Preparation and properties of ionically cross-linked chitosan ...

Research ArticleReceived: 31 December 2007, Revised: 26 June 2008, Accepted: 11 September 2008, Published online in Wiley InterScience: 17 November 2008(www.interscience.wiley.com) DOI: 10.1002/pat.1306Preparation and properties of ionicallycross-linked chitosan nanoparticlesHui Liu a,b and Changyou Gao a,b *Chitosan nanoparticles were fabricated by a method of tripolyphosphate (TPP) cross-linking. The influence offabrication conditions on the physical properties and drug loading and release properties was investigated bytransmission electron microscopy (TEM), dynamic light scattering (DLS), and UV–vis spectroscopy. The nanoparticlescould be prepared only within a zone of appropriate chitosan and TPP concentrations. The particle size and surfacezeta potential can be manipulated by variation of the fabrication conditions such as chitosan/TPP ratio andconcentration, solution pH and salt addition. TEM observation revealed a core–shell structure for the as-preparednanoparticles, but a filled structure for the ciprofloxacin (CH) loaded particles. Results show that the chitosannanoparticles were rather stable and no cytotoxicity of the chitosan nanoparticles was found in an in vitro cell cultureexperiment. Loading and release of CH can be modulated by the environmental factors such as solution pH andmedium quality. Copyright ß 2008 John Wiley & Sons, Ltd.Keywords: chitosan; nanoparticles; ionic crosslinking; properties; drug deliveryINTRODUCTIONDrug delivery systems are an effective means to realize thesustained release of many kinds of drugs. Chitosan andits derivative materials, in particular with a range of particles,have been diversely employed in the field of drug delivery.More recently, chitosan nanoparticles have attracted muchattention by virtue of their large drug loading capacity, goodadsorption performance, and long shelf life. Several techniqueshave been developed to prepare chitosan nanoparticles, such asemulsion cross-linking, [1,2] ionic gelation, [3–5] spray drying, [6] andso on.Chitosan contains abundant amino and hydroxyl groups,which enable nanoparticle formulation via both physical andchemical cross-linking. [7] Covalent cross-linking is usuallyachieved by treatment of glutaraldehyde, which reacts withthe amino groups to form Schiff bases. Ionic cross-linking ofchitosan is a typical non-covalent interaction, which can berealized by association with negatively charged multivalent ionssuch as tripolyphosphate (TPP). [8,9] For pharmaceutical applications,physical cross-linking is more promising since thecross-linking is reversible and may largely avoid potential toxicityof the reagents.Although diverse efforts have been made to obtain thechitosan nanoparticles via TPP cross-linking following thepioneering work of Calvo et al. [10] and to explore the potentialpharmaceutical applications, [11–14] optimization of the fabricatingconditions and the comprehensive properties of the resultantchitosan nanoparticles is still an ongoing important topic in thisfield. These parameters not only affect the storage stability,loading, and release performance but also govern the interactionof the particles with different biological tissues where they areintroduced. In this context, special attention is paid to the drugdelivery and release properties. [10,15,16] The effects of pH of TPPPolym. Adv. Technol. 2009, 20 613–619solution, TPP concentration, and ionic strength on the entrapmentefficiency, release, and activity of lipase in chitosanhydrogel beads were also studied. [17] However, there are manydiscrepancies in terms of the alteration principles versusfabrication conditions, and explanation of the observedphenomena. [18,19] Moreover, little attention is paid to the storagestability and biocompatibility of the chitosan nanoparticles,which are critical issues for practical applications in thepharmaceutical field.In this work, the influence of fabrication conditions issystematically investigated. Stability and biocompatibility ofthe prepared chitosan nanoparticles are assessed. Loading andrelease of a model drug, ciprofloxacin (CH), is also performed totestify the potential application of these nanoparticles. CH is akind of antiseptic, which has an effective anti-bacterial effectagainst both Gram-positive and negative bacteria. To reduce the* Correspondence to: Prof. C. Gao, Department of Polymer Science and Engineering,Zhejiang University, Hangzhou 310027, ChinaE-mail: cygao@mail.hz.zj.cnabH. Liu, C. GaoKey Laboratory of Macromolecular Synthesis and Functionalization, Ministryof Education, Zhejiang University, Hangzhou 310027, ChinaH. Liu, C. GaoDepartment of Polymer Science and Engineering, Zhejiang University,Hangzhou 310027, ChinaContract/grant sponsors: Major State Basic Research Program of China;contract/grant number: 2005CB623902.Contract/grant sponsor: The Natural Science Foundation of China; contract/grant number: 20434030.Contract/grant sponsor: The National Science Fund for Distinguished YoungScholars of China; contract/grant number: 50425311.Copyright ß 2008 John Wiley & Sons, Ltd.613

H. LIU AND C. GAOdrug toxicity as a result of long term application, for examplearthritis damage for children, various sustained release carrierssuch as microcapsules, gels and liposomes have been developed.Yet so far less attention has been given to the nanoparticlecarriers, in particular the chitosan nanoparticles.EXPERIMENTALMaterialsChitosan (CS, M h ¼ 620 kDa, degree of deacetylation ¼ 90%,viscosity ¼ 115 cps) is a commercial product of Haidebei HalobiosCo., Ltd. (Ji’nan, China). Sodium tripolyphosphate (TPP) waspurchased from Dongsheng Chemical Reagent Factory (Wenzhou,China). Ciprofloxacin hydrochloride (CH, C 17 H 18 FN 3 O 3 HClH 2 O)was purchased from Zhejiang Medicine Co. Ltd., XinchangPharmaceutical Factory (Shaoxing, China). Phosphotungstic acidwas purchased from Sinopharm Group Chemical Reagent Co., Ltd.(Shanghai, China). All other chemicals were of analytical grade andused as received. The water used in all experiments was tripledistilled.Preparation of the chitosan nanoparticlesThe nanoparticles were obtained based on ionic gelation of TPPwith chitosan. [10] Briefly, chitosan and TPP were dissolved inacetic acid and triple distilled water to obtain solutions of 1.0–5.0and 0.25–2.0 mg/ml, respectively. The chitosan nanoparticleswere obtained upon addition of 14 ml TPP solution into 35 mlchitosan solution under mild mechanical stirring (550 rpm) atroom temperature. The drug-loaded nanoparticles were similarlyprepared by using a chitosan solution containing the drug.CharacterizationMorphological examination of the chitosan nanoparticles wasperformed by transmission electron microscopy (TEM) (JEOLJEM-200CX, Japan). The sample was stained with 2% (w/v)phosphotungstic acid and a drop of the sample was applied ontoa carbon-coated copper meshwork, and dried at roomtemperature. The acceleration voltage was 100 kV. Zeta potentialof the particles was recorded on a Zetasizer 2000. Each value wasaveraged from three parallel measurements.The particle size was measured by dynamic light scattering(DLS) (90 Plus/BI-MAS). The wavelength of laser is 658 nm.Brookhaven’s unique dust filter algorithm can remove theinfluence of a contaminating fraction of large particles (dust).Processing the fluctuating signal with a state-of-the-art digitalauto-correlator yields the particle’s diffusion coefficient, fromwhich the equivalent spherical particle size is calculated using theStokes–Einstein equation. When a broad distribution of spheres ispresent in the samples, the following auto-correlation function isused:ZgðtÞ ¼ GðGÞe Gt dGan average diameter and a distribution width (polydispersity) areobtained. It has to be noted that the most important pieces ofinformation obtained in photon correlation spectroscopy usingany multimodal distribution analysis are the positions of thepeaks and the ratio of the peak areas. The widths of the peaks,however, are not particularly reliable, since they will narrow,typically, with increasing duration of experiment until a limit isreached. For more details, the readers can refer to the machinemanual. Each measurement was taken for 1 min with a 908 fixedangle detector. The size was averaged from five parallelmeasurements.Stability of the chitosan nanoparticlesThe stability of the chitosan nanoparticles obtained at differentfabricating conditions was examined in terms of the particle size.The nanoparticle suspensions were stored at 48C and the sizes ofthe nanoparticles were monitored by DLS.Biocompatibility of the chitosan nanoparticlesIn vitro cell culture was performed to determine the biocompatibilityof the as-prepared chitosan nanoparticles. The chitosannanoparticles were sterilized by UV irritation before cell culture.Human dermal fibroblasts were isolated from foreskins androutinely expanded. [20] The cells were cultured at 378C, in 5% CO 2 ,and 95% humidity in DMEM supplemented with penicillin (100 U/ml), streptomycin (100 mg/ml), and 10% fetal bovine serum (FBS)(complete medium), which was changed every 3 days. Cells werepassaged at confluence and the 4–8th passage of fibroblasts witha density of 2 10 5 /ml were seeded in a 96-well polystyreneplate. After 24 hr, 60 ml of nanoparticle suspension was addedinto each well. The final nanoparticle concentration was highenough for close contact between the cells and the nanoparticles.The culture medium was changed every 2 days. Meanwhile,normal culture medium was used in parallel as a negative control.The cell proliferation was measured by methylthiazoletetrazolium(MTT) assay. [21] The absorbance was recorded at a wavelength of570 nm by a microplate reader (Bio-Rad model 550). Threeparallel samples were analysed and data were expressed asmean standard deviation.Quantification of the relative CH amount in the chitosannanoparticlesThe prepared nanoparticle suspension containing the CH wascentrifugated at 12,000 rpm for 70 min to obtain the supernatant,which was diluted over hundreds of times and the absorbance at276 nm was recorded by a UV–vis spectrophotometer (ShimadzuUV-2550). The supernatant of blank nanoparticles was obtainedunder the same conditions and used as a control. Theconcentration of CH in the supernatant was quantified byreferring to a calibration curve recorded from a known amount ofCH at the same condition. A lower concentration found in thesupernatant implies a higher loading efficiency in the chitosannanoparticles.614where G is related to the relaxation of the fluctuations. Forobtaining information from this equation, a method of Cumulantanalysis is applied. To get the solution, an approximation is usedby the non-negative constrained least squares (NNLS) algorithm.Moreover, Mie theory is used to calculate the weight and numberfractions from the intensity fractions. For routine determinations,Release of CHCH release was performed using a dialysis technique. The CHloaded chitosan nanoparticles were centrifugated at 12,000 rpmfor 70 min to obtain the nanoparticle precipitate, which waswashed with water twice. The nanoparticles were transferred intowww.interscience.wiley.com/journal/pat Copyright ß 2008 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2009, 20 613–619

IONICALLY CROSS-LINKED CHITOSAN NANOPARTICLESa dialytic-bag with a cut-off molecular weight of 3.5 kDa. Thedialytic-bag was then immersed into 100 ml phosphate bufferedsaline (PBS, pH 7.4) or H 2 Oat378C under continuous shaking. Atpredetermined time, 1 ml dialytic medium was taken out foranalysis and an equal volume of freshly prepared medium wassupplemented. Absorbance of the dialytic medium at 270 nm wasrecorded by an UV–vis spectroscopy. The released amount of CHwas quantified by referring to a calibration curve recorded fromknown amounts of CH at the same condition.RESULTS AND DISCUSSIONFabrication and properties of chitosan nanoparticleThe chitosan molecules are gelled when they encounter TPPmolecules via electrostatic interaction. [10] Firstly, the zone ofparticle formation is investigated. Chitosan was dissolved in 1.5%acetic acid solution to form a 10 mg/ml solution, which wasdiluted with water to various concentrations: 1.0, 2.0, 3.0, 4.0, and5.0 mg/ml. The pH value of these solutions was adjusted to 4.8.TPP was dissolved in water at various concentrations: 0.50, 0.75,1.0, 1.25, 1.50, 1.75, 2.0, and 2.5 mg/ml. Chitosan nanoparticleswere prepared by the addition of 14 ml TPP solution into 35 mlchitosan solution at room temperature under mechanicalagitation. By visual observation, the systems were classified asclear solution (), opalescent suspension (H), and aggregates (#).The results are summarized in Table 1. A clear solution wasobserved when both the chitosan and TPP concentrations weretoo small, whereas aggregates were formed spontaneously whenthey were too large. The zone of the opalescent suspension,which should represent a suspension of colloidal particles, wasfound when the chitosan and the TPP concentrations wereappropriate. Illuminating that formation of the nanoparticles isonly possible for some specific concentrations of chitosan andTable 1. Conditions for formation of the chitosan nanoparticlesCS (mg/ml)TPP (mg/ml)0.25 0.5 0.75 1.0 1.25 1.5 2.0 2.51.0 H H H H # # #2.0 H H H H H H #3.0 H H H H H H4.0 H H H H H H, clear solution; H, opalescent suspension; #, aggregates.TPP. Confirmed by DLS measurement and microscopy observation,the chitosan nanoparticles with a size of 300 nm wereobtained and well dispersed in the solution when theconcentration of chitosan and TPP was 2.0 and 1 mg/ml,respectively.Figure 1a shows that the number average size of the chitosannanoparticles was measured as 321 nm by DLS, which isconsistent with the TEM observation (Fig. 1c, d). No severeagglomeration of the particles was observed. A magnified TEMimage (Fig. 1d) illustrates that the chitosan nanoparticles had acore–shell structure, i.e. with denser skirts and looser interiors.After loading with CH, the particle size increased slightly to337 nm (Fig. 1b) with a homogeneous structure (Fig. 1e, f),suggesting that the drug was surely loaded.For the next study, the chitosan concentration was fixed at2 mg/ml with a pH value of 4.8. Fig. 2 shows that both the particlesize (Fig. 2a) and zeta potential (Fig. 2b) decreased along with theincrease of TPP concentration. This alteration tendency isconsistent with the results of Gan et al., [18] but is the oppositeFigure 1. (a,b) Size distribution and (c–f) TEM images of (a,c,d) chitosan nanoparticles and (b,e,f) CH loaded chitosan nanoparticles, respectively. Thechitosan and TPP concentrations were 2 and 1 mg/ml, respectively. This figure is available in color online at www.interscience.wiley.com/journal/pat.Polym. Adv. Technol. 2009, 20 613–619 Copyright ß 2008 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/pat615

H. LIU AND C. GAOFigure 2. Influence of TPP concentration on: (a) the particle size, (b) zeta potential, and (c) chitosan concentration on the particle size at a fixed CS to TPPmass ratio of 4:1. The error bar indicates the standard deviation averaged from five measurements unless specifically indicated.to the results of Pan. [22] We think the decrease is more acceptable,since at higher TPP concentration, the cross-linking degree ofthe chitosan nanoparticles is also higher. This will result in a morecompact particle structure. Moreover, the neutralization degreeof the charged amino groups is simultaneously improved, leadingto smaller net charge in the particles as shown in Fig. 2b. Due tothe compact structure and weakened charge repulsion, theparticles prepared at higher TPP concentration a have smallersize.At a fixed ratio of CS to TPP concentration, the particle sizeincreased along with the increase of CS and TPP concentration (c)(Fig. 2c). At a given spatial interaction distance, the particle sizeshould be roughly increased by a factor of c 1/3 . Apparently theparticle size shown in Fig. 2c does not follow this law. The reasoncan be both the change of spatial interaction distance and theparticle inner structure, e.g. the compactness of the molecules.As the association is driven by ionic interaction, and thecharging degree of both chitosan and TPP is influenced by pHvalue, a pH variation during the fabrication is expected toinfluence the formulation and structure of the chitosannanoparticles. Figure 3a shows that the nanoparticle sizeincreased rapidly until pH 3.5, and then decreased slowlyuntil pH 5.5, the measured pH value so far. A similar alterationtendency for the zeta potential was recorded with a slightlyFigure 3. Influence of chitosan solution pH value on: (a) the particle size and (b) zeta potential.616Figure 4. Influence of particle suspension pH value on: (a) the size and (b) zeta potential.www.interscience.wiley.com/journal/pat Copyright ß 2008 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2009, 20 613–619

IONICALLY CROSS-LINKED CHITOSAN NANOPARTICLESFigure 5. Influence of NaCl concentration on: (a) the particle size and (b) zeta potential.Figure 6. Alteration of average size of chitosan nanoparticles with: (a) different TPP concentration and (b) at different pH during a storage time of 385days at 48C.higher transition pH value (pH 4) (Fig. 3b). This alteration behavioris dependent on charging properties and the interplay betweenthe chitosan and the TPP molecules. At a critical low pH, most ofthe amino groups of chitosan are protonated, enabling thechitosan molecule with an extension confirmation due to strongcharge repulsion. The TPP molecule is also protonated, leading tolower charging density of the molecule. In this case, the chitosanmolecules cannot be sufficiently cross-linked by TPP to formstable particles. Along with increase of the pH value, thedeprotonation degree of TPP is increased gradually, while theprotonation degree of chitosan is less influenced when the pHvalue is below 5. At a proper pH such as pH 3.5, the chargeinteraction between these two molecules becomes strongenough, thus stable chitosan nanoparticles are obtained witha relatively larger size. At a still higher pH value, the chargingdegree of TPP molecule is enhanced again, which neutralizes thechitosan to a greater extent. Consequently, the particles shrinktheir size again.Besides the fabrication process, the pH variation also influencesthe size of the formed nanoparticles as shown in Fig. 4a. The samedecreasing tendency of the particle size was found exactly whenthe pH value was increased from 4.5 to 6.0. Meanwhile, the zetapotential was decreased too (Fig. 4b). However, the zeta potentialshowed an increasing regime when the pH value was increasedfrom 1 to 3.5, the same phenomenon as shown in Fig. 3b. Ifconsidering only the interaction of chitosan and TPP, the zetapotential should be larger or at least should remain constant atlow pH. The reason for this abnormal low value at low pH is notvery clear so far. Nevertheless, this phenomenon implies that thepositive surface charge of the chitosan particles is shielding,which can be a result of reorganization of the molecule structure,or/and adsorption of other negative charged ions at low pH. Ofcourse, ionic strength effect at a critical low pH can not beexcluded.Figure 7. Cytoviability of fibroblasts measured by MTT assay as a functionof culture time. Blank control was made by culture medium. Threeparallel samples were analysed and data were expressed as mean standard deviation.Polym. Adv. Technol. 2009, 20 613–619 Copyright ß 2008 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/pat617

H. LIU AND C. GAOFigure 8. (a) Influence of chitosan solution pH value on the drug loading. (b) Release profiles of CH from the chitosan nanoparticles in different media at378C as a function of time. This figure is available in color online at www.interscience.wiley.com/journal/pat.618To clarify this point, the chitosan particles were incubated inNaCl solution with different concentrations. Figure 5 shows thatboth the particle size and zeta potential were decreased alongwith the increase in NaCl concentration. It is known that thecharge interaction can be effectively screened by electrolytessuch as NaCl. Charge screening weakens charge repulsionbetween the chitosan molecules, leading to shrinkage of thenanoparticles. The sample explanation can be applied to the zetapotential too.Stability and biocompatibility of the nanoparticlesFor pharmaceutical applications, the storage stability andbiocompatibility of the nanoparticles is a great concern. It isknown that tiny particles are inclined to agglomerate with eachother to reduce the surface area, and hence to reduce the freesurface energy. Figure 6 presents the sizes of the nanoparticlesprepared with different TPP concentration (Fig. 6a) and atdifferent pH values (Fig. 6b) during a storage time of 385 days.Here the storage temperature was kept constant at 48C. All theparticles were rather stable with neglectable size fluctuationwithin 15 days, and then some of them showed size increment atday 85 by a factor of

IONICALLY CROSS-LINKED CHITOSAN NANOPARTICLESdescribed here is not restricted to CH, but can also be extended toother types of drugs, proteins, enzymes, and growth factors fordifferent applications.REFERENCES[1] B. C. Thanoo, M. C. Sunny, A. Jayakrishnan, J. Pharm. Pharmacol. 1992,44, 283–286.[2] I. Genta, M. Costantini, A. Asti, B. Conti, L. Montanari, Carbohydr.Polym. 1998, 36, 81–88.[3] Y. Pan, Y. J. Li, H. Y. Zhao, J. M. Zheng, H. Xu, G. Wei, J. S. Hao, F. D. Cui,Int. J. Pharm. 2002, 249, 139–147.[4] J. Berger, M. Reist, J. M. Mayer, O. Felt, N. A. Peppas, R. Gurny, Eur. J.Pharm. Biopharm. 2004, 57, 19–34.[5] J. A. Ko, H. J. Park, S. J. Hwang, J. B. Park, J. S. Lee, Int. J. Pharm. 2002,249, 165–174.[6] P. He, S. S. Davis, L. Illum, Int. J. Pharm. 1999, 187, 53–65.[7] T. López-León, E. L. S. Carvalho, B. Seijo, J. L. Ortega-Vinuesa, D.Bastos-González, J. Colloid Interface Sci. 2005, 283(2), 344–351.[8] Y. Kawashima, T. Handa, H. Takenaka, S. Y. Lin, Y. Ando, J. Pharm. Sci.1985, 74(3), 264–268.[9] Y. Kawashima, T. Handa, A. Kasai, H. Takenaka, S. Y. Lin, Chem. Pharm.Bull. 1985, 33(6), 2469–2474.[10] P. Calvo, C. Remuñán-López, J. L. Vila-Jato, M. J. Alonso, J. Appl. Polym.Sci. 1997, 63(1), 125–132.[11] S. Shiraishi, T. Imai, M. Otagiri, J. Control. Release 1993, 25, 217–225.[12] Z. Aydin, J. Akbuğa, Int. J. Pharm. 1996, 131(1), 101–103.[13] P. Calvo, C. Remuñán-López, J. L. Vila-Jato, M. J. Alonso, Pharm. Res.1997, 14(10), 1431–1436.[14] X. Z. Shu, K. J. Zhu, Int. J. Pharm. 2000, 201(1), 51–58.[15] K. A. Janes, M. P. Fresneau, A. Marazuela, A. Fabra, M. J. Alonso,J. Control. Release 2001, 73(2–3), 255–267.[16] A. Vila, A. Sánchez, M. Tobío, P. Calvo, M. J. Alonso, J. Control. Release2002, 78(1–3), 15–24.[17] I. A. Alsarra, S. H. Neau, M. A. Howard, Biomaterials 2004, 25,2645–2655.[18] Q. Gan, T. Wang, C. Cochrane, P. McCarron, Colloids Surf. B. Biointerfaces2005, 44, 65–73.[19] L. Y. Chen, M. Subirade, Biomaterials 2005, 26, 6041–6053.[20] Z. W. Mao, L. Ma, C. Y. Gao, J. C. Shen, J. Control. Release 2005, 104,193–202.[21] Y. Y. Liu, A. Yoshikoshi, B. C. Wang, A. Sakanishi, Colloid Surf. B.Biointerfaces 2003, 27, 287–293.[22] Y. Pan, Y. J. Li, H. Y. Zhao, J. S. Hao, J. M. Zheng, J.Chin. Pharmaceut. Sci.2002, 11(3), 97–100.Polym. Adv. Technol. 2009, 20 613–619 Copyright ß 2008 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/pat619