Formulation development and antitumor activity of a filter-sterilizable emulsion of paclitaxel

Article abstract of DOI:10.1023/A:1007565230130

Purpose. Paclitaxel is currently administered i.v. as a slow infusion of a solution of the drug in an ethanol:surfactant:saline admixture. However, poor solubilization and toxicity are associated with this drug therapy. Alternative drug delivery systems, including parenteral emulsions, are under development in recent years to reduce drug toxicity, improve efficacy and eliminate premedication. Methods. Paclitaxel emulsions were prepared by high- shear homogenization. The particle size of the emulsions was measured by dynamic light scattering. Drug concentration was quantified by HPLC and in vitro drug release was monitored by membrane dialysis. The physical stability of emulsions was monitored by particle size changes in both the mean droplet diameter and 99% cumulative distribution. Paclitaxel potency and changes in the concentration of known degradants were used as chemical stability indicators. Single dose acute toxicity studies were conducted in healthy mice and efficacy studies in B 16 melanoma tumor-bearing mice. Results. QW8184, a physically and chemically stable sub-micron oil-in-water (o/w) emulsion of paclitaxel, can be prepared at high drug loading (8-10 mg/mL) having a mean droplet diameter of <100 nm and 99% cumulative particle size distribution of <200 nm. In vitro release studies demonstrated low and sustained drug release both in the presence and absence of human serum albumin. Based on single dose acute toxicity studies, QW8184 is well tolerated both in mice and rats with about a 3-fold increase in the maximum-tolerated-dose (MTD) over the current marketed drug formulation. Using the B16 mouse melanoma model, a significant improvement in drug efficacy was observed with QW8184 over Taxol. Conclusions. QW8184, a stable sub-micron o/w emulsion of paclitaxel has been developed that can be filter-sterilized and administered i.v. as a bolus dose. When compared to Taxol, this emulsion exhibited reduced toxicity and improved efficacy most likely due to the composition and dependent physicochemical characteristics of the emulsion.

Full text of DOI:10.1023/A:1007565230130

Pharmaceutical Research, Vol. 17, No. 2, 2000
Research Paper
Yew tree and like other natural products has limited aqueous
solubility (3). The drug is only administered intravenously since
it is orally inactive due to membrane transport and liver metabo-
lism limitations (4).
Formulation Development and
Antitumor Activity of a Filter-
Sterilizable Emulsion of Paclitaxel
The commercially available product, Taxol᭨ (paclitaxel
injection, Bristol-Myers Squibb Oncology), is currently formu-
lated in a vehicle containing approximately a 1:1 v/v mixture
of polyoxyethylated castor oil (Cremophor EL) and ethanol.
Cremophor EL, a commonly used surfactant for lipophilic com-
pounds, has been associated with bronchospasms, hypotension,
and other manifestations of hypersensitivity particularly follow-
ing rapid administration (5,6). Long infusion times upon a
10-fold dilution of the ethanol:Cremophor EL solution and
premedication are therefore required to reduce these adverse
effects. Furthermore, this formulation is associated with a num-
ber of issues such as stability with the possibility for drug
precipitation upon dilution, filtering requirements and use of
non-plasticized containers and administration sets (6). It is thus
apparent that there is a need for new formulations of paclitaxel
that are efficacious and less toxic than the commercial product
and can alleviate drug administration issues.
Panayiotis P. Constantinides,1,2 Karel J. Lambert,1
Alex K. Tustian,1 Brian Schneider,1 Salima Lalji,1
Wenwen Ma,1 Bryan Wentzel,1 Dean Kessler,1
Dilip Worah,1 and Steven C. Quay1
Received October 11, 1999; accepted November 12, 1999
Purpose. Paclitaxel is currently administered i.v. as a slow infusion of a
solution of the drug in an ethanol:surfactant:saline admixture. However,
poor solubilization and toxicity are associated with this drug therapy.
Alternative drug delivery systems, including parenteral emulsions, are
under development in recent years to reduce drug toxicity, improve
efficacy and eliminate premedication.
Methods. Paclitaxel emulsions were prepared by high-shear homogeni-
zation. The particle size of the emulsions was measured by dynamic
light scattering. Drug concentration was quantified by HPLC and in
vitro drug release was monitored by membrane dialysis. The physical
stability of emulsions was monitored by particle size changes in both
the mean droplet diameter and 99% cumulative distribution. Paclitaxel
potency and changes in the concentration of known degradants were
used as chemical stability indicators. Single dose acute toxicity studies
were conducted in healthy mice and efficacy studies in B16 melanoma
tumor-bearing mice.
Results. QW8184, a physically and chemically stable sub-micron oil-
in-water (o/w) emulsion of paclitaxel, can be prepared at high drug
loading (8–10 mg/mL) having a mean droplet diameter of Ͻ100 nm
and 99% cumulative particle size distribution of Ͻ200 nm. In vitro
release studies demonstrated low and sustained drug release both in
the presence and absence of human serum albumin. Based on single
dose acute toxicity studies, QW8184 is well tolerated both in mice
and rats with about a 3-fold increase in the maximum-tolerated-dose
(MTD) over the current marketed drug formulation. Using the B16
mouse melanoma model, a significant improvement in drug efficacy
was observed with QW8184 over Taxol᭨.
Conclusions. QW8184, a stable sub-micron o/w emulsion of paclitaxel
has been developed that can be filter-sterilized and administered i.v.
as a bolus dose. When compared to Taxol᭨, this emulsion exhibited
reduced toxicity and improved efficacy most likely due to the composi-
tion and dependent physicochemical characteristics of the emulsion.
In recent years considerable emphasis has been given to
the development of new formulations of paclitaxel that are
suitable for intravenous administration to address the aforemen-
tioned drug solubility and toxicity issues. These include dis-
persed systems such as emulsions (7–11), liposomes (4,12,13),
mixed micelles (14), cyclodextrins (15) and microspheres (16).
Water-soluble prodrugs such as polyethylene glycol- and polyg-
lutamate-paclitaxel with promising antitumor activity have also
been developed (17,18).
An o/w emulsion of paclitaxel using triacetin and ethyl
oleate as the oil phase, lecithin and pluronic F-68 as the surfac-
tant has been reported earlier by Tarr et al. (7). Although high
levels of paclitaxel have been solubilized in this emulsion (10–
15 mg/mL), no antitumor activity was reported. This is due to
slow precipitation of the drug upon dilution with a dextrose
solution and toxicity of triacetin at concentrations required for
delivering therapeutic doses of paclitaxel, with a reported LD50
of 1.2 mL/kg (7).
Oil-in-water emulsions using safflower or soybean oil,
lecithin and cholesterol and incorporating paclitaxel up to 5
mg/mL have been reported by Kaufman et al; (8). The mean
droplet diameter of these emulsions was reported to remain
KEY WORDS: paclitaxel; emulsions; filter-sterization; particle between 0.2 to 0.4 ␮m upon a six-week storage at 40ЊC. An
size; stability
o/w emulsion of paclitaxel and Taxol᭨ were evaluated for
toxicity in rats upon dilution to 1 mg of paclitaxel per mL
INTRODUCTION
followed by i.v. infusion over 30 min. At 42 mg/kg (42 mL/
kg of the diluted formulation) there were signs of severe toxicity
in the animals treated with Taxol᭨ whereas animals treated
with the emulsion showed low toxicity that was manifested
primarily as a body weight loss (8). When the emulsion and
the Taxol᭨ formulation were evaluated in vitro for efficacy
against mouse lymphocytic leukemia (L1210) or rat mammary
adenocarcinoma (NMU) cell lines, similar antitumor activity
was observed (8).
Paclitaxel is an important chemotherapeutic agent with a
wide spectrum of activity against solid tumors primarily breast,
ovarian, colon and non-small cell lung carcinomas (1). The
drug exerts its antitumor activity by binding to tubulin and
stabilizing microtubules and thus blocking cell mitosis (2).
Paclitaxel is a natural product present in the bark of the Pacific
Lundberg (9) reported on the preparation and evaluation
of complex submicron paclitaxel emulsions incorporating tri-
glycerides, phospholipids, polysorbate 80 and pegylated phos-
pholipids to prolong the circulation of the emulsion particles
1 SONUS Pharmaceuticals, Bothell, Washington 98021.
2 To whom correspondence should be addressed. (e-mail: panosc@
sonuspharma.com)
175
0724-8741/00/0200-0175$18.00/0 ᭧ 2000 Plenum Publishing Corporation

176
Constantinides et al.
in the blood. These emulsions, triolein:dipalmitoylphosphati- Emulsiflex C-50 (0.5–3.0 L batches) high-pressure homogeniz-
dylcholine: polysorbate 80: PEG-phosphatidylethanolamine: ers (Ottawa, Canada). Following homogenization, the finished
paclitaxel (1:1:0.4:0.1:0.03, mass ratio) with a mean droplet product was terminally sterilized by filtration through a 0.2
diameter of about 40 nm (9) showed good physical and chemical ␮m Posidyne filter (Pall Corp., East Hills, NY). The resulting
stability at low temperature and/or as a lyophilized powder. emulsions contained (%, w/w): a-tocopherol, TPGS (1–10%),
Paclitaxel-loaded emulsions demonstrated good antitumor Poloxamer 407 (1–5%), PEG 400 (1–10%), paclitaxel (0.5–
activity (in vitro) when compared to Taxol᭨. No in vivo studies 1.0%) and water for injection (60–90%).
were reported.
Recently, some prototype o/w emulsions using benzyl ben-
zoate and tributyrin as the oil phase, Tween 80 and Arlacel 20
Physical and Chemical Stability
as surfactants and incorporating up to 5 mg/mL paclitaxel have
been reported by Simamora et al. (10). Although these emul-
sions were found to be stable upon dilution with a variety of
i.v. fluids with no evidence of local irritation, studies on vehicle
toxicity and drug efficacy have not been reported.
Emulsions were analyzed for paclitaxel content by high
performance liquid chromatography (21) on a Phenosphere CN
column (5 microns, 150 ϫ 4.6 mm). The mobile phase consisted
of a methanol/water linear gradient beginning with 40:60 mix-
ture (v/v) at a flow rate of 1.0 mL/min and reaching 100%
methanol after 15 minutes. A UV detector at 227 nm was used
to detect and quantitate paclitaxel. A single peak was detected
which had a retention time and mass spectrogram consistent
with reference paclitaxel obtained from Hauser Laboratories
(Boulder, CO).
Emulsions were stored at 4ЊC or 25ЊC and samples were
removed at predetermined time points and assayed for physical
and chemical stability. The mean droplet diameter and 99%
cumulative distribution of the particles were used as indicators
of physical stability. Paclitaxel potency and the change in con-
centration of the known degradants were monitored for chemi-
cal stability. The following degradants were monitored during
the stability program: 7-epi-paclitaxel, 10-deacetyl-paclitaxel
and baccatin III.
The objective of this study was to develop an injectable
emulsion formulation of paclitaxel using vitamin E as the oil
phase and having the following characteristics: a) incorporate
high levels of paclitaxel (8–10 mg/mL) and is physically and
chemically stable; b) has mean droplet diameter and 99% cumu-
lative distribution of less than 0.2 ␮m and thus can be filter-
sterilized, and c) less toxic and at least as efficacious as the
commercial formulation Taxolᮋ. The use of vitamin E as the
oil phase in the present emulsions may have beneficial effects
in reducing drug toxicity since vitamin E has been reported to
be therapeutic of mucositis (19), one of the principal side effects
of Taxolᮋ therapy. Vitamin E is widely reported to reduce the
generalized toxicity of other cytotoxics (20).
MATERIALS AND METHODS
Materials
Drug Release
In vitro release of paclitaxel from different formulations
in the presence and absence of human serum albumin was
monitored by membrane dialysis at 37ЊC (22) using phosphate-
buffered saline as a sink solution at pH 7.0. The concentration
of paclitaxel in Taxolᮋ and QW8184, injectable paclitaxel
emulsion, was 6.0 and 8.8 mg/mL, respectively, and the drug
to albumin molar ratio was 1000 to 1. The sample volume in
the dialysis bag (cassette) was 1 mL and the sink volume
500 mL. Assuming 100% drug release, the concentration of
paclitaxel in the sink solution will be Ͻ20 ␮g/mL which is
within the reported aqueous paclitaxel solubility range of 1–40
␮g/mL (9). Pierce Slide-A-Lyzerᮋ dialysis cassettes with a
MW cutoff of 10K, and thus freely permeable to paclitaxel (MW
ϭ 853), were used and temperature control was maintained with
jacketed beakers via circulating bath. Concentrations of the
drug in pre-/post-dialysis samples and aliquots at various time
intervals were determined and drug release profiles (% drug
released vs time) were generated.
Vitamin E (DL-␣-tocopherol) USP/FCC grade was pur-
chased from Roche (Belvidere, NJ). Eastman’s vitamin E-TPGS
(␣-tocopherylpolyethyleneglycol-1000 succinate) was supplied
by B&D Nutritional Ingredients, Inc. (Carlsbad, CA). Polyeth-
ylene glycol 400 N.F. (PEG 400) and Poloxamer 407 (Pluronic
F-127) were purchased from Spectrum Quality Products, Inc.
(Gardena, CA) and BASF Corporation (Mount Olive, NJ),
respectively. Paclitaxel (Ͼ99% purity) was supplied by Hauser
Laboratories (Boulder, CO) and Hande Tech. Development Co.
(Houston, TX) and stored desiccated. Taxolᮋ, 5 mL vial at 6 mg/
mL, was purchased from a pharmacy and stored refrigerated.
Particle Size
Mean droplet diameter and particle size distribution were
determined with a Nicomp 370 Submicron Particle Sizer using
a 5 mW laser beam at 632.8 nm (Particle Sizing Systems, Santa
Barbara, CA). Polystyrene bead standards were used to verify
the calibration of the instrument. Data was analyzed in terms
of intensity, volume and number distributions and reported here
as volume weighted distribution.
Single Dose Acute Toxicity
The maximum tolerated dose (MTD) of paclitaxel emul-
sions was determined in mice following a tail vein injection
using 3–6 animals per group. Emulsions were administered i.v
at bolus doses ranging from 20–90 mg/kg or 2–10 mL/kg of
Emulsion Preparation
A pre-emulsion was first prepared by adding the oil phase vehicle. Taxolᮋ was first diluted with saline and then adminis-
(␣-tocopherol) in PEG 400 containing the surfactants (TPGS tered as 1–2 minute i.v infusion at doses ranging from 5–25
and Poloxamer 407) with or without paclitaxel to degassed mg/kg. A saline control group was also included. The determi-
water with vigorous mixing at 45ЊC followed by homogeniza- nation of the MTD was based on body weight loss (Ͼ15%)
tion in the Avestin Emulsiflex C-5 (20–200 mL batches) or and mortality, i.e. number of deaths per group (23,24).

A Filter-Sterilizable Emulsion of Paclitaxel
Efficacy: B16 Melanoma
177
The aforementioned factors were considered during formu-
lation development and optimization of the paclitaxel emulsion
in order to develop a stable and filter-sterilizable emulsion. The
mean droplet diameter and 99% cumulative distribution of a
typical stability lot of QW8184, lot A, are shown in Fig. 1.
There were no significant changes in particle size either at
4ЊC or 25ЊC over twelve months. The actual volume-weighted
particle size distribution of QW8184 incorporating 9 mg/mL
of paclitaxel is shown in the insert with a mean droplet diameter
of 62 nm. No precipitation or other gross changes were observed
during storage. Essentially the same particle size was obtained
with a small (50 mL) and a large (3 L) batch of QW8184.
When the two batches were compared after one month storage
at 4ЊC, the mean droplet/99% cumulative distribution were
67/150 nm and 64/149 nm for the small and large batches,
respectively. An independently manufactured lot of QW8184,
lot B, produced the following physical stability data in terms
of the mean droplet diameter and 99% cumulative distribution
(mean Ϯ SD, n ϭ 3): 63 Ϯ 2.1 nm (initial), 62.9 Ϯ 3.5 nm
(4ЊC, 9 months), 59.1 Ϯ 0.7 nm (25ЊC, 9 months) and 147 Ϯ
1.9 nm (initial), 146.1 Ϯ 2.8 nm (4ЊC, 9 months), 144.8 Ϯ 1.3
nm (25 ЊC, 9 months), respectively.
The chemical stability of the two independently manufac-
tured lots of QW8184, A and B, in terms of paclitaxel potency
and levels of known degradants at 4ЊC and 25ЊC and at time
zero, nine and twelve months is shown in Table 1. Intermediate
stability data was very similar to that shown in Table 1, for
clarity however, is not shown. There is no major change in the
concentration of the active ingredient at least during a 12-month
storage at either temperature. No significant changes in the
levels of any of the degradants were observed under these
storage conditions. Long-term stability is ongoing.
Female B6D2F mice (8 mice per group) were subcutane-
ously (sc) implanted with 107 B16 melanoma tumor cells
(23,24). Four days after implantation, mice were randomly
sorted into treatment groups and were administered i.v with
saline, Taxolᮋ, or QW8184 emulsion with and without incorpo-
rated paclitaxel on a schedule of either q3d ϫ 5 or q4d ϫ 5,
that is, once every three or four days, respectively, for a total
of 5 doses. QW8184 was administered as a bolus injection and
Taxolᮋ was infused over 2 minutes following 10-fold dilution
with saline. The administered volume of saline and drug-free
emulsion (vehicle) were 7 and 7 or 8 mL/kg, respectively. For
specific dosages and schedules see the Figs. 4 and 5 legends.
Antitumor activity was assessed according to the guidelines
established by the National Cancer Institute (24).
RESULTS AND DISCUSSION
Development of Filter-Sterilizable Emulsions
The formulation development strategy of using ␣-tocoph-
erol as the oil phase and TPGS as the primary surfactant was
followed to develop a stable and efficacious emulsion of pacli-
taxel at high drug loading that can be filter-sterilized. Ensuring
product sterility of parenteral emulsions is crucial and terminal
heat sterilization has been generally used for this purpose. Alter-
natively, certain emulsions can be lyophilized and reconstituted
prior to administration without loss of viability and drug potency
(25). If, however, the components of a particular drug-emulsion
are heat labile, filter sterilization of the product may be a viable
option, provided that the emulsion droplets pass through 0.2
micron pore. However, filter sterilization of emulsions is quite
challenging and use of this non-invasive method of sterilization
has been limited (26).
In Vitro Drug Release
Several prototype formulations were developed which,
upon further optimization, yielded the lead development candi-
date QW8184. Formulation optimization was based on the use
of an alternate co-surfactant or co-solvent, and the resulting
emulsions were evaluated for acute and cumulative toxicity,
filterability, drug loading, and where appropriate, stability and
efficacy. Several of the prototype formulations were found to
be stable and efficacious; however, only data with the most
advanced formulation tested to date, QW8184, is presented.
One of the desired characteristics of a drug delivery vehicle
is to provide sustained release of the incorporated drug, a charac-
teristic quite often correlated with improved pharmacokinetics
and efficacy (28,29). In particular, long-circulating emulsions
Physical and Chemical Stability
Physical stability of an emulsion is one of the most
important desired product characteristics. Emulsions are hetero-
geneous systems and thermodynamically unstable and, there-
fore, have a significant tendency to lose physical stability on
storage. The extent of this process is dependent on the character-
istics of each formulation and storage conditions. Some of the
factors that affect the physical stability of emulsions include
the type and levels of surfactant(s) used to stabilize the droplets,
the phase volume ratio, droplet size, compatibility of drug and
excipients with the emulsion, and storage conditions of the
emulsion (25,27). In general, the smaller the particle size of
the emulsion, the better its physical stability (25,27). The assess-
ment of the physical stability of emulsions and of dispersed
systems in general, in addition to being extremely useful during
formulation development and optimization, also allows setting
up of product specifications and expiration dates.
Fig. 1. Mean droplet diameter, [D]mean, and 99% cumulative distribu-
tion, [D]99%, of QW8184 as a function of storage time and temperature.
This particular lot of QW8184 (lot A) incorporated 9.0 mg/mL pacli-
taxel and was stored at 4ЊC or 25ЊC. The actual volume-weighted
particle size distribution of QW8184 is shown in the insert with a mean
droplet diameter of 62 nm.

178
Constantinides et al.
Table 1. QW8184 Chemical Stability: Paclitaxel Potency and Degradants
Degradants (%, mean Ϯ SD)a
Storage time, months
(Lot, temp.)
Paclitaxel potency, mg/mL
(mean Ϯ SD)a
7-Epi-paclitaxel
Baccatin-3
10-Deacetyl-paclitaxel
0.0 (A, 4 ЊC)
9.2 (A, 4 ЊC)
12.2 (A, 4 ЊC)
9.2 (A, 25 ЊC)
12.2 (A, 25 ЊC)
0.0 (B, 4 ЊC)
9.2 (B, 4 ЊC)
9.2 (B, 25 ЊC)
8.4 Ϯ 0.6
9.2 Ϯ 0.1
8.8 Ϯ 0.3
9.0 Ϯ 0.7
8.3 Ϯ 0.8
9.1 Ϯ 0.0
9.4 Ϯ 0.0
9.2 Ϯ 0.1
0.61 Ϯ 0.29
0.36 Ϯ 0.02
0.30 Ϯ 0.04
0.40 Ϯ 0.02
0.35 Ϯ 0.07
0.50 Ϯ 0.07
0.43 Ϯ 0.03
0.41 Ϯ 0.04
0.16 Ϯ 0.01
0.17 Ϯ 0.00
0.21 Ϯ 0.04
0.18 Ϯ 0.01
0.21 Ϯ 0.04
0.07 Ϯ 0.01
0.08 Ϯ 0.01
0.08 Ϯ 0.01
0.15 Ϯ 0.02
0.18 Ϯ 0.01
0.20 Ϯ 0.03
0.18 Ϯ 0.01
0.21 Ϯ 0.04
0.11 Ϯ 0.01
0.13 Ϯ 0.02
0.13 Ϯ 0.01
a
n ϭ 10 (lot A); n ϭ 3 (lot B).
of paclitaxel can improve the delivery of the drug to cancer concentration of 6 mg/mL and prior to administration it requires
sites in the body. The release kinetics of paclitaxel from dilution with commonly used i.v. fluids to a final concentration
QW8184, both in the absence and presence of human serum of 0.3 to 1.2 mg/mL that corresponds to a 20-fold and 5-fold
albumin (HSA) at a drug to albumin molar ratio of 1000:1 and dilution, respectively. When Taxolᮋ was administered i.v. at
from the commercial formulation Taxolᮋ are shown in Fig. 2. 24 mg/kg, either as a bolus dose or 2 min infusion, after a 10-
The concentration of paclitaxel in Taxolᮋ and QW8184 was fold dilution with saline, severe prostration and catatonia were
6.0 and 8.8 mg/mL, respectively. The release of paclitaxel from observed. When Taxolᮋ was first diluted 1:4 with saline and
the emulsion was slow both in the absence and presence of then administered i.v. at 24 mg/kg as a 2 or 5 min infusion,
serum albumin with less than 5% drug being released over 24 100% mortality was observed as compared to 50% mortality
hours, whereas about 12% was released from the commercial when the drug was administered as i.v. bolus. Weight loss was
formulation.
less than 15% even in the group receiving the highest dose of
QW8184 and the animals recovered or gained weight over a
period of 10 days post injection. Vehicle toxicity was also
evaluated. Animals receiving drug-free emulsion at 7 or 8 mL/
kg (the equivalent paclitaxel dose of 63–72 mg/kg at drug
loading of 9.0 mg/mL) grew rapidly and gained slightly more
weight than animals receiving saline. This may be attributed
to the vitamin and calorie content of the formulation. The
clinical dose of paclitaxel is 175 mg/m2 (300 mg total per
person) or 4.3 mg/kg for a 70 kg average human body weight
(1,23). The equivalent dose/volume of QW8184 is about 0.5
mL/kg at a drug loading of 9.0 mg/mL which is about an order
of magnitude below the dose where the emulsion vehicle may
show signs of toxicity. The high drug loading in QW8184
Acute Single Dose Toxicity
Single dose toxicity studies of paclitaxel emulsions in
healthy mice were carried out and representative data with
QW8184 is shown in Fig. 3. The paclitaxel emulsion was well
tolerated and the maximum tolerated dose (MTD) was deter-
mined to be approximately 70 mg/kg for QW8184 as compared
to approximately 20 mg/kg for Taxolᮋ. The marketed drug
product Taxolᮋ is primarily supplied in 5 mL vials at a drug
Fig. 2. Percent paclitaxel released as a function of time at 37ЊC. Drug
release from the lipid emulsion, QW8184, in the presence and absence
of HSA (human serum albumin) and the commercial formulation Tax-
olᮋ was monitored by membrane dialysis as described in Methods Fig. 3. Maximum tolerated dose of paclitaxel in mice. The determina-
using PBS (phosphate-buffered saline) as a sink solution at pH 7.0. tion of the MTD was based on body weight loss (Ͼ15% is considered
The drug to albumin molar ratio was 1000:1.
toxic) and mortality (number of deaths per group of 6 animals).

A Filter-Sterilizable Emulsion of Paclitaxel
179
without prior dilution enables one to administer much lower
volumes, therefore reducing or potentially eliminating issues
of vehicle toxicity.
Efficacy Against B16 Melanoma
B16 Melanoma is a fast growing solid murine tumor that
has been commonly used in early screening of different antican-
cer agents (23,24). Figures 4, A and B present tumor growth
regression and percent survival, respectively, for the dosing
regimen of q3d ϫ 5. QW8184 was evaluated at different doses
and compared to the reference Taxolᮋ formulation at its MTD
(20 mg/kg). The corresponding data for the dosing schedule of
q4d ϫ 5 are shown in Figure 5, A and B, respectively. QW8184
illustrated a definitive dose response in both schedules. The
q3d ϫ 5 schedule, however, seems to be more effective. Admin-
istration of QW8184 at dosages of 20 mg/kg (63 mg/m2) and
40 mg/kg (125 mg/m2) on a schedule of q3d ϫ 5 resulted in
mean increases in survival times of 65% and 94%, respectively,
as compared to control group (Table 2).
Log-cell kill values of 1.8 and 3.0 were observed with
QW8184 at dosages of 20 and 40 mg/kg, respectively (Table
2), as compared to a value of 0.5 obtained with Taxolᮋ at 20
mg/kg (q3d ϫ 5). In addition, there was a significant reduction
in tumor growth in these animals as depicted in Fig. 4A. A
significant reduction in tumor growth was also observed in
animals administered QW8184 on a schedule of q4d ϫ 5 as
illustrated in Fig. 5A. A dosage of 60 mg/kg with QW8184
was toxic in the q3d ϫ 5 regimen with only 1 survivor at the
end of the dosing schedule (Fig. 4B). In the q4d ϫ 5 dosing
schedule, 70 mg/kg of QW8184 was toxic with only 2 survivors
Fig. 4. A) B16 melanoma mean tumor regression to QW8184 and
Taxolᮋ as a function of time on a q3d ϫ 5 schedule. Error bars
at the end of the dosing schedule (Fig. 5B). Statistical analysis represent SEM (n ϭ 8). B) Percent B16 melanoma survival in response
to QW8184 and Taxolᮋ as a function of time on a q3d ϫ 5 schedule.
(t-test) of the survival time between groups was performed.
The results indicate that Taxol᭨ was not statistically different
from the saline control at the q3d ϫ 5 (p ϭ 0.25) or q4d ϫ 5 (p
ϭ 0.09) schedule. QW8184 resulted in a statistically significant
increase in survival time in the q3d ϫ 5 groups at dosages of
20 mg/kg (p ϭ 0.005) and 40 mg/kg (p Ͻ 0.005) as compared
Table 2. Antitumor Activity of QW8184 vs Taxolᮋ in the B16 Melanoma Model
Dose
mg/kg
n ϭ 8
Median tumor weight on day
Survival
(mean Ϯ SD)
%
Log
cell
killd
Test
article
Schedule
days
Mortality
day 20
% T/Ca
day 20
% TGIb
day 20
T-Cc
days
3
15
20
23
34
days
Saline
Vehicle Control
Taxolᮋ
QW8184
QW8184
QW8184
Vehicle Control
Taxolᮋ
Control
q3d x 5
q3d x 5
q3d x 5
q3d x 5
q3d x 5
q3d x 5
q4d x 5
q4d x 5
q4d x 5
q4d x 5
q4d x 5
0
0
0
0
0
0
0
0
0
0
0
699
388
306
50
7
—
400
184
223
49
2422
2358
1871
270
6
—
—
—
903
139
—
—
—
1079
407
15
—
—
—
—
1340
—
—
—
—
—
17 Ϯ 2
20 Ϯ 1
19 Ϯ 5
28 Ϯ 7
33 Ϯ 5
11 Ϯ 3
16 Ϯ 6
12 Ϯ 7
20 Ϯ 3
31 Ϯ 4
17 Ϯ 16
87.5
0e
50
—
93
77
11
0
—
100
69
35
10
0
—
3
—
3
3
10
17
—
0
3
3
7
17
—
—
0.5
1.8
3.0
—
20
20
40
60
23
89
100
—
0
31
65
90
100
0
0f
—
100
62.5
75
50
0
2658
1679
855
246
2
0
20
20
50
70
0.5
0.5
1.2
3.0
QW8184
QW8184
QW8184
2
1577
62.5
a
% T/C ϭ (Median Tumor Wt of treated / Median Tumor Wt of control) ϫ 100.
%TGI ϭ 100 Ϫ (%T/C).
T-C ϭ Tumor Growth Delay Value (median time for the treatment group (T) and control (C) to reach a predetermined size (Ͼ 750 mg).
Log Cell Kill ϭ (T-C value)/(3.32 ϫ tumor doubling time).
All animals were sacrificed due to excessive tumor size.
All animals but three survived to day 27 when they were sacrificed due to tumor size (Ͼ10% of body weight).
b
c
d
e
f

180
Constantinides et al.
Therefore, the observed minimal therapeutic effect from Tax-
ol᭨ in the B16 melanoma model is not believed to be due to
sub-optimal dosing.
The effects observed in the present study have not been
reported with previous injectable emulsions of paclitaxel (7–
10). The improved efficacy of QW8184 may be related to its
preferential uptake by tumor cells as a result of its physicochem-
ical characteristics, particularly lipid composition and particle
size. The in vitro drug release data suggest that emulsion drop-
lets may serve as long-circulating drug reservoirs, thus improv-
ing the delivery of paclitaxel to tumor sites, due to increased
cellular/droplet interactions (34). To penetrate tumors in tissue,
particles must be small enough to pass through endothelial
fenestrae, i.e. Ͻ 70 nm. This mechanism is consistent with
literature reports on long-circulating emulsions and liposomes
(29,34) where lipid composition, particularly the inclusion of
pegylated surfactants and small droplet size, direct lipophilic
drugs away from RES in the liver and spleen to other targeting
tissues, such as inflammatory tissues.
In addition to the studies reported with drug emulsions,
toxicity and antitumor activity was reported with other lipid-
based carriers of paclitaxel, such as liposomes and mixed
micelles. The most extensive work with liposomal formulations
of paclitaxel has been reported by Straubinger et al; (3,12,13).
They developed over 300 liposome formulations of various lipid
compositions and evaluated them for stability and antitumor
activity. Both in vitro and in vivo activity using Colon-26, a
Taxol-resistant murine tumor, were demonstrated (12). Pacli-
taxel liposomes, however, were both unstable and toxic at high
drug loading.
Paclitaxel was solubilized in mixed micelles formed by a
mixture of bile salts and phospholipids followed by a spontane-
ous transformation into drug-loaded liposomes, thus avoiding
drug precipitation (14). These formulations, incorporating less
than 1.0 mg of paclitaxel per mL, produced significant antitumor
activity in vitro and appeared to be less toxic than the Cremophor
EL vehicle (14). It is anticipated however, that at high drug
loading these formulations can be both unstable and toxic with
limited clinical use.
Fig. 5. A) B16 melanoma mean tumor regression to QW8184 and
Taxolᮋ as a function of time on a q4d ϫ 5 schedule. Error bars
represent SEM (n ϭ 8). B) Percent B16 melanoma survival in response
to QW8184 and Taxolᮋ as a function of time on a q4d ϫ 5 schedule.
to the saline control. The animals administered QW8184 at q4d
ϫ 5 resulted in a statistically significant increase in survival
time at dosages of 20 mg/kg (p ϭ 0.05) and 50 mg/kg (p Ͻ
0.005) as compared to the saline control.
Emulsion and liposomal formulations of paclitaxel are
more biocompatible and thus less toxic and at least as efficacious
as the marketed Taxol᭨. Emulsions have higher drug solubiliza-
tion capacity than liposomes and are easier to process and
manufacture in a sterile form. The injectable paclitaxel emulsion
QW8184 described in the present studies exhibits these advan-
tages. Due to the high paclitaxel loading (8–10 mg/mL),
QW8184 can be administered clinically undiluted at high dose
without drug precipitation. As a result of its small droplet size
(Ͻ200 nm for both the mean droplet and cumulative distribu-
tion), QW8184 can be filter-sterilized, exhibits improved shelf-
life, efficacy and targeting to solid tumors. Furthermore,
QW8184 can result in better quality of life, since it can be
administered to patients as a bolus dose and has low toxicity.
Emulsion formulations offer an appealing alternative for the
bolus administration of paclitaxel and other poorly soluble drugs
due to their effectiveness for drug solubilization, improved
efficacy and patient quality of life arising from reduced side
effects. A real therapeutic gain with a particular new formulation
Table 2 summarizes the overall results of the efficacy study
including mean survival times, median tumor weights and log
cell kills for each group and treatment schedule. By all end
points of efficacy (24), QW8184 exhibited superior antitumor
activity in mice at doses that included or well exceeded the MTD
of Taxol᭨ but which were well tolerated. In our laboratory, the
MTD for Taxol᭨ was approximately 20 mg/kg and it is sup-
ported by other literature reports. Bissery et. al; (30) reported
that the highest dose of Taxol᭨ that could be administered to
mice without causing death or undue toxicity was 21.7 mg/kg/
injection. Furthermore, recent reports on the MTD of Taxol᭨
upon i.v. administration in mice indicate that it is 20 mg/kg on
a q1d ϫ 3 (31) or 25 mg/kg on a q1d ϫ 5 (32). From Table
2, it is evident that Taxol᭨ did result in a slight reduction in
tumor growth at both the 15 and 20 days post tumor implant
as compared to the controls. It should also be noted that Taxol᭨
has been shown to be minimally effective in the B16 melanoma
tumor model (33). In addition, drug related mortality is evident
at day 20 in the Taxol᭨ group thus providing further evidence
that doses higher than 20 mg/kg could not have been tolerated. of paclitaxel, including QW8184, can be realized through its

A Filter-Sterilizable Emulsion of Paclitaxel
181
13. A. Sharma and R.M. Straubinger. Novel taxol formulations: prepa-
ration and characterization of taxol-containing liposomes. Pharm.
Res. 11:889–896 (1994).
14. H. Alkan-Onyuksel, S. Ramakrishnan, H.-B. Chai, and J. M.
Pezzuto. A mixed micellar formulation suitable for the parenteral
administration of taxol. Pharm. Res. 2:206–212 (1994).
15. U. S. Sharma, S. V. Balasubramanian and R. M. Straubinger.
Pharmaceutical and physical properties of paclitaxel (Taxol)
complexes with cyclodextrins. J. Pharm. Sci. 84:1223–1230
(1995).
16. Y. M. Wang, H. Sato, and I. Horikoshi. In vitro and in vivo
evaluation of taxol release from poly(lactic-co-glycolic acid)
microspheres containing isopropyl myristate and degradation of
the microspheres. J. Contr. Rel. 49:157–166 (1997).
17. C. Li, D-F., Yu, T. Inoue, D. J. Yang, L. Milas, N. R. Hunter, E.
E. Kim, and S. Wallace. Synthesis and evaluation of water-soluble
polyethylene glycol paclitaxel conjugate as paclitaxel prodrug.
Anticancer Drugs 7:642–648 (1996).
progression to the clinic. As new and safer drug delivery meth-
ods of paclitaxel emerge, it is likely that the clinical use of the
drug will expand.
CONCLUSIONS
QW8184, a stable, injectable and filter sterilizable o/w
emulsion of paclitaxel has been developed at high drug loading
(8–10 mg/mL) with a mean droplet diameter and 99% cumula-
tive particle size distribution of Ͻ 0.2 ␮m. Compared to the
in vitro drug release from Taxol᭨, drug release from QW8184
at 37ЊC is slow both in the presence and absence of human
serum albumin with less than 5% drug being released within
24 hours. In the B16 melanoma tumor model in mice the drug
emulsion was better tolerated and was more efficacious than
Taxol᭨. The composition and dependent physicochemical char- 18. C. Li, D-F., Yu, R. A. Newman, C. Fernando, L. C. Stephens,
N. Hunter, L. Milas, and S. Wallace. Complete regression of well-
acteristics of the emulsion may be related to reduced toxicity
established tumors using a novel water-soluble poly (L-glutamic
and improved efficacy.
acid)-paclitaxel conjugate. Cancer Res. 58:2404–2409
(1998).
19. R. G. Waldeigh, R. S. Redman, M. L. Graham, S. H. Krasnow,
A. Anderson, and M. H. Cohen. Vitamin E in the treatment of
chemotherapy-induced mucositis. Am. J. Med. 92:481–484
(1992).
20. A. Geetha, R. Sankar, T. Marar, and C. S. Shyamala-Devi. ␣-
Tocopherol reduces doxorubicin-induced toxicity in rats-histolog-
ical and biochemical evidences. Ind. J. Physiol. Pharmac. 34:
94–100 (1990).
21. S. D. Harvey, J. A. Campbell, R. G. Kesley, and N. C. Vance.
Separation of paclitaxel from related taxanes in Taxus brevifolia
extracts by isocratic elution reverse-phase microcolumn high-
performance liquid chromatography. J. Chromatogr. 582:273–
278 (1991).
22. N. S. Santos Magalhaes, G. Cave, M. Seiller, and S. Benita. The
stability and in vitro release kinetics of a clofibride emulsion. Int.
J. Pharm. 76:225–237 (1991).
23. W. Rose. Preclinical Antitumor Activity of Taxanes, in Taxolᮋ:
Science and Applications (M. Suffness ed.), CRC Press, Boca
Raton, pp. 237–258 (1996).
24. J. Plowman, D. J. Dykes, M. Hollingshead, L. Simpson-Herren,
and M. C. Alley. Human tumor xenograft models in NCI develop-
ment, in Anticancer Drug Development Guide: Preclinical Screen-
ing, Clinical Trials and Approval (B. A. Teicher ed.), Humana
Press, Totowa, New Jersey, pp. 101–125 (1997).
25. L. C. Collins-Gold, R. T. Lyons, and L. C. Bartholow. Parenteral
emulsions for drug delivery. Adv. Drug Del. Rev. 5:189–208
(1990).
26. D. M. Lidgate, T. Trattner, R. M. Shultz, and R. Maskiewicz.
Sterile filtration of a parenteral emulsion. Pharm. Res. 9:860–
863 (1992).
ACKNOWLEDGMENTS
We thank Dr. Polly R. Pine of SRI International, Pharma-
ceutical Discovery Division, for her assistance with the B16
Melanoma study. The authors also like to thank Dr. Eric K.
Rowinsky of the Institute of Drug Development, San Antonio
Cancer Therapy and Research Center, for his valuable
comments.
REFERENCES
1. E. K. Rowinsky and Donehower, R. C. Paclitaxel (Taxol). N.
Engl. J. Med. 332:1004–1014 (1995).
2. S. B. Horwitz. Mechanism of action of taxol. Trends in Pharma.
Sci. (TIPS) 13:134–135 (1992).
3. R. M. Straubinger. Biopharmaceutics of paclitaxel (taxol): formu-
lation, activity and pharmacokinetics, in Taxolᮋ: Science and
Applications (M. Suffness ed.). CRC Press, Boca Raton, pp. 237–
258 (1996).
4. A. Sparreboom, J. Van Asperen, U. Meyer, A. H. Schinkel, J. W.
Smit, D. K. F. Meijer, P. Borst, W. J. Nooijen, J. H. Beijen, and
O. Van Tellingen. Limited oral bioavailability and active epithelial
excretion of paclitaxel (Taxol) caused by p-glycoprotein in the
intestine. Proc. Natl. Acad. Sci. USA 94:2031–2035 (1997).
5. W. Lorenz, H. J. Reimann, A. Schmal, H. Schult, S. Lang, C.
Ohmann, D. Weber, B. Kapp, L. Luben, and A. Doenicke. Hista-
mine release in dogs by Cremophor EL and its derivatives: oxye-
thylated oleic acid is the most effective constituent. Agents Actions
7:63–67 (1977).
27. A. Kurihara, Y. Shibayama, A. Mizota, A. Yasuno, M. Ikeda,
K. Sasagawa, T. Kobayashi, and M. Hisaoka. Enhanced tumor
delivery and antitumor activity of palmitoyl rhizoxin using stable
lipid emulsions in mice. Pharm. Res. 13:11305–11320 (1996).
28. C. Washington. Drug release and interfacial structure in emul-
sions, in Emulsions and Nanosuspensions for the Formulation of
Poorly Soluble Drugs (R. H. Muller, S. Benita and B. Bohm,
eds), MedPharm Scientific Publishers, Stuttgart, Germany, pp.
101–117 (1998).
29. F. Liu and D. Liu. Long-circulating emulsions (oil-in-water)
as carriers for lipophilic drugs. Pharm. Res. 12:1060–1064
(1995).
30. M. C. Bissery, D. Guenard, F. Gueritte-Voegelein, and F. Lavelle.
Experimental antitumor activity of Taxotere (RP 56976, NSC
628503), the marketed formulation of docetaxel, Cancer Res.
51:4845–4852 (1991).
6. D. Dye and J. Watkins. Suspected anaphylactic reaction to Cremo-
phor EL. Br. Med. J. 280:1353 (1980).
7. B. D. Tarr, T. G. Sambandan, and S. H. Yalkowsky. A new
parenteral emulsion for the administration of taxol. Pharm. Res.
4:162–165 (1987).
8. R. J. Kaufman, T. J. Richard, and R. W. Fuhrhop. Stable oil-in-
water emulsions incorporating a taxine (taxol) and method of
making same. U.S. Patent 5,616,330, April 1, 1997.
9. B. Lundberg. A submicron lipid emulsion coated with amphi-
pathic polyethylene glycol for parenteral administration of pacli-
taxel (Taxol). J. Pharm. Pharmacol. 49:16–21 (1997).
10. P. Simamora, R-M. Dannenfelser, S. E. Tabibi, and S. H. Yalkow-
sky. Emulsion formulations for intravenous administration of
paclitaxel. PDA J. Pharm. Sci. Technol. 52:170–172 (1998).
11. P. P. Constantinides, K. Lambert, A. K. Tustian, W. Ma, B. Schnei-
der, S. Lalji, B. Wentzel, D. Kessler, D. Worah, and S. C. Quay.
Reduced toxicity and improved efficacy of paclitaxel incorporated
in oil-in-water emulsions. Pharm Sci. 1:S-100 (1998).
12. A. Sharma, E. Mayhew, and R. M. Straubinger. Antitumour effect
of taxol containing liposomes in a taxol-resistant murine tumor
model. Cancer Res. 53:5877–5881 (1993).
31. P. C. Gokhale, J.T. Newsome, A. Dritschilo, U. Kasid and A.
Rahman. Toxicity, histopathologic and pharmacokinetic evalua-
tion of liposome-encapsulated paclitaxel and taxol in mice and
rabbits. Proc. Amer. Assoc. Cancer Res. 40:#2759 (1999).
32. I. Ahmad, W. Perkins, G. R. Masters, J. Nguyen, X. Li, J. J.

182
Constantinides et al.
Schupsky, S. Ali, A. S. Janoff, and Mayhew, E. Therapeutic
efficacy of a novel polyethyleneglycol (PEG)-lipid formulation
of 2Ј [alpha]-bromo-hexedecanoyl paclitaxel (BrC16HT) against
human tumor xenografts in SCID mice. Proc. Amer. Assoc. Cancer
Res. 40:#3849 (1999).
Science and Applications (M. Suffness ed.) CRC Press, Boca
Raton, pp. 209–235 (1996).
34. D. D. Lasic, F. J. Martin, A. Gabizon, S. K. Huang, and D.
Papahadjopoulos. Sterically stabilized liposomes: a hypothesis on
the molecular origin of the extended circulation times. Biochim.
Biophys. Acta 1070:187–192 (1991).
33. W. C. Rose. Preclinical antitumor activity of taxanes, in Taxolᮋ