Main

In clinical practice, synthetic corticosteroids (such as dexamethasone and betamethasone) are widely used in the perinatal period to enhance lung maturation. Antenatal corticosteroid therapy has proven to significantly reduce morbidity and mortality from the respiratory distress syndrome and periventricular hemorrhage(1,2), but when started after birth, the effects of this therapy are equivocal [for review, see Bos et al.(3)]. Well known side effects of corticosteroid treatment are hypertension and left ventricle hypertrophy(4,5) and metabolic effects such as hyperglycemia and increased plasma levels of amino acids, the latter possibly being the result of an increased catabolism(6,7).

Perhaps even more alarming are the effects on CNS functioning as demonstrated in both the fetus and the neonate. After maternal betamethasone treatment (24 mg in two doses, with an interval of 24 h), considerable reductions were observed in fetal body movements, breathing activity, and fetal heart rate variability(8–10), whereas the side effects of dexamethasone (24 mg in two doses) were found to be less dramatic(10). In the neonate, immediate effects of dexamethasone treatment have been described on the heart rate and its variation as well as on the quantity and quality of general movements. Decreased speed and amplitude of the extremity and trunk movements during these general movements indicate a compromised CNS(11,12). These effects might be explained by down-regulation of the corticoid receptor expression and changed responsiveness of these receptors in the CNS(13). However, long-term follow-up studies up to the age of 12 y failed to show adverse effects of antenatal corticosteroid therapy on physical growth or psychomotor and cognitive development(14–16).

In animal research it has repeatedly been reported that steroid treatment during neuro-ontogeny leads to abnormalities in brain development. In fetal monkeys, dexamethasone treatment leads to a decreased arborization of pyramidal neurons in the hippocampus(17). Research in rats and mice has demonstrated that corticosteroid administration leads to a decreased proliferation of neural and glial elements, to a retarded myelination, and to increased cell death(18,19). In addition, abnormal motor behavior, such as hyperactivity and impaired passive avoidance reactions, were demonstrated after postnatal administration of corticosterone from the 2nd until the 14th d after birth(20). In another study, rats became hyperactive after a single dose of dexamethasone at the 7th d(21). However, no data are presently available on the development of motor behavior after corticosteroid administration. Therefore, we decided to systematically investigate in the rat the development of motor behavior and neurologic reflexes after corticosteroid treatment. As it was reported that important differences occur after treatment with dexamethasone in comparison with betamethasone(10), we studied two groups of rats treated with either betamethasone or dexamethasone and compared their development with that of a group of control rats. Rats were injected at the 3rd and 4th postnatal d with doses that equal those normally given in clinical practice. At these ages their stage of brain development is comparable to that in human fetuses or preterm babies at around 7-8 mo postmenstrual age(22). The rat's motor development is relatively well documented(23–27), and in this study, we particularly concentrated on the development of walking. In addition, we studied a few aspects of their neurologic development. The emergence or disappearance of reflexes was previously studied in normal and undernourished rats(28,29). Both of these aspects were studied from the 4th until the 21st d at which age rats walk in an adult-like fashion, and important transformations have taken place(30,31).

METHODS

Rats of the white and black hooded Lister strain were studied. One female rat was mated to two male rats, and inspection for new litters occurred twice daily. The day of birth was referred to as the first day (d 1).

Rats were injected intraperitoneally both on the 3rd and 4th postnatal d either with dexamethasone (0.20 mg/kg, 0.04 mg/mL in saline; injection Formularium Dutch Pharmacists), or with betamethasone (0.20 mg/kg, 0.03 mg/mL in saline, Schering-Plough). In a group of control animals we injected 0.5 mL of saline.

In the initial phase of these experiments a dose of 0.20 mg/kg bethamethasone was lethal to all animals from two different litters. This led us to study the lethality of corticosteroids. For betamethasone, the dose-response effects were studied of 0.10 mg/kg (which is 50% of the dosage applied initially), 0.125 mg/kg, 0.15 mg/kg, and 0.175 mg/kg. As none of these experimental animals died, we then returned to the original dosage of 0.20 mg/kg and increased it in other experiments to dosages of 0.40 mg/kg and 0.60 mg/kg. The majority of the animals died after 0.40 mg/kg at the 5th and the 6th d but remarkably none after 0.20 or 0.60 mg/kg at these ages. Therefore, we returned to the original dosage of 0.20 mg/kg betamethasone. None of the animals died in this experiment. Dosages of dexamethasone of 0.40 mg/kg and 0.60 mg/kg at the 3rd and 4th d did not lead to death of any of the animals (see Table 1 for details). We have no explanation for the lethal effects of injections of 0.20 mg/kg betamethasone at these ages in the first set of experiments.

Table 1 Survival rates after betamethasone or dexamethasone injections at the 3rd and 4th postnatal d
Full size table

All three groups consisted of 12 rats, those injected with dexamethasone or betamethasone from three litters each and the control group injected with saline from two litters. Litter size in all groups was limited to eight pups. The rats used for this investigation were marked by a nontoxic ink until the 10th d, and after this age they could be individually recognized by their black and white pattern. This enabled to us to distinguish the experimental rats in all three groups from their untreated littermates.

The rats were investigated each day from the 4th until the 21st d, and the ambient temperature was kept at 23°C. The rats were weighed daily, and the age at eye opening was documented. Several reflexes (grasp reflex, hopping reaction, tactile placing, negative geotaxis, and free-fall righting) were tested (Table 2).

Table 2 Reflexes included in the neurologic examination
Full size table

In addition, the locomotor performance of the rats was recorded during sessions lasting 3-5 min. To this, the rats were placed in a Perspex walking alley (100 × 15 × 30 cm high) with a mirror underneath. The lateral and ventral views (via the mirror) of their locomotor movements were recorded simultaneously (Panasonic type F11 VHS video camera with stroboscopic shutter at 25 frames/s and Panasonic AG 6200 videotape recorder). During several playback runs of the video tape we evaluated the stage of their locomotor abilities and the presence of any abnormalities. Attention was particularly directed toward crawling with either the forepaws alone or with all four paws, locomotion with the ventral body surface in contact with the floor or free from the floor, posture of the trunk during walking, presence or absence of postural tremor, a staggering or fluent walking pattern, hindlimb abduction or adduction during the stance phase of the step cycle, absence or presence of horizontal and vertical head movements while walking [for further details, see Westerga and Gramsbergen(27)].

Transitions in reflexes were characterized by the age at which the cumulative distribution of rats with the changed reactions reached the 50% level [(see also Smart and Dobbing(29)]. We also calculated the 10% and the 90% levels of these distributions to characterize the speed of the transition. A similar calculation characterized the age at eye opening. The data on neurologic and physical development in the three groups were per day subjected to nonparametric statistical testing(Mann-Whitney U test, Fisher exact test).

RESULTS

Physical development. Mean body weights were similar at the 4th d in the three groups of rats (Fig. 1). From this day onward, the weight increases of the rats in both groups injected with corticosteroids were less than those in the control group. From the 7th d the mean weights in both experimental groups were on average 20-30% lower than those in the control group. Statistical testing indicated that the weights of both the dexamethasone- and betamethasone-treated rats differed significantly from control rats from the 5th through the 21st d (Mann-Whitney U test, p < 0.01). Differences between the weights of the betamethasone- and the dexamethasone-treated rats were not statistically significant.

Figure 1
figure 1

Body weight increases (mean and SD) in betamethasone- and dexamethasone-treated rats and in control rats;arrows indicate ages at injections. Vertical axis, weight in grams; horizontal axis, age in days.

Full size image

Eye opening in normal rats occurred at the age of d 14.9 (10-90% range, d 13.9-15.9) (Fig 2). Remarkably, in the group of dexamethasone-treated rats, eyes opened as early as d 13.8 (range, d 12.6-15.9), which was 1 d earlier than in the control rats. The differences between dexamethasone-treated and control rats were significant at the 14th d(Fisher's exact test, p < 0.02). In the betamethasone group, the eyes opened at the normal age (d 14.9; range, d 13.5-15.4), and at the 14th d the results in this group differed significantly from the dexamethasone group (Fisher's exact test; p < 0.01).

Figure 2
figure 2

Age at eye opening in rats treated with betamethasone or dexamethasone and in control rats. Bars, percentages of rats with opened eyes; horizontal axis, age in days.

Full size image

Neurologic testing. The palmar and the plantar grasp reflex were positive in all three groups at the 4th d, and they remained positive until the 21st d (Table 3). The same findings were obtained for the tactile placing response for fore- and hindpaws as well as for the hopping response for the forepaws. The hopping response for the hindpaws was slightly advanced in the dexamethasone and betamethasone groups. In both groups, the median age at which this response could be elicited was d 4.2 (90% values, d 5.4 and 5.2, respectively), whereas in control rats this was the case at d 5.0 (90% value, d 6.2). However, these differences were not statistically significant.

Table 3 Results of the neurologic examination (ages in days)
Full size table

In contrast, the negative geotaxis reaction was retarded in both the betamethasone- and dexamethasone-treated groups. In the control rats this reaction became positive in 50% of the cases at or before the 4th d(Fig. 3; because our research started at this age, no data on earlier ages were available). The development of this reflex was retarded by at least half a day in rats of the betamethasone group (in 50% of the rats positive at d 4.5). The retardation in the dexamethasone-treated rats even amounted to at least 1.5 (in 50% of the rats positive at d 5.5). The differences between the betamethasone-treated rats and control rats were significant at the 4th day (p < 0.05); those between the dexamethasone-treated rats and control rats were statistically significant at the 4th and 5th d (p < 0.02; Fisher's exact test). Also the development of the free-fall righting reflex was retarded after steroid treatment (Fig. 4). In the control rats, this reflex became positive at the 15th d (range, d 14.7-17.9). In the betamethasone group, 50% landed on all four feet at the age of d 18.2 (range, d 15.7-19.9), and in the dexamethasone group this occurred only at the age of d 18.7(range, d 16.7-20.3). The differences between the betamethasone group and the control group were statistically significant on the 15th, 16th, and 17th d(p < 0.05); those between the dexamethasone group and the control group at the 15th (p < 0.05), the 16th, 17th, and 18th d (p < 0.02; Fisher's exact test).

Figure 3
figure 3

Development of the negative geotaxis response in betamethasone- and dexamethasone-treated rats and in control rats. Bars, percentages of rats presenting a positive reflex;horizontal axis, age in days.

Full size image
Figure 4
figure 4

Development of the free-fall righting response in betamethasone- and dexamethasone-treated rats and in control rats. Bars, percentages of rats presenting a positive reflex;horizontal axis, age in days.

Full size image

Development of locomotion. The control rats crawled with their ventral body surface in contact with the floor at the 4th d. During this type of locomotion the ventral body surface and the chin of the head slid over the floor of the alley. In two control rats from the 8th d and in 10 out of the 12 rats from their 9th d the belly was lifted from the floor during the mid-stance phase of either hindpaw. Because at these ages the head still remained in contact with the floor, this lifting of the hindquarter of the body gave walking a staggering and "bumping" character. This pattern continued until the 10th or 11th d, but on the following days, the ventral surface was kept off the floor for longer periods, and the rats also kept the head from the floor. From the 13-14th d, control rats were observed standing still for the first time with the belly off the floor, and from the 14th d rats groomed while sitting on the hindpaws or reared (standing on the hind feet only) with or without support of the wall of the cage for short periods. Head movements in the vertical plane were frequently observed at this age. From the 15th d walking speed increased, and locomotion was remarkably fluent compared with that on earlier ages. The back remained straight throughout the step cycle and the hind legs remained adducted during the stance phase, but at the 15th d the feet were still exorotated. From the 16-17th d the feet were aligned to the longitudinal axis, walking became variable in its speed, and no further changes in this adult-like pattern could be noted until the 21st d. No postural tremor was noted at any age in the control rats.

Although the development of walking in the betamethasone and dexamethasone groups roughly followed the normal time schedule, certain aspects were distinctly abnormal. Locomotion with the ventral body surface off the floor occurred in six rats of the betamethasone group (50%) and in three rats of the dexamethasone-treated group (25%) as early as the 8th d. Although this development tended to be somewhat advanced compared with control rats, the pattern of walking on this day and the following days was distinctly less elegant. Particularly, rats of the betamethasone group walked staggeringly until the 14th d and two rats as late as the 15th d (control rats until the 10-11th d). Unlike in control rats, the back of treated rats was markedly curved in the sagittal plane, and the trunk was lifted high during the stance phase of either hindpaw. These same characteristics were observed during walking in the dexamethasone-treated rats, although less pronounced and only until the 12-14th d. A remarkable finding in both groups of corticosteroid-treated rats, was a postural tremor with a low frequency during walking from the 10th until the 15th d and in four rats until the 16th d. The tremor involved axial muscles in head and trunk and waned when the rats stopped moving. The postural tremor, the irregular stride length, and the clearly insecure gait gave the walking pattern an "atactic" appearance in both groups. From the 15-16th d, locomotion became more fluent and hindpaws were adducted. However, in both groups of rats, the hind feet remained markedly exorotated until the 20th d, leading to an increased gait width. Movement patterns such as grooming without support and rearing developed in the corticosteroid-treated rats at similar ages compared with control rats.

In summary, the development of walking essentially followed similar temporal patterns in the three groups of rats. However, rats in the betamethasone and dexamethasone groups walked abnormally, they staggered and walked atactically until the end of the 2nd wk and thereafter, and until the 20th d their gait width was increased. These abnormalities were particularly pronounced in the betamethasone-treated rats and occurred also but less prominently in the dexamethasone-treated rats.

DISCUSSION

Based on several indices of brain development, Romijn et al.(22) concluded that the rat's brain development at the 8th postnatal d is comparable to that of the human baby at term. We injected rats with betamethasone or dexamethasone at the 3rd and 4th postnatal d. Extrapolated, these ages can be considered analogous to the stage of brain development of fetuses or prematurely born babies at 27-34 wk of gestational age. At that stage, both in humans and in rats, e.g. the proliferation and migration of neurons in the cerebral hemispheres is almost finished(32–34), and the proliferation of the extragranular matrix of the cerebellum has just started in both species(35).

In our investigation, we observed a growth retardation in the groups of rats treated with corticosteroids. The mean body weights of both groups from the 10th d were around 20-30% lower than those in the control group, and these differences were statistically significant from the 7th d. In the dexamethasone-treated group of rats, a catch-up seemed to occur after the 17th d, but the differences between both corticosteroid-treated groups did not reach significance. Growth retardation after postnatal corticosterone administration in rats was reported previously(18,20). These results indicated a 25-30% decrease, which is in agreement with our data.

A striking observation was that eye opening in the dexamethasone-treated rats was advanced by 1 d. No such acceleration was observed in the betamethasone-treated rats, although a few rats in this group opened their eyes earlier than did control rats. The clinical motivation for administering adrenocortical steroids is to accelerate lung maturation(36). Slotkin et al.(37) observed in rats that the development of the norepinephrine turnover after prenatal administration of dexamethasone was advanced during the weaning period. These data on developmental processes that are speeded up, whereas other processes develop abnormally after corticoid administration at early stages point to a complicated influencing of metabolic processes after steroid administration during development(17–19).

Most of the neurologic tests were positive by the 4th d, and they remained so thereafter. Interestingly, the hopping response of the hindpaws in both corticosteroid-treated groups became positive as early at d 4.2 (in 50% of the rats), which is advanced by almost 1 d. This result might indicate that the maturation of spinal mechanisms involved in rhythmic stepping responses is accelerated.

Two other responses, however, were retarded. The negative geotaxis response by about 0.5-1.5 d and the free-fall righting response by even more than 3 d. Both responses have been classically attributed to vestibular functioning. Undoubtedly, the vestibulum plays an important role as the sensory system which induces these reactions. However, both reactions require complicated patterns of muscle contractions (this holds particularly for the free-fall righting reaction), in which cerebellar functioning is involved. Previously, it was demonstrated that the cerebellar development is disturbed after corticosteroid treatment(18,21). Particularly from the 14th d, the cerebellum plays an important role in the regulation of complex movement patterns(38,39). Apart from vestibular involvement, therefore, abnormal functioning of the cerebellum also might be a causal factor underlying a delayed development of these responses.

The development of locomotion was not changed as far as its temporal aspects is concerned. The onset of lifting the ventral body surface from the floor during part of the step cycle (at the 8-9th d) occurred at the normal age, and the age at which the adult-like walking pattern with adducted hind legs emerged (at the 15-16th d) was also not retarded. However, remarkable deviances became apparent, particularly in the group of betamethasone-treated rats and less prominently in the dexamethasone-treated rats. During walking from the 9th d the back was curved in the sagittal plane, a marked postural tremor occurred in the majority of the cases, and the gait often showed atactic traits until the 14-15th d. In both groups of corticosteroid-treated rats the gait width was abnormally increased due to a striking exorotation of the feet of the hind legs.

Rhythmic hind leg movements normally occur from before birth, and they last until after the 2nd wk of life before the adult type of fluent walking develops(26,27). This development is dependent upon the emergence of feet-forward control of posture(30). Probably, the cerebellum plays an important role in this coupling of walking and postural control, as cerebellar hemispherectomy at early postnatal ages leads to an atactic gait and consistent hindpaw abduction(38). In this perspective, it seems plausible that the abnormalities in walking pattern and the retarded development of the free-fall righting response might be induced by an abnormal cerebellar maturation(18).

In conclusion, the results of the present investigation indicate that both betamethasone and dexamethasone, injected into young rats at a maturational stage which is comparable to that of prematurely born human babies at 27-34 wk postmenstrual age, lead to neurologic abnormalities. These abnormalities include a retarded development of vestibular reflexes and abnormalities during the development of walking. Although these abnormalities in a critical period of motor development seem transient, they may have an impact on the timely acquisition of motor skills. Further experimental research on the long-term effects of corticosteroid treatment on neurologic development is needed, and this research should also expand to the effects of treatment at earlier maturational stages. In the light of the present results and those of other animal studies, it seems justified to advise cautiousness regarding the administration of these corticosteroids in clinical practice.