Synthesis, structure-activity relationships, and RARγ-ligand interactions of nitrogen heteroarotinoids

Article abstract of DOI:10.1021/jm9900974

Three heteroarotinoids containing a nitrogen atom in the first ring and a C-O linking group between the two aryl rings were synthesized and evaluated for RAR and RXR retinoid receptor transactivation, tumor cell growth inhibition, and transglutaminase (TG

Full text of DOI:10.1021/jm9900974

3602
J . Med. Chem. 1999, 42, 3602-3614
Syn th esis, Str u ctu r e-Activity Rela tion sh ip s, a n d RARγ-Liga n d In ter a ction s of
Nitr ogen Heter oa r otin oid s
Arindam Dhar,† Shengquan Liu,‡ J ozef Klucik,‡ K. Darrell Berlin,*,‡ Matora M. Madler,‡ Shennan Lu,†,
R. Todd Ivey,† David Zacheis,∇ Chad W. Brown,‡ E. C. Nelson,| Paul J . Birckbichler,§ and Doris M. Benbrook*,†,
Departments of Obstetrics & Gynecology, of Biochemistry & Molecular Biology, of Otorhinolaryngology, and of Urology,
University of Oklahoma Health Sciences Center, P.O. Box 26901, Oklahoma City, Oklahoma 73190, and Departments of
Chemistry and of Biochemistry & Molecular Biology, Oklahoma State University, Stillwater, Oklahoma 74078
Received March 4, 1999
Three heteroarotinoids containing a nitrogen atom in the first ring and a C-O linking group
between the two aryl rings were synthesized and evaluated for RAR and RXR retinoid receptor
transactivation, tumor cell growth inhibition, and transglutaminase (TGase) induction. Ethyl
4-(N,4,4-trimethyl-1,2,3,4-tetrahydroquinolinyl)benzoate (1) contained an N-CH3 group and
activated all retinoid receptors except for RARγ. Inceasing the hydrophobicity around the rings
with analogues ethyl 4-(N,4,4,7-tetramethyl-1,2,3,4-tetrahydroquinolin-6-oyloxy)benzoate (2)
[7-methyl group added] and ethyl 4-(4,4-dimethyl-N-isopropyl-1,2,3,4-tetrahydroquinolin-6-
oyloxy)benzoate (3) [NCH(CH3)2 group at C-4] increased the potency and specificity for RARR,
RARâ, and RXRR, compared to 1, but had little effect on RXRâ and RXRγ activation. Although
1 and 3 were unable to activate RARγ, 2 did activate this receptor with efficacy and high potency
equal to that of 9-cis-retinoic acid (9-c-RA). All three heteroarotinoids exhibited 5-8-fold greater
specificities for RARâ over RARR. In addition, esters 1-3 inhibited the growth of two cell lines
each derived from cervix, vulvar, ovarian, and head/neck tumors with similar efficiencies to
that of 9-c-RA through a mechanism independent of apoptosis. The vulvar cell lines were the
most sensitive, and the ovarian lines were the least sensitive. Ester 2 was similar to 1 and 3
except that 2 was a much more potent growth inhibitor of the two vulvar cell lines, which is
consistent with strong RARγ activation by 2 (but not by 1 and 3) and the high levels of RARγ
expression in skin. All three heteroarotinoids induced production of TGase, a marker of retinoid
activity in human erythroleukemic cells. Esters 2 and 3 were the more potent TGase activators
than 1, in agreement with the stronger activation of the RAR receptors by 2 and 3. The biological
activities of these agents, and the RARγ potency of 2 in particular, demonstrate the promise
of these compounds as pharmaceutics for cancer and skin disorders.
In tr od u ction
appear to exhibit significantly higher biological activity
than the oxygen-containing counterparts.3 In an effort
to enhance useful activity in the heteroarotinoid system,
the nitrogen-containing compounds 1-3 were obtained
and evaluated with respect to biological activity and
receptor specificity. These systems possess 2-atom link-
Retinoids are analogues of vitamin A with promise
as pharmaceuticals for cancer and other diseases, but
there are limitations due to toxicity even for the endog-
enous trans-retinoic acid (t-RA), 9-cis-retinoic acid (9-
c-RA), and 13-cis-retinoic acid (13-c-RA).1 Therapeutic
potential of structurally related stilbene-like analogues,
such as TTNPB, is compromised by a maximum toler-
ated dose (MTD) which is 1000-fold smaller compared
to that of t-RA, a metabolite of vitamin A.2,3 It has been
shown that the insertion of an oxygen atom into the
partially saturated ring in TTNPB decreased the toxic-
ity to an MTD equal to that of t-RA.3 However, the
biological activity of the oxygen-containing heteroaro-
tinoid was only 31% as efficacious in transactivating a
retinoic acid response element (RARE) as compared to
t-RA.3 Heteroarotinoids (structurally related to the
clinical Etretinate)1 containing a sulfur or nitrogen atom
* Corresponding authors. For biology: e-mail, Doris-Benbrook@
OUHSC.edu. For chemistry and modeling: e-mail, kdberlin@bmb-
fs1.biochem.okstate.edu.
† Department of Obstetrics & Gynecology.
‡ Department of Chemistry.
§ Department of Urology.
∇ Department of Otorhinolaryngology.
Department of Biochemistry & Molecular Biology, OUHSC.
| Department of Biochemistry & Molecular Biology, OSU.
10.1021/jm9900974 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/19/1999

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3603
Sch em e 1. Synthesis of 2
ers (C-O) between the two rings, a bonding situation
which increases the flexibility of the heteroarotinoids
and increases biological activity in comparison with
analogues containing alkenyl linkers.3
Ca2+-dependent enzymes which catalyze posttranscrip-
tional modifications of proteins by introducing intermo-
lecular ꢀ-(γ-glutamyl)lysine isopeptide bonds.10,11 Re-
cently, it was observed that certain nitrogen- and sulfur-
containing heteroarotinoids also induced TGase produc-
tion in HEL cells and, in fact, exerted greater induction
than some oxygen-containing analogues.3
The effectiveness of the three nitrogen-containing
heteroarotinoids described herein against solid tumors
was assessed by determining their abilities to inhibit
growth and induce a natural form of cell death (called
apoptosis) in each of the two cell lines derived from
cervical, vulvar, ovarian, and head and neck tumors.
The size of the tumor depends on the rate of growth
versus apoptosis in its cell population. Retinoids have
the potential to cause tumor shrinkage by inhibiting
growth and inducing apoptosis. Some retinoids are good
growth inhibitors and poor inducers of apoptosis, while
others are poor growth inhibitors but potent inducers
of apoptosis.12 Both types of retinoids have the potential
to shrink tumors by altering the balance between
growth and apoptosis.11
Retinoids act through two classes of nuclear receptors,
RARs and RXRs, each of which contain subtypes clas-
sified as R, â, and γ. The arotinoid TTNPB, like t-RA,
activates the RARs but not the RXRs. Substitution of a
hydrogen atom by a methyl group at C-3 (position
equivalent to C-7 in the heteroarotinoids) of TTNPB
confers RXR activation capability to the molecule.4,5 The
RXR selectivity of the 3-methylated stilbene analogues
may result from the change in energy barrier for
rotation by the aryl rings around the linker group, thus
generating a slight change in the average conformation
of the entire system. In this work we report the effect
of the corresponding methyl group at C-7 on the receptor
activation profile of heteroarotinoid 2. Expectedly, the
presence of such a hindering methyl group will decrease
the flexibility around the C-O linker and alter receptor
specificity.
The presence of a lipophilic group at C-4 extending
outward and away from the plane of the ring appears
to be necessary for maximum activation of RAR recep-
tors.6 Substitution of an oxygen atom for C-4 resulted
in weak binding3 to the RAR receptors, while substitu-
tion with a methyl, isopropyl, or sulfur moiety at the
same position might cause strong binding to RAR
receptors.3,7 Therefore, more effective screening of a
nitrogen atom at the same position by a nonpolar
isopropyl group, such as in 3, was reasoned to result in
reduced polarity exposed from this position and im-
proved hydrophobic bonding to the receptor.
Ch em istr y
The synthesis of 1 was described previously.3 Syn-
theses for compounds 2 and 3 are outlined in Schemes
1 and 2, respectively. A major modification in Scheme
2 as shown for conversions 4 f 5 f 6 f 7 f 8 f 9 f
10 f 2 resulted when direct lithiation of 8 failed, and
an alternative approach was required to eventually
produce acid 10. N-Acylation of 4 (Scheme 1), followed
by ring closure of 5 via a solid-phase reaction with AlCl3,
gave lactam 6. Reduction of 6 to amine 7 was facile
although the purification of 7 required chromatography.
N-Methylation of 7 was modestly successful and gener-
ated low-melting 8. An improved procedure to obtain 8
was also realized by first converting lactam 6 to the
N-methyl analogue 11 which was more readily reduced
to 8. This proved the structure of 8. Replacement of the
bromine atom in 8 by a nitrile group was achieved using
CuCN/DMF as shown and gave 9 in high yield. Hy-
drolysis of 9 to acid 10 required several days, and the
yield was modest (38%). Esterification of 10 occurred
In this report, three nitrogen-containing heteroaro-
tinoids were evaluated for anticancer activity against
cell lines derived from leukemia and solid tumors.
Leukemia is currently being treated with t-RA which
causes remission of acute promyelocite leukemia (APL)
by inducing differentiation and retarding growth of APL
cells.8 It has been found that treating some human
erthyroleukemia (HEL) cells with t-RA results in cell
differentiation accompanied by a 9-fold increase in
tissue transglutaminase (TGase) activity.9 TGases are

3604 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Sch em e 2. Synthesis of 3
Dhar et al.
Ta ble 1. Potency (EC50, nM) and Efficacy (% of 9-c-RA) of Retinoid Receptor Transactivation and TGase Induction by the Three
Nitrogen-Containing Heteroarotinoids 1-3
RAR
RXR
heteroarotinoid
R
â
γ
R
â
γ
TGase induction
a
1 EC50
1128
45
256
64
NA
0
601
47
33
53
20
52
ND
37
% effb
a
2 EC50
796
64
92
63
6
103
102
53
70
55
49
52
ND
49
% effb
a
3 EC50
217
59
41
71
NA
0
12
62
47
45
27
47
ND
49
% effb
a
b
The potency is the concentration of the agent that induces half of the agent’s maximal activity (not relative to 9-c-RA). The efficacy
was derived by dividing the maximal activity of the heteroarotinoid by the maximal activity of 9-c-RA for the receptor transactivation
assay and by the maximal activity of t-RA for the TGase assay. The maximal activity occurred at the maximal concentration used (10 µM)
for all three heteroarotinoids and for both assays. ND, not done; NA, not active.
by the usual method, but again the conversion was
apparently hindered by the 7-methyl group which in
turn reduced the yield of ester 2.
Heteroarotinoid 3 was obtained as described in Scheme
2. The sequence of reactions involving 12 f 13 f 14 f
15 f 16 f 17 f 18 f 3 paralleled to some degree that
shown in Scheme 1 for 2. The conversion of bromide 16
to the nitrile 17 was much more facile than the
counterpart in Scheme 1, apparently due to the absence
of the C-7 methyl group in 16.
The rationale for the comparison of 1 and 2 was based
on the premise that a decrease in flexibility around the
C-O bond due to the presence of the 7-methyl group,
as in 2, would alter the RAR receptor binding capability
of the molecule and therefore the activity of the ligand-
receptor complex formed. More effective screening of the
nitrogen atom by an isopropyl group, as in 3, was
reasoned to reduce the polarity at this position for
improved hydrophobic bonding in the receptor for RARs
as activated by t-RA. TTNBP, an arotinoid with strong
anticancer activity, is also nonpolar at C-1 and C-4, for
example. Etretinate, a clinically used retinoid,1 is also
rather nonpolar at the ring end of the molecule as is
true for 1-3.
at the cellular level for inhibition of tumor cell growth
and induction of TGase activity in HEL cells. The
standards used for comparison purposes was 9-c-RA and
TTNPB.
To evaluate the receptor activation, CV-1 cells were
cotransfected with the specific retinoic acid receptor
expression constructs and the âRARE-tk-CAT reporter
plasmid which has been shown to be activated by both
RAR/RXR heterodimers and RXR/RXR homodimers.13
Structural variations in 1-3 had distinct effects on
receptor-dependent transactivation as illustrated in
Table 1. Ester 1 transactivated the RARE through
RARR and RARâ but not through RARγ. The activation
of all three RXR receptors by this compound was not
entirely surprising since prior heteroarotinoids had
confirmed RXR activation capacity.3
Substitution of hydrogen by a methyl group at the
7-position significantly increased RARγ efficacy for 2 but
had minimal effects on RARR and RARâ efficacies as
compared to that noted in 1 and 3. Although RXRâ and
RXRγ were only slightly altered by the presence of the
7-methyl group in 2, RXRR potency was markedly
increased in comparison to 1. The latter is consistent
with the effects of this same substitution on TTNPB.3
The presence of the N-isopropyl group in 3 also in-
creased RARR, RARâ, and RXRR potencies compared
to 1, while exerting minimal effects on RARγ, RXRâ,
and RXRγ activation. Thus, the most striking effect of
Biology
The biological activities of 1-3 were assessed at the
molecular level for activation of nuclear receptors and

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3605
the 7-methyl substitution in 2 was conference of RARγ
activation with a potency and efficacy equal to that of
9-c-RA. All three compounds exhibited nearly equal or
greater specificity for RARâ over RARR.
Heteroarotinoids 1-3 were appraised for ability to
induce TGase (Table 1), a marker of retinoid efficacy in
leukemia cells.3 In agreement with the strong activation
of the retinoid receptors, 2 and 3 exerted a 49%
induction of TGase activity in HEL cells in comparison
to the 37% induction by 1. Treatment of HEL cells with
10 µM t-RA increased cellular differentiation, inhibited
cell proliferation, and increased tissue TGase.9-11 Bio-
logical effects were observed at retinoid concentrations
of 10 µM, and cell viability was greater then 90% in HEL
cultures with t-RA or the heteroarotinoids.9
Compounds 1-3 were screened for ability to inhibit
growth on a panel of two cell lines each derived from
cervix, vulva, ovary, and head and neck tumors (Figure
1). The vulvar cells were the most sensitive, and the
ovarian cells were the least sensitive to retinoid treat-
ment. In a majority of cases, the heteroarotinoids were
similar in efficacy to 9-c-RA. The efficacy for individual
compounds varied depending upon the cell line. The
most dramatic growth inhibition was exerted by 2 on
the two vulvar carcinoma cell lines, consistent with the
potent RARγ specificity of this compound and with the
fact that the retinoid effects in keratinizing skin, such
as vulva, are mediated mainly through the high levels
of RARγ expressed in this type of epithelium.14,15
The greater than 75% growth inhibition observed for
2 with the vulvar cell line indicates that cell loss might
be occurring due to apoptosis or necrosis. It is more
likely, however, that the mechanism in this experiment
is due to a slowing down of the cell cycle because there
were more cells present at the end of the 3-day treat-
ment period than at the initiation. This high degree of
growth inhibition without cell loss is possible for the
vulvar cell lines because such cells exhibited a rapid
growth rate with a doubling time of 30 h. Therefore, at
the end of 72 h of treatment in the experiment, the
control cultures had increased in cell number by ap-
proximately 4.5 times (12 000 cells). The cultures treated
with 2 contained approximately 25% of the cells present
in the control culture (3 000 cells) which is greater than
the 2 000 cells plated before the initiation of treatment.
F igu r e 1. Data from the percent of growth inhibition of
certain cells lines by 1-3, TTNPB, and 9-c-RA. Such cells from
various carcinoma cell lines were plated in microtiter plates
and treated with 10 µM of agents 1-3 or with TTNPB (T) or
9-c-RA (9-c) for 72 h. The concentration of cells at the end of
3 days was determined by SRB assay. Growth inhibition was
measured as the percent of cell loss in the treated cultures
relative to control cultures treated with solvent (DMSO).
To determine if induction of apoptosis was part of the
mechanism by which nitrogen-containing heteroaroti-
noids decrease tumor cell culture density, the effects of
2 and 3 were elevated on the most sensitive cell line,
namely SW962 (Figure 2). Agent 4HPR was used as a
positive control for retinoid induction of apoptosis, and
t-RA was used as a control for retinoid growth inhibition
in the absence of apoptosis.12 The apoptotic cells in these
cultures were quantitated using an ELISA assay for
enrichment of mono- and oligonucleosomes that result
from fragmentation of chromosomes by endogenous
nucleases during apoptosis. As expected in the treated
cells, 4HPR enriched apoptosis by more than 5-fold,
while 2, 3, and t-RA had little effect. This observation
classifies these nitrogen-containing heteroarotinoids as
members of the nonapoptotic retinoid group along with
t-RA.
F igu r e 2. Comparison of apoptosis in SW962 cells treated
with various retinoids or solvent. The absorbance indicates the
presence of mono- and oligonucleosomes in the cyctoplasm,
which is characteristic of apoptotic cells.
tions has not been documented with heteroarotinoids
to any appreciable extent. The availability of data from
the X-ray analysis of RARγ16 and RXRR17 was an
Molecu la r Mod elin g. The utility of molecular mod-
eling in attempts to determine ligand-receptor interac-

3606 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Dhar et al.
Ta ble 2. Energies of Interaction of Flexible Ligand-Fixed RARγ (kcal/mol)a
ligand
run1
run2
run3
13000
15000
tour
20
-34.71
16
16
run4
18000
30000
roul
15
-75.73
44
3
run5
seed•num
gen•num
meth•sco
sol•num
int•ener
% conv
sol•num
int•ener
% conv
5000
8000
23000
25000
roul
20
-64.67
40
10
20000
roul
20
-58.29
35
12
10000
roul
20
-33.56
7
20
t-RA
9-c-RA
-15.48
60
-9.321
47
-7.439
12
-17.33
82
-5.38
9
1
2
3
sol•num
int•ener
% conv
sol•num
int•ener
% conv
sol•num
int•ener
% conv
6
12
12.583
40
12
21.486
15
10
21.486
21
20
-29.51
14
2
14
22.98
11
15
22.98
3
8.351
61
8
2.473
40
18
5.371
74
7
5.371
78
20
8
3.618
33
20
-30.63
24
20
-37.52
40
-45.29
53
-25.6
0.0
a
Seed•num, seed number; gen•num, generation number; meth•sco, method of scoring; sol•num, solution number; int•ener, interaction
energy; % conv, percent convergence; roul, roulette method of scoring; tour, tournament method of scoring.
incentive to examine 1-3, t-RA, and 9-c-RA as ligands
for the rigid form of RARγ, as exemplified by the docking
site in RARγ with t-RA as the ligand in the X-ray
structure. In addition, more flexible forms of both the
receptor and the ligands were evaluated in terms of the
lowest energy conformations for both ligand and recep-
tor. The crystallographic structure of the crystallized
RARγ-ligand (t-RA) system was obtained from the Pro-
tein Data Bank (see Experimental Section). The binding
pocket of the RARγ was defined with the following
amino acid residues: Phe201, Trp227, Phe230, Ser231,
Leu233, Ala234, Lys236, Cys237, Ile238, Leu268, Leu271,
Met272, Arg274, Ile275, Arg278, Phe288, Ser289, Gly303,
Phe304, Pro306, Leu307, Gly393, Ala394, Arg396,
Ala397, Leu400, Met408, Ile412, and Met415 with a
radius of 10 Å around these amino acids.18
for all ligands. The energy minimization of ligands was
accomplished with SYBYL 6.5 using Tripos force field
with the Powell method of line direction search. Charges
were computed before each minimization and used as
“current” in the energy setup option box. The rmsd was
set to 0.001, and the gradient was set to 0.005 kcal/mol
Å. The conformational search was done with 9-c-RA
using simulated annealing where the molecule was
allowed to anneal at 300 K. More than 100 conforma-
tions of 9-c-RA were evaluated from this process, and
one conformer was selected which most closely re-
sembled the normal conformation of t-RA. The confor-
mational energy difference between the normal 9-c-RA
and that conformer which most closely mimics t-RA in
general appearance was 0.87 kcal/mol.
The calculations were performed using SYBYL 6.5
and employing the licensed “Biopolymer” which pos-
sesses a program designated “FlexiDock”. The Flexi-
Dock, SYBYL 6.5 release, was used for docking all
ligands into the binding pocket of RARγ. The param-
eters for the genetic logarithm of FlexiDock were used
as set in default by the manufacturer with the exception
of using Tournament instead of Roullete Wheel as a
scoring method in one round of calculations for each
ligand. For comparison purposes, the t-RA in the bind-
ing site of the crystallogaphic structure of RARγ con-
taining the crystallographic structure of t-RA was re-
moved, and independently t-RA was inserted and docked
as a check on the validity of the technique. For this
superimposition an rmsd value of 0.3464 (SD ) 0.1362)
was obtained which lends credibility to the methodology.
The energy of interaction between the crystal structure
of t-RA in the RARγ binding pocket was calculated to
be -73.592 kcal/mol with the crystal structure confor-
mational energy of t-RA at 64.243 kcal/mol. Taking the
crystal structure of RARγ and docking t-RA gave an
energy of interaction of -75.728 kcal/mol and a confor-
mational energy for t-RA of 23.872 kcal/mol.
Discu ssion
Results reported herein demonstrate that specific
substitutions on the heteroarotinoid structure signifi-
cantly alter the biological activity of the agents. For
example, modifications of structures 2 and 3, compared
to heteroarotinoid 1, resulted in increased receptor
activation with varying effects on growth inhibition,
depending upon the cell line. The variability in the
responses of the different cell lines represent the
complexity of the retinoid response pathway. Retinoids
have been shown to regulate not only gene expression
but also growth factor expression and the activities of
several protein kinases.19-21 Therefore, the response of
a cell line to retinoid treatment depends on the expres-
sion patterns of retinoid receptors and certain cofactors,
growth factor receptors, kinases, and phosphatases. The
addition of the 7-methyl group in 2 conferred RARγ
activation on a structure otherwise devoid of this type
of activity as in 1. The increased susceptibility of the
vulvar cell lines to agent 2, in comparison to the other
epithelial carcinomas, can be explained by the higher
levels of RARγ expression in skin over other epithelial
types.22 This observation is consistent with the RARγ
mediation of retinoid pharmacological action in skin.14,15
The EC50 value of 6 nM and the 103% efficacy of 2 (Table
2), in comparison to that of 9-c-RA, indicate that 2 may
be useful as a pharmaceutical agent for disorders of the
skin.
All ligands were drawn with carboxylate anions using
the fragment library of Insight Discover 97. The
Gasteiger-Huckel method was employed using SYBYL
6.5 for the charge assignment of each ligand, with
formal charges of -0.5 assigned to each oxygen atom of
the carboxylate group resulting in a total charge of -1.0

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3607
Ta ble 3. Energies of Interaction of Flexible Ligand-Flexible
RARγ (kcal/mol)a
Although agent 2 displayed strong RARγ activation,
2 did not dock rigid RARγ with an interaction energy
better than that of 3 or 9-c-RA. Possible explanations
for this is that the crystal structure may not reflect the
active form of RARγ-ligand or that some of the confor-
mations of the amino acids in the binding domain are
not so important for selectivity and activity of RARγ.
Changing the rigid RARγ-ligand system to a flexible
RARγ system, via an option in the FlexiDock program,
and then docking t-RA as a flexible ligand in the flexible
RARγ molecule gave a flexible RARγ-ligand molecule
with a conformation of RARγ which differed only
slightly from that in the crystal structure of RARγ-
ligand. The flexible RARγ system in our work refers to
the binding domain in which all amino acids have
freedom to move in all dimensions. An interaction
energy of -40.34 kcal/mol was noted for the flexible
RARγ-flexible ligand. This is strong indication of the
high stability of the complex. Superimposition of this
flexible-flexible system on the rigid RARγ-ligand
system produced an rmsd of 0.01 with respect to the
majority of amino acids. Some deviations from this rmsd
value were noted with certain amino acids, namely
Phe201 (0.1931), Lys236 (0.752), Cys237 (0.420), Arg278
(0.185), Ser289 (0.523), Leu400 (0.128), Ile412 (0.236),
and Met415 (0.121).
ligand
run1
run2
seed•num
gen•num
meth•sco
sol•num
int•ener
% conv
sol•num
int•ener
% conv
sol•num
int•ener
% conv
sol•num
int•ener
% conv
18000
23000
30000
roul
20
-40.34
60
12
-20.358
65
20
-57.008
88
20
-22.799
80
20
10000
roul
20
-23.56
33
t-RA
9-c-RA
20
- 9.756
9
20
-46.406
7
20
-17.94
15
20
-75.53
40
1
2
3
sol•num
int•ener
% conv
-78.18
85
a
Seed•num, seed number; gen•num, generation number;
meth•sco, method of scoring; sol•num, solution number; int•ener,
interaction energy; % conv, percent convergence; roul, roulette
method of scoring.
TGase induction in human myeloid leukemic cells is
mediated through a pathway involving RARs.23 On the
other hand, TGase induction in rat tracheobronchial
epithelial cells is mediated through a pathway involving
RXRR.14 These results may not be considered conflicting
since nuclear receptors function in cooperation as
dimers.24 While the RXRs can function both as homo-
dimers (RXR/RXR) and as heterodimers with RARs
(RAR/RXR), the RARs function only as heterodimers
(RAR/RXR). In agreement with this fact, myeloid cells
are induced to differentiate by a retinoid which selec-
tively activates RAR/RXR heterodimers, but not by a
retinoid which selectively activates RXR/RXR homo-
dimers.25 Therefore, if the major mediator of retinoid
effects on TGase induction is the RAR/RXR heterodimer,
then 9-c-RA, which activates both partners, would be
expected to exhibit greater activity than t-RA, which
activates only RARs. The more potent TGase induction
by 2 and 3, in comparison to that by 1, is consistent
with the greater receptor activation by 2 and 3. The
nearly equal efficacies of 2 and 3 on TGase induction,
despite the strong activation of RARγ by 2 but not by
3, suggest that RARγ either is not required or is
redundant in TGase induction of HEL cells. In conclu-
sion, the potency of the nitrogen-containing heteroaro-
tinoids 1-3 in receptor activation, tumor cell growth
inhibition, and TGase induction warrants toxicity test-
ing as the next step prior to clinical trials.
Illustrations from the modeling work are shown in
Figures 3 and 4. The docking experiments with 1-3
using parameters obtained from X-ray data (devoid of
the 119 water molecules) for the rigid RARγ system16
did not indicate a good correlation with the activity
observed. In addition, docking experiments with the 119
water molecules present in the rigid RARγ system did
not result in significant conformational changes in any
of the ligands or with the energy of interaction with the
exception of small differences in electrostatic energies.
Therefore, the following discussion is primarily based
upon results with a flexible RARγ system devoid of
water.
As further support for the contention that not all of
the amino acids in the region of the binding pocket are
responsbile for activation f binding, the flexible RARγ-
t-RA system was made “fixed” and the t-RA was
removed while agents 1-3 were then docked in succes-
sion as flexible ligands. No improvement was observed
in the interaction energies (data not included). In
contrast, using the flexible RARγ-flexible ligand model
resulted in all ligands docking favorably with favorable
negative interaction energies for ligand and receptor.
Figure 3 illustrates the various docking experiments
performed with flexible RARγ. For comparison purposes,
A (Figure 3) shows the binding pocket with amino acids
of rigid RARγ based on crystallographic data. Docking
flexible t-RA, 9-c-RA, 1, 2, and 3 with flexible RARγ led
to the various flexible ligand-bound RARγ systems B-F ,
respectively (Figure 3). Extraction of the ligands from
B-F exposed the individual binding pockets (not shown),
and superimposition of the binding pocket of t-RA on
those binding pockets for 9-c-RA and 1-3 gave images
G-J , respectively, as described in Figure 4. Small con-
formational differences in the binding pocket of RARγ
were observed after docking t-RA, both receptor and
ligand being in the flexible modes, as compared to the
resulting conformations observed when 1-3 were docked.
The flexible RARγ-ligand flexible 9-c-RA (annealed)
matched well with B to give the dual images in G. Like-
wise, agent 2 system matched well (I) with the exposed
flexible RARγ devoid of t-RA. In contrast 1 (H) and 3
(J ) clearly possess major deviations in terms of interac-
tions of these ligands with the amino acids in the
docking pocket of flexible RARγ. With 2, only deviations
greater than a rmsd value of 0.001 were noted, such as
with certain amino acids as Phe201 (0.397), Leu233
(0.805), Lys236 (1.192), Cys237 (0.42), Ile275 (0.185),
Arg278 (0.595), Ser289 (0.339), Phe304 (0.134), and
Met408 (0.478). Differences in the amino acid resi-
dues using 9-c-RA as the flexible ligand were Phe231

3608 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Dhar et al.
F igu r e 3. RARγ with bound ligands. A, Binding pocket with the amino acids listed of RARγ from X-ray diffraction data with t-RA extracted. B, Binding
pocket of flexible RARγ bound with flexible t-RA. C, Binding pocket of flexible RARγ bound with flexible 9-c-RA. D, Binding pocket of flexible RARγ bound
with flexible 1. E, binding pocket of flexible RARγ bound with flexible 2. F, Binding pocket of flexible RARγ bound with flexible 3.
(0.202), Lys236 (0.698), Phe230 (0.501), Ser231 (0.415),
Ser289 (0.378), Phe304 (0.143), Leu307 (0.263), Arg396
(0.151), Ile412 (0.462), and Ile275 (0.435). The larger
number of amino acid residues involved with larger
deviations in orientations observed with docked 1 and
3 is in agreement with both agents being RARγ inactive.
Superimposition of the ligand binding domains of RARγ,
each with the previously cited low-energy conformer of
9-c-RA and t-RA removed, had a total rmsd ) 0.22.
Cross-checking the superimposition results with vi-
sual inspection of the final conformations of RARγ and
ligands leads to the speculation that amino acids
Phe230, Met272, Leu271, and Ala234 are the most
important for RARγ activation which agrees with previ-
ous suggestions.16 The phenyl ring of Phe230 assumes
a different orientation with respect to the remaining
amino acid residues and with regard to inactive ligands
as compared to active t-RA and 2, for example (Figure

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3609
F igu r e 4. Binding pocket of flexible RARγ with t-RA removed and superimposed upon the binding pockets of RARγ with the ligands 9-c-RA, 1, 2, and 3
removed. G, Superimpositions of the binding pocket of B (magenta) upon the binding pocket of C (black) both devoid of the ligands t-RA and 9-c-RA, respectively.
The conformations of the following amino acids from the two pockets did n ot match with respect to Lys236, Phe201, Phe230, Ser231, and Ala394 with minor
deviations in Leu271 and Ile412. H, Superimpositions of the binding pocket of B (magenta) upon the binding pocket of D (black) both devoid of the ligands
t-RA and 1, respectively. The conformations of the following amino acids from the two pockets did n ot match with respect to Lys236, Phe201, Phe230, and
Arg396 with minor deviations in Ala234, Phe304, Ile412, and Met415. I, Superimpositions of the binding pocket of B (magenta) upon the binding pocket of
E (black) both devoid of the ligands t-RA and 2, respectively. The conformations of the following amino acids from the two pockets did n ot match with respect
to Lys236 with minor deviations in Met408, Ile412, Ser231, Leu271, Ile412, and Phe304. J , Superimpositions of the binding pocket of B (magenta) upon the
binding pocket of F (black) both devoid of the ligands t-RA and 3, respectively. The conformations of the following amino acids from the two pockets did n ot
match with respect to Phe201, Phe230, and Trp227 with minor deviations in Ala394, Ile412, Ala234, Arg278, and Lys236.

3610 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Dhar et al.
F igu r e 5. Listing of RARγ active and RARγ inactive structures as related to 1-3.
4). Possibly Phe230 acts as a “switch” for activation by
assuming different orientations with specific ligands.
With 9-c-RA, the phenyl ring of Phe230 lies in a position
somewhat between that found with t-RA and 2 (which
match well) as compared to the position observed in
inactive 1 and 3.
proper activation of the type initiated with t-RA, the
endogenous ligand. Consequently, the merit of this
hypothesis lies in the design and evaluation of future
agents to activate RARγ for specific transcription pro-
cesses, all of which await trial. It is interesting to note
that certain synthetic and highly modified retinoids with
an internal amide linkage, which is para to a carboxyl
group on a benzene ring, have been demonstrated to
induce differentiation in human promyelocytic leukemia
cells (HL-60), but no references were made to RARγ
activation.28,29
In conclusion, a decrease in flexibility around the
C-O bond in 2, due to the presence of the methyl group
at the 7-position, conferred RARγ activation capacity
in contrast to the parent 1. Molecular modeling using
X-ray crystallographic data of the ligand binding domain
of RARγ confirmed that 2, not 1 or 3, interacted with
the receptor in an energetically favorable fashion. The
potency of the nitrogen-containing heteroarotinoids 1-3
in receptor activation, tumor cell growth inhibition, and
TGase induction is comparable to that of 9-c-RA, and
the decreased toxicity of the heteroarotinoid class of
retinoids warrants toxicity testing of these agents as the
next step prior to clinical trials.
The role of Ala234, Met272, and Leu271 would appear
to be in the selectivity of the ligand for RARγ through
a series of van der Waal interactions with part of the
ligand. Since 2 differs from 1 by only the methyl group
at the 7-position, the additional nonpolar methyl group
may interact with hydrophobic regions of Ala234, Leu271,
and Met272. Such interaction may be responsible for
positioning the ligand in the binding cavity in such a
manner that the hydrophobic region of 2 is identical or
quite similar in position and interaction experienced by
the hydrophobic region in the binding of t-RA to RARγ.
With 1 and 3, this interaction does not occur with
Ala234 and Leu271, and the ligands are free to reposi-
tion themselves in the cavity. The end result is that the
ligands interact with the receptor in what appears to
be a favorable arrangement, but the phenyl ring of
Phe230 is rotated approximately 120°. Perhaps because
of this new orientation additional changes occur within
and on the surface of the receptor making the receptor
incapable of binding to transcription cofactors. As a
defense for this reasoning, attempts were made to dock
the retinoids shown in Figure 5 with our model flexible
RARγ. All ligands which were reported as RARγ active
docked well, and those reported26 to be inactive did not
dock well. Indeed, the position of the Phe230 matched
well in systems with ligands t-RA, 9-c-RA, and 2 as well
as with CD564, CD367, CD437, and CD666 but did not
match well with 1, 3, and Am80 (Figure 5).
Exp er im en ta l Section
Gen er a l. IR spectra were recorded on a Perkin-Elmer 2000
FT-IR as films or as KBr pellets. GC-MS data were obtained
from a HP G1800A GCD System, GC electron ionization
detector. FAB MS experiments were performed on a VG ZAB-
2SE HRMS unit, while EI MS experiments were accomplished
1
with an HP GC-MS ENGINE, model 5989B unit. All H and
13C NMR spectra were taken on a Varian Inova 400 MHz or a
Varian XL-400 MHz BB spectrometer at 399.92 and 100.57
Hz, respectively, and signals are referenced to TMS. Melting
points were determined with a Thomas-Hoover melting point
apparatus and are uncorrected. Syntheses were executed,
unless otherwise indicated, under an atmosphere of N2.
Elemenal analyses were performed by Galbraith Laboratories,
Knoxville, TN, or by Atlantic Microlab, Inc., Norcross, GA.
Although the intermediates and final heteroarotinoids ap-
peared to be relatively stable in light, precautions were taken
to minimize exposure of all products to any light source as
well as to the atmosphere. Pure heteroarotinoids 1,3 2, and 3
were stored in the cold and dark for indefinite periods without
significant decomposition.
The very recent report27 on the structure, via electron
diffraction studies, of the solid RARγ-9-c-RA complex
revealed that 9-c-RA does indeed assume a bent con-
formation which resembles very much the structure
proposed from the modeling work described herein.
Thus, it is encouraging that our modeling approach has
important potential for guiding the synthesis of future
modified retinoids with lower toxicity than the endog-
enous members. Thus, the modeling experiments appear
to be instructive with respect to “turning on” RARγ for

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3611
22.3 (Ar-CH3), 27.5 (Ar-CCH2), 33.6 (2 CH3), 45.0 (CH2),
Eth yl 4-(N,4,4,7-Tetr a m eth yl-1,2,3,4-tetr a h yd r oqu in o-
lin -6-oyloxy)ben zoa te (2). To a mixture of acid 10 (0.155 g,
0.665 mmol) and ethyl 4-hydroxybenzoate (0.165 g, 0.997
mmol) were added dicyclohexylcarbodiimide (DCC; 0.274 g,
1.33 mmol) and DMAP (10 mg). The resulting solution was
stirred at room temperature (48 h) and then filtered. Concen-
tration of the solution produced a heavy oil which was chro-
matographed over silica gel (H2CCl2:H3C-OH, 100:1) to give
a solid which was recrystallized (hexane:EtOAc, 4:1) to yield
white crystals of 2 (0.065 g, 24%): mp 69-71 °C; IR (KBr)
118.0-137.0 (Ar-C), 171.8 (CdO). MS (EI) Calcd for C12H12
-
BrNO (M+.): 267. Found: 267. Lactam 6 was converted to
amine 7 immediately.
6-Br om o-4,4,7-tr im eth yl-1,2,3,4-tetr ah ydr oqu in olin e (7).
To a cooled (0 °C) solution of lactam 6 (1.0 g, 3.73 mmol) in
dried, distilled toluene (20 mL) was added dropwise a solution
of 10 M borane-dimethyl sulfide (0.88 g, 9.75 mmol). The
mixture was stirred at 0 °C (0.25 h) and then was boiled for 6
h during which time the mixture became gray in color. After
cooling to room temperature, aqueous 10% Na2CO3 (20 mL)
was added, and the solution was stirred for 0.5 h. The organic
layer separated, was dried (MgSO4), and was concentrated to
a light yellow oil which was chromatographed on silica gel (H2-
CCl2) to give 7 as a colorless oil (0.74 g, 78%). 1H NMR (DCCl3)
δ 1.26 (s, 6 H, 2 CH3), 1.69 (t, 2 H, J ) 4.8 Hz, CCH2), 2.30 (s,
3 H, Ar-CH3), 3.27 (t, 2 H, J ) 4.8 Hz, CH2N), 6.35 (s, 1 H,
Ar-H), 7.26 (s, 1 H, Ar-H); 13C NMR (DCCl3) δ 22.3 (Ar-
CH3), 30.8 (2 CH3), 31.5 (ArCCH3), 36.8 (CCH2), 38.3 (CH2N),
1
1718 cm-1; H NMR (DCCl3) δ 1.31 (s, 6 H, 2 CH3), 1.40 (t, 3
H, CH2CH3), 1.74 (t, 2 H, J ) 6.0 Hz, CCH2), 2.60 (s, 3 H,
Ar-CH3), 3.00 (s, 3 H, NCH3), 3.36 (t, 2 H, J ) 6.0 Hz, CH2N),
4.39 (q, 2 H, CH2CH3), 6.38 (s, 1 H, Ar-H), 7.25 (d, 2 H, Ar-
H), 8.09 (s, 1 H, Ar-H), 8.11 (d, 2 H, Ar-H); 13C NMR (DCCl3)
δ 14.2 (CH2CH3), 22.7 (Ar-CH3), 30.0 (ArCCH2), 31.5 (2 CH3),
36.2 (ArCCH2), 38.8 (NCH3), 47.7 (NCH2), 61.2 (OCH2), 112.9-
155.1 (Ar-C), 165.1 (CdO), 166.0 (CdO). Anal. (C23H27NO4)
C, H, N.
47.5 (NCH3), 111.4-142.6 (Ar-C). MS (EI) Calcd for C12H16
-
Eth yl 4-(4,4-Dim eth yl-N-isop r op yl-1,2,3,4-tetr a h yd r o-
qu in olin -6-oyloxy)ben zoa te (3). To acid 17 (0.200 g, 0.810
mmol), ethyl 4-hydroxybenzoate (0.200 g, 1.2 mmol), and H2-
CCl2 (30 mL) were added DCC (∼0.500 g, 3.0 mmol) and a
catalytic amount of DMAP (10 mg) with stirring. After stirring
at room temperature for 24 h, the clear solution was evapo-
rated to a heavy oil which was chromatographed (silica gel;
hexane:EtOAc, 4:1) to a colorless oil. Treatment of the oil with
dry hexane and chilling the mixture (0-4 °C) for 15 h induced
solid formation. Filtration gave a white solid 3 (0.201 g, 63%):
mp 87-89 °C; IR (KBr) 1722 cm-1; 1H NMR (DCCl3) δ 1.24 (d,
6 H, J ) 6.4 Hz, CH(CH3)2), 1.28 (s, 6 H, 2 CH3), 1.38 (t, 3 H,
J ) 7.2 Hz, CH2CH3), 1.68 (t, 2 H, J ) 6.0 Hz, CCH2), 3.28 (t,
2 H, J ) 6.0 Hz, CH2N), 4.21 (m, 1 H, J ) 6.4 Hz, CH(CH3)2),
4.36 (q, 2 H, J ) 7.2 Hz, CH2CH3), 6.68 (d, 1 H, Ar-H), 7.24
(d, 2 H, Ar-H), 7.86 (dd, 1 H, Ar-H), 7.95 (d, 1 H, Ar-H),
8.04 (d, 2 H, Ar-H); 13C NMR (DCCl3) δ 14.3 (CH2CH3), 18.9
(2 CH3), 29.5 (CH(CH3)2), 31.9 (ArCCH2), 35.9 (CH2C), 36.8
(CH2N), 47.5 (CH(CH3)2), 60.9 (CH2CH3), 109.5-155.2 (Ar-
C), 165.0 (CdO), 166.0 (CdO). Anal. (C24H29NO4) C, H, N.
BrN (M+•): 253. Found: 253. Amine 7 was subjected to
N-methylation directly to obtain 8.
6-Br om o-N,4,4,7-tetr a m eth yl-1,2,3,4-tetr a h yd r oqu in o-
lin e (8). A mixture of amine 7 (3.0 g, 0.01 mol), NaHCO3 (1.78
g, 0.02 mol), and water (3 mL) was cooled to 15 °C and then
was treated dropwise with dimethyl sulfate (1.94 g, 0.015 mol).
After stirring at room temperature for 5 h, the mixture was
diluted with HCCl3 (25 mL) which caused a white precipitate
to form. After filtration and washing (HCCl3) the solid, the
filtrate and original organic layer were combined and washed
with brine. Drying (MgSO4) and concentration of the solution
1
gave a solid 8 (1.10 g, 35%): mp 56-58 °C; H NMR (DCCl3)
δ 1.25 (s, 6 H, 2 CH3), 1.73 (t, 2 H, J ) 4.8 Hz, CCH2), 2.30 (s,
3 H, Ar-CH3), 2.87 (s, 3 H, NCH3), 3.19 (t, 2 H, J ) 4.8 Hz,
CH2N), 6.43 (s, 1 H, Ar-H), 7.25 (s, 1 H, Ar-H); 13C NMR
(DCCl3) δ 22.8 (Ar-CH3), 30.8 (2 CH3), 31.9 (ArCCH2), 37.0
(CCH2), 39.3 (CH2N), 47.5 (NCH3), 110.6-144.5 (Ar-C). MS
(EI) Calcd for C13H18BrN (M+.): 267. Found: 267.
Although the above procedure was easy to perform, the
modest yield of 8 was an inducement to develop an alternative
method. A mixture of powdered KOH (0.732 g, 0.013 mol) in
DMSO (15 mL) was stirred at 0 °C for 5 min. Lactam 6 (3.51
g, 0.013 mol) was added cautiously, followed immediately by
the addition of H3C-I (4.79 g, 0.34 mol). Stirring was continued
for 0.5 h (0 °C), after which the mixture was poured into water.
Extracts (H2CCl2, 3 × 20 mL) of the final mixture were com-
bined, washed with water and brine, and then dried (MgSO4).
Evaporation of the solvent gave solid lactam 11 (2.67 g, 73%):
mp 69-71 °C; IR (KBr) 1679 cm-1; 1H NMR (DCCl3) δ 1.27 (s,
6 H, 2 CH3), 2.39 (s, 3 H, Ar-CH3), 2.48 (s, 2 H, CH2CdO),
3.36 (s, 3 H, NCH3), 6.86 (s, 1 H, Ar-H), 7.40 (s, 1 H, Ar-H);
13C NMR (DCCl3) δ 22.8 (Ar-CH3), 27.2 (ArCCH2), 29.4 (2
CH3), 32.7 (NCH3), 45.7 (CH2), 117.3-138.4 (Ar-C), 169.3 (Cd
O). MS (EI) Calcd for C13H14BrNO (M+.): 281. Found: 281.
Lactam 11 was transformed directly to 8 as follows.
Lactam 11 (4. 0 g, 0.014 mol) in dry, distilled toluene (25
mL) was cooled to 0 °C and was then treated with borane-
dimethyl sulfide (2.8 mL, 0.028 mol) over 0.25 h. The mixture
was stirred at 0 °C (0.25 h) and then was boiled for 2 h. After
cooling to room temperature, the reaction mixture was treated
with aqueous Na2CO3 (10%, 20 mL), and the resulting mixture
was stirred for 0.5 h. Separation and drying (MgSO4) of the
organic layer, followed by concentration, gave 8 as an oil.
Placing the oil in EtOAc (2 mL) and chilling (0 °C) overnight
induced crystallization of 8 as a white solid (3.06 g, 82%),
identical to that prepared above. This latter identification
confirms the structure of 8, and the procedure is a slightly
more productive method.
N-(4-Br om o-3-m eth ylp h en yl)-3-m eth yl-2-bu ten a m id e
(5). To the biphasic mixture of 4-bromo-3-methylaniline (4;
15.00 g, 0.081 mol), 10% NaOH (80 mL), and H2CCl2 (50 mL)
was slowly (2 h) added 3,3-dimethylacryloyl chloride (11.47 g,
0.097 mol) in H2CCl2 (30 mL). After stirring at room temper-
ature for 16 h, the reaction mixture was diluted with water
(30 mL), and the resulting aqueous layer was extracted (H2-
CCl2, 4 × 30 mL). The combined extracts were dried (Na2SO4)
and concentrated in vacuo to give a brown solid which was
recrystallized (hexane:EtOAc, 4:1) to give 5 (14.12 g, 65%): mp
97-98 °C; IR (KBr) 3284, 1641 cm-1; 1H NMR (DCCl3) δ 1.89
(s, 3 H, CH3), 2.21 (s, 3 H, CH3), 2.35 (s, 3 H, Ar-CH3), 5.68
(m, 1 H, CdCH), 7.17-7.20 (m, 2 H, Ar-H), 7.41 (d, 2 H, Ar-
H), 7.51 (s, 1 H, Ar-H); 13C NMR (DCCl3) δ 19.9 (CH3), 22.97
(Ar-CH3), 27.42 (CH3), 118.25-138.4 (Ar-C), 118.2 (C)CH),
154.1 (CH2CdC), 164.9 (CdO). MS (EI) Calcd for C12H12BrNO
(M+.): 267. Found: 267. Amide 5 was cyclized at once to lactam
6.
6-Br om o-4,4,7-tr im eth yl-2-oxo-1,2,3,4-tetr a h yd r oqu in -
olin e (6). Amide 5 (6.77 g, 0.025 mol) was heated in a standard
system to 110-120 °C, and then AlCl3 (5.05 g, 0.38 mol) was
added portionwise over 1 h. The resulting viscous mixture was
then stirred for 0.5 h at 110 °C. After cooling to room
temperature, ice water was added and a brown solid formed.
Chloroform (30 mL) was added, and the solid dissolved. This
solution was stirred for 0.25 h, and 2 M HCl (30 mL) was added
cautiously. Two layers separated, and the aqueous layer was
extracted (HCCl3, 5 × 30 mL). The combined organic layers
were washed with aqueous 10% NaHCO3 (30 mL) and brine.
After drying (MgSO4), the solution was concentrated to a
brown solid which was recrystallized (95% ethanol) to white
crystals of 6 (3.48 g, 50%): mp 179-180 °C; IR (KBr) 3209,
6-Cya n o-N,4,4,7-tetr a m eth yl-1,2,3,4-tetr a h yd r oqu in o-
lin e (9). A mixture of amine 8 (2.90 g, 0.011 mol), copper(I)
cyanide (1.34 g, 0.022 mol), and DMF (50 mL) was boiled
vigorously for 6 h, during which time the color became dark
yellow. After cooling to room temperature, the mixture was
diluted with 30% aqueous NaCN (50 mL), and the resulting
1
1686 cm-1; H NMR (DCCl3) δ 1.31 (s, 6 H, 2 CH3), 2.33 (s, 3
H, Ar-CH3), 2.48 (s, 2 H, CH2-CdO), 6.78 (s, 1 H, Ar-H),
7.40 (s, 1 H, Ar-H), 9.93 (s, 1 H, NH); 13C NMR (DCCl3) δ

3612 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Dhar et al.
solution was extracted (benzene, 4 × 30 mL). The combined
extracts were dried (Na2SO4) and concentrated to give 9 as a
light orange solid (2.13 g, 92%): mp 135-137 °C; IR (KBr)
The mixture was stirred (0.25 h), two layers were separated,
and the aqueous layer was extracted (H2CCl2, 3 × 40 mL). The
combined oganic extracts were washed with 10% aqueous
saturated NaHCO3 and brine. After drying (Na2SO4), the
solution was concentrated to an oil (5.34 g, 96%) which was
purified by chromatography (silica gel, H2CCl2) to give pure
1
2205 cm-1; H NMR (DCCl3) δ 1.24 (s, 6 H, 2 CH3), 1.70 (t, 2
H, J ) 6.0 Hz, CCH2), 2.41 (s, 3 H, Ar-CH3), 2.96 (s, 3 H,
NCH3), 3.34 (t, 2 H, J ) 6.0 Hz, CH2N), 6.34 (s, 1 H, Ar-H),
7.32 (s, 1 H, Ar-H); 13C NMR (DCCl3) δ 20.5 (Ar-CH3) 29.9
(ArCCH2), 31.5 (2 CH3), 36.0 (CCH2), 38.8 (CH2N), 47.3
(NCH3), 97.4 (CtN), 110.9-147.8 (Ar-C). MS (EI) Calcd for
1
15 as a light yellow oil (5.15 g, 88%); IR (KBr) 1670 cm-1; H
NMR (DCCl3) δ 1.27 (s, 6 H, 2 CH3), 1.51 (d, 6 H, J ) 7.2 Hz,
2 CH3), 2.41 (s, 2 H, CH2), 4.68 (m, 1 H, CH(CH3)2), 6.99 (s, 1
H, Ar-H), 7.32 (dd, 1 H, Ar-H), 7.37 (d, 1 H, Ar-H); 13C NMR
(DCCl3) δ 20.1 (2 CH3), 26.5 (CH(CH3)2), 32.9 (ArCCH2), 46.9
(CH2), 48.2 (CH(CH3)2), 110.1-138.6 (Ar-C), 169.5 (CdO). MS
(EI) Calcd for C14H18BrNO (M+•): 295. Found: 295. Lactam
15 was used immediately to prepare 16.
C
14H18N2 (M+.): 214. Found: 214. Amine 9 was hydrolyzed at
once to acid 10.
N,4,4,7-Tetr a m eth yl-1,2,3,4-tetr a h yd r oqu in olin e-6-ca r -
boxylic Acid (10). A solution of nitrile 9 (0.150 g, 0.070
mmol), KOH (1.0 g), H3C-OH (3 mL), and diethylene glycol
(DEG; 5 mL) was boiled for 6 days during which time water
was added on occasion to maintain volume. The reaction
mixture was allowed to cool to room temperature, diluted with
water (5 mL), and then acidified with concd HCl (pH ∼ 4).
Extracts (H2CCl2, 3 × 25 mL) of the aqueous layer and the
organic layer were combined, washed with water and then
with brine, and finally dried (Na2SO4). Evaporation of the
solvent gave white, solid 10 (0.062 g, 38%): mp 191-192 °C;
IR (KBr) 3453-2530, 1671 cm-1; 1H NMR (DCCl3) δ 1.28 (s, 6
H, 2 CH3), 1.73 (t, 2 H, J ) 6.0 Hz, CCH2), 2.60 (s, 3 H, Ar-
CH3), 2.99 (s, 3 H, NCH3), 3.34 (t, 2 H, J ) 6.0 Hz, CH2N),
6.34 (s, 1 H, Ar-H), 7.96 (s, 1 H, Ar-H); 13C NMR (DCCl3) δ
22.7 (Ar-CH3), 30.0 (ArCCH2), 31.4 (2 CH3), 36.2 (CCH2), 38.8
(CH2N), 47.5 (NCH3), 112.9-148.7 (Ar-C), 163.5 (CdO). MS
(EI) Calcd for C14H19NO2 (M+.): 233. Found: 233. Acid 10 was
converted to heteroarotinoid 2 without further purification.
4-Br om o-N-isop r op yla n ilin e (13). A mixture of 12 (1.0
g, 5.81 mmol), acetone (0.505 g, 8.7 mmol), titanium(IV)
isopropoxide (1.65 g, 5.8 mmol), and ethanol (4 mL) was stirred
for 12 h at room temperature. To the light brown solution was
added portionwise sodium borohydride (0.439 g, 11.6 mmol)
over 10 min, and then the resulting solution was stirred for 2
h. The system was cooled (0 °C) and quenched with aqueous
ammonia (2 N, 10 mL). A precipitate formed which was filtered
off and washed with ether. The filtrate was concentrated to
approximately one-half volume, cooled (∼8 °C), and refiltered.
Extracts of the aqueous filtrate (ether, 3 × 20 mL) were
washed with brine and then dried (MgSO4). Evaporation of
the solvent gave an oil (0.981 g, 78%) which could be distilled
(77-78 °C/2 mmHg) to give 13 as a colorless oil: IR (neat)
3407 cm-1; 1H NMR (DCCl3) δ 1.17 (d, 6 H, J ) 6.0 Hz, 2 CH3),
3.45 (bs, 1 H, NH), 3.55 (m, 1 H, CH), 6.43 (dd, 2 H, Ar-H),
7.22 (dd, 2 H, Ar-H); 13C NMR (DCCl3) δ 22.7 (2 CH3), 44.2
(CH), 108.3-146.4 (Ar-C). MS (EI) Calcd for C9H12BrN
(M+.): 213. Found: 213. Compound 13 was used directly to
prepare 14.
N-(4-Br om op h en yl)-N-isop r op yl-3-m eth yl-2-bu ten a m -
id e (14). To the biphasic solution of 13 (7.0 g, 0.033 mol),
aqueous NaOH (10%, 40 mL), and H2CCl2 (10 mL) was added
slowly 3,3-dimethylacryloyl chloride (4.65 g, 0.039 mol) in H2-
CCl2 (10 mL) over 1 h. After stirring for 24 h at room
temperature, the solution was then diluted with water (20 mL),
and the resulting aqueous layer was extracted (H2CCl2, 4 ×
30 mL). The extracts were dried (Na2SO4) and evaporated to
produce a yellow solid which was recrystallized (hexane:
EtOAc, 4:1) to give 14 (6.34 g, 66%): mp 79-81 °C; IR (KBr)
1657 cm-1; 1H NMR (DCCl3) δ 1.05 (d, 6 H, J ) 6.8 Hz, 2 CH3),
1.64 (s, 3 H, CH3), 2.09 (s, 3 H, CH3), 5.02 (m, 1 H, CH(CH3)2),
5.23 (bs, 1 H, CdCH), 6.97 (dd, 2 H, Ar-H), 7.52 (dd, 2 H,
Ar-H); 13C NMR (DCCl3) δ 20.1 (CH3), 21.0 (2 CH3), 27.1
(CH3), 45.3 (CHN), 118.2-138.1 (Ar-C), 118.2 (C)CH), 150.2
(CH2C)CH), 166.4 (CdO). MS (EI) Calcd for C14H18BrNO
(M+.): 295. Found: 295. The amide 14 was used at once to
prepare 15.
6-Br om o-N-isop r op yl-4,4-d im eth yl-1,2,3,4-tetr a h yd r o-
qu in olin e (16). To lactam 15 (4.0 g, 13.6 mmol) in toluene
(25 mL) was added dropwise (0.25 h) borane-dimethyl sulfide
(2.03 mL, 20.3 mmol). After being stirred at room temperature
(1 h) and at its boiling point (1.5 h), the mixture was allowed
to cool to room temperature. A solution of 10% aqueous Na2-
CO3 (20 mL) was added, and the mixture was stirred (0.5 h).
Extracts (toluene) of the solution were combined, washed with
brine, and dried (Na2SO4). Evaporation of the solvent gave a
light yellow oil (3.1 g, 81%) which was chromatographed (silica
gel; hexane:EtOAc, 15:1) to give 16 as a colorless oil: 1H NMR
(DCCl3) δ 1.17 (d, 6 H, J ) 6.4 Hz, CH(CH3)2), 1.24 (s, 6 H, 2
CH3), 1.65 (t, 2 H, J ) 6.4 Hz, CCH2), 3.13 (t, 2 H, J ) 6.4 Hz,
CH2N), 4.68 (m, 1 H, CH(CH3)2), 6.53 (d, 1 H, Ar-H), 7.10
(dd, 1 H, Ar-H), 7.22 (d, 1 H, Ar-H); 13C NMR (DCCl3) δ 18.6
(2 CH3), 30.1 (CH(CH3)2), 32.2 (ArCCH2), 36.3 (CH2C), 36.7
(CH2N), 47.1 (CH(CH3)2), 106.1-143.1 (Ar-C). MS (EI) Calcd
for C14H20NBr (M+•): 281. Found: 281. Amine 16 was con-
verted directly to nitrile 17.
6-Cya n o-N-isop r op yl-4,4-d im eth yl-1,2,3,4-tetr a h yd r o-
qu in olin e (17). A mixture of 16 (0.829 g, 2.95 mmol), copper-
(I) cyanide (0.528 g, 5.2 mmol), and DMF (20 mL) was boiled
(24 h), after which time the resulting maroon-colored mixture
was allowed to cool to room temperature. After dilution of the
mixture with 30% aqueous NaCN (15 mL), the resulting
solution was extracted (benzene, 3 × 30 mL), and the combined
extracts were dried (Na2SO4). Evaporation gave an orange oil
which was chromatographed (silica gel, HCCl3) to give a light
orange-colored oil of 17 (0.504 g, 75%): IR (KBr) 2212 cm-1
;
1H NMR (DCCl3) δ 1.22 (d, 6 H, J ) 6.4 Hz, CH(CH3)2), 1.24
(s, 6 H, 2 CH3), 1.66 (t, 2 H, J ) 4.8 Hz, CCH2), 3.25 (t, 2 H,
J ) 4.8 Hz, CH2N), 4.14 (m, 1 H, CH(CH3)2), 6.63 (d, 1 H, Ar-
H), 7.30 (dd, 1 H, Ar-H), 7.38 (d, 1 H, Ar-H); 13C NMR
(DCCl3) δ 18.3 (2 CH3), 29.3 (CH(CH3)2), 31.8 (ArCCH2), 35.6
(CH2C), 36.6 (CH2N), 47.4 [CH(CH3)2], 95.7 (CtN), 110.0-
147.1 (Ar-C). MS (EI) Calcd for C15H20BrN2 (M+•): 307.
Found: 307. Nitrile 17 was used at once to prepare carboxylic
acid 18.
N-Isop r op yl-4,4-d im eth yl-1,2,3,4-tetr a h yd r oqu in olin e-
6-ca r boxylic Acid (18). A mixture of nitrile 17 (0.800 g, 3.50
mmol), KOH (3.0 g), water (3 mL), and DEG (20 mL) was
boiled for 48 h, during which time water was added to
maintain volume. After cooling to room temperature, the
solution was diluted with water (10 mL) and then acidified
with dilute HCl (pH ∼ 3). Extracts (H2CCl2) of the purple
aqueous mixture were combined with the original organic
layer, and the final solution was washed with water and then
brine. After drying (Na2SO4), the solution was evaporated to
a solid (0.495 g, 57%) which was recrystallized (EtOAc with
decolorizing charcoal) to give 18 as a white solid (0.290 g,
1
34%): mp 184-185 °C; IR (KBr) 3443-2542, 1664 cm-1; H
NMR (DCCl3) δ 1.24 (d, 6 H, J ) 6.4 Hz, CH(CH3)2), 1.29 (s, 6
H, 2 CH3), 1.68 (t, 2 H, J ) 6.4 Hz, CCH2), 3.26 (t, 2 H, J )
6.4 Hz, CH2N), 4.21 (m, 1 H, CH(CH3)2), 6.66 (d, 1 H, Ar-H),
7.82 (dd, 1 H, Ar-H), 7.92 (d, 1 H, Ar-H), 11.9 (s, 1 H, CO2H);
13C NMR (DCCl3) δ 18.9 (2 CH3), 29.5 (CH(CH3)2), 31.9
(ArCCH2), 36.0 (CH2C), 36.8 (CH2N), 47.4 (CH(CH3)2), 109.4-
148.4 (Ar-C), 172.6 (CdO). MS (EI) Calcd for C15H21BrNO2
(M+.): 326. Found: 326. Acid 18 was converted directly to
heteroarotinoid 3.
6-Br om o-N-isopr opyl-4,4-dim eth yl-2-oxo-1,2,3,4-tetr ah y-
d r oqu in olin e (15). A system containing 14 (5.86 g, 19.8
mmol) was heated under N2 to 110 °C, and then AlCl3 (3.96 g,
29.8 mmol) was added portionwise over 1 h. After cooling to
room temperature, ice-water (∼40 mL) was added followed
by H2CCl2 (20 mL) which dissolved a solid which had formed.

Nitrogen Heteroarotinoids
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3613
Recep tor Tr a n sa ctiva tion Assa y. The receptor expres-
sion vectors pECE-hRARR/pECE-hRARâ/pECE-hRARγ and
pSG5-mRXRR/pSG5-mRXRâ/pSG5-mRXRγ used in the cotrans-
fection assay were from Magnus Pfhal (La J olla Cancer
Foundation, La J olla, CA) and Pierre Chambon (Strasbourg,
France), respectively. A basal reporter plasmid (pBL-tk-CAT)
with a RARE from the RARâ gene placed upstream to the tk
promoter, âRARE-tk-CAT, was used as the reporter plasmid
in all transfection studies. The CV-1 cells were plated in 6-well
culture plates at a density of 7.5 × 104 cells/well. The next
day the cells were transiently transfected with plasmids
consisting of a 100 ng receptor expression plasmid, 500 ng of
CAT reporter plasmid (âRARE-tk-CAT), and 100 ng of pJ AT-
Lac (â-galactosidase) expression plasmid as internal control,
using FuGENE6 transfection reagent from Boehringer Mann-
heim and following the manufacturer’s protocol (Boehringer
Mannheim Corp., Indianapolis, IN). About 16-18 h after
transfection, retinoids or DMSO (solvent) was added to the
cells. After incubation for 24 h, the cells were washed with
phosphate-buffered saline, lysed, and assayed for CAT and
âGal expression using ELISA kits from Boehringer Mannheim
and following their protocols (Boehringer Mannheim Corp.,
Indianapolis, IN). The retinoid concentration used was in the
range of 10-5-10-10 M, and the DMSO concentration was 0.1%.
All individual determinants were in triplicate and performed
at least twice in separate experiments.
Gr ow th In h ibition Assa y. Cultures were plated in vol-
umes of 150 µL in 96-well microtiter plates at concentrations
of 2000-4000 cells/well, depending upon the cell line used.
Retinoids were added the next day at 4X concentrations in 50
µL of media resulting in 1, 2, 4, 5, 7, and 10 mM final
concentrations of each retinoid. Control cultures were treated
with the same volume of DMSO. After 3 days of treatment,
the cell density in each well was determined by fixing the cells
in trichloroacetic acid and staining the cytoplasmic proteins
with sulforhodamine B (SRB). After rinsing, the SRB was
solubilized in Tris-HCl, and the optical density of each culture
was determined with a MR600 microtiter plate reader. Each
experiment was performed in triplicate, and the three values
for each treatment were averaged. The average OD of the
treated cultures was divided by that of control cultures treated
with solvent alone. To determine the percent growth inhibition,
this ratio was subtracted from 1 and multiplied by 100.
collected in two separate experiments. FlexiDock (available
from Tripos, Inc., 1699 S Hanley Rd, St. Louis, MO 63144)
operates on a genetic algorithm which is initiated by placing
the ligand in the binding pocket of the receptor. The program
then does a search of a large number of conformers which is
labeled a random seed number. An evaluation of the interac-
tion energies between the receptor and the ligand is performed,
and the individual conformers are “scored”. The best ligand
conformers are selected for mutations in terms of variations
in translational, rotational, and torsional variations with
respect to the parent ligand. A similar operation is done on
the receptor. These two operations then generate the flexible
RARγ-flexible ligand. To explore the conformational space of
the ligand, the five experiments were run with different seed
and different generation numbers. The larger the generation
number the higher the probability of convergence to the best
interaction energies. The ligands were drawn as carboxylate
anions using the fragment library of Insight Discover 97.2
(available from MSI, 9685 Scranton Rd, San Diego, CA 92121-
2777). Additional information on the operations is available
from the manufacturers cited.
Ack n ow led gm en t. We gratefully acknowledge par-
tial support of this work by the National Institutes of
Health through a grant from the National Cancer
Institute (CA-73639). We also are pleased to acknowl-
edge funding for the Varian Gemini 300-MHz NMR
spectrometer in the Oklahoma Statewide Shared NMR
Facility by the National Science Foundation (BIR-
9512269), the Oklahoma State Regents for Higher
Education, the W. M. Keck Foundation, and Conoco, Inc.
The research was approved for publication by the
Director, Oklahoma Agriculture Experiment Station
and supported under Project H-1250. One of us (C.W.B.)
gratefully acknowledges partial support by way of a
fellowship through the NIH Minorities in Biomedical
Research, Grant 5 R25 GM55244-02.
Refer en ces
(1) (a) Armstrong, R. B.; Ashenfelter, K. O.; Eckhoff, C.; Levin, A.
A.; Shapiro, S. S. In The Retinoids- Biology, Chemistry, and
Medicine, 2nd ed.; Sporn, M. B., Roberts, A. B., Goodman, D.
S., Eds., Raven Press: New York, 1994; Chapter 13, pp 545-
572. (b) Dawson, M. I. In Burger’s Medicinal Chemistry and Drug
Discovery, 5th ed.; Burger, A., Ed.; Wiley: New York, 1996;
Chapter 44, pp 631-658.
(2) Frolik, C. A. In The Retinoids, Vol. 2; Sporn, M. B., Roberts, A.
B., Goodman, D. S., Eds.; Academic Press: Orlando, FL, 1984;
Chapter 11, pp 177-208.
TGa se Assa y. HEL cells were plated at (1-2) × 105 cells/
mL, and the appropriate retinoid was added to give a final
concentration of 10 µM. After 48 h, the cells were pelleted by
centrifugation for 3 min at 1000g, washed twice with 127 mM
NaCl, 2.7 mM KCl, 1 mΜ KH2PO4, and 8 mM Na2HPO4 (pH
7.4). The cell pellet was resuspended in 20 mM Tris-HCl, 150
mM NaCl, 1 mM EDTA, and 0.5 mM phenylmethanesulfonyl
fluoride (pH 7.4). The cells were disrupted by a 10-s sonication,
(3) (a) Benbrook, D. M.; Madler, M. M.; Spruce, L. W.; Birckbichler,
P. J .; Nelson, E. C.; Subramanian, S.; Weerasekare, G. M.; Gale,
J . B.; Patterson, J r., M. K.; Wang, B.; Wang, W.; Lu, S.; Rowland,
T. C.; DiSevestro, P.; Lindamood, III, C.; Hill, D. L.; Berlin, K.
D. Biologically Active Heteroarotinoids Exhibit Anticancer
Activity and Decreased Toxicity. J . Med. Chem. 1997, 40, 3567-
3583. (b) Benbrook, D. M.; Subramanian, S.; Gale, J . B.; Liu,
S.; Brown, C. W.; Boehm, M. F.; Berlin, K. D. Synthesis and
Characterization of Heteroarotinoids Demonstrates Structure-
Activity Relationships. J . Med. Chem. 1998, 41, 3753-3757. (c)
Lindamood, III, C.; Giles, H. D.; Hill, D. L. Preliminary Profile
of Arotinoids SMR-2 and SMR-6 in male B6D2F1 Mice. Fundam.
Appl. Toxicol. 1987, 8, 517-530. (d) Lindamood, III, C.; Dillehay,
D. L.; Lamon, E. W.; Giles, H. D.; Shealy, Y. F.; Sani, B. P.; Hill,
D. L. Toxicologic and Immunologic Evaluations of N-(All-trans-
Retinoyl)-d,l-Leucine and N-(All-trans-Retinoyl)glycine. Toxicol.
Appl. Pharmacol. 1988, 96, 279-295.
(4) (a) Beard, R. L.; Chandraratna, R. A. S.; Colon, D. F.; Gillett, S.
J .; Henry, E.; Marler, D. K.; Song, T.; Denys, L.; Garst, M. E.;
Arefieg, T.; Klein, E.; Gil, D. W.; Wheeler, L.; Kochhar, D. M.;
Davies, P. J . A. Synthesis and Structure-Activity Relationships
of Stilbene Retinoid Analogues Substituted with Heteroaromatic
Carboxylic Acids. J . Med. Chem. 1995, 38, 2820-2829. (b) Beard,
R. L.; Colon, D. F.; Song, T. K.; Davies, P. J . A.; Kochhar, D. M.;
Chandraratna, R. A. S. Synthesis and Structure-Activity Re-
lationships of Retinoid X Receptor Selective Diaryl Sulfide
Analogues of Retinoic Acid. J . Med. Chem. 1996, 39, 3556-
3563.
and TGase activity was measured in the sonicate by the Ca2+
-
dependent incorporation of [1,4-14C]putrescine (Amersham,
Arlington Heights, IL) into N,N-dimethylcasein (Calbiochem,
San Diego, CA). Radioactivity in acid-insoluble protein was
determined by liquid scintillation spectrometry (Packard
Instruments, Downers Grove, IL). Protein was estimated by
the method of Lowry.20
Ap op tosis Assa y. The SW962 cells were plated in micro-
titer plates at a density of 1000 cells/well. The next day the
cells were treated with a 10 µM solution of the agent or with
the same volume of DMSO (solvent). After 72 h of treatment,
the cells were assayed for the presence of histones in the
cytoplasm using the Cell Death Detection ELISA PLUS assay
acoording to the manufacturer’s protocol (Boehringer Mann-
heim, Indianapolis, IN). The results are presented as average
and standard deviations of three independent experiments
performed in triplicate.
Molecu la r Mod elin g. The program Protein Data Bank is
available from Brookhaven National Laboratory (Bldg 463,
Brookhaven National Laboratory, Upton, NY 11973). Data for
the energy of interaction (Table 2) are from five experimental
runs for each flexible ligand in fixed RARγ. Similar data for
the RARγ flexible receptor and flexible ligand (Table 3) were

3614 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Dhar et al.
(5) (a) Boehm, M. F.; McClurg, M. R.; Pathirana, C.; Manglesdorf,
D.; White, S. K.; Herbert, J .; Winn, D.; Goldman, M. E.; Heyman,
R. A. Synthesis of High Specific Activity [3H]-9-cis-Retinoic Acid
and Its Application for Identifying Retinoids with Unusual
Binding Properties. J . Med. Chem. 1994, 37, 408-414. (b)
Boehm, M. F.; Zhang, L.; Badea, B. A.; White, S. K.; Mais, D.
E.; Berger, E.; Suto, C. M.; Goldman, M. E.; Heyman, R. A.
Synthesis and Structure-Activity Relationships of Novel Ret-
inoid X Receptor-Selective Retinoids. J . Med. Chem. 1994, 37,
2930-2941.
(16) Renauld, J .-P.; Rochel, R.; Ruff, R.; Vivat, V.; Chambon, P.;
Gronemeyer, H.; Moras, D. Crystal Structure of the RAR-γ
Ligand Binding Domain Bound to All-Trans-Retinoic Acid.
Nature 1995, 378, 681-689.
(17) Bourguet, W.; Ruff, M.; Chambon, P.; Gronemeyer, H.; Moras,
D. Crystal Structure of the Ligand-Binding Domain of Human
Nuclear Receptor RXR-R. Nature 1995, 375, 377-382.
(18) Wurtz, J . M.; Borguet, W.; Renauld, J . P. Vivat, V.; Chambon,
P.; Moras, D.; Gronemeyer, H. A Canonical Structure for the
Ligand-Binding Domain of Nuclear Receptors. Nature Struct.
Biol. 1996, 3, 87-94.
(19) Caelles, C.; Gonzalez-Sancho, J . M.; Munoz, A. Nuclear Hormone
Receptor Antagonism with AP-1 by Inhibition of the J NK
Pathway. Genes Devel. 1997, 11, 3351-3364.
(6) Gale, J . B. Recent Advances in the Chemistry and Biology of
Retinoids. In Progress in Medicinal Chemistry, Vol. 30; Ellis, G.
P., Luscombe, D. K., Eds.; Elsevier Science Publishers B.V.:
Amsterdam, 1993.
(7) Lehmann, J . M.; Dawson, M. I.; Hobbs, P. D.; Husmann, M.;
Pfahl, M. Identification of Retinoids with Nuclear Receptor
Subtype-Selective Activities. Cancer Res. 1991, 51, 4803-
4809.
(8) Warrell, J r., R. P.; Frankel, S. R.; Miller, J r., W. H.; Scheinberg,
A.; J akubowski, A. Differentiation Therapy of Acute Promyelo-
cytic Leukemia With Tretinoin (all trans-Retinoic Acid). N. Engl.
J . Med. 1991, 324, 1385-1393.
(9) (a) Suedhoff, T.; Birckbichler, P. J .; Lee, K. N.; Conway, E.;
Patterson, J r., M. K. Expression of Transglutaminase in Human
Erythroleukemia Cells in Response to Retinoic Acid. Cancer Res.
1990, 50, 7830-7834. (b) J acob, K.; Wach, F.; Holzapfel, U.;
Hein, R.; Lengyel, E.; Buettner, R.; Bosserholl, A. K. In Vitro
Modulation of Human Melanoma Cell Invasion and Proliferation
by All-trans-Retinoic Acid. Melanoma Res. 1998, 8, 211-219.
(c) Waliszewski, P.; Waliszewska, M. K.; Gordon, N.; Hurst, R.
E.; Benbrook, D. M.; Dhar, A.; Hemstreet, G. P. Retinoid
Signaling in Immortalized and Carcinoma-Derived Human
Uroepithelial Cells. Mol. Cell. Endrocrinol. 1999, 148, 55-
65.
(10) Birckbichler, P. J .; Patterson, J r., M. K. Growth Regulation by
Products From Cells. Technol. Cell Biol. 1985, C1, 1-20.
(11) Birckbichler, P. J .; Patterson, J r., M. K. Modulation of Trans-
glutaminase-Catalyzed Isopeptide Bonds in Human Cells In
Hemostatis and Cancer; Muszbeck, L., Ed.; CRC Press: Boca
Raton, FL, 1987.
(12) (a) Oridate, N.; Lotan, D.; Mitchell, M. F.; Hong, W. K.; Lotan,
R. Inhibition of Proliferation and Induction of Apoptosis in
Cervical Carcinoma Cells by Retinoids. Implication for Chemo-
prevention. J . Cell Biochem. 1995, Suppl. 23, 80-86. (b) Oridate,
N.; Lotan, D.; Xu, X.-C.; Hong, W. K.; Lotan, R. Differential
Induction of Apoptosis by All-trans-Retinoic Acid and N-(4-
Hydroxyphenyl)retinamide in Human Head and Neck Squamous
Cell Carcinoma Cell Lines. Clin. Cancer Res. 1996, 2, 855-863.
(c) Oridate, N.; Higuchi, M.; Suzuki, S.; Shroot, B.; Hong, W.
K.; Lotan, R. Rapid Induction of Apoptosis in Human C33-A
Cervical Cell Carcinoma Cells by the Synthetic Retinoid 6-[3-
(1- adamantyl)hydroxyphenyl]-2-naphthalene Carboxylic Acid
(CD437). Int. J . Cancer 1997, 70, 484-487.
(13) Zhang, X.-k.; Lehmann, J .; Hoffmann, B.; Dawson, M. I.;
Cameron, J .; Graupner, G.; Hermann, T.; Tran, P.; Pfahl, M.
Homodimer Formation of Retinoid X Receptor Induced by 9-cis-
Retinoic Acid. Nature 1992, 358, 587-5891.
(14) Look, J .; Landwehr, J .; Bauer, F.; Hoffman, A. S.; Bluethmann,
H.; Lemotte, P. Marked Resistance of RARγ-Deficient Mice to
The Toxic Effects of Retinoic Acid. Am. J . Physiol. 1995, 269,
E91-E98.
(20) Fisher, G. J .; Talwar, H. T.; Wang, Z. Q.; McPhillips, F.; Madore,
S.; Voorhees, J . J . Retinoic Acid Antagonizes Activation of
Transcription Factor AP-1 by Ultraviolet Irradiation Through
a Novel Mechanism in Human Skin In Vivo: Inhibition of c-J UN
Protein Expression. J . Invest. Dermatol. 1997, 108, 578 (Ab-
stract).
(21) Yen, A.; Roberson, M. S.; Varyvayanis, S.; Lee, A. T. Retinoic
Acid-Induced Mitogen-Activated Protein (MAO)/Extracellular
Signal-Regulated Kinase (ERK) Kinase-Dependent MAP Kinase
Activation Needed to Elicit HL-60 Cell Differentiation and
Growth Arrest. Cancer Res. 1998, 58, 3163-3172.
(22) Zelent, A.; Krust, A.; Petkovich, M.; Kastner, P.; Chambon, P.
Cloning of Murine R and â Retinoic Acid Receptors and a Novel
Receptor γ Predominately Expressed in Skin. Nature 1989, 339,
714-717.
(23) Pfahl, M.; Apfel, R.; Bendrik, I. Nuclear Retinoid Receptors and
Their Mechanism of Action. In Vitamins and Hormones; Litwack,
G., Ed.; Academic Press: San Diego, CA, 1994; pp 327-382.
(24) Boehm, M. F.; Zhang, L.; Zhi, L.; McClurg, M. R.; Berger, E.;
Wagoner, M.; Mais, D. E.; Suto, C. M.; Davies, P. J . A.; Heyman,
R. A.; Nadzan, A. M. Design and Synthesis of Potent Retinoid X
Receptor Selective Ligands That Induce Apoptosis in Leukemia
Cells. J . Med. Chem. 1995, 38, 3146-3155.
(25) Zhang, L.-X.; Mills, K. J .; Dawson, M. I.; Collins, S. J .; J etten,
A. M. Evidence For The Involvement of Retinoic Acid Receptor
RARR-Dependent Signaling Pathway in the Induction of Tissue
Transglutaminase and Apoptosis By Retinoids. J . Biol. Chem.
1995, 270, 6022-6029.
(26) Bernard, B. A.; Bernardon, J . M.; Delescluse, C.; Martin, B.;
Lenoir, M.-C.; Maignan, J .; Charpentier, B.; Pilgrim, W. R.;
Reichert, U.; Shroot, B. Identification of Synthetic Retinoids
With Selectivity for Human Nuclear Retinoic Acid Receptor γ.
Biochem. Biophys. Res. Commun. 1992, 186, 977-983.
(27) Klaholz, B. P.; Renauld, J .-P.; Mitschler, A.; Zuis, C.; Chambon,
P.; Gronemeyer, H.; Moras, D. Conformational Adaptation of
Agonists to the Human Nuclear Receptor RARγ. Nature Struct.
Biol. 1998, 5, 013-016.
(28) (a) Kagechika, H.; Kawachi, E.; Hashimoto, Y.; Himi, T.; Shudo,
K. Retinobenzoic Acids. 1. Structure-Activity Relationships of
Aromatic Amides With Retinoidal Activity. J . Med. Chem. 1988,
31, 2182-2192. (b) Kagechika, H.; Himi, T.; Kawachi, E.;
Hashimoto, Y.; Shudo, K. Retinobenzoic Acids. 4. Conformation
of Aromatic Amides With Retinoidal Activity. Importance of
trans-Amide Structure for the Activity. J . Med. Chem. 1989, 32,
2292-2296.
(29) J ones, P.; Villeneuve, G. B.; Fei, C.; DeMarte, J .; Haggatary, A.
J .; Nwe, K. T.; Martin, D. A.; Lebuis, A.-M.; Finkelstein, J . M.;
Gour-Salin, B. J .; Chan, T. H.; Leyland-J ones, B. R. Synthesis
and Structure-Activity Relationships of 2-Pyrazinylcarboxami-
dobenzoates and â-Ionylideneacetamidobenzoates With Retin-
oidal Activity. J . Med. Chem. 1998, 41, 3062-3077.
(15) Chen, S.; Ostrowski, J .; Whiting, G.; Roalsvig, T.; Hammer, L.;
Currier, S. J .; Honeyman, J .; Kwasniewski, B.; Sterzycki, R.;
Kim, C. U.; Starrett, J .; Mansuri, M.; Reczek, P. R. Retinoic Acid
Receptor Gamma Mediates Topical Retinoid Efficacy and Irrita-
tion In Animal Models. J . Invest. Dermatol. 1995, 104, 779-
783.
J M9900974