Lipid nanoparticles for brain targeting II. Technological characterization

Page 1

Lc

PAa

b

ARRAA

KSNBRFS

1

wapsldgip

msPPm

UT

0h

Colloids and Surfaces B: Biointerfaces 110 (2013) 130– 137

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces

jou rn al hom epage: www.elsev ier .com/ locate /co lsur fb

ipid nanoparticles for brain targeting II. Technologicalharacterization

aolo Blasi a,∗, Aurélie Schoubbena, Giovanna Valentina Romanoa, Stefano Giovagnoli a,lessandro Di Micheleb, Maurizio Riccia

Department of Chemistry and Technology of Drugs, via del Liceo 1, University of Perugia, 06123 Perugia, ItalyDepartment of Physics, via A. Pascoli, University of Perugia, 06123 Perugia, Italy

a r t i c l e i n f o

rticle history:eceived 30 October 2012eceived in revised form 18 April 2013ccepted 22 April 2013vailable online xxx

eywords:olid lipid nanoparticlesanoparticlesrain targeting

a b s t r a c t

The aim of this work was to characterize lipid nanoparticles from a rheological point of view, intendedfor drug delivery after parenteral administration. The conditions to obtain a re-dispersible powder usingfreeze-drying and spray-drying have also been investigated. Lipid nanoparticles (179.9 ± 6.2 nm) wereprepared with the high pressure homogenization technique, using previously established optimal con-ditions (lipid volume fraction of 0.121), though particle size increased (285.9 ± 4.3 nm) in suspensionsproduced with higher lipid volume fractions (0.255). Rheology evidenced an expected increase of viscos-ity with the volume fraction and Newtonian behaviour was observed for volume fractions up to 0.161,while higher volume fractions showed shear thinning and shear thickening. In the suspension with a vol-ume fraction of 0.255, a change of the complex modulus was observed at low shear stress. Freeze-drying

heologyreeze-dryingpray-drying

and nano spray-drying were suitable only when trehalose was employed as an additive. In the formercase, particle size was increased by 18% (198.7 ± 1.1 nm) using 20 fold water dilution. With spray-drying,the use of 20 fold dilution in water:ethanol (8:2) led to particle dimensions of 207.7 ±10.0 (�size 20%).In conclusion, cetylpalmitate nanoparticles seem to be suitable for parenteral application, up to volumefractions of 0.16, and pharmaceutical operations, which submit suspensions to shear stress, should not

be a critical issue.

. Introduction

Nanoscience and nanotechnology are continuously evolving,ith a huge impact on everyday human life. Among the numerous

reas of application, nanotechnologies are raising great hopes inharmaceutical and biomedical fields [1]. Since the 1970’s, severalubmicron carriers have been developed with the aim of control-ing drug release in time and space. Sadly, there is still a limitediffusion on the market of colloidal carriers and of nanotechnolo-

ies applied to the pharmaceutical field, nowadays, the challenges to translate these systems from the laboratory to the clinic. Thisartial success is perhaps also due to technological issues affecting

Abbreviations: CP, cetylpalmitate; GDW, Gaussian distribution width; MHD,ean hydrodynamic diameter; NP, nanoparticle; NPs, nanoparticles; NLCTM, nano-

tructured lipid carriers; PACA, polyalkylcyanoacrylate; PEG, polyethylene glycol;LG, polyglycolic acid; PLA, polylactic acid; PLGA, poly(lactic-co-glycolic) acid;80, polysorbate 80; SLNTM, solid lipid nanoparticles; TEM, transmission electronicroscopy; �, volume fraction.∗ Corresponding author at: Dipartimento di Chimica e Tecnologia del Farmaco,niversità degli Studi di Perugia, via del Liceo 1, 06123, Perugia, Italy.el.: +39 0755852057; fax: +39 0755855163.

E-mail address: [email protected] (P. Blasi).

927-7765/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.colsurfb.2013.04.021

© 2013 Elsevier B.V. All rights reserved.

nanoparticle (NP) industrial manufacturing and long-term stability[2].

Among the different nanometric carriers proposed in the scien-tific literature for drug delivery, lipid nanoparticles (NPs), knownas solid lipid nanoparticles (SLNTM) or nanostructured lipid carriers(NLCTM), combine the advantages of liposomes, parenteral emul-sions, and polymer NPs and have only some of their drawbacks.Liposomes, discovered in the late 1960’s and subsequently pro-posed as drug delivery carriers, have shown problems of physicalstability, difficulties in industrial production and high manufac-turing costs [3]. Parenteral emulsions, used in clinics since the1960’s, share with liposomes and SLNTM the relative safety of thematerials employed (i.e., lipids) but lack controlled release, a fun-damental feature for drug delivery and targeting. Unlike SLNTM, theliquid state of lipids does not permit controlled/prolonged release.Polymeric NPs, generally, do not present stability problems but pro-duction scale up and degradation by-product toxicity still representlimiting issues.

SLNTM and NLCTM, particles of nanometric sizes composed of

lipids/lipid mixtures solid at body temperature and stabilized byone or more surfactants, appear to be very versatile carries [4,5].In fact, they seem to solve three of the most serious disadvantageslinked to the use of liposomes, emulsions, and polymeric NPs, such

Page 2

s B: Biointerfaces 110 (2013) 130– 137 131

al

(autatdtwr

htplfst

tAaasfad

pewdcdicrw

oho

cpbefRdaipi

2

2

sd

Table 1Particle size of lipid NPs prepared with different lipid volumetric fractions.

Formulation Volume fraction Meanhydrodynamicdiameter (nm)

Gaussiandistributionwidth (nm)

A 0.121 179.9 ± 6.2 56.0 ± 0.1B 0.161 182.9 ± 0.4 51.3 ± 8.9C 0.194 218.5 ± 4.1 69.9 ± 3.1

P. Blasi et al. / Colloids and Surface

s material/by-product toxicity, uncontrolled release and difficultarge scale production [5].

Even though some polymer materials, such as polylactic acidPLA), polyglycolic acid (PLG), poly(lactic-co-glycolic) acid (PLGA),nd polyalkylcyanoacrylate (PACA), are approved for parenteralse and considered safe, cytotoxicity studies have evidenced loweroxicity for SLNTM when compared to PACA or PLA/PLGA [6]. Inddition, an unsolved problem in the case of intravenous injec-ion, especially when brain accumulation is expected, is wasteisposal [7]. Once accumulated in a tissue, lipids degrade fasterhan biodegradable or bioerodible polymeric materials, limiting theaste disposal problem. High lipid dose, injected intravenously to

odents, has shown low or no toxicity [8].Among the different methods proposed to prepare lipid NPs,

igh pressure homogenization has the highest potential for indus-rial manufacturing [9] since it has been used in the food industry torocess high fat products for longer than a century [10]. However,

ipid NPs and especially SLNTM are not without disadvantages. Inact, drug expulsion during crystallization, polymorphism, suspen-ion gelation and the production of supercooled melts have to beaken into account [4,5,7].

The features discussed here give an idea of the huge poten-ial of lipid carriers in different areas of drug delivery [5,11–13].mong these, one very interesting application is the delivery ofctive pharmaceutical ingredients to the central nervous systemnd in particular to the brain [7,14]. Lipid and polymer NPs, whentabilized with specific polyethylene glycol (PEG)-containing sur-actants (e.g., polysorbate 80, P80 and poloxamer 188), have shownn interesting tropism for the blood-brain barrier leading to brainrug targeting [15–18].

In a previous work, P80 stabilized lipid NPs was prepared by highressure homogenization with 3 different lipid materials and theffect of preparation parameters on particle size and polydispersityere studied with the aid of a computer generated experimentalesign [19]. Cetylpalmitate (CP) NPs showed the best dimensionalharacteristics for brain drug delivery and, by using a fluorescentye, preliminary results confirmed the capability of this carrier to

nteract with the brain capillary endothelium and also the translo-ation of the dye to the brain parenchyma [20]. By studying CP NPat brain acute toxicity, TEM images clearly showed NPs interactingith and endocytosed by the BBB [21].

In order to shed light, at very early stages, on the real potentialf the colloids mentioned and to foresee important issues that mayinder their industrial manufacture and clinical application, a seriesf in vitro and in vivo characterization is under way.

The present paper focuses on the production and technologicalharacterization of CP NPs intended for brain drug delivery afterarenteral administration. In particular, rheological studies haveeen performed to predict the suspension behaviour under differ-nt shear rate/shear stress conditions to which liquid-injectableormulations are subjected during production and administration.heology helps to understand parenteral formulation behavioururing pharmaceutical operations (product pumping or vial filling)nd administration (syringeability and injectability). The possibil-ty to obtain a dispersible dry powder, always convenient in theharmaceutical industry for its superior stability, has also been

nvestigated using both lyophilisation and spray-drying.

. Materials and methods

.1. Materials

CP (batch 120851, purity 93.1%) was a kind gift from Gattefossé.a.s. (Saint Priest, France). P80 (batch 038K00902, purity > 90%),ihydrated D(+)trehalose (batch 090M7357 V, purity ≥ 99%), and

D 0.224 266.6 ± 5.9 62.6 ± 6.3E 0.255 285.9 ± 4.3 82.9 ± 0.6

PEG (MW ≈ 10,000 Da) (batch 022K0039, purity > 90%) were pro-vided by Sigma–Aldrich (Milan, Italy), while absolute ethanolwas supplied by J.T. Baker (Deventer, Holland), ultra pure water(∼18 M� cm) was obtained by a New Human Power system(Human Corporation, Seoul, Korea), provided with ion-exchangeresin and active carbon filters.

2.2. Preparation of solid lipid nanoparticles

Lipid NPs, containing different amounts of CP, were preparedusing the hot high pressure homogenization technique. The waxwas melted at 65 ◦C and slowly added to 35 mL of external phase, aP80 aqueous solution (2% w/v), heated up to the same temperature.This was done using a high-speed stirring device set at 8000 rpmfor 1 min (Ultra Turrax T25 IKA® Werke GmbH & Co. KG, Staufen,Germany). This pre-emulsion was then processed with a high pres-sure homogenizer (Avestin EmulsiFlex C5, Ottawa, Canada) at thesame temperature, using seven homogenization cycles at 1500 bar.The emulsion obtained was immediately cooled in an ice bath,under mild stirring, for 20 min [19]. All the formulations reportedin Table 1 were prepared in triplicate.

2.3. Particle size and morphology determination

Particle size was determined with photon correlation spec-troscopy, using a Nicomp 380 autocorrelator (PSS Inc., SantaBarbara, CA, USA) equipped with a Coherent Innova 70-3 (LaserInnovation, Moorpark, CA, USA) argon ion laser. The scattered lightwas detected al 90◦ from the incident light [22,23]. Samples wereprepared adding 2 �L of lipid NPs suspension to 3 mL of ultrapurewater. Analyses were performed at 20 ◦C, for 15 min, in triplicate.

Lipid NPs were morphologically characterized using transmis-sion electron microscopy (TEM) (Philips EM 400 T microscope,Eindhoven, Nederland). Samples were prepared by allowing a dropof the NP suspension to dry overnight at 20 ◦C on the surface of a 200mesh Formvar® coated copper grid (TAAB Laboratories EquipmentLtd., Aldermaston, England).

2.4. True density determination

CP true density was determined using a helium pycnometerUltrapycnometer 1000 (Quanta-chrome Instruments, USA). Anal-yses were performed in a 10 cm3 cell and the true density (g/cm3)was expressed as the mean of ten successive measures.

2.5. Rheological characterization

A Stresstech HR rheometer (Rheologia Instruments AB, Milano,Italy), equipped with a cone-plate (cone angle 1◦, cone/plate diam-eter 40 mm) and plate-plate (diameter 40 mm) geometry and aPeltier device for temperature control, was used for the rheological

characterization.

In order to determine the lipid NP suspension viscosity, continu-ous flow measurements were performed using a fixed gap betweencone and plate. Oscillation stress sweep tests, useful to determine

Page 3

1 s B: Biointerfaces 110 (2013) 130– 137

csrvttpgmd

2

IT1wgwp�

�

w(M

2

dU

2

(1ra

slbaw

3

3

ootsaatpspdamd

32 P. Blasi et al. / Colloids and Surface

omplex modulus G*, were performed with the above-mentionedetting (cone-plate geometry), at a frequency of 1 Hz and a stressanging from 1 to 100 Pa. Stress sweep test was performed to indi-iduate the linear viscoelasticity region and gain information onhe system structure. All the measurements were carried out inriplicate at a temperature of 20 ± 0.1 ◦C. Measurements were alsoerformed in triplicate using the plate-plate geometry at differentaps at temperatures of 20 ± 0.1 and 37 ± 0.1 ◦C. Continuous floweasurement duration was 140 s while oscillation stress sweep test

uration was 5 min.

.6. Lipid nanoparticle drying

Lipid NPs were dried using both freeze-drying and spray-drying.n the latter, two different spray-dryer instruments were compared.he suspension with a � of 0.121 was appropriately diluted (1:5,:10, and 1:20) and cryoprotectants (i.e., ethanol, trehalose, PEG)ere evaluated for their usefulness in avoiding irreversible aggre-

ation during water removal. After drying, the obtained powderas dispersed in ultrapure water under stirring and subjected toarticle size analysis. Size increase after drying was expressed assize, calculated using the following equation.

size = mhd after drying − Initial mhdInitial mhd

100 (1)

here mhd after drying is the mean hydrodynamic diameterMHD) after freeze-drying or spray-drying and Initial mhd is the

HD just after preparation.

.6.1. Freeze-dryingSamples were rapidly frozen at −80 ◦C and successively freeze-

ried for 24 h using a Benchtop 2K freeze-dryer (VirTis, New York,.S.A.).

.6.2. Spray-dryingSuspensions were processed with a mini spray-dryer B290

Büchi, Italy). The inlet temperature was varied between 100 and20 ◦C according to the employed solvent mixture, the pump flowate was set at 5 mL/min, the aspirator rate at 27 m3/h and the sprayir flow at 536 L/h.

Lipid NP suspensions were alternatively processed using a nanopray-dryer B90 (Büchi, Italy). In this work, 4.0 �m pore size stain-ess steel membrane was used to nebulise the suspension and theest working parameters were an inlet temperature of 110 ◦C, anir flow of 100 L/min, a pressure of 50 mbar in the drying chamber,hile the pump was set at 1 and the spray at 100%.

. Results and discussion

.1. Lipid nanoparticle preparation and characterization

Using a computer generated experimental design, formulationsf P80 stabilized lipid NPs, intended for brain drug delivery, wereptimized in terms of particle size [19]. Even though P80 concen-rations higher than 2% (w/v), i.e., 3 and 4%, led to smaller particleize and distribution, 2% (w/v) P80, that gave particles with MHDround 180 nm, was considered more suitable due to the potentialnd still controversial in vivo toxic effects of P80 at high concen-rations [19]. A TEM photograph of lipid NPs prepared using thereparation parameters mentioned is shown in Fig. 1. Lipid NPseem spherical or slightly oblate with a smooth surface. The TEMhotograph shows particle mean diameters compatible with those

etermined by photon correlation spectroscopy. Small differencesre obviously due to the fact that photon correlation spectroscopyeasures MHD, while from the TEM image only the diameter of

ry particles can be determined.

Fig. 1. Representative transmission electron microscopy photograph of lipid NPsfrom formulation A.

Additional formulations were produced using the same param-eters (i.e., 2% P80, 7 homogenization cycles at 1500 bar) butdifferent lipid volume fractions (�). To characterize the rheologicalbehaviour of this colloidal system as a function of �, the amountof the internal phase was increased to a � of 0.255 (Table 1). Asexpected and previously reported [24], an increase of the lipidamount (internal phase) at fixed external phase volume (35 mL)corresponded to an increase of both MHD and Gaussian distribu-tion width (GDW) (Table 1). This phenomenon can be ascribedto the diminished quantity of surfactant (2% in all the formula-tions). In fact, while the P80 concentration was unmodified, thelipid/water ratio was increased, thereby decreasing the amountof surfactant per unit of lipid volume. The larger the surface areaand the surface free energy generated, the higher the particle size.However, other explanations can not be completely ruled out. Anincrease in the internal phase leads to higher emulsion viscos-ity, the latter reducing the homogenization performance [25]. Inaddition, emulsion growth mechanisms (note that lipid NPs areprepared from an emulsion), such as flocculation (also referred toas agglomeration) and coalescence, may have played a role [26].The higher the lipid �, higher the probability of agglomeration andcoalescence during preparation. On the contrary, the contributionof Ostwald ripening can be ruled out due to the very short timeof the emulsification process and the low water solubility of CP[27].

3.2. Rheological characterization

Suspension characterization was performed with the aim ofunderstanding the behaviour of this colloidal suspension dur-ing manufacturing operations, syringe filling (i.e., syringeability),and injection (i.e., injectability) [28]. During industrial produc-tion, different steps, such as mixing, stirring, product pumping andhigh-speed filling, involve the submission of liquid products (e.g.,solutions, suspensions) to high shear rates [29,30].

Rheological characterization was done in a � range 0.009–0.255but, since the effect of particle size on suspension viscosity iswell known, measurements and theoretical calculations weredone on the formulation with � = 0.255 (MHD, 285.9 ± 4.3; GDW,82.9 ± 0.6) and on its dilution with the external phase, obtaining

diluted (� = 0.009), semi-diluted (0.044 > � > 0.224), and concen-trated (� = 0.255) suspensions [31,32]. This procedure avoided theeffect of both � and particle size on suspension viscosity and vis-coelasticity. Lipid � higher than 0.255 were excluded from the

Page 4

s B: Biointerfaces 110 (2013) 130– 137 133

cn

Narp(itbp

rmttor[iola[

osslisms0attimtpnsavbtt

latatd(

�e

�

wi

essary to add a parameter, representing double hydrodynamic

P. Blasi et al. / Colloids and Surface

haracterization since semisolid systems, applicable to the skin butot parenteral, were produced [33,34].

Viscosity increased with � but suspension behaviour remainedewtonian up to a � of 0.161 for the shear rates investigated while,t higher �, fluids were Newtonian only in the first part of theheogram (data not shown). The Newtonian behaviour of the sus-ensions with lower � was expected since the suspending fluidwater with 2% w/v of P80), � = 0, showed no changes in viscosityn this range. The interpretation of the behaviour of high � par-icle suspensions can be more complex and tedious. While similarehaviours are observed for � 0.194 and 0.255, � 0.224 has its ownarticular rheogram (data not shown).

The behaviour of suspension viscosities as a function of shearate or shear stress is described below. In equilibrium, Brownianotions produce random collisions between particles and make

he latter naturally resistant to flow so, at low shear rates, New-onian behaviour is generally observed [35,36]. With the increasef the shear rate, particles tend to align to the flow, reducing theesistance of the fluid to motion, and a lower viscosity is recorded35,36]. Finally, when shear rate becomes very high, hydrodynamicnteractions predominate over stochastic ones and the formationf clusters makes it harder for single particles to flow [35,36]. Thiseads to a rapid increase of suspension viscosity. These 2 transitionsre referred as shear-thinning and shear-thickening transitions35,36].

Fig. 2 displays the change of viscosity plotted as a functionf shear stress for the four highest �. Since lower � presentedubstantially the same behaviour, but different viscosity, as theuspension with � of 0.161, the latter may be used to describeower � dispersions. While Newtonian behaviour was observedn the range of � between 0.009 and 0.161, shear-thinning andhear-thickening transitions were observed in the remaining for-ulations (Fig. 2). In particular, � 0.194 and 0.255 showed

hear-thickening between 2–3 Pa and 3–4 Pa, respectively, while.224 � showed a moderate shear-thinning and, around 5 Pa, anbrupt but small increase of viscosity and again a very slight shear-hinning [36,37]. The explanation of rheological phenomena athe molecular and macromolecular levels is complex in and oftself. However, shear-thinning and shear-thickening are very com-

only observed in particle dispersions and some explanation ofhis phenomenon has been provided [36,37]. Thermal agitationrovokes random collisions between particles, making the fluidaturally resistant to flow (Newtonian behaviour). By increasinghear stress/shear rate, viscosity starts to decrease due to particlelignment to flow. At very high shear stress/shear rates, viscosityalues increased. At 2–3 Pa, hydrodynamic interactions are proba-ly stronger than stochastic interactions and particle aggregationakes place, making the fluid more resistant to flow (i.e., shearhickening) [36,37].

Measured suspension viscosities were compared to those calcu-ated using previously reported equations that correlate viscositynd � [38–43]. Particles in colloidal suspensions are referredo as Brownian particles while larger particles are referred tos non-Brownian particles [44]. Suspensions are also differen-iated (for theoretical treatment) as a function of the �, iniluted (� < 0.01–0.02), semi-diluted (� < 0.25) and concentrated� > 0.25) suspensions [31,32].

Einstein first established the relationship between viscosity and for diluted suspensions in Newtonian fluids with the following

quation [38]:

(5

)

= �0 1 +

2� (2)

here � is suspension viscosity, �0 is pure suspending fluid viscos-ty and � is the volumetric fraction of rigid monodisperse spheres.

Fig. 2. Viscosity versus shear stress for the four highest volume fraction suspensions.

For semi-diluted suspensions in a Newtonian fluid, it is nec-

interactions, to the Einstein equation [39]:

� = �0

(1 + 5

2� + b �2

)(3)

Page 5

134 P. Blasi et al. / Colloids and Surfaces B: Biointerfaces 110 (2013) 130– 137

Fig. 3. Normalized experimental and theoretical viscosity as function of volumeffo

7

se

�

wcvsa

opaiswdfths[oncdi

Fig. 4. Complex modulus versus shear stress for 0.224 and 0.255 (mean hydro-

raction. Empty rhombus, experimental data (±S.D.); black solid line, data obtainedrom Eq. (2); grey solid line, data obtained from Eq. (3); black dotted line, databtained from Eq. (4).

The exact value of b was calculated by Batchelor and Green as.6 for non-Brownian spheres and 6.2 for Brownian spheres [40].

Another equation employed to correlate viscosity and � inemi-diluted or concentrated suspensions is the Krieger-Doughertyquation [41], subsequently modified by Kitano et al. [42]

r = �

�0=

(1 − �

�m

)−ε

(4)

here �r is the suspension relative viscosity, � is the suspension vis-osity, �0 is the pure suspending fluid viscosity, �m is the maximumolumetric fraction for solids and ε is equal to 2. In an experimentalituation, it has been observed that in a suspension where particlesre randomly arranged �m is equal to 0.63 [43].

Fig. 3 shows the comparison of experimental viscosity with the-retical calculations. All the equations investigated have a goodredictability only in the case of the diluted suspension (� 0.009)nd deviation to higher viscosities is observed for higher �. Whilen the case of the Einstein equation this behaviour was expectedince it was conceived for diluted suspensions, a higher predictivityas expected from Eqs. (3) and (4). The deviation observed could beue to the different assumptions of the mathematical treatment. Inact, lipid NPs should be considered soft spheres and applied equa-ions have been developed and confirmed experimentally only withard spheres [45]. However, it is reported that suspensions of softpheres at � up to ≈0.4 behave like suspensions of hard spheres45]. The equations mentioned have been conceived for dispersionsf monodisperse spheres so, even though the GDW (i.e., 83 nm) is

ot very large, lipid NPs used in the experimental measurementsan not be considered monodispersed. Finally, some deviation mayerive from the slightly negative � potential recorded for the NPs

nvestigated. In fact, particle surface charge has been recognized as

dynamic diameter, 285.9 ± 4.3 nm; Gaussian distribution width, 82.9 ± 0.6 nm)volume fraction suspensions.

a factor increasing suspension viscosity, a phenomenon known aselectroviscous effect [46].

During industrial manufacturing of a liquid formulation, productpumping (>1000 s−1) and high-speed filling (>5000 s−1) [30] maybe critical if the suspension does not behave as a Newtonian fluid.In this specific case, even though extremely high shear rates werenot evaluated, CP suspensions with � up to ≈0.16 should not beproblematic during pharmaceutical operations [5].

In order to gain information on suspension behaviour(0.009 < � < 0.161) during withdrawing from vials using a syringewith a needle, viscosity measurements were repeated usingplate-plate geometry with variable gaps. Since the problem of with-drawing and injecting is linked to needle diameter and length [28],gaps were chosen to simulate the internal radius of four differentneedle sizes (i.e., 31, 27, 22, and 18 Gauge). In particular, consider-ing the internal diameters of the needles mentioned (i.e., 31 Gauge,0.114 mm; 27 Gauge, 0.191 mm; 22 Gauge, 0.394 mm; 18 Gauge,0.838 mm), gaps of 50, 100, 200, and 400 �m were considered goodapproximations of the internal radius and used in this study.

As reported above, measured viscosity was a function of � butan increase in viscosity was also recorded with the rise of the plate-plate gaps for the same �. For instance, for � of 0.194, viscosityincreased from ∼1 to ∼3 cP and, for � of 0.255, it increased from ∼5to ∼10 cP when raising the plate-plate gap from 50 to 400 �m. Forall formulations, the maximum viscosity observed was lower than10 and 5 cP for 20 and 37 ◦C, respectively. Lipid NP suspensions with� between 0.009 and 0.161 did not exceed 50 cP of viscosity, con-firming their suitability for parenteral use and a potentially goodsyringeability and injectability with a wide range of syringe needles[47].

Oscillation stress sweep tests were used to determine the region

of linear viscoelasticity. In Fig. 4, the complex modulus plottedversus the shear stress for the 2 most concentrated suspensions isreported. Complex modulus is useful to evidence structural changes

Page 6

P. Blasi et al. / Colloids and Surfaces B: Biointerfaces 110 (2013) 130– 137 135

Table 2Particle size of lipid NPs after freeze-drying in different conditions of lipid NPs prepared with the same parameters of Formulation A (Table 1).

Additive External phase Dilution Mean hydrodynamic diameterafter treatment ± s.d. (nm)

Gaussian distributionwidth ± s.d. (nm)

�sizea (%)

– Water 1:5 >2000 – –1:10 1740.2 ± 1653.2 1761.5 ± 1985.5 9591:20 878.4 ± 117.2 859.9 ± 141.7 635

Water/Ethanol (8:2) 1:5 >2000 – –1:10 >2000 – –1:20 >2000 – –

Trehalose (10%) Water 1:5 234.8 ± 2.7 81.5 ± 7.8 391:10 215.6 ± 8.2 75.2 ± 3.6 281:20 198.7 ± 1.1 61.9 ± 8.6 18

Water/Ethanol (8:2) 1:5 701.4 ± 611.0 533.4 ± 605.4 3191:10 241.5 ± 1.5 93.0 ± 6.4 441:20 240.6 ± 5.6 88.9 ± 4.1 44

Polyethylene glycol(10%)

Water 1:5 316.8 ± 5.9 138.5 ± 5.2 901:10 621.2 ± 680.5 533.1 ± 663.1 2721:20 336.6 ± 17.3 182.1 ± 33.9 102

Water/Ethanol (8:2) 1:5 274.1 ± 2.5 102.6 ± 1.5 641:10 289.8 ± 10.1 99.7 ± 8.8 731:20 269.3 ± 3.8 80.0 ± 8.5 61

orted

iioclb

fioco

3

ioaa(piprtGPwmh

istvtat[

ast

recovered were much larger (�size of + 242%). However, the morediluted suspensions (i.e., 1:10 and 1:20) led to better results interms of size recovery (�size of + 29% in both cases) and yield (i.e.,

a Please note that particle size before drying may slightly vary from the value rep

n the disperse system [25]. From the results of the oscillatory tests,t seems that only the suspension with the highest � had a changef the G* in the stress range investigated, indicating an irreversiblehange in the fluid structure (Fig. 4) which may be compatible withipid NP aggregation and sintering. Macroscopic aggregates coulde observed after the measurements (data not shown) [5].

All the rheological experiments (i.e., viscosity determination atxed and variable gaps and viscoelasticity) were repeated afterne month (formulations were stored at 4 ◦C) and no signifi-ant changes in the viscosity values and rheology behaviour werebserved.

.3. Suspension drying

The data obtained after freeze-drying lipid NP suspensions,.e., Formulation A (Table 1), are reported in Table 2. As can bebserved, a size increase was recorded in all experiments and, in thebsence of additives, freeze-drying led to big lipid NP aggregatesnd even dilution up to 20 times did not improve the procedureTable 2). Aggregation was strongly reduced using trehalose thaterformed better compared to PEG. As evidenced in Table 2, a size

ncrease as low as 18% was obtained working with a diluted sus-ension (1:20). The other dilutions with water also gave acceptableesults, while the addition of ethanol led to good results for dilu-ions 1:10 and 1:20 (Table 2). In all these cases, the MHD and theDW were low and the process was considered successful. UsingEG as additive to avoid particle size increase, the best resultsere obtained by diluting the suspension with a water/ethanolixture. However, results were less satisfactory than with tre-

alose.Spray-drying was the other technique taken in consideration

n this work to obtain a dispersible powder. A limitation of usingpray-drying instead of freeze-drying is the use of temperatureo remove water, even though the exposition to temperature isery short and the latent heat of vaporization drastically reduceshe thermal stress of the material. Working with lipid materi-ls, the risk is to obtain lipid melting during drying that leadso particle sintering with the loss of the initial particle size48].

With a conventional spray-dryer, only a few assays, using water/ethanol mixture to lower the inlet temperature, wereuccessfully performed. As suspected, after drying, particles stucko the glassware allowing no or extremely low material recovery.

in Table 1 for Formulation A.

In the case of 1:5 dilution, the yield was low (14%) and the particles

Fig. 5. Transmission electron microscopy pictures of NPs resuspended after spray-drying (nano spray-dryer B90). (A) 1:20 water dilution without additive; (B) 1:20dilution with ethanol and trehalose.

Page 7

136 P. Blasi et al. / Colloids and Surfaces B: Biointerfaces 110 (2013) 130– 137

Table 3Particle size of lipid NPs after spray-drying (nano spray-dryer B90) in different conditions of lipid NPs prepared with the same parameters of Formulation A (Table 1).

Additive External phase Dilution Mean hydrodynamic diameterafter treatment ± s.d. (nm)

Gaussian distributionwidth ± s.d. (nm)

�sizea (%) Yield (%)

– Water 1:5 >2000 – – 141:10 >2000 – – 121:20 >2000 – – 17

Water/Ethanol (8:2) 1:5 >2000 – – 321:10 >2000 – – 321:20 677.1 ± 170.0 529.1 ± 207.2 312 33

Trehalose (10%) Water 1:5 >2000 – – 501:10 275.1 ± 6.1 118.9 ± 4.3 59 671:20 242.1 ± 8.6 88.7 ± 4.2 40 55

Water/Ethanol (8:2) 1:5 356.1 ± 1.9 105.7 ± 36.4 106 481:10 232.4 ± 3.8 107.2 ± 4.9 34 521:20 207.7 ± 10.0 81.6 ± 1.1 20 63

Polyethylene glycol(10%)

Water 1:5 811.1 ± 38.3 221.4 ± 14.4 389 591:10 510.6 ± 50.9 243.4 ± 52.0 208 611:20 329.6 ± 15.7 167.7 ± 21.5 98 62

Water/Ethanol (8:2) 1:5 875.5 ± 36.2 425.3 ± 17.8 428 571:10 583.0 ± 16.6 356.2 ± 22.9 247 581:20 421.9 ± 47.6 273.2 ± 15.0 154 60

a orted

3t

(rdeftjdah

dwwtatsaSal[

4

aardpwsabrn

[[[

[[[

[

[[

[

[

[

[

[[

Please note that particle size before drying may slightly vary from the value rep

4 and 39%). As previously reported, suspension dilution reduceshe probability of aggregation [49].

In Table 3, data obtained using a new generation spray-dryernano spray-dryer) are reported. In this instrument, the powderecovery mechanism is completely different and one could expectifferent results compared to conventional spray-dryers. How-ver, as in the case of lyophilisation, the use of an additive wasundamental to limit aggregation as much as possible. Large par-icle aggregates and low yields were obtained after drying withust water or a water/ethanol mixture (Fig. 5A). Trehalose pro-uced better results than PEG and both the dilution and the use

water/ethanol mixture led to a lower size increase (Fig. 5B) and aigher yield (Table 3).

Overall, the best results were therefore obtained with 20 timesiluted suspensions and trehalose as protecting agent, eitherith lyophilisation in water (�size, 18%) or by spray-drying aater/ethanol mixture (�size, 20%). This success is probably related

o the unique feature of trehalose, such as low hygroscopicity, thebsence of internal hydrogen bonds, and a high glass transitionemperature [50–52]. In fact, this sugar has been extensively anduccessfully used to protect biomolecules during freezing, dryingnd contact with organic solvents or hydrophobic surfaces [50,52].o lyophilisation and spray-drying, the latter using a new gener-tion instrument, were the most promising techniques to obtainipid NP dry powders of a suitable size for parenteral administration53–55].

. Conclusions

The results presented here have indirectly confirmed the suit-bility of P80 stabilized CP NPs for industrial manufacturingnd parenteral application for suspensions with a lipid volumet-ic fraction up to 0.16. In addition, the possibility of producingry powders, useful when the suspension shows instabilityhenomena, has been demonstrated using freeze- or spray-drying

hen suitable excipients are employed. Lipid colloidal suspen-

ions, such as SLNTM and NLCTM, have a great potential in manyreas of drug delivery but this potential has to be demonstratedeyond all reasonable doubt and, though more data is needed, theseesults are a valuable contribution to the field of pharmaceuticalanotechnology.

[[[[

[

in Table 1 for Formulation A.

Acknowledgements

The Authors would like to acknowledge Mr. Lanfranco Barberinifor his help in image elaboration and Mrs. Mary Kerrigan for editingthe manuscript English form.

References

[1] S.M. Moghimi, A.C. Hunter, J.C. Murray, FASEB J. 19 (2005) 311.[2] P. Blasi, P. Caliceti, G. De Rosa, H. Haas, M. Maio, P. Matricardi, S. Salmaso, CRS

Newsletter 29 (2) (2012) 20.[3] R.H. Müller, A. Dingler, Euro Cosmetics 7–8 (1998) 19.[4] W. Mehnert, K. Mäder, Adv. Drug Deliv. Rev. 47 (2001) 165.[5] S.A. Wissing, O. Kayser, R.H. Müller, Adv. Drug Deliv. Rev. 56 (2004) 1257.[6] R.H. Müller, D. Rühl, S. Runge, K. Schulze-Forster, W. Mehnert, Pharm. Res. 14

(1997) 458.[7] P. Blasi, S. Giovagnoli, A. Schoubben, M. Ricci, C. Rossi, Adv. Drug Deliv. Rev. 59

(2007) 454.[8] H. Weyers, S. Ehlers, H. Hahn, Pharmazie 61 (2006) 539.[9] V. Jenning, A. Lippacher, S.H. Gohla, J. Microencapsul. 19 (2002) 1.10] W.D. Pandolfe, J. Dairy Sci. 65 (1982) 2035.11] S. Das, A. Chaudhury, AAPS PharmSciTech 12 (2011) 62.12] C. Puglia, R. Filosa, A. Peduto, P. de Caprariis, L. Rizza, F. Bonina, P. Blasi, AAPS

PharmSciTech 7 (2006) E61.13] A. Seyfoddin, J. Shaw, R. Al-Kassas, Drug Deliv. 17 (2010) 467.14] I.P. Kaur, R. Bhandari, S. Bhandari, V. Kakkar, J. Control. Release 127 (2008) 97.15] P. Blasi, A. Schoubben, S. Giovagnoli, C. Rossi, M. Ricci, in: C. Kumar (Ed.), Nano-

materials for the Life Sciences, vol. 10, Wiley-VCH, Weinheim, 2011, Chapter13.

16] J.M. Koziara, P.R. Lockman, D.D. Allen, R.J. Mumper, J. Control. Release 99 (2004)259.

17] J. Kreuter, Int. Congr. Ser. 1277 (2005) 85.18] V. Podio, G.P. Zara, M. Carazzone, R. Cavalli, M.R. Gasco, J. Pharm. Pharmacol.

52 (2000) 1057.19] P. Blasi, S. Giovagnoli, A. Schoubben, C. Puglia, F. Bonina, C. Rossi, M. Ricci, Int.

J. Pharm. 419 (2011) 287.20] P. Blasi, L. Mouldon, L. Barberini, C. Cirotto, M. Ricci, E.A. Neuwelt, CRS Italy

Chapter Workshop, 2011.21] P. Blasi, A. Schoubben, G. Traina, G. Manfroni, L. Barberini, P.F. Alberti, C. Cirotto,

M. Ricci, Int. J. Pharm. Under evaluation.22] D.F. Nicoli, J.S. Wu, Y.J. Chang, D.C. McKenzie, K. Hasapidis, Am. Lab. 27 (1995)

41.23] D.F. Nicoli, D.C. McKenzie, J.S. Wu, Am. Lab. 23 (1991) 32.24] S. Martins, I. Tho, D.C. Ferreira, E.B. Souto, M. Brandl, Drug Dev. Ind. Pharm. 37

(2011) 815.25] A. Lippacher, R.H. Müller, K. Mäder, Eur. J. Pharm. Biopharm. 58 (2004) 561.26] E. Fredrick, Adv. Colloid Interface Sci. 2153 (2010) 30.

27] P. Taylor, Adv. Colloid Interface Sci. 75 (1998) 107.28] F. Cilurzo, F. Selmin, P. Minghetti, M. Adami, E. Bertoni, S. Lauria, L. Montanari,

AAPS PharmSciTech 12 (2011) 604.29] H. Schott, in: A. Gennaro (Ed.), Remington: The Science and Practice of Phar-

macy, Lippincott, Williams & Wilkins, Philadelphia, 2000, Chapter 23.

Page 8

s B: Bi

[

[[[[[

[[[[[[[[

[

[

[[[

[[[[

P. Blasi et al. / Colloids and Surface

30] J.H. Wood, in: L. Lachman, H.A. Lieberman, J.L. Kanig (Eds.), The Theoryand Practice of Industrial Pharmacy, Lea and Febiger, Philadelphia, 1986,Chapter 6.

31] I.R.R. Rutgers, Rheol. Acta 2 (1962) 202.32] D.G. Thomas, J. Colloid Sci. 20 (1965) 267.33] E.B. Souto, R.H. Müller, Pharmazie 62 (2007) 505.34] R.H. Müller, M. Radtke, S.A. Wissing, Adv. Drug Deliv. Rev. 54 (2002) S131.35] D.T.N. Chen, Q. Wen, P.A. Janmey, J.C. Crocker, A.G. Yodh, Annu. Rev. Condens.

Matter Phys. 1 (2010) 301.36] N.J. Wagner, J.F. Brady, Phys. Today 62 (2009) 27.37] J. Mellema, Curr. Opin. Colloid Interface Sci. 2 (1997) 411.38] A. Einstein, Ann. Phys. 34 (1911) 591.

39] R. Pasquino, Panta Rei 12 (2009) 7–19.40] G.K. Batchelor, J.T. Green, J. Fluid Mech. 56 (1972) 401.41] I.M. Krieger, T.J. Dougherty, Trans. Soc. Rheol. 3 (1959) 137–152.42] T. Kitano, T. Kataoka, T. Shirota, Rheol. Acta 20 (1981) 207.43] R.K. Mc Geary, J. Am. Ceram. Soc. 44 (1961) 513.

[

[[

ointerfaces 110 (2013) 130– 137 137

44] P. Romanczuk, M. Bär, W. Ebeling, B. Lindner, L. Schimansky-Geier, Eur. Phys. J.Spec. Top. 202 (2012) 1.

45] R.G. Larson (Ed.), The Structure and Rheology of Complex Fluids, Oxford Uni-versity Press Inc., New York, 1999.

46] P. Kumar, P.C. Kapur, D.N. Saraf, Colloid Polym. Sci. 253 (1975) 738.47] M.A. Miller, J.D. Engstrom, B.S. Ludher, K.P. Johnston, Langmuir 26 (2010) 1067.48] X. Zhang, W. Pan, L. Gan, C. Zhu, Y. Gan, S. Nie, Chem. Pharm. Bull. 56 (2008)

1645.49] C. Freitas, J.S. Lucks, R.H. Müller, Eur. J. Pharm. Sci. 2 (1994) 178.50] L.N. Crowe, D.S. Reid, J.H. Crowe, Biophys. J. 71 (1996) 2087.51] A. Cesarò, O. De Giacomo, F. Sussich, Food Chem. 106 (2008) 1318.52] S. Giovagnoli, P. Blasi, M. Ricci, C. Rossi, AAPS PharmSciTech 5 (4) (2004), Article

51.53] R. Cavalli, O. Caputo, M.E. Carlotti, M. Trotta, C. Scarnecchi, M.R. Gasco, Int. J.

Pharm. 148 (1997) 47.54] C. Freitas, R.H. Müller, Eur. J. Pharm. Biopharm. 46 (1998) 145.55] C. Schwarz, W. Mehnert, Int. J. Pharm. 157 (1997) 171.