Identification of Retinoids with Nuclear Receptor Subtype-selective Activities1 · radical bromination [A'-bromosuccinimide, (C6H5CO2)2,CC14,hu), dis placement of the benzylic bromide

[CANCER RESEARCH 51. 4804-4809. September 15. 1991]

Identification of Retinoids with Nuclear Receptor Subtype-selective Activities1

JÃ¼rgenM. Lehmann, Marcia I. Dawson, Peter D. Hobbs, Matthias Husmann, and Magnus Pfahl2

Cancer Center, La Jolla Cancer Research Foundation, La Jolla, California 92037 [J. M. L., M. H., M. P.], and Life Sciences Division, SRI International,Menlo Park, California 94025 [M. l. D., P. D. H.]

ABSTRACT

Retinole acid (RA) and its synthetic analogues, retinoids, have shownpromising results in the prevention of epithelial carcinogenesis and in thetreatment of acute promyelocytic leukemia and various proliferatile skindisorders. Retinoid action on gene regulation is mediated by three distinctnuclear retinole acid receptor subtypes, RA receptors a, 0, and ->. The

existence of multiple RA receptors has raised the possibility that receptorsubtype-specific retinoids with reduced side effects can be developed. To

analyze the activity of retinoids at the molecular level, we used a receptoractivation assay. RA and 22 retinoids were compared on the three receptorsubtypes. We found the a receptor to be least sensitive to activation byRA and the y receptor to be most sensitive. Compared with RA, one ofthe retinoids showed increased activity for the a and ,? receptors. Threeretinoids revealed no gene activation activity and showed no antagonisticeffects when assayed in the presence of RA. Surprisingly, several of theretinoids were efficient activators of the ÃŸand 7 receptors but pooractivators or nonactivators of the a receptor. Our data demonstrate thatthe three RA receptor subtypes have differential ligand activation specificities and that the design of receptor subtype-selective retinoids is

possible.

INTRODUCTION

RA3 and its synthetic analogues (retinoids) affect processes

as diverse as growth, differentiation, morphogenesis, reproduction, and vision (for a review, see Refs. 1-3) and suppress andreverse malignant transformation induced by either chemicalcarcinogens or ionizing radiation (4, 5). The beneficial effectsof retinoids in the treatment of cancer patients is best reflectedby a recent clinical study by Hong et al. (6). The syntheticretinoid 13-c/'s-retinoic acid (isotretinoin), administered in high

doses, significantly reduced the occurrence of second primarytumors in patients with squamous cell carcinomas of the headand neck. The full potential of retinoids as chemopreventativeagents has been hampered by their undesirable systemic sideeffects (7) and teratogenicity (8). Although many syntheticretinoids have been evaluated (9), analysis of the structure-activity relationships has been difficult because of the differences in activities of individual retinoids observed in the variousin vitro and in vivo test systems (9, 10). These differences cannotbe explained solely by pharmacokinetic arguments and probablyresult from different modes of retinoid action in the varioussystems.

The recent discovery of nuclear RARs (11-16) as mediatorsof RA action on gene regulation provides us with the opportunity to investigate the effects of retinoids at the molecular level.The RARs belong to a large family of ligand-activated transcrip-

Received 4/18/91; accepted 7/9/91.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported in parts by NIH Grants DK35083 and CA50676

(M. P.) and 32428 (M. I. D.). J. M. L. and M. H. were supported by a fellowshipfrom the Deutsche Forschungsgemeinschaft.

1To whom requests for reprints should be addressed, at La Jolla Cancer

Research Foundation, 10901 North Torrey Pines Road. La Jolla. CA 92037.'The abbreviations used are: RA. rctinoic acid; CAT, chloramphenicol acetyl

transferase; ER, estrogen receptor; ERE, estrogen responsive element: RAR.retinoic acid receptor; rfRARE, retinoic acid-responsive element of the RAR/3promoter; tk, thymidine kinase.

tion factors that interact specifically with cognate DNA elements (i.e., hormone response elements; for a review see Refs.17 and 18). Three retinoic acid receptors, RARÂ«, ÃŸ,and 7,encoded by distinct genes, have been identified in humans (lilo). In addition, a related group of receptors (RXRa, ÃŸ,7) wasisolated, which are activated by retinoids as well (19). Fromeach of the three RAR genes a number of receptor isoformscan be generated that contain different amino terminal sequences but are identical in their DNA- and ligand-bindingdomains (15, 16, 20-23). The identification of this complexvariation in RARs suggests that many of the diverse biologicaleffects of retinoids may be due to their differential receptoraffinities, combined with different temporal and spatial expression patterns of receptor subtypes and isoforms (24-26).

To elucidate the molecular mechanisms underlying retinoidtranscriptional activity and to develop more specific retinoidshaving fewer undesirable side effects, we investigated the possibility of identifying receptor-selective retinoids as well asantagonists. To achieve this goal, an assay system was developed in which retinoids were evaluated for their transcriptionalactivation capabilities. Because apparently all mammalian celllines contain endogenous RARs (Ref. 27),4 hybrid receptorswere constructed that consist of the ligand-binding domain andthe carboxy-terminal portion of either RARa, ÃŸ,or 7, and theDNA-binding domain and amino-terminal portion of the ER.These hybrid receptors showed retinoid specificities identicalto those of the wild-type RARs (28). RA and 22 conformation-ally restricted retinoids were evaluated for their transcriptionalactivation activity. The identification of several retinoids withreceptor subtype-selective activity patterns shows for the firsttime that RARs can be activated differentially and that thedevelopment of receptor-specific retinoids is possible.

MATERIALS AND METHODS

Chemicals. All-/rani-retinoic acid (Rl of Fig. 3) was purchased fromSigma Chemical Co. (St. Louis, MO). (Â£)-4-[2-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)propenyl]benzoic acid (R2) wasprepared by the method of Loeliger et al. (29). The following retinoids(R3-R10), which are analogues of R2, were prepared by similar methods, which are described elsewhere (30, 31): (Â£>4-[2-(5,6,7,8-tetrahy-dro-5,5-dimethyl-2-naphthalenyl)propenyljbenzoic acid (R3) (30); (Â£)-7-[1-(4-carboxyphenyl)propen-2-yl]-3,4-dihydro-2//-1 -benzothiopyran(R4) (30); (E)-4-[2-(5,6,7,8-tetrahydro-8,8-dimethyl-2-naphtha-lenyl)propenyl]benzoic acid (R5) (30); (Â£>6-[l-(4-carboxy-phenyl)propen-2-yl]-4,4-dimethyl-3,4-dihydro-2//-1 -benzopyran (R6)(31); (E)-6-[l-(4-carboxyphenyl)-4,4-dimethylpropen-2-yl]-3,4-dihydro-2//-1 -benzothiopyran (R7) (31); (Â£)-4-[2-(3-f-butyi-4-methoxy-phenyl)propenyl]benzoic acid (R8) (30); (Â£)-4-|2-[4-(3-methylbu-tyl)thiophenyl]propenyl|benzoic acid (R9) (30); and (Â£)-4-[2-(4,5,6,7-tetrahydro-4,4-dimethylbenzo[6]thienyl)propenyl] benzoic acid (RIO)(30). (Â£)-4-[l-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyl-2-naphtha-lenyl)propen-2-yl]benzoic acid (RII) was prepared as described (32).(Â£)-6-[2-(4-Carboxyphenyl)propenyl]-4,4-dimethyl-3,4-dihydro-2//-l-benzothiopyran (R12) was prepared by a similar method (30). Thesyntheses of retinoids R13 and R14 have not been previously reported

4 Unpublished observations.

4804

on June 30, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

RETINOIDS WITH RECEPTOR-SELECTIVE ACTIVITIES

and so will be described here. (E)-4-[l-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyl-4-methoxy-2-naphthalenyl)propen-2-yl]benzoic acid (R 13)(30) was prepared by the following route: (a) introduction of functionality at the 4-position of the tetrahydronaphthalene ring by nitration(HNO3, H2SO4, MeNO2, 0Â°C;79%) of ethyl 5,6,7,8-tetrahydro-5,5,8,8-

tetramethyl-2-naphthalenecarboxylate (33) at the 3-position, followedby reduction [H2, Pd(C), Et2O] and acetylation (Ac2O, pyridine, Et2O;100%) to give the amid, and nitration (NO2BF4, MeCN, 0-20Â°C;74%)

at the 4-position, cleavage (KOH, aq. EtOH; aq. HC1; 97%) of theacetyl group, removal of the 3-amino group by deamination (n-C6H,3ONO, CF3CO2H/EtOH; H3PO2; 81 %), and reduction [H2, Pd(C),EtOH; 100%] of the 4-nitro group to the amine to give 4-amino-5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenecarboxylic acid; (b) conversion of the 2-carboxylic acid function to the phosphonium salt byreduction (LiAlH4, tetrahydrofuran, 20Â°C;aq. NaOH), acetylation

(Ac2O, pyridine, Et2O), hydrolysis of the benzyl acetate (K2CO3,MeOH, 20Â°C;76% overall) to the alcohol, and treatment with (C6H5)3P(xylene, 80Â°C;89%); (c) formation of the propen-2-yl bond system byWittig reaction (-10Â°C to 20Â°C;58%) of the ylid derived from treatment of the phosphonium salt with n-BuLi (tetrahydrofuran, -35Â°Ctoâ€”10Â°C)and ethyl 4-acetylbenzoate, and Chromatographie separation

(35% EtOAc/hexane, silica gel; 41%) of the Â£isomer; (d) hydrolysis(KOH, aq. diethyleneglycol, 40Â°C;aq. HC1; 83%) of the acid and amine

protecting groups; and (e) introduction of the 4-methoxy group bynonaqueous diazotization (n-C6H13ONO, CF3CO2H/EtOH) andaqueous thermal decomposition (aq. NaOH; aq. H2SO4) of the resultantdiazonium salt to give the phenol, which was purified as the ethyl ester(CH3CHN2, Et2O; 56% overall), followed by methylation of the phenolunder phase-transfer conditions (Me2SO4, (n-Bu)4NI, 50% aq. NaOH,CH2C12;94%) (34), and ester hydrolysis (NaOH, aq. MeOCH2CH2OH,20Â°C;aq. HC1; 100%). This rather complex route was employed because

of the availability of the starting materials and was not process developed. (Â£)-4-[3-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyl-2-naphthal-enyl)-2-buten-2-yl]benzoic acid (R14), Ã-ranÃ--4-[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)cyclopropyl]benzoic acid (R 15),and 4-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-anthracenyl)benzoicacid (R16) were synthesized as described (32), as were (4-(4-azido-5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-anthracenyl)benzoic acid(R 17) and (Â£>3-azido-4-[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)propenyl]benzoic acid (R19) (35).

4-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyI)naphtho[2,3-b]thiophen-2-yl)benzoic acid (R18) was prepared from the methyl ester of thecorresponding 3-bromo analogue, which was used as an intermediatein another synthesis. This intermediate was obtained by (a) introductionof ring functionality by Friedel-Crafts alkylation of benzo[Â¿>]thiophenewith toluene (A1C13,CH2C12,80*C) (36) at the 2-position to give the 4-

methylphenyl and 2-methylphenyl isomers, followed by Chromatographie isolation of the 4-methylphenyl compound (20%) and bromi-nation (CHC13; 92%) at the 3-position of its benzo[6]thiophene ring,and cyclialkylation (cat. A1C13,CH2C12, -10Â°C; 44%) of the resultant

2-aryl-3-bromobenzo[2>]thiophene with 2,5-dichloro-2,5-dimethylhex-ane; and (b) conversion of the methyl group to the methyl ester byradical bromination [A'-bromosuccinimide, (C6H5CO2)2,CC14,hu), dis

placement of the benzylic bromide by acetate (CH3CO2K, HCONMe2;70%), and transesterification to give the benzyl alcohol (K2CO3, EtOH,80Â°C;45% overall), which was oxidized using Corey's procedure(Mn02, NaCN, HOAc, MeOH, 20Â°C;90%) (37) to give the methylester. The 3-bromo group was removed by hydrogenolysis [H2, Pd(C),Et3N, EtOAc; 95%), and the methyl ester removed (KOH, aq. EtOH;aq. H2SO4; 100%) to give R18.

6-(5,6,7,8-Tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-2-naphthalenecarbo xylic acid (R20), 6-(5,6,7,8-tetrahydro-3,5,5,8,8-pen-tamethyl-2-naphthalenyl)-2-naphthalenecarboxylic acid (R22), and 6-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-5-methyl-2-naphthalenecarboxylic acid (R23) were synthesized as described (32),using methodology that was also employed to synthesize 6-(5,6,7,8-tetrahydro-8,8-dimethyl-2-naphthalenyl)-2-naphthalenecarboxylic acid(R21) (30). All retinoids were fully characterized ('H nuclear magnetic

resonance spectrometry, infrared and UV spectroscopy, melting point,and elemental analysis) and were of at least 99% purity by high-

performance liquid chromatography. Stock solutions (10 2 M) in dimethyl sulfoxide were maintained at -20Â°C. Dilutions were made in

cell culture medium prior to usage.Transient Transfection. For all transient transfection assays we used

green monkey kidney cells (CV-1 ) grown in Dulbecco's modified Eagle's

medium supplemented with 10% fetal calf serum. The fetal calf serumhad been pretreated with charcoal (12 mg/ml serum) to remove retinoids. Transient transfections using the calcium phosphate precipitation procedure were performed as described (Ref. 28 and referencestherein) with the following modifications. Cells (7 x IO4)were seededin a volume of 1 ml/well in 24-well tissue culture plates (Costar,Cambridge, MA). In each well, transfection mixtures contained a totalof 1 /jg supercoiled plasmid DNA including 50 ng of receptor expressionvector and 100 ng of reporter plasmid for each well to be transfected.Cotransfection of 300 ng /3-galactosidase expression vector pCHllO(Pharmacia, Piscataway, NJ) served to correct for variation in transfection and harvesting efficiency. After overnight incubation, mediumexchange, and retinoid treatment for 24 h, lysates were obtained bythree in situ freeze-thaw cycles. Cell debris was removed by centrifu-gation. Quantitation of CAT activity was performed with a sensitiveextraction-liquid scintillation procedure described elsewhere (Ref. 28and references therein). Linearity of the assays with time and ratio ofCAT: /3-galactosidase activity was established.

Hybrid Receptor Constructs. RAR/3 was previously referred to asRARÂ«(14). Isolation of RAR7 (21), as well as the isolation of theRARÂ«ligand-binding domain and the construction of ER-RAR hybridreceptors using the polymerase chain reaction, have been described(28). Briefly, standard polymerase chain reactions (10 cycles, denatu-ration at 94'C for 1.5 min, annealing for 2 min at 37'C, elongation for2 min at 72"C) were carried out using 1 Mg of the corresponding

receptor complementary DNA cloned in pBluescript (Stratagene, LaJolla, CA). The oligonucleotide primers were synthesized according tothe complementary DNA sequence but contained an artificially introduced endonuclease restriction site (i.e., BamHl site). As a commonlink-up site of the NH2-terminal estrogen receptor half and the COOH-terminal half of the RAR subtypes, we have chosen a site in the hingeregion, 71 bases upstream of the beginning of the ligand-bindingdomain.

RESULTS

RAR Hybrid Receptors Show Subtype Differential Response.Because mammalian cell lines generally contain variousamounts of endogenous RARs (27, 38), it is often impossibleto distinguish activation of the transgene by the endogenousreceptor(s) from activation by the cotransfected target RAR.Hybrid receptors that contain the ligand-binding domains ofRARs and the DNA-binding domain of the less ubiquitouslyexpressed receptor allow the usage of a hybrid receptor-specificreporter gene for measuring retinoid activities that cannot beactivated by endogenous RARs. An example is shown in Fig.

ÃŸRARE tk-CAT ERE tk-CAT

RA RA

Fig. 1. Reporter gene response to endogenous RAR activity of CV-1 cells. CV-1 cells were transiently transfected with either the ÃŸRARE-tk-CAT(38) or theERE-tk-CAT reporter gene (41), treated for 24 h with RA as indicated, andassayed for CAT activity as described in "Materials and Methods."

4805




1, CV-1 cells reveal endogenous RAR activity when transfectedwith a reporter construct containing the ÃŸRARE(38) but notwhen a reporter gene is used that contains an ERE. Therefore,we employed hybrid receptors that contained the DNA-bindingdomain and amino-terminal half of the human ER linked tothe ligand-binding domain and carboxy-terminal portion of thethree different RAR subtypes a, ÃŸ,or 7, in order to be able torigorously determine receptor-specific retinoid activities. Domain swapping among nuclear receptors is possible because theligand-binding domains of nuclear receptors maintain theirspecificity when linked to the DNA-binding domains of otherreceptors (14,39) and even when linked to members of differenttranscription factor families (40). In addition we previouslydemonstrated that hybrid receptors containing the RAR ligand-binding domain were activated specifically by RA and otherretinoids (28, 39). The linkage site for all the hybrid-receptorconstructs was at an introduced restriction site in the hingeregion located between the DNA-binding and ligand-bindingdomains as described in "Materials and Methods." As the

reporter gene, a CAT construct (ERE-tk-CAT) was used thatcontained an ERE at the â€”105position of the herpes simplex

virus tk promoter (41).Transcriptional activation of the three hybrid receptors by

RA at concentrations ranging from 0.1 nM to 1 IÂ¿Mshowed adifferential response (Fig. 2A). The y and ÃŸreceptors were themost sensitive and the a receptor the least sensitive. In addition,all three receptors showed different basal level activities in theabsence of ligand, with y having the highest and a the lowestlevel of constitutive activity. Both results are consistent withdata obtained for the wild-type receptors (16, 21). To comparethe retinoid activities in the transcriptional activation assay, thedata were normalized by subtraction of the basal level activityand by expressing the amount of transcriptional activation as apercentage of the activation obtained with each receptor at 10~6

M RA (Fig. 2B).The fact that the three hybrid receptors yielded distinct

induction profiles indicated that the structural differences observed in their ligand-binding domains confer differential responsiveness to RA. These data suggested to us the possibilityof finding retinoids with enhanced selective responsiveness toone of the receptors.

Synthetic Retinoids Reveal Receptor-Selective Activation Profiles. We compared the receptor-mediated transcriptional activation activity of 22 retinoids with that of RA. These retinoidswere conformationally restricted analogues of RA in whichselected bonds of the polyolefinic chain of the retinoid skeletonwere incorporated into aromatic ring systems. Because thesecompounds have fewer degrees of freedom, they provide moreuseful information about the geometry of the ligand-bindingsite than the more conformationally flexible RA. The structuresof these analogues, which are shown in Fig. 3, fall into twogeneral classes, the 4-substituted benzoic acids R2-R19 havingthe 11,13-double bonds of RA replaced by a phenyl ring andthe 6-substituted 2-naphthalenecarboxylic acids R20-R23 having the 9,11,13-double bond system of RA replaced by a naphthalene ring. All analogues have an aromatic ring (phenyl orthienyl) replacing the 5,7-double bond system. Modificationsin the /3-cyclogeranylidene ring included removal of the gem-dimethyl groups at the 1- and 4-positions of the ring, replacement of carbons at these positions with more polar atoms(oxygen and sulfur), and species having the 2-3 and 1-6 bondsbroken. Other modifications were made in the region of the 9-double bond, where the methyl group was shifted to the adjacent10-posi t ion. an additional methyl group was placed at the 10-position, the bond system was replaced by a cyclopropyl ring,

10 ' 10 Â° 10

RA cone. (M)

Fig. 2. Retinole acid-dependent transcriptional activation of ER-RAR hybridreceptors. CV-1 cells were transiently transfected with expression plasmids forthe hybrid receptors ER-RARÂ«,ÃŸ,or y, the reporter gene ERE-tk-CAT, and (forinternal standardization) with the tÃ--galactosidaseexpression plasmid p( III 10.Cells were treated subsequently for 24 h with increasing concentrations of RA, asindicated, and were assayed for CAT activity. A, dose-response curve for receptor-mediated transactivation by RA. Points, average of five experiments; B, receptoractivation expressed as a percentage of maximal induction (measured at 10 " \i

RA) after constitutive receptor activity was substracted.

or the incorporation of the bond system into a phenyl or thienylring. The effect of increased steric bulk on biological activitywas assessed using methoxy, azido, and methyl group substit-uents at various aromatic ring positions.

The transactivation assay results are summarized in Fig. 3.The data represent the average of three independent experiments, and the SD at retinoid concentrations that resulted insubstantial (>25%) receptor activation typically did not exceed10%. To ensure that the hybrid receptors faithfully representedRAR subtype specificity we also evaluated all retinoids for theiractivity with RAR/3 and observed no substantial differencesfrom ER-RAR0 (data not shown). One of the retinoids, R16,showed clearly increased activity for the 0-receptor and a slightbut significant increase for the a-receptor compared with thatof RA (Fig. 4A). The high activity was consistent with highactivities reported for R16 in other in vitro systems (9). Threeretinoids (R9, R14, and R23) displayed no gene activationactivity, which again was consistent with their lack of activityin in vitro assay systems (9). When these compounds wereassayed in the presence of RA, they did not display any antagonist activity either, suggesting that their affinity for the receptors is low (data not shown).

Several retinoids displayed differential activity for the threereceptors. Generally, the greater differences in transcriptionalactivation activity were observed between ÃŸand a and betweeny and a. The strongest examples are R15, R19, R20, and R22,which were efficient activators of the ÃŸand y receptors atcritical concentrations but poor activators or nonactivators ofthe a receptor (Fig. 4B). In contrast, activation of ÃŸby 10~7and 10~6 M R21 was 2-fold higher than that for the a or 7

receptor (Fig. 4C). This was the only example in the two seriestested in which the activation pattern of 7 was clearly closer tothat of a than to .1 In general, our studies indicated that thetranscriptional activation assay described here will permit theidentification of retinoids with differential receptor activationactivity. Receptor activation was dependent on retinoid concentration, except in two cases (R7 and R13), in which activationof the ÃŸreceptor was higher at IO"7M than at IO"6M.

4806




Fig. 3. Summary of transcriptional activation characteristics of RA and 22 retinoids.Structures of RA (Rl) and the retinoids R2 toR23 are shown. The receptor activation capacities were evaluated as described in Fig. 2.Retinoid-induced receptor activation is expressed as a percentage of maximal inductionby RA after constitutive receptor activity wassubstracted. The activation values were dividedinto five activity groups, as indicated by thesizes of the columns: 81%. The data represent theaverage of three experiments. One hundred %receptor activation represents the activity observed for each receptor in the presence of 10~6

M RA.

DC retinoid cone (M)RETINOID 2 10'Â°109 10 8 10 7 10 10'Â° 109 10e 107 106

roroptcr activation ('

â€¢ â€¢81

DISCUSSION

We have analyzed here retinoids as inducers of retinoic acidreceptor subtypes, as measured by gene activation. To facilitatethe interpretation of the retinoid induction data, we have employed hybrid receptors in the transcriptional activation assaythat contain the NH2-terminal half of the ER and COOH-

terminal half of the different RAR subtypes. When comparedto wild-type RARs we observe no significant difference in theresponse of these hybrid receptors to RA and the 22 retinoidstested. Thus, at the molecular level, this system is likely tomeasure the major biological activity of individual RARs. Surprisingly, by having tested only a limited number of compounds,we were able to obtain clear evidence that individual retinoidscan have striking differences in their activation capacity for thedifferent receptor subtypes. This may at least partially explainthe different activity profiles in various biological assays observed previously for the same retinoids (10). The variations indifferent assay systems may thus reflect the activities of distinctRARs or various combinations of them. In our transcriptionalactivation assay we used green monkey kidney CV-1 cells. Many

other cell lines can certainly be used. It will be of interest to seewhether cellular factors, including the retinoid binding proteinsCRBP and CRABP and retinoid-modifying enzymes present insome cell lines, can influence the transcriptional activationcapacity of certain retinoids. From enzymatic activities positiveas well as negative effects can be expected. For CRABP, recentevidence indicates that it may reduce the effective intracellularconcentration of RA (42); thus similar effects can be expectedfor other CRABP-binding retinoids.

Although the three-dimensional structure of the RAR ligand-binding pockets has not yet been elucidated, analysis of thesequence homology of the RARs indicates that the ligand-

binding domains of RARi and ÃŸare more closely related toeach other than to RARa (11-16). This finding is consistentwith our observation that some of these retinoids showed activation profiles that were more similar and higher for RAR/3and 7 than for RARa. Interestingly, the inverse situation inwhich a compound showed higher activity with RARa thanwith ÃŸand y was not found. These results suggest that at leastfor the compounds analyzed here, the ÃŸand 7 receptor ligand-

4807




B

Fig. 4. Dose-response curves for retinoids. The receptor activation patterns were evaluated with the ER-RAR hybrid receptors (Â«,0, y) as described for Rl in Fig.2. One hundred % receptor activation represents the activity measured in the presence of IO"6M RA. A, retinoid R16; A, receptor subtype specific retinoid R15, R19,

R20, and R22; C, retinoids R21.

binding domains are more tolerant of the tetraene chain-modified retinoids than is the a receptor ligand-binding domain.The ligand-binding domain of the RARs has been conservedalmost completely during mammalian evolution, showing a 99to 100% identity between human and mouse (20). Thus, thedifferences in retinoid activities appear to correlate with theminor structural differences in this domain among the receptorsubtypes. Northern blot and in situ hybridization analysis demonstrated ubiquitous expression of RARa, whereas RAR/3 andy expression is spatially and temporally regulated (26-28). Theavailability of RAR/3/7-specific retinoids allows the biologicalfunctions of the ubiquitously expressed RARÂ« to be distinguished from those of RAR/3 and y.

Our study indicates that receptor-selective retinoids can beidentified. RARa appears to be far more sensitive to variationsin the structural geometry of the retinoids in the 4-substitutedbenzoic acid and 6-substituted 2-naphthalenecarboxylic acidclasses than RAR/? and RAR7 are. Modifications at the 4-position of the retinoid skeleton appear to contribute to thisselectivity. An independent study, analyzing modifications of6'-substituted naphthalene-2-carboxylic acid analogues in moredetail, has also revealed striking selectivity of this class of c/'.v-

RA analogues for RAR/3 and RAR7 over RARa.5 Our results

are also consistent with other reports of receptor selectivity.For example, benzoic acid R2 is reported to have 50% of thebinding affinity for recombinant RARa and 70% of the affinityfor RAR/3 that RA has (43). R2 was also found to induce RAR/3mRNA in S91-C2 melanoma cells less efficient than RA (44).This is consistent with our results that show R2 as a lessefficient activator than RA of these two receptors. 2-Naphtha-lenecarboxylic acid R20, which we found is less effective thanRA at RARa and RAR/3 activation, is also reported to bind toboth recombinant receptors less efficiently (43). In addition,our most efficient transcriptional activator, R16, had the highest binding affinity among 13 retinoids evaluated (43). RA andR2 were also among a series of eight retinoids evaluated forstimulation of RAR-dependent transcription using a retinoicacid-responsive reporter gene (TRE3-tk-CAT) in cotransfected

*G. Graupner, G. Malle, G. Lang, M. PruniÃ©ras,and M. Pfahl. 6'-Substitutednaphthalene-2-carboxylic acid analogs, a new class of retinoic acid receptorsubtype-specific ligands. Biochem. Biophys. Res. Commun., in press, 1991.

CV-1 cells (45). Although clear differences in activation activities of compounds for RAR subtypes were observed, the retinoids analyzed were active with all three receptor subtypes.

In conclusion, our molecular approach has yielded resultsthat allow us to further select and develop receptor-specificretinoids. Such compounds will be valuable tools for the dissection of the complex biological pathways of RA action. Theefficacy of retinoids as anticancer drugs is best reflected by tworecent studies, one showing the prevention of second primarycarcinomas in patients with squamous cell carcinomas of thehead and neck (6), the other showing complete remissions inpatients with acute promyelocytic leukemia upon treatmentwith retinoic acid (46). However, intolerable side effects due torelatively high doses and length of the treatment led to highdropout rates. This underscores the importance of developingbetter tolerated retinoids which might be accomplished by theusage of receptor subtype-specific retinoids. Together with theknowledge of the receptor distribution in normal and diseasedtissue, receptor-specific retinoids may in the future permit thedesign of efficient retinoid therapies with minimal side effects.The test systems described here can be easily extended to otherretinoid receptors such as the RXRs, which appear to be activated by RA derivatives rather than by RA itself ( 19).

ACKNOWLEDGMENTS

We thank Elmira Khodapanah for her excellent technical assistance.

REFERENCES

1. Roberts, A. B., and Sporn, M. B. Cellular biology and biochemistry of theretinoids. In: M. B. Sporn, A. B. Roberts, and D. S. Goodman (eds.), TheRetinoids, Vol. 2, pp. 210-286. Orlando, FL: Academic Press, Inc., 1984.

2. Lotan, R. Effect of vitamin A and its analogs (retinoids) on normal andneoplastic cells. Biochim. Biophys. Acta, 605: 33-91, 1980.

3. Thaller,


R. M., Schantz, S. P., Kramer, A. M., Lotan, R., Peters, L. J., Dimer>', I.W., Brown, B. W., and Goepfert, H. Prevention of second primary tumorswith isotretinoin in squamous-cell carcinoma of the head and neck. N. Engl.J. Med., 323: 795-801, 1990.

7. Moon, R. G., and Itri, L. M. Retinoids and cancer. In: M. B. Sporn, A. B.Roberts, and D. S. Goodman (eds.). The Retinoids, Vol. 2, pp. 327-371.New York: Academic Press, 1984.

8. Lammer, E. J., Chen, D. T., Hoar, R. M., Agnisti, N. D., Benke, P. J.,Braun, J. T., Curry, C. J., Fernhoff, P. M., Grix, Jr., A. W., Lott, I. T.,Richard, J. M., and Shyan, C. S. Retinoic acid embryopathy. N. Engl. J.Med., 313: 837-841, 1985.

9. Dawson, M. I., Chao, W., Hobbs, P. D., and Delair, T. The inhibitory effectsof retinoids on the induction of ornithine decarboxylase and the promotionof tumors in mouse epidermis. In: M. I. Dawson and W. H. Okamura (eds.),Chemistry and Biology of Synthetic Retinoids, pp. 385-466. Boca Raton,FL: CRC Press, Inc., 1990.

10. Sporn, M. B., and Roberts, A. B. Biological methods for analysis and assayof retinoidsâ€”relationship between structure and activity. In: M. B. Sporn,A. B. Roberts, and D. S. Goodman (eds.), The Retinoids, Vol. 2, pp. 235-279. Orlando, FL: Academic Press, Inc., 1984.

11. Petkovich, M., Brand, N. J., Krust, A., and ChambÃ³n, P. A human retinoicacid receptor belongs to the family of nuclear receptors. Nature (Lond.), 330:444-450, 1987.

12. ( i ignore, V., Ong, E. S., Segui, P., and Evans, R. M. Identification of areceptor for the morphogen retinoic acid. Nature (Lond.), 330: 624-629,1987.

13. Brand, N., Petkovich, M., Krust, A., de The, H., Marchio, A., Tiollais, P.,and Dejean, A. Identification of a second human retinoic acid receptor.Nature (Lond.), 332:850-853. 1988.

14. Benbrook, D., Lernhardt, E., and Pfahl, M. A new retinoic acid receptoridentified from a hepatocellular carcinoma. Nature (Lond.), 333: 669-672,1988.

15. Krust, A., Kastner, P. H., Petkovich, M., Zelent, A., and ChambÃ³n, P. Athird retinoic acid receptor, hRAR-i. Proc. Nati. Acad. Sci. USA, 86: 5310-5314, 1989.

16. Giguere, V., Shago, M., Zirngibl, R., Tate, P., Rossant, J.. and Varmuza, S.Identification of a novel isoform of the retinoic acid receptor-y expressed inthe mouse embryo. Mol. Cell. Biol., 10: 2335-2340, 1990.

17. Evans, R. M. The steroid and thyroid hormone superfamily. Science (Washington DC), 240: 889-895, 1988.

18. Green, S., and ChambÃ³n, P. Nuclear receptors enhance our understandingof transcription regulation. Trends Genet., 4: 309-314, 1988.

19. Mangelsdorf, D. J., Ong, E. S., Dyck, J. A., and Evans, R. M. Nuclearreceptor that identified a novel retinoic acid response pathway. Nature(Lond.), 345: 224-229, 1990.

20. Kastner, P. H., Krust, A., Mendelsohn, C., Gamier, J. M., Zelent, A., Leroy,P., Staub, A., and ChambÃ³n, P. Murine isoforms of retinoic acid receptor ywith specific patterns of expression. Proc. Nati. Acad. Sci. USA, 87: 2700-2704, 1990.

21. 1dimani], J. M., Hoffmann, B., and Pfahl, M. Genomic organization of theretinoic acid receptor y gene. Nucleic Acids Res., 19: 573-578, 1991.

22. Leroy, P., Krust, A., Zelent, A., Mendelsohn, C., Gamier, J-M., Kastner, P.,Dierich, A., and ChambÃ³n, P. Multiple isoforms of the mouse retinoic acidreceptor a are generated by alternative splicing and differential induction byretinoic acid. EMBO J., 10: 59-69, 1991.

23. Zelent, A., Mendelsohn, C., Kastner, P., Krust, A., Gamier, J-M., Ruffenach,F., Leroy, P., and ChambÃ³n, P. Differentially expressed isoforms of themouse retinoic acid receptor .t are generated by usage of two promoters andalternative splicing. EMBO J., 10: 71-81, 1991.

24. Dolle, P., Ruberie, E., Kastner, P., Petkovich, M., Stoner, C. M., Gudas, L.J., and ChambÃ³n, P. Differential expression of genes encoding a, ÃŸand yretinoic acid receptors and CRABP in the developing limbs of the mouse.Nature (Lond.), 342: 702-704, 1989.

25. Ruberie, E., Dolle, P., Krust, A., Zelent, A., Morriss-Kay, G., and ChambÃ³n,P. Specific spatial and temporal distribution of retinoic acid receptor ytranscripts during mouse embryogenesis. Development (Camb.), Â¡08:213-222, 1990.

26. Dolle, P., Ruberie, E., Leroy, P., Morriss-Kay, G., and ChambÃ³n, P. Retinoic

acid receptors and cellular retinoid binding proteins. Development (Camb.),7/0:1133-1151. 1990.

27. Sucov, H. M., Murakami. K. K., and Evans. R. M. Characterization of anautoregulated response element in the mouse retinoic acid receptor type .',gene. Proc. Nati. Acad. Sci. USA, 87: 5392-5396, 1990.

28. Pfahl, M., Tzukerman, M., Zhang, X-K., Lehmann, J. M., Hermann, T.,Wills, K. N., and Graupner, G. Rapid procedures for nuclear retinoic acidreceptor cloning and their analysis. Methods Enzymol., 153: 256-270, 1990.

29. Loeliger, P., Bollag, W., and Mayer, H. Arotinoids, a new class of highly-active retinoids. Eur. J. Med. Chem., IS: 9-15, 1980.

30. Schiff, L. J., Okamura, W. H., Dawson, M. I., and Hobbs, P. D. Structurebiological activity relationships of new synthetic retinoids on epithelial differentiation of cultured hamster trachea. In: M. I. Dawson and W. H.Okamura (eds.). Chemistry and Biology of Synthetic Retinoids, pp. 308-363. Boca Raton, FL: CRC Press, Inc., 1990.

31. Dawson, M. I., Hobbs, P. D., Derdzinski, K., Chan, R. L., Gruber, J., Chao,W., Smith, S., Thies, R. W., and Schiff, L. J. Conformationally restrictedretinoids. J. Med. Chem., 27: 1516-1531, 1984.

32. Dawson, M. I., Hobbs. P. D., Derdzinski, K. A., Chao. W.. Frenking, G.,Loew, G. H., Jetten, A. M., Napoli, J. L., Williams, J. B., Sani, B. P., Wille,J. J., Jr., and Schiff, L. J. Effect of structural modifications in the C7-C11region of the retinoid skeleton on biological activity in a series of aromaticretinoids. J. Med. Chem., 32: 1504-1517, 1989.

33. Myhre, P. C., and Schubert, W. M. Isolation and proof of structure ofl,l,4,4-tetramethyl-6-/-butyl-l,2,3,4-tetrahydronaphthalene. J. Org. Chem.,25:708-711,1960.

34. McKillop, A., Fiaud, J. C., and Hug, R. P. The use of phase-transfer catalysisfor the synthesis of phenol ethers. Tetrahedron, 30: 1379-1382, 1974.

35. Sani, B. P., Wille, J. J., Jr., Dawson, M. I., Hobbs, P. D., Bupp, J., Rhee,S., Chao, W., Dorsky, A., and Morimoto, H. Biologically active aromaticretinoids bearing azido photoaffinity-labeling groups and their binding tocellular retinoic acid-binding protein. Chem.-Biol. Interact.. 75: 293-304,1990.

36. Clark, P. D., Clarke, K., Ewing, D. F., and Scrowston, R. M. Additionreactions of benzo[ojthiophen. Part 1. Self-addition of simple aromatichydrocarbons. J. Chem. Soc. Perkin Trans. I, /: 677-685, 1980.

37. Corey, E. J., Gilman, N. W., and Ganem, B. E. New methods for theoxidation of aldehydes to carboxylic acids and esters. J. Am. Chem. Soc., 90:5616-5617, 1968.

38. Hoffmann, B., Lehmann, J. M., Zhang, X-k., Hermann, T., Husmann, M.,Graupner, G., and Pfahl, M. A retinoic acid receptor specific element controlsthe retinoic acid receptor-ii promoter. J. Mol. Endo., 4: 1734-1743, 1990.

39. Graupner, G., Wills, K. N., Tzukerman, M., Zhang, X-K., and Pfahl, M.Dual regulatory role for thyroid-hormone receptors allows control of retinoic-acid receptor activity. Nature (Lond.), 340: 653-656, 1989.

40. Elliston, J. F., Tsai, S. Y., O'Malley, B. W., and Tsai, M-J. Superactiveestrogen receptors. J. Biol. Chem., 265: 11517-11521, 1990.

41. Klein-Hitpass, L., Schorpp, M., Wagner, U., and Ryffel, G. U. An estrogen-responsive element derived from the 5' flanking region of the Xenopusvitellogenin A2 gene functions in transferred human cells. Cell, 46: 1053-1061. 1986.

42. Boylan, J. F., and Gudas, L. J. Overexpression of the cellular retinoic acidbinding protein-l (CRABP-I) results in a reduction in differentiation-specificgene expression in F9 teratocarcinoma cells. J. Cell Biol.. 112: 965-979,1991.

43. Crettaz, M., Baron, A., Siegenthaler, G., and Hunziker, W. Ligand specificities of recombinant retinoic acid receptors RARÂ«and RAR/1. Biochem. J.,272:391-397, 1990.

44. Clifford, J. L., Petkovich, M., ChambÃ³n, P., and Lotan, R. Modulation byretinoids of mRNA levels for nuclear retinoic acid receptors in murinemelanoma cells. J. Mol. Endo., 4: 1546-1555, 1990.

45. AstrÃ¶m,A., Pettersson, U., Krust, A., ChambÃ³n, P., and Voorhees, J. J.Retinoic acid and synthetic analogs differentially activate retinoic acid receptor dependent transcription. Biochem. Biophys. Res. Commun., 173: 339-345, 1990.

46. Huang, M. E., Ye, Y. I., Chen, S. R., Chai, J. R., Lu, J-x., Zhoa, L., Gu, L-j., and Whang, Z-y. Use of all-frans-retinoic acid in the treatment of acutepromyoelocytic leukemia. Blood, 72: 567-571, 1988.

4809



1991;51:4804-4809. Cancer Res Jürgen M. Lehmann, Marcia I. Dawson, Peter D. Hobbs, et al. Subtype-selective ActivitiesIdentification of Retinoids with Nuclear Receptor

Updated version

http://cancerres.aacrjournals.org/content/51/18/4804

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/51/18/4804To request permission to re-use all or part of this article, use this link


http://cancerres.aacrjournals.org/content/51/18/4804http://cancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://cancerres.aacrjournals.org/content/51/18/4804http://cancerres.aacrjournals.org/

Documents

Identification of Retinoids with Nuclear Receptor Subtype-selective Activities1 · radical bromination [A'-bromosuccinimide, (C6H5CO2)2,CC14,hu), dis placement of the benzylic bromide