Journal of Controlled Release 51 (1998) 13–22

Methoxy poly(ethylene glycol) and e-caprolactone amphiphilicblock copolymeric micelle containing indomethacin.II. Micelle formation and drug release behaviours

a a a , b c*So Yeon Kim , IL Gyun Shin , Young Moo Lee , Chong Su Cho , Yong Kiel SungaDepartment of Industrial Chemistry, College of Engineering, Hanyang University, Seoul 133-791, Korea

bDepartment of Polymer Engineering, Chonnam National University, Kwangiu 550-757, KoreacDepartment of Chemistry, Dongguk University, Seoul 100-715, Korea

Received 10 December 1996; received in revised form 4 June 1997; accepted 11 June 1997

Abstract

Amphiphilic diblock copolymer composed of methoxy poly(ethylene glycol) (MePEG) and ´-caprolactone (´-CL) wasprepared by polymerization of ´-CL initiated with MePEG. MePEG/´-CL block copolymeric micelles containingindomethacin (IMC) were prepared by a dialysis method and evaluated as a novel drug carrier. The size of micelle formedwas less than 200 nm, and the size distribution of the micelle showed a narrow and monodisperse unimodal pattern. Also, themicelles formed by a dialysis method exhibited spherical structures. The indomethacin content in nanospheres was about42.2%, for those prepared using copolymer, having molecular weight of about 12 000 and polymer / IMC weight ratio of 1 /1.A release rate of indomethacin from nanospheres was slow, and thus the release continued over 15 days. As the molecularweights of the copolymer and the amount of drug entrapped increased, the release rate decreased. These results indicated thatthe drug-loaded nanospheres could be useful as a novel drug carrier in injectable delivery systems for hydrophobic drugs. 1998 Elsevier Science B.V.

Keywords: Methoxypolyethylene /´-caprolactone block copolymer; Nanosphere; Micelle; Controlled release system;Indomethacin

1. Introduction ment of new procedures to prepare well-character-ized particulate drug carriers [2–8]. An ideal drug

Drug delivery systems (DDS) have generated carrier should satisfy many necessary requirementsgreat interest in the scientific community and DDS such as the binding of a drug, a controlled drugusing various kinds of polymers has been extensively release, the preservation of drug activity duringstudied [1,2]. In particular, over the last two decades, delivery to a target site, and a desired interactionconsiderable progress has been made in the develop- with protein, cells, and tissues. With an aim of

achieving these requirements, one way of modifyingthe biodistribution of drugs is to entrap them inmicroscopic or sub-microscopic drug carriers [9–16].

*Corresponding author. Particulate drug delivery systems offer a number of

0168-3659/98/$19.00 1998 Elsevier Science B.V. All rights reserved.PII S0168-3659( 97 )00124-7

14 S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22

advantages over conventional dosage forms [2–8]. particulate systems have been extensively studied.Due to their small particle size, particulate systems Among these carriers, we focused on the amphiphiliclend themselves to parenteral administrations and block copolymeric micelle-type drug carrier. Advan-may be useful for the delivery to a specific organ or tages of these system are an easy control of particletarget site as sustained release injections [3]. Target- size, good structural stability, good water solubility,ing a drug to a desired site of action would not only and the separated functionality due to formation of aimprove the therapeutic efficiency but also enable a microcontainer to protect transporting substancesreduction of the amount of drug which must be from the outer environment [18–20].administered to achieve a therapeutic response. To prepare degradable polymeric nanospheresTherefore, it would be possible to minimize an displaying the desired feature of long-circulatingunwanted toxic effects. drug carriers, we prepared amphiphilic block co-

For these reasons, it may be stated that nanopar- polymeric micelles based on methoxy (polyethyleneticulates hold a promise as drug delivery systems not glycol) (MePEG) and ´-caprolactone. Poly(´-cap-only for parenteral, but also for oral, ocular, and rolactone) is well-known as a biocompatible, bio-transdermal administrations. Nanoparticulate systems degradable material and also its biodegradability cancan be classified by the nature of materials used for a be enhanced greatly by copolymerization [21]. As apreparation process [8]: hydrophillic segment, we chose polyethylene glycol

(PEG) because of its nontoxic nature and because it1. Amphiphilic macromolecules that undergo a is approved by the Food and Drug Administration

cross-linking reaction during the preparation of (FDA) for internal use in the human body [25]. Also,the nanoparticles. PEG is known to impart a protein and cellular stealth

2. Monomers that polymerize during the formation properties to surfaces and interfaces [22–24]. In ourof the nanoparticles. amphiphilic block copolymeric micelle system, hy-

3. Hydrophobic polymers, which are initially dis- drophobic cores surrounded by water-soluble polarsolved in organic solvents and then precipitated groups which extended into an aqueous medium,under controlled conditions. produced a core-shell type polymeric carrier [26–

35]. Therefore, the corona region of a micelleTo achieve a long blood circulation half-life which is composed of PEG interacts with the biologicala major issue in a particulate drug delivery system milieu [36]. Indeed, it has been reported that PEG-[4–6,17], it is necessary to consider two factors such modified proteins, liposomes, and nanoparticles [37–as particle size and surface characteristics. The size 39] effectively inhibit reticuloendothelial systemof a particle should be small enough to avoid any (RES) sequestration and prolong circulation times inmechanical clearance by filtration in the lungs or in blood [25].the spleen. Most of these systems are immediately In our previous report [35], the dialysis methodrecognized as foreign products and removed from a can be used successfully to make block copolymericblood circulation through phagocytosis by cells of micelles containing up to about 42% indomethacinthe reticuloendothelial system (RES). RES uptake (IMC). In the present study, the same procedure wasgenerally increases with particle size. Thus, a re- applied to prepare indomethacin-loaded nanoparticlesduced carrier size (possibly smaller than 200 nm) is from MePEG/´-CL block copolymer. However, wedesired for a long-circulating drug carrier. In addi- further investigated the effects of solvent on micelletion, the surface of a particle should be ‘stealthy’ formation, release behaviours of indomethacin fromrelative to RES cells. As described above, appro- nanoparticles and possible effects on release be-priate methods for preparing drug-loaded nanoparti- haviours. It is our objective to study the influence ofcles have been developed depending on the physico- drug loading content, molecular weight and hydro-chemical properties of a polymer and a drug. In phillic /hydrophobic chain length of block copolymeraddition, the procedure and the formulation con- used to make micelles on the indomethacin releaseditions which determine the inner structure of these behaviours.

S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22 15

2. Experimental thane (DCM) and then subsequently added IMC(100 mg). The organic solution was stirred at room

2.1. Materials temperature. To form IMC-loaded micelles andremove free IMC, the solution was dialysed using

Indomethacin (IMC) was obtained from Sigma (St cellulose dialysis membranes (molecular weight cutLouis, MO). Methoxy poly(ethylene glycol) off: 6000|8000, Spectrum and 12 000|14 000, size:(MePEG, M 55 000 by supplier, M 55 541 by our 21/35, Sigma, USA) with three litres of ultra-puren n

GPC measurements) was supplied by Fluka. ´-Cap- water for 24 h. The micellar solution was sonicatedrolactone was purchased from Tokyo Kasei Organic using a Branson 2210 sonicator (Branson UltrasonicsChemicals. All other chemicals used were reagent Co., USA), and then centrifuged (Jouan BP403,grade and used as purchased without further purifica- France) to eliminate unloaded IMC and aggregatedtion. In the dialysis method, ultrapure water was used particles. The supernatants, micellar solutions, ob-by purifying with a Mili-Q plus (Waters, Milipore, tained in this process, were frozen and lyophilized byUSA). freeze dryer system (Labconco, USA), to obtain

dried nanosphere products.2.2. Preparation of drug-loaded nanospheres

2.3. Determination of drug efficiencyMePEG/´-CL diblock copolymers were prepared

by polymerization of ´-caprolactone initiated with To remove an unbound IMC and aggregated IMC-MePEG, as described in our previous paper [35]. In loaded nanospheres, the micelle solution obtained byorder to obtain a satisfactory formula, as shown in a dialysis method was sonicated, centrifuged andTable 1, various copolymer compositions were test- then lyophilized. Following this, the nanospheresed. ‘‘MEP series’’ has a fixed molecular weight of which were obtained by freeze-drying the micelles,5000 in the MePEG chain, and PCL portion varies were disrupted by the addition of ethanol and THFfrom M /I535 to 150. Nanospheres of MePEG/´-CL (1:1 v/v), the amount of IMC entrapped was de-block copolymer containing IMC were prepared termined by measuring the UV absorbance at 319using the process described in our previous paper nm. An indomethacin content entrapped into the[35]. MePEG/´-CL block copolymer (100 mg) was ´-caprolactone portion of nanospheres was calcu-dissolved in 10 ml of selected organic solvent such lated from the weight of the initial drug loadedas dimethylformamide (DMF), tetrahydrofuran nanospheres and the amount of drug incorporated(THF), dimethylacetamide (DMA) and dichlorome- from the following equation.

Table 1Characterization of MePEG/´-caprolactone block copolymer and its micelle

Sample Feed Molar composition Composition Number-average Polydispersity Critical micellea b b ¯ ¯molar ratio in copolymer weight% Molecular weight (M /M ) concentrationw n

(CMC)c d

´-CL/MePEG ´-CL/MePEG MePEG:´-CL Calc. Expt’l.

MePEG 0 0 100:0 5000 5541 1.128

MEP35 35 21.8 68.9:31.1 8995 8037 1.25627MEP50 50 40.1 54.8:45.2 10 707 10 116 1.250 3.4731027MEP70 70 54.2 47.2:52.8 12 990 11 734 1.102 1.2131027MEP100 100 81.5 37.3:62.7 16 414 14 839 1.102 0.63310

MEP150 150 109.9 30.6:69.4 22 121 18 085 1.178

a ¯Determined on the basis of M of MePEG calculated in GPC experiments.nb ¯Estimated as the difference between the experimental total M of copolymer and MePEG homopolymer in GPC experiments.ncCalculated from MePEG (MW55000, Fluka).dMeasured by GPC analysis.

16 S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22

Drug loading efficiency(DLE)(%) weights, relaxation times, electron densities, and thelike).amount of indomethacin in nanospheres

]]]]]]]]]]]]] To assure the correct alignment, we measured5 amount of indomethacin 2 loaded nanospherespolystyrene standard sample, provided by Duke

IMC Scientific Co. in various conditions. We varied the]]]]]3 100 5 3 100.IMC 1 Polymer light intensity to 50, 100 and 200 mW, the pinhole

size to 100 and 200 mm, the duration time to 5, 4, 3,2 and 1 mins, and the analysis method with cumul-2.4. In vitro drug release studiesant, exponential, and CONTIN methods. From theresults of the particle size measurement of theThe in vitro IMC release profiles of IMC fromstandard sample for various conditions, we chose theMePEG/´-CL block copolymeric micelles were de-CONTIN method as a best analysis method. Intermined as follows. The appropriate amount ofCONTIN method, under the most of the variedIMC-loaded nanospheres were precisely weighed andconditions, the results were within permitted errorsuspended in 5 ml of a phosphate buffer solutionrange given by manufacturer. From these measure-(PBS, 0.1 M, pH 7.4). The micellar solution wasments we obtained the diameter and variance of theintroduced into a dialysis membrane bag and the bagmicelles.was placed in 250 ml of phosphate buffer solution

release media, and the media were stirred at 378C. Atpredetermined time intervals, 3 ml aliquots of the 2.6. Determination of critical micelle concentrationaqueous solution were withdrawn from the releasemedia. After the concentration of IMC released was In order to determine the critical micelle con-monitored using a UV spectrophotometer at 319 nm, centration (CMC) of MePEG/´-CL block copoly-the solution taken as a sample was replaced into the meric micelle, fluorescence measurements were car-release media. ried out by fluorescence spectrophotometer (JASCO

Model FP-777, Japan Spectroscopic Co. Ltd.) usingpyrene as a fluorescent probe. We obtained the

2.5. Measurement of size and size distributionfluorescence excitation spectra of pyrene at variousconcentrations of MePEG/´-CL block copolymers.

An average size and the size distribution ofExcitation wavelength was 390 nm and pyrene

nanospheres were estimated by a dynamic light 27concentration was kept constant at 6.0310 M.scattering (DLS) using a Model 95 ION Lager(Lexel Laser Inc., USA) at a wavelength of 514 nmat 208C. The intensity of a scattered light was 2.7. Characterizationdetected at 908 to an incident beam. Measurementswere made after the aqueous micellar solution was The molecular weight distribution of block co-filtered with a microfilter having an average pore size polymers was determined by gel permeation chroma-of 0.8 (Milipore, USA). The average size distribution tography apparatus (Waters Model 510 HPLC pump,of aqueous micellar solutions was determined based Milford, USA). In addition, the composition and theon CONTIN programs of Provencher et al., [40]. number-average molecular weight of each copolymer

1CONTIN is widely used as a data analysis technique in CDCl solution were determined by 500 MHz H3

in many area of dynamic light scattering experiments NMR (Bruker AMX-500).as a FORTRAN package for inverting noisy linear The nanospheres were coated with gold under aoperator equations. This method is completely auto- vacuum of 0.1 Torr and a voltage of 1.2 kV and 10matic in that no initial estimate or prior information mA. The nanospheres obtained by this procedureabout the distribution is needed. It is called CONTIN were examined with a scanning electron microscopebecause it is often applied to solve integral equations (Jeol Model JSM-35CF). Indomethacin was assayedof the first kind (for effectively CONTINuous dis- by UV-visible spectrophotometer (Shimadzu Modeltributions of diffusion coefficients, molecular UV-2101 PC).

S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22 17

3. Results and discussion others, resemble, in many respect, that of the lowmolecular mass amphiphiles. [26–33] However,

3.1. Characterization of nanospheres critical micelle phenomena in block copolymer sys-tems occur at very much lower concentrations than

In our previous paper, [35] we reported that the in low molecular mass amphiphiles. [26] At con-MePEG/´-CL diblock copolymer could be prepared centrations below the CMC, all the block copolymersby ring opening polymerization mechanism of ´-CL are in a single chain form. At the CMC, thein the presence of MePEG, containing hydroxyl collapsed blocks begin to associate to form loosefunctional group at one end of the chain, without any aggregates. At this stage the insoluble blocks remaincatalysts. in their individual collapsed states in order to

To achieve long blood circulation half-lives, par- maintain an equilibrium with single chains. Micelleticles should be small enough to avoid a mechanical formation requires the presence of two opposingclearance by filtration in lungs, with particles larger forces. One is an attractive force between am-than 5–7 mm, or in the spleen. The RES uptake phiphiles leading to aggregation, the other is agenerally increases with increasing a particle size. repulsive force that prevents an unlimited growth ofThus, a reduced carrier size smaller than 200 nm is micelles into a distinct macroscopic phase. [26]desired for a long-circulating drug carrier. [4–6,17] As the concentration of block copolymer mole-From the dynamic light scattering measurements, the cules increases, an equilibrium of state shifts to aaverage diameter of nanospheres was smaller than micellar form. Insoluble blocks in micellar cores200 nm and size distribution showed a narrow and rearrange to find their low-energy conformation, andmonodisperse unimodal pattern as shown in Fig. 1. the solvent molecules are gradually driven out of theThese results indicated that all the block copolymer micellar cores. At higher polymer concentrations,chains existed as uniform micelles without a large large collapsed micellar cores consisting of manyaggregation. Therefore, these micelles could be insoluble blocks are present, surrounded by diffuseuseful as a long-circulating drug carrier in injectable outer shells (coronas) formed from the solubledelivery systems. blocks.

In our MePEG/´-CL diblock copolymer system,3.2. Micelle formation mechanism of amphiphilic we examined the CMC to get information on theblock copolymer influence of copolymer component ratio. Especially,

the relationship of hydrophobic chain length and theThe micellization process of block copolymers in onset of micellization were investigated. As de-

selective solvents, which is thermodynamically scribed in Table 1, a copolymer with identicalfavourable for one block, but unfavourable for the hydrophillic MePEG block length and varying hy-

drophobic ´-CL block length were examined. Figureseven of our previous study [35] plotted the intensityratio of I /I from pyrene excitation spectra vs.337.5 334

log C for various MePEG/´-CL block copolymers.Below the CMC, there were no micelles present inthe system. The pyrene fluorescence spectrum had alow intensity and showed a mild slope of I /I .337.5 334

At CMC, however, the curve showed a sharp in-crease in the I /I ratio. This fact suggests that337.5 334

we are able to determine the experimental CMCvalue in the fluorescence spectra. We measured CMCvalues from the intersection of straight line seg-ments, drawn through the points at the lowestFig. 1. Typical size distribution measured by dynamic lightpolymer concentrations, which lie on a nearlyscattering (MEP70): (a) before loading IMC, (b) after loading

IMC horizontal line, with that going through the points on

18 S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22

the rapidly rising part of the plot. These CMC values micelles was relatively identical to that before drug27 27were calculated to be 3.47310 , 1.21310 and loading, maintaining the narrow and monodisperse

270.63310 mole /, for MEP50, MEP70 and unimodal distribution.MEP100, respectively. Namely, as the hydrophobiccomponents in a copolymer increased, critical mi- 3.4. Solvent effect on micelle formationcelle concentration values deceased.

These results are in a good agreement with those We investigated the relationship between the sizeby other researches [26,27]. The onset of micelliza- of micelle and the selected initial solvent used totion is determined mainly by the nature and the prepare micelle by dialysis in water. Despite oflength of hydrophobic block, where the nature of extensive reported works, there are only a fewsoluble hydrophillic blocks has only a slight depen- systematic comparisons between the experimentaldence on the onset of micellization. The effect of data and theoretical calculations. In our study, thepolystyrene block lengths on the thermodynamic copolymer was dissolved in organic solvent andstability of polystyrene–polyisoprene in n-hexane subsequent micellar solution was readily formed bywas also studied by Z. Gao et al., [27].Values of DG8 dialysis in a water medium.and DH8 were found to be strongly dependent on the These copolymers are not soluble in pure water.molecular mass of polystyrene block, both becoming But their aqueous micellar solutions can be preparedmore negative as the molecular mass increased. using a dialysis technique. Table 2 shows the dy-Many other investigations indicate that the copoly- namic light scattering results of micelles prepared bymer structure and the solvent composition and dialysis against water using different solvents suchtemperature markedly influence the free chain-mi- as, dimethylformamide, tetrahydrofuran, dimethyl-celle equilibrium, the micelle structure, and the acetamide and dichloromethane.dynamic of the unimer–micelle exchange. From these results, the solvent used to dissolve the

copolymer significantly affects the size and the size3.3. Factors influencing size and size distribution distribution of micelles. To form micelle by dialysis

against water, a solvent should be miscible withIn our previous study [35], we already described water. Thus, it could be expected that miscibilities of

that the size of micelle depended on the loading a polymer and a solvent, or water and a solventamount of drug in micelles and the molecular weight would affect micelle formation.of copolymer. As the molecular weight of copolymerincreases, the size of micelle formed increases. 3.5. Microscopic characterization of nanospheresParticularly, they are 5460.082, 7760.010,11460.089, 13060.105 (mean6S.D. (nm)), for Fig. 2 (a) and (b) show scanning electron mi-MEP35, MEP50, MEP70 and MEP100, respectively. crographs (SEM) of MePEG/´-CL block copoly-Also, the size of micelles increases with the loading meric nanoparticle (MEP70). This figure is a resultamounts of drug. When we used MEP70 as a of SEM measurements of the sample MEP70. As cancopolymer and THF as a solvent, the size of micelle be seen, it was confirmed that nanoparticles ofwithout loaded drug was about 120 nm. However, invariably nonporous spherical shape with a smooththey became 145 and 165 nm, for feed weight ratioof polymer to IMC being 1.0: 0.5 and 1.0: 1.0, Table 2

Effect of the solvents on the size and size distribution of micellesrespectively. From these results, we found out thata bthe molecular weight of copolymer and loading Sample Solvent Size6S.D. (nm)

amount of drug affected the micelle size.1 Dimethylformamide 11460.089

Fig. 1 shows the typical size distribution of 2 Dimethylacetamide 11660.351micelles before and after drug loading. As mentioned 3 Tetrahydrofuran 12060.006

4 Dichloromethan 18160.674above, after the IMC loading in micelle, the size ofamicelles somewhat increases according to loading MEP 70.bamount of drug. However, the size distribution of Dynamic light scattering measurement.

S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22 19

Fig. 2. Surface morphology and characteristics by scanning electron microphotographs: (a) nanoparticles prepared with MEP70 (310 000),(b) nanoparticles prepared with MEP70 (330 000)

surface had been prepared. Also, we observed sub- various polymer /drug weight ratio in the preparativemicron size particles which confirmed similar results process. That is, a plot of relative release percentagesfrom DLS measurement. of indomethacin based on loading amount versus

time is shown in Fig. 4. In contrast to the rapid3.6. Influence of drug loading content on release release of the free indomethacin from cellulosebehaviours membrane, the release of a drug from nanospheres

through the membrane was slow and showed sus-The amount of indomethacin introduced into the tained release characteristics. As loading amount of

micelle by controlling the weight ratio between drug in nanospheres increases, the release rate de-polymer and drug is shown in Table 3. Also, the

Table 3drug loading efficiencies depending on molecularIndomethacin loading contents of MePEG/´-caprolactone blockweights of block polymer were described. Thecopolymer micellesdesignation ‘‘DIP series’’ was used for drug-loadedNo. Sample Feed weight ratio Drug loading contentsamples varying the feed ratio of IMC to polymer

b(DLE) (%)(Drug IMC Polymer), for example, in the case of aIMC : polymerMEP25, feed ratio of IMC to polymer was 0.25: 1.00

1 DIP 25 0.25: 1.0 16.33in loading process). The DMEP was used for drug-2 DIP 50 0.50: 1.0 20.99loaded sample prepared by fixed feed ratio of IMC to3 DIP 75 0.75: 1.0 31.96

polymer and varied molecular weight of copolymer. 4 DIP 100 1.00: 1.0 41.98For example, DMEP35 was drug-loaded sample 5 DMEP 35 1.00: 1.0 25.83

6 DMEP 50 1.00: 1.0 34.08prepared by using the MEP35 copolymer.7 DMEP 70 1.00: 1.0 41.98As shown in Table 3, the drug loading efficiency8 DMEP100 1.00: 1.0 42.03increased with the ratio of drug to polymer, molecu-aIndomethacin.lar weights and hydrophobic chain lengths of a block amount of indomethacin in nanospheresb ]]]]]]]]]]]DLE(%) 5copolymer. amount of indomethacin loaded in nanospheres

Fig. 3 shows a release profile of indomethacinIMC

]]]]from MePEG/´-CL nanospheres prepared by using 3 100 5 3 100.IMC 1 Polymer

20 S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22

indomethacin was examined. The results were dem-onstrated by plotting the relative release percentagesof drug versus time in Fig. 4.

The release rate of indomethacin from nanos-pheres seems to decrease with an increase in molecu-lar weight of MePEG/´-CL block copolymer. Forexample, if we compared DMEP 70 and DMEP 100which have similar DLE(%) as 41.98 and 42.02%and different molecular weight, the release rate ofindomethacin from MEP70, nanospheres made ofcopolymer with a molecular weight of 12 990 (calc.),was faster than that from MEP100, nanospheresmade of copolymer with a molecular weight 16 414(calc.). Also, in the case of DIP 75 and DMEP50,DMEP50 had even greater DLE(%) value (34.08%)Fig. 3. In vitro release profiles of indomethacin from micelles withthan DIP 75 (31.96%) but the release rate fromvarious drug loading content in pH 7.4 PBS buffer solution atDMEP 50 (calc. M 510 707) was faster than DIP75378C: percentage of release amoung to total loading amount (%) n

vs. time (calc. M 512 990). These results was clearly due ton

the influence of the molecular weight of the polymer.creases. From these results, the level of drug release In the release patterns of all the samples, we couldseems to depend on the amount of drug entrapped not observe an initial burst release effect. However,inside of micelle. An increase of indomethacin with at a later stage, the drug release rate was reduced.hydrophobic property in nanospheres enhanced an Furthermore, it showed a significant sustained re-interaction between indomethacin and ´-CL as a lease characteristic. Even in the MEP35, nanosphereshydrophobic part, leading to a decreased drug re- showing the fastest indomethacin release rate, morelease. than 50% (w/w) of indomethacin remained unre-

leased over seven days (not shown in plotted data).3.7. Influence of molecular weights on release From these results, we can consider that the drugbehaviours binding affinity and polymer degradation are two

related phenomena. It is generally assumed that aThe effect of molecular weights of a copolymer drug is released by several processes [25]: (a) a

used to prepare nanospheres on release profiles of diffusion through the polymer matrix, (b) a releaseby polymer degradation (either by surface or bulkerosion), and (c) a solubilisation and a diffusionthrough microchannels that exist in the polymermatrix or are formed by an erosion.

Indomethacin, owing to its moderate lipophiliccharacter, is physically entrapped in hydrophobiccore of a micelle. Accordingly, the in vitro releasebehaviours of a lipophilic compound from thesepolymeric micellar systems are largely affected by itsinner core with hydrophobic properties. As shown inTable 1, a copolymer used to prepare nanosphereswas synthesized by ring opening polymerization of´-caprolactone with MePEG prepolymer havingfixed molecular weight of 5000. Thus, as the molecu-Fig. 4. In vitro release profiles of indomethacin from micelles withlar weight of a copolymer increases, the hydrophobicvarious molecular weights in pH 7.4 PBS buffer solution at 378C:

percentage of release amount to total loading amount (%) vs. time segments in a block copolymer increases, resulting in

S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22 21

the binding affinity between indomethacin and ´- Engineering Foundation Grant [ 95-0300-16-3.caprolactone increased. SYK and IGS are grateful to the Graduate School of

The other factor affecting release behaviours is Advanced Materials and Chemical Engineering atpolymer degradation. In general, degradation rate of Hanyang University for a fellowship.a polymer increases with increasing the molecularweight of polymer. However, it might not be largelyaffected because poly(´-caprolactone) degrades quite

Referencesslowly. It is concluded that the effect of bindingaffinity is greater than that of polymer breakdown on

[1] M.A. El-Nokaly, D.M. Piatt, B.A. Charpentier, Polymerica release profile in this system. Consequently, in thisdelivery systems, American chemical society, Washington,

micelle system with amphiphilic block copolymer, DC, 1993.the release rate of a drug is inversely proportional to [2] J. Kreuter, Nanoparticle-based drug delivery systems, J.

Control. Release 16 (1991) 169–176.the molecular weight of a block copolymer.[3] B. Seijo, E. Fattal, L.R. Treupel, P. Couvreur, Design of

nanoparticles of less than 50 nm diameter: preparation,characterization and loading, Int. J. Pharm. 62 (1990) 1–7.

4. Conclusions [4] N. Kossovsky, D. Millet, A. Gelman, E. Sponsler, H.J.Hnatyszyn, Self-assembling nanostructures, Biol. Technolgy11 (1993) 1534–1536.We successfully prepared drug-loaded block co-

[5] G.M. Whitesides, J.P. Mathias, C.T. Seto, Molecular self-polymeric nanospheres by dialysis method. The sizeassembly and nanochemistry: a chemical strategy for theof the nanospheres is less than 200 nm, which issynthesis of nanostructures, Science 254 (1991) 1312–1319.

similar to that of viruses. Therefore, a micelle may [6] E. Allemann, R. Gurny, E. Doelker, Drug-laoded nanoparti-also penetrate through the sinusoidal and fenestrated cles preparation methods and drug targeting issues, Eur. J.

Pharm. Biopharm. 39 (1993) 173–191.capillaries without any uptake by reticuloendothelial[7] L.B. Peppas, Recent advances on the use of biodegradablesystem. Preliminary investigations of the structural

microparticles and nanoparticles in controlled drug delivery,characteristics of the nanoparticles based onInt. J. Pharm. 116 (1995) 1–9.

MePEG/´-CL block copolymer have shown that [8] M.J. Alonso, Drug and the pharmaceutical sciences vol. 77:indomethacin is entrapped in nanoparticles with a Microparticulate systems for the delivery of proteins andrelatively high loading content (up to about 42%). In vaccines, Marcel Dekker, Inc., USA, 1996, pp. 203–242.

[9] S.S. Guterres, H. Fessi, G. Barratt, J.P. Devissaguet, F.the release pattern of all the samples, we could notPuisieux, Poly(DL-lactide) nanocapsules containing di-observe any initial burst release. It rather showed aclofenac: I. formulation and stability study, Int. J. Pharm.

significant sustained release characteristics. An in- 113 (1995) 57–63.crease of the content of indomethacin having hydro- [10] T. Uchida, K. Yoshida, A. Ninomiya, S. Goto, Optimizationphobic property in nanospheres enhanced the inter- of preparative conditions for polylactide (PLA) microspheres

containing ovalbumin, Chem. Pharm. Bull. 43 (1995) 1569–action of indomethacin and ´-caprolactone as a1573.hydrophobic part, leading to a decreased drug re-

[11] G. Spenlehauer, M. Vert, J.P. Benoit, F. Chabot, M. Veillard,lease. Also, the release rate of indomethacin de- Biodegradable cisplatin microspheres prepared by the solventcreases with an increase in molecular weight of evaporation method: molphology and release characteristics,MePEG/´-CL block copolymer. These results sug- J. Control. Release 7 (1988) 217–229.

[12] S. Miyazaki, N. Hashiguchi, W.M. Hou, C. Yokouchi, M.gest that this nanosphere system has a considerableTakada, Preparation and evaluation in vitro and in vivo ofpotential for sustained release of hydrophobic drugfibrinogen microspheres containing adriamycin, Chem.

with a minimal side effects and providing a conveni- Pharm. Bull. 34 (1986) 3384–3393.ent method of drug delivery while minimizing drug [13] Y. Ohya, T. Takei, H. Fukushima, T. Ouchi, Preparation oftoxicity and maximizing drug effectiveness. albumin microspheres grafted galactose residues through

polyethylene-glycol spacers, release behavior of 5-fluoro-uracil from them, and their lectin-mediated aggregation, J.Macromol. Sci. 8 (1991) 743–760.

Acknowledgements [14] C. Verdun, P. Couvreur, H. Vranckx, V. Lenaerts, M. Roland,Development of a nanoparticle controlled-release formation

This research was supported by Korea Science and for human use, J. Control. Release 3 (1986) 205–210.

22 S.-Y. Kim et al. / Journal of Controlled Release 51 (1998) 13 –22

[15] G. Cavallaro, M. Fresta, G. Giammona, G. Puglisi, A.Villari, [28] J.R. Quintana, M.D. Janez, I. Katime, Formation of blockEntrapment of b- lactams antibiotics in poly- copolymer micelles in solution of linear homopolymer in aethylcyanoacrylate nanoparticles: studies on the possible in good solvent, Macromol. Rapid Commun. 16 (1995) 607–vivo application of this colloidal delivery system, Int. J. 612.Pharm. 111 (1994) 31–41. [29] J.R. Quintana, M. Villacampa, M. Munoz, A. Andrio, A.

[16] R. Jalil, J.R. Nixon, Biodegradable poly(lactic acid) and Katime, Micellization of a polystyrene-block-poly(ethylene /poly(lactide-co-glycolide) microcapsules: problems associ- propylene) copolymer in n-alkanes. 1: Thermodynamicated with preparative techniques and release properties, J. study. Macromolecules 25 (1992) 3125–3128.Microencapsulation 7 (1990) 297–325. [30] P. Munk, A. Qin, M. Tian, C. Ramireddy, S.E. Webber, Z.

[17] M. Yokoyama, T. Okano, Y. Sakurai, H. Ekimoto, C. Tuzar, K. Prochazka, Characterization of block copolymerShibazaki, K. Kataoka, Toxicity and antitumor activity micelles in aqueous media, J. Appl. Polym. Sci. Appl.against solid tumors of micelle-forming polymeric anticancer Polym. Symposium. 52 (1993) 45–54.drug and its extremely long circulation in blood, Cancer [31] E. Hamad, S. Qutubuddin, Theory of micelle formation byResearch 51 (1991) 3229–3236. amphiphilic side-chain polymers, Macromolecules 23 (1990)

[18] K. Kataoka, G.S. Kwon, M. Yokoyama, T. Okano, Y. 4185–4191.Sakurai, Block copolymer micelles as vehicles for drug [32] M. Malmsten, B. Lindman, Self-assembly in aqueous blockdelivery, J. Control. Release 24 (1993) 119–132. copolymer solutions, Macromolecules 25 (1992) 5440–5445.

[19] C.S. Cho, J.W. Na, Y.I. Jeong, S.H. Kim, Y.M. Lee, Y.K. [33] C. Price, R.B. Stubbersfield, S.E. Kafrawy, K.D. Kendall,Sung, Micelle formation of the hexablock copolymer consist- Thermodynamics of micellization of polystyrene-block-poly-ing of poly(g-benzyl L-glutamate) as the hydrophobic part (ethylene /propylene) copolymers in decane, Br. Polym. J. 21and poly(ethylene oxide) as the hydrophilic part, Polymer (1989) 391–394.(Korea) 19 (1995) 926–931. [34] J.L. Jones, C.M. Marques, J.F. Joanny, Shear-induced micel-

[20] Z. Gao, S.K. Varshney, S. Wong, A. Eisenberg, Block lization of diblock copolymers, Macromolecules 28 (1995)copolymer ‘‘crew-cut’’ micelles in water, Macromolecules 136–142.27 (1994) 7923–7927. [35] I.G. Shin, S.Y. Kim, Y.M. Lee, C.S. Cho, Y.K. Sung,

[21] H. Huatan, J.H. Collett, D. Attwood, The microencapsulation Methoxy Poly (ethylene Glycol) /´-Caprolactone Am-of protein using a novel ternary blend based on poly (´- phiphilic Block Copolymeric Micelle Containing In-caprolactone), J. Microencapsulation 12 (1995) 557–567. domethacin: I. Preparation and Characterization, J. Control.

[22] D. Bazile, C. Prud’homme, M.T. Bassoullet, M. Marlard, G. Release submitted.Spenlehauer, M. Villard, Stealth Me.PEG-PLA nanoparticles [36] P. Cerri, M. Tricoli, L. Lelli, G.D. Guerra, R.S. Guerra, M.G.avoid uptake by the mononuclear phagocytes system, J. Cascone, P. Giusti, Block copolymers of L-lactide andPharm. Sci. 84 (1995) 493–498. poly(ethylene glycol) for biomedical applications, J. Mater.

[23] M. Yokoyama, M. Miyauchi, N. Yamada, T. Okano, Y. Sci.: Materials in Medicine 5 (1994) 308–313.Sakurai, K. Kataoka, S. Inoue, Polymer micelles as novel [37] F. Puisieux, G. Barratt, G. Couarraze, P. Couvreur, J.P.drug carrier: adriamycin-conjugated poly(ethylene glycol)- Devissaguet, C. Dubernet, E. Fattal, H. Fessi, C. Vauthier,poly(aspartic acid) block copolymer, J. Control. Release 11 Polymeric Biomaterials, Marcel Dekker, Inc., pp. 749–794.(1990) 269–278. [38] T. Niwa, H. Takeuchi, T. Hino, N. Kunou, Y. Kawashima, In

[24] G.S. Kwon, M. Yokoyama, T. Okano, Y. Sakurai, K. Kata- vitro release behavior of D,L-lactide /glycolide copolymeroka, Biodistribution of micelle-forming polymer-drug conju- (PLGA) nanospheres with nafarelin acetate prepared by agates, Pharm. Res. 10 (1993) 970–973. novel spontaneous emulsification solvent diffusion method,

[25] R. Gref, Y. Minamitake, M.T. Peracchia, R. Langer, Mi- J. Pharm. Sci. 36 (1988) 1095–1103.croparticulate systems for the delivery of protein and vac- [39] S.B. La, T. Okano, K. Kataoka, Preparation and characteriza-cines, Marcel Dekker, Inc., USA, 1996, pp. 279–306. tion of the micelle-forming polymeric drug indomethacin

[26] I. Astafieva, X.F. Zhong, A. Eisenberg, Critical micellization -incorporated poly (ethylene oxide)-poly(b-benzyl L-aspara-phennmena in block polyelectrolyte solutions, Macromole- tate) block copolymer micelles, J. Pharm. Sci. 85 (1996)cules 26 (1993) 7339–7352. 85–90.

[27] Z. Gao, A. Eisenberg, A model of micellization for block [40] S.W. Provencher, J. Hendrixs, Direct determination of molec-copolymers in solutions, Macromolecules 26 (1993) 7353– ular weight distributions of polystyrene in cyclohexane with7360. photon correlation, J. Chem. Phys. 69 (1978) 4273–4276.