Modeling, synthesis, and biological evaluation of potential retinoid X receptor (RXR) selective agonists: Novel analogues of 4-[1-(3,5,5,8,8- Pentamethyl-5,6,7,8-tetrahydro-2-naphthyl)ethynyl]benzoic Acid (Bexarotene) and (E)-3-(3-(1,2,3,4-tetrahydro-1,1,

Article

pubs.acs.org/jmc

Modeling, Synthesis, and Biological Evaluation of Potential Retinoid

X Receptor (RXR) Selective Agonists: Novel Analogues of 4‑[1-

(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydro-2-

naphthyl)ethynyl]benzoic Acid (Bexarotene) and (E)‑3-(3-(1,2,3,4-

tetrahydro-1,1,4,4,6-pentamethylnaphthalen-7-yl)-4-

hydroxyphenyl)acrylic Acid (CD3254)

Peter W. Jurutka,^†,∥Ichiro Kaneko,^†,∥Joanna Yang,^†Jaskaran S. Bhogal,^†Johnathon C. Swierski,^†

Christa R. Tabacaru,^†Luis A. Montano,^†Chanh C. Huynh,^†Rabia A. Jama,^†Ryan D. Mahelona,^†

Joseph T. Sarnowski,^†Lisa M. Marcus,^†Alexis Quezada,^†Brittney Lemming,^†Maria A. Tedesco,^†

Audra J. Fischer,^†Said A. Mohamed,^†Joseph W. Ziller,^‡Ning Ma,^§Geoﬀrey M. Gray,^§

Arjan van der Vaart,^§Pamela A. Marshall,^†and Carl E. Wagner*^,†

^†School of Mathematical and Natural Sciences, New College of Interdisciplinary Arts and Sciences, Arizona State University, 4701

West Thunderbird Road, Glendale, Arizona 85306, United States

^‡Department of Chemistry, University of California, Irvine, 576 Rowland Hall, Irvine, California 92697, United States

^§Department of Chemistry, University of South Florida, 4202 East Fowler Avenue, CHE 205, Tampa, Florida, 33620, United States

S

* Supporting Information

ABSTRACT: Three unreported analogues of 4-[1-(3,5,5,8,8-

pentamethyl-5-6-7-8-tetrahydro-2-naphthyl)ethynyl]benzoic

acid (1), otherwise known as bexarotene, as well as four novel

analogues of (E)-3-(3-(1,2,3,4-tetrahydro-1,1,4,4,6-pentame-

thylnaphthalen-7-yl)-4-hydroxyphenyl)acrylic acid (CD3254),

are described and evaluated for their retinoid X receptor

(RXR) selective agonism. Compound 1 has FDA approval as a

treatment for cutaneous T-cell lymphoma (CTCL), although

treatment with 1 can elicit side-eﬀects by disrupting other

RXR-heterodimer receptor pathways. Of the seven modeled

novel compounds, all analogues stimulate RXR-regulated transcription in mammalian 2 hybrid and RXRE-mediated assays,

possess comparable or elevated biological activity based on EC₅₀proﬁles, and retain similar or improved apoptotic activity in

CTCL assays compared to 1. All novel compounds demonstrate selectivity for RXR and minimal crossover onto the retinoic acid

receptor (RAR) compared to all-trans-retinoic acid, with select analogues also reducing inhibition of other RXR-dependent

pathways (e.g., VDR-RXR). Our results demonstrate that further improvements in biological potency and selectivity of

bexarotene can be achieved through rational drug design.

speciﬁc hormone responsive element (HRE) on DNA. While

an increasing number of HREs have been located at a

considerable distance either upstream or downstream from

their regulated genes, many HREs are often found within or

close to the promoter region of the regulated gene. Typically,

HREs consist of minimal core hexad sequences comprising two

half-sites punctuated by a nucleotide spacer of variable length

between direct, inverted, or everted repeats.²Nuclear receptor

proteins, as homodimers or heterodimers, regulate transcription

when binding to their corresponding half-sites within the

HREs.

INTRODUCTION

■

The human retinoid X receptors (hRXRs) consist of three

identiﬁed isoforms (α, β, γ)¹that function as transcription

factors, often in partnership with other members of a larger

nuclear receptor (NR) superfamily of transcriptional regulators

including the thyroid hormone receptor (TR), the vitamin D

receptor (VDR), the liver X receptor (LXR), the peroxisome

proliferator-activated receptor (PPAR), and the retinoic acid

receptor (RAR), to name a few. All of these nuclear receptors

generally promote or modulate gene transcription in the

presence of the appropriate receptor ligand. This ligand, usually

comprising an endogenous molecule, binds to the receptor

ligand-binding domain (LBD), eﬀecting a conformational shift

that enables the receptor to associate with high aﬃnity to its

Received: June 10, 2013

Published: November 1, 2013

8432

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

While RAR, VDR, and TR were initially thought to bind to

their HREs as homodimers,³they actually bind as heterodimers

with RXR.⁴When RXR binds to its endogenous 9-cis-retinoic

acid (9-cis-RA) agonist ligand, an RXR homodimer forms that

subsequently binds to the RXR responsive element (RXRE).

However, when RXR operates as a heterodimeric partner with

other nuclear receptors, it can do so either in the presence or

absence of a ligand in the RXR LBD. For the RXR-VDR

heterodimer, evidence suggests that RXR is unoccupied.⁵

Alternatively, data suggest that RXR may possess a ligand

when it operates as a heterodimer with LXR.⁶Notably, RXR

has been described as the central NR regulator because it often

plays a critical role in heterodimer complex formation with the

other NRs to allow DNA binding to their respective response

elements.⁷

benzoic acid (1)¹⁷was shown to be a highly speciﬁc RXR

agonist that Ligand Pharmaceuticals Inc. further developed to

become an FDA approved drug, commonly known as

bexarotene, for treatment of cutaneous T-cell lymphoma.

Several analogues of bexarotene, some quite similar such as the

disilabexarotene (2),¹⁸have since been reported and shown to

possess similarly speciﬁc RXR agonism.

These and other observations employing a number of studies

with RXR, rexinoids, and a wide range of nuclear receptors,

have led to the classiﬁcation of at least two types of RXR-

heterodimer complexes. These have been termed non-

permissive or permissive heterodimers. Purely nonpermissive

heterodimers can only be activated by the ligand of the RXR

partner receptor, while permissive heterodimers that can be

activated either by an RXR ligand or by a partner-speciﬁc

ligand.⁸The nonpermissive heterodimers include RXR−TR,

RXR−VDR, and RXR−RAR. Under most (but not all)

conditions, RXR is completely “silent” in TR and VDR

heterodimers, while RXR−RAR heterodimers can be further

activated by select RXR ligands in the presence of an RAR

ligand. Furthermore, in some cases, unique rexinoids have also

been identiﬁed that stimulate RXR−RAR, even in the absence

of a RAR ligand.⁹Thus, RXR−RAR can also be considered

conditionally nonpermissive. In contrast, the fully permissive

heterodimers include the RXR−LXRs, RXR−PPARs, and

RXR−FXR.

The development of RXR-speciﬁc ligands for therapeutic use

is complicated by the observation that not only are some RXR

heterodimers permissive but also that under certain cellular

conditions, or in particular cell types, the pool of RXR may be

somewhat limiting. This has led to the observation that RXR

ligands, like 9-cis RA for example, can inhibit nonpermissive

nuclear receptor transactivation by such receptors as VDR¹⁰

and TR.¹¹Moreover, nuclear receptor ligands like thyroid

hormone or 1,25-dihydroxyvitamin D₃(1,25D) that stimulate

RXR heterodimerization with their cognate receptor are

thought to participate in crosstalk inhibition of signaling by

other receptors that also utilize the limited intracellular RXR

pool, and this has been reported for T₃−TR−RXR-mediated

inhibition of VDR transactivation¹²and 1,25D−VDR−RXR-

directed inhibition of TR activity^12b,13although the mechanism

involved in cross-receptor squelching likely extends beyond just

sequestration of RXR. Nonetheless, a central issue in the

development of novel therapeutic rexinoids is the creation of

compounds that are likely to achieve a greater degree of

heterodimer selectivity. Thus, changing the structure of the

central RXR heteropartner ligand may result in speciﬁc NR

modulators (SNuRMs) that possess unique characteristics that

may exert a novel combination of NR control.¹⁴

While 1 is primarily used to treat CTCL, it has also been

explored as a potential treatment for colon cancer¹⁹and breast

cancer,²⁰and it has even been applied as an oﬀ-label treatment

for lung cancer²¹because promoting the transcription of RXR-

regulated genes seems to suppress uncontrolled cell prolifer-

ation and predispose cancer cells to undergo apoptosis during

chemotherapy.²⁰Several RXR selective agonists, including

compound 1, have been explored in mouse models of

noninsulin-dependent diabetes mellitus (NIDDM) because of

the ability of RXR to partner with PPAR.²²Despite the

speciﬁcity that compound 1 displays for RXR in comparison to

RAR, the known side eﬀects of 1 include hypothyroidism,²³

cutaneous toxicity, and hyperlipidemia. These side eﬀects of 1

are likely incurred either by nonpermissive receptor antago-

nism, as has been noted for TR²⁴to explain RXR agonist

induced hypothyroidism, or permissive receptor agonism, such

as LXR to eﬀect hyperlipidemia²⁵or RAR to eﬀect cutaneous

toxicity²⁶at typical dose concentrations. Thus, there is

compelling motivation to develop novel RXR selective agonists

whose eﬀects on other receptor mediated pathways are

attenuated or altogether absent. In addition to this motivation,

several RXR agonists are being studied to treat other diseases

through RXR-impacted pathways, now including two separate

clinical trials of 1 in Alzheimer’s disease (AD) based on up-

regulation of apoE in mouse models of AD.²⁷

There are several RXR speciﬁc agonists modeled on 9-cis RA

and 1 reported in literature. For example, the cyclopropyl

dienoic acid (3)²⁸is a potent RXR agonist. Novel aza-retinoids,

exempliﬁed by compound 4,²⁹as well as amide retinoids,³⁰have

been described. RXR agonists based on aryl-trienoic acid

compounds either unbranched³¹or locked with one³²or

multiple-fused³³ring systems were developed by Boehm and

co-workers, of which 5³³demonstrates the latter. Our group

has reported that the addition of a single ﬂuorine atom ortho to

the carboxylic acid group of 1, giving compound 6,³⁴and the

addition of two ﬂuorine atoms, giving compound 7,³⁵increases

RXR agonism in human colon cancer cell lines, Caco-2 and

HCT-116, respectively. Incorporating a pyridine ring, com-

pound 8³⁶has been reported to be equally potent as 1, and

substituting the alkene bridging group in 8 for a cyclopropyl

bridging group yields LGD100268 (9),³⁶with a reported

potency 1 order of magnitude greater than 1 in a CV-1 cell

RXR selective ligand (rexinoid) SNuRMs present attractive

medicinal targets because selectively activating RXR versus

RAR confers chemotherapeutic eﬀects for several human

cancers without incurring negative crossover RAR eﬀects.¹⁵

Following lengthy synthesis and testing¹⁶of molecules modeled

after 9-cis RA, 4-[1-(3,5,5,8,8-pentamethyltetralin-2-yl)ethynyl]-

8433

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

line.³⁶An unsaturated analogue of 1, compound 10,³⁷has a

reported synthetic route that can provide preclinical trial

quantities,³⁸although to our knowledge, the ability of 10 to

bind and activate RXR has not been reported. Finally, acrylic

acid 11³⁹demonstrates potent, selective RXR agonism, and

acrylic acid CD2915 (12)⁴⁰is also a known RXR agonist.

Figure 1. Docking results for compounds 8−19 relative to 1.

Calculated Autodock binding free energies were used to score the

ligands, with negative values indicating better binding than 1. Circles,

docking to 1MVC; diamonds, docking to 1H9U; triangles, docking to

3FUG; open symbols, using Babel charges; ﬁlled symbols, using

Autodock charges.

Because of the potent and selective RXR agonism reported

for compounds 8, 9, and 11 as well as a lack of reported

biological evaluation of 10, recent work in our laboratories has

focused on synthesizing seven novel analogues, 13−19,

modeled on compounds 8−12 and subsequently assessing

these novel compounds for RXR-selective activation alongside

1 and their parent compounds, 8−12.

RESULTS AND DISCUSSION

■

Molecular Modeling. Docking results for compounds 8−

19 relative to 1 are given in Figure 1. Compounds were scored

by the Autodock binding free energy estimate, with negative

values indicating better binding than 1. Docks were performed

with both the Autodock and OpenBabel charges; the latter tend

to show less overpolarization.³⁵All compounds were predicted

to bind nearly as well or better than 1, with predicted binding

aﬃnities generally within 0.5 kcal/mol of 1. Overlays of docking

poses in the L-shaped binding pocket showed that the phenyl

ring was held in nearly identical position for most ligands, while

small adjustments were observed in the position of the double

ring system. This is illustrated in Figure 2, which shows an

overlay of the top docking binding poses of 1 and 13 in the

binding pocket. The observation that chemical substitutions are

accommodated by small shifts of the fused ring system indicates

the presence of available space around the fused ring, which

Figure 2. Overlay of best poses for 1 (orange) and compound 13

(blue) in 1MVC using OpenBabel charges. Ile268, which makes

important hydrophobic contacts, and was treated ﬂexible in the

docking studies, is shown as ball-and-stick.

suggests that the fused ring might be a good target for future

chemical modiﬁcations.

8434

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Scheme 1

Scheme 2

8435

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Chemistry. The synthesis of pyrimidine-bexarotene ana-

logue 13 and the pyrimidine-9 analogue 14 follows a scaled

down methodology³⁸that was recently reported for the

synthesis of pyridine-bexarotene analogues 8 and 9 (Scheme

1). We also undertook the synthesis of bexarotene analogue 8

and 9 to have standards against which to test 13 and 14.

Commercially available dimethyl pyridine-2,5-dicarboxylate was

selectively saponiﬁed by the addition of solid sodium hydroxide

pellets to a suspension in methanol that was then reﬂuxed for 4

h followed by acidiﬁcation to give crude 5-(methoxycarbonyl)-

pyridine-2-carboxylic acid (20)³⁸in 74% after ﬁltration and

recrystallization from boiling water. Carboxylic acid 20 was

converted to acid chloride 21³⁸by treatment with thionyl

chloride, and acid chloride 21 was used in the aluminum

chloride catalyzed Friedel−Crafts acylation reaction with

compound 22³⁴to give ketone 23³⁸in 54% yield over two

steps. Ketone 23 was treated with a 3.0 M solution of methyl

magnesium chloride in THF at 0 °C, followed by quenching

with a 1 N solution of hydrochloric acid and extraction, and the

concentrated residue was reﬂuxed with para-toluene sulfonic

acid monohydrate to give alkene 24³⁸in 55% yield over three

steps. Alkene 24 was converted to the cyclopropyl methyl ester

25³⁸in 82% yield by adding a solution of 24 in THF to a

DMSO solution of trimethylsulfoxonium iodide treated with 22

wt % potassium tert-butoxide in THF. Both alkene 24 and

cyclopropyl methyl ester 25 were saponiﬁed by treatment with

aqueous potassium hydroxide and reﬂuxing in methanol to give,

after precipitation with 20% HCl, the corresponding carboxylic

acids 8 and 9, respectively, in near quantitative crude yields.

The carboxylic acids were puriﬁed by recrystallization from hot

ethyl acetate and hexanes.

Fortunately, compound 14 yielded single, transparent

crystals from ethyl acetate, suitable for X-ray diﬀraction, and

an X-ray diﬀraction study was performed (Figure 3).

Figure 3. X-ray crystal structure of 14. The thermal ellipsoid plot is

shown at the 50% probability level.

To synthesize a sample of compound 10, a modiﬁed

literature procedure³⁷was used. Toluene was reacted with

dihydro-2,2,5,5-tetramethylfuran-3(2H)-one (32) with alumi-

num chloride as a catalyst to provide 3,4-dihydro-1,1,4,4,7-

pentamethylnaphthalen-2(1H)-one (33)³⁷in 44% yield.

Ketone 33 was reacted with bromine and an aluminum

chloride catalyst to aﬀord 6-bromo-3,4-dihydro-1,1,4,4,7-

pentamethylnaphthalen-2(1H)-one (34). Ketone 34 was

condensed with para-toluenesulfonyl hydrazide to give

compound 35 in 50% yield. Compound 35 was treated with

methyl lithium to provide 6-bromo-1,4-dihydro-1,1,4,4,7-

pentamethylnaphthalene (36)³⁸in a Shapiro reaction⁴¹in

90% yield (Scheme 3).

Commercially available pyrimidine-2,5-dicarboxylic acid was

reﬂuxed in thionyl chloride with a few drops of catalytic DMF

for 4 h, and excess thionyl chloride was removed in vacuo to

give the solid diacid chloride that was dissolved in toluene and

added to a solution of triethyl amine in methanol to yield the

dimethyl pyrimidine-2,5-dicarboxylate 26 in 72% yield over two

steps. A suspension of dimethyl ester 26 in methanol was then

treated with sodium hydroxide pellets and reﬂuxed for 4 h,

followed by acidiﬁcation with 2N HCl, to give 5-(methox-

ycarbonyl) pyrimidine-2,5-dicarboxylate (27) in 89% yield after

ﬁltration. Acid 27 was converted to acid chloride 28 by

treatment with thionyl chloride, and acid chloride 28 was used

in the aluminum chloride catalyzed Friedel−Crafts acylation

reaction with compound 22 to give ketone 29 in 68% yield over

two steps. Ketone 29 was treated with a 3.0 M solution of

methyl magnesium chloride in THF at 0 °C followed by

quenching with a 1 N solution of hydrochloric acid and then

extracted, and the concentrated residue was reﬂuxed with para-

toluene sulfonic acid monohydrate to give alkene 30 in 40%

yield over three steps. Alkene 30 was converted to the

cyclopropyl methyl ester 31 in 68% yield by adding a solution

of 30 in THF to a DMSO solution of trimethylsulfoxonium

iodide treated with 22 wt % potassium tert-butoxide in THF.

Both alkene 30 and cyclopropyl methyl ester 31 were

saponiﬁed by treatment with aqueous potassium hydroxide

and reﬂuxing in methanol to give, after precipitation with 20%

HCl, the corresponding carboxylic acids 13 and 14,

respectively, in near quantitative crude yields. The carboxylic

acids were puriﬁed by recrystallization from hot ethyl acetate

and hexanes (Scheme 2).

With alkene 36 in hand, we proceeded to make the

appropriate coupling partner for it en route to compound 10

by reacting 4-cyanobenzoyl chloride (37) with N,O-hydroxyl-

amine hydrochloride to give the Weinreb amide (38)³⁸in 88%

yield (Scheme 4).

To complete the synthesis of compound 10, we treated

alkene 36 with butyl lithium followed by amide 38 to give

ketone 39³⁸after aqueous acidic workup in 82% overall yield.

The nitrile group of ketone 39 was hydrolyzed with potassium

hydroxide to give carboxylic acid 40³⁸after aqueous acidic

workup in 96% yield (Scheme 5).

Acid 40 was treated with methyl magnesium chloride,

followed by aqueous acidic workup to give alcohol 41³⁸that

was subsequently reﬂuxed with para-toluenesulfonic acid

monohydrate in toluene to give alkene 10 in 91% overall

yield (Scheme 5).

To synthesize compound 15, a strategy using the Suzuki−

Miyura coupling reaction⁴²as the key step was followed.

Compound 22 was treated with bromine in DCM to give

compound 42^16bin 60% yield. Bromide 42 was treated with

butyl lithium followed by triisopropylborate and aqueous acidic

workup to give boronic acid 43 in 45% yield over three steps

(Scheme 6).

To prepare the Suzuki coupling partner for boronic acid 43³⁸

en route to 15, 3-bromo-4-methylbenzoic acid (44) was treated

with borane·THF to give a benzyl alcohol that was

subsequently oxidized to aldehyde 45⁴³with manganese(IV)-

oxide in 74% overall yield. Aldehyde 45 was converted to the

alkene ester 46 by a Horner−Wadsworth−Emmons reaction.⁴⁴

With the alkene ester 46 in hand, we proceeded to couple it

8436

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Scheme 3

Scheme 4

Scheme 6

with boronic acid 43 via a Suzuki−Miyura reaction to give

biphenyl 47 in 80% yield. The ester group of biphenyl 47 was

saponiﬁed with KOH to give the carboxylic acid 15 in 75%

yield after aqueous acidic workup (Scheme 7).

Next, we targeted the synthesis of a triﬂuoro-methyl analogue

of 11, compound 16, because we wished to compare the

substitution of the hydroxyl group of 11 with a nonpolar

methyl group (15) versus a much more polar triﬂuoro-methyl

group (16). Hence, we performed the Horner−Wadsworth−

Emmons reaction on commercially available aldehyde 48 to

give ester 49 in 98% yield. With aromatic bromine 49 in hand,

we proceeded to couple it with boronic acid 43 via a Suzuki−

Miyura reaction to give biphenyl 50 in 71% yield. The ester

group of biphenyl 50 was saponiﬁed with KOH to give the

carboxylic acid 16 in 75% yield after aqueous acidic workup

(Scheme 8).

Next, we set out to synthesize an analogue of 15 with an

alkene in the aliphatic ring: compound 17. Thus, we converted

alkene 36 to boronic acid 51³⁸by treating 36 with butyl lithium

followed by triisopropyl borate and an aqueous acidic workup.

With boronic acid 51 in hand, we proceeded to couple it with

ester 46 via a Suzuki−Miyura reaction to give biphenyl 52 in

72% yield. The ester group of biphenyl 52 was saponiﬁed with

KOH to give the carboxylic acid 17 in 75% yield after aqueous

acidic workup (Scheme 9).

To compare the eﬀect of placing the alkene in the aliphatic

ring of 11, we prepared compound 18. First, we made the

appropriate aromatic halide coupling partner for boronic acid

51 by acetylating 3-bromo-4-hydroxybenzaldehyde (53) to give

aldehyde 54³⁹in 90% yield, and then we carried out the

Horner−Wadsworth−Emmons reaction on 54 to give ester

55³⁹in 84% yield. With aromatic bromide 55 in hand, we

proceeded to couple it with boronic acid 51 via a Suzuki−

Miyura reaction to give biphenyl 56 in 54% yield. The ester

group of biphenyl 56 was saponiﬁed with KOH to give the

carboxylic acid 18 in 89% yield after aqueous acidic workup

(Scheme 10).

Because of our recent ﬁnding³⁴that the addition of a ﬂuorine

atom ortho to the carboxylic acid group of 1 increases its

binding and activation of RXR, we prepared 19 to investigate

both the addition of an alkene in the aliphatic ring of 1 and

ﬂuorination ortho to the carboxylic acid. The synthesis of 19

begins with the sodium carbonate mediated conversion of

methyl-4-acetyl-2-ﬂuorobenzoate 57 to the enol triﬂate 58 in

27% yield. The enol triﬂate 58 was then coupled to 51 to give

Scheme 5

8437

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Scheme 7

human colon cancer (HCT-116) cells (Figure 4). This assay

tests for ligand binding to a recombinant, full length RXR

containing an activation domain. If the ligand−RXR complex

then homodimerizes with an RXR−Gal4 fusion protein, the

luciferase reporter gene will be transcribed, as the luciferase

gene is downstream of Gal4p DNA binding elements.

Scheme 8

We tested the indicated compounds at two diﬀerent

concentrations (25 and 100 nM) in this system. The 100 nM

assay (Figure 4B) revealed that ﬁve compounds (8, 9, 11, 13,

and 15) exhibited high activity compared with compound 1

(bexarotene). Moreover, the 25 nM assay of compounds shows

an even wider range of receptor binding and RXR

homodimerization ability because the 100 nM assay likely

approaches RXR binding saturation while the 25 nM assay

represents a subsaturating concentration of ligand (see also

EC₅₀values in Table 1). Compounds 8, 9, 11, 13, and 15

possess signiﬁcantly (using a one-way ANOVA, p < 0.01)

greater activity than compound 1 in the 25 nM assay (Figure

4A), with compound 18 also approaching statistical signiﬁ-

cance. Additionally, compounds 12, 17, and 19 display a

smaller degree of RXR transcriptional response (using a one-

way ANOVA, p < 0.01) relative to compound 1 in the 25 nM

assay, and compounds 12 and 17 also possess signiﬁcantly

lower binding than compound 1 in the 100 nM assay, but 12,

17, and 19 still have high activity compared to the ethanol

vehicle control (Figure 4A,B). These results imply that

compounds modeled after 1 can be synthesized successfully

and can mediate RXR transcriptional activation, and we are

interested in elucidating the factors responsible for eliciting the

diﬀerent response-ranges observed. We have also utilized the

mammalian two-hybrid assay in human HCT-116 cells to assess

dose−response relationships (see Supporting Information

Figures 1−3), and EC₅₀values as described below.

Novel Analogues of 1 Bind to RXR and Mediate

Transactivation in the Context of an RXRE. It is important

to note that the use of the mammalian two-hybrid assay as an

initial screen for agonist binding to and activation of RXR is

useful because of the speed, convenience, and sensitivity of the

assay. However, one potential drawback of the mammalian two-

hybrid assay is that the RXR−agonist complex may be

dependent on the correct biologically relevant DNA platform,

or retinoid X receptor responsive element (RXRE), that

speciﬁcally associates with the RXR homodimer in vivo. The

RXRE DNA sequence is usually found in the promoter region

and also up- or downstream of the genes controlled by the RXR

homodimer in response to the endogenous 9-cis RA ligand or

when RXR is bound to a synthetic rexinoid. It is possible that

Scheme 9

methyl ester 59 in 55% yield, and ester 59 was saponiﬁed with

KOH in methanol to give 19 in 70% yield (Scheme 11).

Finally, we prepared a sample of acrylic acid 12 to compare

against 11 and the other novel acrylic acid analogues (15−18)

in this study. Thus, 3-bromobenzaldehyde 60 was converted to

ethyl acrylate 61 in 98% yield. The ethyl acrylate 61 was

coupled to boronic acid 43 to give ethyl ester 62 in 70%, and

62 was subsequently saponiﬁed with KOH in methanol to yield

12 in 84% yield (Scheme 12).

Biological Assays and Rationale. A Mammalian Two-

Hybrid Assay to Assess RXR Transcriptional Response of

Analogues Compared to 1. Biological assessment of a subset

of analogues described above (compounds 8−19) was ﬁrst

carried out employing the mammalian two-hybrid assay in

8438

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Scheme 10

low expression of RXRα⁴⁷and other receptors of interest to

this study so we could quantitate most of the measurable eﬀects

as resulting from the expression of the transfected, rather than

endogenous, receptor. The expression levels (as assessed by

Western blotting) and ligand responsiveness of the HCT-116

cell line transfected with and without each of the receptors

evaluated in this study are shown in Supporting Information

Figure 4.

Scheme 11

We tested the indicated compounds at two diﬀerent

concentrations (25 and 100 nM) in the RXRE assay. The

results in Figure 5 reveal that compounds 8, 9, 10, 11, 13, 15,

and 18 displayed transcriptional activity consistently above the

compound 1 level in both assays (Figure 5A,B). In particular,

compounds 8, 11, 15, and 18 possess a signiﬁcantly greater

activity than compound 1 at both concentrations tested (using

a one-way ANOVA, p < 0.05 for compound 8, 15, and 18, and

p < 0.01 for compound 11). Importantly, the proﬁle of

analogue activity in the RXRE-based assays (Figure 5) is almost

identical to the proﬁle of analogues identiﬁed in the mammalian

two-hybrid assay (Figure 4), thus validating the mammalian

two-hybrid assay, and reinforcing the identiﬁcation and

reproducibility of those analogues with the greatest potency

in terms of RXR activation.

Scheme 12

Evaluation of EC₅₀Values and Quantitation of RAR

Agonist Activity. We utilized our mammalian two-hybrid assay

in human colon cancer cells (Figure 4) to evaluate a much

larger array of 1 and analogue concentrations. In addition to

employing a single ligand concentration as in Figure 4, dose−

response assays were carried out with ligand concentrations

ranging from 10⁻⁹M up to 0.3 × 10⁻⁵M of each compound as

well as 1 (Supporting Information Figures 1−3). Utilizing this

collection of dose−response experiments, we were able to

calculate EC₅₀values, which are listed in Table 1. Our EC₅₀

value for 1, which is an estimate of ligand aﬃnity for RXR, is

similar to values obtained previously by another group.^16b,17In

fact, Boehm and co-workers show that the diﬀerent isoforms of

RXR have similar EC₅₀values, within the same order of

magnitude, not only for 1, but several analogues of 1.¹⁷

Moreover, the data in Figures 4 and 5 suggest that compounds

12, 16, 17, and 19 possess consistently lower RXR-mediated

transcriptional activity while compounds 8, 9, 11, 13, 15, and

18 are consistently more active, observations that are entirely

consistent with the EC₅₀values in Table 1. We also performed

an analysis of the “residual” retinoic acid receptor (RAR)

agonist activity of 1, as well as for each analogue versus the

known RAR ligand (all-trans retinoic acid). The results of this

the RXRE may allosterically inﬂuence the aﬃnity and/or

selectivity of the RXR protein toward potential ligands. Thus, a

second screening protocol for our collection of possible RXR

agonists included transfection of HCT-116 cells with an

expression vector for wild-type human RXRα along with a

reporter construct that contains an authentic RXRE driving the

expression of the luciferase reporter gene. The most frequent

side eﬀect of bexarotene administration is an increase in plasma

triglycerides as a result of liver RXR/LXR-directed activation of

lipid metabolism.⁴⁵Thus, we chose to ﬁrst assess the ability of

our analogues to bind/activate RXRα because the liver

expresses predominantly the α isoform,⁴⁶and this isoform is

likely responsible for the elevated triglyceride synthesis in the

liver. We selected HCT-116 as a cell line that has a relatively

8439

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Figure 4. Identiﬁcation of RXR agonists via a mammalian two-hybrid screening assay in human colon cancer cells. HCT-116 human colon cancer

cells were cotransfected using a pCMV-hRXR binding domain vector (BD) as well as an hRXR-activation domain (AD) plasmid along with a pFR-

Luc reporter gene containing BD-binding sites and renilla control plasmid. Cells were transfected for 24 h utilizing a liposome-mediated transfection

protocol and then treated with ethanol vehicle and either 25 or 100 nM of the indicated compound. After a 24 h incubation, cells were lysed and a

dual luciferase assay was completed. Analogue-mediated RXR binding and homodimerization, as measured by luciferase output, was compared to the

RXR agonist parent compound 1 (value set to 1.0). Values are means SD *p < 0.05, **p < 0.01 vs compound 1.

assay, which employed expression of human RARα (the most

widely distributed isoform of RAR in human tissues) and a

retinoic acid responsive element (RARE)-luciferase reporter

system, revealed that compounds 8, 10, 13, 14, 15, 17, and 19

possess greater RAR agonist activity, compounds 9 and 18 are

approximately equal to 1 in its activation of RAR, and

compounds 11, 12, and 16 possess slightly lower RAR-directed

activity. Taken together, these results suggest that modiﬁcation

of 1 with more polar atoms in the aromatic ring that bears the

carboxylic acid may not only increase the analogues’ ability to

activate RAR (compounds 8, 13, 14, 19), but also increase its

ability to activate RXR (analogues 8 and 13).

Analysis of Apoptotic Activity by Bexarotene and Novel

Analogues in a CTCL System. One potential mechanism to

explain the eﬃcacy of 1 in treating CTCL eﬀectively is that this

rexinoid induces apoptosis and/or cytotoxicity in the T-

lymphocyte.⁴⁸Thus, we next tested CTCL (Hut-78) cells

treated with 1 and the indicated analogues for their ability to

induce classic apoptosis, using an assay for caspases 3 and 7,

two caspases that mediate apoptosis.⁴⁹We tested each

compound (8−19) by treatment for either 24 or 48 h (Figure

6). The results of the apoptosis assay suggest that most of the

analogues possess some pro-apoptotic activity. Two com-

pounds (analogues 8 and 13) that exhibit higher RXR binding

aﬃnity (lower EC₅₀, Table 1) and greater potency in the RXR

transcription assays (Figures 4 and 5) are also more active

(Figure 6) than compound 1 in the apoptotic assay (using a

one-way ANOVA, p < 0.05 for compound 8, p < 0.01 for

compound 13 both at 48 h, and p < 0.05 for compound 13 at

24 h). Several analogues reveal activity that is greater than the

ethanol vehicle control (Figure 6) in eliciting apoptosis at 24 h

(using a one-way ANOVA, p < 0.05 for compounds 1, 9, 10,

12, 14, and 18, p < 0.01 for compounds 8, 13, 16, 17, and 19)

or 48 h (using a one-way ANOVA, p < 0.05 for compounds 1,

9, 14, 17, 18, and 19, p < 0.01 for compounds 8, 13). A potent

known apoptotic inducer (sodium butyrate, NaBu) serves as a

positive control in this system.

Evaluation of 1 and Analogues As Inhibitors of VDR−RXR

Activity Due to Analogue-Mediated Formation of RXR−RXR

Homodimers. 1 binds to RXR and induces RXR−RXR

homodimerization, thus there may be less RXR available in

the intracellular pool for the creation of heterodimers with

other nuclear receptors such as the formation of VDR−RXR

heterodimers. This disruption of RXR heterodimer pathways is

one of several potential molecular mechanisms that may be

responsible for side eﬀects of 1 in humans. Therefore, we seek

to develop RXR analogues that are not only more robust in

their RXR biopotency via RXRE-mediated transcription but

also that are less disruptive to RXR heterodimerization

pathways. As a result, we measured vitamin D responsive

element (VDRE)-mediated transcriptional activity in the

presence of the active 1,25-dihydroxyvitamin D₃metabolite

(1,25D; 10⁻⁹M) to assess the level of RXR−RXR “diversion”

and thus likely inhibition of VDR−RXR-directed transcription

at two diﬀerent concentrations of RXR analogues (Figure 7A,

rexinoid 10⁻⁷M; Figure 7B, rexinoid 10⁻⁶M). Compound 1

inhibited VDR activity by approximately 50% in the 10⁻⁶

M

assay (Figure 7B) and by 30% in the 10⁻⁷M assay (Figure 7A).

Compounds 8, 10, 13, 14, and 17 inhibited VDR activity less

compared with compound 1 in both assays (Figure 7A,B).

Statistically, compounds 1, 9, 11, and 15 are signiﬁcantly lower

(using a one-way ANOVA, p < 0.05 for compounds 1, p < 0.01

for compounds 9, 11, and 15) than 1,25D alone in the 100 nM

assay (Figure 7A). Most compounds (except for 10) are

signiﬁcantly lower (using a one-way ANOVA, p < 0.05 for

compounds 17, p < 0.01 for compounds 1, 8, 9, 11, 12, 13, 14,

15, 16, 18, and 19) than 1,25D alone in the 1000 nM assay

(Figure 7B). These results suggest that the expected side eﬀects

of compound 8, 10, 13, 14, and 17 due to RXR−RXR diversion

may be smaller than compound 1. It is important to note that

some previous reports⁵⁰have demonstrated a stimulatory eﬀect

8440

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

a

Table 1. Determination of EC₅₀Values in RXR M2H Assay and Quantitation of RAR Agonist Activity

a

(1) EC₅₀values were determined from full dose−response experiments with each compound in the range of 10⁻⁹to 10⁻⁵M in transfected HCT-

116 cells using an RXR mammalian two-hybrid system. (2) RAR agonist activity was derived from an RAR/RARE reporter system in transfected

HCT-116 cells treated with all-trans RA or indicated analogue at 100 nM or 1 μM. The activity with analogue (or compound 1) divided by the

activity with all-trans RA expressed as a percentage represents the RAR agonist activity.

of 9-cis RA/rexinoids on 1,25D−VDR−VDRE-mediated tran-

scription. However, several other studies have revealed the

opposite eﬀect, i.e., suppression of VDR−RXR activity (see

Introduction). Therefore, this remains a controversial theme in

the vitamin D ﬁeld. To address this important question, we

conducted additional transcriptional assays with VDR−RXR

employing diﬀerent classes of vitamin D responsive elements

(VDREs) beyond the ER6 VDRE used in Figure 7A,B. These

experiments (Figure 7C−F) reveal that the eﬀect of rexinoids

(1/analogues) on 1,25D−VDR−RXR transcriptional response

8441

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Figure 5. Evaluation of RXR agonists via an RXRE-luciferase reporter-based assay in human colon cancer cells. HCT-116 cells were transfected with

human RXRα, RXRE-luciferase reporter gene, and renilla control plasmid. Cells were transfected for 24 h utilizing a liposome-mediated transfection

protocol as in Figure 4 and then treated with ethanol vehicle and 25 or 100 nM of the indicated compound. After a 24 h incubation, cells were lysed

and a luciferase assay was completed. Analogue-stimulated, RXR-mediated transcription, as measured by luciferase output, was compared to the RXR

agonist parent compound 1 (value set to 1.0). Values are means SD *p < 0.05, **p < 0.01 vs compound 1.

Figure 6. Assessment of bexarotene and analogues for apoptotic activity in CTCL cells. Human cutaneous T lymphocytes (Hut-78) were plated and

subsequently dosed with the indicated treatments. Cells were allowed to incubate for 24 or 48 h, followed by measurement of apoptosis with a

commercial kit (see Experimental Section). Sodium butyrate (NaBu), a known inducer of apoptosis in this system, was used as a positive control.

#

Values are means SD *p < 0.05, **p < 0.01 vs compound 1. p < 0.05, ^##p < 0.01 vs ethanol.

(e.g., repression, no eﬀect, stimulation) is dependent on the

nature of the VDRE. For example, as shown in Figure 7A,B,

there is a reproducible inhibition with 1 and most analogues on

VDR activity (as described above) when the VDR DNA

platform is an everted repeat-6 (ER6) from the CYP3A4 gene.

In contrast, there is a stimulation in VDR-mediated activity

when using the more classical direct repeat-3 (DR3) VDRE

(Figure 7C,D) in the presence of 1,25D and some of the

analogues compared to treatment with 1,25D alone. Com-

pounds 1 and 8 are signiﬁcantly higher (using a one-way

ANOVA, p < 0.05) than 1,25D alone, and compound 12 and

17 are signiﬁcantly lower (using a one-way ANOVA, p < 0.05)

than 1,25D alone in the 100 nM assay (Figure 7C).

Additionally, compounds 1, 8, and 15 are signiﬁcantly higher

(using a one-way ANOVA, p < 0.05 for compounds 8 and 15, p

< 0.01 for compounds 1) than 1,25D alone in the 1000 nM

assay (Figure 7D). The CYP24 natural promoter, which

contains 2 VDREs, has a net zero eﬀect of most rexinoids on

1,25D−VDR transactivation (Figure 7E,F) with only a few

analogues showing signiﬁcant inhibiton (using a one-way

ANOVA, p < 0.05 for compounds 18 and 19, p < 0.01 for

compounds 12 in the 100 nM assay, p < 0.05 for compounds

12 and 18 in the 1000 nM assay). Related ﬁndings have been

reported for a TR−RXRβ−TRE system where RXRβ can either

enhance, not alter, or inhibit TR-dependent transcription

depending on the nature of TRE.⁵¹These novel results with

VDR not only address the need to evaluate the eﬀect of 1/

rexinoids on 1,25D−VDR−RXR transcriptional activity in the

context of multiple VDREs, but they may also provide a

compelling resolution to the conﬂicting data that have been

reported in the literature with respect to the actions of

rexinoids on vitamin D signaling.

While much of the comparative discussion has focused on

how a given novel analogue (13−19) performs in these assays

8442

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

Figure 7. Evaluation of bexarotene and selected analogues for modulation of VDR−VDRE-mediated transcription via formation and/or diversion of

RXR homodimers. HCT-116 cells were transfected with hVDR, VDRE-luciferase reporter plasmids (proximal ER6 or distal DR3 from the human

CYP3A4 gene or the natural human CYP24A1 promoter-5.5kb), and a renilla control plasmid. Cells were transfected for 24 h utilizing a liposome-

mediated transfection protocol and then treated with ethanol vehicle, 1 nM 1,25D, and 100 or 1000 nM of the indicated compound in the presence

of 1 nM 1,25D. After a 24 h incubation, cells were lysed and a luciferase assay was completed. 1,25D-stimulated, VDRE-mediated transcription in the

presence of rexinoids (as measured by luciferase output) was compared to the 1,25D-only group (value set to 1.0). (A) ER6-mediated activity in the

presence of 1 nM 1,25D and 100 nM analogues. (B) ER6-mediated activity in the presence of 1 nM 1,25D and 1000 nM analogues. (C) DR3-

mediated activity in the presence of 1 nM 1,25D and 100 nM analogues. (D) DR3-mediated activity in the presence of 1 nM 1,25D and 1000 nM

analogues. (E) hCYP24A1 promoter activity in the presence of 1 nM 1,25D and 100 nM analogues. (F) hCYP24A1 promoter activity in the

presence of 1 nM 1,25D and 1000 nM analogues. Values are means SD *p < 0.05, **p < 0.01 vs 1 nM 1,25D only (black bar).

versus 1, it is also useful to examine how each novel analogue

performs relative to a known compound (8−12) possessing a

correspondingly more similar structure. Substituting a

pyrmidine ring in 13 and 14 versus a pyridine ring in 8 and

9 resulted in higher RXRα EC₅₀values, whereas the RAR

agonist activity was similar, as was the ability to modulate VDR

activity on ER6 and CYP24, and the ability of 13 and 14 to

stimulate apoptosis. On the other hand, the ﬂuorine

substitution in novel analogue 19 versus the hydrogen atom

in 10 resulted in similar RXRα EC₅₀values, whereas the RAR

agonist activity for 19 was much greater than 10, but the

apoptosis and ER6/CYP24 VDRE transcription levels were

similar for both 10 and 19. Substituting a methyl group, 15, or

triﬂuoromethyl group, 16, for the hydroxyl group of 11 led to

higher RXRα EC₅₀values, whereas the RAR agonist activity was

higher in 15 but nearly identical in 16. The apoptosis proﬁle

and activity on all three VDREs were similar for 11, 15, and 16.

Interestingly, introducing an unsaturation in the aliphatic ring

of 11 to give analogue 18 conferred similar RXRα EC₅₀values

and RAR agonist activities between 18 and 11 as well as similar

apoptosis and VDRE transcription levels. However, substituting

a methyl group for the hydroxyl group in 11 while

simultaneously introducing an unsaturation in the aliphatic

ring to give analogue 17 resulted in a higher RXRα EC₅₀value

as well as higher RAR agonist activities than observed in 11.

Bexarotene and the Novel Analogues Are Not Mutagenic.

Utilizing the EPA-approved Saccharomyces cerevisiae muta-

genesis assay,³⁴all compounds were tested for their ability to

damage DNA. DNA damage in this assay leads to either a point

mutation (reversion of a phenotype of isoleucine auxotroph to

isoleucine prototroph) or substantial damage leading to

subsequent repair. In this way, DNA mutagenesis in this strain

can be quantiﬁed by colony formation on selective media,

indicative of a phenotype change induced by DNA damage. All

compounds were tested at concentrations up to 0.15% w/v and

8443

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

ng of the renilla control plasmid and 2 μL/well of Express-IN was

again used for the liposome-mediated delivery. The cells were

incubated for 24 h post-transfection and then treated with media

containing ethanol vehicle (0.1%), or analogues (ﬁnal concentration of

100 nM or 1 μm). After a 24 h incubation period, the amount of

retinoid activity was measured using the same luciferase assay

described above. Three independent assays were conducted with

triplicate samples for each treatment group.

EC₅₀Determination. EC₅₀values were determined from full

dose−response curves ranging from 1 × 10⁻⁹to 0.3 × 10⁻⁵M in

transfected HCT-116 cells using an RXR mammalian two-hybrid

system. HCT-116 cells were plated overnight at 80000 cells/well in a

24-well plate and maintained as described above. The cells were

cotransfected using a human RXR binding domain (BD) vector, a

human RXR activation domain (AD) vector, a luciferase reporter gene

containing BD-binding sites, and renilla control plasmid. Transfection

was achieved via 2 μL/well of Express-IN transfection reagent which

was allowed to incubate for 24 h with the cells. Then, the cells were

treated with ethanol vehicle (0.1%) or analogues (1.0, 2.5, 5.0, 7.5, 10,

25, 50, 75, 100, 250, 500 nM, 1, 2, 3 μM) and incubated for 24 h. The

amount of rexinoid activity at each concentration was measured using

the same luciferase assay described above, and EC₅₀values were

derived from dose−response curves of ligand concentration versus

normalized luciferase activity

RARE-Mediated Transcription Assay. The RARE assay was

completed using HCT-116 cells plated at 80000 cells/well in a 24-well

plate and maintained as described above. The cells were cotransfected

using 250 ng of pTK-DR5(x2)-luciferase reporter gene. This RARE is

an optimized element that has been described previously⁵⁴and is

responsive to the RAR ligand, all-trans retinoic acid. Also included in

the transfection were 50 ng of pCMX-human RARα and 20 ng of the

renilla control plasmid, along with 2 μL/well of Express-IN for

liposome-mediated delivery of the DNA into the cells. The cells were

incubated for 24 h post-transfection and then treated with ethanol

(0.1%) or analogues (ﬁnal concentration of 100 nM or 1 μM). After a

24 h incubation period, the amount of retinoid activity was measured

using the same luciferase assay described above. The activity with

analogue (or compound 1) divided by the activity with all-trans RA

(expressed as a percentage) represents the RAR agonist activity. Three

independent assays were conducted with triplicate samples for each

treatment group.

none increased colony formation above vehicle control (data

not shown).

CONCLUSIONS

■

1 is an FDA approved drug to treat cutaneous T-cell lymphoma

and is being explored to treat a myriad of additional types of

cancer. Moreover, recent data suggest that 1 is also a promising

compound in the continued ﬁght against AD.^27,52Our work

was motivated by the untoward side eﬀects of 1 because the

drug appears to be safe and eﬃcacious in treatment of some

CTCL patients, but the side eﬀects include hyperlipidemia and

hypothyroidism due to the ability of 1 to serve as an RXR

ligand with little heterodimer-speciﬁc selectivity. The potential

beneﬁt of exploring additional RXR ligands for therapeutic

treatments is further motivated by the opportunities of rational

drug design, and as we have shown,^34,35modeling can help

drive synthesis of additional potent RXR ligands. Our

continued work has produced seven novel compounds, four

of which possess EC₅₀values lower than 1 and two of which

possess superior pro-apoptotic proﬁles in CTCL cells. We are

thus hopeful that an RXR ligand with a lower EC₅₀will translate

into a pharmaceutical that would allow for a lower eﬀective

dose regime. Furthermore, side eﬀects of hyperlipidemia and

hypothyroidism are an ongoing consideration in treatment with

1. We address concerns about cross-talk by RXR ligands with

other RXR nuclear receptor pathways, as well as the potential

for new rexinoids to titrate away RXR from nonpermissive

heteropartners by assessing our novel compounds for their

ability to modulate VDR−RXR-directed transcription utilizing

multiple VDREs. We demonstrate that many of our novel

analogues are superior at mitigating RXR heterodimerization

disruption compared to 1. Taken together, our novel RXR

ligands appear more potent, demonstrate higher eﬃcacy, and

are less likely to lead to side eﬀects than the parent compound.

Work is ongoing to analyze further biological parameters of

these compounds and develop additional novel analogues of 1

with improved side eﬀect proﬁles.

Apoptosis Assay. Apoptotic activity was assessed by the Caspase-

Glo 3/7 assay (Promega, Madison, WI) according to the

manufacturer’s instructions. The Caspase-Glo 3/7 assay is based on

the cleavage of the DEVD sequence of a luminogenic substrate by

caspases 3 and 7, which results in a luminescent signal. HuT-78

(human T-cell lymphoma) cells were distributed (1 × 10⁴cells/well)

in white-walled 96-well microplates (Corning, NY) in 100 μL of

medium and incubated with 500 μM sodium butyrate (NaBu), 100

nM 1, or rexinoid analogues for 24 and 48 h. The Caspase-Glo 3/7

reagent was then added to each well and incubated for an additional 1

h at room temperature. The luminescence was measured in a

luminometer (Saﬁre2, Tecan, US). NaBu, a known inducer of

apoptosis in HuT-78, was used as a positive control. Each treatment

group was dosed in triplicate, and at least two independent

experiments were performed. Numbers were standardized to sodium

butyrate (set at 100%).

VDR−VDRE Assay. The VDR−VDRE assay was completed using

HCT-116 cells plated at 80000 cells/well in a 24-well plate and

maintained as described above. The cells were cotransfected using 250

ng of the indicated VDRE-luciferase reporter gene (proximal ER6 or

distal DR3 VDRE from the human CYP3A4 gene, or the natural

human CYP24A1 promoter-5.5kb), 50 ng of pSG5-hVDR, and 20 ng

of the renilla control plasmid with 2 μL/well of Express-IN used for

liposome-mediated DNA delivery. The cells were incubated for 24 h

post-transfection and then further treated with ethanol vehicle (0.1%),

10⁻⁹M 1,25-dihydroxyvitamin D₃(1,25D) alone, or 10⁻⁹M 1,25D in

combination with analogues (ﬁnal concentration of 10⁻⁶M or 10⁻⁷

M). After a 24 h incubation period, the amount of luciferase activity

was measured using the luminescence assay described above. Three

EXPERIMENTAL SECTION

■

Mammalian Two-Hybrid Assay. HCT-116 colorectal carcinoma

cells were plated overnight at 80000 cells/well in a 24-well plate and

maintained in DMEM/high glucose (Hyclone) enhanced with 10%

FBS (Invitrogen), 1 mM sodium pyruvate (Invitrogen), 100 μg/mL

streptomycin, and 100 unit/mL penicillin. The cells were cotransfected

using a human RXR binding domain (BD) vector, a human RXR

activation domain (AD) vector, a luciferase reporter gene containing

BD-binding sites, and renilla control plasmid. A liposome-mediated

transfection was completed according to the manufacturer’s protocol

using 2 μL/well of Express-IN transfection reagent (Thermo Fisher

Scientiﬁc, Lafayette, CO) and allowed to incubate for 24 h. The cells

were then treated with media containing ethanol vehicle (0.1%),

bexarotene or analogues at a ﬁnal concentration of 25 nM or 100 nM

and incubated for 24 h. The amount of rexinoid activity was measured

by luciferase output utilizing a dual-luciferase reporter assay system

according to the manufacturer’s protocol (Promega, Madison, WI) in a

Sirus luminometer (Berthold Detection System, Pforzheim, Germany).

Three independent assays were conducted with triplicate samples for

each treatment group.

RXRE-Mediated Transcription Assay. The RXRE assays were

completed using HCT-116 cells plated at 80000 cells/well in a 24-well

plate and maintained as described above. The cells were cotransfected

using 250 ng of RXRE-luciferase reporter gene (RXRE from the

naturally occurring responsive element in the rat cellular retinol

binding protein II gene:⁵³AAAATGAACTGTGACCTGT-

GACCTGTGACCTGTGAC), 50 ng of pSG5-human RXRα, and 20

8444

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

independent assays were conducted with triplicate samples for each

treatment group. Modulation of vitamin D activity was determined by

evaluation of 1,25D-treated cells compared to cells treated with 1,25D

plus bexarotene/analogue.

Mutagenicity Assays. Mutagenicity assays utilized the D7 strain

of Saccharomyces cerevisiae.⁵⁵D7 cells were incubated at 30 °C with

shaking for 3 h with the compound solubilized in DMSO in increasing

concentrations up to 0.15% w/v, and then aliquots were plated on

selective media as described in Marshall⁵⁶and allowed to incubate for

4 days. Colony number was then scored and compared to DMSO

alone (negative) and ethidium bromide (positive) controls.

Docking Studies. Docking studies on the human RXRα (Protein

Data Bank codes 1MVC⁵⁷and 3FUG,³⁹respectively) and RXRβ

(Protein Data Bank code 1H9U⁵⁸) were performed using the

Autodock 4.2 software.⁵⁹As in our previous studies,³⁴protein residues

Arg316 and Ile268 (residue numbering as in 1MVC) were treated

ﬂexibly and both Autodock and OpenBabel 2.3.0 charges⁶⁰were used.

Docking used the Lamarckian genetic algorithm with 25 000 000

energy evaluations per dock and a total of 250 docks per compound.

Calculated binding free energies were used to score the compounds.

Instrumentation. A 400 MHz Bruker spectrometer was used to

acquire ¹H NMR and ¹³C NMR spectra. Chemical shifts (δ) are listed

in ppm against residual nondeuterated solvent peaks in a given

deuterated solvent (e.g., CHCl₃in CDCl₃) as an internal reference.

Coupling constants (J) are reported in Hz, and the abbreviations for

splitting include: s, single; d, doublet; t, triplet; q, quartet; p, pentet; m,

multiplet; br, broad. All ¹³C NMR spectra were acquired on a Bruker

instrument at 100.6 MHz. Chemical shifts (δ) are listed in ppm against

deuterated solvent carbon peaks as an internal reference. High

resolution mass spectra were recorded using either a JEOL GCmate

(2004), a JEOL LCmate (2002) high resolution mass spectrometer, or

an ABI Mariner (1999) ESI-TOF mass spectrometer.

remove residual thionyl chloride. The acid chloride 21 was dried on

high vacuum to remove residual benzene. To a two-neck, 50 mL

round-bottom ﬂask equipped with a reﬂux condenser and magnetic

stir-bar was added 22 (1.40 g, 6.91 mmol) followed by a solution of

crude acid chloride 21 (6.62 mmol) in DCM (15 mL). Aluminum

chloride (2.20 g, 16.5 mmol) was added to the reaction solution at

room temperature slowly, with stirring, and the reaction solution

turned from colorless to red accompanied by the evolution of gas and

heat. The reaction was stirred for 5 min then heated to reﬂux for 15

min. The reaction was judged to be complete by TLC, and the

solution was poured into an ice solution (25 mL) acidiﬁed with a 20%

HCl solution (8 mL), and ethyl acetate was added (13 mL). The

aqueous and organic layers were separated, and the aqueous layer was

extracted with ethyl acetate (15 mL, twice). The combined organics

were washed with water and brine, dried over sodium sulfate, ﬁltered,

and concentrated to give crude 23. Crude 23 was puriﬁed by column

chromatography (250 mL SiO₂, hexanes:ethyl acetate 95:5 to 85:15)

to give 23 (1.30 g, 54%) as an oﬀ-white, crystalline solid, mp 122−125

1

°C. H NMR (400 MHz, CDCl₃) δ 9.26 (dd, J = 2.0, 0.8, 1H), 8.49

(dd, J = 8.0, 2.0, 1H), 8.10 (dd, J = 8.0, 0.8, 1H), 7.42 (s, 1H), 7.21 (s,

1H), 3.99 (s, 3H), 2.39 (s, 3H), 1.68 (s, 4H), 1.29 (s, 6H), 1.20 (s,

6H). ¹³C NMR (100.6 MHz, CDCl₃) δ 195.8, 165.1, 158.6, 149.9,

149.2, 141.6, 138.1, 135.7, 133.2, 130.3, 129.7, 127.5, 123.8, 52.7, 34.8,

34.8, 34.4, 33.8, 31.6, 31.5, 20.5. IR (neat) 2919, 1720, 1688 cm⁻¹. LC-

MS (M + H)⁺calcd for C₂₃H₂₈NO₃366.2069, found 366.2070.

Methyl 6-[(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaph-

thalen-2-yl)ethenyl]nicotinate (24). The procedure of Faul and

co-workers was followed.³⁸To a 100 mL round-bottom ﬂask charged

with 23 (1.0077 g, 2.757 mmol) was added toluene (10 mL), and the

solution was cooled in a salt-water ice bath to −15 °C with stirring,

under nitrogen. To this solution was added a 22 wt % solution of

MeMgCl in THF (1.20 mL, 3.60 mmol), and the reaction was stirred

for 15 min at −15 °C and then warmed to room temperature and

stirred for 35 min before quenching with 1N HCl (7 mL, 7 mmol).

The reaction mixture was then separated, and the aqueous layer was

extracted with ethyl acetate. The combined organic layers were ﬁltered

and a small yellow ﬁlter-cake was dissolved in chloroform and added to

the organic ﬁltrate, and the combined organic ﬁltrate was concentrated

in vacuo to give a crude alcohol intermediate that was used without

further puriﬁcation. To this crude intermediate in a 100 mL round-

bottom ﬂask was added p-toluenesulfonic acid monohydrate (0.5247 g,

2.76 mmol) and toluene (40 mL), and the reaction was reﬂuxed for 3

h into a Dean−Stark apparatus half-ﬁlled with toluene (6 mL). After

the reaction had cooled to room temperature, it was added to a

solution of sodium carbonate (0.78 g) in water (15 mL), shaken

vigorously, and the layers were separated. The aqueous layer was

extracted with ethyl acetate, and the combined organic layers were

washed with brine, dried over sodium sulfate, ﬁltered, and

concentrated in vacuo to give a crude brown product that was

puriﬁed by column chromatography (150 mL SiO₂, hexanes:ethyl

acetate 97.5:2.5 to 95:5) to give 24 (0.5595 g, 55%) as a white, ﬁber-

like solid, mp 166−167 °C. ¹H NMR (400 MHz, CDCl₃) δ 9.22 (dd, J

= 2.0, 0.8, 1H), 8.15 (dd, J = 8.0, 2.0, 1H), 7.14 (s, 1H), 7.11 (s, 1H),

7.02 (dd, J = 8.4, 0.8, 1H), 6.54 (d, J = 2.0, 1H), 5.51 (d, J = 2.0, 1H),

3.94 (s, 3H), 1.98 (s, 3H), 1.69 (s, 4H), 1.31 (s, 6H), 1.26 (s, 6H). ¹³C

NMR (100.6 MHz, CDCl₃) δ 165.8, 161.1, 150.6, 148.0, 144.5, 142.5,

137.6, 136.7, 132.7, 128.0, 124.1, 121.1, 121.0, 52.2, 35.1, 35.0, 33.9,

33.8, 31.9, 31.8, 19.8. IR (neat) 2959, 1722, 1591 cm⁻¹. LC-MS (M +

H)⁺calcd for C₂₄H₃₀NO₂364.2277, found 364.2272.

General Procedures. Removal of volatile solvents transpired

under reduced pressure using a Buchi rotary evaporator and is referred

̈

to as removing solvents in vacuo. Thin layer chromatography was

conducted on precoated (0.25 mm thickness) silica gel plates with

60F-254 indicator (Merck). Column chromatography was conducted

using 230−400 mesh silica gel (E. Merck reagent silica gel 60). All

tested compounds were analyzed for purity by combustion analysis

through Columbia Analytical Services (formerly Desert Analytics in

Tucson, AZ) and were found to be >95% pure.

5-(Methoxycarbonyl)pyridine-2-carboxylic Acid (20). The

protocol of Faul et al. was followed.³⁸To a suspension of dimethyl

pydrine-2,5-dicarboxylate (10.09 g, 51.7 mmol) in methanol (130 mL)

was added sodium hydroxide pellets (2.20 g, 55.0 mmol), and the

heterogeneous reaction was stirred and reﬂuxed for 4 h. After cooling

to 65 °C, 2.0 N HCl (38 mL, 76 mmol) was slowly added, and over

the course of addition, the solid dissolved and a precipitate formed.

The reaction solution was allowed to cool to room temperature, and

then it was cooled in an ice-bath and ﬁltered. The ﬁlter-cake was

washed with water to give an oﬀ-white product (7.71 g, 82%). A

sample of this crude, ﬁltered product (1.09 g) was dissolved in boiling

water (70 mL), and the solution was slowly cooled to room

temperature and then cooled in an ice-bath before it was ﬁltered to

give a white, crystalline powder (0.98 g) at a 74% overall yield, mp

189−190 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 9.15 (dd, J = 2.0, 0.8,

1H), 8.44 (dd, J = 8.0, 2.0, 1H), 8.15 (dd, J = 8.0, 0.8, 1H), 3.91 (s,

3H), 3.34 (br s, 1H). ¹³C NMR (100.6 MHz, DMSO-d₆) δ 165.5,

164.6, 151.7, 149.8, 138.3, 127.9, 124.6. IR (neat) ν 3440, 2969, 1717,

1694 cm⁻¹. LC-FAB-MS (M + H)⁺calcd for C₈H₈NO₄182.0453,

found 182.0458.

Methyl 6-[(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaph-

thalen-2-yl)carbonyl]nicotinate (23). The procedure of Faul and

co-workers was followed.³⁸Methyl 6-(chlorocarbonyl)-pyridine-3-

carboxylate (21) was synthesized by reﬂuxing 5-(methoxycarbonyl)-

pyridine-2-carboxylic acid (20) (1.20 g, 6.62 mmol) in thionyl chloride

(12.0 mL, 165 mmol) in a 100 mL one-neck round-bottom ﬂask ﬁtted

with a water-cooled reﬂux condenser. Excess thionyl chloride was

removed in vacuo to give crude 21 as an oﬀ-white solid, and this solid

was dissolved in dry benzene (ca. 20 mL) and evaporated to dryness to

Methyl 6-[(3,5,8,8-Pentamethyl-5.6,7,8-tetrahydronaphtha-

len-2-yl)cyclopropyl]nicotinate (25). The procedure of Faul and

co-workers was followed.³⁸To a suspension of trimethylsulfoxonium

iodide (0.365 g, 1.66 mmol) in DMSO (1.2 mL) was added a 20 wt %

solution of potassium tert-butoxide in THF (0.94 mL, 1.67 mmol). A

solution of 24 (0.40 g, 1.10 mmol) in THF (4.8 mL) was added

dropwise over 20 min at 30−34 °C with stirring. The reaction was

stirred for 60 min at 35 °C, then cooled to room temperature and

quenched with 1N HCl (5 mL). The layers were separated, and the

aqueous layer was extracted with ethyl acetate. The combined organic

layers were washed with brine, dried over sodium sulfate, and

8445

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

concentrated to give an oﬀ-white crude solid that was puriﬁed by

column chromatography (150 mL SiO₂, hexanes:ethyl acetate 97.5:2.5

to 90:10) to give 25 (0.3411 g, 82%) as a white crystalline solid, mp

5-(Methoxycarbonyl)pyrimidine-2-carboxylic Acid (27). To a

suspension of of dimethyl pyrimidine-2,5-dicarboxylate (26) (3.34 g,

17.0 mmol) in methanol (30 mL) was added sodium hydroxide pellets

(0.748 g, 18.7 mmol), and the heterogeneous reaction was stirred and

reﬂuxed for 4 h. After cooling to 65 °C, 2.0 N HCl (14 mL, 28 mmol)

was slowly added. The reaction solution was allowed to cool to room

temperature, and then it was concentrated in vacuo to give a solid. The

solid was ﬁltered with cold water, and the ﬁlter-cake was washed with a

small amount of cold water to give 27 as a white crystalline solid (mp

1

168−170 °C. H NMR (400 MHz, CDCl₃) δ 9.08 (dd, J = 2.4, 0.8,

1H), 7.98 (dd, J = 8.4, 2.4, 1H), 7.27 (s, 1H), 7.11 (s, 1H), 6.74 (dd, J

= 8.4, 0.8, 1H), 3.90 (s, 3H), 2.11 (s, 3H), 1.82−1.83 (m, 2H), 1.69 (s,

4H), 1.35−1.36 (m, 2H), 1.30 (s, 6H), 1.27 (s, 6H). ¹³C NMR (100.6

MHz, CDCl₃) δ 169.2, 166.1, 150.4, 143.8, 142.6, 137.1, 136.5, 135.7,

129.2, 128.3, 122.1, 120.7, 52.0, 35.1, 34.0, 33.9, 31.9, 31.8, 30.3, 20.2,

19.2. IR (neat) 2958, 1714, 1595 cm⁻¹. LC-MS (M + H)⁺calcd for

C₂₅H₃₂NO₂378.2433, found 378.2422.

1

152−154 °C) (2.765 g, 89%). H NMR (400 MHz, DMSO-d₆) δ

13.89 (br s, 1H), 9.39 (s, 2H), 3.97 (s, 3H). ¹³C NMR (100.6 MHz,

DMSO-d₆) δ 164.3, 163.4, 159.6, 158.4, 124.8, 52.9. IR (neat) 3493,

2970, 1716, 1584, 1558 cm⁻¹. GC-MS (M)⁺calcd for C₇H₆N₂O₄

182.0328, found 182.0335.

Methyl 6-(1-(1,2,3,4-Tetrahydro-1,1,4,4,6-pentamethylnaph-

thalen-7-yl)vinyl)pyridine-3-carboxylate (8).³⁶To a suspension

of 24 (0.2898 g, 0.797 mmol) in methanol (5.0 mL) was added a

solution of KOH (0.116 g) in water (0.18 mL), and the reaction was

reﬂuxed at 85 °C for 1 h. The reaction solution was cooled to room

temperature and quenched with 20% HCl (26 mL). The crude

precipitate was ﬁltered and dried to give a crude white product (0.2753

g, 98%) that was recrystallized from hexanes:ethyl acetate 4:1 to give

the pure 8 as a white crystalline solid, mp 243−245 °C. ¹H NMR (400

MHz, CDCl₃) δ 13.51 (br s, 1H), 9.47 (s, 1H), 8.82 (dd, J = 8.4, 2.0,

1H), 7.41 (d, J = 8.4, 1H), 7.23 (s, 1H), 7.15 (s, 1H), 7.10 (s, 1H),

6.08 (s, 1H), 2.01 (s, 3H), 1.68 (s, 4H), 1.29 (s, 6H), 1.24 (s, 6H). ¹³C

NMR (100.6 MHz, CDCl₃) δ 161.8, 154.3, 146.4, 146.3, 143.6, 139.9,

133.0, 132.5, 130.0, 128.8, 128.6, 128.1, 125.6, 34.8, 34.1, 33.9, 31.8,

31.7, 19.7. IR 2948, 1706, 1590 cm⁻¹. LC-MS (M + H)⁺calcd for

C₂₃H₂₈NO₂350.2120, found 350.2111. Anal. Calcd for C₂₄H₂₉NO₂·

HCl: C, 71.58; H, 7.31; N, 3.63. Found: C, 72.64; H, 7.35; N, 3.62.

6-[1-(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaphthalen-

2-yl)cyclopropyl]nicotinic Acid (9).³⁶To a suspension of 25

(0.3043 g, 0.806 mmol) in methanol (8.0 mL) was added a solution of

KOH (0.1287 g) in water (0.18 mL), and the reaction was reﬂuxed at

85 °C for 1 h. The reaction solution was cooled to room temperature

and quenched with 20% HCl (27 mL). The crude precipitate was

ﬁltered and dried to give a crude white product (0.2543 g, 86%) that

was recrystallized from hexanes:ethyl acetate 4:1 to give pure 9 as a

white crystalline solid (0.147 g, 50%), mp 273−274 °C. ¹H NMR (400

MHz, CDCl₃) δ 11.69 (br s, 1H), 9.18 (dd, J = 2.0, 0.8, 1H), 8.05 (dd,

J = 8.4, 2.0, 1H), 7.27 (s, 1H), 7.12 (s, 1H), 6.79 (dd, J = 8.4, 0.8, 1H),

2.13 (s, 3H), 1.86−1.87 (m, 2H), 1.69 (s, 4H), 1.39−1.40 (m, 2H),

1.31 (s, 6H), 1.27 (s, 6H). ¹³C NMR (100.6 MHz, CDCl₃) δ 170.8,

170.2, 151.0, 143.9, 142.7, 137.2, 136.9, 135.7, 129.2, 128.3, 121.4,

120.9, 35.1, 34.0, 33.9, 31.9, 31.8, 30.5, 20.5, 19.2. IR (neat) 2952,

1686, 1593 cm⁻¹. LC-MS (M + H)⁺calcd for C₂₄H₃₀NO₂364.2277,

found 364.2265. Anal. Calcd for C₂₄H₂₉NO₂: C, 79.30; H, 8.04; N,

3.85. Found: C, 78.76; H, 7.93; N, 3.76.

Methyl 2-[(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaph-

thalen-2-yl)carbonyl]pyrimidine-5-carboxylate (29). Methyl 2-

(chlorocarbonyl)-pyrimidine-5-carboxylate (28) was synthesized by

reﬂuxing 5-(methoxycarbonyl)pyrimidine-2-carboxylic acid (27) (1.22

g, 6.70 mmol) in thionyl chloride (12.0 mL, 165 mmol) in a 100 mL

one-neck round-bottom ﬂask ﬁtted with a water-cooled reﬂux

condenser. Excess thionyl chloride was removed in vacuo to give

crude 28 as an oﬀ-white solid, and this solid was dissolved in dry

benzene (ca. 20 mL) and evaporated to dryness to remove residual

thionyl chloride. The acid chloride 28 was dried on high vacuum to

remove residual benzene. To a two-neck, 50 mL round-bottom ﬂask

equipped with a reﬂux condenser and magnetic stir-bar was added 22

(1.45 g, 7.16 mmol) followed by a solution of crude acid chloride 28

(6.70 mmol) in DCM (15 mL). Aluminum chloride (2.20 g, 16.5

mmol) was added to the reaction solution at room temperature slowly,

with stirring, and the reaction solution turned from colorless to red

accompanied by the evolution of gas and heat. The reaction was stirred

for 5 min then heated to reﬂux for 15 min. The reaction was judged to

be complete by TLC, and the solution was poured into an ice solution

(25 mL), acidiﬁed with a 20% HCl solution (8 mL), and ethyl acetate

added (13 mL). The aqueous and organic layers were separated, and

the aqueous layer was extracted with ethyl acetate (15 mL, twice). The

combined organics were washed with water and brine, dried over

sodium sulfate, ﬁltered, and concentrated to give crude 29. Crude 29

was puriﬁed by column chromatography (250 mL SiO₂, hexanes:ethyl

acetate 95:5 to 85:15) to give 29 (1.677 g, 68%) as an oﬀ-white,

1

crystalline solid, mp 94−96 °C. H NMR (400 MHz, CDCl₃) δ 9.42

(s, 2H), 7.40 (s, 1H), 7.21 (s, 1H), 4.03 (s, 3H), 2.43 (s, 3H), 1.66 (s,

4H), 1.28 (s, 6H), 1.17 (s, 6H). ¹³C NMR (100.6 MHz, CDCl₃) δ

192.8, 165.9, 163.5, 158.4, 150.3, 142.0, 136.7, 131.9, 130.8, 130.0,

124.0, 52.9, 34.7, 34.6, 34.4, 33.8, 31.6, 31.4, 20.9. IR (neat) 2951,

1715, 1683, 1577, 1549 cm⁻¹. LC-MS (M + H)⁺calcd for C₂₂H₂₇N₂O₃

367.2022, found 367.2017.

Dimethyl Pyrimidine-2,5-dicarboxylate (26). To a 100 mL

round-bottom ﬂask charged with pyrimidine-2,5-dicarboxylic acid

(5.067 g, 30.1 mmol) was slowly added thionyl chloride (21 mL, 290

mmol) and 3 drops of DMF, and the reaction was reﬂuxed in an oil

bath at 85 °C for 3 h. Excess thionyl chloride was removed in vacuo

and benzene (20 mL) was added to the crude solid and the benzene

was removed in vacuo to give crude pyrimidine-2,5-dicarboyl

dichloride that was used without further puriﬁcation. To the crude

pyrimidine-2,5-dicarboyl dichloride was added toluene (13.6 mL), and

this solution was added dropwise to a solution of triethylamine (16.5

mL, 118 mmol) in methanol (86 mL). After stirring for 1 h at room

temperature, the reaction was quenched with 1 N HCl (120 mL). The

reaction was poured into ethyl acetate (80 mL), the layers were

separated, and the aqueous layer was extracted with ethyl acetate. The

combined organic layers were washed with saturated NaHCO₃(75

mL) and brine, dried over sodium sulfate, and removed in vacuo to

give crude 26. Crude 26 was puriﬁed by column chromatography (250

mL SiO₂, hexanes:ethyl acetate 2:3) to give 26 (4.26 g, 72%) as a

Methyl 2-[(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaph-

thalen-2-yl)ethenyl]pyrimidine-5-carboxylate (30). To a 100

mL round-bottom ﬂask charged with 29 (1.0373 g, 2.83 mmol) was

added toluene (10 mL), and the solution was cooled in a salt-water ice

bath to −15 °C with stirring, under nitrogen. To this solution was

added a 22 wt % solution of MeMgCl in THF (1.20 mL, 3.60 mmol),

and the reaction was stirred for 15 min at −15 °C and then warmed to

room temperature and stirred for 35 min before quenching with 1N

HCl (7 mL, 7 mmol). The reaction mixture was then separated, and

the aqueous layer was extracted with ethyl acetate. The combined

organic layers were ﬁltered and a small yellow ﬁlter-cake was dissolved

in chloroform and added to the organic ﬁltrate and the combined

organic ﬁltrate and combined organic ﬁltrate was concentrated in

vacuo to give a crude alcohol intermediate that was used without

further puriﬁcation. To this crude intermediate in a 100 mL round-

bottom ﬂask was added p-toluenesulfonic acid monohydrate (0.5247 g,

2.76 mmol) and toluene (40 mL), and the reaction was reﬂuxed for 3

h into a Dean−Stark apparatus half-ﬁlled with toluene (6 mL). After

the reaction had cooled to room temperature, it was added to a

solution of sodium carbonate (0.78 g) in water (15 mL), shaken

vigorously, and the layers separated. The aqueous layer was extracted

with ethyl acetate, and the combined organic layers were washed with

brine, dried over sodium sulfate, ﬁltered, and concentrated in vacuo to

1

colorless, crystalline solid (mp 140−142 °C). H NMR (400 MHz,

CDCl₃) δ 9.40 (s, 2H), 4.07 (s, 3H), 3.99 (s, 3H). ¹³C NMR (100.6

MHz, CDCl₃) δ 163.0, 162.9, 158.8, 158.3, 125.4, 53.8, 53.0. IR (neat)

2969, 1709, 1579. 1551 cm⁻¹. GC-MS (M)⁺calcd for C₈H₈N₂O₄

196.0484, found 196.0485.

8446

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

give a crude brown product that was puriﬁed by column

chromatography (150 mL SiO₂, hexanes:ethyl acetate 97.5:2.5 to

95:5) to give 30 (0.4127 g, 40%) as a white, ﬁber-like solid, mp 189−

for 30 min, then warmed to room temperature and stirred for 2 h, and

then heated to 50−55 °C with stirring for 15 min. The reaction was

cooled to room temperature and then poured onto ice water (50 mL)

and extracted with ethyl acetate. The combined organic layers were

washed with brine, dried over sodium sulfate, ﬁltered, and

concentrated in vacuo. The resulting crude solid was puriﬁed by

column chromatography (150 mL SiO₂, hexanes:ethyl acetate 98:2) to

1

191 °C. H NMR (400 MHz, CDCl₃) δ 9.25 (s, 2H), 7.17 (s, 1H),

7.11 (s, 1H), 6.83 (d, J = 2.0, 1H), 5.80 (d, J = 2.0, 1H), 3.97 (s, 3H),

1.98 (s, 3H), 1.69 (s, 4H), 1.31 (s, 6H), 1.28 (s, 6H). ¹³C NMR (100.6

MHz, CDCl₃) δ 168.6, 164.4, 158.2, 148.3, 144.4, 142.1, 136.3, 132.7,

128.0, 127.9, 126.7, 121.1, 52.5, 35.1, 34.0, 33.8, 31.9, 31.8, 20.0. IR

(neat) 2959, 1717, 1582 cm⁻¹. LC-MS (M + H)⁺calcd for

C₂₃H₂₉N₂O₂365.2229, found 365.2232.

1

give 33 (1.6348 g, 44%) as a white crystalline solid, 49−51 °C. H

NMR (400 MHz, CDCl₃) δ 7.26 (d, J = 8.0, 1H), 7.13 (s, 1H), 7.05

(d, J = 8.0, 1H), 2.63 (s, 2H), 2.35 (s, 3H), 1.44 (s, 6H), 1.31 (s, 6H).

¹³C NMR (100.6 MHz, CDCl₃) δ 214.5, 142.9, 140.6, 136.3, 127.7,

Methyl 2-[(3,5,8,8-Pentamethyl-5.6,7,8-tetrahydronaphtha-

len-2-yl)cyclopropyl]pyrimidine-5-carboxylate (31). To a sus-

pension of trimethylsulfoxonium iodide (0.365 g, 1.66 mmol) in

DMSO (1.2 mL) was added a 20 wt % solution of potassium tert-

butoxide in THF (0.94 mL, 1.67 mmol). A solution of 30 (0.4062 g,

1.11 mmol) in THF (4.8 mL) was added dropwise over 20 min at 30−

34 °C with stirring. The reaction was stirred for 60 min at 35 °C, then

cooled to room temperature and quenched with 1N HCl (5 mL). The

layers were separated, and the aqueous layer was extracted with ethyl

acetate. The combined organic layers were washed with brine, dried

over sodium sulfate, and concentrated to give an oﬀ-white crude solid

that was puriﬁed by column chromatography (150 mL SiO₂,

hexanes:ethyl acetate 97.5:2.5 to 90:10) to give 31 (0.2884 g, 68%)

127.4, 124.3, 51.5, 47.9, 37.6, 30.6, 28.4, 21.0. IR (neat) 2968, 1705

cm⁻¹. GC-MS (M)⁺calcd for C₁₅H₂₀O 216.1514, found 216.1523.

6-Bromo-3,4-dihydro-1,1,4,4,7-pentamethylnaphthalen-

2(1H)-one (34). A DCM (22.0 mL) solution of 33 (3.193 g, 14.76

mmol) at 0 °C was stirred during the addition of aluminum chloride

(3.993 g, 29.95 mmol), and the mixture was stirred at 0 °C for 5 min.

To this solution was added a solution of bromine (0.90 mL, 17.47

mmol) in DCM (11.0 mL) dropwise, and the reaction solution was

stirred for 30 min at 0 °C. The reaction solution was poured onto ice

(50 mL) and extracted with ethyl acetate. The combined organic layers

were washed with brine, dried over sodium sulfate, and concentrated

in vacuo to give a crude product that was puriﬁed by column

chromatography (150 mL SiO₂, hexanes:ethyl acetate 98:2) to give 34

(4.36 g, 100%) as a white crystalline solid, mp 114−117 °C. ¹H NMR

(400 MHz, CDCl₃) δ 7.50 (s, 1H), 7.15 (s, 1H), 2.60 (s, 2H), 2.38 (s,

3H), 1.42 (s, 6H), 1.28 (s, 6H). ¹³C NMR (100.6 MHz, CDCl₃) δ

213.6, 143.2, 142.4, 136.3, 129.5, 128.3, 123.1, 51.2, 47.6, 37.5, 30.4,

28.3, 22.6. IR (neat) 2951, 1706 cm⁻¹. GC-MS (M)⁺calcd for

C₁₅H₁₉BrO 294.0619, found 294.0617.

(E)-1-(6-Bromo-3,4-dihydro-1,1,4,4,7-pentamethylnaphtha-

len-2(1H)-ylidene)-2-tosylhydrazine (35). To a suspension of 34

(3.0664 g, 10.4 mmol) and p-toluenesulfonylhydrazide (2.2158 g, 11.9

mmol) in methanol (61 mL) was added p-toluenesulfonic acid

monohydrate (0.4977 g, 2.616 mmol), and the reaction solution was

stirred and reﬂuxed under nitrogen for 24 h. The reaction was then

cooled to room temperature and then it was stirred in a salt-water ice-

bath at −15 °C for 1 h and the resulting precipitate was ﬁltered and

rinsed with cold methanol to aﬀord 35 as a white crystalline solid (2.41

g, 50%), mp 181−184 °C. ¹H NMR (400 MHz, CDCl₃) δ 8.03 (br s,

1H), 7.84 (d, J = 8.0, 2H), 7.34 (s, 1H), 7.28 (d, J = 8.4, 2H), 7.12 (s,

1H), 2.41 (s, 2H), 2.38 (s, 3H), 2.33 (s, 3H), 1.37 (s, 6H), 1.08 (s,

6H). ¹³C NMR (100.6 MHz, CDCl₃) δ 163.2, 144.0, 143.1, 142.7,

136.0, 134.9, 129.3, 128.3, 128.0, 122.7, 42.8, 36.8, 36.2, 29.9, 29.89,

22.5, 21.5. IR (neat) 3251, 2971 cm⁻¹. LC-MS (M + H)⁺calcd for

C₂₂H₂₈N₂O₂SBr 463.1055, found 463.1048.

1

as a white crystalline solid, mp 202−203 °C. H NMR (400 MHz,

CDCl₃) δ 9.07 (s, 2H), 7.23 (s, 1H), 7.09 (s, 1H), 3.92 (s, 3H), 2.12

(s, 3H), 1.87−1.88 (m, 2H), 1.67 (s, 4H), 1.46−1.47 (m, 2H), 1.29 (s,

6H), 1.27 (s, 6H). ¹³C NMR (100.6 MHz, CDCl₃) δ 176.8, 164.7,

157.7, 143.2, 141.8, 136.9, 135.9, 128.6, 127.8, 119.9, 52.3, 35.2, 34.0,

33.9, 31.9, 31.8, 31.7, 21.6, 19.4. IR (neat) 2951, 1719, 1588 cm⁻¹. LC-

MS (M + H)⁺calcd for C₂₄H₃₁N₂O₂379.2386, found 379.2385.

Methyl 2-(1-(1,2,3,4-Tetrahydro-1,1,4,4,6-pentamethylnaph-

thalen-7-yl)vinyl)pyridine-5-carboxylate (13). To a suspension of

30 (0.828 g, 2.27 mmol) in methanol (15.0 mL) was added a solution

of KOH (0.3627 g) in water (0.54 mL), and the reaction was reﬂuxed

at 85 °C for 1 h. The reaction solution was cooled to room

temperature and quenched with 20% HCl (60 mL). The crude

precipitate was ﬁltered and dried to give a crude white product (0.7741

g, 97%) that was recrystallized from hexanes:ethyl acetate 4:1 to give

the pure 13 as a white crystalline solid (0.3869 g, 48%), mp 254−255

1

°C. H NMR (400 MHz, CDCl₃) δ 10.11 (br s, 1H), 9.31 (s, 2H),

7.17 (s, 1H), 7.12 (s, 1H), 6.86 (d, J = 1.6, 1H), 5.86 (d, J = 1.6, 1H),

1.99 (s, 3H), 1.68 (s, 4H), 1.28 (s, 6H), 1.27 (s, 6H). ¹³C NMR (100.6

MHz, CDCl₃) δ 169.0, 168.2, 158.8, 148.0, 144.5, 142.2, 136.0, 132.7,

128.0, 127.9, 127.3, 120.7, 35.1, 34.0, 33.8, 31.9, 31.8, 20.0. IR (neat)

2963, 1682, 1580 cm⁻¹. LC-MS (M + H)⁺calcd for C₂₂H₂₇N₂O₂

351.2073, found 351.2082. Anal. Calcd for C₂₂H₂₆N₂O₂: C, 75.40; H,

7.48; N, 7.99. Found: C, 75.35; H, 7.54; N, 7.91.

2-[1-(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahydronaphthalen-

2-yl)cyclopropyl]pyrimidine-2-carboxylic Acid (14). To a

suspension of 31 (0.287 g, 0.758 mmol) in methanol (5.0 mL) was

added a solution of KOH (0.126 g) in water (0.18 mL), and the

reaction was reﬂuxed at 85 °C for 1 h. The reaction solution was

cooled to room temperature and quenched with 20% HCl (32 mL).

The crude precipitate was ﬁltered and dried to give a crude white

product (0.2568 g, 92%) that was recrystallized from hexanes:ethyl

acetate 4:1 to give the pure 14 as a white crystalline solid (0.2433 g,

88%), mp 266−267 °C. ¹H NMR (400 MHz, CDCl₃) δ 9.18 (s, 2H),

7.23 (s, 1H), 7.09 (s, 1H), 2.13 (s, 3H), 1.90−1.91 (m, 2H), 1.66 (s,

4H), 1.50−1.51 (m, 2H), 1.27 (s, 6H), 1.26 (s, 6H). ¹³C NMR (100.6

MHz, CDCl₃) δ 177.6, 168.6, 158.3, 143.3, 141.9, 136.6, 135.9, 128.6,

127.9, 119.2, 35.2, 34.0, 33.9, 32.0, 31.9, 31.8, 22.1, 19.4. IR (neat)

2954, 1679, 1586 cm⁻¹. LC-MS (M + H)⁺calcd for C₂₃H₂₉N₂O₂

365.2229, found 365.2225. Anal. Calcd for C₂₃H₂₈N₂O₂: C, 75.79; H,

7.74; N, 7.69. Found: C, 75.67; H, 7.95; N, 7.32.

6-Bromo-1,1,4,4,7-pentamethyl-1,4-dihydronaphthalene

(36).³⁸To a suspension of 35 (1.0127 g, 2.19 mmol) in MTBE (20

mL) was added a methyl lithium LiBr complex solution (1.5 M) in

ether (4.40 mL, 6.60 mmol) at room temperature with stirring under

nitrogen. The solution turned yellow with the evolution of gas

(presumably nitrogen), and a ﬁne oﬀ-white precipitate formed. The

heterogeneous solution was stirred for 1 h, cooled to 0 °C, and then

quenched with water (25.0 mL). The reaction was extracted with ethyl

acetate, and the organic layers were combined, washed with brine,

dried over sodium sulfate, ﬁltered, and concentrated in vacuo to give a

crude oﬀ-white solid that was puriﬁed by column chromatography

(150 mL SiO₂, hexanes) to give 36 (0.5463 g, 90%) as a white solid

(mp 109−111 °C). ¹H NMR (400 MHz, CDCl₃) δ 7.50 (s, 1H), 7.21

(s, 1H), 5.49 (s, 2H), 2.38 (s, 3H), 1.32 (s, 12H). ¹³C NMR (100.6

MHz, CDCl₃) δ 142.3, 141.8, 135.0, 132.7, 132.6, 129.8, 128.5, 122.4,

35.0, 34.9, 32.5, 32.4, 22.6. IR (neat) 3013, 2958, 1483, 1455, 1079,

889 cm⁻¹. GC-MS (M)⁺calcd for C₁₅H₁₉Br 278.0670, found

278.0655.

(4-Cyanophenyl)-N-methoxy-N-methylformamide (38).³⁸

The method of Faul and co-workers was used.³⁸To a suspension of

N,O-dimethylhydroxyamine hydrochloride (7.07 g, 72.5 mmol) and

K₂CO₃(10.0 g, 72.5 mmol) in ACN (100 mL) and water (50 mL) was

added 4-cyanobenzoyl chloride (37) (8.00 g, 48.3 mmol), and the

reaction was stirred for 2 h at room temperature. The reaction solution

3,4-Dihydro-1,1,4,4,7-pentamethylnaphthalen-2(1H)-one

(33).³⁷A modiﬁed procedure of Boehm and co-workers was used.³⁷

To a 100 mL round-bottom ﬂask charged with dihydro-2,2,5,5-

tetramethylfuran-3(2H)-one (32) (2.4 g, 17 mmol) and toluene (10.0

mL, 94 mmol) at 0 °C was added aluminum chloride (4.55 g, 34

mmol) in one portion with stirring. The reaction was stirred at 0 °C

8447

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

was poured into water (50 mL) and extracted with ethyl acetate. The

combined organic layers were washed with brine, dried over sodium

sulfate, and concentrated in vacuo to give a crude solid that was

puriﬁed by column chromatography (250 mL SiO₂, hexanes:ethyl

acetate 45:55 to 1:1) to give 38 (8.12 g, 88%) as a white crystalline

solid, mp 78−79 °C. ¹H NMR (400 MHz, CDCl₃) δ 7.76 (d, J = 8.4,

2H), 7.69 (d, J = 8.4, 2H), 3.50 (s, 3H), 3.36 (s, 3H). ¹³C NMR (100.6

MHz, CDCl₃) δ 167.8, 138.2, 131.8, 128.7, 118.1, 114.0, 61.2, 33.0. IR

(neat) 2947, 1627 cm⁻¹. GC-MS (M)⁺calcd for C₁₀H₁₀N₂O₂

190.0742, found 190.0739.

column chromatography (25 mL SiO₂, ethyl acetate) to give 10

(0.3628 g, 91%) as a white crystalline solid (mp 210−213 °C). H

1

NMR (400 MHz, CDCl₃) δ 11.98 (br s, 1H), 8.05 (d, J = 8.8, 2H),

7.40 (d, J = 8.4, 2H), 7.20 (s, 1H), 7.15 (s, 1H), 5.86 (d, J = 1.2, 1H),

5.54 (s, 2H), 5.38 (d, J = 1.2, 1H), 1.99 (s, 3H), 1.37 (s, 6H), 1.34 (s,

6H). ¹³C NMR (100.6 MHz, CDCl₃) δ 172.0, 149.0, 146.3, 142.0,

140.1, 138.2, 133.1, 133.1, 133.0, 130.3, 128.0, 127.7, 127.7, 126.6,

117.3, 35.0, 34.9, 32.7, 32.6, 20.0. IR (neat) 2953, 1688, 1606 cm⁻¹.

LC-MS (M + H)⁺calcd for C₂₄H₂₇O₂347.2011, found 347.1997. Anal.

Calcd for C₂₄H₂₆O₂: C, 83.20; H, 7.56. Found: C, 82.73; H, 7.54.

6-Bromo-1,2,3,4-tetrahydro-1,1,4,4,7-pentamethylnaphtha-

lene (42).^16bThe method of Dawson and co-workers was used to

synthesize 42.^16bTo a solution of 22 (1.32 g, 6.52 mmol) in

chloroform (6.0 mL) was added bromine (0.5 mL, 9.71 mmol) at

room temperature, and the reaction was stirred for 30 min and then

diluted with ethyl acetate. The solution was poured into an aqueous

saturated solution of Na₂SO₃, and the biphasic mixture was shaken, the

layers were separated, and the aqueous layer was extracted with ethyl

acetate. The combined organic layers were washed with brine, dried

over sodium sulfate, and concentrated in vacuo to give an oil that was

puriﬁed by column chromatography (150 mL SiO₂, ethyl acetate) to

give an inseparable 2:1 mixture of 42:22 (1.679 g, 60% yield for 42) as

a white crystalline solid. ¹H NMR (400 MHz, CDCl₃) δ 7.42 (s, 1H),

7.14 (s, 1H), 2.34 (s, 3H), 1.66 (s, 4H), 1.27 (s, 12H). GC-MS (M)⁺

calcd for C₁₅H₂₁Br 280.0827, found 280.0812.

4-[(3,5,5,8,8-Pentamethyl-2-5,8-dihydronaphthyl)carbonyl]-

benzenecarbonitrile (39).³⁸The method of Faul and co-workers

was followed.³⁸To a solution of 35 (2.00 g, 7.16 mmol) in THF (25

mL) at −78 °C under nitrogen was added a 1.6 M solution of n-BuLi

in hexanes (5.40 mL, 8.60 mmol) over 10 min, and the solution was

stirred for 20 min at −78 °C. This reaction solution was transferred via

airtight syringe to a solution of 38 (1.23 g, 6.47 mmol) in THF (10

mL) at −78 °C, and the combined mixture was stirred for 15 min at

−78 °C and then warmed to room temperature before 1.0 N HCl (75

mL) was added to quench the reaction. The solution was poured into

ethyl acetate, the layers were separated, and the aqueous layer was

extracted with ethyl acetate. The combined organic layers were washed

with brine, dried over sodium sulfate, and concentrated in vacuo to

give a crude product that was puriﬁed by column chromatography

(150 mL SiO₂, hexanes:ethyl acetate 95:5) to give 39 (1.7865 g, 82%)

1

5,6,7,8-Tetrahydro-3,5,5,8,8-pentamethylnaphthalen-2-yl-2-

boronic Acid (43).³⁸The method of Faul and co-workers was used.³⁸

To a 100 mL round-bottom ﬂask containing THF (30 mL) was added

a 1.6 M solution of n-BuLi in hexanes (8.0 mL, 12.8 mmol), and the

resulting solution was cooled in a dry ice acetone bath to −78 °C with

stirring, under nitrogen. To this solution was added a solution of the

2:1 mixture of 42:22 (3.3587 g, 7.88 mmol) in THF (8 mL) over 20

min and the reaction was stirred at −78 °C for 10 min and a mixture of

triisopropylborate (4.9 mL, 21.3 mmol) in THF (10 mL) was added

dropwise over 20 min. The reaction was stirred at −78 °C for 1 h and

then warmed to room temperature and stirred for 2 h. The reaction

was then quenched with 3 N HCl (35 mL), and after stirring for 2 h, it

was poured into ethyl acetate, the layers were separated, and the

aqueous layer was extracted with ethyl acetate. The combined organic

layers were washed with brine, dried over sodium sulfate, and

concentrated in vacuo to give a crude product that was puriﬁed by

column chromatography (150 mL SiO₂, ethyl acetate:hexanes 1:3) to

give 43 (0.8838 g, 45%) as a white crystalline solid, mp 220−225 °C.

¹H NMR (400 MHz, CDCl₃) δ 8.29 (s, 1H), 7.21 (s, 1H), 2.82 (s,

as a white crystalline solid, mp 117−119 °C. H NMR (400 MHz,

CDCl₃) δ 7.91 (d, J = 8.4, 2H), 7.75 (d, J = 8.4, 2H), 5.52 (s, 2H),

2.37 (s, 3H), 1.37 (s, 6H), 1.26 (s, 6H). ¹³C NMR (100.6 MHz,

CDCl₃) δ 196.6, 146.2, 141.6, 139.8, 135.0, 134.4, 132.6, 132.6, 129.2,

128.0, 118.0, 115.9, 35.3, 34.8, 32.4, 32.3, 20.0. IR (neat) 2960, 1669

cm⁻¹. GC-MS (M)⁺calcd for C₂₃H₂₃NO 329.1780, found 329.1788.

4-[(3,5,5,8,8-Pentamethyl-2-5,8-dihydronaphthyl)carbonyl]-

benzoic Acid (40).³⁸The method of Faul and co-workers was

followed.³⁸To a heterogeneous solution of 39 (1.62 g, 4.92 mmol) in

2-methoxyethanol (20 mL) was added a solution of KOH (1.64 g, 24.5

mmol) in water (10 mL). The reaction was heated in an oil bath at

reﬂux temperature and stirred under nitrogen for 16 h. The reaction

was allowed to cool to room temperature before it was quenched with

1 N HCl (50 mL). The solution was poured into ethyl acetate, the

layers were separated, and the aqueous layer was extracted with ethyl

acetate. The combined organic layers were washed with brine, dried

over sodium sulfate, and concentrated in vacuo with additional toluene

to azeotrope oﬀ 2-methoxyethanol to give 40 (1.65 g, 96%) as a white

powder, mp 199−201 °C. ¹H NMR (400 MHz, CDCl₃) δ 13.36 (br s,

1H), 8.08 (d, J = 8.4, 2H), 7.80 (d, J = 8.4, 2H), 7.42 (s, 1H), 7.34 (s,

1H), 5.54 (s, 2H), 2.25 (s, 3H), 1.33 (s, 6H), 1.22 (s, 6H). ¹³C NMR

(100.6 MHz, CDCl₃) δ 197.0, 166.6, 145.1, 140.7, 139.3, 135.2, 134.5,

133.9, 132.6, 132.5, 129.7, 129.6, 128.8, 127.1, 34.9, 34.5, 32.0, 19.4. IR

(neat) 2970, 1745 cm⁻¹. LC-MS (M + H)⁺calcd for C₂₃H₂₅O₃

349.1804, found 349.1805.

3H), 1.72 (s, 4H), 1.34 (s, 6H), 1.33 (s, 6H). ¹³C NMR (100.6 MHz,

CDCl₃) δ 149.4, 142.8, 141.4, 136.3, 128.5, 35.1, 35.0, 34.3, 33.8, 31.8,

31.5, 22.6. IR (neat) 2958, 1602, 1335 cm⁻¹.

3-Bromo-4-methylbenzaldehyde (45).⁴³The method of Adams

and co-workers was followed.⁴³To a solution of 3-bromo-4-

methylbenzoic acid (44) (5.08 g, 23 mmol) in THF (50 mL) stirring

under nitrogen at 0 °C was added a 1 M borane·THF solution (34.6

mL, 34.6 mmol) dropwise. The reaction was warmed to room

temperature, stirred for 18 h, then cooled to 0 °C and quenched by the

slow addition of water (10 mL). The reaction was warmed to room

temperature, and the solvents were removed in vacuo. The crude

product was dissolved in ethyl acetate, washed with a 1 M aqueous

sodium carbonate solution then brine, and the organic layer was dried

over sodium sulfate and concentrated in vacuo to give a benzyl alcohol

product (4.75 g, 100%) that was used without further puriﬁcation. ¹H

NMR (400 MHz, CDCl₃) δ 7.52 (d, J = 0.8, 1H), 7.20 (d, J = 8.0, 1H),

7.17 (dd, J = 7.6, 1.6, 1H), 4.61 (s, 2H), 2.38 (s, 3H), 2.05 (br s, 1H).

¹³C NMR (100.6 MHz, CDCl₃) δ 140.1, 137.0, 130.8, 130.7, 125.8,

124.9, 64.2, 22.5. To a solution of the benzyl alcohol intermediate (4.5

g, 22.4 mmol) in chloroform (100 mL) was added manganese dioxide

(15 g, 172 mmol). The reaction was reﬂuxed with stirring in an oil

bath at 70 °C for 18 h. Then it was ﬁltered through Celite, and

solvents were removed in vacuo to give a crude product that was

puriﬁed by column chromatography (150 mL SiO₂, ethyl acetate:hex-

anes 1:9) to give 45 (3.2974 g, 74%) as a white crystalline solid, mp

4-[1-(3,5,5,8,8-Pentamethyl-2-5,8-dihydronaphthyl)vinyl]-

benzoic Acid (10).³⁸The procedure of Faul and co-workers was

followed.³⁸To a 100 mL round-bottom ﬂask charged with a 3.0 M

solution of MeMgCl (1.53 mL, 4.60 mmol) was added THF (3 mL),

and the solution was cooled to −10 °C in a salt−water ice bath with

stirring under nitrogen. To this solution was added a solution of 40

(0.40 g, 1.15 mmol) in THF (4 mL), dropwise, and the reaction was

stirred at 0 °C for 4 h. The reaction was quenched with 1.0 N HCl (15

mL), the solution was extracted with ethyl acetate, and the combined

organic layers were washed with brine, dried over sodium sulfate,

ﬁltered, and concentrated in vacuo to give an intermediate alcohol that

was used without further puriﬁcation. The intermediate alcohol was

dissolved in toluene (30 mL), and to this solution was added p-

toluenesulfonic acid monohydrate (0.02 g, 0.116 mmol) and the

solution was reﬂuxed into a Dean−Stark apparatus preﬁlled with

toluene. After the solution was reﬂuxed for 2 h, it was cooled to room

temperature and poured into ethyl acetate and water. The aqueous

layer was extracted with ethyl acetate, and the combined organic layers

were washed with brine, dried over sodium sulfate, ﬁltered, and

concentrated in vacuo to give a crude product that was puriﬁed by

1

45−51 °C. H NMR (400 MHz, CDCl₃) δ 9.91 (s, 1H), 8.02 (d, J =

8448

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

1.6, 1H), 7.70 (dd, J = 7.6, 1.6, 1H), 7.39 (d, J = 8.0, 1H), 2.47 (s, 3H).

¹³C NMR (100.6 MHz, CDCl₃) δ 190.4, 145.1, 135.8, 133.4, 131.3,

128.3, 125.6, 23.4. IR (neat) 2980, 1682, 1598 cm⁻¹. GC-MS (M)⁺

calcd for C₈H₇OBr 197.9680, found 197.9665.

(E)-Ethyl 3-(3-Bromo-4-(triﬂuoromethyl)phenyl)acrylate

(49). To a solution of a 60% dispersion of NaH in mineral oil (0.31

g, 7.75 mmol) in DME (2 mL) at −30 °C was added a solution of

ethyl 2-phosphonoacetate (1.46 mL, 7.29 mmol) in DME (13 mL),

and the mixture was stirred at this temperature for 30 min. To this

solution was added a solution of 3-bromo-4-(triﬂuoromethyl)-

benzaldehyde (48) (1.68 g, 6.63 mmol) in DME (3 mL), and the

reaction was stirred at −30 °C for 1.5 h and then poured into water

(50 mL) and extracted with ethyl acetate. The combined organic layers

were washed with an aqueous saturated NH₄Cl solution and then

brine, dried over sodium sulfate, ﬁltered, and concentrated in vacuo to

give a crude product that was puriﬁed by column chromatography

(150 mL SiO₂, ethyl acetate:hexanes 5:95) to give 49 (2.1188 g, 98%)

(E)-Ethyl 3-(3-Bromo-4-methylphenyl)acrylate (46). To a

solution of a 60% dispersion of NaH in mineral oil (0.29 g, 7.25

mmol) in DME (2 mL) at −30 °C was added a solution of ethyl 2-

phosphonoacetate (1.46 mL, 7.29 mmol) in DME (13 mL), and the

mixture was stirred at this temperature for 30 min. To this solution

was added a solution of 45 (1.32 g, 6.63 mmol) in DME (3 mL), and

the reaction was stirred at −30 °C for 1.5 h and then poured into

water (50 mL) and extracted with ethyl acetate. The combined organic

layers were washed with an aqueous saturated NH₄Cl solution and

then brine, dried over sodium sulfate, ﬁltered, and concentrated in

vacuo to give a crude product that was puriﬁed by column

chromatography (150 mL SiO₂, ethyl acetate:hexanes 1:9) to give

1

as a colorless crystalline solid, mp 75−76 °C. H NMR (400 MHz,

CDCl₃) δ 7.84 (s, 1H), 7.69 (d, J = 8.0, 1H), 7.59 (d, J = 16.0, 1H),

7.52 (d, J = 8.0, 1H), 6.50, (d, J = 16.0, 1H), 4.27 (q, J = 7.2, 2H), 1.34

(t, J = 7.2, 3H). ¹³C NMR (100.6 MHz, CDCl₃) δ 165.9, 141.0, 139.2,

133.8, 131.4, 130.7, 130.4, 128.3, 128.2, 128.1, 128.1, 126.4, 123.9,

122.2, 121.2, 120.6, 120.5, 60.9, 14.2. IR (neat) 2923, 1711, 1638, 1603

cm⁻¹. GC-MS (M)⁺calcd for C₁₂H₁₀F₃O₂Br 321.9816, found

321.9805.

1

46 (1.576 g, 88%) as a colorless crystalline solid, mp 46−47 °C. H

NMR (400 MHz, CDCl₃) δ 7.69 (d, J = 1.6, 1H), 7.57 (d, J = 16.0,

1H), 7.35 (dd, J = 8.0, 1.6, 1H), 7.23 (d, J = 7.6, 1H), 6.38, (d, J =

16.0, 1H), 4.25 (q, J = 7.2, 2H), 2.41 (s, 3H), 1.33 (t, J = 7.2, 3H). ¹³C

NMR (100.6 MHz, CDCl₃) δ 166.7, 142.8, 140.1, 133.9, 131.6, 131.1,

126.8, 125.3, 118.6, 60.5, 22.9, 14.2. IR (neat) 2984, 1718, 1698, 1634

cm⁻¹. LC-MS (M + H)⁺calcd for C₁₂H₁₄O₂Br 269.0177, found

269.0171.

(E)-Ethyl 3-(4-(Triﬂuoromethyl)-3-(1,2,3,4-tetrahydro-

1,1,4,4,6-pentamethylnaphthalen-7-yl)phenyl)acrylate (50).

To a 50 mL Schlenk ﬂask charged with bromide 49 (1.03 g, 3.20

mmol), boronic acid 43 (0.8040 g, 3.27 mmol), TBAB (1.04 g),

Na₂CO₃(1.02 g, 9.62 mmol), and water (7.4 mL) was added

Pd(OAc)₂(0.0406 g, 0.18 mmol), and the ﬂask was evacuated and

backﬁlled with nitrogen three times. The reaction was stirred at room

temperature for 15 min and then placed in an oil bath preheated to

150 °C and stirred for 5 min. The reaction was allowed to cool to

room temperature, and the black residue was taken up in ethyl acetate

and water. The layers were separated, and the aqueous layer was

extracted with ethyl acetate. The combined organic layers were washed

with brine, dried over sodium sulfate, ﬁltered, and concentrated in

vacuo to give a crude product that was puriﬁed by column

chromatography (150 mL SiO₂, ethyl acetate:hexanes 1:9) to give

(E)-Ethyl 3-(3-(1,2,3,4-Tetrahydro-1,1,4,4,6-pentamethyl-

naphthalen-7-yl)-4-methylphenyl)acrylate (47). To a 50 mL

Schlenk ﬂask charged with bromide 46 (0.4317 g, 1.60 mmol), boronic

acid 43 (0.4010 g, 1.63 mmol), TBAB (0.52 g), Na₂CO₃(0.51 g, 4.81

mmol), and water (3.7 mL), was added Pd(OAc)₂(0.0203 g, 0.09

mmol), and the ﬂask was evacuated and backﬁlled with nitrogen three

times. The reaction was stirred at room temperature for 15 min and

then placed in an oil bath preheated to 150 °C and stirred for 5 min.

The reaction was allowed to cool to room temperature, and the black

residue was taken up in ethyl acetate and water. The layers were

separated, and the aqueous layer was extracted with ethyl acetate. The

combined organic layers were washed with brine, dried over sodium

sulfate, ﬁltered, and concentrated in vacuo to give a crude product that

was puriﬁed by column chromatography (150 mL SiO₂, ethyl

acetate:hexanes 1:9) to give 47 (0.5032 g, 80%) as a colorless oil.

¹H NMR (400 MHz, CDCl₃) δ 7.68 (d, J = 16.0, 1H), 7.41 (dd, J =

8.0, 2.0, 1H), 7.33 (d, J = 1.6, 1H), 7.27 (d, J = 8.0, 1H), 7.16 (s, 1H),

7.00 (s, 1H), 6.39, (d, J = 16.0, 1H), 4.25 (q, J = 7.2, 2H), 2.09 (s, 3H),

2.01 (s, 3H), 1.70 (s, 4H), 1.33 (t, J = 7.2, 3H), 1.32 (s, 6H), 1.26 (s,

3H), 1.24 (s, 3H). ¹³C NMR (100.6 MHz, CDCl₃) δ 171.1, 167.2,

144.6, 143.8, 143.1, 142.5, 142.1, 141.6, 139.0, 138.8, 137.7, 132.8,

132.3, 131.7, 130.3, 129.3, 127.9, 127.6, 127.4, 127.2, 126.6, 117.2,

60.4, 60.3, 35.2, 35.1, 34.0, 33.9, 32.0, 31.9, 31.8, 31.8, 21.0, 20.0, 19.8,

19.5, 14.3, 14.1. IR (neat) 2957, 1711, 1635 cm⁻¹. LC-MS (M + H)⁺

calcd for C₂₇H₃₅O₂391.2637, found 391.2655.

(E)-3-(3-(1,2,3,4-Tetrahydro-1,1,4,4,6-pentamethylnaphtha-

len-7-yl)-4-methylphenyl)acrylic Acid (15). To a 100 mL round-

bottom ﬂask containing 47 (0.3288 g, 0.84 mmol) suspended in

methanol (5.0 mL) was added a solution of KOH (0.1412 g, 2.5

mmol) in water (0.18 mL), and the solution was reﬂuxed in an oil bath

preheated to 85 °C for 1 h. The reaction was allowed to cool to room

temperature and acidiﬁed with an aqueous 20% HCl solution (28 mL).

The resulting precipitate was ﬁltered and washed with copious

amounts of water, and the crude white powder was puriﬁed by column

chromatography (25 mL SiO₂, ethyl acetate:hexanes 15:85) to give 15

(0.2295 g, 75%) as a white crystalline solid, mp 189−196 °C. ¹H NMR

(400 MHz, CDCl₃) δ 7.79 (d, J = 16.0, 1H), 7.44 (dd, J = 8.0, 1.6,

1H), 7.37 (d, J = 1.6, 1H), 7.29 (d, J = 7.6, 1H), 7.17 (s, 1H), 7.01 (s,

1H), 6.42, (d, J = 16.0, 1H), 2.11 (s, 3H), 2.02 (s, 3H), 1.71 (s, 4H),

1.33 (s, 3H), 1.32 (s, 3H), 1.27 (s, 3H), 1.24 (s, 3H). ¹³C NMR (100.6

MHz, CDCl₃) δ 172.6, 147.1, 143.8, 142.7, 142.1, 139.7, 137.5, 132.3,

131.4, 130.4, 129.6, 127.7, 126.9, 116.3, 35.2, 35.1, 34.0, 33.9, 32.0,

31.9, 31.8, 20.1, 19.5. IR (neat) 2957, 1680, 1627 cm⁻¹. LC-MS (M +

H)⁺calcd for C₂₅H₃₁O₂363.2324, found 363.2311. Anal. Calcd for

C₂₅H₃₀O₂: C, 82.83; H, 8.34. Found: C, 81.84; H, 8.28.

1

50 (1.0138 g, 71%) as a colorless oil. H NMR (400 MHz, CDCl₃) δ

7.75 (d, J = 8.4, 1H), 7.68 (d, J = 16.0, 1H), 7.57 (d, J = 8.0, 1H), 7.45

(s, 1H), 7.13 (s, 1H), 7.02 (s, 1H), 6.51 (d, J = 16.0, 1H), 4.26 (q, J =

7.2, 2H), 1.98 (s, 3H), 1.69 (s, 4H), 1.32 (t, J = 7.2, 3H), 1.31 (s, 6H),

1.24 (s, 3H), 1.22 (s, 3H). ¹³C NMR (100.6 MHz, CDCl₃) δ 166.4,

144.5, 142.6, 141.9, 141.3, 137.1, 135.1, 132.3, 131.1, 130.2, 129.9,

127.6, 127.3, 126.7, 126.6, 126.3, 125.0, 122.3, 120.8, 60.7, 35.1, 35.0,

33.9, 33.8, 31.9, 31.8, 31.7, 19.7, 14.2. IR (neat) 2960, 1716, 1641

cm⁻¹. GC-MS (M)⁺calcd for C₂₇H₃₁F₃O₂444.2276, found 444.2253.

(E)-3-(4-(Triﬂuoromethyl)-3-(1,2,3,4-tetrahydro-1,1,4,4,6-

pentamethylnaphthalen-7-yl)phenyl)acrylic Acid (16). To a 100

mL round-bottom ﬂask containing 50 (0.4196 g, 1.00 mmol)

suspended in methanol (5.0 mL) was added a solution of KOH

(0.1706 g, 3.04 mmol) in water (0.22 mL), and the solution was

reﬂuxed in an oil bath preheated to 85 °C for 1 h. The reaction was

allowed to cool to room temperature and acidiﬁed with an aqueous

20% HCl solution (33 mL). The resulting precipitate was ﬁltered and

washed with copious amounts of water, and the crude white powder

was puriﬁed by column chromatography (25 mL SiO₂, ethyl

acetate:hexanes 15:85) to give 16 (0.3445 g, 88%) as a white

crystalline solid, mp 217−221 °C. ¹H NMR (400 MHz, CDCl₃) δ 7.79

(d, J = 15.6, 1H), 7.78 (d, J = 8.4, 1H), 7.61 (d, J = 8.4, 1H), 7.48 (s,

1H), 7.15 (s, 1H), 7.03 (s, 1H), 6.52, (d, J = 16.0, 1H), 2.00 (s, 3H),

1.70 (s, 4H), 1.36 (s, 6H), 1.25 (s, 3H), 1.23 (s, 3H). ¹³C NMR (100.6

MHz, CDCl₃) δ 171.7, 145.2, 144.6, 142.1, 141.3, 136.6, 134.9, 132.3,

131.4, 130.7, 130.4, 127.6, 127.3, 126.7, 126.6, 124.9, 122.2, 119.8,

35.1, 35.0, 34.0, 33.8, 31.9, 31.8, 31.7, 19.7. IR (neat) 2961, 1688, 1635

cm⁻¹. LC-MS (M + H)⁺calcd for C₂₅H₂₈F₃O₂417.2041, found

417.2068. Anal. Calcd for C₂₅H₂₇F₃O₂: C, 72.10; H, 6.53; F, 13.69.

Found: C, 71.42; H, 6.45; F, 14.3.

5,8-Dihydro-3,5,5,8,8-pentamethylnaphthalen-2-yl-2-bor-

onic Acid (51).³⁸The method of Faul and co-workers was used.³⁸To

a 100 mL round-bottom ﬂask containing THF (20 mL) was added a

1.6 M solution of n-BuLi in hexanes (4.94 mL, 7.90 mmol), and the

8449

dx.doi.org/10.1021/jm4008517 | J. Med. Chem. 2013, 56, 8432−8454

Journal of Medicinal Chemistry

Article

resulting solution was cooled in a dry ice acetone bath to −78 °C with

stirring, under nitrogen. To this solution was added a solution of 36

(2.1244 g, 7.61 mmol) in THF (5 mL) over 20 min and the reaction

was stirred at −78 °C for 10 min and a mixture of triisopropylborate

(3.5 mL, 15.2 mmol) in THF (5 mL) was added dropwise over 20

min. The reaction was stirred at −78 °C for 2 h and then warmed to

room temperature and stirred for 1 h. The reaction was then quenched

with 3 N HCl (30 mL), and after stirring for 30 min, it was poured

into ethyl acetate, the layers were separated, and the aqueous layer was

extracted with ethyl acetate. The combined organic layers were washed

with brine, dried over sodium sulfate, and concentrated in vacuo to

give a crude product that was puriﬁed by column chromatography

(150 mL SiO₂, ethyl acetate:hexanes 1:3) to give 51 (1.1639 g, 62%)

vacuo to give a crude product that was puriﬁed by column

chromatography (150 mL SiO₂, ethyl acetate:hexanes 1:9) to give

54 (2.44 g, 90%) as a white crystalline solid, mp 58−60 °C. ¹H NMR

(400 MHz, CDCl₃) δ 9.94 (s, 1H), 8.13 (d, J = 2.0, 1H), 7.85 (dd, J =

8.4, 2.0, 1H), 7.31 (d, J = 8.4, 1H), 2.38 (s, 3H). ¹³C NMR (100.6

MHz, CDCl₃) δ 189.5, 167.8, 152.8, 135.2, 134.5, 129.8, 124.5, 117.4,

20.7. IR (neat) 2944, 1743, 1697, 1686 cm⁻¹. GC-MS (M)⁺calcd for

C₉H₇O₃Br 241.9579, found 241.9574.

(E)-Ethyl 3-(4-Acetoxy-3-bromophenyl)acrylate (55).³⁹The

method of Gronemeyer and co-workers was followed.³⁹To a solution

of a 60% dispersion of NaH in mineral oil (0.31 g, 7.75 mmol) in

DME (2 mL) at −30 °C was added a solution of ethyl 2-

phosphonoacetate (1.46 mL, 7.29 mmol) in DME (13 mL), and the

mixture was stirred at this temperature for 30 min. To this solution

was added a solution of 54 (1.61 g, 6.62 mmol) in DME (3 mL), and

the reaction was stirred at −30 °C for 1.5 h and then poured into

water (50 mL) and extracted with ethyl acetate. The combined organic

layers were washed with an aqueous saturated NH₄Cl solution and

then brine, dried over sodium sulfate, ﬁltered, and concentrated in

vacuo to give a crude product that was puriﬁed by column

chromatography (150 mL SiO₂, ethyl acetate:hexanes 1:9) to give

1

as a white crystalline solid, mp 162−166 °C. H NMR (400 MHz,

CDCl₃) δ 8.36 (s, 1H), 7.28 (s, 1H), 5.56 (s, 2H), 2.87 (s, 3H), 1.41

(s, 6H), 1.39 (s, 6H). IR (neat) 3213 (br), 2958, 1605 cm⁻¹.

(2E)-Ethyl 3-(3-(1,4-Dihydro-1,1,4,4,6-pentamethylnaphtha-

len-7-yl)-4-methylphenyl)acrylate (52). To a 50 mL Schlenk ﬂask

charged with bromide 46 (0.6506 g, 2.41 mmol), boronic acid 51

(0.2032 g, 0.83 mmol), TBAB (0.26 g), Na₂CO₃(0.256 g, 2.42 mmol),

and water (1.85 mL) was added Pd(OAc)₂(0.0136 g, 0.061 mmol),

and the ﬂask was evacuated and backﬁlled with nitrogen three times.

The reaction was stirred at room temperature for 15 min and then

placed in an oil bath preheated to 150 °C and stirred for 5 min. The

reaction was allowed to cool to room temperature, and the black

residue was taken up in ethyl acetate and water. The layers were

separated, and the aqueous layer was extracted with ethyl acetate. The

combined organic layers were washed with brine, dried over sodium

sulfate, ﬁltered, and concentrated in vacuo to give a crude product that

was puriﬁed by column chromatography (150 mL SiO₂, ethyl

acetate:hexanes 2.5:97.5) to give 52 (0.2256 g, 72%) as a colorless

oil. ¹H NMR (400 MHz, CDCl₃) δ 7.69 (d, J = 16.0, 1H), 7.42 (dd, J

= 8.0, 2.0, 1H), 7.36 (d, J = 2.0, 1H), 7.29 (d, J = 8.0, 1H), 7.22 (s,

1H), 7.06 (s, 1H), 6.41, (d, J = 16.0, 1H), 5.53 (s, 2H), 4.25 (q, J = 7.2,

2H), 2.10 (s, 3H), 2.04 (s, 3H), 1.38 (s, 6H), 1.33 (t, J = 7.2, 3H), 1.32

(s, 3H), 1.30 (s, 3H). ¹³C NMR (100.6 MHz, CDCl₃) δ 167.8, 144.5,

142.4, 141.5, 139.8, 139.0, 138.1, 133.1, 133.0, 132.7, 131.8, 130.4,

129.3, 127.3, 126.8, 126.6, 117.3, 60.3, 35.0, 34.9, 32.7, 32.6, 32.6, 20.0,

19.6, 14.3, 14.1. IR (neat) 2948, 1698, 1634 cm⁻¹. GC-MS (M)⁺calcd

for C₂₇H₃₂O₂388.2402, found 388.2414.

1

55 (1.7598 g, 84%) as a colorless crystalline solid, mp 71−72 °C. H

NMR (400 MHz, CDCl₃) δ 7.76 (d, J = 2.0, 1H), 7.58 (d, J = 16.0,

1H), 7.46 (dd, J = 8.0, 2.0, 1H), 7.14 (d, J = 8.4, 1H), 6.38, (d, J =

16.0, 1H), 4.25 (q, J = 7.2, 2H), 2.36 (s, 3H), 1.33 (t, J = 7.2, 3H). ¹³C

NMR (100.6 MHz, CDCl₃) δ 168.2, 166.4, 149.3, 141.9, 133.9, 132.5,

127.9, 124.0, 119.8, 116.8, 60.6, 20.7, 14.2. IR (neat) 2981, 1758, 1715

cm⁻¹. GC-MS (M)⁺calcd for C₁₃H₁₃O₂Br 311.9997, found 311.9988.

(2E)-Ethyl-3-(4-acetoxy-3-(1,4-dihydro-1,1,4,4,6-pentame-

thylnaphthalen-7-yl)phenyl)acrylate (56). To a 50 mL Schlenk

ﬂask charged with bromide 55 (0.5038 g, 1.60 mmol), boronic acid 51

(0.4040 g, 1.65 mmol), TBAB (0.52 g), Na₂CO₃(0.51 g, 4.81 mmol),

and water (3.70 mL) was added Pd(OAc)₂(0.0277 g, 0.123 mmol),