- Email: [email protected]
Drying of crystalline drug nanosuspensions—The importance of surface hydrophobicity on dissolution behavior upon redispersion
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
Drying of crystalline drug nanosuspensions—The importance of surface hydrophobicity on dissolution behavior upon redispersion Bernard Van Eerdenbrugh a , Ludo Froyen b , Jan Van Humbeeck b , Johan A. Martens c , Patrick Augustijns a , Guy Van den Mooter a,∗ a b c
Laboratory for Pharmacotechnology and Biopharmacy, K.U. Leuven, Gasthuisberg O&N2, Herestraat 49, Box 921, 3000 Leuven, Belgium Metallurgy and Materials Engineering Department, K.U. Leuven, Kasteelpark Arenberg 44, 3001 Leuven, Belgium Center for Surface Chemistry and Catalysis, K.U. Leuven, Kasteelpark Arenberg 23, 3001 Heverlee, Belgium
a r t i c l e
i n f o
a b s t r a c t
Article history:
d-␣-Tocopherol polyethylene glycol 1000 succinate (TPGS)-stabilized nanosuspensions
Received 15 April 2008
(25 wt%, relative to the drug weight) were produced by media milling for 9 model drug
Received in revised form
compounds [cinnarizine, griseofulvin, indomethacin, itraconazole, loviride, mebendazole,
20 May 2008
naproxen, phenylbutazone and phenytoin]. After 3 months of storage at room temperature,
Accepted 23 June 2008
Ostwald ripening occurred in all of the samples, except for indomethacin. Whereas lowering
Published on line 3 July 2008
the temperature could slow down the ripening, it markedly increased upon storage at 40 ◦ C. As for ripening, settling generally became more pronounced at 40 ◦ C compared to 4 ◦ C. As the
Keywords:
nanosuspensions were afflicted by Ostwald ripening and settling, we explored nanosuspen-
Nanosuspensions
sion drying as a strategy to circumvent these stability issues. Spray-drying and freeze-drying
Stability
were evaluated for nanosuspensions and coarse reference suspensions of the compounds.
Spray-drying
Nanoparticle agglomeration could be visually observed in all of the powders. To evaluate
Freeze-drying
the effect of agglomeration on the key characteristic of drug nanocrystals (i.e. rapid dissolution), dissolution experiments were performed under poor sink conditions. It was found that the compounds could be categorized into 3 groups: (i) compounds for which it was impossible to differentiate between coarse and nanosized products (griseofulvin, mebendazole, naproxen), (ii) compounds that gave clear differences in dissolution profiles between the nanosized and the coarse products, but for which drying of the nanosuspensions did not decrease the dissolution performance of the product (indomethacin, loviride, phenytoin) and (iii) compounds that showed differences between coarse and nanosized products, but for which drying of the nanosuspensions resulted in a significant decrease of the dissolution rate (cinnarizine, itraconazole, phenylbutazone). To gain insight on the influence of the drug compound characteristics on the dissolution of the dried products, the dissolution behavior of the compounds of the second and the third group was linked to the compound’s characteristics. It was found that compounds with a more hydrophobic surface resulted in agglomerates which were harder to disintegrate, for which dissolution was compromised upon drying. The same was found for compounds having higher log P values. © 2008 Elsevier B.V. All rights reserved.
∗
Corresponding author. Tel.: +32 16330304; fax: +32 16330305. E-mail address: [email protected] (G. Van den Mooter). 0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2008.06.009
128
1.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Introduction
Oral drug delivery is the preferred way of drug administration, since it avoids the pain and discomfort associated with injections and is more attractive from a marketing and patient compliance perspective (Fasano, 1998). Fuelled by the fact that, over the last decade, new drug candidates have become more hydrophobic and less water-soluble (Lipinski, 2002), designing an adequate oral dosage form is becoming challenging and formulation scientists frequently have to consider more complex drug delivery platforms including solid dispersions, cyclodextrins, salt formation and lipid dosage forms. Within this scaffold of drug delivery tools, formulation of drugs as nanocrystals, submicron-sized drug crystals, has rapidly grown to a mature drug delivery strategy during the last decade, with currently 5 products on the market. For the manufacture of nanocrystals, media milling is currently the most popular approach (Liversidge et al., 1991). The major advantage of nanocrystals for oral delivery is the increase in the dissolution velocity and hence oral absorption, based on the increased specific surface area of the particles. Also often advantages may accompany this formulation approach, such as reduced fasted/fed state variability and ease of administration (Kesisoglou et al., 2007). The particle size reduction processes used to produce nanocrystals (media milling and high pressure homogenization) are conducted in the suspended state, hence forming a nanosuspension. However, solid dosage forms are considered more attractive, due to their convenience (marketing aspects) and possible stability issues associated with nanoparticles in their suspended state. These stability issues can be both physical (e.g. Ostwald ripening and agglomeration) and chemical (e.g. hydrolysis), although examples exist for which formulation as a nanosuspension actually prevents the latter, compared with formulation as a solution (e.g. Möschwitzer et al., 2004; Merisko-Liversidge and Linden, 2003). Technically, transformation of nanosuspensions into solid products can be achieved using established unit-operations such as freeze-drying, spray-drying, pelletization and granulation (Müller et al., 2006). Often matrix formers are added to the suspensions prior to the drying step, in order to achieve adequate redispersion in water. Whatever technique used for this transformation step, water removal occurs either by evaporation or sublimation. In contrast with the importance of this transformation step, the current available literature is scarce (e.g. Müller et al., 2006; Vergote et al., 2001; Lee, 2003; Möschwitzer and Müller, 2006; Van Eerdenbrugh et al., 2007). Furthermore, reports addressing the key attribute of a nanocrystal system intended for oral delivery, i.e. preservation of the beneficial dissolution characteristics of the nanosuspension upon redispersion, are even more limited (e.g. Van Eerdenbrugh et al., 2007). To date, to the best of our knowledge, the influence of the drug compound on the dissolution characteristics after drying of nanosuspensions has never been investigated. The aim of this study is to evaluate the effect of drying on the dissolution performance of drug nanosuspensions. Therefore, a screening approach was followed, using 9 different model drug compounds. The structures and abbreviations of
the model compounds are provided in Fig. 1. The stabilizing system chosen for this study was 25 wt% d-␣-tocopherol polyethylene glycol 1000 succinate (TPGS) relative to the drug weight, as a previous study proved that nanosuspension production was feasible for all drug compounds using this stabilizing system (submitted manuscript). The first part of the study consists of a brief description of short-term physical stability of the nanosuspensions (Ostwald ripening and agglomeration). Ostwald ripening and, in some cases, agglomeration, further stresses on the importance of the second part, a study on drying of the nanosuspensions. Both evaporation and sublimation techniques were evaluated, by using spray-drying and freeze-drying, respectively. No matrix formers were added prior to drying as the aim and focus of the present study is to gather insight on the influence of the drug compounds’ characteristics on the dissolution characteristics of the dried products. The resulting powders were evaluated using dissolution, as rapid dissolution can be considered as the main advantage of a nanosized system intended for oral delivery. Finally, the dissolution behavior of the products was related to the compounds’ characteristics.
2.
Materials and methods
2.1.
Materials
d-␣-Tocopherol polyethylene glycol 1000 succinate (TPGS, Eastman Chemical Company, Kingsport, TN, USA) was a gift from the manufacturer. Loviride and itraconazole were kindly provided by Johnson & Johnson Pharmaceutical Research and Development (Beerse, Belgium). Cinnarizine (Certa n.v., Braine-l’Alleud, Belgium), griseofulvin (Certa n.v.), indomethacin (Certa n.v.), mebendazole (Certa n.v.), naproxen (Certa n.v.), phenylbutazone (Fagron NV, Waregem, Belgium), phenytoin (Fagron NV), acetonitrile gradient grade far UV (Fisher Scientific UK Limited, Loughborough, UK), dimethylsulfoxide (Acros Organics, Geel, Belgium), methanol 99.9% for HPLC gradient grade (Acros Organics), potassium dihydrogen phosphate GR for analysis (Merck KGaA, Darmstadt, Germany), dipotassium hydrogen phosphate trihydrate GR for analysis (VWR International bvba/sprl, Leuven, Belgium), sodium dodecyl sulfate (SLS, Merck Schuchardt OHG, Hohenbrunn, Germany), tetrabutyl ammonium hydrogen sulfate (TBAH, Acros Organics), 1 M HCl (Titrinorm® , VWR International, Fontenay Sous Bois, France), yttrium-stabilized zirconia beads (diameter 0.5 mm, YTZ® grinding media, Tosoh Corporation Advanced Ceramics Department, Tokyo, Japan) were obtained commercially. Demineralized water was used for all formulations (Elga, maxima ultra pure water, ≥18 M).
2.2.
Nanosuspension production
Nanosuspensions were prepared using media milling. Suspensions of 1 g of drug compound in 5 ml water were made in 10 ml vials using 250 mg of TPGS as a stabilizer. Subsequently, 15 g of zirconia beads (Ø 0.5 mm) were added as a milling agent and vials were placed in self-designed nylon holders (6 vials per holder) that fitted into 500 ml mixing bowls.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
129
Fig. 1 – Chemical structures of the drug compounds used in the study, with the abbreviations as used throughout the manuscript.
The four mixing bowls were then placed into the grinding stations of a planetary mill (Retsch PM 400 MA, Retsch GmbH, Haan, Germany) and milling was performed at 250 rpm for 24 h. Six batches, each containing 1 gram of drug, were prepared for each compound. After milling, batches were pooled and diluted with water to a concentration of 20 mg/ml, after content determination by HPLC (Section 2.3).
2.3.
Quantitative assays by HPLC
For the determinations of the drug content in the nanosuspensions (Section 2.2) and the dried products (Section 2.7), the determination of SLS solubilization for the design of relevant dissolution conditions (Section 2.8.1), and the analysis of samples in the dissolution experiments (Section 2.8.2), HPLC-UV was used. Samples were analyzed, after dilution into the validated range, on a Merck-Hitachi-Lachrom instrument (Hitachi Ltd., Tokyo, Japan) using a Merck Chromolith Performance RP18-e (100–4.6 mm) column (Merck, Darmstadt, Germany).
Each determination was performed in triplicate and the average and standard deviation was calculated.
2.4. (DLS)
Size determination by dynamic light scattering
For the measurement of the particle size of the nanosuspensions, dynamic light scattering was used. DLS was performed using a Nanophox instrument in autocorrelation mode (SympaTec GmbH, Clausthal-Zellerfeld, Germany). Prior to measurement, samples were diluted with a saturated solution of the compound under investigation. Saturated solutions were prepared by filtration of a drug suspension, stirred overnight, through a 0.45 m nylon 47 mm membrane filter (Millipore Corporation, Bedford, MA, USA). For griseofulvin, a saturated solution of the drug containing 5 mg/ml TPGS was used, since severe agglomeration occurred upon dilution in a saturated solution without TPGS. Detection was carried out at a scattering angle of 90◦ , sample temperature was set at
130
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
25 ◦ C and 3 runs of 30 s were performed on each sample. From the resulting correlation curves, a 2nd order analysis was performed to calculate the mean particle size and standard deviation.
2.5.
Stability evaluation of the nanosuspensions
Three temperature conditions were applied in the stability study of the nanosuspensions: 4 ◦ C (refrigerator), ambient conditions, and 40 ◦ C (thermostatic oven). Physical stability of the nanosuspensions was evaluated after 3 months of storage in sealed vials. Particle size measurements were performed by DLS (Section 2.4) and settling behavior by visual examination. Particle size results were expressed relative to the mean particle size obtained after production (%).
2.6. Freeze-drying and spray-drying of the nanosuspensions 2.6.1.
Freeze-drying
20 ml of the 20 mg/ml nanosuspensions corresponding to 400 mg of drug compound was transferred into a glass vial. Subsequently, the suspensions were frozen by immersing the vials in liquid nitrogen. Freeze-drying of the products was performed with a Christ model Alpha freeze-dryer (type 1050, Van Der Heyden, Brussels, Belgium) at a shelf temperature of −50 ◦ C with a pressure below 1 mbar and the vials were removed after 48 h of drying. After freeze-drying, residual water was removed by placing the vials overnight in a vacuum oven at 25 ◦ C (Mazzazi Systems, Italy). The same procedure was followed on 50 ml of non-nanosized suspensions, corresponding to 1 g of drug compound.
2.6.2.
Spray-drying
Spray-drying was performed with a Büchi Mini Spray Dryer B191 (Flawil, Switzerland). 200 ml of nanosuspension (4 g of drug compound) was spray-dried using the following process parameters: the inlet temperature was 120 ◦ C, the aspirator air flow were set at 100%; the pump speed was 40%. During spray-drying the suspension was continuously stirred with a magnetic stirrer. The spray-dried solid product was isolated and subsequently dried overnight in a vacuum oven at 25 ◦ C (Mazzazi Systems). The same procedure was followed on 300 ml of non-nanosized suspensions, corresponding to 6 g of drug compound.
2.7.
in the filtrate of an equilibrated aqueous drug suspension. Therefore, an excess of drug compound was added to a test tube containing 10 ml of 0.5, 1 or 2% (w/v) SLS and 0.01N HCl in water. For cinnarizine, a medium without HCl was used, as pronounced agglomeration and irreproducible results were observed upon using HCl in the medium. After equilibration overnight on a horizontal shaker (Köttermann type 4020, Haenigsen, Germany), 5 ml of suspension was filtrated through a 0.1 m PTFE syringe filter (Whatman Inc., Clifton, NJ, USA). Subsequently, the drug content in the filtrate was assessed by HPLC (Section 2.3). Each experiment was performed in triplicate and the average and standard deviations were calculated. The SLS concentration for the dissolution experiments, corresponding to a drug compound solubility of 0.375 mg/ml was determined from the linear fit of the concentration of solubilized compound as a function of SLS concentration, taking into account the 0.5, 1 and 2% (w/v) data points. All linear fits resulted in a R2 value higher than 0.995.
Determination of drug content
Drug content in the obtained powders was determined by dissolving about 10 mg of powder in 10 ml of DMSO. Samples were analyzed by HPLC-UV (Section 2.3) from which the drug content [% (w/w)] was determined. The procedure was performed in triplicate and the average and standard deviations were calculated.
2.8.
Design of dissolution experiments
2.8.1.
Determination of relevant dissolution conditions
2.8.2.
3.
Results and discussion
It should be pointed out that, to conduct a study on nanosuspension drying for a set of 9 drug compound, one first has to identify a stabilizing system that is able to yield nanosuspensions for all of the compounds. The choice of TPGS as a stabilizing system was based on a screening study, using 13 different stabilizers in 3 concentrations for these drug compounds. In this study, only TPGS in 25 or 100 wt%, relative to the drug weight, proved to be able to yield nanosuspensions for all of the drug compounds (submitted manuscript). The 25% concentration was selected to minimize possible effects of the stabilizer on the products’ constitution after drying.
3.1. Micellar solubilization of the drug compound by SLS at ambient conditions was performed by analyzing the drug content
Dissolution experiments
Dissolution experiments were performed at room temperature using a miniaturized design. An amount of formulation corresponding to 5 mg of drug compound (0.25 mg/ml) was transferred in a 20 ml disposable syringe, together with an Al2 O3 ball of 10.5 mm diameter (to provide hydrodynamics during the experiment). Subsequently, 20 ml of dissolution medium was gently sucked into the syringe, a 0.1 m PTFE syringe filter (Whatman Inc.) was mounted on the syringe and the syringe was fixed into a rotary mixer (Labinco BV, the Netherlands), making 15 rotations/min. After 1.5, 5, 15, 30, 60, 90 and 120 min, 300 l of medium was filtrated, from which the initial 100 l was discarded and the remaining fraction was collected in a 1.5 ml eppendorf tube (Eppendorf, Hamburg, Germany). The exact concentration corresponding to 100% release was determined by assaying the total (dissolved + undissolved) drug concentration in the suspension after the experiment. Samples were analyzed by HPLC (Section 2.3). Each dissolution experiment was performed in triplicate and the average values and standard deviations were calculated.
Physical stability of drug nanosuspensions
Particle size data of the nanosuspensions after production and physical stability data after 3 months
131
e
d
c
b
DLS analysis of the PHB nanosuspension resulted in correlation curves with shapes that deviated significantly from the theoretical fitting. SEM indicated that the particles had a large aspect ratio, which can explain this phenomenon. Hence, DLS evaluation of the stability samples is of limited value. T/C indicates a nanosuspension that shows settling, but is still turbid. C 0.9 indicates a settled nanosuspension for which about 90% of the total volume was clear. T indicates a completely turbid nanosuspension. C 0.5 indicates a settled nanosuspension for which about 50% of the total volume was clear. a
T/Cb T/Cb C 0.5e T/Cb T/Cb T/Cb Td T/Cb Td T/Cb Td C 0.5e Td T/Cb T/Cb T/Cb Settling (visual)
4 ◦C 40 ◦ C
C 0.9c C 0.9c
106 ± 5% 107 ± 4% 122 ± 8% –a –a –a 107 ± 5% 110 ± 3% 146 ± 6% 113 ± 3% 193 ± 3% 267 ± 11% 103 ± 2% 109 ± 1% 139 ± 2% 98 ± 2% 115 ± 2% 140 ± 2% 97 ± 2% 100 ± 1% 122 ± 2% 106 ± 5% 111 ± 2% 223 ± 17% 110 ± 4% 113 ± 4% 120 ± 5% 4 ◦C Ambient 40 ◦ C Stability data (3 months) Relative size (DLS)
406 ± 17 498 ± 10 288 ± 4% 190 ± 2 156 ± 2 220 ± 4
Phenytoin Phenylbutazone Naproxen Mebendazole Loviride Itraconazole Indomethacin
193 ± 2 256 ± 1 366 ± 12 Mean (nm)
To study the effect of drying on the nanosuspensions, water was eliminated from the nanosuspensions by evaporation (spray-drying) and sublimation (freeze-drying), since these processes form the elementary mechanisms of water removal in all unit-operations that could be typically applied for nanosuspension drying. Drug content of the powders after drying as well as drug yield of the spray-drying process for the different compounds are summarized in Table 2. A general observation is that the drug content in the powders is slightly lower than the expected 80% (w/w) (as TPGS was added as 25% of the drug weight). An exception to this are the spray-dried coarse suspensions, that always resulted in content values above 80% (w/w) (88 ± 5% on average). This was due to a pronounced loss of TPGS in the instrument’s cyclone, as witnessed by the formation of a viscous liquid that solidified upon cooling to room temperature in the cyclone during spray-drying. Agglomeration could be visually observed after drying of all nanosized products. Although there is no consensus on the mechanism leading to nanoparticle agglomeration, the capillary pressure theory is currently the main theory to explain agglomeration during drying (Wang et al., 2005). According to
Starting suspension Size by DLS
Drying of the drug nanosuspensions
Griseofulvine
3.2.
Cinnarizine
of storage of the nanosuspensions are summarized in Table 1. The first part of the evaluation of the physical stability is the investigation of Ostwald ripening. This process by which smaller particles are consumed in the growth of larger particles originates from the enhanced solubility arising from the higher curvature of smaller particles, as described by the Ostwald-Freundlich equation (Grant and Brittain, 1995). As can be seen from the table, Ostwald ripening can be observed at room temperature in all of the nanosuspensions, except for indomethacin. Whereas storage at lower temperature (4 ◦ C) could slow down (cinnarizine, griseofulvin, loviride, mebendazole, naproxen, phenytoin, although not always significantly) or even stop (ITR) the ripening, it markedly increased at higher temperature (40 ◦ C), most pronounced for griseofulvin and mebendazole. Although temperature has been described to exert its effect on Ostwald ripening through a complex combination of effects (Madras and McCoy, 2003), it clearly increases with temperature for the systems studied. As for ripening, settling generally increased at an elevated temperature of 40 ◦ C (except for phenylbutazone). Whereas almost half of the suspensions stayed completely turbid at 4 ◦ C (indomethacin, itraconazole, loviride, mebendazole), at least some settling could be observed in all of the samples at 40 ◦ C. The severe settling behavior of griseofulvin is most likely due to agglomeration caused by dilution of the nanosuspension to 20 mg/ml prior to storage, as upon dilution, settling occurred within hours at ambient conditions, whereas this was not the case in the undiluted nanosuspension. The above results clearly indicate that physical stability issues (Ostwald ripening and agglomeration) occur in the nanosuspensions studied. They advocate further transformation of these systems into solid powders, in order to obtain a product with adequate physical stability.
Table 1 – Size of the nanosuspensions after production (DLS), relative size after 3 months of storage at 4 ◦ C, room temperature and 40 ◦ C and settling behavior after 3 months of storage at 4 and 40 ◦ C
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
132
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Table 2 – Drug content of the spray-dried and freeze-dried Content [% (w/w)] Spray-dried Nano ± ± ± ± ± ± ± ± ±
Cinnarizine Griseofulvin Indomethacin Itraconazole Loviride Mebendazole Naproxen Phenylbutazone Phenytoin
75 74 74 77 79 81 73 74 75
1 1 1 1 1 1 0 0 1
Overall
76 ± 3
Freeze-dried
Coarse 80 85 88 83 94 85 94 88 92
± ± ± ± ± ± ± ± ±
1 1 1 1 2 2 3 1 1
88 ± 5
Nano 83 73 78 75 80 79 77 67 66
± ± ± ± ± ± ± ± ±
Coarse
1 2 1 0 1 1 1 1 2
75 ± 6
84 74 81 78 78 79 77 74 71
± ± ± ± ± ± ± ± ±
1 3 2 2 3 2 1 0 2
77 ± 4
this theory, agglomeration occurs as a result of the capillary forces encountered during the drying process.
3.3.
Fig. 2 – Dissolution curves of mebendazole. Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
Dissolution of the dried products
Since the main attribute of a nanosuspension is the increased dissolution rate resulting from the high specific surface area of the particles, dissolution of the powders was performed as a pharmaceutical evaluation. It is important to realize that nanoparticles agglomeration is likely to be inevitable during drying of a nanosuspension to which no additional matrix formers are added. From the point of dissolution performance, the key question is therefore not whether nanoparticle agglomeration will occur, rather the central question is how easily the agglomerates break up during dissolution. In other words, will desagglomeration become a rate-controlling step in the overall dissolution process? To provide an answer to this question, poor sink conditions were designed for each of the compounds in which the solubility of the compound was 150% (0.375 mg/ml) of the concentration upon complete dissolution (0.250 mg/ml). Dissolution was performed on nanosized and coarse products, in the suspended state and on spray-dried and freeze-dried powders. Upon analyzing the dissolution curves of the different compounds, three distinct groups of profiles were found. In the first group (griseofulvin, mebendazole, naproxen), no differences
could be observed between both the nanosized and the nonnanosized products, as illustrated in Fig. 2 for mebendazole. Since the dissolution medium was not able to discriminate between nanosized and non-nanosized products, these compounds were disregarded for further evaluation. All compounds of the second and third group showed clear differences between nanosized and non-nanosized products. The second group (indomethacin, loviride, phenytoin), for which dissolution profiles are given in Fig. 3A–C are characterized by the fact that spray-drying and freeze-drying of the nanosuspensions did not compromise the dissolution rate to a significant extent. This is an indication that the agglomerates easily break up upon redispersion in water. For these powders, addition of a matrix former prior to drying is unnecessary from the point of view of preservation of the dissolution rate of the nanoparticles. For the third group (cinnarizine, itraconazole, phenylbutazone), on the other hand, dissolution rate clearly decreased in the dried nanosized products (Fig. 4A–C). The agglomerates formed upon drying are apparently so strong that their disintegration compromises the overall dissolution process. Therefore, addition of matrix formers prior to drying is nec-
Table 3 – Key physicochemical properties of the model compounds Compound CIN GRI IND ITR LOV MEB NAP PHB PHT a b
Molecular weight (g/mol)
Melting point (onset, ◦ C)
369 353 358 706 351 295 230 308 252
119.5 ± 0.1 220a 159.9 ± 0.08 165.2 ± 0.1 226.3 ± 0.1 289a 153a 106.0 ± 0.1 295.03 ± 0.48
log P (calculated) 6.1 1.1 3.3 8.5 3.7 3.3 3.0 4.2 2.3
Aqueous solubility (g/ml) <2b 7.7 ± 0.2 16 ± 1 <0.4b <2b 0.3 ± 0.0 31 ± 2 24 ± 0 17 ± 1
Values were taken from the Merck Index as the compound showed polymorphism and/or degradation upon heating. Below the limit of quantification.
Density (g/ml) 1.13 1.45 1.37 1.35 1.44 1.38 1.24 1.19 1.26
± ± ± ± ± ± ± ± ±
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Fig. 3 – Dissolution curves of indomethacin (A), loviride (B), and phenytoin (C). Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
essary in this case to maintain rapid dissolution of the drug. The performance of various matrix formers is the subject of a separate manuscript.
3.4.
Influence of the drug compound
An important trend that can be seen in the dissolution results provided in Figs. 3 and 4 is that spray-drying and freeze-drying always resulted in very similar dissolution profiles. This suggests that the agglomeration of the nanoparticles is governed by the compounds characteristics rather than by the dry-
133
Fig. 4 – Dissolution curves of cinnarizine (A), itraconazole (B), and phenylbutazone (C). Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
ing procedure applied. Agglomeration tendency and strength can be expected to be a result of the surface properties of the compound, more precisely the surface hydrophobicity of the particles. In a previous study (submitted manuscript), surface hydrophobicity of the nanosuspensions of these compounds was probed by determination of the amount of TPGS adsorbed per unit of surface area, based on the fact that the hydrophobic d-␣-tocopherol part of TPGS can be expected to adsorb on the nanoparticle surfaces. The outcome of this study indicated that nanosuspension stabilization, i.e. the prevention of agglomeration, was easier for less hydrophobic
134
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Fig. 6 – log P as a function of TPGS adsorption for the compounds of the first (), second () and third () group.
Fig. 5 – Release after 15 min as a function of (A) TPGS adsorption per unit of surface area and (B) log P of the compounds of the second and third group. Release values are the averages of the spray-dried and freeze-dried nanosuspension of that compound. From left to right: (A) indomethacin, phenytoin, loviride, phenylbutazone, cinnarizine and itraconazole; (B) phenytoin, indomethacin, loviride, phenylbutazone, cinnarizine and itraconazole.
compounds, adsorbing less TPGS per unit of surface area. For other physicochemical properties (molecular weight, melting point, log P, solubility and density), no correlation with the ease of nanosuspension stabilization could be observed. Considering the importance of drug compound properties on nanosuspension stabilization, the question arises to what extent these properties can be brought in relation with the dissolution behavior of the compounds of the second and third group. Key physicochemical properties of the compounds are summarized in Table 3. The average % release of the spraydried and freeze-dried products after 15 min of dissolution was plotted as a function of the different physicochemical property values. No correlation could be found for molecular weight, melting point, solubility and density (data not shown). As for nanosuspension stabilization, this was not surprising, since molecular and/or bulk properties cannot be expected to correlate with a desagglomeration process, determined by surface properties. The plots for the amount of TPGS adsorbed per unit of surface area as well as for log P are provided in Fig. 5A and B. For TPGS adsorption (Fig. 5A), the trend can be observed that for the compounds having more hydrophobic surfaces,
adsorbing more than 5 mg TPGS/m2 surface area, the release after 15 min tends to decrease. For log P (Fig. 5B), a similar trend can be observed for more lipophilic compounds, having log P values above 4. Whereas the former property reflects the hydrophobicity of the surfaces, the latter is a molecular property. log P, however, is much more straightforward to computate or determine, making it attractive for prediction purposes. The correlation of log P of the drug compounds with their surface hydrophobicity is provided in Fig. 6. For the 6 compounds of the second and third group, more hydrophobic surfaces in general correspond to higher log P values. Extension of the evaluation to the 3 compounds of the first group, however, learns that exceptions exist. Therefore, we suggest that log P can be used as a quick but rough prediction tool. Measurement of surface hydrophobicity might be better to predict the product’s performance, although less straightforward in its determination. Therefore, as for nanosuspension production, surface hydrophobicity determined the extent to which preservation of the dissolution characteristics after drying is feasible. More hydrophobic compounds will result in more severe and harder-to-disintegrate agglomerates that will lower the dissolution rate of the product. In view of this, the hydrophobicity of the compound particle surface dictates the need of addition of matrix formers prior to drying, in order to prevent agglomeration of the nanoparticles. An evaluation of the performance of a number of alternative matrix formers is the subject of a separate study.
4.
Conclusion
Physical stability observed during short-term storage of TPGS-stabilized nanosuspensions, together with patient convenience arguments, stress the importance of drying of nanosuspensions. Freeze-drying and spray-drying were performed on the nanosuspensions and resulted in agglomerated powders. The compound rather than the drying process applied determined the resulting dissolution profile. Upon evaluation of surface hydrophobicity, it was found that compounds with more hydrophobic surfaces, adsorbing more than
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
5 mg TPGS/m2 surface area, resulted in agglomerates for which disintegration could become rate-limiting in the overall dissolution process. The same trend was found for compounds having a log P value higher than 4. It was noted, however, that exceptions exist to the correlation between log P and surface hydrophobicity, which might limit its predictive power. These data suggest that the need for addition of matrix formers prior to drying is largely dictated by the hydrophobicity of the compound for which the nanosuspension is made.
Acknowledgement The work was carried out within the framework of an interdisciplinary research project sponsored by K.U. Leuven (IDO-project IDO/04/009).
references
Fasano, A., 1998. Innovative strategies for the oral delivery of drugs and peptides. TIBTECH 16, 152–157. Grant, D.J.W., Brittain, H.G., 1995. Solubility of pharmaceutical solids. In: Brittain, H.G. (Ed.), Physical Characterization of Pharmaceutical Solids, Drugs and the Pharmaceutical Sciences, vol. 70. Marcel Dekker, Inc., New York, USA, pp. 321–386. Kesisoglou, F., Panmai, S., Wu, Y., 2007. Nanosizing—oral formulation development and biopharmaceutical evaluation. Adv. Drug Deliv. Rev. 59, 631–644. Lee, J., 2003. Drug nano- and microparticles processed into solid dosage forms: physical properties. J. Pharm. Sci. 92, 2057–2068.
135
Lipinski, C.A., 2002. Poor aqueous solubility—an industry wide problem in drug discovery. Am. Pharm. Rev. 5, 82–85. Liversidge, G.G., Cundy, K.C., Bishop, J., Czekai, D., 1991. Surface modified drug nanoparticles. US Patent, US5145684. Madras, G., McCoy, B.J., 2003. Temperature effects during Ostwald ripening. J. Chem. Phys. 119, 1683–1693. Merisko-Liversidge, E., Linden, W., 2003. Stabilization of active agents by formulation into nanoparticulate form. World Patent, WO03024424. Möschwitzer, J., Achleitner, G., Pomper, H., Müller, R.H., 2004. Development of an intravenously injectable chemically stable aqueous omeprazole formulation using nanosuspension technology. Eur. J. Pharm. Biopharm. 58, 615–619. Möschwitzer, J., Müller, R.H., 2006. Spray coated pellets as carrier system for mucoadhesive drug nanocrystals. Eur. J. Pharm. Biopharm. 62, 282–287. Müller, R.H., Möschwitzer, J., Bushrab, F.N., 2006. Manufacturing of nanoparticles by milling and homogenization techniques. In: Gupta, R.B., Kompella, U.B. (Eds.), Nanoparticle Technology for Drug Delivery, Drugs and the Pharmaceutical Sciences, vol. 159. Taylor & Francis Group LLC, New York, USA, pp. 21–51. Van Eerdenbrugh, B., Froyen, L., Martens, J.A., Blaton, N., Augustijns, P., Brewster, M., Van den Mooter, G., 2007. Characterization of physico-chemical properties and pharmaceutical performance of sucrose co-freeze-dried solid nanoparticulate powders of the anti-HIV agent loviride prepared by media milling. Int. J. Pharm. 338, 198–206. Vergote, G.J., Vervaet, C., Van Driesche, I., Hoste, S., De Smedt, S., Demeester, J., Jain, R.A., Ruddy, S., Remon, J.P., 2001. An oral controlled release matrix pellet formulation containing nanocrystalline ketoprofen. Int. J. Pharm. 219, 81–87. Wang, B.H., Zhang, W.B., Zhang, W., Mujumdar, A.S., Huang, L.X., 2005. Progress in drying technology for nanomaterials. Drying Technol. 23, 7–32.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
Drying of crystalline drug nanosuspensions—The importance of surface hydrophobicity on dissolution behavior upon redispersion Bernard Van Eerdenbrugh a , Ludo Froyen b , Jan Van Humbeeck b , Johan A. Martens c , Patrick Augustijns a , Guy Van den Mooter a,∗ a b c
Laboratory for Pharmacotechnology and Biopharmacy, K.U. Leuven, Gasthuisberg O&N2, Herestraat 49, Box 921, 3000 Leuven, Belgium Metallurgy and Materials Engineering Department, K.U. Leuven, Kasteelpark Arenberg 44, 3001 Leuven, Belgium Center for Surface Chemistry and Catalysis, K.U. Leuven, Kasteelpark Arenberg 23, 3001 Heverlee, Belgium
a r t i c l e
i n f o
a b s t r a c t
Article history:
d-␣-Tocopherol polyethylene glycol 1000 succinate (TPGS)-stabilized nanosuspensions
Received 15 April 2008
(25 wt%, relative to the drug weight) were produced by media milling for 9 model drug
Received in revised form
compounds [cinnarizine, griseofulvin, indomethacin, itraconazole, loviride, mebendazole,
20 May 2008
naproxen, phenylbutazone and phenytoin]. After 3 months of storage at room temperature,
Accepted 23 June 2008
Ostwald ripening occurred in all of the samples, except for indomethacin. Whereas lowering
Published on line 3 July 2008
the temperature could slow down the ripening, it markedly increased upon storage at 40 ◦ C. As for ripening, settling generally became more pronounced at 40 ◦ C compared to 4 ◦ C. As the
Keywords:
nanosuspensions were afflicted by Ostwald ripening and settling, we explored nanosuspen-
Nanosuspensions
sion drying as a strategy to circumvent these stability issues. Spray-drying and freeze-drying
Stability
were evaluated for nanosuspensions and coarse reference suspensions of the compounds.
Spray-drying
Nanoparticle agglomeration could be visually observed in all of the powders. To evaluate
Freeze-drying
the effect of agglomeration on the key characteristic of drug nanocrystals (i.e. rapid dissolution), dissolution experiments were performed under poor sink conditions. It was found that the compounds could be categorized into 3 groups: (i) compounds for which it was impossible to differentiate between coarse and nanosized products (griseofulvin, mebendazole, naproxen), (ii) compounds that gave clear differences in dissolution profiles between the nanosized and the coarse products, but for which drying of the nanosuspensions did not decrease the dissolution performance of the product (indomethacin, loviride, phenytoin) and (iii) compounds that showed differences between coarse and nanosized products, but for which drying of the nanosuspensions resulted in a significant decrease of the dissolution rate (cinnarizine, itraconazole, phenylbutazone). To gain insight on the influence of the drug compound characteristics on the dissolution of the dried products, the dissolution behavior of the compounds of the second and the third group was linked to the compound’s characteristics. It was found that compounds with a more hydrophobic surface resulted in agglomerates which were harder to disintegrate, for which dissolution was compromised upon drying. The same was found for compounds having higher log P values. © 2008 Elsevier B.V. All rights reserved.
∗
Corresponding author. Tel.: +32 16330304; fax: +32 16330305. E-mail address: [email protected] (G. Van den Mooter). 0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2008.06.009
128
1.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Introduction
Oral drug delivery is the preferred way of drug administration, since it avoids the pain and discomfort associated with injections and is more attractive from a marketing and patient compliance perspective (Fasano, 1998). Fuelled by the fact that, over the last decade, new drug candidates have become more hydrophobic and less water-soluble (Lipinski, 2002), designing an adequate oral dosage form is becoming challenging and formulation scientists frequently have to consider more complex drug delivery platforms including solid dispersions, cyclodextrins, salt formation and lipid dosage forms. Within this scaffold of drug delivery tools, formulation of drugs as nanocrystals, submicron-sized drug crystals, has rapidly grown to a mature drug delivery strategy during the last decade, with currently 5 products on the market. For the manufacture of nanocrystals, media milling is currently the most popular approach (Liversidge et al., 1991). The major advantage of nanocrystals for oral delivery is the increase in the dissolution velocity and hence oral absorption, based on the increased specific surface area of the particles. Also often advantages may accompany this formulation approach, such as reduced fasted/fed state variability and ease of administration (Kesisoglou et al., 2007). The particle size reduction processes used to produce nanocrystals (media milling and high pressure homogenization) are conducted in the suspended state, hence forming a nanosuspension. However, solid dosage forms are considered more attractive, due to their convenience (marketing aspects) and possible stability issues associated with nanoparticles in their suspended state. These stability issues can be both physical (e.g. Ostwald ripening and agglomeration) and chemical (e.g. hydrolysis), although examples exist for which formulation as a nanosuspension actually prevents the latter, compared with formulation as a solution (e.g. Möschwitzer et al., 2004; Merisko-Liversidge and Linden, 2003). Technically, transformation of nanosuspensions into solid products can be achieved using established unit-operations such as freeze-drying, spray-drying, pelletization and granulation (Müller et al., 2006). Often matrix formers are added to the suspensions prior to the drying step, in order to achieve adequate redispersion in water. Whatever technique used for this transformation step, water removal occurs either by evaporation or sublimation. In contrast with the importance of this transformation step, the current available literature is scarce (e.g. Müller et al., 2006; Vergote et al., 2001; Lee, 2003; Möschwitzer and Müller, 2006; Van Eerdenbrugh et al., 2007). Furthermore, reports addressing the key attribute of a nanocrystal system intended for oral delivery, i.e. preservation of the beneficial dissolution characteristics of the nanosuspension upon redispersion, are even more limited (e.g. Van Eerdenbrugh et al., 2007). To date, to the best of our knowledge, the influence of the drug compound on the dissolution characteristics after drying of nanosuspensions has never been investigated. The aim of this study is to evaluate the effect of drying on the dissolution performance of drug nanosuspensions. Therefore, a screening approach was followed, using 9 different model drug compounds. The structures and abbreviations of
the model compounds are provided in Fig. 1. The stabilizing system chosen for this study was 25 wt% d-␣-tocopherol polyethylene glycol 1000 succinate (TPGS) relative to the drug weight, as a previous study proved that nanosuspension production was feasible for all drug compounds using this stabilizing system (submitted manuscript). The first part of the study consists of a brief description of short-term physical stability of the nanosuspensions (Ostwald ripening and agglomeration). Ostwald ripening and, in some cases, agglomeration, further stresses on the importance of the second part, a study on drying of the nanosuspensions. Both evaporation and sublimation techniques were evaluated, by using spray-drying and freeze-drying, respectively. No matrix formers were added prior to drying as the aim and focus of the present study is to gather insight on the influence of the drug compounds’ characteristics on the dissolution characteristics of the dried products. The resulting powders were evaluated using dissolution, as rapid dissolution can be considered as the main advantage of a nanosized system intended for oral delivery. Finally, the dissolution behavior of the products was related to the compounds’ characteristics.
2.
Materials and methods
2.1.
Materials
d-␣-Tocopherol polyethylene glycol 1000 succinate (TPGS, Eastman Chemical Company, Kingsport, TN, USA) was a gift from the manufacturer. Loviride and itraconazole were kindly provided by Johnson & Johnson Pharmaceutical Research and Development (Beerse, Belgium). Cinnarizine (Certa n.v., Braine-l’Alleud, Belgium), griseofulvin (Certa n.v.), indomethacin (Certa n.v.), mebendazole (Certa n.v.), naproxen (Certa n.v.), phenylbutazone (Fagron NV, Waregem, Belgium), phenytoin (Fagron NV), acetonitrile gradient grade far UV (Fisher Scientific UK Limited, Loughborough, UK), dimethylsulfoxide (Acros Organics, Geel, Belgium), methanol 99.9% for HPLC gradient grade (Acros Organics), potassium dihydrogen phosphate GR for analysis (Merck KGaA, Darmstadt, Germany), dipotassium hydrogen phosphate trihydrate GR for analysis (VWR International bvba/sprl, Leuven, Belgium), sodium dodecyl sulfate (SLS, Merck Schuchardt OHG, Hohenbrunn, Germany), tetrabutyl ammonium hydrogen sulfate (TBAH, Acros Organics), 1 M HCl (Titrinorm® , VWR International, Fontenay Sous Bois, France), yttrium-stabilized zirconia beads (diameter 0.5 mm, YTZ® grinding media, Tosoh Corporation Advanced Ceramics Department, Tokyo, Japan) were obtained commercially. Demineralized water was used for all formulations (Elga, maxima ultra pure water, ≥18 M).
2.2.
Nanosuspension production
Nanosuspensions were prepared using media milling. Suspensions of 1 g of drug compound in 5 ml water were made in 10 ml vials using 250 mg of TPGS as a stabilizer. Subsequently, 15 g of zirconia beads (Ø 0.5 mm) were added as a milling agent and vials were placed in self-designed nylon holders (6 vials per holder) that fitted into 500 ml mixing bowls.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
129
Fig. 1 – Chemical structures of the drug compounds used in the study, with the abbreviations as used throughout the manuscript.
The four mixing bowls were then placed into the grinding stations of a planetary mill (Retsch PM 400 MA, Retsch GmbH, Haan, Germany) and milling was performed at 250 rpm for 24 h. Six batches, each containing 1 gram of drug, were prepared for each compound. After milling, batches were pooled and diluted with water to a concentration of 20 mg/ml, after content determination by HPLC (Section 2.3).
2.3.
Quantitative assays by HPLC
For the determinations of the drug content in the nanosuspensions (Section 2.2) and the dried products (Section 2.7), the determination of SLS solubilization for the design of relevant dissolution conditions (Section 2.8.1), and the analysis of samples in the dissolution experiments (Section 2.8.2), HPLC-UV was used. Samples were analyzed, after dilution into the validated range, on a Merck-Hitachi-Lachrom instrument (Hitachi Ltd., Tokyo, Japan) using a Merck Chromolith Performance RP18-e (100–4.6 mm) column (Merck, Darmstadt, Germany).
Each determination was performed in triplicate and the average and standard deviation was calculated.
2.4. (DLS)
Size determination by dynamic light scattering
For the measurement of the particle size of the nanosuspensions, dynamic light scattering was used. DLS was performed using a Nanophox instrument in autocorrelation mode (SympaTec GmbH, Clausthal-Zellerfeld, Germany). Prior to measurement, samples were diluted with a saturated solution of the compound under investigation. Saturated solutions were prepared by filtration of a drug suspension, stirred overnight, through a 0.45 m nylon 47 mm membrane filter (Millipore Corporation, Bedford, MA, USA). For griseofulvin, a saturated solution of the drug containing 5 mg/ml TPGS was used, since severe agglomeration occurred upon dilution in a saturated solution without TPGS. Detection was carried out at a scattering angle of 90◦ , sample temperature was set at
130
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
25 ◦ C and 3 runs of 30 s were performed on each sample. From the resulting correlation curves, a 2nd order analysis was performed to calculate the mean particle size and standard deviation.
2.5.
Stability evaluation of the nanosuspensions
Three temperature conditions were applied in the stability study of the nanosuspensions: 4 ◦ C (refrigerator), ambient conditions, and 40 ◦ C (thermostatic oven). Physical stability of the nanosuspensions was evaluated after 3 months of storage in sealed vials. Particle size measurements were performed by DLS (Section 2.4) and settling behavior by visual examination. Particle size results were expressed relative to the mean particle size obtained after production (%).
2.6. Freeze-drying and spray-drying of the nanosuspensions 2.6.1.
Freeze-drying
20 ml of the 20 mg/ml nanosuspensions corresponding to 400 mg of drug compound was transferred into a glass vial. Subsequently, the suspensions were frozen by immersing the vials in liquid nitrogen. Freeze-drying of the products was performed with a Christ model Alpha freeze-dryer (type 1050, Van Der Heyden, Brussels, Belgium) at a shelf temperature of −50 ◦ C with a pressure below 1 mbar and the vials were removed after 48 h of drying. After freeze-drying, residual water was removed by placing the vials overnight in a vacuum oven at 25 ◦ C (Mazzazi Systems, Italy). The same procedure was followed on 50 ml of non-nanosized suspensions, corresponding to 1 g of drug compound.
2.6.2.
Spray-drying
Spray-drying was performed with a Büchi Mini Spray Dryer B191 (Flawil, Switzerland). 200 ml of nanosuspension (4 g of drug compound) was spray-dried using the following process parameters: the inlet temperature was 120 ◦ C, the aspirator air flow were set at 100%; the pump speed was 40%. During spray-drying the suspension was continuously stirred with a magnetic stirrer. The spray-dried solid product was isolated and subsequently dried overnight in a vacuum oven at 25 ◦ C (Mazzazi Systems). The same procedure was followed on 300 ml of non-nanosized suspensions, corresponding to 6 g of drug compound.
2.7.
in the filtrate of an equilibrated aqueous drug suspension. Therefore, an excess of drug compound was added to a test tube containing 10 ml of 0.5, 1 or 2% (w/v) SLS and 0.01N HCl in water. For cinnarizine, a medium without HCl was used, as pronounced agglomeration and irreproducible results were observed upon using HCl in the medium. After equilibration overnight on a horizontal shaker (Köttermann type 4020, Haenigsen, Germany), 5 ml of suspension was filtrated through a 0.1 m PTFE syringe filter (Whatman Inc., Clifton, NJ, USA). Subsequently, the drug content in the filtrate was assessed by HPLC (Section 2.3). Each experiment was performed in triplicate and the average and standard deviations were calculated. The SLS concentration for the dissolution experiments, corresponding to a drug compound solubility of 0.375 mg/ml was determined from the linear fit of the concentration of solubilized compound as a function of SLS concentration, taking into account the 0.5, 1 and 2% (w/v) data points. All linear fits resulted in a R2 value higher than 0.995.
Determination of drug content
Drug content in the obtained powders was determined by dissolving about 10 mg of powder in 10 ml of DMSO. Samples were analyzed by HPLC-UV (Section 2.3) from which the drug content [% (w/w)] was determined. The procedure was performed in triplicate and the average and standard deviations were calculated.
2.8.
Design of dissolution experiments
2.8.1.
Determination of relevant dissolution conditions
2.8.2.
3.
Results and discussion
It should be pointed out that, to conduct a study on nanosuspension drying for a set of 9 drug compound, one first has to identify a stabilizing system that is able to yield nanosuspensions for all of the compounds. The choice of TPGS as a stabilizing system was based on a screening study, using 13 different stabilizers in 3 concentrations for these drug compounds. In this study, only TPGS in 25 or 100 wt%, relative to the drug weight, proved to be able to yield nanosuspensions for all of the drug compounds (submitted manuscript). The 25% concentration was selected to minimize possible effects of the stabilizer on the products’ constitution after drying.
3.1. Micellar solubilization of the drug compound by SLS at ambient conditions was performed by analyzing the drug content
Dissolution experiments
Dissolution experiments were performed at room temperature using a miniaturized design. An amount of formulation corresponding to 5 mg of drug compound (0.25 mg/ml) was transferred in a 20 ml disposable syringe, together with an Al2 O3 ball of 10.5 mm diameter (to provide hydrodynamics during the experiment). Subsequently, 20 ml of dissolution medium was gently sucked into the syringe, a 0.1 m PTFE syringe filter (Whatman Inc.) was mounted on the syringe and the syringe was fixed into a rotary mixer (Labinco BV, the Netherlands), making 15 rotations/min. After 1.5, 5, 15, 30, 60, 90 and 120 min, 300 l of medium was filtrated, from which the initial 100 l was discarded and the remaining fraction was collected in a 1.5 ml eppendorf tube (Eppendorf, Hamburg, Germany). The exact concentration corresponding to 100% release was determined by assaying the total (dissolved + undissolved) drug concentration in the suspension after the experiment. Samples were analyzed by HPLC (Section 2.3). Each dissolution experiment was performed in triplicate and the average values and standard deviations were calculated.
Physical stability of drug nanosuspensions
Particle size data of the nanosuspensions after production and physical stability data after 3 months
131
e
d
c
b
DLS analysis of the PHB nanosuspension resulted in correlation curves with shapes that deviated significantly from the theoretical fitting. SEM indicated that the particles had a large aspect ratio, which can explain this phenomenon. Hence, DLS evaluation of the stability samples is of limited value. T/C indicates a nanosuspension that shows settling, but is still turbid. C 0.9 indicates a settled nanosuspension for which about 90% of the total volume was clear. T indicates a completely turbid nanosuspension. C 0.5 indicates a settled nanosuspension for which about 50% of the total volume was clear. a
T/Cb T/Cb C 0.5e T/Cb T/Cb T/Cb Td T/Cb Td T/Cb Td C 0.5e Td T/Cb T/Cb T/Cb Settling (visual)
4 ◦C 40 ◦ C
C 0.9c C 0.9c
106 ± 5% 107 ± 4% 122 ± 8% –a –a –a 107 ± 5% 110 ± 3% 146 ± 6% 113 ± 3% 193 ± 3% 267 ± 11% 103 ± 2% 109 ± 1% 139 ± 2% 98 ± 2% 115 ± 2% 140 ± 2% 97 ± 2% 100 ± 1% 122 ± 2% 106 ± 5% 111 ± 2% 223 ± 17% 110 ± 4% 113 ± 4% 120 ± 5% 4 ◦C Ambient 40 ◦ C Stability data (3 months) Relative size (DLS)
406 ± 17 498 ± 10 288 ± 4% 190 ± 2 156 ± 2 220 ± 4
Phenytoin Phenylbutazone Naproxen Mebendazole Loviride Itraconazole Indomethacin
193 ± 2 256 ± 1 366 ± 12 Mean (nm)
To study the effect of drying on the nanosuspensions, water was eliminated from the nanosuspensions by evaporation (spray-drying) and sublimation (freeze-drying), since these processes form the elementary mechanisms of water removal in all unit-operations that could be typically applied for nanosuspension drying. Drug content of the powders after drying as well as drug yield of the spray-drying process for the different compounds are summarized in Table 2. A general observation is that the drug content in the powders is slightly lower than the expected 80% (w/w) (as TPGS was added as 25% of the drug weight). An exception to this are the spray-dried coarse suspensions, that always resulted in content values above 80% (w/w) (88 ± 5% on average). This was due to a pronounced loss of TPGS in the instrument’s cyclone, as witnessed by the formation of a viscous liquid that solidified upon cooling to room temperature in the cyclone during spray-drying. Agglomeration could be visually observed after drying of all nanosized products. Although there is no consensus on the mechanism leading to nanoparticle agglomeration, the capillary pressure theory is currently the main theory to explain agglomeration during drying (Wang et al., 2005). According to
Starting suspension Size by DLS
Drying of the drug nanosuspensions
Griseofulvine
3.2.
Cinnarizine
of storage of the nanosuspensions are summarized in Table 1. The first part of the evaluation of the physical stability is the investigation of Ostwald ripening. This process by which smaller particles are consumed in the growth of larger particles originates from the enhanced solubility arising from the higher curvature of smaller particles, as described by the Ostwald-Freundlich equation (Grant and Brittain, 1995). As can be seen from the table, Ostwald ripening can be observed at room temperature in all of the nanosuspensions, except for indomethacin. Whereas storage at lower temperature (4 ◦ C) could slow down (cinnarizine, griseofulvin, loviride, mebendazole, naproxen, phenytoin, although not always significantly) or even stop (ITR) the ripening, it markedly increased at higher temperature (40 ◦ C), most pronounced for griseofulvin and mebendazole. Although temperature has been described to exert its effect on Ostwald ripening through a complex combination of effects (Madras and McCoy, 2003), it clearly increases with temperature for the systems studied. As for ripening, settling generally increased at an elevated temperature of 40 ◦ C (except for phenylbutazone). Whereas almost half of the suspensions stayed completely turbid at 4 ◦ C (indomethacin, itraconazole, loviride, mebendazole), at least some settling could be observed in all of the samples at 40 ◦ C. The severe settling behavior of griseofulvin is most likely due to agglomeration caused by dilution of the nanosuspension to 20 mg/ml prior to storage, as upon dilution, settling occurred within hours at ambient conditions, whereas this was not the case in the undiluted nanosuspension. The above results clearly indicate that physical stability issues (Ostwald ripening and agglomeration) occur in the nanosuspensions studied. They advocate further transformation of these systems into solid powders, in order to obtain a product with adequate physical stability.
Table 1 – Size of the nanosuspensions after production (DLS), relative size after 3 months of storage at 4 ◦ C, room temperature and 40 ◦ C and settling behavior after 3 months of storage at 4 and 40 ◦ C
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
132
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Table 2 – Drug content of the spray-dried and freeze-dried Content [% (w/w)] Spray-dried Nano ± ± ± ± ± ± ± ± ±
Cinnarizine Griseofulvin Indomethacin Itraconazole Loviride Mebendazole Naproxen Phenylbutazone Phenytoin
75 74 74 77 79 81 73 74 75
1 1 1 1 1 1 0 0 1
Overall
76 ± 3
Freeze-dried
Coarse 80 85 88 83 94 85 94 88 92
± ± ± ± ± ± ± ± ±
1 1 1 1 2 2 3 1 1
88 ± 5
Nano 83 73 78 75 80 79 77 67 66
± ± ± ± ± ± ± ± ±
Coarse
1 2 1 0 1 1 1 1 2
75 ± 6
84 74 81 78 78 79 77 74 71
± ± ± ± ± ± ± ± ±
1 3 2 2 3 2 1 0 2
77 ± 4
this theory, agglomeration occurs as a result of the capillary forces encountered during the drying process.
3.3.
Fig. 2 – Dissolution curves of mebendazole. Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
Dissolution of the dried products
Since the main attribute of a nanosuspension is the increased dissolution rate resulting from the high specific surface area of the particles, dissolution of the powders was performed as a pharmaceutical evaluation. It is important to realize that nanoparticles agglomeration is likely to be inevitable during drying of a nanosuspension to which no additional matrix formers are added. From the point of dissolution performance, the key question is therefore not whether nanoparticle agglomeration will occur, rather the central question is how easily the agglomerates break up during dissolution. In other words, will desagglomeration become a rate-controlling step in the overall dissolution process? To provide an answer to this question, poor sink conditions were designed for each of the compounds in which the solubility of the compound was 150% (0.375 mg/ml) of the concentration upon complete dissolution (0.250 mg/ml). Dissolution was performed on nanosized and coarse products, in the suspended state and on spray-dried and freeze-dried powders. Upon analyzing the dissolution curves of the different compounds, three distinct groups of profiles were found. In the first group (griseofulvin, mebendazole, naproxen), no differences
could be observed between both the nanosized and the nonnanosized products, as illustrated in Fig. 2 for mebendazole. Since the dissolution medium was not able to discriminate between nanosized and non-nanosized products, these compounds were disregarded for further evaluation. All compounds of the second and third group showed clear differences between nanosized and non-nanosized products. The second group (indomethacin, loviride, phenytoin), for which dissolution profiles are given in Fig. 3A–C are characterized by the fact that spray-drying and freeze-drying of the nanosuspensions did not compromise the dissolution rate to a significant extent. This is an indication that the agglomerates easily break up upon redispersion in water. For these powders, addition of a matrix former prior to drying is unnecessary from the point of view of preservation of the dissolution rate of the nanoparticles. For the third group (cinnarizine, itraconazole, phenylbutazone), on the other hand, dissolution rate clearly decreased in the dried nanosized products (Fig. 4A–C). The agglomerates formed upon drying are apparently so strong that their disintegration compromises the overall dissolution process. Therefore, addition of matrix formers prior to drying is nec-
Table 3 – Key physicochemical properties of the model compounds Compound CIN GRI IND ITR LOV MEB NAP PHB PHT a b
Molecular weight (g/mol)
Melting point (onset, ◦ C)
369 353 358 706 351 295 230 308 252
119.5 ± 0.1 220a 159.9 ± 0.08 165.2 ± 0.1 226.3 ± 0.1 289a 153a 106.0 ± 0.1 295.03 ± 0.48
log P (calculated) 6.1 1.1 3.3 8.5 3.7 3.3 3.0 4.2 2.3
Aqueous solubility (g/ml) <2b 7.7 ± 0.2 16 ± 1 <0.4b <2b 0.3 ± 0.0 31 ± 2 24 ± 0 17 ± 1
Values were taken from the Merck Index as the compound showed polymorphism and/or degradation upon heating. Below the limit of quantification.
Density (g/ml) 1.13 1.45 1.37 1.35 1.44 1.38 1.24 1.19 1.26
± ± ± ± ± ± ± ± ±
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Fig. 3 – Dissolution curves of indomethacin (A), loviride (B), and phenytoin (C). Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
essary in this case to maintain rapid dissolution of the drug. The performance of various matrix formers is the subject of a separate manuscript.
3.4.
Influence of the drug compound
An important trend that can be seen in the dissolution results provided in Figs. 3 and 4 is that spray-drying and freeze-drying always resulted in very similar dissolution profiles. This suggests that the agglomeration of the nanoparticles is governed by the compounds characteristics rather than by the dry-
133
Fig. 4 – Dissolution curves of cinnarizine (A), itraconazole (B), and phenylbutazone (C). Key: () coarse suspension, () nanosuspension, () spray-dried coarse suspension, (♦) spray-dried nanosuspension, ( ) freeze-dried coarse suspension, and ( ) freeze-dried nanosuspension.
ing procedure applied. Agglomeration tendency and strength can be expected to be a result of the surface properties of the compound, more precisely the surface hydrophobicity of the particles. In a previous study (submitted manuscript), surface hydrophobicity of the nanosuspensions of these compounds was probed by determination of the amount of TPGS adsorbed per unit of surface area, based on the fact that the hydrophobic d-␣-tocopherol part of TPGS can be expected to adsorb on the nanoparticle surfaces. The outcome of this study indicated that nanosuspension stabilization, i.e. the prevention of agglomeration, was easier for less hydrophobic
134
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
Fig. 6 – log P as a function of TPGS adsorption for the compounds of the first (), second () and third () group.
Fig. 5 – Release after 15 min as a function of (A) TPGS adsorption per unit of surface area and (B) log P of the compounds of the second and third group. Release values are the averages of the spray-dried and freeze-dried nanosuspension of that compound. From left to right: (A) indomethacin, phenytoin, loviride, phenylbutazone, cinnarizine and itraconazole; (B) phenytoin, indomethacin, loviride, phenylbutazone, cinnarizine and itraconazole.
compounds, adsorbing less TPGS per unit of surface area. For other physicochemical properties (molecular weight, melting point, log P, solubility and density), no correlation with the ease of nanosuspension stabilization could be observed. Considering the importance of drug compound properties on nanosuspension stabilization, the question arises to what extent these properties can be brought in relation with the dissolution behavior of the compounds of the second and third group. Key physicochemical properties of the compounds are summarized in Table 3. The average % release of the spraydried and freeze-dried products after 15 min of dissolution was plotted as a function of the different physicochemical property values. No correlation could be found for molecular weight, melting point, solubility and density (data not shown). As for nanosuspension stabilization, this was not surprising, since molecular and/or bulk properties cannot be expected to correlate with a desagglomeration process, determined by surface properties. The plots for the amount of TPGS adsorbed per unit of surface area as well as for log P are provided in Fig. 5A and B. For TPGS adsorption (Fig. 5A), the trend can be observed that for the compounds having more hydrophobic surfaces,
adsorbing more than 5 mg TPGS/m2 surface area, the release after 15 min tends to decrease. For log P (Fig. 5B), a similar trend can be observed for more lipophilic compounds, having log P values above 4. Whereas the former property reflects the hydrophobicity of the surfaces, the latter is a molecular property. log P, however, is much more straightforward to computate or determine, making it attractive for prediction purposes. The correlation of log P of the drug compounds with their surface hydrophobicity is provided in Fig. 6. For the 6 compounds of the second and third group, more hydrophobic surfaces in general correspond to higher log P values. Extension of the evaluation to the 3 compounds of the first group, however, learns that exceptions exist. Therefore, we suggest that log P can be used as a quick but rough prediction tool. Measurement of surface hydrophobicity might be better to predict the product’s performance, although less straightforward in its determination. Therefore, as for nanosuspension production, surface hydrophobicity determined the extent to which preservation of the dissolution characteristics after drying is feasible. More hydrophobic compounds will result in more severe and harder-to-disintegrate agglomerates that will lower the dissolution rate of the product. In view of this, the hydrophobicity of the compound particle surface dictates the need of addition of matrix formers prior to drying, in order to prevent agglomeration of the nanoparticles. An evaluation of the performance of a number of alternative matrix formers is the subject of a separate study.
4.
Conclusion
Physical stability observed during short-term storage of TPGS-stabilized nanosuspensions, together with patient convenience arguments, stress the importance of drying of nanosuspensions. Freeze-drying and spray-drying were performed on the nanosuspensions and resulted in agglomerated powders. The compound rather than the drying process applied determined the resulting dissolution profile. Upon evaluation of surface hydrophobicity, it was found that compounds with more hydrophobic surfaces, adsorbing more than
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 127–135
5 mg TPGS/m2 surface area, resulted in agglomerates for which disintegration could become rate-limiting in the overall dissolution process. The same trend was found for compounds having a log P value higher than 4. It was noted, however, that exceptions exist to the correlation between log P and surface hydrophobicity, which might limit its predictive power. These data suggest that the need for addition of matrix formers prior to drying is largely dictated by the hydrophobicity of the compound for which the nanosuspension is made.
Acknowledgement The work was carried out within the framework of an interdisciplinary research project sponsored by K.U. Leuven (IDO-project IDO/04/009).
references
Fasano, A., 1998. Innovative strategies for the oral delivery of drugs and peptides. TIBTECH 16, 152–157. Grant, D.J.W., Brittain, H.G., 1995. Solubility of pharmaceutical solids. In: Brittain, H.G. (Ed.), Physical Characterization of Pharmaceutical Solids, Drugs and the Pharmaceutical Sciences, vol. 70. Marcel Dekker, Inc., New York, USA, pp. 321–386. Kesisoglou, F., Panmai, S., Wu, Y., 2007. Nanosizing—oral formulation development and biopharmaceutical evaluation. Adv. Drug Deliv. Rev. 59, 631–644. Lee, J., 2003. Drug nano- and microparticles processed into solid dosage forms: physical properties. J. Pharm. Sci. 92, 2057–2068.
135
Lipinski, C.A., 2002. Poor aqueous solubility—an industry wide problem in drug discovery. Am. Pharm. Rev. 5, 82–85. Liversidge, G.G., Cundy, K.C., Bishop, J., Czekai, D., 1991. Surface modified drug nanoparticles. US Patent, US5145684. Madras, G., McCoy, B.J., 2003. Temperature effects during Ostwald ripening. J. Chem. Phys. 119, 1683–1693. Merisko-Liversidge, E., Linden, W., 2003. Stabilization of active agents by formulation into nanoparticulate form. World Patent, WO03024424. Möschwitzer, J., Achleitner, G., Pomper, H., Müller, R.H., 2004. Development of an intravenously injectable chemically stable aqueous omeprazole formulation using nanosuspension technology. Eur. J. Pharm. Biopharm. 58, 615–619. Möschwitzer, J., Müller, R.H., 2006. Spray coated pellets as carrier system for mucoadhesive drug nanocrystals. Eur. J. Pharm. Biopharm. 62, 282–287. Müller, R.H., Möschwitzer, J., Bushrab, F.N., 2006. Manufacturing of nanoparticles by milling and homogenization techniques. In: Gupta, R.B., Kompella, U.B. (Eds.), Nanoparticle Technology for Drug Delivery, Drugs and the Pharmaceutical Sciences, vol. 159. Taylor & Francis Group LLC, New York, USA, pp. 21–51. Van Eerdenbrugh, B., Froyen, L., Martens, J.A., Blaton, N., Augustijns, P., Brewster, M., Van den Mooter, G., 2007. Characterization of physico-chemical properties and pharmaceutical performance of sucrose co-freeze-dried solid nanoparticulate powders of the anti-HIV agent loviride prepared by media milling. Int. J. Pharm. 338, 198–206. Vergote, G.J., Vervaet, C., Van Driesche, I., Hoste, S., De Smedt, S., Demeester, J., Jain, R.A., Ruddy, S., Remon, J.P., 2001. An oral controlled release matrix pellet formulation containing nanocrystalline ketoprofen. Int. J. Pharm. 219, 81–87. Wang, B.H., Zhang, W.B., Zhang, W., Mujumdar, A.S., Huang, L.X., 2005. Progress in drying technology for nanomaterials. Drying Technol. 23, 7–32.