- Email: [email protected]
Simple low-cost miniaturization approach for pharmaceutical nanocrystals production
Accepted Manuscript Title: Simple low-cost miniaturization approach for pharmaceutical nanocrystals production Author: Gregori B. Romero Cornelia M. Keck Rainer H. M¨uller PII: DOI: Reference:
S0378-5173(15)30394-X http://dx.doi.org/doi:10.1016/j.ijpharm.2015.11.047 IJP 15385
To appear in:
International Journal of Pharmaceutics
Received date: Revised date: Accepted date:
28-7-2015 25-11-2015 26-11-2015
Please cite this article as: Romero, Gregori B., Keck, Cornelia M., M¨uller, Rainer H., Simple low-cost miniaturization approach for pharmaceutical nanocrystals production.International Journal of Pharmaceutics http://dx.doi.org/10.1016/j.ijpharm.2015.11.047 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Simple low-cost miniaturization approach for pharmaceutical nanocrystals production Gregori B. Romeroa [email protected] [email protected], Cornelia M. Keckb, Rainer H. Müllera* [email protected] a
Pharmaceutics, Pharmaceutical Nanotechnology & NutriCosmetics, Freie Universität Berlin, Kelchstr. 31,
12169 Berlin, Germany b
Applied Pharmacy, University of Applied Sciences Kaiserslautern, Campus Pirmasens, Carl-Schurz-Str. 10-16,
66953 Pirmasens, Germany *
Corresponding author: Tel.: +49 30 838 506 78.
1
Graphical Abstract
2
Abstract Systematic screening for optimal formulation composition and production parameters for nanosuspensions consumes a lot of time and also drug material when performed at lab scale. Therefore, a cost-effective miniaturized scale top down approach for nanocrystals production by wet bead milling was developed. The final set-up consisted of 3 magnetic stirring bars placed vertically one over the other in a 2 mL glass vial and agitated by a common magnetic stirring plate. All of the tested actives (cyclosporin A, resveratrol, hesperitin, ascorbyl palmitate, apigenin and hesperidin) could be converted to nanosuspensions. For 4 of them, the particles sizes achieved were smaller than previously reported on the literature (around 90 nm for cyclosporin A; 50 nm for hesperitin; 160 nm for ascorbyl palmitate and 80 nm for apigenin). The “transferability” of the data collect by the miniaturized method was evaluated comparing the production at larger scale using both wet bead milling and high pressure homogenization. Transferable information obtained from the miniaturized scale is minimum achievable size, improvements in size reduction by reduction of beads size, diminution kinetics and potentially occurring instabilities during processing. The small scale batches also allow identification of optimal stabilizer types and concentrations. The batch size is 0.5 mL, requiring approximately 50 mg or 5 mg of drug (5% and 1% suspension, respectively). Thus, a simple, accessible, low-cost miniaturized scale method for the production of pharmaceutical nanocrystals was established. Keywords: nanocrystals; nanosuspension; nanoparticle; miniaturization; downscaling; bead milling
3
1. Introduction Nanocrystals are meanwhile well established for the formulation of poorly soluble actives, preferentially for oral (Keck and Müller, 2006) and dermal delivery (Shegokar and Müller, 2010). Products are on the market in pharma (e.g. Rapamune, Tricor), but also in cosmetics (e.g. platinum rare by la prairie). The typical size is above 100 nm (but below 1,000 nm), thus they are no nanomaterial/nanoparticle product according to e.g. European Commission recommendation on the definition of nanomaterial (2011/696/EU) and FDA guidance “Considering Whether an FDARegulated Product Involves the Application of Nanotechnology”. This fact eases product registration. In addition, this “submicron” size range is more easily accessible by the various production methods used. The most important industrially used methods are bead milling (Alkermes, prev. élan/Nanosystems) (Liversidge et al., 1991) and high pressure homogenization (HPH) (SkyePharma PLC) (Kruss et al., 1996; Parikh and Selvaraj, 1999). There are also combination methods described consisting of a pre-treatment step followed by a main step of crystal disintegration, e.g. NANOEDGE technology by Baxter (Kipp et al., 2006); H69 process: spray drying and subsequent HPH (Müller and Möschwitzer, 2009); H96 process: lyophilisation and subsequent HPH (Lemke and Moeschwitzer, 2007). Here often one aim is to access also the particle size range below 100 nm (H69, H96). For assessing the comminution ability of crystals, formulation screening (e.g. type and concentrations of stabilizers, stabilizer mixtures) and first physical stability investigations, a downscaling to small volumes is desirable. This saves time, and often very important, saves active. This is especially important in case of new chemical entities with potentially very limited amount available (often rather milligrams than grams). Commercial bead mills in their small volume version have suspension volumes rather in the range 50 ml to 100 ml (e.g. Bühler PML-2, 200 ml chamber and required suspension volume about 120 ml). Assuming a 5% concentration of active, a density of 1.5 g/ml, about 10 g of active are required. Most of the high pressure homogenizers require 40-200 ml minimum volume, i.e. at least about 3 g of active (e.g. Micron LAB 40, APV Deutschland,
4
Germany: 40 ml). There are also homogenizers with smaller volumes (e.g. 3-5 ml), but the process parameter pressure cannot be controlled very precisely (e.g. Avestin EmulsifFlex-B15, minimum volume 3 mL). One approach could be to reduce the drug concentration (e.g. to 1%), but this makes the milling process less effective. The drug crystals move against each other during the milling process, which contributes to the comminution efficiency. Thus rather about 20% suspension concentration is ideal. Independent on the concentration aspect, all these instruments require a certain processing time, which also might be shortened by using a small scale “multiple” system (i.e. small scale and running many samples in parallel). Thus there is clearly a need for effective down scaling, i.e. having a small volume and being at the same time cost-effective. A very simple approach in bead milling is filling of a 20 ml injection vial with milling beads, adding a magnetic stirrer bar and the suspension and placing it on a magnetic stirrer plate. However, this milling process is not very controlled, sometimes not effective due to uncontrolled movement of the stirring bar. In case of coated stirrer bars, erosion from the polymer coat can take place. To have a more controlled process, e.g. a glass-vial-based system has been described, where vials with different volumes were placed in a Retsch PM 400 MA planetary mill (Van Eerdenbrugh et al., 2009). However, this requires the investment of a mill, or a few mills (20 vials per holder in the mill). In parallel, milling in a 96-well plate was investigated, fixed on a shaker (25-340 µl/well) (Van Eerdenbrugh et al., 2009). In case of plates made from plastic material, erosion from the walls cannot be excluded. The suspensions used were about 16%, and drug amounts down to 1 mg could be processed. In this study, a down scaled system was developed based on the preferable glass vials, having a special optimized arrangement of the stirrer bars, and avoiding the use of a planetary mill by using a multiple magnetic stirring plate. Systematically investigated were the effect of number and arrangement of stirring bars, the effect of the type of active to be diminuted, and effect of milling bead size in this miniaturized system. The results obtained from the small scale were compared to
5
data from larger scale high pressure homogenization and the Bühler bead mill PML-2, to judge the transferability of data from miniaturized to larger scale. 2. Materials and methods 2.1. Materials Resveratrol, hesperitin, ascorbyl palmitate, apigenin and hesperidin were purchased from Denk Ingredients GmbH (Germany). Cyclosporin A was a donation from PharmaSol GmbH (Germany). Vitamin E polyethylene glycol succinate (TPGS) (trade name Kolliphor® TPGS), alkyl polyglucoside C8-C10 (trade name Plantacare® 2000 UP) and poloxamer 188 (trade name Kolliphor® P 188) were donations from BASF SE (Germany). Double distilled and ultrapurified water was obtained from a Milli-Q apparatus (Millipore GmbH, Germany). All other reagents were from analytical grade. 2.2. Nanosuspensions production 2.2.1. Miniaturized wet bead milling using a magnetic stirring plate The production principle was a top-down approach, in this case wet bead milling with stirring bars in a super reduced scale. The milling chamber consisted of a 2 mL glass vial (i.e. 10x lower than the namely used 20 mL injection vials). The final set-up was composed of 3 cylindrical stirring bars (9.5 x 6 mm) (VWR International, Germany), disposed vertically one over the other (Fig. 3). The volume of the 3 stirring bars was 0.7 mL (0.233 mL/stirring bar). From the remaining space (1.3 mL), 1 mL was destined for the milling beads and the suspensions of the raw powder of the actives (i.e. 0.3 mL headspace). The milling beads possessed various sizes (diameters of 0.05, 0.1, 0.2 and 0.4-0.6 mm) and were yttria stabilized zirconium oxide beads (Hosokawa Alpine, Germany).
6
All the formulations investigated were aqueous suspensions and contained 5% active and 1% stabilizer (all w/w). The proportion of beads to suspension was 1:1 (volume), in other words, 0.5 mL (1.9 g) of milling beads (density 3.8 Kg/L) and 0.5 mL of the suspension. The final 0.3 mL was left empty as headspace. The actives used and their respective stabilizers are shown in table 1. The stabilizer type for each active was selected based on previous experiences producing nanosuspensions.
The vials were stirred on a magnetic stirring plate RCT basic (IKA-Werke GmbH & Co. KG, Germany) at 1,200 rpm and 5 oC. Samples were drawn after defined intervals up to 120 hours. 2.2.2. Bench scale wet bead milling using a Bühler PML-2 The bench scale production by wet bead milling was performed using a bead mill PML-2 (Bühler, Switzerland). Milling beads of 0.2 or 0.4-0.6 mm identical to the ones described in 2.2.1 were used. The milling time was up to 60 minutes, at a speed of 2,000 rpm and 5°C. Samples were drawn after 10, 20, 30 and 60 minutes of processing. 2.2.3. Bench scale high pressure homogenization with a Micron LAB 40 The bench scale production by high pressure homogenization was performed using a homogenizer Micron LAB 40 (APV Deutschland GmbH, Germany). The suspension was first pre-processed by 2 high pressure homogenization (HPH) cycles at 150, 500 and 1,000 bar, respectively, followed by 20 cycles at 1,500 bar. 2.3. Particle characterization 2.3.1.Photo correlation spectroscopy The particle size of the nanocrystals was analyzed by photon correlation spectroscopy (PCS), using a Zetasizer Nano SZ (Malvern Instruments, UK). The results are the hydrodynamic 7
diameter ( z-average, z-ave), which is the intensity weighted mean diameter of the bulk population, and the polydispersity index (PdI), which is a measure for the width of the size distribution. Samples were diluted in water to a suitable concentration and the mean values were calculated from 10 single measurements. 2.3.2.Laser diffraction Potential larger particles or aggregates (> 3-5 µm) which cannot be detected by PCS were investigated by laser diffractrometry (LD) using a Mastersizer 2000 (Malvern Instruments, UK). Samples have been analyzed using the Mie theory and optical parameters - refractive index (RI) and imaginary refractive index (IRI) - were used according to the active being analyzed. Results were expressed by volume weighted diameters 10, 50, 90, 95 and 99%. 2.3.3.Light microscopy To complement laser diffraction regarding the presence of large particles or agglomerates, light microscopy using a microscope Orthoplan (Leitz, Germany) was performed at 160, 600 and 1,000 fold magnifications. 3. Results and discussion 3.1. Rationale for 3 magnetic stirring bars – miniaturized mill design Stirring using 1, 2 and 3 magnetic stirring bars was compared. The objective here was to determine how many stirring bars were necessary to reduce or even completely avoid the presence of a “dead volume” of the suspension not being processed. An optimal stirrer arrangement creates a homogeneous energy dissipation in the milling volume, the movement of all the beads and suspension parts is similar. The presence of a dead volume represents loss in milling efficiency (since part of the suspension is not being processed) and could result in problems regarding reproducibility and up-scaling. This step was realized both visually and by evaluating the particle size of nanosuspensions produced by different milling set-ups. 8
In figure 1, the milling beads were agitated together with pure water (for better visualization of the milling process) using 1, 2 and 3 stirring bars, placed one over the other.
When only 1 stirring bar was used, it is evident that part of the “suspension” was not being agitated together with the milling beads (clearer supernatant) (Fig. 1-left). This happens due to the very high density of the material of which the milling beads are made. It is yttria stabilized zirconium oxide (ZrO2 / Y2O3) and it has a density of 3.8 Kg/L, therefore the beads have a very strong tendency to sediment. When 2 stirring bars where used, the dead volume was macroscopically reduced but still present (Fig. 1-middle). Finally, when 3 stirring bars were used, the rotation power of the 3 stirring bars together was enough to keep all of the beads moving. The complete the suspension was being processed and no dead volume was identified (Fig. 1right). The next step was to use the 3 different milling set-ups (1, 2 or 3 stirring bars) and evaluate the particle size of the resulting nanosuspensions. What had been hypothesized by visually evaluating the different set-ups was confirmed by particle size measurements (Fig. 2-top). Already after 1 hour milling, the PCS mean diameter of a cyclosporin A suspension processed by the miniaturized method using 3 stirring bars was around 550 nm, whereas the size of the particles produced with 1 stirring bar was around 900 nm. Also the smallest PdI after 1 hour processing was found for the set-up with 3 stirring bars (Fig.2-bottom). After 24 hours milling, the set-up containing 3 stirring bars proved to be most efficient and was able to produce the smallest nanocrystals, with 175 nm (approx. half the size produced by the other set-ups) and PdI 0.18, indicating a narrow size distribution. The bead milling process is frequently used for industrial production. The production processes for the cosmetic nanocrystals rutin and hesperidin were developed by our group (now products of Dr. Rimpler GmbH, Germany; INCI: rutin submicron crystals, smartCrystal-lemon extract). The
9
reproducibility of the final crystal size at identical production conditions was found to be very high in the validation process, typically ± 20 nm. Thus the observed difference between the 3 bars (175 nm) and about 350 nm with less bars is clearly significant, even with just one experiment in this study. Therefore, for further investigations, the set-up with 3 stirring bars was chosen, as shown in figure 3 (detailed set-up picture). This set-up also proved not to corrode the stirring bars (checked by weighting all stirrer bars before and after completion of milling process). Corrosion of the stirring bars is often observed when milling using a magnetic stirring plate with vials having just one stirring bar. This erosion is often already macroscopically clearly visible. In the present set up, the surfaces looked perfect without any damage. To assess a potential non-visible erosion, the weight loss was determined. Mean weight of the bars was 793 mg before milling and 793 mg after milling, thus no really detectable loss by weight. The suspensions are thus considered having a sufficiently low contamination to be used for preliminary investigations in product development. Of course they are not fulfilling the requirements of a commercial product for human use, because the typical max. 10 ppm contamination will not be met with milling material coated with standard polymers, such as the bars. The 10 ppm limit is also not reliably detectable by weighing of used bars. 3.2. Effect of active type The miniaturized bead milling method was tested by milling other selected actives with poor solubility, all of which could benefit from the special features of nanocrystals. All of the processed suspensions consisted of 5 % drug and 1 % stabilizer (all w/w), details cf. table 1. The milling process (beads size 0.05 mm) was monitored by PCS measurements and light microscopy. Laser diffractometry (LD) was not possible to be performed due to insufficient amount of sample. Milling during only 1 hour was already enough to reduce the mean particle size of all formulations to the nano range (< 1,000 nm), except for resveratrol, which reached the nano range after 24 hours (Fig. 4).
10
Cyclosporin A was one of the actives milled using this self developed method. In previous studies, cyclosporin A nanosuspensions were obtained by bead milling and high pressure homogenization and had mean diameters of around 200 nm (Nakarani et al., 2010) and 950 nm (Müller et al., 2006), respectively. When a cyclosporin A suspension was processed by the present miniaturized method, the particle size gradually reduced as a function of time during the first 72 hours (Fig. 4-A). The smallest PCS diameter was 93 nm with a PdI of 0.14 (after 72 hours). Light microscopy pictures are in agreement with PCS measurements, showing particle size reduction as a function of milling time. After 24 hours milling, the particles were too small and below the size limit of LM. It was also observed that after 72 hours milling, due to the Tyndall effect, the nanosuspension turned slightly bluish. The antioxidant resveratrol was also investigated. It has been previously formulated as nanocrystals by high pressure homogenization and had a mean particle size of around 200 nm (Kobierski et al., 2011). The same active was processed by the minituarized method and the smallest particle size achieved was 202 nm, PdI 0.46. This size was achieved after 48 hours milling (Fig. 4-B). Here it can be nicely observed that for resveratrol, the miniaturized method could well predict the smallest particle size achievable. Hesperitin was the active which yielded the smallest particle size of all compounds tested. After 24 hours milling, the mean particle size was around 50 nm (Fig. 4-C). This is an extremely small particle size given the production method is a top-down approach. Hesperitin nanocrystals previously produced by HPH had a remarkable bigger mean particle size of around 300 nm (Mishra et al., 2009). This suggests that the maximum reduction in particle size achievable for this active cannot be reached by the production method HPH, despite 20 cycles at 1,500 bar. If smaller particles sizes are required (< 300 nm), bead milling is probably a good production method. It should also be noted that, although particle size was extremely small, the suspension still preserved its physical integrity (absence of gelling/aggregation).
11
Teeranachaideekul et al., 2008, reported ascorbyl palmitate nanocrystals with mean particle size of 365 nm produced by high pressure homogenization. Using the investigated miniaturized method, a mean particle size of 159 nm and PdI index of 0.19 was achieved after 120 hours milling (Fig. 4-D). It should be however noted, that a small size of 279 nm was already achieved after 6 hours of milling. In sections 3.3, 3.4 and 3.5, ascorbyl palmitate was further investigated and the miniaturized production method was compared to the standard production methods. Apigenin was only milled up to 24 hours (Fig. 4-E). Longer milling was not possible because the suspension gelled and the stirring bars were not able to move anymore. When the resulting “gel” was analyzed by PCS, the mean particle size was 1,000 nm. Obviously, the crystals had formed a network comparable to the “pearl network” of e.g. Aerosil. Using the stabilizer solution instead of water for dilution prior to PCS analysis revealed a mean diameter of 80 nm. When diluting with the stabilizer solution, the nanocrystals can be truly measured because the surfactant samples de-aggregates the structure formed in the gel. This information can be very useful when extrapolating these results to bench and industrial scale. Gelling of suspensions is known to occur during bead milling under certain circumstances. A better understanding of what might influence the occurrence of gelling such as stabilizer type/concentration, milling time or critical particle size, prior to upscaling, can be obtained by this miniaturized method. In this case, for apigenin, after 6 hours milling the particle size was 210 nm (PdI was 0.26) and the nanosuspension was still liquid. This information can guide the upscaling process, meaning that when this particle size has been achieved, further milling might represent a risk of gelling. Therefore, the production should be terminated when this particle size (around) has been reached and is sufficient for the envisaged application. However, when smaller crystals are required, milling can be continued and the final gelled product turned again into a liquid suspension by dilution with surfactant solution. Apigenin had been previously formulated as nanosuspension using the smartCrystal technology, which is a combination of bead milling and HPH, and the final particle size achieved was around 400 nm (Al Shaal et al., 2011). 12
Hesperidin reached its smallest particle size after 24 hours, 294 nm with PdI 0.27 (Fig.4-F). Interestingly, this particle size and PdI were practically the same ones obtained for this active when produced in industrial scale by bead milling (Romero et al., 2015a). They reported that after 5 passages in the Bühler PML-2 using the continuous mode (milling chamber of 1.2 L), the mean diameter was 290 nm and PdI was 0.23. This shows that for hesperidin, the data collected by the miniaturized method was coherent with what was obtained when producing on industrial scale. Further milling with the miniaturized method did not further decrease particle size; on the contrary, it caused the nanosuspension to aggregate and particle size increased to 863 nm. Unlike apigenin, dilution with the stabilizer solution did not enable the measurement of individual particles, the aggregation was therefore not reversible. There are obviously two different underlying mechanisms during bead milling of apigenin and hesperidin. For the apigenin suspension, the crystal size reduction continues, and large surface provides particleparticle interaction leading to gelling. The interactions are weak, dilution can reverse the structure leading to separate crystals in a fluid suspension. For the hesperidin suspension, the input of further milling energy (e.g. from 1 to 6 hours) seems to create rather very firm single aggregates, no network-like structure. Dilution with surfactant solution is not able to deaggregate these single aggregates. The different actives showed different maximum reduction in particle size and different particle size reduction velocities (Table 2). This reflects the differences in the lattice structure of each compound, which affects its hardness and consequently, milling patterns. All actives tested could be converted into nanosuspensions, with mean particle sizes ranging from 49 nm until 294 nm.
Based on the data in table 2, three classes of model drugs can be differentiated: a) Easy disintegratable: cyclosporin A, hesperitin and apigenin (minimum size < 100 nm, milling time 24-48 hours).
13
b) Medium disintegratable: resveratrol and ascorbyl palmitate (minimum size > 100 – 200 nm, milling times 48-72 hours). c) Disintegratable: hesperidin (minimum size bout 300 nm, milling time 1 hour). Interesting is also the “kinetics” of size reduction (Fig. 5). Hesperidin reaches 294 nm already after 1 hour; cyclosporin A reaches approximately the same size (346 nm) after 6 hours – despite being the compound yielding one of the smallest sizes of all (93 nm). Hesperitin reaches also the same size as hesperidin (291 nm) after 1 hour. But in contrast to hesperidin, this is not the maximum reachable dispersitivity. Continuation of milling leads to 49 nm. Of course, the rotation speed influences the velocity of size reduction. In general, increase in stirring speed accelerates crystal diminution due to the higher input of energy, but at the same time increases erosion from the production equipment. Using a stirring speed as high as possible is desirable in production lines (e.g. continuous bead mills) to minimize industrial production costs. However, in this study time is no critical factor, critical factor is however to avoid erosion from the milling bars. Very distinct size reduction was achieved for most samples within 1-6 hours (which might be sufficient for many applications, or even better than the smallest possible size), only minimum achievable sizes required 24-72 hours. Increasing speed with potential increased erosion is in this case thus mostly no issue for this lab scale method. The bead size affects also the minimum achievable size, nanocrystal size decreases with decreasing milling bead size. A very rough rule of thumb is that the minimum achievable nanocrystals size is 1,000 less than the bead size, e.g. 0.1 mm beads yield about 0.1 µm nanocrystals. Thus the choice of beads is dependent on the target size of the nanocrystals. The choice also depends on the industrial processability of the beads, that means preferentially, their separation from the nanocrystal suspension. This is much more tedious with 50 µm beads compared to 0.2 mm beads. Thus too small beads should be avoided in formulation development with the proposed small scale unit, having already in mind the ease of industrial production. 14
The small scale production provides also first hints regarding aggregation phenomena during milling, and potential gelling (affected by crystal size). Fig. 4 shows a clear increase after 48 hours (= minimum size) up to 120 hours. Thus in scale up, a fine tuning of optimum milling time needs to be performed. 3.3. Effect of bead size It is known that the particle size achieved by bead milling is related to the size of the milling beads used. The smaller the milling beads, the smaller will be the particle size obtained. Smaller beads mover faster and consequently their capacity to erode the surface and/or break the crystals being milled is stronger. There is a rule of thumb that a 1,000 fold smaller crystal size is obtained than the size of the beads used (i.e. 0.4 mm beads can lead to crystals around 400 nm) However, choosing the smallest beads as possible is not always the smartest option. The same way smaller beads are more powerful in milling the crystals, they also cause a stronger wearing of the equipment itself due to erosion of the parts in contact with the beads. Besides, when the beads as very small, the chance that they are accumulating in undesired sites of the milling machine (such as internal moving parts) is increased, and thus leading to higher risk of damage of the mechanical parts. Very small milling beads could also result in a very fast particle size reduction, reaching very quickly the smallest possible particle size possible for that drug. The excess energy input not being used for particle size reduction is then converted into kinetic energy, making the particles move faster. Faster movement of the particles makes them collide to each other more strongly, which represents a risk of aggregation. Gelling is another issue that could result from very small milling beads. Bellow a certain particle size limit, the suspension, once liquid, becomes more viscous due to a 3-D espacial organization of the excessively small nanocrystals. When the particle size reduces very abruptly, a fine tuning of the size via the milling time is less possible. Milling times differences of some minutes or even seconds can transform a liquid suspension into a viscous gel. Therefore, to find the best cost-benefit relationship of bead size and particle size achieved is so important. When milling an ascorbyl 15
palmitate suspension with the miniaturized method using 3 different milling beads sizes (0.1, 0.2 and 0.4-0.6 mm), the influence of the milling bead size could be verified. For instance, after 6 hours milling, the particle size achieved with milling beads of 0.1 mm was 239 nm, compared to 327 and 411 nm for milling beads of 0.2 and 0.4-0.6 mm, respectively (Figure 6). These bead sizes (0.2 and 0.4-0.6 mm) are the ones generally preferred for lab scale production with the Bühler PML-2. For industrial scale production, the bigger milling beads of 0.4-0.6 mm or bigger are safer to be used. The sizes obtained after 6 hours reflect to some extent the “rough” rule of thumb. However, after 24 hours of milling the size differences are getting reduced. Obviously, the higher milling efficiency of smaller beads can be compensated to some extent by longer milling times with larger beads (e.g. size obtained with 0.4-0.6 mm beads after 24 hours is similar to size with 0.1 mm beads after just 1 hour). For industrial production, the shorter milling time of smaller beads has to be weighed against their more problematic handling (e.g. separation from nanosuspension, easier interference with machine parts such as fittings).
3.4. Super small scale vs. high pressure homogenization Sometimes nanosuspensions are required which are sterile or have at least a very low microbial load. This is more tedious to produce with a bead mill arrangement, thus alternatively nanocrystals can be produced with high pressure homogenization (HPH). HPH yields a sterile product; production can be made in a laminar air flow cabinet. To show the production ability, ascorbyl palmitate was used as example. Previous studies showed that ascorbyl palmitate could be formulated as nanocrystals using HPH (Teeranachaideekul et al., 2008). However, the stabilizer used was Tween 80, a PEG-containing molecule. There is a trend nowadays to avoid PEG-containing ingredients in cosmetics. Therefore, the use a PEG-free stabilizer for stabilization of ascorbyl palmitate nanocrystals for dermal use was investigated using the miniaturized
16
method. It was compared to high pressure homogenization (HPH) using a Micron LAB 40 and standard bead milling using the Bühler PML-2. When the same ascorbyl palmitate formulation was processed by HPH, it was observed that there was a strong particle size reduction to 290 nm after 5 cycles at 1,500 bar and the particle size remained practically unchanged up to 20 cycles (Fig. 7). After 20 cycles, the mean particle size was 254 nm (PdI was 0.15). The approximate same PCS diameter was achieved by the miniaturized milling method after 24 hours milling with 0.4-0.6 mm beads (264 nm). This time was shortened to around 6 hours when milling was performed with 0.05 and 0.1 mm beads (239 and 283 nm, respectively). The maximum particle size reduction was achieved after 120 hours milling with 0.05 mm beads (as previously discussed, the smaller the beads, the more powerful in reducing the particle size they are) and it was 159 nm. Normally 0.05 mm beads are not used in common practice because of the problems such small beads might cause, as previously discussed in section 3.3. In summary, after 24 hours milling with 0.4-0.6 or 0.2 mm and 6 hours with 0.1 mm beads applying the miniaturized milling, a particle size was achieved similar to the one achieved after 5 cycles of HPH. Size results from the small scale cannot be transferred straightforward to HPH, because production parameters in both methods strongly affect the result (e.g. pressure and cycles number in HPH). However, the miniaturized milling with these small beads gives the information of what is the maximum reduction in particle size achievable while still preserving the physical integrity of the nanosuspension. Similar small sizes should also be achievable – at least in most cases – when applying optimal production conditions in HPH.
3.5. Super small scale vs. Bühler PML-2 Ascorbyl palmitate was also milled using the discontinuous mode of a Bühler PML-2 to verify if the data generated by the miniaturized method could be extrapolated to the commercial 17
available equipment. For this, the same ascorbyl palmitate formulation was processed during 60 minutes using a PML-2 with milling beads of 0.2 mm and 0.4-0.6 mm (for the reasons previously discussed in item 3.3). Maximum reduction in particle size with the PML-2 was achieved after 20 minutes milling with 0.2 mm beads being 286 nm (PdI was 0.21), LD diameter 50% was 0.12 µm and the LD 99% was 0.62 µm. Longer milling did not further reduce the particle size (Fig. 8). For the same milling time (20 minutes) with 0.4-0.6 mm beads, the particle size was still 539 nm. After 60 minutes the maximum particle size reduction using this bead size was 442 nm (PdI was 0.39), LD 50% and 99% were 0.15 and 5.3 µm, respectively. Considering the bead size of 0.2 mm, the time necessary for achieving the smallest particle size with the PML-2 (286 nm) was 20 minutes in comparison to approximately 24 hours with the miniaturized method. Obviously, the energy density in the PML-2 is much higher than in the miniaturized method. As outlined above, the smallest achievable size will be similar for miniaturized and large scale, but the time to get this size is mainly determined by the production parameters. On large scale, these parameters are mainly flow through velocity and rotation speed of the agitator, controlling energy density (=energy dissipated per time and dissipation volume).
The physical stability of the bead milled nanosuspensions can be further increased by applying an additional, second processing step: HPH. This step can be made not only on larger scale but also on the proposed miniaturized scale, e.g. processing the bead milled nanosuspension through a small scale high pressure homogenizer, e.g. Avestin B3 with 3 ml batch size.Romero et al., 2015a reported long-term physical and chemical stable hesperidin nanocrystals produced in industrial scale by the smartCrystal® combination technology. The smartCrystal® combination technology (CT) comprises a wet bead milling step followed by high pressure homogenization to produce more stable nanosuspensions with reduced particle size (in most cases) and reduced production time (Petersen, 2006). Therefore, as a first step towards upscaling, the 18
nanosuspension with best results produced using the PML-2 (20 minutes milling with 0.2 mm beads) was further processed by HPH and its short-term physical stability were evaluated, to verify if this is a real formulation candidate and what the best parameters for further upscaling investigations are. After 1 subsequent HPH cycle at 1,500 bar, the mean diameter of the nanosuspension did not change and remained 283 nm, LD 50% was 0.12 µm and LD 99% was 0.57 µm. However, the PdI considerably reduced to 0.13 (compared to 0.21) after initial first bead milling step. This means that the uniformity of the particles increased, which is beneficial for physical stability. A narrower size distribution helps avoid instability issues such as crystals growth by Ostwald ripening (Müller et al., 2001). Samples with and without further HPH processing were stored at 4 and 25 °C (investigations at 40 °C were interrupted after 1 week due to strong degradation of the samples) and evaluated during 6 months. It was observed that for both cases (with and without a HPH step) the particle sizes slightly increased already after 1 week. However this increase was more pronounced in the samples without the HPH step. After the initial increase, the particle sizes remained stable up to 6 months and were around 360 nm for the samples without a HPH step and around 300 nm for the samples with a HPH step (Fig. 9). The samples with a HPH step also showed the smallest PdI’s after 6 months, around 0.13 compared to 0.26.
4. Conclusions It was possible to develop a miniaturized method using only 0.5 mL (approx. 500 mg) suspension and – important – being cost-effective. Based on a 5% or 1% suspension, the 500 mg suspension corresponds to approx. 50 mg or 5 mg of the active, compared to 4 g and 0.4 g required, e.g. for one 40 mL batch in a Micron LAB 40 homogenizer. Most important information from the miniaturized scale are the minimum achievable size of a compound, the diminution kinetics, effect of bead sizes on both minimum size and kinetics and changes potentially occurring in the constitution of the suspension during processing (e.g.
19
aggregates). From this, conclusions can be drawn for defining production parameters on large scale. This accelerates large scale process development. The developed miniaturized method clearly accelerates formulation screening compared to present lab scale and allows estimation of large scale production parameters, and this at simultaneously lowcost and high throughput of test formulations. 5. Acknowledgments The authors would like to thank the CAPES foundation (a federal agency under the Ministry of Education from Brazil) and the German Academic Exchange Service (DAAD) for financial support through the CAPES/DAAD doctoral program (Grant nr. 12416/12-6).
20
6. References Al Shaal, L., Shegokar, R., Müller, R.H., 2011. Production and characterization of antioxidant apigenin nanocrystals as a novel UV skin protective formulation. Int. J. Pharm. 420, 133-140. Keck, C.M., Müller, R.H., 2006. Drug nanocrystals of poorly soluble drugs produced by high pressure homogenisation. Eur. J. Pharm. Biopharm. 62, 3-16. Kipp, J.E., Wong, J.C.T., Doty, M.J., Rebbeck, C.L., 2006. Microprecipitation method for preparing submicron suspensions. US patent 7037528 B2. Kobierski, S., Ofori-Kwakye, K., Muller, R.H., Keck, C.M., 2011. Resveratrol nanosuspensions: interaction of preservatives with nanocrystal production. Die Pharmazie 66, 942-947. Kruss, B., Peters, K., Becker, R., Muller, R.H., 1996. Pharmaceutical nanosuspensions for medicament administration as systems with increased saturation solubility and speed of dissolution. Patent CA2205046 A1. Lemke, A., Moeschwitzer, J., 2007. Method for carefully producing ultrafine particle suspensions and ultrafine particles and use thereof. Patent WO2006108637 A3. Liversidge, G.G., Cundy, K.C., Bishop, J.F., Czekai, D.A., 1991. Surface modified drug nanoparticles. Patent US5145684 A. Mishra, P.R., Al Shaal, L., Muller, R.H., Keck, C.M., 2009. Production and characterization of Hesperetin nanosuspensions for dermal delivery. Int. J. Pharm. 371, 182-189. Müller, R.H., Jacobs, C., Kayser, O., 2001. Nanosuspensions as particulate drug formulations in therapy: Rationale for development and what we can expect for the future. Adv. Drug Delivery Rev. 47, 3-19. Müller, R.H., Möschwitzer, J., 2009. Method and device for producing very fine particles and coating such particles. Patent US20090297565 A1. Müller, R.H., Runge, S., Ravelli, V., Mehnert, W., Thünemann, A.F., Souto, E.B., 2006. Oral bioavailability of cyclosporine: Solid lipid nanoparticles (SLN®) versus drug nanocrystals. Int. J. Pharm. 317, 82-89. Nakarani, M., Patel, P., Patel, J., Patel, P., Murthy, R.S.R., Vaghani, S.S., 2010. Cyclosporine ANanosuspension: Formulation, Characterization and In Vivo Comparison with a Marketed Formulation. Sci. Pharm. 78, 345-361. Parikh, I., Selvaraj, U., 1999. Composition and method of preparing microparticles of water-insoluble substances. Patent US5922355 A. Petersen, R., 2006. Nanocrystals for use in topical cosmetic formulations and method of production thereof. Patent US608866233. Romero, G.B., Chen, R., Keck, C.M., Müller, R.H., 2015a. Industrial concentrates of dermal hesperidin smartCrystals® – production, characterization & long-term stability. Int. J. Pharm. 482, 54-60. Romero, G.B., Keck, C.M., Müller, R.H., Bou-Chacra, N.A., 2015b. Cationic nanocrystal formulation for improved ocular delivery (submitted). Eur. J. Pharm. Biopharm., Submission ID: EJPB-D15-00505.
21
Shegokar, R., Müller, R.H., 2010. Nanocrystals: Industrially feasible multifunctional formulation technology for poorly soluble actives. Int. J. Pharm. 399, 129-139. Teeranachaideekul, V., Junyaprasert, V.B., Souto, E.B., Müller, R.H., 2008. Development of ascorbyl palmitate nanocrystals applying the nanosuspension technology. Int. J. Pharm. 354, 227-234. Van Eerdenbrugh, B., Stuyven, B., Froyen, L., Van Humbeeck, J., Martens, J.A., Augustijns, P., Van den Mooter, G., 2009. Downscaling Drug Nanosuspension Production: Processing Aspects and Physicochemical Characterization. AAPS PharmSciTech 10, 44-53.
22
Figure Captions Figure 1. Left: milling vial with 1 stirring bar; middle: milling vial with 2 stirring bars; right: milling vial with 3 stirring bars, which showed no dead volume (milling beads size: 0.05 mm; rotation speed: 1,200 rpm).
23
Figure 2. PCS diameter (top) and polydispersity indices (PdI) (bottom) of a cyclosporin A nanosuspension after milling with different number of stirring bars during 1-24 hours (0.05 mm beads and 1,200 rpm).
24
Figure 3. Left: schematic representation of the miniaturized method; right: picture of the experimental set-up without the milling beads and suspension.
25
Figure 4. Particle size as a function of milling time using the miniaturized method for 6 different drugs. A: cyclosporin A; B: resveratrol; C: hesperitin; D: ascorbyl palmitate; E: apigenin; F: hesperidin; (milling time 120 hours; beads size: 0.05 mm). For resveratrol (B), the y-axis is extended. *due to aggregation or gelling, the milling time for apigenin and hesperidin was 24 hours.
26
Figure 5. Principle of size reduction kinetics during milling as observed for the model actives hesperidin (Hd), hesperitin (Ht) and cyclosporin A (Cyc A).
27
Figure 6. Effect of different sized milling beads during 24 hours milling using the miniaturized set-up for ascorbyl palmitate.
28
Figure 7. Particle size reduction of ascorbyl palmitate as a function of high pressure homogenization cycles at 1,500 bar.
29
Figure 8. Particle size of ascorbyl palmitate as a function of milling time using a mill PML-2 (discontinuous mode) and two different bead sizes (0.2 and 0.4-0.6 mm) during 60 minutes.
30
Figure 9. PCS diameter of ascorbyl palmitate nanosuspensions produced with the PML-2 with and without a HPH step, after 6 months storage at 4 and 25 °C.
31
Tables Table 1. Actives and their respective stabilizer processed by the miniaturized milling method (alkyl polyglucoside C8-C10 = Plantacare® 2000 UP). ACTIVE Cyclosporin A Resveratrol Hesperitin Ascorbyl palmitate Apigenin Hesperidin
STABILIZER TPGS alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 Poloxamer 188
32
Table 2. Overview of the size reduction ability (size after 1 hour milling, smallest achievable size) and constitution (aggregation, gelling) during milling. SMALLEST SIZE PCS size after NO AGGREGATION/ GELLING/ MODEL DRUG size hours of 1 hour (nm) AGGREGATION SIZE (nm) SIZE (nm) (nm) milling cyclosporin A resveratrol hesperitin ascorbyl palmitate apigenin hesperidin
668 5,057 291 749 238 294
97 202 49 171 85 294
48 h 48 h 24 h 72 h 24 h 1h
Ø + / 729 + / 97 Ø + / 85 + / 844
33
S0378-5173(15)30394-X http://dx.doi.org/doi:10.1016/j.ijpharm.2015.11.047 IJP 15385
To appear in:
International Journal of Pharmaceutics
Received date: Revised date: Accepted date:
28-7-2015 25-11-2015 26-11-2015
Please cite this article as: Romero, Gregori B., Keck, Cornelia M., M¨uller, Rainer H., Simple low-cost miniaturization approach for pharmaceutical nanocrystals production.International Journal of Pharmaceutics http://dx.doi.org/10.1016/j.ijpharm.2015.11.047 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Simple low-cost miniaturization approach for pharmaceutical nanocrystals production Gregori B. Romeroa [email protected] [email protected], Cornelia M. Keckb, Rainer H. Müllera* [email protected] a
Pharmaceutics, Pharmaceutical Nanotechnology & NutriCosmetics, Freie Universität Berlin, Kelchstr. 31,
12169 Berlin, Germany b
Applied Pharmacy, University of Applied Sciences Kaiserslautern, Campus Pirmasens, Carl-Schurz-Str. 10-16,
66953 Pirmasens, Germany *
Corresponding author: Tel.: +49 30 838 506 78.
1
Graphical Abstract
2
Abstract Systematic screening for optimal formulation composition and production parameters for nanosuspensions consumes a lot of time and also drug material when performed at lab scale. Therefore, a cost-effective miniaturized scale top down approach for nanocrystals production by wet bead milling was developed. The final set-up consisted of 3 magnetic stirring bars placed vertically one over the other in a 2 mL glass vial and agitated by a common magnetic stirring plate. All of the tested actives (cyclosporin A, resveratrol, hesperitin, ascorbyl palmitate, apigenin and hesperidin) could be converted to nanosuspensions. For 4 of them, the particles sizes achieved were smaller than previously reported on the literature (around 90 nm for cyclosporin A; 50 nm for hesperitin; 160 nm for ascorbyl palmitate and 80 nm for apigenin). The “transferability” of the data collect by the miniaturized method was evaluated comparing the production at larger scale using both wet bead milling and high pressure homogenization. Transferable information obtained from the miniaturized scale is minimum achievable size, improvements in size reduction by reduction of beads size, diminution kinetics and potentially occurring instabilities during processing. The small scale batches also allow identification of optimal stabilizer types and concentrations. The batch size is 0.5 mL, requiring approximately 50 mg or 5 mg of drug (5% and 1% suspension, respectively). Thus, a simple, accessible, low-cost miniaturized scale method for the production of pharmaceutical nanocrystals was established. Keywords: nanocrystals; nanosuspension; nanoparticle; miniaturization; downscaling; bead milling
3
1. Introduction Nanocrystals are meanwhile well established for the formulation of poorly soluble actives, preferentially for oral (Keck and Müller, 2006) and dermal delivery (Shegokar and Müller, 2010). Products are on the market in pharma (e.g. Rapamune, Tricor), but also in cosmetics (e.g. platinum rare by la prairie). The typical size is above 100 nm (but below 1,000 nm), thus they are no nanomaterial/nanoparticle product according to e.g. European Commission recommendation on the definition of nanomaterial (2011/696/EU) and FDA guidance “Considering Whether an FDARegulated Product Involves the Application of Nanotechnology”. This fact eases product registration. In addition, this “submicron” size range is more easily accessible by the various production methods used. The most important industrially used methods are bead milling (Alkermes, prev. élan/Nanosystems) (Liversidge et al., 1991) and high pressure homogenization (HPH) (SkyePharma PLC) (Kruss et al., 1996; Parikh and Selvaraj, 1999). There are also combination methods described consisting of a pre-treatment step followed by a main step of crystal disintegration, e.g. NANOEDGE technology by Baxter (Kipp et al., 2006); H69 process: spray drying and subsequent HPH (Müller and Möschwitzer, 2009); H96 process: lyophilisation and subsequent HPH (Lemke and Moeschwitzer, 2007). Here often one aim is to access also the particle size range below 100 nm (H69, H96). For assessing the comminution ability of crystals, formulation screening (e.g. type and concentrations of stabilizers, stabilizer mixtures) and first physical stability investigations, a downscaling to small volumes is desirable. This saves time, and often very important, saves active. This is especially important in case of new chemical entities with potentially very limited amount available (often rather milligrams than grams). Commercial bead mills in their small volume version have suspension volumes rather in the range 50 ml to 100 ml (e.g. Bühler PML-2, 200 ml chamber and required suspension volume about 120 ml). Assuming a 5% concentration of active, a density of 1.5 g/ml, about 10 g of active are required. Most of the high pressure homogenizers require 40-200 ml minimum volume, i.e. at least about 3 g of active (e.g. Micron LAB 40, APV Deutschland,
4
Germany: 40 ml). There are also homogenizers with smaller volumes (e.g. 3-5 ml), but the process parameter pressure cannot be controlled very precisely (e.g. Avestin EmulsifFlex-B15, minimum volume 3 mL). One approach could be to reduce the drug concentration (e.g. to 1%), but this makes the milling process less effective. The drug crystals move against each other during the milling process, which contributes to the comminution efficiency. Thus rather about 20% suspension concentration is ideal. Independent on the concentration aspect, all these instruments require a certain processing time, which also might be shortened by using a small scale “multiple” system (i.e. small scale and running many samples in parallel). Thus there is clearly a need for effective down scaling, i.e. having a small volume and being at the same time cost-effective. A very simple approach in bead milling is filling of a 20 ml injection vial with milling beads, adding a magnetic stirrer bar and the suspension and placing it on a magnetic stirrer plate. However, this milling process is not very controlled, sometimes not effective due to uncontrolled movement of the stirring bar. In case of coated stirrer bars, erosion from the polymer coat can take place. To have a more controlled process, e.g. a glass-vial-based system has been described, where vials with different volumes were placed in a Retsch PM 400 MA planetary mill (Van Eerdenbrugh et al., 2009). However, this requires the investment of a mill, or a few mills (20 vials per holder in the mill). In parallel, milling in a 96-well plate was investigated, fixed on a shaker (25-340 µl/well) (Van Eerdenbrugh et al., 2009). In case of plates made from plastic material, erosion from the walls cannot be excluded. The suspensions used were about 16%, and drug amounts down to 1 mg could be processed. In this study, a down scaled system was developed based on the preferable glass vials, having a special optimized arrangement of the stirrer bars, and avoiding the use of a planetary mill by using a multiple magnetic stirring plate. Systematically investigated were the effect of number and arrangement of stirring bars, the effect of the type of active to be diminuted, and effect of milling bead size in this miniaturized system. The results obtained from the small scale were compared to
5
data from larger scale high pressure homogenization and the Bühler bead mill PML-2, to judge the transferability of data from miniaturized to larger scale. 2. Materials and methods 2.1. Materials Resveratrol, hesperitin, ascorbyl palmitate, apigenin and hesperidin were purchased from Denk Ingredients GmbH (Germany). Cyclosporin A was a donation from PharmaSol GmbH (Germany). Vitamin E polyethylene glycol succinate (TPGS) (trade name Kolliphor® TPGS), alkyl polyglucoside C8-C10 (trade name Plantacare® 2000 UP) and poloxamer 188 (trade name Kolliphor® P 188) were donations from BASF SE (Germany). Double distilled and ultrapurified water was obtained from a Milli-Q apparatus (Millipore GmbH, Germany). All other reagents were from analytical grade. 2.2. Nanosuspensions production 2.2.1. Miniaturized wet bead milling using a magnetic stirring plate The production principle was a top-down approach, in this case wet bead milling with stirring bars in a super reduced scale. The milling chamber consisted of a 2 mL glass vial (i.e. 10x lower than the namely used 20 mL injection vials). The final set-up was composed of 3 cylindrical stirring bars (9.5 x 6 mm) (VWR International, Germany), disposed vertically one over the other (Fig. 3). The volume of the 3 stirring bars was 0.7 mL (0.233 mL/stirring bar). From the remaining space (1.3 mL), 1 mL was destined for the milling beads and the suspensions of the raw powder of the actives (i.e. 0.3 mL headspace). The milling beads possessed various sizes (diameters of 0.05, 0.1, 0.2 and 0.4-0.6 mm) and were yttria stabilized zirconium oxide beads (Hosokawa Alpine, Germany).
6
All the formulations investigated were aqueous suspensions and contained 5% active and 1% stabilizer (all w/w). The proportion of beads to suspension was 1:1 (volume), in other words, 0.5 mL (1.9 g) of milling beads (density 3.8 Kg/L) and 0.5 mL of the suspension. The final 0.3 mL was left empty as headspace. The actives used and their respective stabilizers are shown in table 1. The stabilizer type for each active was selected based on previous experiences producing nanosuspensions.
The vials were stirred on a magnetic stirring plate RCT basic (IKA-Werke GmbH & Co. KG, Germany) at 1,200 rpm and 5 oC. Samples were drawn after defined intervals up to 120 hours. 2.2.2. Bench scale wet bead milling using a Bühler PML-2 The bench scale production by wet bead milling was performed using a bead mill PML-2 (Bühler, Switzerland). Milling beads of 0.2 or 0.4-0.6 mm identical to the ones described in 2.2.1 were used. The milling time was up to 60 minutes, at a speed of 2,000 rpm and 5°C. Samples were drawn after 10, 20, 30 and 60 minutes of processing. 2.2.3. Bench scale high pressure homogenization with a Micron LAB 40 The bench scale production by high pressure homogenization was performed using a homogenizer Micron LAB 40 (APV Deutschland GmbH, Germany). The suspension was first pre-processed by 2 high pressure homogenization (HPH) cycles at 150, 500 and 1,000 bar, respectively, followed by 20 cycles at 1,500 bar. 2.3. Particle characterization 2.3.1.Photo correlation spectroscopy The particle size of the nanocrystals was analyzed by photon correlation spectroscopy (PCS), using a Zetasizer Nano SZ (Malvern Instruments, UK). The results are the hydrodynamic 7
diameter ( z-average, z-ave), which is the intensity weighted mean diameter of the bulk population, and the polydispersity index (PdI), which is a measure for the width of the size distribution. Samples were diluted in water to a suitable concentration and the mean values were calculated from 10 single measurements. 2.3.2.Laser diffraction Potential larger particles or aggregates (> 3-5 µm) which cannot be detected by PCS were investigated by laser diffractrometry (LD) using a Mastersizer 2000 (Malvern Instruments, UK). Samples have been analyzed using the Mie theory and optical parameters - refractive index (RI) and imaginary refractive index (IRI) - were used according to the active being analyzed. Results were expressed by volume weighted diameters 10, 50, 90, 95 and 99%. 2.3.3.Light microscopy To complement laser diffraction regarding the presence of large particles or agglomerates, light microscopy using a microscope Orthoplan (Leitz, Germany) was performed at 160, 600 and 1,000 fold magnifications. 3. Results and discussion 3.1. Rationale for 3 magnetic stirring bars – miniaturized mill design Stirring using 1, 2 and 3 magnetic stirring bars was compared. The objective here was to determine how many stirring bars were necessary to reduce or even completely avoid the presence of a “dead volume” of the suspension not being processed. An optimal stirrer arrangement creates a homogeneous energy dissipation in the milling volume, the movement of all the beads and suspension parts is similar. The presence of a dead volume represents loss in milling efficiency (since part of the suspension is not being processed) and could result in problems regarding reproducibility and up-scaling. This step was realized both visually and by evaluating the particle size of nanosuspensions produced by different milling set-ups. 8
In figure 1, the milling beads were agitated together with pure water (for better visualization of the milling process) using 1, 2 and 3 stirring bars, placed one over the other.
When only 1 stirring bar was used, it is evident that part of the “suspension” was not being agitated together with the milling beads (clearer supernatant) (Fig. 1-left). This happens due to the very high density of the material of which the milling beads are made. It is yttria stabilized zirconium oxide (ZrO2 / Y2O3) and it has a density of 3.8 Kg/L, therefore the beads have a very strong tendency to sediment. When 2 stirring bars where used, the dead volume was macroscopically reduced but still present (Fig. 1-middle). Finally, when 3 stirring bars were used, the rotation power of the 3 stirring bars together was enough to keep all of the beads moving. The complete the suspension was being processed and no dead volume was identified (Fig. 1right). The next step was to use the 3 different milling set-ups (1, 2 or 3 stirring bars) and evaluate the particle size of the resulting nanosuspensions. What had been hypothesized by visually evaluating the different set-ups was confirmed by particle size measurements (Fig. 2-top). Already after 1 hour milling, the PCS mean diameter of a cyclosporin A suspension processed by the miniaturized method using 3 stirring bars was around 550 nm, whereas the size of the particles produced with 1 stirring bar was around 900 nm. Also the smallest PdI after 1 hour processing was found for the set-up with 3 stirring bars (Fig.2-bottom). After 24 hours milling, the set-up containing 3 stirring bars proved to be most efficient and was able to produce the smallest nanocrystals, with 175 nm (approx. half the size produced by the other set-ups) and PdI 0.18, indicating a narrow size distribution. The bead milling process is frequently used for industrial production. The production processes for the cosmetic nanocrystals rutin and hesperidin were developed by our group (now products of Dr. Rimpler GmbH, Germany; INCI: rutin submicron crystals, smartCrystal-lemon extract). The
9
reproducibility of the final crystal size at identical production conditions was found to be very high in the validation process, typically ± 20 nm. Thus the observed difference between the 3 bars (175 nm) and about 350 nm with less bars is clearly significant, even with just one experiment in this study. Therefore, for further investigations, the set-up with 3 stirring bars was chosen, as shown in figure 3 (detailed set-up picture). This set-up also proved not to corrode the stirring bars (checked by weighting all stirrer bars before and after completion of milling process). Corrosion of the stirring bars is often observed when milling using a magnetic stirring plate with vials having just one stirring bar. This erosion is often already macroscopically clearly visible. In the present set up, the surfaces looked perfect without any damage. To assess a potential non-visible erosion, the weight loss was determined. Mean weight of the bars was 793 mg before milling and 793 mg after milling, thus no really detectable loss by weight. The suspensions are thus considered having a sufficiently low contamination to be used for preliminary investigations in product development. Of course they are not fulfilling the requirements of a commercial product for human use, because the typical max. 10 ppm contamination will not be met with milling material coated with standard polymers, such as the bars. The 10 ppm limit is also not reliably detectable by weighing of used bars. 3.2. Effect of active type The miniaturized bead milling method was tested by milling other selected actives with poor solubility, all of which could benefit from the special features of nanocrystals. All of the processed suspensions consisted of 5 % drug and 1 % stabilizer (all w/w), details cf. table 1. The milling process (beads size 0.05 mm) was monitored by PCS measurements and light microscopy. Laser diffractometry (LD) was not possible to be performed due to insufficient amount of sample. Milling during only 1 hour was already enough to reduce the mean particle size of all formulations to the nano range (< 1,000 nm), except for resveratrol, which reached the nano range after 24 hours (Fig. 4).
10
Cyclosporin A was one of the actives milled using this self developed method. In previous studies, cyclosporin A nanosuspensions were obtained by bead milling and high pressure homogenization and had mean diameters of around 200 nm (Nakarani et al., 2010) and 950 nm (Müller et al., 2006), respectively. When a cyclosporin A suspension was processed by the present miniaturized method, the particle size gradually reduced as a function of time during the first 72 hours (Fig. 4-A). The smallest PCS diameter was 93 nm with a PdI of 0.14 (after 72 hours). Light microscopy pictures are in agreement with PCS measurements, showing particle size reduction as a function of milling time. After 24 hours milling, the particles were too small and below the size limit of LM. It was also observed that after 72 hours milling, due to the Tyndall effect, the nanosuspension turned slightly bluish. The antioxidant resveratrol was also investigated. It has been previously formulated as nanocrystals by high pressure homogenization and had a mean particle size of around 200 nm (Kobierski et al., 2011). The same active was processed by the minituarized method and the smallest particle size achieved was 202 nm, PdI 0.46. This size was achieved after 48 hours milling (Fig. 4-B). Here it can be nicely observed that for resveratrol, the miniaturized method could well predict the smallest particle size achievable. Hesperitin was the active which yielded the smallest particle size of all compounds tested. After 24 hours milling, the mean particle size was around 50 nm (Fig. 4-C). This is an extremely small particle size given the production method is a top-down approach. Hesperitin nanocrystals previously produced by HPH had a remarkable bigger mean particle size of around 300 nm (Mishra et al., 2009). This suggests that the maximum reduction in particle size achievable for this active cannot be reached by the production method HPH, despite 20 cycles at 1,500 bar. If smaller particles sizes are required (< 300 nm), bead milling is probably a good production method. It should also be noted that, although particle size was extremely small, the suspension still preserved its physical integrity (absence of gelling/aggregation).
11
Teeranachaideekul et al., 2008, reported ascorbyl palmitate nanocrystals with mean particle size of 365 nm produced by high pressure homogenization. Using the investigated miniaturized method, a mean particle size of 159 nm and PdI index of 0.19 was achieved after 120 hours milling (Fig. 4-D). It should be however noted, that a small size of 279 nm was already achieved after 6 hours of milling. In sections 3.3, 3.4 and 3.5, ascorbyl palmitate was further investigated and the miniaturized production method was compared to the standard production methods. Apigenin was only milled up to 24 hours (Fig. 4-E). Longer milling was not possible because the suspension gelled and the stirring bars were not able to move anymore. When the resulting “gel” was analyzed by PCS, the mean particle size was 1,000 nm. Obviously, the crystals had formed a network comparable to the “pearl network” of e.g. Aerosil. Using the stabilizer solution instead of water for dilution prior to PCS analysis revealed a mean diameter of 80 nm. When diluting with the stabilizer solution, the nanocrystals can be truly measured because the surfactant samples de-aggregates the structure formed in the gel. This information can be very useful when extrapolating these results to bench and industrial scale. Gelling of suspensions is known to occur during bead milling under certain circumstances. A better understanding of what might influence the occurrence of gelling such as stabilizer type/concentration, milling time or critical particle size, prior to upscaling, can be obtained by this miniaturized method. In this case, for apigenin, after 6 hours milling the particle size was 210 nm (PdI was 0.26) and the nanosuspension was still liquid. This information can guide the upscaling process, meaning that when this particle size has been achieved, further milling might represent a risk of gelling. Therefore, the production should be terminated when this particle size (around) has been reached and is sufficient for the envisaged application. However, when smaller crystals are required, milling can be continued and the final gelled product turned again into a liquid suspension by dilution with surfactant solution. Apigenin had been previously formulated as nanosuspension using the smartCrystal technology, which is a combination of bead milling and HPH, and the final particle size achieved was around 400 nm (Al Shaal et al., 2011). 12
Hesperidin reached its smallest particle size after 24 hours, 294 nm with PdI 0.27 (Fig.4-F). Interestingly, this particle size and PdI were practically the same ones obtained for this active when produced in industrial scale by bead milling (Romero et al., 2015a). They reported that after 5 passages in the Bühler PML-2 using the continuous mode (milling chamber of 1.2 L), the mean diameter was 290 nm and PdI was 0.23. This shows that for hesperidin, the data collected by the miniaturized method was coherent with what was obtained when producing on industrial scale. Further milling with the miniaturized method did not further decrease particle size; on the contrary, it caused the nanosuspension to aggregate and particle size increased to 863 nm. Unlike apigenin, dilution with the stabilizer solution did not enable the measurement of individual particles, the aggregation was therefore not reversible. There are obviously two different underlying mechanisms during bead milling of apigenin and hesperidin. For the apigenin suspension, the crystal size reduction continues, and large surface provides particleparticle interaction leading to gelling. The interactions are weak, dilution can reverse the structure leading to separate crystals in a fluid suspension. For the hesperidin suspension, the input of further milling energy (e.g. from 1 to 6 hours) seems to create rather very firm single aggregates, no network-like structure. Dilution with surfactant solution is not able to deaggregate these single aggregates. The different actives showed different maximum reduction in particle size and different particle size reduction velocities (Table 2). This reflects the differences in the lattice structure of each compound, which affects its hardness and consequently, milling patterns. All actives tested could be converted into nanosuspensions, with mean particle sizes ranging from 49 nm until 294 nm.
Based on the data in table 2, three classes of model drugs can be differentiated: a) Easy disintegratable: cyclosporin A, hesperitin and apigenin (minimum size < 100 nm, milling time 24-48 hours).
13
b) Medium disintegratable: resveratrol and ascorbyl palmitate (minimum size > 100 – 200 nm, milling times 48-72 hours). c) Disintegratable: hesperidin (minimum size bout 300 nm, milling time 1 hour). Interesting is also the “kinetics” of size reduction (Fig. 5). Hesperidin reaches 294 nm already after 1 hour; cyclosporin A reaches approximately the same size (346 nm) after 6 hours – despite being the compound yielding one of the smallest sizes of all (93 nm). Hesperitin reaches also the same size as hesperidin (291 nm) after 1 hour. But in contrast to hesperidin, this is not the maximum reachable dispersitivity. Continuation of milling leads to 49 nm. Of course, the rotation speed influences the velocity of size reduction. In general, increase in stirring speed accelerates crystal diminution due to the higher input of energy, but at the same time increases erosion from the production equipment. Using a stirring speed as high as possible is desirable in production lines (e.g. continuous bead mills) to minimize industrial production costs. However, in this study time is no critical factor, critical factor is however to avoid erosion from the milling bars. Very distinct size reduction was achieved for most samples within 1-6 hours (which might be sufficient for many applications, or even better than the smallest possible size), only minimum achievable sizes required 24-72 hours. Increasing speed with potential increased erosion is in this case thus mostly no issue for this lab scale method. The bead size affects also the minimum achievable size, nanocrystal size decreases with decreasing milling bead size. A very rough rule of thumb is that the minimum achievable nanocrystals size is 1,000 less than the bead size, e.g. 0.1 mm beads yield about 0.1 µm nanocrystals. Thus the choice of beads is dependent on the target size of the nanocrystals. The choice also depends on the industrial processability of the beads, that means preferentially, their separation from the nanocrystal suspension. This is much more tedious with 50 µm beads compared to 0.2 mm beads. Thus too small beads should be avoided in formulation development with the proposed small scale unit, having already in mind the ease of industrial production. 14
The small scale production provides also first hints regarding aggregation phenomena during milling, and potential gelling (affected by crystal size). Fig. 4 shows a clear increase after 48 hours (= minimum size) up to 120 hours. Thus in scale up, a fine tuning of optimum milling time needs to be performed. 3.3. Effect of bead size It is known that the particle size achieved by bead milling is related to the size of the milling beads used. The smaller the milling beads, the smaller will be the particle size obtained. Smaller beads mover faster and consequently their capacity to erode the surface and/or break the crystals being milled is stronger. There is a rule of thumb that a 1,000 fold smaller crystal size is obtained than the size of the beads used (i.e. 0.4 mm beads can lead to crystals around 400 nm) However, choosing the smallest beads as possible is not always the smartest option. The same way smaller beads are more powerful in milling the crystals, they also cause a stronger wearing of the equipment itself due to erosion of the parts in contact with the beads. Besides, when the beads as very small, the chance that they are accumulating in undesired sites of the milling machine (such as internal moving parts) is increased, and thus leading to higher risk of damage of the mechanical parts. Very small milling beads could also result in a very fast particle size reduction, reaching very quickly the smallest possible particle size possible for that drug. The excess energy input not being used for particle size reduction is then converted into kinetic energy, making the particles move faster. Faster movement of the particles makes them collide to each other more strongly, which represents a risk of aggregation. Gelling is another issue that could result from very small milling beads. Bellow a certain particle size limit, the suspension, once liquid, becomes more viscous due to a 3-D espacial organization of the excessively small nanocrystals. When the particle size reduces very abruptly, a fine tuning of the size via the milling time is less possible. Milling times differences of some minutes or even seconds can transform a liquid suspension into a viscous gel. Therefore, to find the best cost-benefit relationship of bead size and particle size achieved is so important. When milling an ascorbyl 15
palmitate suspension with the miniaturized method using 3 different milling beads sizes (0.1, 0.2 and 0.4-0.6 mm), the influence of the milling bead size could be verified. For instance, after 6 hours milling, the particle size achieved with milling beads of 0.1 mm was 239 nm, compared to 327 and 411 nm for milling beads of 0.2 and 0.4-0.6 mm, respectively (Figure 6). These bead sizes (0.2 and 0.4-0.6 mm) are the ones generally preferred for lab scale production with the Bühler PML-2. For industrial scale production, the bigger milling beads of 0.4-0.6 mm or bigger are safer to be used. The sizes obtained after 6 hours reflect to some extent the “rough” rule of thumb. However, after 24 hours of milling the size differences are getting reduced. Obviously, the higher milling efficiency of smaller beads can be compensated to some extent by longer milling times with larger beads (e.g. size obtained with 0.4-0.6 mm beads after 24 hours is similar to size with 0.1 mm beads after just 1 hour). For industrial production, the shorter milling time of smaller beads has to be weighed against their more problematic handling (e.g. separation from nanosuspension, easier interference with machine parts such as fittings).
3.4. Super small scale vs. high pressure homogenization Sometimes nanosuspensions are required which are sterile or have at least a very low microbial load. This is more tedious to produce with a bead mill arrangement, thus alternatively nanocrystals can be produced with high pressure homogenization (HPH). HPH yields a sterile product; production can be made in a laminar air flow cabinet. To show the production ability, ascorbyl palmitate was used as example. Previous studies showed that ascorbyl palmitate could be formulated as nanocrystals using HPH (Teeranachaideekul et al., 2008). However, the stabilizer used was Tween 80, a PEG-containing molecule. There is a trend nowadays to avoid PEG-containing ingredients in cosmetics. Therefore, the use a PEG-free stabilizer for stabilization of ascorbyl palmitate nanocrystals for dermal use was investigated using the miniaturized
16
method. It was compared to high pressure homogenization (HPH) using a Micron LAB 40 and standard bead milling using the Bühler PML-2. When the same ascorbyl palmitate formulation was processed by HPH, it was observed that there was a strong particle size reduction to 290 nm after 5 cycles at 1,500 bar and the particle size remained practically unchanged up to 20 cycles (Fig. 7). After 20 cycles, the mean particle size was 254 nm (PdI was 0.15). The approximate same PCS diameter was achieved by the miniaturized milling method after 24 hours milling with 0.4-0.6 mm beads (264 nm). This time was shortened to around 6 hours when milling was performed with 0.05 and 0.1 mm beads (239 and 283 nm, respectively). The maximum particle size reduction was achieved after 120 hours milling with 0.05 mm beads (as previously discussed, the smaller the beads, the more powerful in reducing the particle size they are) and it was 159 nm. Normally 0.05 mm beads are not used in common practice because of the problems such small beads might cause, as previously discussed in section 3.3. In summary, after 24 hours milling with 0.4-0.6 or 0.2 mm and 6 hours with 0.1 mm beads applying the miniaturized milling, a particle size was achieved similar to the one achieved after 5 cycles of HPH. Size results from the small scale cannot be transferred straightforward to HPH, because production parameters in both methods strongly affect the result (e.g. pressure and cycles number in HPH). However, the miniaturized milling with these small beads gives the information of what is the maximum reduction in particle size achievable while still preserving the physical integrity of the nanosuspension. Similar small sizes should also be achievable – at least in most cases – when applying optimal production conditions in HPH.
3.5. Super small scale vs. Bühler PML-2 Ascorbyl palmitate was also milled using the discontinuous mode of a Bühler PML-2 to verify if the data generated by the miniaturized method could be extrapolated to the commercial 17
available equipment. For this, the same ascorbyl palmitate formulation was processed during 60 minutes using a PML-2 with milling beads of 0.2 mm and 0.4-0.6 mm (for the reasons previously discussed in item 3.3). Maximum reduction in particle size with the PML-2 was achieved after 20 minutes milling with 0.2 mm beads being 286 nm (PdI was 0.21), LD diameter 50% was 0.12 µm and the LD 99% was 0.62 µm. Longer milling did not further reduce the particle size (Fig. 8). For the same milling time (20 minutes) with 0.4-0.6 mm beads, the particle size was still 539 nm. After 60 minutes the maximum particle size reduction using this bead size was 442 nm (PdI was 0.39), LD 50% and 99% were 0.15 and 5.3 µm, respectively. Considering the bead size of 0.2 mm, the time necessary for achieving the smallest particle size with the PML-2 (286 nm) was 20 minutes in comparison to approximately 24 hours with the miniaturized method. Obviously, the energy density in the PML-2 is much higher than in the miniaturized method. As outlined above, the smallest achievable size will be similar for miniaturized and large scale, but the time to get this size is mainly determined by the production parameters. On large scale, these parameters are mainly flow through velocity and rotation speed of the agitator, controlling energy density (=energy dissipated per time and dissipation volume).
The physical stability of the bead milled nanosuspensions can be further increased by applying an additional, second processing step: HPH. This step can be made not only on larger scale but also on the proposed miniaturized scale, e.g. processing the bead milled nanosuspension through a small scale high pressure homogenizer, e.g. Avestin B3 with 3 ml batch size.Romero et al., 2015a reported long-term physical and chemical stable hesperidin nanocrystals produced in industrial scale by the smartCrystal® combination technology. The smartCrystal® combination technology (CT) comprises a wet bead milling step followed by high pressure homogenization to produce more stable nanosuspensions with reduced particle size (in most cases) and reduced production time (Petersen, 2006). Therefore, as a first step towards upscaling, the 18
nanosuspension with best results produced using the PML-2 (20 minutes milling with 0.2 mm beads) was further processed by HPH and its short-term physical stability were evaluated, to verify if this is a real formulation candidate and what the best parameters for further upscaling investigations are. After 1 subsequent HPH cycle at 1,500 bar, the mean diameter of the nanosuspension did not change and remained 283 nm, LD 50% was 0.12 µm and LD 99% was 0.57 µm. However, the PdI considerably reduced to 0.13 (compared to 0.21) after initial first bead milling step. This means that the uniformity of the particles increased, which is beneficial for physical stability. A narrower size distribution helps avoid instability issues such as crystals growth by Ostwald ripening (Müller et al., 2001). Samples with and without further HPH processing were stored at 4 and 25 °C (investigations at 40 °C were interrupted after 1 week due to strong degradation of the samples) and evaluated during 6 months. It was observed that for both cases (with and without a HPH step) the particle sizes slightly increased already after 1 week. However this increase was more pronounced in the samples without the HPH step. After the initial increase, the particle sizes remained stable up to 6 months and were around 360 nm for the samples without a HPH step and around 300 nm for the samples with a HPH step (Fig. 9). The samples with a HPH step also showed the smallest PdI’s after 6 months, around 0.13 compared to 0.26.
4. Conclusions It was possible to develop a miniaturized method using only 0.5 mL (approx. 500 mg) suspension and – important – being cost-effective. Based on a 5% or 1% suspension, the 500 mg suspension corresponds to approx. 50 mg or 5 mg of the active, compared to 4 g and 0.4 g required, e.g. for one 40 mL batch in a Micron LAB 40 homogenizer. Most important information from the miniaturized scale are the minimum achievable size of a compound, the diminution kinetics, effect of bead sizes on both minimum size and kinetics and changes potentially occurring in the constitution of the suspension during processing (e.g.
19
aggregates). From this, conclusions can be drawn for defining production parameters on large scale. This accelerates large scale process development. The developed miniaturized method clearly accelerates formulation screening compared to present lab scale and allows estimation of large scale production parameters, and this at simultaneously lowcost and high throughput of test formulations. 5. Acknowledgments The authors would like to thank the CAPES foundation (a federal agency under the Ministry of Education from Brazil) and the German Academic Exchange Service (DAAD) for financial support through the CAPES/DAAD doctoral program (Grant nr. 12416/12-6).
20
6. References Al Shaal, L., Shegokar, R., Müller, R.H., 2011. Production and characterization of antioxidant apigenin nanocrystals as a novel UV skin protective formulation. Int. J. Pharm. 420, 133-140. Keck, C.M., Müller, R.H., 2006. Drug nanocrystals of poorly soluble drugs produced by high pressure homogenisation. Eur. J. Pharm. Biopharm. 62, 3-16. Kipp, J.E., Wong, J.C.T., Doty, M.J., Rebbeck, C.L., 2006. Microprecipitation method for preparing submicron suspensions. US patent 7037528 B2. Kobierski, S., Ofori-Kwakye, K., Muller, R.H., Keck, C.M., 2011. Resveratrol nanosuspensions: interaction of preservatives with nanocrystal production. Die Pharmazie 66, 942-947. Kruss, B., Peters, K., Becker, R., Muller, R.H., 1996. Pharmaceutical nanosuspensions for medicament administration as systems with increased saturation solubility and speed of dissolution. Patent CA2205046 A1. Lemke, A., Moeschwitzer, J., 2007. Method for carefully producing ultrafine particle suspensions and ultrafine particles and use thereof. Patent WO2006108637 A3. Liversidge, G.G., Cundy, K.C., Bishop, J.F., Czekai, D.A., 1991. Surface modified drug nanoparticles. Patent US5145684 A. Mishra, P.R., Al Shaal, L., Muller, R.H., Keck, C.M., 2009. Production and characterization of Hesperetin nanosuspensions for dermal delivery. Int. J. Pharm. 371, 182-189. Müller, R.H., Jacobs, C., Kayser, O., 2001. Nanosuspensions as particulate drug formulations in therapy: Rationale for development and what we can expect for the future. Adv. Drug Delivery Rev. 47, 3-19. Müller, R.H., Möschwitzer, J., 2009. Method and device for producing very fine particles and coating such particles. Patent US20090297565 A1. Müller, R.H., Runge, S., Ravelli, V., Mehnert, W., Thünemann, A.F., Souto, E.B., 2006. Oral bioavailability of cyclosporine: Solid lipid nanoparticles (SLN®) versus drug nanocrystals. Int. J. Pharm. 317, 82-89. Nakarani, M., Patel, P., Patel, J., Patel, P., Murthy, R.S.R., Vaghani, S.S., 2010. Cyclosporine ANanosuspension: Formulation, Characterization and In Vivo Comparison with a Marketed Formulation. Sci. Pharm. 78, 345-361. Parikh, I., Selvaraj, U., 1999. Composition and method of preparing microparticles of water-insoluble substances. Patent US5922355 A. Petersen, R., 2006. Nanocrystals for use in topical cosmetic formulations and method of production thereof. Patent US608866233. Romero, G.B., Chen, R., Keck, C.M., Müller, R.H., 2015a. Industrial concentrates of dermal hesperidin smartCrystals® – production, characterization & long-term stability. Int. J. Pharm. 482, 54-60. Romero, G.B., Keck, C.M., Müller, R.H., Bou-Chacra, N.A., 2015b. Cationic nanocrystal formulation for improved ocular delivery (submitted). Eur. J. Pharm. Biopharm., Submission ID: EJPB-D15-00505.
21
Shegokar, R., Müller, R.H., 2010. Nanocrystals: Industrially feasible multifunctional formulation technology for poorly soluble actives. Int. J. Pharm. 399, 129-139. Teeranachaideekul, V., Junyaprasert, V.B., Souto, E.B., Müller, R.H., 2008. Development of ascorbyl palmitate nanocrystals applying the nanosuspension technology. Int. J. Pharm. 354, 227-234. Van Eerdenbrugh, B., Stuyven, B., Froyen, L., Van Humbeeck, J., Martens, J.A., Augustijns, P., Van den Mooter, G., 2009. Downscaling Drug Nanosuspension Production: Processing Aspects and Physicochemical Characterization. AAPS PharmSciTech 10, 44-53.
22
Figure Captions Figure 1. Left: milling vial with 1 stirring bar; middle: milling vial with 2 stirring bars; right: milling vial with 3 stirring bars, which showed no dead volume (milling beads size: 0.05 mm; rotation speed: 1,200 rpm).
23
Figure 2. PCS diameter (top) and polydispersity indices (PdI) (bottom) of a cyclosporin A nanosuspension after milling with different number of stirring bars during 1-24 hours (0.05 mm beads and 1,200 rpm).
24
Figure 3. Left: schematic representation of the miniaturized method; right: picture of the experimental set-up without the milling beads and suspension.
25
Figure 4. Particle size as a function of milling time using the miniaturized method for 6 different drugs. A: cyclosporin A; B: resveratrol; C: hesperitin; D: ascorbyl palmitate; E: apigenin; F: hesperidin; (milling time 120 hours; beads size: 0.05 mm). For resveratrol (B), the y-axis is extended. *due to aggregation or gelling, the milling time for apigenin and hesperidin was 24 hours.
26
Figure 5. Principle of size reduction kinetics during milling as observed for the model actives hesperidin (Hd), hesperitin (Ht) and cyclosporin A (Cyc A).
27
Figure 6. Effect of different sized milling beads during 24 hours milling using the miniaturized set-up for ascorbyl palmitate.
28
Figure 7. Particle size reduction of ascorbyl palmitate as a function of high pressure homogenization cycles at 1,500 bar.
29
Figure 8. Particle size of ascorbyl palmitate as a function of milling time using a mill PML-2 (discontinuous mode) and two different bead sizes (0.2 and 0.4-0.6 mm) during 60 minutes.
30
Figure 9. PCS diameter of ascorbyl palmitate nanosuspensions produced with the PML-2 with and without a HPH step, after 6 months storage at 4 and 25 °C.
31
Tables Table 1. Actives and their respective stabilizer processed by the miniaturized milling method (alkyl polyglucoside C8-C10 = Plantacare® 2000 UP). ACTIVE Cyclosporin A Resveratrol Hesperitin Ascorbyl palmitate Apigenin Hesperidin
STABILIZER TPGS alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 alkyl polyglucoside C8-C10 Poloxamer 188
32
Table 2. Overview of the size reduction ability (size after 1 hour milling, smallest achievable size) and constitution (aggregation, gelling) during milling. SMALLEST SIZE PCS size after NO AGGREGATION/ GELLING/ MODEL DRUG size hours of 1 hour (nm) AGGREGATION SIZE (nm) SIZE (nm) (nm) milling cyclosporin A resveratrol hesperitin ascorbyl palmitate apigenin hesperidin
668 5,057 291 749 238 294
97 202 49 171 85 294
48 h 48 h 24 h 72 h 24 h 1h
Ø + / 729 + / 97 Ø + / 85 + / 844
33