Tricyclic isoxazoles are novel inhibitors of the multidrug resistance protein (MRP1)

Article abstract of DOI:10.1016/S0960-894X(02)00051-3

Tricyclic isoxazoles were identified from a screen as a novel class of selective multidrug resistance protein (MRP1) inhibitors. From a screen lead, SAR efforts resulted in the preparation of LY 402913 (9h), which inhibits MRP1 and reverses drug resistanc

Full text of DOI:10.1016/S0960-894X(02)00051-3

Bioorganic & Medicinal Chemistry Letters 12 (2002) 883–886
Tricyclic Isoxazoles are Novel Inhibitors of the Multidrug
Resistance Protein (MRP1)
Bryan H. Norman,a,* Joseph M. Gruber,a Sean P. Hollinshead,c Joseph W. Wilson,c
James J. Starling,b Kevin L. Law,b Tracy D. Self,b Linda B. Tabas,b
Daniel C. Williams,b Donald C. Paul,b Margaret M. Wagnerb and Anne H. Dantzigb
aDiscovery Chemistry Research, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
bCancer Research, Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center,
Indianapolis, IN 46285, USA
cDepartment of Chemistry, Sphinx Laboratories, Eli Lilly and Company, 20 T. W. Alexander,
Research Triangle Park, NC 27709, USA
Received 16 November 2001; accepted 7 January 2002
Abstract—Tricyclic isoxazoles were identified from a screen as a novel class of selective multidrug resistance protein (MRP1) inhi-
bitors. From a screen lead, SAR efforts resulted in the preparation of LY 402913 (9h), which inhibits MRP1 and reverses drug
resistance to MRP1 substrates, such as doxorubicin, in HeLa-T5 cells (EC50=0.90 mM), while showing no inherent cytotoxicity.
Additionally, LY 402913 inhibits ATP-dependent, MRP1-mediated LTC4 uptake into membrane vesicles prepared from the
MRP1-overexpressing HeLa-T5 cells (EC50=1.8 mM). LY 402913 also shows selectivity (ꢀ22-fold) against the related transporter,
P-glycoprotein, in HL60/Adr and HL60/Vinc cells. Finally, when dosed in combination with the oncolytic MRP1 substrate vin-
cristine, LY 402913 delays the growth of MRP1-overexpressing tumors in vivo. # 2002 Published by Elsevier Science Ltd.
The multidrug resistance protein (MRP1) is a member
of the family of ATP-binding cassette (ABC) transpor-
ters and is structurally and functionally related to the
multidrug resistance (MDR) protein, P-glycoprotein (P-
gp).1 ABC transporters are known to have a number of
functions in normal human physiology. While the P-gp
pump has been linked to the removal of toxic xenobio-
tics,2 MRP1 seems to play an important role in the
transport of leukotrienes.3À7 The most thoroughly stud-
ied MRP1 substrate is leukotriene C4 (LTC4), which is
transported by MRP1 in an ATP-dependent fashion
with high efficiency (Km approx. 100 nM).4,5 Like P-gp,
MRP1 has been shown to confer cellular resistance to
various natural product oncolytics by transporting the
agents out of the cells. The resulting lower intracellular
concentration of the oncolytic produces a drug-resistant
phenotype.8 These drug-resistant cells require higher
concentrations of the oncolytic to achieve cytotoxicity.
MRP1-associated oncolytics include the anthracyclines
(doxorubicin and daunarubicin), epipodophylotoxins
(etoposide), and vinca alkoloids (vincristine). While P-
gp confers resistance to taxanes, MRP1 is apparently
not involved in taxane resistance.8
Drug resistance is a tremendous problem in the treat-
ment of many types of cancer. In the clinical setting,
patients who overexpress multidrug resistance proteins
such as P-gp and MRP1 in their tumors are usually not
responsive to associated anticancer agents.9 Thus, many
of these patients progress to advanced disease and have
poorer prognoses. It has been hypothesized that inhibi-
tion of MDR transporters can restore sensitivity to
some oncolytics, allowing patients with drug-resistant
tumors to become responsive to their drug therapy. This
hypothesis is presently being tested in the clinic for P-
gp.10 However, the search for MRP1 modulators just
began a few years ago and few potent and selective
modulators are known.11,12 We would like to report
here on the discovery of tricyclic isoxazoles, a new class
of selective MRP1 modulators.
A screen of potential MRP1 modulators uncovered the
tricyclic isoxazole indane 1, which had an EC50 of 10
mM in the MRP1-transfected cell line HeLa-T5 (when
*Corresponding author. Fax: +1-317-277-3652; e-mail: norman@
lilly.com
0960-894X/02/$ - see front matter # 2002 Published by Elsevier Science Ltd.
PII: S0960-894X(02)00051-3

884
B. H. Norman et al. / Bioorg. Med. Chem. Lett. 12 (2002) 883–886
dosed in the presence of a sublethal dose of doxo-
rubicin).13 The tricyclic moiety was formed, somewhat
serendipitously, during the base-promoted deprotection
of the indane alcohol functionality (Scheme 1). Under
the saponification conditions, the 2-chloro-6-fluoro-
phenyl-isoxazole functionality reacted via an intramole-
cular nucleophilic aromatic substitution reaction to give
the tricyclic compound 1. It was later determined that
this tricyclic functionality was absolutely required, as all
non-cyclized materials were inactive in our assays. From
this lead compound, we began to explore the SAR of
tricyclic isoxazoles and related compounds.
the corresponding acid chloride (Scheme 3). While
similar to the earlier route, we were able to employ effi-
cient parallel synthesis techniques in the final step of this
sequence. Specifically, by treating the acid chloride 8
with excess aniline and triethylamine, all amine-con-
taining impurities could be easily removed by passing
the crude mixture down a strong cation exchange (SCX)
column. The resulting solutions could be concentrated
to give extremely pure products in >80% yield. Early
results on the aniline series were promising, both from a
chemistry and biological activity perspective. Thus, it
seemed logical to pursue an initial SAR around the
more synthetically friendly phenyl linker region with
aniline substitution.
Using the method outlined in Scheme 3, an SAR was
conducted and the highlights are summarized in Table 1.
While the activities of most compounds ranged from
about 1 to >10 mM, one compound, 9h (LY 402913),
demonstrated activity below 1 mM (EC50=0.90 mM).
Thus, the 3,4,5-trimethoxyphenyl substitution was
Scheme 1.
At first, we investigated the importance of the indane
functionality by preparing the simplified compound 4
(Scheme 2). The synthesis began with 3-nitrobenzyla-
mine. Acylation with the isoxazole acid chloride, fol-
lowed by reduction with tin (II) chloride, resulted in the
aniline 2. Treatment of this material with the commer-
cially available 3-(2-chloro-6-fluorophenyl)-5-methyl-4-
isoxazoyl chloride resulted in the acyclic compound 3.
When this material was treated with sodium hydroxide
in DMF at room temperature, the intramolecular
nucleophilic aromatic substitution reaction occurred to
give the tricyclic product 4. The related ortho and para
substituted analogues 5 and 6 were prepared in similar
fashion. While the meta material 4 had an EC50 value
(7.9 mM) similar to the indane, both ortho and para
compounds were inactive (EC50 >> 10 mM).
Scheme 3. (a) 3-(2-Chloro-6-fluorophenyl)-5-methyl-4-isoxazolyl chlor-
ide, pyr, DMAP (cat.), CH2Cl2, 25ꢁC, 3 h (94%); (b) 2 N NaOH/
methanol, DMF, 25 ꢁC, 2 h (53%); (c) (COCl)2, CH2Cl2, 1 h (100%);
(d) selected aniline, Et3N, CH2Cl2, 15 h, purified on SCX column (80–
95%).
At the same time, we explored a route towards the iso-
meric amide series, where anilines could be reacted with
ꢁ
ꢁ
.
Scheme 2. (a) 5-Methyl-3-isoxazolyl chloride, Et3N, CH2Cl2, 25 C, 1 h (79%); (b) SnCl2 2H2O, THF, 25 C, 18 h (86%); (c) 3-(2-chloro-6-fluoro-
phenyl)-5-methyl-4-isoxazolyl chloride, pyr, DMAP (cat.), CH2Cl2, 25 ꢁC, 3 h (90%); (d) 2 N NaOH/methanol, DMF, 25 ꢁC, 2 h (97%).

B. H. Norman et al. / Bioorg. Med. Chem. Lett. 12 (2002) 883–886
885
Table 1. Structures and in vitro activities of tricyclic isoxazoles in the
MRP1-transfected HeLa-T5 cell line in the presence of doxorubicin
activity (EC50 in the low mM range in the HeLa-T5
assay), comparable to LY 402913. One change in this
region that did result in a slight increase in potency was
the N-methyl analogue 13 (EC50=0.64 mM) (Fig. 1).16
Based on the in vitro data in the HeLa-T5 cellular assay,
several compounds were evaluated for their ability to (a)
inhibit the target directly in an LTC4 transport assay
and (b) selectively inhibit the desired target (MRP1) in
preference to the related transporter (P-gp). These data
are summarized in Table 2. The LTC4 transport assay
measures the ability to specifically inhibit the transport
of the known substrate (LTC4) by MRP1 in an ATP-
dependent fashion.17 Compounds that can reverse the
drug-resistant phenotype in a cellular assay (such as the
HeLa-T5 assay) by some other mechanism will not
show activity in this assay. Thus, for our program, it
was critical that compounds display activity here.
Selectivity was determined in the HL60 panel of drug
selected cells, similar to the HeLa-T5 experiments. The
HL60/Adr cells overexpress MRP1, while the HL60/
Vinc cells overexpress P-gp.18,19 Therefore, the ability
for a compound to reverse drug resistance in the HL60/
Adr cell line, but not in the HL60/Vinc cell line, would
be indicative of selective MRP1 inhibition. A selectivity
ratio was calculated using the EC50 values in these two
assays. While all of the compounds showed good activ-
ity and selectivity in these assays, LY 402913 demon-
strated the best profile of potency in the cell lines,
specificity at the target and selectivity for MRP1 versus
P-gp.
Compd
Ar
EC50 (mM)
9a
9b
9c
9d
9e
9f
9g
9h
9i
9j
9k
9l
9m
9n
9o
Ph
4-Me-Ph
2.3
3.7
1.8
3-OMe-Ph
4-OMe-Ph
3,5-(OMe)2-Ph
3,4-(OMe)2-Ph
3,4-(-OCH2O-)-Ph
3,4,5-(OMe)3-Ph
3,5-(OMe)2-4-OH-Ph
4-F
>10
2.1
4.1
4.0
0.90 (Æ0.03)a
1.3
3.2
4-t-Bu
4-CN
4-CF3
4-CO2Me
4-SO2NH2
>10
4.9
>10
>10
>10
aStandard error for 9h (n=81), which was the control compound for
all experiments. Other compounds were tested in duplicate.
pursued in subsequent SAR studies. It should be poin-
ted out that compounds were tested for cytotoxicity in
this cell line and all were found to have no inherent
cytotoxicity. This is important to note, as additive
cytotoxic effects would confound the interpretation of
these data.
We developed an MRP1-dependent in vivo model by
implanting the drug-resistant HeLa-T5 cell line into
nude mice.20 This produced tumors that were resistant
to the MRP1-associated oncolytic vincristine. It should
be pointed out that the vector control cell line HeLa-C1,
which does not overexpress MRP1, was found to be
responsive to vincristine treatment. Thus, a successful
MRP1 inhibitor should have a synergistic effect on
reducing tumor growth when dosed in combination
with vincristine in the HeLa-T5 in vivo model. Figure 2
shows the results of an in vivo study using LY 402913
and vincristine. Animals were dosed orally with 10 mg/
kg LY 402913, 30 min before and after a bolus ivinfu-
sion (0.5 mg/kg) of vincristine, for 5 days. Additionally,
several controls were included in this study. First, the
vehicle control shows the tumor growth with no therapy
and, second, the vincristine-treated animals (0.5 mg/kg,
Several amide replacements were pursued. Initially, we
prepared 10, which contained our original amide orien-
tation, using the procedure shown in Scheme 2. We also
prepared ether isostere 1114 and the reduced amide iso-
stere 12.15 All three of these compounds showed good
Table 2. Specificity and selectivity of MRP1 inhibitors
Compd
LTC4
transport
EC50 (mM)a
HL60/Adr HL60/Vinc
Selectivity
ratiob
9h
(LY 402913)
1.8
0.07 (Æ0.01) 1.54 (Æ0.08)
22
9i
11
13
1.7
4.4
4.2
0.19 (Æ0.08) 2.58 (Æ0.42)
0.16 (Æ0.02) 2.34 (Æ0.63)
0.32 (Æ0.11) 1.33 (Æ0.15)
14
14
4
aStandard errors are shown in parentheses.
bSelectivity ratios were determined by calculating the ratio of average
EC50 values in the HL60/ADR and HL60/Vinc cell lines.
Figure 1. Amide isosteres of LY 402913.

886
B. H. Norman et al. / Bioorg. Med. Chem. Lett. 12 (2002) 883–886
6. Loe, D. W.; Almquist, K. C.; Cole, S. P. C.; Deeley, R. G.
J. Biol. Chem. 1996, 271, 9683.
7. Leier, I.; Jedlitschky, G.; Buchholz, U.; Cole, S. P. C.;
Deeley, R. G.; Keppler, D. J. Biol. Chem. 1994, 269, 27807.
8. Cole, S. P. C.; Sparks, K. E.; Fraser, K.; Loe, D. W.;
Grant, C. E.; Wilson, G. M.; Deeley, R. G. Cancer Res. 1994,
54, 5902.
9. Bellamy, W. T.; Dalton, W. S. Adv. Clin. Chem. 1994, 31, 1.
10. Shustik, C.; Dalton, W.; Gros, P. Mol. Aspects Med. 1995,
16, 1.
11. Norman, B. H. Drugs Future 1998, 23, 1001.
12. Norman, B. H.; Dantzig, A. H.; Kroin, J. S.; Law, K. L.;
Tabas, L. B.; Shepard, R. L.; Palkowitz, A. D.; Hauser, K. L.;
Winter, M. A.; Sluka, J. P.; Startling, J. J. Bioorg. Med. Chem.
Lett. 1999, 9, 3381.
13. Compounds were tested against the HeLa-T5 cell line in
vitro using an MTT-like assay. A fixed concentration of
doxorubicin (0.5 mg/mL) and varying concentrations of test
compounds, starting at 10 mM, were added to triplicate wells
of a 96-well plate and incubated for 72 h at 37 ꢁC. EC50 values
were calculated from level of inhibition of cell growth using
curve fitting software. Each compound was also tested for
inherent toxicity in the same assay in the absence of doxo-
rubicin. Test compounds showed <20% growth inhibition at
10 mM.
14. Ether isostere 11 was prepared using the following condi-
tions: (a) 3,4,5-trimethoxyphenol, 3-nitrophenethyl alcohol,
DEAD, PPh3, toluene, 1 h (75%); (b) H2/Pd on C, methanol,
30 psi, 2 h (98%); see Scheme 2 (c) and (d).
15. Reduced amide 12 was prepared according to Scheme 2,
beginning with BOC-protected 3-nitrobenzylamine. Upon
formation of the tricyclic moiety, the BOC group was removed
(TFA, 100%) and reductive amination with 3,4,5-trimethoxy-
benzaldehyde (NaCNBH3, methanol, 87%) gave the product.
16. The N-methyl analogue 13 was prepared according to
Scheme 2 and was methylated after the first acylation (NaH,
MeI, DMF, 1 h, 90%).
Figure 2. In vivo activity of LY 402913.
iv) showed the effect of drug without inhibitor. A final
control was performed where the animals were dosed
with inhibitor alone (same regimen). This curve, which
overlaps the vehicle control curve (data not shown),
indicates that the inhibitor has no antitumor activity of
its own. As shown in Figure 2, the combination treat-
ment of vincristine and LY 402913 gave a statistically
significant improvement in efficacy, relative to the con-
trols. We believe that this is the first report of a selective
MRP1 modulator with in vivo efficacy in an MRP1-
dependent animal model.
In summary, these studies show that the tricyclic isox-
azoles are a promising series of selective MRP1 inhibitors.
As more is learned on the tissue specificity of ABC
transporters, and their relationship to clinical drug
resistance, we feel that it is essential that selective drug
resistance modulators be developed. Due to the fact that
most of the known MRP1 modulators also inhibit the
related transporter, P-gp, this series is particularly
exciting.
17. Dantzig, A. H.; Shepard, R. L.; Tabas, L.; Pratt, S.; Gil-
lesppie, J. I.; Binkley, S.; Wrighton, S. A. J. Pharmacol. Exp.
Ther. 1999, 290, 854.
18. McGrath, T.; Latoud, C.; Arnold, S. T.; Safa, A. R.;
Felsted, R. L.; Center, M. S. Biochem. Pharmacol. 1989, 38,
3611.
19. Modulator EC50 in HL60/Adr and HL60/Vinc cells were
determined by assessing the concentration-dependent ability of
the modulator to enhance the antiproliferative response of the
cells to an IC20 concentration of doxorubicin using alamar-
Blue1 reduction as a surrogate measure for cell number. The
selectivity ratio was calculated by dividing the average EC50
for HL60/Vinc cells by the average EC50 for HL60/Adr cells.
20. The ability of the compounds to increase the antitumor
activity of co-administered vincristine in the HeLa-T5 xeno-
graft model was tested. Test compounds were orally adminis-
tered for 5 consecutive days to female nude mice with and
without intravenous administration of vincristine sulfate.
Caliper measurements of tumor growth were taken at various
time points and plotted to indicate antitumor response.
References and Notes
1. Cole, S. P. C.; Bhardwaj, G.; Gerlach, J. H.; Mackie, J. E.;
Grant, C. E.; Almquist, K. C.; Stewart, A. J.; Kurz, E. U.;
Duncan, A. M. V.; Deeley, R. G. Science 1992, 258, 1650.
2. Endicott, J. A.; Ling, V. Annu. Rev. Biochem. 1989, 58, 137.
3. Jedlitschky, G.; Leier, I.; Buchholz, U.; Center, M.; Kep-
pler, D. Cancer Res. 1994, 54, 4833.
4. Jedlitschky, G.; Leier, I.; Buchholz, U.; Barnouin, K.;
Kurz, G.; Keppler, D. Cancer Res. 1996, 56, 988.
5. Loe, D. W.; Almquist, K. C.; Deeley, R. G.; Cole, S. P. C.
J. Biol. Chem. 1996, 271, 9675.