Synthesis and biological activity of retinoic acid receptor-α specific amides

Article abstract of DOI:10.1016/S0960-894X(02)00647-9

Retinoids are analogues of all-trans-retinoic acid, a powerful hormone that mediates many fundamental biological processes. Cancer and other serious hyperproliferative diseases are attractive therapeutic targets for retinoids, but the therapeutic use of retinoids is limited due to severe toxicity. We report here the design of retinoid receptor-α specific ligands with growth inhibitory activity in breast cancer cell lines, and which do not cause the cutaneous toxicity associated with the currently available nonselective retinoid agonists.

Full text of DOI:10.1016/S0960-894X(02)00647-9

Bioorganic & Medicinal Chemistry Letters 12 (2002) 3145–3148
Synthesis and Biological Activity of Retinoic Acid
Receptor-ꢀ Specific Amides
Richard L. Beard,a,* Tien T. Duong,a Min Teng,a Elliott S. Klein,b
Andrew M. Standevanb and Roshantha A. S. Chandraratnaa,b,
*
aRetinoid Research, Department of Chemistry, Allergan Inc., Irvine, CA 92623-9534, USA
bRetinoid Research, Department of Biology, Allergan Inc., Irvine, CA 92623-9534, USA
Received 7May 2002; accepted 28 July 2002
Abstract—Retinoids are analogues of all-trans-retinoic acid, a powerful hormone that mediates many fundamental biological pro-
cesses. Cancer and other serious hyperproliferative diseases are attractive therapeutic targets for retinoids, but the therapeutic use of
retinoids is limited due to severe toxicity. We report here the design of retinoid receptor-a specific ligands with growth inhibitory
activity in breast cancer cell lines, and which do not cause the cutaneous toxicity associated with the currently available nonselective
retinoid agonists.
# 2002 Elsevier Science Ltd. All rights reserved.
Retinoids are analogues of all-trans-retinoic acid
(ATRA), a powerful hormone that is capable of med-
iating many fundamental biological processes, including
cell proliferation and differentiation and apoptosis.1
Thus, retinoids have biological properties that suggest
potential therapeutic uses in the treatment and pre-
vention of cancer. Indeed, commercially available
retinoids, including ATRA, have been approved for
the treatment of some cancers.2 Unfortunately, the
longer term use of retinoids, particularly in the
treatment of premalignant lesions, is severely limited
by serious side effects such as mucocutaneous toxi-
city, hypertriglyceridemia, and headache.3 In order
for retinoids to reach their full therapeutic potential
in treating cancer and other serious diseases, new
compounds with reduced toxicity and improved
therapeutic indices must be discovered.
the RXRs, and it binds to both RXRs and RARs with
high affinity.5 In biological systems, RARs and RXRs
form functional RAR–RXR heterodimers which are
effectively activated by RAR-specific agonists but not
by RXR-specific agonists. Upon binding an RAR ago-
nist, these RAR–RXR heterodimers, which are asso-
ciated with specific response elements in the promoter
regions of retinoid responsive genes, recruit co-activator
proteins to the receptor and induce gene transcription.
Retinoids induce their powerful biological effects by
binding to and activating nuclear receptors. There are six
known retinoid receptors, the retinoic acid receptors—
RARa, b and g—and the retinoid X receptors—RXRa, b,
and g. ATRA binds with high affinity to all three RARs
but does not bind to the RXRs.4 A geometric isomer of
ATRA, 9-cis-retinoic acid, is the putative hormone for
It is now evident that each of the RAR subtypes can
activate distinct sets of genes. In addition, it is known
that the tissue distribution of the RAR subtypes is not
uniform, and that distinct RAR subtypes are expressed
in different tissues. For example, it is known that the
*Corresponding authors. Fax:+1-714-246-6207; e-mail: beard_richard@
allergan.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00647-9

3146
R. L. Beard et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3145–3148
most abundant receptor in the skin is RARg,6 and
studies strongly suggest that the mucocutaneous toxicity
caused by retinoids is through activation of RARg.7
These factors imply that retinoids that specifically acti-
vate a single RAR subtype will elicit a narrower range
of biological effects in fewer tissues than nonselective
retinoids. Thus, it may be possible to design RAR sub-
type specific compounds that have increased beneficial
biological effects and decreased toxicity relative to non-
selective retinoids. Such compounds with improved
therapeutic indices will be much more useful for the
treatment of retinoid responsive diseases than the cur-
rently available nonselective retinoid drugs.
immediately with either ethyl 2-fluoro-4-aminobenzoate
or ethyl 2,6-difluoro-4-aminobenzoate11 and pyridine to
produce the corresponding amide derivatives. The ester
group was then hydrolyzed under standard conditions
(NaOH, ethanol) to produce the desired retinoid analo-
gues, 1 (X=H) and 2 (X=F).
Scheme 2 shows the synthesis of compounds 3 and 4.
The known aryl bromide 613 was subjected to a lithium-
halogen exchange reaction, and the resulting anion was
quenched with carbon dioxide to give carboxylic acid 7
after acidic workup. Compound 7 was treated as above
with thionyl chloride and ethyl 2,6-difluorobenzoate to
form amide analogue 8. The methyl protecting group
was then removed by treating 8 with boron tribromide,
producing compound 9. Compound 9 was then hydro-
lyzed under the usual conditions to give retinoid analo-
gue 3. Alternatively, compound 9 was chlorinated with
sulfuryl chloride to give compound 10, which was then
hydrolyzed to carboxylic acid 4.
Several groups have reported compounds that specifi-
cally activate either the RAR or RXR families of
receptors.8 In addition, several compounds with mod-
erate selectivity for two of the three RAR subtypes have
been described.8 Among the first RAR subtype selective
retinoids discovered was the amide linked compound,
Am-580, which was reported by Shudo and co-workers
in 1988.9 This compound has higher affinity for the
RARa subtype than for RARb or RARg. The amide
linker group is a key structural feature for the RARa
selectivity displayed by these arotinoids. This is pre-
sumably due to a favorable hydrogen-bonding interac-
tion between the amide group of the ligand and the
hydroxyl group on serine 232 residue present in the ligand
binding pocket of RARa.10 The space occupied by serine
232 in RARa is occupied by lipophilic alanine residues in
RARb and RARg, leading to destabilized protein–ligand
interactions and decreased binding affinities. We recently
reported on AGN 193836, an amide-linked compound,
that has true pharmacological specificity (>1000-fold) in
binding affinity for RARa over RARb and RARg.11 In
cotransfection assays, AGN 193836 transactivates
exclusively through RARa, and is completely inactive at
RARb and RARg. We have continued to develop this
series of compounds and report here the discovery of 4
(AGN 195183), a compound with improved RARa
binding selectivity relative to AGN 193836. The struc-
ture–activity relationships displayed by these com-
pounds are also examined. Compound 4 inhibited the
growth of breast cancer cell lines, and was inactive in an
in vivo model of topical irritation.
Table 1 summarizes the RAR binding and transactiva-
tion data of these compounds, which were tested as
previously described.14 These compounds had no bind-
ing or transactivation activity at the RXRs (data not
shown). Am-580 binds to RARa with about 35-fold
higher affinity than to RARb and about 105-fold higher
affinity than to RARg. In transactivation assays, Am-
580 is able to fully activate all three RAR subtypes, and
it is moderately more potent at RARa (EC50=2 nM)
than at RARb (9 nM) and RARg (19 nM). A significant
improvement in binding selectivity was achieved with
the monofluorobenzoic acid, compound 1. This com-
pound is about 190-fold selective in binding to RARa
than to RARb, and almost 500-fold selective in binding
to RARa than to RARg. Similar to Am-580, compound
1 is about 5-fold more potent in activating RARa than
The mono-and di-fluorinated retinoids, 1 and 2, were
prepared from tetralin carboxylic acid 59 by a three step
protocol as shown in Scheme 1.12 Tetralin 5 was
refluxed in thionyl chloride to convert the carboxylic
acid to an acid chloride. This compound was reacted
Scheme 2. Synthesis of RARa-specific retinoid agonists 3 and 4: (a)
t-BuLi, THF, À78 ꢀC; CO2; H3O+, 93%; (b) SOCl2, reflux; (c) ethyl
2,6-difluoro-4-aminobenzoate, pyridine, CH2Cl2, 85%; (d) BBr3,
CH2Cl2, À78 À25 ꢀC, 92%; (e) NaOH, EtOH; H+, 85–90%; (f)
SO2Cl2, CH2Cl2, Et2O, 25 ꢀC, 70%.
Scheme 1. Synthesis of mono- and difluorobenzoic acid compounds 1
and 2: (a) SOCl2, reflux; (b) ethyl 2-fluoro-4-aminobenzoate or ethyl
2,6-difluoro-4-aminobenzoate, pyridine, CH2Cl2, 82–90%; (c) NaOH,
EtOH; H+, 80–93%.

R. L. Beard et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3145–3148
3147
Table 1. RAR transcriptional activation and competitive binding data for RARa-selective retinoid agonists14
Compd
RAR bind. Kd (nM)
RAR Trans. EC50 (nM) (% efficacy)
a
b
g
a
b
g
ATRA
Am 580
193836
1
2
3
4
14
36
4
2*
7*
3
11
16
240 (100)
2 (101)
290 (50)
8 (69)
94 (81)
112 (77)
200 (80)
38 (100)
9 (105)
NA
6 (100)
19 (80)
NA
350 (85)
NA
1400
6400
380*
1470*
>104
>105
3800
>104
990*
5600*
5600
>105
45 (78)
>1000 (25)
>1000 (10)
NA
>1000 (10)
NA
3
Values represent the mean of three determinations. Errors in this assay are approximately 15% of the mean value. NA indicates <10% efficacy at
the highest dose tested (1 mM). Efficacy (%) is the percent transcriptional activity of a compound at the highest dose tested relative to transcriptional
activity produced by 1 mM ATRA.
RARb in the transactivation assay. More significantly,
compound 1 activates gene transcription through RARa
with >40-fold higher potency than through RARg. The
difluorobenzoic acid, compound 2, was even more
selective than compound 1, having Kd values of 7nM at
RARa and 1470 nM (210-fold selectivity) at RARb and
5600 nM (800-fold selectivity) at RARg. In terms of
transcriptional activity, compound 2 is a full agonist of
RARa-mediated gene transcription, having an EC50 of
94 nM. This compound has only partial agonist activity
at RARb at the highest dose tested (1000 nM), and it
has no transcriptional activity at RARg. Thus, the
diflurobenzoic acid moiety is a key structural unit
necessary for achieving high selectivity in the binding
and transactivation assays.
We then determined the growth inhibitory properties of
compound 4 in two human breast cancer cell lines,
T-47D, which is estrogen receptor (ER) positive, and
SK-BR-3, which is ER negative. Both of these cell lines
are known to express relatively high levels of RARa and
should be responsive to RARa-specific agonists.15
Compound 4 inhibited of the growth of T-47D cells by
approximately 60% with an IC50 value of 1.4 nM. By
comparison, ATRA inhibited the growth of T-47D cells
by about 70% with an IC50 value of 200 nM. In SK-BR-
3 cells, 4 was even more efficacious showing 93%
growth inhibition with an IC50 value of 11 nM. This
compares well with ATRA, which induced 99% growth
inhibition with an IC50 value of approximately 2 nM.
These data indicate that RARa-specific ligands such as 4
are potentially useful for the treatment of breast cancer.
We mentioned earlier that the hydrogen-bonding char-
acteristics of the amide group in these compounds are
responsible for maintaining strong binding to RARa
while destabilizing binding to RARb and RARg. We
wished to examine if this property could be exploited
further by introducing a hydrogen-bonding group adja-
cent to the amide linker in the tetrahydronaphthalene
moiety. To this end, compound 3, which has a hydroxyl
group in the a-position relative to the amide group, was
synthesized. Indeed, compound 3 binds to RARa with
very high affinity, having a Kd value of 3 nM, while
being unable to bind to RARb and binding only very
weakly (Kd=5600 nM) to RARg. Furthermore, com-
pound 3 maintained full transcriptional activity through
RARa, with an EC50 value of 112 nM, but only acti-
vated RARb and RARg with about 10% efficacy at the
highest dose (1000 nM) tested. Another beneficial
structural modification was to add halogens (Cl and Br)
to the C18 position (ATRA numbering) to further
improve RARa selectivity. Like AGN 193836, which is
substituted with bromine in the C18-position, com-
pound 4, which is substituted with chlorine, has true
pharmacological specificity for RARa over RARb and
RARg in the binding assays. Compound 4 binds to
RARa with high affinity, having a Kd value of 3 nM,
and it does not bind to either RARb or RARg, even at
doses as high as 100 mM. In the transactivation assays,
compound 4 activates RARa exclusively, with an EC50
value of 200 nM. Compound 4 was unable to induce
transcriptional activity through either RARb or RARg
at the doses tested. To our knowledge, compound 4 is
the most selective RARa agonist known.
With regard to retinoid toxicity, we wanted to deter-
mine if RARa-specific agonists such as 4 had any
advantage over RARa-selective compounds such as
Am-580. We tested 4 and Am-580 in our previously
described topical irritation assay, a clinically relevant in
vivo model of retinoid toxicity.3b In this assay, the reti-
noids were applied at a dose of 1000 nmol/25 g to the
backs of hairless mice for an 8 day period. The severity
of the induced topical toxicity was scored from 0 to 21
(0=no irritation, 21=severe irritation) based on the
amount of flaking, abrasion and body weight loss
observed in each animal. As can be seen from the data
summarized in Table 2, Am-580 caused severe topical
irritation at this dose whereas 4 caused no topical irri-
tation whatsoever. Thus, the topical toxicity caused by
RARa subtype specific agonists such as 4 is very differ-
ent from that caused by compounds with only moderate
RARa subtype selectivity.
In summary, we have described the synthesis and struc-
ture–activity relationships of new RARa-selective ago-
nists which led to the discovery of the RARa-specific
agonist 4. We have shown that this compound is a
potent growth inhibitor of the ER positive and ER
negative human cancer breast cell lines, T-47D and
SK-BR-3, respectively. In addition, in a hairless mouse
model of retinoid induced topical irritation, compound
4 was non-toxic while Am-580 caused severe topical
irritation. This confirms our hypothesis that compounds
with true RAR subtype specificity will have very differ-
ent biological properties than RAR subtype selective or

3148
R. L. Beard et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3145–3148
Table 2. Compound 4 and ATRA inhibit growth of the human
breast cancer cell lines, T-47D and SK-BR-3,15 compound 4 does not
cause the topical irritation induced by the RARa-selective retinoid,
Am-5803b
Practice; Korean, G., Ed.; Mercel Dekker: New York, 1992; p 47 .
(b) Standeven, A. M.; Johnson, A. T.; Escobar, M.; Chan-
draratna, R. A. S. Toxicol. Appl. Pharmacol. 1996, 138, 169.
4. (a) Petkovich, M.; Brand, N. J.; Krust, A.; Chambon, P.
Nature 1987, 330, 444. (b) Giguere, V.; Ong, E. S.; Segui, P.;
Evans, R. M. Nature 1987, 330, 624. (c) Krust, A.; Kastner, P.;
Petkovich, M.; Zelent, A.; Chambon, P. Proc. Natl. Acad.
U.S.A. 1989, 86, 5310.
Compd
T-47D
IC50 (nM)
(% efficacy)
SK-BR-3
IC50 (nM)
(% efficacy)
Topical irritation
score
(0–21)
ATRA
Am 580
4
200Æ71 (74)
NDa
1.4Æ1.8 (57)
1.7Æ0.5 (99)
NDa
11Æ7(93)
NDa
16Æ1
0 Æ0
5. (a) Heyman, R. A.; Mangelsdorf, D. J.; Dyck, J. A.; Stein,
R. B.; Eichele, G.; Evans, R. M.; Thaller, C. Cell 1992, 68,
397. (b) Levin, A. A.; Sturzenbecker, L. J.; Kazmer, S.; Bosa-
kowski, T.; Huselton, C.; Allenby, G.; Speck, J.; Kratzeisen,
C.; Rosenberger, M.; Lovey, A.; Grippo, J. F. Nature 1992,
355, 359.
6. Elder, J. T.; Fisher, G. J.; Zhang, Q.-Y.; Eisen, D.; Krust,
A.; Kastner, P.; Chambon, P.; Voorhees, J. J. J. Invest. Der-
matol. 1991, 96, 425.
7. Reczek, P. R.; Ostrowski, J.; Yu, K.-L.; Chen, S.; Hammer,
L.; Roalsvig, T.; Starrett, J. E., Jr.; Driscoll, J. P.; Whiting, G.;
Spinazze, P. G.; Tramposch, K. M.; Mansuri, M. M. Skin
Pharmacol. 1995, 8, 292.
8. Beard, R. L.; Chandraratna, R. A. In Handbook of
Experimental Pharmacology, Retinoids The Biochemical and
Molecular Basis of Vitamin A and Retinoid Action; Nau, H.,
Blaner, W. S., Eds.; Springer: Berlin Heidelberg, 1999; Vol.
139, p 185.
aND, not determined. Efficacy (%) is the percent inhibition of BrdU
incorporated in dividing cells at the highest dose tested (1 mM) relative
to vehicle control.
non-selective retinoid agonists. Such compounds will
have improved therapeutic indices relative to the non-
selective retinoids currently available and greater utility
in treating cancers and other serious retinoid responsive
diseases. Compound 4 (AGN 195183) is currently in
Phase I/IIA clinical trials in cancer patients.
Acknowledgements
9. Kagechika, H.; Kawachi, E.; Hashimoto, Y.; Himi, T.;
Shudo, K. J. Med. Chem. 1988, 31, 2182.
We are grateful to Mr. Hai Nguyen for mass spectral
analyses.
10. Renaud, J.-P.; Rochel, N.; Ruff, M.; Vivat, V.; Chambon,
P.; Gronemeyer, H.; Moras, D. Nature 1995, 378, 681.
11. Teng, M.; Duong, T. T.; Klein, E. S.; Pino, M. E.; Chan-
draratna, R. A. S. J. Med. Chem. 1996, 39, 3035.
1
12. All new compounds gave H NMR, 13C NMR, IR, and
References and Notes
HRMS data consistent with their structures.
13. Qing, F.-L.; Fan, J.; Sun, H.-B.; Yue, X.-J. J. Chem. Soc.,
Perkin. Trans. 1 1997, 20, 3053.
14. (a) Transactivation assay: Benbrook, D.; Lernhardt, E.;
Pfahl, M. Nature 1988, 333, 669. (b) Graupner, G.; Malle, G.;
´
Maignan, J.; Lang, G.; Prunieras, M.; Pfahl, M. BBRC 1991,
179, 1554. Binding assay: Allegretto, E. A.; McClurg, M. R.;
Lazarchik, S. B.; Clemm, D. L.; Kerner, S. A.; Elgrot, M. G.;
Boehm, M. F.; White, S. K.; Pike, J. W.; Heyman, R. A.
J. Biol. Chem. 1993, 268, 26625.
15. Fitzgerald, P.; Teng, M.; Chandraratna, R. A. S.; Hey-
man, R. A.; Allegretto, E. A. Cancer Res. 1997, 57, 2642.
1. Review: Sporn, M. B., Roberts, A. B., Goodman, D. S.,
Eds. The Retinoids: Biology, Chemistry, and Medicine; 2nd ed.;
Raven: New York, 1994.
2. (a) Huang, M. E.; Ye, Y. C.; Chen, S. R.; Chai, J. R.; Lu,
J. X.; Zhao, L.; Gu, L. J.; Wang, Z. Y. Blood 1988, 72, 567. (b)
Motzer, R. J.; Schwartz, L.; Law, T. M.; Murphy, B. A.;
Hoffman, A. D.; Albino, A. P.; Vlamis, V.; Nanus, D. M. J.
Clin. Oncol. 1995, 13, 1950. (c) Duvic, M.; Hymes, K; Heald,
P.; Breneman, D.; Martin, A. G.; Myskowski, P.; Crowley, C.;
Yocum, R. C. J. Clin. Oncol. 2001, 19, 2456.
3. (a) Agnish, N. D.; Kochhar, D. M. In Retinoids and Clinical