3 unsaturated fatty acid ethyl

Journal of Microencapsulation Downloaded from informahealthcare.com by Saarlaendische Universitaets- und Landesbibliothek on 07/20/15 For personal use only.

j. microencapsulation, 2001, vol. 18, no. 3, 347±357

In¯uences of process parameters on preparation of microparticle used as a carrier system for O ¡ 3 unsaturated fatty acid ethyl esters used in supplementary nutrition A. LAMPRECHT, U. SCHAÈ FER and C.-M. LEHR* Department of Biopharmaceutics and Pharmaceutical Technology, Saarland University, Im Stadtwald, D-66123 SaarbruÈcken, Germany (Received 10 January 2000; accepted 18 April 2000 ) Microparticles were prepared by complex coacervation to encapsulate eicosapentaenoic acid ethyl ester (EPA-EE) for incorporation into foods as a nutrition supplement. Gelatin and acacia were used in the coacervation process. With an increasing oil/polymer ratio, both yield and encapsulation rate decreased; with an increasing homogenization time, the yield remained constant while the encapsulation rate slightly increased. Several particle hardening techniques were examined and their in¯uence on particle structure, yield and encapsulation rate were examined. Ethanol hardening was compared to cross-linking with dehydroascrobic acid with respect to both yield and encapsulation rate. The particle diameters for both formulations were similar (ethanol: 38:4 § 4:1 mm; cross-linking: 41:8 § 3:0 mm). Spray-drying of the coacervates led to the smallest particles (5:2 § 1:1 mm), lowest yield and encapsulation rate. All microencapsulation products were assayed for their storage stability over 4 weeks with respect to the oxidation of the encapsulated O ¡ 3 unsaturated fatty acid ester inside the particles. Hardening with ethanol showed the lowest amount of peroxides: particle wall cross-linking by dehydroascorbic acid and spray-drying were observed to be less protective. All microparticles were characterized for their internal structure with confocal laser scanning microscopy (CLSM) after ¯uorescence labelling of the polymers, in order to localize the oil phase and visualize the distribution of the polymers in the coacervates. With increasing homogenization time, the internal structure changed stepwise from a capsule structure (core/wall) towards a matrix structure. For all experiments, a homogeneous distribution for both polymers, gelatin and acacia was observed inside the particle wall. No in¯uence of the di erent particle hardening procedures on the polymer distribution was found. Keywords: Microencapsulation, complex coacervation, particle hardening, nutrition supplement, confocal laser scanning microscopy.

Introduction O ¡ 3 unsaturated fatty acids play various physiological roles in the body, especially in in¯ammatory processes. They are thought to be implicated * To whom correspondence should be addressed; e-mail: [email protected] Journal of Microencapsulation ISSN 0265±2048 print/ISSN 1464±5246 online # 2001 Taylor & Francis Ltd http://www.tandf.co.uk/journals DOI: 10.1080/02652040010000433


348

A. Lamprecht et al.

in several chronic diseases, from thrombosis, asthma, hypertension, heart failure, rheumatoid arthritis, to Crohn’s disease and others (Singer 1989). A problem of the therapy with O ¡ 3 unsaturated fatty acids is the high dose required per day and also the frequent intake, if the compound is applied as cold water ®sh oil. A higher purity grade of O ¡ 3 unsaturated fatty acids extracted from ®sh oil should allow reduced daily dosage regimens and, therefore, improve patient compliance. The increased purity of the O ¡ 3 unsaturated fatty acids can be obtained by a supercritical ¯uid chromatography extraction method (Treitz 1997). This technique removes saturated fatty acids from the oil phase and concentrates the pharmacologically active O ¡ 3 unsaturated fatty acids, leading to a higher e ciency of the applied ®sh oil amount. With the administration of pure O ¡ 3 unsaturated fatty acids, the regular intake of a high number of gelatin capsules containing O ¡ 3 unsaturated fatty acids is necessary (Aûmann 1995). An alternative approach is the incorporation of the O ¡ 3 unsaturated fatty acids into food, where the entrapment in microparticles is advantageous with respect to taste masking and oxidative protection. Possible microencapsulation procedures have to rely on excipients, such as polymers and hardening agents that are approved for food requirements. This limits the normally rich variety of methods. Di erent techniques for the preparation of microparticles and microcapsules for incorporation in food have been reported (Arshady 1993, Shahidi and Han 1993). Complex caocervation is a very common approach, requiring only a moderate expenditure on equipment. The gelatin/acacia system has been widely used for the encapsulation of drugs in the area of pharmaceutical sciences (Luzzi and Gerraughty 1967, Jizomoto et al. 1993, Ertan et al. 1997, Burgess and Ponsart 1998). The e ects of polymer concentration and pH have been extensively studied (Daniels and Mittermaier 1995). For the encapsulation of oily liquids, the addition of surfactants has also been considered to modify the encapsulation process and the properties of the microparticles (Tirkkonen et al. 1994, Rabiskova and Valaskova 1998). However, such additives may be prohibitive for use in food supplements, and, therefore, other parameters have to be considered for process optimization of complex coacervation in such applications. In this study, the in¯uences of di erent parameters, such as oil/polymer ratio and homogenization speed, on the encapsulation process were examined in order to increase the process yield and encapsulation rate. Particles were characterized for their structure, i.e. oil phase distribution and polymer wall composition using confocal laser scanning microscopy (CLSM). To elucidate the hardening step, which represents a particularly critical problem of this microencapsulation process, several hardening techniques were applied, using food compatible compounds to the coacervate, and the particle characteristics were compared. A very important aim in this study was the determination of the ability of the di erent microencapsulation products to increase the stability of O ¡ 3 unsaturated fatty acids against oxidation. Due to the higher purity of the unsaturated fatty acid in the oil phase, its storage stability inside the microparticles becomes a critical parameter and is, therefore, determined for all di erent hardening procedures.

Microparticles for supplementary nutrition

349

Materials and methods


Materials Gelatin, type A and B, with di erent bloom numbers was obtained from Deutsche Gelatine Fabrik Stoess & Co. (Eberbach, Germany), acacia was obtained from Caesar & Loretz (Hilden, Germany). Eicosapentaenoic acid ethyl ester (EPA-EE) was obtained with a purity of >96% as a gift from K. D. Pharma (Bexbach, Germany). Fluorescein isothiocyanate (FITC) and rhodamine B isothiocyanate (RBITC) were purchased from Fluka (Deisenhofen, Germany); nile red and ethanolamine were purchased from Sigma (Deisenhofen, Germany). Savinase1 was obtained from Novo Nordisk Biotechnologie (Mainz, Germany). All other compounds were obtained from Merck AG (Darmstadt, Germany) and were of analytical grade purity. Methods Preparation of EPA-EE microparticles by complex coacervation. Gelatin (2.5 g) was dissolved in 100 ml water at 508C. Then, the oil phase (EPA-EE) was emulsi®ed in the gelatin solution. This emulsion was prepared in di erent ways, either by using an ordinary magnetic stirrer or a four-blade stirrer (500±1000 rpm, Typ RW 18, Janke & Kunkel, IKA-Werk, Staufen, Germany) or by using an Ultraturrax (8000 rpm, Model T 25, Janke & Kunkel, IKA-Labortechnik, Staufen, Germany). The emulsion was poured into a beaker with 100 ml water containing 2.5% (w/v) of acacia, 400 ml of water were added, the pH was lowered with hydrochloric acid (1N) or acetic acid (10%) to 4.0±4.3 and the system was cooled to 48C. Thereafter, one of the particle wall hardening steps described below was applied to the system. All preparation steps were carried out under argon or CO2 atmosphere to avoid oxidation of the unsaturated fatty acid ester. Hardening step by ethanol extraction. After sedimenting the microcapsules for 2 hours, the supernatant was decanted and the particles were hardened by adding di erent quantities of ethanol to the sediment. Finally, the microcapsules were ®ltered and dried overnight at room temperature to constant weight. Hardening step by spray drying. After coacervation, the particles were sedimented and the supernatant was removed before starting the spray drying process. Cooling the emulsion down to 48C was optional. For the spray drying process, a BuÈchi 190 Mini Spray Dryer was used. Apparatus parameters were chosen as follows, after optimization: aspiration: 20; pressure: 300 NI/h; temperature: inlet: 1408C, outlet: 808C. Hardening step by dehydroascorbic acid. An alternative hardening procedure was performed by chemical cross-linking of the particle wall polymer. L-ascorbic acid was stirred for 48 h in phosphate bu er (pH 7.4) at 408C, the e cient oxidation of ascorbic acid was proven by UV spectrometry. The coacervation was processed until the cooling step, which lowered the temperature to 48C. Thereafter, 2.5 mm dehydroascorbic acid were added to the coacervate, and the system was stirred at room temperature overnight. The microparticles were separated by ®ltration and washed twice with distilled water

350

A. Lamprecht et al.


to eliminate the cross-linking molecules present on the microparticles’ surface, but not covalently linked to the coacervated polymers. Hardening step by glutardialdehyde. All processes were performed analogue to the hardening procedure by dehydroascorbic acid. Instead of using dehydroascorbic acid as a hardening agent, however, 2.5 mm of glutardialdehyde was added to the coacervate. Fluorescence labelling of the microparticle components. In order to be visualized by confocal laser scanning microscopy, the microparticle components had to be labelled by appropriate ¯uorescent dyes. The oil phase was labelled before preparing the oil-in-water emulsion; nile red was dissolved in the oil phase as a ¯uorescent market in a concentration of 0.5 mg/ml nile red. The labelling of the polymers was performed as follows: 100 ml of the 2.5% (w/v) aqueous gelatin solution were adjusted at pH 8.5 by sodium hydroxide (1N). The ¯uorescent dye (FITC or RBITC) was dissolved in DMSO at a concentration of 1 mg/ml. One hundred microlitres of the dye solution were added to the polymer solution and stirred for 1 h at 408C. The labelling reaction was stopped by adding 50 ml ethanolamine. Acacia was labelled by following the same procedure, but at a higher pH (11.0). Confocal laser scanning microscopy. A Biorad MRC 1024 Laser Scanning Confocal Imaging System (Hemel Hempstead, UK), equipped with an argon ion laser (American Laser Corp., Salt Lake City, USA) and a Zeiss axiovert 100 microscope (Carl Zeiss, Oberkochen, Germany), was used to investigate the structure and morphology of the microparticles. Dry microparticles were dispersed in neutral oil to avoid swelling of the polymeric material during the imaging procedure. For the oil phase localization, one ¯uorescence channel was used in the single red ¯uorescence mode at 514 nm excitation wavelength. The ®nal pictures were composed from the red ¯uorescence and transmitted light channels to visualize the oil phase and capsule wall in the same image. To determine the polymer deposition in the particle wall qualitatively, the laser was adjusted in the green/red ¯uorescence mode which yielded two excitation wavelengths at 488 and 514 nm. The emission ®lter blocks VHS/A1 and A2 were used. Green and red ¯uorescence images were obtained from two separate channels and a third picture from the transmitted light detector was optional. Determination of encapsulation e ciency. Dry microparticles (100 mg) prepared by coacervation methods, were degraded enzymatically by incubation with a mixture of 5.0 ml of NH3 solution (2 m) and 0.5 ml of Savinase1 overnight in a water bath at 408C. The microcapsule degradation was veri®ed by light microscopy. Afterwards, the EPA-EE was extracted with n-hexane and injected into the gas chromatograph (Hewlett-Packard 5880; detector: FID; capillary column: Permabond FFAP, Machery & Nagel, id: 0.25 mm, L ˆ 25 m, ®lm thickness: 0.1 mm) without diluting. Alternatively, after extraction of the oil phase with n-hexane and removing the solvent by evaporation, the recovered amount of EPA-EE was determined gravimetrically (Rabiskova et al. 1994). The extraction procedure was carried out under argon or CO2 atmosphere to avoid oxidation of the unsaturated fatty acid ester. All batches were analysed in triplicate.



351

Quanti®cation of the peroxides. The extracted oil phase of 5 g microparticles was dissolved in a mixture of 18 ml acetic acid (96%) and 12 ml chloroform. To this mixture, 0.5 ml of saturated potassium iodide solution were added and, after incubation for 1 min, 30 ml distilled water were added to stop the reaction. Titration of free potassium iodide was performed with sodium thiosulphate (0.01N) against starch as indicator (European Pharmacopeia 1999). The quantity of peroxides were calculated as a quotient of mmol peroxides per mmol EPA-EE. Results and discussion Optimization of process parameters In¯uence of the oil/polymer ratio. The in¯uence of the oil/polymer ratio in the microencapsulation process was examined using the ethanol particle hardening. The oil/polymer ratio was changed from 1:5 to 5:1 by adding di erent amounts of oil to a ®xed amount of particle wall polymer solution (5 g). The absolute mass of entrapped oil per gram microparticles increased, but, at the same time, a decrease of process yield and encapsulation rate was observed (®gure 1). For high amounts of oil phase (>10 g), it was found that a certain quantity of oil was left nonencapsulated, which was recovered as free ¯oating oil during the decantation step. At higher oil amounts (>5 g), the risk arose that ¯oating microparticles could be separated by the decantation step due to their low density. This observation was considered to be a speci®c problem of the ethanol hardening procedure and, therefore, this was further evaluated in the application of other hardening procedures. In¯uence of the emulsi®cation step. The in¯uence of the emulsi®cation step during the preparation of the oil-in-water emulsion was examined. As magnetic stirring for the emulsi®cation process was not a success (the encapsulation rate was less

Figure 1. In¯uence of oil/polymer ratio on process yield or encapsulation rate (* ˆ process yield in per cent; ~ ˆ encapsulation rate in per cent). The homogenization step was performed by an Ultraturrax.


352

A. Lamprecht et al.

Figure 2. In¯uence of homogenization time (Ultraturrax) on process yield and encapsulation rate (* ˆ process yield in per cent; ! ˆ encapsulation rate in per cent).

than 5%) and the application of a four-blade stirrer limited the encapsulation rate for the highest stirring speed (1500 rpm) to ¹32%, the Ultraturrax was used to increase the emulsi®cation e ciency. The experiments were performed at 8000 rpm with di erent time intervals for the homogenization step. A nearly constant microparticle yield was observed, independent of the homogenization time, whereas an increase of the encapsulation rate for homogenization times longer than 5 min was found (®gure 2). The prolonged homogenization led to a more e cient emulsi®cation, resulting in a highly e cient incorporation into the coacervate and less non-encapsulated oil. Short homogenization times led to the typical core/wall structure, while longer homogenization time reduced the mean diameter of the oil phase droplets, thus leading to a change from the core/wall structure into a matrix structure (®gure 3). This was paralleled by an increased emulsi®cation e ciency. Comparison of the di erent hardening procedures One of the major problems in the use of coacervates for food supplements has always been the ®nal particle hardening, which was performed by formaldehyde or glutardialdehyde in pharmaceutical research (Luzzi and Gerraughty 1964, Desoize et al. 1986), but these compounds are not allowed for food applications. So, the aim was the comparison of some alternative particle hardening procedures available to the ®eld of food supplements, in order to ®nd the most promising approach. Hardening step by ethanol. As mentioned before by others, alcohol is applicable for hardening the particle wall after coacervation (Jizomoto et al. 1993, Rabiskova et al. 1994). The use of ethanol led to a successful hardening of the particle wall and free ¯owing particles were obtained after drying, for ethanol/polymer ratios higher than 30:1. Ethanol extracted water from the coacervate, and hardened the particle wall as a non-solvent of the polymers.



353

Figure 3. CLSM images of the internal microparticle structure after oil phase staining with nile red. These two pictures show the change of particle structure from core/wall (a) to a matrix structure ((b) + (c)) due to the increased homogenization time applied to the o/w emulsion.

Hardening step by spray drying. Since the ethanol hardening method is a relatively complex procedure which involves various steps, other particle drying or hardening techniques had to be considered. Among those, spray drying was tested as a promising approach for the evaporation of water from the coacervate after phase separation. EPA-EE was emulsi®ed in the gelatin solution, coacervated after adding acacia, and ®nally spray dried. The average particle size was much smaller than found after applying other hardening techniques (table 1). The oil droplets in the

354

A. Lamprecht et al.


Table 1. Particle size, yields, encapsualtion rates and free oil recovered from the particle surface after incorporation of EPA-EE with complex coacervation and using di erent hardening techniques.

Ethanol Spray drying Dehydroascorbic acid Glutardialdehyde

Particle size (mm)

Process yield (%)

Encapsulation rate (%)

Surface oil (%)

38.2 § 4.1 5.2 § 1.1 41.8 § 3.0 35.0 § 4.9

84.2 § 1.9 32.0 § 9.5 79.5 § 5.2 80.9 § 3.4

55.2 § 6.6 15.7 § 4.9 53.8 § 5.2 51.9 § 7.0

1 28.2 § 4.7 4.3 § 2.2 n.d.

emulsion were observed to be larger before hardening than the total microcapsule obtained after spray drying. So, the emulsion was arranged in structure by the spraying process and, therefore, the coacervate at the oil±water interface changed place during its passage through the sprayer nozzle. An explicit core/wall structure was observed to be independent of droplet size. Additionally, it was shown by CLSM imaging that the general homogeneous distribution of gelatin was retained throughout the capsule (®gure 4). Acacia was proven to have the same distribution. The major disadvantage of the spray drying method was the low yield and encapsulation rate (table 1). Moreover, some non-encapsulated EPA-EE remained on the particle surface, leading to an inactivation of the compound by oxidation and also to a very unpleasant odour. This made an additional washing step necessary. By the washing step with ethanol, 28:2 § 4:7% was recovered as a non-encapsulated oil phase. Another problem which had to be solved was the hot air used during the hardening step. The heat causes the oxidation of the unsaturated fatty acid during the preparation step.

Hardening step by dehydroascorbic acid. The attraction of this cross-linking approach is the possible application to nutrition supplements without toxicity. Dehydroascorbic acid was selected as a cross-linker due to the possibility of covalent bond formation to functional groups of the polymers by its carbonyl groups (Aiedeh et al. 1997). The diameter of the microparticles obtained by this method was not signi®cantly di erent from those after ethanol hardening, indicating that the particle structure is not in¯uenced by these two hardening procedures. The particle wall showed a homogeneous polymer composition which was proven by CLSM (®gure 4). In general, this hardening method was as e cient as the ethanol hardening technique concerning yield and encapsulation rate (table 1). The major advantage of the use of dehydroascrobic acid concerns hardening by cross-linking, as no decantation step is required with the risk of losing microparticles. From the surface, 4:3 § 2:2% residues of non-encapsulated EPA-EE were recovered, referred to the total oil mass. For this hardening approach, it was also observed that non-encapsulated oil remained on the particle surface, requiring an additional ethanol washing step.



355

Figure 4. CLSM images representing the microparticle wall composition with a view to gelatin labelled with RBITC. Internal structure of microparticles after the di erent hardening procedures: ethanol extraction (a); hardening by cross-linking with dehydroascrobic acid (b); and hardening by spray drying (c). In all cases, the homogenization step was performed by an Ultraturrax prior to the hardening step.

Also, the non-oxidized ascorbic acid was applied to the cross-linking reaction, which proved its capability to cross-link due to the oxidation taking place during the microparticle preparation step.

Storage stability The ability to protect the incorporated oil from oxidation was examined for microparticles containing di erent oil volumes. A distinct decrease in stability was found for higher oil amounts (®gure 5).


356

Figure 5.

A. Lamprecht et al.

In¯uence of the increased oil phase volume in the oil/polymer ratio on PO (after 2 weeks).

Figure 6. In¯uence of storage time (within 4 weeks) on the stability of the encapsulated oil phase against oxidation of the EPA-EE is represented by the PO (& ˆ pure nonencapsulated oil; * ˆ ethanol hardening; ~ ˆ spray drying; ^ ˆ cross-linking by dehydroascorbic acid).

This may re¯ect the decreasing amount of protecting wall polymer as the oil phase mass was rising. The distinct matrix structure of the microparticles after longer homogenization was found to have no signi®cant e ect on the peroxide formation inside the microparticles (data not shown). All formulations prepared with di erent hardening procedures were assayed for their storage stability over 4 weeks at room temperature (®gure 6). Ethanol and dehydroascorbic acid hardening showed no signi®cant di erence, probably due to their equivalent particle structure, similar size and wall thickness. After hardening by spray drying, the protection of EPA-EE from oxidation was less than with the


357


other techniques. The relatively small size of the particles result in an increased area of the air/particle interface, and the very thin protecting polymer wall promote oxidation. Acknowledgements This project was supported by K. D. Pharma GmbH, Bexbach, Germany. The `Fonds der Chemischen Industrie’ is thanked for ®nancial support. References Aiedeh, K., Gianasi, E., Orienti, I., and Zecchi, V., 1997, Chitosan microcapsules as controlled release systems for insulin. Journal of Microencapsulation, 14, 567±576. Arshady, R., 1993, Microcapsules for food. Journal of Microencapsulation, 10, 413±435. Assmann, C., 1995, Pharmakologische Wirkungen von Omega-3-FettsaÈuren. Deutsche Apotheker Zeitung, 135, 2530±2532. Burgess, D. J., and Ponsart, S., 1998, -Glucuronidase activity following complex coacervation and spray drying microencapsulation. Journal of Microencapsulation, 15, 569±579. Daniels, R., and Mittermaier, E. M., 1995, In¯uence of pH adjustment on microcapsules obtained from complex coacervation of gelatin and acacia. Journal of Microencapsulation , 12, 591±599. Desoize, B., Jardillier, J. C., Kanoun, K., Guerin, D., and Levy, M. C., 1986, In-vitro cytotoxic activity of cross-linked protein microcapsules. Journal of Pharmaceutics and Pharmacology, 38, 8±13. Ertan, G., Ozer, O., Baloglu, E., and Guneri, T., 1997, Sustained-release microcapsules of nitrofurantoin and amoxicillin; preparation, in-vitro release rate, kinetic and micromeritic studies. Journal of Microencapsulation, 14, 379±388. European Pharmacopoeia, 3rd edn, 1999 (Germany: Govi-Verlag Eschborn). Jizomoto, H., Kanaoka, E., Sugita, K., and Hirano, K., 1993, Gelatin-acacia microcapsules for trapping micro oil droplets containing lipophilic drugs and ready disintegration in the gastrointestianl tract. Pharmaceutical Research, 10, 1115± 1122. Luzzi, L. A., and Gerraughty, R. J., 1964, E ects of selected variables on the extractability of oils from coacervate capsules. Journal of Pharmaceutical Science, 53, 429±431. Luzzi, L. A., and Gerraughty, R. J., 1967, E ects of selected variables on the microencapsulation of solids. Journal of Pharmaceutical Science, 56, 634±638. Rabiskova, M., and Valaskova, J., 1988, The in¯uence of HLB on the encapsulation of oils by complex coacervation. Journal of Microencapsulation, 15, 747±751. Rabiskova, M., Song, J., Opawale, F. O., and Burgess, D. J., 1994, The in¯uence of surface properties on uptake of oil into complex coacervation microcapsules. Journal of Pharmacy and Pharmacology, 46, 631±635. Shahidi, F., and Han, X. Q., 1993, Encapsulation of food ingredients. Critical Reviews in Food Science and Nutrition, 33, 501±547. Singer, P., 1989, Zur EssentialitaÈt von Omega-3 FettsaÈuren, Aktuelle ErnaÈ hrungsmedizin, 14, 293±303. Tirkkonen, S., Turakka, L., and Paronen, P., 1994, Microencapsulation of indomethacin by gelatin-acacia complex coacervation in the presence of surfactants. Journal of Microencapsulation , 11, 615±626. Treitz, M., 1997, UÈberkritisches CO2 in Extraktion und Chromatographie, PhD thesis, SaarbruÈcken.