Paclitaxel-loaded composite fibers: Microstructure and emulsion stability

Paclitaxel-loaded composite fibers: Microstructure andemulsion stability

Amir Kraitzer, Meital ZilbermanDepartment of Biomedical Engineering, Faculty of Engineering, Tel-Aviv University, Tel-Aviv 69978, Israel

Received 13 June 2006; revised 16 August 2006; accepted 29 August 2006Published online 21 November 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.31068

Abstract: New core/shell fiber structures loaded withpaclitaxel were developed and studied. These compositefibers are ideal for forming thin, delicate, biomedically im-portant structures for various applications. Possible applica-tions include fiber-based endovascular stents that mechani-cally support blood vessels while delivering drugs for pre-venting restenosis directly to the blood vessel wall, or drugdelivery systems for cancer treatment. The core/shell fiberstructures were formed by ‘‘coating’’ nylon fibers with po-rous paclitaxel-containing poly(DL-lactic-co-glycolic acid)structures. Shell preparation (‘‘coating’’) was performed byfreeze-drying water in oil emulsions. The present studyfocused on the effects of the emulsion’s formulation (com-position) and processing conditions on the porous shellstructure, which actually reflects the emulsion’s stability

and also the drug release profile from the fibers. In general,extremely porous ‘‘shell’’ structures were obtained withgood adhesion to the core fiber. An increase in the emul-sion’s drug content and copolymer composition demon-strated a significant effect on pore size and distribution,because of enhanced emulsion instability, whereas the ho-mogenization rate and duration had only a slight effect onthe pores’ microstructure. The thermodynamic parametersin the studied system are thus more important than the ki-netic parameters in determining the emulsion’s stabilityand the shell’s porous structure. � 2006 Wiley Periodicals,Inc. J Biomed Mater Res 81A: 427–436, 2007

Key words: paclitaxel; composite fibers; controlled drugrelease; poly(DL-lactic-co-glycolic acid); porous structure

INTRODUCTION

Organ or tissue failure or loss is one of the most fre-quent and devastating problems in human healthcare.Principles of biomaterials, engineering, and biology areapplied to the development and study of implantablemedical devices or substitutes for damaged tissues.These may be based on fiber structures and may con-tain bioactive molecules that enhance the healing ofthe surrounding tissues or help cure certain diseases.

Few controlled-release fiber systems based on poly-mers have been investigated to date.1–8 The two basictypes of drug-loaded fibers that have been reported aremonolithic fibers and reservoir fibers. In systems thatuse monolithic fibers, the drug is dissolved or dis-persed throughout the polymer fiber. For example,curcumin, paclitaxel, and dexamethasone have beenmelt-spun with poly(L-lactic acid) (PLLA) to generatedrug-loaded fibers1 and aqueous drugs have been so-

lution-spun with PLLA.2 Various steroid-loaded fibersystems have demonstrated the expected first-orderrelease kinetics.3–5 In systems that use hollow reservoirfibers, drugs such as dexamethasone and methotrexanehave been added to the internal section of the fiber.6–8

The advantages of drug-loaded fibers include ease offabrication, high surface area for controlled release,and localized delivery of bioactive agents to their tar-get. Disadvantages include poor mechanical propertiesdue to drug incorporation and limitations in drugloading. Furthermore, many drugs and all proteins donot tolerate melt processing and organic solvents.

In one of our recent studies, we presented a new con-cept of core/shell fiber structures which successfullymet these challenges.9 These composite fibers combinea dense polymer core fiber and a protein-loaded po-rous shell structure, that is, the drug or protein islocated in a separate compartment (a ‘‘shell’’) around amelt-spun ‘‘core’’ fiber. The shell is prepared usingmild processing conditions. A schematic representationof the composite fiber structure is presented in Figure1(a). This results in good mechanical properties as wellas the desired drug release profile.10 Our new fibersare ideal for forming thin, delicate, biomedically im-portant structures for various applications, such asfiber-based endovascular stents that mechanically sup-port blood vessels while delivering drugs for prevent-

Correspondence to: M. Zilberman; e-mail: [email protected] grant sponsor: Israeli Ministry of Health; con-

tract grant number: 5821Contract grant sponsor: RAMOT (Horowitz) Foundation,

Tel-Aviv University

' 2006 Wiley Periodicals, Inc.

ing restenosis directly to the blood vessel wall, or drugdelivery systems for cancer treatment. Paclitaxel is apotent cell proliferation inhibitor and is known to bevery effective in the treatment of cancer as well as inpreventing restenosis.11,12 It is highly hydrophobic, andits release from polymeric devices in an aqueous me-dium is therefore relatively slow.12 Porous structuresthat encapsulate the drug, such as our fiber’s shell, maytherefore be beneficial in such cases.

Emulsions aremetastable colloidsmade of two immis-cible fluids, one being dispersed in the other in the pres-ence of surface-active agents (surfactants).13 Invertedemulsions are composed of water droplets dispersed ina continuous oil (organic) phase. Emulsions are obtainedby shearing two immiscible fluids, leading to the frag-mentation of one phase into the other. They are metasta-ble, and their life-timemay vary considerably dependingon the temperature and their composition. The instabil-ity is due to the large interfacial area, and therefore alarge surface energy which is associated with finely dis-persed systems.14 The stability of emulsions is highly im-portant for their use as drug delivery systems. Separat-ing a two-phase system produces a large surface areaper drop, leading to a high excess Gibbs energy perdrop, and thus to a tendency in the direction of decreas-ing the Gibbs energy.13,14 The destabilization of a two-phase emulsion goes through several consecutive andparallel steps before reaching the final stage of separatedlayers. Several types of interaction patterns may arisewhen two particles driven by random Brownian motionapproach each other, including flocculation, coalescence,Ostwald ripening, and creaming.13,15,16

The present study focuses on composite core/shellfiber structures loaded with paclitaxel. The porousshell (drug-containing section) was prepared usingthe technique of freeze drying an inverted emulsion.The effects of the emulsion’s formulation (compo-nents) and processing conditions on the shell’s micro-structure were examined. The shell’s porous structureis important, since it may affect the drug release pro-file. It is also a good measure of the emulsion’s stabil-ity. Furthermore, these poly(DL-lactic-co-glycolic acid(PDLGA)-based inverted, freeze-dried emulsions areunique and can be loaded with many bioactive agentsand may serve in a wide variety of biomedical andtissue regeneration applications.

MATERIALS AND METHODS

Materials

EthilonTM monofilament nylon sutures (model W597),Ethicon, USA, were used as core fibers.

Bioresorbable porous structures (the shell coating) weremade of 75/25 poly(DL-lactic-co-glycolic acid) (PDLGA), in-herent viscosity (i.v.) ¼ 0.65 dL/g (in CHCl3 at 308C,�97,100 g/mole), Absorbable Polymer Technologies, USA.

Paclitaxel (GenexolTM) was purchased from Sam YangCorp., Seoul, Korea.

Surface active agents

1. Pluronic L121TM, a triblock copolymer of ethyleneoxide and propylene oxide, with a mean molecular

Figure 1. The structure of the core/shell composite fibers: (a) A schematic representation showing the concept, (b) SEMfractograph of the reference specimen, (c) and (d) SEM micrographs of the surface.

428 KRAITZER AND ZILBERMAN

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

weight of �4400 Da was received as a gift fromBASF, USA.

2. Poly(vinyl alcohol) (PVA), 87–89% hydrolyzed, mo-lecular weight ¼ 13,000–23,000 Da was purchasedfrom Sigma.

Preparation of core/shell fiber structures

Fiber surface treatment

The sutures were surface-treated in order to dispose ofthe original fiber’s coating and to enhance the adhesionbetween the core fiber and the coating. The nylon fiberswere slightly stretched on special holders and dipped in a75/25 v/v formic acid/ethanol solution for 15 s. The fiberswere then washed and dried in a vacuum oven at 658C for80 min.

Emulsion formation

A known amount of PDLGA was dissolved in chloro-form to form an organic solution, and paclitaxel was addedto the solution. Double-distilled water was then poured intothe organic phase (in a test tube) and homogenization ofthe emulsion was performed using a hand-held homoge-nizer (OMNI TH, 7-mm rotor) operating at 16,500 rpm (me-dium rate) for 3 min, for most investigated samples. Inorder to investigate the effect of processing conditions onthe porous shell structure, certain samples were preparedusing homogenization rates of 5500 rpm (low rate) or 25,000rpm (high rate), and homogenization durations of 1 and4 min. As a reference sample, we chose an emulsion formu-lation containing 17.5% w/v polymer in the organic solu-tion, 1.43% w/w paclitaxel (relative to the polymer load),and an organic to aqueous (O:A) phase ratio of 2:1 v/v. Allother formulations are presented in Table I. Surface activeagents were added to the emulsion in some of the samples:pluronic (1% w/w relative to the polymer quantity) wasadded to the polymer solution and PVA (1% w/v relativeto the water quantity) was added to the water.

Core/shell fiber structure formation

The treated core nylon fibers were dip-coated (whileplaced on holders) in fresh emulsions and then frozen im-mediately in a liquid nitrogen bath. The holders þ sampleswere then placed in a precooled (�1058C) freeze drier(Virtis 101 equipped with a nitrogen trap) capable of work-ing with organic solvents (freezing temperature of the con-denser was ��1058C), and freeze dried in order to preservethe microstructure of the emulsion-based core/shell fiberstructures. That is, this process of freeze drying enablessublimation of water from the aqueous phase and solventfrom the organic phase, leaving solid polymeric structurewith pores inside.

Drying was performed in two stages:

1. The freeze drier chamber pressure was reduced to 100mTorr, while the temperature remained at �1058C.

2. The condenser was turned off and its plate tempera-ture slowly increased to room temperature, while thepressure was monitored between 100 and 700 mTorr.During this step, the liquid nitrogen trap condensedthe excess water and solvent vapors.

The samples were stored in desiccators until use.

Morphological characterization

The morphology of the composite core/shell fiber struc-tures (cryogenically fractured surfaces) was observed usinga Jeol JSM-6300 scanning electron microscope (SEM) at anaccelerating voltage of 5 kV. The SEM samples were Au-sputtered prior to observation. The structure of the shell(coating) surface was also observed for certain samples. Themean pore diameter of the observed morphologies was ana-lyzed using Sigma Scan Pro software, and statistics weredrawn using SPSS. Statistical significance was determinedusing the ANOVA (Tukey-Kramer) method. The effects ofthe emulsion’s composition and processing parameters onthe microstructure were studied by examining the follow-ing parameters:

1. Paclitaxel content (% w/w, measured relative to thepolymer weight).

TABLE IThe Examined Specimens and Their Shell’s

Structural Characteristics

AmountMean PoreSize (mm) Porositya

CoatingThickness (mm)

Polymer content (% w/v)15 5.8 6 2.3 85 27.7 6 3.617.5 6.5 6 2.3 85.2 104 6 31.422.5 5.4 6 2.1 82 64.2 6 32.4

Paclitaxel content (% w/w)0 6.9 6 1.9 N/A 42.2 6 30.71 5.4 6 2.6 89 74.2 6 9.91.43 6.5 6 2.3 85.2 104 6 31.42.86 21.2 6 6 85 81 6 37.77.14 79.1 6 17 N/A 192.8 6 90.7

Organic to aqueous phase ratio (v/v)4:1 6.1 6 3.1 87.6 52.3 6 12.52:1 6.5 6 2.3 85.2 104 6 31.41.3:1 7.8 6 3.8 94.2 64.6 6 24.1

Surfactant content (1% w/v)None 6.5 6 2.3 85.2 104 6 31.4Pluronic 8.2 6 3.0 88 204.1 6 129.3PVA 6.2 6 2.8 87.5 77.5 6 24.7

Homogenization duration (s)60 7 6 3.7 86.8 23.8 6 1.3180 6.5 6 2.3 85.2 104 6 31.4240 5.9 6 2.6 81.6 90.2 6 44.7

Homogenization rate (rpm)5500 7.7 6 3.5 92.7 114.6 6 33.216,500 6.5 6 2.3 85.2 104 6 31.425,000 5.8 6 1.9 86 65.7 6 20.7

aThe measurement error of the porosity is 10%.

PACLITAXEL-LOADED COMPOSITE FIBERS 429


2. Polymer content (% w/v, measured relative to the sol-vent volume).

3. Aqueous to organic phase ratio (v/v).4. PDLGA copolymeric ratio.5. Addition of surface active agents.6. Duration and rate of homogenization.

Microstructure characterization included the followingparameters:

1. Mean pore diameter and distribution.2. Porosity and pore structure.3. Interconnection between the pores.4. Coating thickness and adhesion quality.

In vitro paclitaxel release studies

Four samples of a chosen composite core/shell fiber wereimmersed in PBS at 378C for 112 days. The medium was(completely) removed periodically and fresh medium wasintroduced. The paclitaxel content of each medium samplewas determined using Agilent 1100 high performance liquidchromatography, as described in detail elsewhere.10

RESULTS AND DISCUSSION

A SEM fractograph showing the bulk morphologyof the reference specimen is presented in Figure 1(b).The quality of the interface between the fiber and theporous coating is high, that is, the surface treatmentenabled good adhesion between core and shell. Theshell’s porous structure contains round-shaped poresof 6.5 6 2.3 mm, with a porosity that exceeds 85%(Table I). It can be noted that the shell’s microstruc-ture is uniform, probably due to the rapid quenchingof the emulsion which enabled the preservation of itsmicrostructure. The surface of the coating is homoge-nously and continuously porous [Fig. 1(c)]. A pore hi-erarchy is observed, in which the primary structuredemonstrates 20–30 mm regions composed of 5-mmpores [Fig. 1(d)]. This may be explained by the factthat the surface is more reactive than the bulk, thusproducing a strong tendency to coalesce. The smallpores may indicate that a finer surface structure couldhave been formed if the surface was more stable. Cu-mulative release of paclitaxel from a selected speci-men for 4 months is presented in Figure 2. It can beseen that paclitaxel is released in an exponential man-ner, that is, the rate decreases with time. Such arelease profile is typical of diffusion-controlled sys-tems. A minor burst effect of less than 3% wasobtained during the first day of release.

The shell’s microstructure may affect the drugrelease profile and can also serve as a good measureof the emulsion’s stability. The microstructure isdetermined by both thermodynamic and kinetic pa-

rameters. Thermodynamic stabilization is governedby the emulsion’s formulation, that is by the polymerand drug contents, the O:A ratio, surfactants and co-polymer composition, whereas the kinetic considera-tions are actually the processing conditions, whichinclude the duration and rate of homogenization. Theeffects of these thermodynamic and kinetic parame-ters on the shell’s microstructure and the emulsion’sstability are described later in detail.

Polymer content

A relatively narrow polymer content range of 15–22.5% w/v was used in this study. Polymer contentsbelow 15% w/v produce a nonhomogenous structurereflected by a partially porous structure, because ofthe emulsion’s instability, whereas early phase sepa-ration of the emulsion occurs at polymer contentsabove 22.5% w/v, since the organic phase is rapidlyrejected by the water. An optimal polymer contentrange of 15–22.5% w/v was therefore used, in whicha relatively stable and homogenous porous structurewas achieved. The effects of polymer content on themorphological characteristics of the shell in this rangeare presented in Figure 3 and Table I. The polymerexhibited only a minor effect on the shell’s porosityand mean pore size. In contradistinction, the coatingthickness was extremely thin (27.7 mm) when the poly-mer content was 15% w/v and increased to 100 mmwhen the polymer content was increased to 17.5%w/v. The interface quality also improved with the in-crease in polymer content, probably due to frictionforces that were enhanced by the increased viscosity.

According to the theory of emulsion stability, poly-mer content affects an emulsion’s stability either byincreasing its viscosity or by promoting surface stabil-ity.13 The viscosity of the organic phase is exponen-tially dependent on the density of the organic polymersolution. Higher viscosity reduces the tendency ofdroplets to move up, which leads to breaking of theemulsion. In addition, higher viscosity reduces the dif-

Figure 2. Paclitaxel release profile from a selected core/shell fiber specimen.



fusion of water through the organic phase, and there-fore the rate of larger drops creation on the expanse ofsmaller drops is reduced (Ostwald ripening).13–16 Thus,high viscosities produce structures with a smaller meanpore diameter. Furthermore, the polymer can act as asurfactant, stabilizing the emulsion by adsorbing to thesurface of the droplets and creating a steric interfer-ence.16 Therefore, increasing either the polymer contentor its chain length may reduce flocculation and coales-cence. Polymer content would thus be expected toaffect the shell’s pore size. However, in the presentstudy, the viscosity probably does not change so muchdue to the chosen specific range of polymer contents,and the fact that the PDLGA stabilizing effect is muchlower than that of a functional surfactant due to lack ofspecific binding regions (such as those found in blockcopolymers).

Drug content

The effects of drug content on the shell’s morphologi-cal characteristics are presented in Figure 4 and Table I.The mean pore diameter increased significantly from�6 mm for samples with a drug content of 0–1.43%, to79 mm for samples with 7.14% paclitaxel. Since pacli-taxel is a hydrophobic drug, its presence in the emul-sion’s organic phase changes the emulsion’s hydropho-bic–hydrophilic balance. Furthermore, specific interac-tions such as hydrogen bonds can form between thedrug and the host polymer. Higher paclitaxel contentsthus decrease the surface tension of the organic phase,

that is, increase the interfacial tension between thewater and the organic phase, thus resulting in loweremulsion stability. The dispersed droplets tend to eitheraggregate irreversibly or accumulate at an interface inorder to reduce the interfacial energy. Larger pores aretherefore created with a higher paclitaxel content.

The emulsion achieved using a high drug contentwas more viscous and exhibited better adhesion tothe core fiber. Consequently, the coating thicknessgreatly increased with drug load. However, the poreslost their round shape at drug contents that exceeded1.43% w/w and were distorted or collapsed.

Organic:Aqueous (O:A) phase ratio

The effects of the O:A phase ratio on the coating’sstructural characteristics are presented in Figure 5 andTable I. At the microstructure level, the size of thepores, the pore distribution, and the porosity pre-sented a trend of an increase with the increase in theaqueous phase content. The increase in pore sizebetween 4:1 samples and 2:1 samples is insignificant,while a significant increase in pore size and distribu-tion is achieved between 4:1 and 1.3:1 samples. Itshould be mentioned that relatively narrow range ofO:A values (2:1 and 4:1) are allowed in this system, inorder to achieve stable emulsions.

Theoretically, the overall pore volume as reflectedby the size and area of the pores is expected to de-crease with the increase in the O:A ratio. Thus, lowerporosity is obtained when using a low water content,

Figure 3. SEM fractographs of composite fibers showing the effect of polymer content on the shell’s microstructure. Poly-mer contents used: (a) 15% w/v, (b) 17.5% w/v, (c) 22.5% w/v, (d) Pore diameter distribution of samples with: u 15% w/vpolymer, n 17% w/v polymer, 22.5% w/v polymer.



and as a result the wall thickness (between pores)increases. The large difference in the surface tensionsof the two phases (enhanced by paclitaxel moleculesin our formulation) did not allow the creation of highO:A ratios emulsions, and therefore this parameterdid not have a significant effect on the microstructure.Whang et al.17,18 also investigated inverted PDLGAemulsions, and their results correlated with theoreticalexpectations when using proteins as a release agent.Furthermore, one of our previous studies9 showedthat using higher O:A ratios, such as 8:1 v/v and 16:1v/v, significantly decreased the mean pore diameter.In that research the released agent was a relativelyhigh molecular weight protein that acted as a surfac-tant with a stabilizing effect, and high O:A ratios were

beneficial for the release profile of the water-solubleprotein. However, in the present study, high O:A val-ues are not required, since low porosity will reducethe release rate of the hydrophobic paclitaxel. If highO:A ratios were required, surfactants could be addedin order to stabilize the emulsion.

A specimen of 1.3:1 v/v was fabricated and charac-terized in order to measure the maximal possiblequantity of the aqueous phase, and consequentlydefine the maximum porosity. SEM observationsindicated a maximal porosity of 94% and a nonho-mogenous coating with a very large pore size distri-bution, reflecting low emulsion stability. Hence, inthis specific system, O:A ratios lower than 2:1 cannotcreate a stable inverted emulsion.

Figure 4. SEM fractographs of composite fibers showing the effect of paclitaxel content on the shell’s microstructure. Pacli-taxel contents used: (a) 0.71% w/w, (b) 1.43% w/w, (c) 2.86% w/w, (d) 7.14% w/w. (e) Pore diameter distribution: u 0.71%w/w paclitaxel, n 1.43% w/w paclitaxel, 2.86% w/w paclitaxel, 7.14 % w/w paclitaxel.



Incorporation of surfactants

Surfactants were incorporated in order to studytheir effect on the emulsion’s stability. It was assumedthat a decrease in mean pore diameter would be ob-tained with the addition of surfactants. Therefore,Pluronic L121 and PVA were chosen based on previ-ous studies that included surfactants in paclitaxel-con-taining PDLGA emulsions.19–23

The effects of Pluronic on the shell’s structure arepresented in Figure 6 and Table I. It can be seen thatthe morphology changed completely and exhibitedvery dense pore populations with a small mean porediameter, surrounded by very large voids where theporosity remained constant. The microstructure givesthe impression that emulsion aggregates are formed,rather than the familiar continuous emulsion structure.Water was probably rejected from the dense porous

regions and formed the secondary structure with largevoids between areas of primary structure. Further-more, the coating is relatively thick, and better adhe-sion between core and shell is achieved.

The effects of PVA as a surfactant on the shell’smicrostructure are presented in Table I. The emulsionstability and mean pore diameter were hardlyaffected by the addition of 0.5–10% w/v PVA in theaqueous phase. Higher PVA concentrations resultedin a direct emulsion in the form of microspheresrather than an inverted emulsion.

According to Bancroft’s rule, a direct emulsion istypically obtained with a water-soluble surfactant,whereas an inverted emulsion is more easily obtainedwith an oil-soluble surfactant.13 PVA is soluble inwater, whereas Pluronic L121 is soluble in oil becauseof its low hydrophilic lipophilic balance (HLB < 7).PVA therefore did not stabilize the ‘‘water in oil’’

Figure 5. SEM fractographs of composite fibers showing the effect of the O:A phase ratio on the shell’s microstructure. O:A(v/v) used: (a) 4:1, (b) 2:1, (c) 1.3:1. (d) Pore diameter distribution of samples with O:A values of:u 1:4 v/v, n 1:2 v/v, 3:4 v/v.

Figure 6. SEM fractographs of core/shell fiber structures showing the effect of Pluronic (as surfactant) on the shell’s micro-structure of the standard sample: (a) 0.5% w/w Pluronic, (b) no Pluronic.



inverted emulsion. Another explanation is in the spa-tial ring structure of the large hydrophilic vinyl alco-hol chain in PVA, which favors an o/w system ratherthan a w/o system.20 In contradistinction, PluronicL121’s long hydrophobic poly(propylene oxide)(PPO) segments anchor themselves in the organicphase, whereas the hydrophilic poly(ethylene oxide)segments (10% w/w) extend into the aqueous me-dium producing a curvature favoring a w/o emul-sion, stabilizing the emulsion and reducing the meanpore diameter. Unfortunately, Pluronic’s hydrophobicPPO blocks interact with the PDLGA clews to formPDLGA aggregates, resulting in early phase separa-tion between the aqueous and organic phases.19,23

Therefore, in our study, the aggregated PDLGA clewsresulted in the formation of large aqueous-richdomains, which transform into large voids followingfreeze drying. Consequently, the achieved pore struc-ture is uniform in certain regions of very small meanpore diameter surrounded by large voids. A less

hydrophobic higher molecular weight surfactantshould be selected in order to eliminate these voidsand further reduce the size of the pores.

Copolymer composition

The effect of the copolymer composition, that is, therelative quantities of lactic acid (LA) and glycolic acid(GA) in the copolymer, was studied on films. Thesethin films were prepared using the same emulsionpreparation technique and were then cast into smallaluminum molds and freeze-dried. The resulting struc-tures are presented in Figure 7. The mean pore diame-ter increased with the increase in LA content in the co-polymer. For example, copolymer compositions of 50/50, 75/25, and 100/0 exhibit pore diameters of 1–2 mm,3–7 mm, and 20–40 mm, respectively. An increase in thecopolymer’s LA content actually increases the surfacetension, prompting the water droplets toward floccula-

Figure 7. SEM fractographs showing the effect of the copolymer composition on the porous structure of PDLGA, with anLA/GA composition (w/w) of: (a) 50/50, (b) 65/35, (c) 75/25, (d) 85/15, (e) 100/0. All polymers had an initial i.v. of0.65 dL/g.



tion and coalescence, and therefore increasing themean pore diameter in the same manner as addition ofpaclitaxel affects the emulsion.

Processing conditions

The effects of the duration and rate of homogeniza-tion on the structural characteristics of the porous shellare presented in Table I. The mean pore diameter andits distribution slightly decreased with the increase inhomogenizing time from 60 to 240 s. Thus, a longer du-ration of homogenization produces a more homoge-nous pore structure with a smaller mean pore diameter,which occurs because more droplets are fragmentedand refragmented. The homogenization rate also slightlyaffects the mean pore diameter and distribution, bothof which decreased at higher stirring rates (p < 0.05between all specimens). Higher fragmentation energyincreases the shearing rate producing finely dispersedwater droplets, whereas energy is continuously lostdue to friction between the viscous emulsion and thehomogenizer blades. These results are in agreementwith those of Bibette et al.13 and Liu and McGrath,24

who investigated ‘‘organic in aqueous’’ emulsions. Incan be concluded that in our systems the effect of theemulsion formulation (drug and polymer contents,O:A ratio and surfactants) on the resulting microstruc-ture is more important than the processing conditions.

Control of the emulsion’s stability

The emulsion’s stability has a significant effect on thescaffold’s microstructure. Emulsion destruction mecha-nisms, such as creaming, flocculation, Ostwald ripen-ing, and coalescence, result in an increased pore diame-ter as well as in poor dispersion of the pores. Asreflected in our microstructure characterization, themean pore diameter can be decreased using threeapproaches: increasing the viscosity of the emulsion,promoting the surface stability of the water droplets,and increasing the mechanical fragmentation. Thesemethods were applied separately or in combinationand resulted in an overall trend of reducing the meanpore diameter of the fiber’s shell while maintaininghigh porosity levels.

Emulsion viscosity

The emulsion’s viscosity is determined by the i.v. ofthe host polymer, the polymer content of the organicphase, and the O:A phase ratio. The organic phase vis-cosity is exponentially dependent on the density of thepolymer solution, which is determined by the i.v. aswell as by the polymer concentration within the or-ganic phase. When viscosity increases over a certain

threshold, the velocity of a water droplet is reduced,reducing the creaming rate and consequently promot-ing stability. High viscosities also reduce the Ostwaldripening rate due to slow diffusion.13 Consequently,high viscosities increase the emulsion’s stability andreduce the obtained microstructure’s pore diameter.Higher viscosity increases the friction forces betweenthe fiber and the coating, and therefore results in bet-ter adhesion between core and shell.

Surface stability

Controlling the rate of flocculation and coalescencemay be achieved by introducing steric forces betweenthe water droplets and the oil (polymeric) medium.Ionic surfactants are not effective within the organic me-dium due to the low dielectric constant of the organicsolvents. Thus, steric stabilization is the key approach inthis system.13 Steric stabilization may be achieved usinga low HLB surfactant (i.e., Pluronic L121) or by increas-ing the polymer content or molecular weight.25 Anincrease in polymer content or surfactant incorporationcauses the polymer to bind at the interface between thewater and oil phases and stabilizes the emulsion. Sur-face stability is also obtained with low paclitaxel loadsand a lower LA content of the copolymer composition.We have shown that both the drug content and the co-polymer composition affect the emulsion’s stability inthe same manner. The mean pore diameter increasedsignificantly with the increase in the polymer’s drugcontent or LA content. This may be explained by thefact that both paclitaxel and LA are hydrophobic. Add-ing these hydrophobic materials to the emulsion in-creases the interfacial tension. The dispersed phase(water) tends to reduce the surface tension by aggrega-tion, which is caused by flocculation and coalescence.

Mechanical fragmentation

Homogenizing energy (i.e., homogenization rate orduration) causes fragmentation, which reduces themean pore diameter and distribution. However, themicrostructure is less affected by homogenizationenergy than by the emulsion’s formulation. The homog-enization rate may exhibit a greater effect on the micro-structure if a more powerful homogenizer is used.

It is not always clear which instability mechanismdominates the system. This fact emphasizes the com-plexity of our emulsion-based system.

SUMMARY AND CONCLUSIONS

New bioresorbable core/shell fiber structures forbiomedical and tissue regeneration applications were



developed and studied. These structures were com-posed of a nylon core and a porous PDLGA shellloaded with the antiproliferative agent, paclitaxel, pre-pared using freeze drying of inverted emulsions. Sincethe drug was loaded only in the porous shell, thesenew fibers are designed to combine good mechanicalproperties with a versatile drug release profile. Investi-gation of the composite fibers focused on the effects ofthe emulsion’s composition (formulation) and process-ing conditions on the porous shell structure, whichreflects the emulsion’s stability and also may affect thedrug release profile from the fibers.

In general, extremely porous ‘‘shell’’ structures (meanporosity of �85% and mean pore size of 6 mm) wereobtained with good adhesion to the core fiber. The fol-lowing emulsion parameters were chosen in order toobtain a stable emulsion that will result in a homogene-ous porous shell structure and a feasible release profileof the water-insoluble drug: 15–22.5% w/v polymercontent in the organic phase, 0.71–2.86% w/w pacli-taxel (relative to the polymer), O:A phase ratio inthe range of 2:1–4:1, homogenization rates of 5500–25,000 rpm and homogenization durations of 60–240 s.

The emulsion’s drug content and copolymer com-position exhibited a significant effect on the shell’sstructure. An increase in drug content or in the LAcontent of the PDLGA copolymer resulted in anincrease in pore diameter, mainly due to the morehydrophobic nature of the organic phase. Thisincreased the difference in interfacial surface tensionbetween the two phases and therefore enhanced theemulsion’s instability. A decrease in the O:A ratioresulted in some increase in pore size and porosity.An increase in homogenization rate and durationresulted in a small decrease in porosity and distribu-tion. Thus, in the studied system, the thermodynamicparameters are more important than the kinetic pa-rameters in determining the emulsion’s stability andthe shell’s porous structure.

References

1. Su SH, Landau CL, Chao RY, Timmons RB, Meidell RS, Tang L,Eberhart RC. Expandable bioresorbable endovascular stent withanti-platelet and anti-inflammation treatments. Circulation 2001;104:500–507.

2. Alikacem N, Yoshizawa T, Wilson C, Nelson KD. QuantitativeMR Imaging study of intravitreal sustained release of VEGF inrabbits. Invest Ophthalmol Vis Sci 2000;41:1561–1569.

3. Dunn RL, Lewis DH, Goodson JM. Monolithic fibers for con-trolled delivery of tetracycline. Proc Int Symp Control Rel BioactMater 1982;9:157–163.

4. Dunn RL, English JP, Stoner WC, Potter AG, Perkins BH. Bio-degradable fibers for the controlled release of tetracycline in

treatment of peridontal disease. Proc Int Symp Control RelBioact Mater 1987;14:289–294.

5. Dunn RL, Lewis DH, Beck LR. Fibrous polymer for the deliv-ery of contraceptive steroids to the female reproductive track.In: Lewis DH, editor. Controlled Release of Pesticides andPharmaceuticals. New York: Plenum; 1981. pp 125–146.

6. Eenink MDJ, Feijen J, Oligslanger J, Albers JHM, Rieke JC,Greidonus PJ. Biodegradable hollow fibers for the controlledrelease of hormones. J Control Release 1987;6:225–237.

7. Polacco G, Cascone MG, Lazzeri L, Ferrara S, Giusti P. Bio-degradable hollow fibers containing drug-loaded nanopar-ticles as controlled release systems. Polym Int 2002;51:1464–1472.

8. Lazzeri L, Cascone MG, Quiriconi S, Morabito L, Giusti P. Bio-degradable hollow microfibers to produce bioactive scaffolds.Polym Int 2005;54:101–107.

9. Levy Y, Zilberman M. Novel bioresorbable composite fiberstructures loaded with proteins for tissue regeneration appli-cations. J Biomed Mater Res A. Forthcoming.

10. Zilberman M, Kraitzer A. Paclitaxel-loaded composite fibers:Drug release and mechanical properties. 2006. Submitted forpublication.

11. Silber S, Grube E. The Boston scientific antipropliferativepaclitaxel eluting stent (TAXUS). In: Serruys, PW editor.Handbook of Coronary Stents. London: Martin Dunitz; 2001.pp 311–319.

12. Dhanikula AB, Panchagnula R. Localized paclitaxel delivery.Int J Pharm 1999;183:85–100.

13. Bibette J, Calderon FL, Poulin P. Emulsions: Basic principles.Rep Prog Phys 1999;62:969–1033.

14. Bencher P. Encyclopedia of Emulsion Technology, Vol. 3. NewYork: Marcel Dekker; 1988.

15. Sjoblom J, editor. Emulsion and Emulsion Stability. New York:Marcel Dekker; 1996.

16. Washington C. Stability of lipid emulsions for drug delivery.Adv Drug Deliv Rev 1996;20:131–140.

17. Whang K, Thomas CH, Healy KE, Nuber G. A novel methodto fabricate bioabsorbable scaffolds. Polymer 1995;36:837–842.

18. Whang K, Goldstick TK, Healy KE. A biodegradable polymerscaffold for delivery of osteotropic factors. Biomaterials2000;21:2545–2551.

19. DesNoyer JR, McHugh AJ. The effect of Pluronic on the pro-tein release kinetics of an injectable drug delivery system.J Control Release 2003;86:15–24.

20. Feng SS, Huang G. Effects of emulsifiers on the controlledrelease of paclitaxel (Taxol) from nanospheres of biodegrad-able polymers. J Control Release 2001;71:53–69.

21. Mu L, Feng SS. PLGA/TPGS nanoparticles for controlledrelease formulation of Paclitaxel: Effects of the emulsifier anddrug loading ratio. Pharm Res 2003;20:1864–1872.

22. Mu L, Feng SS, Kiang TW. Comparison of Paclitaxel loadedPLGA nanospheres fabricated by the single emulsion-freezedry technique under various preparation conditions. Proc IntSymp Control Rel Bioact Mater 2001;28:23–27.

23. Zhang R, Weng W, Du P, Zhao G, Shen G, Han G. Effect ofPluronic F127 on the pore structure of macrocellular biode-gradable polylactic foams. Polym Adv Technol 2004;15:425–430.

24. Liu EH, McGrath KM. Emulsion microstructure and energyinput, roles in emulsion stability. Colloids Surf A 2005;262:101–112.

25. Hiemenz PC. Principles of Colloid and Surface Chemistry.New York: Marcel Dekker; 1986.



Documents

Paclitaxel-loaded composite fibers: Microstructure and emulsion stability