Practical synthesis of a chromene analog for use as a retinoic acid receptor alpha antagonist lead compound

Article abstract of DOI:10.1016/j.ejmech.2013.02.012

Retinoic acid receptor alpha (RARα) selective compounds may guide the design of drugs that can be used in conjunction with hormonal adjuvant therapy in the treatment of breast cancer. Herein we report a modified synthesis of a known RARα antagonist, 2-fluoro-4-[[[8-bromo-2,2-dimethyl-4-(4- methylphenyl)chroman-6-yl]carbonyl]amino]benzoic acid and a synthesis of its unknown, desfluoro analog, 4-[[[8-bromo-2,2-dimethyl-4-(4-methylphenyl)chroman- 6-yl]carbonyl]amino]benzoic acid. The modified route allows for facile reaction workups, increased yields, lower cost and incorporates a green alternative step. Structure-activity relationship studies determined through functional cell-based assays, demonstrated antagonism to RARα for both compounds. Molecular modeling within the RARα binding pocket was used to compare binding interactions of the desfluoro analog to a known RAR antagonist.

Full text of DOI:10.1016/j.ejmech.2013.02.012

European Journal of Medicinal Chemistry 63 (2013) 104e108
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Preliminary communication
Practical synthesis of a chromene analog for use as a retinoic acid receptor alpha
antagonist lead compound
,
Rachael Jetson a, Neha Malik a, Amarjit Luniwal a, Venkatesh Chari b, Manohar Ratnam b, Paul Erhardt a
*
a Center for Drug Design and Development, Department of Medicinal and Biological Chemistry, College of Pharmacy and Pharmaceutical Sciences, University of Toledo, Toledo,
OH 43606, USA
b Department of Biochemistry and Cancer Biology, College of Medicine and Life Sciences, University of Toledo, Toledo, OH 43606, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Retinoic acid receptor alpha (RARa) selective compounds may guide the design of drugs that can be used
Received 4 December 2012
Received in revised form
7 February 2013
Accepted 12 February 2013
Available online 19 February 2013
in conjunction with hormonal adjuvant therapy in the treatment of breast cancer. Herein we report a
modified synthesis of known RAR antagonist, 2-fluoro-4-[[[8-bromo-2,2-dimethyl-4-(4-
methylphenyl)chroman-6-yl]carbonyl]amino]benzoic acid and a synthesis of its unknown, desfluoro
analog, 4-[[[8-bromo-2,2-dimethyl-4-(4-methylphenyl)chroman-6-yl]carbonyl]amino]benzoic acid. The
modified route allows for facile reaction workups, increased yields, lower cost and incorporates a green
alternative step. Structureeactivity relationship studies determined through functional cell-based assays,
demonstrated antagonism to RARa for both compounds. Molecular modeling within the RARa binding
pocket was used to compare binding interactions of the desfluoro analog to a known RAR antagonist.
Ó 2013 Elsevier Masson SAS. All rights reserved.
a
a
Keywords:
Breast cancer
Hormone therapy
Retinoic acid receptor
Synthesis of antagonists
1. Introduction
In functional assays for the three retinoic acid receptors, the
des- and mono-fluoro compounds 1 and 2 exhibited nearly equal
Hormonal adjuvant therapy using estrogen antagonists, selec-
tive estrogen receptor modulators (e.g. tamoxifen) or inhibitors of
estrogen synthesis (e.g. aromasin) are a mainstay in the treatment
of breast cancer in the majority of patients [1,2]. However, in many
cases hormonal adjuvants are only tumoristatic wherein the re-
sidual tumor cells can sustain a basal level of cell growth that may
enable the generation and progression of genetic or epigenetic
changes that eventually lead to the emergence of drug resistant
tumors [2e5]. Recent evidence indicates that in estrogen-
dependent breast cancer cells, the retinoic acid receptor alpha 1
inhibitory potency toward RAR
greater than that of the di-fluoro compound 3. Higher dosing was
required to inhibit RAR and RAR . Between the chromene pair,
the mono-fluoro analog 4 was again about 10 times more potent
than the di-fluoro analog 5 at RAR . Both of the chromenes
exhibited enhanced selectivity compared to series 1e3, with 5
showing the highest overall [6]. Taking advantage of this pre-
ceding work, we wanted to further assess the possibility that a
chromene type of compound might be able to serve as a lead for
a that was approximately 10-fold
b
g
a
elaboration of antagonists at RARa. For practical reasons we
(RAR
support a basal level growth in the absence of hormone or in the
presence of tamoxifen. RAR 1 is genetically redundant in normal
tissues. It is consistently expressed in breast tumors. Targeted
inactivation or depletion of RAR 1 in cancer cells could potentially
a
1) activates growth genes in a ligand-insensitive manner to
sought a compound that could be produced from readily avail-
able starting materials. The preceding structureeactivity re-
lationships suggest that the des-fluoro analog of 4 could
potentially fulfill the desired biological profile while taking
advantage of ethyl p-aminobenzoate as a common starting ma-
terial amenable to considerable chemical modification. We
report herein: (i) An improved synthesis of 4 along with the first-
time synthesis of its des-fluoro analog; (ii) Comparison of their
activities as antagonists of the classical transcriptional mecha-
a
a
enhance current hormonal adjuvant therapies [2].
Teng et al. have identified an interesting series of antagonists
that are selective for RARa compared to RARb and RARg [6]. These
structures are shown in Fig. 1.
nism of RAR
A tentative binding mode for the des-fluoro analog within the
RAR 1 pocket derived from X-ray data coupled with preliminary
molecular modeling studies.
a using a cell-based promoter activity assay; and, (iii)
* Corresponding author. 2801 W. Bancroft, Toledo, OH 43606, MS 606, USA.
Tel.: þ1 419 530 2167; fax: þ1 419 530 7946.
a
E-mail address: paul.erhardt@utoledo.edu (P. Erhardt).
0223-5234/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2013.02.012

R. Jetson et al. / European Journal of Medicinal Chemistry 63 (2013) 104e108
105
As anticipated from analogous literature methods [9] and
previous experiments within our labs, coupling of the isoprene
synthon 17 with phenol 18 was accompanied by intramolecular
cyclization to afford chromene 19 in excellent yield [7]. Catalytic
hydrogenation readily afforded 11. This two-step sequence can be
accomplished conveniently in 4 days to provide pure material in
70% yield after 1 flash column, whereas in our hands the published
route across 5 steps required 7 days and 2 flash columns to pro-
vide pure 11 in 56% yield. Subsequent steps paralleled those of the
published method in terms of conditions and reasonable yields,
namely bromination followed by oxidation to afford 13. The next
step involves the Grignard reaction and dehydration procedure
that occurs in less than 10% yield, namely step (h) within Scheme
1. Consideration of the mechanism for this reaction led us to
suspect that the first equivalent of Grignard reagent becomes
spent by serving as a base to extract the acidic proton alpha to the
carbonyl which, in turn, leads to ring opening. Our proposed
mechanism is depicted in Fig. 2. Intentional production and
isolation of the otherwise transient ring-opened intermediate 22
was confirmed by proton NMR (in the protonated form provided
through an aqueous work-up). By encouraging its formation
before adding the entire compliment of Grignard reagent followed
by the dehydration and concomitant re-closure of the ring, we
were able to significantly increase the yield of 14. In the end, these
steps can be accomplished within a single pot while using TLC to
conveniently monitor the formation and then disappearance of 22.
While the prior process required 2 column chromatography pu-
rifications and resulted in a 10% yield, the modified procedure
provides a 55% yield after crystallization without the need for a
column.
Y
O
X
Y
O
X
O
OH
O
OH
N
H
N
H
O
Br
1 X = Y = H
2 X = H, Y = F
3 X = Y = F
4 X = H, Y = F
5 X = Y = F
Fig. 1. Selective RARa1 antagonists developed by Teng et al. [6].
2. Results and discussion
2.1. Synthetic chemistry
The reported synthetic route [6] to 4 as deployed within our lab
is depicted in Scheme 1. We were able to repeat this sequence in an
uneventful manner, achieving similar yields at each step except for
the bromination and Grignard reactions (steps f and h, respec-
tively). While lower than the reported quantitative yield, we felt
that 71% was still quite acceptable during bromination. Alterna-
tively, we suspected that the Grignard reaction might be able to be
enhanced by optimizing conditions conducive to its overall mech-
anism. We also noted that pyridinium p-toluenesulfonate (PPTS)
was better for dehydration to 14. Although reported at 25%, in our
case we were able to achieve only a 10% yield of crude intermediate
14 (NMR), making this process the lowest yielding step in the
overall sequence. In addition to addressing the Grignard reaction,
we imagined that the initial sequence of reactions leading to in-
termediate 11 might be accomplished in fewer steps by taking
advantage of a high-yielding approach to various chromenes that
we had utilized previously within our labs [7]. Finally, we also felt
that it would be worthwhile to substitute the reagents used in step
(j) with a composite that reflects greener chemistry [8]. Our
modified synthetic route is depicted in Scheme 2.
Hydrolysis of ester 14 followed reported conditions. Activation
of the exposed carboxyl group, as a prelude to coupling in step (g),
utilized propanephosphonic acid anhydride (T3PÒ) rather than
ethyl-dimethylaminopropyl-carbodiimide (EDC). T3PÒ is more
environmentally friendly because it leads to easier workup with
fewer side-products [8]. Yields were comparable when either
O
(b)
(d)
HO
(a)
(c)
O
O
O
O
HO
9
8
6
7
O
O
O
O
O
O
OEt
OEt
OH
OEt
(e)
(f)
(g)
O
O
O
Br
13
Br
12
10
11
O
F
O
F
O
OEt
O
OH
O
(j)
(k)
(h)
N
H
N
H
OR
O
O
O
O
OEt
16
Br
4
Br
Br
H2N
F
R = Et, 14
R = H, 15
(i)
Scheme 1. Synthesis of 4 [6]. Reagents, conditions and [yields]: (a) MeMgCl, THF, overnight, [100%]; (b) 15% H2SO4, toluene, 2 h, [95%]; (c) CH3COCl, AlCl3, DCM, 30 min, [65%]; (d)
5% NaOCl, dioxanes, NaOH, 85 ꢀC, 4 days [86%]; (e) H2SO4, EtOH, 95 ꢀC, overnight, [95%]; (f) AcOH, Br2, overnight, [71%]; (g) CrO3, AcOH, Ac2O, benzene, 0 ꢀC, 4 h, [70%]; (h) (1) p-
tolylMgBr, THF, ꢁ78 ꢀC to rt; (2) PPTS, MeOH, reflux, 4 h, [<10% over 2 steps]; (i) 20% NaOH, EtOH, THF, [95%]; (j) EDC, DMAP, DCM, rt, overnight, [39%]; (k) 2 M NaOH, EtOH, [65%].

106
R. Jetson et al. / European Journal of Medicinal Chemistry 63 (2013) 104e108
O
O
O
O
O
OEt
OEt
OEt
(a)
(b)
(c)
+
HO
O
O
18
11
19
17
O
O
O
O
O
OEt
OEt
OEt
(e)
(d)
(g)
O
O
O
R
Br
13
Br
Br
OEt
12
R = Et, 14
R = H, 15
(f)
H2N
R = H, F
O
O
O
OEt
(h)
O
OH
NH
R
N
H
R
O
O
Br
Br
R = F, 16
R = H, 20
R = F, 4
R = H, 21
Scheme 2. Modified synthesis of 4 and first synthesis of 21. Reagents, conditions and [yields]: (a) 3-Picoline, anhyd. xylene, 125 ꢀC, 70 h, [86%]; (b) 35 psi H2, Pd/C, rt, overnight,
EtOAc, [82%]; (c) AcOH, Br2, overnight, [71%]; (d) CrO3, AcOH, Ac2O, benzene, 0 ꢀC, 4 h, [70%]; (e) (1) p-tolylMgBr, THF, ꢁ78 ꢀC; (2) p-tolylMgBr, THF, ꢁ78 ꢀC; (3) PPTS, MeOH, reflux,
4 h, [55% over 3 steps]; (f) EtOH, THF, NaOH, 45 ꢀC, 2 h, [95%]; (g) T3P/EtOAc, reflux, overnight, [35%, 65%]; (h) 2 M NaOH, EtOH, [65%, 90%].
reagent was tried with the same aryl amine partner. Significant
differences were noted between the latter, however, with the
fluorinated analog being less reactive and suffering from much
lower overall yield (ca. 35%) than the des-fluoro analog (ca. 65%).
This difference in reactivity is not surprising since the aryl amine
becomes much less nucleophilic when the ring is substituted
by an electron withdrawing fluorine atom. The accompanying
improvement in overall yield for the des-fluoro analog, however,
became quite fortuitous with our initial desire to deploy the latter
as a coupling partner that would itself be highly amenable to
chemical modification. Finally, hydrolysis of the penultimate ethyl
esters, 16 and 20, in step (h) followed the reported conditions to
provide known product 4 and new product 21 in about 65% and 90%
yields, respectively.
Fig. 2. Proposed mechanism for the Grignard reaction shown in step (e) of Scheme 2.

R. Jetson et al. / European Journal of Medicinal Chemistry 63 (2013) 104e108
107
800000
700000
600000
500000
400000
300000
200000
100000
0
derivatives that can selectively attenuate the ligand-insensitive
non-classical mechanism of action of RARa.
2.3. Molecular modeling
Molecular modeling studies were conducted to dock 21 into a
model of the RAR 1 pocket that was derived from an X-ray struc-
a
ture of the protein bound with a known antagonist 23 [10]. The
latter’s chemical structure is depicted within Panel A of Fig. 4. Its
bound arrangement is shown within Panel B. Panel C shows a
super-position of both 21 and 23 docked into this pocket while
Panel D shows just 21. Both compounds orient their carboxylic acid
moiety so as to be aligned for interactions with Arg-276 and Ser-
287 located in the Northwest corner of the receptor pocket. Both
compounds likewise take advantage of their amide NeH to form a
ꢀ
Fig. 3. Antagonist effects of 4 and 21 on the classical ligand-sensitive promoter acti-
vation by RAR
vehicle control. Error bars reflect 1 standard deviation from the mean derived from
triplicate experiments.
strong (<3 A) hydrogen bond with Ser-232 which is centrally
a in MCF7 breast cancer cells. AtRA is all trans-retinoic acid; veh is
located within the pocket. Interestingly, moving to the Eastern edge
of the pocket each compound then displays its more lipophilic aryl
moiety in either a Northerly (23) or a Southerly (21) direction,
perhaps the bromine on the latter sufficing to capture any hydro-
phobic bonding advantage gained from a Northern orientation.
While intended as only a preliminary examination of the binding
process, these studies further demonstrate the utility of deploying
21 as a prototype for the rational design of antagonists that can
2.2. Biology
Target compounds 4 and 21 were tested for their antagonist
effects on the classical, retinoid-dependent, mechanism of tran-
scriptional activation in a promoter activity assay in MCF7 cancer
cells. This assay uses all-trans-retinoic acid (AtRA) (10 nM) as the
agonist. The degree of antagonism by the test compounds was
measured by quantifying the decrease in the activity of the
promoter-driven reporter luciferase. The results are depicted in
Fig. 3. While the des-fluoro analog 21 is about 100 times less potent
than the mono-fluoro compound 4, it follows a stepped dosee
response curve starting from 0.1 nM and is able to essentially
abolish the agonist effects of AtRA at a dose of 10 nM. The latter
represents a useful potency range for the future development of
interact with RARa1 within its receptor pocket.
3. Conclusions
In conclusion, we devised a modified synthesis for 4 and
implemented it for the preparation of the novel des-fluoro analog
21. In lieu of the previous reported synthesis we were able to
modify the procedure by: reducing the number of steps to form the
key chromene intermediate; mechanistically revaluating the
Grignard reaction so as to prompt a change in conditions that
Fig. 4. Panel A: RAR antagonist 23. Panel B: 23 docked within a model of RAR
model of RAR 1.
a1. Panel C: 21 and 23 superimposed in the RARa1 binding pocket model. Panel D: 21 docked within a
a

108
R. Jetson et al. / European Journal of Medicinal Chemistry 63 (2013) 104e108
significantly enhanced this low yielding step; and, adopting a
greener reagent for the amide coupling reaction. Furthermore, by
synthesizing the non-fluorinated analog we were able to reduce
reagent cost as well as increase reaction yields. Biological studies in
MCF7 breast cancer cells showed that 4 is a potent antagonist of
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2013.02.012.
ligand-dependent promoter activation by RARa at sub-nanomolar
References
potency while 21 is effective at low nanomolar potency. The
latter is within a range that is useful as a parent compound for the
synthesis of appropriate analogs to target the non-classical (ligand-
[1] M.D. Salazar, M. Ratnam, M. Patki, I. Kisovic, R. Trumbly, M. Iman, M. Ratnam,
During hormone depletion or tamoxifen treatment of breast cancer cells the
estrogen receptor apoprotein supports cell cycling through the retinoic acid
insensitive) mechanism of action of RARa.
receptor a1 apoprotein, Breast Cancer Res. 13 (2011) 1e17.
[2] A. Luniwal, R. Jetson, P. Erhardt, Selective estrogen receptor modulators
(SERMs), in: J. Fischer, C. Ganellin, D. Rotella (Eds.), Analogue Based Drug
Discovery, third ed., Wiley-VCH, Weinheim, Germany, 2013, pp. 165e185.
[3] R.L. Sutherland, E.A. Musgrove, Biological determinants of endocrine resis-
tance in breast cancer, Nat. Rev. Cancer 9 (2009) 631e643.
Molecular modeling suggests how the des-fluorinated analog
may be able to associate at the pertinent binding domains of the
RARa1 receptor binding pocket in a manner that is similar to the
known antagonist 23.
[4] B. Kuske, C. Naughton, K. Moore, K.G. Macleod, W.R. Miller, R. Clarke,
S.P. Langdon, D.A. Cameron, Endocrine therapy resistance can be associated with
high estrogen receptor alpha (ERalpha) expression and reduced ERalpha phos-
phorylationinbreastcancermodels, Endocr. Relat. Cancer 13(2006) 1121e1133.
[5] A.W. Millar, K.P. Lynch, Rethinking clinical trials for cytostatic drugs, Nat. Rev.
Cancer 3 (2003) 540e545.
Funding
Funding was provided by the National Institute of Health under
a NIH R01 CA140690 grant to Dr. Manohar Ratnam.
[6] M. Teng, T.T. Duong, A.T. Johnson, E.S. Klein, L. Wang, B. Khalifa,
R.A.S. Chandraratna, Identification of highly potent retanoic acid receptor
selective antagonists, J. Med. Chem. 40 (1997) 2445e2451.
a-
[7] A. Luniwal, R. Khupse, M. Reese, J. Liu, M. El-Dakdouki, N. Malik, L. Fang, P. Erhardt,
Multigram synthesis of glyceollin I, Org. Process Res. Dev. 15 (2011) 1149e1162.
[8] F.L. Zumpe, M. Flüß, K. Schmitz, A. Lender, Propane phosphonic acid anhy-
Notes
dride:
a new promoter for the one-pot biginelli synthesis of 3,4-
The authors declare no competing financial interest.
dihydropyrimidin-2(1H)-ones, Tetrahedron Lett. 48 (2007) 1421e1423.
[9] J.T. North, D.R. Kronenthal, A.J. Pullockaran, S.D. Real, H.Y. Chen, Synthesis of
6-cyano-2,2-dimethyl-2H-1-benzopyran and other substituted 2,2-dimethyl-
2H-1-benzopyrans, J. Org. Chem. 60 (1995) 3397e3400.
[10] P. Germain, C. Gaudon, V. Pogenberg, S. Sanglier, A. Van Dorsselaer, C.A. Royer,
M.A. Lazar, W. Bourguet, H. Gronemeyer, Differential action of coregulator
interaction defines inverse retinoid agonists and neutral antagonists, Chem.
Biol. 16 (2009) 479e489.
Acknowledgments
The authors of this paper would like to thank Dr. Yong Wah Kim
for maintenance of the NMR instruments and guidance in their use.