Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

All-trans-retinoic acid (RA) is a naturally occurring retinoid responsible for growth and differentiation of epithelial tissues in mammals (Sporn et al., 1994). RA exerts its activity through binding with transcription-regulatory factors, known as the retinoic acid receptors, RAR-, -, and -. On interaction with RA, these RARs bind to specific DNA sequences and either repress or activate the transcription of adjacent target genes that encode for a variety of proteins, enzymes, and membrane receptors (Giguère, 1994; Chambon, 1996). Given the multitude of RAR-responsive genes, RA elicits some profound effects on epithelial growth and differentiation, such as induction of epidermal hyperplasia in mice and humans (Plewig and Braun-Falco, 1975; Connor, 1986), suppression of keratinization in the stratified squamous epithelium of rats (Geiger and Weiser, 1989), reduction of the utriculus size in rhino mice (Mezick et al., 1984), attenuation of skin wrinkling and hyperpigmentation in humans (Weiss et al., 1988; Rafal et al., 1992), and normalization of desquamation of the follicular epithelium and reduction of comedone formation in acne patients (Orfanos et al., 1997). Because of these pleiotropic actions, it is not surprising that plasma and tissue levels of RA are kept under tight homeostatic control. Contributing to this homeostatic control is the inactivation of RA through oxidative metabolism of the retinoid, which is initiated by the 4-hydroxylation of RA to form 4-hydroxy-RA and 4-keto-RA (Roos et al., 1998). The 4-hydroxylation of RA is carried out by the microsomal cytochrome P450 (CYP) isozyme system. Several CYP isozymes have been shown to be capable of metabolizing RA via this reaction (Leo et al., 1984, 1989; Roberts et al., 1992; Martini and Murray, 1993; Raner et al., 1996), but CYP26 appears to be the most dedicated RA 4-hydroxylase by far (White et al., 1996, 1997). CYP26 recognizes only RA as its substrate, and the expression and/or activity of this isozyme can be induced by RA both in vitro and in vivo (Ray et al., 1997; Abu-Abed et al., 1998; Sonneveld et al., 1998).

Irrespective of the CYP isozyme(s) involved, inappropriate metabolism of RA could generate a condition of retinoid deficiency, which is characterized by hyperkeratinization and desquamation as seen in acne, psoriasis, and ichthyosis (Orfanos et al., 1997). Reasoning that pharmacological modulation of RA metabolism would represent a viable approach to treat cutaneous disorders, we have been searching for inhibitors of the CYP-mediated metabolic breakdown of RA. Several years ago, this search resulted in the identification of liarozole, an imidazole derivate capable of inhibiting the CYP-dependent metabolism of RA by hamster liver microsomes (Van Wauwe et al., 1990). Furthermore, in rats, the compound enhanced plasma RA levels and exerted RA-mimetic effects, exemplified by its inhibition of vaginal keratinization (Van Wauwe et al., 1992). In humans, liarozole reduced the plasma elimination of exogenously administered RA (Miller et al., 1994). In open clinical studies, liarozole was found to be therapeutically effective in patients with psoriasis (Dockx et al., 1995; Van Pelt et al., 1998) and with ichthyosis (Lucker et al., 1997). However, liarozole lacked CYP isozyme specificity and also inhibited the CYP-mediated biosynthesis of adrenal and gonadal steroid hormones (Vanden Bossche, 1992).

This study reports on the pharmacological characterization of R115866 as a potent and selective inhibitor of the CYP26-mediated metabolism of RA. Oral treatment of rats with R115866 induced transient increases of plasma and tissue RA levels and generated retinoidal effects, such as suppression of vaginal keratinization, induction of pinnal hyperplasia, conversion of caudal para- to orthokeratosis, and up-regulation of hepatic CYP26 mRNA expression. Additionally, the keratinization-reducing and CYP26 mRNA up-regulating activities of R115866 could be reversed by concomitant treatment with the prototypical RAR antagonist, AGN193109 (Johnson et al., 1995). The chemical structure of R115866, or (B)-N-[4-[2-ethyl-1-(1H-1,2,4-triazol(-1-yl)butyl]phenyl]-2-benzothiazolamine, is presented in Fig. 1.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Chemical structure of R115866.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Wistar rats of both sexes (160-220 g) and female skh:HR1 hairless mice (24-30 g) were purchased from Charles River (Sulzfeld, Germany). Female rats were ovariectomized as described previously (Sietsema and DeLuca, 1982).

Chemicals. [11,12-³H]RA (49.6 Ci/mmol) was obtained from New England Nuclear (Boston, MA). Unlabeled RA was obtained from Sigma Chemical Co. (Bornem, Belgium). R115866, or (B)-N-[4-[2-ethyl-1-(1H-1,2,4-triazol-1-yl)butyl]phenyl]-2-benzothiazolamine; liarozole; and the RAR antagonist, AGN193109, were prepared at the Chemical Synthesis Department, Janssen Research Foundation, Val-de-Reuil, France. For in vitro experiments, stock solutions of R115866 (10 mM) were prepared in dimethyl sulfoxide (DMSO) and then appropriately diluted with assay buffer. Control samples contained an equal amount of DMSO (0.1%). For in vivo experiments, drugs were dissolved in polyethylene glycol (PEG) 200 (vehicle) and orally administered to animals in a volume of 0.5 ml/100 g b.wt. (rats) or 0.1 ml/25 g b.wt. (mice). All other chemicals or solvents were of analytical grade or HPLC grade when required. Protein concentrations were determined using the Bio-Rad protein assay (Bio-Rad, Rockville Center, NY) with BSA as standard.

Laboratory Precautions. All laboratory manipulations involving RA were carried out in a darkened room under dim yellow light to avoid isomerization of the retinoid.

In Vitro Metabolism of RA. Human CYP26 was expressed in yeast cells using standard procedures (Sanglard et al., 1996). Briefly, the coding sequence of human CYP26 [identified by homology search of the Incyte Lifeseq EST database with the zebra fish CYP26 homolog (White et al., 1996) as template] was inserted into the yeast PMA91 expression vector (Biomedical Research Center, Dundee, Scotland). This vector was then introduced into the Saccharomyces cerevisiae strain INVSc1 (Invitrogen, Groningen, the Netherlands), engineered to overexpress the human NADPH-CYP reductase (Biomedical Research Center, Dundee, Scotland) when grown on glucose as a carbon source. Transformed cells were cultured at 30°C in selection medium containing 6.7 mg/ml yeast nitrogen base without amino acids, 20 mg/ml glucose, 20 µg/ml L-histidine, and 20 µg/ml uracil. The cells were collected during the late exponential phase, rinsed, and homogenized in a 10-fold volume (w/v) of 0.1 M potassium phosphate/0.65 M sorbitol (pH 7.4) containing a mixture of the protease inhibitors: phenylmethylsulfonyl fluoride (40 µg/ml), 4-amidinophenylmethane-sulfonyl fluoride (1 µg/ml), and leupeptin (1 µg/ml). Microsomes were isolated by differential centrifugation (2,000g, 5 min; 15,000g, 15 min; 105,000g, 90 min). The microsomal pellet, suspended in 0.5 mM Tris-HCl buffer (pH 7.4) containing 1 mM EDTA and 20% glycerol, was kept at 70°C.

In Vitro CYP Selectivity. The conversion of androstendione to estrone by aromatase (CYP19) in human placental microsomes, the conversion of 17-hydroxy-20-dihydroprogesterone to testosterone by 17,20-lyase (CYP17) in rat testicular S10 fractions, and the 2-,6-, and 7-hydroxylation of testosterone (CYP2C11, CYP3A, CYP2A1) in rat liver microsomes were carried out as described previously (Vanden Bossche et al., 1990).

In Vivo RA Metabolism. Rats were treated orally with R115866 (2.5 mg/kg) or vehicle (PEG 200) and sacrificed 2, 4, 6, 8, and 18 h later. Blood was collected in heparin (50 U/ml), and plasma was obtained by centrifugation. Samples of clipped belly skin, liver, epididymal fat, kidney, and spleen were recovered. Plasma and tissues were immediately stored at 70°C and extracted within 1 week. Before homogenization, skin samples (stripped of muscular and adipose layers) were frozen in liquid nitrogen and ground into a fine powder with a Mikro-Dismembrator S (Braun Biotech International, Meslungen, Germany). RA was extracted and quantified by UV absorbance after HPLC separation as described previously (Van Wauwe et al., 1992, 1994).

Vaginal Keratinization. Ovariectomized rats were injected with estradiol undecylate to induce vaginal keratinization (Van Wauwe et al., 1992). On days 1, 2, and 3 after the estrogenic stimulation, animals were treated orally once daily with vehicle (PEG 200), R115866 (0.04-10 mg/kg), or RA (0.6-20 mg/kg). In some experiments, the RAR antagonist, AGN193109 (5 mg/kg), was administered together with R115866 (5 mg/kg) or RA (20 mg/kg). One day after the last treatment, rats were decapitated. Their vaginae were excised and processed for light microscopic examination (Van Wauwe et al., 1992). The intensity of vaginal keratinization was scored in a blinded fashion according to the following arbitrary scale: grade ++ = presence of keratinized squamae along the entire vaginal epithelium; grade + = presence of keratinized squamae along part of the vaginal epithelium; grade 0 = absence of keratinized squamae.

Induction of Ear Epidermal Hyperplasia. Mice were treated orally once daily for 14 days with vehicle (PEG 200), R115866 (0.3-2.5 mg/kg), or RA (2.5 mg/kg). One day after the last treatment, the animals were sacrificed. Four-millimeter-diameter ear punches (one sample per animal) were fixed (overnight at 4°C) in 2% paraformaldehyde/2.5% glutaraldehyde/0.1 M phosphate buffer (pH 7.4), postfixed (1 h) in 2% osmiumtetroxide/0.1 M phosphate buffer (pH 7.4), dehydrated in ethanol, and embedded in LC-112 epoxy resin (Ladd Research Industries, Burlington, VT). Sections (2 µm thick, one per fixed sample) were cut and stained with toluidine blue. Thickness and the number of viable epidermal cell layers were determined in five randomly selected interfollicular regions per section.

Induction of Caudal Para- to Orthokeratotic Transformation. Mice were treated orally once daily for 14 days with vehicle (PEG 200), R115866 (0.3-2.5 mg/kg), or RA (2.5 mg/kg). One day after the last administration, the animals were sacrificed. Tail skin samples were fixed (overnight at 4°C) in 4% paraformaldehyde/0.1% glutaraldehyde/0.1 M phosphate buffer (pH 7.4), dehydrated in ethanol, and embedded in Unicryl resin (British BioCell International, Cardiff, UK). Sections (2 µm thick, one per fixed sample) were cut and stained with hematoxylin-eosin. Para- to orthokeratotic transformation was quantified by counting the number of orthokeratotic scale regions (i.e., parts of the tail skin displaying a clearly visible continuous granular layer and orthokeratotic horny layer) and expressing this value as a percentage of the total number of scale regions.

Induction of CYP26 mRNA Expression. Female rats were treated orally once daily for 4 days with vehicle (PEG 200), R115866 (0.04-2.5 mg/kg), or RA (0.04-2.5 mg/kg). In additional experiments, the RAR antagonist, AGN193109 (2.5 mg/kg), was coadministered orally with 2.5 mg/kg of R115866 or RA. Four hours after the last treatment, rats were sacrificed, and liver samples (80-120 mg) were removed and frozen immediately in liquid nitrogen. Total RNA was isolated using the SV Total RNA Isolation System following the instructions of the manufacturer (Promega, Madison, WI). RNA content was determined by absorbance measurement at 260 nm. Analysis for the presence of CYP26 mRNA was carried out using the Titan One Tube RT-PCR (reverse transcription-polymerase chain reaction) System (Roche Molecular Biochemicals, Brussels, Belgium). Briefly, 200 ng of RNA was reverse-transcribed and amplified in 50 µl RT-PCR buffer (20 mM Tris-HCl, pH 7.5; 100 mM KCl; 1.5 mM MgCl₂) containing 0.2 mM dNTP, 5 mM DDT, 8 U of RNase inhibitor, enzyme mix (AMV reverse transcription and Expand High Fidelity enzymes; Roche Molecular Biochemicals), and 0.4 µM 5'- and 3'-specific primers. To account for quantitative and/or qualitative differences in the RNA preparation, primers for the rat ₂-microglobulin gene were used in each experiment. Samples were transferred to a thermocycler (PCT-200 Peltier Thermal Cycler; MJ Research, Watertown, MA), incubated at 50°C for 30 min, and subjected to a denaturation step (94°C, 2 min) and to 27 cycles consisting of 30 s at 94°C, 1 min at 55°C, and 30 s at 72°C, followed by a final elongation step of 4 min at 72°C. After amplification, PCR products were subjected to electrophoresis on a 2% agarose gel, visualized by UV light illumination after ethidium bromide staining, and quantified using the Lumi-Imagen F1 Workstation (Roche Molecular Biochemicals). Specific primers were obtained from Eurogentec (Seraing, Belgium). Their sequences were as follows: CYP26: sense, ACGCACTGCAGCTCTTGATTG and antisense, CATGTCTAACTTGTCCTCGTG; ₂-microglobulin: sense, CGTCGTGCTTGCCATTCAGA and antisense, GGTGTGAATTCAGTGTGAGC.

Data Analysis. Data are expressed as means ± S.E. IC₅₀ values were calculated by probit analysis. Statistical significance between parallel treatment groups was assessed by applying a two-sided Student's t test, two-sided Dunnett's test, or Wilcoxon-Mann-Whitney rank-sum test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of R115866 on In Vitro Metabolism of RA. [11,12-³H]RA was incubated for 30 min at 37°C in the presence of NADPH and yeast microsomes expressing human CYP26. The formation of radioactive reaction products was analyzed by reverse-phase HPLC. As shown in Fig. 2A, RA was converted into several radioactive products having high (retention time, 3-10 min) or medium (retention time, 15-25 min) polarity. No attempts were made to identify the structure of these products, but the peaks with retention times of 17.5 and 19.5 min eluted at the same position as authentic 4-hydroxy-RA and 4-keto-RA, respectively. Coincubation of the microsomes with R115866 (1 µM) resulted in an almost complete inhibition of the RA conversion (Fig. 2B). Dose-response experiments indicated that R115866 inhibited the RA conversion with an IC₅₀ value of 4 nM. For comparison, liarozole suppressed the reaction with an IC₅₀ value of 3 µM. To define the selectivity of R115866 toward CYP26, we tested whether the compound affected the biosynthesis of estradiol by CYP19 (aromatase) and 17-hydroxy-20-dihydroprogesterone by CYP17 (17,20-lyase), and the 2-, 6-, and 7-hydroxylations of testosterone by CYP2C11, CYP3A, and CYP2A1, respectively. As shown in Table 1, micromolar concentrations of R115866 were needed to inhibit these CYP-mediated reactions.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Reverse-phase HPLC analysis of radiolabeled metabolites formed by incubation of [11,12-3H]RA (20 pmol) with yeast microsomes expressing human CYP26. Microsomes (2 mg of protein) were incubated (30 min at 37°C) with [11,12-3H]RA (1 µCi) and NADPH in the presence of 0.1% DMSO (A) or 1 µM R115866 (B). Separation was done on a Zorbax 5C8 column eluted with a mixture of methanol/2% acetic acid/acetonitrile at a flow rate of 1 ml/min. At 40 min, unreacted RA was eluted by changing the solvent to 100% methanol. The elution positions of 4-hydroxy-RA and 4-keto-RA are indicated.

Effects of R115866 on Plasma and Tissue Levels of RA. To determine the ability of R115866 to affect the in vivo turnover of RA, rats were treated orally with vehicle or R115866 (2.5 mg/kg). The animals were sacrificed 2, 4, 6, 8, and 18 h later, and the RA content in plasma, skin, liver, fat, kidney, and spleen was measured. The results are shown in Fig. 3. Mean RA levels (expressed as nanogram per milliliter of plasma or nanogram per gram of tissue) in vehicle-treated animals varied between 0.2 and 0.6 (plasma), 1.3 and 1.9 (skin), 3.5 and 5.1 (liver), 1.4 and 1.7 (fat), 0.5 and 1.0 (kidney), and 0.5 and 0.7 (spleen). From 2 to 8 h after dosing, R115866 enhanced these values significantly (P < .01) to levels of 0.8 to 1.6 (plasma), 3.0 to 7.8 (skin), 2.6 to 3.7 (fat), 1.3 to 1.8 (kidney), and 1.1 to 1.8 (spleen). However, in the liver, R115866 elicited a significant (P < .05) RA increase only at the 6 h time point, raising the retinoid concentration from 4.3 ± 0.6 to 6.2 ± 0.5 ng/g of organ. In all tissues, RA levels returned to basal levels at 18 h after R115866 administration. The ability of R115866 to increase systemic levels prompted us to examine the compound for potential retinoidal activities. For this we compared the oral effects of R115866 and RA in a number of retinoid-responsive rodent models.

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Endogenous RA content in plasma and tissues at the indicated times after oral administration of vehicle () or R115866 (2.5 mg/kg) (). Each value represents the mean ± S.E. using six to ten rats per treatment group. **P < .01, *P < .05 versus corresponding vehicle group.

Retinoidal Effects of R115866 on Vaginal Keratinization. Estrogen-exposed ovariectomized rats were treated once daily for 3 days with vehicle, R115866 (0.04-10 mg/kg), or RA (0.6-20 mg/kg). On the fourth day, the animals were sacrificed, and their vaginae were histologically examined for signs of keratinization. As shown in Table 2, both compounds dose dependently inhibited the formation of keratinized squamae with ED₅₀ values of 1.0 (0.6-1.9) mg/kg and 5.1 (3.5-7.6) mg/kg for R115866 and RA, respectively.

Retinoidal Effects of R115866 on Pinnal Epidermis. Hairless mice were treated once daily for 14 days with vehicle, R115866 (0.3-2.5 mg/kg), or RA (2.5 mg/kg). One day after the last treatment, the hyperplastic response of the ear epidermis was assessed by histology. As shown in Fig. 4A, the pinnal epidermis in vehicle-treated animals consisted of three to five viable cell layers with a thickness of 16.5 ± 1.5 µm. Treatment with R115866 (2.5 mg/kg) or RA (2.5 mg/kg) induced a strong hyperplasia (Fig. 4, B and C), resulting in the formation of a six- to seven-cell-layered epidermis with a height of 35.5 ± 2.1 and 31.8 ± 2.1 µm, respectively. As shown in Fig. 5, the ear epidermis was still significantly (P < .05) thickened after dosing with 0.3 to 0.6 mg/kg of R115866.

View larger version (98K): [in this window] [in a new window]   Fig. 4.   Morphological appearance of mouse pinnal epidermis after repetitive (once daily for 14 days) oral treatment with vehicle (A), R115866 (2.5 mg/kg) (B), or RA (2.5 mg/kg) (C). The viable epidermis is denoted by a double arrow.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Thickness of the pinnal epidermis from hairless mice treated orally once daily for 14 days with vehicle or indicated doses of R115866 or RA. Each value represents the mean ± S.E. using eight mice per treatment group. *P < .05 versus corresponding vehicle group.

Retinoidal Effects of R115866 on Caudal Epidermis. As above, hairless mice were treated once daily for 14 days with R115866 (0.3-2.5 mg/kg) or RA (2.5 mg/kg). On day 15, the level of orthokeratotic scale type in the tail epidermis was then assessed by histology. In vehicle-treated animals (Fig. 6A), the tail epidermis displayed a typical differentiation pattern, characterized by a succession of parakeratotic scale (lacking a granular cell layer) and orthokeratotic (having a granular layer) interscale regions. As shown in Fig. 6, B and C, treatment with R115866 (2.5 mg/kg) and RA (2.5 mg/kg) transformed the caudal epidermis into a continuous orthokeratic skin type. Figure 7 shows that the percentage of tail orthokeratosis increased from 30.3 ± 4.0% (vehicle-treated) to more than 90% after dosing with R115866 (1.25-2.5 mg/kg) or RA (2.5 mg/kg). At 0.6 mg/kg, R115866 still significantly (P < .01) enhanced the fraction of orthokeratotic scale regions.

View larger version (104K): [in this window] [in a new window]   Fig. 6.   Morphological appearance of mouse tail epidermis after repeated (once daily for 14 days) oral administration of vehicle (A), R115866 (2.5 mg/kg) (B), or RA (2.5 mg/kg) (C). Ortho, orthokeratotic interscale region; para, parakeratotic scale region.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Percentage of orthokeratotic scale regions in tail epidermis from hairless mice treated once daily for 14 days with vehicle or indicated dose of R115866 or RA. Each value represents the mean ± S.E. using eight mice per treatment group. **P < .01 versus corresponding vehicle group.

Retinoidal Effects of R115866 on CYP26 mRNA Expression. Rats were treated once daily for 4 days with vehicle, R115866 (0.04-2.5 mg/kg), or RA (0.04-2.5 mg/kg). Three hours after the last treatment, liver RNA was extracted, and RT-PCR was used to obtain a qualitative assessment of the CYP26 gene expression. The RT-PCR data are given in Fig. 8. Both compounds dose dependently induced an increase of the hepatic CYP26 mRNA expression. RA was the stronger inducer: at 0.6 and 2.5 mg/kg, it elicited a 1.8-fold higher CYP26 mRNA expression than R115866. Even at 0.16 mg/kg, RA induced a clear elevation of CYP26 transcripts (~10 times higher than the vehicle level), whereas R115866 proved inactive.

View larger version (37K): [in this window] [in a new window]   Fig. 8.   RT-PCR analysis of CYP26 mRNA induction. RNA was extracted from livers of rats treated once daily for 4 days with vehicle or indicated doses of R115866 or RA. After RT-PCR, amplified products were analyzed by agarose gel electrophoresis and ethidium bromide staining. Representative data from one of two independent experiments are shown.

Retinoidal Effects of R115866: Reversal by RAR Antagonism. The ability of R115866 to raise systemic levels of RA, together with its retinoidal actions, pointed to a mediating role of endogenous RA. To substantiate such a scenario, we compared the level of vaginal keratinization in estrogen-stimulated ovariectomized rats after treatment with R115866 (5 mg/kg) or RA (20 mg/kg) alone, or combined with the RAR antagonist, AGN193109 (5 mg/kg). As shown in Table 3, whereas administration of R115866 or RA resulted in a complete antikeratinizing effect in all animals tested, concomitant treatment with AGN193109 markedly reduced their efficacy, resulting in keratinization grades of + and ++ in the combination-treated animals. When dosed singly, AGN193109 had no effect on vaginal keratinization. Additionally, we determined the hepatic CYP26 mRNA expression in rats treated with R115866 (2.5 mg/kg) or RA (0.6 mg/kg) alone or together with AGN193109 (2.5 mg/kg). The RT-PCR results in Fig. 9 indicate that combined administration with the RAR antagonist markedly reduced the induction of CYP26 transcripts by both R115866 and RA. To exclude pharmacokinetic artifacts, we verified that AGN193109 cotreatment had not attenuated the uptake of R115866. Rats were therefore treated p.o. with R115866 (2.5 mg/kg) alone or in combination with AGN193109 (2.5 mg/kg), and at various times after treatment, plasma levels of R115866 were determined. Plasma levels (expressed as nanograms per milliliter, n = 3 animals/treatment group/time point) were 14.8 ± 4.6 (0.5 h), 61.4 ± 1.0 (1 h), 67.7 ± 15.6 (3 h), and 20.9 ± 9.3 (8 h) after R115866 alone treatment and were highly comparable with the corresponding values of 23.1 ± 6.8 (0.5 h), 67.7 ± 15.6 (1 h), 46.9 ± 20.4 (3 h), and 18.9 ± 2.7 (8 h) after the combined R115866-AGN193109 administration.

View larger version (46K): [in this window] [in a new window]   Fig. 9.   RT-PCR analysis of CYP26 mRNA expression. RNA was extracted from livers of rats treated once daily for 4 days with vehicle, R115866, and RA and AGN193109 either alone or combined as indicated. RNA extracts were subjected to RT-PCR, and amplified products were analyzed by agarose gel electrophoresis with ethidium bromide staining. Representative data from one of two separate experiments are shown.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The first part of this study characterizes R115866 as a potent and selective inhibitor of the in vitro CYP26-mediated metabolism of RA. In several aspects, the triazole-containing R115866 can be distinguished from the imidazole derivate, liarozole, a first generation RA metabolism inhibitor (Van Wauwe et al., 1992). With regard to potency, R115866 is a nanomolar (IC₅₀ = 4 nM) inhibitor of the CYP26-dependent RA conversion and about three orders of magnitude more powerful than liarozole (IC₅₀ = 3 µM). As for its selectivity, R115866 shows trivial inhibitory effects on the CYP-dependent formation of testosterone and estradiol, whereas liarozole suppressed the biosynthesis of these gonadal hormones more potently (IC₅₀ = 0.2-0.3 µM) (Vanden Bossche et al., 1990) than the RA conversion by CYP26. In line with these in vitro observations, a single p.o. administration of liarozole (2.5 mg/kg) to dogs reduced plasma testosterone concentrations to castration levels (Vanden Bossche et al., 1990), whereas treatment with R115866 (10 mg/kg) had no effect on blood levels of this hormone (data not shown). Taken together, R115866 behaves as a selective CYP26 inhibitor and should be considered less likely to produce unwanted side effects.

The second part of this study presents in vivo evidence that R115866 is able not only to enhance endogenous RA levels but also to mimic the effects of RA. In rats, a single p.o. dosing of R115866 (2.5 mg/kg) stimulated a marked and protracted (from 2 to 8 h) surge of endogenous RA levels in plasma, skin, fat, kidney, and spleen. In all probability, R115866 derives this activity from interference with the turnover rate of endogenous RA. In rats, RA turnover is estimated to be fast, with body pool replacement taking place once every 10 h (Kurlandsky et al., 1995). Thus, in tissues where RA is broken down via the CYP pathway, R115866 would delay RA metabolism and increase local RA levels. To reduce the potentially harmful RA accumulation, tissues could commit other CYP-independent pathways of RA breakdown and/or export the RA surplus into the bloodstream for redistribution to other tissues. This chain of events is compatible with the appearance of RA in the plasma and explains the RA elevations seen in parts of the body, such as fat tissue, where RA-metabolizing enzymes are not expressed.

Paradoxically, in liver, the organ considered to be a major site of RA formation and CYP-dependent metabolism, R115866 induced only a modest and momentary (at the 6-h time point) RA enhancement. Perhaps, exposed to rather high basal levels of RA (3-5 ng/g of tissue), rat liver may be under a particularly strict control, which readily neutralizes any soaring of the retinoid RA. The ability of R115866 to exert oral retinoidal activities is documented in four RA-responsive rodent models: inhibition of vaginal keratinization, induction of pinnal hyperplasia, transformation of caudal para- to orthokeratosis, and induction of hepatic CYP26 mRNA expression. The first model uses estrogen-treated ovariectomized rats, whose vaginal epithelium transforms into a stratified keratinizing epithelium that expresses RAR-, RAR-, and RXR- (Boehm et al., 1997). Through interaction with these receptors, retinoids inhibit vaginal keratinization; and, in fact, this effect on vaginal differentiation is considered to be an obligatory property for a compound to qualify as a retinoidal drug (Geiger and Weiser, 1989; Chateau et al., 1996). Based on data obtained with liarozole (Van Wauwe et al., 1992), the finding that R115866 was able to suppress the vaginal keratinization process was expected, but its 5-fold higher potency over RA was surprising. Indeed, whereas R115866 (2.5 mg/kg) barely tripled vaginal RA concentration from 0.8 to 2.0 (vehicle) to 3.0 to 5.3 ng/200 mg of tissue (data not shown), treatment with RA (20 mg/kg) increased these values to 80 to 200 ng of RA/200 mg of vaginal tissue (Van Wauwe et al., 1992). Obviously, through its intracellular site of action, R115866 allows RA to linger inside cells, increasing its potential to move into the nucleus and modulate gene expression. By contrast, exogenously administered RA would largely remain in the extracellular space, unable to produce a biological response. The second model is based on the observations that topically applied retinoids induce a strong hyperplastic response in mouse ear epidermis (Connor, 1986). In hairless mice, the pinnal epidermis consists of three to five viable cell layers (known as basal, prickle, and granular cell layers), which are covered by about six layers of cornified cells. After retinoid treatment, the epidermis becomes hyperplastic, typically as a result of the increased number of prickle and granular cell layers. Oral treatment of hairless mice with R115866 induced hyperplasia of the ear epidermis that was morphologically indiscernible from the effect generated by RA treatment. The third model relies on the particular sort of metaplasia that retinoids can induce in mouse tail epidermis (Schweizer et al., 1987). In adult mice, the tail epidermis displays a regular succession of parakeratotic scale and orthokeratotic interscale regions. The parakeratotic scale epidermis lacks a granular cell layer and is covered by a thick cornified cell layer, whereas the orthokeratotic scale epidermis displays a granular cell layer and is characterized by a loose layer of cornified cells. Repetitive topical treatment with retinoids results in a complete para- to orthokeratotic conversion of the scale regions. Here, we showed that oral dosing of hairless mice with R115866 induced a similar orthokeratotic transformation of the tail. The fourth model relies on findings that RA is a strong inducer of CYP26 mRNA expression and/or enzyme activity. In vitro, RA exposure of several human and murine tumor cell lines results in CYP26 expression, and this phenomenon is dependent on the presence of RAR- and RAR- (White et al., 1997; Abu-Abed et al., 1998; Sonneveld et al., 1998). Moreover, in vivo, acute administration of RA increased steady-state levels of CYP26 mRNA in adult mouse liver. As demonstrated here, four once-daily administrations of RA also enhanced the formation of CYP26 transcripts in rat liver. Subchronic dosing with R115866 generated the same hepatic response, although the compound turned out to be approximately twice less potent than RA. The weaker CYP26 mRNA induction by R115866 may represent an advantage. Because CYP26 mRNA levels and RA metabolic activity are directly related (White et al., 1997; Abu-Abed et al., 1998), strong enhancement of CYP26 transcription should be considered counterproductive. Indeed, inducible RA metabolic activity has been implicated in the relapse and RA resistance seen during treatment of patients with acute promyelocytic leukemia (Muindi et al., 1992). Thus, given the relatively modest CYP26-inductive effect of R115866 combined with its strong (5-fold more potent than RA) activity to inhibit vaginal keratinization, one can argue that R115866 may be a more efficient therapeutic agent than RA. The last part of this study shows that the antikeratinizing and CYP26 mRNA-inducing activities of both R115866 and RA are reversed by the prototypical RAR antagonist, AGN193109. This compound is a nanomolar-affinity blocker of RA-induced function at each of the three RA receptor subtypes, RAR-, -, and - (Johnson et al., 1995), and has been shown to be topically active, capable of abrogating the hyperplasia of the dorsal epidermis induced in mice by RA (Thacher et al., 1997). As shown here, AGN193109 possesses potent oral activity as well. In rats, for instance, CYP26 mRNA induction in rat liver by treatment with RA (4 × 0.6 mg/kg) or R115866 (4 × 2.5 mg/kg) could be suppressed by equal dosing with AGN193109. Taken together, these data support the concept that the in vivo activities of R115866 are channeled through enhancement of endogenous RA levels and subsequent triggering of nuclear RA receptors.

The beneficial effects of topical and oral retinoids in the treatment of hyperkeratotic skin disorders, such as psoriasis and acne, are well established (Orfanos et al., 1997). Growing insight into the metabolic processing of RA has led to the development of compounds that suppress the CYP-dependent RA metabolism and, as a consequence, generate RA action from within. Liarozole, a first generation inhibitor of RA metabolism (Van Wauwe et al., 1992), has been demonstrated to be an effective antipsoriatic drug (Dockx et al., 1995; Van Pelt et al., 1998). The compound described in this study may provide clinicians with a more potent and specific drug for dermatological therapy.