Comparative teratogenicity of three retinoids: the arotinoids Ro 13-7410, Ro 13-6298 and Ro 15-1570

A transgenic reporter mouse model for in vivo assessment of retinoic acid receptor transcriptional activation

Background: Vitamin A is essential for a wide range of life processes throughout embryogenesis to adult life. With the aim of developing an in vivo model to monitor retinoic acid receptor (RAR) transactivation real-time in intact animals, we generated transgenic mice carrying a luciferase (luc) reporter gene under the control of retinoic acid response elements (RAREs) consisting of three copies of a direct repeat with five spacing nucleotides (DR5). Methods: Transgenic mice carrying a RARE dependent luciferase reporter flanked with insulator sequence were generated by pronuclear injection. RARE dependent luciferase activity was detected by in vivo imaging or in tissue extracts following manipulations with RAR/retinoid X receptor (RXR) agonists, RAR antagonists or in vitamin A deficient mice. Results: We found a strong induction of luciferase activity in a time and dose dependent manner by retinoic acid as well as RAR agonists, but not by the RXR agonist (using n=4-6 per group; 94 mice). In addition, luciferase activity was strongly reduced in vitamin A-deficient mice (n=6-9; 30 mice). These observations confirm that luciferase activity was controlled by RAR activation in the RARE-luc mouse. Luciferase activity was detectable in various organs, with high activity especially in brain and testis, indicating strong retinoid signalling in these tissues. Conclusion: The RARE-luc transgenic mice, which enabled real-time in vivo assessment of RAR activation, will be useful in understanding the normal physiology of vitamin A, the role of retinoid signalling in pathologies as well as to evaluate pharmacological ligands for RARs.

Retinoic acid regulation by CYP26 in vertebrate lens regeneration

Xenopus laevis is among the few species that are capable of fully regenerating a lost lens de novo. This occurs upon removal of the lens, when secreted factors from the retina are permitted to reach the cornea epithelium and trigger it to form a new lens. Although many studies have investigated the retinal factors that initiate lens regeneration, relatively little is known about what factors support this process and make the cornea competent to form a lens. We presently investigate the role of Retinoic acid (RA) signaling in lens regeneration in Xenopus. RA is a highly important morphogen during vertebrate development, including the development of various eye tissues, and has been previously implicated in several regenerative processes as well. For instance, Wolffian lens regeneration in the newt requires active RA signaling. In contrast, we provide evidence here that lens regeneration in Xenopus actually depends on the attenuation of RA signaling, which is regulated by the RA-degrading enzyme CYP26. Using RT-PCR we examined the expression of RA synthesis and metabolism related genes within ocular tissues. We found expression of aldh1a1, aldh1a2, and aldh1a3, as well as cyp26a1 and cyp26b1 in both normal and regenerating corneal tissue. On the other hand, cyp26c1 does not appear to be expressed in either control or regenerating corneas, but it is expressed in the lens. Additionally in the lens, we found expression of aldh1a1 and aldh1a2, but not aldh1a3. Using an inhibitor of CYP26, and separately using exogenous retinoids, as well as RA signaling inhibitors, we demonstrate that CYP26 activity is necessary for lens regeneration to occur. We also find using phosphorylated Histone H3 labeling that CYP26 antagonism reduces cell proliferation in the cornea, and using qPCR we find that exogenous retinoids alter the expression of putative corneal stem cell markers. Furthermore, the Xenopus cornea is composed of an outer layer and inner basal epithelium, as well as a deeper fibrillar layer sparsely populated with cells. We employed antibody staining to visualize the localization of CYP26A, CYP26B, and RALDH1 within these corneal layers. Immunohistochemical staining of these enzymes revealed that all 3 proteins are expressed in both the outer and basal layers. CYP26A appears to be unique in also being present in the deeper fibrillar layer, which may contain cornea stem cells. This study reveals a clear molecular difference between newt and Xenopus lens regeneration, and it implicates CYP26 in the latter regenerative process.

Relative developmental toxicity potencies of retinoids in the embryonic stem cell test compared with their relative potencies in in vivo and two other in vitro assays for developmental toxicity

The present study determines the relative developmental toxicity potencies of retinoids in the embryonic stem (ES)-D3 cell differentiation assay of the embryonic stem cell test, and compares the outcomes with their relative potencies in in vivo and two other in vitro assays for developmental toxicity. The results reveal that the potency ranking obtained in the ES-D3 cell differentiation assay is similar to the reported potency rankings in the two other in vitro assays for developmental toxicity. TTNPB ((E)-4[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1-propenyl]benzoic acid) was the most potent retinoid, whereas etretinate and retinol had the lowest potency. All-trans-retinoic acid, 13-cis-retinoic acid, 9-cis-retinoic acid and acitretin showed an intermediate potency. In vivo potency rankings of the developmental toxicity of retinoids appear to be dependent on the species and/or exposure regimens used. The obtained in vitro potency ranking does not completely correspond with the in vivo potency rankings, although TTNPB is correctly predicted to be the most potent and retinol the least potent congener. The lack of in vivo kinetic processes in the ES-D3 cell differentiation assay might explain the deviating potency predictions of some retinoids. Therefore, knowledge on the species-dependent in vivo kinetics is essential when using in vitro toxicity data for the estimation of in vivo developmental toxicity potencies within series of related compounds.

The role of CYP26 enzymes in defining appropriate retinoic acid exposure during embryogenesis

Retinoic acid (RA) is a pleiotropic derivative of vitamin A, or retinol, which is responsible for all of the bioactivity associated with this vitamin. The teratogenic influences of vitamin A deficiency and excess RA in rodents were first observed more than 50 years ago. Efforts over the last 15-20 years have refined these observations by defining the molecular mechanisms that control RA availability and signaling during murine embryonic development. This review will discuss our current understanding of the role of RA in teratogenesis, with specific emphasis on the essential function of the RA catabolic CYP26 enzymes in preventing teratogenic consequences caused by uncontrolled distribution of RA. Particular focus will be paid to the RA-sensitive tissues of the caudal and cranial regions, the limb, and the testis, and how genetic mutation of factors controlling RA distribution have revealed important roles for RA during embryogenesis.

Multiple factors contribute to the toxicity of the aromatic retinoid TTNPB (Ro 13-7410): interactions with the retinoic acid receptors

The aromatic retinoid, (E)-4-[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthylenyl)-1 -propenyl] benzoic acid (TTNPB) is 1000-fold more teratogenic than all trans-retinoic acid (tRA) in several species. Factors that partially explain the potency of this retinoid include binding affinities to retinoid nuclear receptors (RARs) in the nanomolar range, reduced affinities for the cytosolic binding proteins (CRABPs), and slow rate of metabolism (M. A. Pignatello, F. C. Kauffman, and A. A. Levin, Toxicol. Appl. Pharmacol. 142, 319-327, 1997). The present work investigates the possible involvement of longer receptor occupancy and increased transcriptional activity of the ligand receptor complex in the greater toxicity of TTNPB. Ligand off-rates from nuclear receptors were determined in nucleosol fractions prepared from COS-1 cells transfected with cDNA encoding the appropriate RAR subtype. When assayed at 10 degrees C, [3H]TTNPB was displaced from the RARs at a significantly faster rate than that of [3H]tRA. The difference in displacement was reduced at 4 degrees C. These observations are consistent with the 10-fold lower affinity of TTNPB vs tRA for RARs and, therefore, do not explain the greater potency of TTNPB. The ability of TTNPB and tRA to activate the RARs was determined using a luciferase reporter gene transfected into JEG-3 cells with the appropriate RAR subtype. The expression of the reporter was driven by a retinoic acid response element (RARE) from the RAR beta gene, which was incorporated into the reporter plasmid. Dose-response for gene activation indicated that the potency of TTNPB and tRA in activating mRAR alpha, beta, and gamma was similar after 24 h with comparable EC50s in the nanomolar range. However, after 72 h, activation by TTNPB was greater than that of tRA as indicated by EC50s and threshold for activation. This study indicates that the higher potency of TTNPB in activating the RARs may be due to slower disappearance of the retinoid and, therefore, is a contributing factor to its greater toxicity.

Multiple factors contribute to the toxicity of the aromatic retinoid, TTNPB (Ro 13-7410): binding affinities and disposition

The aromatic retinoid (E)-4-[2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthylenyl)-1 -propenyl] benzoic acid (TTNPB) is 1000-fold more potent as a teratogen than all trans-retinoic acid (tRA) in several species and in the inhibition of chondrogenesis in the mouse limb bud cell culture. Factors responsible for the potency of TTNPB were investigated including binding to nuclear retinoic acid receptors (RARs and RXRs), cytosolic binding proteins (CRABPs), and metabolic disposition of TTNPB. For competitive binding assays and saturation kinetics, nucleosol or cytosol fractions were obtained from COS-1 cells transfected with cDNAs encoding the appropriate nuclear receptor or binding protein. TTNPB binds to RAR alpha, beta, and gamma with Kds in the nanomolar range; however, these binding affinities are 10-fold less than those of tRA. Although the affinities are high for TTNPB, it is unlikely that the binding affinities to nuclear receptors alone account for the potency of TTNPB. The binding affinities of TTNPB for the CRABPs are significantly lower than those of tRA. TTNPB did not compete with [3H]9-cis RA for binding to RXR alpha, beta, or gamma. Mouse limb bud cell cultures, a well characterized model for retinoid teratogenesis, were used to compare the metabolic disposition of TTNPB and tRA. In the media of limb bud cell cultures treated with either retinoid, the disappearance of TTNPB was significantly slower than that of tRA over 72 hr. Both retinoids reached approximately equal concentrations in cell uptake experiments; however, TTNPB disappeared from the limb bud cell at a significantly slower rate than did tRA. Collectively, these results indicate that high affinity binding to RARs, lower affinity to CRABPs, and resistance to metabolism contribute to the potency of TTNPB.

Characterization of synthetic retinoids with selectivity for retinoic acid or retinoid X nuclear receptors

The broad spectrum of physiological activities of retinoids is mediate d by two types of receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs). Though they have 9-cis retinoic acid as a common ligand, the amino acid sequence of their ligand binding domains is only distantly related (27%). This fact makes it probable that the ligand binding pockets of RARs and RXRs differ significantly with respect to their three dimensional structure. Therefore, one can expect that selective ligands for these receptor subclasses do exit. A clear example of a naturally existing RAR-selective retinoid is all-trans retinoic acid. Here we report on two synthetic retinoids which are very closely related to retinoic acid in structure yet show good receptor subclass selectivity. These compounds have a saturated double bond in the polyene side chain between either the 7, 8 or 9, 10 carbon atoms and are highly RAR or RXR selective, respectively (as shown by receptor binding, transactivation activity and the ability to induce RXR homodimer formation). In addition, we present compounds of the synthetic arotinoid class which are highly RAR selective. Interestingly the corresponding '9-cis analogs' are not able to bind or activate RXR alpha and show greatly reduced activity on the RARs.

Comparative disposition, receptor affinity, and teratogenic activity of sulfon arotinoids

To investigate the relationship between sulfon arotinoid biotransformation and teratogenic activity, the potency of the ethyl (Ro 15-1570) and methyl (Ro 14-9706) arotinoid sulfones and their in vivo disposition in pregnant hamsters were studied. Administration of Ro 15-1570 was teratogenic, but Ro 14-9706 showed no such activity. Total absorbed doses of the ethyl and methyl sulfones (measured as maternal plasma AUC) were very similar. Total delivered dose of Ro 14-9706 to liver and lung was 120-160% that of Ro 15-157, and Ro 14-9706 was transferred in greater amounts to the embryo as well. Placenta AUC for parent sulfon arotinoids was 160-250% that in the embryo. Plasma analyses by HPLC suggested that the ethyl sulfone was oxidized and appeared in maternal plasma as the corresponding sulfinic (Ro 14-9572) and sulfonic (Ro 14-3899) acids, amounting to 10% and 16%, respectively, of the mean maternal ethyl sulfone Cmax value. The concentrations of sulfinic and sulfonic metabolites were always less than the analytical limit of detection in placenta and embryo after maternal ethyl sulfone intubation. Neither the sulfinic nor the sulfonic acid were ever detected in maternal circulation, placenta, or embryo after methyl sulfone intubation. Comparisons of their binding affinities found that neither the ethyl nor the methyl arotinoid sulfone could act as a ligand for cellular retinoic acid-binding protein (CRABP), nor could these compounds bind retinoid nuclear receptors (RAR). Transcriptional activation of RARs was weak and similar for both compounds. The sulfinic and sulfonic acid arotinoids bind and transactivate RARs, and bind CRABP with efficiencies similar to all-trans-retinoic acid. Furthermore, they are active in cultured limb bud chondrocytes. The results suggest that the methyl sulfone (in accord with its lack of activity in cultured limb bud chondrocytes) is of no toxicologic significance in hamster embryo--even after relatively high delivered dose. Teratogenicity of the ethyl sulfone (which shows marked inhibition of chondrogenesis in cultured limb bud) does not appear to depend on measurable concentrations of these sulfinic/sulfonic acid metabolites in the hamster embryo.

Transforming growth factor beta 2 in epithelial differentiation of developing teeth and odontogenic tumors

Dysregulation of TGF beta 2, a modulator of cell growth and differentiation, can result in uncontrolled growth and tumor formation. Our comparative studies on the expression of TGF beta 2 mRNA and protein indicate that TGF beta 2 may primarily be a regulator of epithelial differentiation during tooth development (between 13 and 20 gestational wk) and tumorigenesis of odontogenic neoplasms. A paracrine mode of action for TGF beta 2 in early human tooth germ (cap/early bell stage) is suggested by location of mRNA in the mesenchyme surrounding the tooth germ, whereas protein is found in the epithelial dental lamina and enamel organ. During the late bell stage, TGF beta 2 gene expression shifted from the mesenchyme to the odontogenic epithelium and was colocalized with protein, suggesting an autocrine role for the terminal differentiation of ameloblasts. In odontogenic tumors of epithelial origin (ameloblastomas) and epithelial-ectomesencymal origin (ameloblastic fibromas), TGF beta 2 mRNA was mostly located in the mesenchymal tumor component and protein in the epithelial tumor component. Odontogenic ectomesenchymal tumors (myxomas) were not associated with TGF beta 2 mRNA and protein expression. The results imply that TGF beta 2 may play an important role in epithelial-mesenchymal interactions in human tooth morphogenesis and development of odontogenic tumors.

Analysis of high dysmorphogenic activity of Ro 13-6307, a tetramethylated tetralin analog of retinoic acid

Certain synthetic retinoids differ widely from retinoic acid (RA) in teratogenic potency, being much more or much less effective than RA. It is assumed that the potency of a retinoid may depend on the nature of its interaction with cellular binding components (nuclear retinoic acid receptors or cytoplasmic binding proteins) and, as in the case of retinoids that are mammalian teratogens, on factors that determine its accessibility to the embryo. To investigate some of the factors that contribute to potency, we used a new synthetic retinoid Ro 13-6307 that differs in structure from RA in having an aromatic ring inserted in its side chain along with gem dimethyl modification of the natural cyclohexenyl ring. Pregnant ICR mice were given a single oral dose (0, 1, or 10 mg/kg) on day 11 of gestation, and the resultant teratogenic outcome was monitored on day 17. Direct effects on cell differentiation were obtained by exposing high density cultures of limb bud mesenchymal cells to a range of concentrations (0.3 ng/ml-3 micrograms/ml) of Ro 13-6307 and scoring for chondrogenic suppression. Concentrations reaching the embryo after maternal administration of Ro 13-6307 were measured by HPLC to quantify the analog for a period of 4 h after administration of the oral dose. We found that this retinoid was 40-fold as active as RA in both inducing teratogenesis and suppressing chondrogenesis, yet its concentration in the affected embryo was only a fraction of that achieved after an equivalent dose of RA was employed in a similar protocol.(ABSTRACT TRUNCATED AT 250 WORDS)

Teratogenicity of arotinoids (retinoids) in vivo and in vitro

The effect of structural modifications on the arotinoid molecule, a new class of retinoids, on their teratogenicity in mice was studied. Animals were treated on days 8 and 9 of gestation, the most susceptible stages to retinoid-induced malformations in rodents. The teratogenic potency of the 13 arotinoids tested varied over a dose range of more than five orders of magnitude. Next, we tested whether the quantitative differences in the teratogenicity of these arotinoids correlates with their activity in high density (micromass) cultures of rat embryonic limb bud and midbrain cells. There was a good quantitative correlation between the in vivo teratogenicity and the in vitro activity in limb bud cells but no correlation was found in midbrain cells. Thus, the limb bud cell culture system may be useful for a preliminary testing to select non-teratogenic retinoids. For the risk assessment in humans, however, the in vitro results should be verified in animals studies.