Single-site chemical modification at C10 of the baccatin III core of paclitaxel and Taxol C reduces P-glycoprotein interactions in bovine brain microvessel endothelial cells

Article abstract of DOI:10.1016/j.bmcl.2005.10.063

A single-site modification of paclitaxel analogs at the C10 position on the baccatin III core that reduces interaction with P-glycoprotein in bovine brain microvessel endothelial cells is described. Modification and derivatization of the C10 position were

Full text of DOI:10.1016/j.bmcl.2005.10.063

Bioorganic & Medicinal Chemistry Letters 16 (2006) 495–498
Single-site chemical modification at C10 of the baccatin III core of
paclitaxel and Taxol C reduces P-glycoprotein interactions
in bovine brain microvessel endothelial cells
Jared T. Spletstoser,a Brandon J. Turunen,a Kelly Desino,b Antonie Rice,b Apurba Datta,a
Dinah Dutta,a Jacquelyn K. Huff,c Richard H. Himes,c,* Kenneth L. Audus,b,*
Anna Seeligd,* and Gunda I. Georga,*
aDepartment of Medicinal Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence, Kansas 66045, USA
bDepartment of Pharmaceutical Chemistry, University of Kansas, Lawrence, Kansas 66047, USA
cDepartment of Molecular Biosciences, University of Kansas, Lawrence, Kansas 66045, USA
dDepartment of Biophysical Chemistry, Biocenter of the University of Basel, Klingelbergstrasse 70, 4056 Basel, Switzerland
Received 24 September 2005; revised 18 October 2005; accepted 19 October 2005
Available online 10 November 2005
Abstract—A single-site modification of paclitaxel analogs at the C10 position on the baccatin III core that reduces interaction with
P-glycoprotein in bovine brain microvessel endothelial cells is described. Modification and derivatization of the C10 po\sition were
carried out using a substrate controlled hydride addition to a key C9 and C10 diketone intermediate. The analogs were tested for
tubulin assembly and cytotoxicity, and were shown to retain potency similar to paclitaxel. P-glycoprotein interaction was examined
using a rhodamine assay and it was found that simple hydrolysis or epimerization of the C10 acetate of paclitaxel and Taxol C can
reduce interaction with the P-glycoprotein transporter that may allow for increased permeation of taxanes into the brain.
Ó 2005 Elsevier Ltd. All rights reserved.
Chemotherapy of CNS localized tumors has been limit-
ed by the inability of drugs to bypass the blood–brain
barrier (BBB) and little progress has been made since
the advent of the nitrosoureas.1 While primary CNS tu-
mors are relatively uncommon, metastatic tumors of the
brain outnumber primary lesions at least 10–1.2 There-
fore, the development of agents able to permeate the
BBB would provide clinicians with additional options
for the treatment of brain localized tumors, especially
those with multiple lesions.
Paclitaxel does not cross the blood–brain barrier and is
not effective against cancers localized in the cranial cav-
ity.5 The primary mechanism of decreased paclitaxel
BBB penetration is active efflux by P-glycoprotein
(Pgp), which is a membrane bound protein produced
by the multi-drug resistant gene cassette (MDR1). Pgp
knockout mice show nearly a fourfold increase in the
Type I
weak
Type II
Type I
Paclitaxel (1, Fig. 1) is a microtubule-binding chemo-
therapeutic agent that causes cell cycle arrest at the
G2/M phase leading to cellular apoptosis.3 Paclitaxel is
approved for the treatment of breast, ovarian, and
non-small cell lung cancers, Kaposiꢀs sarcoma, and is
active against a wide variety of cancer cell types.4
O
O
O
O
NH
O
OH
10
1`
13
Ph 3`
O
7
Type II
weak
2
OH
4
HO
H
BzO
AcO
O
Keywords:
P-glycoprotein; Blood–brain barrier.
Paclitaxel;
10-Deacetylpaclitaxel;
10-epi-Paclitaxel;
Type I
Type I
*
Corresponding authors. Tel.: +1 785 864 4498; fax: +1 785 864
5836; e-mail: georg@ku.edu
Figure 1. Paclitaxel (1) recognition elements for Pgp.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.063

496
J. T. Spletstoser et al. / Bioorg. Med. Chem. Lett. 16 (2006) 495–498
concentration of brain paclitaxel levels as compared to
normal mice.6 Paclitaxel has a relatively low toxicity to
primary cortical neurons and has been shown to protect
them from b-amyloid protein-induced toxicity, which
has been implicated in Alzheimerꢀs disease.7 Thus, if
paclitaxel could be modified to evade Pgp, it would pro-
vide a potential means for the treatment of brain-related
cancers and other CNS diseases.
tion-dependent manner.15 In the work presented here,
we sought to transform the type II interaction to a
weaker type I element via C10 ester hydrolysis or by
altering the spatial distance between the oxygen hetero-
atoms through C10 epimerization in paclitaxel and Tax-
ol C.16
Our group has developed an efficient route to C10 mod-
ified taxanes that utilizes a stereoselective hydride addi-
tion from the b-face of the baccatin III core to afford the
a-C10 epimeric products 10-epi-paclitaxel (5) and 10-
epi-10-deacetyl-paclitaxel 6 (shown in Scheme 1).17 Pro-
pionyl analog 4 was prepared similarly by reacting 3
with propionic anhydride, followed by HF-mediated
deprotection (Scheme 1).
A series of recognition elements required for Pgp inter-
action have been described recently by us.8,9 This analy-
sis identified clusters common to Pgp substrates that
consist of hydrogen bond acceptors organized in a de-
fined spatial arrangement, and their relative frequency
was correlated with the strength of interaction with
Pgp. Two motifs, termed type I and type II, were iden-
tified that possess different relative recognition affinities
for Pgp. We analyzed paclitaxel (Fig. 1) and identified a
strongly and a weakly interacting type II unit, three
standard type I units, and one weakly interacting type
I unit that consists of an electron-donating heteroatom
and a weak p-system in close proximity to one another.
The type II unit is considered to have the highest Pgp
affinity.
Selective protection of the C20 hydroxyl group of Taxol
C (2) as the silyl ether furnished 8 in good yield (Scheme
2). Hydrazine-mediated cleavage18 of the C10 acetate
provided the C10 hydroxyl analog 9, which was regiose-
lectively protected at the C7 hydroxyl group with trieth-
ylsilyl chloride to afford 10 in excellent yield.
Deprotection of the b-hydroxy taxane 9 provided the
10b-deacetyl-Taxol C analog 11 in 84% yield. Dess–
Martin periodinane oxidation of the C10 hydroxy group
in 10 occurred smoothly resulting in the C9–C10 dike-
tone 12.
Ojima et al. have demonstrated, using photoaffinity-la-
beling experiments, that Pgp contains a binding site spe-
cific for taxanes.10 They have also shown that the NCI/
ADR-RES (formerly named MCF-7ADR)11 cell line (a
multi-drug resistant breast cancer phenotype known to
overexpress Pgp) is very sensitive to substitution of the
C10 moiety in 30-dephenyl-30-isobutylpaclitaxel ana-
logs,12 although this effect was not observed when the
same modifications were made on paclitaxel.13 Since
the C9 carbonyl-C10 ester motif is a type II unit, we sur-
mised that it might be largely responsible for Pgp inter-
action. All Pgp inducers carry at least one type II unit
and it has been suggested by us that this unit interacts
more strongly with the efflux protein.8 Thus, we rea-
soned that the chemical modification of this motif
through changes of the C10 oxidation state and/or abso-
lute configuration might decrease the affinity of certain
taxanes for Pgp. This hypothesis is supported by the re-
cent report that TXD-258 (7-O-methyl-10-O-methyldo-
cetaxel) is able to cross the BBB.14 This analog has a
C10 –OMe group in place of the acetoxy group usually
found at this position in paclitaxel analogs (this alters
the C10 type II ester functionality). We have also found
that deletion of the C10 ester unit in certain paclitaxel
analogs decreases interaction with Pgp in a concentra-
Reduction of 12 using sodium borohydride complexed
in diglyme was required to provide sufficient amounts
of the a-C10 compound 13, as solid sodium borohydride
gave mostly the C9–C10 diol resulting from reduction of
the C9 carbonyl group. Acetylation and propionylation
occurred as expected in good yields affording 14 and 15
in 75% and 85%, respectively, using the corresponding
anhydrides. Hydrogen fluoride deprotection furnished
target compounds 16, 17, and 18 in 75%, 71%, and
80% yield, respectively. All compounds displayed spec-
troscopic properties in agreement with their structures.
Table 1 displays the effective dose ratios [ED50/
ED50(paclitaxel)] for each compound in a tubulin assem-
bly assay and cytotoxicity studies in the MCF-7
(parental cell line) and NCI/ADR-RES (multi-drug
resistant phenotype) cell lines. Our original hypothesis
was that the NCI/ADR-RES line would serve as an
initial indicator of the Pgp evading potential as this
line has been shown to express high levels of MDR-
1 and Pgp.19 Modification of the C10 center did not
significantly affect cytotoxicity or tubulin assembly
across the range of taxanes. In addition, none of the
compounds in Table 1 showed a large increase in
cytotoxicity for the multi-drug resistant phenotype.
This was unexpected, as we had anticipated significant
potency of these compounds in this cell line.
O
O
1
2
R
OH
R
O
O
Ph
NH
O
Ph
NH
O
OTES
OH
a, b
O
O
OTBS
OH
HO
BzO
H
AcO
HO
BzO
H
AcO
O
O
Each analog was evaluated for potential interactions
with Pgp relative to the negative (no addition) and posi-
tive controls [cyclosporine A (CsA) and paclitaxel, both
of which are known Pgp substrates and cause a net in-
crease in cellular rhodamine concentrations] in primary
cultures of bovine brain microvessel endothelial cells
(BMECs),20 which we have previously demonstrated to
4: R1 = H, R2 = OCOCH2CH3
3
5: R1 = H, R2 = OAc
6: R1 = H, R2 = OH
7: R1 = OH, R2 = H
Scheme 1. Synthesis and chemical structures of paclitaxel analogs.
Reagents and conditions: (a) (CH3CH2CO)2O, DMAP, 0 °C; 82%; (b)
HF/pyridine, pyridine, 0 °C; 75%.

J. T. Spletstoser et al. / Bioorg. Med. Chem. Lett. 16 (2006) 495–498
497
O
O
OR2
O
O
O
C5H11
NH
O
C5H11
NH
O
OR3
OTES
e
f
O
O
OTBS
OR1
(10 → 12)
HO
BzO
H
HO
BzO
H
AcO
AcO
O
O
2: R1 = R3 = H, R2 = Ac
a
c
12
8:
9:
R
R
1 = TBS, R2 = Ac, R3 = H
1 = TBS, R2 = R3 = H
b
d
10: R1 = TBS, R2 = H, R3 = TES
11: R1 = R2 = R3 = H
O
O
OR
OR
O
O
C5H11
NH
O
C5H11
NH
O
OH
OTES
i
O
O
OH
OTBS
HO
BzO
H
HO
BzO
H
AcO
O
AcO
O
16: R = H, 75%
17: R = Ac, 71%
13: R = H
14: R = Ac
g
h
18: R = -COCH2CH3, 80%
15: R = -COCH2CH3
Scheme 2. Synthesis of Taxol C analogs. Reagents and conditions: (a) TBSCl, imidazole, CH2Cl2; 87%; (b) N2H4ÆH2O, EtOH; 89%; (c) TESCl,
DMAP, CH2Cl2; 85%; (d) HF/pyridine, pyridine, 0 °C; 84%; (e) Dess–Martin periodinane, CH2Cl2; 99%; (f) NaBH4 (solution in diglyme); 50%;
(g) Ac2O, DMAP, pyridine, 0 °C; 75%; (h) (CH3CH2CO)2O, DMAP, pyridine, 0 °C; 86%; (i) HF/pyridine, pyridine, 0 °C.
Thus, it is conceivable that the acetate unit is no longer
able to interact with the complementary amino acid res-
idue in the Pgp binding site. Additionally, it is feasible
that the epimerization of this unit changes the spatial
distance of the outer two electron donors, disrupting
or weakening its hydrogen bonding ability. Interesting-
ly, as the size of the a-substituent increases (Fig. 2, com-
pounds 4 and 18), Pgp interaction is observed at levels
near that of the b-acetate compounds or positive
controls.
Table 1. ED50 ratios (compound/paclitaxel) for in vitro tubulin
assembly and cytotoxicity for compounds 1–2, 4–7, 11, and 16–1813
Compounds
Tubulin assembly
MCF-7
NCI/ADR-RES
1
2
1
1
1
2.0
4.1
0.8
3.2
2.1
2.2
1.6
0.21
2.1
2.2
2.6
0.7
1.4
7.5
4.8
9.2
3.2
3.8
0.48
1.1
1.3
0.7
6.6
3.8
3.1
1.6
2.9
4
5
6
7
11
16
17
18
We previously reported that C10-deacetoxypaclitaxel
and
C10-deacetoxy-C7-deoxypaclitaxel
interacted
Paclitaxel has a mean ED50 of 3.23 nM 1.84 and 1.53 lM 1.26 in
the MCF-7 and NCI/ADR-RES lines, respectively.
appreciably with Pgp in the rhodamine assay at 10 lM
but not at 5 lM.15 However, C7-deoxypaclitaxel inter-
express Pgp.21 A compound that does not interact, or
interacts weakly, with Pgp should not significantly
change the rhodamine level relative to the negative
control.
The results with paclitaxel (1) and Taxol C (2) (Fig. 2)
demonstrate that compounds containing the b-C10 ace-
tate group strongly interact with Pgp. As is evident,
there is a marked loss of interaction when the b-C10 ace-
tate is hydrolyzed (Fig. 2, compounds 7 and 11) for both
series. This is likely due to the conversion of the strongly
interacting type II unit to a weaker type I motif. The epi-
merization of the b-OH to the a configuration does not
cause a significant further change in interaction with
Pgp in this assay (Fig. 2, compounds 6 and 16). This
decreased interaction with the efflux protein is mostly
retained upon acetylation of the a-hydroxyl moiety
causing just a slight increase for compound 5 and a de-
crease for compound 17 in rhodamine accumulation.
This is significant as for both series the b-acetate com-
pounds interact appreciably with the Pgp transporter
relative to rhodamine. C10 a-substituents reside within
the concave ꢁcupꢀ of the baccatin portion of the taxanes.
Figure 2. Rhodamine uptake results for compounds 4–7 (A, paclitaxel
series) and 2, 11, and 16–18 (B, Taxol C series), in BMECs. Paclitaxel
and the derivatives were present at a concentration of 10 lM. The
concentrations of cyclosporin A (CsA) and rhodamine were 5 lM.

498
J. T. Spletstoser et al. / Bioorg. Med. Chem. Lett. 16 (2006) 495–498
3. (a) Woods, C. M.; Zhu, J.; McQueney, P. A.; Bollag, D.;
Lazarides, E. Mol. Med. 1995, 1, 506; (b) Jordan, M. A.
Curr. Med. Chem.: Anti-Cancer Agents 2002, 2, 1.
4. Suffness, M.; Wall, M. E. Discovery and Development of
Taxol. In Taxol: Science and Applications; Suffness, M.,
Ed.; CRC Press, Inc.: Boca Raton, FL, 1995; pp 3–25.
5. (a) Schinkel, A. H. Adv. Drug Deliv. Rev. 1999, 36, 179; (b)
Schinkel, A. H.; Smit, J. J. M.; Tellingen, O. van.; Beijnen,
J. H.; Wagenaar, E.; Deemter, L. van.; Mol, C. A. A. M.;
Valk, M. A. Van Der; Robanus-Maandag, E. C.; Riele, H.
P. J. Te.; Berns, A. J. M.; Borst, P. Cell 1994, 77, 491; (c)
Schinkel, A. H.; Wagenaar, E.; Mol, C. A. A. M.; Van
Deemter, L. J. Clin. Invest. 1996, 97, 2517; (d) 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. J. Clin.
Invest. 2002, 110, 1309; (e) Kemper, E. M.; van Zandber-
gen, A. E.; Cleypool, C.; Mos, H. A.; Boogerd, W.;
Beijnen, J. H.; van Tellingen, O. Clin. Cancer Res. 2003, 9,
2849.
acted with Pgp at both concentrations. Thus, it seems
that the C10 acetate plays a pivotal role in Pgp recogni-
tion. It is interesting to note that the compounds pre-
sented in this study (5–7, 11, 15, and 17) seem to do
little to rhodamine accumulation even at the 10 lM
concentration.
Although the rhodamine assay is an indirect measure-
ment of Pgp interaction, we have previously established
it as a viable indicator of relative BBB permeation by
comparison to bidirectional permeation data obtained
from BMEC monolayers of related compounds.20 Fur-
ther studies regarding these compounds in the BMEC
permeation assay are underway.
Although paclitaxel has six purported recognition ele-
ments, the simple modification of only one of these
groups is sufficient to decrease interaction with Pgp.
Although we do not yet know the relative contributions
of each recognition element in binding to Pgp, it is clear
that the element involving C9 and C10 plays a signifi-
cant role. This is underscored by observations that cer-
tain modifications at C10 resulted in compounds with
higher activity against Pgp-overexpressing cancer cell
lines12 or with ability to cross the BBB.14 We did not
find a correlation between our compoundsꢀ activity
against the NCI/ADR-RES cell line and rhodamine up-
take in BMECs. The discrepancy suggests the possibility
that other efflux systems may be present in the MDR cell
line that recognizes these compounds.22
6. van Asperen, J.; Mayer, U.; van Tellingen, O.; Beijnen, J.
H. J. Pharm. Sci. 1997, 86, 881.
7. (a) Michaelis, M.-L.; Ranciat, N.; Chen, Y.; Bechtel, M.;
Ragan, R.; Hepperle, M.; Liu, Y.; Georg, G. I. J.
Neurochem. 1998, 70, 1623; (b) Michaelis, M. L.; Ansar,
S.; Chen, Y.; Reiff, E.; Seyb, K. I.; Himes, R. H.; Audus,
K.; Georg, G. I. J. Pharmacol. Exp. Ther. 2005, 48, 832.
8. Seelig, A. Eur. J. Biochem. 1998, 251, 252.
9. (a) Seelig, A.; Landwojtowicz, E.; Fischer, H.; Blatter, X.
L. Methods Princ. Med. Chem. 2003, 18, 461; (b) Seelig,
A.; Li Blatter, X.; Wohnsland, F. Int. J. Clin. Pharmacol.
Ther. 2000, 251, 252.
10. Ojima, I.; Duclos, O.; Dorman, G.; Simonot, B.; Prest-
wich, G. D.; Rao, S.; Lerro, K. A.; Horwitz, S. B. J. Med.
Chem. 1995, 38, 3891.
In summary, we have described a novel single point
chemical modification of the baccatin portion of taxanes
that significantly decreases affinity for the P-glycopro-
tein transporter system. Additionally, we have begun
to elucidate a preliminary structure–activity relationship
for Pgp and taxanes with respect to the recognition ele-
ments postulated. The concept and data presented illus-
trate that it might be possible to deliver chemically
modified taxanes across the blood–brain barrier. This
is in contrast to current strategies for brain delivery of
non-CNS permeable drugs, including Pgp inhibition,5a
and BBB disruption.1 Long-term application of this re-
search has the potential to lead to effective agents for
CNS cancer treatment.
11. This cell line has since been renamed NCI/ADR-RES as
DNA fingerprinting has shown it is not derived from the
MCF-7 cell line: Scudiero, D. A.; Monks, A.; Sausville, E.
A. J. Natl. Cancer Inst. 1998, 90, 862.
12. 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. J. Med. Chem. 1996, 39, 3889.
13. Liu, Y.; Boge, T. C.; Victory, S.; Ali, S. M.; Zygmunt, J.;
Georg, G. I.; Marquez, R. T.; Himes, R. H. Comb. Chem.
High Throughput Screening 2002, 5, 39.
14. For review: Beckers, T.; Mahboobi, S. Drugs Future 2003,
28, 767.
15. Ge, H.; Vasandani, V.; Huff, J. K.; Audus, K. L.; Himes,
R. H.; Seelig, A.; Georg, G. I. Bioorg. Med. Chem. Lett.
2005, Manuscript Accepted.
16. (a) Hepperle, M. Isolation, Identification, Preparation,
Conformational Aspects and Biological Activities of Tax-
anes; University of Kansas: Lawrence, KS, 1995; (b)
Acknowledgments
´
Senhilh, V.; Blechert, S.; Colin, M.; Guenard, D.; Picot,
F.; Potier, P.; Varenne, P. J. Nat. Prod. 1984, 47, 131.
17. Datta, A.; Vander Velde, D. G.; Georg, G. I. Tetrahedron
Lett. 1995, 36, 1985.
18. Datta, A.; Hepperle, M.; Georg, G. I.; Himes, R. H.
J. Org. Chem. 1995, 60, 761.
This work was supported by a grant from the National
Cancer Institute (CA82801). J.T.S. and B.J.T. would
like to acknowledge the Department of Defense Breast
Cancer Research Program for predoctoral fellowships
(DAMD17-99-1-92530
Paclitaxel and Taxol C were generously donated by Tap-
estry Pharmaceuticals, Boulder, CO.
and
DAMD17-02-1-0435).
19. (a) Alvarez, M.; Paull, K.; Monks, A.; Hose, C.; Lee, J. S.;
Weinstein, J. J. Clin. Invest. 1995, 95, 2205; (b) Lee, J. S.;
Paull, K.; Alvarez, M.; Hose, C.; Monks, A.; Grever, M.
Mol. Pharmacol. 1994, 46, 627.
20. Rice, A.; Liu, Y.; Michaelis, M.-L.; Himes, R. H.; Georg,
G. I.; Audus, K. L. J. Med. Chem. 2005, 48, 832.
21. Rose, J. M.; Peckham, S. L.; Scism, J. L.; Audus, K.
Neurochem. Res. 1998, 23, 203.
References and notes
1. Fortin, D. Prog. Drug Res. 2003, 61, 127.
2. Patchell, R. A. Cancer Treat. Rev. 2003, 29, 533.
22. Sun, H.; Dai, H.; Shaik, N.; Elmquist, W. F. Adv. Drug
Deliv. Rev. 2003, 55, 83.