Nanostructured microspheres produced by supercritical fluid extraction of emulsions

ARTICLE

Nanostructured Microspheres Produced bySupercritical Fluid Extraction of Emulsions

Giovanna Della Porta,1,2 Ernesto Reverchon1,2

1Department of Chemical & Food Engineering, University of Salerno,

Via Ponte Don Melillo 1, 84084 Fisciano-Salerno, Italy; telephone: 39-089-964104;

fax: 39-089-964057; e-mail: [email protected]_MATES, Research Centre for NANOMAterials and

nanoTEchnology at Salerno University, Fisciano-Salerno, Italy

Received 29 October 2007; revision received 23 January 2008; accepted 7 February 2008

Published online 27 February 2008 in Wiley InterScience (www.interscience.wiley.com
). DOI 10.1002/bit.21845
ABSTRACT: The system poly(lactic-co-glycolic) acid/piroxicam (PLGA/PX) was selected, as a model system, toevaluate the effectiveness of supercritical carbon dioxide(SC-CO2) extraction of the oily phase (ethyl acetate) fromoil-in-water emulsions used in the production of polymer/drug microspheres for sustained drug release applications.The influence of process parameters like operating pressureand temperature, flow rate and contacting time betweenthe emulsion and SC-CO2 was studied with respect to themicrosphere size, distribution and solvent residue. Differentpolymer concentrations in the oily phase were also testedin emulsions formulation to monitor their effects ondroplets and microspheres size distribution at fixed mixingconditions. Spherical PLGA microspheres loaded withPX (10% w/w) with mean sizes ranging between 1 and3 mm and very narrow size distributions were obtaineddue to the short supercritical processing time (30 min) thatprevents the aggregation phenomena typically occurringduring conventional solvent evaporation process. A solventresidue smaller than 40 ppm was also obtained at optimizedoperating conditions. DSC and SEM-EDX analyses con-firmed that the produced microparticles are formed by asolid solution of PLGA and PX and that the drug isentrapped in an amorphous state into the polymericmatrix with an encapsulation efficiency in the range of90–95%. Drug release rate studies showed very uniformdrug concentration profiles, without any burst effect, con-firming a good dispersion of the drug into the polymerparticles.

Biotechnol. Bioeng. 2008;100: 1020–1033.

� 2008 Wiley Periodicals, Inc.

KEYWORDS: supercritical carbon dioxide; poly(lactic-co-glycolic) acid; piroxicam; oil-in-water emulsion; micro-spheres; controlled release

Correspondence to: G. Della Porta

1020 Biotechnology and Bioengineering, Vol. 100, No. 5, August 1, 2008

Introduction

Controlled release formulations provide prolonged drugdelivery, maintaining its blood concentration within thera-peutic limits. Controlled release dosage can circumventproblems related to the conventional formulations such as,drug absorption or metabolism, optimize the therapy itselfand improve patient comfort and compliance (Hickey et al.,2002; Kim et al., 2002; King and Patrick, 2000; Meinel et al.,2001).

Drug–polymer microcomposites are the most promisingform for controlled release devices development, especially ifproduced with a narrow particle size distribution (PSD).They can also be used for site specific controlled delivery ofsmall molecular weight drugs and proteins in a large varietyof applications as chemotherapy, cardiovascular disease,hormone therapy and vaccines.

Sustained drug release from biodegradable matrices isobtained by different mechanisms: polymer swelling,diffusion through the polymer, polymer degradation or acombination of these mechanisms (Washington, 1996). Therelease kinetics are governed by microsphere particle sizeand surface morphology, as well as, drug and polymerphysical chemistry. Particularly, sphere size and distributionwill determine the surface area/volume ratio and, thereby,the amount of surface available for drug release. Mono-disperse microspheres may exhibit the most uniform drugrelease avoiding any peak related side effect (Berkland et al.,2002, 2003; Dunne et al., 2000). It was also observed thatthere is an ideal sphere size which provides a desired releaserate. Spheres that are ‘‘too small’’ exhibit poor sustainedrelease efficiency, they also may migrate from the site ofinjection producing undesirable drug release; whereas,spheres that are ‘‘too large’’ may not easily pass througha syringe needle. Thus, the typically poly-disperse samplesgenerated by conventional fabrication techniques must befiltered or sieved to isolate particles within the desiredsize range. Precise size distributions may also allow the

� 2008 Wiley Periodicals, Inc.

preparation of advanced delivery systems formulation thatare not possible using poly-disperse microspheres; forexample, uniform microspheres approximately of 1–5 mmin diameter would be ideal for passive targeting of specificantigen-presenting cells (APCs) such as macrophages anddendritic cells (Akhtar and Lewis, 1997; Evora et al., 1998).Similarly, microspheres of 10–20 mm in diameter could beused to target the tortuous capillary bed of tumor tissues bychemo-embolization (Dass and Burton, 1999).

Microspheres are conventionally prepared using varioustechniques which are the modifications of three basicprocesses: solvent evaporation or extraction of emulsions,phase separation and spray-drying. Solvent evaporation/extraction technology uses a simple vessel/stirrer setup, butmay exhibit difficulties in producing large amounts ofmicrospheres in a robust and well-controlled manner.Solvent evaporation may also require elevate temperaturesor reduced pressures to eliminate the liquid solvent. Solventextraction uses relatively large amounts of a second solventand, then, the mixture of these two solvents has to berecycled. They also require long processing times (severalhours) to be completed and, as a consequence, aggregationphenomena occur between the droplets producing micro-spheres with a larger poly-dispersity respect to thestarting emulsions (Yang et al., 2000). Phase separationby coacervation is frequently impaired by residual solvent orcoacervating agents found in the microspheres; furthermoreit is not well suited for producing particles in the lowmicrometer size range (Freitas et al., 2005). Spray-drying isrelatively simple and of high throughput, but, cannotbe used for highly temperature sensitive compounds;moreover, particle size control is difficult and yields forsmall batches can be reduced. The emulsion evaporation/extraction and spraying approaches have been used both atthe bench and industrial scale (Maa and Hsu, 1996; Tracy,1998). However, large standard deviations are common; asa consequence, in several cases the production of micro-spheres with a controlled and narrow particle size is still achallenge.

It is well known that near to the critical point, smallchanges in temperature or pressure can produce largechanges in the density/solvation ability of fluids; in addition,the lower viscosity and the higher diffusivity of supercriticalfluids with respect to liquid solvents, improve the masstransfer, which is often a limiting factor for the extractionprocesses. Due to these favorable properties, supercriticalfluids (mainly supercritical carbon dioxide, SC-CO2) arecurrently proposed in a wide range of extraction applica-tions (Reverchon and De Marco, 2006), microparticleformation (Della Porta and Reverchon, 2007a; Della Portaet al., 2005, 2006b; Reverchon and Adami, 2006; Reverchonand Antonacci, 2007) and membrane drying technologies(Reverchon et al., 2006). Recently, SC-CO2 has been alsoproposed to extract the organic phase of oil-in-water (O/W)or water-in-oil (W/O) emulsions. Particularly, Zhang et al.(2002, 2004) synthesized nanoparticles of Ag and TiO2

using a water-in-isooctane emulsion; SC-CO2 was used to

eliminate the organic phase and most of surfactants bysolubilization of the continuous phase. Chattopadhyayand Gupta (2003) used a W/O microemulsion for thepreparation of silica nanoparticles by injecting the emulsioninto a batch reactor containing SC-CO2, used to eliminatethe organic solvent and surfactants. Chattopadhyay et al.(2006) also used SC-CO2 to eliminate the organic solventfrom O/W emulsions prepared with some drugs suchas, megestrol acetate and cholesterol acetate or lipidmicrospheres charged with ketoprofen and indometacin.Various authors also proposed the use of SC-CO2 as thecontinuous phase of an emulsion, producing water-in-CO2

emulsions (W/C). Using these w/c reverse micelles noorganic solvents are used and the presence of the SC-CO2

adds further flexibility to the process since the variation ofdensity with pressure and temperature allows to use it as atunable medium for reaction and separation processes (Limet al., 2004; Zhang et al., 2001). Very recently new surfactantsto better stabilize W/C emulsions were also proposed (Ryooet al., 2006). Therefore, the SC-extraction of emulsions hasbeen applied only to few of the many possible systems.Moreover, the control of microsphere size, distribution anddrug loading by varying both emulsion and SC-CO2 processconditions has been scarcely explored.

In this work a tentative to understand the mechanism ofthe SC-CO2 extraction of an O/W emulsion was proposed inorder to produce tailored microspheres for advanceddrug delivery system formulations. Poly-lactic-co-glycolicacid (PLGA) was chosen, as biopolymer, due to its excellentbiocompatibility, its frequent use as biomaterial andbecause PLGA microspheres have been widely investigatedas delivery devices for a variety of therapeutics (Andersonand Shive, 1997). Piroxicam (PX) was chosen amongthe non-steroidal anti-inflammatory drugs (NSAFD), as amodel compound.

The preparation of PLGA microspheres, using SC-CO2 asanti-solvent, has previously proposed in the literature. Wanget al., (2005) proposed a supercritical anti-solvent (SAS)coating process using a suspension of silica particles inacetone/PLGA solution that was sprayed through a capillarynozzle into SC-CO2, which acted as anti-solvent for acetone.The diffusion of SC-CO2 in acetone causes the super-saturation of the polymer solution, leading PLGA pre-cipitation on silica particles. As reported by the authors, insome cases, nice PLGA microspheres were produced;however, one of the main problem of the SAS approachwas particle agglomeration, especially when operating atpressures higher than 110 bar (also because the polymer Tgwas depressed by the presence of SC-CO2). Moreover, thePSD depended on droplets produced during the sprayformation; it means that a quite large PSD was obtained. Theproposed approach of emulsion extraction, can have twomain advantages respect to the SAS process; first, narrowersize distributions can be produced, because they are strictlyrelated to the droplet size and distribution of the startingemulsion; second, particles agglomeration can be preventedby the water/surfactant external phase.

Della Porta and Reverchon: Nano- and Micro-Spheres 1021

Biotechnology and Bioengineering

The proposed process should also overcome several ofthe disadvantages of conventional emulsion evaporationtechnology and produce micro- and submicrospheres with aprecise control of the PSD. The influence of operatingparameters, such as operating pressure, temperature andflow rate was explored and a comparison between themicrosphere size distributions obtained using the solventevaporation and SC-CO2 extraction was proposed. Otheroperating parameters such as, polymer concentration andoily phase viscosity in the emulsion versus emulsion dropletsizes and their influence on the microsphere size werealso studied. Chemical and physical analyses as well asdissolution studies were used to verify the efficiency of themicrocomposite device produced.

Experimental Methods

Materials and Emulsion Preparation

CO2 (99.9% SON, Naples, Italy), polyvinyl alcohol (PVA,Mol wt: 30,000–55,000, Aldrich Chemical Co., Milan, Italy),ethyl acetate (EA, purity 99.9%, Aldrich Chemical Co.),piroxicam (purity 99.9% Sigma-Aldrich Co., Milan, Italy),poly(lactic/glycolic) acid (PLGA, 85:15 Mol wt; 60,000–120,000, Aldrich Chemical Co.) were used as received. Aknown amount of drug–polymer mixture was dissolved intowater-saturated ethyl acetate (EA) to form an organicsolution. Then, this solution was added into a knownamount of the EA-saturated aqueous PVA solution (0.8–1.0% w/w water) to form an emulsion using a high-speedstirrer (mod. L4RT Silverson Machines Ltd, Waterside,Chesham Bucks, UK) operating for 3 min at 2,800 rpm.

Microspheres Preparation by ConventionalSolvent Evaporation

EA was evaporated from the different emulsions undercontrolled and mild vacuum (170 mmHg, rotatingevaporator) for 60 min at 308C. During the evaporation,the emulsions were swept by a continuous nitrogen flow atconstant flow rate (70 L/h).

Figure 1. Description of the apparatus: C¼CO2 cylinder, DP¼ diaphragm pump,

V¼ high pressure vessel, SL¼ liquid separator.

Viscosity Measurements

The determination of PLGA/EA solutions viscosity wasobtained using a controlled stress rheometer (CSL100,Carri-Med, RHEO, Chaylan, France) using a cone and plategeometry. The temperature was fixed at 58C and controlledto within 0.18C with a Peltier device, which was in-corporated into the plate. The surroundings of the platewere saturated with EA.


Apparatus for SC Extraction

The detailed description of the apparatus used is shown inFigure 1. In a typical experiment, 40 g of the O/W emulsionwere placed into the 0.25 dm3 cylindrical stainless steelvessel. SC-CO2 was delivered using a high pressurediaphragm pump (Milton Roy, model Milroyal B, PointSaint Pierre, France) and was bubbled into the extractionvessel at a constant flow rate (0.1–0.5 kg/h), through acylindrical stainless steel dispenser located at the bottomof the extractor. The cylindrical dispenser maximizes thecontact between the two phases during the extraction.Temperature was maintained constant using an air-heatedthermostated oven. A separator located downstream themicrometering valve was used to recover the liquid solventextracted and the pressure in the separator was regulated bya backpressure valve. At the exit of the separator a rotameterand a dry test meter were used to measure the CO2 flow rateand the total quantity of CO2 delivered, respectively. Whenthe extraction process was complete, the suspension wasremoved from the bottom of the extractor vessel for furtherprocessing. Particles were washed several times by cen-trifugation with distilled water, recovered by membranefiltration and dried at air for morphological studies. Aschematic representation of the proposed process for SC-CO2 extraction of O/W emulsions is reported in Figure 2.

Droplets and Microspheres Morphology

The droplets formed in the emulsions were observed usingan optical microscope (mod. BX 50 Olympus, Tokyo, Japan)

Figure 2. Schematic representation of the proposed process for SC-CO2

extraction of O/W emulsion. [Color figure can be seen in the online version of this

article, available at www.interscience.wiley.com.]

equipped with a phase contrast condenser. Field Emission-Scanning Electron Microscope (FE-SEM mod. LEO 1525,Carl Zeiss SMT AG, Oberkochen, Germany) was usedto study the morphology of the collected microspheres.Powders were dispersed on a carbon tab previously stuckto an aluminum stub. Samples were coated with gold-palladium (layer thickness 250 A) using a sputter coater(mod. 108 A, Agar Scientific, Stansted, UK).

Droplets and Microspheres Size Distributions

Droplet size distributions (DSDs) and PSDs were measuredby static light scattering (mod. Mastersizer S, MalvernInstruments Ltd, Worcestershire, UK). The MastersizerS software uses Mie theory to produce an optimal analysis ofthe light energy distribution and to obtain the sizedistribution of the particles. Analyses were performed justafter the preparation of emulsions and of microspheresuspensions using several milligrams of each sample(corresponding to more than one million of droplets orparticles) and repeated ten times.

Solvent Residue Analysis

At the end of each run, the EA content in the collectedsuspensions was analyzed to determine the efficiency ofsolvent removal from the emulsion. The EA residue wasevaluated using a head space sampler (mod. 50 Scan,Hewlett Packard, Palo Alto, CA) coupled to a gaschromatograph interfaced with a flame ionization detector(GC-FID, mod. 6890 Agilent Series, Agilent Technologies,Inc., Wilmington, DE). EA was separated using a fused-silicacapillary column 30 m length, 0.25 mm internal diameter,0.25 mm film thickness (mod. DB-1, J&W, Folsom, CA). GCconditions were: oven temperature at 408C for 8 min. Theinjector was maintained at 1808C (split mode, ratio 1:1) andhelium was used as the carrier gas (7 mL/min). Head spaceconditions were: equilibration time 60 min at 1008C,pressurization time 2 min, loop fill time 1 min. Head space

samples were prepared in 10 mL vials filled with 4 mL ofsuspension. Analyses were performed on each sample inthree replicates.

Drug Dispersion in the Microspheres

Drug dispersion in the polymer microspheres wasevaluated by Differential Scanning Calorimetry (DSC mod.TC11, Mettler, Toledo, Inc., Columbus, OH) comparing thethermograms of the microspheres with the ones of purecomponents (PLGA and PX). Temperature and enthalpy offusion were calibrated with indium standard materials(melting point 156.68C). Five milligrams of sample (þ0.5)was accurately weighed, crimped in an aluminum panand heated from 25 to 4008C under a nitrogen purge at108C min�1.

Drug dispersion in the microspheres was also evaluatedusing an Energy Dispersive X-Ray analyzer (EDX mod.INCA Energy 350, Oxford Instruments, Witney, UK), usingthe signal of the sulfur atoms that are present only in thedrug. Before the evaluation of the elemental composition,the samples were coated with chromium (layer thickness150 A) using a turbo sputter coater (mod. K575X, EmiTechAshford, Kent, UK).

Drug Loading

PX loading was determined by dissolving a known mass (2.5 mg)of microspheres in 1 mL of 0.25 M sodium hydroxide.Samples were rotated for at least 24 h at 10 rpm to ensurethe complete dissolution of the polymer. Blank (PX free)microspheres of the same size were treated identically.The concentration of PX in the resulting solution wasdetermined by an UV–vis spectrophotometer (mod. Cary50, Varian, Palo Alto, CA) measuring the absorbance at276 nm in a quartz cuvette and then subtracting absorbancevalues obtained for the blank microspheres. The measure-ments were performed in three replicates and a goodreproducibility was observed.

Drug Release

Microspheres (5–10 mg) containing PX were pre-suspendedin 1 mL of distilled water and charged into dialysis sack.Then the PX release profiles were determined in 250 mL ofphosphate-buffered saline (PBS, pH 7.4) containing 0.5% ofTween-20 continuously inverted at 10 rpm in a 378Cincubator to maintain adequate sink conditions. The addi-tion of 0.5% Tween-20 was necessary to enhance the PXsolubility in the release medium to 100 mg/mL, as reportedby Berkland et al. (2003). PX released was determined bymeasuring the absorbance at 276 nm (UVvis mod. Cary 50,Varian).



Results and Discussion

Process Parameters Optimization

The experiments performed were aimed at studying thecharacteristics of the extraction process through theselection of opportune operating pressures and tem-peratures, SC-CO2 flow rates and processing times for thecomplete solvent removal and for the production of welldefined microparticles of controlled size. The emulsionstability during the process and the influence of some ofemulsification process parameters on the particles dimen-sions were also evaluated. The emulsion composition wasfixed at an O/W ratio of 20:80. The oily phase contained 10%w/w of PLGA in EA and the water phase contained PVA0.8% w/w, as surfactant. The droplets produced using thisformulation were stable, non-coalescing and exhibited, bylaser scattering measurement, a mean droplet diameter(MD) of 5 mm and a standard deviation (SD) of 3 mm, asillustrated in the optical microscope image reported inFigure 3.

Operating pressure and temperature conditions wereselected to enhance the extraction of the oily dispersed phaseof the emulsion, but also, to avoid drug or polymer losses bydissolution in SC-CO2 and to avoid the emulsion loss bywashing out in the SC-CO2 stream. High-pressure vapor–liquid equilibria (VLE) of the binary system EA/CO2 definethe conditions at which it is possible the formation of ahomogeneous mixture that maximizes the solvent extrac-tion from the emulsion. This data can be found in theliterature and it is reported that EA/CO2 mixture criticalpressure at 388C is of 85 bar with a CO2 molar fraction of0.9 (Della Porta et al., 2006a; Smith et al., 1998). At these

Figure 3. Optical microscope image of a O/W emulsion (20:80); oil phase formed

by EA containing PLGA 10% w/w (with PX 10% w/w of PLGA); water phase contains

PVA 0.8% w/w. [Color figure can be seen in the online version of this article, available

at www.interscience.wiley.com.]


operating conditions water is only slightly soluble in SC-CO2 (King et al., 1992; Wiebe and Gaddy, 1940); whereas,EA is fully miscible. As a consequence, operating pressuresexplored in this work were selected in the range 85–150 bar;whereas, temperature was fixed at 388C. Temperatureshigher than 388C were avoided due to the vicinity of PLGAglass transition that occurs at 408C (De and Robinson,2004). At operating temperatures higher than 388C andpressures higher than 100 bar, microspheres coalescence wasobserved, as illustrated in the SEM image reported inFigure 4. This behavior is probably due to a modification ofthe surfactant/organic phase interaction or to a change inthe emulsion properties leading to a droplets coalescence.Very well defined and non-coalescing microsphereswere produced in the pressure range of 85–100 bar at388C, as illustrated in the SEM image reported in Figure 5. Inthis case the particles showed a MD of 3.5 mm with a SD of2 mm.

The fact that SC-CO2 can produce ‘‘swelling’’ and ‘‘Tglowering’’ effects on PLGA is well known in the literature(Koushik and Kompella, 2004; Liu and Tomasko, 2007;Young and Johnston, 1999), confirming that PLGA micro-spheres can be quite unstable in the presence on SC-CO2.Nevertheless, in the proposed process, the presence of waterplus surfactant, in the external phase, may prevent PLGAmicrospheres coalescence or swelling, in a certain rangeof pressure and temperatures conditions. The coalescenceprevention is one of the important benefit of the proposedprocess.

Increasing SC-CO2 flow rates from 0.1 to 0.5 kg/h anincrease of the extraction rate of the organic solventphase was observed; however, no significant variations ofthe microspheres morphology and of the mean PSs weremeasured. This result confirms that the PS is influencedmore by the nature of the emulsion than by the mass transfer

Figure 4. SEM image of PLGA microspheres collapsed when an operating

temperature of 408C was used.

Figure 5. SEM image of PLGA microspheres obtained operating at 80 bar and 388C, with a SC-CO2 flow rate of 0.5 kg/h for 30 min from a O/W emulsion 20:80; oil phase formed

by EA containing PLGA 10% w/w; water phase contains PVA 0.8% w/w.

conditions. SC-CO2 flow rates higher than 0.5 kg/h inducedthe emulsion wash out from the extraction vessel and partof the water was lost in the downstream separator. As aconsequence, the best operating conditions in our apparatuswere found at 80 bar and 388C, with a SC-CO2 flow rate of0.5 kg/h for 30 min. At these operating conditions, SC-CO2

is an excellent solvent for EA and, contacting the aqueousphase of the emulsion, extracts the EA from the continuousphase and causes the progressive release of EA from thedroplets. It is also possible that SC-CO2 can diffuse intothe solvent droplets leading the solubilization of theorganic phase. Both processes are very fast and causepolymer supersaturation and nucleation inside the dropletsresulting in the formation of spherical polymer–drugmicrospheres.

able I. Laser scattering measured distribution of microspheres obtained

y solvent evaporation (SE) and by supercritical extraction (SC-CO2)

arting from the same emulsion prepared with 5% PLGA and 0.5% PX

the oily phase.

Droplets size distribution

Microspheres size

distributions

SC-CO2 SE

10 mm 1.33 0.88 0.93

50 mm 2.27 1.52 1.60

90 mm 3.40 2.25 2.38

100 mm 4.90 3.54 6.56

The droplets size distribution is also reported for comparison.

Comparison With Solvent Evaporation

A comparison between the characteristics of the micro-spheres obtained by SC-CO2 extraction at the previouslydefined process conditions and by conventional solventevaporation technique was performed, starting from thesame emulsion containing the 5% of PLGA and 0.5% of PXin the oily phase. Microspheres obtained with SC-CO2

showed a PSD narrower that the ones obtained by theconventional process, as illustrated in Table I, where thedistributions obtained by SC-CO2 extraction and by solventevaporation, are reported. Particularly, both PSDs showalmost the same D50 (1.52 and 1.60 mm for the SC-CO2 and

SE, respectively) and D90 (2.25 and 2.38 mm for the SC-CO2

and SE, respectively), but, looking at the D100 (3.54 and6.56 mm for the SC-CO2 and SE, respectively) is clear thatthe microspheres obtained by solvent evaporation alsoinclude a 10% of particles from 3 to 6 mm that are notpresent in the SC-CO2 sample. The observed behavior canbe justified considering the faster precipitation routeobtained in SC-CO2 processing (only 30 min against 8 h)that may prevent droplet coalescence or aggregationphenomena typically occurring during the solvent evapora-tion process.

Moreover, at the optimized SC-CO2 extraction condi-tions, the EA level in the microsphere suspension is lowerthan 40 ppm and about 30 g of CO2 were used for each gramof EA removed. The very low solvent residue is another

Tb

st

in

D

D

D

D



Table II. Laser scattering measured distribution and standard deviations of

the droplets and microspheres produced at different PLGA concentrations

in the oily phase.

PLGA (%) w/w

contained in EA

Droplets SC-microspheres

1 2.5 5 10 1 2.5 5 10

MS (mm) 1.3 1.9 3.02 5.02 0.89 1.26 2.07 3.46

SD (mm) 0.74 1.09 1.74 2.9 0.51 0.73 1.19 1.99

CV (%) 56.9 57.3 57.6 57.7 57.3 57.9 57.4 57.5

D 10 (mm) 0.56 1.02 1.33 1.62 0.26 0.70 1.14 1.16

D 50 (mm) 0.84 1.62 2.27 4.84 0.56 1.02 1.95 3.10

D 90 (mm) 1.18 2.32 3.40 9.45 0.96 1.38 2.88 6.19

MS, mean size; SD, standard deviation; CV, coefficient of variation.

important advantage of the proposed process with respect tothe conventional solvent evaporation that showed an EAresidual content of 500 ppm. The very high solventresidue obtained can be explained in the hypothesis thatin the conventional process as long as EA is evaporatedfrom the aqueous phase of the emulsion, a shift in theemulsion equilibrium is caused, that leads to the diffusionof the organic solvent from the emulsion droplets to thecontinuous phase. As a consequence, the maximum amountof EA that can be evaporated from water is related to theVLE behavior of the water/EA mixture at the operatingtemperature and pressure. Therefore, a lower solvent residueis very difficult to be obtained by this technology (Yang et al.,2000).

Particle Size Control

The PS is mainly related to the emulsion droplet sizeand, therefore, will depend on the method of emulsionformulation. Therefore, different emulsions were preparedvarying the PLGA concentration in EA from 1%, 2.5%,5%, and 10% w/w; whereas, the PX concentrationwas maintained fixed to 10% w/w (PX/PLGA). The DSDsobtained are illustrated in Figure 6. The diagram shows that,fixing all the other parameters, larger droplets are obtainedwhen the polymer concentration in the oily phase isincreased; for example, mean droplet sizes from 1.3 mm (SD0.74) to 5.02 mm (SD 2.9) were obtained for a PLGAconcentration of 1% and 10%, respectively. The distributiondata and standard deviations of the diameter of the emulsiondroplets obtained at different PLGA/PX concentrations are

Figure 6. DSDs of O/W emulsion (20:80); oil phase formed by EA containing

PLGA concentration from 1% to 10% w/w; water phase contains PVA 0.8% w/w, as

surfactant.


reported in Table II. According to Desgouilles et al. (2003),the observed behavior can be related to the increase of theoily phase viscosity when the polymer–drug concentration isincreased. This hypothesis was confirmed measuringthe viscosities (Pa s) of the different EA/PLGA/PX solutionsat different concentrations. The viscosity data are reportedin Table III as a function of the weight concentrationof the polymer–drug in EA (% w/w) and confirm a quasi-linear relationship between PLGA/PX content and solutionviscosity in the range of compositions explored. Theincrease of the emulsion droplets size with polymer–drugconcentration produces a significant increase in PSDs ofthe obtained microspheres, as illustrated in Figure 7a–dwhere SEM images, taken at the same enlargement, of themicrospheres obtained at different PLGA concentrationsare reported. From these SEM images is also clear that, in allcases, the microspheres produced are spherical and notagglomerated. PSDs obtained by laser scattering analysisare reported in Figure 8 and confirm the general trendevidenced from Figure 7a–d. The MD varied from 0.89 mm(SD 0.51 mm), to 2.07 mm (SD 1.19 mm) and 3.46 mm (SD1.99 mm) when the PLGA concentration was varied to1%, 5%, and 10%, respectively. These results confirm thepossibility of PLGA microspheres production with a specificPSD varying the polymer concentration in the oily phase ofthe emulsion. All the distribution data and standarddeviations of the microspheres produced at differentconcentrations are also reported in Table II. In the sametable also the coefficient of variation (CV) of thedistributions it is reported. CV is defined as the ratiobetween the standard deviation and the mean size ofthe same distribution and it is expressed in percentage(CV¼ SD/MS %). It gives an indication of the sharpness ofthe distribution, since put the variability in relation to the

Table III. Variation of the solution viscosity (Pa s) as a function of the

PLGA concentration in EA (PX 10% of PLGA w/w).

PLGA (%) w/w Viscosity (Pa s)

1 0.0078

2.5 0.0123

5 0.0325

10 0.0842

magnitude of the measured diameters. In Table II the CVvalues of the droplets distributions are equal to the CV of thecorresponding microspheres and the CV of the particles isthe same, whatever the polymer concentrations used. Thisdata means that the particle distributions maintain the samedispersion around the mean value; that is, the polydispersityof the microspheres distributions is always the same. Theobserved result is not common since the polydispersity of

Figure 7. a–d: SEM images of PLGA microspheres obtained varying the polymer c

respectively, (PX is 10% w/w of PLGA). Operating conditions: 80 bar and 388C, with a SC

the particles conventionally tends to increase with theparticle mean size.

Looking at the distribution data is also evident that themean sizes of the microspheres are always smaller and thedistributions narrower than the droplet mean sizes fromwhich they were generated. This behavior is better evidencedin Figure 9a and b, where the cumulative distribution of theemulsion droplets and of the related suspensions obtained

oncentration in EA from 1% w/w (a), 2.5% w/w (b), 5% w/w (c), and 10 (d) % w/w,

-CO2 flow rate of 0.5 kg/h for 60 min.



Figure 7. (Continued )

for different PLGA concentrations of 1 and 2.5% are plottedtogether. The shrinking factor (SF) defined as the ratiobetween the MD of the microspheres in suspension and theMD of the droplet in emulsion was measured for all thedifferent PLGA/PX concentrations explored in the oilyphase and the obtained values are reported in Table IV. Inthe case of the microspheres produced by SC-CO2, it isinteresting to observe that the SF value is of about 0.65 in allcases, indicating that the particles produced are almost


always the 35% smaller than the original droplets. DifferentSFs were measured in the case of samples produced bysolvent evaporation and, in some cases, the microspheremean sizes were larger than the original mean droplet sizesdue to aggregation phenomena (see Table IV). These resultssuggest that, departing from an emulsion, the PSD of theproduced particles is mainly related to the emulsion dropletsize; but, also the solvent elimination process can play arelevant role in determining the final size of the produced

Figure 8. PSDs of PLGA microspheres obtained from O/W emulsion (20:80); oil

phase formed by EA containing PLGA in concentrations from 1% to 10% w/w (PX is

10% w/w of PLGA); water phase contains PVA 0.8% w/w, as surfactant. Operating

conditions: 80 bar and 388C, with a SC-CO2 flow rate of 0.5 kg/h for 60 min.

microspheres. In this sense, SC-CO2 extraction process islargely preferable with respect to solvent evaporationbecause no droplet coalesce was observed.

The shrinking of droplets during particles formation, alsosuggests that the microspheres are not empty and that the

Figure 9. a and b: Cumulative diameter distributions of the droplets and of the relative

PLGA in concentrations of 1% w/w (a) and 2.5 (b) % w/w (PX is 10% w/w of PLGA); water pha

SC-CO2 flow rate of 0.5 kg/h for 60 min.

drug could be uniformly dispersed inside. This observationwas confirmed by SEM analysis on cut particles, reportedin Figure 10, where is evident the microspheres are full ofmaterial.

Drug Loading and Release Rates

The presence of PX inside the microspheres can bequalitatively ascertained by visual observation: the producedpowder and PX alone are light yellow; whereas, purePLGA microparticles are white. However, to evaluate the PXspatial distribution inside the microcomposite particles, theelemental composition of microspheres was studied byEnergy Dispersive X-Ray (EDX) analyzer integrated inthe SEM apparatus. PX molecule (C15H13N3O4S) contains asulfur atom (S), that is not present in the polymer; therefore,it can be used to indicate the location of the drug in thepowder. In Figure 11a a SEM image of the microsphereswith the elemental maps of carbon (green) and sulfur(red), is reported. The sulfur is uniformly spread overall the microspheres and the maps of the two elementspractically overlap; that is, the drug and the polymer donot precipitate separately, but co-precipitate into the micro-particles forming a solid solution of the two components.Figure 11b reports the quantitative data of the sulfur atomwith respect to the other elements. The quantitative analysisindicates the presence of almost 10% of sulfur respect to theoverall carbon atoms identified in the sample, that confirmsthe efficiency of drug loading in the produced microspheresand gives the indication that PX is nanodispersed in thePLGA structure.

microspheres obtained from a O/W emulsion (20:80); oil phase formed by EA containing

se contains PVA 0.8% w/w, as surfactant. Operating conditions: 80 bar and 388C, with a



Table IV. Shrinking factor (SF) of microsphere (MDS, mean diameter

suspensions) with respect to droplets (MDE, mean diameter emulsions)

measured at different PLGA concentration in EA (PX 10% of PLGA w/w).

PLGA (%) w/w

SF (MDS/MDE)

SC-CO2 SE

1 0.68 0.82

2.5 0.66 0.98

5 0.68 0.70

10 0.68 1.2

Thermal analysis (DSC) was performed on SC-micro-spheres containing the 10% w/w of PX, on the physicalmixture prepared with the same PX content in PLGA andon raw PX and PLGA, for comparison. The obtainedthermograms are reported in Figure 12. PLGA shows a glasstransition that remains relatively constant (onset 39.988C;endset 45.328C) in all the observed samples and a polymerdecomposition that occurs above 3008C; raw PX shows anendothermic peak at about 2008C, due to the melting ofthe crystals (Drebushchak et al., 2006). PLGA/PX physicalmixture (PX 10% w/w) shows two thermal events related tothe two components: melting of the crystalline PX at 2008Cand the polymer decomposition above 3008C; whereas, theproduced microspheres exhibit only the PLGA degradationevent. The absence of the PX endothermic peak in thesamples produced by SC-CO2 extraction confirms theformation of an amorphous PX intimate nanodispersioninside PLGA microspheres.

Figure 10. SEM image of cut microsph


Summarizing, DSC and SEM-EDX analysis confirmedthat the produced microparticles are formed by a solidsolution of PLGA and PX, and that the drug is entrapped inan amorphous state into the polymeric matrix, thus forminga nanostructured microparticle. However, none of the pre-vious analysis is able to determine the amount of the drugcharged and how it will be released from the polymer.Therefore, drug loading and release tests were performedon two batches of microspheres with different mean sizes(1.3 mm with SD¼ 0.73 and 3.5 mm with SD¼ 2) tomonitor the efficiency in drug release of the microcompositeparticles produced. Theoretical PX load in the two batchestested was maintained always at 10% w/w; the measuredmean encapsulation efficiency was of 90% and 95%,respectively. A smaller encapsulation efficiency was mea-sured in the microspheres with a smaller mean size. One ofthe reasons why we obtained high encapsulation efficiency isthat PX solubility in water is less than 50 mg/mL at 258C andits log POW (at 258C) is reported to be 1.6 (Giaginis et al.,2007). However, we have also to take into account thatthe SC-CO2 emulsion extraction is a very fast process, ifcompared to the conventional SE; therefore, a higherencapsulation efficiency can be expected, since PX has lesstime to migrate in the continuous phase.

The obtained drug release profiles are reported inFigure 13; in the same figure the dissolution rate of purePX is also reported, for comparison. The release rate isinfluenced by the drug spatial distribution into the particles;indeed, if the drug is concentrated on the particle surface, a‘‘burst effect’’ occurs followed by a secondary slower release

eres revealing that the particles are full.

Figure 11. a and b: EDX image of PLGA microspheres charged with 10% w/w of PX, obtained operating at 80 bar and 388C, with a SC-CO2 flow rate of 0.5 kg/h for 60 min: red

points represent the sulfur atoms, green points represent the carbon atoms (a); EDX quantitative determination of the sample (b). [Color figure can be seen in the online version of

this article, available at www.interscience.wiley.com.]

rate; if the drug is concentrated in the core of the particle, anearly zero initial release occurs (delayed release), followedby a secondary faster release once the dissolution mediumhas reached the core. Finally, if the drug is uniformlydistributed in the polymeric matrix, neither burst effect nordelayed release occurs (Washington, 1996).

The release curves reported in Figure 13 show that PLGAmicrospheres produce a smooth PX release profile withoutany burst effect, thus confirming the good dispersion of PXinto the particles and the formation of a drug/polymersolid solution. A moderate faster release was observed forthe smaller microspheres with respect to larger particles;particularly, after the first 800 min of testing the 43% of PXwas released from the particles with a MD of 1 mm, whereas,the 37% was released from the ones with a MD of 3 mm. The

pure PX was dissolved in 350 min. The observed behaviorcan be due to several factors such as, smaller microspheresoffer higher specific surface area and shorter diffusiondistance that influences drug diffusion, as well as, PLGAdegradation kinetics.

Conclusions

The majority of papers on emulsions and SC-CO2 is relatedto W/O emulsions and to the extraction of the continuoussolvent phase and the surfactants by supercritical solvent. Inthis work, SC-CO2 has confirmed its ability to operate as asolvent also in the case in which the emulsion is an O/W andthe organic solvent to be extracted is the dispersed phase.



Figure 12. DSC traces of untreated PLGA and PX, their physical mixture (PLGA

with 10% PX w/w) and SC-extracted microspheres (PLGA with 10% PX w/w).

Indeed, in the system studied, the SC-CO2 is an excellentsolvent for EA upon contact with the aqueous phase ofthe emulsion leading to rapid diffusion of solvent from theemulsion droplets and probably by diffusion/solubilization

Figure 13. In vitro PX release profiles from PLGA microspheres with different

mean diameters (MD) and charged with 10% of PX w/w.


into the solvent droplets. The process is faster than theconventional solvent evaporation resulting in the precipita-tion of spherical PLGA/PX nanostructured microsphereswith a narrower PSDs respect to the ones produced byconventional solvent evaporation. Indeed, the proposedprocess will prevents droplets coalescence or aggregationphenomena that typically occur during the solvent evapo-ration process. Therefore, it can be used to producemicrospheres for controlled drug delivery system formula-tion with a drug well dispersed into the polymeric matrix, asconfirmed by the PX release profiles.

A continuous operating process is also under develop-ment using a packed bed continuous tower in order toovercome the limits of batch processing (Reverchon andDella Porta, 2007b).

References

Akhtar S, Lewis KJ. 1997. Antisense oligonucleotide delivery to cultured

macrophages is improved by incorporation into sustained-release

biodegradable polymer microspheres. Int J Pharm 151:57–67.

Anderson JM, Shive MS. 1997. Biodegradation and biocompatibility of PLA

and PLGA microsferes. Adv Drug Deliv Rev 28:5–24.

Berkland C, King M, Cox A, Kim K, Pack D. 2002. Precise control of PLG

microsfere size provides enhanced control of drug release rate.

J Control Release 82:137–147.

Berkland C, Kim K, Pack D. 2003. PLG microsphere size controls drug

release rate through several competing factors. Pharm Res 20(7):1055–

1062.

Chattopadhyay P, Gupta RB. 2003. Supercritical CO2-based formation silica

nano-particles using water-in-oil microemulsions. Ind Eng Chem Res

42:465–472.

Chattopadhyay P, Huff R, Shekunov BY. 2006. Drug encapsulation using

supercritical fluid extraction of emulsions. J Pharm Sci 95:667–679.

Dass CR, Burton MA. 1999. Microsphere-mediated targeted gene therapy of

solid tumors. Drug Deliv 6:243–252.

De S, Robinson DH. 2004. Particle size and temperature effect on the

physical stability of PLGA nano-spheres and micro-spheres containing

Bodipy. AAPS Pharm Sci Tech 5(4): article 53.

Della Porta G, Reverchon E. 2007a. Supercritical fluid-based technologies

for particulate drug delivery. In: Ravi Kumar MNV, editor. Handbook

of particulate drug delivery. New York: American Scientific Pub. ISBN

1-58883-079-9. p 35–59.

Della Porta G, De Vittori C, Reverchon E. 2005. Supercritical assisted

atomization: A novel technology for micro-particles preparation of an

asthma-controlling drug. AAPS Pharm Sci Tech 6(3): article 52.

Della Porta G, Volpe MC, Reverchon E. 2006a. Supercritical cleaning of

rollers for printing and packaging industry. J Supercrit Fluids 37:409–

416.

Della Porta G, Ercolino SF, Parente L, Reverchon E. 2006b. Corticosteroids

micro-particles produced by supercritical assisted atomization: Process

optimization, product characterization and ‘‘in vitro’’ performance.

J Pharm Sci 95(9):2062–2076.

Desgouilles S, Vauthier C, Bazile D, Vacus J, Grossiord JL, Veillard M,

Couvreur P. 2003. The design of nanoparticles obtained by solvent

evaporation: A Comprehensive Study. Langmuir 19:9504–9510.

Drebushchak VA, Shakhtshneider TP, Apenina SA, Medvedeva AS, Safro-

nova LP, Boldyrev VV. 2006. Thermoanalytical investigation of drug

excipient interaction. Part II. Activated mixtures of piroxicam with

cellulose and chitosan. J Thermal Anal Calorimetry 86(2):303–309.

Dunne M, Corrigan OI, Ramtolla Z. 2000. Influence of particle size and

dissolution conditions on the degradation properties of polylactide-co-

glycolide particles. Biomaterials 21:1659–1668.

Evora C, Soriano I, Rogers RA, Shakesheff KM, Hanes J, Langer R. 1998.

Relating the phagocytosis of microparticles by alveolar macrophages to

surface chemistry: The effect of 1,2-dipalmitoylphosphatidylcholine.

J Control Release 5:143–152.

Freitas S, Merkle HP, Gander B. 2005. Microencapsulation by solvent

extraction/evaporation: Reviewing the state of the art of microsphere

preparation process technology. J Control Release 102:313–332.

Giaginis C, Theocharis S, Tsantili-Kakoulidou A. 2007. Octanol/water

partitioning simulation by reversed-phase high performance liquid

chromatography for structurally diverse acidic drugs: Effect of

n-octanol as mobile phase additive. J Chromatogr 1166:116–125.

Hickey T, Kreutzer D, Burgess DJ, Moussy F. 2002. Dexamethasone/PLGA

microspheres for continuous delivery of an anti-inflammatory drug for

implantable medical devices. Biomaterials 23:1649–1656.

Kim H, Lee H, Park TG. 2002. Pegylated recombinant human epidermal

growth factor for sustained release from biodegradable PLGA micro-

spheres. Biomaterials 23:2311–2317.

King TW, Patrick CW. 2000. Development and in vitro characterization of

vascular endothelial growth factor (VEGF)-loaded poly(lactic-co-gly-

colic acid)/poly(ethylene glycol) microspheres using a solid encapsula-

tion/single emulsion/solvent extraction technique. J Biomed Mater Res

51:383–390.

King MB, Mubarak A, Kim JD, Bott TR. 1992. The mutual solubilities of

water with supercritical and liquid carbon dioxide. J Supercrit Fluids

5:296–302.

Koushik K, Kompella K. 2004. Preparation of large porous deslorelin-PLGA

microparticles with reduced residual solvent and cellular uptake using a

supercritical carbon dioxide process. Pharm Res 21(3):524–535.

Lim KT, Hwang HS, Ryoo W, Johnston KP. 2004. Synthesis of TiO2

nanoparticles utilizing hydrated reverse micelles in CO2. Langmuir

20:2466–2471.

Liu D, Tomasko DL. 2007. Carbon dioxide sorption and dilation of

poly(lactide-co-glycolide). J Supercrit Fluids 39:416–425.

Maa YF, Hsu C. 1996. Microencapsulation reactor scale-up by dimensional

analysis. J Microencapsul 13(1):53–66.

Meinel L, Illi OE, Zapf J, Malfanti M, Merkle HP, Gander B. 2001.

Stabilizing insulin-like growth factor-I in poly (lactide-co-glycolide)

microspheres. J Control Release 70:193–202.

Reverchon E, Adami R. 2006. Nanomaterials and supercritical fluids.

J Supercrit Fluids 37:1–22.

Reverchon E, Antonacci A. 2007. Drug–polymer microparticles produced

by supercritical assisted atomization. Biotech Bioeng 97(6):1626–1637.

Reverchon E, De Marco I. 2006. Supercritical fluid extraction and fractio-

nation of natural matter. J Supercrit Fluids 38:146–166.

Reverchon E, Della Porta G. 2007b. Processo continuo per la produzione di

microsfere mediante liquidi espansi. Italian Patent no. SA02007/

A000004. PCT submitted 2008.

Reverchon E, Cardea S, Schiavo Rappo E. 2006. Production of loaded

PMMA structures using the supercritical CO2 phase inversion process.

J Membr Sci 273:97–105.

Ryoo W, Webber SE, Bonnecaze RT, Johnston KP. 2006. Long-ranged

electrostatic repulsion and crystallization of emulsion droplets in an

ultralow dielectric medium supercritical carbon dioxide. Langmuir 22:

1006–1015.

Smith RL, Yamaguchi T, Sato T, Suzuki H, Arai K. 1998. Volumetric

behavior of ethyl acetate, ethyl octanoate, ethyl laurate, ethyl linoleate

and fish oil ethyl esters in the presence of supercritical CO2. J Supercrit

Fluids 13:29–36.

Tracy MA. 1998. Development and scale-up of a microsphere protein

delivery system. Biotechnol Prog 14:108–115.

Wang Y, Pfeffer R, Dave R, Enick R. 2005. Polymer encapsulation of fine

particles by a supercritical antisolvent process. AIChE J 51(2):440–

455.

Washington C. 1996. Drug release from microparticulate system. In:

Benita S, editor. Microencapsulation, methods and industrial applica-

tion. New York: Marcel Dekker Inc. ISBN 0-8247-9703-5. p 155–

181.

Wiebe R, Gaddy V. 1940. The solubility of carbon dioxide in water at

various temperatures from 12 to 408C and at pressures to 500 atmo-

spheres. Critical phenomena. J Am Chem Soc 62:815–817.

Yang YY, Chia HH, Chung TS. 2000. Effect of preparation temperature on

the characteristics and release profiles of PLGA microspheres contain-

ing protein fabricated by double-emulsion solvent extraction/evapora-

tion method. J Control Release 69:81–96.

Young TJ, Johnston KP. 1999. Encapsulation of lysozyme by precipitation

with a vapor-over-liquid antisolvent. J Pharm Sci 88:640–650.

Zhang H, Lu J, Han B. 2001. Precipitation of lysozyme solubilized in reverse

micelles by dissolved CO2. J Supercrit Fluids 20:65–71.

Zhang J, Han B, Liu J, Zhang X, He J, Liu Z, Jiang T, Yang G. 2002. Recovery

of silver nano-particles synthesized in AOT/C12E4 mixed reverse

micelles by anti-solvent CO2. Chem Eur J 8:3879–3883.

Zhang J, Han B, Liu J, Zhang X, Yang G, Zhao H. 2004. Size tailoring of

ZnS nanoparticles synthesized in reverse micelles and recovered by

compressed CO2. J Supercrit Fluids 30:89–95.



Date post:	10-Nov-2023
Category:	Documents
Upload:	independent
View:	1 times
Download:	0 times

Nanostructured microspheres produced by supercritical fluid extraction of emulsions

Documents