TOXICOLOGY

AND

APPLIED

PHARMACOLOGY

112, 144- 153 ( 1992)

Structure-Affinity Relationships of Retinoids with Embryonic Cellular Retinoic Acid-Binding Protein’ CALVIN

C. WILLHITE,*

ANDRZEJ JUREK,~ RAGHUBIR P. SHARMA,~,~ AND MARCIA I. DAWSON$

*Department of Toxic Substances Control, California Environmental Protection Agency, Berkeley, California 94710; t Toxicology Program, Department of Animal, Dairy & Veterinary Sciences, Utah State University, Logan, Utah 84322-5600; and $Bio-Organic Chemistry Laboratory, SRI International, Menlo Park, California 94025 Received April 29, 1991; accepted August 29, 199 1

Structure-Affinity Relationships of Retinoids with Embryonic Cellular Retinoic Acid-Binding Protein. WILLHITE, C. C., JUREK, A., SHARMA, R. P., AND DAWSON, M. I. (1992). Toxicol. Appl. Pharmacol. 112,144-153. Separation and quantitation of cellular retinoic acid-binding protein (CRABP) in embryonic and fetal hamster tissues was accomplished with high-performance size-exclusion chromatography. Binding affinity of 26 retinoids was established by in vitro displacement of high specific activity all-trans-[3H,]retinoic acid from fetal CRABP. The CRABP concentration in presomite-to-early somite (Day 8) hamster embryos was 1.9 pmol/mg cytosolic protein and increased to 7.5 pmol/mg protein in Day 13 fetuses; CRABP concentrations subsequently declined as gestation progressed. CRABP waslocated primarily in fetal brain and skin (5.8 + 0.3 and 2.2 + 0.1 pmol/mg protein, respectively), whereas only trace concentrations were found in fetal liver, placenta, and maternal uterus. Retinoids that could displace alltrans-retinoic acid from CRABP had a free acid at the polar terminus (or were carboxylate esters that were readily hydrolyzed to the corresponding free acid) and had a hydrophobic ring at the distal position. The ligand specificity of the CRABP studied here suggests that this protein was analogous to the CRABP I isoform. The in vitro binding affinities of teratogenic retinoids that competed for embryonic CRABP failed to correlate directly with relative teratogenic potency. In some instances, the latter observation can be related to extensive in vivo biotransformation of retinoids to multiple teratogenic metabolites and to retinoid persistence in the embryo. Three analogs containing a free carboxy terminus, SRI 589821, SRI 7323-78, and SRI 6153-40, were identified with high teratogenic potency but failed to bind fetal hamster CRABP. The structure-activity and binding data of the analogs studied here indicate that many, if not most, teratogenic retinoids (or their acidic metabolites) bind with embryonic/fetal CRABP, but the present data question the role for CRABP in their teratogenic mechanism of action. 0 1992 Academic Press, Inc.

I This paper was presented in part at the 1989 Annual Meeting of the American Society for Pharmacology & Experimental Therapeutics (ASPET) (Pharmacologist 31, 139). * To whom correspondence should be addressed. 0041-008X/92 $3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

144

It is now widely accepted that families of binding proteins and nuclear receptors participate in the diverse pharrnacologic activities of the retinoids (Blomhoff et al., 1990; Sani, 1990). Considerable attention has been focused on the cellular retinoic acid-binding proteins (CRABP) and their role in control of cell differentiation and embryonic morphogenesis (Willhite et al., 1989; Holfrnann, 1990). Nevertheless, current evidence for a CRABP requirement in the biochemical mechanism of retinoid action has not been compelling (Sherman, 1985; Jetten, 1990). Certain acidic retinoids are potent teratogens in humans (Lammer et al., 1985; Howard and Willhite, 1986), nonhuman primates (Hummler et al., 1990), and rodents. Craniofacial terata are the hallmarks in humans, macaques, and hamsters. Retinoid teratogenic potency spans five orders of magnitude in hamsters, and even relatively minor chemical substitution can have pronounced influences on retinoid pharmacokinetics, metabolic fate, toxicity, and efficacy (Willhite, 1990). The purpose of the present investigation was to examine the affinity of 26 analogs of retinoic acid (RA) for hamster embryonic/fetal CRABP. The results are then discussed with respect to retinoid teratogenic potency, the temporal and anatomic expression of embryonic/fetal CRABP, and retinoid localization in the hamster conceptus. MATERIALS

AND METHODS

The retinoids studied here (Table I) and their sources are as described previously (Willhite, 1990; Willhite and Dawson, 1990). Purity of all retinoids was >95% as measured by reversed-phase high-performance liquid chromatography (HPLC) (Willhite and Book, 1990). Radioactive all-trans-[ 11,12‘HJretinoic acid (a11-trans-[3H]RA, 52.5 Ci/mmol; 98.8%) was purchased from NEN Research Products (Boston, MA). Precautions against retinoid thermal and photochemical degradation were followed (Willhite and Book, 1990).

Adult male and female hamsters [Lak:LVG(SYR)] were purchased from Charles River Breeding Laboratories (Wilmington, MA); the animals were maintained and bred as noted previously (Willhite and Book, 1990). The day following the evening of breeding was considered Day I of gestation. Embryos (Gestation Day S), fetuses (Gestation Days 9- 15) or Day I4 maternal uterus and placenta were collected at laparotomy and homogenized (Biospec Products. Bartlesville, OK) or selected tissues were dissected free

STRUCTURE-AFFINITY

RELATIONSHIPS

WITH EMBRYONIC

145

CRABP

TABLE 1 Retinoid Nomenclature, Structure, CRABP Binding Affinity, and Teratogenicity Common Name/ Code No. all-nans-retinoic acid SRI 2712-24

all-nans-retinol

13-cis-4-0x0retinoic acid/R0 226595

N-(2-hydroxyethyl) 13-cisretinamide

all-transretinylidene methyl nitcone

hydroxenin

Ro 12-0995

Ro 12-7554

Smlctm

Name (E)-3,7-dimethyl-9(2,6,6-trimethyl- 1-cyclohexen- l-yl)-2,4,6,8nonatetraenoic acid (E)-9-(2-ethyl-6,6dimethyl-l-cyclohexen-lyl)-3,7-dimethyl-2,4,6,8nonatetraenoic acid (E)-3,7-dimethyl-9(2,6,6-trimethyl- l-cyclohexen- l-yl)-2,4,6,8nonatetraenal (E)-3,7-dimethyl-9(2,6,6-trimethyl-l-cyclohexen- l-yl)-2,4,6,8nonatetraenol (22,4E,6E,gE)-3,7-dimethyl-9-(2,6,6-u+ methyl-3-oxo- l-cyclohexen- l-yl)-2,4,6,8nonatetraenoic acid N-1-(2-hydroxyethyl) (2Z,4E,6E,8E)-3,7dimethyl-9-(2,6,6-tmethyl- 1-cyclohexen- lyl)-2,4,6,8-nonatetraenamide (Z)-a-[(E)-2,6-dimethyl8-(2,6,6-trimethyl- lcyclohexen- I-yl)1,3,5,7-octatetraenyll-Nmethyl&one (Z)-3,7-dimethyl-9(2,6,6-trimethyl-l-cyclohexen-1-yl)nona-2,4,6triene- 1,6-diol ethyl (E)-9-(2-chloro-3,6dimethyl-4-methoxyphenyl)-3,7-dimethyl2,4,6,8-nonateuaenoate ethyl (E)-9-(2,~dichloro4-methoxy-3-methylphenyl)-3,7-dimethyl2,4,6,8-nonatetraenoate

DCsoa (Ml 5.1 x

10-7

6.5 x 10-G

EDsob (pmol/kg) 35

20

NCC

g&FJHzoH Cl

NC

76

3.9 x 10-G

35

NC

Inactive

NC

130

NC

>510

6

1.1 x 10-4

+COOWs ‘I

NC

ND

Ct-43

and homogenized at 4°C in an equal volume (w/v) of 3[Nmorpholino]propanesulfonic acid buffer (Mops IO mM, IO ItIM KCI. 2 mM 2-mercaptoethanol, 1 mM EDTA, pH 7.5). Cytosol was isolated by centrifugation of the crude homogenate (4°C) at 10,OOOgfor 30 min, followed by lO5,OOOgcenttifugation of the supernatant for 60 min (Beckman SW 41 Ti). Cytosolic protein was diluted to 1.04 mg/ml with fresh Mops buffer. and protein was analyzed (Pierce Protein Reagent, Rockford, IL) using bovine serum albumin as the standard (Howard ef al., 1990). Dilute cytosol (480 ~1) and 10 ~1 100 nM all-lrans-[‘H]RA (2.63 PCi in 95% ethanol) were pipetted into incubation tubes. Each control tube received an additional 10 ~11ethanol. The solutions were mixed thoroughly. Com-

petition assayswith each of the analogs were carried out by addition of 10 ~195% ethanol containing 0. l- to 200-fold molar excessof unlabeled retinoid. After incubation at 4°C for I2 hr in the dark, 100 pi Mops buffer containing 2.5% charcoal/0.25% dextran was added and vortexed for 10 min. The mixture was centrifuged at 1000s for 10 min and the supematant passedthrough a 0.45 pm filter. High-performance size-exclusion chromatography of the resulting solution was carried out after Rainier and co-workers (1983). and the radioactivity in appropriate fractions quantified by liquid scintillation counting. Displacement of all-trun@H]RA from CRABP was calculated by comparing the radioactivity in corresponding CRABP fractions (eluting at the location of myoglobin) with that of the concurrent control containing

146

WILLHITE

ET AL.

TABLE 1-Continued

Common Name/ Code No.

Name

(E)-3,7-dimethyl-9-[5,5dimethyl-2-(l-methoxyethyl)-l-cyclopenten-lyl]-2,4,6,8-nonatetraenoic acid all-trans-5.6 2-[Q-3,7-dimethyl-9(2,2,6-trimethyl-5,6-oxaepoxy-Z retinylidenecyclohex-l-yl)-2,4,6,8cyclohexanedione nonatetraenylidenel- 1,3cyclohexanedione Ro 21-6667 (E)-8-(4-methoxy-2,3,6trimethylphenylthio)-3,7dimethyl-2,4,6-octatrienoic acid SRI 5898-21 (E)-6-[2-(2,6,6&methyll-cyclohexen-l-yl)ethenI-yl]-Znaphthalenecarboxylic acid (E)-7-(5,6,7,8-tetraRo 13-6307 hydro-5,5,8,8-tetramethyl-2-naphthalenyl)3-methyl-2,4,6-octauienoic acid Ro 13-2389 ethyl (E)-7-(5,6,7,8tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)3-methyl-2,4,6-octatrienoate Ro 13-4306 (E)-3-methyl-7-(1,1,3,3tetramethyl-5-indanyl)2,4,6-octatrienoic acid

DCs$ 0

smlcture

Ro lo-1770

2.4 x 10-T

NC

COOH

ND

Inactive

NC

38

3.0 x 10-T

1.95

6.4 x IO-6

2.3

1.7 x 10-6

14.5

co.11

(E)-5,6,7,8-tetrahydro-2[ l-(Cmethylphenyl)propen-2-yl]-5,5,8,8tetramethylnaphthalene

NC

Ro 13-8320

(E)-4-[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)l-propen-l-yllphenylmethanol ethyl (E)-4-[2-(5,6,7,8tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)1-propen-1-yllphenyl sulfone

NC

an identical volume of ethanol alone. Scatchard analyses were performed and a maximum likelihood estimate of the retinoid concentration required to displace 50% of the bound all-trans-[3H]RA was made (Howard et al., 1990). Where notable displacement of label was observed in the case of the retinoid ethyl esters (Ro 12-0955, Ro 13-6298, Ro 13-2389), either eserine sulfate or neostigmine bromide (3 X lo-’ M) was added to the incubation buffer; an identical series of studies with these esterase inhibitors with 200fold molar excess of unlabeled all-iruns-RA and aIl-truns-[3H]RA was included as an additional control.

0.03

NC

Ro 13-9272

Ro 15-1570

ED& (cLmol/kg)

NC

0.085

82

Retinoid teratogenic potency was derived based on previous investigations (Willhite, 1990; Willhite and Dawson, 1990). In those few cases where an insufficient quantity of drug was available for in vivo development toxicity bioassay, the retinoid teratogenic potential was estimated using published rat and mouse teratology data (Kistler, 1987) and established structure-activity relationships. Maximal amount of all-fmns[‘H]RA binding provided the basis for calculation of the concentration of CRABP in whole embryo or fetuses and selected tissues from 14-day fetuses.

STRUCTURE-AFFINITY

RELATIONSHIPS

WITH EMBRYONIC

CRABP

147

TABLE l-Continued

Common Name/ Code No.

Name

DCsoa CM>

structure

Ro 13-6298

ethyl (E)-4-[2-(5,6,7,8tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)1-propen-1-yllbenzoate

1.9 x IO-5

Ro 151550

(E)-4-[2-(1,1,3,3-tetramethylindan-5-yl)- 1-propen- 1-yl]benzoic acid

4.3 x 10-h

SRI 6409-94

N-(4-carbethoxyphenyl) 5,6,7,8-tetrahydro5,5,8,8-tetramethyl-2naphthalenecarboxamide

SRI 7323-78

(E)-6-[2-(4-carboxyphenyl)-1-propen-1-yl]-3,4dihydro-4,4dimethyl2/Z-1-benzothiopyran 6-(5,6,7,8-tetrahydro8,8-dimethyl-2-naphthalenyl)-2-naphthalenecarboxylic acid 6-(5,6,7,8-tetrahydro5,5,8,8-tetramethyl-2naphthalenyl)-2-naphthalenecarboxylic acid

SRI 6153-40

SRI 5898-7 1

d-6

DSOb (pmoVkg)