Journal of Controlled Release 138 (2009) 177–184

Contents lists available at ScienceDirect

Journal of Controlled Release

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

NANOMEDICIN

E

Novel self-assembled core–shell nanoparticles based on crystalline amorphousmoieties of aliphatic copolyesters for efficient controlled drug release

Sofia Papadimitriou, Dimitrios Bikiaris ⁎Laboratory of Organic Chemical Technology, Department of Chemistry, Aristotle University of Thessaloniki, GR-541 24, Thessaloniki, Macedonia, Greece

⁎ Corresponding author. Tel.: +30 2310 997812; fax:E-mail address: dbic@chem.auth.gr (D. Bikiaris).

0168-3659/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.jconrel.2009.05.013

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 February 2009Accepted 9 May 2009Available online 14 May 2009

Keywords:ε-Caprolactone/poly(propylene succinate)NanoparticlesTiboloneRelease behavior

Poly(propylene succinate-co-caprolactone) copolymers [P(PSu-co-CL)] with different ε-caprolactone (ε-CL)to propylene succcinate (PSu) monomer ratios were synthesized using ring opening polymerization. Thesepolymers consisted of crystalline poly(ε-caprolactone) (PCL) and amorphous poly(propylene succinate)(PPSu) moieties, as shown by WAXD. In vitro biocompatibility studies showed that these copolyesters arebiocompatible. Drug-loaded nanoparticles, using tibolone as a model drug, were prepared by the solventevaporation method. Nanoparticle size ranged between 150 and 190 nm and decreased with increasingpropylene succinate (PSu) ratio in the copolymers. Nanoparticle yield, encapsulation efficiency, and drugloading increased with increasing PSu ratio. Scanning Electron Microscopy (SEM) revealed that the preparednanoparticles had a spherical shape and Transmission Electron Microscopy (TEM) showed that they wereself-assembled in core–shell structures. Amorphous PPSu and crystalline PCL comprised the core and shell,respectively. The drug is mainly located into the amorphous core in the form of nanocrystals. Drug releasestudies showed that complete release of the drug from the nanoparticles occurs over a period of 50 h. Therelease rate is greatly influenced by the copolymer composition, nanoparticle size, and encapsulationefficiency. Among the main advantages of the nanoparticles produced in this study is the absence of bursteffect during drug release.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

During the past few decades there has been an increasing interestin the development of biodegradable nanoparticles for effective drug,peptide, protein and DNA delivery [1]. Incorporation of the drug into aparticulate carrier canprotect the active substance against degradationin vivo and in vitro, improve therapeutic effect, prolong biologicalactivity, control drug release rate, and decrease administrationfrequency [2,3]. One of the major problems associated with the useof nanoparticles as polymeric drug carriers is their rapid eliminationfrom the blood stream through phagocytosis after intravenousadministration and recognition by the macrophages of the mono-nuclear phagocyte system [4]. In order tomaintain the required level ofthe active substance in the blood stream for longer time periods, long-circulating polymer nanoparticles may be designed and used [5–7].Currently, one of the most used techniques to prepare suchnanoparticles is PEGylation [8–16], in which a large number ofpolymer structures can be used as the hydrophobic segment [17].Poly(ε-caprolactone) (PCL), owing to its biodegradable/biocompatiblecharacteristics and advantage of being high permeable to drugs, hasattracted attention and became an important candidate for drugdelivery applications [18]. However, the application of PCL as drug

+30 2310 997769.

ll rights reserved.

delivery system reveals certain drawbacks, such as slow rate ofbiodegradation in human tissue due to the polymer's high degree ofhydrophobicity and crystallinity [19], and very slow drug release rates.For example, the release of thewhole amount of a drug encapsulated inPCL often requires weeks or months. In order to overcome theseproblems it is essential to modify the polymer properties in away thatmeets the requirements of the drug release application. The mostcommon and relatively easyway to enhance the properties of PCL is bymeans of copolymerization or blending with other polymers. In thepresent study, improvement of the drug release properties of PCL isattempted through the synthesis of PCL copolymers with poly(propylene succinate) (PPSu).

Poly(propylene succinate), a relatively new biodegradable poly-ester that is fast biodegradable [20] and biocompatible [21], can beproduced from monomers derived from renewable sources usingenvironmentally friendly methods and a variety of microorganisms[22]. PPSu is a very soft material with a lowmelting point (Tm=44 °C)and glass transition temperature (Tg=−36 °C) and exhibits highbiodegradation rate due to its lower degree of crystallinity [23].

In the present study, copolymers of ε-CL with propylene succinate(PSu) were synthesized, characterized, and used as candidatepolymers towards the development of effective nanoparticle drugdelivery systems. Our main objective is to prepare nanoparticleswithout the appearance of burst effects that can release the drug overa relatively short period of time, typically less than 2 days.

178 S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

Tibolonewas used as amodel hydrophobic drug to be encapsulatedinto the polymeric nanoparticles. Tibolone, [(7a,17a)-17-hydroxy-19-nor-17-pregn-5(10)-en-20-yn-3-one], is a synthetic steroid known tohave combined oestrogenic, progestogenic and androgenic character-istics and is structurally related to the progestogens norathindrone andnorethynodrel. It is a known tissue specific and effective agent that canbe used in hormone replacement therapy (HRT) in (post)menopausalwoman, for the treatment of menopausal and postmenopausaldisorders, including climacteric complaints, vasomotor symptoms,osteoporosis and vaginal atrophy [24–27].

2. Experimental

2.1. Materials

ε-Caprolactone (ε-CL) (99% Sigma-Aldrich) was dried over CaH2

and purified by distillation under reduced pressure prior to use.Succinic acid (purum 99+%), 1,3-Propanediol (1,3-PD) (purity:N99.6%) and Tetrabutyl Titanate (TBT), used as catalyst, werepurchased from Aldrich Chemical Co. Polyphosphoric acid (PPA)used as heat stabilizer was supplied from Fluka. Sodium cholate wasobtained from Sigma. Crystalline tibolone (TIBO) with an Assay of99.59% was supplied from Zhejiang Xianju Junye Pharmaceutical Co.LTD.

2.2. Synthesis of polyesters

2.2.1. Synthesis of PPSu and polycaprolactone (PCL)Synthesis of aliphatic polyester PPSu was performed following

the two-stage melt polycondensation method (esterification andpolycondensation) in a glass batch reactor [28]. The bulk polymer-ization of ε-caprolactone in order to synthesize neat PCL, was carriedout in 250 cm3 round-bottomed flask equipped with a mechanicalstirrer and a vacuum apparatus as described elsewhere [29].

2.2.2. Synthesis of P(PSu-co-CL) copolymersP(PSu-co-CL) copolymerswith variousmolar ratios, such as2.5/97.5,

5/95, 10/90, and 75/25, were synthesized according to the proceduresdescribed by Seretoudi et al. [29]. Briefly, purified PPSu was addedinto the same apparatus used for PCL synthesis and the proper amountsof ε-CL monomer were added as well as TBT (1×10−4 mol TBT/molε-CL). Polymerization took place at 180 °C under nitrogen flow and astirring rate of 500 rpm while the reaction was completed after 2 h.Unreacted monomer was removed through distillation by applying ahigh vacuum (≈5 Pa) over a time period of 15min. Polymerizationwasstopped by rapid cooling to room temperature.

2.3. Polymer characterization

Intrinsic viscosity [η] measurements on the isolated polymerswereperformed using an Ubbelohde viscometer cap. Oc at 25 °C inchloroform at a solution concentration of 1 wt.%.

Molecular weight determinations were performed by gel perme-ation chromatography (GPC) method using a Waters 150C GPCequipped with differential refractometer as detector and threeultrastyragel (103, 104, 105 Å) columns in series. Tetrahydrofuran(THF) was used as the eluent (1 ml/min) and the measurements wereperformed at 35 °C. Calibration was performed using polystyrenestandards with a narrow molecular weight distribution.

1H NMR and 13C NMR spectra of polyesters were obtained with aBruker spectrometer operating at a frequency of 400 MHz for protons.Deuterated chloroform (CDCl3) was used as solvent in order toprepare solutions of 5% w/v. The number of scans was 10 and thesweep width was 6 kHz.

A Setaram DSC141 differential scanning calorimeter (DSC),calibrated with indium and zinc standards, was used, for the

identification of the thermal properties of the copolyesters. A sampleof about 10mgwas used for each test, placed in an aluminium pan andheated from −100 °C up to 100 °C at a heating rate of 20 °C/min. Thesample remained at that temperature for 5 min in order to erase anythermal history. After that it was cooled down to−100 °C at a coolingrate of−150 °C/min and scanned again (second heating) up to 100 °Cusing 20 °C/min as heating rate. The glass transition temperature (Tg),the melting temperature (Tm) and crystallization temperature (Tc)were recorded.

Wide Angle X-Ray Diffractrometry (WAXD) was used for theidentification of the crystal (structure and changes) of the polymersand also of the drug used in case of nanoparticle samples.WAXD studywas performed over the range 2θ from 5 to 50 °C, using a Philips PW1710 diffractometer with Bragg–Brentano geometry (θ, 2θ) and Ni-filtered CuKa radiation.

2.4. Biocompatibility study of the prepared polyesters

2.4.1. Cell cultureHuman umbilical vein endothelial cells (HUVEC) were grown

routinely in RPMI-1640 medium supplemented with 15% fetal bovineserum (FBS), 15 mg ECGS,100 U/ml penicillin, 100 µg/ml streptomycin,50 µg/ml gentamycin and 2.5 µg/ml amphotericin B. The cultures weremaintained at 37 °C, 5% CO2 and 100% humidity.

2.4.2. In vitro biocompatibility studyThe biocompatibility of P(PSu-co-CL) was evaluated by measuring

the viability of HUVEC cells in the presence of different polymerconcentrations and comparing results with those obtained from poly(lactic acid) (PLA). Cell viability was determined by means of the MTTassay. HUVEC cells were seeded in 24-well plates at a density of30,000 cells per well in 500 µl cell culturemedium. Twenty-four hoursafter plating, different amounts of P(PSu-co-CL) nanoparticles(suspended inwater) were added in thewells. After 24 h of incubationat 37 °C, 50 µl of MTT solution (5 mg/ml in PBS pH 7.4) was added intoeach well and plates were incubated at 37 °C for 2 h. The mediumwaswithdrawn and 200 µl acidified isopropanol (0.33 ml HCl in 100 mlisopropanol) was added in each well and agitated thoroughly todissolve the formazan crystals. The solutionwas transferred to 96-wellplates and immediately read by a microplate reader (Biorad, Hercules,CA, USA) at 490 nm wavelength. The experiments were performed intriplicates. Biocompatibility of polymers was expressed as % cellviability, calculated as the ratio between the number of cells treatedwith the nanoparticles and that of non-treated cells (control).

2.5. Preparation of P(PSu-co-CL) nanoparticles loaded with tibolone

P(PSu-co-CL) copolymer nanoparticles were prepared by o/wsolvent evaporation method. Copolymer (50 mg) and tibolone (5 mg)were dissolved in 2 ml of dichloromethane. The solution wastransferred to an aqueous solution of sodium cholate (V=6 ml,C=12 mV) and it was probe-sonicated for 1 min [30]. The o/wemulsion formed was gently stirred until the evaporation of theorganic solvent was complete. Nanoparticles were purified bycentrifugation (9500 rpm for 20 min). The samples were reconsti-tuted with deionized water. Polymer aggregates were removed byfiltering the suspension through a 1.2 µm pore size microfilter.

2.6. Characterization of drug-loaded nanoparticles

2.6.1. SEM measurementsThemorphology of the prepared nanoparticles was examined with

a Scanning Electron Microscope (JEOL, JMS-840). The samples werecoated with carbon black to avoid charging under the electron beam.Operating conditions were: accelerating voltage 20 kV, probe current45 nA, and counting time 60 s.

Table 2Thermal properties of prepared copolymers.

Polyester Feed Tm1 Tm2 Tg Tc ΔHm

(°C) (°C) (°C) (°C) (J/g)

P(PSu-co-CL) 0/100 56.3 – −61.6 27.8 65.3P(PSu-co-CL) 2.5/97.5 55.9 – −60 26.8 61.2P(PSu-co-CL) 5/95 54.8 −58.6 25.1 57.3P(PSu-co-CL) 10/90 50.9 – −58.2 21.7 46.9P(PSu-co-CL) 25/75 40.5 45.4 −57.9 11.5 39.5P(PSu-co-CL) 100/0 44 – −35.0 – 44.3

179S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

2.6.2. TEM measurementsThe morphological examination of nanoparticles and tibolone-

loaded nanoparticles was performed using a JEOL 120 CX transmissionelectron microscope (TEM) operating at 120 kV. The samples wereimmobilized on copper grids, dried at room temperature, andexamined using TEM without being stained.

2.6.3. Mean size determinationNanoparticle size distribution was determined by dynamic light

scattering (DLS) using a Zetasizer Nano Instrument (MalvernInstruments, Nano ZS, ZEN3600, UK) operating with a 532 nm laser.A suitable amount of nanoparticles was dispersed in distilled watercreating a total concentration of 1‰ and was kept at 37 °C underagitation at 100 rpm. All measurements were performed in triplicatesand the results were reported in terms of mean diameter±SD.

2.6.4. Nanoparticle yield, drug loading and entrapment efficiencyThe obtained micellar solutions were frozen and lyophilized in a

freeze-drier system in order to obtain the nanoparticles. The weighednanoparticles were dispersed and properly diluted in the mobilephase used (methanol/water 77/23 v/v). The quantitative analysiswas carried out by using a Shimadzu HPLC (model LC-20AD). Thecolumn used was a Hypersil BDS, 5 µm, 200×4.6 mm with columntemperature of 30 °C. The mobile phase consisted of methanol/water77/23 v/v at a flow rate of 0.8 ml/min. Concentration determinationwas performed using UV detection at 205 nm, and was based on apreviously created calibration curve. Nanoparticle yield, drug loadingand drug entrapment efficiency were calculated from Eqs. (1)–(3),respectively:

Nanoparticle yield kð Þ = weight of nanoparticlesweight of polymer anddrug fed initially

× 100

ð1Þ

Drug loading kð Þ = weight of drug innanoparticlesweight of nanoparticles

× 100 ð2Þ

Entrapment efficiency kð Þ = weight of drug innanoparticlesweight of drug fed initially

× 100: ð3Þ

2.6.5. Fourier Transformation-Infrared Spectroscopy (FT-IR)FT-IR spectra of freeze-dried nanoparticles were obtained using a

Perkin-Elmer FT-IR spectrometer, model Spectrum 1000. In order tocollect the spectra, a small amount of each material was used (1 wt.%)and compressed in KBr tablets. The IR spectra, in absorbance mode,were obtained in the spectral region of 450 to 4000 cm−1 using aresolution of 4 cm−1 and 64 co-added scans.

2.6.6. In vitro drug release studiesIn vitro drug release studies were performed by using the

dissolution apparatus II basket method. Drug-loaded nanoparticlesuspensions corresponding to 2.5 mg tibolone were placed in adialysis cellulose membrane bag, with a molecular weight cut-off of12,400, tied, and placed into the baskets. The dissolution mediumwas

Table 1Chemical composition and molecular weight of prepared polyesters.

Polyester Feed 1H NMR [η] Mn Mw Mw/Mn(dL/g) (g/mol) (g/mol)

P(PSu-co-CL) 0/100 0/100 1.25 72,360 113,600 1.57P(PSu-co-CL) 2.5/97.5 2.72/97.28 1.22 47,800 69,400 1.45P(PSu-co-CL) 5/95 3.95/96.05 0.85 41,800 56,500 1.35P(PSu-co-CL) 10/90 7.1/92.9 0.94 39,600 52,000 1.31P(PSu-co-CL) 25/75 19.5/80.5 0.96 53,100 87,400 1.64P(PSu-co-CL) 100/0 100/0 0.5 23,000 41,500 1.8

consisted of 500 ml water and SLS (PH=7, Sodium Lauryl Sulfate0.25%), the stirring rate was kept constant at 100 rpm, and thetemperature at 37 °C [31]. At predetermined time intervals, 4 ml ofaqueous suspension was withdrawn from the release media, filtered,and subjected to the analysis described above using the same HPLCmethod and an appropriate calibration curve.

3. Results and discussion

3.1. Synthesis preparation and characterization of P(PSu-co-CL) copolymers

The preparation of the P(PSu-co-CL) copolymers was achieved bycombining two different processes. In the first one, PPSu wassynthesized by esterification reactions of 1,3-propanediol and succinicacid at an elevated temperature (190 °C). The prepared oligomers arebrought to a higher temperature (230 °C) in order to increase themolecular weight [21]. The resulting PPSu has an average numbermolecular weight of 23,300 g/mol and an average weight molecularweight of 41,500 g/mol (as found by GPC), which is satisfactory foraliphatic polyester. In the second process, copolymers of P(Psu-co-CL)at theoretical molar ratios of 25/75, 10/90, 5/95 and 2.5/97.5,respectively, were synthesized using the prepared PPSu and differentε-CL amounts. The reaction of ε-CL with PPSu took place at 180 °C, asdescribed previously [32]. The molecular weight distribution of theprepared polyesters was determined by intrinsic viscosity values aswell as by GPC (Table 1). It can be seen from Table 1 that, as theamount of PCL increases in the copolymer, the molecular weight alsoincreases.

Thermal analysis of the prepared copolyesters was performed forthe characterization of their physical state. Melting point tempera-tures, glass transition temperatures and heat of fusions (ΔHm) arepresented in Table 2. Neat PPSu and PCL have melting points of 44 °Cand 62 °C, respectively. The DSC traces of copolymers show also acrystalline character since endothermic peaks are also recorded.Comparing the melting points of the copolymers it can be seen thatthese are shifted to lower temperatures by increasing PPSu content(Table 2) due to the lower perfection and purity of the formed crystals.In the case of P(PSu-co-CL) 25/75 copolymer, a double melting pointwith maximum at Tm1=40.5 °C and Tm2=45.4 °C is recorded, whichmay be attributed to the different blocks (PCL, PPSu) in themacromolecular backbone and different microphases. The creationof block copolymer seems to be the case as has been previously provedfrom the analysis of the corresponding NMR spectra [32]. Bothmeltingpoints are shifted to lower temperatures compared to the corre-sponding melting points of the homopolymers. In this particularcopolymer it seems that PPSu blocks are in appropriate length toprepare crystalline structures, while in other copolymers only PCL canbe crystallized. However, the addition of PPSu moieties affects thecrystal formation of PCL, thus the degree of crystallinity is reducedfrom 48% for neat PCL to about 29% for the P(CL-co-PSu) 75/25. Forcompletely crystalline PCL, the enthalpy of fusion has been reported tobe 136 J/g [33], while for completely crystalline PPSu it was found tobe 140 J/g [34].

Fig. 1 shows theWAXD patterns of the neat PPSu and PCL polymersas well as their P(PSu-co-CL) copolyesters. PPSu shows the following

Fig. 1. XRD patterns of neat PPSu and PCL polymers and their copolymers.

Table 3Nanoparticle yield, drug loading and entrapment efficiency of tibolone-loaded PCL/PPSucopolymer nanoparticles.

Sample Drugloading

Entrapmentefficiency

Micelleyield

Meandiameter

PdI

(%) (%) (%) (nm)

P(PSu-co-CL)0/100

5.14±0.7 33.40±5.7 58.9±6.52 190±3.5 0.22

P(PSu-co-CL)2.5/97.5

4.55±0.45 34.08±3.23 68.09±4.31 188±2.6 0.32

P(PSu-co-CL)5/95

4.26±0.62 29.14±5.66 62.09±7.6 175±7.6 0.35

P(PSu-co-CL)10/90

8.48±0.43 74.25±1.33 79.54±2.45 163±1.2 0.33

P(PSu-co-CL)25/75

8.76±0.65 82.79±2.47 85.91±1.22 152±3.4 0.23

P(PSu-co-CL)100/0

2.35±1.2 7.10±5.67 4.36±1.34 180±5.3 0.18

180 S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

characteristic peaks at 2θ=18.32, 19.42, 20.65, 22.63 and 26.07° andPCL at 2θ=15.71, 21.32, 21.96, 23.67 and 29.75°. In their copolymersonly the characteristic peaks of PCL crystalline structure are recorded,except in the case of P(PSu-co-CL) containing 25 wt.% PPSu, where asmall peak is recorded at 2θ=18.34°. This is evidence that only at thiscopolymer PPSu can crystallize while in the others it remains inamorphous form. However, even in this case the degree of crystallinityof PPSu should be very small. These observations are in agreementwith DSC data since only at this copolymer two melting points wererecorded.

3.2. In vitro biocompatibility of polyesters

The PPSu polymer and the P(PSu-co-CL) copolymers exhibited lowtoxicity against HUVEC cells, with appreciable cytotoxicity (higherthan 20% reduction of cell viability) being observed only afterexposing the cells at high nanoparticle concentrations, i.e. higherthan 800 µg/ml (Fig. 2). In terms of polymer toxicity on HUVEC cells,PPSu and P(PSu-co-CL) copolymers were comparable to PLA, apolymer of high biocompatibility that is widely used in biomedicalapplications [35].

Fig. 2. Viability of HUVEC cells as a function of polymer type and concentration (cellincubation time: 24 h).

3.3. Morphology and size distribution of drug-loaded nanoparticles

The particle size and size distribution were measured by dynamiclight scattering (DLS). The mean diameter varied from 150 to 200 nm.All nanoparticle samples show a unimodal size distribution. It can beseen that the particle size decreases with increasing PPSu ratio in thecopolymer (Table 3). This result is in agreement with the character-istics of many copolymer micelles, namely the shorter the hydro-phobic component, the smaller the size of micelles [36]. In the studiedsamples, PCL is the hydrophobic segment, since the presence of twocarbonyl groups in the chemical structure of PPSu increase thehydrophilic character of the latter.

Scanning Electron Microscopy (SEM) photographs from twonanoparticle samples are shown in Fig. 3. These images are alsorepresentative for the rest of the samples. SEMmicrographs establishedthat the drug-loaded nanoparticles had a discrete spherical shape with

Fig. 3. SEM photographs of P(PSu-co-CL) nanoparticles with molar ratios a) 10/90 andb) 25/75.

Fig. 4. TEM photographs of P(PSu-co-CL) 25/75 nanoparticles.

181S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

different sizes ranging from 150 up to 200 nm, a fact that is inagreement with the measurements of dynamic light scattering.However, the ability of SEM to study in detail the nanoparticle structureis limited and for this reason Transmission Electron Microscopy (TEM)was also used.

TEM micrographs of the prepared nanoparticles also revealed thatthese have spherical shape (Fig. 4), with dimensions similar to thosefound through SEM and DLS measurements. Examination of thesenanoparticles at higher magnifications (Fig. 5) shows that they have acore–shell structure. In the outer shell layer the macromolecules arearranged in concentric rings, forming an “onion” structure. Moreover,the outer shell of each nanoparticle appears as a dark region, a factthat indicates the presence of a crystalline region. On the contrary, thecolor in the centre of the nanoparticle changes and gives a bright

Fig. 5. TEM photographs of a) P(PSu-co-CL) 25/75 nanoparticles, b) P(PSu-co-CL) 10/90 nad) schematic presentation of core/shell nanoparticles with encapsulated drug nanocrystals

region which should correspond to an amorphous phase. The shellthickness ranged between 40 and 60 nm (Fig. 5b).

As PPSu is slightly more hydrophilic than PCL, one would expectthat PPSu should form the shell and PCL the nanoparticle core.However, the above results indicate that the copolymer adopt theopposite arrangement. Therefore, a possible explanation for theobserved structure should not be based purely on hydrophilicity vs.hydrophobicity arguments, since, overall, both segments of thecopolymer have a rather hydrophobic nature. Examination of thecrystallization behavior of the two different copolymer segments,shows that PPSu exhibits a very low crystallization rate as compared toPCL [20,37]. Furthermore, from WAXD patterns it was found that onlyPCL can be crystallized in all copolymers, expect in the case of P(PSu-co-CL) 25/75. For this reason it is believed that during solventevaporation and nanoparticle formation, PCL segments of thecopolymers crystallize first, creating the nanoparticle shell. In contrastthe PPSu segment is suppressed into the inner part of the nanoparticleand finally remains there in amorphous state. These copolymers, dueto the high differences in their crystallization rates, are self organizedduring nanoparticle formation in a crystalline shell and an amorphouscore. The drug is encapsulated mainly into the amorphous core in theform of nanocrystals, probably due to the presence of PPSu. Thesenanocrystals appear as dark spotsmainly inside the core nanoparticles(Fig. 5a). The drug crystallization was also verified with WAXD andwill be discussed below. Schematic presentations of these core/shellnanoparticles and the dispersed crystals are depicted in Fig. 5c.

3.4. Evaluation of nanoparticle yield, drug loading and entrapmentefficiency

Table 3 summarizes yield, drug loading, and entrapment efficiencyof P(PSu-co-CL) nanoparticles. These parameters depend mainly onthe composition of the copolymer, and can be affected by factors such

noparticles, c) P(PSu-co-CL) 5/95 nanoparticles revealing the core–shell structure andinto the polymer matrix.

Fig. 7. a) FT-IR spectra of PCL/PPSu nanoparticles with tibolone and b) expanded –OH

182 S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

as molecular weight of the polymer and crystallinity. From the resultsshown in Table 3 it appears that, in these copolyesters, increases in thePPSu weight ratio result in concomitant increases in micelle yield,entrapment efficiency, and drug loading. The only exception seems tobe the sample of P(PSu-co-CL) 5/95 where the drug loading appearsto be lower. There is also an indication that the mean diameter of thenanoparticles tends to decrease with increasing PPSu content,however, this cannot be easily concluded from the present data,since the relatively high values of polydispersity index mean thatactual sizes of different batches are similar and overlapping may beexpected [38].

3.5. Physicochemical characterization of copolymer nanoparticles

3.5.1. Wide Angle Diffractometry (WAXD)Wide Angle X-Ray Diffractrometry was also used in order to study

the physical and crystal state of the drug [39]. In Fig. 6 thecorresponding patterns of tibolone, re-crystallized tibolone, andtibolone-loaded P(PSu-co-CL) nanoparticles are presented. WAXDpattern of pure tibolone showed that the drug is in crystalline statewith characteristic diffraction peaks at 2θ=14.15, 15.3, 16.31, 17.71,18.65, 20.73, 21, 23.74, 24.32, 27.67, 23.6 and 35.37° which correspondto the crystalline pure form known as form II. Re-crystallized tibolonehas a completely different pattern, where characteristic peaksrecorded at 2θ=13.74, 15.26, 16.1, 17.45, 17.81, 18.4, 19.34, 20.32,21.23, 21.67, 23.88, 28.07, 33.56 and 37.37° indicate that tibolone wascrystallized in a different form, known as form I [40]. From WAXDpatterns of the drug-loaded nanoparticles, one can detect thecharacteristic peaks of the copolyesters, mainly PCL, and also themost intense peaks of tibolone at 2θ=15.36, 16.31, and 17.71°. Thisconstitutes strong evidence that the drug exists in a crystalline form,as observed from TEM micrographs. However, because of the lowintensity and the broadness of the characteristic peaks of the drug inthe nanoparticles, it was not possible to identify the exact crystallineform, in which the drug exists. In order to collect some moreinformation on this subject, FT-IR spectra of the drug-loadednanoparticles were obtained.

3.5.2. Fourier Transformation-Infrared Spectroscopy (FT-IR)Tibolone spectra show some characteristic peaks (Fig. 7). At

1714 cm−1 the stretching of carbonyl group of tibolone appears. The –

OH stretching frequencies give a single absorption band at 3492 cm−1

in the case of crystallized tibolone in the initial crystal form II and a

Fig. 6. XRD patterns of pure copolymers, tibolone, re-crystallized tibolone and P(PSu-co-CL) nanoparticles loaded with tibolone.

region of nanoparticles spectra with tibolone.

double absorption band at 3492 and 3410 cm−1 in the case of re-crystallized tibolone (crystal form I) [40].

In the case of the nanoparticle samples, the characteristic peaks ofthe drug confirm that tibolone exists in crystalline form and, inparticular, polymorphous form since the characteristic peaks of bothcrystal forms (I and II) are recorded in all nanoparticles (Fig. 7a).Furthermore, the characteristic peak of crystal form I is shifted from3410 cm−1 at 3446 cm−1 (Fig. 7b) giving an indication that a kind ofphysicochemical interaction may take place, either between thepolymermatrix and the drugmolecules (intermolecular interactions),or between the molecules of drug tibolone (intramolecular interac-tions). This seems to be a more plausible explanation considering theresults obtained by TEM measurements, where the drug was found tobe encapsulated in the polymer matrix in the form of nanocrystals(Fig. 5).

3.6. In vitro drug release behavior

Fig. 8 shows the release profile of tibolone from PCL, PPSu andP(PSu-co-CL) nanoparticles. It is obvious that the drug release is

Fig. 8. Tibolone release profile from copolymer nanoparticles PCL/PPSu and neatpolyesters.

183S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

faster from nanoparticles created from homopolymers thancopolymers. A similar behavior has been previously reported forthe case of PCL/PLLA copolymers [41]. The PPSu nanoparticles hadthe fastest release rate for tibolone, when compared with the othercopolymers and neat PCL. PPSu has a lower melting point than PCLand, almost the lowest melting point compared to the studiedcopolymers (Table 2). Maybe this is the main factor for its highrelease rate, since the experiments were carried out at 37 °C.However, as the ratio of PPSu in the copolymers increases, thedrug release rate seems to decrease. Hence, the copolymer P(PSu-co-CL) 25/75, which exhibits even lower than pure PPSu meltingpoint, has the lowest release rate. From this it can be inferred thatthe melting point of a polymer does not constitute a determiningfactor in predicting a polymer's drug release behavior.

As has been previously reported, the release of a drug fromnanoparticles made of biodegradable polymers is a rather complicatedprocess, where different parameters can drastically affect the releaseprofile. As a result, many different factors may sometimes play acontradictory role. Such factors are polymer degradation rate,molecular weight, crystallinity, glass transition and melting tempera-tures, binding affinity between the polymer and the drug, drugloading capacity, nanoparticle size, the hydrophilicity or hydrophobi-city of the drug, and others [14,42,43].

One of the most decisive factors for the release profile may be thecrystallinity of the polymer matrix. As has been previously reported,the higher the crystallinity of the polymer matrix, the faster therelease of the drug from the nanoparticles [41,44]. This fact is beingexplained by the hypothesis that the high crystallinity leads to theformation of amicrochannel structure [18] that causes faster release ofthe drug. This hypothesis in our case may be able to explain the highrelease rate of tibolone from PCL nanoparticles. Furthermore, as theratio of PPSu increases, the crystallinity is reduced, as shown by theDSC in Table 2. The same trend is followed by the release rate of thedrug. At the same time these polymers do not show the characteristicburst effect [43], a fact that strengthens the hypothesis madepreviously that the drug is dispersed as nanocrystals into the polymermatrix, and mainly in the amorphous core that consists of PPSu. As aresult, water can penetrate easily the microchannels of PCL, however,when it comes in contact with the nanocrystals of the lipophilic drugin the nanoparticle core, its diffusion slows down significantly.However, crystallinity alone cannot explain the release rate of tibolonefrom neat PPSu, which exhibits almost the lowest degree of crystal-linity but the highest release rate.

Another factor that seems to play a key role on the release profile isthe capability of the polymer to incorporate the drug. Many groups havealso stated the fact that the higher the hydrophobic drug concentration

in the nanoparticles, the lower the release rate [14,42,45]. A similarsituation is observed in the case of P(PSu-co-CL) nanoparticles, wherethe fastest release is obtained by the homopolymer PPSu, which has thelowest drug loading while the opposite situation occurs for PCL/PPSu75/25 which has the highest drug loading (cf. Table 3). This is alsoobvious for the samples of P(PSu-co-CL) 2.5/97.5 and 5/95, which havevery similar drug loading and, as a result, similar release rates arerecorded (Fig. 8). When the sample P(PSu-co-CL) 25/75 is comparedwith the sample P(PSu-co-CL) 10/90 it seems that the key factor for therelease rate of the drug is the crystallinity of the sample, as they seemtohave similardrug loading, but the release rate of tibolone in the case ofP(PSu-co-CL) 10/90 is much faster. These results clearly show that anoptimization of drug release properties is possible by adjusting the PCL-to-PPSu ratio.

4. Conclusions

Copolymers of P(PSu-co-CL) having different propylene succinate-to-caprolactone molar ratios were prepared, resulting in materialswith different molecular weights and thermal properties. The degreeof crystallinity in the copolymer was found to decrease withincreasing PPSu content. Moreover, it was shown that thesecopolymers are biocompatible and can be used for the synthesis(through o/w solvent evaporation) of nanoparticles that encapsulatethe drug tibolone. It is found that the nanoparticle size, yield,encapsulation efficiency, and drug loading are mainly affected bythe PPSu content. TEM and WAXD examination revealed that thesenanoparticles are self organized into core/shell structures, driven bydifferences in hydrophobicity and, mainly, physical state (amorphousvs. crystalline, respectively) between the PPSu and PCL segments.More specifically, PCL moieties, which are mainly crystalline, createthe nanoparticle shell during solvent evaporation, while the PPSumoieties are amorphous and form the core. Tibolone is mainlyencapsulated into the shell as nanocrystals. The drug release rate isdirectly dependent on the P(PSu-co-CL) copolymer ratio and is freefrom burst effects.

References

[1] G. Barratt, Review. Colloidal drug carriers: achievements and perspectives, CMLS,Cell. Mol. Life Sci. 60 (1) (2003) 21–37.

[2] P. Couvreur, C. Vauthier, Nanotechnology: intelligent design to treat complexdisease, Pharm. Res. 23 (7) (2006) 1417–1450.

[3] M. Rawat, D. Singh, S. Saraf, S. Saraf, Nanocarriers: promising vehicle for bioactivedrugs, Biol. Pharm. Bull. 29 (9) (2006) 1790–1798.

[4] S. Stolnik, L. Illum, S.S. Davis, Long circulating microparticulate drug carriers, Adv.Drug Deliv. Rev. 16 (2–3) (1995) 195–214.

[5] K. Kataoka, A. Harada, Y. Nagasaki, Block copolymer micelles for drug delivery:design, characterization and biological significance, Adv. Drug Deliv. Rev. 47 (1)(2001) 113–131.

[6] V.P. Torchilin, PEG-based micelles as carriers of contrast agents for differentimaging modalities, Adv. Drug Deliv. Rev. 54 (2) (2002) 235–252.

[7] V.P. Torchilin, V.S. Trubetskoy, Which polymers can make nanoparticulate drugcarriers long-circulating? Adv. Drug Deliv. Rev. 16 (2–3) (1995) 141–155.

[8] G. Orive, A.R. Gascon, R.M. Hernandez, A.D. Gil, J.L. Perdaz, Techniques: newapproaches to the delivery of biopharmaceuticals, Trends Pharmacol. Sci. 25 (7)(2004) 382–387.

[9] D.E. Owens III, N.A. Peppas, Review. Opsonization, biodistribution and pharma-cokinetics of polymeric nanoparticles, Int. J. Pharm. 307 (1) (2006) 93–102.

[10] C. Gryparis, G. Mattheolabakis, D. Bikiaris, K. Avgoustakis, Effect of conditions ofpreparation on the size and encapsulation properties of PLGA–mPEG nanoparticlesof cisplatin E, Drug Deliv. 14 (6) (2007) 371–380.

[11] E. Pierri, K. Avgoustakis, Poly(lactide)–poly(ethylene glycol) micelles as a carrierfor griseofulvin, J. Biomed. Mater. Res. — Part A 75 (3) (2005) 639–647.

[12] Zhang, Y. Hu, X. Jiang, C. Yang, W. Lu, Y.H. Yang, Camptothecin derivative-loadedpoly(caprolactone-co-lactide)-b–PEG-b–poly(caprolactone-co-lactide) nanopar-ticles and their biodistribution in mice, J. Control. Release 96 (1) (2004) 135–148.

[13] Y.I. Jeong, M.-K. Kang, H.-S. Sun, S.-S. Kang, H.-W. Kim, K.-S. Moon, K.-J. Lee, S.-H. Kim,S. Jung, All-trans-retinoic acid release from core–shell type nanoparticles of poly(ε-caprolactone)/poly (ethylene glycol) diblock copolymer, Int. J. Pharm. 273 (1–2)(2004) 95–107.

[14] A. Layre, P. Couvreur, H. Chacun, J. Richard, C. Passirani, D. Requier, J.P. Benoit, R.Gref, Novel composite core–shell nanoparticles as busulfan carriers, J. Control.Release 111 (3) (2006) 271–280.

184 S. Papadimitriou, D. Bikiaris / Journal of Controlled Release 138 (2009) 177–184

NANOMEDICIN

E

[15] Y. Zhang, R.-X. Zhuo, Synthesis and in vitro drug release behaviour of amphiphilictriblock copolymer nanoparticles based on poly(ethylene glycol) and polycapro-lactone, Biomaterials 26 (33) (2005) 6736–6742.

[16] Y. Zhang, C. Wang, W. Yang, B. Shi, S. Fu, Tri-component diblock copolymers of poly(ethylene glycol)–poly(ε-caprolactone-co-lactide): synthesis, characterizationand loading captothecin, Colloid Polym. Sci. 283 (11) (2005) 1246–1252.

[17] K.S. Soppimath, T.M. Aminabhavi, A.R. Kulkarni, W.E. Rudzinski, Biodegradablepolymeric nanoparticles as drug delivery devices, J. Control. Release 70 (1–2)(2001) 1–20.

[18] V.R. Sinha, K. Bansal, R. Kaushic, R. Kumria, A. Trehan, Poly-ε-caprolactonemicrospheres and nanospheres an overview, Int. J. Pharm. 278 (1) (2004) 1–23.

[19] F. Barbato, M.I. La Rotonda, G. Maglio, R. Palumbo, F. Quaglia, Biodegradablemicrospheres of novel segmented poly(ether-ester-amide)s based on poly(ε-caprolactone) for the delivery of bioactive compounds, Biomaterials 22 (11)(2001) 1371–1378.

[20] S.A. Papadimitriou, G.Z. Papageorgiou, D.N. Bikiaris, Crystallization and enzymaticdegradation of novel poly(ε-caprolactone-co-propylene succinate) copolymers,Eur. Polym. J. 44 (7) (2008) 2356–2366.

[21] D.N. Bikiaris, G.Z. Papageorgiou, S.A. Papadimitriou, E. Karavas, K. Avgoustakis,Novel biodegradable polyester poly(propylene succinate): synthesis and applica-tion in the preparation of solid dispersions and nanoparticles of fluvastatin, AAPSPharm. Sci. Tech. 10 (1) (2009) 138–146.

[22] S. Papanikolaou, P. Ruiz-Sanchez, B. Pariset, F. Blanchard, M. Fick, High productionof 1,3-propanediol from industrial glycerol by a newly isolated Clostridiumbutyricum strain, J. Biotechnol. 77 (2–3) (2000) 191–208.

[23] D.N. Bikiaris, G.Z. Papageorgiou, D.S. Achilias, Synthesis and comparativebiodegradability studies of three poly(alkylene succinate)s, Polym. Degrad. Stab.91 (1) (2006) 31–43.

[24] K. Sator, M.O. Sator, P.G. Sator, C. Egarter, J.C. Huber, Effects of tibolone on selectinsin postmenopausal women, Maturitas 53 (2) (2006) 166–170.

[25] H.J. Kloosterboer, Tissue-selectivity: the mechanism of action of tibolone,Maturitas 48 (1) (2004) S30–S40.

[26] M.J. Reeda, H.J. Kloosterboer, Tibolone: a selective tissue estrogenic activityregulator (STEAR), Maturitas 48 (1) (2004) S4–S6.

[27] S.A. Papadimitriou, D. Bikiaris, K. Avgoustakis, Microwave-induced enhancementof the dissolution rate of poorly water-soluble tibolone from polyethylene glycolsolid dispersions, J. Appl. Polym. Sci. 108 (2) (2008) 1249–1258.

[28] C.P. Roupakias, D.N. Bikiaris, G.P. Karayannidis, Synthesis, thermal characteriza-tion, and tensile properties of alipharomatic polyesters derived from 1,3-propanediol and terephthalic, isophthalic, and 2,6-naphthalenedicarboxylic acid,J. Polym. Sci., A, Polym. Chem. 43 (17) (2005) 3998–4011.

[29] G. Seretoudi, D. Bikiaris, C. Panayiotou, Synthesis, characterization and biodegrad-ability of poly(ethylene succinate)/poly(1-caprolactone) block copolymers,Polymer 43 (20) (2002) 5405–5415.

[30] K. Avgoustakis, A. Beletsi, Z. Panagi, P. Klepetsanis, A.G. Karydas, D.S. Ithakissios, PLGA–mPEGnanoparticles of cisplatin: invitronanoparticle degradation, invitrodrug releaseand in vivo drug residence in blood properties, J. Control. Release 79 (1–3) (2002)123–135.

[31] De Haan, Pieter, Immediate-release pharmaceutical dosage form comprisingpolymorphous Tibolone. EP 1 499 278 B1, 2004.

[32] S. Papadimitriou, D.N. Bikiaris, K. Chrissafis, K.M. Paraskeuopoulos, S. Mourtas,Synthesis, characterization, and thermal degradation mechanism of fast biode-gradable PPSu/PCL copolymers, J. Polym. Sci., A, Polym. Chem. 45 (22) (2007)5076–5090.

[33] P.J. Barham, A. Keller, E.L. Otum, P.A. Holmes, Crystallization and morphology of abacterial thermoplastic: poly-3-hydroxybutyrate, J. Mater. Sci. 19 (9) (1984)2781–2794.

[34] G.Z. Papageorgiou, D.N. Bikiaris, Crystallization and melting behavior of threebiodegradable poly(alkylene succinates). A comparative study, Polymer 46 (26)(2005) 12081–12092.

[35] K.A. Athanasiou, G.G. Niederauer, C.M. Agrawal, Sterilization, toxicity, biocompat-ibility, and clinical applications of polylactic acid/polyglycolic acid copolymers,Biomaterials 17 (2) (1996) 93–102.

[36] H. Ge, Y. Hu, X. Jiang, D. Cheng, Y. Yuan, H. Bi, C. Yang, Preparation, characterizationand drug release behaviors of drug nimodipine-loaded poly(ε-caprolactone)–poly(ethylene oxide)–poly(ε-caprolactone) amphiphilic triblock copolymer micelles,J. Pharm. Sci. 91 (6) (2002) 1463–1473.

[37] D.N. Bikiaris, G.Z. Papageorgiou, D.S. Achilias, E. Pavlidou, A. Stergiou, Miscibilityand enzymatic degradation studies of poly(ε-caprolactone)/poly(propylenesuccinate) blends, Eur. Polym. J. 43 (6) (2007) 2491–2503.

[38] M. Gaumet, A. Vargas, R. Gurny, F. Delie, Nanoparticles for drug delivery: the needfor precision in reporting particle size parameters, Eur. J. Pharm. Biopharm. 69 (1)(2008) 1–9.

[39] D. Bikiaris, G.Z. Papageorgiou, A. Stergiou, E. Pavlidou, E. Karavas, F. Kanaze, M.Georgarakis, Physicochemical studies on solid dispersions of poorly-water solubledrugs. Evaluation of capabilities and limitations of thermal analysis techniques,Thermochim. Acta 439 (1–2) (2005) 58–67.

[40] A.J.M.T. Genard Sas, E.M. van Doornum, Pharmaceutical Composition whichContains a Pharmaceutically Suitable Carrier and the Compound Having theStructure (7a,17a)-17-hydroxy-7-methyl-19-nor-17-pregn-5(10)-en-20-yn-3-one, US Patent 5,037,817, 1991.

[41] H. Ge, Y. Hu, S. Yang, X. Jiang, C. Yang, Preparation, characterization and drugrelease behaviors of drug-loaded ε-caprolactone/L-lactide copolymer nanoparti-cles, J. Appl. Polym. Sci. 75 (7) (2000) 874–882.

[42] R. Gref, Y. Minamitake, M.T. Peracchia, V.S. Trebetskoy, V.P. Torchilin, R. Langer,Biodegradable long-circulating polymeric nanospheres, Science 263 (5153)(1994) 1600–1603.

[43] Y. Hu, X. Jiang, Y. Ding, L. Zhang, C. Yang, J. Zhang, J. Chen, Y. Yang, Preparation anddrug release behaviors of nimodipine-loaded poly(caprolactone)–poly(ethyleneoxide)–polylactide amphiphilic copolymer nanoparticles, Biomaterials 24 (13)(2003) 2395–2404.

[44] J.S. Chawla, M.M. Amiji, Biodegradable poly(ε-caprolactone) nanoparticles fortumor targeted delivery of tamoxifene, Int. J. Pharm. 249 (2002) 127–137.

[45] P.L. Soo, L. Luo, D. Mayasinger, A. Eisenberg, Incorporation and release ofhydrophobic probes in biocompatible polycaprolactone-block–poly(ethyleneoxide) micelles: implications of drug delivery, Langmuir 18 (25) (2002)9996–10004.