Journal of Controlled Release 146 (2010) 363–369

Contents lists available at ScienceDirect

Journal of Controlled Release

j ourna l homepage: www.e lsev ie r.com/ locate / jconre l

Novel antibacterial nanofibrous PLLA scaffolds

Kai Feng a, Hongli Sun b, Mark A. Bradley c, Ellen J. Dupler c, William V. Giannobile d,e, Peter X. Ma a,b,d,⁎a Macromolecular Science and Engineering Center, University of Michigan, Ann Arbor, MI 48109, United Statesb Department of Biologic and Materials Sciences, University of Michigan, Ann Arbor, MI 48109-1078, United Statesc Undergraduate Research Opportunity Program, University of Michigan, Ann Arbor, MI 48109, United Statesd Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI 48109, United Statese Department of Periodontics and Oral Medicine, School of Dentistry, University of Michigan, Ann Arbor, MI 48109-1078, United States

⁎ Corresponding author. Department of Biologic and MUniversity Ave., Room 2211, The University of MichigaUnited States. Tel.: +1 734 764 2209; fax: +1 734 647

E-mail address: mapx@umich.edu (P.X. Ma).

0168-3659/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.jconrel.2010.05.035

a b s t r a c t

a r t i c l e i n f o

Article history:Received 11 February 2010Accepted 31 May 2010Available online 4 June 2010

Keywords:DoxycyclineControlled deliveryNanofiberScaffoldTissue engineering

In order to achieve high local bioactivity and low systemic side effects of antibiotics in the treatment ofdental, periodontal and bone infections, a localized and temporally controlled delivery system is crucial. Inthis study, a three-dimensional (3-D) porous tissue engineering scaffold was developed with the ability torelease antibiotics in a controlled fashion for long-term inhibition of bacterial growth. The highly solubleantibiotic drug, doxycycline (DOXY), was successfully incorporated into PLGA nanospheres using a modifiedwater-in-oil-in-oil (w/o/o) emulsion method. The PLGA nanospheres (NS) were then incorporated intoprefabricated nanofibrous PLLA scaffolds with a well interconnected macro-porous structure. The releasekinetics of DOXY from four different PLGA NS formulations on a PLLA scaffold was investigated. DOXY couldbe released from the NS-scaffolds in a locally and temporally controlled manner. The DOXY release iscontrolled by DOXY diffusion out of the NS and is strongly dependent upon the physical and chemicalproperties of the PLGA. While PLGA50-6.5K, PLGA50-64K, and PLGA75-113K NS-scaffolds discharge DOXYrapidly with a high initial burst release, PLGA85-142K NS-scaffold can extend the release of DOXY to longerthan 6 weeks with a low initial burst release. Compared to NS alone, the NS incorporated on a 3-D scaffoldhad significantly reduced the initial burst release. In vitro antibacterial tests of PLGA85 NS-scaffolddemonstrated its ability to inhibit common bacterial growth (S. aureus and E. coli) for a prolonged duration.The successful incorporation of DOXY onto 3-D scaffolds and its controlled release from scaffolds extends theusage of nano-fibrous scaffolds from the delivery of large molecules such as growth factors to the delivery ofsmall hydrophilic drugs, allowing for a broader application and a more complex tissue engineering strategy.

aterials Sciences, 1011 Northn, Ann Arbor, MI 48109-1078,2110.

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Chronic inflammation diseases caused by bacterial infection suchas periodontitis and septic arthritis remain common in the UnitedStates and throughout the world. For example, approximately half ofthe adult population in the US suffers chronic periodontitis. Thisdisease is initiated by pathogenic infection on the tooth's surface as amicrobial biofilm which gradually leads to the destruction of theperiodontal tissue, supporting alveolar bone, and eventually leadingto tooth loss [1]. In the treatment of periodontitis, antibiotics such astetracycline (TCL) and metronidazole have been frequently used.

Doxycycline (DOXY) is a well known broad-spectrum antibiotic,which is effective against both Gram-positive and Gram-negativebacteria, protozoa, and various anaerobes [2]. As a tetracyclineanalogue, it can work as a bacteriostatic which is capable of inhibiting

the bacterial protein synthesis at the ribosomal sites [3]. Compared toTCL, DOXY has certain advantages such as a longer half-life, higherlipid solubility, and greater oral absorption [4]. It also shows stronginhibition of matrix metalloproteinases (MMPs) both in vitro and inclinical studies [5–9]. After being introduced to the clinical area in1967 [10], DOXY has been frequently used in treating destructiveperiodontal diseases such as juvenile periodontitis and acuteperiodontal abscesses. It has been used against periodontal infectionand enhancing bone regeneration after periodontal disease [11]. It hasalso been used to prevent bacterial infection related to septic arthritis[10,12]. However, there are some concerns over possible side effectssuch as gastro-intestinal disturbance, esophageal erosion, andphotosensitivity when administrated orally [13–18]. Therefore, inorder to reach the infection deep inside tissue with an effective drugconcentration and to circumvent the systemic side effects, controlledlocal delivery of DOXY is desired.

Many attempts have been made in the area of controlled deliveryof the DOXY/TCL family and certain progress has been achieved. Drugencapsulation methods such as mixing, emulsification, and electro-spinning have been used. Delivery systems in the forms of particles,

364 K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

stripes, hollow fibers, gels, and cements have been fabricated [19–25].However, most systems release all the encapsulated DOXY from hoursup to 2 weeks. While a high initial burst release is acceptable for ahighly hydrophilic antibiotic such as DOXY/TCL in order to immedi-ately achieve effective concentration, it is also necessary to attain along-term sustained release to fight against chronic infection andavoid repeated local drug administration. How to achieve bothreduced initial burst and extended release duration remains achallenge.

Controlled release of proteins has been widely investigated [26–31]. Our laboratory has previously developed a novel growth factordelivery system. The growth factors were encapsulated into PLGAnanospheres and the nanospheres were then incorporated into three-dimensional (3-D) macro-porous and nano-fibrous PLLA scaffolds[32–34]. The function of this delivery system is two fold: first, it servesas an osteoconductive scaffold. With its suitable pore structure andnano-fibrous architecture similar to collagen, it improves cellattachment, proliferation and differentiation [35,36]. Second, itstimulates vascularization or bone formation by localized controlleddelivery of a bioactive angiogenic or osteogenic growth factor [32,33].In this paper, we incorporated DOXY containing PLGA nanospheresinto the nano-fibrous scaffolds to evaluate their antibacterial capacity.The combination of the osteoconductive properties of the macro-porous and the nano-fibrous scaffold with its ability to releaseantibiotics could be a good candidate for local treatment of oral andperiodontal diseases.

2. Materials and methods

2.1. Materials

Doxycycline hyclate, D-fructose, mineral oil, and sorbitanmo-nooleate (span 80) were obtained from the Sigma-Aldrich ChemicalCompany (St. Louis, MO). PLGA copolymers with LA/GA ratio of 50:50(Lakeshore Biomaterials™, PLGA50-6.5K, Mw=6.5 kDa; PLGA50-64K, Mw=64KDa); PLGA copolymers with a LA/GA ratio of 75:25(Lakeshore Biomaterials™, PLGA75-113K, Mw=113 kDa) and thePLGA copolymers with a LA/GA ratio of 85:15 (Lakeshore Biomater-ials™, PLGA85-142K, Mw=142 kDa) were purchased from SurMo-dics Pharmaceuticals Inc. (Birmingham, AL). Poly(L-lactic acid) (PLLA)with inherent viscosity of 1.6 dl/g was purchased from BoehringerIngelheim (Ingelheim, Germany). Poly(acrylic acid) (MW=5000) andpoly(vinyl alcohol) (PVA) (88 mol% hydrolyzed, MW=25,000) wereobtained from Polysciences Inc. Hexaglyceryl polyricinolate (Hex-aglyn PR-15) was provided by the Barnet Products Corp (EnglewoodCliffs, NJ). Acetonitrile and petroleum ether (BP 60-95 °C) wereobtained from Acros Organic (Morris Plains, NJ). Dichloromethane(DCM), cyclohexane, hexane, tetrahydrofuran (THF) and Fisherbrandregenerated cellulose dialysis bags (nominal MWCO 3500) werepurchased from Fisher Scientific (Pittsburgh, PA).

2.2. DOXY containing PLGA nanospheres (NS) and NS-incorporatednano-fibrous scaffold (NS-scaffold) preparation

Doxycycline-containing PLGA NS were fabricated using a modifiedwater-in-oil-in-oil (w/o/o) emulsion method [37,38]. Briefly, 80 mgPLGA was dissolved in 1 ml solvent mixture of acetonitrile/DMC (3:2in volume ratio). 100 μl aqueous solution of certain amount of DOXYwas added into the above solution and emulsified with a probesonicator at 50 W (Virsonic 100, Gardiner, NY). This primary w/oemulsion was then emulsified into 30 ml of mineral oil containing 4%hexaglyceryl polyricinolate under sonication at 90 W to form the w/o/o emulsion. The resulting secondary emulsion was magneticallystirred at room temperature for 12 h to evaporate the solvents. Thesphereswere collected by centrifugation andwashed three timeswithpetroleum ether and freeze dried. Four PLGA formulations: PLGA50-

6.5K, PLGA50-64K, PLGA75-113K and PLGA85-142Kwere used for thein vitro release study. The macro-porous PLLA nano-fibrous scaffoldswith dimensions of 7.2 mm in diameter and 2 mm in thickness wereprefabricated by a combination of phase separation and sugarleaching techniques previously described [34]. The PLGA NS wereincorporated onto PLLA scaffolds using a post seedingmethod [32,33].Briefly, PLGANSwere dispersed in hexanewith 0.1% span 80 andwereseeded onto NS-scaffolds drop wise. Then, the scaffolds were subjectto vapor of a mixed solvent of hexane/THF (volume ratio 9:1) for30 min. The scaffolds were vacuum-dried for 3 days to remove thesolvent.

2.3. Characterization of PLGA NS and NS-scaffolds

The morphology and size of the nanospheres and the scaffolds(before and after NS incorporation) were examined using scanningelectron microscopy (Philips XL30 FEG SEM). The nanospheres weredispersed in hexane and deposited onto a metal stub. SEM figures ofthese spheres were taken after gold coating. Average diameters of alltype of PLGA spheres studied were calculated based on the sizesmeasured from the SEM images using ImageJ software (NationalInstitutes of Health). The sizes of PLGA50-6.5K and PLGA85-142Kspheres were also measured by dynamic light scattering (DLS) in anaqueous solution for comparison (Delsa Nano C Particle Size Analyzer,Beckman Coulter Inc). The drug content and encapsulation efficiencywere determined by UV spectrophotometric analysis. Briefly, 10 mgDOXY-encapsulated nanospheres (or one NS-scaffold) were dissolvedin dichloromethane. Then, 10 ml PBS was added to extract DOXY. Theextraction process was repeated several times. The solution wasfiltered through a 0.45 μm filter to remove polymer debris and thenthe DOXY concentrationwasmeasured using a UV spectrophotometer(Hitachi U2910) at the λmax value of 274 nm. PBS was used as thecontrol.

2.4. In vitro drug release studies

The release profiles of DOXY loaded NS or NS-scaffolds with fourPLGA formulations were investigated in PBS (0.1 M). 10 mg NS weresuspended in 1 ml PBS and placed within a dialysis bag. The dialysisbag was kept in a glass flask containing 20 ml PBS. One NS-scaffoldwas suspended in 1 ml PBS in a glass vial. Both the glass vial and theglass flask were covered in aluminum foil and shaken at 50 rpm at37 °C. At designated time points (1, 3, 5, 7, 10, 14, 21, 28, 35, 42 days),the release medium was withdrawn and replaced with pre-warmedfresh PBS. The released amounts of DOXY were analyzed using a UVspectrophotometer at the λmax value of 274 nm [23]. For DOXY releasethrough the dialysis bag from NS alone, PBS was used as the control.For DOXY release from NS-incorporated scaffolds, blank PLGA NS(without DOXY) were immobilized onto the scaffolds and were usedas the controls. The absorbance values of the controls were subtractedfrom those of the DOXY containing samples.

2.5. In vitro antibacterial activity of NS-scaffold

The antibacterial assay was performed as previously described[39]. The overnight incubated bacteria solution (S. aureus and E. coli)was diluted with sterile Müeller–Hinton Broth and reached anabsorbance value between 0.1 and 0.2 at 625 nm (equivalent to 0.5McFarland standard). First, the bacterial clone formation assay wasperformed using the Müeller–Hinton agar (1.5%) plate. The scaffoldswith or without DOXY nanospheres were placed in the center of themelting agar just before solidification. The DOXY was allowed todiffuse from the scaffold into the agar for 4 h before 100 μL inoculumwas spread over the agar plate. The plates were incubated at 37 °C for5 days. The bacterial growth on plates was visualized directly on theplate. Second, the released drug solutions, collected from different

365K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

scaffold groups at different time points, were mixed with 2XMüeller–Hinton Broth and overnight inoculum (the bacterial final absorbancevalue was around 0.1 at 625 nm). All tested solutions were incubatedat 37 °C on a shaker (50 rpm). The absorbance at 625 nm wasmonitored after 12 h of incubation. The percentage of bacterialinhibition was calculated using the following equation.

Bacterial inhibition (%)=abs(Ic− Is)/(Ic)×100, where Ic and Is arethe absorbances of the control bacterial solution without drug andbacterial solutions with released drug respectively at 625 nm after12 h [39].

3. Results

3.1. PLGA nanospheres loaded PLLA nano-fibrous scaffolds

PLGA NS prepared from amodified w/o/o double emulsion processwere of smooth surface (Fig. 1A). The combination of the phaseseparation and sugar sphere templating techniques allowed theformation of PLLA scaffolds with a high porosity (98%), multi-levelpore structures and nano-fibrous architecture of the pore walls. Thescaffolds prepared for this study have regular spherical macro-pores250–425 μm in diameter with well interconnected inter-pore open-ings of ∼100 μm (Fig. 1B). The scaffold walls are entirely made of thePLLA nanofibers with the diameter range of 50–500 nm (Fig. 1C),which is similar in size to native collagen fibers. Fig. 1D and E show across section of the scaffolds after the DOXY loaded NS incorporation,and the pore wall surface at a higher magnification. From SEMobservation, it is clear that the macro-pores and inter-pore openingstructure of the scaffolds were well preserved and the nanosphereswere uniformly distributed throughout the nano-fibrous pore walls.

Four different PLGA NS formulations were examined for encapsu-lation efficiency and release behavior. Two of the copolymers had thesame LA/GA ratio of 50/50 but different molecular weights, PLGA50-6.5K (Mw: 6.5 K) and PLGA50-64K (Mw: 64 K). The other two haddifferent LA/GA ratios: PLGA75-113K with a LA/GA ratio of 75/25(Mw: 113 K) and PLGA85-142K with a LA/GA ratio of 85/15(Mw:142 K). The effects of processing parameters such as polymerconcentration, volume ratio of drug solution to inner oil phase (w/o1),volume ratio of inner oil to outer oil phase (o1/o2), the addition of PAAon the DOXY encapsulation efficiency (EE) of PLGA85-142K NS arepresented in Table 1. It was observed that the EE of DOXY increasedwith an increase in the polymer concentration from 3% to 8%, nosignificant difference between 8% and 10%. With the increase ofaqueous drug volume, the encapsulation efficiency decreased. Thehigher polymer concentrationmay entrap the aqueous DOXY dropletsmore efficiently (higher EE), while the greater aqueous DOXY volumedestabilizes the emulsion (lower EE). The increase in the volume ratioof outer oil phase to inner oil phase did not change EE significantly. Itwas found that the addition of poly(acrylic acid) (PAA) as a co-encapsulation factor did not have significant effect on the EE of theDOXY.

The diameter and the DOXY encapsulation efficiency of NS of fourPLGA formulations with the same other processing parameters arelisted in Table 2. Based on SEMmeasurement, the average diameter ofthe spheres increased with the increase of LA/GA ratio and/or PLGAmolecular weight, with PLGA50-6.5K NS had the smallest diameter of300 nm and PLGA85-142K NS had the largest diameter of 730 nm.Based on DLS measurement, the diameter of PLGA50-6.5K NS andPLGA85-142K NS was 400 nm and 890 nm respectively in aqueoussolution, possibly due to some degree of swelling or/and certain

Fig. 1. Characterization of PLGA85-142 K nanospheres (NS) and PLLA nano-fibrousscaffold (NS-scaffold) before and after nanospheres incorporation. (A) SEM view ofDOXY containing PLGA85-142 K NS; (B, C) SEM views of the PLLA scaffold before NSincorporation at 100× and 10,000× magnifications respectively; and (D, E) SEM viewsof PLLA scaffold after NS incorporation at 100× and 10,000× magnificationsrespectively.

Table 2Effects of PLGA composition on sphere diameter with the same other processingparameters (8% polymer concentration, 4% PR-15, 1:10 w/o1, and 1:30 o1/o2). Each datapoint represents a mean±standard deviation. Sphere diameters were measured usingSEM.

Polymer LA/GA ratio Mw (kDa) EE (%) Sphere diameter(nm)

PLGA50-6.5K 50/50 6.5 26.1±4.8 300±110PLGA50-64K 50/50 64 24.4±3.5 520±150PLGA75-113K 75/25 113 23.1±4.4 570±130PLGA85-142K 85/15 142 25.3±3.0 730±160

366 K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

discrepancy between the two different methods. The changes in sizeand shape of similar PLGA spheres in aqueous solution over long-termculture were reported previously by our lab [40] and others [37]. SuchPLGA spheres swelled first and then deformed from their originalspherical shape into irregular shape with a rough surface texture overtime. Eventually, they collapsed and disintegrated as the degradationcontinued [40]. EE of DOXY was not changed significantly amongdifferent PLGA NS formulations. The larger diameter of PLGA85-142KNS was possibly resulted from the higher viscosity of the DOXY/PLGA85-142K polymer emulsion compared to those of other polymers(Table 2).

3.2. In vitro DOXY release kinetics

In vitro release profiles of DOXY from PLGA NS with four PLGAformulations were displayed in Fig. 2A and release profiles of DOXYfrom NS-scaffolds were shown in Fig. 2B. Compared to release profilesof DOXY from PLGA NS alone, the release from the scaffolds showed asimilar release pattern with decreased initial burst release andprolonged release period. In general, the DOXY release profiles ofNS-scaffolds showed two phases: an initial burst release within thefirst day followed by a slow and sustained release over time. Therewere also significant differences in both the burst release rate andsustained release rate of different PLGA formulations. Both PLGA50-6.5K NS-scaffold and PLGA50-64K NS-scaffold released approximately70% of DOXY within the first day and 95% of DOXY within 2 weeks,while the PLGA75-113K NS-scaffold (with a higher LA/GA ratio andhigher molecular weight) had a lower burst release (50%) in the firstday and a slower sustained release of DOXY (85% within 2 weeks, 95%within 6 weeks). The PLGA85-142K NS-scaffold (with the highest LA/GA ratio and the highest molecular weight) had the lowest burstrelease (25%) in the first day and kept a continuous slow release forlonger than 6 weeks (80% within 6 weeks). This two-phase releaseprofile of PLGA nanospheres, characterized by an initial burst releasefollowed by a sustained slow release, indicated that DOXY wasdispersed in a polymeric matrix inside the PLGA nanospheres [41,42].The initial burst release was reducedwith the increase of the PLGA LA/GA ratio and the PLGA molecular weight. PLGA85-142K NS-scaffoldswith the lowest burst release and the longest release period wasselected for in vitro antibacterial test.

3.3. In vitro antibacterial effect of DOXY NS-scaffolds

The antibacterial activity of the released DOXY was investigatedusing bacteria inhibition experiments. S. aureus and E. coli were usedin the current study since they represent the most common Gram-

Table 1Effects of various processing parameter on encapsulation efficiency of DOXY in PLGA85-142K nanospheres. Each EE data point represents a mean±standard deviation (n=3).Note: w: water phase; o1: inner oil phase (PLGA solution); and o2: outer oil phase(mineral oil solution).

PLGA85-142K concentration w/o1 o1/o2 PAA concentration EE (%)

3% 1:10 1:30 0 13.3±5.25% 1:10 1:30 0 15.3±4.18% 1:10 1:30 0 25.3±3.010% 1:10 1:30 0 25.9±3.98% 1:4 1:30 0 8.6±2.18% 1:5 1:30 0 10.8±3.28% 1:10 1:30 0 25.3±3.08% 1:20 1:30 0 25.4±3.58% 1:10 1:10 0 22.6±4.88% 1:10 1:15 0 24.4±3.88% 1:10 1:30 0 25.3±3.08% 1:10 1:30 2% 24.4±3.38% 1:10 1:30 4% 23.5±4.58% 1:10 1:30 8% 23.1±5.68% 1:10 1:30 16% 25.7±3.1

positive and Gram-negative bacteria in the human body. The growthof S. aureus and E. coli bacteria can be visualized directly from the plateto assess the antibacterial function of all scaffold groups (shown inFig. 3). It was observed that the agar plates with the BSA scaffoldswere totally covered with S. aureus or E. coli after 5 days of incubation.In contrast, the growth of S. aureus and E. coli were both inhibited atregions around the DOXY adsorbed scaffolds and DOXY NS-scaffolds.The bacteria inhibited areas were much larger on the plates with theDOXY NS-scaffolds than DOXY adsorbed scaffolds.

The data from the long-term liquid bacterial culture (S. aureus datashown in Fig. 4 and E. coli data shown in Fig. 5) demonstrated that theantibacterial activity of DOXY lasted for more than 6 weeks in the NScontrolled release group (showed 70% inhibition for S. aureus and 40%inhibition for E. coli at Day 42), while that of the adsorbed DOXY

Fig. 2. In vitro release kinetics of DOXY from NS (A) and from NS-incorporated nano-fibrous PLLA scaffolds (B): in 10 mM PBS with DOXY loading of 100 μg/scaffold. Eachdata point represents a mean±standard deviation (n=3).

Fig. 4. Long-term S. aureus growth inhibition test for release solution from three groupsof scaffolds. Each data point represents a mean±standard deviation (n=3).

367K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

decreased rapidly after 14 days (showed 18% inhibition for S. aureusand 9% inhibition for E. coli at Day 42) and the BSA scaffold had noantibacterial effect even from the very beginning (showed less than20% inhibition for S. aureus and less than 10% inhibition for E. coli). Theantibacterial activity of the scaffold measured over time was themeasurement of the overall antibacterial capacity of the scaffold,including the reduced DOXY amount released at a later time andpossibly also some biological activity loss over time. These resultsindicated that the encapsulation of DOXY in the nanospheres not onlycan control the DOXY release rate and prolong its release duration butalso retain its antibacterial activity.

4. Discussion

The main objectives of this work were to fabricate the DOXYloaded NS with good encapsulation efficiency, to incorporate the NSonto a nano-fibrous scaffold, to release DOXY in a controlled manner,and to retain the antibacterial activity. In previous publications fromour lab, novel nano-fibrous scaffolds capable of delivering growthfactors were successfully developed [32,33]. While water-in-oil-in-water (w/o/w) double emulsionmethod can be readily used to entrapproteins into nanospheres, it is difficult to entrap very hydrophilicsmall molecules such as doxycycline into NS with an acceptableencapsulation efficiency (less than 5% showed in Table 3) since thehigh solubility of the drug in the external phase and its small sizemakes it easy for the majority of the drug to diffuse to the externalphase during the emulsion and solvent evaporation process andresults in very low encapsulation efficiency [42]. Some researchersmade modifications to w/o/w method [23] or used other encapsula-tion method [21] to achieve a higher DOXY encapsulation efficiency.However, the EE increase was limited and the size of the spheresobtained becamemuch larger, usually in the range of 100 μm [23]. TheDOXY loaded spheres with this size may be used alone forantibacterial treatment, but they are too large to be incorporatedinto the nano-fibrous scaffolds [34]. To overcome these limitations, amodified w/o/o emulsion and solvent evaporation method wassuccessfully adopted to encapsulate DOXY into spheres. By changingthe external continuous phase from water to an organic solventmixture, the diffusion of the drug was reduced due to the decreasedsolubility of the drug in the external phase, and the encapsulationefficiency was greatly improved. It was also discovered that therelease profile of DOXY from the PLGA NS-scaffolds was mainlydetermined by the release from NS which can be adjusted by tailoringthe chemical composition and molecular weight of the PLGAcopolymers. The PLGA85 NS-scaffolds with the highest molecular

Fig. 3. Agar petri dish cultivated with S. aureus (upper) and E. coli (lower) with scaffoldsamples in the center after 5 days of incubation. Scaffolds on the left side were seededwith BSA containing NS. Scaffolds in the center absorbed 100 μg DOXY. Scaffolds on theright side were seeded with NS containing 100 μg DOXY.

weight and highest LA/GA had the lowest initial burst release (25%)and longest release duration of DOXY (longer than 6 weeks). Thistrend is consistent with our lab's previous publications on the releaseprofiles of growth factors from NS-scaffolds [32,33]. The difference inrelease profiles of PLGA NS-scaffold was possibly resulted from boththe difference in the NS diameter and PLGA degradation rate. PLGA85-142K NSwith the largest NS diameter had the smallest specific surfacearea and the longest average travel distance for the encapsulatedDOXY to diffuse out of the spheres, resulting in the lowest burstrelease during the earlier release stage. It is also possible that thehighest viscosity of PLGA85 polymer solution also reduced thediffusion of DOXY to the outer oil phase during the emulsion andsolvent evaporation process so that the least amount of the drug wassimply absorbed on or near the NS surface, resulting in a reducedinitial burst release. PLGA85-142K NS also had the slowest degrada-tion rate due to the highest molecular weight and LA/GA ratio, whichcontributed to the slower release rate of DOXY. It was also discoveredthat incorporating DOXY loaded PLGA NS into PLLA nano-fibrousscaffold effectively reduced burst release and slowed down the DOXYrelease rate. The reduced burst release of DOXY from nano-fibrousscaffold may have resulted from the adsorption and desorption ofDOXY on the nanofibers before diffusing out of the scaffolds due to thevery high surface area (100 m2/g) of the nano-fibrous scaffolds.

The antimicrobial experiment demonstrated the biological func-tion of the DOXY released from PLGA NS-scaffolds for a long duration(6 weeks), which is a great improvement for the long-term treatmentof dental, periodontal and bone infection or MMP-inhibition since the

Fig. 5. Long-term E. coli growth inhibition test for release solution from three groups ofscaffolds. Each data point represents a mean±standard deviation (n=3).

Table 3Encapsulation efficiencies (10% polymer concentration, 1% PVA, w1/o=1:10, o/w2=1:30). Each data point represents a mean±standard deviation. Note: w1: aqueousdrug solution, o: PLGA solution, and w2: aqueous PVA solution.

Polymer LA/GA ratio Mw (kDa) EE (%) Sphere diameter(nm)

PLGA50-6.5K 50/50 6.5 3.2±1.8 280±70PLGA50-64K 50/50 64 2.7±0.5 320±100PLGA75-113K 75/25 113 3.1±1.4 370±90PLGA85-142K 85/15 142 2.3±1.0 410±120

368 K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

commercially available DOXY releasing gels typically release all DOXYwithin 2 weeks [43]. These novel DOXY NS-incorporated nano-fibrousscaffolds were intended for tissue engineering applications wheremicrobial infection may be involved. While the macro-porous andnano-fibrous scaffolds may promote the tissue regeneration, theDOXY NS in the scaffolds may also protect the tissue from microbialinvasion. Since DOXY is also an inhibitor of multiple MMPs such asMMP-8, -9 and -13, DOXY NS-scaffold may also be able to protectengineered bone from MMP destruction in situations with chronicinflammation in diseases such as periodontitis and rheumatoidarthristis [44].

The 3-D NS-scaffolds developed in this work have shown thecapability of controlled delivery of DOXY and inhibition of bacterialgrowth for a prolonged period. This success has demonstrated thescaffolds to be capable of controlled delivering small hydrophilic drugmolecules as well as large proteins. With the versatile deliverycapability and macro-porous and nano-fibrous structure, the scaffoldsmay allow for the implementation of more comprehensive strategiesin tissue engineering.

5. Conclusion

Results of in vitro release and antibacterial experiments suggestthat the developed drug-containing nano-fibrous scaffolds arecapable of effectively delivering doxycycline in a controlled fashionwith prolonged duration. These biodegradable PLLA scaffolds havewell interconnected macro-porous and nano-fibrous structure andcan inhibit common bacterial growth for more than 6 weeks with theincorporation of DOXY containing PLGA NS. Release kinetics from thescaffolds was found to be determined by PLGA formulation. Theincorporation of the NS into scaffolds reduced the burst release.PLGA85-142K NS-scaffold has the lowest initial burst release and thelongest release period compared to other formulations investigated.Taken together with earlier studies, the NS-scaffold system can beused as a delivery system for both large proteins and small drugmolecules for various complex tissue engineering applications.

Acknowledgement

The authors would like to acknowledge the financial support fromthe National Institutes of Health (research grants DE015384 andDE017689: PXM).

References

[1] A. JM, Epidemiology and risk factors of periodontal diseases, Dent. Clin. North Am.(49) (2005) 517–532.

[2] M. Bokor-Bratic, T. Brkanic, Clinical use of tetracyclines in the treatment ofperiodontal diseases, Med. Pregl. 53 (5–6) (2000) 266–271.

[3] V.L., C.W. Stratton, Mechanisms of action for antimicrobial agents: generalprincipals and mechanisms for selected classes of antibiotics, Antibiotics inLaboratory Medicine, 4th ed, Williams & Wilkins, Baltimore, 1996.

[4] R.A. Seymour, P.A. Heasman, Tetracyclines in the management of periodontaldiseases. A review, J. Clin. Periodontol. 22 (1) (1995) 22–35.

[5] A.R. Amin, M.G. Attur, G.D. Thakker, P.D. Patel, P.R. Vyas, R.N. Patel, I.R. Patel, S.B.Abramson, A Novel Mechanism of Action of Tetracyclines: Effects on Nitric OxideSynthases, Proc. Natl. Acad. Sci. U. S. A. 93 (24) (1996) 14,014–14,019.

[6] R. Gapski, H. Hasturk, E. Van Dyke Thomas, J. Oringer Richard, S. Wang, M. BraunThomas, V. Giannobile William, Systemic MMP inhibition for periodontal woundrepair: results of a multi-centre randomized-controlled clinical trial, J. Period-ontol. 36 (2) (2009) 149–156.

[7] W.V. Giannobile, Host-response therapeutics for periodontal diseases, J. Period-ontol. 79 (8, Suppl) (2008) 1592–1600.

[8] R.A. Greenwald, S.A. Moak, N.S. Ramamurthy, L.M. Golub, Tetracyclines suppressmatrix metalloproteinase activity in adjuvant arthritis and in combination withflurbiprofen, ameliorate bone damage, J. Rheumatol. 19 (6) (1992) 927–938.

[9] B.V. Shlopov, J.M. Stuart, M.L. Gumanovskaya, K.A. Hasty, Regulation of cartilagecollagenase by doxycycline, J. Rheumatol. 28 (4) (2001) 835–842.

[10] N. Joshi, D.Q. Miller, Doxycycline revisited, Arch. Intern. Med. 157 (13) (1997)1421–1428.

[11] J. Slots, T.E. Rams, Antibiotics in periodontal therapy: advantages and disadvan-tages, J. Periodontol. 17 (7 ( Pt 2)) (1990) 479–493.

[12] N.C. Klein, B.A. Cunha, Tetracyclines, Med. Clin. North Am. 79 (4) (1995) 789–801.[13] R.A. Seymour, P.A. Heasman, Tetracyclines in the management of periodontal-

diseases — a review, J. Periodontol. 22 (1) (1995) 22–35.[14] S.M. Attar, Tetracyclines: what a rheumatologist needs to know? International

Journal of Rheumatic Diseases 12 (2) (2009) 84–89.[15] W.J. Loesche, The antimicrobial treatment of periodontal disease: changing the

treatment paradigm, Critical Reviews in Oral Biology & Medicine 10 (3) (1999)245–275.

[16] A. Mombelli, L.P. Samaranayake, Topical and systemic antibiotics in themanagement of periodontal diseases, Int. Dent. J. 54 (1) (2004) 3–14.

[17] C.W. McIlwraith, Treatment of infectious arthritis, Veterinary Clinics of NorthAmerica—Large Animal Practice, vol. 5,(2), 1983, pp. 363–379.

[18] F.J. van der Ouderaa, Anti-plaque agents. Rationale and prospects for preventionof gingivitis and periodontal disease, J. Periodontol. 18 (6) (1991) 447–454.

[19] E. Esposito, R. Cortesi, F. Cervellati, E. Menegatti, C. Nastruzzi, Biodegradablemicroparticles for sustained delivery of tetracycline to the periodontal pocket:formulatory and drug release studies, J. Microencapsul. 14 (2) (1997) 175–187.

[20] J.M. Goodson, D. Holborow, R.L. Dunn, P. Hogan, S. Dunham, Monolithictetracycline-containing fibers for controlled delivery to periodontal pockets, J.Periodontol. 54 (10) (1983) 575–579.

[21] M.C. Haerdi-Landerer, M.M. Suter, A. Steiner, M.M. Wittenbrink, A. Pickl, B.A.Gander, In vitro cell compatibility and antibacterial activity of microencapsulateddoxycycline designed for improved localized therapy of septic arthritis, J.Antimicrob. Chemother. 61 (2) (2008) 332–340.

[22] E.-R. Kenawy, G.L. Bowlin, K. Mansfield, J. Layman, D.G. Simpson, E.H. Sanders, G.E.Wnek, Release of tetracycline hydrochloride from electrospun poly(ethylene-co-vinylacetate), poly(lactic acid), and a blend, J. Control Release. 81 (1–2) (2002) 57–64.

[23] R.C. Mundargi, S. Srirangarajan, S.A. Agnihotri, S.A. Patil, S. Ravindra, S.B. Setty, T.M. Aminabhavi, Development and evaluation of novel biodegradable micro-spheres based on poly(D,L-lactide-co-glycolide) and poly(e-caprolactone) forcontrolled delivery of doxycycline in the treatment of human periodontal pocket:in vitro and in vivo studies, J. Control Release. 119 (1) (2007) 59–68.

[24] K. Schwach-Abdellaoui, N. Vivien-Castioni, R. Gurny, Local delivery of antimicro-bial agents for the treatment of periodontal diseases, Eur. J. Pharm. Biopharm. 50(1) (2000) 83–99.

[25] F. Tamimi, J. Torres, R. Bettini, F. Ruggera, C. Rueda, M. Lopez-Ponce, E. Lopez-Cabarcos, Doxycycline sustained release from brushite cements for the treatmentof periodontal diseases, J. Biomed. Mater. Res. A. 85A (3) (2008) 707–714.

[26] R.C. Mundargi, V.R. Babu, V. Rangaswamy, P. Patel, T.M. Aminabhavi, Nano/microtechnologies for delivering macromolecular therapeutics using poly(D,L-lactide-co-glycolide) and its derivatives, J. Control. Release. 125 (3) (2008) 193–209.

[27] X.Q. Wang, E. Wenk, X.H. Zhang, L. Meinel, G. Vunjak-Novakovic, D.L. Kaplan,Growth factor gradients via microsphere delivery in biopolymer scaffolds forosteochondral tissue engineering, J. Control. Release. 134 (2) (2009) 81–90.

[28] Y. Kakizawa, R. Nishio, T. Hirano, Y. Koshi, M. Nukiwa, M. Koiwa, J. Michizoe, N. Ida,Controlled release of protein drugs from newly developed amphiphilic polymer-based microparticles composed of nanopcarticles, J. Control. Release. 142 (1)(2010) 8–13.

[29] S.E. Bae, J.S. Son, K. Park, D.K. Han, Fabrication of covered porous PLGAmicrospheres using hydrogen peroxide for controlled drug delivery andregenerative medicine, J. Control. Release. 133 (1) (2009) 37–43.

[30] H.K. Kim, H.J. Chung, T.G. Park, Biodegradable polymeric microspheres with“open/closed” pores for sustained release of human growth hormone, J. Control.Release. 112 (2) (2006) 167–174.

[31] A. Goraltchouk, V. Scanga, C.M. Morshead, M.S. Shoichet, Incorporation of protein-eluting microspheres into biodegradable nerve guidance channels for controlledrelease, J. Control. Release. 110 (2) (2006) 400–407.

[32] G.Wei, Q. Jin,W.V. Giannobile, P.X. Ma, Nano-fibrous scaffold for controlled deliveryof recombinant human PDGF-BB, J. Control. Release. 112 (1) (2006) 103–110.

[33] G. Wei, Q. Jin, W.V. Giannobile, P.X. Ma, The enhancement of osteogenesis bynano-fibrous scaffolds incorporating rhBMP-7 nanospheres, Biomaterials 28 (12)(2007) 2087–2096.

[34] G. Wei, P.X. Ma, Macroporous and nanofibrous polymer scaffolds and polymer/bone-like apatite composite scaffolds generated by sugar spheres, J. Biomed.Mater. Res. A. 78A (2) (2006) 306–315.

[35] K.M. Woo, V.J. Chen, P.X. Ma, Nano-fibrous scaffolding architecture selectivelyenhances protein adsorption contributing to cell attachment, J. Biomed. Mater.Res. A. 67A (2) (2003) 531–537.

[36] K.M. Woo, J.-H. Jun, V.J. Chen, J. Seo, J.-H. Baek, H.-M. Ryoo, G.-S. Kim, M.J.Somerman, P.X. Ma, Nano-fibrous scaffolding promotes osteoblast differentiationand biomineralization, Biomaterials 28 (2) (2006) 335–343.

369K. Feng et al. / Journal of Controlled Release 146 (2010) 363–369

[37] Y. Li, H.L. Jiang, K.J. Zhu, J.H. Liu, Y.L. Hao, Preparation, characterization and nasaldelivery of alpha-cobrotoxin-loaded poly(lactide-co-glycolide)/polyanhydridemicrospheres, J. Control. Release. 108 (1) (2005) 10–20.

[38] S.P. Schwendeman, M. Tobio, M. Joworowicz, M.J. Alonso, R. Langer, Newstrategies for the microencapsulation of tetanus vaccine, J. Microencapsul. 15(3) (1998) 299–318.

[39] K. Kim, Y.K. Luu, C. Chang, D. Fang, B.S. Hsiao, B. Chu, M. Hadjiargyrou,Incorporation and controlled release of a hydrophilic antibiotic using poly(lactide-co-glycolide)-based electrospun nanofibrous scaffolds, J. Control. Re-lease. 98 (1) (2004) 47–56.

[40] G.B. Wei, G.J. Pettway, L.K. McCauley, P.X. Ma, The release profiles and bioactivityof parathyroid hormone from poly(lactic-co-glycolic acid) microspheres, Bioma-terials 25 (2) (2004) 345–352.

[41] Y. Kawashima, H. Yamamoto, H. Takeuchi, T. Hino, T. Niwa, Properties of a peptidecontainingDL-lactide/glycolide copolymer nanospheres prepared bynovel emulsionsolvent diffusion methods, Eur. J. Pharm. Biopharm. 45 (1) (1998) 41–48.

[42] J. Yin, Y. Noda, T. Yotsuyanagi, Properties of poly(lactic-co-glycolic acid)nanospheres containing protease inhibitors: camostat mesilate and nafamostatmesilate, Int. J. Pharm. 314 (1) (2006) 46–55.

[43] N.H. Stoller, L.R. Johnson, S. Trapnell, C.Q. Harrold, S. Garrett, The pharmacokineticprofile of a biodegradable controlled-release delivery system containing doxycy-cline compared to systemically delivered doxycycline in gingival crevicular fluid,saliva, and serum, J. Periodontol. 69 (10) (1998) 1085–1091.

[44] G.N. Smith Jr., E.A. Mickler, K.A. Hasty, K.D. Brandt, Specificity of inhibition ofmatrix metalloproteinase activity by doxycycline. relationship to structure of theenzyme, Arthritis Rheum. 42 (6) (1999) 1140–1146.