Main

It is generally accepted that the potential for teratogens to produce malformations is determined by several factors including the stage of exposure, genetic variability, and the dose or the magnitude of the exposure(1). The present investigation concerns the dose response relationship and the question of whether a threshold or no-effect dose exists for the production of teratogenic effects. The threshold dose has been defined as “the level of exposure below which the incidence of death, malformation, growth retardation, or functional deficiency is not statistically greater than that of controls”(1). This dose is usually less than one order of magnitude below the minimal teratogenic dose. All teratogens that have been appropriately tested have exhibited threshold phenomena during organogenesis(2).

Recently Gaylor et al.(3) suggested that the observed threshold for a teratogen might represent a “pragmatic threshold” rather than a true threshold. They postulated that evidence for an effect below this dose might be found if the database were simply large enough. Although this is a reasonable hypothesis, it is not testable by experiment because any number of animals may not be large enough to detect an increased risk.

Gaylor et al.(3) also reasoned that if a substance produces malformations via the same mechanism as that responsible for the spontaneous incidence in control animals, no threshold for that teratogen can exist. This is because any additional amount of the teratogen in question could only increase the magnitude of the teratogenic milieu. Examples given of endogenous factors that may not have a no-effect dose include hormones and vitamins(3, 4). We used this concept to form an experimentally testable hypothesis: if an endogenous substance is responsible for all or a portion of the spontaneous incidence of a particular malformation in a population, then reduction of that substance should decrease the incidence of the malformation. Furthermore, if an exogenous teratogen contributes to the detrimental action of an endogenous substance, then reduction of the endogenous substance should decrease the effect of the exogenous teratogen.

Glucocorticoid-induced CP in mice(5) provides an ideal model in which to test this hypothesis, especially as it has been suggested that endogenously produced corticosterone may contribute both to the action of exogenously administered teratogens and the background incidence of CP in mice(6–8).

In our first series of experiments, we examined whether endogenous levels of maternal plasma corticosteroids contribute to the spontaneous incidence of CP in mice: A/J mice were adrenalectomized to determine whether maternal endogenous corticosteroids contribute to the spontaneous incidence of CP. A/J mice are highly susceptible to corticosteroid-induced CP and the majority of reports linking endogenous corticosteroids and CP have used this strain. In the second series of experiments we determined whether endogenous levels of maternal corticosteroids were sufficient to influence induced CP resulting from administration of corticosteroids: adrenalectomized and intact pregnant CD-1 mice were used, rather than A/J mice, because they have similar susceptibility to corticosteroid-induced CP but tolerate surgery earlier in pregnancy.

METHODS

Adult A/J and CD-1 mice 6-10 wk of age were obtained from Jackson Laboratories (Bar Harbor, ME). Two females were housed with a male of the same strain overnight. The presence of a copulation plug on the following morning indicated 0.5 d p.c. Pregnant A/J mice were divided into four groups;1) control, 2) sham adrenalectomized, 3) adrenalectomized, and 4) adrenalectomized with plasma sampling. Pregnant CD-1 mice were divided into four groups; 1) control,2) control with plasma sampling, 3) sham adrenalectomized, and 4) adrenalectomized. Mice in the CD-1 sham adrenalectomized and adrenalectomized groups were further divided into five subgroups based on the dose of cortisone administered. All animals were housed in the accredited Alfred I. duPont Institute's Life Science Center. All experimental protocols were approved by the Alfred I. duPont Animal Care and Use Committee and the Thomas Jefferson University's Institutional Animal Care and Use Committee.

Adrenalectomies and sham adrenalectomies. Adrenalectomies performed on A/J mice early in gestation resulted in a high rate of maternal and fetal mortality. Therefore, adrenalectomies and sham adrenalectomies were performed on 12.5 d p.c. for A/J mice. Adrenalectomies and sham adrenalectomies were performed on CD-1 mice on 6.5 d p.c. Mice were anesthetized with a solution of ketamine, 37.5 mg/mL, and xylazine, 5 mg/mL, at a dose of 0.01 mL/5 g body weight. Mice were shaved and scrubbed with Betadine (Purdue Frederick, Norwalk, CT). An incision was made on the flank, the adrenal gland and periadrenal fat were exteriorized and, in the adrenalectomized group, the adrenal gland was removed. The muscle wall was sutured, and the procedure was repeated on the contralateral side. Skin incisions were closed with 9-mm stainless steel wound clips. Adrenalectomized mice received isotonic sodium chloride solution ad libitum to assist in maintaining sodium balance.

Determination of plasma corticosterone in A/J mice. Tail vein blood was collected from pregnant A/J mice just before adrenalectomy, and 48 h, 72 h, and 6 d after the operation. The tail was swabbed with 70% ethanol and the vein was incised with a scalpel blade. Approximately 0.2 mL of blood was collected in heparinized capillary tubes. The blood was centrifuged and the plasma stored at -20°C. Plasma samples were analyzed for corticosterone by an independent laboratory (Endocrine Sciences, Tarzana, CA).

Injection of cortisone and determination of plasma corticosterone in CD-1 mice. A 50 mg/mL solution of cortisone acetate (Carter-Glogau Laboratories, Glendale, AR) was diluted in isotonic saline to a final concentration of 12.5 mg/mL. Cortisone was injected s.c. into CD-1 mice in sham adrenalectomized and adrenalectomized groups at a dose of 12.5, 25, 50, or 100 mg/kg body weight at 1100 h on 10.5, 11.5 and 12.5 d p.c. Additional mice were injected with physiologic saline as a control.

Blood was collected from the tail vein of control, sham adrenalectomized and adrenalectomized CD-1 dams on 10.5 d p.c. before injection of cortisone, and again on 12.5 d p.c. 5 h after the injection of cortisone or the vehicle. Plasma corticosterone levels were determined by RIA using a commercial kit(ICN Biochemicals, Costa Mesa, CA).

Assessment of fetal outcome. On 18.5 d p.c. mice from all groups were killed by cervical dislocation. Adrenalectomized mice were examined for the presence of any remaining adrenal tissue; mice with visible adrenal tissue were discarded. The uterus was removed from each mouse, and the number of implantation sites, live and resorbed fetuses, and their location were recorded. The fetuses and placentae were weighed. Live offspring were examined for external malformations, killed with CO2, and placed in Bouin's fixative. Fetuses were examined for visceral and skeletal anomalies using the procedure of Wilson(9) as modified by Barrow and Taylor(10).

Statistical analysis. Fetal and placental weights were compared using one-way analysis of variance. Pairwise comparisons were made using Student-Newman-Keuls method. The incidence of resorptions and malformations in offspring were compared on a per litter basis using Kruscal Wallis one way ranked analysis of variance. Pairwise comparisons were carried out with the Mann-Whitney nonparametric U test. Analysis of the number of clefts or other malformations in the total number of surviving fetuses was byχ2 using Yate's correction. Levels of maternal corticosterone for each group were compared using one-way analysis of variance. Pre- and postinjection levels of maternal corticosterone were analyzed using a pairedt test. The level of significance was set at p ≤ 0.05 for all tests.

RESULTS

Experiments Using A/J Mice

Maternal plasma corticosterone in A/J mice. The maternal plasma corticosterone was significantly reduced at 48 and 72 h but not at 6 d after adrenalectomy when compared with levels before surgery(Table 1). At 72 h after adrenalectomies, 15.5 d p.c., corticosterone concentrations were significantly higher than 48 h values, but still significantly lower than presurgery levels.

Table 1 Plasma corticosterone concentrations from pregnant A/J mice before and after adrenalectomy
Full size table

Fetal outcome in A/J mice. The offspring of A/J mice from adrenalectomized, sham adrenalectomized, and control groups were examined on 18.5 d p.c. The number of implantation sites, number of surviving fetuses, and the incidence of orofacial clefting in surviving fetuses were determined for each group and are presented in Table 2. The number of implantation sites did not vary significantly among groups. The incidence of intrauterine death in litters from sham adrenalectomized and adrenalectomized dams was significantly greater than for control litters but not from each other. Because of this the number of live offspring was significantly smaller in the sham and adrenalectomized groups. A small but significant decrease in fetal body weight was also observed for the sham and adrenalectomized groups in comparison with the nonoperated controls.

Table 2 Summary of fetal outcome after subjecting pregnant A/J mice to different treatments
Full size table

Malformations in A/J mice. Offspring of A/J dams that had been adrenalectomized, sham adrenalectomized, or that had nothing done to them(control) were examined for CP, CL, and CLCP. Adrenalectomy or sham adrenalectomy of A/J dams did not alter the incidence of CLCP, nor the overall incidence of orofacial malformations (CLCP and CP) in offspring of the three groups (Table 2). One occurrence of CL without associated CP was observed in the control group. There was a significant increase in the incidence of isolated CP in the offspring of adrenalectomized A/J dams when compared with controls (Table 2). Isolated CP did not occur in offspring from the control group.

A 7% incidence of digit malformations (adactyly and syndactyly) was also noted in the offspring of sham adrenalectomized and adrenalectomized A/J dams. There was no significant correlation between the occurrence of digit malformations and either CLCP or CP. Although the pathology of the digit malformations is not known, they appear to be unrelated to orofacial clefting. No malformations other than orofacial clefting occurred in fetuses from dams in the nonoperated controls.

Effect of Maternal Adrenalectomy and Cortisone Injection on Offspring of CD-1 Mice

Plasma corticosterone in CD-1 mice. Maternal plasma corticosterone concentrations were determined for sham adrenalectomized and adrenalectomized CD-1 mice on 10.5 d p.c., just before the injection of cortisone or saline, and again on 12.5 d p.c., 5 h after the injection of cortisone. Plasma corticosterone concentrations were determined at the same time points for control animals (Table 3). There was a significant decrease in the concentration of plasma corticosterone in CD-1 dams that were adrenalectomized when compared with those that were sham adrenalectomized or to control dams on 10.5 d p.c. No significant difference was noted between corticosterone concentrations in sham adrenalectomized animals when compared with controls.

Table 3 Plasma corticosterone concentrations (ng/ml) from pregnant CD-1 mice subjected to different treatments
Full size table

On 12.5 d p.c. a significant decrease in plasma corticosterone occurred in sham adrenalectomized animals that received injections of cortisone at all doses administered when compared with controls (Table 3). In contrast, a significant increase in plasma corticosterone occurred in sham adrenalectomized CD-1 dams that were injected with saline 5 h before blood sampling when compared with values from control animals. Plasma corticosterone in adrenalectomized dams remained low and did not differ significantly from concentrations before injections on d 10.5.

Fetal outcome in CD-1 mice. Nine or more litters in each treatment group were examined for the number of implantation sites, resorptions, fetal body weight and the incidence of CP and CL(Table 4). No significant difference was noted in the number of implantation sites among the groups. A significant increase in intrauterine death occurred at a dose of 100 mg of cortisone/kg in both the sham and adrenalectomized groups. No difference in the incidence of intrauterine death was noted between sham adrenalectomized and adrenalectomized groups administered the same dose of cortisone.

Table 4 Summary of fetal outcome after subjecting pregnant CD-1 mice to different treatments
Full size table

A significant decrease in fetal body weight was observed in litters from all experimental groups when compared with fetal weights in the control group from which no plasma samples were taken. Fetal weights from sham adrenalectomized dams injected with 50 mg of cortisone/kg or greater and fetal weights from all adrenalectomized groups were also significantly reduced from those of fetuses from the control group dams from which blood samples were taken. Fetal weights were further reduced with increasing doses of cortisone. However, with the exception of dams administered 25 mg of cortisone/kg, no significant difference in fetal body weight occurred between cortisone-matched groups of adrenalectomized and sham adrenalectomized mice.

Malformations in CD-1 mice. The frequency of spontaneously occurring CP in CD-1 mice did not vary significantly between control groups with and without plasma sampling (Table 4). In general the incidence of spontaneously occurring CP in CD-1 mice was higher than the incidence of 0.2% for this strain reported by Rosenzweig and Blaustein(11). Offspring of control CD-1 mice that were not subjected to blood sampling, operated on, injected or treated in any way had a frequency of 2.2% CP (based on total number of fetuses), whereas control animals that were subjected to blood sampling had a frequency of 4.8%. No CP occurred in the offspring of sham adrenalectomized dams which had also been subjected to blood sampling. There was no observed increase in malformations other than CP in any group.

The incidence of CP in the offspring of CD-1 mice from control groups did not differ significantly from offspring of dams that had been sham adrenalectomized or adrenalectomized and were injected with isotonic saline(Table 4). The incidence of CP in offspring of mice in sham adrenalectomized and adrenalectomized groups administered a dose of 12.5 mg of cortisone/kg on 10.5-12.5 d p.c. did not differ from control groups or from each other. Litters from both sham adrenalectomized and adrenalectomized dams injected with 25 mg of cortisone/kg or greater had an increased incidence of CP when compared with controls. The incidence of CP increased with increasing doses above 25 mg of cortisone/kg and reached 100% in offspring of some mice injected with 100 mg of cortisone/kg. The incidence of CP did not differ significantly between cortisone dose-matched groups of adrenalectomized and sham adrenalectomized mice. In almost all cases the clefting occurred as isolated CP; only two fetuses had associated CL.

DISCUSSION

Since the discovery that exogenous corticosteroids induced CP in mice(5, 6) it has been hypothesized that endogenous corticosterone levels might also be teratogenic. Evidence linking maternal corticosterone, the major corticosteroid found in mice(12), and CP induction is largely derived from studies in which experimental stresses that elevated maternal corticosterone levels were associated with an increase in the incidence of CP in offspring. These stresses included food and water deprivation and/or physical restraint(11, 13–16), transportation(17), avoidance conditioning, and increased handling(13). Strain susceptibility to induction of CP by these means also correlated with that of exogenous corticosteroid-induced CP, with A/J mice being particularly susceptible to these effects(6, 17). The minimum concentration of endogenous corticosteroid sufficient to increase the incidence of CP was not determined in any of the above studies.

To assess the contribution of endogenous corticosteroids to the spontaneous incidence of orofacial clefting, we adrenalectomized pregnant A/J dams to reduce endogenous levels of corticosteroids. It is well documented that susceptibility to corticosterone-induced CP varies with gestational age and that the fetus is most susceptible on d 11-14 p.c. Once the the palate has closed it is no longer subject to corticosteroid induced teratogenesis(6). For these experiments adrenalectomy on d 12.5 p.c. significantly reduced levels of maternal corticosterone before and during the period of palatal closure, which in this strain occurs on late d 14 to early d 15 p.c.(6). This confirms earlier reports that demonstrated that the majority of plasma corticosterone during this period of gestation is of maternal origin(18). Fetal secretion of corticosterone is not detectable before d 15, and adrenalectomy does not appear to induce precocious secretion of fetal corticosteroids(19). By 15.5 d p.c. maternal plasma levels of corticosterone are increased from levels found on d 14.5 but are still significantly reduced from levels present in intact animals. Corticosterone detected in the maternal plasma on d 15.5 and subsequently is of fetal origin and is readily transferred to the maternal plasma(18). Although the fetus is susceptible to corticosteroid CP before d 12.5 when the adrenalectomies were performed, any effects of maternal corticosterone on palate development or tissue commitment before adrenalectomy would presumably be reflected in control and sham adrenalectomized as well as adrenalectomized animals. Thus we believe that adrenalectomy was effective in significantly reducing endogenous corticosterone such that the effect of endogenous corticosterone on palate closure could be assessed.

Reduction of maternal corticosterone by adrenalectomy did not alter the incidence of CLCP in offspring when compared with sham adrenalectomized and control groups. These results are not surprising because CLCP occurs spontaneously in A/J mice and the incidence is not increased by administration of corticosteroid(5, 6, 20). The incidence of CLCP in this study, i.e. 4.8% for control, 1.9% for sham adrenalectomized, and 3.5% for adrenalectomized, was somewhat below the incidence reported by others of approximately 5-14%(20). In contrast, a significantly increased incidence of isolated CP (4.1%) was observed in the offspring of adrenalectomized dams.

Unlike CLCP, isolated CP is a relatively infrequent defect in A/J mice, with a reported incidence of 0.25% for newborns(21) and 1% for fetuses taken before birth(14, 22). It is considered to be etiologically distinct from CLCP and is the major defect associated with corticosteroid teratogenesis in this strain(5, 6, 20). If endogenous maternal corticosterone were a contributing factor to the incidence of isolated CP, one would expect the incidence of this defect to be greatest in offspring of sham adrenalectomized dams as these animals would have the highest levels of corticosterone produced in response to the stress of surgery and handling. Instead, we found that although isolated CP was increased in the sham adrenalectomized group, the incidence of isolated CP was highest in offspring of the adrenalectomized mice. Therefore it is unlikely that endogenous maternal corticosterone contributed to the incidence of isolated CP in the A/J mice. Rather, the increased incidence of isolated CP observed in offspring of some mice may have resulted from an adverse effect of the surgery as both sham and adrenalectomized groups were effected. Both groups also had significant growth retardation when compared with controls. The ability of the secondary palate to fuse in mice is influenced by both genetic and environmental factors including the time of shelf elevation in relation to the growth of the head, shelf width and the ability of the shelves to fuse once elevation has occurred(23). It is likely that trauma from surgery influenced one or more of these parameters and delayed palate closure to the point where the palate was no longer capable of closing in a number of offspring.

The second series of experiments was designed to assess whether endogenous corticosteroids contribute to the incidence of CP induced by an exogenous corticosteroid. CD-1 mice were used in these experiments as this strain has a susceptibility to cortisone-induced CP similar to that of the A/J strain(11) but was less susceptible than the A/J mice to the maternal complications resulting from surgeries performed early in gestation. Because of this, adrenalectomies could be performed well before palate development was initiated. Adrenalectomy of pregnant CD-1 dams on 6.5 d p.c. resulted in an almost complete removal of endogenous corticosterone; by 12.5 d p.c. the levels of maternal plasma corticosterone remained at or below detectable levels. No difference was observed in the incidence of CP in the offspring of control, sham adrenalectomized or adrenalectomized dams. Sham adrenalectomized dams injected with saline had significantly increased corticosterone levels compared with those of pregnant dams that were only subjected to blood sampling and yet no CP occurred in this group. Adverse effects on fetal development seen in the A/J mice in response to surgery were not observed in CD-1 mice, possibly because adrenalectomies were performed well before palatogenesis was initiated. This finding also indicates that significantly reduced maternal endogenous corticosteroids do not adversely effect fetal development in this strain.

Elevated maternal endogenous corticosterone levels have been implicated as a potential mechanism by which teratogenic agents such as PHT, haloperidol, 2,4,5-trichlorophenoxyacetic acid, and others induce CP in mice. These agents elevate maternal plasma corticosterone and the strain susceptibility of mice to CP induced by these agents correlates with corticosteroid-induced CP(7, 8, 24). However, adrenalectomy in pregnant CD-1(19) or Swiss Webster(25) mice does not reduce the incidence of PHT-induced CP in offspring when compared with mice that had not been adrenalectomized. These studies demonstrate that PHT is capable of inducing CP in the absence of endogenous corticosteroid production. The level of maternal plasma corticosterone reportedly produced in response to a teratogenic dose of PHT in CD-1 or A/J mice(8, 19) is higher than the levels reported for other drugs(8) but is similar to the value(approximately 1000 ng/mL) we observed for sham adrenalectomized CD-1 mice injected with saline; CP was not observed in the offspring of these mice. Although the level of endogenous corticosteroid in the latter group may have been transiently elevated in response to stress from the saline injection 5 h before sampling and levels induced by drug administration may persist longer, taken together the studies indicate that this concentration of maternal endogenous corticosterone is not teratogenic. Thus the primary mechanism of action of these drugs is probably not related to maternal plasma corticosterone levels.

Results from the present study also demonstrate that the levels of endogenous maternal corticosterone present in pregnant CD-1 mice are not high enough to affect the dose response of exogenous corticosteroid-induced CP. Both adrenalectomized and sham adrenalectomized pregnant CD-1 dams were injected with various doses of cortisone during the time of palatal closure. No difference was observed in the susceptibility of fetuses from either group to cortisone-induced CP. Thus the concentration of endogenous maternal corticosterone in sham adrenalectomized mice is not high enough to influence the induction of CP by exogenous corticosteroids.

Although these studies were not designed to define the highest no-effect dose, they confirm that cortisone, like other teratogens, has a no-effect level. The incidence of CP in offspring of sham adrenalectomized or adrenalectomized dams injected with 12.5 mg/kg of cortisone was not significantly different from the spontaneous incidence in controls. However this dose was high enough to suppress maternal corticosterone secretion in sham adrenalectomized dams, presumably due to a negative feedback of the pituitary-adrenal axis. A dose of 25 mg of cortisone/kg or greater was teratogenic and the incidence of CP increased with increasing dose.

Because cortisone is a more potent inducer of CP than corticosterone(26), it follows that a higher concentration of corticosterone would be needed to induce the same incidence of CP as cortisone. Given the present data, it is unlikely that maternal endogenous corticosterone can reach teratogenic levels in the mouse. Although Hieberget al.(28) reported that high doses of ACTH given during pregnancy to stimulate endogenous corticosterone secretion resulted in an increase in the incidence of isolated CP in A/J mice (9.2%), ACTH was also embryotoxic at the doses used. Other studies failed to demonstrate an increased incidence of CP resulting from ACTH administration to either A/J(27) or CD-1 mice(26). In the present study we observed that A/J mice are susceptible to isolated CP in the absence of maternal endogenous corticosteroids when subjected to trauma/stress such as that resulting from surgery. Thus, factors other than corticosterone secretion cannot be rejected when interpreting the ACTH induced CP reported by Hieberg et al.(28).

The results from these studies using A/J and CD-1 mice clearly show that cortisone, and presumably other glucocorticoids, have a no-effect level for CP induction. We have also shown that corticosterone does not contribute to the spontaneous incidence of this malformation in the highly susceptible A/J mouse strain. These studies support the concept of a threshold in the dose-response relationship for corticosteroid-induced CP in mice. This concept is extremely valuable as any exception to the threshold concept would considerably limit our ability to assess potential teratogenic risk of environmental agents.