of drug content detection was the same as described in the preceding section. At the end of the dissolution study, the undissolved solid left in the medium was dried at room temperature using a vacuum concentrator (Savant SPD131DDA SpeedVac Concentrator, Thermo Fisher Scientific, MA, USA) before being checked by DSC or XRPD for possible polymorphic transition. Viscosity Measurement. The steady shear viscosities of amorphous ITZ at different temperatures were measured using an ARES rheometer (TA Instruments, USA). The method was essentially the same as that reported previously.18 Briefly, measurements were conducted using two 25 mm-diameter parallel plates with a 0.8 mm gap size between them. Nitrogen gas was used to control the temperature inside the heating hood. ITZ crystalline powder was melted between the two parallel plates, and the sample was held isothermally at a temperature 5 °C higher than the melting point of ITZ (∼165 °C) for a few minutes to ensure complete melting. Afterward, the temperature was lowered to a predefined temperature, and the upper plate was then lowered, and the excess material was trimmed off along the border of the plate. The sample was equilibrated B

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 1. Morphologies of ITZ crystals grown under different conditions. (A) Form I crystal grew by diffusion from the side of the bulk at 100 °C. (B) Bulk Form II crystals at 100 °C. (C) Bulk Form I, II, and III crystals at 100 °C; arrows indicate respective polymorphs. (D) Bulk Form III crystals at 100 °C. (E) Bulk Form I crystals grown at 130 °C. (F) Bulk Form II crystals grown at 130 °C. (G) Surface Form I crystals at 100 °C. (H) Surface Form II crystals at 100 °C. Scale bar equals 1000 μm. at the test temperature for 15 min, and then a shear deformation was applied at a rate of 0.1 s−1 to 10 s−1 at a constant temperature. The shear viscosity value was recorded when the shear rate reading became steady.

■

conditions, as observed under a polarized light microscope. Figure 2A−C shows the XRPD patterns for the three different growth zones, respectively. The diffraction pattern of the first zone was consistent with that of the commercial Form I,19,20 while the second one exhibited the same diffraction pattern as that of Form II, as first documented in the U.S. Patent Application Publication No. 2003/0100568.14 The third zone, tentatively termed Form III, was characterized by signature peaks at approximately 11.9°, 16.7°, 17.4°, 18.7°, 19.1°, 20.0°,

RESULTS

X-ray Powder Diffraction and Thermal Analyses. As shown in Figure 1, amorphous ITZ recrystallized from melt between two glass coverslips into three distinct spherulite growth zones with different morphologies under different C

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

formation to Form II, which was completed at 145 °C (Figure 3B,C). Further heating led to melting of Form II at 158 °C (Figure 3D) and eventually melting of Form I at around 167 °C (Figure 3E). To gain further insight into the thermal behaviors and melting properties of the different polymorphs, DSC analyses were conducted, and the respective thermograms are presented in Figure 4A. Form I melted at 165.1 °C, which is in excellent agreement with the melting point of the commercial ITZ form (Form I).12,13 The melting point of Form II (155.7 °C) is the same as that presented in the U.S. Patent Application Publication No. 2003/0100568 and U.S. Patent No. 7,193,084.14,15 The measured molar enthalpy of fusion, ΔHf, of Form II was 60.6 ± 1.7 kJ/mol, which is less than that (ΔHf = 68.6 ± 1.5 kJ/mol) of Form I. The molar entropy of fusion, ΔSf, of Form II (141 ± 4 J/mol·K) was also less than that of Form I (ΔSf = 156 ± 4 J/mol.K). A well-defined melting endotherm and reliable ΔHf estimate could not be obtained for Form III because it rapidly converted to Form II upon heating, as evidenced by the progressive phase changes of Form III observed with increasing heating rate in DSC (Figure 4B). At low scanning rate, e.g., 10 °C/min, a small exothermic peak appeared at ∼120 °C (which possibly reflects the transition of Form III to Form II), followed by an endotherm at ∼156 °C, corresponding to the melting of Form II. At higher heating rates, e.g., 40 °C/min, the melting endotherm of Form III started to emerge at 119.4 °C, followed by a relatively small endotherm at a higher temperature, which can be attributed to the melting of recrystallized Form II from the Form III melt. Because of the coexistence of Form III and/or its melt, this Form II was seen to melt at a lower temperature than its phase pure counterpart (mp ∼156 °C). Such a transition to Form II through the molten state appeared to occur extremely rapidly since the Form II endotherm was still discernible at the highest scanning speed (100 °C/min). Consequently, in order for the melting event of Form III to become visible, the melting of Form III needs to occur faster than the concurrent recrystallization of Form II from the Form III melt, which would only be possible at sufficiently high heating rates, as substantiated by the current DSC results. However, the possibility of a transition of Form III to Form II in the solid state rather than via the molten state cannot be entirely ruled out. To verify such a possibility, Form III was subjected to grinding using a pestle and mortar for different time periods prior to DSC analysis at 40 °C/min. Grinding is known to be effective for facilitating polymorphic conversion in the solid state by weakening the intermolecular bonding in the material and lowering the activation energy barrier against the conversion. As shown in Figure 4C, no melting endotherm of Form III was observable for the ground sample; the thermogram only displayed an exothermic transition at ∼90 °C, and the typical melting endotherm of pure Form II at ∼156 °C. The exothermic event is probably a result of the solid phase transition of Form III to Form II since the Form III material remained unchanged in crystal structure after grinding, as confirmed by XRPD. Infrared Spectral Analysis. The three ITZ polymorphs also displayed distinct differences in two defined fingerprint regions of the FTIR spectra, viz. 1650−1750 cm−1 and 800 cm−1−1100 cm−1 (Figure 5). In the first region, Form I showed a strong absorption peak at 1699 cm−1, ascribable to the CO stretch, while the corresponding peak of Form III occurred at 1696 cm−1. As for Form II, peak splitting appeared at 1692 and 1705 cm−1. In the second fingerprint region, Form I showed

Figure 2. PXRD patterns of various ITZ polymorphs. Samples recrystallized from melts were scraped from glass coverslips for the PXRD analysis. (A) Form I. (B) Form II. (C) Form III. (D) Form III powder annealed at 100 °C for 10 h. (E) Form III powder annealed at 120 °C for 10 h. (F) Amorphous ITZ prepared by rapid melt quenching.

and 21.1°, indicative of a new crystal form, which has not been reported previously. As with other drug materials such as felodipine16 and nifedipine,21 ITZ exhibited temperature-dependent recrystallization behaviors. All of the three forms recrystallized from melt as spherulites at temperatures below 110 °C (Figure 1C). Form III grew only in the melt bulk as relatively small, dark spherulites with distinct concentric rings on their surface (Figure 1D). While the morphologies of Forms I and II were similar and almost indistinguishable, they could still be differentiated by their differences in optical properties and surface features under the polarized light microscope. Form I spherulites appeared darker and displayed a more striated surface with clearly discernible radial lines (Figure 1C). At temperatures above 110 °C, Forms I and II grew as separate faceted crystals (Figure 1E,F), whereas Form III transformed to Form II. The latter was further substantiated by annealing Form III at 100 and 120 °C. As verified by XRPD (Figure 2D,E), annealing of Form III at 120 °C for 10 h led to its conversion to Form II, while similar treatment at 100 °C afforded no such phase transition. Form III was found to be stable at ambient conditions since scraping of the sample from the glass coverslip for XRPD analysis did not induce any solid− solid transition. Samples of ITZ polymorphs recrystallized from melt on glass coverslips were first examined by HSM. As shown in Figure 3, upon heating above 100 °C, Form III underwent transD

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 3. Hot-stage microscopic examination of ITZ polymorphs at different temperatures. (A) ITZ crystals at 100 °C; arrows indicate different forms of ITZ. (B) ITZ crystals at 120 °C. (C) ITZ crystals at 145 °C; Form II converted from Form III in (A). (D) ITZ crystals (Form I) at 158 °C. (E) Melting of Form I at 167 °C. (F) Complete melting of Form I at 170 °C. Heating rate of hot stage is 3 °C/min. Scale bar equals 2000 μm.

two strong absorption peaks at 998 and 899 cm−1, which were absent in both Forms II and III. In addition, both Forms I and II exhibited an absorption peak at 1010 cm−1 whose intensity decreased markedly in Form III. All these IR spectral differences provide a useful means of differentiating the various ITZ polymorphs. Solubility and Dissolution Studies. In addition to the enthalpy of fusion, equilibrium solubility and the related enthalpy of solution in a given solvent system can offer useful information on the bonding strength within the crystal lattice of a particular polymorph. Figure 6 shows the van’t Hoff solubility plot of natural logarithms of mole fraction solubility (ln x2) versus the reciprocals of absolute temperature (1/T) for Form I and Form II in 50% v/v aqueous ethanol at different

temperatures. According to the van’t Hoff equation, the plot should be linear if the enthalpy of solution term in the equation is truly temperature-independent.22 The molar enthalpies of solution determined for Form I and Form II from linear regression of the solubility curves in Figure 6A were 55.8 (±2.8) kJ/mol and 51.3 (±2.9) kJ/mol, respectively. The extrapolation of two linear plots yielded a transition temperature, Tr, of 497.8 °C at the intersection point for the two polymorphs. The enthalpy, ΔHII→I, and entropy ΔSII→I of polymorphic transition from Form II to Form I were estimated to be −4470.8 J/mol and −5.7 J/K·mol from the slope of the log solubility ratio plot in Figure 6B and from the relation, ΔSII→I = ΔHII→I /Tr (i.e., when ΔGII→I = 0), respectively. The free energies of polymorphic transition, ΔGII→I, from Form II E

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

to Form I at 25 and 37 °C, as defined by the equation, ΔGII→I = ΔHII→I − TΔSII→I, were calculated to be −2771 J/mol and −2703 J/mol, respectively. The solubility of Form II was about 2−3 fold higher than that of Form I in the temperature range studied. Unlike Form I and Form II which remained stable throughout the solubility study, Form III underwent conversion to Form I, as verified by DSC and XRPD analyses of the excess solid left after the study, and therefore, its saturation solubility could not be determined. As an alternative means to compare the solubilities of the three ITZ polymorphs, the dissolution−time profiles of these crystal forms were determined in the same solvent system. According to the Noyes Whitney equation, the dissolution rate per unit surface area of a drug solid in contact with the dissolution medium is directly proportional to its solubility under sink conditions at a constant temperature. Thus, the amount of drug dissolved initially over a defined time period can provide an indirect measure of the drug solubility. Figure 7 shows the dissolution−time profiles of the different ITZ forms in 50% v/v aqueous ethanol at 25 °C. The concentrations of Form I and Form II in dissolution media increased gradually with time and plateaued off after ∼10 h. The maximum drug concentration attained was 2−3 fold higher for Form II than for Form I, consistent with the equilibrium solubility data. In contrast, Form III dissolved much more rapidly than the other two forms during the initial time period, reaching a peak drug concentration about 7 fold than that of Form I at ∼15 h before the drug concentration started to decline. In comparison with all three polymorphs, the dissolved drug concentration with the amorphous form increased most rapidly with time during the initial 10 min time period, but quickly dropped thereafter to an essentially constant level, equivalent to the equilibrium solubility of Form I after 50 h. DSC and XRPD analyses of all undissolved solids at the end of the dissolution studies revealed that both Form III and the amorphous form were converted to Form I in the dissolution media, while Form I and Form II remained unchanged in their crystal structures. Crystallization Kinetic Studies. Figure 8 shows the plots of bulk crystal growth distances of the three polymorphs against time. The apparent linearity in these plots (r2 > 0.999; n = 5) is consistent with that observed in other studies on bulk crystallization from melts,17,23 which would suggest that the bulk crystal growth rate is independent of time. The temperature dependence of the bulk crystal growth rate is shown in Figure 9. The bulk crystal growth rate of Form III at very high temperatures (>120 °C) could not be measured as it readily converted to Form II after melting at around this temperature, as discussed earlier. On the other hand, the bulk crystal growth rate at around the glass transition temperature of ITZ (∼59 °C) was too slow to be measured within a reasonable time frame. Compared to similar data obtained with other drugs such as indomethacin,23 nifedipine,21 the crystal growth rate of ITZ was relatively slow. Within the workable temperature range, the crystal growth rate of Form II was significantly higher than those of Form I and Form III, while the crystal growth rate of Form III was the slowest among the three polymorphs. To examine the impact of nucleation site on crystal growth, the crystal growth rates at the surface and in the bulk of the ITZ melt were determined. Typically, crystal growth rate at the surface is of several orders of magnitude higher than that in the bulk.21,24,25 At temperatures above 80 °C, the crystal growth rates of Form I and Form II at the surface were comparable to

Figure 4. (A) DSC analysis of different polymorphs and amorphous form of ITZ at 3 °C/min. Form I and Form II show single melting endotherms. Form III shows split peaks indicative of partial conversion to Form II during heating; the small exothermic peak at around 120 °C indicates transition of Form III to Form II. Amorphous form of ITZ was prepared by quenching of ITZ melt through rapid cooling to room temperature in a DSC pan. (B) DSC analysis of Form III samples at different heating rates. The melting endotherm of Form III began to emerge at higher heating rates. (C) DSC analysis of Form III samples ground for different time periods at 40 °C/min; the exothermic peak at ∼90 °C indicates solid phase transition from Form III to Form II. F

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 5. FTIR spectra of ITZ polymorphs. Arrows indicate the characteristic peaks (1699 cm−1, 1010 cm−1, 998 cm−1, 899 cm−1 for Form I; 1705 cm−1, 1692 cm−1, 1013 cm−1 for Form II; 1696 cm−1, 1010 cm−1 for Form III) used to identify the polymorphs.

Figure 7. Dissolution−time profiles of ITZ Forms I, II, III, and amorphous form in 50% v/v aqueous ethanol at 25 °C. Form I and Form II maintained their original crystal structures, while Form III and the amorphous material converted to Form I after the 3-day dissolution study. Each data point is the mean of three independent determinations.

those in the bulk (data not shown). The surface growth rate of Form III could not be determined as only Form I and Form II could nucleate at the surface. While there were no significant differences in the growth rate for both surface and bulk crystallization between Form I and Form II in the temperature range employed (70−140 °C), the nucleation rate at the surface was apparently higher than that in the bulk. For instance, at 90 °C, Form I appeared at the surface within 5 h, but it was still not apparent in the bulk after 2 days. Furthermore, as reported in other studies, nucleation of ITZ usually initiated at the surface and crystal growth then

Figure 6. van’t Hoff plots of (A) log solubility for ITZ Forms I and II and (B) log solubility ratio of Form II and Form I in 50% v/v aqueous ethanol. Each data point is the mean of three independent determinations.

G

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

The flow of ITZ in the amorphous supercooled state was found to be Newtonian; i.e., the shear viscosity does not change with the shear rate. To account for the influence of temperature on the crystallization of the three ITZ polymorphs, the crystal growth kinetic data obtained at different temperatures were further analyzed using the following equation derived from the twodimensional nucleation model:18,27 v=

⎛ ΔG* ⎞ k exp⎜ − ⎟ η ⎝ kBT ⎠

(1)

where v is the crystal growth rate; k is a constant; η is the viscosity of the supercooled liquid; ΔG* is the energy barrier to the nucleation of a new crystal monolayer; kB is the Boltzmann constant, and T is the absolute temperature. The activation energy, ΔG*, can be estimated from

Figure 8. Growth distances of different ITZ polymorphs plotted as a function of time at 100 °C. The data were analyzed by linear regression and the slope was taken as the crystal growth rate, u. Each data point is the mean of at least four separate measurements.

ΔG* = −

πbγ 2 ΔGv

(2)

where b is the height of the two-dimensional nuclei, γ is the crystal/amorphous interface free energy, and ΔGv is the free energy change per unit volume from the amorphous phase to the crystalline phase, which can be computed from the Hoffman equation as a function of temperature as follows: ΔGv =

−ΔH f (Tm − T ) Tm

(3)

where ΔHf is the enthalpy of fusion of the crystalline phase, Tm is the melting point and ΔT = Tm − T. Combining eqs 1, 2, and 3, we obtain Figure 9. Crystal growth kinetics of different ITZ polymorphs in the melt bulk as a function of temperature. Each data point is the mean of at least four separate measurements.

v=

proceeded along the boundary and penetrated into the bulk.21,26 To evaluate the crystallization kinetics of ITZ, the zero-shear (or near zero-shear) viscosity of ITZ in the supercooled liquid region was measured as a function of time at 80−130 °C (Figure 10). Within this temperature range, the viscosity measurement was typically completed within 30 min, and the absence of crystalline ITZ was confirmed by optical inspection.

⎛ πbγ 2Tm ⎞ k ⎟⎟ exp⎜⎜ − f η ⎝ kBT ΔH ΔT ⎠

(4)

Equation 4 can be transformed to the following linear form: ln(vη) = ln k −

πbγ 2Tm kBT ΔH f ΔT

(5)

Typical linear plots of ln(vη) versus 1/TΔTbased on eq 5 of the three ITZ polymorphs (r2 = 0.84−0.97; n = 5) are shown in Figure 11. The free energies of activation, ΔG*, of crystal growth of the three ITZ polymorphs at different temperatures were calculated from the slopes of the plots, and the results are summarized in Figure 12.

Figure 10. Zero-shear rate viscosities of pure ITZ in supercooled amorphous state at various temperatures.

Figure 11. Plot of ln(vη) versus1/TΔT for different ITZ polymorphs. The straight lines were obtained by linear regression of the data. H

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Typical Recrystallization Behaviors of Amorphous ITZ. Because of the high Tg and the slow crystal growth rate of ITZ, it would not be practical or even feasible to measure the crystal growth rate of ITZ at around its Tg or below. Hence, all our crystal growth measurements had to be conducted at elevated temperatures (80−140 °C). The recrystallization behavior of amorphous ITZ is fairly complicated compared with other amorphous drugs.21,23,29 Below 90 °C, instead of recrystallization directly from an isotropic liquid state like other drugs, amorphous ITZ would initially form intermediate liquid crystals which have a certain orientational but without any translational order.30 Two mesophase transitions (Figure 4A) for ITZ were identified in our study. The one near 90 °C could result from an isotropic liquid-chiral nematic mesophase transition, and the other at 74 °C was probably due to the rotational restriction of the molecule. These findings are in excellent agreement with the results of a similar study reported earlier.13 The XRPD of amorphous ITZ in the present study also indicates the existence of the typical nematic mesophase, as evidenced by the presence of distinct peaks at 2θ = 7° and 18° (Figure 2F).13 Here, the kinetic barrier of the isotropic liquid to liquid crystal transition is likely to be very low, as liquid crystals still formed even when the molten ITZ was quenched by liquid nitrogen. However, the energy barrier against the liquid crystalto-crystal transition appeared to be much greater, since the amorphous sample was still crystal-free after being stored at room temperature for two months. The reason for the existence of such liquid crystals is not immediately apparent but may be due to the high asphericity of the molecule.30 It is still unclear how this particular behavior would impact the recrystallization of amorphous ITZ. However, maintaining ITZ in a liquid crystal form may also serve as an effective strategy to improve its dissolution rate and bioavailability.31 Recrystallization Kinetics of ITZ from Melt. It has been widely observed that crystal growth rate at the surface of melt is of several orders of magnitude higher than that in the bulk.17,23,32 Several mechanisms have been proposed to explain this phenomenon, including release of crystallization-induced tension at the surface, surface molecular mobility, and so forth.26,33 However, it was found in a previous study employing felodipine that at 70−90 °C, the crystal growth rates at the surface and in the bulk of melt did not differ significantly.16 Indeed, our present study with ITZ also yielded results similar to felodipine. This can be explained by a different crystal growth mechanism being in operation at elevated temperatures when the viscosity of supercooled liquids becomes very low. Despite the fact that our study also demonstrated insignificant differences in crystal growth rate between the surface and bulk of the melt at elevated temperatures, the nucleation process was apparently faster at the surface than in the bulk between 60 and 140 °C. This is reflective of a relatively low energy barrier to nucleation at the surface,21,25 which explains why nucleation usually occurs at the surface, followed by crystal growth along the surface and into the bulk.26,32 To probe into the crystal growth mechanism of the different polymorphs, we had further analyzed the crystal growth data of ITZ obtained at different temperatures based on the wellestablished two-dimensional nucleation model.18,27 This data treatment generally yielded good data fitting with the model and revealed an inverse relationship between the crystal growth rates and the calculated ΔG* for the three polymorphs (Figures 9 and 12), reflecting the suitability of this model for delineating the crystal growth of ITZ from melt.

Figure 12. Relationship between the activation energies, ΔG*, of crystal growth of different ITZ polymorphs and temperature.

■

DISCUSSION Characterization of ITZ Polymorphs. Amorphous ITZ prepared by melt quenching recrystallized into three different polymorphs (Forms I, II, and III). Form I and Form II are monotropically related, as suggested by the enthalpy/entropy of fusion rule (i.e., the higher melting form has the higher enthalpy/entropy of fusion). This has been further substantiated by the intersection of the solubility curves of Form I and Form II in the van’t Hoff plot being at a temperature (i.e., transition temperature) higher than the melting points of both forms. The fact that Form I has a higher ΔHf and a higher ΔHs (in 50% v/v aqueous ethanol) than Form II suggests that Form I has a stronger crystal lattice and is the more stable form. In addition, the difference in ΔHf (at Tm) between Form I and Form II is larger than that of ΔHs (determined at 4−40 °C), reflecting the difference in molar heat capacity, ΔCp, between the two forms.28 On the basis of the different characteristic peaks in XRPD, distinctly different IR spectra, and phase changes at elevated temperatures in DSC, Form III is likely to be a novel crystal form of ITZ, which, to the best of our knowledge, has never been reported previously. Unfortunately, characterization of Form III and its polymorphic relationship with the other two forms by conventional solubility measurement or thermal analysis has not been fruitful due to its conversion to Form I upon dissolution in aqueous ethanol or to Form II upon heating in DSC. However, on the basis of its rapid exothermic transition to Form II in DSC at elevated temperatures (>100 °C) and the induction of this transition at a lower temperature by grinding treatment (Figure 4B,C), Form III should be a metastable form and its relationship with Form II is likely to be monotropic. The relatively low thermodynamic stability of Form III has been further attested by its much faster dissolution in 50% v/v aqueous ethanol compared to the other two forms. As shown in Figure 7, the peak dissolved drug concentration attained over a defined time period with Form III was about 7 fold and 2.2 fold higher than those with Form I and Form II respectively. In addition, although Form III readily transforms to the more stable Form II at elevated temperatures or Form I in solution, it is kinetically stable at ambient conditions, being relatively insensitive to grinding treatment. Taken together, the observed thermal behaviors and properties of the three polymorphs suggest that the rank order of decreasing stability for these polymorphs should be Form I > Form II > Form III. I

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Polymorph Selectivity of Recrystallized ITZ from Melt. It should be noted that polymorph selectivity depends not only on crystal growth rate, but also on nucleation rate. Although the metastable Form II displays the fastest crystal growth rate among the three polymorphs in the temperature range of our investigation (70−140 °C), it may not be the predominant or most prevalent form upon spontaneous crystallization. Nucleation rate may also be a crucial determinant of the final crystal form, since nucleation precedes crystal growth in any crystallization process.34,35 In our studies, nuclei of Form I started to appear at a lower temperature compared with Form II, and the Form I nuclei tended to outnumber those of Form II (data not shown), indicative of a lower activation energy barrier to nucleation and a faster nucleation rate for Form I than for Form II at the surface. Furthermore, only Form I was observed in most of the previously reported studies on the physical stability of amorphous ITZ.19,20,36,37 Consequently, although Form II grows faster than Form I, Form I may still be the predominant form during spontaneous recrystallization from melt. As for Form III, nucleation was only detected in the bulk but not at the surface, and it showed the slowest crystal growth rate, which would render it the least probable form during crystallization. In short, polymorph selectivity of ITZ recrystallized from melt is a function of both nucleation and crystal growth rates.

N.T. Hong Kong. E-mail: [email protected]. Tel: 852 39439795. Fax: 852 26035295. *(A.H.L.C.) Address: School of Pharmacy, The Chinese University of Hong Kong, Rm 801E, Lo Kwee-Seong Integrated Biomedical Sciences Building, Area 39, Shatin, N.T. Hong Kong. E-mail: [email protected]. Tel: 852 39436829. Fax: 852 26035295. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support from The Chinese University of Hong Kong (Research Seed Grant Ref. No. 4930033 for T.W.Y.L. and postgraduate studentship for S.Z.) is gratefully acknowledged. We also thank Prof. Y. Zheng, Institute of Chinese Medical Sciences, University of Macau, for assistance with the hot-stage microscopy studies.

■

■

CONCLUSIONS Three polymorphs of ITZ, namely, Forms I, II, and III, were successfully obtained by recrystallization from melt. Form I and Form II are monotropically related, as determined by solubility and thermal analyses. Form III and Form II possibly represent another monotropic pair, as suggested by the rapid exothermic transition of Form III to Form II at elevated temperatures (>100 °C) and the ability of grinding treatment to induce this transition at a lower temperature. In terms of thermodynamic stability, Form I is the most stable polymorph, while Form III is the least stable. The bulk crystal growth rates of the three forms followed the rank order: Form II > Form I > Form III, whereas Form I displayed the fastest nucleation, followed by Form II and then Form III. Analysis of the crystal growth data using the two-dimensional nucleation model generally afforded good data fitting with the model and showed an inverse relationship between the crystal growth rates of the three ITZ polymorphs and the corresponding activation energies, suggestive of the appropriateness of this kinetic model for delineating the crystal growth of ITZ from melt. All these observations suggest that the polymorph selectivity of ITZ recrystallized from the amorphous form is governed by both the nucleation and crystal growth rates of the polymorphs. The present findings may have important implications for crystal polymorphism control as well as the design and development of kinetically stable amorphous solid dispersions of ITZ. Additionally, being apparently stable at ambient conditions and having an aqueous solubility about 7-fold higher than that of Form I (i.e., the commercially available form), Form III may hold promise for development into a more bioavailable and efficacious oral ITZ dosage form.

■

REFERENCES

(1) Lee, T. W.-Y.; Boersen, N. A.; Yang, G.; Hui, H.-W. Drug Dev. Ind. Pharm. 2014, 40, 1072−1083. (2) Serajuddin, A. T. M. J. Pharm. Sci. 1999, 88, 1058−1066. (3) Vasconcelos, T.; Sarmento, B.; Costa, P. Drug Discovery Today 2007, 12, 1068−1075. (4) Laitinen, R.; Lobmann, K.; Strachan, C. J.; Grohganz, H.; Rades, T. Int. J. Pharm. 2013, 453, 65−79. (5) Sun, Y.; Zhu, L.; Wu, T.; Cai, T.; Gunn, E. M.; Yu, L. AAPS J. 2012, 14, 380−388. (6) Yu, L. J. Am. Chem. Soc. 2003, 125, 6380−6381. (7) Andronis, V.; Zografi, G. J. Non-Cryst. Solids 2000, 271, 236−248. (8) Rodríguez-hornedo, N.; Murphy, D. J. Pharm. Sci. 1999, 88, 651− 660. (9) Yang, J.; Grey, K.; Doney, J. Int. J. Pharm. 2010, 384, 24−31. (10) Baird, J. A.; Taylor, L. S. Adv. Drug Delivery Rev. 2012, 64, 396− 421. (11) Taupitz, T.; Dressman, J. B.; Buchanan, C. M.; Klein, S. Eur. J. Pharm. Biopharm. 2013, 83, 378−387. (12) Six, K.; Leuner, C.; Dressman, J.; Verreck, G.; Peeters, J.; Blaton, N.; Augustijns, P.; Kinget, R.; Van den Mooter, G. J. Therm. Anal. Calorim. 2002, 68, 591−601. (13) Six, K.; Verreck, G.; Peeters, J.; Binnemans, K.; Berghmans, H.; Augustijns, P.; Kinget, R.; Van den Mooter, G. Thermochim. Acta 2001, 376, 175−181. (14) Werling, J.; Doty, M. J.; Rebbeck, C. L.; Wong, J. C. T.; Kipp, J. E. U.S. Patent Application Publication 2003/0100568 A1, 2003. (15) Werling, J.; Doty, M. J.; Rebbeck, C. L.; Wong, J. C. T.; Kipp, J. E. U.S. Patent 7,193,084 B2, 2007. (16) Kestur, U. S.; Taylor, L. S. Cryst. Growth Des. 2013, 13, 4349− 4354. (17) Cai, T.; Zhu, L.; Yu, L. Pharm. Res. 2011, 28, 2458−2466. (18) Kestur, U. S.; Lee, H.; Santiago, D.; Rinaldi, C.; Won, Y.-Y.; Taylor, L. S. Cryst. Growth Des. 2010, 10, 3585−3595. (19) Janssens, S.; de Armas, H. N.; Roberts, C. J.; Van den Mooter, G. J. Pharm. Sci. 2008, 97, 2110−20. (20) Janssens, S.; de Armas, H. N.; D’Autry, W.; Van Schepdael, A.; Van den Mooter, G. Eur. J. Pharm. Biopharm. 2008, 69, 1114−1120. (21) Zhu, L.; Wong, L.; Yu, L. Mol. Pharmaceutics 2008, 5, 921−926. (22) Hancock, B.; Parks, M. Pharm. Res. 2000, 17, 397−404. (23) Wu, T.; Yu, L. Pharm. Res. 2006, 23, 2350−2355. (24) Backus, E.; Grecea, M.; Kleyn, A.; Bonn, M. Phys. Rev. Lett. 2004, 92, 236101. (25) Zhu, L.; Brian, C. W.; Swallen, S. F.; Straus, P. T.; Ediger, M. D.; Yu, L. Phys. Rev. Lett. 2011, 106, 256103. (26) Sun, Y.; Zhu, L.; Kearns, K. L.; Ediger, M. D.; Yu, L. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 5990−5995. (27) Mullin, J. W. Crystallization, 4th ed.; Butterworth Heinemann: London, 2001.

AUTHOR INFORMATION

Corresponding Authors

*(T.W.Y.L.) Address: School of Pharmacy, The Chinese University of Hong Kong, Rm 801H, Lo Kwee-Seong Integrated Biomedical Sciences Building, Area 39, Shatin, J

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

(28) Grant, D.; Higuchi, T. Solubility Behavior of Organic Compounds; John Wiley & Sons: New York, 1990; p 656. (29) Shtukenberg, A.; Freundenthal, J.; Gunn, E.; Yu, L.; Kahr, B. Cryst. Growth Des. 2011, 11, 4458−4462. (30) Agarwal, U.; Escobedo, F. A. Nat. Mater. 2011, 10, 230−235. (31) Stevenson, C. L.; Bennett, D. B.; Lechuga-Ballesteros, D. J. Pharm. Sci. 2005, 94, 1861−1880. (32) Zhu, L.; Jona, J.; Nagapudi, K.; Wu, T. Pharm. Res. 2010, 27, 1558−1567. (33) Farrance, O. E.; Jones, R. A. L.; Hobbs, J. K. Polymer 2009, 50, 3730−3738. (34) Shi, F. G.; Tong, H. Y.; Ayers, J. D. Appl. Phys. Lett. 1995, 67, 350. (35) Yoshihashi, Y.; Yonemochi, E.; Maeda, Y.; Terada, K. J. Therm. Anal. Calorim. 2010, 99, 15−19. (36) Six, K.; Berghmans, H.; Leuner, C.; Dressman, J.; Van Werde, K.; Mullens, J.; Benoist, L.; Thimon, M.; Meublat, L.; Verreck, G.; Peeters, J.; Brewster, M.; Van den Mooter, G. Pharm. Res. 2003, 20, 1047−1054. (37) Janssens, S.; Roberts, C.; Smith, E. F.; Van den Mooter, G. Int. J. Pharm. 2008, 355, 100−107.

K

DOI: 10.1021/acs.cgd.6b00342 Cryst. Growth Des. XXXX, XXX, XXX−XXX