A sustained elevation in retinoic acid receptor-beta 2 mRNA and protein occurs during retinoic acid-induced fetal dysmorphogenesis.

We have previously shown that oral treatment of pregnant mice with all-trans retinoic acid (RA) at doses which cause 100% fetal dysmorphogenesis results in a rapid elevation in the mRNA of one specific isoform of the RA receptor-beta, RAR-beta 2, in susceptible embryonic regions. To further investigate the involvement of RAR-beta 2 mRNA in teratogenesis, we have examined its expression in mouse embryos exposed to marginal/nonteratogenic and teratogenic dosing regimens of both 13-cis RA and all-trans RA. We have found that the mere elevation in embryonic RAR-beta 2 mRNA levels and free retinoid levels is not sufficient to result in dysmorphogenesis. Rather, retinoid-induced dysmorphogenesis of embryos appears to occur only when RAR-beta 2 mRNA and unbound retinoid levels remain elevated for at least 6-9 h following retinoid treatment resulting in a significant and prolonged elevation in RAR-beta protein levels.

The use of a retinoid receptor antagonist in a new model to study vitamin A-dependent developmental events.

Multiple fetal anomalies occur in vitamin A deficient animals as well as in retinoic acid receptor gene 'knockout' mice, indicating that retinoic acid (an active metabolite of vitamin A) performs some essential functions in normal development. Additional approaches are needed to probe directly the stages and sites in the embryo where a presence of endogenous retinoic acid is indispensable. We have employed a new strategy for this purpose which involved an intervention in retinoic acid receptor (RAR)-dependent functions at specific developmental stages by means of a highly effective RAR antagonist, AGN 193109. We report that in an in vitro cell differentiation bioassay, AGN 193109 completely reversed the inhibitory action of a potent RAR agonist, AGN 190121. In pregnant mice, a single oral 1 mg/kg dose of the antagonist given on 8 day post coitum (dpc) produced a severe craniofacial anomaly (median cleft face or frontonasal dysplasia) and eye malformations in virtually all exposed fetuses. On the other hand, treatment on 11 dpc, a time in development when RARs are strategically expressed in the limb bud primordium, no limb anomalies could be induced by the antagonist. Even after a high dose of 100 mg/kg, limb development progressed normally in spite of the fact that measurable concentrations of the antagonist were present. Because retinoids are long known to influence skin morphology, we next monitored the effects of the antagonist on skin development. When given late in gestation, on 14 dpc, we found that the antagonist delayed differentiation and maturation of the fetal skin and hair follicles. We conclude that this model provides a convenient and pertinent system which enables us to seek and clarify true functions of retinoic acid and its cognate receptors in embryogenesis and in adult animals.

Modulation of limb bud chondrogenesis by retinoic acid and retinoic acid receptors.

An excess of retinoic acid (RA) in the mouse embryo in utero produces hypochondrogenesis and severe limb bone deformities. Since one of the RA receptors--RAR-beta 2, is specifically induced in the limb bud cells upon treatment of embryos with teratogenic doses of RA, we investigated if this receptor played a role in teratogenesis by regulating the process of chondrogenesis. In micromass cultures of mouse limb bud mesenchymal cells, we found that a downregulation of RAR-beta 2 as well as several other RAR isoforms by supplementation of the culture medium with specific oligodeoxynucleotides stimulated chondrogenesis: cartilage nodule number, sulfated proteoglycans, and synthesis of collagen type IIB were all enhanced in a dose-dependent manner. However, only the antisense RAR-beta 2 probe efficiently prevented the strong inhibitory effects of exogenous RA on chondrogenesis in these cells. The data suggest that the RAR-RA complexes play a role in position-dependent patterning of the limb skeleton in normal development and that, in particular, RAR-beta 2 serves to prevent the mesenchymal cells from expressing their chondrogenic bias. Our results further strengthen the argument that RA-dependent elevation in RAR-beta 2 levels plays a unique role in RA-induced teratogenesis.

Synthesis and structure-activity relationships of retinoid X receptor selective diaryl sulfide analogs of retinoic acid.

Retinoids exert their biological effects by binding to and activating nuclear receptors that interact with responsive elements on DNA to promote gene transcription. There are two families of retinoid receptors, the retinoic acid receptor (RAR) family and the retinoid X receptor (RXR) family, which are each further divided into three subclasses: RAR alpha, beta, gamma and RXR alpha, beta, gamma. Herein we describe the synthesis and structure-activity relationships of a new series of diaryl sulfide retinoid analogs that specifically bind and transactivate the RXRs. Furthermore, the sulfoxide and sulfone derivatives of these analogs are partial agonists which activate the RXRs only at high concentrations. Thus, these compounds possess a potential site of metabolic deactivation and may have less prolonged systemic effects than other compounds with arotinoid-like structures. We show also that these compounds have activity in nontransfected cells as demonstrated by their ability to induce TGase activity in HL-60 cells. Finally, we corroborate our earlier report that RXR-specific agonists may possess reduced teratogenic toxicity compared to RAR-specific agonists since these compounds are much less potent inhibitors of chondrogenesis than RAR-specific agonists such as TTNPB.

Diminished teratogenicity of retinoid X receptor-selective synthetic retinoids.

One feature that contraindicates the wide therapeutic use of retinoids is their teratogenicity. Synthetic retinoids are distinguishable from each other on the basis of their partial or exclusive preference in binding and activation of all-trans retinoic acid receptors (RARs) or retinoid X receptors (RXRs). Using mouse embryo limb bud cells in micromass cultures as a bioassay, we examined the inhibitory activities of a number of standard and novel retinoids on chondrogenic cell differentiation. Transient cotransfection of HeLa cells was used to measure the ability of each retinoid to induce transcription of a reporter gene by activating RAR alpha, RAR beta, RAR gamma, or RXR alpha chimeric constructs. All retinoids in this study that activated RARs to any degree in the cotransfection assay also inhibited chondrogenesis in vitro, whereas retinoids that were either specific for RXR or inactive in the cotransfection assay did not. The activity of RAR-selective agonists and the inactivity of RXR-specific agonists in the cotransfection assay correlated well with the relative teratogenicity of six of the representative retinoids studied when orally administered at day 11 to pregnant ICR mice.

Effects of retinoic acid treatment on release of arachidonic acid by chondrogenic cells in response to ionophore A23187.

Chondrogenic differentiation in mouse limb bud mesenchymal cells cultured at high density was suppressed by supplementation of the medium with retinoic acid in a dose-dependent fashion. Cells prelabeled with (3H) arachidonic acid were treated with 0.3 microgram/ml retinoic acid. Treatment with retinoic acid increased the (3H) fatty acid in the triglyceride fraction. Furthermore, treatment with retinoic acid enhanced the release of (3H) fatty acid upon stimulation of these cells with the divalent ionophore A23187. These data permit the suggestion that there may be a correlation between altered lipid metabolism and retinoic acid's ability to disrupt chondrogenic differentiation.

Differential teratogenic response of mouse embryos to receptor selective analogs of retinoic acid.

Early events that initiate teratogenesis by Accutane or other retinoids in mammalian embryos remain unknown. It would be helpful for mechanistic considerations to know whether or not retinoids act through retinoid receptor-dependent pathways, and if they do, which of the two families of receptors (retinoic acid receptors - RARalpha, beta, gamma or retinoid X receptors - RXRalpha, beta, gamma) are more likely involved. We previously used an in vitro bioassay to demonstrate that those retinoid analogs with binding affinity and transactivational activity limited only to the RXRs have a low potential as teratogens. Here, we have extended the study to examine teratogenicity, in pregnant mice, of a number of synthetic retinoids with varying degrees of receptor selectivity. The ability of each compound to induce fetal limb and craniofacial defects after a single exposure on day 11 of gestation was assessed and compared to that of all-trans retinoic acid (RA). The highest dose selected was 100 mg/kg maternal body weight since such a regimen of all-trans RA affects virtually every exposed embryo without any indication of maternal toxicity. We found that although all RAR agonists were strong teratogens, their potencies varied over a wide magnitude. The teratogenic potencies and receptor transactivation profiles of RAR agonists were not directly correlated since compounds with similar receptor activities presented major differences in potencies. Three compounds were exclusively RXR agonists, and these were not teratogenic under our experimental conditions. Two additional compounds which turned out to be non-teratogenic were distinguished by the fact that they activated neither RARs nor RXRs. These data indicate that although RAR-dependent mechanisms are likely involved in retinoid-induced teratogenesis, there are additional factors which determine teratogenic potency. The absence of teratogenic response in the case of RXR agonists suggests that risk-benefit analyses of such receptor-selective compounds may be fruitful in further studies.

Production of congenital limb defects with retinoic acid: phenomenological evidence of progressive differentiation during limb morphogenesis.

Maternal administration of a single dose of retinoic acid (vitamin A acid, 100 mg/kg) on either the 11th, 11 1/2, 12th, 12 1/2, 13th or 13 1/2 day of gestation produced phocomelia or partial phocomelia in ICR/DUB fetuses. The results depended upon the time of treatment and two gradients of effect were produced: 1) cranio-caudal gradient, since forelimb defects resulted from between days 11 and 13, while similar hindlimb abnormalities were produced by administration of retinoic acid 12 to 24 hours later: 2) proximo-distal gradient, due to the heterogenous sensitivity among individual bones of the limb. In the forelimb, early treatment (11th day) produced humero-unlar defects and later treatment (12th day) ulnoradial defects. A similar proximo-distal gradient was observed in the hindlimb. The use of teratological studies as a tool to assist morphogenetic investigation is discussed.

Transplacental pharmacokinetics of teratogenic doses of etretinate and other aromatic retinoids in mice.

The transplacental pharmacokinetics of single teratogenic doses of etretinate and motretinide were compared with particular emphasis on distribution and concentrations in the exposed embryos of the free acid metabolite, etretin. The three aromatic retinoids were also tested for their direct inhibitory effect on chondrogenesis in the limb bud mesenchymal cell "micromass" culture assay. After a standard dose of 100 mg/kg administered on day 11 of gestation in NMRI mice, all three compounds were teratogenic, but they differed from each other in potency. Etretinate was most active as a teratogen, equalling the potency of our standard all-trans-retinoic acid; every exposed fetus was deformed with severe shortening of all limb bones as well as cleft palate. Etretin was less potent than etretinate, and motretinide was considerably less active as a teratogen than the other two. In the in vitro assay, only etretin suppressed chondrogenesis and this activity was equivalent to that of all-trans-retinoic acid (IC50 of 12 ng/ml). Both etretinate and motretinide (which contain an ethyl ester and ethylamide terminal group, respectively) were essentially inactive in vitro, demonstrating the fact that a free carboxylic group may be a requirement for the in vitro suppression of chondrogenesis. These differences between the results obtained in vivo and in vitro could be resolved by pharmacokinetic investigations using HPLC methods. Both etretinate and motretinide were metabolized in vivo to etretin, their likely common teratogenic metabolite. The high teratogenic potency of etretinate was probably the result of high concentrations as well as AUC values of its metabolite etretin in the embryo. On the other hand, the comparatively low teratogenicity of motretinide could be related to approximately 5 x lower embryonic peak levels as well as AUC values of etretin. A comparison of these results with those previously obtained for all-trans- and 13-cis-retinoic acids confirms the correlation between embryonic exposure and teratogenic potency in the mouse. Our results indicate that pharmacokinetic studies are essential for the interpretation of relative teratogenic potencies of retinoids as well as apparent differences between in vivo and in vitro teratogenesis. A free carboxyl group at the terminal end of the tetraene chain was necessary for high activity of the retinoids studied.

Criteria for identifying and listing substances known to cause developmental toxicity under California's Proposition 65.

Because of the automatic restrictions and warning requirements imposed on substances identified by the state as "known to cause developmental toxicity," the Expert Committee recommends the use of criteria that emphasize human relevancy, biological plausibility, and evidence in support of a selective, adverse developmental effect at non-maternally-toxic doses. In many instances, data for substances of public concern will be insufficient at present to meet these criteria. The fact that a substance is not listed as "known to cause developmental toxicity" does not create a presumption that the substance is safe. The Expert Committee, therefore, urges that these substances be recommended for further testing and that high priority be given to conducting the necessary tests. The Expert Committee reiterates its concern that substances listed by the SAP be identified according to the toxic endpoints (cancer, male reproductive toxicity, female reproductive toxicity, and/or developmental toxicity) that led to listing. Further, the Expert Committee recommends that the state Health and Welfare Agency institute education programs emphasizing appropriate courses of action for citizens informed of exposures to substances known to the state to cause cancer, birth defects, or reproductive toxicity.

Skeletal morphogenesis: comparative effects of a mutant gene and a teratogen.

As a teratogenic agent retinoic acid (RA) produces severe limb reduction defects if administered at a certain stage of embryonic development. In vitro, RA is able to prevent chondrogenesis and this inhibitory effect is accompanied by the absence of cartilage specific proteoglycans in treated cultures. Such an effect is ruled out as a direct causative factor in teratogenesis for two reasons. First, the limbs of treated embryos show extensive chondrogenesis and this cartilage is normal as far as the expression of biochemical markers of differentiation are concerned. Second, the morphogenetic effects of a mutant gene, cmd, where there is a functional deficit of the proteoglycan core protein are very different from those associated with RA-induced teratogenesis. The differences between the two are not wholly reconciled by the fact that the effects of the mutant gene are cumulative and progressive while those of the RA insult are transitory. There are a number of developmental events which are, however, altered by RA in the mesenchymal cells of the early limb bud such as cell proliferation, cell death, and hyaluronic acid metabolism. Not only any one or more of these factors may secondarily inhibit chondrogenesis but, more importantly, may also have a number of other consequences in the developing embryo. Since a number of cell types besides mesenchymal cells respond to RA by altering their pattern of differentiation, it is conceivable that some fundamental molecular step in the process of differentiation provides a target for its action. In a recent review, Sporn and Roberts (1983) have suggested that to be compatible with the wide ranging effects of retinoids documented so far, any hypothesis put forward for its molecular mechanism of action must include a role in gene expression. No experimental work has yet directly addressed how retinoids might modify gene expression. We believe that along with teratocarcinoma stem cell lines, the use of retinoids as selective teratogens may open up another avenue in search of molecular mechanisms of cell differentiation.

Cellular expression of a mutant gene (cmd/cmd) causing limb and other defects in mouse embryos.

The cmd mutation is associated with a systemic defect in all parts of the cartilaginous skeleton. Since limb buds isolated from the embryos and cultured in vitro develop the defect, the gene expression is independent of the humoral factors in the maternal/placental/embryonic environment. The limb development in the mutant progresses normally through early events in morphogenesis including the formation of mesenchymal cell condensations preparatory to chondrogenesis. The deviation first becomes apparent as soon as the chondrogenic cells begin the process of accumulation of extracellular matrix; cmd/cmd chondrogenic cells remain close to each other and lack the abundant extracellular matrix which accumulates between normal cells. Quantitatively normal levels of chondroitin sulfate proteoglycans are synthesized by the mutant limbs during precartilaginous stages. Subsequently, however, the mutant fails to attain the normally high levels of chondroitin sulfate synthesis. Its growth rate also slows down, as judged by the lowered protein synthesis in the mutant cultured limb buds. The lack of at least one species of protein molecules, ie, proteoglycan core-protein, is already known from previous studies on this mutation; an abnormal or a deficient core-protein was shown to lead to a virtually complete shut off of the biosynthesis of cartilage-specific proteoglycans in another mutation in the chick embryo [Goetinck, 1982]. It may be important to note that the cmd mutation does not seem to interfere with the process of determination of cartilage even though it interrupts virtually completely one important biosynthetic aspect of the chondrogenic cell differentiation pathway. The mutant chondrocytes, embedded as they are in an abnormal and proteoglycan-deficient matrix, begin to degenerate prematurely without first undergoing hypertrophy. Also, the process of ossification begins precociously in the shortened cartilage models of the mutant, hence resulting in overall shortening of the limbs. As assessed from the HA:s-GAG ratios during early embryonic limb development, some of the phenotypically normal embryos could be distinguished as recessive carriers of the mutation. Even though these carriers have an intermediate level of chondroitin sulfate proteoglycan synthesis, this does not interfere with their normal development during prenatal stages. It will be of practical importance to follow these carriers through subsequent postnatal stages and adult life to assess any long-term effects.

Transplacental passage of label after administration of (3H) retinoic acid (vitamin A acid) to pregnant mice.

To determine whether there is transplacental passage of retinoic acid (vitamin A acid) or its metabolites levels of radioactivity were measured in embryos after administration to 12th-day pregnant ICR/DUB mice of nonteratogenic or teratogenic quantities of (3H)retinoic acid. After the former treatment radioactivity was demonstrable in embryos, placentas, and visceral yolk sacs (VYS) at 1 h after treatment. Expressed in terms of radioactivity per unit protein content, embryos at 1 h had about one-third as much radioactivity as maternal livers. The placentas had maximum radioactivity at 2 h, whereas embryos and VYS accumulated maximum levels by 6 h after treatment. At 6 h the level in embryos was twice that in placentas. After teratogenic treatment placentas and embryos attained peak levels of radioactivity at 6 and 12 h later, respectively. Again, embryos at peak level contained 2-3 times as much radioactivity as placentas or VYS. The amount of retinoic acid, metabolites, or both, in embryos after the teratogenic treatment, calculated on the basis of specific activity of the injected label, was 26 times greater than after nonteratogenic treatment; the amount in maternal liver was 75 times greater. The results support previous evidence that the teratogenic effects of retinoic acid are associated with raised levels of retinoic acid, its metabolites, or both, in embryos.

The use of in vitro procedures in teratology.

The capabilities of investigators in the fields of teratology and toxicology are greatly enhanced by the use of tissue culture procedures in unraveling the mechanisms of drug action. Techniques currently available for the culture of postimplantation mammalian embryos permit evaluation of their metabolic responses to potential teratogens even when the length of time embryos survive and develop in culture is too short to allow a conventional teratologic survey of malformations. A simple procedure for culturing mouse embryos during early organogenetic stages is described in this report that will be of value to teratologists; it avoids the requirements of special glassware and equipment by using ordinary capped test tubes which are rotated tomaintain and efficient nutritional and gaseous evnironment. Some studies concucted with this procedure to monitor the metabolism of embryo during the first 24 h of culture are summarized. Another aspect of tissue culture, organ culture, provides further manipulative capability by which embryonic organs can be maintained for long periods of time during which they develop and differentiate to an extent that their morphological and biochemical responses to a teratogen can usually be made. Comparative effects of several teratogenic agents and the relative concentration of each that produces a similar degree of response are summarized. It is concluded that organs are more sensitive to teratogens in culture than they are in vivo, and that different teratogens possess enough specificity to isolate their simple growth-retarding effect from the role they play in distrubing other specific developmental events.

Embryonic limb bud organ culture in assessment of teratogenicity of environmental agents.

The limb bud organ culture system offers a variety of endpoints which may be monitored in the screening process. These are: cell proliferation, differential growth, morphogenetic cell death, size and shape of limb parts, chondrogenesis, collagen or proteoglycan biosynthesis, etc. Essentials of the system including various parameters of normal limb bud development in vitro are described. These parameters serve to gauge the effects of test chemicals with unknown hazard potential. Validation has been carried out only to a limited extent. Further, it needs to be combined with an efficient drug metabolizing preparation before it can achieve its full potential as a short-term screening system.

In vitro testing of teratogenic agents using mammalian embryos.

The need for efficient methods to screen new chemicals, drugs, and environmental pollutants for their teratogenic activity is obvious. The method currently available, ie, pregnant-animal testing, is of considerable value but there are certain drawbacks which prevent reliance on this method alone as the predictive device. Nutritional state of the dam, variability in the developmental age of embryos from litter to litter and even within the same litter, metabolic differences between species, placental function, and a host of other factors must be taken into account before data obtained from animal testing can be logically extrapolated to human situation. Many of these variables are either eliminated altogether or at least can be controlled by the use of tissue culture techniques. After surveying a variety of in vitro systems, it is our opinion that organ culture, whole embryo culture, and a combination of the two offer at present the best potential for screening of suspected teratogens. These culture techniques provide a much better simulation of in vivo situations than isolated cells grown as monolayers. Among other advantages, these procedures allow one to exercise control over the effective concentration of the suspected teratogen to which an embryo is exposed and also the duration of this exposure. Since maternal metabolism or modification of the drug is routinely eliminated in these experiments, there is a need for exploring the use of drug-metabolizing preparations as additives to the culture medium. The choice of limb bud in the screening system is promising since during its development, the limb progresses through a succession of embryonic processes that are generally relevant to other organ systems as well. Hence, such a screening system may not only predict teratogenicity but also provide insight into the mechanisms by which a test chemical is teratogenic.

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)

Induction of tissue transglutaminase and apoptosis by retinoic acid in the limb bud.

Mesenchymal cells in the limb buds of midgestation mouse embryos suffer prominent cell death upon exposure to retinoic acid (RA), an event likely associated with the micromelic and phocomelic anomalies of the resultant fetuses. It has been suggested, but not yet shown, that cells die by an active process termed apoptosis rather than by necrotic cytolysis. In certain cell types, investigators have previously observed a specific and early effect of RA on transcriptional activation of the gene for tissue transglutaminase (tTG), an enzyme suspected to play a role in apoptosis. We report here a distinct but transient increase in tTG activity which accompanied the initiation of cell death in the mesenchymal cells located in the central core of RA-treated limb buds. We also ascertained microscopically that the cytological appearance of the affected cells was consistent with a characterization of the process of cell death as apoptosis.