Chemical modification of paclitaxel (taxol) reduces P-glycoprotein interactions and increases permeation across the blood-brain barrier in vitro and in situ

Article abstract of DOI:10.1021/jm040114b

The purpose of this work was to introduce a chemical modification into the paclitaxel (Taxol) structure to reduce interactions with the product of the multidrug resistant type 1 (MDR1) gene, P-glycoprotein (Pgp), resulting in improved blood-brain barrier

Full text of DOI:10.1021/jm040114b

832
J. Med. Chem. 2005, 48, 832-838
Chemical Modification of Paclitaxel (Taxol) Reduces P-Glycoprotein
Interactions and Increases Permeation across the Blood-Brain Barrier in Vitro
and in Situ
Antonie Rice,† Yanbin Liu,‡ Mary Lou Michaelis,§ Richard H. Himes,| Gunda I. Georg,‡ and Kenneth L. Audus*,†
Department of Pharmaceutical Chemistry, University of Kansas, 226 Simons, 2095 Constant Avenue, Lawrence, Kansas 66047,
Department of Pharmacology and Toxicology, University of Kansas, Malott Hall 5064, 1251 Wescoe Hall Drive,
Lawrence, Kansas 66045, Department of Medicinal Chemistry, University of Kansas, Malott Hall 4070,
1251 Wescoe Hall Drive, Lawrence, Kansas 66045, and Department of Molecular Biosciences, University of Kansas,
Haworth Hall 2034, 1200 Sunnyside Avenue, Lawrence, Kansas 66045
Received June 10, 2004
The purpose of this work was to introduce a chemical modification into the paclitaxel (Taxol)
structure to reduce interactions with the product of the multidrug resistant type 1 (MDR1)
gene, P-glycoprotein (Pgp), resulting in improved blood-brain barrier (BBB) permeability.
Specifically, a taxane analogue, Tx-67, with a succinate group added at the C10 position of
Taxol, was synthesized and identified as such a candidate. In comparison studies, Tx-67 had
no apparent interactions with Pgp, as demonstrated by the lack of enhanced uptake of
rhodamine 123 by brain microvessel endothelial cells (BMECs) in the presence of the agent.
By contrast, Taxol exposure substantially enhanced rhodamine 123 uptake by BMECs through
inhibition of Pgp. The transport across BMEC monolayers was polarized for both Tx-67 and
Taxol with permeation in the apical to basolateral direction greater for Tx-67 and substantially
reduced for Taxol relative to basolateral to apical permeation. Taxol and cyclosporin A
treatments also did not enhance Tx-67 permeation across BMEC monolayers. In an in situ rat
brain perfusion study, Tx-67 was demonstrated to permeate across the BBB at a greater rate
than Taxol. These results demonstrate that the Taxol analogue Tx-67 had a reduced interaction
with Pgp and, as a consequence, enhanced permeation across the blood-brain barrier in vitro
and in situ.
Introduction
ties with respect to crossing the BBB. The hypothesis
by Seelig1 also holds that there are differences in affinity
for Pgp among structurally similar molecules. These
differences are characterized by fixed spatial separa-
tions of pairs of recognition elements, type I and type
II, that have different strengths to be electron donors
or acceptor groups. It is believed that these differences
could be exploited by chemical modification of the
recognition elements. That is, by chemical modification,
the recognition elements on a particular molecule would
be decreased and the entity’s strength for Pgp binding
thus reduced. Seelig further observed that type I units
with negative charge, such as the carboxylic acids group
(e.g., S-farnesylcysteine and reserpic acid), the sulfonyl
group (e.g., trypan blue), or the mesomeric nitro group
(e.g., flunitrazepam), do not interact with Pgp. On the
other hand, chemical structures that bear type I units
in addition to the negatively charged groups are sub-
strates (e.g., cefoperazone or morphine-6-glucoronide).1
The objective of these studies was to characterize the
transport of Taxol and a novel taxane, Tx-67, with
potentially reduced Pgp interactions, both in vitro and
in situ. The structures of Taxol and Tx-67 are shown in
Figure 1. In this new taxane, the introduction of the
succinate removed apparent Pgp interactions, despite
the fact that the molecule has other type I and type II
units. This could be due to the fact that the interactions
at the C10 position are particularly important for Pgp
interaction.
Paclitaxel (Taxol) is one of the most active cancer
chemotherapeutic agents known.1-5 At present, it is
commonly used for the treatment of breast and ovarian
cancers. Taxol causes cells to arrest at the G2/M phase
of the cell cycle through drug-induced tubulin polym-
erization and microtubule stabilization, thus leading to
induced cell apoptosis.6-9
Previous studies have established that Taxol is not
significantly absorbed across the gastrointestinal epthe-
lium after oral administration10,11 and does not cross
the blood-brain barrier (BBB).12-14 A primary mecha-
nism limiting Taxol distribution into the brain is active
efflux by the multidrug resistant gene product 1 (MDR1)
or P-glycoprotein (Pgp) localized on the blood side of the
microcerebrovascular endothelium comprising the BBB.
Permeation of Taxol across these tissues may be sub-
stantially improved by inhibiting Pgp-mediated ef-
flux.13,14
Ojima et al.15 have observed that there is a specific
binding site for taxoids on Pgp. Accordingly, by modify-
ing specific moieties on the Taxol molecule, new taxanes
might be identified with reduced binding or recognition
by Pgp and potentially enhanced permeability proper-
* Corresponding author. Phone: 785-864-3591. Fax: 785-864-5265.
E-mail: audus@ku.edu.
† Department of Pharmaceutical Chemistry.
‡ Department of Medicinal Chemistry.
§ Department of Pharmacology and Toxicology.
| Department of Molecular Biosciences.
10.1021/jm040114b CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/05/2005

Modification of Taxol Reduces Interactions with Pgp
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3 833
37.6, 43.6, 46.8, 56.6, 58.1, 71.8, 73.2, 74.4, 75.4, 75.5, 79.3,
81.5, 84.6, 126.8, 127.4, 129.1, 129.2, 130.6, 132.1, 134.0, 134.4,
136.5, 138.5, 138.7, 167.3, 167.4, 170.6, 171.7, 210.2; HRMS
(FAB) m/z calcd for C57H78NO13Si2 (M + 1) 1040.5012, found
1040.5019.
[
14C]-10-O-Succinyl Mono Ester of 10-O-Deacetyl-
paclitaxel ([14C]TX-67). Step 1. Using dry toluene (3 × 0.5
mL), commercially available 1,4-14C-succinic anhydride (spe-
cific activity 50 mCi/mmol, approximately 1.0-0.4 mg, 0.010-
0.004 mmol) was transferred into a 10 mL round-bottom
reaction flask containing a stirring bar. To this solution were
added 2′-O-(tert-butyldimethylsilyl)-10-O-deacetyl-7-O-trieth-
ylsilylpaclitaxel (2, 62 mg, 0.0596 mmol) and DMAP (4.9 mg,
0.04 mmol). The reaction vial was capped and the reaction was
heated at 85-90 °C for 20 h until the completion of the reaction
(monitored by TLC). After cooling to room temperature, the
reaction mixture was placed into a top-screwed test tube
containing CH2Cl2 (5 mL) and aqueous HCl (0.2%, 5 mL). The
reaction vessel was washed with additional CH2Cl2 (2 mL),
which was also collected in the test tube. The aqueous layer
was removed and extracted twice with CH2Cl2 (2 mL each).
The aqueous layer was discarded. The organic layer was dried
over sodium sulfate. The solvent was then removed and the
residue was purified by column chromatography using silica
gel (3% methanol in chloroform) to yield 3.9 mg of the acylated
reaction product.
Figure 1. Chemical structures for Taxol and Tx-67.
Scheme 1
Step 2. The acylated reaction product was transferred into
a Teflon vial using CH2Cl2 (5 × 1 mL). The solvent was
removed slowly from the vial by a flow of argon. The residue
was dissolved in pyridine (1.0 mL) and cooled to 0 °C. After
the addition of six drops of HF/pyridine, the ice bath was
removed and the mixture was stirred at 25 °C for 5 h. The
reaction mixture was then placed into a top-screwed test tube
containing CH2Cl2 (5 mL) and 2% aqueous HCl (5 mL). The
aqueous layer was discarded. The organic layer was washed
twice with aqueous HCl (2%, 2 × 5 mL) and then dried over
magnesium sulfate. Removal of the solvent under reduced
pressure, followed by drying on a vacuum pump overnight,
yielded 2.9 mg (specific activity 42 mCi/mmol) of the target
compound [14C]Tx-67.
Tx-67 was first prepared using the conditions described
above with nonradioactive succinic anhydride. All spectroscopic
data for Tx-67, prepared under those conditions, were identical
to those previously reported for Tx-67.17 The radioactive
product [14C]Tx-67 was compared to TX-67 using thin-layer
chromatography and several different solvent systems.
Materials and Methods
[14C]Sucrose (specific activity 401 mCi/mmol), [3H]Taxol
(specific activity 10.5 Ci/mmol), [3H]sucrose (specific activity
40 mCi/mmol), cell culture medium, 100-mm culture dishes,
0.4-µm-pore polycarbonate filters (13 mm diameter), cyclospor-
in A (CsA), penicillin G, amphotericin B, endothelial cell
growth supplement, and fibronectin were all commercially
available.
Synthesis of [14C]TX-67. The radioactive C10 monosucci-
nate [14C]TX-67 was prepared in two steps (Scheme 1) from
2′-O-(tert-butyldimethylsilyl)-10-deacetylpaclitaxel (1).16 Selec-
tive silylation of 1 at O7 with triethylsilyl chloride provided
2′,7-bis-silylated intermediate 2, which was reacted with 1,4-
14C-succinic anhydride, followed by desilylation to yield [14C]-
TX-67. In this reaction, a large excess of the taxane relative
to the 14C-labeled succinic anhydride was used because of the
high cost of the radioactive reagent. [14C]TX-67 (2.9 mg) was
prepared with a specific activity of 42 mCi/mmol.
2′-O-(tert-Butyldimethylsilyl)-10-O-deacetyl-7-O-tri-
ethylsilylpaclitaxel (1). To a solution of 2′-O-(tert-butyldi-
methylsilyl)-10-O-deacetylpaclitaxel16 (0.025 g, 0.027 mmol) in
CH2Cl2 (1.0 mL) were added DMAP (0.033 g, 0.27 mmol) and
chlorotriethylsilane (0.045 mL, 0.27 mmol). The mixture was
stirred at room temperature for 5 h and then diluted with CH2-
Cl2 (10 mL). This mixture was washed with a saturated NH4-
Cl solution (5.0 mL) and water (5.0 mL). The organic layer
was dried over Na2SO4. Removal of solvent followed by flash
chromatography on silica gel (EtOAc-hexane, 1:3) provided
the title compound as a colorless solid (0.024 g, 89% yield):
1H NMR (300 MHz, CDCl3) δ -0.26 (s, 3H, SiCH3), -0.02 (s,
3H, SiCH3), 0.50-0.62 (m, 6H, Si(CH2CH3)3), 0.80 (s, 9H, tert-
butyl), 0.94 (t, J ) 7.8 Hz, 9H, Si(CH2CH3)3), 1.10 (s, 3H, H17),
1.22 (s, 3H, H16), 1.76 (s, 3H, H19), 1.93 (s, 3H, H18), 1.92-
2.14 (m, 2H, H6), 2.3 and 2.56 (m, 2H, H14), 2.58 (s, 3H, 4-Ac),
3.91 (d, J ) 6.9 Hz, 1H, H3), 4.23 (d, J ) 8.1 Hz, 1H, H20),
4.30 (d, J ) 1.8 Hz, 1H, 10-OH), 4.35 (d, J ) 8.1 Hz, 1H, H20),
4.42 (dd, J ) 6.6 and 10.8 Hz, 1H, H7), 4.68 (d, J ) 2.1 Hz,
1H, H2′), 4.98 (dd, J ) 1.5 and 9.3 Hz, 1H, H5), 5.13 (d, J )
1.8 Hz, 1H, H10), 5.69 (d, J ) 6.9 Hz, 1H, H2), 5.73 (dd, J )
1.8 and 9 Hz, 1H, H3′), 6.34 (t, J ) 8.7 Hz, 1H, H13), 7.11 (d,
J ) 9 Hz, 1H, NH), 7.33-8.16 (m, 15H); 13C NMR (75 MHz,
CDCl3) δ 5.6, 7.2, 10.5, 14.7, 18.5, 21.4, 23.5, 25.9, 27.0, 36.3,
Analytical data for Tx-67, 10-O-Deacetylpaclitaxel 10-
1
monosuccinyl ester: H NMR (300 MHz, CDCl3) δ 1.12 (s,
3H), 1.25 (s, 3H), 1.64 (s, 3H), 1.78 (s, 3H), 1.81-1.87 (m, 1H),
2.20-2.72 (m, 7H), 2.35 (s, 3H), 3.76 (d, J ) 6.9 Hz, 1H), 4.20
and 4.27 (d, J ) 8.4 Hz, 2H), 4.36 (dd, J ) 6.9 and 10.5 Hz,
1H), 4.78 (d, J ) 3 Hz, 1H), 4.92 (d, J ) 8.7 Hz, 1H), 5.65 (d,
J ) 6.9 Hz, 1H), 5.75 (dd, J ) 2.4 and 8.7 Hz, 1H), 6.19 (t, J
) 9.0 Hz, 1H), 6.25 (s, 1H), 7.18 (d, J ) 9.0 Hz, 1H), 7.27-
8.11 (m, 15H); 13C NMR (75 MHz, CDCl3) δ 9.6, 14.7, 21.0,
22.6, 28.9, 29.0, 35.6, 43.0, 45.7, 55.1, 58.4, 72.0, 72.1, 73.1,
75.0, 75.8, 78.9, 81.1, 84.4, 127.0, 127.1, 128.2, 128.6, 128.7,
128.9, 129.2, 130.2, 131.9, 133.0, 133.6, 138.0, 142.2, 166.8,
167.5, 168.0, 170.3, 172.3, 172.6, 203.5; HRMS m/z calcd for
C49H54NO16 (M + 1) 912.3364, found 912.3350. The purity was
verified as 96-99% by HPLC. HPLC conditions were as
follows: Vydac C18 column (5 µm, 10 × 250 mm), UV detection
at 280 nM, and gradient elution with 0.1% aqueous trifluor-
acetic acid and 0-100% acetonitrile in 0.1% trifluoroacetic acid
for 40 min. Retention time was 24.8 min.
Cell Culture. Bovine brain microvessel endothelial cells
(BMECs) were isolated and grown in primary culture and
characterized as detailed by Audus et al.18,19
Rhodamine 123 Uptake Assay. BMECs were seeded onto
12-well cluster dishes at a density of 50 000 cells/cm2. The
culture medium was changed every other day after seeding
until a confluent monolayer was formed as determined by light
microscopy. Experiments were performed in phosphate-
buffered saline supplemented with calcium and glucose (PBSA),

834 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3
Rice et al.
pH 7.4, essentially as detailed earlier by Rose et al.20 Briefly,
the growth medium was first aspirated off and then the cells
were rinsed three times with prewarmed (37 °C) PBSA. The
monolayers were then either equilibrated in 1 mL of PBSA
for 1 h at 37 °C for control experiments or equilibrated for 30
min in PBSA alone, followed by another 45 min preincubation
with the a taxane. Cyclosporin A (CsA; 10 µM) was used as a
known Pgp inhibitor and as the positive control. Rhodamine
123 accumulation was then performed for 45 min with or
without a potential inhibitor present with gentle agitation
(∼30 rpm) at 37 °C. At the end of the experiment, the drug
solution was removed by aspiration, and the monolayers were
immediately rinsed three times with ice-cold PBSA. Each
monolayer was solubilized for 30 min (37 °C) with 1 mL of
lysing solution (0.5% v/v Triton X-100 in 0.2 N NaOH). Cell
lysates were assayed using a microplate fluorescence reader
(Bio-Tek Instruments, Winooski, VT) at excitation/emission
wavelengths of 485 nm/520 nm and then quantified against
standard curves of rhodamine 123 in lysing solution. The
fluorescence of the cell lysates was corrected for autofluores-
cence of untreated cells. The protein content of each monolayer
was then determined using the BCA protein assay reagent kit.
Results were expressed as total moles of rhodamine 123
accumulation per microgram of cellular protein.
and Use Committee. Adult male Sprague-Dawley rats (350-
400 g) were anesthetized with a mixture containing 1.5 mL/
kg of solution consisting of 37.5 mg/mL ketamine, 1.9 mg/mL
xylazine, and 0.37 mg/mL acepromazine. A full dosage was
given for the initial anesthesia and one-third of the full dose
given for repeated dose at 1-h intervals. After the rat was
anesthetized, it was placed on a warm-pad and the tempera-
ture was monitored by inserting a thermometer into the
rectum, and body temperature was maintained at 37 °C.
Briefly, either the right or left common carotid artery was
surgically exposed and cannulated for delivery of perfusion
fluid (0.9% saline) to the brain. The external carotid was
ligated to reduce the amount of perfusion fluid distribution to
extracerebral tissues.22 An individual rat was then perfused
with the buffer containing a tracer (sucrose) and the sample
(Taxol or Tx-67) for specific time intervals (i.e., for 30, or 60,
or 120 s). After the perfusion, the rat was decapitated and the
brain was removed for sampling of regions. The brain samples
were digested in Solvable for 24 h, and the radioactivity was
quantified via liquid scintillation spectrometry. The capillary
permeability-surface area product (PA) (mL/s/g) was calcu-
lated for both Taxol and Tx-67 by the following equation:
PA ) -F ln [1 - Cbr(T)/FTCpf]
Permeation of Taxanes across BBMEC Monolayers.
BMECs were grown on 0.4 µm polycarbonate membranes.
Upon confluency, the cells were transferred to side-by-side
diffusion chambers as detailed by Audus et al.18,19 to charac-
terize the transport of [3H]Taxol or [14C]Tx-67. Transport
studies were performed in pH 7.4 standard buffer solutions,
consisting of either Hank’s balanced salt solution (HBSS) or
phosphate-buffered saline (PBS) supplemented with 0.63 mM
CaCl2, 0.74 mM MgSO4, 5.3 mM glucose, and 0.1 mM ascorbic
acid (PBSA). Briefly, all studies were performed in 3 mL of
PBSA in each donor and receiver chamber of the side-by-side
diffusion chambers and stirred at 600 rpm at 37 °C. The cells
were allowed to equilibrate in PBSA for 30 min prior to each
experiment and oriented such that the apical or blood side of
the cells face the donor chamber and the basolateral or brain
side of the cells adhere to the polycarbonate membranes and
face the receiver chamber of the diffusion apparatus. At each
time point, a 100 µL sample was taken from the receiver
compartment and immediately replaced with an equal volume
of PBSA. The transport studies were performed in the presence
or absence of CsA (5-10 µM). Monolayers were checked for
trypan blue exclusion after experiments as well to assess
general cell viability. The permeability of all monolayers used
in these experiments was monitored for [14C] sucrose leakage
as a measure of potential taxane-induced loss of integrity. The
radioactivity was quantified using liquid scintillation spec-
trometry. Apparent permeability coefficients (Papp) were cal-
culated as detailed by Adson et al.21
Determination of Concentration Dependence for Tax-
ol and Tx-67 Efflux by Pgp Efflux Pump. Cells were grown
and transport studies were set up as described in the previous
section. Permeation studies were performed in the presence
of ranges of concentrations of [3H]Taxol (1 nM to 25 µM) and
[14C]Tx-67 (1-250 nM). Sucrose transport was determined for
each permeation study. Transport studies were performed for
2 h. The radioactivity was quantified using liquid scintillation
spectrometry. Apparent permeability coefficients (Papp) were
calculated.
In other experiments, permeation studies were performed
to determine if Tx-67 transport was inhibited by P-glycoprotein
inhibitors. The bidirectional flux was evaluated for [14C]Tx-
67 in the presence of unlabeled Taxol (25 µM) and CSA (5-10
µM).
where F is the regional cerebral blood flow (mL/s/g), Cbr
represents the concentration of tracer in the brain parenchyma
(dpm/g), Cpf is the concentration in the perfusion fluid (dpm/
mL), and T is the total perfusion time.22
Results
Effects of Taxol and Tx-67 on Rhodamine 123
Uptake. The potential interactions of Taxol and Tx-67
with P-glycoprotein, relative to a positive control, CsA,
were investigated by assessing rhodamine 123 uptake
by BMECs in the presence and absence of the agents.
As shown in Figure 2A, Taxol exposure enhanced the
BMEC uptake of the P-glycoprotein substrate rhodamine
123. Similarly, in Figure 2B, the positive control CsA
enhanced rhodamine 123 uptake by BMECs. In con-
trast, Tx-67 had no effect on rhodamine 123 uptake by
BMECs, suggesting an absence of interactions with
P-glycoprotein.
Bidirectional Permeation of Taxol and Tx-67
across BMEC Monolayers. To more directly examine
the interactions of Taxol and Tx-67 with the P-glyco-
protein directly, the concentration-dependent perme-
ation of Taxol and Tx-67 across BMEC monolayers was
investigated for asymmetric, bidirectional permeation
that is typical for P-glycoprotein substrates. The apical
to basolateral permeation of 1 nM Tx-67 across BMEC
monolayers was approximately 9.7 ( 0.4 × 10-5 cm/s
and was higher than the inert background permeation
marker sucrose at 5.3 ( 0.8 × 10-5 cm/s. The apical to
basolateral permeation of 1 nM Taxol across BMEC
monolayers was approximately 0.88 ( 0.9 × 10-5 cm/s
and was significantly lower than the inert background
permeation marker sucrose. Figure 3A shows the per-
meation curves with increasing concentrations of Taxol
in the side-by-side diffusion chamber. At lower concen-
trations of Taxol (∼50-100 nM), the permeability rates
were greater for the basolateral to apical direction. At
higher concentrations of Taxol, the asymmetry of per-
meation across the monolayers was lost. Over the
concentration range examined, increasing the concen-
tration of Taxol appeared to saturate the transport
mechanism and eventually resulted in increased ef-
ficiency of permeation (i.e., increased permeability coef-
ficients).
In Situ Rat Brain Perfusion of Taxol and Tx-67. Adult
male Sprague-Dawley rats, weighing 250-350 g, were ob-
tained from Sasco. The in situ rat brain perfusion technique
originally described by Smith and Allen22 was employed to
determine the BBB permeability of Taxol and Tx-67. Animal
subjects were maintained and used in these studies with full
approval of the University of Kansas Institutional Animal Care

Modification of Taxol Reduces Interactions with Pgp
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3 835
Figure 2. Rhodamine 123 uptake alone and in the presence
of (A) Taxol and (B) cyclosporin A and Tx-67 (B). Bars
represent the mean ( standard deviation for n ) 3 different
monolayers. *Significant difference, p < 0.05, compared to
rhodamine uptake in the absence of any treatment (control)
as determined by ANOVA.
Figure 4. Effect of 10 µM cyclosporine A (CsA) (open bars)
on the apparent permeability coefficients (Papp) for (A) 10 nM
[3H]Taxol (solid bars) and (B) 1 µM [3H]Taxol (solid bars)
permeation across brain microvessel endothelial cell mono-
layers. Data points represent the mean ( standard deviation
for n ) 3 different monolayers. *Significant difference, p <
0.05, compared to Taxol alone as determined by ANOVA.
ation across the monolayers was greater than that of
Taxol over the entire concentration range examined for
the agent. Increasing concentrations of Tx-67 appeared
to result in a decreased permeation efficiency (i.e.,
decreased permeability coefficients) and suggested in-
teractions with a rate-limiting mechanism, possibly
another transporter.
Sucrose permeability was not affected by the increas-
ing concentration of Taxol or Tx-67 at the concentrations
used in Figure 3 (data not shown). Additionally, mono-
layers showed trypan blue exclusion at the conclusion
of the transport studies, further confirming that the cells
remain viable under the experimental conditions (data
not shown).
Pgp has been localized to the apical or blood side of
brain microvessel endothelium.23 To verify in our system
that an apically located mechanism consistent with Pgp
was involved, we treated the cell monolayers with the
Pgp inhibitor CsA. Treatment of BMEC monolayers
with the Pgp inhibitor CsA (10 µM) increased the
permeation of 10 nM Taxol in the apical to basolateral
direction when added to the donor chamber, as shown
in Figure 4A, or both donor and receiver chambers (data
not shown) during the transport experiments. No sig-
nificant increase in Taxol permeation was observed in
the apical to basolateral direction when CsA was added
only to the receiver chamber (data not shown). Taxol
permeation was not significantly altered in the baso-
lateral to apical direction when CsA was added either
Figure 3. Concentration-dependence of apparent perme-
ability coefficients for the bidirectional permeation of (A) [3H]-
Taxol and (B) [14C]Tx-67 across brain microvessel endothelial
cell monolayers: - [ -, apical to basolateral; -9-, basolateral
to apical. Data points represent the mean ( standard deviation
for n ) 3 different monolayers.
Figure 3B shows that in the lower concentration
range of Tx-67 (<50 nM) permeation was greater in the
apical to basolateral direction. Moreover, Tx-67 perme-

836 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3
Rice et al.
Table 1. Apparent Permeability Coefficients for [14C]Tx-67 and
[3H]Taxol in the in Situ Rat Brain Perfusion Techniquea
Papp × 107 cm/s
compound
14C]Tx-67
[3H]Taxol
30 s
60 s
120 s
[
8.47
0.845
13.71
1.700
10.75
1.574
a Time points were for 30, 60, and 120 s (each time point was
done in duplicate, n ) 2).
Figure 6. Cerebrovascular permeability-surface area product
(PA) for [3H]Taxol (open bars) and [14C]Tx-67 (solid bars), as
measured by the brain perfusion technique (60 s data). Data
points represent the mean ( standard deviation for n ) 3
different rats. *Significant difference, p < 0.05, compared to
Taxol as determined by ANOVA.
Figure 5. Effect of (A) 10 µM cyclosporine A (CsA; open bars)
and (B) 25 µM Taxol (open bars) on the apparent permeability
coefficient (Papp) for the permeation of 25 nM [14C]Tx-67 (solid
bars) across brain microvessel endothelial cell monolayers.
Data points represent the mean ( standard deviation for n )
3 different monolayers. *Significant difference, p < 0.05,
compared to Taxol alone as determined by ANOVA.
In Situ Rat Brain Perfusion Studies with Tx-67
and Taxol. To verify the in vitro observations regarding
enhanced BBB permeation of Tx-67, in situ perfusion
experiments were performed with [14C]Tx-67 (300 nM
or 0.03 mg/kg) and [3H]Taxol (100 nM or 0.01 mg/kg)
at the lowest concentrations allowed by the specific
activity of the agents. As summarized in Table 1, we
observed a greater Papp for Tx-67 than Taxol at each
perfusion time. Figure 6 also shows in a separate series
of studies, with at least three animals sampled for each
brain region, significant increases in cerebrovascular
permeability-surface area products (PAs) determined
from 60 s perfusion data for brain uptake in selected
regions of the brain where the BBB generally restricts
the free passage of blood-borne agents. The data sug-
gested that Taxol was not retained in the rat brain with
increasing time; only minimal amounts (2.5% of total
control) were detectable in various regions of the brain.
Approximately 6-8% of the total control was retained
in the brain for Tx-67, suggesting that this chemically
modified taxane was able to cross the BBB.
to the donor compartment, as illustrated in Figure 4A,
or to both compartments (data not shown). These effects
were consistent with an apical localization of Pgp and
influence on Taxol transport.
Figure 4B shows that treatment of BMEC monolayers
with the Pgp inhibitor CsA (10 µM) in the donor
chamber of the diffusion apparatus had no effect on the
permeation of 1 µM Taxol across BMEC monolayers in
either the apical to basolateral direction or the baso-
lateral to apical direction. Addition of CsA to either
donor and receiver or receiver alone also was not
effective in altering the permeation of the higher
concentration of Taxol across BMEC monolayers. These
results were consistent with high concentrations of
Taxol saturating Pgp and an absence of effects of CsA
on permeation.
We investigated the permeation properties of Tx-67
in the presence of CsA and Taxol to further examine
the interaction of Tx-67 with Pgp. Figure 5A shows the
absence of effects of CsA in the donor chamber of the
diffusion apparatus on the permeation of 25 nM Tx-67
across BMEC monolayers in both apical to basolateral
and basolateral to apical directions. Similarly, Figure
5B shows the absence of effects of Taxol in the donor
chamber of the diffusion apparatus on the permeation
of 25 nM Tx-67 across BMEC monolayers in both apical
to basolateral and basolateral to apical directions. These
results were consistent with a lack of influence of Pgp
on Tx-67 permeation across BMEC monolayers.
Discussion
We have previously demonstrated that Pgp is ex-
pressed in primary cultures of BMECs.20 Other re-
searchers have demonstrated that Taxol does not cross
the endothelial lining of the in vivo BBB due to the
presence of Pgp12-14 localized on the blood-facing sur-
face.23 A potential approach to overcoming the BBB and
achieving delivery of Taxol into the brain is to inhibit
Pgp, a strategy that has been well demonstrated by
Fellner et al.13 and Kemper et al.14 Analogues of CSA
as inhibitors of Pgp, for instance, enhance brain Taxol
levels by about 3-fold in mice and rats.13,14 An alterna-

Modification of Taxol Reduces Interactions with Pgp
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3 837
tive approach is to introduce chemical modifications that
reduce interactions of Taxol with Pgp to achieve the
same goal of enhanced brain levels. Examples in which
this latter approach may be productive have been
reported for Taxol derivatives in both Pgp-expressing
cells in vitro and in drug resistant tumors in mice.24-27
Thus, our goal was to identify a taxane that could avoid
Pgp at the BBB and achieve enhanced brain levels.
Tx-67 is a Taxol derivative that has not been examined
in any previous study for interactions with Pgp in cell
or animal models.
Significant numbers of taxanes were generated by
combinatorial chemistry techniques17 and tested for
cytotoxicity and interactions with Pgp by a rhodamine
123 assay. We were able to identify select taxanes such
as Tx-67 which retained cytotoxic effects against the
breast cancer cell line MCF7 with an IC50 of 37.3 nM,
an efficacy that was comparable to Taxol, with an IC50
of 1.8 nM in the MCF7 cell line.17 In this study, Tx-67
exhibited no effect on rhodamine 123 uptake, suggesting
no interactions with Pgp. Since the rhodamine 123 assay
was an indirect indicator of Tx-67 interactions with Pgp,
further studies were conducted to look at the trans-
cellular transport of Tx-67 relative to Taxol both in vitro
and in situ.
Our data support the previous findings that Pgp is
localized to the apical side of the BMECs,23 as indicated
by the asymmetric permeation of Taxol across the
monolayers. Taxol permeation across BMEC monolayers
was saturable and increased by the typical Pgp inhibitor
CsA.28 Moreover, results indicated that Taxol perme-
ation was increased more effectively when CsA was
applied to the apical surface of the monolayers. These
observations are entirely consistent with the behavior
of Taxol in other recent studies demonstrating the role
of Pgp in limiting access of CsA or its analogues into
the brain.12-14
In other recent studies, colleagues have provided
pharmacological evidence that Tx-67 but not Taxol is
able to enter brain tissue. Li et al.30 have shown that
Taxol stabilized cyclin-dependent kinase 5 activator
(cdk5) (p35) and decreased â-amyloid peptide toxicity
(Aâ25-35 and Aâ1-42) in cortical neuron cultures gener-
ated from untreated mice. However, when mice were
acutely pretreated with equivalent amounts (i.e., ad-
ministered i.p. 8 mg/kg every 2 days for 16 days) of
either Taxol or Tx-67, Aâ-induced cdk5 activation in
subsequently dissociated cortical neuron cultures was
blocked only in the neurons from the Tx-67-pretreated
animals and not from the Taxol-pretreated animals.
These findings imply that the efficacy of Tx-67 in mice
was due to improved BBB penetration and entry into
neuronal tissue, which provided the block of attempted
cdk5 activation in the dissociated neurons in culture by
Aâ peptides.30
In summary, our studies have demonstrated that a
taxane analogue, Tx-67, can be generated with reduced
interactions with Pgp at the BBB. Tx-67 appears to
cross the BBB both in vitro and in situ by avoiding Pgp
interactions, perhaps by exploiting an alternative influx
transport mechanism that facilitates passage movement
of the agent across the endothelium. As is supported
by reports for absorption and distribution into tumor
cells,24-27 our Tx-67 permeation studies are consistent
with a hypothesis that the approach of incorporating
chemical modifications into Pgp substrates, e.g.,
taxanes, can enhance permeation across the BBB.
Acknowledgment. This work was supported by a
grant from the National Cancer Institute (1 RO1
CA82801) and the Department of the Army through a
Research Predoctoral Fellowship (DAMD17-99-1-92430)
to Y.L.
References
In contrast to Taxol, the permeation rates for Tx-67
exceeded those of the parent agent at all concentrations
examined in vitro. In fact, the asymmetric permeation
of Tx-67 across BMEC monolayers favored the apical-
to-basolateral passage, even at low concentrations where
Taxol was subject to the influence of Pgp. Taxol and the
typical Pgp inhibitor, CsA, failed to affect Tx-67 per-
meation across BMEC monolayers as further evidence
that the influence of efflux mechanisms was substan-
tially reduced and or absent completely.
(1) Taxus: The genus Taxus. In Medicinal and Aromatic Plantss
Industrial Profiles; Itokawa, H., Lee, K.-H., Eds.; Taylor &
Francis: London, 2003.
(2) Kingston, D. G. I.; Jagtap, P. G.; Yuan, H.; Samala, L. The
Chemistry of Taxol and Taxol Related Taxoids. In Progress in
the Chemistry of Natural Products; Herz, W., Falk, H., Kirby,
G. W., Eds.; Springer: Wien, 2002; pp 56-225.
(3) Taxane Anticancer Agents: Basic Science and Current Status;
Georg, G. I., Chen, T. T., Ojima, I., Vyas, D. M., Eds.; ACS
Symposium Series 583; American Chemical Society: Washing-
ton, DC, 1995.
(4) Taxol Science and Applications; Suffness, M., Ed.; CRC: Boca
Raton, FL, 1995.
(5) The Chemistry and Pharmacology of Taxol and Its Derivatives;
Farina, V., Ed.; Elsevier: Amsterdam, 1995.
The in situ rat brain perfusion data correlated well
with the in vitro permeability data. Our Taxol data was
also comparable to other perfusion data typical of other
anticancer drugs, such as vincristine and vinblastine,
which penetrate the BBB poorly due in part due to the
presence of Pgp.29 However, the PAs and Papps for
Tx-67 were substantially improved over Taxol perme-
ation into all rat brain regions sampled. The Tx-67 data
were consistent with other reports in which certain
newly synthesized taxane analogues have increased
absorption and decreased interactions with P-glyco-
protein in absorption models and tumor cell lines when
modifiedinC7andC10positionsontheTaxolstructure.24-27
However, that we are aware, this study is the first to
demonstrate a C10-modified taxane with improved
distribution across the BBB without the coadministra-
tion of Pgp inhibitors.
(6) Jordan, M. A.; Toso, R. J.; Thrower, D.; Wilson, L. Mechanism
of mitotic block and inhibition of cell proliferation by Taxol at
low concentrations. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 9552-
9556.
(7) Schiff, P. B.; Fant, J.; Horwitz, S. B. Promotion of microtubule
assembly in vitro by Taxol. Nature 1979, 277, 665-667.
(8) Parness, J.; Horwitz, S. B. Taxol binds to polymerized tubulin
in vitro. J. Cell Biol. 1981, 91, 479-487.
(9) Blagosklonny, M. V.; Fojo, T. Molecular effects of paclitaxel:
Myths and reality (a critical review). Int. J. Cancer 1999, 83,
151-156.
(10) Eiseman, J. L.; Eddington, N. D.; Leslie, J.; MacAuley, C.; Sentz,
D. L.; Zuhowski, M.; Kujawa, J. M.; Young, D.; Egorin, M. J.
Plasma pharmacokinetics and tissue distribution of paclitaxel
in CD2F1 mice. Cancer Chemother. Pharmacol. 1994, 34, 465-
471.
(11) Anderson, C. D.; Wang, J.; Kumar, G. N.; McMillan, J. M.; Walle,
U. K.; Walle, T. Dexamethasone induction of Taxol metabolism
in the rat. Drug. Metab. Dispos. 1995, 23, 1286-1290.
(12) van Asperen, J., Mayer, U., van Tellingen, O., Beijnen, J. H.
The functional role of P-glycoprotein in the blood-brain barrier.
J. Pharm. Sci. 1997, 86, 881-884.

838 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 3
Rice et al.
(13) Fellner, S.; Bauer, B.; Miller, D. S.; Schaffrik, M.; Frankhanel,
M.; Sprub, T.; Berhardt, G.; Graeff, C.; Farber, L.; Gschaidmeier,
H.; Buschauer, A.; Fricker, G. Transport of paclitaxel (Taxol)
across the blood-brain barrier in vitro and in vivo. J. Clin.
Invest. 2002, 110, 1309-1318.
(22) Smith, Q. R.; Allan, D. D. In situ brain perfusion technique. In
The Blood-Brain Barrier; Nag, S., Ed.; Humana Press: Totowa,
NJ, 2003; pp 209-218.
(23) Bendayan, R.; Lee, G.; Bendayan, M. Functional expression and
localization of P-glycoprotein at the blood-brain barrier. Microsc.
Res. Technol. 2002, 57, 365-380.
(14) Kemper, E. M.; van Zandbergen, A. E.; Cleypool, C.; Mos, H. A.;
Boogerd, W.; Beijnen, J. H.; van Tellingen, O. Increased penetra-
tion of paclitaxel into the brain by inhibition of P-glycoprotein.
Clin. Cancer Res. 2003, 9, 2849-2855.
(24) Ojima, I.; Slater, J. C.; Michaud, E.; Kuduk, S. D.; Bounaud, P.
Y.; Vrignaud, P.; Bissery, M. C.; Veith, J. M.; Pera, P.; Bernacki,
R. J. Syntheses and structure-activity relationships of the
second-generation antitumor taxoids: Exceptional activity against
drug-resistant cancer cells. J. Med. Chem. 1996, 39, 3889-3896.
(25) Ojima, I.; Bounaud, P. Y.; Ahern, D. G. New photoaffinity
analogues of paclitaxel. Bioorg. Med. Chem. Lett. 1999, 9, 1189-
1194.
(26) Vredenburg, M. R.; Ojima, I.; Veith, J.; Pera, P.; Kee, K.; Cabral,
F.; Sharma, A.; Kanter, P.; Greco, W. R.; Bernacki, R. J. Effects
of orally active taxanes on P-glycoprotein modulation and colon
and breast carcinoma drug resistance. J. Natl. Cancer Inst. 2001,
93, 1234-1245.
(27) Shionoya, M.; Jimbo, T.; Kitagawa, M.; Soga, T.; Tohgo, A. DJ-
927, a novel oral taxane, overcomes P-glycoprotein-mediated
multidrug resistance in vitro and in vivo. Cancer Sci. 2003, 94,
459-466.
(28) Cordon-Cardo, C.; O’Brien, J. P.; Casals, D.; Rittman-Grauer,
L.; Biedler, J. L.; Melamed, M. R.; Bertino, J. R. Multidrug-
resistance gene (P-glycoprotein) is expressed by endothelial cells
at blood-brain barrier sites. Proc. Natl. Acad. Sci. U.S.A. 1989,
86, 695-698.
(29) Greig, N. H.; Soncrant, T. T.; Shetty, H. U.; Momma, S.; Smith,
Q. R.; Rapoport, S. I. Brain uptake and anticancer activities of
vincristine and vinblastine are restricted by their low cere-
brovascular permeability and binding to plasma constituents in
rat. Cancer Chemother. Pharmacol. 1990, 26, 263-268.
(30) Li, G.; Faibushevich A.; Turunen, B. J.; Yoon, S. O.; Georg, G.;
Michaelis, M. L.; Dobrowsky, R. T. Stabilization of the cyclin-
dependent kinase 5 activator, p35, by paclitaxel decreases
â-amyloid toxicity in cortical neurons. J. Neurochem. 2003, 84,
347-362.
(15) Ojima, I.; Duclos, O.; Dorman, G.; Simonot, B.; Prestwich, G.
D.; Rao, S.; Lerro, K. A.; Horwitz, S. B. A new paclitaxel
photoaffinity analogue with
a 3-(4-benzoylphenyl)propanoyl
probe for characterization of drug-binding sites on tubulin and
P-glycoprotein. J. Med. Chem. 1995, 38, 3891-3894.
(16) Datta, A.; Hepperle, M.; Georg, G. I. Selective deesterification
studies on taxanes: Simple and efficient hydrazinolysis of C-10
and C-13 Ester Functionalities. J. Org. Chem. 1995, 60, 761-
763.
(17) Liu, Y.; Boge, T. C.; Victory, S.; Ali, S. M.; Zygmunt, J.; Georg,
G. I.; Marquez, R. T.; Himes, R. H. A systematic SAR study of
C10 modified paclitaxel analogues using a combinatorial ap-
proach. Combi. Chem. High Through. Screen 2002, 5, 39-48.
(18) Audus, K. L.; Ng, L.; Wang, W.; Borchardt, R. T. Brain microves-
sel endothelial cell culture systems. In Model Systems for
Biopharmaceutical Assessment of Drug Absorption and Metabo-
lism; Borchardt, R. T., Smith, P. L., Wilson, G., Eds.; Plenum:
New York, 1996; pp 239-258.
(19) Audus, K. L.; Rose, J. M.; Wang, W.; Borchardt, R. T. Brain
microvessel endothelial cell culture systems. In An Introduction
to the Blood-Brain Barrier: Methodology and Biology;
Pardridge, W., Ed.; Cambridge University: Cambridge, 1998;
pp 86-93.
(20) Rose, J. M.; Peckham, S. L.; Scism, J. L.; Audus, K. L. Evaluation
of the role of P-glycoprotein in ivermectin uptake by primary
cultures of bovine brain microvessel endothelial cells. Neuro-
chem. Res. 1998, 23, 203-209, 1998.
(21) Adson, A.; Raub, T. J.; Burton, P. S.; Barsuhn, C. L.; Hilgers, A.
R.; Audus, K. L.; Ho, N. F. H. Quantitative approaches to
delineate paracellular diffusion in cultured epithelial cell mono-
layers. J. Pharm. Sci. 1994, 83, 1529-1536.
JM040114B