Advanced Drug Delivery Reviews 58 (2006) 1009 – 1029

www.elsevier.com/locate/addr

Particle engineering techniques for inhaled biopharmaceuticals☆

Sunday A. Shoyele ⁎, Simon Cawthorne
3M Health Care Ltd, Morley Street, Loughborough, LE11 1EP, UK
Received 14 March 2006; accepted 25 July 2006
Available online 12 August 2006

Abstract

Formulation of biopharmaceuticals for pulmonary delivery is faced with the challenge of producing particles with the
optimal properties for deep lung deposition without altering the native conformation of these molecules. Traditional techniques
such as milling are continuously being improved while newer and more advanced techniques such as spray drying, spray freeze
drying and supercritical fluid technology are being developed so as to optimize pulmonary delivery of biopharmaceuticals.
While some of these techniques are quite promising, some are harsh and impracticable. Method scale up, cost-effectiveness and
safety issues are important factors to be considered in the choice of a technique. This paper reviews the presently developed
techniques for particle engineering biopharmaceuticals.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Biopharmaceuticals; Protein/peptide; Particle engineering; Aerodynamic diameter; Spray drying; Supercritical fluid

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1010
2. Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1011
2.1. Jet milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1011
2.2. Media milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1012
3. Spray drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013
3.1. CO2-assisted nebulization with a bubble dryer™ (CAN–BD) . . . . . . . . . . . . . . . . . . . . . . 1016
3.2. Supercritical fluid-assisted atomization (SAA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016
3.3. Nebulization, air drying and electrostatic collection (NAE) . . . . . . . . . . . . . . . . . . . . . . . . 1016
4. Spray freeze-drying (SFD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017
4.1. Spray freezing into liquid (SFL). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018
4.2. Spray-freezing with compressed CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018

☆
This review is part of the Advanced Drug Delivery Reviews theme issue on ʽʽChallenges & Innovations in Effective Pulmonary Systemic &
Macromolecular Drug Delivery", Vol. 58/9–10, 2006.
⁎ Corresponding author. Tel.: +44 1509613921; fax: +44 1509613944.
E-mail address: sashoyele@mmm.com (S.A. Shoyele).

0169-409X/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.addr.2006.07.010
1010 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029

5. Supercritical fluid technology (SF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018

5.1. RESS process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1020
5.2. PGSS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1020
5.3. GAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
5.4. ASES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
5.5. SEDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022
5.6. PCA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022
6. Microcrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
6.1. Microcrystallization and sustained release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1024
6.2. Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1024
6.3. Controlling polymorphism with SF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1024
7. Aerodynamic particle diameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025

1. Introduction Although, the oral route has showed some promise

A biopharmaceutical is a protein or nucleic acid-based
pharmaceutical substance used for therapeutic or in vivo
diagnostic purposes, which is produced by means other
than direct extraction from a native (non-engineered)
biological source [1]. In other words, biopharmaceuticals
include recombinant proteins, monoclonal antibodies
and nucleic acid-based therapeutics.
The biopharmaceutical industry started in the late
1970s [2] with the development of the twin techniques
of genetic engineering and hybridoma technology, but it
was not until 1982 that the first biopharmaceutical,
‘Humulin™’ (recombinant human insulin, developed
by Genentech and Eli Lilly) gained marketing approval
in the US [3]. Since this period, the market for
biopharmaceuticals for use in human therapy has
continued to grow at a high rate with 270 new
biopharmaceuticals being evaluated in clinical studies
currently and estimated sales in excess of
30 billion euros annually [4]. Presently, up to 160 bio-
pharmaceuticals have gained medical approval and most
are protein-based, although two nucleic acid-based
products are now on the US/European market [5].
Most of these approved biopharmaceuticals are
being delivered via injection. Apart from the inconve-
nience this might cause to the patient (pain, abscess,
etc.), injections, in the majority of cases, cannot be
administered by patients themselves. There is a need
for some expertise of a healthcare professional. These
factors do not help compliance and so there is a need
for alternative methods of administering these drugs to
the patient.
[6,7], delivering biopharmaceuticals via this route is
still quite challenging because of the vulnerability of
these molecules to some gastro-intestinal enzymes and
reduced bioavailability caused by first pass metabo-
lism. Alternative routes such as transdermal, nasal,
buccal and ocular have been investigated [8–12].
Although, these routes showed little success, the search
for a suitable alternative to injection still continues.
Systemic delivery of biopharmaceuticals by inha-
lation is currently attracting considerable attention
because a number of these molecules are more
efficiently absorbed from the lungs compared to oral,
nasal or transdermal routes [13].
The inhalation route offers an enormous absorptive
surface area in the range 35–140 m2, thin (0.2 μm) and
highly vascularized epithelium, which leads to high
bioavailability [14]. Furthermore, in most societies oral
inhalation is well accepted by the general population
[15].
Formulating biopharmaceuticals for inhalation pre-
sents new challenges to scientists. It is important that
these molecules be size reduced without altering their
native conformations and also have optimal aerody-
namic properties for them to get to the deep lungs
when inhaled [16].
Since the mucocilliary apparatus and alveolar
macrophages are both involved in the clearance of
drug molecules from the lung [17]; there is the
possibility of producing particles to avoid these
mechanisms. For instance, uptake of particles by the
alveolar macrophages has been found to be size-
dependent [17]. Particles 6 μm are phagocytosed to a
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
much smaller extent than those 3 μm in diameter [18],
while particles of diameter less than 0.26 μm are
minimally taken up by macrophages. However, if the
objective of delivery were to target the alveolar
macrophages as a medium for drug delivery [19,20],
particles of mass median diameter of around 3 μm would
be the ideal particles to target. Depending on what the
objectives of delivery are, particles of a specific size
distribution will have to be produced for effective
delivery.
Biopharmaceuticals are relatively large molecules;
studies have shown that molecules with molecular
weight above 30 kDa may need an absorption enhancer
to be absorbed in the alveoli [21]. Some of the absorption
enhancers used involve entrapping or encapsulating
these molecules within them [22]. This can potentially
increase the size of the particles produced. There is
therefore a need to reduce these particles to within the
respirable range.
Whilst it is crucial to control the aerodynamic particle
size, it is also important to control the surface properties
of reduced particles as micronized particles tend to be
highly charged leading to agglomeration. Furthermore,
milling damages the crystal surface of particles leading
to partially amorphous parts, which leads to physical
instability of the drug [23].
Inhaled biopharmaceuticals therefore need to be
size reduced by suitable techniques that not only
produce particles within the target range of 1–5 μm but
also with optimal surface properties to aid pulmonary
delivery without loss of biological activity.
Because most biopharmaceuticals are quite labile
[24], care should be exercised during their processing.
Conventional micronization techniques such as milling
have been used with limited success [25,26]. Alterna-
tive techniques are continuously evolving leading to
the production of particles with the required optimal
properties which include among others, narrow size
distribution, optimal aerodynamic diameter and low
surface energy.
2. Milling
Milling is the traditional method of drug powder
micronization [27]. Various mills like ball mills,
colloid mills, hammer mills and jet or fluid energy
mills have been studied [28], however, the majority of
inhalation powders are prepared with a jet mill [29].
1011
2.1. Jet milling
Jet milling process involves micronization by inter-
particle collision and attrition [30]. The milling medium
is a high velocity gas (air), which passes through the
nozzles and carries along coarse particles into the jet mill
where they are suspended in the turbulent gas stream.
Coarse particles are fractured into smaller once in the gas
stream due to inter-particle collisions [31]. The spherical
and elliptical path taken by the gas stream moving at high
speed, facilitates the exit of fine particles while larger
particles remain in the mill for further micronization by
more inter-particle collision. Jet mills can typically
produce particles within the size range of 1–20 μm [25].
Despite the possibility of having fine materials
through jet milling, the disadvantages such as lack of
control over parameters such as size, shape, morphol-
ogy and surface properties of the milled particles
coupled with the high energy input which can promote
chemical degradation [29], make this technique less
promising for biopharmaceuticals. Furthermore, it has
been observed that the high energy required for such
processes often damage the crystal surface, leading to
highly charged and cohesive particles which result in
chemical and physical instability of the drug [23].
These limitations, however, have not dissuaded
scientists from exploring this cost-effective technique.
It has been studied as a potential particle engineering
technique for various proteins.
The model protein horseradish peroxidase has been
co-jet milled with poly(acrylate) to produce protein-
loaded microparticles [32] after co-precipitating the two
compounds from demineralized water. Precipitating
with petroleum ether led to a remaining activity of 67±
4.7% while isopropanol was the best precipitant leading
to a remaining activity of approximately 90%. It was
concluded that co-precipitation has a much greater
impact on protein stability than the milling process itself.
Leuprolide acetate, a synthetic nonapeptide has been
jet-milled and formulated into a metered dose inhaler
(MDI) [33,34]. It was observed that suspension MDI of
leuprolide had a higher bioavailability (27.9% relative to
subcutaneous administration) compared to solution
MDI, using 20% ethanol as a co-solvent (6.6% relative
to subcutaneous administration). This low bioavailabil-
ity could be attributed to increased chemical degradation
of the protein facilitated by the ethanol or coarse aerosol
plume in the ethanol system leading to reduced
1012 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
deposition in the deep lung, resulting in reduced uptake
into the systemic circulation.
Insulin that is available as a crystalline powder [25]
has been jet-milled using nitrogen as the gas, to an
average particle size of 2.4 μm [35]. The occurrence of
insulin as crystals makes it amenable to jet milling due
to good flow properties.
The nature of the particles fed into the jet mill has a
substantial effect on the efficiency of the mill [25]. If
particles fed in are too large, they can block the hopper
and pipe-work in the mill while too fine powder may
make it difficult to obtain a smooth feed rate into the mill,
with the optimal size of most materials for jet milling
being in the 75–100 μm range. Brittle materials tend to
fracture during inter-particle collision while ductile
materials tend to undergo plastic deformation rather
than fracture [28]. However, Tomihisa [36] proposed that
chilling ductile materials before and during milling
makes them more brittle, which facilitate size reduction.
Biopharmaceuticals that are not available as crystals
or forms suitable for jet milling can be lyophilized to
prepare coarse powders [25]. Kobayashi et al. jet-milled
salmon calcitonin powders that had to be lyophilized
with lactose in order to improve the flow of the powder
prior to milling [37]. However, lyophilization followed
by milling can lead to loss of activity of the biomolecule
unless a suitable excipient is used as a lyoprotectants
during lyophilization. Lyoprotectants are important
during lyophilization because at sub-zero temperatures,
water crystallizes and leads to a strong increase in the
concentration of the molecules dispersed in it. This so-
called freeze concentration forces particles towards each
other, which facilitates aggregation. [38]. Lyoprotectants
prevent such aggregation. Preventing aggregation during
lyophilization is further important because once in a
stable dry state, increased shelf life can be expected even
at ambient temperature [38,39]. For instance, a formu-
lation of interferon-β containing sorbitol, human serum
albumin and NaCl, did not significantly reduce its
activity after lyophilization and jet milling, while the
same formulation without sorbitol showed significant
loss of activity [40,41].
Although the use of jet milling in size reduction of
biopharmaceuticals has been relatively successful, the
two-step approach of either crystallizing or lyophilizing
before milling can make this technique more expensive
and time-consuming. It also sometimes lead to degra-
dation of the drug as shown above.
2.2. Media milling
Milling as a technique has been improved and
adapted for the engineering of biopharmaceutical
molecules. Media milling is a concept that has found
some application in biopharmaceutical processing. It
involves using a liquid as a medium during milling and
was first recommended by Adjei et al. using non-CFC
propellants for the preparation of peptide suspensions
for pressurized metered dose inhaler (pMDI) [42].
Lizio et al., however, improved the technique using a
modified and abrasive resistant pearl mill [43]. They
were able to produce fine protein particles with a
volumetric mean diameter of 3.1 μm without gener-
ating degradation products or contaminants from the
milling process. An important feature of this technique
is the direct manufacturing of fine drug suspension for
systemic aerosol delivery by pMDI.
Media milling has also been used by Zijlstra et al. to
produce protein with comparable qualities to particle
engineering techniques such spray drying and spray
freeze drying [33]. They worked with cetrorelix
acetate, a decapeptide used for the controlled ovarian
super stimulation for assisted reproductive technique.
Using a pearl mill equipped with a cryostat, the system
was cooled to − 60 °C, cetrorelix lyophilisate in HFA
227 was milled with ZrO2-Ir stabilized pearls for
15 min. It was observed that only minute amounts of
degradation products (b1.17%) were produced. This is
indicative of the stability of the milled cetrorelix,
although minimal metal contaminants were observed.
Scanning electron micrograph (Fig. 1) showed that
milled cetrorelix consist of agglomerates of very small
primary particles which are mostly smaller than 1 μm.
Dynamic vapour sorption studies showed that the
surface properties of milled cetrorelix was comparable
to those from spray drying or spray freeze drying. The
aerosol performance of the dry powder formulation
was assessed by comparing the fine particle fraction
(FPF) (b3.75 μm) of emitted powder from both the
ISF inhaler and the Novolizer. The ISF inhaler
produced an FPF of 30% while the novolizer produced
an FPF of 40%.
Irngartinger et al. also worked on media milling
cetrorelix [44]. Using the same modified pearl mill as
described above, but incorporating 96% ethanol as a
coolant, the cetrorelix was suspended within fluid
propellant and after the milling, the suspension of
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
Fig. 1. SEM showing media-milled cetrorelix acetate. Reprinted
from Zijlstra et al. [33] with permission from Elsevier.
milled cetrorelix acetate in propellant was poured into
a flask and rotated at room temperature until the
propellant was completely evaporated. The dry
powder obtained can be formulated as a dry powder
inhaler, which makes this technique quite versatile.
The use of media milling has improved the use of
milling for size reducing proteins and other biomole-
cules as the low temperature of milling could reduce
deformation of these biomolecules and possibly reduce
the number of steps involved in milling.
3. Spray drying
In the spray drying process a feed solution containing
the biopharmaceutical is atomized into droplets that dry
rapidly due to their high surface area and intimate
contact with the drying gas (compressed air) [45]. The
drying time for droplets depends on the process
conditions such as flow rate, pump rate, aspiration rate
and heat. This drying time can range from less than
100 min to a few seconds [25]. The temperature
experienced by the droplets is considerably lower than
the temperature of the drying air due to evaporative
cooling. The dried powder is protected from overheating
by rapid removal of solvent from the drying zone. The
final product can be removed from the air stream by the
use of cyclones or filters.
Although thermal degradation can possibly be
avoided in spray drying, degradation during atomiza-
tion may constitute a problem to biopharmaceuticals.
Atomization requires high shear rates, which can
denature proteins [25]. The large surface area of the
droplets improves mass transfer but can sometimes be
1013
a site for degradation of biopharmaceuticals. The use
of suitable excipients can reduce denaturation during
spray drying. For instance, polysorbate-20 has been
used to reduce spray drying induced denaturation of
human growth hormone [25].
Careful selection of operating parameters can play a
significant role in obtaining high quality product
during spray drying. Experimental design techniques
have been used to optimize the spray drying process
for β-galactosidase [46].
The yield obtained from spray drying also constitutes a
limitation to the use of this technique. Spray drying has
relatively low cyclone collection efficiency for particles
below 2 μm [25]. The typical yield from a spray dryer is
between 20% and 50% [47]. However, the introduction of
high-performance cyclone by Büchi, Switzerland, has
improved the yield from spray drying to greater than 70%
[48].
Other important issues in spray drying are the lack of
control over the mean droplet size, possibility of a broad
droplet distribution and the risk of clogging if pneumatic
nozzles are used [44]. The use of an ultrasonic nozzle,
which is able to generate droplets with more uniform size,
leading to a relatively homogenous size distribution of
the produced powder, has been suggested [49]. The
problem envisaged here is that the energy needed for
atomization may further contribute to the denaturation of
proteins.
Spray drying has however been recognized as a
successful process for generating protein-containing
powder in a single step from solutions [50]. It is a
technique that has found widespread use in particle
engineering of inhaled biopharmaceuticals, from pro-
duction of aerodynamic particles to controlled release
delivery systems. Particles from spray drying may be
spherical, have convoluted surfaces, asperities, holes or
voids [45,51].
Patton et al. have been able to prepare respirable
insulin powders by spray drying a solution of the drug
with sodium citrate and mannitol [52]. A powder with
particles typical of spray-dried particles, with a mass me-
dian diameter ranging from 2.0 to 2.8 μm was achieved.
Recombinant human granulocyte colony-stimulating
factor (rhG-CSF), a protein with a molecular weight of
18.8 kDa used in cancer chemotherapy has been
successfully spray-dried and tested for pulmonary
absorption in rabbits [53]. The product obtained had a
mass median diameter of less than 3 μm with a
1014 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
corrugated surface. It was observed that the rhG-CSF
was absorbed more rapidly from the lungs of the rabbit
than the subcutaneous dose. The inhalation powder also
had the desired pharmaceutical effect of increasing the
white blood cell count, showing that the biological
activity of the spray-dried powder was maintained.
Sutton et al. have succeeded in producing hollow
protein microcapsules using spray drying [54]. Al-
though the intended use for this fine powder is ultrasonic
imaging contrast enhancement, many of the particles
prepared had dimensions in the respirable range. This
study was also able to show the effect of nozzle type on
the dimensions of particle obtained. When a two-fluid
nozzle was used for atomization, the particle size ranged
from 4 to 6 μm, while when a centrifugal nozzle was
used, particles in the range 1–8 μm were obtained.
Flake-like, corrugated aerodynamic, cetrorelix parti-
cles (Fig. 2), have been produced by spray drying [33].
Lyophilized cetrorelix was dissolved in a mixture of
40 ml tert-butanol and 60 ml of distilled water. Tert-
butanol was added because it improves dispersion
characteristics of water [55]. The powder produced had
a fine particle fraction (FPF) of 66%, which is consistent
with the morphology of the particles produced.
Salmon calcitonin has also been spray-dried and the
effect of mannitol on aerosol performance of the powder
produced studied [56]. It was observed that surface
enrichment by mannitol increases surface energy
measured by inverse gas chromatography (iGC) and
emitted dose was improved as the mannitol concentra-
tion decreased.
Irngartinger et al. [44], studied the effect of outlet
temperature on the stability of spray-dried cetrorelix
Fig. 2. SEM showing spray-dried cetrorelix acetate. Reprinted from
Zijlstra et al. [33] with permission from Elsevier.
acetate and although particles with median diameter of
3.5 μm with 83% of the particles smaller than 5 μm
were produced, it was discovered that the outlet
temperature had an enormous effect on the rate of
degradation of the peptide. A sample exposed to nearly
60 °C at the product filter contained 1.3% of ornithin
derivatives while a sample exposed to nearly 50 °C
generated a just 0.6% of ornithin derivative.
The effect of co-spraying recombinant human
DNase (rhDNase) with NaCl on the aerodynamic
properties of the protein has been studied [57]. Spray-
dried pure rhDNase powder was quite cohesive with a
fine particle fraction of about 20% while powders
containing NaCl showed a linear relationship between
the NaCl content and FPF for a similar primary size
(approximately 3 μm volume median diameter) of
particles. The particle morphology of these powders
varied systematically with the salt content.
The effect of various excipients on the biological
activity and aerosol performance of spray-dried
lysozyme and catalase has also been studied [58].
Sucrose, trehalose and polyvinyl alcohol (PVA) were
studied as potential stabilizers for proteins during
spray drying. It was observed that pure lysozyme, lost
just over 10% of its original biological activity, while
the incorporation of sucrose, trehalose or a combina-
tion of trehalose and PVA with lysozyme preserved
approximately 100% of its original biological activity.
Pure catalase retained 55% of its biological activity
upon spray drying but this retention of activity was
increased to approximately 93% when either sucrose or
trehalose was used to stabilize the enzyme during the
manufacturing process. Catalase retained almost full
activity when it was stabilized using a mixture of PVA-
trehalose. SEM showed that the morphology of the
spray-dried powders appeared to be similar regardless of
the type and quantities of excipients used to stabilize the
protein.
Lactose, trehalose, dipamitoylphosphatidylcholine
(DPPC) and/or albumin have been studied as potential
excipients for spray drying proteins [59]. Lactose is of
particular interest because it had been established earlier
that lactose, being a reducing sugar, has the potential to
react with the primary amine of lysine residue and
generate Schiff-base by-product [60]. Using parathyroid
hormone (PTH) as model protein, after the spray drying
process, the mass spectrum of the powder containing
lactose did not show any by-product resulting from a
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
covalent bond between the sugar and PTH; this was also
the case for the spectrum of powder containing trehalose
a non-reducing sugar. No peaks from other degradation
products were observed by either HPLC or mass
spectrometry. It was further demonstrated that albumin
as an excipient could markedly decrease systemic
absorption of the hormone through binding. However,
all the powders produced presented an average primary
particle diameter between 3.9 and 5.9 μm and achieved
up to 98% emitted dose and up to 61% FPF using a DPI
device.
Further studies have been performed on the use of
lactose and DPPC as excipients for spray drying
proteins. Bosquillon et al. [61] prepared a dry powder
by spray drying human growth hormone (hGH) with
lactose and DPPC in proportion 20:20:60 w/w/w in a
0.1% total excipient concentration.
The proportion of soluble hGH dimers after powder
production was evaluated by size-exclusion chroma-
tography. The dimer content was b1% in native hGH as
well as in hGH powder, indicating that incorporation of
hGH in the a solution containing lactose and DPPC did
not induce dimerization. In contrast, 3.9% of hGH
dimers were obtained following spray drying with
lactose in the absence of DPPC. This shows that DPPC
has a stabilizing effect on proteins when co-spray-dried
with lactose. This assumption was further demostrated
by the fact that when a potential shift in iso-electric
point (pI) was analysed by iso-electric focussing, a dry
powder containing trehalose, a non-reducing sugar, as a
negative control, no modification to the protein pI was
observed after formulation of both lactose/DPPC
powder and trehalose/DPPC powder and up to 5 months
of storage at 4 °C for the powder containing lactose and
DPPC.
Stabilizing proteins with a surface acting agent
(surfactant) is a crucial step during spray drying of
proteins because proteins being amphipathic tend to
adsorb at the air–liquid interface of droplets in spray,
unfold and aggregate at the droplet surface. Excipients
have been found to exclude proteins from this interface
and promote their stability [62].
Spray drying as a technique has also found great use
in controlled release of biopharmaceuticals.
The PEG–insulin covalent complex developed for
controlled released insulin by Nektar has been prepared
for pulmonary delivery by spray drying [63]. Using a
Büchi 190 mini spray dryer, the ED of the aerosol from
1015
Nektar Pulmonary Delivery System was approximately
90% and a fine particle mass, defined as particles
b3.3 μm, was 56%.
Spray drying has also been used to prepare encapsu-
lated bovine serum albumin (BSA) from aqueous two-
phase systems (ATPs) containing polyvinyl alcohol and
dextran solutions in different proportions [64]. Using a
laboratory-scale spray dryer, the dryer was operated in
co-current mode, with a jacketed two-fluid nozzle. The
emulsion produced from the ATPs was continuously
stirred as it was fed into the spray dryer. Particles were
separated from the dry air by a membrane filter. SEM
micrograph of the powders produced showed that the
shape and surface morphology of particles were strongly
affected by the composition of the two-phase system. As
the PVA content increased, the particles turned from
highly wrinkled to smooth spheres. Thread-like struc-
tures and small doughnut shaped particles were also
observes as well as a slight increase of the particle size
when the PVA content increased.
Controlled release anti-influenza antibody has been
produced for inhalation using immunoglobulin (Ig)
incorporated spray-dried lipid micro-particles (SDLM)
[65]. DPPC was used as the major excipient because of
its biocompatibility and being a component of normal
lung surfactant. Either lactose or hydroxyethylstarch
(HES) was used as an additional excipient. Prior to
spray drying, the lipid was complexed with Ca2+ in a
1:1 molar ratio, yielding SDLM of increased stability.
Two categories of particles were generated:
• Retentive powder, containing 10% (w/w) HES in
addition to 60% (w/w) DPPC-Ca2+/DPPE-Texas
Red and 30% (w/w) of human IgG
• Non-retentive particles that were constructed by
replacing HES with lactose plus 1% tyloxapol, a
biocompatible surfactant detergent approved for
human use in surfactant formulations [65]
Co-spray drying proteins with biocompatible sur-
factant lipid, such as DPPC (with or without additives
such as tyloxapol) dramatically reduced the particle
size and improved the aerodynamic characteristics and
homogeneity of aerosols (MMAD of 3.42 μm with
FPF in excess of 70%). In addition, such surfactant–
lipid excipients make the microparticles compatible
with non-aqueous propellants, offering a degree of
flexibility in the strategy of delivery [53].
1016 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
It was concluded that controlled spray drying of IgG,
with DPPC-Ca2+ as a major excipient, results in the
generation of microparticles with good aerodynamic
properties.
Microencapsulated chitosan nanoparticles for lung
delivery of proteins have been produced by spray
drying [66]. The process involves, firstly, preparing the
protein loaded nanoparticles by a process described by
Calvo et al. [67] and then encapsulating them in lactose
and mannitol. The encapsulation was performed by
spray drying. Aqueous solutions of the excipients,
lactose and mannitol, were prepared by dissolution in
purified water to obtain concentration of 7.5%–10%.
Chitosan nanoparticles prepared were resuspended in
purified water and mixed with the excipients solution in
order to achieve excipient/nanoparticle ratio of 95:5,
90:10, 80:20, and 50:50 (w/w). The final suspension
concentration ranged from 0.84% to 8.4% (w/v). Dry
powders were obtained by spray drying aqueous
solutions of the excipients and suspension of chitosan
nanoparticles in the excipients using Büchi mini spray
dryer, B290 model with a feed rate of 2.5 ml/min.
The morphology of the particles produced was
dependent on powder composition. Mannitol solution
without nanoparticles produced mostly spherical but
aggregated microparticles while lactose solution pro-
duced well-defined, non-aggregated spherical micro-
spheres. Combination of mannitol (90%) and lactose
(10%) also produced spherical particles which were
less aggregated than the mannitol particles with better
defined limits. Spherical microparticles with more
defined limits were produced upon incorporation of
the protein-containing chitosan nanoparticles. This
might suggest that the incorporation of nanoparticles
as a solid structure enhances the morphology of the
microspheres. In the formulation comprising mannitol
with 10% lactose, no improvement in particle morpho-
logy was observed. In addition, no differences in
morphology were found between microparticles made
of mannitol-containing nanoparticles and microparti-
cles of both mannitol and lactose with nanoparticles.
Despite the seeming success of spray drying as a
technique for particle engineering of biopharmaceu-
ticals, further modifications have been performed on it
to further produce respirable particles with total
retention of biological activity.
Supercritical carbon dioxide (CO2)-assisted spray
drying is a technique where supercritical CO2 is used
to assist the nebulization of the solution [68].
Advantages attributed to this technique include: the
minimal decomposition of thermally labile drugs such
as proteins, the absence of a high pressure vessel as
opposed to normal supercritical fluid technologies, and
the small size of particle produced (below 3 μm) [68].
Two variants of this technique have been reported.
3.1. CO2-assisted nebulization with a bubble dryer™
(CAN–BD)
In this method, the drug is dissolved in water, alcohol
or a mixture of both. The solution produced is mixed with
CO2 in either a sub-critical or supercritical state by
pumping both fluids through a low volume tee to generate
an emulsion [69]. The emulsion produced expands
through a nozzle into the drying chamber, which is held
at atmospheric pressure to generate aerosol microbubbles,
and microdroplets that are dried by a heated nitrogen gas
source.
Sievers et al. have applied this technique to prepare
rhDNase aerosol for nebulization [70]. Buffered
aqueous solution of rhDNase was mixed with SFCO2
in a low dead-volume tee. The mixture formed was
expanded through a restrictor to form aerosol particles
with diameter less than 5 μm. The aerosol formed
could be diluted with air for inhalation or collected for
formulation in other pharmaceutical products such as
pMDI. Following the processing, 95% of the respira-
ble protein was in an active unagglomerated form.
3.2. Supercritical fluid-assisted atomization (SAA)
Reverchon et al. [71] described a process where
supercritical CO2 and a solution of the drug (in water
or organic solvent) are mixed and sprayed into a vessel
at conditions near atmospheric pressure and a flow of
hot nitrogen.
This technique has been used in the processing of
protein powders such as recombinant human deoxy-
ribonuclease (rhDNase), lysozyme, lactate dehydroge-
nase (LDH), ovalbumin, and trypsogen [72–75].
3.3. Nebulization, air drying and electrostatic
collection (NAE)
Dalby et al. have also described a process similar to
spray drying [76]. Fine protein particles were produced
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
by combining nebulization, air drying and electrostatic
collection. Using alkaline phosphatase as a model
protein, a solution of this protein was prepared with
lactose and nebulized in a small-particle aerosol
generator (SPAG-2). This aerosol entered a secondary
drying chamber followed by an electrostatic collector,
which could be operated with a potential difference
ranging from 0 to 20 kV. Particles were recovered from
the precipitator by rinsing with a propellant. Although
this technique produced fine particles, there was some
loss in the activity of the alkaline phosphatase.
4. Spray freeze-drying (SFD)
SFD was first introduced in 1994 [77]. It was
classified as a variant of dry milling. It involves spraying
a solution containing the macromolecule into a vessel
containing a cryogenic liquid such as nitrogen, oxygen or
argon. Since the normal boiling point for such a liquid is
very low, the droplets are quickly frozen. Lyophilizing
these frozen droplets results in porous spherical particles
suitable for inhalation [27]. It has been established that
these droplets may begin to freeze during the time of
flight through the cold vapour phase and then completely
freeze upon contact with cryogenic liquid [78–81].
Although this technique has been used for producing
protein particles, it is faced with the limitations of stresses
associated with freezing and drying, which may cause
irreversible damage to the protein. This is manifested as
structural denaturation, aggregation and loss of biological
activity upon rehydration [82,83]. Similar to spray
drying, loss of stability due to unfolding and aggregation
remains a major challenge.
Webb et al. were able to show that adsorption of
recombinant human interferon-γ (rhIFN-γ) at the air–
liquid interface produced by the atomization resulted
in significant aggregation [81]. This confirms that the
majority of protein activity loss in the SFD process is
due to protein adsorption at the air–liquid interface in
the atomization step and during the freezing of the
droplet in the cold vapour phase rather than during the
lyophilization process.
Further possible limitations to this technique
include the fact that it is time consuming, the nature
of the process has safety issues involved with spraying
into cryogenic fluids and it is expensive. A full SFD
process may take as long as 3 days and the fact that
liquid nitrogen with a boiling point below − 195.8 °C
1017
is used makes this technique hugely impractical.
Combining a spray dryer with a lyophilizer makes
the technique quite expensive.
Using cetrorelix as a model peptide, Zijlstra et al.
[33] were able to produce particles of large surface
area and high porosity which were comparable to those
produced by Costantino et al. [79] and Maa at al. [78]
(as seen in Fig. 3), although degradation of the drug
did not occur, the aerosol performance of the powder
produced in an adhesive mixture was poor. This could
be attributed to the fact that SFD produced very fragile
particles, which cannot withstand the production
process of an adhesive mixture. During the mixing
process, the fragile particles of SFD processed cetro-
relix powder were exposed to impact forces from
colliding lactose particles, which crushed the particles
to smaller fragments. These very small fragments,
once attached to the lactose carrier, are more difficult
to separate again from the carrier during inhalation
because of their small diameter and low density (small
amount of mass per particle).
Maa et al. compared particles produced from spray
and spray freeze drying of rhDNase and anti-IgE
monoclonal antibody (anti-IgEMAb) [79].
Using the same atomization procedure, spray
drying produced small (approximately 3 μm) dense
particles, but SFD resulted in large (approximately 8–
10 μm), porous particles. The FPF of the spray freeze-
dried powder was significantly better than that of the
spray-dried powder, which confirms the concept of
porous particles giving low aerodynamic size.
The enzyme lysozyme has been spray freeze-dried
[84]. Using an aqueous feed containing lysozyme
Fig. 3. SEM showing spray freeze-dried cetrorelix acetate. Reprinted
from Zijlstra et al. [33] with permission from Elsevier.
1018 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
powder in 10 mM phosphate buffer (pH 7.4). The
aqueous feed solution was sprayed through a two-fluid
nozzle. The frozen particles were collected and lyoph-
ilized. It was discovered that the rapid freezing rates
obtained in the SFD process generate significant super-
saturation and thus rapid nucleation rates of dissolved
substances. The rapid nucleation and restricted growth
that occurs after the fine droplets are formed and frozen
leads to extremely small primary particles with a high
surface area following lyophilization [84].
Having confirmed that adsorption of protein at the
air–liquid interface during atomization is mainly
responsible for loss of activity of proteins during spray
drying and SFD, techniques have been developed to
limit the time of exposure to the air–liquid interface
during atomization. Two techniques have been devel-
oped and they include:
• Spray freezing into liquid
• Spray freezing with compressed CO2
4.1. Spray freezing into liquid (SFL)
SFL is a novel particle engineering technique which
was initially developed for particle engineering poorly
water soluble or insoluble drugs by atomizing a feed
liquid containing a poorly water soluble drug and
solubility enhancing excipients below the surface of a
cryogenic liquid to produce rapidly frozen particles
that are subsequently dried [85]. The ultra rapid
freezing prevents phase separation of the drug and
excipient and also prevents crystalline growth in
frozen water. These two factors could result in non-
homogeneous powder aggregates consisting of crys-
talline drug domains. Ultra-rapid freezing rates
produce a porous micronized flowable powder with a
high surface area [85].
This technique has successfully been adapted for
proteins. The liquid solution of protein is sprayed
directly into the liquid nitrogen through an insulated
nozzle, rather than into cold vapour as compared to SFD.
Insulin particles produced by SFL had a low bulk
density, a high surface area, narrow particle size
distribution and no loss in stability as characterized by
absence of covalent dimer [86]. Bovine serum albumin
(BSA) was also studied using this technique [87] and
degradation was lower for SFL than in the case of SFD
reported in an earlier study [79].
4.2. Spray-freezing with compressed CO2
In this process, aqueous solution of the biopharma-
ceutical and excipients is simultaneously atomized and
frozen by dispersing supercritical CO2 in it [88]. The
solution containing the biopharmaceutical is mixed
with compressed CO2 in a static mixer to form droplets
of a CO2-saturated aqueous solution dispersed in CO2.
Joule–Thomson expansion cooling brings about the
rapid freezing of the droplets when injected into the
spray tower through a nozzle [88]. Once the spraying
process is completed, the frozen particles are lyoph-
ilized to obtain dried particles. The main goal of this
method is to obtain stable, porous or hollow protein
particles with a narrow size distribution suitable for
inhalation.
This method has been successfully used in the
processing of trypsinogen [88]. Using different
mathematical models, it was discovered that the flow
rates of the solution and CO2 as well as excipient
concentration could influence the size of solid particles
of protein produced.
5. Supercritical fluid technology (SF)
Supercritical fluids (SFs) are gases and liquids at
temperatures and pressures above their critical points (Tc
– critical temperature; Pc – critical pressure) [89]. The
ranges, 1 b T/Tc b 1.1 and 1 b P/Pc b 2, have been estab-
lished as of particular interest for supercritical fluid
application because in this region, the SF exists as a
single phase with several advantageous properties of
both liquids and gases. They have density values that
enable appreciable solvation power, whilst the viscosity
of solutes in SF is lower than in liquids. The diffusivity of
solutes is higher, which facilitates mass transfer [89].
Most importantly, SFs are highly compressible, partic-
ularly near the critical point and so their densities and
thus solvation power can be controlled by variation in
temperature and pressure [90].
For pharmaceutical applications, the most widely
used SF is CO2 (SFCO2) because of its low critical
temperature (31.2 °C) and pressure (7.4 MPa) [66], it is
non-flammable, non-toxic and inexpensive. The only
limitation facing the use of SFCO2 is its limited
solvation power for compounds of interest although
this can be turned into an advantage in a situation
where the SFCO2 is used as an anti solvent [89].
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
It is possible to modify the solvation power of a
SFCO2 by incorporating a small amount of volatile co-
solvent, such as acetone or ethanol (i.e. organic
modifier) [91]. Although, the critical properties of the
resulting supercritical solutions are changed, the
addition can dramatically improve the loading capacity
or selectivity of the modified SF [89].
Particle engineering by SFs can be divided into
three broad groups (see Fig. 4) [89].
• Precipitation from supercritical solutions composed
of SF and solutions
• Precipitation from gas saturated solutions
• Precipitation from saturated solutions using SF as
anti-solvent.
The first method involves dissolving or solubilizing
the drug in a SF followed by rapid expansion of the SF
solution across a heated orifice to cause a reduction in
the density of the solution, thereby decreasing the
solvation power of the SF which leads to precipitation
of the drug [68]. This process has been christened
‘rapid expansion of supercritical solution’ (RESS)
[89].
Fig. 4. Schematics of various supercritical fluid techniques. Reprinted from Vemavarapu et al. [134] with permission from Elsevier.
1019
The second method is a relatively harsh technique
but it is similar to the first one because they both
involve using SFCO2 as a solvent rather than an anti
solvent. This has been named ‘precipitation from gas-
saturated solution’ (PGSS) [92]. In this process the SF
is dissolved in molten solute and the resulting
supercritical solution fed via an orifice into a chamber
to allow a rapid expansion under ambient conditions
[93].
The third method is the most common method and
different variants of the technique have been devel-
oped. This technique utilizes a similar concept to the
use of anti-solvent in solvent-based crystallization
processes [89]. The high solubility of SFCO2 in
organic solvents leads to volume expansion when the
fluids make contact. This leads to reduction in solvent
density and parallel fall in solvation capacity. Such
reduction causes increased level of supersaturation,
solute nucleation and particle formation [25].
Acronyms such as GAS (gaseous anti-solvent),
ASES (aerosol solvent extraction system), SEDS
(solution enhanced dispersion by SF), and PCA
(precipitation by compressed anti-solvent) have been
assigned to different applications of this technique.
1020 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
5.1. RESS process
Krukonis first applied SFs to produce particles of
drugs with a narrow size distribution [94]. Sufficient
solubility of the material to be processed in the SF is the
major prerequisite for the use of RESS process. The
solubility of a drug powder in SF will depend on the
density of the SF, the drug's chemical structure and the
SF-drug contact time [68]. The morphology and size
distribution of the particles formed are dependent on the
pre-expansion concentration of the solute in SF and the
expansion conditions (e.g. temperature, pressure) [95].
The higher the pre-expansion concentration, the narrower
the particle size distribution [68]. The expansion
conditions depend on the nozzle temperature, geometry
and size [68]. The nozzle has to be maintained at a
suitable pre-expansion temperature to prevent the
premature precipitation of the drug.
The limitations facing the use of RESS for biophar-
maceutical include: high temperature needed for the rapid
expansion (typical temperature of 313 K) which can
destroy proteins, poor predictive control of particle size
and morphology [25]. Scale up of this technique is limited
by particle aggregation and nozzle blockage caused
expansion cooling [25]. It is also limited to molecules that
are soluble in SFCO2.
Despite these limitations RESS has been studied for
the particle engineering of proteins by some authors.
Cyclosporin A, an immunosuppressant drug that has
been used for the treatment of respiratory tract infections,
has been processed by RESS [93]. Cyclosporine was
dissolved in a near critical or SFCO2 and the resulting
solution was depressurized rapidly by spraying the solu-
tion at atmospheric conditions through a 50 μm nozzle
placed inside the expansion chamber. Cyclosporine
particles precipitated using this technique were spherical
with the primary particle size of 150 nm that were highly
aggregated. The effect of the pre-expansion temperature
and pressure on the characteristics of the precipitates was
also investigated. The morphology and particle size of the
precipitates were not significantly affected, however, the
amount of cyclosporine precipitation was observed to
increase as the system pressure was increased.
Lysozyme and lipase have been microencapsulated in
polymers by modifying RESS in a technique called ‘rapid
expansion of supercritical solution with a non-solvent
(RESS-N) [96]. The polymers studied included PEG,
PLA, PLGA and PEG–poly(propylene glycol)–PEG
triblock co-polymer. The objective of the study was to
develop a micro-encapsulation process whereby the
particle size distribution could be controlled whilst
avoiding the use of a toxic solvent. Ethanol was used as
a co-solvent. The polymer and co-solvent were charged
into a pressure vessel. SFCO2 was added to the vessel via
a vessel containing the protein of interest. The resultant
mixture was stirred for about 4 h and kept for 1 h without
agitation to ensure the polymer was dissolved and the
protein suspended. The resultant solution/suspension was
sprayed through a capillary nozzle held at 313.15 K. The
yield of the processed microparticles was found to be
dependent on the solubility of the polymer in the SF/co-
solvent. It was observed that ethanol used as the co-
solvent in this study produced particles that were less
adhesive. This is thought to be due to the low solubility of
high molecular weight PEGs in absolute ethanol.
5.2. PGSS
In the PGSS process, the viscosity of the molten
compound can be decreased due to the dissolved gas [93].
The gas saturated liquid phase is expanded to generate
particles from materials that are not necessarily soluble in
SF and so avoid the limitations of the RESS process.
Compounds can also absorb some of the CO2 and con-
sequently swell (polymers) or melt at temperature signi-
ficantly lower than the normal melting or glass transition
temperature.
Cyclosporine has been processed using this technique
[93]. Melting point depression study was performed on
the cyclosporine by loading 10 mg of cyclosporine in a
glass tube (i.d. = 5.8 mm) and placed inside a view cell in
temperature-controlled bath. CO2 was gradually intro-
duced into the view cell at 5 bar increments. The system
was isolated and equilibrated for at least 30 min at each
pressure interval. The melting point of cyclosporine was
depressed to 25, 35, 40, 45, and 50 °C from 150 °C
(normal melting point of cyclosporine) when exposed to
CO2 at 53, 58, 60, 65 and 77 bar, respectively.
The PGSS was performed by dissolving compressed
CO2 in cyclosporine, an equilibrium cell containing the
cyclosporine was heated up to the required temperature
and allowed to equilibrate. The system was then pres-
surized with CO2 to melt the cyclosporine. The resultant
solution was then expanded through a nozzle and the CO2
evaporated in the expansion chamber; cyclosporine
particles were then collected for analysis. At all
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
conditions of micronization ranging from 160 to 200 bar
and 25 to 45 °C, cyclosporine was precipitated as
microspheres as small as 150 nm and was highly
aggregated as observed by SEM. The average particle
size of the precipitates was not significantly affected as the
pre-expansion pressure increased from 180 to 200 bar.
The effect of nozzle diameter on the micronization of
cyclosporine was also assessed. With a 100-μm nozzle,
the primary particle size increased from 150 nm to 1 μm.
This is believed to be due to a greater reduction in pressure
and density of the fluid at the exit of the nozzle [97]. It was
observed that the particles formed with the 100-μm
nozzle became porous upon the release of CO2 from the
cyclosporine-saturated phase during depressurization.
However, the particle size distribution measured by
coulter and Andersen Cascade Impactor (ACI) was higher
with median diameter of 4.5 μm and FPF of 59%
(b 4.7 μm), respectively.
5.3. GAS
In this process, the SF is added to the particle forma-
tion vessel that contains the protein solution of interest
[98]. The protein precipitates during the dissolution of the
SF in the solvent. The SF is generally introduced through
the bottom of the vessel and bubbled through the protein
solution to achieve better mixing of the solvent and anti-
solvent [68]. Once the protein has precipitated out, the
solvent–SF can be removed. The protein particles are
then washed with subsequent SF washes. Finally, the
pressure in the vessel is released and the protein powder
recovered.
The solvent in which the protein is dissolved should
have a high solvent power for the protein, preferably be
soluble in the SF of choice and be compatible with the
protein. Dimethyl sulfoxide (DMSO) and N,N-dimethyl
formamide (DMFA) have been used since they satisfy
the criteria of solvation of the protein and solubility in
SFCO2. However, there are concerns over the toxicity
and compatibility with protein molecules [99,100].
Alternatives to SFCO2 have also been studied.
Supercritical ammonia and ethane have been used,
while ammonia produced completely denatured pro-
teins [102], ethane produced results comparable to
SFCO2 in terms of particle size and improved
biological activity of insulin [102]. Whilst the perfor-
mance of SF-ethane may be promising, its high
flammability will limit its use.
1021
GAS has been used to process biomolecules such as
lysozyme, insulin and myoglobin from organic
solvents [101–104]. The proteins were dissolved in
DMSO, MeOH, EtOH or DMF and the effect of the
solvents on particle size, morphology and biological
activity studied. For insulin, methanol produced the
smallest particles (0.05–0.45 μm) while DMSO
produced the largest particles (1.4–8 μm). DMSO
was used as the main solvent for lysozyme and the
effect of operating temperature studied. The highest
activity retention was observed at 18–35 °C (92–
100%) while the lowest activity retention was
observed at 45 °C (60%). The particle size range was
similar for all temperatures (0.05–0.3 μm).
The principal disadvantage of this process is the
lack of control of particle formation. This has been
observed to be true in batch operating conditions
because the level of saturation is not maintained [98].
5.4. ASES
This process involves pumping CO2 into a high-
pressure vessel until the system reaches the desired
fixed conditions (pressure and temperature) [68]. The
protein solution is sprayed via an atomization device
into the vessel containing SFCO2. Particles are then
collected on a filter at the bottom of the vessel [98]. In
this process, the anti-solvent concept of GAS remains
but happens at the droplet level. This offers a
favourable higher anti-solvent to solvent ratio, an
increased surface area and mass transfer rate and hence,
an acceleration of the drying process. The rate is
particularly fast when the operating pressure reaches
the mixture critical pressure. The process is then
controlled by mixing of miscible fluids rather than
mass transfer over interface of the droplets in the spray
[98]. When the miscibility of the fluids is poor,
especially in systems containing water and CO2, the
mass transfer requires special attention. The mass
transfer can be improved by increasing the drying
medium to solvent ratio, decreasing the droplet size or
the relative velocity between the solvent and drying
medium [98]. This process enables production of
protein particles with narrow size distribution, uniform
shape and desired physico-chemical characteristics
[105,106].
ASES has been applied in producing insulin loaded
PEG/PLA nanoparticles [107]. PLA and varying levels
1022 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
of 1900 Da PEG were used to prepare nanoparticles
from DMSO. The organic solution of PEG, PLA and
bovine insulin was atomized through a 50-μm nozzle
at a rate of 1 ml/min into a vessel containing the
SFCO2. SEM of the particles indicated that lower PEG
compositions produced well-separated nanoparticles.
In contrast higher PEG compositions formed a
continuous network structure. Light scattering analysis
showed that particles below 1 μm were obtained under
the adopted preparation condition. However, it was
observed that the addition of high amounts of low MW
PEG was essential to preparing products with high
drug loading, high protein stability, effective drug
release and solvent free product [107].
5.5. SEDS
In the earlier discussed ASES process, the mass
transfer of the SF into the sprayed droplet determines
the rate of particle formation, while particle agglom-
eration and aggregation phenomena are influenced by
the rate of solvent mass transfer into the SF from the
droplet [89]. Mass transfer of SF into droplet is
dependent upon atomization efficiency while mass
transfer of solvent into SF is dependent on dispersing
and mixing phenomena between the solution droplet
and the SF [107]. To minimize particle agglomeration
often observed in ASES and other anti-solvent-based
techniques, and to reduce drying times, increased mass
transfer rates are required [89]. This has been
successfully achieved by the SEDS process. In this
process, the drug solution and the SF are introduced
into the particle formation vessel (where temperature
and pressure are controlled) through a co-axial nozzle
with a mixing chamber [68]. The high velocity of the
SF allows the production of droplets of very small
sizes, while the mixing of solvent with the SF inside
the mixing chamber leads to an increase of mass
transfer of SF into the solvent and vice versa [68]. A
high mass transfer allows a faster nucleation and a
smaller particle size with less agglomeration [107].
SEDS has been further optimized to process water-
soluble compounds such as peptides and proteins by
introducing an organic solvent, the SCF, and the
aqueous solution as separate streams into a co-axial
three-component nozzle. This helps in overcoming
problems associated with limited solubility of water in
SFCO2. SEDS is a more controllable and reproducible
technique compared to other anti-solvent-based SCF
process. It is suitable for scaling up and manufacturing
according to GMP requirements [25]. Furthermore, the
use of destructive organic solvents such as DMSO can
be avoided.
Lysozyme has been processed by SEDS with a
micro-fine product showing a spherical morphology
with free flowing powder handling properties [107].
When the retention of enzymatic activity of SEDS-
processed lysozyme was compared to those processed
by spray drying and SFD, the SEDS-processed
enzyme retained the highest enzymatic activity [107].
Supercoiled plasma DNA (pDNA) has been pro-
cessed by SEDS from aqueous solutions [108].
Increased yields pDNAwere obtained when the aqueous
feed containing mannitol was buffered to counteract the
acidic nature of the SFCO2 [107]. This indicates that by
using a buffered aqueous solution, the SEDS process
can be used to control the particle formation and protect
labile biological materials during processing.
Okamoto et al. [108] recently studied the stability of
a chitosan–pDNA complex powder prepared by a
SEDS-type process for pulmonary delivery. An
aqueous chitosan–pDNA complex solution containing
mannitol was injected into the stream of a SFCO2/
ethanol admixture to precipitate a gene powder. The
obtained gene powder and gene solution (0.06 μg
pDNA and 0.33% sodium dodecylsulfate aqueous
solution) were placed in stability chambers at 25 or
40 °C for 4 weeks. The integrity and transfection
potency of the gene were examined by electrophoresis
and in vivo pulmonary transfection study in mice. The
process decreased the supercoiled DNA during the
manufacturing; however, the decrease in the remaining
supercoiled and open circular DNA in the powders
during storage was much slower than in solutions. The
powders had higher transfection potency than the
solutions containing the same amount of DNA. The
effect of chitosan on the stability of DNA in solution
was not obvious but it improved the stability of DNA
in powders during manufacturing and storage.
5.6. PCA
The PCA is a slight modification of the ASES
process. The process has been reported to be a one-step
technique for effectively producing solvent free
particles with a narrow size distribution at mild
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
operating conditions [109,110]. It involves feeding
SFCO2 into a precipitator so as to pressurize the
precipitator to a desired value. Then the CO2 flow rate
into the precipitator is fixed. The solution of protein is
fed into the precipitator through a nozzle. The protein
particles are precipitated due to the anti-solvent effect
of the SFCO2 [111].
Fine protein particles of catalase and insulin have
been prepared using the PCA process [112,113].
Insulin was dissolved in DMSO or DMF and these
solutions were sprayed into a crystallizer with a
continuous feed of SFCO2 [113]. The insulin powder
formed had 90% particles with diameter less than 4 μm
and 10% less than 1 μm. it was also observed that the
processed insulin maintained its biological activity.
Protein-loaded N-trimethyl chitosan (TMC) micro-
spheres have been prepared by the PCA process [113].
It had been observed that using the PCA process to
precipitate proteins and polar polymers from DMSO
using CO2 as solvent has some difficulties [114].
When operating below the mixture critical pressure of
the DMSO–CO2, solution droplets formed by atom-
ization in the nozzle dry slowly due to the low
solubility of DMSO in the surrounding CO2 gas phase
[113]. A wet product or a film over the filter plate is
often obtained when operating at these conditions
[114]. On the other hand, proteins and polar polymers
have an extremely low solubility in CO2 so when
working at conditions above the critical pressure of the
mixture, high supersaturations are induced as soon as
the proteins or polymers come into contact with the
CO2, in which case agglomerated nanoparticles are
formed [113]. It has been suggested that the addition of
a third component to the system solvent–CO2 modifies
its phase behaviour [115,116]. Perez de Diego et al.
used water to modify the phase behaviour of the
system DMSO–CO2 and solved the limitation of the
PCA at conditions below the mixture critical pressure
of the system solvent–CO2. This approach made it
possible to develop a process to produce protein-
loaded TMC particles suitable for inhalation. SFCO2
was fed using into the precipitator to the desired vol-
ume and pressure. The CO2 flow rate was then fixed at
20 kg/h and the pressure in the precipitator automat-
ically regulated. DMSO was filtered and fed into the
precipitator using a Gilson piston pump to avoid the
blockage of the nozzle at the beginning of the injection
of the aqueous DMSO solution of protein (lysozyme)
1023
and polymer. To start precipitation, the solvent flow
was stopped and the protein/polymer DMSO–water
solution fed through the nozzle at the same flow rate as
the solvent. The mixture solvent–CO2 continuously
left the precipitator and the precipitated protein-loaded
particles collected on a 0.2 μm membrane filter. Once
the entire solution had been injected, the SFCO2
continued to flow through the vessel to wash out the
solvent. It was observed that working at 80 bar
(pressure in precipitation), a wet product was always
observed over the filter plate so it was necessary to
increase the pressure to 90 bar to obtain a dry product.
As pressure increased, the extraction of the liquid
solvent to the supercritical phase takes place faster due
to a higher solubility of the solvent in the CO2. As
pressure increased, the atomization of the solution
produced smaller droplets and the drying time de-
creased. The influence of water content of the DMSO–
water solution was also studied and it was observed
that, as the water content of the solution increased, the
morphology changed from nanoparticles to a mixture
of nanoparticles and microspheres. Water concentra-
tions in the DMSO solution above 3.5% w/w produced
a wet product. The presence of water enhances the
mass transfer of SFCO2 through the droplet interface
due to the high solubility of the solution in the SFCO2.
However, there seems to be an optimal water content
for the DMSO solution as above a certain water
concentration, solubility of the solvent mixture in the
vapour phase decreases and the drying is no longer
efficient leading to a wet product [111].
6. Microcrystallization
Crystalline protein particles have been found to be
more active and stable than their amorphous counter-
parts [117,118]. The higher stability arises from the
fact that, unlike crystalline phases, amorphous materi-
als consist of disordered arrangement of molecules and
therefore possess no distinguishable crystal lattice
[68]. Thermodynamically, the absence of crystallinity
causes energy content higher than the crystalline state,
leading to lower stability and higher reactivity [68].
Amorphous protein particles are cleared rapidly from
systemic circulation and are more susceptible to
hydrolytic and enzymatic degradation because of
their higher reactivity [119]. These factors make
crystalline protein desirable as a fine pharmaceutical
1024 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
ingredient. Crystallization of proteins involves a one-
step process and results in high purity. Crystallization
can improve protein handling during processing,
storage and delivery. It can also offer sustained release
of the therapeutic agent for an effective duration by
changing the dissolution characteristics [120].
Despite these advantages, apart from insulin, few
crystalline forms of proteins have been used as
pharmaceutical active ingredient, even though over
90% of pharmaceutical products that contain a drug in
particulate form, do so as a crystalline form [121]. This is
because proteins being large and flexible are difficult to
crystallize [119]. Furthermore, crystallization can lead to
particles with a wide size distribution. Milling has been
applied to size reduce crystallized protein to respirable
size, but milling may lead to a high energy input leading
to particles of reduced crystallinity and may contain
disordered regions, thereby introducing regions of
reduced stability [122].
The concept of microcrystallization has been used
to overcome milling-induced disorder in crystalline
powders.
Lee et al. [119] have succeeded in producing
microcrystals of α-lactalbumin, a 16 kDa glycoprotein.
α-Lactalbumin was dissolved in 0.1 N acetic acid
containing PEG-8000 as a stabilizer. The pH of the
protein solution was adjusted to about 3.0. 10 N, NaOH
solution was added rapidly to the protein solution to
adjust pH to about 4.0 (at this pH, the protein has lowest
solubility). The supersaturated protein solution was
stored at 16 °C. The microcrystals formed were roughly
spherical and about 1–2 μm in diameter. These spherical
crystals have been applied for pulmonary delivery [123].
6.1. Microcrystallization and sustained release
The challenge presently confronting pulmonary
delivery of protein is the short duration of action obtained
via this route due to fast rate of clearance via the
mucocilliary apparatus and alveolar macrophages
[124,125]. Different methods of prolonging insulin
absorption in the lung have been proposed such as
microencapsulation using a biodegradable polymer, but
problem is perceived with the accumulation of these
polymers in the lung [126] and loss of insulin activity
during the preparation of microspheres [127].
Recently, a unique insulin microcrystallization process
using a seed zone method was developed [123]. Insulin
microcrystals with a mean diameter of 3 μm were
prepared using this seed zone method. SEM micrograph
showed the microcrystals to be of a homogenous
rhombohedral shape with some rhombus forms, without
aggregates. Following the administration of 32 U/kg of
the microcrystal suspension to streptozotocin-induced
diabetic rats by intratracheal instillation, the blood glucose
levels were reduced and hypoglycemia was prolonged
over 13 h as compared to the normal insulin solution
[123]. It was suggested that the sustained release effect
was due to the decreased solubility of the microcrystals.
6.2. Polymorphism
Polymorphism is the propensity of a substance to
crystallize into more than one crystal structure [68]. This
is a potential problem for the crystallization of
biopharmaceuticals for pulmonary delivery as regulatory
bodies only approve a specific crystal structure or
polymorph [128]. During crystallization, different poly-
morphs of a compound may be formed [68]. The more
stable polymorph has the higher molecular packing
density and has the lower values for: Gibb's free energy,
fugacity, vapour pressure, thermodynamic activity,
dissolution rate per unit area in any solvent and rate of
reaction, including decomposition rate [68]. For thermo-
dynamic reasons, the less stable polymorph would want
to convert to the stable one. The phase transition can
occur through for different mechanisms: solid–solid
transition, melting, solution and solution mediated [129].
The kinetics of the phase transitions can be influenced by
typical environmental parameters (temperature, pressure,
relative humidity), presence of crystalline defect, impu-
rities and mechanical stress [68].
Polymorph purity is therefore an important parameter
to consider in a drug product since the presence of
differing crystal phases can accelerate the conversion
process by lowering the relevant activation energy
barrier [130]. The phase transformation could lead to a
different polymorph with unwanted physico-chemical
properties. For this reason, regulatory authorities have
long recognized the need to limit polymorphic impuri-
ties in pharmaceutical materials [130].
6.3. Controlling polymorphism with SF
The identification of stable polymorphs with desired
physical properties is very important for product
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
development. Before a drug is submitted for review to the
regulatory authorities, it is important to identify the most
stable or suitably stable polymorph of such a drug and
work with it. The solid phase must be monitored, es-
pecially after processes that are potentially able to modify
the solid-state properties of the components [68]. For
instance, during micronization processes such as milling,
spray drying and spray freeze drying, the substance is
usually exposed to mechanical stress, contact with sol-
vents, heating–cooling cycles that can often lead to
alteration of the solid phase, such as new polymorph
formation, dehydration or melting mechanism [68].
The SEDS process has proved to be an effective
technique in obtaining pure polymorphs of different
drugs based on different operating conditions and
crystallization kinetics [131]. The rapid drying and
cooling experienced in SEDS exclusively gives rise to
a polymorph of a drug [68].
7. Aerodynamic particle diameter
Aerodynamic diameter (dae) particle size is a key
parameter determining the aerosol performance of an
inhaled powder [49]. Unlike the physical particle
diameter, dae is a concept incorporating the size, shape,
and density of particle [27]. For a solid spherical particle,
dae is simply equal to the physical diameter (d) multiplied
by the square root of the particle density (ρ) [49].
pffiffiffi
dae ¼ ð q Þd ð1Þ
This means for a given density, small particles have
lower aerodynamic diameter compared to larger parti-
cles. However porous particles because of their low
density would have low dae based on Eq. (1).
Nektar Therapeutics has adopted this porous
particle approach in the formation of their pulmo-
sphere® [132]. The physical diameter (d) is mostly
measured using laser diffraction technique [133].
Characterizing the dae for non-spherical particles such
as elongated or needle-like object is quite challenging as
Eq. (1) cannot be used. It has been discovered that the dae
of elongated particles is controlled by the short axis rather
than the long axis [47]. It was found that dae increases
with the aspect ratio (i.e. ratio of the long axis to the short
axis) of an elongated particle. When the aspect ratio is
greater than 15 or 20, dae will be about two or three times
the short axis (physical diameter). Any further increase in
1025
the aspect ratio will only increase the dae very small
amounts [49]. This means that dae can be made small as
long as the short axis remains small (e.g. ≤1 μm) [49].
8. Conclusion
The emergence of advanced particle engineering
techniques coupled with the modification of the
traditional methods such as milling has contributed
to the increased possibility of formulating biopharma-
ceutical for pulmonary delivery. Particles with aerosol
properties suited for deep lung delivery have been
engineered without destroying the biological activity
of these sensitive molecules. The wide range of
techniques currently available or being developed
coupled with the increasing knowledge of excipient
use for the protection of biopharmaceuticals now
allows a diverse range of biopharmaceuticals to be
processed for use in inhalation delivery devices. The
challenge to its development is that scientist may now
shift from “can the molecule be processed” to “which
process should be used” since it is yet to be shown if
any one technique can provide the solution to all types
of molecule.
References
[1] G. Walsh, Biopharmaceuticals and biotechnology medicines: an
issue of nomenclature, Eur. J. Pharm. Sci. 15 (2002) 135–138.
[2] G. Walsh, Second generation biopharmaceuticals, Eur. J.
Pharm. Biopharm. 58 (2004) 185–196.
[3] I. Johnson, Human insulin from recombinant DNA technol-
ogy, Science 291 (2004) 632–637.
[4] J. Fielder, N. Wagner, S. Grammaticos, H. Hoffman, H.
Kaufmann, R. Otto, Use of flow-cytometric analysis to optimize
cell banking strategies for production of biopharmaceuticals from
mammalian cells, J. Biotechnol. 120 (2005) 111–120.
[5] G. Walsh, Biopharmaceuticals: recent approvals and likely
directions, Trends Biotechnol. 23 (2005) 553–558.
[6] Y. Bagger, L.B. Tanko, P. Alexadersen, M.A. Karsdal, M.
Olson, L. Mindeholm, M. Azria, C. Christiansen, Oral salmon
calcitonin induced suppression of urinary collagen type II
degradation in post menopausal women: a new potential
treatment of osteoarthritis, Bone 33 (2005) 425–430.
[7] Y.J. Zheng, M. Fulu, Decrease of genital organ weights and
plasma testosterone in rat following oral administration of
leuprolide microemulsion, Int. J. Pharm. 307 (2006) 209–215.
[8] C. Cullander, R.H. Guy, Routes of delivery: case studies (6),
transdermal delivery of peptides and proteins, Adv. Drug
Deliv. Rev. 8 (1992) 291–329.
[9] P. Edman, E. Bjork, Routes of delivery: case studies (1), nasal
deliveryofpeptidedrugs,Adv.DrugDeliv.Rev. 8(1992)165–177.
1026 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
[10] D. Harris, J.H. Liaw, J.R. Robinson, Routes of delivery: case
studies (7) ocular delivery of peptide and protein drugs, Adv.
Drug Deliv. Rev. 8 (1992) 331–339.
[11] N.F.N. Ho, C.L. Barsutin, P.S. Burton, H.P. Merckle, Routes
of delivery: case studies (3), mechanistic insights to buccal
delivery of proteinaceous substance, Adv. Drud Deliv. Rev.
8 (1992) 197–235.
[12] P.L. Smith, D.A. Wall, C.H. Gochoco, G. Wilson, Routes of
delivery: case studies (5), oral absorption of peptides and
proteins, Adv. Drug Deliv. Rev. 8 (1992) 253–290.
[13] V. Codrons, F. Vanderbist, R.K. Vanbeeck, M. Arras, D.
Lison, V. Preat, R. Vanbever, Systemic delivery of parathyroid
hormone (1–34) using inhalation dry powder in rats, J. Pharm.
Sci. 92 (2003) 938–948.
[14] M.A. Hollinger, Respiratory pharmacology and toxicology,
W. B., Saunders, Philadelphia, 1994, pp. 1–20.
[15] P.R. Byron, Determinants of drug and polypeptide bioavail-
ability from aerosols delivered to the lung, Adv. Drug Deliv.
Rev. 5 (1990) 107–132.
[16] T.S. Purewal, Formulation of metered dose inhalers, in: T.S.
Purewal, D.J.W. Grant (Eds.), Metered Dose Inhaler Tech-
nology, Interpharm, Illinois, 1998, pp. 9–68.
[17] U.B. Kompella, V.H. Lee, Delivery systems for the
penetration enhancement of peptide and protein drugs: design
consideration, Adv. Drug Deliv. Rev. 46 (2001) 211–245.
[18] S.P. Sorokin, A morphological and cytochemical study of the
great alveolar cell, J. Histochem. Cytochem. 14 (1961) 884–897.
[19] L. Garcia-Conteras, L.D. Jones, A.J. Hickey, Alveolar
macrophages as a target for drug delivery, Res. Drug Deliv.
IX (2004) (California).
[20] J. Liu, H. Wong, J. Moselly, B. Bowen, X.Y. Wu, M.R.
Johnston, Targeting colloidal particulates to thoracic lymph
nodes, Lung Cancer 51 (2006) 377–386.
[21] K. Corkery, Inhalable drugs for systemic therapy, Respir. Care
45 (2000) 831–835.
[22] A. Hussain, J.J. Arnold, M.A. Khan, F. Ahsan, Absorption
enhancers in pulmonary delivery, J. Control. Release 94
(2004) 15–24.
[23] M. Hanna, P. York. Method and apparatus for the formation of
particles. European Patent 0706421B1, 1998.
[24] E.A.Quinn, R.T.Forbes,A.C.Williams,M.J.Oliver, L.McKenzie,
T. Purewal, Protein conformational stability in the hydro-
fluoroalkane propellants tetraflouroethane and heptafluoropro-
pane analysed by Fourier transform Raman spectroscopy, Int.
J. Pharm. 186 (1999) 31–41.
[25] K.A. Johnson, Preparation of peptide and protein powders for
inhalation, Adv. Drug Deliv. Rev. 26 (1997) 3–15.
[26] E. Phillips, E. Allsopp, T. Christensen, M. Fitzgerald, L. Zhao,
Size reduction of peptides and proteins by jet milling, in: R.N.
Dalby, P.R. Byron, S.J. Farr (Eds.), Proceeding of RDD, vol. VI,
Interpharm, Buffalo, 1998, pp. 161–168.
[27] R.C. Willis, Crystallize to deliver, Mod. Drug Discov. (2003)
12 (March).
[28] G.W. Hallworth, The formulation and evaluation of pressur-
ized metered dose inhalers, in: D. Ganderton, T.M. Jones
(Eds.), Drug Delivery to the Respiratory Tract, Ellis Hor-
wood, Chistester, 1987, pp. 87–118.
[29] R.H. Snow, B.H. Kaye, C.E. Capes, G.C. Sresty, Size
reduction and size enlargement, in: R.H. Perry, D. Green
(Eds.), Perry's Chemical Engineer's Handbook, McGraw
Hill, New York, 1984, pp. 1–59.
[30] L.H. Van Vlack, Elements of Materials Science and Engineer-
ing, Addison Wesley Publishing, Reading, 1980, pp. 185–208.
[31] A. Adjei, Pulmonary bioavailability of peptide drugs: animal
models, preclinical studies and clinical studies, Presented at
Respiratory Drug Delivery II, Keystone, Co, 1990.
[32] W. Schlocker, S. GschlieBer, A. Bernkop-Schniirch, Evalu-
ation of the potential of air jet milling of solid protein–poly
(acrylate) complexes for microparticle preparation, Euro
J. Pharm. Biopharm. 62 (2006) 260–266.
[33] G.S. Zijlstra, W.L.J. Hinrichs, A.H. deBoer, H.W. Frijlink,
The role of particle engineering in relation to formulation and
de-agglomeration principle in the development of a dry
powder formulation for inhalation of cetrorelix, Eur. J. Pharm.
Sci 23 (2004) 139–149.
[34] A. Adjei, D. Sundberg, J. Miller, A. Chin, Bioavailability
of leuprolide acetate following nasal and inhalation deli-
very to the rats and healthy humans, Pharm. Res. 9 (1992)
244–249.
[35] K.G. Backstrom, C.M. Dahlback, P. Edman, A.C.B. Johans-
son, Therapeutic preparation for inhalation. International
Patent Application (1995) WO 95/00127.
[36] K. Tomihisa, Grinding of rubber below the brittle temperature
with supersonic jet air. Japanese Patent (1973) JP73103614.
[37] S. Kobayashi, S. Kondo, K. Juni, Pulmonary delivery of
salmon calcitonin dry powders containing absorption
enhances in rats, Pharm. Res. (1996) 80–83.
[38] W.L.J. Hinrichs, N.N. De Smelt, J. Demeester, H.W. Frijlink,
Insulin is a promising cryo- and lyoprotectant for PEGylated
lipoplexes, J. Control. Release 103 (2005) 465–479.
[39] S. Ohtake, C. Schebor, S.P. Palecek, J.J. de Pablo, Phase
behaviour of freeze-dried phospholipid–cholesterol mixtures
stabilized with trehalose, Biochim. Biophys. Acta, Biomembr.
1713 (2005) 57–64.
[40] R.M. Platz, J. Utsum, Y. Satoh, N. Naruse, Pharmaceutical
aerosol formulation of solid polypeptide microparticles and
method for the preparation thereof. International Patent
Application (1991) WO 91/16038.
[41] R.M. Platz, A. Ip, C.L. Whitham, Process for preparing
micronized polypeptide drugs. US Patent (1994) 5354562.
[42] A.L. Adjei, E.S. Johnson, J.W. Kesterson, LHRH analog
formulations, US Patent (1990) 4897256.
[43] R. Lizio, M. Damm, A.W. Sarlikiotis, H.H. Bauer, C.M. Lehr,
Low temperature micronization of a peptide drug in a fluid
propellant: case study cetrorelix, AAPS PharmSciTech 2
(2001) 1–7.
[44] M. Irngartinger, V. Camuglia, M. Damm, J. Goede, J. Frijlink,
Pulmonary delivery of therapeutic peptides via dry powder
inhalation: effects of micronisation and manufacturing, Eur. J.
Pharm. Biopharm. 58 (2004) 7–14.
[45] K. Masters, Spray Drying Handbook, Longman Scientific and
Technical, Essex, England, 1991.
[46] J. Broadhead, S.K.F. Rouan, I. Hau, C.T. Rhodes, The effect
of process and formulation variable on the properties of spray
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
dried β-galactosidase, J. Pharm. Pharmacol. 46 (1994)
458–467.
[47] P. Labrude, M. Rasolomanana, C. Vigneron, C. Thiron, B.
Chailott, Protective effect of sucrose on spray drying of
oxyhemoglobin, J. Pharm. Sci. 78 (1989) 223–229.
[48] H. Brandenberger, Best@buchi evaporation, Inf. Bull. 27
(2003).
[49] H.K. Chan, Dry powder aerosol drug delivery – opportunities
for colloid and surface scientists, Colloids and Surfaces A:
Physicochem. Eng. Aspects 284–285 (2006) 50–55.
[50] B. Bittner, T. Kissel, Ultra atomization for spray drying: a versatile
technique for the preparation of protein loaded biodegradable
microspheres, J. Microencapsul 16 (1999) 325–341.
[51] Y.F. Maa, P.A. Nguyen, J.D. Andya, T.D. Sweeney, S.J. Shire,
C.C. Hsu, Effects of spray drying and subsequent processing on
residual moisture content and physical/biochemical stability of
protein inhalation powders, Pharm. Res. 15 (1998) 768–775.
[52] J.S. Patton, L.C. Foster, R.M. Platz, Methods and composition
for pulmonary delivery of insulin. International patent
application, WO95/ 24183.
[53] R.W. Niven, F.D. Lott, A.Y. Ip, J.M. Cribbs, Pulmonary
delivery of powders and solutions containing recombinant
human granulocyte colony-stimulating factor, Pharm. Res 11
(1994) 1101–1109.
[54] A.D. Sutton, R.A. Johnson, P.J. Senior, D. Heath, Preparation
of diagnostic agents. European patent application (2000)
0681843A2.
[55] S. Willaya-Areekul, G.F. Needham, N. Milton, M.L. Roy, S.L.
Nail, Freeze-drying of tert-butanol/water co-solvent system: a
case report on formation of a friable freeze dried powder of
tobramycin sulphate, J. Pharm. Sci. 91 (2002) 1147–1155.
[56] H. Chan, D. Lechuga-Ballesteros, R. Vehring, W.R. Foss, J.C.
Feeley, A. Clark, Effect of bulk and surface properties on the
aerosol performance of dry powders for inhalation, Presented
at 2002 AAPS, Annual Meeting and Exposition, Nov. 10–14,
Canada, Toronto, Ontario, 2002.
[57] H.K. Chan, A. Clark, I. Gonda, M. Mumenthaler, C. Hsu,
Spray dried powders and powder blends of recombinant
human deoxyribonuclease (rhDNase) for aerosol delivery,
Pharm. Res. 14 (1997) 431–437.
[58] Y.–H. Liao, M.B. Brown, S.A. Jones, T. Nazir, G.P. Martin,
The effect of polyvinyl alcohol on the in-vitro stability and
delivery of spray-dried protein particles from surfactant-free
HFA 134a-based pressurised metered dose inhalers, Int. J.
Pharm. 304 (2005) 29–39.
[59] V. Codrons, F. Vanderbist, R. Verbeeck, A. Mohammed, D.
Lison, V. Preat, R. Vanbever, Systemic delivery of parathyroid
hormone (1–34) using inhalation dry powder in rats, J. Pharm.
Sci. 92 (2003) 938–950.
[60] J.D. Andya, Y.F. Maa, H.R. Costantino, P.A. Nguyen, N.
Dasovich, T.D. Sweeney, C.C. Hsu, S.J. Shire, The effect of
formulation excipients on protein stability and aerosol per-
formance of spray dried powders of a recombinant humanized
anti-IgE monoclonal antibody, Pharm. Res. 16 (3) 350–358.
[61] C. Bosquillon, V. Preat, R. Vanbever, Pulmonary delivery of
growth hormone using dry powders and visualization of its
local fate in rats, J. Control. Release 96 (2004) 233–244.
1027
[62] M. Adler, M. Unger, G. Lee, Surface composition of spray
dried particles of bovine serum albumin/trehalose/surface,
Pharm. Res. 17 (2000) 863–870.
[63] C.L. Leach, J.S. Patton, K.M. Perkins, M.C. Kwo, B. Bueche,
L. Guo, M.D. Bentley, Dry powder formulation of inhaled
PEG-Insulin yields prolonged systemic activity in dogs,
Presented at the American Diabetes Association 63rd
Scientific Session, June, Louisiana, New Orleans, 2003.
[64] J. Elversson, A. Millqvist-Fureby, Aqueous two-phase system
as a formulation concept for spray-dried protein, Int. J. Pharm.
294 (2005) 73–87.
[65] L. Dellamary, D.J. Smith, A. Bloom, S. Bot, G.–R. Guo, M.
Costello, A. Bot, Rational design of solid aerosols for immu-
noglobulin delivery by modulation of aerodynamic and release
characteristics, J. Control. Release 95 (2004) 489–500.
[66] A. Grenha, B. Seijo, C. Remunan-Lopez, Microencapsulated
chitosan nanoparticles for lung protein delivery, Eur. J. Pharm.
Sci. 25 (2005) 427–437.
[67] P. Calvo, C. Remunan-Lopez, J.L. Vita-Jato, M.J. Alonso,
Novel hydrophilic chitosan–polyethylene oxide nanoparti-
cles as protein carriers, J. Appl. Polym. Sci. 63 (1997)
125–132.
[68] I. Pasquali, R. Bettini, F. Gordono, Solid-state chemistry and
particle engineering with supercritical fluids in pharmaceu-
tics, Eur. J. Pharm. Sci. 27 (2006) 299–310.
[69] R.E. Sievers, U. Karst, P.D. Milewski, S.P. Sellers, B.A.
Miles, J.D. Schaefer, C.R. Stoldt, C.Y. Xu, Formation of
aqueous small droplet aerosol assisted by supercritical carbon
dioxide, Aerosol Sci. Technol. 30 (1999) 3–15.
[70] R.E. Sievers, U. Karst, J.D. Schaefer, C.R. Stoldt, B.A.
Watkins, Supercritical CO2-assisted nebulization for the
production and administration of drugs, J. Aerosol Sci. 27
(1996) S497–S498.
[71] E. Reverchon, I. De Marco, G. Della Porta, Rifampicin
microparticles production by supercritical anti-solvent pre-
cipitation, Int. J. Pharm. 243 (2002) 83–91.
[72] R.E. Sievers, B.A. Miles, S.P. Sellers, P.D. Milewski, K.D.
Kusek, New process for manufacture of 1-micron spherical
drug particles by CO2-assisted nebulization of aqueous
solutions, Proceeding of Resp. Drug Deliv., VI, South
Carolina, 1998, pp. 417–419.
[73] S.P. Sellers, G.S. Clark, R.E. Sievers, J.F. Carpenter, Dry pow-
ders of stable protein formulations from aqueous solutions
prepared using supercritical CO2-assisted aerosolization,
J. Pharm. Sci. 90 (2001) 785–797.
[74] R.E. Sievers, E.T.S. Huang, J.A. Villa, T.R. Walsh, H.V.
Meresman, C.D. Liang, S.P. Cape, Rapid gentle drying using
dense carbon dioxide to form fine dry powders, Proceed-
ings of Resp Drug Deliv: VIII. Tucson, Arizona, 2002,
pp. 675–677.
[75] R.E. Sievers, E.T.S. Huang, J.A. Villa, J.K. Kawamoto, M.M.
Evans, P.R. Brauner, Low temperature manufacturing of fine
pharmaceutical powders with supercritical fluid aerosolization
in bubble dryer, Pure Appl. Chem. 73 (2001) 1299–1303.
[76] R.N. Dalby, V. Naini, P.R. Byron, Droplet drying and electro-
static collection, a novel alternative to conventional commu-
nition techniques, J. Biopharm. Sci. 3 (1992) 91–99.
1028 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
[77] W.R. Gombotz, M.S. Healy, L.R. Brown, H.E. Auer, Process for
producing small particles of biologically active molecules,
European Patent Application (1994) EP0432232B1.
[78] Y.F. Maa, P.A. Nguyen, T. Sweeny, S.J. Shire, C.C. Hsu,
Protein inhalation powders. Spray drying vs spray freeze
drying, Pharm. Res. 16 (1999) 249–254.
[79] H.R. Constantino, L. Firouzabadian, K. Hogeland, C. Wu, C.
Banganski, C.M. Cordova, K.G. Carrasquillo, K. Griebenow,
S.E. Zale, M.A. Tracy, Protein spray freeze-drying. Effect of
atomization conditions on particle size and stability, Pharm.
Res. 17 (2000) 1374–1384.
[80] H.R. Constantino, L. Firouzabadian, C. Wu, K.G. Carras-
quillo, K. Griebenwo, S.E. Zale, M.A. Tracy, Protein spray
freeze-drying: 2. Effect of formulation variables on particle
size and stability, J. Pharm. Sci. 91 (2002) 388–395.
[81] S.D. Webb, S.T. Golledge, J.T. Cleland, J.F. Carpenter, T.W.
Randolph, Surface adsorption of recombinant human interferon-
ã in lyophilized and spray-lyophilized formulations, J. Pharm.
Sci. 91 (2002) 1474–1487.
[82] M.C. Heller, J.F. Carpenter, T.W. Randolph, Protein formu-
lation and lyophilization cycle design: prevention of damage
due to freeze-concentration induced phase separation, Bio-
technol. Bioeng. 63 (1999) 166–174.
[83] R.A. DePaz, D.A. Dale, C.C. Barnett, J.F. Carpenter, A.C.
Gaertner, W. Randolph, Effects of drying methods and
additives on the structure, function and storage stability of
subtilisin: role of protein conformation and molecular
mobility, Enzyme Microb. Technol. 31 (2002) 765–774.
[84] Z. Yu, K.P. Johnston, R.O. Williams III, Spray freezing into
liquid versus spray-freeze drying: influence of atomization on
protein aggregation and biological activity, Eur. J. Pharm. Sci.
27 (2006) 9–18.
[85] T.L. Roger, A.C. Nelsen, J. Hu, J.N. Brown, M. Sarkari, T.J.
Young, K.P. Johnston, R.O. William III, A novel particle
engineering technology to enhance dissolution of poorly
water soluble drugs: spray-freezing into liquid, Eur. J. Pharm.
Biopharm. 54 (2002) 271–280.
[86] Z. Yu, T.L. Rogers, J. Hu, K.P. Johnston, R.O. Williams III,
Preparation and characterization of microparticles containing
peptide produced by a novel process: spray freezing into
liquid, Eur. J. Pharm. Biopharm. 54 (2002) 221–228.
[87] Z. Yu, A.S. Garcia, K.P. Johnston, R.O. Williams, Spray
freezing into liquid for highly stable protein nanostructured
microparticles, Eur. J. Pharm. Biopharm. 58 (2004)
529–537.
[88] M. Henczka, J. Baldyga, B.Y. Shekunov, Modelling of spray
freezing with compressed CO2, Chem. Eng. Sci. 61 (2006)
2880–2887.
[89] P. York, Strategies for particle design using supercritical fluid
technologies, PSST 2 (1999) 430–440.
[90] J.W. Tom, P.G. Debenedetti, J. Aerosol Sci. 22 (1999) 555–584.
[91] W.J. Schmitt, R.C. Reid, Fluid Phase Equilib. 32 (1986) 77–99.
[92] E. Weidner, High-pressure chemical engineering, Process
Technol, Proceedings, vol. 12, Elservier, 1996.
[93] A. Tandya, F. Dehghani, N.R. Foster, Micronization of
cyclosporine using dense gas technique, J. Supercrit. Fluids
37 (2006) 272–278.
[94] V. Krukonis, Supercritical fluid nucleation of difficult to
comminute solids, Proceedings of the Annual Meeting
AlChE, vol. 140 F, 1984, p. T-214.
[95] E.M. Phillips, V.J. Stella, Rapid expansion from supercritical
solutions: applications to pharmaceutical processes, Int. J.
Pharm. 94 (1993) 1–10.
[96] K. Mishinma, K. Matsuyama, D. Tanabe, S. Yamauchi, T.J.
Young, K.P. Johnston, Microencapsulation of protein by rapid
expansion of supercritical solution with a non solvent,
Material, Interfaces and Electrochemical Phenomena,
AlChE Journal, vol. 46, 2000, pp. 857–865.
[97] T.J. Young, S. Mawson, K.P. Johnstone, I.B. Henriksen, G.W.
Pace, A.K. Mishra, Rapid expansion from supercritical to
aqueous solution to produce submicron suspensions of water-
insoluble drugs, Biotechnol. Prog. 16 (2000) 402–407.
[98] N. Jovanovic, A. Bouchard, G.W. Hofland, G. Witkamp, D.J.
Crommelin, W. Jiskoot, Stabilization of proteins in dry
powder formulations using supercritical fluid technology,
Pharm. Res. 21 (2004) 1955–1969.
[99] S.D. Yeo, G.B. Lim, P.G. Debenedetti, H. Bernstein, Formation
of microparticulate protein powders using a supercritical fluid
anti-solvent, Biotechnol. Bioeng. 41 (1993) 341–346.
[100] M. Jackson, H.H. Mantsch, Beware of proteins in DMSO,
Biochim. Biophys. Acta 1078 (1991) 231–235.
[101] R. Thiering, F. Dehghani, N.R. Foster, Micronization of
model proteins using compressed CO2, Proceedings of the 5th
International Symposium on Supercritical Fluids, Atlanta,
2000.
[102] N.R. Foster, H.L. Regtop, F. Dehghani, R.T. Bustami, H.–K.
Chan, Synthesis of small particles, WO Patent No: 0245690.
[103] R. Thiering, F. Dehghani, A. Dillow, N.R. Foster, The
influence of operating conditions on the dense gas precipi-
tation of model proteins, J. Chem. Technol. Biotechnol. 75
(2000) 29–41.
[104] S.D. Yeo, P.G. Debenechi, M. Radosz, H.W. Schmidt,
Supercritical anti-solvent process for substituted para-linked
aromatic polyamides: phase equilibrium and morphology
study, Macromolecules 26 (1993) 6207–6210.
[105] S.D. Yeo, G.B. Lim, P.G. Debenedetti, H. Bernstein, Formation
of microparticulate protein powders using a supercritical fluid
anti-solvent, Biotechnol. Bioeng. 41 (1993) 341–346.
[106] S. Palakodaty, P. York, Phase behaviour effects on particle
formation process using supercritical fluids, Pharm. Res. 16
(1999) 976–985.
[107] P. Caliceti, S. Salmaso, N. Elvassore, A. Bertucco, Effective
protein release from PEG/PLA nanoparticles produced by
compressed anti-solvent precipitation techniques, J. Control.
Release 94 (2004) 195–205.
[108] H. Okamoto, Y. Sakamura, K. Shiraki, K. Oka, S. Nishida, H.
Todo, K. Lida, K. Danjo, Stability of chitosan–pDNA
complex powder prepared by supercritical CO2 process, Int.
J. Pharm 290 (2005) 73–81.
[109] R. Bodmeier, H. Wang, D.J. Dixon, S. Mawson, K.P.
Johnston, Polymeric microspheres prepared by spraying into
compressed CO2, Pharm. Res. 12 (1995) 1211–1217.
[110] R.F. Falk, W. Th. Randolph, Process variable implications for
residual solvent removal and polymer morphology in the
 S.A. Shoyele, S. Cawthorne / Advanced Drug Delivery Reviews 58 (2006) 1009–1029
formation of gentamycin-loaded poly(L-lactide) microparti-
cles, Pharm. Res. 15 (1998) 1233–1237.
[111] Y. Perez deDiego, H.C. Pellikaan, F.E. Wubbolts, G.
Borchard, G.J. Witkamp, P.J. Jansens, Opening new operating
windows for protein micronization using the PCA process,
J. Supercrit. Fluids 36 (2006) 216–224.
[112] P.G. Debenedetti, G. Lim, R.K. PrudHomme, Formation of
protein microparticles by antisolvent precipitation. European
patent application EP052314A1, 1993.
[113] S. Yeo, G. Lim, P.G. Debenedetti, H. Bernestein, Formulation
of microparticulate protein powders using a supercritical fluid
antisolvent, Biotechnol. Bioeng. 41 (1993) 341–346.
[114] Y. Perez de Diego, M.B.P. van Rooijen, F.E. Wubbolts, P.J.,
Effect of process parameters on microparticle formation by
precipitation with compressed anti-solvent (PCA), Proceed-
ings of Eighth International Workshop on Industrial Crystal-
lization, 2001, pp. 98–104.
[115] E. Reverchon, I. DeMarco, Supercritical antisolvent micro-
nization of cefonicid, thermodynamic interpretation of results,
J. Supercrit. Fluids 31 (2004) 207–215.
[116] A. Weber, I.V. Yelash, T. Kraska, Effect of the phase
behaviour of the solvent–anti-solvent systems on the gas
anti-solvent crystallization of paracetamol, J. Supercrit. Fluids
33 (2005) 107–113.
[117] A.A. Elkordy, R.T. Forbes, B.W. Barry, Integrity of crystalline
lysozyme exceeds that of a spray dried form, Int. J. Pharm.
247 (2002) 79–90.
[118] A.A. Elkordy, R.T. Forbes, B.W. Barry, Stability of crystallised
and spray dried lysozyme, Int. J. Pharm. 278 (2004) 209–219.
[119] M.J. Lee, J.–H. Kwon, J.–S. Shin, C.–W. Kim, Microcrys-
tallization of α-lactalbumin, J. Crystal Growth 282 (2005)
434–437.
[120] A. Jen, H.P. Merkle, Pharm. Res. 18 (2001) 1483.
[121] C. Valder, D. Merrifield, Pharmaceutical Technology, Smithk-
line Beecham R&D News, vol. 32, 1996, p. 1.
[122] I. Krycer, J.A. Hersey, A comparative study of communica-
tion in rotatory and vibratory ballmills, Powder Technol. 27
(1980) 137–141.
[123] J.–H. Kwon, B.–H. Lee, J.–J. Lee, C.W. Kim, Insulin
microcrystal suspension as a long acting formulation for
pulmonary delivery, Eur. J. Pharm. Sci. 22 (2004) 107–116.
1029
[124] J.S. Skyler, R.A. Gelfand, I.A. Koirides, Treatment of type 2
diabetes mellitus with inhaled human insulin: a 3 month,
multicenter trial (abstract), Diabetes 47 (1998) A61 (suppl.).
[125] W.T. Cefalu, R.A. Gelfand, I.A. Koirides, Treatment of type 2
diabetes mellitus with inhaled insulin: a 3 month, multicenter
trial (abstract), Diabetes 47 (suppl.) (1998) A61.
[126] J.S. Patton, J. Bukar, S. Nagarajan, Inhaled insulin, Adv. Drug
Deliv. Rev. 35 (1999) 235–247.
[127] A. Sanchez, B. Villamayor, Y. Guo, J. Mclver, M.J. Alonso,
Formulation strategies for the stabilization of tetanus toxoid in
poly(lactide co-glycolide) microspheres, Int. J. Pharm. 185
(1999) 255–266.
[128] R.J. Davey, N. Blagden, G.D. Potts, R. Docherty, Polymor-
phism in molecular crystals: stabilization of a metastable form
by conformational mimicry, J. Am. Chem. Soc. 199 (2000)
1767–1772.
[129] G.G.Z. Zhang, D. Law, E.A. Schmilt, Y. Qiu, Phase
transformation considerations during process development
and manufacture of solid oral dosage forms, Adv. Drug Deliv.
Rev. 56 (2004) 371–390.
[130] P. York, Particle design, a ‘supercritical’ issue for the pharma
industry, Pharm. Vision, (200) 28–30.
[131] A.D. Edward, B.Y. Shekunov, A.I. Kordikowski, R.T. Forbes,
P. York, Crystallization of pure anhydrous polymorphs of
carbamazepine by solution enhanced dispersion with super-
critical fluids (SEDS), J. Pharm. Sci. 90 (2001) 1115–1124.
[132] S.P. Duddu, S.A. Sick, Y.H. Walter, T.E. Tarara, T.R.,
Improved lung delivery from a passive dry powder inhaler
using an engineered pulmosphere powder, Pharm. Res. 19
(2002) 689–695.
[133] A.H. de Boer, D. Gjaltema, P. Hagedoom, M. Schaller, W.
Witt, H.W. Frijlink, Design and application of a new modular
adapter laser diffraction characterisation of inhalation aero-
sols, Int. J. Pharm. 249 (2002) 233–245.
[134] C. Vemavarapu, M.J. Mollan, M. Lodaya, T.E. Needham,
Design and process aspect of laboratory scale SCF particle
formation systems, Int. J. Pharm. 292 (2005) 3.