Conformationally Defined 6-s-trans-Retinoic Acid Analogs. 3. Structure

J. Med. Chem. 1996, 39, 3625-3635

3625

Conformationally Defined 6-s-trans-Retinoic Acid Analogs. 3. Structure-Activity Relationships for Nuclear Receptor Binding, Transcriptional Activity, and Cancer Chemopreventive Activity Donald D. Muccio,*,† Wayne J. Brouillette,*,† Muzaffar Alam,† Michael F. Vaezi,† Brahma P. Sani,‡ Pratap Venepally,‡ Lakshmi Reddy,‡ Ellen Li,§ Andrew W. Norris,§ Linda Simpson-Herren,‡ and Donald L. Hill‡ Department of Chemistry, University of Alabama at Birmingham, Birmingham, Alabama 35294, Southern Research Institute, 2000 Ninth Avenue South, Birmingham, Alabama 35205, and Departments of Medicine and of Biochemistry and Molecular Biophysics, Washington University, St. Louis, Missouri 63110 Received April 26, 1996X

We recently demonstrated that conformationally defined 6-s-trans-retinoic acid (RA) analogs were effective in the prevention of skin papillomas (Vaezi et al. J. Med. Chem. 1994, 37, 44994507) and selective agonists for nuclear receptor binding and activation (Alam et al. J. Med. Chem. 1995, 38, 2302-2310). In order to probe important structure-activity relationships, we evaluated a homologous series of four 6-s-trans-retinoids that are 8-(2′-cyclohexen-1′-ylidene)3,7-dimethyl-2,4,6-octatrienoic acids with different substituents at 2′ (R2) and 3′ (R1) positions on the cyclohexene ring. UAB1 (R1 ) R2 ) H), UAB4 (R1 ) R2 ) Me), UAB7 (R1 ) Me, R2 ) iPr), and UAB8 (R1 ) Et, R2 ) iPr) contain alkyl R groups that mimic, to different extents, portions of the trimethylcyclohexenyl ring of RA. Both 9Z- and all-E-isomers of these retinoids were evaluated in binding assays for cellular retinoic acid-binding proteins (CRABP-I and CRABP-II), a nuclear retinoic acid receptor (RARR), and a nuclear retinoid X receptor (RXRR). The all-E-isomers of UAB retinoids bound tightly to CRABPs and RARR, the binding affinity of the all-E-isomer increased systematically from UAB1 to UAB8, and binding for the latter was comparable to that of all-E-RA. In contrast to RA, the (9Z)-UAB retinoids were at least 200-fold less active than the all-E-isomers in binding to RARR. The (9Z)-UAB isomers exhibited increasingly stronger binding to RXRR, and (9Z)-UAB8 was nearly as effective as (9Z)-RA in binding affinity. The retinoids were also evaluated in gene expression assays mediated by RARR and RXRR homodimers or RARR/RXRR heterodimers. Consistent with the binding affinities, the (all-E)-UAB retinoids activated gene transciption mediated by RARR homodimers or RARR/RXRR heterodimers, while the (9Z)-UAB isomers activated only the RXRR homodimermediated transcription. The all-E- and 9Z-isomers of the UAB retinoids were further evaluated for their capacity to prevent the induction of mouse skin papillomas. When compared to RA, only the (all-E)-UAB retinoids containing bulky R1 and R2 groups were effective in this chemoprevention assay. (9Z)-RA displayed equal capacity as RA to prevent papillomas, while the 9Z-isomers of the UAB retinoids were much less effective. Taken together, these studies demonstrate that the cyclohexenyl ring substituents of 6-s-trans-UAB retinoids are important for their biological activities and that the chemopreventive effect of the all-E-isomers of these retinoids correlates well with their capacity to bind to RARs and activate RAR/RXR-mediated transcription. Introduction Retinoic acid (RA) is essential for diverse biological processes, including the control of cellular differentiation and development.1,2 The pleiotropic effects of RA appear to be related to its interactions with a large family of nuclear receptors. Two classes of nuclear retinoic acid receptors have been identified (RARs and RXRs), and each class contains several different subtypes. Collectively these nuclear receptors, which are related to the steroid/thyroid hormone superfamily of receptors,3 act as ligand-dependent transcription factors for different genes.4 Two RA configurational isomers ((all-E)-RA and (9Z)-RA) are important in its natural function. Both isomers of RA bind to RARs and activate transcription mediated by RAR homodimers or RAR/ RXR heterodimers,5,6 but (9Z)-RA is the only known natural ligand for RXRs.7-9 * Authors to whom correspondence should be addressed. † University of Alabama at Birmingham. ‡ Southern Research Institute. § Washington University. X Abstract published in Advance ACS Abstracts, August 15, 1996.

S0022-2623(96)00312-3 CCC: $12.00

Since RA has a role in controlling gene expression, its use as a cancer chemopreventive agent has been investigated using animal models.10 Even though studies have shown that RA may be effective in cancer prevention, its clinical use has been limited due to toxicity11 and teratogenicity.12 It has been suggested that both the therapeutic and toxicologic effects of RA may be mediated by RARs, RXRs, and binding proteins.13 In order to improve the therapeutic ratio of RA, several groups have designed new retinoid analogs that selectively activate individual nuclear receptors, either RAR-selective14 or RXR-selective agonists.15 Hopefully these receptor-selective retinoids will provide the beneficial effects of RA with reduced toxicity.10b We designed16 a series of 6-s-trans-retinoids based on the observation that certain proteins17 selectively bind this conformation of vitamin A derivatives.18 Previously we reported19 the synthesis of one 6-s-trans-retinoid (UAB7) and showed that it was effective in inhibiting the chemical induction of mouse skin papillomas and that it also had reduced toxicity. Subsequently, we generated20 a more elaborate example (UAB8) and © 1996 American Chemical Society

3626

Journal of Medicinal Chemistry, 1996, Vol. 39, No. 19

Muccio et al.

Scheme 1

Scheme 2

demonstrated selectivity in RAR and RXR nuclear receptor binding and activation. Here we report a thorough comparison of the all-E- and 9Z-isomers for four synthetic homologs of these retinoids (UAB1, UAB4, UAB7, and UAB8). These studies demonstrate the importance of large R1 and R2 substituents in the UAB retinoids for enhanced biological activity and suggest a correlation between nuclear receptor binding/ activation and the cancer chemopreventive effects in skin.

Chemistry Scheme 1 summarizes the methods employed to prepare UAB4. This approach was identical with that which we previously reported for the syntheses of UAB7

and UAB8,19,20 except that a different starting enone (1 in Scheme 1) was used. Some key points of the synthesis are reported here. Unlike previous reports,19,21 a δ-lactone intermediate was not detected but was assumed to be an intermediate in the production of (9Z)-2. (9Z)-5 and (9Z,13Z)-5 were preparatively separated by HPLC on silica gel (0.5% Et2O, 0.1% THF in hexane) using methods similar to those we previously described.19,20,22 Individual isomers of ester 5 were hydrolyzed to the corresponding acids in KOH, without E/Z-isomerization.23 In order to produce (all-E)-UAB4, intermediate aldehyde (9Z)-4 was isomerized using I2 to yield a 1:2 ratio of all-E- and 9Z-isomers. As for (9Z)4, a Horner-Emmons condensation using (all-E)-4 produced a 2:1 mixture of all-E- and 13Z-ester 5. These two isomers were also separated by HPLC on silica gel (1% Et2O, 0.5% THF in hexane). The isomeric purities for the final acids (>97%) were determined by NMR and reverse-phase HPLC.20 Experimental yields and selected data for the intermediates and products in Scheme 1 are summarized in Table 1. The chemical shift assignments (Table 2) for the UAB4 isomers were made using methods previously described.19,24,25 As shown in Scheme 2, the synthesis of UAB1 was accomplished by a modification of the approach in Scheme 1. Since Horner-Emmons condensations on 2-cyclohexenone gave only 1,4-addition products, we first generated ester 7, a dihydro precursor to UAB1.

Conformationally Defined 6-s-trans-RA Analogs

Journal of Medicinal Chemistry, 1996, Vol. 39, No. 19 3627

Table 1. Selected Data for New Compounds Produced in Schemes 1 and 2 compd

yield (%)

(9Z)-2 (9Z)-3 (9Z)-4 (all-E)-4 (9Z)-5f (9Z,13Z)-5f (all-E)-5f (13Z)-5f (9Z)-UAB4 (9Z,13Z)-UAB4 (all-E)-UAB4 (13Z)-UAB4 (all-E)-8 (all-E)-9 (all-E)-10 (all-E)-11g (7Z)-11g (13Z)-11g (all-E)-UAB1 (7Z)-UAB1

79 93 87 51 81e 81e 75e 75e 98 95 98 96 63 91 78 80e 80e 80e 99 94

Rf

UV/visb λmax e

IRc GC/MS CdO CdC (m/z)

0.28 0.25 0.46 0.38 0.69 0.71 0.67 0.73 0.34 0.29 0.20 0.24 0.80 0.21 0.39 0.68 0.68 0.68 0.25 0.25

320 266 298 324 329 331 362 369 341 338 378 378 298 272 326 368 366 370 370 366

1672 1580 1630 1672 1607 1660 1577 1706 1602 1708 1605 1706 1598 1709 1604 1672 1594 1684 1601 1672 1598 1673 1596 1712 1633 1600 1650 1580 1702 1605 1712 1605 1709 1607 1682 1595 1675 1593

a

11 500 13 500 7500 16 000 23 000 22 000 27 500 24 000 19 000 17 500 26 000 21 000 17 000 NDd ND 39 000 36 000 29 000 32 000 25 000

206 192 190 190 300 300 300 300 272 272 272 272 206 164 162 272 272 272 244 244

a Values obtained on silica gel using either diethyl ether or ethyl acetate in hexane as an eluent: 10% Et2O (4 and 5), 20% Et2O (2, UAB4, 11, and UAB1), 30% Et2O (3), 10% EtOAc (8-10). b The wavelength maximum (nm) and extinction coefficients (M-1 cm-1) were obtained in cyclohexane (2-5, 8, 11, and UAB1), chloroform (UAB4), and MeOH (9 and 10) at room temperature. c The IR stretching frequencies (cm-1) were obtained as thin films on NaCl plates. The hydroxyl group frequency was observed at 3326 cm-1 for (9Z)-3 and at 3340 cm-1 for (all-E)-9. d ND ) not determined. e Yield for the E,Z-product mixture. f The separation of retinoid 5 isomers was performed on a Whatman Partisil 10 M20/50 column (500 × 22 mm i.d.). Retention times (min): 91 (all-E), 83 ((9Z)and (9Z,13Z)-isomer mixture), 80.5 (13Z) using 1% Et2O, 0.5% THF in hexane with a flow rate of 5 mL/min. Retention times (min): 181 (9Z); 175 (9Z,13Z) using 0.5% Et2O, 0.1% THF in hexane with a flow rate of 5 mL/min. g The ratio of (all-E)- to (13Z)11 was 6:1 with a trace amount of (7Z)-11. The separation of retinoid 11 isomers was performed on a Whatman Partisil 10 M20/ 50 column (500 × 22 mm i.d.). Retention times (min): 33 (13Z), 53 (7Z), 56 (all-E) using 0.5% Et2O, 0.2% THF in hexane with a flow rate of 10 mL/min.

This was accomplished using a Horner-Emmons condensation between cyclohexanone and diethyl 3-(ethoxycarbonyl)-2-methylprop-2-enylphosphonate, which we previously reported in modest yield during a reactivity study.21 The modified procedure described here utilized HMPA as cosolvent, produced much higher yields, and gave nearly exclusively (9E)-7. The cyclohexenyl double bond was then introduced via regioselective bromination with NBS followed by dehydrobromination to give (allE)-8; this was carried on as in Scheme 1 to give UAB1. Biology The IC50 and Kd′ values were determined for the E/Zisomers of UAB retinoids with CRABPs, RARs, and RXRs. The IC50 values were evaluated in a chick skin CRABP radioligand binding assay.26 The Kd′ values were determined using a fluorometric titration of ligand and apo-CRABP-I or apo-CRABP-II.27 This approach was also used to determine Kd′ values to the ligandbinding (DEF) domain of human RXRR.28 The IC50 values for each E/Z-isomer of the UAB retinoids were next determined in an assay for the inhibition of the binding of (all-E)-RA to RARR.8a Similar studies were also performed for the inhibition of the binding of (9Z)RA to RXRR.29 Each isomer was evaluated for the efficiency of activating RARR homodimers, RXRR ho-

modimers, or RXRR/RARR heterodimers using a CAT reporter gene containing the TREpal.30 The UAB retinoids were evaluated in a skin antipapilloma assay which measured the capacity of retinoids to prevent chemically induced papillomas in mice.31 Results UAB retinoids employ a dimethylene bridge between C18 and C7 of the polyene chain to maintain a 6-s-transconformation. When the polyene chain of RA is overlaid with that of the UAB retinoids, the two methylene groups in the bridge do not occupy space which is in common with that of RA.19 R1 and R2 alkyl substituents are used to mimic the steric volume of the carbon atoms within the trimethylcyclohexenyl ring of RA. UAB1 (R1 ) R2 ) H) thus represents a minimal structure without any R groups to mimic the alkyl portion of the RA ring. The homologous series of UAB4 (R1 ) R2 ) Me), UAB7 (R1 ) Me, R2 ) iPr), and UAB8 (R1 ) Et, R2 ) iPr) adds hydrophobic groups to the terminal end of the polyene chain which may mimic the RA ring. Computational Studies. Molecular mechanics calculations were used to evaluate the structural similarity of the UAB retinoids to RA. Using the MM3(94) force field, the low-energy structures of (all-E)- or (9Z)-UAB retinoids containing an 8-s-trans-conformation were computed (Table 3). The ψ5,6,7,8 torsional angles were nearly planar for either (all-E)- or (9Z)-UAB1. As the size of R1 and R2 increased, this torsional angle became increasingly less planar (about 35° out of plane for UAB8). The ψ7,8,9,10 torsional angles were nonplanar by about 40°, due to steric interactions between the C2′ methylene and the C19 methyl group. For the 8-s-cisconformer, the ψ5,6,7,8 torsional angles were nearly identical with those of the 8-s-trans-conformer and the ψ7,8,9,10 angles were nonplanar by about 50°. Further, the 8-s-cis-conformers were slightly more stable (∆G ≈ 1 kcal/mol) than the corresponding 8-s-trans-conformers. RA adopts primarily a 6-s-cis-conformation in solution,18 a conformation17f,g found for the vitamin bound to CRABP-I/CRABP-II/RAR-γ. The C7-C15 portion of the polyene chain adopts an s-trans-conformation for the low-energy structure.18 Structures of (all-E)- and (9Z)RA were calculated using torsional angles from the X-ray crystal structures32 of (all-E)-RA and (9Z)-retinal, both of which crystallized in a 6-s-cis-conformation. The ψ5,6,7,8 torsional angles of either (all-E)- or (9Z)-RA were substantially different than those found in the UAB retinoids; this is due to the methylene bridge used in the latter compounds to maintain 6-s-trans-geometry. The ψ7,8,9,10 torsional angles of the RAs were nearly 180°, which was significantly different from those of the UAB retinoids (Table 3). Unlike UAB retinoids, the energies for 8-s-cis-RA conformers were about 1-2 kcal/mol higher than those for the corresponding 8-s-transconformers. While (all-E)-RA adopted a completely planar 8-s-cis-orientation, the corresponding UAB compounds had ψ7,8,9,10 torsional angles that were skewed by about 50°. This is presumably due to steric interactions between the H10 methine and H2′ methylene protons. Taken together, these studies suggest that UAB compounds adopt very similar structures to RA along the polyene chain between C9 and C15, but they differ in the terminal end of the molecule (C5-C9). In order to evaluate electrostatic properties, the dipole moments of RA and UAB retinoids were calculated. The

3628


Table 2.

1H

Muccio et al.

Chemical Shifts (ppm, TMS) of UAB1 and UAB4 Retinoids and Intermediates in CDCl3

compd

H-1′

H-2′

H-1

H-4

H-8

H-10

(9Z)-2 (9Z)-3 (9Z)-4 (all-E)-4 (9Z)-5 (9Z,13Z)-5 (13Z)-5 (all-E)-5 (9Z)-UAB4 (9Z,13Z)-UAB4 (13Z)-UAB4 (all-E)-UAB4

1.65 1.64 1.65 1.65 1.62 1.62 1.64 1.63 1.62 1.62 1.65 1.65

2.35 2.15 2.26 2.53 2.14 2.18 2.53 2.54 2.16 2.18 2.56 2.54

1.81 1.79 1.84 1.82 1.81 1.81 1.81 1.81 1.82 1.81 1.82 1.82

1.84 1.79 1.82 1.84 1.85 1.84 1.81 1.81 1.86 1.85 1.82 1.82

6.50 5.74 5.93 5.88 5.84 5.84 5.89 5.91 5.84 5.84 5.91 5.91

5.72 5.47 5.92 5.93 6.03 6.13 6.20 6.09 6.04 6.15 6.22 6.11

4.00 9.54 10.04 6.57 6.57 6.93 6.93 6.62 6.62 6.97 6.98

H-1′

H-2′

H-5

H-6

H-8

H-10

H-11

1.74 1.70 1.75 1.71 1.77 1.70 1.77

2.59 2.55 2.60 2.64 2.35 2.64 2.36

5.96 5.83 5.95 5.88 5.91 5.89 5.92

6.08 6.05 6.10 6.09 6.67 6.09 6.67

5.71 5.65 5.75 5.79 5.69 5.79 5.70

5.71 5.56 5.97 6.16 6.14 6.17 6.16

4.25 10.05 6.92 6.92 6.98 6.98

(all-E)-8 (all-E)-9 (all-E)-10 (all-E)-11 (7Z)-11 (all-E)-UAB1 (7Z)-UAB1

H-11

H-12

H-14

H-18

H-19

H-20

6.18 7.67 7.74 6.23 6.21 7.64 7.69 6.26

5.74 5.59 5.62 5.76 5.76 5.62 5.64 5.79

2.13 2.13 2.14 2.17 2.15 2.15 2.13 2.14 2.15 2.13 2.13 2.14

2.06 1.79 2.01 2.28 1.91 1.91 1.99 1.99 1.91 1.92 2.01 2.01

2.27 1.98 2.07 2.35 2.27 2.01 2.10 2.37

H-12

H-14

H-18

H-19

H-20

5.78 5.78 5.79 5.78

2.15 2.14 2.15 2.16 2.19 2.15 2.18

2.28 1.82 2.30 2.04 2.02 2.05 2.03

2.35 2.35 2.36 2.36

6.26 6.26 6.28 6.29

Table 3. Torsional Angles and Relative Energies of 8-s-trans/8-s-cis-Conformers of the UAB Retinoids Using Molecular Mechanics Calculations (MM3(94) Force Field) 8-s-trans

8-s-cis

retinoid

ψ5,6,7,8

ψ7,8,9,10

energya

RA (all-E)-UAB8 (all-E)-UAB7 (all-E)-UAB4 (all-E)-UAB1 (9Z)-RA (9Z)-UAB8 (9Z)-UAB7 (9Z)-UAB4 (9Z)-UAB1

-59 -146 -150 -159 -175 -60 -146 -149 -159 -176

-180 -137 -139 -140 -143 176 -130 -129 -135 -138

25.1 34.5 26.8 23.7 18.3 24.9 34.5 26.8 23.8 18.6

ψ5,6,7,8

ψ7,8,9,10

energya

∆G° b

µ (D)c

-59 -146 -150 -159 -175 -56 -146 -149 -158 -174

-4 52 50 49 48 -46 62 62 62 64

26.3 33.9 26.3 26.3 17.8 26.5 33.6 26.0 23.3 18.3

1.9 -0.9 -0.9 -1.0 -0.9 1.2 -1.2 -1.1 -1.1 -1.0

1.6 1.5, 1.6 1.5, 1.6 1.5, 1.6 1.2, 1.3 1.6, 1.7 1.7, 1.8 1.6, 1.8 1.6, 1.8 1.9, 2.1

a The strain energy is calculated in kcal/mol. b ∆G° is the difference in free energy (kcal/mol) between 8-s-cis- and 8-s-trans-conformers (G°trans - G°cis) calculated at 300 K. c Dipole moments are measured in Debye (D); two values appear for 8-s-cis- and 8-s-trans-conformers.

dipole moments of UAB4, UAB7, and UAB8 were very similar (Table 3) and close to that of RA. The dipole moment of UAB1 was slightly smaller for the all-Eisomer. The Connolly surface area of these structures was also determined using a 1.4 Å probe. For the (allE)-UAB retinoids, the surface area increased systematically: 265 (UAB1), 296 (UAB4), 324 (UAB7), and 338 (UAB8) Å2 as compared to 325 Å2 for RA. The 9Zisomers had similar values and identical trends. CRABP Binding Affinity. UAB retinoids were evaluated in the CRABP binding assay26 which utilizes [3H]-(all-E)-RA as the radioligand. The IC50 values of the UAB retinoids are compared to those of (all-E)- and (9Z)-RA (Table 4). At 100-fold excess, (all-E)-UAB1 inhibited the binding of [3H]RA by only 50%. (all-E)UAB7 and UAB8, with larger R1 and R2 substituents, had improved binding profiles relative to UAB1 and had IC50 values similar to (all-E)-RA in this assay. (9Z)RA competes weakly for the CRABP binding site, inhibiting only 25% of the binding of labeled RA at 100fold excess.33 Likewise, the (9Z)-UAB retinoids were not effective binders to these proteins. Interestingly (7Z)UAB1 was as effective as (all-E)-UAB1. Fluorescence titrations27 of the all-E- and 9Z-isomers were measured with CRABP-I and CRABP-II (mouse) at micromolar protein concentrations to determine the stoichiometry of the ligand-protein complex (1:1). The 9Z-isomers were poor binders to these proteins (Kd′ >

Table 4. Structure-Activity Relationships for UAB Retinoids and Binding to Mouse and Chick Skin CRABPs chick skin CRABP Kd (nM)

retinoid

% inhibtn at 100fold excess liganda

IC50 (µM)

CRABP-I CRABP-II

(all-E)-RA (all-E)-UAB8 (all-E)-UAB7 (all-E)-UAB4 (all-E)-UAB1 (9Z)-RA (9Z)-UAB8 (9Z)-UAB7 (9Z)-UAB4 (7Z)-UAB1

100 100 100 90 50 25 15 43 10 45

0.6 0.7 0.6 1.3 >10 >10 >10 >10 >10 >10

0.4 ( 0.3 0.3 ( 0.02 1 ( 0.6 1.2 ( 1 3(2 >200 >200 >200 >200 3(1

2(1 1.4 ( 0.2 14 ( 2 13 ( 5 13 ( 4 >200 >200 >200 >200 10 ( 2

a The percent inhibition of binding of 100-fold excess of test retinoids with [3H]-(all-E)-RA (n ) 1). IC50 values were determined from five doses run in duplicate and have error of less than 20% of their mean.

200 nM). Due to the poor solubility of retinoids in aqueous solutions,34 accurate Kd′ values can not be measured for weakly bound ligands. Fluorometric titrations of the tight binding retinoid isomers were performed at 10-25 nM protein concentrations to more accurately determine Kd′ values (Table 4).35 The Kd′ values for the protein-ligand complex with (all-E)UAB8 were 0.3 nM (mCRABP-I) and 1.4 nM (mCRABPII), which were in excellent agreement with previous reports on RA complexes27 (Table 4). For CRABP-I, the



Table 5. Structure-Activity Relationships for the Binding of UAB Retinoid Isomers to Mouse Retinoic Acid Receptors (mRAR), Mouse Retinoid X Receptors (mRXR), and Human Retinoid X Receptor DEF Domain (hRXR-(DEF)) RARR retinoid isomer (all-E)-RA (all-E)-UAB8 (all-E)-UAB7 (all-E)-UAB4 (all-E)-UAB1 (9Z)-RA (9Z)-UAB8 (9Z)-UAB7 (9Z)-UAB4 (7Z)-UAB1

RXRR

IC50 % IC50 hRXRR-(DEF) % Kd (nM)c inhibtna (nM)a inhibtnb (nM)b 100 98 100 90 55 95 42 37 10 45

6 11 8 55 600 31 >1000 >1000 >2000 >2000

10 0 28 0 5 100 50 98 40 25

>2000 >2000 >2000 >2000 >2000 82 1000 620 >2000 >2000

>200 >200 >200 >200 >200 3 ( 0.5 16 ( 0.6 33 ( 5 120 >200

a The percent inhibition was determined by competition of 100fold excess of unlabeled retinoid with 5 nM [3H]-(all-E)-RA (n ) 1). The IC50 values were determined from nine doses run in duplicate as reported by Alam et al.20 The standard deviation was less than 10% of the IC50 value. b The percent inhibition was determined by competition of 100-fold excess of unlabeled retinoid with 20 nM [3H]-(9Z)-RA (n ) 1). The IC50 values were determined from nine doses run in duplicate as reported in Alam et al.20 c Kd values were determined by the method of Cheng et al.28

binding affinity decreased 3-fold in going from UAB7 to UAB1; a 10-fold decrease in binding affinity occurred between UAB8 and UAB7 for CRABP-II. The binding affinity of (7Z)-UAB1 to either protein was much better than that of the (9Z)-UAB compounds and comparable to that of (all-E)-UAB1. Nuclear Receptor Binding Affinity. Each isomer was evaluated for its capacity to inhibit the binding of [3H]-(all-E)- and [3H]-(9Z)-RA to the RARR and RXRR subtypes. To survey the inhibition process, RARR was first exposed to 1000 nM UAB retinoids and 5 nM [3H](all-E)-RA. This was compared to a positive control using 1000 nM unlabeled (all-E)-RA and a control using no retinoid. (all-E)-UAB1 displaced the radioactive ligand only 55% as well as (all-E)-RA, and the IC50 was 100-fold greater than that of (all-E)-RA. The percent inhibition of binding increased and the IC50 values decreased systematically as alkyl substituents were added to this minimal structure (Table 5). The binding affinities of (all-E)-UAB7 and (all-E)-UAB8 were in the same range as that of (all-E)-RA. Therefore, as observed for binding to CRABPs (Table 4), large alkyl groups at R1 and R2 are important for the efficient binding of UAB retinoids to RARR. Unlike (9Z)-RA, the (9Z)-UAB retinoids bound weakly to RARR (IC50 > 1000 nM). (9Z)-UAB4 was the weakest binder to RARR; its IC50 value was over 100-fold higher than that for (9Z)RA. (7Z)-UAB1 also displayed poor binding affinity to RARR. The UAB retinoids were evaluated for their capacity to inhibit the binding of [3H]-(9Z)-RA to RXRR. A preliminary screen was performed using 2000 nM retinoid and 20 nM [3H]-(9Z)-RA, which was efficiently inhibited by unlabeled (9Z)-RA at 100-fold excess. None of the all-E-isomers had appreciable affinity for RXRR, consistent with the results observed for (all-E)-RA. Among the 9Z-isomers, (9Z)-UAB4 was the poorest binder. (9Z)-UAB7 and (9Z)-UAB8 had significant activity in this assay but less than that of (9Z)-RA. These retinoids were also evaluated in a binding assay for the apo-hRXRR-(DEF) domain using fluorometric titrations of the ligand. A systematic trend in the Kd′

Figure 1. (A) Evaluation of RARR receptor-mediated transcriptional activation by the all-E-isomers of RA, UAB1, UAB4, UAB7, and UAB8 at 10-6-10-9 M concentrations. (B) Evaluation of RARR receptor-mediated transcriptional activation by (9Z)-RA, (7Z)-UAB1, (9Z)-UAB4, (9Z)-UAB7, and (9Z)-UAB8 at 10-6-10-9 M concentrations. The measurements were made in triplicate at each dose. Error was estimated at 10%.

values was observed for the (9Z)-UAB retinoids; an increased binding affinity occurred for retinoids with large alkyl groups at R1 and R2 (Table 5). The dissociation constant for (9Z)-UAB8 was only 5-fold greater than that observed for (9Z)-RA. Nuclear Receptor Transcriptional Activity. To determine if the UAB retinoids exhibited functional activity within the cell, their ability to induce RAR- and RXR-mediated transcriptional activity was examined and compared to those of (all-E)- and (9Z)-RA.36 (allE)-UAB8 was equivalent to or better than (all-E)-RA in inducing the receptor-activated transcription of the reporter gene at 10-6 M (Figure 1A). For the all-Eisomers, activation of transcriptional activity decreased in going from UAB7 to UAB1, consistent with the IC50 values found for the binding to the RARs (Table 5). Activation of RARs by (9Z)-RA was slightly better than by (all-E)-RA (Figure 1B). This is consistent with previous studies6 reporting that (9Z)-RA had equal or better activity than (all-E)-RA in inducing transcriptional activation in GAL4-RAR chimeric constructs. In contrast, (9Z)-UAB4-UAB8 at 10-7 M and lower were dramatically less effective than (9Z)- or (all-E)-RA. Even at micromolar doses, these retinoids were much

3630


Muccio et al.

Figure 2. (A) Evaluation of RARR/RXRR receptor-mediated transcriptional activation by the all-E-isomers of RA, UAB1, UAB4, UAB7, and UAB8 at 10-6-10-9 M concentrations. (B) Evaluation of RARR/RXRR receptor-mediated transcriptional activation by (9Z)-RA, (7Z)-UAB1, (9Z)-UAB4, (9Z)-UAB7, and (9Z)-UAB8 at 10-6-10-9 M concentrations. The measurements were made in triplicate at each dose.

Figure 3. (A) Evaluation of RXRR receptor-mediated transcriptional activation by the all-E-isomers of RA, UAB1, UAB4, UAB7, and UAB8 at 10-6-10-9 M concentrations. (B) Evaluation of RXRR receptor-mediated transcriptional activation by (9Z)-RA, (7Z)-UAB1, (9Z)-UAB4, (9Z)-UAB7, and (9Z)-UAB8 at 10-6-10-9 M concentrations. The measurements were made in triplicate at each dose.

less effective than RA in inducing the RARR-mediated response. The activity of UAB retinoids on the transcription mediated by RARR/RXRR heterodimers was evaluated. Both (all-E)-UAB8 and (all-E)-UAB7 were nearly as efficient as (all-E)-RA in activating transcription mediated by RARR/RXRR heterodimers (Figure 2A). As observed for RARR activation, the other (all-E)-UAB retinoids displayed decreased activity. Further, induction of the reporter gene by (9Z)-UAB8 was apparent only at 10-6 M, unlike (9Z)-RA (Figure 2B). Thus, the (9Z)-UAB retinoids are extremely poor binders and activators of both RARR homodimer and RARR/RXRR heterodimers. When the effect on RXRR homodimermediated transcriptional activity was studied, (all-E)RA did not efficiently activate the reporter gene (Figure 3A), nor did any of the (all-E)-UAB retinoids. In contrast, (9Z)-RA efficiently activated this reporter gene (Figure 3B), which is consistent with previous observations.9 In this assay, the (9Z)-UAB retinoids were also highly active in inducing transcription at 10-6 M. However, relative to (9Z)-RA, transcriptional activation by (9Z)-UAB retinoids was less efficient at doses less than 10-6 M. Only (9Z)-RA caused significant activation

at concentrations as low as 10-8 M. The trends in the transactivational data are consistent with the larger Kd′ values and IC50 values, relative to (9Z)-RA, reported for (9Z)-UAB retinoids (Table 5). Taken together, these results demonstrate that, unlike (9Z)-RA, (9Z)-UAB retinoids, especially UAB7 and UAB8, are RXR-selective ligands. Mouse Skin Antipapilloma Assay. The UAB retinoids were next evaluated for their activity in preventing the chemical induction of papillomas on mouse skin.31 (all-E)-RA was very effective in this chemopreventive assay (Table 6). At a 45.9 nmol dose, it prevented 95% of tumor formation. The ED50 value obtained in the present study was 3 nmol, which is consistent with previously reported values.37 (all-E)UAB7 and UAB8 were as effective as RA in this assay (Table 6). Other (all-E)-UAB retinoids were much less effective in this assay. (all-E)-UAB4 prevented only 32% of tumor formation (45.9 nmol dose); the ED50 value was estimated at greater than 100 nmol. In comparison to RA, (all-E)-UAB1 was only marginally effective at the 45.9 nmol dose. This assay was also used to evaluate (9Z)-RA. As suggested from RAR and RXR binding and transcrip-

Conformationally Defined 6-s-trans-RA Analogs Table 6. Structure-Activity Comparisons for the Activities of UAB Retinoids in the Prevention of Mouse Skin Papillomas retinoid (all-E)-RA (all-E)-UAB8 (all-E)-UAB7 (all-E)-UAB4 (all-E)-UAB1 (9Z)-RA (9Z)-UAB8 (9Z)-UAB7 (9Z)-UAB4 (13Z)-RA (13Z)-UAB8

% papilloma reductiona at 45.9 nmol

ED50 (papilloma)b (nmol)

95 99 95 32 22 95 43 30 50

3.0 2.5 5.0 490 >1000 2.1 >100 >100 >50 64c 4.7

81

a

Percentage of papilloma reduction from the application of 45.9 nmol dose of test retinoid 1 h prior to TPA application (n ) 1). The percent papilloma reduction was calculated as (1 - tumor suppression) × 100, where tumor suppression is defined as (mean number of tumors in retinoid-treated animals)/(mean number of tumors in control). b ED50 values were determined by a probit analysis of a dose-response curve. Estimated standard error is 20% of the mean. c Value determined by Dawson et al.37

tional activational studies, (9Z)-RA also prevented papilloma formation as well as RA (Table 6). In constrast, (9Z)-UAB retinoids were uniformily poor in papilloma reduction; they prevented only 30-50% of tumor formation at the 45.9 nmol dose. Unlike (9Z)-RA, (all-E)-RA, or (all-E)-UAB retinoids, the (9Z)-UAB retinoids did not respond in a dose-dependent manner; their ED50 values are estimated to be 10-fold greater than that of (9Z)RA (Table 6). Considering the trends noted above for the (all-E)-UAB retinoids and RA, the low activities of the (9Z)-UAB compounds correlate with their poor capacity to activate RARs and not with their RXRselective actions. Interestingly, (13Z)-UAB8 exhibited very high activity in the antipapilloma assay, comparable to that for RAs and (all-E)-UAB8 (Table 6). Previously it was reported37 that (13Z)-RA is about 12-fold less active than RA in this assay. The surprisingly high activity of (13Z)-UAB8 in the antipapilloma assay correlates well with its capacity to bind to and activate RARs, as we previously demonstrated.20 Discussion We report here a strong correlation between the biological activities of the UAB retinoids and the increased size of the R1 and R2 substituents in their structures. These assays included CRABP-I, CRABPII, RARR, and RXRR. Recently, Jones and co-workers17f determined the X-ray crystal structures of (all-E)-RA bound to both CRABP-I and CRABP-II, and Moras and co-workers17g reported the structure of this hormone bound to the RARγ ligand binding domain. Despite differences in sequence homology and folds, the RA binding sites in CRABPs and RARs are primarily composed of hydrophobic amino acids. Positively charged amino acids (K and R) occupy one end of the RA binding pocket, and these residues form salt bridges and hydrogen bonds with the RA carboxylate group. The RA ring is in a well-defined hydrophobic binding site at the other end of the pocket. The RA binding pocket for RXRR and the anticipated RA binding sites in other RARs are expected to have similar features due to the high sequence homology of these proteins.17g Together the hydrophobic and ionic forces stabilize RA in a tight


protein-ligand complex (nanomolar Kd′ values). A positive electrostatic potential near the carboxylate binding site is thought, at least for RARs, to be involved in the proper docking of RA. (all-E)-RA adopts a surprisingly similar conformation in each site. The polyene chain of RA (C6-C15) is flat, and the conformation is in an extended s-trans-conformation. The trimethylcyclohexenyl ring adopts a distorted 6-s-cis-geometry relative to the chain. For CRABP-I/II the C5-C6C7-C8 torsional angle of RA is -33°. (This torsional angle was not reported by Moras and co-workers,17g but RA was depicted in a 6-s-cis-geometry.) This is slightly more planar than the torsional angle reported in the structures of (all-E)-RA crystallized in a triclinic crystal32a and for other related 6-s-cis-retinoids.32c Through the incorporation of a dimethylene bridge, the UAB retinoids were designed to maintain a 6-strans-conformation between terminal double bonds of the polyene chain. Even with this structural constraint, UAB7 and UAB8 are very active in several biological assays (comparable to RA). Since RA adopts a 6-s-cisgeometry in these binding sites, the high affinity of the UAB retinoids is unexpected in the protein binding assays. In order to explore possible explanations for the high activity of UAB7 and UAB8 in these assays, their low-energy structures were calculated from MM3(94) (Table 3) and overlaid with that of 6-s-cis-RA (Figure 4). When 8-s-trans-UAB8 was overlaid with RA, the ethyl group at R1 in UAB8 occupies similar space as the C3 and C4 of the RA ring and the isopropyl group at R2 is near C1 and the gem-dimethyl groups of the RA ring. The correspondence of these groups is due to the distorted nature of the polyene chain at the C7-C8C9-C10 bond (-140° torsional angle) in the UAB retinoids and the nonplanar orientation about the C5C6-C7-C8 bond (-59° torsional angle) in 6-s-cis-RA. When the structure of 8-s-cis-UAB8 is overlaid with that of RA, there is little agreement between the positions of R1 and R2 and the RA ring. The structure of (9Z)-RA bound to RXRs has not been determined. Using ligand-based design on conformationally rigid retinoids and resulting structure-activity studies, Dawson et al.15d suggested that RXR-selective ligands need to maintain a 10 Å distance between the C15 carboxylate and C1 and C4 of the RA ring. Longer and more planar retinoids tended to also bind to RARs, and shorter ones were not very active as RXR agonists. They also identified, among other effects, that retinoids like SR11217 with nonplanar structures (corresponding to C8-C11 of the polyene chain of RA) were potent and selective RXR agonists. We determined these distances in (9Z)-UAB7 and (9Z)-UAB8, two retinoids which also display RXR selectivity (Table 4). In the 8-s-trans-UAB retinoids, the C1-C15 and C4-C15 distances were 9.7 and 10.8 Å, respectively. For the 8-s-cis-UAB retinoids, these distances were slightly shorter (9.6 and 10.3 Å). Due to steric interactions, both the 8-s-cis- and 8-s-transconformations are expected to adopt twisted low-energy conformations about the C8-C9 bond (Table 3). The structural properties of the UAB retinoids with bulky R1 and R2 groups are consistent with known properties of other RXR-selective ligands. Since (9Z)-RA also binds tightly to RARs, interest has centered on how this configuration of RA interacts with the receptors. Moras and co-workers17g were not able

3632


Muccio et al.

Figure 4. Stereoview (relaxed) for (all-E)-RA (top) in a 6-s-cis-conformation with (all-E)-UAB8 (bottom) in an 8-s-transconformation. The structures were energy minimized using MM3(94), and the C9-C15 atoms were fit onto one another. See text and Table 3 for details.

to determine the X-ray structure of (9Z)-RA bound to RARγ. However, by allowing similar ionic interactions of the carboxylate group of (9Z)-RA and minimizing steric interactions, they suggested that (9Z)- and (allE)-RA occupy similar binding sites in RARs. On the basis of these criteria, the rings of (9Z)- and (all-E)-RA were suggested to occupy similar positions, and the single bonds for (9Z)-RA were shown in an extended s-trans-conformation like in (all-E)-RA. If (9Z)-RA, like (all-E)-RA, adopts a planar s-trans-conformation in this region, then retinoids like (9Z)-UAB8, which are significantly skewed about the C8-C9 bond (Table 3), may

introduce steric interactions with the receptor not present in (9Z)-RA. Indeed, Moras and co-workers17g suggested a close contact between the C19 methyl group in (9Z)-RA and the H5 helix in RARγ. These steric interactions may be more significant in the (9Z)-UAB retinoids resulting in their poor binding affinity relative to (9Z)-RA and their RXR-selective characteristics. The activation of RARs and RXRs by retinoids is a powerful tool for evaluating their capacity to act as novel pharmacological agents. However, to evaluate their potential in cancer chemoprevention, the mouse skin antipapilloma model is often used.37 Beard et al.15c



recently reported that RAR-selective retinoids, but not RXR-selective retinoids, are active as antiproliferative agents in skin. Chandranatna et al.14e showed that RARβ- and RARγ-selective retinoids are excellent inhibitors of ODC activity in skin, an assay which correlates well with the skin antipapilloma assay.31 We demonstrate here that (9Z)-RA is as effective as RA for the prevention of skin papillomas, and we report that (all-E)-UAB7 and UAB8 are as active as (all-E)- and (9Z)-RA in this assay. We also show that RXR-selective retinoids, like (9Z)-UAB8, are much less effective in skin cancer prevention. Zhang et al.38 demonstrated that this effect is observed in other cancer models. (all-E)UAB8 is as effective as RA and (9Z)-RA in inhibiting the growth of hormone-dependent breast cancer cells, presumably through RARβ activation; however, (9Z)UAB8 was ineffective in these actions.

(E)-Ethyl 3-Methyl-4-cyclohexylidenebut-2-enoate ((9E)-7). NaH (60% dispersion in mineral oil, 1.24 g, 31.0 mmol) was washed under N2 with dry THF (3 × 10 mL) to remove the mineral oil. Anhydrous THF (10 mL) was added, the suspension was cooled to 0 °C, and freshly distilled diethyl 3-(ethoxycarbonyl)-2-methylprop-2-enylphosphonate (5.40 g, 20.4 mmol) was added. After stirring for 30 min, HMPA (7.30 g, 40.7 mmol) followed by cyclohexanone (2.00 g, 20.4 mmol) was introduced and stirring was continued at 0 °C for 2 h. At that time the reaction was carefully quenched by adding, dropwise, ice-cold water (10 mL). The resulting mixture was extracted with ether (3 × 20 mL); the combined ether layers were washed once with brine (15 mL), dried (Na2SO4), and concentrated under vacuum to give crude ester 7 (4.32 g). A hexane solution of crude 7 was purified by flash chromatography on silica gel (400 g, 2% Et2O in hexane eluent, Rf 0.64) to provide pure 7 (4.10 g, 96.0%) as a pale yellow oil: UV (λmax) 275 nm; 1H NMR, IR, and mass spectra were identical with those previously reported21 for 7. (2E,4E)-Ethyl 4-(2′-Cyclohexen-1′-ylidene)-3-methyl-2butenoate ((all-E)-8). To a solution of (9E)-7 (3.00 g, 14.4 mmol) in anhydrous CH2Cl2 (30 mL), under N2 at 0 °C, were added NBS (2.80 g, 15.7 mmol), CaO (0.660 g, 11.8 mmol), and NaHCO3 (2.18 g, 26.0 mmol). The cooling bath was removed and the reaction mixture stirred at room temperature for 8 h. The mixture was filtered through a sintered glass funnel to remove succinimide, and the filtrate was concentrated under vacuum. Freshly distilled quinoline (1.97 g, 15.2 mmol) was added to the residue and the mixture was stirred at 100 °C for 2 h. The brownish mixture was cooled to room temperature and dissolved in ether (250 mL), and the resulting solution was washed with 5% H2SO4 (2 × 150 mL), water (100 mL), and saturated NaHCO3 (150 mL). The ether solution was then dried (Na2SO4) and concentrated under vacuum to provide (allE)-8 (1.88 g, 63.0%) as an oil. While (all-E)-8 was sufficiently pure to carry on in this form, a sample (0.48 g) was further purified by flash chromatography (1% Et2O/hexane) to give pure (all-E)-8 (0.40 g, Rf 0.80, 10% EtOAc/hexane). (2E,4E)-4-(2′-Cyclohexen-1′-ylidene)-3-methyl-2-buten1-ol ((all-E)-9). To a solution of (all-E)-8 (1.33 g, 6.45 mmol) in anhydrous ether (10 mL) under N2 at -78 °C was added, dropwise, a solution of 1 M LiAlH4 in ether (6.50 mL, 6.50 mmol). The mixture was allowed to warm to 0 °C, stirred for 1 h, and cooled to -78 °C. To this mixture was added, dropwise, methanol (5 mL) followed by 5% H2SO4 (20 mL). The reaction mixture was allowed to warm to room temperature and extracted with ether (2 × 20 mL). The combined ether layers were washed with brine (30 mL), dried (Na2SO4), and concentrated under vacuum to provide (all-E)-9 (0.98 g, 91%) as an oil. This alcohol was not further purified but immediately oxidized to the aldehyde. Biology. The chick skin CRABP binding assay measured IC50 values for retinoid binding to CRABP-II using a radiolabeled competition assay.26 The Kd′ values used a fluorescence titration method27,28 to measure the affinities of the retinoids to recombinant mCRABP-I, mCRABP-II, and hRXRR-(DEF). The IC50 values for retinoids with RARs and RXRs were measured with a radioligand competition assay.20 The nuclear receptor transcriptional activity assays were performed using CV-1 cells. Transient transfection of these cells with a DNA plasmid was performed essentially as described in Alam et al.20 The mouse skin antipapilloma assay measured the ED50 values for retinoid tumor inhibition on the dorsal skin of mice. The inhibition of mouse skin papilloma by retinoids was performed according to a modification of the procedure developed by Verma and Boutwell31 as reported previously.39 Molecular Modeling. Retinoid structures were generated with Sybyl version 6.2 (Tripos Inc., St. Louis, MO) on a Silicon Graphics Indigo 2 workstation. RA structures were built using the x-ray crystal structure torsional angles.32 The RA ring is in a half-chair conformation with C1, C6, C5, and C4 essentially planar and C2 and C3 above and below this plane.32a Another ring conformation occurs in solution due to conversion to the other half-chair conformer. Since ring inversion does not occur in the minimization, the calculated RA structures contained the half-chair conformation reported in the crystal

Conclusions We synthesized two new examples (UAB1 and UAB4) from a new class of conformationally constrained 6-strans-retinoids to form a homologous series: UAB1, UAB4, UAB7, and UAB8. We find that both protein affinity and nuclear receptor activation improve with increasing size at R1 and R2. Assuming that these R groups mimic the steric space of the RA ring, these results demonstrate the importance of this ring in achieving high biological activities in several assays. We also show that the structural changes contained in the UAB retinoids generate either RAR- or RXR-selective retinoids and that only RAR-selective ligands are effective agents for skin cancer prevention. If RA mediates its cancer chemoprevention and toxicity effects through nuclear receptors, the similar activity and lower toxicity20,39 of UAB7 and UAB8 (relative to RA) may result from their nuclear receptor-selective actions. Experimental Section Chemistry. General Methods. 1H NMR spectra were obtained at 300.1 MHz (Bruker ARX spectrometer) and NOE experiments were performed on degassed samples.19 UV/vis spectra were recorded on an AVIV 14DS spectrophotometer in cyclohexane solution (Fisher, Spectrograde). IR spectra were recorded using a Nicolet FT IR spectrometer on films. Electron-impact mass spectra were obtained on a Hewlett Packard 5985 GC/MS instrument with an ultraperformance fused silica gel column. HPLC separations were performed on a Gilson HPLC gradient system using 25 mL pump heads and an ISCO V′4 variable wavelength detector. The column employed was a Whatman Partisil 10 M20/50 (500 × 22 mm i.d.). Reverse-phase HPLC separations on carboxylic acids were performed as described previously20 using a Chromanetics Spherisorb ODS 5 µm column (250 × 4.6 mm i.d.). TLC chromatography was performed on precoated 250 µm silica gel GF glass plates (Analtech, Inc.; 5 × 10 cm). Solvents and liquid starting materials were distilled prior to use. Reactions and purifications were conducted with deoxygenated solvents, under inert gas (N2) and subdued lighting. Diethyl 3-(ethoxycarbonyl)-2-methylprop-2-enylphosphonate was synthesized according to Iqbal et al.40 2,3-Dimethyl-2cyclohexenone (1) (Scheme 1) was prepared from 1,3-dimethoxybenzene using the method of Jung et al.41 All synthetic procedures for the conversion of 1 to UAB4 (Scheme 1) were the same as those we previously reported in detail for UAB7 and UAB8.19,20 Referring to Scheme 2, the preparation of ester (9E)-7 was reported by us in modest yields.21 A modified procedure giving much better yield is described here. All procedures involved in converting alcohol (all-E)-9 to UAB1 were the same as those utilized in Scheme 1 for comparable transformations.

3634


structure. The structures were minimized using Maximin (Sybyl), MM2, MM2(91), and MM3(94). Allinger’s MM3(94) force field produced the structure that most closely corresponded to the X-ray structures. The 8-s-cis-RA structures were generated by rotating just the C7-C8-C9-C10 bond to the s-cis-conformation followed by energy minimization. The (9Z)-RA structure was obtained by changing the configuration of this bond to Z followed by energy minimization. The resulting torsional angles agreed well with those found in the X-ray structure of (9Z)-retinal. The 8-s-cis-conformers were generated directly from the 8-s-trans-conformers. The structures of the UAB retinoids were generated in a similar manner. Since X-ray crystal structures of final compounds are not available, a closely related synthetic intermediate which had a solved structure (manuscript in preparation) was chosen for validation of the structure calculation. This energy-minimized structure was used to generate the corresponding UAB retinoids as described for RA. The thermodynamic parameters (G, H, and S) of each structure were calculated at 300 K by MM3(94) using a fullmatrix diagonalization. To generate overlapping structures, the C9-C15 carbons including C19 and C20 of RA and the UAB retinoids were fit onto each other. The Connolly surface area of each structure was obtained using a 1.4 Å probe and default parameters in Sybyl.

Acknowledgment. These studies were supported in part by grants PO1 CA34968 (D.D.M., W.J.B., D.L.H., B.P.S.), DK40172 (E.L.), RCDADK02072 (E.L.), AICR 93B39 (B.P.S.), and RO1 CA59446 (B.P.S.). References (1) For recent reviews, see: (a) Dawson, M. A., Okamura, W. H., Eds. Chemistry and Biology of Synthetic Retinoids; CRC Press: Boca Raton, FL, 1990. (b) Sporn, M. B., Roberts, A. B., Goodman, D. S., Eds. THE RETINOIDS Biology, Chemistry and Medicine, 2nd ed.; Raven Press: New York, 1994. (2) Thaller, C.; Eichele, G. Identification and Spatial Distribution of Retinoids in the Developing Chick Limb Bud. Nature 1987, 327, 624-628. (3) Rosen, J.; Day, A.; Jones, T. K.; Jones, T. T.; Nadzan, A. M.; Stein, R. B. Intracellular Receptors and Signal Transducers and Activators of Transcription Superfamilies: Novel Tragets for Small-Molecule Drug Discovery. J. Med. Chem. 1995, 38, 48554878. (4) For recent reviews, see: (a) Mangelsdorf, D. J.; Umesons, K.; Evans, R. M. The Retinoid Receptors. In THE RETINOIDS Biology, Chemistry and Medicine, 2nd ed.; Sporn, M. B., Roberts, A. B., Goodman, D. S., Eds.; Raven Press: New York, 1994; pp 319-349. (b) Gudas, L. Retinoids and Vertebrate Development. J. Biol. Chem. 1994, 269, 15399-15402. (5) (a) Petkovich, M.; Brand, N. J.; Krust, A.; Chambon, P. A Human Retinoic Acid Receptor which Belongs to the Family of Nuclear Receptors. Nature (London) 1987, 330, 444-450. (b) Giguere, V.; Ong, E. S.; Prudimar, S.; Evans, R. M. Identification of a Receptor for the Morphogen Retinoic Acid. Nature (London) 1987, 330, 624-629. (6) Allenby, G.; Janocha, R.; Kazmer, S.; Speck, J.; Grippo, J. F.; Levin, A. A. Binding of 9-cis-Retinoic Acid and All-trans-Retinoic Acid to Retinoic Acid Receptors R, β, and γ. J. Biol. Chem. 1994, 269, 16689-16695. (7) Mangelsdorf, D. J.; Ong, E. S.; Dyck, J. A.; Evans, R. M. A Nuclear Receptor that Identifies a Novel Retinoic Acid Response Pathway. Nature (London) 1990, 345, 224-229. (8) (a) Levin, A. A.; Sturzenbecker, L. J.; Kazmer, S.; Bosakowski, T.; Huselton, C.; Allenby, G.; Speck, J.; Kratzeisen, C.; Rosenberger, M.; Lovey, A.; Grippo, J. F. 9-Cis Retinoic Acid Stereoisomer Binds and Activates the Nuclear Receptor RXRR. Nature (London) 1992, 355, 359-361. (b) Heyman, R. A.; Mangelsdorf, D. J.; Dyck, J. A.; Stein, R. B.; Eichele, G.; Evans, R. M.; Thaller, C. 9-Cis Retinoic Acid is a High Affinity Ligand for the Retinoid X Receptor. Cell 1992, 68, 397-406. (9) Zhang, X.-k.; Lehmann, J. M.; 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 (London) 1992, 358, 587-591. (10) (a) Moon, R. C.; Mehta, R. G.; Rao, K. V. N. Retinoids and Cancer in Experimental Animals. In THE RETINOIDS Biology, Chemistry and Medicine, 2nd ed.; Sporn, M. B., Roberts, A. B., Goodman, D. S., Eds.; Raven Press: New York, 1994; pp 573630. (b) Nadzan, A. M. Retinoids for the Treatment of Oncological Diseases. Annual Reports in Medicinal Chemistry; Academic Press, Inc.: New York, 1995; Vol. 30, pp 119-128.

Muccio et al. (11) (a) Hixson, E. J.; Denine, E. P. Comparative Subacute Toxicity of all-trans- and 13-cis-Retinoic Acid in Swiss Mice. Toxicol. Appl. Pharmacol. 1978, 44, 29-40. (b) Kamm, J. J. Toxicology, Carcinogenicity, and Teratogenicity of some Orally Administered Retinoids. J. Am. Acad. Dermatol. 1982, 6, 652-659. (c) Cohen, M. Tretinoin: A Review of Preclinical Toxicological Studies. Drug Dev. Res. 1993, 30, 244-251. (12) (a) Kochhar, D. M. Teratogenic Activity of Retinoic Acid. Acta Pathol. Microbiol. Immunol. 1967, 70, 398-404. (b) Willhite, C. C. In Chemistry and Biology of Synthetic Retinoids; Dawson, M. I., Okamura, W. H., Eds.; CRC Press: Boca Raton, FL, 1990; pp 539-573. (c) Adams, J. Structure-Activity and Dose-Response Relationships in the Neural and Behavioral Teratogenesis of Retinoids. Neurotoxicol. Teratol. 1993, 15, 193-202. (13) Armstrong, R. B.; Ashenfelter, K. O.; Eckhoff, C.; Levin, A. A.; Shapiro, S. General and Reproductive Toxicology of Retinoids. In THE RETINOIDS Biology, Chemistry and Medicine, 2nd ed.; Sporn, M. B., Roberts, A. B., Goodman, D. S., Eds.; Raven Press: New York, 1994; pp 545-572. (14) (a) Crettaz, M.; Baron, A.; Siegenathaler, G.; Hunziker, W. Ligand Specificities of Recombinant Retinoic Acid Receptors RARR and RARβ. Biochem. J. 1990, 272, 391-397. (b) Graupner, G.; Malle, G.; Maignan, J.; Lang, G.; Prunieras, M.; Pfahl, M. 6′-Substituted Naphtalene-2-carboxylic Acid Analogs, A New Class of Retinoic Acid Receptor Subtype-Specific Ligands. Biochem. Biophys. Res. Commun. 1991, 179, 1554-1561. (c) Delescluse, C.; Cavey, M. T.; Martin, B.; Bernard, B. A.; Reichert, U.; Maignan, J.; Darmon, M.; Shroot, B. Selective High Affinity Retinoic Acid Receptor R or β-γ Ligands. Mol. Pharmacol. 1991, 40, 556-562. (d) Fukasawa, H.; Iijima, T.; Kagechika, H.; Hashimoto, Y.; Shudo, K. Expression of the Ligand-Binding Domain-Containing Region of Retinoic Acid Receptors R, β, and γ in Escherichia coli and Evaluation of Ligand-Binding Selectivity. Biol. Pharm. Bull. 1993, 16, 343-348. (e) Chandraratna, R. A. S.; Gillett, S. J.; Song, T. K.; Attard, J.; Vuligonda, S.; Garst, M. E.; Arefieg, T.; Gil, D. W.; Wheeler, L. Synthesis and Pharmacological Activity of Conformationally Restricted, Acetylenic Retinoid Analogs. BioMed. Chem. Lett. 1995, 5, 523-527. (15) (a) Jong, L.; Lehmann, J. M.; Hobbs, P. D.; Harlev, E.; Huffmann, J. C.; Pfahl, M.; Dawson, M. A. Conformational Effects on Retinoid Receptor Selectivity. 1. Effect of 9-Double Bond Geometry on Retinoid X Receptor Activity. J. Med. Chem. 1993, 36, 2605-2613. (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 Retinoid X Receptor-Selective Retinoids. J. Med. Chem. 1994, 37, 2930-2941. (c) 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.; Gill, D. W.; Wheeler, L.; Kochhar, D. M.; Davies, P. J. A. Synthesis and Structure-Activity Relationships of Stilbene Retinoid Analogs Substituted with Heteroaromatic Carboxylic Acids. J. Med. Chem. 1995, 38, 2820-2829. (d) Dawson, M. I.; Jong, L.; Hobbs, P. D.; Cameron, J. F.; Chao, W.-r.; Pfahl, M.; Lee, M.-O.; Shroot, B.; Pfahl, M. Conformational Effects on Retinoid Receptor Selectivity. 2. Effects of Retinoid Bridging Group on Retinoid X Receptor Activity and Selectivity. J. Med. Chem. 1995, 38, 33683383. (16) Muccio, D. D.; Brouillette, W. J. (UAB Research Foundation) Retinoid Compounds. U.S. Patent 5,094,783, Mar. 10, 1992. (17) (a) Newcomer, M. E.; Jones, T. A.; Aqvist, J.; Sundelin, J.; Eriksson, U.; Rask, L.; Petersen, P. A. The Three-dimensional Structure of Retinol Binding Protein. EMBO J. 1984, 3, 14511454. (b) van der Steen, R.; Biesheuvel, P. L.; Mathias, R. A.; Lugtenburg, J. Retinal Analogues with Locked 6-7 Conformations Show that Bacteriorhodopsin Requires the 6-s-Trans Conformation of the Chromophore. J. Am. Chem. Soc. 1986, 108, 6410-6411. (c). Smith, S. O.; Palings, I.; Copie, V.; Raleigh, D. P.; Courtin, J.; Pardoen, J. A.; Lugtenburg, J.; Mathias, R. A.; Griffin, R. G. Low-Temperature Solid-State 13C NMR Studies of the Retinal Chromophore in Rhodopsin. Biochemistry 1987, 26, 1606-1611. (d) Winter, N. S.; Bratt, J. W.; Banaszak, L. J. Crystal Structures of Holo and Apo-cellular Retinol-binding Protein II. J. Mol. Biol. 1993, 230, 1247-1259. (e) Newcomer, M. E.; Pappas, R. S.; Ong, D. E. X-ray Crystallographic Identification of a Protein-binding Site for both all-trans- and 9-cisRetinoic Acid. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 9223-9227. (f) Keywegt, G. J.; Bergfor, T.; Senn, H.; Le Motte, P.; Gsell, B.; Shudo, K.; Jones, T. A. Crystal Structures of Cellular Retinoic Acid Binding Proteins I and II in complex with all-trans-Retinoic Acid and a Synthetic Retinoid. Structure 1994, 2, 1241-1258. (g) Renaud, J.-P.; Rochel, N.; Ruff, M.; 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. (18) Honig, B.; Hudson, B.; Sykes, B.; Karplus, M. Ring Orientation in β-Ionone and Retinals. Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1289-1293.

Conformationally Defined 6-s-trans-RA Analogs (19) Vaezi, M. F.; Alam, M.; Sani, B. P.; Rogers, T. S.; SimpsonHerren, L.; Wille, J. J.; Hill, D. L.; Doran, T. I.; Brouillette, W. J.; Muccio, D. D. A Conformationally Defined 6-s-trans-Retinoic Acid Isomer. Synthesis, Chemopreventive Activity, and Toxicity. J. Med. Chem. 1994, 37, 4499-4507. (20) Alam, M.; Zhestkov, V.; Sani, B. P.; Venepally, P.; Levin, A. A.; Kazmer, S.; Li, E.; Norris, A. W.; Zhang, X.-k.; Lee, M.-O.; Hill, D. L.; Lin, T.-H.; Brouillette, W. J.; Muccio, D. D. Conformationally Defined 6-s-trans-Retinoic Acid Analogs. 2. Selective Agonists for Nuclear Receptor Binding and Transcriptional Activity. J. Med. Chem. 1995, 38, 2302-2310. (21) Robinson, C. Y.; Brouillette, W. J.; Muccio, D. D. Reactions of Vinylogous Phosphonate Carbanions and Reformatsky Reagents with a Sterically Hindered Ketone and Enone. J. Org. Chem. 1989, 54, 1992-1997. (22) Vaezi, M. F.; Robinson, C. Y.; Hope, K. D.; Brouillette, W. J.; Muccio, D. D. Preparation of the 9-cis, 13-cis, and All transIsomers of R- and β2-Retinal. Org. Prep. Proc. Int. 1987, 19, 187195. (23) Hale, R. L.; Burger, W.; Perry, C. W.; Liebman, A. A. Preparation of High Specific Activity All Trans-R-Retinyl-11-3H Acetate. J. Labelled Compds. Radiopharm. 1977, 13, 123-135. (24) Vaezi, M. F.; Brouillette, W. J.; Muccio, D. D. 13C NMR Assignments of Conformationally Defined 6-s-trans-Retinoids. Magn. Reson. Chem. 1995, 33, 497-499. (25) Hope, K. D.; Robinson, C. Y.; Vaezi, M. F.; Brouillette, W. J.; Muccio, D. D. 13C NMR Assignments of the Isoprenoid Chain Carbons of Retinoids from Empirical Chemical Shift Differences. Magn. Reson. Chem. 1987, 25, 1040-1045. (26) Sani, B. P.; Titus, B. C.; Banerjee, C. K. Determination of Binding Affinities of Retinoids to Retinoic Acid-Binding Protein and Serum Albumin. Biochem. J. 1978, 171, 711-717. (b) Sani, B. P. Cellular Retinoic Acid-Binding Protein and the Action of Retinoic Acid. In Chemistry and Biology of Synthetic Retinoids; Dawson, M. I., Okamuar, W. H., Eds.; CRC Press, Inc.: Boca Raton, FL, 1990; pp 365-384. (27) Norris, A. W.; Cheng, L.; Giguere, V.; Rosenberger, M.; Li, E. Measurement of Subnanomolar Retinoic Binding Affinities for Cellular Retinoic Acid Binding Proteins by Fluorometric Titration. Biochim. Biophys. Acta. 1994, 1209, 10-18. (28) Cheng, L.; Norris, A. W.; Tate, B. F.; Rosenberger, M.; Grippo, J. F.; Li, E. Characterization of the Ligand Binding Domain of Human Retinoid X Receptor R Expressed in Escherichia coli. J. Biol. Chem. 1994, 269, 18662-18667. (29) Allenby, G.; Bocquel, M.-T.; Sounders, M.; Karmer, S.; Speck, J.; Rosenberger, M.; Lovey, A.; Kastner, P.; Grippo, J. F.; Chambon, P.; Levin, A. A. Retinoic Acid Receptors and Retinoid X Receptors: Interactions with Endogenous Retinoic Acids. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 30-34. (30) Zhang, X.-k.; Hoffmann, B.; Tran, P.; Graupner, G.; Pfahl, M. Retinoid X Receptor is an Auxilliary Protein for Thyroid Hormone and Retinoic Acid Receptors. Nature (London) 1992, 355, 441-446.

Journal of Medicinal Chemistry, 1996, Vol. 39, No. 19 3635 (31) Verma, A.; Boutwell, R. Vitamin A Acid (Retinoic Acid), a Potent Inhibitor of 12-O-Tetradecanoylphorbol-13-acetate-induced Ornithine Decarboxylase Activity in Mouse Epidermis. Cancer Res. 1977, 37, 2196-2201. (32) (a) Stam, C. H.; MacGillavry, C. H. The Crystal Structure of the Triclinic Modification of Vitamin-A Acid. Acta Crystallogr. 1963, 16, 62-68. (b) Simmons, C. The Structures of 9-cis-Retinal and 19,19,19-Trifluoro-9-cis-retinal. Acta Crystallogr. 1986, C42, 1558-1563. (c) Simmons, C. J.; Liu, R. S. H.; Denny, M.; Seff, K. The Crystal Structure of 13-cis-Retinal. The Molecular Structures of its 6-s-cis and 6-s-trans Conformers. Acta Crystallogr. 1981, B37, 2197-2205. (33) Sani, B. P.; Grippo, J. F.; Levin, A. A. 9-cis-Retinoic Acid Binds Cellular Retinoic Acid-binding Protein with Low Affinity. Oncol. (Life Sci. Adv.) 1994, 13, 21-27. (34) Szuts, E. Z.; Harosi, F. I. Solubilities of Retinoids in Water. Arch. Biochem. Biophys. 1991, 287, 297-304. (35) Birdsall, B.; King, R. W.; Wheeler, M. R.; Lewis, C. A., Jr.; Goode, S. R.; Dunlap, R. B.; Roberts, G. C. K. Correction for Light Absorption in Fluorescence Studies of Protein-Ligand Interations. Anal. Biochem. 1983, 132, 353-361. (36) Graupner, G.; Wills, K. N.; Tzukerman, M.; Zhang, X.-k.; Pfahl, M. Dual Regulatory Role for Thyroid Hormone Receptors Allows Control of Retinoic Acid Receptor Activity. Nature (London) 1989, 340, 653-656. (37) Dawson, M. I.; Chao, W.-R.; Hobbs, P. D.; Delair, T. The Inhibitory Effects of Retinoids on the Induction of Ornithine Decarboxylase and the Promotion of Tumors in Mouse Epidermis. In Chemistry and Biology of Synthetic Retinoids; Dawson, M. I., Okamura, W. H., Eds.; CRC Press, Inc.: Boca Raton, FL, 1990; pp 384-466. (38) Liu, Yi.; Lee, M.-O.; Li, Y.; Zhang, X.-k. RARβ Mediates the Growth Inhibitory Effect of Retinoic Acid by Promoting Apotosis and is Abnormally Expressed in Hormone-independent Human Breast Cancer Cells. Mol. Cell. Biol. 1996, 16, 1138. (39) Lin, T.-H.; Rogers, T. S.; Hill, D. H.; Simpson-Herren, L.; Farnell, D. R.; Kochhar, D. M.; Alam, M.; Brouillette, W. J.; Muccio, D. D. Murine Toxicology and Pharmacology of UAB-8, a Conformationally Constrained Analog of Retinoic Acid. Toxicol. Appl. Pharmacol. 1996, 139, in press. (40) Iqbal, M.; Copan, W. G.; Muccio, D. D.; Mateescu, G. D. Synthesis of 13C Single and Double Labelled Retinals, Precursors for NMR Studies of Visual Pigments and Related Systems. J. Labelled Compds. Radiopharm. 1985, 22, 807-817. (41) Jung, M. E.; McCombs, C. A.; Takeda, Y.; Pan, Y. Use of Silyloxydienes in Synthesis. Total Synthesis of the Sesquiterpene (()-Seychellene. J. Am. Chem. Soc. 1981, 103, 6677-6685.

JM9603126

Conformationally Defined 6-s-trans-Retinoic Acid Analogs. 3. Structure

Recommend Documents