European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

Contents lists available at SciVerse ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics

journal homepage: www.elsevier.com/locate/ejpb

Research paper

Crosslinked chitosan nanoparticle formulations for delivery from pressurized

metered dose inhalers
Ketan Sharma a, Satyanarayana Somavarapu a, Agnes Colombani b, Nayna Govind b, Kevin M.G. Taylor a,
a
b

UCL School of Pharmacy, London, UK

AstraZeneca R&D Charnwood, Loughborough, UK

a r t i c l e

i n f o

Article history:
Received 5 August 2011
Accepted in revised form 22 December 2011
Available online 9 January 2012
Keywords:
Chitosan
Inhalation
Nanocarrier
Nanoparticle
Polyethylene glycol
Pressurized metered dose inhaler

a b s t r a c t
Crosslinked chitosan nanoparticles, prepared using ionic gelation, have been successfully formulated into
pressurized metered dose inhalers (pMDIs) with potential for deep lung delivery of therapeutic agents.
Nanoparticles were prepared from crosslinked chitosan alone and incorporating PEG 600, PEG 1000 and
PEG 5000 for dispersion in aerosol propellant, hydrofuoroalkane (HFA) 227. Spherical, smooth-surfaced,
cationic particles of mean size less than 230 nm were produced. Nanoparticles were positively charged
and non-aggregated at the pH of the airways. Crosslinked chitosanPEG 1000 nanoparticles demonstrated
greatest dispersibility and physical stability in HFA-227, whereas other formulations readily either
creamed or sedimented. Following actuation from pMDIs, the ne particle fraction (FPF) for crosslinked
chitosanPEG 1000 nanoparticles, determined using a next generation impactor, was 34.0 1.4% with a
mass median aerodynamic diameter of 4.92 0.3 lm. The FPFs of crosslinked chitosan, crosslinked chitosanPEG 600 and crosslinked chitosanPEG 5000 nanoparticles were 5.7 0.9%, 11.8 2.7% and
17.0 2.1%, respectively. These results indicate that crosslinked chitosanPEG 1000-based nanoparticles
are promising candidates for delivering therapeutic agents, particularly biopharmaceuticals, using pMDIs.
2011 Elsevier B.V. All rights reserved.

1. Introduction
Pulmonary drug delivery may be employed for therapeutic
agents having local or systemic activity. It provides advantages
over other delivery routes as it is non-invasive, avoid rst-pass
metabolism and the lung offers a highly vascularized, large surface
area for drug absorption [1]. Pressurized metered dose inhalers
(pMDIs) are widely used inhalation devices, being convenient to
use and offering a sealed environment, providing protection from
air, light, moisture and microbial degradation. These medical devices comprise a therapeutic agent either suspended or dissolved
in a hydrouoroalkane (HFA) propellant. To achieve deep lung
deposition of particles, one successful approach has been to formulate low density hollow particles, which have relatively large physical diameters, corresponding to a much smaller aerodynamic
diameter [2]. An alternative approach is the use of nanoparticles
that are particularly attractive for pulmonary delivery, as their size
not only permits access to the peripheral airways [3] but also ensures that they escape both phagocytic and mucociliary clearance
mechanisms [4]. Incorporating drugs into, or onto, nanoparticles
potentially provides protection against intracellular and extracellular barriers, degradation and may overcome formulation chal Corresponding author. UCL School of Pharmacy, 29-39 Brunswick Square,
London WC1N 1AX, UK. Tel.: +44 207753 5853; fax: +44 207753 5942.
E-mail address: kevin.taylor@pharmacy.ac.uk (K.M.G. Taylor).
0939-6411/$ - see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejpb.2011.12.014

lenges, such as delivery of poorly aqueous soluble and unstable

drugs without compromising the native conformation of these
molecules [5,6]. The small size of dry nanoparticles leads to high
inter-particle cohesive forces that negatively impact on their
aggregation behaviour, which is particularly problematic for dry
powder inhaler (DPI) formulation. However, in a pMDI formulation, the presence of propellants such as HFAs offers potential for
deaggregating the nanoparticles, though other excipients may be
necessary [7]. A number of groups have studied pMDI nanoparticle
delivery, though adequate dispersion of such small particles in liqueed aerosol propellants is a major formulation challenge [3,8,9].
Oleic acid, sorbitan trioleate, dipalmitoylphosphatidylcholine and
volatile oils have been used within HFA propellants, permitting
the successful dispersion of protein nanoparticles [10] which
maintained protein integrity, and were successfully delivered in
an aerosol having appropriate aerodynamic characteristics for
therapeutic activity.
Chitosan has gained considerable interest as a polymer for preparing nanoparticles because of its biodegradable, biocompatible,
non-toxic and mucoadhesive properties [11,12]. Chitosan has been
reported to increase the uptake of macromolecules through opening of tight junctions of epithelial cells [13] and has also been formulated as nanoparticles designed to improve the delivery of
therapeutically active molecules across mucosal surfaces [14].
The successful application of chitosan for in vitro and in vivo gene
delivery has demonstrated its potential for pharmaceutical and

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

biomedical applications [15]. The potential for pulmonary delivery

has been recognized, and chitosan nanoparticles encapsulated in
mannitol microspheres have been demonstrated to be biocompatible with Calu-3 and A549 human respiratory epithelial cell lines
for up to 48 h [16]. Crosslinked chitosan microparticulates containing the bronchodilator, salbutamol sulphate prepared using spray
drying, were able to achieve controlled release of the drug [17]
and could be formulated into DPIs, with good aerosolization properties. Chitosan has not been extensively studied as a carrier for
delivery from pMDI systems; hence, in this study, the formulation
and characterization of crosslinked chitosan nanoparticles for
delivery using pMDIs are explored. Previous studies [9] suggested
that tripolyphosphate (TPP)-crosslinked chitosan microspheres
were not suitable for HFA-134a-based pMDI systems, because of
the density difference between the particles and propellant. The
dispersibility of such particles within a non-polar medium, such
as HFA-227 (1,1,1,2,3,3,3-heptauoropropane, which has higher
density than HFA-134a), can be improved by controlling the steric
repulsive forces between particles [18], for instance by inclusion of
hydrophilic polymers, such as polyethylene glycol (PEG) and polyvinylpyrrolidone (PVP) [19,20]. A recent study [19] has demonstrated that PEG acts as a polymeric surfactant and helps to
reduce the cohesive interactive forces between drug particles that
are suspended in a uorinated solvent (2H,3H-peruoropentane)
used as a non-volatile model for HFA propellants. PEG was chosen
in the formulation of crosslinked chitosan nanoparticles investigated in this study, since it has appreciable solubility in HFAs
[21] and is approved by the FDA as an excipient for use in pMDIs
[22]. In a previous study [23], inclusion of PEG300 within particles
of salbutamol sulphate prevented cohesive interactions when particles were dispersed in propellant HFAs (HFA-227 and HFA-134a).
Together, these studies indicate that polymers such as PEGs can
adsorb at an interface in uorinated liquid solvents, including HFA
propellants, and that crosslinked chitosanbased microparticles
can be formulated into pMDIs. The objective of the current study
was to prepare and characterize crosslinked chitosanbased nanoparticles, which might have future application in the delivery of
small molecules and biopharmaceuticals to the peripheral airways.
Dispersion of nano-sized particles is a considerable challenge, and
the study is the rst to explore the applicability of PEG of different
molecular weights as stabilizers for such particles dispersed in
HFA-227. The study describes the use of combinations of FDA-approved propellant and PEG to produce a functional and potentially
therapeutically useful nanoparticle formulation for pulmonary
delivery, via a pMDI.
2. Materials and methods
2.1. Materials
Chitosan (Protasan UP G 113, m.w 150200 kDa, degree of
deacetylation 7590%) was purchased from Novamatrix (Norway).
Fluorescein 5-isothiocyanate (FITC) with P90% purity, glycerol
(P99% purity), polyethylene glycol 1000 and sodium tripolyphosphate 85% were purchased from SigmaAldrich (Germany). Polyethylene glycol 5000 monomethyl ether was obtained from Fluka
(US). Polyethylene glycol 600 and HPLC grade water were purchased from Fischer Scientic (UK). HFA-227 (1,1,1,2,3,3,3-heptauoropropane) was obtained from INEOS Fluor (UK).
2.2. Preparation of nanoparticles
Nanoparticles were produced using an ionic gelation method
[24], whereby sodium tripolyphosphate (TPP) solution (0.5 mg/
ml) was added drop-wise to a chitosan solution (1 mg/ml) in the

75

ratio of 1:5 (w/w) using a peristaltic pump (Gilson, France) under

constant stirring. Three molecular weights of PEG (600/1000/
5000) were used, and the same protocol was followed to prepare
chitosanTPPPEG (5:1:30) nanoparticles by dissolving PEG with
TPP prior to its addition to chitosan. The ratio of chitosan/TPP/
PEG was established in preliminary experiments (data not shown),
which demonstrated that this ratio was optimal for preparation of
non-aggregated, nano-sized particles.
2.3. Preparation of uorescent crosslinked chitosan nanoparticles
Chitosan was labelled with FITC by conjugating the primary
amine group of the chitosan with the isothiocyanate group of FITC
[25,26]. Briey, 1% (w/v) of chitosan was dissolved using 0.1 M acetic acid, followed by an equal volume of methanol (16 ml). To this,
2.5 ml of FITC in methanol (2 mg/ml) was added drop-wise with
constant stirring for 3 h. The solution was adjusted to pH 10.0 by
the addition of 0.5 M sodium hydroxide solution, which resulted
in a precipitate, and then centrifuged at 25,000  g for 10 min.
The FITC-labelled chitosan precipitate was washed with methanol/water (70:30) mixture and centrifuged until no uorescence
was detected in the supernatant at excitation and emission maxima of 492 and 518 nm, respectively (LS 55 Fluorescence Spectrometer, PerkinElmer, UK). The precipitate was freeze-dried
(Virtis Advantage, SP Scientic, USA) to obtain a dry powder. Labelling efciency (percent weight of FITC to weight of the FITC-chitosan) was calculated by measuring the uorescence intensity of the
FITC-labelled chitosan solution against a standard solution of FITC.
Chitosan-conjugated FITC was used to prepare uorescent nanoparticles as described in Section 2.2, and these were used in
in vitro aerosolization studies.
2.4. Size and surface charge of nanoparticles
The hydrodynamic diameter and zeta potential of nanoparticles
were determined using photon correlation spectroscopy (PCS) and
laser Doppler electrophoresis (LDE), respectively, using a Zetasizer
(Nano ZS, Malvern Instruments, UK). The instrument parameters,
such as refractive index (1.333) and viscosity (0.8872 cP), were
set for the dispersion medium (water) in which the nanoparticles
were dispersed at 25 C. Data are presented as the mean of three
independent experiments.
2.5. Morphology of nanoparticles
Each formulation was freeze-dried, and a small sample of the
dried particles was placed on a scanning electron microscopy
(SEM) stub. Samples were sputter-coated with gold and examined
by SEM (FEI XL30 TMP, Philips, the Netherlands). The morphology
and integrity of nanoparticles post-aerosolization were determined by actuating pMDI formulations of nanoparticles from a distance of approximately 15 cm onto a glass slide, which was cut into
pieces, mounted on an SEM stub and visualized by SEM as described above.
2.6. Effect of pH on nanoparticle properties
The effect of pH on the size and zeta potential of nanoparticles
was studied using a Zetasizer (Nano ZS, Malvern Instruments, UK)
equipped with an auto-titration unit (MPT-2, Malvern Instruments,
UK). The aqueous dispersion of nanoparticles (12 ml) was titrated
with 0.1 M sodium hydroxide solution (NaOH) under constant stirring over a range of pH (5.58). The titrated dispersion was transferred to a measuring capillary cell by a spinning disc, and changes
in the properties of the nanoparticles were measured as a function
of pH.

76

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

2.7. Preparation of nanoparticle dispersions in HFA propellant

A uniform dispersion of nanoparticles in propellant HFA-227
was prepared. Briey, 7 mg of freeze-dried nanoparticles was
added to a clear transparent polyethylene terephthalate (PET) vial.
Vials were crimped with a continuous valve (Valois, France) using a
pneumatic crimper (P2002/20, Pamasol, Switzerland). 0.5 g of propellant was manually added through the valve. The mixture was
vortexed for 30 s and sonicated for 5 min at 20 C. The process
was repeated twice, and at each step, additional propellant was
added until the desired weight of 8.4 g was achieved. These PET
vials were used for studies of nanoparticle dispersion. A similar
process was followed to manufacture pMDI batches containing
FITC-labelled nanoparticle formulations. In this case, coated aluminium canisters (supplied by AstraZeneca, UK) were employed.
After lling with nanoparticle formulation (7 mg), the canisters
were crimped with metering valves (Valois, France) having a delivery volume of 50 lL per actuation. Additional propellant HFA-227
was added through the valve in a step-wise manner using automatic pressure lling equipment (P2011, Pamasol, Switzerland)
to achieve a total content weight of 8.4 g. This corresponds to a
dose of 59 lg nanoparticles per 50 lL actuation. All pMDI canisters
were vigorously shaken for 10 s; stored inverted for one month at
ambient temperature, and later, after insertion into a plastic actuator (supplied by AstraZeneca, UK), investigated for ex-actuator
aerosol particle size distribution and ne particle fraction (FPF).
2.8. Analysis of nanoparticle suspension stability
PET vials containing the HFA-227 nanoparticle suspensions
were vigorously hand-shaken and then placed in a light box for visual inspection to analyse their physical characteristics such as
sedimentation, creaming, occulation and coalescence. The stability and dispersion behaviour were monitored by naked eye observation after manual shaking ceased, until sedimentation or
creaming was apparent. Photographic images were taken at 10 s
and 1 min time points.
HFA-based formulations in PET vials were further investigated
using an optical analyzer; Turbiscan (MA2000, Formulaction,
France) [27]. The content of each vial was transferred, via the valve,
into a pressure-sealed glass tube, using purpose-built apparatus
(courtesy of AstraZeneca, UK), then studied using the Turbiscan.
After shaking, the entire length of the glass tube was scanned four
times at 1-min intervals using a light source (NIR, k = 850 nm) and
the backscattered light measured.
2.9. Aerosol particle size analysis
Aerosol particle size analysis was performed using a Sympatec
particle size analyzer (Sympatec GmbH System-Partikel-Technik,
Germany). pMDI suspensions were shaken 10 times and actuated
via a sealed central adapter into the Sympatec at a ow rate of
60 L/min for 10 s. The aerosol cloud generated passed through
the laser beam. An optical lens (0.4587.5 lm size range) was used
to collect the diffracted light for calculation of size distribution. The
rst ten doses of each new pMDI were red to waste and a time
interval of 60 s elapsed between each actuation to prevent excessive cooling of the pMDI metering chamber. The data generated
are presented as 10th (D10), 50th (D50, volume median diameter;
VMD) and 90th (D90) percentile of the cumulative particle undersize frequency distribution.
2.10. Determination of aerosol parameters using the next generation
impactor
The aerosol performance of FITC-labelled pMDI formulations
was determined using a next generation impactor (NGI; Copley

Scientic, UK), operated in accordance with the specications for

pMDIs described in the European Pharmacopoeia [28]. The collection plates of the NGI were uniformly coated with glycerol prior to
the measurement, with stage 8 comprising a micro-orice lter, to
collect very ne particles. The NGI was operated at an air ow rate
of 30 L/min for preset 10 s intervals following pMDI actuation, to
allow particle deposition on the plates. The pMDI was vigorously
shaken for 10 s before actuation. Following actuation of ve shots
to waste, the pMDI was actuated into the NGI setup, with a 60-s
interval between each actuation. Twenty actuations were cumulatively collected on glycerol-coated plates for each pMDI canister.
The collection plates, micro-orice lter, throat and actuator were
rinsed with water and washings collected and made up to volume.
The amount of FITC-labelled crosslinked chitosan nanoparticles
collected from each stage was determined using a uorescence
spectrophotometer as described previously. The parameters calculated for the aerosols produced were ne particle fraction (FPF;
stage 3 to lter, i.e. <6.4 lm) as a percentage of emitted dose, mass
median aerodynamic diameter (MMAD) and geometric standard
deviation (GSD) [29].
2.11. Statistical analysis
The data are presented as a mean standard deviation. The
physicochemical properties of non-PEG and PEG-based crosslinked
chitosan nanoparticles were compared statistically using a nonparametric KruskalWallis test followed by a Nemenyis post hoc
test. A MannWhitney non-parametric test was performed to compare samples. The results were considered signicantly different
based upon 95% probability values (p < 0.05).
3. Results and discussion
3.1. Nanoparticle size and zeta potential
The size and zeta potential of crosslinked chitosan nanoparticles with and without PEG are shown in Table 1. The mean hydrodynamic diameter of the nanoparticles ranged from 166 to 221 nm.
Sizes in this range are essential if such particles are intended for
the delivery of biopharmaceuticals such as plasmid DNA and siRNA, since nanoparticles have a signicantly higher intracellular uptake than microparticles [30]. The mean zeta potential for all
formulations was positive, ranging from +24 to +30 mV as a result
of the amine groups in chitosan. The cationic nature of these nanoparticles would be useful for production of complexes with negatively charged nucleic acids for bio-therapeutic applications [31]
and for promoting high transfection efciency within negatively
charged cells, as a result of electrostatic interactions [32]. There
was no signicant difference (p > 0.05) in zeta potential between
the four formulations, whether FITC-labelled or unlabelled. A signicant difference (p < 0.05) in nanoparticle size was observed.
Post hoc Nemenyis test showed that crosslinked chitosanPEG
5000 nanoparticles (both FITC-labelled and unlabelled) were signicantly larger than crosslinked chitosan particles without PEG.
This indicates that PEG (molecular weights 6005000) had no
inuence on nanoparticle surface charge, but that the highest
molecular weight PEG 5000 increased particle size, presumably
due to the presence of the large chain polymer on the particle surface. Labelling of particles had no signicant effect on size or surface charge characteristics (p > 0.05) in comparison to unlabelled
nanoparticles, and hence, both would be predicted to behave in a
similar manner when incorporated into pMDI formulations. The
polydispersity index (PDI) was slightly higher for unlabelled nanoparticles compared to labelled nanoparticles. For all formulations,
the PDI was less than 0.3, which is an indication of a narrow parti-

77

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

Table 1
Size and zeta potential of FITC-labelled and unlabelled chitosan and chitosanPEG nanoparticles, (mean S.D., n = 3).
Formulation

Chitosan
ChitosanPEG 600
ChitosanPEG 1000
ChitosanPEG 5000

Unlabelled nanoparticles

FITC-labelled nanoparticles

Hydrodynamic diameter
(nm S.D) (PDI S.D)

Zeta potential
(mV S.D)

Hydrodynamic diameter
(nm S.D) (PDI S.D)

Zeta potential
(mV S.D)

169.6 5.9
174.7 4.5
193.9 7.2
210.1 9.0

+25.8 3.3
+23.7 0.8
+28.2 3.7
+25.0 3.4

166.0 3.0 (0.194 0.015)

183.0 8.2 (0.213 0.015)
203.0 8.5 (0.141 0.021)
221.0 10.6 (0.217 0.009)

+30.0 1.7
+25.0 3.6
+24.0 5.6
+27.0 6.1

(0.259 0.025)
(0.267 0.002)
(0.248 0.003)
(0.255 0.018)

cle size distribution in all instances. A relatively high proportion of

PEG was employed during particle manufacture to ensure that it
was trapped within the crosslinked chitosan nanoparticles during
ionic gelation, permitting modication of their surface properties.
3.2. Morphology of nanoparticles
The nanoparticles produced by the described methodology
were spherical and of approximately uniform size with a smooth
surface. Fig. 1a, for example, shows smooth, spherical crosslinked
chitosanPEG 1000 nanoparticles having a narrow size distribution. Similar SEM images were obtained for non-PEG, PEG 600
and PEG 5000 based crosslinked chitosan nanoparticles. Fig. 1b
shows that crosslinked chitosanPEG 1000 nanoparticles collected
post-actuation from a pMDI canister, via the metering valve, were
similar morphologically to freeze-dried nanoparticles prior to
pMDI manufacture.
3.3. Effect of pH on nanoparticle properties
In order to explore the potential of chitosan-based nanoparticles for lung delivery, it is important to consider their behaviour
in different physiological conditions, especially at lung pH; 6.5
[33]. The change in zeta potential of nanoparticles over a pH range
5.58.0 is shown in Fig. 2a. The measured surface charge for all four
formulations was very similar at any single measured pH. The high

positive surface charge density for crosslinked chitosan at lower

pH is due to the free surface amine groups of chitosan. As the pH
of the nanoparticle suspension was increased, a greater proportion
of amine groups were deprotonated resulting in a decrease in the
measured positive zeta potential for the particles.
The inuence of pH on nanoparticle size is shown in Fig. 2b. At a
pH range of 5.56.5, the mean nanoparticle size was constant. The
positive charge of chitosan in acidic medium results in repulsion
between nanoparticles [34]. However, as the pH was increased,
the mean measured particle size increased, which suggests the
occurrence of aggregation. Such increases were more marked for
non-PEG-based formulations, compared to those with PEG, suggesting that the association of PEG with crosslinked chitosan nanoparticles provides steric hindrance, preventing nanoparticles from
aggregating. At pH 7.5 and greater, the measured particle size increased due to decreased surface charge (Fig. 2a), leading to aggregation of all formulations, and both PEG- and non-PEG-based
dispersions became turbid in appearance. This agrees with another,
parallel study [35], where chitosan and enoxaparin complexes
showed physicochemical stability at a pH range of 36.5, but
aggregated at higher pH.
From these studies, it was observed that the physical stability of
crosslinked chitosan nanoparticles is pH dependent, but signicant
aggregation does not occur at lung pH; 6.5 [33]. Further, these nd-

Fig. 1. Scanning electron micrographs of chitosanPEG 1000 nanoparticles; (a)

freeze-dried and (b) post-actuation from a pMDI.

Fig. 2. pH-induced variation in physicochemical properties of nanoparticles

(mean S.D., n = 4); (a) zeta potential and (b) hydrodynamic diameter. (For
interpretation of the references to colour in this gure legend, the reader is
referred to the web version of this article.)

78

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

ings suggest that PEG-based crosslinked chitosan nanoparticles are

more resistant to aggregation as a result of changing pH than nonPEG-based formulations.

3.4. Visual analysis of nanoparticle dispersion stability

The physical stability of the nanoparticles following dispersion
in propellant HFA-227 is shown in Fig. 3. Preliminary experiments
suggested that nanoparticle dispersions in HFA-227 had greater
physical stability than in HFA-134a (data not shown), possibly
due to the greater density of HFA-227, and hence, this propellant
was used in all further studies.
Chitosan nanoparticles without PEG aggregated readily, and
phase separation occurred with clearly visible aggregates settling
rapidly (Fig. 3A). The inclusion of PEG 600 within the nanoparticles
did not result in any apparent change in their dispersion properties
(Fig. 3B). Both formulations were physically unstable, resulting in
aggregation in less than one minute. Previously, the inability of
PEG 600 to stabilize silica particles dispersed in a uorinated solvent (2H,3H-peruoropentane) has been reported [20]. Chitosan
PEG 1000 nanoparticles produced a homogeneous translucent dispersion in HFA-227, which exhibited no phase separation for up to
10 min (Fig. 3C). In addition to exhibiting the slowest rate of sedimentation, these nanoparticles readily re-dispersed upon shaking.
This property is essential for acceptable dose reproducibility from a
pMDI. A good stabilization excipient for pMDI formulation should
be well solvated and block particleparticle interaction [19]. PEG
1000 is able to solvate well within HFA-227 [36], though PEG solubility is likely to decrease with increasing molecular weight [37].
ChitosanPEG 5000 nanoparticles produced an opaque, milky dispersion in HFA-227, which separated into two phases (Fig. 3D),
with particles creaming to the surface of the liquid propellant, presumably as the density of these particles is less than the density of

Fig. 3. Images of formulations dispersed within pressure-sealed HFA-227 propellant at 10 s and 1 min time points following hand shaking: (A) Chitosan, (B)
ChitosanPEG 600, (C) ChitosanPEG 1000 and (D) ChitosanPEG 5000
nanoparticles.

the propellant. Although this formulation creamed in less than

1 min, it was re-dispersible on gentle shaking.
These PEG and non-PEG crosslinked chitosan nanoparticles did
not visibly adhere to the walls of PET vials, even after one months
storage. ChitosanPEG 1000 composition greatly improves the
nanoparticle physicochemical characteristics. From these ndings,
the polymeric chain length of PEG 1000 seems optimal to provide
steric stabilization and to minimize particleparticle interaction
between crosslinked chitosan nanoparticles in HFA-227. Previous
studies have revealed the stabilizing effect of PEG 1000 and polyvinylpyrrolidone (PVP) within the model propellant 2H, 3H-peruoropentane, resulting in reduced particleparticle and
particlecanister wall surface interactions [38]. Long-term stability, dispersibility and ease of redispersion of HFA-based pMDI formulations are key parameters for the quality of an inhalation
product. These ndings, undertaken using pressurized apparatus
to investigate properties in the clinically relevant propellant
HFA-227, indicate that PEG 1000, incorporated into crosslinked
chitosan nanoparticles, is an effective formulation strategy to produce viable nanoparticle dispersions for delivery from a pMDI.
3.5. Optical analysis of nanoparticle dispersion stability
The experimental data obtained using the Turbiscan are a qualitative or semi-quantitative indication of nanoparticle suspension
behaviour, in a pressurized system, with respect to time. A formulation is deemed to be unstable if the variation in its scan intensity
on a time scan graph is >10% [39]. The acquired scans for the percentage of backscattered light (%BS) obtained for each formulation
as a function of time and sample height are shown in Fig. 4. Turbiscan results obtained with the nanoparticle/HFA suspension
showed a sharp increase in backscattering (>10%) between the rst
scan (time = 0) and all the consecutive scans with greatest increase
at the lower regions of the tube (Fig. 4a). This is due to particle size
variation and particle migration to the bottom of the glass tube,
which causes a variation in light transmission through the glass
tube and an increase in backscattering. The nanoparticle suspension signal was modied in the presence of PEG 600, indicated
by a step-wise increase in BS signal level between each successive
scan (Fig. 4b). This signies a delay in the separation between two
different phases compared to the nanoparticle/HFA formulation
(without PEG). The variation in the percentage of backscattering
between the rst and the nal scan was greater than 10%, indicating that PEG 600 does not prevent aggregation, but only delays the
onset of aggregation.
The scan for PEG 1000crosslinked chitosan dispersions in HFA227 showed a 2% variation in backscattering for the entire scan at
all time points (Fig. 4c), indicating high stability for this formulation, that is, no settling or creaming behaviour within the HFA.
There was no apparent change in particulate size and volume fraction for the dispersion, even after 3-min post-shaking, leaving the
product homogeneous. It is likely that the nanoparticle-bound PEG
1000 covers the surface of the nanoparticles suspended within HFA
and reduces the interfacial energy between particulates and propellant improving suspension stability. The dispersion of crosslinked chitosanPEG 5000 nanoparticles in HFA-227 was opaque
and milky in appearance. Due to the turbidity of the suspension,
it was difcult to monitor the onset of creaming at the start of
phase separation using the naked eye. The Turbiscan was a useful
technique to identify the appearance of a clear layer at the bottom
of the glass tube and/or creaming at the top. The backscattering
signal increased to 25% between 40 and 45 mm, reecting the
creaming behaviour of the sample (Fig. 4d), and suggesting that
crosslinked chitosanPEG 5000 nanoparticles had lower density
than HFA-227 causing them to cream. These results demonstrate
that producing crosslinked chitosan nanoparticles that incorporate

79

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

PEG 1000 was successful in stabilizing the particle dispersion in

HFA-227. However, PEG 600 and PEG 5000 were not able to
achieve the same degree of physical stability when particles containing these polymers were dispersed in the propellant, such that
the particles either sedimented or creamed. These results are in
line with the data obtained by visual analysis of dispersions, as described in the previous section.

3.6. Aerosol particle size distribution

The post-actuation size distribution of the emitted aerosol

clouds from pMDIs is shown in Fig. 5, with each data point representing the mean size following actuation from four canisters containing the same formulation. The nanoparticles showed a wide
size distribution, with some multi-modality and a VMD of
54.24 lm (Table 2). This suggests that the nanoparticles were
highly aggregated within HFA-227 propellant and largely remained aggregated following actuation and subsequent rapid
evaporation of the propellant. This rapid evaporation or ashing
of the propellant results in dispersion of ne particles into the air
if formulation is appropriate [40]. PEG 600-based formulations
showed a bi-modal distribution, having a VMD of 23.16 lm
(Table 2). The inability of PEG 600 to stabilize silica particles has
previously been reported [20], which may be related to the relatively short length of the polymer chains, or the physical state of
PEG 600, which is liquid at room temperature, while PEG 1000
and 5000 are solids. There was a small subpopulation of particles
having a mode less than 3 lm, and similar observations were made
for crosslinked chitosanPEG 5000 particles, which had a multimodal size distribution, having a VMD of 28.43 lm (Table 2). These
data suggest that aerosolization of these three formulations using a
pMDI is not appropriate for lung delivery, as their median size
greatly exceed the size (15 lm) required for peripheral lung
deposition. In contrast, the crosslinked chitosanPEG 1000 particle

Fig. 5. Post-actuation size distribution data for pMDI systems containing chitosan
and chitosanPEG nanoparticles (mean S.D., n = 4).

Fig. 4. Back scattering proles of formulations dispersed within pressure-sealed

HFA 227 as a function of time (1 min) with percentage change on Y-axis and sample
height on X-axis, (a) Chitosan, (b) ChitosanPEG 600, (c) ChitosanPEG 1000 and (d)
ChitosanPEG 5000 nanoparticles. The increase or decrease in scan intensity on a
time scan graph correlating to the sample height below 2.5 mm (convex bottom of
glass tube) and above the meniscus (shown by major change in backscattering of
plots) of the sample is because of multiple light diffractions and has no physical
meaning. Hence, it is not taken into account for stability analysis. (For interpretation of the references to colour in this gure legend, the reader is referred to the
web version of this article.)

Table 2
Post-actuation cumulative size distribution for chitosan and chitosanPEG nanoparticles, (mean S.D., n = 4). D10, D50 and D90 represent the 10%, 50% and 90%
cumulative undersize diameter determined by laser diffraction.
Formulation

D10 S.D (lm)

D50 S.D (lm)

D90 S.D (lm)

Chitosan
ChitosanPEG 600
ChitosanPEG 1000
ChitosanPEG 5000

1.58 0.31
2.33 1.03
0.72 0.16
1.29 0.05

54.24 4.66
23.16 3.03
1.53 0.07
28.43 6.52

78.56 1.57
57.81 2.27
2.76 0.27
67.69 5.04

80

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

formulation produced a more uniform size distribution, with a

much smaller VMD of 1.53 lm, which is appropriate for deep lung
alveolar delivery [41]. The improvement in dispersibility of
crosslinked chitosan nanoparticles, when incorporating PEG 1000
as observed in previous sections, is thus reected in an improved
performance in delivery from a pMDI.
3.7. In vitro assessment of aerosolization performance
FITC was used as a uorescent marker in order to quantify low
concentrations of chitosan following aerosolization and deposition
in the NGI. FITC percentage content was 2.8%. The measured FPF
was 5.7 0.9%, 11.8 2.7% and 17.0 2.1% for crosslinked chitosan
nanoparticles, crosslinked chitosanPEG 600 and crosslinked
chitosanPEG 5000 nanoparticles, respectively (Fig. 6). The poor
performance of these formulations as aerosols may be attributed
to aggregation of particles within the liqueed HFA and/or an
inability to disperse during actuation and propellant evaporation
subsequent to actuation, resulting in poor deposition in the lower
stages of the NGI. ChitosanPEG 1000 nanoparticles showed highest (p < 0.05) FPF of 34.0 1.4%, with MMAD of 4.92 0.29 lm and
GSD of 3.35 0.92, indicating this was the most suitable crosslinked chitosan nanoparticle formulation for lung delivery, using
a pMDI. These results correlate well with the aerosol sizing by laser
diffraction described earlier. Chitosan deposition in all stages of the
NGI and accessories (actuator, mouthpiece and throat) was between 75% and 125% of the anticipated label claim, which is in
accordance with the criteria of the European Pharmacopoeia [28].
In comparison with the other formulations, the crosslinked chitosanPEG 1000 formulation in HFA demonstrated high and consistent aerosolization performance and good dose recovery (mass
balance).
The mass deposition of chitosan at each stage of the NGI is presented in Fig. 7, for the crosslinked chitosanPEG 1000 formulation. NGI is an instrument used to assess inhalation aerosol
performance and gives an indication of the likely performance of
an inhalation product in vivo. The FPF of 34% for this formulation
is predictive of the proportion of therapeutically useful aerosol
likely to reach the deep lung. This may be considered adequate
to exhibit a therapeutic effect, as commercially available pMDI
products deliver about 30% of the total emitted dose to the lungs
[42,43]. A similar deposition prole, with FPF of 31.5% was reported for semi-interpenetrating polymeric network microspheres
loaded with bovine serum albumin delivered from a DPI [44]. Previous reports of nanoparticle delivery from HFA-134a based pMDI
systems have reported a FPF of approximately 45% for insulin
loaded nanoparticles, employing a volatile oil as a dispersant [10]
and lysozyme nanoparticles with oleic acid, sorbitan trioleate or
dipalmitoylphosphatidylcholine as dispersants [8]. Lecithin-based
nanoparticles were successfully dispersed in HFA-227 and
achieved a high ne particle fraction (>58%) [7]. Successful pMDI
formulation depends on the properties of the dispersed material,
dispersant and propellant. This study has demonstrated the poten-

Fig. 6. Fine particle fraction of pMDI formulations of chitosan and chitosanPEG

nanoparticles determined using the NGI (mean S.D., n = 3).

Fig. 7. Deposition of chitosanPEG 1000 nanoparticles delivered from a pMDI into

the NGI, (mean S.D., n = 3).

tial of a novel crosslinked chitosanPEG 1000 formulation, capable

of delivery from a pMDI, as a nanocarrier for pulmonary drug
delivery.

4. Conclusion
This study has successfully demonstrated a formulation approach, potentially capable of delivering crosslinked chitosan
based nanoparticles to the lung using a pMDI. Nanoparticles prepared from crosslinked chitosan alone showed aggregation, and
the preparation was physically unstable when suspended within
an HFA-227 pressurized system. Inclusion of PEG during particle
production modied their properties. Due to the amphiphilic nature of PEG, it is likely to be inside the nanoparticle and at the particle surface; this is supported by the increased size of
nanoparticles with PEG 5000 and the improved dispersion properties, of PEG 1000 and 5000 containing particles. The presence of
PEG 1000, in particular, provided steric stabilization when incorporated into crosslinked chitosan nanoparticles, prior to dispersion in
HFA-227. Dispersibility of these particles and subsequent deposition from a pMDI into an NGI were greatly improved compared
to crosslinked chitosan particles alone, or those incorporating
PEG 600 and PEG 5000. PEG1000 may have produced the best results of the three PEGS investigated for a combination of reasons:
PEG 600 is a small liquid molecule, which may not be well retained
in the nanoparticles. PEG 1000 is well solvated in HFA, thus providing steric stabilization of the particles. PEG 5000 has a much longer
chain length than PEG 1000, which is likely to be less well solvated,
and association of the long PEG chains may result in aggregation.
The relatively high FPF and ease of redispersion, combined with a
small primary particle size and positive charge, suggest that the
crosslinked chitosanPEG 1000 nanoparticles have potential application in delivery of drugs and biopharmaceuticals, such as nucleic
acids to the lungs. However, these ndings pertain to a system that
does not include an active pharmaceutical ingredient. Further
studies are required to determine whether a nanoparticle formulation of crosslinked chitosan and PEG 1000 is optimal in the presence of an incorporated therapeutic agent.
This study will help in early phase new system development for
polymeric nanoparticle delivery of therapeutic agents using pMDIs.
Future work will involve the evaluation of dispersion stability and
aerosolization behaviour over long-term storage of nanoparticles
dispersed within pressurized propellant-based systems and the
inclusion of therapeutic molecules.

K. Sharma et al. / European Journal of Pharmaceutics and Biopharmaceutics 81 (2012) 7481

Acknowledgements
Financial support from AstraZeneca is gratefully acknowledged.
We also thank David McCarthy, UCL School of Pharmacy for assistance with electron microscopy. Special thanks to Hamid Merchant
(University of London) and Varsha Thakoersing (Leiden University,
The Netherlands) for numerous stimulating discussions.
References
[1] J. Birchall, Pulmonary delivery of nucleic acids, Expert Opin. Drug Deliv. 4
(2007) 575578.
[2] L.A. Dellamary, T.E. Tarara, D.J. Smith, C.H. Woelk, A. Adractas, M.L. Costello, H.
Gill, J.G. Weers, Hollow porous particles in metered dose inhalers, Pharm. Res.
17 (2000) 168174.
[3] M.M. Bailey, C.J. Berkland, Nanoparticle formulations in pulmonary drug
delivery, Med. Res. Rev. 29 (2009) 196212.
[4] N. Tsapis, D. Bennett, B. Jackson, D.A. Weitz, D.A. Edwards, Trojan particles:
large porous carriers of nanoparticles for drug delivery, Proc. Natl. Acad. Sci.
USA 99 (2002) 1200112005.
[5] C. Vauthier, D. Labarre, Modular biomimetic drug delivery systems, J. Drug
Deliv. Sci. Technol. 18 (2008) 5968.
[6] C. Mayer, Nanocapsules as drug delivery systems, Int. J. Artif. Organs 28 (2005)
11631171.
[7] P.A. Dickinson, S.W. Howells, I.W. Kellaway, Novel nanoparticles for
pulmonary drug administration, J. Drug Target. 9 (2001) 295302.
[8] B.K. Nyambura, I.W. Kellaway, K.M.G. Taylor, The processing of nanoparticles
containing protein for suspension in hydrouoroalkane propellants, Int. J.
Pharm. 372 (2009) 140146.
[9] R.O. Williams, M.K. Barron, J.M. Alonso, C. Remun-Lpez, Investigation of a
pMDI system containing chitosan microspheres and P134a, Int. J. Pharm. 174
(1998) 209222.
[10] B.K. Nyambura, I.W. Kellaway, K.M.G. Taylor, Insulin nanoparticles: stability
and aerosolization from pressurized metered dose inhalers, Int. J. Pharm. 375
(2009) 114122.
[11] L. Illum, Chitosan and its use as a pharmaceutical excipient, Pharm. Res. 15
(1998) 13261331.
[12] O. Felt, P. Buri, R. Gurny, Chitosan: a unique polysaccharide for drug delivery,
Drug Dev. Ind. Pharm. 24 (1998) 979993.
[13] H.L. Luessen, B.J. de Leeuw, M.W. Langemeyer, A.B. de Boer, J.C. Verhoef, H.E.
Junginger, Mucoadhesive polymers in peroral peptide drug delivery. VI.
Carbomer and chitosan improve the intestinal absorption of the peptide
drug buserelin in vivo, Pharm. Res. 13 (1996) 16681672.
[14] R. Fernandez-Urrusuno, P. Calvo, C. Remunan-Lopez, J.L. Vila-Jato, M.J. Alonso,
Enhancement of nasal absorption of insulin using chitosan nanoparticles,
Pharm. Res. 16 (1999) 15761581.
[15] S.Y. Chae, M.K. Jang, J.W. Nah, Inuence of molecular weight on oral absorption
of water soluble chitosans, J. Contr. Rel. 102 (2005) 383394.
[16] A. Grenha, C.I. Grainger, L.A. Dailey, B. Seijo, G.P. Martin, C. Remun-Lpez, B.
Forbes, Chitosan nanoparticles are compatible with respiratory epithelial cells
in vitro, Eur. J. Pharm. Sci. 31 (2007) 7384.
[17] D.O. Corrigan, A.M. Healy, O.I. Corrigan, Preparation and release of salbutamol
from chitosan and chitosan co-spray dried compacts and multiparticulates,
Eur. J. Pharm. Biopharm. 62 (2006) 295305.
[18] R.J. Pugh, T. Matsunaga, F.M. Fowkes, The dispersibility and stability of carbon
black in media of low dielectric constant. 1. Electrostatic and steric
contributions to colloidal stability, Colloids Surfaces 7 (1983) 183207.
[19] D. Traini, P.M. Young, P. Rogueda, R. Price, Investigation into the inuence of
polymeric stabilizing excipients on inter-particulate forces in pressurised
metered dose inhalers, Int. J. Pharm. 320 (2006) 5863.
[20] A. Paul, P.C. Grifths, P.G. Rogueda, Towards an understanding of adsorption
behaviour in non-aqueous systems: adsorption of poly(vinyl pyrrolidone) and
poly(ethylene glycol) onto silica from 2H,3H-peruoropentane, J. Pharm.
Pharmacol. 57 (2005) 13831387.

81

[21] C. Vervaet, P.R. Byron, Drug-surfactantpropellant interactions in HFAformulations, Int. J. Pharm. 186 (1999) 1330.
[22] G. Pilcer, K. Amighi, Formulation strategy and use of excipients in pulmonary
drug delivery, Int. J. Pharm. 392 (2010) 119.
[23] L. Wu, M. Al-Haydari, S.R.P. Da Rocha, Novel propellant-driven inhalation
formulations: engineering polar drug particles with surface-trapped
hydrouoroalkane-philes, Eur. J. Pharm. Sci. 33 (2008) 146158.
[24] A. Trapani, J. Sitterberg, U. Bakowsky, T. Kissel, The potential of glycol chitosan
nanoparticles as carrier for low water soluble drugs, Int. J. Pharm. 375 (2009)
97106.
[25] C. Colonna, B. Conti, P. Perugini, F. Pavanetto, T. Modena, R. Dorati, P. Iadarola,
I. Genta, Ex vivo evaluation of prolidase loaded chitosan nanoparticles for
the enzyme replacement therapy, Eur. J. Pharm. Biopharm. 70 (2008)
5865.
[26] M. Huang, Z. Ma, E. Khor, L.-Y. Lim, Uptake of FITCchitosan nanoparticles by
A549 cells, Pharm. Res. 19 (2002) 14881494.
[27] P. Bru, L. Brunel, H. Buron, I. Cayr, X. Ducarre, A. Fraux, O. Mengual, G.
Meunier, A. de Sainte Marie, P. Snabre, Particle sizing and characterisation, in:
Particle Sizing and Characterisation, 2004. pp. 4560.
[28] European Pharmacopoeia, Preparations for Inhalation (2.9.18), in: European
Pharmacopoeia 6.0. 2008, Directorate for the Quality of Medicines of the
Council of Europe: Strasbourg, pp. 287300.
[29] T.C. Carvalho, J.I. Peters, R.O. Williams, Inuence of particle size on regional
lung deposition what evidence is there?, Int J. Pharm. 406 (2011) 110.
[30] M. Bivas-Benita, S. Romeijn, H.E. Junginger, G. Borchard, PLGAPEI
nanoparticles for gene delivery to pulmonary epithelium, Eur. J. Pharm.
Biopharm. 58 (2004) 16.
[31] S. Mao, W. Sun, T. Kissel, Chitosan-based formulations for delivery of DNA and
siRNA, Adv. Drug Deliv. Rev. 62 (2010) 1227.
[32] T. Sato, T. Ishii, Y. Okahata, In vitro gene delivery mediated by chitosan, Effect
of pH, serum, and molecular mass of chitosan on the transfection efciency,
Biomaterials 22 (2001) 20752080.
[33] K. Mohri, T. Okuda, A. Mori, K. Danjo, H. Okamoto, Optimized pulmonary gene
transfection in mice by spray-freeze dried powder inhalation, J. Contr. Rel. 144
(2010) 221226.
[34] T. Kiang, J. Wen, H.W. Lim, K.W.K.W. Leong, The effect of the degree of chitosan
deacetylation on the efciency of gene transfection, Biomaterials 25 (2004)
52935301.
[35] W. Sun, S. Mao, D. Mei, T. Kissel, Self-assembled polyelectrolyte
nanocomplexes between chitosan derivatives and enoxaparin, Eur. J. Pharm.
Biopharm. 69 (2008) 417425.
[36] R.P. Peguin, S.R. da Rocha, Solvent-solute interactions in hydrouoroalkane
propellants, J. Phys. Chem. B 112 (2008) 80848094.
[37] K.B. Ridder, C.J. Davies-Cutting, I.W. Kellaway, Surfactant solubility and
aggregate orientation in hydrouoroalkanes, Int. J. Pharm. 295 (2005) 5765.
[38] R. Ashayer, P.F. Luckham, S. Manimaaran, P. Rogueda, Investigation of the
molecular interactions in a pMDI formulation by atomic force microscopy, Eur.
J. Pharm. Sci. 21 (2004) 533543.
[39] C. Celia, E. Trapasso, D. Cosco, D. Paolino, M. Fresta, Turbiscan Lab expert
analysis of the stability of ethosomes and ultradeformable liposomes
containing a bilayer uidizing agent, Colloids Surf. B Biointerfaces 72 (2009)
155160.
[40] K.M.G. Taylor, Pulmonary drug delivery, in: Aultons Pharmaceutics: The
Design and Manufacture of Medicines, Churchill Livingstone, Edinburgh, 2007,
pp. 539554.
[41] W. Stahlhofen, J. Gebhart, J. Heyder, Experimental determination of the
regional deposition of aerosol particles in the human respiratory tract, Am.
Ind. Hyg. Assoc. J. 41 (1980) 385398.
[42] J.L. Rau, The inhalation of drugs: advantages and problems, Respir. Care 50
(2005) 367382.
[43] P.H. Hirst, G.R. Pitcairn, J.G. Weers, T.E. Tarara, A.R. Clark, L.A. Dellamary, G.
Hall, J. Shorr, S.P. Newman, In vivo lung deposition of hollow porous particles
from a pressurized metered dose inhaler, Pharm. Res. 19 (2002) 258264.
[44] I.M. El-Sherbiny, H.D.C. Smyth, Biodegradable nano-micro carrier systems for
sustained pulmonary drug delivery: (I) self-assembled nanoparticles
encapsulated in respirable/swellable semi-IPN microspheres, Int. J. Pharm.
395 (2010) 132141.