Synthesis and antitumor activity of novel paclitaxel-chlorambucil hybrids

Article abstract of DOI:10.1016/S0960-894X(01)00067-1

The syntheses and antitumor activity of three paclitaxel-chlorambucil hybrids are presented. Hybrid 3 showed significant in vivo efficacy.

Full text of DOI:10.1016/S0960-894X(01)00067-1

Bioorganic & Medicinal Chemistry Letters 11 (2001) 811±814
Synthesis and Antitumor Activity of Novel
Paclitaxel±Chlorambucil Hybrids
Mark D. Wittman,a,* John F. Kadow,a Dolatrai M. Vyas,a Francis L. Lee,b
William C. Rose,b Byron H. Long,b Craig Fairchildb and Kathy Johnstonb
aDiscovery Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 5100, Wallingford, CT 06492-7660, USA
bOncology Drug Discovery, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 4000 Princeton, NJ 08543-4000, USA
Received 1 November 2000; accepted 25 January 2001
AbstractÐThe syntheses and antitumor activity of three paclitaxel±chlorambucil hybrids are presented. Hybrid 3 showed signi®cant
in vivo ecacy. # 2001 Elsevier Science Ltd. All rights reserved.
The synergistic combination of Taxol1 with other anti-
tumor agents has been well recognized to provide clini-
cally e€ective treatment regimes. Our own preclinical
evidence suggested that the combination of paclitaxel
and an alkylating agent could yield supraadditive anti-
tumor e€ects.1 M109taxlR,2 a cell line that was made
resistant to paclitaxel, was found to be 2- to 5-fold
hypersensitive in vitro to cisplatin, Mitomycin C, Mel-
phalan and BCNU versus the parental cell line (M109).
This hypersensitivity to alkylating agents was also
demonstrated in vivo. This suggested that a hybrid
molecule combining an alkylating agent with paclitaxel
might extend the therapeutic utility of paclitaxel to
resistant or paclitaxel insensitive tumors.
Two reports describing paclitaxel hybrid molecules have
appeared3,4 in which the individual molecules are con-
nected via a chemically stable linker. In this study, we
envisioned using hybrids with a range of stability to
esterase cleavage capable of releasing both paclitaxel
and the alkylating agent simultaneously. In general,
hybrid molecules are not as e€ective as combination
therapy since optimization of individual drug con-
centration and schedule is not possible. Nevertheless,
there may be potential advantages if both drugs can be
deactivated as a result of the linking function and then
released together. To test this hypothesis three pacli-
taxel±chlorambucil hybrid molecules were prepared
(Fig. 1). Hybrid 1 is a simple ester of chlorambucil and
*Corresponding author. Tel.:+1-203-677-6318; fax: +1-203-677-
7702; e-mail: mark.wittman@bms.com
Figure 1.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00067-1

812
M. D. Wittman et al. / Bioorg. Med. Chem. Lett. 11 (2001) 811±814
paclitaxel and as such is expected to be quite stable to
esterase cleavage.5 Extending the ester by means of an
acetal linkage, as in 2, renders the ester more susceptible
to esterase cleavage. Hybrid 3 takes advantage of the
acetal extension to provide two molecules of chlor-
ambucil and one molecule of paclitaxel upon esterase
cleavage. Preliminary experiments on the bis silyl pro-
tected diethanolamine prodrug 5, showed paclitaxel was
generated rapidly upon silyl ether cleavage (bu€ered
Bu4NF) presumably from self-immolative cyclization of
the hydroxyl group on the extended ester (Fig. 2). In
comparison, the directly linked prodrug 6 generated
paclitaxel much more slowly, and the desilylated inter-
mediate was observed. The rapid degradation of 5 sug-
gested this would be an ideal platform on which to
introduce an alkylating agent. Prodrug 4, with two
acetates rather than two molecules of chlorambucil, was
also prepared as a control.
The hybrid molecules 1±4 were prepared following
Scheme 1. Starting with 20-O-TES paclitaxel, hybrid 1
was prepared by simple acylation of the 7-hydroxyl
group followed by 20 desilylation. The extended ester 2
was prepared using the MTM intermediate 7. Activa-
tion of the MTM ether using NIS with AgOTf as a cat-
alyst provided the extended ester 2. Treatment of the
MTM ether with chloroacetic acid under similar activa-
tion conditions provided the extended chloroacetyl ester
8 which could be reacted with the amine 9 to give hybrid
3. The diacetoxy prodrug 4 was prepared using the
sequence used to prepare 3. Amine 9 was prepared in
straightforward fashion from diethanolamine. Silylation
of the hydroxyls (TBSCl, imidazole, DMF) followed by
protection of amine as a Cbz carbamate (CbzCl, iPr2-
NEt, CH2Cl2) and deprotection of the silyl ethers
(Dowex 50W-X8, MeOH) provided the N-protected
diethanolamine. Esteri®cation of the hydroxyls with
chlorambucil using DCC and DMAP followed by Cbz
removal provided the amine 9 in 70% overall yield. The
N,N diacetoxyethyl amine was prepared similarly.
The hybrid molecules 1±4 were evaluated in vitro for
inhibition of tubulin polymerization6 and cytotoxicity
against the HCT-116 human colon tumor cell line.7
Figure 2.
Scheme 1. Reagents and conditions: (a) chlorambucil, DCC, DMAP, toluene (79%); (b) 1 N HCl, CH3CN, OꢀC; (c) DMS, Bz2O2, CH3CN (93%);
(d) chlorambucil, NIS, AgOTf, 4 A sieves, THF (26%); (e) ClCH2CO2H, NIS, AgOTf, 4 A sieves, THF (51%); (f) NaI, acetone, 9 (47%).

M. D. Wittman et al. / Bioorg. Med. Chem. Lett. 11 (2001) 811±814
813
Figure 3. Activity of hybrid 3 versus im bilaterally implanted M109 and M109/txlr.
Table 1. Biological activity of compounds 1±4
To better understand what was driving the antitumor
activity of 3, it was tested in the A2780 ovarian carci-
noma xenograft tumor model10 (paclitaxel-sensitive,
chlorambucil-resistant), HCT/pk human ovarian carci-
noma xenograft model11 (paclitaxel-resistant, chlor-
ambucil-resistant), and the M5076 murine sarcoma
model12 (paclitaxel-insensitive, chlorambucil-sensitive).
In the paclitaxel-sensitive A2780, hybrid 3 was equally
active to paclitaxel but not as ecacious as the optimal
combination treatment. In the HCT/pk model (pacli-
taxel and chlorambucil-resistant) compound 3 did not
achieve an active result. In the M5076 model (paclitaxel-
insensitive, chloroambucil-sensitive) hybrid 3 showed
signi®cant antitumor activity that was equivalent to
chlorambucil or the optimal combination treatment.
These results show that hybrid 3 shows signi®cant anti-
tumor ecacy in tumor models that are responsive to
either paclitaxel or chlorambucil, suggesting that it
retains both paclitaxel and chloroambucil-like activity.
However, in the HCT/pk model which is resistant to
both agents no activity was observed for compound 3.
The HCT/pK results show that not all paclitaxel resis-
tant cell lines are hypersensitive to alkylating agents but,
in cell lines that are (such as M109/taxlR), hybrid
molecules such as 3 can yield supra-additive antitumor
e€ects superior to combination treatment.
Compound
Tubulina
IC50 HCT-166 (nM)
Ip/ip M109b
Paclitaxel
1
53
62
70
6.5
2±4
161
4.6
24
194±246
127
155
224
17
1
2
3
4
3.9
aTubulin assay measures the initial rate of polymerization and is
expressed as a ratio to that obtained for paclitaxel. Ratios less than 1
re¯ect analogues that show a more rapid polymerization rate than
paclitaxel.6
bValues are %T/C; lifespan of treatment group divided by the lifespan
of the the control group multiplied by 100.
Compounds were tested in vivo using the M109 Madi-
son murine lung carcinoma M109 tumor model8 (Table
1). The lack of activity in the cell-free tubulin poly-
merization assay suggested that compounds 1±3 are
inactive until cleavage of the ester takes place. The
cytotoxicity of compounds 2 and 4 suggests that these
analogues may undergo cleavage to generate pacli-
taxel. All four molecules showed in vivo activity in the
paclitaxel-sensitive M109 model.
Compound 3 was tested in a bilateral9 distal site model
against M109 (paclitaxel and chlorambucil-sensitive)
and M109/taxlR (paclitaxel-resistant, chlorambucil-sen-
sitive). Compound 3 was curative in both the paclitaxel-
sensitive M109 arm and paclitaxel-resistant M109/taxlR
arm (Fig. 3). The clear superiority of hybrid 3 relative to
the optimal combination treatment demonstrates that a
paclitaxel±alkylating agent hybrid can yield super-addi-
tive antitumor activity in paclitaxel-resistant tumors
that are responsive to alkylating agents such as M109/
taxlR.
In summary, a series of three paclitaxel±chlorambucil
hybrids was prepared and their antitumor activity eval-
uated in vitro and in vivo. All hybrids showed in vivo
antitumor activity with compound 3 being the most
active. Compound 3 showed excellent in vivo activity in
M109 and the paclitaxel resistant M109/taxlR models
that was superior to single agent or combination treat-
ment. Compound 3 showed in vivo ecacy in the
paciltaxel-sensitive, A2780 ovarian carcinoma model
and the chlorambucil-sensitive, M5076 murine sarcoma
model.

814
M. D. Wittman et al. / Bioorg. Med. Chem. Lett. 11 (2001) 811±814
6. Schi€, P. B.; Fant, J.; Horwitz, S. B. Nature 1979, 277, 665.
References and Notes
An E0.01was determined for each analogue and expressed as a
ratio relative to paclitaxel.
7. Scudiero, D. A.; Shoemaker, R. H.; Paull, K. D.; Monks,
A.; Tieney, S.; Nofziger, T. H.; Currens, M. J.; Seni€, D.;
Boyd, M. R. Cancer Res. 1988, 48, 4827.
1. Lee, F. Y. F.; Wittman, M. D.; Burke, C.; Clark, J.; Fager,
K.; Vanderen, C.; Rose, W. C. Proc. Am. Assoc. Cancer Res.
2000, 41, 214.
2. M109/taxlR model was developed by treating M109 tumor
bearing mice with the maximum tolerated dose of paclitaxel
(daily Â5) allowing the tumor to re-grow. The tumors were
removed, disassociated, re-implanted, and treated with a sec-
ond cycle of paclitaxel. This cycle was repeated a total of 10
times. The tumors that survived this treatment were then used
to establish a cell line in vitro. Paclitaxel resistance in vitro was
observed to be 4-fold. Lee, F. Y. F.; Fager, K.; Fairchild, C.;
Carboni, J. Proc. Am. Assoc. Cancer Res. 1995, 36, 319.
3. Kar, A. K.; Braun, P. D.; Wandless, T. J. Bioorg. Med.
Chem. Lett. 2000, 10, 261.
4. Metha, G.; Maiya, G. B.; Muthusamy, S.; Chanon, M.; Jul-
liard, M. Fr. Pat. Appl. 2764605. Chem. Abstr. 1998, 130, 153827.
5. Taxane Anticancer Agents, Basic Science and Current Status;
Vyas, D. M., Ueda, Y., Wong, H., Matiskella, J. D., Hauck,
S., Mikkilineni, A. B., Farina, V., Rose, W. C., Casazza,
A. M., Georg, G. I., Chen, T. T., Ojima, I., Vyas, D. M., Eds.;
ACS Symposium Series 583; American Chemical Society:
Washington, DC, 1995; 99124.
8. %T/C values of greater than 125 are considered active.
Rose, W. C. Cancer Treat. Rep. 1981, 65, 299.
9. In this experiment the tumor was implanted intra-muscu-
larly (hind leg) and the drug given intra-venously. The same
animal had both an M109 tumor and a M109/taxlR tumor.
10. A2780 human ovarian carcinoma xenograft model is
highly sensitive to paclitaxel. Rose, W. C. Anti-Cancer Drugs
1992, 3, 311.
11. HCT/pk human ovarian carcinoma xenograft model
(multidrug resistant tumor model). HCT/pk was established
by exposing parental HCT-116 cells to concentrations of
paclitaxel that mimic clinically achievable plasma levels. This
resulted in a multidrug resistant tumor model that expressed
elevated levels of P-glycoprotein, was cross resistant to a
number of natural product anticancer agents, and was resis-
tant to paclitaxel in vivo.
12. M5076 murine sarcoma model is insensitive to paclitaxel.
Rose, W. C. Anti-Cancer Drugs 1992, 3, 311.