Acceleration of the crystal growth rate of low molecular weight organic compounds in supercooled liquids in the presence of polyhydroxybutyrate

Abstract

Modification of the crystallization kinetics of amorphous systems is an important topic, in particular in the area of drug delivery. Herein, we studied the impact of polyhydroxybutyrate (PHB), a biocompatible polymer with a low glass transition temperature (T_g), on crystal growth rates from the supercooled liquid phase of a variety of organic compounds. Low levels of PHB were mixed with the supercooled liquid and growth rates were monitored using optical microscopy at several temperatures above the T_g of the mixture. Growth rates were determined from a plot of crystal size as a function of time. It was found that PHB inhibited the growth of low T_g compounds, while accelerated crystal growth was observed for compounds that had a much higher T_g (>50 °C) than the polymer. The observations were rationalized based on the extent of mobility differences between the crystallizing compound and the polymer. This is the first reported example of a polymer having compound specific inhibitory or accelerator effects on crystal growth.

---

Introduction

Active pharmaceutical ingredients are often dispersed in a polymer matrix for the purpose of enhancing drug delivery. One area of particular interest is amorphous solid dispersions, where the drug is formulated in an amorphous state by mixing at a molecular level with a pharmaceutically acceptable polymer.^1–3 Recent trends of new drug candidates having very low aqueous solubility make amorphous formulations an attractive approach to enhance bioavailability. As the amorphous form is a high energy state, it temporarily exhibits a higher dissolution rate and solubility and thus can enhance the absorption of the drug in vivo.^2,4 However, because of its high energy state, it tends to transform to thermodynamically stable crystalline phase and the advantage of the amorphous form is lost.⁵ Therefore, in order to successfully develop an amorphous formulation, the crystallization of the amorphous compound must be prevented for the shelf life of the product.

Typically, polymers are added to the formulation in order to reduce the rate of crystallization. There are numerous examples demonstrating that improved physical stability against drug crystallization can be achieved through the use of polymers.^1,6–9 The stabilizing mechanism of polymers in terms of preventing crystallization of the drug is currently poorly understood, although several mechanisms have been suggested to be important. These include increasing the glass transition temperature (T_g) of the dispersion and thus reducing the mobility at the storage temperature,¹⁰ acting as a diluent thereby decreasing the chemical potential of the drug, and by forming specific drug–polymer interaction and thus restricting the mobility at a local scale.^6,8,11

For oral drug delivery, drugs are typically dispersed in water soluble polymers in order to ensure rapid release of the drug during the gastrointestinal transit time period. However, drugs may also be dispersed with water-insoluble polymers for more sustained drug delivery from implantable formulations. The physical state of the drug in the formulation will be highly dependent on the production method, and approaches such as melt extrusion are likely to lead to an amorphous formulation. PHB (poly-3-hydroxybutyrate) is liner polyester produced by many strains of bacteria.^12,13 This polymer is semi-crystalline and has numbers of properties comparable to petroleum based plastics, such as high melting temperature (175 °C) and relatively high tensile strength (30–35 MPa).¹⁴ Because of its biodegradability and biocompatibility, PHB has recently attracted interest for medical applications, including as a material for tissue engineering,^15,16 target specific therapy for cancer and tuberculosis^17,18 and nano or micro beads for drug delivery.^19–21 Biodegradable polymers can be loaded with a drug and then used for localized drug delivery whereby the release of the drug can take place over a period of months. As such, the phase behavior of the drug in the matrix is of interest, since changes such as crystallization may impact the release rate. While there have been several studies evaluating the impact of cellulose derivatives and synthetic polymers such as polyvinyl pyrrolidone on the crystal growth rates of organic compounds, PHB does not appear to have been evaluated to date.

In this study, several model drug compounds were mixed with PHB to form amorphous solid dispersions. The crystallization behavior of the compound from the resultant amorphous film was then investigated as a function of temperature. An unexpected acceleration of the compound crystal growth rate was found for some systems. Additional studies were then conducted in order to better understand this phenomenon.

Materials and methods

Acetaminophen (APAP), flutamide (FLU), griseofulvin (GRI), (R)-(−)-3,5-dinitro-N-(1-phenylethyl)benzamide (DPB) and (S)-(−)-2-amino-1,1,3-triphenyl-1-propanol (ATP) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Celecoxib (Cel) was a kind gift from Pfizer Inc. (Groton, CT). Polyhydroxybutyrate (molecular weight 50 000) was purchased from Advanced Polymer Materials (Montreal, QC, Canada). Poly(vinylacetate) (molecular weight 350 000) was purchased from Scientific Polymer Products Inc.

The chemical structures of the compounds and the polymers are given in Fig. 1.

Fig. 1 Chemical structures of (a) GRI, (b) CEL, (c) DPB, (d) ATP, (e) APAP, (f) FLU, (g) PHB and (h) PVAc.

Sample preparation

GRI–PHB mixtures were prepared at a 5% and 20% w/w polymer concentration by first mixing 500 mg of powder in a mortar and pestle and then cryomilling for 10 min in a 6750 freezer mill (Spex Sampleprep, Metuchen, New Jersey). The same method was used to prepare blends of PHB with CEL, APAP and FLU and a blend of PVAc with GRI. DPB–PHB mixtures were prepared at 5%, 10%, 20%, and 50% w/w polymer concentration by first dissolving the compound and polymer in a 1 : 1 mixture of dichloromethane and ethanol followed by solvent removal using rotary evaporation. The resultant powder was lightly triturated in a mortar and pestle. ATP–PHB mixtures were prepared using a similar method.

Growth rate measurements

Crystal growth rate measurements were performed by melting approximately 3 mg of sample between two clean coverslips at around 5 °C above melting point of the compound followed by immediate quenching to room temperature to yield a clear film. The samples were confirmed to be noncrystalline by visual observation under cross-polarized light using a microscope (Nikon Eclipse E600 POL microscope, Nikon Corp, Tokyo, Japan). The samples were subsequently heated to and maintained at different temperatures using a hot stage (Linkam THMS 600, Surrey, UK). The growth of the spherulitic crystals (Fig. 2) was recorded by time lapse photography and the growth rate was determined from the slope of the line obtained by plotting the diameter versus time. The change in diameter with time was found to be linear. All growth rates experiments were performed in triplicate. Two crystal morphologies of APAP were observed in the presence of PHB, but only the growth rate for the same morphology as in the absence of polymer is reported.

Fig. 2 Photomicrographs of FLU without 5% PHB (a) and with 5% PHB (b) at 40 °C and DPB without 5% PHB (c) and with 5% PHB (d) at 100 °C.

Fourier transform infrared (FTIR) spectroscopy

FTIR spectroscopy was performed in transmission mode using a Bio-Rad FTS 6000 (Bio-Rad, Cambridge, MA). The samples were prepared by melting the compound and polymer mixture between a KRS-5 substrate and a glass coverslip, immediately quenching to room temperature followed by removal of the coverslip.

The spectra were then obtained by coadding 128 scans over the wavenumber region of 4000–400 cm⁻¹ while maintaining close to 0% RH with a dry air purge.

Thermal analysis

Thermal analysis of the compound and its mixtures was performed using a differential scanning calorimeter (DSC) (TA Q2000, TA Instruments, New Castle, DE, USA) attached to a refrigerated cooling system. Dry nitrogen was used as the purge gas and was maintained at a flow rate of 50 mL min⁻¹. The instrument was calibrated for temperature using indium and tin, and enthalpy was calibrated using indium. The samples were placed in aluminum pans with a pinhole in the lid. Determination of the glass transition temperature (T_g) was performed by cooling the melt at 10 °C min⁻¹ to at least 50 °C below the T_g, and then reheating at 10 °C min⁻¹. The melting temperature was also measured from physical mixtures of the compound and polymer at a heating rate of 10 °C min⁻¹. The onset glass transition temperature and the peak melting temperatures are reported.

Results

Growth rate kinetics

The growth morphology of flutamide and DPB in the presence and absence of 5% PHB is shown in Fig. 2. The compounds crystallized as spherulites with a well-defined interface with the supercooled liquid, both in the presence and absence of the polymer. The crystal growth rates of each compound alone or in the presence of PHB, as a function of temperature, are shown in Fig. 3 and 4. Growth rate experiments were also conducted with another polymer, PVAc. This polymer was selected since it has a low T_g, and is likely to have weak interactions with the compounds. The impact of PHB addition on crystal growth could be separated into two categories: a) little effect or a slight inhibition of growth rate, b) an increase of growth rate. For FLU, APAP and ATP, the growth rate was either delayed in the presence of PHB, or was essentially the same as for the pure compound, depending on PHB concentration. In contrast, GRI, CEL and DPB showed an increase in crystal growth rate in the presence of PHB relative to the compound alone. The inhibitory effect of PHB on FLU and the acceleratory impact on GRI crystal growth rates are shown in Fig. 5A and B respectively. In general, differences in growth rate in the presence and absence of the polymers were greater for growth rate measurements performed at lower temperatures. PVAc had little impact on the crystal growth rate of GRI.

Fig. 3 Growth rates of FLU, ATP and APAP crystals with and without PHB as a function of temperature showing a small reduction in crystal growth rate in the presence of the polymer. Percentages shown in the legend represent the amount of PHB added to the system. Error bars represent standard deviations, n = 3.

Fig. 4 Growth rates of GRI, CEL, DPB with and without polymer as a function of temperature showing increases in crystal growth rate in the presence of PHB. Percentages represent the amount of polymer added to the system. Error bars represent standard deviations, n = 3.

Fig. 5 A: Change in size of crystals as a function of time at 30 °C for Flu pure and Flu with 5% PHB showing the reduction in growth rate in the presence of the polymer. B: Change in size of crystals as a function of time at 120 °C for Gri pure and Gri with 5% PHB showing the increase in growth rate in the presence of the polymer.

Thermal analysis

Thermal analysis of each compound, physical mixtures of crystalline drug and the polymer, as well as the dispersions was carried out in order to determine the effect of polymers on the glass transition temperature (T_g) and melting temperature (T_m). The values of T_g and T_m are summarized in Table 1. The T_g of PHB, a partially crystalline polymer, was found to be around 0 °C, while PVac had a T_g of around 40 °C. The T_g of compounds ranged from −0.6 °C (FLU) to 89.8 °C (GRI). The T_gs of the mixtures were dependent on the amount of polymer that was present, being unchanged at low amounts of polymer and decreasing as the mass of polymer increased, with the exception of FLU systems, where the T_g was invariant across all compositions (Fig. 5). The melting point of the various compounds was not influenced to a large extent by the presence of the polymers.

Table 1 Glass transition temperatures (T_g) and melting temperatures (T_m) in the presence and absence of the polymers

Sample	T g (°C)	T m (°C)
PHB	0	176
PVAc	40	—
Pure GRI	90	221
GRI 1% PHB	85	220
GRI 5% PHB	78	221
GRI 20% PHB	58	219
GRI 5% PVAc	84	220
Pure APAP	21	171
APAP 5% PHB	22	170
Pure ATP	17	145
ATP 5% PHB	16	146
Pure DPB	53	161
DPB 5% PHB	46	161
DPB 10% PHB	43	160
DPB 20% PHB	35	160
Pure CEL	57	164
CEL 5% PHB	52	163
CEL 20% PHB	41	160
Pure FLU	−1	113
FLU 5% PHB	−2	112
FLU 10% PHB	0	112

---

FTIR spectroscopy

Infrared spectroscopy is a well-established technique for characterizing intermolecular interactions such as hydrogen bonding and has been extensively applied to probe drug–polymer interactions. Fig. 7–10 show the IR spectra of various solid dispersion in the NH stretching region and the C=O stretching region. In the NH stretching region, there was no change in the peak position of the functional groups. In the C=O stretching region, the spectrum of the dispersion represents the simple summation of the spectra of the two pure components. From these results, there seems to be no significant drug–polymer interaction in the solid dispersions prepared in this study. However, it was noted that PHB, a semi-crystalline polymer in the pure state, underwent a considerable change in IR spectrum when dispersed with the drugs (Fig. 6), indicating that it was in amorphous form in the presence of the drug.

Fig. 6 Variation of T_g as a function of polymer concentration.

Fig. 7 Infrared spectra of the DPB/PHB solid dispersion showing the NH stretching region (2800–3600 cm⁻¹, left plot) and the carbonyl stretching region (1550–1800 cm⁻¹, right plot). Percentages represent the weight percentage of PHB in the solid dispersion. Pure amorphous DPB and semicrystalline PHB are shown for reference.

Fig. 8 Infrared spectra of FLU/PHB solid dispersion showing the NH stretching region (2800–3600 cm⁻¹, left plot) and the carbonyl stretching region (1550–1800 cm⁻¹, right plot). Percentages represent the weight percentage of PHB in the solid dispersion.

Fig. 9 Infrared spectra of GRI/PHB solid dispersion showing the NH stretching region (2800–3600 cm⁻¹, left plot) and the carbonyl stretching region (1550–1800 cm⁻¹, right plot). Percentages represent the weight percentage of PHB in the solid dispersion.

Fig. 10 Infrared spectra of CEL/PHB solid dispersion showing the NH stretching region (2800–3600 cm⁻¹, left plot) and the carbonyl stretching region (1550–1800 cm⁻¹, right plot). Percentages represent the weight percentage of PHB in the solid dispersion.

Discussion

The most notable finding of this study is that the polymer, PHB, can either accelerate or inhibit the crystal growth rate in the supercooled liquid depending on the compound evaluated and the concentration of the polymer. These results are of interest, since to date, only one other polymer has been noted to accelerate crystal growth of low molecular weight organic materials,²² and in this instance only an acceleratory effect was noted rather than the variation in behavior that we observed in this study. The crystal growth rate (u) from the supercooled liquid is expected to depend on both thermodynamic and kinetic factors as shown by eqn (1).²³where ΔG is the free energy difference between the liquid and crystalline forms, k is a temperature independent constant, η is the viscosity, R is the gas constant and T is temperature. From eqn (1), it is apparent that the two factors that might be influenced by polymeric additives, and hence have an impact on crystal growth, are the thermodynamic driving force for phase transformation, and the viscosity of the liquid. These two factors give rise to the “bell” shaped curve typically observed for crystal growth as a function of temperature.²⁴ At temperatures close to the melting point, the degree of undercooling, T_m–T, which represents the thermodynamic driving force for phase transformation, is small while the viscosity of the liquid is low and hence mass transport is rapid. Thus the growth rate is under thermodynamic control. At lower temperatures, the undercooling is high but the viscosity increases, and mass transport is the limiting factor for crystal growth. A maximum in crystal growth rate typically occurs at a temperature equivalent to 0.94 × T_m.²⁵ For the growth data shown in Fig. 3 and 4, the growth rate is expected to be dependent on the mass transport rate of the drug to the crystal surface since the experiments were conducted at larger undercoolings, below the temperature at which the maximum crystal growth rate occurs. In addition to the thermodynamic and kinetic parameters, drug–polymer specific interactions are also considered to play a role in determining the effectiveness of a given polymer at inhibiting crystal growth. Thus small amounts of polymer which do not have significant effects on T_g or viscosity have been found to considerably delay crystal growth.^1,26–28 Interesting, in this study, the opposite phenomenon has been observed, whereby small amounts of the polymer, PHB, are found to accelerate the crystal growth rate of some compounds. For example, a PHB level of only 5% increases the growth rate of griseofulvin crystals by a factor of about 10 at 120 °C. Eqn (1) serves as a basis to try understand how the presence of PHB might result in an acceleration of crystal growth. First, we can confirm that the polymer has little impact on the thermodynamic driving force for crystallization. This can be inferred based on the lack of melting point depression (Table 1) observed in the presence of the polymer. Hence the degree of supercooling (T_m–T) was not substantially affected by the presence of the polymer. This observation is in accordance with the IR spectroscopy results which showed no specific interactions between the drug and polymers; specific interactions would be expected to depress the melting point. Hence, the polymers most likely impact the growth rate through impacting the mobility of the crystallizing solute in the supercooled liquid. Variations in T_g as a function of polymer composition support the conjecture that changes in viscosity play a role in the observed changes in crystal growth rates. Variations in T_g have been used in the polymer field to assess the impact of polymeric additives on the growth rate. Wu et al.²⁹ reported that to a first approximation, the growth rates of polyoxyethylene in the presence and absence of different polymeric additives could be superimposed when normalized to both the degree of supercooling and the change in the glass transition temperature (T–T_g, taken as an indicator of a change in the mass transport properties) where T is the crystallization temperature. Fig. 11 shows the growth rate data of DPB, GRI and ATP with various concentrations of PHB vs. (T–T_g) in an attempt to account for mass transport effects. At a concentration of 5% or 10% PHB, the growth rates of the crystals from the systems containing polymer become very similar to the growth rates of the pure compounds. At a concentration of 20% PHB, or for the PVAc system, however, the growth rate is lower than would be predicted after accounting for changes in T_g.

Fig. 11 Growth rates of DPB, GRI and CEL with different amounts of added PHB as a function of T–T_g. At low polymer concentrations (5 and 10%), the data tend to collapse onto the same line. At a higher polymer concentration (20%), the crystal growth rate is lower than would be predicted based on changes in T_g.

It is apparent from the data shown in Fig. 3 and 4 that compounds with a higher T_g appear to show an acceleration of growth rate in the presence of PHB, while the compounds with a lower T_g do not show this phenomenon. The growth rate data was further evaluated by plotting the ratio of the growth rate in the presence of polymer to that in the absence of polymer, as a function of compound T_g in order to further explore this relationship, with results shown in Fig. 12. This plot is very revealing regarding the impact of PHB on crystal growth rates, in that it is clear that the polymer can accelerate or retard the crystal growth rate in a manner that depends on the T_g of the compound evaluated. It should be noted that the growth rates were evaluated from the liquid at temperatures 30–40 °C above the compound T_g. Yu and coworkers evaluated the impact of low levels of polymers with different pure T_g values on the crystal growth rates of nifedipine in both the glass and liquid states. They found that the polymer with the lowest T_g, polyethylene glycol, (PEG) resulted in growth acceleration, whilst all other polymers either had minimal impact or retarded growth.²² A log–linear relationship between nifedipine crystal growth rate and polymer T_g was observed in their study. The acceleration or retardation of the growth rate by different polymers was attributed to polymer mobility whereby a low T_g polymer is expected to have a high segmental mobility and vice versa. Polymer mobility in turn is postulated to impact the ability of molecules of the crystallizing component to join the crystal resulting in the observed growth rate acceleration by the lowest T_g polymer, PEG. The authors further pointed out the effectiveness of the polymer on modifying the solvent mobility (herein the low molecular weight organic compound is considered as analogous to a solvent) should depend on the difference between the mobility of the polymer and the host molecules. Thus the observations of the current agreement are in broad agreement with the findings and explanation given by Yu and coworkers. The difference in mobility between PHB and the model drugs is greatest for the high T_g compounds, leading to an acceleration of growth rate. As the mobility difference between PHB and the drug decreases, as for APAP, ATP and FLU, all of which have much lower T_g's, the polymer no longer accelerates crystal growth and begins to have a retardation effect. This retardation effect most likely stems from the polymer molecules physically impeding transport of the crystallizing molecules to the crystal surface.

Fig. 12 Ratio of growth rate in the presence of the polymer to that in the absence of the polymer versus the T_g of the pure compound. The growth rate values are for analysis temperatures 30–40 °C above the compound T_g.

Interestingly, PVAc did not exhibit the same acceleratory effect on the growth rate of GRI. This drug–polymer combination has a T_g difference of around 50 °C, which is comparable to that observed for PHB and CEL, however in the latter system the polymer was found to accelerate the growth rate. Clearly T_g is only a crude indicator of mobility and more sophisticated indicators²² are likely to provide greater mechanistic insight into the observed differences between the two polymers.

Conclusions

A small amount of polyhydroxybutyrate was found to have a different effect on crystal growth rates from the supercooled liquid depending on the compound investigated. Compounds with T_g's much higher than that of the polymer (>50 °C higher) showed acceleration of crystal growth in the presence of the polymer suggesting that the polymer increased the molecular mobility of the crystallizing compound. In contrast, for compounds with T_g's closer to the T_g of the polymer, the polymer was able to act as a crystal growth inhibitor, reducing the transport rate of molecules to the growing crystal. This is the first reported example of a polymer having compound specific inhibitory or accelerator effects on crystal growth. These observations are of importance in the context of the stabilization of organic compounds to crystallization for applications such as drug delivery.

Acknowledgements

We acknowledge the financial support provided by Kyowa Hakko Kirin Co., Ltd. (Tokyo, Japan). Dr. N. S. Trasi is thanks for helpful discussions.

References

1. T. Matsumoto and G. Zografi, Pharm. Res., 1999, 16, 1722–1728.
2. W. L. Chiou and S. Riegelman, J. Pharm. Sci., 1971, 60, 1281–1302.
3. Y. Kawabata, K. Wada, M. Nakatani, S. Yamada and S. Onoue, Int. J. Pharm., 2011, 420, 1–10.
4. B. C. Hancock and G. Zografi, J. Pharm. Sci., 1997, 86, 1–12.
5. V. Andronis and G. Zografi, J. Non-Cryst. Solids, 2000, 271, 236–248.
6. L. Taylor and G. Zografi, Pharm. Res., 1997, 14, 1691–1698.
7. K. Khougaz and S. D. Clas, J. Pharm. Sci., 2000, 89, 1325–1334.
8. T. Miyazaki, S. Yoshioka, Y. Aso and S. Kojima, J. Pharm. Sci., 2004, 93, 2710–2717.
9. B. Van Eerdenbrugh and L. S. Taylor, Mol. Pharmaceutics, 2010, 1058–1066.
10. G. Van den Mooter, M. Wuyts, N. Blaton, R. Busson, P. Grobet, P. Augustijns and R. Kinget, Eur. J. Pharm. Sci., 2001, 12, 261–269.
11. U. S. Kestur and L. S. Taylor, CrystEngComm, 2010, 12, 2390.
12. H.-M. Müller and D. Seebach, Angew. Chem., Int. Ed. Engl., 1993, 32, 477–502.
13. M. Lemoigne, Bull. Soc. Chim. Biol., 1926, 8, 770–782.
14. V. Thakur and M. Thakur, Handbook of sustainable polymers: Processing and applications, 2016.
15. L. Ricotti, A. Polini, G. G. Genchi, G. Ciofani, D. Iandolo, H. Vazão, V. Mattoli, L. Ferreira, A. Menciassi and D. Pisignano, Biomed. Mater., 2012, 7, 035010.
16. E. I. Shishatskaya, T. G. Volova, A. P. Puzyr, O. A. Mogilnaya and S. N. Efremov, J. Mater. Sci.: Mater. Med., 2004, 15, 719–728.
17. A. Althuri, J. Mathew, R. Sindhu, R. Banerjee, A. Pandey and P. Binod, Bioresour. Technol., 2013, 145, 290–296.
18. N. A. Parlane, K. Grage, J. Mifune, R. J. Basaraba, D. N. Wedlock, B. H. A. Rehm and B. M. Buddle, Clin. Vaccine Immunol., 2012, 19, 37–44.
19. K. Grage, A. C. Jahns, N. Parlane, R. Palanisamy, I. A. Rasiah, J. A. Atwood and B. H. A. Rehm, Biomacromolecules, 2009, 10, 660–669.
20. A. C. Kassab, K. Xu, E. B. Denkbaş, Y. Dou, S. Zhao and E. Pişkin, J. Biomater. Sci., Polym. Ed., 1997, 8, 947–961.
21. E. I. Shishatskaya, A. V. Goreva, O. N. Voinova, E. V. Inzhevatkin, R. G. Khlebopros and T. G. Volova, Bull. Exp. Biol. Med., 2008, 145, 358–361.
22. C. T. Powell, T. Cai, M. Hasebe, E. M. Gunn, P. Gao, G. Zhang, Y. Gong and L. Yu, J. Phys. Chem. B, 2013, 117, 10334–10341.
23. P. C. Hiemenz and T. P. Lodge, Polymer chemistry, CRC press, 2007.
24. S. Umemoto and N. Okui, Polymer, 2002, 43, 1423–1427.
25. K. Naito, Chem. Mater., 1994, 6, 2343–2350.
26. U. Kestur and L. S. Taylor, CrystEngComm, 2010, 14, 1691–1698.
27. B. Van Eerdenbrugh and L. S. Taylor, CrystEngComm, 2011, 13, 6171.
28. K. Kothari, V. Ragoonanan and R. Suryanarayanan, Mol. Pharmaceutics, 2015, 12, 162–170.
29. L. Wu, M. Lisowski, S. Talibuddin and J. Runt, Macromolecules, 1999, 32, 1576–1581.

---