Nano-engineering block copolymer aggregates for drug delivery

Colloids and Surfaces B: Biointerfaces 16 (1999) 3–27

Nano-engineering block copolymer aggregates for drugdelivery

Christine Allen a, Dusica Maysinger b, Adi Eisenberg a,*a Department of Chemistry, McGill Uni6ersity, 801 Sherbrooke Street West, Montreal PQ, Canada H3A 2K6

b Department of Pharmacology and Therapeutics, McGill Uni6ersity, 3655 Drummond St., Montreal PQ, Canada H3G 1Y6

Abstract

This review describes the properties of block copolymer micelles which influence their efficiency as drug deliveryvehicles for hydrophobic drugs. The key performance related properties we discuss are loading capacity, releasekinetics, circulation time, biodistribution, size, size distribution and stability. Each of the properties is discussed indetail with specific attention given to the way in which they may be changed or controlled, the aim being to allowthe reader to tailor-make block copolymer micelles for a particular application. In addition, the last section of thereview focuses on the morphology of the micelles as another performance related property which, to this point,remains unexplored in this connection. © 1999 Elsevier Science B.V. All rights reserved.

Keywords: Block copolymer micelles; Hydrophobic drugs; Morphology of micelles

www.elsevier.nl/locate/colsurfb

1. Introduction

Micelles formed from amphiphilic di or triblock copolymers have been explored in recentyears as carriers for hydrophobic drugs [1–32]. In

an aqueous environment, the hydrophobic blocksof the copolymer form the core of the micellewhile the hydrophilic blocks form the corona orouter shell. The hydrophobic micelle core servesas a microenvironment for the incorporation oflipophilic drugs, while the corona shell serves as astabilizing interface between the hydrophobic coreand the external medium.

To date, the greatest contributions in this areahave been made by the groups of Kabanov [1–8]and Kataoka [9–21]. Kabanov’s work initiallyfocused mostly on micelles formed fromPluronic™ triblock copolymers as delivery vehi-cles for drug targeting across the blood brainbarrier [1,2]. The Kabanov group was the first todemonstrate that Pluronic™ unimers inhibit theP-gp-mediated drug efflux system in both bovine

Abbre6iations: AD, Radriamycin; CMC, critical micelle con-centration; DMA, dimethylacetamide; DMF, dimethylfor-mamide; DMSO, dimethylsulfoxide; DNA, deoxyribonucleicacid; IMC, indomethacin; Nagg, aggregation number; NGF,nerve growth factor; oMMA, oligo(methyl methacrylate);PAA, poly(acrylic acid); P(Asp), poly(aspartic acid); PBLA,poly(b-benzyl-L-aspartate); PCL, poly(o-caprolactone); PEO,poly(ethylene oxide); PLA, polylactic acid; PPO, poly(propy-lene oxide); RES, reticuloendothelial system; Tg, glass transi-tion temperature; THF, tetrahydrofuran.

* Corresponding author. Tel.: +1-514-398-6934; fax: +1-514-398-3797.

E-mail address: [email protected] (A. Eisenberg)

0927-7765/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.

PII: S0927 -7765 (99 )00058 -2

C. Allen et al. / Colloids and Surfaces B: Biointerfaces 16 (1999) 3–274

brain microvessel endothelial cells and Caco-2monolayers [3–5]. In recent years, his researchhas also expanded to include block ionomer com-plexes as carriers for DNA [6,7]. The work ofKataoka’s group has largely focused on micellesformed from copolymers containing a poly(aminoacid) core-forming block, as delivery vehicles foranti-cancer drugs [9–13]. However, his researchhas also centered on the use of ‘polyion complexmicelles’ as carriers for charged molecules such asDNA or enzymes [14,15]. Recently, the groups ofKataoka and Okano have developed thermo-re-sponsive micelles from poly(N-isopropylacry-lamide)-b-poly(styrene) copolymers [16].

Several other groups have also done research inthis area [23–32]. Such a high level of activity hasbrought a great deal of diversity to this field, sincemost groups have introduced their own micellesystem formed from copolymers which contain aunique hydrophilic–hydrophobic block combina-tion. In almost all cases, the hydrophilic block hasbeen poly(ethylene oxide) for reasons which willbe discussed in Section 3.0. By contrast to theuniversal use of poly(ethylene oxide) as the hy-drophilic block, a much wider range of hydropho-bic blocks have been explored (Table 1). Thus theuniqueness associated with the different copoly-mer systems largely originates from the choice ofthe hydrophobic block. The various PEO-hydro-phobic block combinations have given rise to anumber of micelle systems which have distinctphysico–chemical properties and different charac-teristics important to their suitability as drug car-

riers. Also, micelles containing a specifichydrophobic block have been found to be moreeffective as carriers of some specific drug ratherthan other drugs. Unfortunately, the synthesisand manipulation of many of the biocompatibleblock copolymers is difficult. For this reason, fewsystematic studies have been performed wherebyone parameter (e.g. core block length) is variedand the effect of this variation on the micelle’scharacteristics such as size, stability, loading ca-pacity and release kinetics, is measured. Also, dueto the fact that there are such a large number ofvariables which influence micelle properties, itmay be unrealistic for groups in the area of drugdelivery to explore all of the relevant parametersin a systematic way. However, the PPO–PEObased block copolymers are available commer-cially, and for this reason many studies have beendone on this system.

In studies completely unrelated to drug deliv-ery, a number of groups have investigated manyof the physico–chemical parameters of block co-polymers micelles in a very systematic way [33–56]). Several reviews have resulted from this work[57–61]. These studies involve micelles which aremost often formed from non-biocompatible co-polymers, yet the information obtained may beused to gain insight into the biocompatiblesystems.

The aim of this review is 2-fold; first we aim tooutline many of the properties of micelles whichdetermine their ability to deliver various drugs,and second, we address the factors which influ-ence or control each of these individual proper-ties. Several of the key performance relatedproperties are loading capacity, release kinetics,circulation time, biodistribution, size, size popula-tion distribution and stability, all of which arediscussed in this review. In addition, the lastsection is dedicated to the discussion of the mor-phology of block copolymer aggregates as an-other potential performance related parameter.The morphology of the block copolymer aggre-gates, in so far as it affects drug delivery, has notbeen addressed to date, primarily because it wasnot possible to prepare reproducibly a wide rangeof morphologies. This has now become possible,and Section 5.0 discusses the preparation and

Table 1A list of several of the biocompatible polymers that have beenemployed as core-forming blocks

Biocompatible core-forming polymers References

Poly(aspartic acid) [9,15]Poly(b-benzyl-L-aspartate) [13,17]Polycaprolactone [25,32]Poly(gamma-benzyl-L-glutamte) [27]

[24,30]Poly(D,L-lactide)[14]Poly(L-lysine)

Poly(propylene oxide) [1,4]Polyspermine [7]Oligo(methyl methacrylate) [26]

C. Allen et al. / Colloids and Surfaces B: Biointerfaces 16 (1999) 3–27 5

potential relevance of block copolymer aggregatesof different morphologies in drug delivery [61].

The method and extent to which each of themicelle parameters can be manipulated has beendescribed most systematically and thoroughly inliterature which is unrelated to drug delivery. Forthis reason, this review draws from both the liter-ature on block copolymer micelles which is relatedto drug delivery and that which is more generaland unrelated to that topic. In this way, we hopeto bring together information which will allow forblock copolymer micelles to be tailor-made ascarriers for the effective delivery of specific drugs.

The review is divided into five major sections,the first of which is the introduction. The secondsection covers performance related properties ofthe micelle as a whole, including methods ofpreparation, stability, as well as size and sizepopulation distributions of the micelles. The thirdand fourth sections deal with the two principalworking parts of the micelle, the corona whichacts as the stabilizing interface between the coreand the external medium and the micelle corewhich act as the cargo space for lipophilic drugs.More specifically, in section three, aspects relatedto the micelle corona are discussed; these includethe use of poly(ethylene oxide) as the coronaforming block, the use of hydrophilic polymerblocks other than PEO and the concept of stericstabilization. In Section 4, the two core-relatedparameters, loading capacity and release kinetics,are discussed in detail with special emphasisplaced on factors which may be used to alter bothof these parameters. In Section 5.0 we discuss therelevance of block copolymer aggregates of differ-ent morphologies in drug delivery and the variousmethods by which aggregates with these mor-phologies can be prepared.

2. The micelle as a whole

2.1. Methods of micelle preparation

There are two principal methods for the prepa-ration of block copolymer micelles, the directdissolution method and the dialysis method, asoutlined in Fig. 1. The choice of which method to

use depends mostly on the solubility of the blockcopolymer in water. To this point, mostly star-type micelles have been investigated as drug carri-ers. Star-type micelles are formed from blockcopolymers which have corona-forming blocksthat are longer than the core-forming blocks. Ifthe copolymer is marginally soluble in water, thedirect dissolution method is employed, whereas ifthe copolymer is poorly soluble in water, thedialysis method is usually employed.

The direct dissolution simply involves addingthe copolymer to water or another aqueousmedium such as phosphate buffer saline. Themicelles formed from the PEO-b-PPO-b-PEO co-polymers are routinely formed by direct dissolu-tion, but in some cases the copolymer and waterare mixed at elevated temperatures to ensure mi-cellization [8].

The dialysis method is often used when micellesare to be formed from a copolymer that is noteasily soluble in water [17,25]. In this case, thecopolymer is first dissolved in a common organicsolvent that is miscible with water such asdimethylformamide, tetrahydrofuran, or dimethy-lacetamide. The copolymer solvent mixture isstirred and then dialyzed against bidistilled water.During the process of dialysis micelle formation isinduced and the organic solvent is removed.

The size and size population distribution ofmicelles produced using the dialysis method mayvary depending on the organic solvent employed[17,25]. In addition, the weight fraction or yield ofmicelles obtained was also found to vary with thechoice of organic solvent. For example, in a studyby La et al. [17], the use of DMSO as the organicsolvent gave rise to PEO-b-PBLA micelles whichwere only 17 nm in size; however, only 6% of thecopolymer formed micelles. Yet, when DMAc wasused, the micelles were obtained in high yield,with an average particle size of 19 nm and anarrow size distribution (dw/dn=1.27) [17]. In thisway, the dialysis method provides a means oftailoring the size and size population distributionof the micelles.

Recently, our group has been working on crew-cut micelle systems formed from a variety ofcopolymers such as PS-b-PAA, PS-b-PEO andPCL-b-PEO [32,62,63]. Crew-cut aggregates are


Fig. 1. A schematic of the two principal methods employed for the preparation of block copolymer micelles.

formed from copolymers which have core-formingblocks that are longer than the corona-formingblocks. These copolymers are thus insoluble inwater and therefore must first be dissolved in acommon organic solvent. For this reason, themethod of preparation employed involves the ini-tial dissolution of the copolymer in a commonorganic solvent followed by the slow addition ofwater at a very slow rate. Self-assembly occurs atsome critical water content which depends on thephysical properties of the block copolymer, pri-marily the length of the hydrophobic block andthe copolymer concentration. The copolymer inthe organic/water solvent mixture is then dialyzedagainst bidistilled water. Our studies have foundthat the size, size distribution and morphology ofthe micelles can depend on both the commonorganic solvent employed and the rate of wateraddition to the copolymer solvent mixture [64].Once again this demonstrates the many parame-ters of the micelles (size, size population distribu-tion and morphology) that can be manipulated by

simple variations within the method ofpreparation.

2.1.1. Methods of drug incorporationThe method of drug incorporation employed

will depend mostly on the method of micellepreparation used for the particular block copoly-mer in question. If the micelles are formed bydirect dissolution in water, than an aliquot of acopolymer water stock solution is often added toa vial which contains the drug to be incorporated.For example, a drug stock solution in acetone ismade and then an aliquot is added to an emptyvial, the acetone is allowed to evaporate, and thenthe copolymer/water mixture is added. However,the drug may also be incorporated by the oil inwater emulsion method, in which case the drug isadded dropwise in a solvent such as chloroform tothe micelle solution in water. The drug is incorpo-rated as the solvent evaporates.

Finally, if the micelles are prepared by thedialysis method, then the drug is added with the


copolymer to the common organic solvent andthen the preparation proceeds as described abovefor the micelles alone. In some cases, the oil inwater emulsion method is also used for the incor-poration of drugs into micelles prepared by thedialysis method [18].

In a study by La et al., the amount of in-domethacin (IMC) entrapped into PEO-b-PBLAmicelles was measured when both the dialysismethod and the oil in water emulsion methodwere employed as methods of drug incorporation[18]. The amount of IMC entrapped into thePEO-b-PBLA micelles was found to be 20.4%(w/w) and 22.1% (w/w) when the dialysis methodand oil in water emulsion method were employed,respectively [18].

For the incorporation of drugs into crew-cutmicelle systems, the slow addition of watermethod may be employed, as described previ-ously. For example, the copolymer and drug aredissolved in the organic solvent and stirred forseveral hours. Water is then added at a slow rateand then the solutions are dialyzed against bidis-tilled water.

2.2. Micelle stability

The stability of block copolymer micelles in-cludes two different concepts thermodynamic sta-bility and kinetic stability. A micelle isthermodynamically stable relative to disassemblyto single chains in pure water if the total copoly-mer concentration is above the critical micelleconcentration (CMC). The critical micelle concen-

tration (CMC) is the copolymer concentrationbelow which only single chains exist but abovewhich both micelles and single chains are present.However, even if a micelle system is below itsCMC, it may still be kinetically stable and surviveat least for some period or time, if the core islarge and the core material is below the Tg or if itis crystalline and thus physically crosslinked.Table 2, discusses the way in which a number ofparameters affect the stability of the micelle as adrug delivery vehicle.

2.2.1. Thermodynamic stabilityA delivery system is subject to ‘sink conditions’

or severe dilution upon intravenous injection intoan animal or human subject. In an average indi-vidual, the total blood volume is approximately 5l. For example, following the intravenous injec-tion of 100 ml (i.e. 0.3 ml kg−1 min−1 for 5 min.)of a 2% (w/w) PCL21-b-PEO44 micelle solution,the concentration of copolymer in the bloodwould be 400 mg l−1. Therefore it is very impor-tant to know the critical micelle concentration ofa particular copolymer. The CMC for PCL21-b-PEO44 is 2.8×10−7 or 1.2 mg l−1 [31]. However,the copolymer concentration of 400 mg l−1 maybe below the value of the CMC of many of theother block copolymers that have been exploredas micellar delivery vehicles. The CMC values forPBLA-b-PEO have been reported to range be-tween 5–18 mg l−1 [17,21] while the CMC for aPLA-b-PEO system was found to be 35 mg l−1

[24]. The CMC values for several PEO-b-PPO-b-PEO systems were reported to range between 10

Table 2The various factors which influence the thermodynamic or kinetic stability of block copolymer micelles

Micelle stability ReferenceParameter

[31]CMC Low ¡High

Tg [18–20]¡LowHigh

LowHydrophobic–hydrophilic block ratio ¡ [48] High

LowConjugated drug content ¡ [12]High


and 1000 mg l−1 [6]. In some cases, injecting alarger volume or a more concentrated micellarsolution would prevent the copolymer concentra-tion from falling below the CMC immediatelyupon injection.

However, it may prove to be more advanta-geous to begin with a copolymer system with alower CMC value (Table 2). The CMC of acopolymer is determined by many factors, someof which are the nature and length of the core-forming block, length of the hydrophilic blockand the presence of hydrophobic solubilizates.The nature and length of the core-forming blockhave the most profound effect on the CMC. Am-phiphilic copolymers which contain a highly hy-drophobic block have lower CMC values in waterthan those which include the less hydrophobicblocks. The CMC values for PS-b-PEO copoly-mers, which contain the highly hydrophobicpolystyrene block, range between 1 and 5 mg l−1

[46].For a series of copolymers, if the corona-form-

ing block is kept constant, an increase in themolecular weight of the core-forming block willdecrease the CMC [52]. To a lesser extent, if thelength of the core-forming block is maintained ata constant length, than an increase in the lengthof the hydrophilic block will cause an increase inthe value of the CMC [52,53].

The use of a copolymer system with a lowCMC value may increase the in vivo stability ofthe micelles. However, in many papers, the disas-sembly of micelles into single chains is mentionedto be advantageous since this will facilitate elimi-nation of the copolymer material from the bodyvia the kidneys. Therefore, the ideal micelle sys-tem will be stable to sink conditions encounteredupon injection and will facilitate elimination byeventual disassembly into single chains.

2.2.2. Kinetic stabilityThe disassembly of micelles at copolymer con-

centrations below the CMC has been reported tobe quite slow for some copolymer systems [18–20]. The rate of disassembly depends, among oth-ers, upon the physical state of the micelle core[18]. Micelles formed from copolymers containinga hydrophobic block which has a high glass tran-

sition temperature will tend to disassemble moreslowly than those with a low glass transitiontemperature.

The rate of disassembly is likely affected bymany of the same factors which affect the rate ofunimer exchange between micelles. The unimerexchange rate has been found to be dependent onmany factors such as content of solvent within thecore, the hydrophobic content of the copolymerand the lengths of both the hydrophilic and hy-drophobic blocks [37,38,41,43,44,48,49]. For ex-ample, Creuz et al. studied micelles formed frompoly((dimethylamino)alkyl methacrylate)-b-sodium methacrylate and found that the rate ofunimer exchange decreased with an increase in thehydrophobic/hydrophilic balance of the copoly-mer [48].

In addition, there is also evidence that theincorporation of hydrophobic compounds intoblock copolymer micelles may enhance micellestability (Table 2). For example, in a study byKataoka’s group, they found that both the physi-cal entrapment and/or chemical conjugation ofadriamycin (ADR) into the micelle core increasedthe structural stability of the poly(ethylene gly-col)–poly(aspartic acid) (PEG–P(Asp)) micelles[12]. In their study, they assessed the stability ofthe micelle by gel exclusion chromatography.They found that the stability of the micelle in-creased as the amount of chemically conjugatedadriamycin was increased, and also that the phys-ical entrapment of adriamycin into the PEG–P(Asp)ADR micelles further enhanced micellarstability. They suggested that the presence of boththe physically entrapped and chemically conju-gated drug increased the hydrophobic interactionswithin the core, producing micelles which weremore tightly packed [12].

2.3. Micelle size

The size of colloidal particles is one of theproperties which largely influences the circulationtime and organ distribution of the vehicle. Parti-cles which are less than 200 nm are said to be lesssusceptible to RES clearance, and those less than5 mm have access to small capillaries [36]. Also,the size of the carrier may influence its mechanism


Fig. 2. The key physical properties of the micelle corona thatinfluence factors important in terms of the capabilities ofmicelles as drug carriers.

individual primary micelles which have aggre-gated due to hydrophobic–hydrophobic interac-tions between the cores of the micelles [16]. Thisaspect will be discussed further in Section 3.0.

A bimodal population distribution may be un-favorable for a drug delivery system; however,this depends upon the stability of the aggregates.Clearly, micelles of approximately 10 nm in sizeand those of 200 nm in size are likely to havedifferent circulation times and biodistributions.However, if the large aggregates are not stableand break into individual primary micelles upondilution, then they may not present a problem.For example, in our study with PCL-b-PEO mi-celles, we found that as the micelle solution wasdiluted from 0.2% (w/w) to 0.01% (w/w) the pop-ulation of larger aggregates decreased in favor ofthe smaller population [32].

3. The micelle corona

The micelle shell acts as a stabilizing interfacebetween the hydrophobic micelle core and theexternal medium (Fig. 2). The properties of theouter shell will predominantly affect the biodistri-bution of the micelle and thereby that of theincorporated drug as well as its pharmacokineticparameters. In most cases the hydrophilic shell-forming block is poly(ethylene oxide) with amolecular weight which is usually between 1000and 12 000 g mol−1 and a length which is greaterthan or equal to the length of the core-formingblock. Poly(ethylene oxide) has also been used toincrease the biocompatibility and enhance the col-loidal stability of other types of delivery vehiclessuch as nanoparticles [67], microspheres [68], andliposomes [69]. The stability imparted bypoly(ethylene oxide) is due to an effect termedsteric stabilization; this effect is most likely en-hanced by PEO’s unique solution properties [70–72].

3.1. PEO as the corona-forming block

In the bulk, PEO is a non-ionic crystalline,thermoplastic water-soluble polymer [70,71]. Itswater solubility is reported to be unlimited at

of entry into cells, which may, in turn, influencethe kinetics and extent of cell uptake.

The size of micelles is controlled by severalfactors, among which are the length of the core-forming block and the length of the corona form-ing block [36]. Several different groups havecontributed to this area of research, and as aresult scaling relations have been developed[36,50,51,65].

2.4. Size population distribution

A common problem with most block copoly-mer micelle systems has been the presence of abimodal population distribution [16,32]. Thelarger aggregates are believed to be clusters of


room temperature for all degrees of polymeriza-tion [70]. Many close structurally relatedpolyethers cannot achieve nearly the same degreeof water solubility as PEO. The high degree ofhydration and large excluded volume induce re-pulsive forces which contribute to the stabilizationof a PEO coated surface.

3.1.1. Steric stabilizationThe steric stabilization of colloidal particles in

an aqueous medium may be achieved by surfacefixation of neutral water-soluble stabilizing moi-eties such as long chains of poly(ethylene oxide)[72]. The stabilizing moieties will create stericrepulsive forces which will compete with the inter-particle van der Waals attractive forces. Coagula-tion is prevented if the repulsive forces overwhelmthe attractive forces operative between theparticles.

The outer PEO shell of micelles may inhibit thesurface adsorption of biological components.Proteins adsorb to the surface of a foreign mate-rial within the first few minutes of exposure to theblood, especially if the surface is charged or hy-drophobic [23]. The adsorption of proteins to thesurface of a drug delivery vehicle can cause dam-age and lysis of the vehicle, inducing leakage ofthe drug entrapped within the carrier. This makesit nearly impossible to predict the pharmacokinet-ics of a drug incorporated in such a system. Also,protein surface adsorption can lead to the inter-ference of many biochemical pathways such as thecomplement cascade, the coagulation cascade andthe fibrinolysis cascade [70]. The interference ofthese pathways can have serious implicationssince they play a central role in maintaininghomeostasis. Surface opsonization of proteinswhich are part of the reticuloendothelial systemwill attract phagocytic macrophages which takeup the vehicle and the net result is accumulationin the liver, spleen and lungs [66]. RES clearancedecreases the circulation time of the vehicle in thebloodstream, reducing the vehicle’s chance of suc-cessfully reaching its target site.

The PEO shell is also important in micelles as itmay influence the extent to which primary mi-celles aggregate to form large secondary clusters.The aggregation of micelles leads to a larger

average particle size for a specific micelle popula-tion; this, in turn, may have a profound effect onthe overall biodistribution of the vehicles. Largerparticles are said to be more susceptible to clear-ance by the RES and may also be unable to entersmall capillaries and other sites which are accessi-ble to smaller particles.

The extent to which the PEO corona is able tosterically stabilize the micelle depends on both thesurface density of PEO and the thickness of thePEO shell [18]. The surface density of the PEOshell will influence the amount of core materialwhich remains exposed to the aqueous medium. Itis believed that van der Waals interactions be-tween the exposed cores are responsible for sec-ondary aggregation of micelles resulting in theformation of large clusters [17]. The density ofPEO at the surface of the micelle will be deter-mined by the aggregation number or the numberof copolymer chains per micelle. The larger theaggregation number of the micelle the more PEOblocks at the micelle surface [24].

The biodistribution of the micelle-incorporateddrug will largely be determined by the surfaceproperties of the vehicles such as charge, degree ofhydrophilicity and steric stability. This effect isclearly seen in a study by Hagan et al. on thebiodistribution of two block copolymer micellesystems formed from PLA-b-PEO copolymerswhere one system had a PLA:PEO ratio of 1.5:2and the other had a PLA: PEO ratio of 2:5 [24].The PLA-b-PEO system with a ratio of PLA:PEOof 1.5:2 has a higher density of PEO at the surfaceof the micelle due to the large aggregation numberof this system in comparison to the 2:5 PLA-b-PEO system. The high PEO surface density of the1.5:2 PLA-b-PEO system led to an improvedbiodistribution with reduced liver uptake [24].

3.2. Corona-forming blocks other than PEO

There has been little investigation into the useof hydrophilic corona-forming blocks other thanPEO in drug delivery. For many of the reasonsdescribed previously, PEO has been the most pop-ular choice of hydrophilic polymer in block co-polymers used to form micellar drug carriers.However, Inoue et. al., for example, reported on


the development of micelles formed from the am-phiphilic copolymer oligo(methyl methacrylate)(oMMA) and poly(acrylic acid) (PAAc) [26]. Inan aqueous environment, this block copolymerproduces micelles with a negative charge at thesurface of the micelle.

The use of a charged corona-forming blockleads to the question of whether or not it isbiologically and/or physically more favorable forthe drug carrier to have a neutral or chargedsurface. Previous studies performed on liposomeswith a charged surface have revealed that follow-ing intravenous injection, their fate in vivo islargely dependent on their surface charge. Workby Gregoriadis et al. and Senior et al. found thatthe presence of negatively charged groups at thesurface of liposomes decreased their circulationtime and enhanced their accumulation in both theliver and the spleen [73,74]. Liposomes with apositive charge at the surface were found to haveenhanced uptake in both the lungs and the liver[75]. However, in later studies, it was found that ifthe negative charge at the surface of the liposomeswas shielded by ‘bulky hydrophilic groups’ theircirculation time was actually enhanced [76]. Papa-hadjopoulos suggested that the negative charge atthe surface of the liposomes requires direct con-tact with certain biological components (plasmaproteins or cell-surface receptors) in order to re-sult in a decrease in the circulation time of thevehicle [76,77].

In many studies, it has been found that thepresence of negative charge at the surface of thevehicle will enhance the extent of in vitro uptakeinto various cell lines. For example, Bajoria et al.,found that both anionic and neutral unilamellarliposomes were taken up into trophoblast cellsmore readily than cationic liposomes of the samesize [78]. Also, Yu et al. found that the in vitrocell uptake of negatively charged liposomes intohepatocytes was greater than the uptake of neu-tral or cationic liposomes [79]. In addition, theyfound that the cationic liposomes gave rise to thehighest plasma concentrations. For this reasonthey suggested that cationic liposomes may bemost appropriate for use when liver uptake is tobe avoided while negatively charged liposomes aremore useful for delivery to the liver [79].

Several delivery systems with charged surfaceshave been designed [80–84] for delivery to mu-cosal surfaces which include the lower and upperrespiratory tracts, the gastrointestinal tract andthe urogenitary tract [85]. Many groups havefound that for delivery to mucosal surfaces, it isadvantageous to use a delivery system with abioadhesive surface. In a study by Akiyama et al.the use of mucoadhesive microspheres with poly(-acrylic acid) at the surface resulted in the accumu-lation of a greater amount of drug in the stomachand higher drug plasma concentration levels incomparison to those obtained with non-mucoad-hesive microspheres [82]. The mucoadhesive prop-erties allow for increased gastrointestinalresidence times which may aid gastrointestinalabsorption. Also, discussions by Bailey et al.demonstrate that the particle size and charge oftherapeutic aerosols may provide a means oftargeting specific regions of the lung [86].

Therefore, if we consider the information avail-able on other delivery systems (liposomes, mi-croparticles, hydrogels etc.) with charged surfaces,it is clear that there is a need for the explorationof micelle systems with negatively chargedcorona-forming blocks such as the system de-scribed by Inoue et al. [26]. The charged systemsmay be most useful for delivery to mucosal sur-faces and may afford the development of effectiveoral and aerosol block copolymer micellarformulations.

4. The micelle core

4.1. Loading capacity

The micelle core serves as the cargo space forvarious lipophilic drugs (Fig. 3). However, thiscargo space is limited; for instance, a typical 1%(w/w) PCL-b-PEO (20-b-44) micelle solution(Nagg=125) contains only approximately 0.5%core volume [32]. This means that in a 1 mlaliquot of this 1% (w/w) micelle solution only 5 mlis core volume. In order to exploit maximally theminimal loading space available, we must manip-ulate the many factors which control the loadingcapacity and loading efficiency.


Several of the major factors which influenceboth the loading capacity and loading efficiencyof block copolymer micelles are nature of thesolute, nature of the core-forming block, coreblock length, total copolymer molecular weight,solute concentration and, to a lesser extent, thenature and block length of the corona. Manystudies have indicated that the overriding factor isthe compatibility between the solubilizate and thecore-forming block.

4.1.1. Compatibility between the solubilizate andthe core-forming block

Many studies have explored the influence ofproperties of the solubilizate on their extent ofincorporation into block copolymer micelles. Sev-eral of the important properties of the solubilizatethat have been identified are molecular volume,interfacial tension against water, polarity, hydro-phobicity, charge and degree of ionization

Studies by Nagarajan et al. demonstrated thatthe amount of solubilizate incorporated decreaseswith an increase in the molecular volume of the

solubilizate [34]. They also found a correlationbetween the interfacial tension of the solubilizateagainst water and the extent of incorporation intoblock copolymer micelles. Aromatic hydrocarbonsare incorporated to a greater extent in comparisonto aliphatic hydrocarbons owing to their lowerinterfacial tension against water [34].

The nature of the solute, including polarity,hydrophobicity, charge and degree of ionization,have also been found to influence greatly theincorporation; however, this is entirely dependenton the nature of the core-forming block. It is thecompatibility between the solute and the core-forming block that can be used to enhance incor-poration most effectively. One parameter whichhas been used to assess compatibility between thepolymer and the solubilizate is the Flory–Hug-gins interaction parameter. This interactionparameter (xsp) between the solubilizate and thecore-forming block is described by the followingequation:

xsp= (ds−dp)2 Vs

RT

where xsp= interaction parameter between thesolubilizate (s) and the core-forming polymerblock (p), ds= the Scatchard–Hildebrand solubil-ity parameter of the solubilizate, dp= theScatchard–Hildebrand solubility parameter of thecore-forming polymer block and Vs= the molarvolume of the solubilizate [34,54].

The lower the positive value of the interactionparameter (xsp) the greater the compatibility be-tween the solubilizate and the core-forming block.The highest degree of compatibility will bereached when ds=dp. Note, that if specific inter-actions are present, e.g. ionic interactions, thevalue of x may even be negative. The possiblecomplexity of polymer drug interactions suggeststhat the largest amount of drug loaded per micellewill be reached when the core-forming block ismost suitably matched with the drug to be loaded.Due to the fact that each drug is unique, thissuggests that no one core-forming block will en-able maximum loading levels to be achieved forall drugs. For this reason, it is unlikely that anyone micelle system will serve as a universal deliv-ery vehicle for all drugs. In the same way, it is

Fig. 3. The properties of the core and drug which have themost profound influence on both the loading capacity of themicelle and release kinetics of the drug.


equally unlikely that any one drug will be asefficiently delivered by all block copolymer micellesystems. For instance, the partition coefficient forthe model hydrophobic compound, pyrene, be-tween micelles and water has been measured inseveral different block copolymer systems. Theresults obtained revealed the partition coefficientsto be of the order of 102 for PEO-b-PPO-b-PEO[8], 103 for PCL-b-PEO [31], 104 for PBLA-b-PEO [19], 105 for PS-b-PEO (and PS-b-PAA)[46,56]. From the results we see that the partitioncoefficient for pyrene is higher in systems withcores which are less polar [19]. Therefore, it ap-pears reasonable to expect that more polar com-pounds will be most easily loaded into micelles inwhich the cores are more polar. For example,Kim et al. [25] obtained an indomethacin contentof 42.2% in micelles formed from PEO-b-PCLwhile La et al. obtained an indomethacin contentof only 22.1 (wt)% for micelles formed from PEO-b-PBLA [17]. These studies demonstrate the needto match a drug with certain properties with acopolymer containing a core–forming blockwhich has similar properties.

The environment within the core may be madeto be more compatible for a specific drug byattaching drug molecules to the core-formingblock. This was shown in studies by Yokoyama etal. who found that the physical entrapment ofadriamycin was higher in micelles formed fromadriamycin conjugated copolymer (PEO-b-P(As-p(ADR))) than in micelles formed from the PEO-b-P(Asp) copolymer alone [12].

Aside from polymer–drug compatibility andinteractions, there are other factors which influ-ence the extent of incorporation of a solubilizateinto block copolymer micelles. These will be dis-cussed below.

4.1.2. Other factors which influence drug loading

4.1.2.1. Length of the core-forming block. For acopolymer with a constant hydrophilic blocklength, an increase in the length of the core-form-ing block has been found to increase the partitioncoefficient of the solubilizate between the micellesand the external medium [39,42]. Yu et al. mea-sured the partition coefficient for pyrene between

water and PCL-b-PEO micelles formed from co-polymers with constant PEO block length anddifferent PCL block lengths [31]. The partitioncoefficient was found to increase from 240, 760,1450, for micelles formed from PCL-b-PEO withthe same PEO block length and 14, 21 and 40units of caprolactone, respectively [31].

An increase in the length of the core-formingblock has been found to decrease the criticalmicelle concentration (CMC) and also cause anincrease in the core size per micelle which, in turn,results in an increased loading capacity per mi-celle [54]. For example, for any one block copoly-mer system, as one increases the length of thecore-forming block the aggregation number in-creases, resulting in a larger core size and thus alarger cargo space per micelle. However, due tothe increase in the aggregation number per mi-celle, the total number of micelles in solution willdecrease per unit mass of polymer. In addition,Kabanov’s group has found a clear reciprocalrelationship between the cmc and the length of thecore-forming block [6].

4.1.2.2. Length of the corona-forming block. Asignificant increase in the length of the hy-drophilic block can both increase the critical mi-celle concentration and decrease the aggregationnumber. The increase in the cmc will cause fewerchains to be present as micelles and thus availablefor drug loading. The smaller fraction of copoly-mer chains in micelle form decreases the degree ofsolubilization since there is a smaller hydrophobicvolume [42].

4.1.2.3. Nature of the corona-forming block. Theeffect of the nature of the corona-forming blockon incorporation is once again a question ofcompatibility. The Flory–Huggins interactionparameter between the drug and the corona-form-ing block may be considered in this context. If theinteraction between the corona-forming block andthe drug is favorable, even if only slightly, someof the drug may be present in the outer shell. Instudies by Gadelle et al. PEO-b-PPO-b-PEO mi-celles were found to have a higher affinity forchlorobenzene than for benzene; this was unex-pected, since the xsp for benzene (x=0.0014) is


lower than that for chlorobenzene (xsp=0.0068)[54]. However, some of the cholorobenzene mayhave been solubilized in the outer shell since thexsp (PEO) for chlorobenzene (xsp(PEO)=0.1711)is lower than that for benzene (xsp(PEO)=0.2483)[54].

Likely, amphiphilic drugs with a low xsp(coreblock) for the hydrophobic portion of themolecule and a low xsp(corona block) for thehydrophilic portion of the molecule will be wellsolubilized at the interface of the micelle and inthe near-corona region.

4.1.2.4. Copolymer concentration. In theoreticalstudies by Xing et al. the solubilization capacitywas found to increase to the saturation level(maximum loading level) with an increase in thecopolymer concentration [42]. The copolymerconcentration at which the saturation level isreached is largely influenced by the interactionbetween the solubilizate and the core-formingblock, where stronger interactions enable satura-tion to be reached at lower polymer concentra-tions [42]. Several studies indicate that there aretwo patterns of solubilization, one in which theextent of solubilization increases with copolymerconcentration, and another in which the degree ofsolubilization is independent of the copolymerconcentration [42,54]. Hurter and Hatton foundthat the solubilization of napthalene into micellesformed from a range of PEO-b-PPO-b-PEO co-polymers was independent of copolymer concen-tration for micelles formed from copolymers witha high hydrophobic content. In the same studies,it was found that the extent of solubilization didincrease with copolymer concentration for themicelles formed from copolymers with high hy-drophilic block contents [87–89]

4.1.2.5. Solute concentration. The presence of ahydrophobic solubilizate has been found to en-hance the aggregation of amphiphilic copolymermolecules. Studies by Mattice et al. employingMonte Carlo simulations to study the solubiliza-tion of small molecules into triblock copolymermicelles revealed that the solubilizate can enhancethe aggregation of copolymer molecules [42]. Theenhancement of the aggregation lowers the CMC

value, leading to an increased number of micellespresent in solution [42,54]. Increasing the soluteconcentration has also been found to increase theaggregation number of the copolymer whichcauses larger micelles to be produced. This isreported to increase the solubilization capacityper micelle [42].

Clearly, there are a large number of factorswhich influence both the loading efficiency andloading capacity of block copolymer micelles.However, the most important factor identified todate is the compatibility between the drug and thecore-forming block. For this reason, the choice ofwhich polymer to employ as the core-formingblock is most crucial. The copolymer-drug matchshould be made on the basis of many of theparameters described above. Once the copolymersystem has been chosen the other properties(block copolymer length, copolymer mol. Wt.,copolymer concentration) may be modified asneeded.

4.2. Release kinetics

The release of drugs from block copolymermicelles will depend upon the rate of diffusion ofthe drug from the micelles, micelle stability andthe rate of biodegradation of the copolymer. Ifthe micelle is stable and the rate of biodegrada-tion of the copolymer is slow, the release rate willbe mostly influenced by several of the followingfactors: the strength of the interactions betweenthe drug and the core-forming block [17], thephysical state of the micelle core [13], the amountof drug loaded [25], the molecular volume of thedrug, the length of the core-forming block [25]and the localization of the drug within the micelle.

4.2.1. Polymer–drug interactionsThe stronger the interaction between the drug

and the core-forming block the slower the releaseof the drug from the micelle. For example, studiesby La et al. found that the in vitro release rate ofindomethacin from PBLA-b-PEO micelles in-creased with increasing pH of the externalmedium. As the pH of the medium is increased,more of the carboxylic acid groups become ion-ized, weakening the hydrophobic–hydrophobic


interaction which holds the drug within the mi-celle core [17].

Strong polymer–drug interactions will enhanceloading and decrease the release rate of the drugfrom the micelles. For this reason, a compromisemust be achieved such that both the loading leveland the release kinetics of the drug are optimized.

4.2.2. Localization of the drug within the micelleThe incorporated drug may lie within the mi-

celle core, at the interface between the micelle coreand the corona or even within the corona itself.The localization of the drug will largely dependupon its physical properties and the interactionparameters between the drug and the micelle coreand corona – forming blocks [40,45]. It has beenfound that the more soluble compounds are local-ized in the inner corona or the core–corona inter-face while the more hydrophobic compounds aresituated mostly in the micelle core [40].

The release rate of the drug will largely be afunction of its localization within the micelle. Theouter corona region of the micelle is quite mobile;as a result, release from this area will be rapid.The release of drug localized in the corona or atthe interface is said to account for ‘burst release’from the micelle [40].

4.2.3. Physical state of the micelle coreThe state of the micelle core, whether it be

liquid-like or solid-like, will have a large influenceon the rate of release of drugs from the micellecore. The physical state of the core under normalphysiological conditions (37°C) will dependlargely upon the glass transition temperature ofthe core-forming block and the degree of crys-tallinity, if any. Below the glass transition temper-ature the polymer is in a solid-like state, whileabove it the polymer is in a liquid-like state,although the changes in properties such as viscos-ity are usually gradual [90]. Micelles containingpolystyrene cores are glassy at room temperaturesince the glass transition temperature ofpolystyrene is ca. 100°C; however, if a solvent candiffuse into the core, plasticization can lower theTg appreciably. Many of the biocompatible poly-mers have glass transition temperatures which arewell below that of polystyrene. For instance, the

glass transition temperatures for several of thebiocompatible polymers are 35°C for Poly(gly-colic acid) (mol. wt. 50 000), 54°C for Poly(L-lac-tic acid) (mol. wt. 50 000), −62°C forPolycaprolactone (mol. wt. 44 000) and −75°Cfor Poly(propylene oxide) [91]. The movement ofa drug from a glassy core will be slower in com-parison to the movement of drugs out of coreswhich are more mobile. This effect was seen, forexample, in studies by Teng et al. who found thatthe diffusion constant for pyrene in poly(styrene)is less than it is in poly(tert-butyl acrylate). Thisresult was expected since the Tg of polystyrene is100°C while that of poly(tert-butyl acrylate) is42°C [40].

In addition, micelle cores formed from themore hydrophilic polymers may contain apprecia-ble amounts of water. The release of drugs fromcores of this nature has been found to proceedrapidly [40].

4.2.4. Other factors influencing release kinetics

4.2.4.1. The length of the core-forming block. Theproperties of the micelle core, including core ra-dius, influence the release rate if the drug isprimarily situated in the micelle core. The longerthe core-forming block, the larger the core, andthe slower the release of the drug from themicelle.

However, the release kinetics of drugs whichreside at the core–corona interface or in the outershell do not have to diffuse through the core, andare thus not influenced by the core radius orlength of the core-forming block [45]. For exam-ple, Hruska et al. found that the exit rate coeffi-cient for the fluorescent probe BAN wasindependent of the size of the micelle since itresides close to the core surface [45].

4.2.4.2. Molecular 6olume of the drug. The size ormolecular volume of the drug will affect its rate ofdiffusion from the micelle [40]. Drugs with largermolecular volumes will have a smaller diffusionconstant resulting in a decreased release rate.Once again, the burst release or release of drugslocalized at the micelle surface should be indepen-dent of the molecular volume.


4.2.4.3. The physical state of the drug in the mi-celle. The physical state of the drug within thedelivery vehicle can have a significant effect on therelease kinetics. The drug which has been incorpo-rated into the micelles may not be well dissolvedor solubilized in the copolymer. Several groupshave identified cases where drugs which are incor-porated into delivery vehicles are not dissolved inthe polymer; instead, they exist in another form[92–94], e.g. as crystals dispersed within the mi-celle core. If the drug is dissolved molecularly, itmay act as a plasticizer and lower the glass transi-tion temperature of the core-forming block; this,in turn, may accelerate the release. However, ifthe drug is present as a crystal, it may instead actas a reinforcing filler, especially if strong interac-tions between the polymer and the surface of thecrystallite are present, which may cause an in-crease in the glass transition temperature.

5. Morphologies of block copolymer aggregates

To this point, the effect of the morphology ofthe aggregates on their efficiency as drug carriershas remained virtually unexplored. In most pa-pers, if the morphology was reported, it wasspherical. This morphology is most likely to beencountered, since most of the micelle systemswhich have been tried as drug carriers are star-type systems which are known to produce spheres.

In contrast, amphiphilic block copolymer sys-tems which form crew cut aggregates have beenshown to produce a wide range of morphologies[61–65,95–108]. The crew-cuts are formed fromhighly asymmetric block copolymers in which thelength of the core-forming block is much greaterthan that of the corona-forming block. The mor-phologies include spheres, rods, vesicles, lamellae,large compound micelles, tubules, hexagonallypacked hollow hoops and many more [61–65,95–105]. The number of morphologies that can bereproducibly formed in a highly controlled man-ner is of some interest, because the usefulness ofaggregates other than spheres in drug delivery,while not established, is certainly an intriguingpossibility.

5.1. The potential rele6ance of morphologies ofblock copolymer aggregates in drug deli6ery

There are good reasons for suspecting thatmorphologies other than spheres may be of inter-est in drug delivery. The physical parameters ofdrug delivery systems, such as size and surfacecharge, have been shown to have a substantialinfluence on the effectiveness of the drug deliverysystem in terms of its capability to deliver certaindrugs. As previously mentioned, the biodistribu-tion, circulation time and mechanism and extentof cell uptake are altered by changing the surfaceproperties and/or size of the drug carrier. Thestriking influence of these two physical parameterssuggests that it may be worthwhile to explore themorphology of the carrier as another performancerelated physical parameter.

The various morphologies could be used fordifferent applications in drug delivery. For in-stance, rod shaped aggregates will likely havedifferent loading capacities and release kineticsthan their spherical counterparts. The rod-likeaggregates may be most useful in the preparationof aerosol formulations whereby their thin tubularstructure may facilitate access to the differentparts of the lung, because the aerodynamic prop-erties of rods have been shown to be more suit-able than those of spheres [109].

Block copolymer vesicles can be made to con-tain hydrophilic compounds, as has been shownfor liposomes [69,73–76] and a copolymer system[110]. In this way a cocktail of vesicles and mi-celles could be used to deliver a combination ofhydrophilic and hydrophobic drugs. There aremany hydrophobic/hydrophilic drug combina-tions which are synergistic, one example is NGFand FK506. The use of sphere/vesicle combina-tions will first require the complete in vivo charac-terization of both systems. These studies arenecessary to ensure that the release kinetics,biodistribution and pharmacokinetic parametersof drugs incorporated into both systems are suit-able for this purpose.

If crew-cut copolymer systems are to be usedfor drug delivery, one should be at least aware ofwhat conditions produce which morphologies.This will be discussed in Section 5.2., which will


also mention the conditions under which somepitfalls may be encountered.

5.2. Morphologies of crew-cut aggregates

An aqueous solution of stable crew-cut aggre-gates cannot be formed by direct dissolution intowater due to the large hydrophobic content of thecore-forming block. The preparation requires ini-tial dissolution of the block copolymer in a com-mon organic solvent (DMF, dioxane or THF),and then slow addition of a material which is agood solvent for the hydrophilic blocks but aprecipitant for the hydrophobic blocks (e.g. wateror methanol) [61–65,95–105]. An aqueous solu-tion can then be obtained by dialyzing againstwater in order to remove the common organicsolvent. The addition of the precipitant increasesthe Flory–Huggins parameter between the solventmixture and the core-forming block. This increas-ingly uncomfortable environment for the hydro-phobic blocks provides the driving force formicelle formation. If, during the process of wateraddition, one crosses stability regions for differentmorphologies, as is frequently the case with crew-cut copolymer systems, then mixtures of mor-phologies can result. However, this can sometimesbe avoided by the dissolution of the copolymer ina specific solvent/precipitant mixture at a particu-lar copolymer concentration that is known toyield a single morphology. Frequently, however,regions are encountered where morphologies cancoexist under thermodynamic equilibrium. Thus,a knowledge of both the thermodynamics andkinetics of the micellization process are needed,and these are under active study.

A large number of morphological studies havebeen carried out by our group, primarily usingpolystyrene-b-poly(acrylic acid) and polystyrene-b-poly(ethylene oxide) copolymers with a widerange of copolymer compositions [61–65,95–105].These studies have revealed that the formation ofvarious morphologies of crew-cut aggregates maybe explained by a force balance effect involvingthe following three components: the degree ofstretching of the core-forming blocks, the interfa-cial energy between the micelle core and the sol-vent, and the inter-corona chain interactions

[62,96]. The force balance also determines thestructural parameters of the existing morphologysuch as size, core dimensions, Nagg etc. A verybrief description of the three force componentsmay be useful at this point.

The degree of stretching (Sc) may be defined asthe ratio of the core radius to the copolymer chainend to end distance in the unperturbed state. Anincrease in the degree of stretching of the core-forming blocks results in a loss in entropy for thesystem. For spherical micelles the degree ofstretching will increase with an increase in thecore radius.

The inter-corona interactions provide anothermechanism for morphological change. In an ion-izable block such as PAA, the inter-corona inter-actions may be both steric and electrostatic innature. The strength of the interactions is in-versely proportional to the area per corona chainat the core–corona interface. For spherical mi-celles the surface area per chain increases withdecreasing aggregation number of the micelles.For each individual morphology, spheres, rods,vesicles, the area per corona chain decreases withan increase in the dimension of the PS region.However, in comparing different morphologieswith the same dimension of the PS region (e.g.sphere or rod diameter, vesicle wall thickness), thearea per corona chain decreases from spheres torods to vesicles [61]. The interfacial energy be-tween the core and corona regions generally in-creases with an increase in the water content inthe solvent mixture. This progressive increase is adriving force for the minimization of the totalinterfacial area, and leads to an increase in theaggregation number per micelle but a decrease inthe total number of micelles [61].

The structural parameters of the resulting mor-phology will be a result of the balance achievedbetween these three forces. For instance, forspherical crew-cut micelles, the addition of water(one of the morphologenic factors) increases theinterfacial energy between the core and thecorona; in response to such a perturbation, thesystem will tend to decrease the total interfacialarea. This is accomplished by increasing the di-ameter of the micelles through increasing the ag-gregation number, and thus results in a


Table 3Several of the morphogenic factors which control the morphology of block copolymer micelles and examples of each

Morphogenic factors Block copolymer Morphology Reference

Copolymer composition (5.3.1) PS-b-PAA200-b-21 Spheres [61,96]

Rods200-b-15 [61,96]Vesicles [61,96]200-b-8

Copolymer Concentration (5.3.2) PS-b-PAA410-b-25

Spheres2 wt% [61,101]2.6 wt% Rods [61,101]4 wt% Vesicles [61,101]

Presence of added acid, base or salt (5.3.3) PS-b-PAA410-b-25 (1 wt%) Spheres

RodsAcid [63]+HCl (210 mM)Vesicles(240 mM)

Base PS-b-PAAVesicles410-b-13 (2wt%)Spheres (dia.=38 nm)+NaOH (28 mM) [97]

(56 mM) Spheres (dia.=35 nm) [97](115 mM) Spheres (dia.=33nm) [97]

Salt PS-b-PAA410-b-25 (1 wt%) Vesicles

Spheres 63+CaCl2 (120 mM)

Common organic solvent (5.3.4) PS-b-PAA390-b-40

Spheres [100,102]DMFVesicles [100,102]THF or Dioxane

consequent decrease in the total number of mi-celles. However, in the process, the degree ofstretching increases and the inter-corona repul-sions increase, and at some point one of twothings can happen. The process comes to an endand the aggregation number can not increasefurther due to the presence of strong inter-coronarepulsions; this occurs if the corona chains arelong as is the case in the star-type micelles. How-ever, if the corona chains are short and the repul-sions are weak, the core-size expansion continuesuntil some critical point is reached at which thedegree of stretching is too large to support thespherical morphology, at which point the mor-phology changes. The occurrence of either ofthese two events depends not only on the relative

lengths of the two blocks, but also on the othermorphogenic factors to be discussed later.

The balance of forces which results in mor-phological control can be exercised by controll-ing a number of different parameters. Theresulting morphology depends on coil dimen-sions and interfacial energy; there are severaldifferent variables which may be used tocontrol this. These variables include, among oth-ers, block copolymer composition, copoly-mer concentration, type and concentration ofadded ions and the nature of the common sol-vent used in micelle preparation. The studieson the PS-b-PAA system have enabled severalof the morphogenic factors and their effectsto be identified (Table 3), these will now be dis-cussed.


5.3. Morphogenic factors

5.3.1. Copolymer composition

5.3.1.1. Effect on the degree of stretching. Thevalue for the degree of stretching of the PS coreforming block has been found to depend on thecopolymer composition by the following equa-tion: Sc:NPS

−0.1 NPAA−0.15. This equation applies to

spherical micelles in aqueous solution which donot contain any remaining organic solvent in theircores [96]. It should be recalled that, prior todialysis, the micelle cores are swollen with com-mon organic solvent; as a result, the value of Sc isgreater than it is following dialysis. The equationdemonstrates that an increase in the length of thecore and/or corona forming blocks, while keepingthe other block constant, will lead to a decrease inthe degree of stretching of the core-formingblocks. For example, consider a spherical micellecontaining a long poly(acrylic acid) block and apolystyrene block of some particular length. If thelength of the poly(acrylic acid) is decreased, thedegree of stretching of the core will increase, andat some point the morphology can change toproduce rod shaped aggregates. The decrease inthe length of the poly(acrylic acid) block decreasesinter-corona repulsions and thus allows more co-polymer chains to pack into one micelle. By con-trast, if one were to begin with a spherical micelleand decrease the polystyrene block length, themorphology will not change. Yet, if one takes arod shaped micelle and then decrease thepolystyrene block length, eventually the decreasein the degree of stretching will result in a changeto a spherical morphology.

5.3.1.2. Effect on inter-corona interactions. Thearea per corona chain for spherical micelles isrelated to the copolymer composition by the fol-lowing relation: Ac�NPS

0.6NPAA0.5 . The area per

corona chain increases with an increase in thelength of the core and/or corona forming blocks[96]. Since the inter-corona interactions are in-versely proportional to the area per corona chain,this means that an increase in the length of eitherblocks will lead to an increase in the inter-coronachain repulsions.

Consider a PS-b-PAA copolymer with a con-stant PS block length and a decreasing PAA blocklength. Let us begin with spherical micellesformed from copolymers with long PAA blocklengths, in which the aggregation number ismostly restricted by the inter-corona repulsions.As the PAA block length is decreased, the inter-corona repulsions become less important, and in-stead, the aggregation number is restricted by thedegree of stretching of the core-forming blocks.Eventually, a further decrease in the length of thePAA blocks will lead to a change in morphologyfrom spheres to rods and even to vesicles. Forexample [61,96], for a series of PS-b-PAA copoly-mers with a constant PS block length but differentPAA block lengths, 200-b-21, 200-b-15, 200-b-8the morphologies obtained were spheres, cylindersand vesicles respectively, for which the degrees ofstretching of the PS blocks were calculated to be1.41, 1.26, and 0.99. If the three copolymer sys-tems had each formed spherical morphologies, thedegree of stretching in each case would have been1.41, 1.47, 1.62. The change in morphology fromspheres, to rods to vesicles led to a decrease in thedegree of stretching of the core-forming blocks,and, in turn, to a decrease in the free energy ofthe system.

5.3.2. Copolymer ConcentrationAn increase in the copolymer concentration in

solution will increase the aggregation number ofthe micelles [61,101]. The increased aggregationnumber means the chains in the core will bestretched, which results in the production of rodsunder conditions that would previously have pro-duced spheres, or vesicles under conditions thatwould previously have produced rods. Therefore,using the same PS-b-PAA copolymer in the samesolvent, one can obtain different morphologies bysimply changing the copolymer concentration.For example, it was reported that the aggregatesformed from PS(410)-b-PAA(25) at an initial co-polymer concentration of 2 wt% were spherical, ifthe concentration was 2.6 wt% rod-like micelleswere formed, and for a 4 wt% solution vesicleswere formed (with a few rods also present insolution).


5.3.3. Presence of added acid, base or saltThe presence of ions influences the morphology

of the micelles by effecting the strength of theinter-corona chain interactions [63,97]. Severalstudies were performed on solutions of PS-b-PAAcopolymers in the presence of various ions includ-ing HCl, CaCl2, NaCl, NaOH. The addition ofelectrolytes to the PS-b-PAA solutions was eitherfound to strengthen or weaken the inter-coronarepulsive interactions, depending on the nature ofthe ions added. PAA in a DMF/water mixture isslightly ionized, so the addition of HCl protonatesthe ionic sites resulting in a decrease in the inter-corona repulsions; this decrease in turn, will causethe morphology to change from spheres to rods tovesicles. On addition of CaCl2, the Ca2+ ions willbind to the carboxylate groups and in this waydecrease inter-corona repulsions, which will havethe same effect as the addition of HCl. The addi-tion of NaCl will also decrease inter-corona repul-sions, but in this case by providing electrostaticshielding of the few ionic sites on the chain. Bycontrast, the addition of NaOH in small quanti-ties ionizes the poly(acrylic acid), which increasesinter-corona repulsions for electrostatic reasons.As a result, the increase in the concentration ofNaOH changes the morphology from vesicles tospheres.

5.3.4. Common organic sol6ent used in micellepreparation

For the preparation of the crew-cut aggregates,the block copolymer is initially dissolved in acommon organic solvent prior to the slow addi-tion of water. The choice of common organicsolvent employed has been found to have a pro-found effect on the morphology obtained. Thedetailed explanation of this effect is quite compli-cated, because a change in the solvent employedwill have an effect on three of the six interactionparameters involved. The six interaction parame-ters are: polymer A – polymer B, solvent –polymer A, solvent – polymer B, solvent – pre-cipitant, precipitant – polymer A and precipitant– polymer B. The change in the interactionparameters will change the coil dimensions butusually to a different extent for the corona andcore blocks. The precipitant may also be changed

from water to, for example, methanol; this willalso change three interaction parameters. For thisreason, it is difficult to predict which morphologywill be produced by changing either the commonorganic solvent or the precipitant employed.However, if we know that by using one commonorganic solvent we get, for example, spheres, andwith another we get vesicles; then by using amixture of the two organic solvents we can obtainthe intermediate morphology.

It is important to keep in mind that an inter-play between thermodynamics and kinetics is op-erative in all of the examples described above. Therate of morphological change is a function of thesolvent content in the core, which, in turn, isdetermined by water content in the solvent mix-ture outside the core. The higher the water con-tent the slower the kinetics. This can be illustratedby a recent study [106], which involved the mixingof two different copolymer (PS(1140)-b-PAA(165)and PS(170)-b-PAA(33)) micelle solutions inDMF/water solvent mixtures with differentDMF/water ratios. For the PS(1140)-b-PAA(165)copolymer system alone the micelles produced are38 nm in diameter, while for the PS(170)-b-PAA(33) copolymer system the micelles are 25 nmin diameter. When the two copolymer micellesolutions are mixed at a 5 wt% water content, asingle micelle population is obtained following a24 h period. By contrast when the added watercontent is 11 wt% the individual micelle popula-tions remained intact following the 24 h period.Thus, for a particular rate of water addition, awater content will be reached above which themorphology no longer changes over the experi-mental time scale, and the system can be consid-ered to be frozen. At that point the morphologyreflects the equilibrium that was prevalent underconditions (water content) when equilibrium wasstill operative. It should also be recognized thatbecause of the generally slower kinetics in blockcopolymer systems when compared to smallmolecule amphiphiles, it is very easy to trap mor-phologies which are intermediates in the transi-tion from one morphology to another. Forexample, a morphology consisting of vesicles withhollow rods in the walls running parallel to thewall was recently documented, this morphology is


an intermediate in the transition from vesiclesto hexagonally packed hollow rods or hoops[105].

To this point we have discussed only threemorphologies in detail, spheres, rods and vesicles.However, a wide range of other morphologieshave been prepared in our studies, several ofwhich may be of interest in drug delivery. Illustra-tions are provided in Figs. 4, 5, and 6. Fig. 4,includes a few of the ‘simple’ morphologies ob-tained with PS-b-PAA copolymers, i.e. spheres(a), rods (b), vesicles (c); a more complex mor-phology is shown in 4 (d) and consists of acompartmented sphere; we refer to it as a largecompound vesicle (LCV). Several of the interest-ing morphologies obtained with PS-b-PEO co-polymers are shown in Fig. 5 they include: atubule (a), a tubular bicontinuous structure

(plumber’s nightmare) (b), pincushions (c) and avesicle with hollow rods in the walls (d). In Fig. 6,two of the most striking morphologies from PS-b-PAA are shown, i.e. a bicontinuous rod (a) andhexagonally packed hollow hoop structures (b).The latter is analogous to a hollow rod structureexcept that on the size scale of the present featurethe endcap energy is larger than the energy ofcurvature which makes the rods close up intohoops. The hexagonal packing is preserved. Theseaggregates of various morphologies are stable,low free energy structures; and it is thus notsurprising that several of them show a great simi-larity to biological structures commonly found innature. The biological structures include for ex-ample; the smooth endoplasmic reticulum (bicon-tinuous rods), microtubules (tubules) and vesicles(endosomes, cells and several others).

Fig. 4. (A) spherical micelles from PS(500)-b-PAA(58); (B) rod-like micelles from PS(190)-b-PAA(20); (C) vesicles from PS(410)-b-PAA(20); and (D) large compound vesicles from PS(410)-b-PAA(13) (all as prepared in Ref. [61]).


Fig. 5. (A) tubule from PS-b-PEO (reprinted from Ref. [103]); (B) tubular bicontinuous from PS-b-PEO (reprinted from Ref. [99]);(C) pincushions from PS-b-PEO (reprinted from Ref. [99]) (D) vesicles with hollow rods in the walls from PS-b-PEO (reprinted fromRef. [104]).

6. Conclusion

Fig. 7 summarizes many of the most favorablecharacteristics of the ‘ideal micelle system’. Themicelle as a whole should be less than 100 nm indiameter in order to avoid RES uptake and tofacilitate access to cells and tissues. A narrow sizedistribution is favored over a broad distributiondue to the unpredictable change in both the drug’spharmacokinetic parameters and organ biodistri-bution which may result from a broad micelle sizedistribution. The micelle should be stable to ‘sinkconditions’ encountered upon injection. The life-time of the micelle should be longer than the timerequired for it to reach its target site. The micellecorona must act as an effective steric barrier whilealso providing sufficient coverage of the core.Also, modification of the corona may also enablethe micelle to be targeted to a specific site. Themicelle core should have a high loading capacityand allow for a controlled drug release profile.

Core-drug compatibility has been found to be oneof the key determinants for the loading capacityof the micelle.

It is difficult to produce micelles which possessall of the ideal performance related properties.Largely, this is due to the fact that many of theperformance related properties are influenced bythe same factors. As demonstrated in Table 4, achange in one factor can influence many differentproperties of the micelle. For example, the load-ing capacity per micelle may be enhanced byincreasing the length of the core-forming block,yet this may also increase the size of the micelle.For this reason, it may be necessary to assignpriority to the different properties of the micelleand in this way compromise in order to create thebest possible micellar delivery vehicle.

To this point, the morphology of the micellardelivery vehicles has not yet been explored as aperformance related property. However, due tothe overwhelming influence of other physical

C.

Allen

etal./

Colloids

andS

urfacesB

:B

iointerfaces16

(1999)3

–27

23

Table 4The factors which influence the many performance related properties of block copolymer micelles as drug carriers

Influencing fac- Performance related propertiestors

Surface properties -steric stability -hy- Size -Nagg Stability -CMC -rate of dis-Release profile MorphologyLoading capacity ordrophilicityefficiency assembly

Drug:× ××Nature (ds)

×Molecular vol- ×ume

× ××Concentration ×Core-forming

block:××× ×××Nature (dp-core)

×Tg ×××Length ×× ×

Corona-formingblock:

× × ××Nature (dp-

corona)××××Length

Copolymer:×× ×Block Ratio ×× ×

×× ×× ×Total Mol. Wt.× ×Concentration ×


Fig. 6. (A) bicontinuous rods from PS(190)-b-PAA(20) (as prepared in Ref. [61]); (B) hexagonally packed hollow hoop structurefrom PS-b-PAA (as prepared in [105]).

Fig. 7. A schematic of the ‘ideal micelle system’ including the most favorable properties for the micelle as a whole, the core and thecorona.


properties such as size and surface charge, itseems necessary to investigate morphology as an-other potentially important property. The extentto which the micelles with various morphologiesmay be reproducibly prepared also encouragesexploration in this area.

Acknowledgements

The authors would like to thank NSERC forsupport of their research; C. Allen is grateful toNSERC for financial assistance provided in theform of a fellowship. Thanks are also due to bothDr Kui Yu and Dr Lifeng Zhang for providingillustrations of the block copolymer aggregates ofmultiple morphologies.

References

[1] A.V. Kabanov, V.P. Chekhonin, V.Y. Alakhov, E.V.Batrakova, A.S. Lebedev, N.S. Melik-Nubarov, S.A.Arzhakov, A.V. Levashov, G.V. Morozov, E.S. Severin,V.A. Kabanov, FEBS Lett. 258 (1989) 343.

[2] A.V. Kabanov, E.V. Batrakova, N.S. Melik-Nubarov,N.A. Fedoseev, T.Y. Dorodnich, V.Y. Alakhov, V.P.Chekhonin, I.R. Nazarova, A.V. Kabanov, J. Controll.Rel. 22 (1992) 141.

[3] V.Y. Alakhov, E.Y. Moskaleva, E.V. Batrakova, A.V.Kabanov, Bioconjugate Chem. 7 (1996) 209.

[4] D.W. Miller, E.V. Batrakova, T.O. Waltner, V.Y.Alakhov, A.V. Kabanov, Bioconjugate Chem. 8 (1997)649.

[5] E.V. Batrakova, H.Y. Han, V.Y. Alakhov, D.W. Miller,A.V. Kabanov, Pharm. Res. 15 (1998) 850.

[6] V.Y. Alakhov, A.V. Kabanov, Exp. Opin. Invest. Drugs7 (1998) 1453.

[7] A.V. Kabanov, V.A. Kabanov, Advan. Drug DeliveryRev. 30 (1998) 49.

[8] A.V. Kabanov, I.R. Nazarov, I.V. Astafieva, E.V. Ba-trakova, V.Y. Alakhov, A.A. Yaroslavov, V.A. Ka-banov, Macromolecules 28 (1995) 2303.

[9] M. Yokoyama, M. Miyauchi, N. Yamada, T. Okano, Y.Sakurai, K. Kataoka, S. Inoue, Cancer Res. 50 (1990)1693.

[10] M. Yokoyama, G.S. Kwon, T. Okano, Y. Sakurai, M.Naito, K. Kataoka, J. Controlled Release 28 (1994) 59.

[11] G. Kwon, S. Suwa, M. Yokoyama, T. Okano, Y.Sakurai, K. Kataoka, J. Controlled Release 29 (1994)17.

[12] M. Yokoyama, S. Fukushima, R. Uehara, K. Okamoto,

K. Kataoka, Y. Sakurai, T. Okano, J. Controlled Re-lease 50 (1998) 79.

[13] G. Kwon, M. Naito, M. Yokoyama, T. Okano, Y.Sakurai, K. Kataoka, J. Controlled Release 48 (1997)195.

[14] S. Katayose, K. Kataoka, J. Pharm. Sci. 87 (1998) 160.[15] A. Harada, K. Kataoka, Macromolecules 31 (1998) 288.[16] S. Cammas, K. Suzuki, C. Sone, Y. Sakurai, K.

Kataoka, T. Okano, J. Controlled Release 48 (1997) 157.[17] S.B. La, T. Okano, K. Kataoka, J. Pharm. Sci. 85 (1996)

85.[18] G.S. Kwon, T. Okano, Advan. Drug Delivery Rev. 21

(1996) 107.[19] G.S. Kwon, M. Naito, K. Kataoka, M. Yokoyama, Y.

Sakurai, T. Okano, Colloids and Surfaces B: Biointer-faces 2 (1994) 429.

[20] G.S. Kwon, K. Kataoka, Advan. Drug Delivery Rev. 16(1995) 295.

[21] G. Kwon, M. Naito, M. Yokoyama, T. Okano, Y.Sakurai, K. Kataoka, Langmuir 9 (1993) 945.

[22] D. Maysinger, A. Morinville, Trends Biochem. Sci. 15(1997) 410.

[23] R.H. Muller, Colloidal Carriers for Controlled DrugDelivery and Targeting, CRC Press, Boca Raton, FL,1991.

[24] S.A. Hagan, A.G.A. Coombes, M.C. Garnett, S.E.Dunn, M.C. Davies, L. Illum, S.S. Davis, Langmuir 12(1996) 2153.

[25] S.Y. Kim, I.G. Shin, Y.M. Lee, C.S. Cho, Y.K. Sung, J.Controlled Release 51 (1998) 13.

[26] T. Inoue, G. Chen, K. Nakamae, A.S. Hoffman, J.Controlled Release 51 (1998) 221.

[27] Y.I. Jeong, J.B. Cheon, S.H. Kim, J.W. Nah, Y.M. Lee,Y.K. Sung, T. Akaike, C.S. Cho, J. Control. Release 51(1998) 169.

[28] J.W. Nah, Y.I. Jeong, C.S. Cho, J. Polym. Sci.: Part B:Polym. Phys. 36 (1998) 415.

[29] X. Zhang, H.M. Burt, D. Von Hoff, D. Dexter, G.Mangold, D. Degen, A.M. Oktaba, W.L. Hunter, Can-cer Chemother. Pharmacol. 40 (1997) 81.

[30] M. Ramaswamy, X. Zhang, H.M. Burt, K.M.- Wasan,J. Pharm. Sci. 86 (1997) 460.

[31] Y. Yu, C. Allen, S. Chijiwa, D. Maysinger, A. Eisen-berg, Macromolecules, submitted for publication.

[32] C. Allen, Y. Yu, D. Maysinger, A. Eisenberg, Bioconju-gate Chem. 9 (1998) 564.

[33] R. Nagarajan, K. Ganesh, J. Colloid and Interface Sci.184 (1996) 489.

[34] R. Nagarajan, M. Barry, E. Ruckenstein, Langmuir 2(1986) 210.

[35] R. Nagarajan, Current Opinion in Colloid and InterfaceSci. 2 (1997) 282.

[36] R. Nagarajan, K. Ganesh, J. Chem Phys. 90 (1989)5843.

[37] M. Tian, A. Qin, C. Ramireddy, S.E. Webber, P. Munk,Z. Tuzar, K. Prochazka, Langmuir 9 (1993) 1741.


[38] K. Prochaska, D. Kiserow, C. Ramireddy, Z. Tuzar, P.Munk, S.E. Webber, Macromolecules 25 (1992) 454.

[39] M. Tian, E. Arca, Z. Tuzar, S.E. Webber, P. Munk, J.Polymer Science: Part B: Polymer Physics 33 (1995)1713.

[40] Y. Teng, M.E. Morrison, P. Munk, S.E. Webber,Macromolecules 31 (1998) 3578.

[41] M. Pacovska, K. Prochazka, Z. Tuzar, P. Munk, Poly-mer 34 (1993) 4585.

[42] L. Xing, W.L. Mattice, Macromolecules 30 (1997) 1711.[43] Y. Wang, C.M. Kausch, M. Chun, R.P. Quirk, W.L.

Mattice, Macromolecules 28 (1995) 904.[44] T. Haligolu, I. Bahar, B. Erman, W.L. Mattice, Macro-

molecules 29 (1996) 4764.[45] Z. Hruska, M. Piton, A. Yekta, J. Duhamel, M.A.

Winnik, G. Riess, M.D. Croucher, Macromolecules 26(1993) 1825.

[46] M. Wilhelm, C.L. Zhao, Y. Wang, R. Xu, M.A. Winnik,Macromolecules 24 (1991) 1033.

[47] R. Xu, M.A. Winnik, Macromolecules 24 (1991) 87.[48] S. Creutz, J. van Stam, F.C. De Schryver, R. Jerome,

Macromolecules 31 (1998) 681.[49] S. Creutz, J. van Stam, S. Antoun, F.C. De Schryver, R.

Jerome, Macromolecules 30 (1997) 4078.[50] P.G. deGennes, In Solid State Physics, L. Liebert (Ed.),

Academic Press, New York, 1978, supplement 14, p. 1.[51] A. Halperin, Macromolecules 20 (1987) 2943.[52] (a) I. Astafieva, X.F. Zhong, A. Eisenberg, Macro-

molecules 26 (1993) 7339. (b) I. Astafieva, K. Khougaz,A. Eisenberg, Macromolecules (1995) 7127–7134.

[53] (a) P. Alexandridis, J.F. Holzwarth, T.A.- HattonMacromolecules 27 (1994) 2414. (b) P. Alexandridis, V.Athanassiou, S. Fukuda, T.A. Hatton, Langmuir 10(1994) 2604.

[54] F. Gadelle, W.J. Koros, R.S. Schechter, Macromolecules28 (1995) 4883.

[55] A. Halperin, S. Alexander, Macromolecules 22 (1989)2403.

[56] J. Zhao, C. Allen, A. Eisenberg, Macromolecules 30(1997) 7143.

[57] C. Price, in: I. Goodman (Ed.), Developments in blockcopolymers, vol. 1, Applied Science Publishers, London,1982, p. 39.

[58] J. Selb, Y. Gallot, in: I. Goodman (Ed.), Developmentsin block copolymers, Applied Science Publishers, Vol 2,London, 1985, p. 27.

[59] G Riess, G Hurtrez, P. Bahadur, Encyclopedia of Poly-mer Science and Engineering’, Vol 2, 2nd edition, Wiley,New York, 1985, p. 324.

[60] B. Chu, Z. Zhou, In: V.M. (Ed.), Non-ionic Surfactants:Polyoxyalkylene Block Copolymers, Marcel Dekker,New York, 1996, p. 67.

[61] L. Zhang, A. Eisenberg, P.A.T. 9 (1998) 677.[62] L. Zhang, A. Eisenberg, Science 268 (1995) 1728.[63] L. Zhang, K. Yu, A. Eisenberg, Science 272 (1996) 1777.[64] L. Chen, A. Eisenberg, in preparation.[65] L. Zhang, J. Barlow, A. Eisenberg, Macromolecules 28

(1995) 6055.

[66] R. Gref, A. Domb, P. Quellec, T. Blunk, R.H. Muller,J.M. Verbavatz, R. Langer, Advan. Drug Delivery Rev.16 (1995) 215.

[67] R. Gref, Y. Minamitake, M.T. Peracchia, V. Trubeskoy,V. Torchilin, R. Langer, Science 263 (1994) 1600.

[68] M. Tobio, R. Gref, A. Sanchez, R. Langer, M. Alonso,J, Pharm. Res. 15 (1998) 270.

[69] T.M. Allen, Adv. Drug Deliv. Rev. 13 (1994) 285.[70] D.L. Elbert, J.A. Hubbell, Ann. Rev. Mater. Sci. 26

(1996) 365.[71] J.H. Lee, H.B. Lee, J.D. Andrade, Prog. Polym. Sci. 20

(1995) 1043.[72] R.J. Hunter, Foundations of Colloid Science, vol. 1,

Oxford University Press, New York, 1991.[73] G. Gregoriadis, J. Senior, FEBS Lett. 119 (1980) 43.[74] J. Senior, J.C.W. Crawley, G. Gregoriadis, Biochim

Biopyhys. Acta 839 (1985) 1.[75] M.R. Mauk, R.C. Gamble, J.D. Baldeschwieler, Proc.

Natl. Acad. Sci. USA. 77 (1980) 4430.[76] D. Papahadjopoulos, J. Liposome Res. 6 (1996) 3.[77] K.D. Lee, K. Hong, D. Papahadjopoulos, Biochim.

Biophys. Acta 1103 (1992) 185.[78] R. Bajoria, S.R. Sooranna, S.F. Contractor, Human

Reproduction 12 (1997) 1343.[79] H.Y. Yu, C.Y. Lin, J. Formosan Medical Assoc. 96

(1997) 409.[80] R.M. Kuntz, W.M. Saltzman, Trends in Biotechnol. 15

(1997) 364.[81] T. Inoue, G.H. Chen, A.S. Hoffman, K. Nakamae, J.

Bioactive and Compatible Polym. 13 (1998) 50.[82] Y. Akiyama, N. Nagahara, E. Nara, M. Kitano, S.

Iwasa, I. Yamamoto, J. Azuma, Y. Ogawa, J. Pharm.Pharmacol. 50 (1998) 159.

[83] T. Gershanik, S. Benzeno, S. Benita, Pharm. Res. 15(1998) 863.

[84] R. Calvo, C. Remunanlopez, J.L. Vilajato, M.J. Alonso,Colloid and Polym. Sci. 275 (1997) 46.

[85] J.A. Lowrey, N.D.L. Savage, D. Palliser, M. Corsin-jimenez, L.M.G. Forsyth, G. Hall, S. Lindley, G.A.Stewart, K.A.L. Tan, G.F. Hoyne, J.R. Lamb, Int.Arch. Aller. Immunol. 116 (1998) 93.

[86] A.G. Bailey, A.H. Hashish, T.J. Williams, J. Electrostat.44 (1998) 3.

[87] P.N. Hurter, T.A. Hatton, Langmuir 8 (1992) 1291.[88] P.N. Hurter, J.M.H.M. Scheutjens, T.A. Hatton,

Macromolecules 26 (1993) 5592.[89] P.N. Hurter, J.M.H.M. Scheutjens, T.A. Hatton,

Macromolecules 26 (1993) 5030.[90] J.M.G. Cowie, Polymers: Chemistry and Physics of

Modern Materials, 2nd Edition, Blackie Academic andProfessional, New York, 1994.

[91] B.D. Ratner, A.S. Hoffman, F.J. Schoen, J.E. Lemons,Biomaterials Science: An Introduction to Materials inMedicine, Academic Press, New York, 1996.

[92] J.P. Benoit, F. Courteille, C. Thies, Int. J. Pharm. 29(1986) 95.


[93] M. Cavalier, J.P. Benoit, C. Thies, J. Pharm. Pharmacol.38 (1986) 249.

[94] M. Donbrow, In vitro Release from Microcapsules andMicrospheres In Microcapsules and Nanoparticles inMedicine and Pharmacy, CRC Press, Boca Raton, 1992.

[95] A. Desjardins, A. Eisenberg, Macromolecules 24 (1991)5779.

[96] L. Zhang, A. Eisenberg, J. Am. Chem. Soc. 118 (1996)3168.

[97] L. Zhang, A. Eisenberg, Macromolecules 29 (1996) 8805.[98] K. Yu, A. Eisenberg, Macromolecules 29 (1996) 6359.[99] K. Yu, L. Zhang, A. Eisenberg, Langmuir 12 (1996)

5984.[100] Y. Yu, L. Zhang, A. Eisenberg, Langmuir 13 (1997)

2578.[101] L. Zhang, A. Eisenberg, Macromol. Symp. 113 (1997)

221.

[102] Y. Yu, A. Eisenberg, J. Am. Chem. Soc. 119 (1997)8383.

[103] K. Yu, A. Eisenberg, Macromolecules 31 (1998) 3509.[104] K. Yu, C. Bartels, A. Eisenberg, Macromolecules, in

press.[105] L. Zhang, C. Bartels, Y. Yu, H. Shen, A. Eisenberg,

Phys. Rev. Lett. 79 (1997) 5034.[106] L. Zhang, H. Shen, A. Eisenberg, Macromolecules 30

(1997) 1001.[107] J. Massey, K.N. Power, I. Manners, M.A. Winnik, J.

Am. Chem. Soc. 120 (1998) 9533.[108] J. Tao, S. Stewart, G. Liu, M. Yang, Macromolecules 30

(1997) 2738.[109] D.L. Johnson, M. Polikandritou-Lambros, T.B. Mar-

tonen, J. Drug Delivery and Targeting of TherapeuticAgents 3 (1996) 9.

[110] J.F. Ding, G.J. Liu, J. Phys. Chem. B 102 (1998) 6107.

Documents

Nano-engineering block copolymer aggregates for drug delivery