- Email: [email protected]
Chitosan microparticles for oral vaccination:
Biomaterials 22 (2001) 687}694
Chitosan microparticles for oral vaccination: preparation, characterization and preliminary in vivo uptake studies in murine Peyer's patches I.M. van der Lubben , J.C. Verhoef , A.C. van Aelst, G. Borchard , H.E. Junginger * Leiden/Amsterdam Center for Drug Research, Division of Pharmaceutical Technology, P.O. Box 9502, 2300 RA Leiden, Netherlands Laboratory of Experimental Plant Morphology and Cell Biology, Arboretumlaan 4, 6703 BD Wageningen, Netherlands Received 15 December 1999; accepted 3 July 2000
Abstract Although oral vaccination has numerous advantages over parenteral injection, degradation of the vaccine in the gut and low uptake in the lymphoid tissue of the gastrointestinal tract still complicate the development of oral vaccines. In this study chitosan microparticles were prepared and characterized with respect to size, zeta potential, morphology and ovalbumin-loading and -release. Furthermore, the in vivo uptake of chitosan microparticles by murine Peyer's patches was studied using confocal laser scanning microscopy (CLSM). Chitosan microparticles were made according to a precipitation/coacervation method, which was found to be reproducible for di!erent batches of chitosan. The chitosan microparticles were 4.3$0.7 lm in size and positively charged (20$1 mV). Since only microparticles smaller than 10 lm can be taken up by M-cells of Peyer's patches, these microparticles are suitable to serve as vaccination systems. CLSM visualization studies showed that the model antigen ovalbumin was entrapped within the chitosan microparticles and not only associated to their outer surface. These results were veri"ed using "eld emission scanning electron microscopy, which demonstrated the porous structure of the chitosan microparticles, thus facilitating the entrapment of ovalbumin in the microparticles. Loading studies of the chitosan microparticles with the model compound ovalbumin resulted in loading capacities of about 40%. Subsequent release studies showed only a very low release of ovalbumin within 4 h and most of the ovalbumin (about 90%) remained entrapped in the microparticles. Because the prepared chitosan microparticles are biodegradable, this entrapped ovalbumin will be released after intracellular digestion in the Peyer's patches. Initial in vivo studies demonstrated that #uorescently labeled chitosan microparticles can be taken up by the epithelium of the murine Peyer's patches. Since uptake by Peyer's patches is an essential step in oral vaccination, these results show that the presently developed porous chitosan microparticles are a very promising vaccine delivery system. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Chitosan microparticles; Ovalbumin; Oral vaccination; Confocal laser scanning microscopy; Scanning electron microscopy; Peyer's patches
1. Introduction Although the introduction of an e$cient oral vaccine would diminish costs, patient uncomfort and the need for quali"ed personnel to administer the vaccine, most vaccines still have to be administered by injection. The reasons for unsuccessful development of a potent oral vaccine carrier are still hindered due to degradation of the antigen in the gastrointestinal (GI) tract and ine$-
* Corresponding author. Tel.: #131-715274308; fax: #131715274565. E-mail address: [email protected] (H.E. Junginger).
cient targeting to the site of action in the gut. Before the antigen reaches the M-cells of the Peyer's patches in the GI tract, it has to pass the stomach with its low pH and several inactivating enzymes. Even if the vaccine has arrived near the Peyer's patches, an immune response is not always elicited. The antigen might not reach the dome of the Peyer's patch due to ine$cient uptake, or is not potent enough to provoke a su$cient immune reaction [1]. Several studies have shown that by associating the vaccine with a number of microparticulate drug carrier systems, the uptake by M-cells is enhanced and the degradation of the vaccine in the GI tract is prevented [2}6]. In this way, a &Trojan horse' principle is applied: the vaccines are delivered hidden inside biodegradable
0142-9612/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 2 3 1 - 3
688
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
microspheres that can be taken up by endocytosis by these intestinal M-cells [7,8]. Microparticles smaller than 10 lm are taken up by the M-cells and transported into the Peyer's patches. Most microparticles larger than 5 lm stay in the Peyer's patches, while microparticles smaller than 5 lm are transported through the e!erent lymphatics [6]. Chitosan has been shown to be a suitable vaccine carrier. Chitosan is the deacetylated form of chitin (polyb-(1P4)-N-acetyl-D-glucosamin), the second most abundant polymer in nature. Chitin is mostly obtained from exoskeletons of marine arthropods. Molecular weight and degree of acetylation determine the properties of chitosan. Because of its biocompatibility, biodegradability, low cost and ability to open intercellular tight junctions, this polymer is a valuable excipient for oral drug delivery systems [9]. Recently, numerous studies on chitosan as a drug absorption enhancer have been published. Chitosan formulations are used for ocular [10], oral [11], parenteral [12] and nasal delivery [13}15], as well as for DNA transfection studies [16]. Furthermore, chitosan can easily form microparticles. Advances in microparticulate drug delivery research have opened up the way to apply these techniques for oral vaccination. Due to the high protein binding properties of some types of chitosan microparticles, they are also potential candidates for oral delivery of antigens [10]. Mild preparation can protect the proteins when they are incorporated during preparation of the microparticles [10,18]. In order to circumvent protein denaturation conditions, chitosan microparticles can be loaded passively. Such a mild loading procedure was described by Jameela et al. [12]. Besides antigens, also DNA coding for antigens can be taken up by the M-cells of Peyer's patches. Transcription of this DNA leads to the production of the antigen. Roy et al. [16] reported that in this way chitosan nanospheres can also be used as an oral vaccine delivery system by means of transfection. Chen et al. [17] showed that poly (lactide-coglycolide) microparticles loaded with rotavirus VP6 DNA provoked an immune reaction after oral administration. In this study chitosan microparticles for oral vaccination were prepared. Because of the non-toxic features and potent antigen binding properties, chitosan polymers are expected to be promising candidates for oral vaccination. Chitosan microparticles were made by a coacervation/precipitation method and characterized with respect to oral vaccination. The morphology of the microparticles was evaluated using light microscopy, confocal laser scanning microscopy (CLSM) and scanning electron microscopy. The size, charge, loading and release characteristics for the model antigen ovalbumin were investigated. Furthermore, initial in vivo uptake studies of #uorescently labeled chitosan microparticles by murine Peyer's patches were performed and vizualized by CLSM.
2. Materials and methods 2.1. Materials Chitosan was a generous gift from Primex (Avaldsnes, Norway). The viscosity of the chitosan used was measured as 1% (w/v) chitosan in 1% (v/v) acetic acid in MilliQ water on a rotation viscosimeter (Haake, Karlsruhe, Germany) in our laboratory and appeared to be 13 mpas. This di!ered from the viscosity mentioned by the manufacturer (20 mpas). The degree of deacetylation, as determined by the supplier was 93%. FITC-ovalbumin (494/520) and Bodipy威 (665/767) were purchased from Molecular Probes (Leiden, The Netherlands). Ovalbumin and Tween威 80 were obtained from Sigma (St. Louis, USA). All other reagents were of analytical grade. PhastGels (gradient 8}25) were purchased from Pharmacia Biotech (Uppsala, Sweden). Female Balb/c mice (9}12 weeks old) were ordered from Charles River (Sulzfeld, Germany). 2.2. Preparation of microparticles A 0.25% (w/v) chitosan solution was prepared in a mixture of 2% (v/v) acetic acid and 1% (w/v) Tween威80. Then 2 ml of 10% (w/v) sodium sulfate was added dropwise (about 1 ml/min) to 200 ml chitosan solution under magnetic stirring and continuous sonication. After adding the sodium sulfate, stirring and sonication were continued for 20 min. The microparticle suspension was subsequently centrifuged for 25 min (2750 rpm). The pellet was resuspended in MilliQ water to wash the microparticles and centrifuged again. This washing procedure was repeated twice before freeze-drying of the pellet overnight. Freeze-drying was performed using a Christ freeze-dryer (Osterode am Harz, Germany). 2.3. Characterization of chitosan microparticles 2.3.1. Size and zeta potential The size of the chitosan microparticles was determined using an Accusizer 770 (PSS, Santa Barbara, CA, USA). The zeta potential of the chitosan microparticles was measured in demineralized water at neutral pH, using a Malvern 2000 zetasizer (Herrenberg, Germany). 2.3.2. Morphology Light microscopy (Axioskop-type microscope; Zeiss, Weesp, The Netherlands) was performed for initial visualization of the chitosan microparticles. Morphology was evaluated using the following parameters: approximate size, form, uniformity, and the formation of "bers, gel or aggregates. Chitosan microparticles were analyzed after resuspension in water without staining or "xation. Further morphology studies were carried out by means of confocal laser scanning microscopy (CLSM) and
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
scanning electron microscopy (SEM). For CLSM studies, 1 ml of 1% (w/v) chitosan microparticles was incubated with 3 mg FITC-ovalbumin. A MRC-600 Lasersharp system (Bio-Rad Laboratories, Richmond, CA, USA) linked to a Zeiss IM 35 inverted microscope (Zeiss, Oberkochen, Germany) was used for this purpose, and several z-scans over a depth of 10 lm were made. For "eld emission SEM analysis, lyophilized chitosan microparticles were adhered to double-sided tape (Carbon adhesive tabs, 12 mm, electron microscopy studies, Fort Washington, USA). The microparticles were sputter coated with 5 nm platina in a preparation chamber (CT 1500 HF, Oxford Instruments, Oxfordshire, UK). Then the sputtered chitosan microparticles were studied with a JEOL 6300F "eld emission SEM (JEOL, Schiphol, The Netherlands) at 5 kV. Digital images were made and stored. 2.3.3. Ovalbumin loading The ovalbumin loading of microparticles was performed by incubating 1% (w/v) chitosan microparticles and 0.5}2.5% (w/v) ovalbumin in phosphate bu!ered saline (PBS; pH 7.3) under shaking at 253C. After incubation for 180 min, the suspension was centrifuged (1400 rpm for 30 min) to remove the unloaded ovalbumin. The loading degree was determined by quantifying the non-bound ovalbumin in the supernatant with the Lowry protein assay method [19]. Both loading capacity (LC) and loading e$cacy (LE) were determined: LC"[(total amount ovalbumin)!(free ovalbumin)] /weight microparticles, LE"[(total amount ovalbumin)!(free ovalbumin)] /total ovalbumin. 2.3.4. Ovalbumin release Ovalbumin release from chitosan microparticles was determined in PBS (pH 7.3). To load the microparticles, 3.5 ml of a 1% (w/v) chitosan microparticle suspension containing 0.5% (w/v) ovalbumin was incubated for 3 h. After centrifuging (1400 rpm for 30 min) the loaded microparticles were resuspended in PBS (pH 7.3) to make a 1% (w/v) microparticle suspension. Samples were incubated at 373C under mild shaking. After 15, 30, 45, 60, 90, 120, 180 and 240 min, the tubes were given a spin-o! and samples of 250 ll of the supernatant were taken. These samples were replaced by PBS (pH 7.3). The non-bound ovalbumin in PBS was determined with the Lowry protein assay [19]. 2.3.5. Degradation of ovalbumin-loaded microparticles To study if the loaded microparticles still contained ovalbumin after release studies, the chitosan microparticles were dissolved. This was done by incubating ovalbumin without microparticles, non-loaded micro-
689
particles and ovalbumin-loaded microparticles in 20% (v/v) acetic acid (pH 2.0) for a period of 2 h. Thereafter, the degraded microparticles were separated from the digestion solution by centrifuging (1400 rpm for 20 min). The ovalbumin in the supernatant was detected with the Lowry protein assay [19]. 2.3.6. Gel electrophoresis of released ovalbumin Released ovalbumin from loaded microparticles was analyzed by phast system威 gel electrophoresis to determine possible protein degradation by the loading and release processes. As controls both empty (non-loaded) microparticles and ovalbumin was used. Since hardly any ovalbumin was released over a period of 240 min (Fig. 5), forced release was also applied. This was done by loading the microparticles in MilliQ water with ovalbumin and subsequently performing release studies in PBS (pH 7.3). The ovalbumin loading of microparticles was performed by incubating 1% (w/v) chitosan microparticles and 0.5% (w/v) ovalbumin in MilliQ water under shaking at 253C. Release was performed as described in Section 2.3.4. Gel electrophoresis was performed on an 8}25 gradient gel and run for 87 Vh. Approximately 0.4 mg ovalbumin per lane was applied. 2.4. In vivo uptake of chitosan microparticles by murine Peyer+s patches For in vivo studies 1 ml of 1% (w/v) chitosan microparticles was incubated with 3 mg FITC-ovalbumin in PBS (pH 7.3) for three hours at 253C. After incubating, the suspension was centrifuged for 20 min (2750 rpm) to remove the non-bound FITC-ovalbumin. The pellet was resuspended in 300 ll PBS (pH 7.3) and fed intragastrically to the mice. In order to stain the membranes of the intestinal epithelial cells, the mice were also supplied with a non-speci"c lipophilic dye for cell membranes (Bodipy威). Groups of female Balb/c mice were fasted overnight and then fed intragastrically with a 150 ll Bodipy威 solution (0.1% in PBS : Glycerin : PEG400"1 : 1 : 1 (v/v/v)). After 3 h, the animals were also fed with either FITCovalbumin loaded microparticles (300 ll) or a solution of FITC-ovalbumin (3 mg in 300 ll PBS (pH 7.3)). The following 3 groups were used in this in vivo experiment: group 1 consisting of 6 mice fed with Bodipy威 and subsequently with FITC-ovalbumin loaded microparticles, group 2 consisting of 4 mice fed with Bodipy威 and subsequently with FITC-ovalbumin solution, and group 3 consisting of 3 mice only fed with Bodipy威. Three hours after the last feeding, the mice were sacri"ed by cervical dislocation. Then the Peyer's patches were dissected and immediately analyzed using CLSM. Since the spectra of the labels are not overlapping, both FITC-ovalbumin and Bodipy威 could be analyzed using two di!erent lasers at the same time. FITC-ovalbumin
690
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
was analyzed using a 633 nm laser and Bodipy威 using a 488 nm laser.
3. Results 3.1. Size and zeta potential of chitosan microparticles The size of the prepared non-loaded chitosan microparticles was found to be 4.23$0.67 lm (mean$SD; n"3) and the zeta potential was positive (20$1 mV). The small di!erences in size and zeta potential between the di!erent batches of microparticles demonstrate the good reproducibility of the method of preparation used. 3.2. Morphology of chitosan microparticles The morphology of the presented chitosan microparticles was analyzed by light microscopy after washing and freeze-drying. As depicted in Fig. 1, chitosan microparticles neither form a gel when resuspended in water, nor were forming big aggregates of microparticles, were not accompanied by "bers, and all showed reasonable uniformity. Fig. 2 shows #uorescently labeled chitosan microparticles after FITC-ovalbumin labeling. Three images at di!erent depths are presented, and it is evident that the ovalbumin is not only associated to the surface of the microparticles, but also entrapped within the microparticles. In all 3 scans the #uorescent label is clearly visible in the entire micoparticle. Field emission SEM demonstrated that pores are present at the rough surface of the chitosan microparticles (Fig. 3A). By focusing in the pores, the inside of the chitosan microparticles was visualized. As evident from Fig. 3B, the inside of the chitosan microparticles also has a very porous structure.
Fig. 2. Z-scans of #uorescently labeled chitosan microparticles using CLSM (60;). The #uorescent label is taken up by the microparticles and visible in all scans at di!erent depths ((A) 0.0 lm; (B) 1.9 lm; (C) 2.9 lm).
Fig. 1. Light microscopic image of chitosan microparticles (400;).
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
691
Fig. 4. Ovalbumin loading of chitosan microparticles at di!erent concentrations of total ovalbumin (LC loading capacity, LE loading e$cacy). Data are expressed as mean$SD of 10 experiments.
ovalbumin in demineralized water at neutral pH was !30$3 mV (mean$SD; n"3). 3.4. Ovalbumin release from chitosan microparticles
Fig. 3. Surface visualization of the chitosan microparticles using "eld emission SEM. (A) A detail of the surface of the microparticles. (B) A detail inside a pore of the chitosan microparticles. Scale bars are 1 lm.
3.3. Ovalbumin loading of chitosan microparticles The microparticles were loaded with di!erent amounts of ovalbumin by incubation of a 1% (w/v) chitosan microparticle suspension with 0.5}2% (w/v) ovalbumin. Both loading e$cacies (LE) and loading capacities (LC) were determined. Fig. 4 shows that the LC is not substantially in#uenced by the amount of ovalbumin available in the loading solution. Thus, the more ovalbumin is o!ered to the microparticles, the more ovalbumin remains unbound during the loading process. Therefore, 0.5% (w/v) ovalbumin in the loading solution was selected as the optimal concentration. When 1% of microparticles was incubated with 0.5% of ovalbumin, a loading percentage of 32$4% (n"10) for independently made batches was obtained. Under these conditions, a very high LE of 85$3% was obtained, indicating that only a small amount of ovalbumin was lost during the loading process. The size of the loaded chitosan microparticles was measured to be 5.7$0.6 lm. The zeta potential of loaded microparticles could not be determined, since the loading was performed in PBS. The ions present in this bu!er highly in#uence the zeta potential. The zeta potential of
Ovalbumin release from chitosan microparticles in PBS (pH 7.3) was determined over a time span of 240 min. After a release of about 10% in the "rst 60 min, no ovalbumin was released during the following 180 min. This indicates that most ovalbumin remained entrapped in the microparticles under these conditions. To verify whether the entrapped ovalbumin was still present in the chitosan microparticles after the loading and subsequent release studies, the loaded microparticles were dissolved. Non-loaded microparticles, ovalbuminloaded microparticles and ovalbumin suspensions were incubated in acetic acid at pH 2. After 2 h of incubation, the chitosan microparticles were found to be completely disintegrated and about 30% of the loaded ovalbumin could be determined in the suspension with the digested microparticles. When the microparticles were incubated for longer time periods at pH 2, the ovalbumin became denatured and appeared as a visible clot in the suspension. 3.5. Gel electrophoresis To investigate if the ovalbumin was not degraded during the loading and release processes, phast system威 gel electrophoresis was performed to compare released ovalbumin to non-treated ovalbumin. Because ovalbumin release from chitosan microparticles in PBS was very low, release in MilliQ water (forced release) was applied to analyze the ovalbumin associated with the microparticles. Fig. 5 shows the gel electrophoresis of released ovalbumin after normal and forced release, empty (non-loaded) chitosan microparticles and ovalbumin solutions in PBS bu!er and MilliQ water. No di!erence could be observed between the released ovalbumin and the ovalbumin solutions and no protein staining in the lane with empty microparticles. Since these results did not show any di!erences in molecular weight,
692
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
Furthermore, no uptake of FITC-ovalbumin was found in intestinal epithelium without Peyer's patches.
4. Discussion
Fig. 5. Gel electrophoresis of released ovalbumin (lane 1, supernatant of empty microparticles; lane 2, released ovalbumin from microparticles loaded and released in PBS; lane 3, released ovalbumin form microparticles loaded in MilliQ water and released in PBS; lane 4, released ovalbumin from microparticles loaded in PBS and released in MilliQ; lane 5, ovalbumin dissolved in PBS and lane 6, ovalbumin dissolved in MilliQ).
it is assumed that no substantial degradation of ovalbumin occurred during the loading and release processes. 3.6. In vivo uptake studies Fig. 6 shows typical examples of the uptake of chitosan microparticles present in mouse intestinal epithelium of the Peyer's patches. In the left image the epithelial cells are visualized and it is evident that the cell membranes have taken up some of the intragastrically fed Bodipy威. In this way the intestinal epithelial barrier could clearly be visualized. In the right image, transport of FITCovalbumin loaded chitosan microparticles across the epithelial barrier is observed. At the bottom of this same image also some #uorescent labeling is visible. This may be due to earlier transport of loaded particles across the cells. No transport could be determined in mice fed with ovalbumin solutions without chitosan microparticles.
Recently, chitosan nano- or microparticles have been prepared according to several precipitation/coacervation methods and some of these particles showed good antigen binding capacities [10]. Chitosan microspheres were also designed for colonic drug delivery [18,20]. For oral vaccination, microparticulate vaccine carrier systems not only need to associate a high amount of antigen, but also require speci"c release properties. After oral administration of such systems the vaccine should be well entrapped and protected from degradation in the GI-tract, and should only be released from the carrier system after uptake by the M-cells of the Peyer's patches. Besides, nano- or microparticles should not exceed 10 lm in size, and the hydrophobicity and antigens presented on the surface of the carrier system are important parameters [1,6]. The chitosan microparticles described in this study are much smaller than 10 lm, being a suitable size for M-cells uptake. Although they are hydrophilic compared to poly(D,L-lactide-co-glycolide), our preliminary in vivo studies showed that in vivo uptake of chitosan microparticles in Peyer's patches could be visualized. These observations indicate that hydrophilicity of chitosan microparticles may not be a principal factor. Microparticles (10 lm are taken up by the M-cells and transported to the dome of the Peyer's patches. Microparticles (5 lm are then transported to the spleen and lymph nodes, where speci"c IgM and IgG are produced. Since the cumulative size distribution showed that also microparticles between 5 and 10 lm were formed, these microparticles might stay in the Peyer's patches. In this case additional antigen speci"c IgM is formed [6]. The zeta potential of the prepared non-loaded microparticles is positive (20$1 mV) at neutral pH. This probably explains why microparticles do not aggregate in an
Fig. 6. Uptake of FITC-ovalbumin loaded chitosan microparticles by the intestinal epithelium of the Peyer's patches. (left) The epithelial cells were stained with Bodipy威 (arrow); in this way the epithelial barrier could be located. (right) Transport of FITC-ovalbumin loaded chitosan microparticles across the epithelial barrier (arrow). At the bottom of the image earlier transported ovalbumin is also visible. In none of the mice, fed with FITC-ovalbumin in PBS, intestinal uptake of FITC-ovalbumin was observed.
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
aqueous solution. Since cell membranes are negatively charged, these microparticles are also expected to be more easily associated and subsequently taken up by the membranes than negatively charged microparticles. Because the zeta potential is highly in#uenced by the presence of salts, the zeta potential could not be determined after loading the microparticles in PBS. In contrast to earlier studies with chitosan microspheres [18], no glutaraldehyde was required during the preparation of the present microparticles. Berthold et al. [21] found that crosslinking with glutardialdehyde makes the microspheres more stable to acids. However, the chitosan microparticles in the present study were only disintegrating after incubation for more than 2 h at a pH value of 2, compared to 30 min for the microspheres prepared by Berthold et al. [21] at similar ionic strength and pH values. Possible reasons for the increased stability of the present chitosan microparticles are yet unknown. Results from the in vitro CLSM studies showed that ovalbumin not only associates to the surface, but also enters within the chitosan microparticles. With this technique it was possible to visualize both the inside and the surface, and the #uorescent label appeared to be present throughout the microparticles. The observation that ovalbumin was able to enter the microparticles was con"rmed using "eld emission SEM. These studies demonstrated that the chitosan microparticles had a very porous structure. It is therefore likely that the ovalbumin enters the pores during the loading process and associates by electrostatic interactions to the inside of the chitosan microparticles. Because the present chitosan microparticles have a high loading capacity, large quantities of antigen can be transported to the Peyer's patches. As evident from the release studies, most of this associated ovalbumin will only be released after disintegration of the chitosan microparticles. Since chitosan is biodegradable, this might happen after M-cell uptake and chitinases are expected to play an important role in this degradation process [22]. Electrophoretic studies did not show any signi"cant degradation of the released ovalbumin after the loading and release process. Although in vivo studies are needed to exclude protein degradation, it is assumed that the association between chitosan microparticles and ovalbumin does not a!ect the protein structure and its immunogenic properties. The in vivo studies showed that the present chitosan microparticles can be targeted to the Peyer's patches. Since uptake by M-cells is the "rst step in oral vaccination, our results suggest that these chitosan microparticles are promising as an e$cient oral vaccine delivery system. Since real vaccines in general provoke stronger immune responses than model antigens like ovalbumin, the chitosan microparticles will be further investigated after loading with vaccines against diphtheria or polio.
693
Furthermore, the potential of cholera toxoids bound to the surface of these microparticles will be evaluated with regard to the enhancement of M-cell uptake.
5. Conclusion The chitosan microparticles, as developed in the present study, have suitable size and zeta potential to be taken up by the M-cells of the Peyer's patches. These microparticles show a high loading capacity and loading e$cacy for the model antigen ovalbumin. About 90% of the ovalbumin remained in the microparticles after release studies for 4 h in PBS. Since both the digestion experiments, the CLSM and "eld emission SEM visualization studies revealed that the loaded ovalbumin was entrapped in the microparticles, high amounts of the model antigen are expected to be transported into the Peyer's patches. These microparticles show a suitable release pro"le for oral vaccination, and most of the ovalbumin remains associated with the microparticles after resuspending in PBS. As expected, these chitosan microparticles can also be taken up in vivo by murine Peyer's patches. In conclusion, these chitosan microparticles show suitable in vitro and in vivo characteristics for oral vaccination and are therefore a promising carrier system for this particular purpose.
References [1] Aizpurua HJ, Russell-Jones GJ. Identi"cation of classes of proteins that provoke an immune response upon oral feeding. J Exp Med 1988;67:440}51. [2] Alpar HO, Eyles JE, Williamson ED. Oral and nasal immunization with microencapsulated clinically relevant proteins. STP Pharma Sci 1998;8:31}3. [3] Shalaby WSW. Development of oral vaccines to stimulate mucosal and systemic immunity: barriers and novel strategies. Clin Immunol Immunopath 1995;74:127}34. [4] Kreuter J. Nanoparticles and microparticles for drug and vaccine delivery. J Anat 1996;189:503}5. [5] Ko#er N, Ruedl C, Rieser C, Wick G, Wolf H. Oral immunization with poly-(D,L-lactide-co-glycolide) and poly-(L-lactic acid) microspheres containing pneumotropic bacterial antigens. Int Arch Allergy Immunol 1997;113:424}31. [6] Eldridge JH, Hammond CJ, Meulbroek JA, Staas JK, Gilley RM, Tice TR. Controlled vaccine release in the gut-associated lymphoid tissues. I. Orally administered biodegradable microspheres target the Peyer's patches. J Control Release 1990;11:205}14. [7] Fasano A. Innovative strategies for the oral delivery of drugs and peptides. TiBtech 1998;16:152}7. [8] Lydyard P, Grossi C. Secondary lymphoid organs and tissues. In: Roitt I, Brosto! J, Male D, editors. Immunology. London: Mosby, 1998. p. 33}8. [9] Illum L. Chitosan and its use as a pharmaceutical excipient. Pharm Res 1998;15:1326}31. [10] Calvo P, Remunan-Lopez C, Vila-Jato JL, Alonso MJ. Novel hydrophilic chitosan-polyethylene oxide nanoparticles as protein carriers. J Appl Polym Sci 1997;16:125}32.
694
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
[11] Artursson P, Lindmark T, Davis SS, Illum L. E!ect of chitosan on the permeability of monolayers of intestinal epithelial cells (Caco-2). Pharm Res 1994;11:1358}61. [12] Jameela SR, Kumary TV, Lal AV, Jayakrishnan A. Progesterone-loaded chitosan microspheres: a long acting biodegradable controlled delivery system. J Control Release 1998;52:17}24. [13] Illum L, Farraj NF, Davis SS. Chitosan as a novel nasal delivery system for peptide drugs. Pharm Res 1994;11:1186}9. [14] Jabbar-Gill I, Neil-Faber A, Rappuoli R, Davis SS, Illum L. Stimulation of mucosal and systemic antibody responses against Bordella pertussis "lamentous haemagglutinin and recombinant pertussis toxin after nasal administration with chitosan in mice. Vaccine 1998;16:2039}46. [15] Witschi C, Mrsny RJ. In vitro evaluation of microparticles and polymer gels for use as nasal platforms for protein delivery. Pharm Res 1999;16:382}90. [16] Roy K, Mao H-Q, Huang S-K, Leong KW. Oral gene delivery with chitosan DNA nanoparticles generates immunologic protection in a murine model of peanut allergy. Nature Med 1999;5:387}91.
[17] Chen SC, Jones DH, Fynan EF, Farrar GH, Clegg JCS, Greenberg HB, Herrmann JE. Protective immunity by oral immunization with a rotavirus DNA vaccine encapsulated in microparticles. J Virol 1998;72:5757}61. [18] Berthold A, Cremer K, Kreuter J. Preparation and characterization of chitosan microspheres as drug carrier for prednisolone sodium phosphate as model for anti-in#ammatory drugs. J Control Release 1996;39:17}25. [19] Lowry O, Rosebrough N, Farr A, Randall R. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265. [20] Lorenzo-Lamosa ML, Remunan-Lopez C, Vila-Jato JL, Alonso MJ. Design of microencapsulated chitosan microspheres for colonic drug delivery. J Control Release 1998;52:109}18. [21] Berthold A, Cremer K, Kreuter J. In#uence of crosslinking on the acid stability and physicochemical properties of chitosan microspheres. STP Pharma Sci 1996;6:358}64. [22] Muzzarelli RA. Human enzymatic acitivities related to the therapeutic administration of chitin derivatives. Cell Mol Life Sci 1997;53:131}40.
Chitosan microparticles for oral vaccination: preparation, characterization and preliminary in vivo uptake studies in murine Peyer's patches I.M. van der Lubben , J.C. Verhoef , A.C. van Aelst, G. Borchard , H.E. Junginger * Leiden/Amsterdam Center for Drug Research, Division of Pharmaceutical Technology, P.O. Box 9502, 2300 RA Leiden, Netherlands Laboratory of Experimental Plant Morphology and Cell Biology, Arboretumlaan 4, 6703 BD Wageningen, Netherlands Received 15 December 1999; accepted 3 July 2000
Abstract Although oral vaccination has numerous advantages over parenteral injection, degradation of the vaccine in the gut and low uptake in the lymphoid tissue of the gastrointestinal tract still complicate the development of oral vaccines. In this study chitosan microparticles were prepared and characterized with respect to size, zeta potential, morphology and ovalbumin-loading and -release. Furthermore, the in vivo uptake of chitosan microparticles by murine Peyer's patches was studied using confocal laser scanning microscopy (CLSM). Chitosan microparticles were made according to a precipitation/coacervation method, which was found to be reproducible for di!erent batches of chitosan. The chitosan microparticles were 4.3$0.7 lm in size and positively charged (20$1 mV). Since only microparticles smaller than 10 lm can be taken up by M-cells of Peyer's patches, these microparticles are suitable to serve as vaccination systems. CLSM visualization studies showed that the model antigen ovalbumin was entrapped within the chitosan microparticles and not only associated to their outer surface. These results were veri"ed using "eld emission scanning electron microscopy, which demonstrated the porous structure of the chitosan microparticles, thus facilitating the entrapment of ovalbumin in the microparticles. Loading studies of the chitosan microparticles with the model compound ovalbumin resulted in loading capacities of about 40%. Subsequent release studies showed only a very low release of ovalbumin within 4 h and most of the ovalbumin (about 90%) remained entrapped in the microparticles. Because the prepared chitosan microparticles are biodegradable, this entrapped ovalbumin will be released after intracellular digestion in the Peyer's patches. Initial in vivo studies demonstrated that #uorescently labeled chitosan microparticles can be taken up by the epithelium of the murine Peyer's patches. Since uptake by Peyer's patches is an essential step in oral vaccination, these results show that the presently developed porous chitosan microparticles are a very promising vaccine delivery system. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Chitosan microparticles; Ovalbumin; Oral vaccination; Confocal laser scanning microscopy; Scanning electron microscopy; Peyer's patches
1. Introduction Although the introduction of an e$cient oral vaccine would diminish costs, patient uncomfort and the need for quali"ed personnel to administer the vaccine, most vaccines still have to be administered by injection. The reasons for unsuccessful development of a potent oral vaccine carrier are still hindered due to degradation of the antigen in the gastrointestinal (GI) tract and ine$-
* Corresponding author. Tel.: #131-715274308; fax: #131715274565. E-mail address: [email protected] (H.E. Junginger).
cient targeting to the site of action in the gut. Before the antigen reaches the M-cells of the Peyer's patches in the GI tract, it has to pass the stomach with its low pH and several inactivating enzymes. Even if the vaccine has arrived near the Peyer's patches, an immune response is not always elicited. The antigen might not reach the dome of the Peyer's patch due to ine$cient uptake, or is not potent enough to provoke a su$cient immune reaction [1]. Several studies have shown that by associating the vaccine with a number of microparticulate drug carrier systems, the uptake by M-cells is enhanced and the degradation of the vaccine in the GI tract is prevented [2}6]. In this way, a &Trojan horse' principle is applied: the vaccines are delivered hidden inside biodegradable
0142-9612/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 2 3 1 - 3
688
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
microspheres that can be taken up by endocytosis by these intestinal M-cells [7,8]. Microparticles smaller than 10 lm are taken up by the M-cells and transported into the Peyer's patches. Most microparticles larger than 5 lm stay in the Peyer's patches, while microparticles smaller than 5 lm are transported through the e!erent lymphatics [6]. Chitosan has been shown to be a suitable vaccine carrier. Chitosan is the deacetylated form of chitin (polyb-(1P4)-N-acetyl-D-glucosamin), the second most abundant polymer in nature. Chitin is mostly obtained from exoskeletons of marine arthropods. Molecular weight and degree of acetylation determine the properties of chitosan. Because of its biocompatibility, biodegradability, low cost and ability to open intercellular tight junctions, this polymer is a valuable excipient for oral drug delivery systems [9]. Recently, numerous studies on chitosan as a drug absorption enhancer have been published. Chitosan formulations are used for ocular [10], oral [11], parenteral [12] and nasal delivery [13}15], as well as for DNA transfection studies [16]. Furthermore, chitosan can easily form microparticles. Advances in microparticulate drug delivery research have opened up the way to apply these techniques for oral vaccination. Due to the high protein binding properties of some types of chitosan microparticles, they are also potential candidates for oral delivery of antigens [10]. Mild preparation can protect the proteins when they are incorporated during preparation of the microparticles [10,18]. In order to circumvent protein denaturation conditions, chitosan microparticles can be loaded passively. Such a mild loading procedure was described by Jameela et al. [12]. Besides antigens, also DNA coding for antigens can be taken up by the M-cells of Peyer's patches. Transcription of this DNA leads to the production of the antigen. Roy et al. [16] reported that in this way chitosan nanospheres can also be used as an oral vaccine delivery system by means of transfection. Chen et al. [17] showed that poly (lactide-coglycolide) microparticles loaded with rotavirus VP6 DNA provoked an immune reaction after oral administration. In this study chitosan microparticles for oral vaccination were prepared. Because of the non-toxic features and potent antigen binding properties, chitosan polymers are expected to be promising candidates for oral vaccination. Chitosan microparticles were made by a coacervation/precipitation method and characterized with respect to oral vaccination. The morphology of the microparticles was evaluated using light microscopy, confocal laser scanning microscopy (CLSM) and scanning electron microscopy. The size, charge, loading and release characteristics for the model antigen ovalbumin were investigated. Furthermore, initial in vivo uptake studies of #uorescently labeled chitosan microparticles by murine Peyer's patches were performed and vizualized by CLSM.
2. Materials and methods 2.1. Materials Chitosan was a generous gift from Primex (Avaldsnes, Norway). The viscosity of the chitosan used was measured as 1% (w/v) chitosan in 1% (v/v) acetic acid in MilliQ water on a rotation viscosimeter (Haake, Karlsruhe, Germany) in our laboratory and appeared to be 13 mpas. This di!ered from the viscosity mentioned by the manufacturer (20 mpas). The degree of deacetylation, as determined by the supplier was 93%. FITC-ovalbumin (494/520) and Bodipy威 (665/767) were purchased from Molecular Probes (Leiden, The Netherlands). Ovalbumin and Tween威 80 were obtained from Sigma (St. Louis, USA). All other reagents were of analytical grade. PhastGels (gradient 8}25) were purchased from Pharmacia Biotech (Uppsala, Sweden). Female Balb/c mice (9}12 weeks old) were ordered from Charles River (Sulzfeld, Germany). 2.2. Preparation of microparticles A 0.25% (w/v) chitosan solution was prepared in a mixture of 2% (v/v) acetic acid and 1% (w/v) Tween威80. Then 2 ml of 10% (w/v) sodium sulfate was added dropwise (about 1 ml/min) to 200 ml chitosan solution under magnetic stirring and continuous sonication. After adding the sodium sulfate, stirring and sonication were continued for 20 min. The microparticle suspension was subsequently centrifuged for 25 min (2750 rpm). The pellet was resuspended in MilliQ water to wash the microparticles and centrifuged again. This washing procedure was repeated twice before freeze-drying of the pellet overnight. Freeze-drying was performed using a Christ freeze-dryer (Osterode am Harz, Germany). 2.3. Characterization of chitosan microparticles 2.3.1. Size and zeta potential The size of the chitosan microparticles was determined using an Accusizer 770 (PSS, Santa Barbara, CA, USA). The zeta potential of the chitosan microparticles was measured in demineralized water at neutral pH, using a Malvern 2000 zetasizer (Herrenberg, Germany). 2.3.2. Morphology Light microscopy (Axioskop-type microscope; Zeiss, Weesp, The Netherlands) was performed for initial visualization of the chitosan microparticles. Morphology was evaluated using the following parameters: approximate size, form, uniformity, and the formation of "bers, gel or aggregates. Chitosan microparticles were analyzed after resuspension in water without staining or "xation. Further morphology studies were carried out by means of confocal laser scanning microscopy (CLSM) and
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
scanning electron microscopy (SEM). For CLSM studies, 1 ml of 1% (w/v) chitosan microparticles was incubated with 3 mg FITC-ovalbumin. A MRC-600 Lasersharp system (Bio-Rad Laboratories, Richmond, CA, USA) linked to a Zeiss IM 35 inverted microscope (Zeiss, Oberkochen, Germany) was used for this purpose, and several z-scans over a depth of 10 lm were made. For "eld emission SEM analysis, lyophilized chitosan microparticles were adhered to double-sided tape (Carbon adhesive tabs, 12 mm, electron microscopy studies, Fort Washington, USA). The microparticles were sputter coated with 5 nm platina in a preparation chamber (CT 1500 HF, Oxford Instruments, Oxfordshire, UK). Then the sputtered chitosan microparticles were studied with a JEOL 6300F "eld emission SEM (JEOL, Schiphol, The Netherlands) at 5 kV. Digital images were made and stored. 2.3.3. Ovalbumin loading The ovalbumin loading of microparticles was performed by incubating 1% (w/v) chitosan microparticles and 0.5}2.5% (w/v) ovalbumin in phosphate bu!ered saline (PBS; pH 7.3) under shaking at 253C. After incubation for 180 min, the suspension was centrifuged (1400 rpm for 30 min) to remove the unloaded ovalbumin. The loading degree was determined by quantifying the non-bound ovalbumin in the supernatant with the Lowry protein assay method [19]. Both loading capacity (LC) and loading e$cacy (LE) were determined: LC"[(total amount ovalbumin)!(free ovalbumin)] /weight microparticles, LE"[(total amount ovalbumin)!(free ovalbumin)] /total ovalbumin. 2.3.4. Ovalbumin release Ovalbumin release from chitosan microparticles was determined in PBS (pH 7.3). To load the microparticles, 3.5 ml of a 1% (w/v) chitosan microparticle suspension containing 0.5% (w/v) ovalbumin was incubated for 3 h. After centrifuging (1400 rpm for 30 min) the loaded microparticles were resuspended in PBS (pH 7.3) to make a 1% (w/v) microparticle suspension. Samples were incubated at 373C under mild shaking. After 15, 30, 45, 60, 90, 120, 180 and 240 min, the tubes were given a spin-o! and samples of 250 ll of the supernatant were taken. These samples were replaced by PBS (pH 7.3). The non-bound ovalbumin in PBS was determined with the Lowry protein assay [19]. 2.3.5. Degradation of ovalbumin-loaded microparticles To study if the loaded microparticles still contained ovalbumin after release studies, the chitosan microparticles were dissolved. This was done by incubating ovalbumin without microparticles, non-loaded micro-
689
particles and ovalbumin-loaded microparticles in 20% (v/v) acetic acid (pH 2.0) for a period of 2 h. Thereafter, the degraded microparticles were separated from the digestion solution by centrifuging (1400 rpm for 20 min). The ovalbumin in the supernatant was detected with the Lowry protein assay [19]. 2.3.6. Gel electrophoresis of released ovalbumin Released ovalbumin from loaded microparticles was analyzed by phast system威 gel electrophoresis to determine possible protein degradation by the loading and release processes. As controls both empty (non-loaded) microparticles and ovalbumin was used. Since hardly any ovalbumin was released over a period of 240 min (Fig. 5), forced release was also applied. This was done by loading the microparticles in MilliQ water with ovalbumin and subsequently performing release studies in PBS (pH 7.3). The ovalbumin loading of microparticles was performed by incubating 1% (w/v) chitosan microparticles and 0.5% (w/v) ovalbumin in MilliQ water under shaking at 253C. Release was performed as described in Section 2.3.4. Gel electrophoresis was performed on an 8}25 gradient gel and run for 87 Vh. Approximately 0.4 mg ovalbumin per lane was applied. 2.4. In vivo uptake of chitosan microparticles by murine Peyer+s patches For in vivo studies 1 ml of 1% (w/v) chitosan microparticles was incubated with 3 mg FITC-ovalbumin in PBS (pH 7.3) for three hours at 253C. After incubating, the suspension was centrifuged for 20 min (2750 rpm) to remove the non-bound FITC-ovalbumin. The pellet was resuspended in 300 ll PBS (pH 7.3) and fed intragastrically to the mice. In order to stain the membranes of the intestinal epithelial cells, the mice were also supplied with a non-speci"c lipophilic dye for cell membranes (Bodipy威). Groups of female Balb/c mice were fasted overnight and then fed intragastrically with a 150 ll Bodipy威 solution (0.1% in PBS : Glycerin : PEG400"1 : 1 : 1 (v/v/v)). After 3 h, the animals were also fed with either FITCovalbumin loaded microparticles (300 ll) or a solution of FITC-ovalbumin (3 mg in 300 ll PBS (pH 7.3)). The following 3 groups were used in this in vivo experiment: group 1 consisting of 6 mice fed with Bodipy威 and subsequently with FITC-ovalbumin loaded microparticles, group 2 consisting of 4 mice fed with Bodipy威 and subsequently with FITC-ovalbumin solution, and group 3 consisting of 3 mice only fed with Bodipy威. Three hours after the last feeding, the mice were sacri"ed by cervical dislocation. Then the Peyer's patches were dissected and immediately analyzed using CLSM. Since the spectra of the labels are not overlapping, both FITC-ovalbumin and Bodipy威 could be analyzed using two di!erent lasers at the same time. FITC-ovalbumin
690
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
was analyzed using a 633 nm laser and Bodipy威 using a 488 nm laser.
3. Results 3.1. Size and zeta potential of chitosan microparticles The size of the prepared non-loaded chitosan microparticles was found to be 4.23$0.67 lm (mean$SD; n"3) and the zeta potential was positive (20$1 mV). The small di!erences in size and zeta potential between the di!erent batches of microparticles demonstrate the good reproducibility of the method of preparation used. 3.2. Morphology of chitosan microparticles The morphology of the presented chitosan microparticles was analyzed by light microscopy after washing and freeze-drying. As depicted in Fig. 1, chitosan microparticles neither form a gel when resuspended in water, nor were forming big aggregates of microparticles, were not accompanied by "bers, and all showed reasonable uniformity. Fig. 2 shows #uorescently labeled chitosan microparticles after FITC-ovalbumin labeling. Three images at di!erent depths are presented, and it is evident that the ovalbumin is not only associated to the surface of the microparticles, but also entrapped within the microparticles. In all 3 scans the #uorescent label is clearly visible in the entire micoparticle. Field emission SEM demonstrated that pores are present at the rough surface of the chitosan microparticles (Fig. 3A). By focusing in the pores, the inside of the chitosan microparticles was visualized. As evident from Fig. 3B, the inside of the chitosan microparticles also has a very porous structure.
Fig. 2. Z-scans of #uorescently labeled chitosan microparticles using CLSM (60;). The #uorescent label is taken up by the microparticles and visible in all scans at di!erent depths ((A) 0.0 lm; (B) 1.9 lm; (C) 2.9 lm).
Fig. 1. Light microscopic image of chitosan microparticles (400;).
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
691
Fig. 4. Ovalbumin loading of chitosan microparticles at di!erent concentrations of total ovalbumin (LC loading capacity, LE loading e$cacy). Data are expressed as mean$SD of 10 experiments.
ovalbumin in demineralized water at neutral pH was !30$3 mV (mean$SD; n"3). 3.4. Ovalbumin release from chitosan microparticles
Fig. 3. Surface visualization of the chitosan microparticles using "eld emission SEM. (A) A detail of the surface of the microparticles. (B) A detail inside a pore of the chitosan microparticles. Scale bars are 1 lm.
3.3. Ovalbumin loading of chitosan microparticles The microparticles were loaded with di!erent amounts of ovalbumin by incubation of a 1% (w/v) chitosan microparticle suspension with 0.5}2% (w/v) ovalbumin. Both loading e$cacies (LE) and loading capacities (LC) were determined. Fig. 4 shows that the LC is not substantially in#uenced by the amount of ovalbumin available in the loading solution. Thus, the more ovalbumin is o!ered to the microparticles, the more ovalbumin remains unbound during the loading process. Therefore, 0.5% (w/v) ovalbumin in the loading solution was selected as the optimal concentration. When 1% of microparticles was incubated with 0.5% of ovalbumin, a loading percentage of 32$4% (n"10) for independently made batches was obtained. Under these conditions, a very high LE of 85$3% was obtained, indicating that only a small amount of ovalbumin was lost during the loading process. The size of the loaded chitosan microparticles was measured to be 5.7$0.6 lm. The zeta potential of loaded microparticles could not be determined, since the loading was performed in PBS. The ions present in this bu!er highly in#uence the zeta potential. The zeta potential of
Ovalbumin release from chitosan microparticles in PBS (pH 7.3) was determined over a time span of 240 min. After a release of about 10% in the "rst 60 min, no ovalbumin was released during the following 180 min. This indicates that most ovalbumin remained entrapped in the microparticles under these conditions. To verify whether the entrapped ovalbumin was still present in the chitosan microparticles after the loading and subsequent release studies, the loaded microparticles were dissolved. Non-loaded microparticles, ovalbuminloaded microparticles and ovalbumin suspensions were incubated in acetic acid at pH 2. After 2 h of incubation, the chitosan microparticles were found to be completely disintegrated and about 30% of the loaded ovalbumin could be determined in the suspension with the digested microparticles. When the microparticles were incubated for longer time periods at pH 2, the ovalbumin became denatured and appeared as a visible clot in the suspension. 3.5. Gel electrophoresis To investigate if the ovalbumin was not degraded during the loading and release processes, phast system威 gel electrophoresis was performed to compare released ovalbumin to non-treated ovalbumin. Because ovalbumin release from chitosan microparticles in PBS was very low, release in MilliQ water (forced release) was applied to analyze the ovalbumin associated with the microparticles. Fig. 5 shows the gel electrophoresis of released ovalbumin after normal and forced release, empty (non-loaded) chitosan microparticles and ovalbumin solutions in PBS bu!er and MilliQ water. No di!erence could be observed between the released ovalbumin and the ovalbumin solutions and no protein staining in the lane with empty microparticles. Since these results did not show any di!erences in molecular weight,
692
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
Furthermore, no uptake of FITC-ovalbumin was found in intestinal epithelium without Peyer's patches.
4. Discussion
Fig. 5. Gel electrophoresis of released ovalbumin (lane 1, supernatant of empty microparticles; lane 2, released ovalbumin from microparticles loaded and released in PBS; lane 3, released ovalbumin form microparticles loaded in MilliQ water and released in PBS; lane 4, released ovalbumin from microparticles loaded in PBS and released in MilliQ; lane 5, ovalbumin dissolved in PBS and lane 6, ovalbumin dissolved in MilliQ).
it is assumed that no substantial degradation of ovalbumin occurred during the loading and release processes. 3.6. In vivo uptake studies Fig. 6 shows typical examples of the uptake of chitosan microparticles present in mouse intestinal epithelium of the Peyer's patches. In the left image the epithelial cells are visualized and it is evident that the cell membranes have taken up some of the intragastrically fed Bodipy威. In this way the intestinal epithelial barrier could clearly be visualized. In the right image, transport of FITCovalbumin loaded chitosan microparticles across the epithelial barrier is observed. At the bottom of this same image also some #uorescent labeling is visible. This may be due to earlier transport of loaded particles across the cells. No transport could be determined in mice fed with ovalbumin solutions without chitosan microparticles.
Recently, chitosan nano- or microparticles have been prepared according to several precipitation/coacervation methods and some of these particles showed good antigen binding capacities [10]. Chitosan microspheres were also designed for colonic drug delivery [18,20]. For oral vaccination, microparticulate vaccine carrier systems not only need to associate a high amount of antigen, but also require speci"c release properties. After oral administration of such systems the vaccine should be well entrapped and protected from degradation in the GI-tract, and should only be released from the carrier system after uptake by the M-cells of the Peyer's patches. Besides, nano- or microparticles should not exceed 10 lm in size, and the hydrophobicity and antigens presented on the surface of the carrier system are important parameters [1,6]. The chitosan microparticles described in this study are much smaller than 10 lm, being a suitable size for M-cells uptake. Although they are hydrophilic compared to poly(D,L-lactide-co-glycolide), our preliminary in vivo studies showed that in vivo uptake of chitosan microparticles in Peyer's patches could be visualized. These observations indicate that hydrophilicity of chitosan microparticles may not be a principal factor. Microparticles (10 lm are taken up by the M-cells and transported to the dome of the Peyer's patches. Microparticles (5 lm are then transported to the spleen and lymph nodes, where speci"c IgM and IgG are produced. Since the cumulative size distribution showed that also microparticles between 5 and 10 lm were formed, these microparticles might stay in the Peyer's patches. In this case additional antigen speci"c IgM is formed [6]. The zeta potential of the prepared non-loaded microparticles is positive (20$1 mV) at neutral pH. This probably explains why microparticles do not aggregate in an
Fig. 6. Uptake of FITC-ovalbumin loaded chitosan microparticles by the intestinal epithelium of the Peyer's patches. (left) The epithelial cells were stained with Bodipy威 (arrow); in this way the epithelial barrier could be located. (right) Transport of FITC-ovalbumin loaded chitosan microparticles across the epithelial barrier (arrow). At the bottom of the image earlier transported ovalbumin is also visible. In none of the mice, fed with FITC-ovalbumin in PBS, intestinal uptake of FITC-ovalbumin was observed.
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
aqueous solution. Since cell membranes are negatively charged, these microparticles are also expected to be more easily associated and subsequently taken up by the membranes than negatively charged microparticles. Because the zeta potential is highly in#uenced by the presence of salts, the zeta potential could not be determined after loading the microparticles in PBS. In contrast to earlier studies with chitosan microspheres [18], no glutaraldehyde was required during the preparation of the present microparticles. Berthold et al. [21] found that crosslinking with glutardialdehyde makes the microspheres more stable to acids. However, the chitosan microparticles in the present study were only disintegrating after incubation for more than 2 h at a pH value of 2, compared to 30 min for the microspheres prepared by Berthold et al. [21] at similar ionic strength and pH values. Possible reasons for the increased stability of the present chitosan microparticles are yet unknown. Results from the in vitro CLSM studies showed that ovalbumin not only associates to the surface, but also enters within the chitosan microparticles. With this technique it was possible to visualize both the inside and the surface, and the #uorescent label appeared to be present throughout the microparticles. The observation that ovalbumin was able to enter the microparticles was con"rmed using "eld emission SEM. These studies demonstrated that the chitosan microparticles had a very porous structure. It is therefore likely that the ovalbumin enters the pores during the loading process and associates by electrostatic interactions to the inside of the chitosan microparticles. Because the present chitosan microparticles have a high loading capacity, large quantities of antigen can be transported to the Peyer's patches. As evident from the release studies, most of this associated ovalbumin will only be released after disintegration of the chitosan microparticles. Since chitosan is biodegradable, this might happen after M-cell uptake and chitinases are expected to play an important role in this degradation process [22]. Electrophoretic studies did not show any signi"cant degradation of the released ovalbumin after the loading and release process. Although in vivo studies are needed to exclude protein degradation, it is assumed that the association between chitosan microparticles and ovalbumin does not a!ect the protein structure and its immunogenic properties. The in vivo studies showed that the present chitosan microparticles can be targeted to the Peyer's patches. Since uptake by M-cells is the "rst step in oral vaccination, our results suggest that these chitosan microparticles are promising as an e$cient oral vaccine delivery system. Since real vaccines in general provoke stronger immune responses than model antigens like ovalbumin, the chitosan microparticles will be further investigated after loading with vaccines against diphtheria or polio.
693
Furthermore, the potential of cholera toxoids bound to the surface of these microparticles will be evaluated with regard to the enhancement of M-cell uptake.
5. Conclusion The chitosan microparticles, as developed in the present study, have suitable size and zeta potential to be taken up by the M-cells of the Peyer's patches. These microparticles show a high loading capacity and loading e$cacy for the model antigen ovalbumin. About 90% of the ovalbumin remained in the microparticles after release studies for 4 h in PBS. Since both the digestion experiments, the CLSM and "eld emission SEM visualization studies revealed that the loaded ovalbumin was entrapped in the microparticles, high amounts of the model antigen are expected to be transported into the Peyer's patches. These microparticles show a suitable release pro"le for oral vaccination, and most of the ovalbumin remains associated with the microparticles after resuspending in PBS. As expected, these chitosan microparticles can also be taken up in vivo by murine Peyer's patches. In conclusion, these chitosan microparticles show suitable in vitro and in vivo characteristics for oral vaccination and are therefore a promising carrier system for this particular purpose.
References [1] Aizpurua HJ, Russell-Jones GJ. Identi"cation of classes of proteins that provoke an immune response upon oral feeding. J Exp Med 1988;67:440}51. [2] Alpar HO, Eyles JE, Williamson ED. Oral and nasal immunization with microencapsulated clinically relevant proteins. STP Pharma Sci 1998;8:31}3. [3] Shalaby WSW. Development of oral vaccines to stimulate mucosal and systemic immunity: barriers and novel strategies. Clin Immunol Immunopath 1995;74:127}34. [4] Kreuter J. Nanoparticles and microparticles for drug and vaccine delivery. J Anat 1996;189:503}5. [5] Ko#er N, Ruedl C, Rieser C, Wick G, Wolf H. Oral immunization with poly-(D,L-lactide-co-glycolide) and poly-(L-lactic acid) microspheres containing pneumotropic bacterial antigens. Int Arch Allergy Immunol 1997;113:424}31. [6] Eldridge JH, Hammond CJ, Meulbroek JA, Staas JK, Gilley RM, Tice TR. Controlled vaccine release in the gut-associated lymphoid tissues. I. Orally administered biodegradable microspheres target the Peyer's patches. J Control Release 1990;11:205}14. [7] Fasano A. Innovative strategies for the oral delivery of drugs and peptides. TiBtech 1998;16:152}7. [8] Lydyard P, Grossi C. Secondary lymphoid organs and tissues. In: Roitt I, Brosto! J, Male D, editors. Immunology. London: Mosby, 1998. p. 33}8. [9] Illum L. Chitosan and its use as a pharmaceutical excipient. Pharm Res 1998;15:1326}31. [10] Calvo P, Remunan-Lopez C, Vila-Jato JL, Alonso MJ. Novel hydrophilic chitosan-polyethylene oxide nanoparticles as protein carriers. J Appl Polym Sci 1997;16:125}32.
694
I.M. van der Lubben et al. / Biomaterials 22 (2001) 687}694
[11] Artursson P, Lindmark T, Davis SS, Illum L. E!ect of chitosan on the permeability of monolayers of intestinal epithelial cells (Caco-2). Pharm Res 1994;11:1358}61. [12] Jameela SR, Kumary TV, Lal AV, Jayakrishnan A. Progesterone-loaded chitosan microspheres: a long acting biodegradable controlled delivery system. J Control Release 1998;52:17}24. [13] Illum L, Farraj NF, Davis SS. Chitosan as a novel nasal delivery system for peptide drugs. Pharm Res 1994;11:1186}9. [14] Jabbar-Gill I, Neil-Faber A, Rappuoli R, Davis SS, Illum L. Stimulation of mucosal and systemic antibody responses against Bordella pertussis "lamentous haemagglutinin and recombinant pertussis toxin after nasal administration with chitosan in mice. Vaccine 1998;16:2039}46. [15] Witschi C, Mrsny RJ. In vitro evaluation of microparticles and polymer gels for use as nasal platforms for protein delivery. Pharm Res 1999;16:382}90. [16] Roy K, Mao H-Q, Huang S-K, Leong KW. Oral gene delivery with chitosan DNA nanoparticles generates immunologic protection in a murine model of peanut allergy. Nature Med 1999;5:387}91.
[17] Chen SC, Jones DH, Fynan EF, Farrar GH, Clegg JCS, Greenberg HB, Herrmann JE. Protective immunity by oral immunization with a rotavirus DNA vaccine encapsulated in microparticles. J Virol 1998;72:5757}61. [18] Berthold A, Cremer K, Kreuter J. Preparation and characterization of chitosan microspheres as drug carrier for prednisolone sodium phosphate as model for anti-in#ammatory drugs. J Control Release 1996;39:17}25. [19] Lowry O, Rosebrough N, Farr A, Randall R. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265. [20] Lorenzo-Lamosa ML, Remunan-Lopez C, Vila-Jato JL, Alonso MJ. Design of microencapsulated chitosan microspheres for colonic drug delivery. J Control Release 1998;52:109}18. [21] Berthold A, Cremer K, Kreuter J. In#uence of crosslinking on the acid stability and physicochemical properties of chitosan microspheres. STP Pharma Sci 1996;6:358}64. [22] Muzzarelli RA. Human enzymatic acitivities related to the therapeutic administration of chitin derivatives. Cell Mol Life Sci 1997;53:131}40.