Prenatal ethanol exposure: Effects on androgen and nonandrogen dependent behaviors and on gonadal development in male rats

Neurotoxicologyand Teratology,Voi. 16, No. 1, pp. 31-39, 1994 Copyright©1994ElsevierScienceLtd Printedin the USA.All riots reserved 0892-0362/94 $6.00 + .00

Pergamon

Prenatal Ethanol Exposure: Effects on Androgen and Nonandrogen Dependent Behaviors and on Gonadal Development in Male Rats B E T T Y A . B L A N C H A R D .1 A N D J O H N H . H A N N I G A N t

*Department o f Anatomy and Neurobiology, School o f Medicine, University o f Missouri, Columbia, MO 65212 ~fFetal Alcohol Research Center, C. S. Mott Center for Human Growth and Development, Departments o f Obstetrics & Gynecology, and Psychology, Wayne State University, School o f Medicine, Detroit, All 48201 Received 20 A u g u s t 1992; Accepted 27 July 1993 BLANCHARD, B. A. AND J. H. HANNIGAN. Prenatalethanol exposure: Effects on androgen and nonandrogen dependent behaviors and on gonadaldevelopment in male rats. NEUROTOXICOL TERATOL 16(1) 31-39, 1994.-Prenatal exposure to alcohol may impair gonadal and behavioral development in male rats, possibly via reduction of perinatal androgenization. We examined locomotor activity on postnatal day 18 (PND 18), which is not influenced by perinatal androgens and juvenile play and testicular development (testes weight), which are dependent on perinatal androgen exposure, in rats whose dams consumed ethanol during pregnancy. Male offspring of pair-fed and lab chow-fed dams served as controls. Despite reduced anogenital distance at birth, indicating compromised perinatal androgenization, fetal ethanol-exposed males did not exhibit demasculinization of play behavior. Hyperactivity in fetal ethanol-exposed males indicated that the treatment regimen was sufficient to produce behavioral deficits. Testes weight was reduced in both ethanol-exposed and pair-fed offspring, indicating that nutritional deficits associated with maternal ethanol intake may impair normal gonadal development in male rats. The findings suggest that fetal ethanol exposure may influence gonadal development but not necessarily affect a gonadal hormone-dependent behavior. Fetal alcohol exposure Juvenile play

Testosterone

Anogenital distance

CHRONIC alcohol exposure has a suppressive effect on gonadal growth and function in adolescent and adult males, resulting in testicular atrophy and decreased testosterone (T) production in several mammalian species, including humans, rats, and mice (2,14,34,35). Similarly, McGivern et al. (22) showed that prenatal alcohol exposure reduced T levels of fetal male rats on gestation day 18 (GD 18), an age at which normal male fetuses show a surge in T concentration (38). Redei and McGivern (31) also found that a second surge in T levels that normally occurs just after birth in male rats (11) was reduced by prenatal ethanol exposure. Because adequate perinatal androgen exposure is essential for the organization

Demascufinization

Hyperactivity

of normal masculine neurobehavioral development (4,5), ethanol-induced suppression of T production by the fetal testes may be a mechanism by which alcohol exerts some of its teratogenic effects. Treatments other than alcohol that disrupt the level or timing of perinatal androgen exposure (e.g., neonatal castration, pre- or postnatal anti-androgen treatment, prenatal stress) have been shown to compromise subsequent neurobehavioral development (4,5,37,38). Several studies examining the effects of prenatal ethanol exposure in rats indicate that some of the effects seen in males are similar to those produced by other treatments that reduce perinatal androgen levels. Fetal etha-

1 Requests for reprints should be addressed to Betty A. Blanchard, Department of Anatomy and Neurobiology, School of Medicine, University of Missouri, Columbia, MO 65212 31

32

BLANCHARD AND H A N N I G A N

nol-exposed male rats have been reported to show demasculinization a n d / o r feminization of both reproductive and nonreproductive sex-influenced behaviors and of sex-influenced brain structures. For example, male rats exposed prenatally to ethanol were reported to show feminized saccharin consumption (20; but cf. 1,7) and juvenile play behavior (26); feminized (19) and demasculinized sexual behavior (29,33); feminized a n d / o r demasculinized neuroanatomy (3,32,39,40). In addition, fetal ethanol-exposed male but not female rats showed altered psychopharmacological profiles relative to controls (9,16). The experiments reported here represent the initial stage of a larger study (8); the goal was to examine the efficacy of neonatal androgen replacement in alleviating fetal ethanolinduced deficits in androgen-dependent juvenile play behavior in male rats that have been reported previously (26). In rodents, juvenile play is sexually dimorphic, with males showing greater levels of play than females (6,23,24,27). The sex difference in play is thought to be entirely dependent on perinatal androgens, since manipulations o f androgen levels (e.g., castration of males) after about PND 10 are ineffective in altering levels of social play (6). It has been suggested that fetal ethanol exposure alters play behavior in male rats by suppressing testosterone production during the perinatal period (26). Before attempting to manipulate play behavior in prenatal ethanolexposed male rats by administering supplemental testosterone, we wished to establish prenatal ethanol effects on juvenile play behavior. Although we observed a number o f fetal ethanol effects, including hyperactivity, we found that juvenile play behavior was unchanged in fetal ethanol-exposed male rats. METHOD

Breeding Procedures Long-Evans rats (Blue Spruce Farms, Altamont, NY) maintained on a 12L : 12D cycle (lights on at 0700 h) with free access to food and water were mated overnight. Females with vaginal plugs the following morning were weighed, housed individually in plastic breeding cages, and assigned to one of three prenatal treatment groups. The ethanol group (ETOH; n = 31) received free access to a liquid diet consisting of 35% ethanol-derived calories on days 6-20 of pregnancy. A second, pair-fed group (PF; n = 31) received liquid diet with sucrose substituted isocalorically for ethanol and served as a nutritional control for the effects of reduced caloric intake associated with the ethanol diet. Each animal in the PF group received the same dally caloric intake on a body weight basis as a yoked ETOH rat. On pregnancy day 5, both liquid diet groups received 30 ml liquid diet without ethanol, in addition to lab chow and water, to accustom animals to the novel flavor. A third, chow-fed control group (CONT; n = 27) received ad lib lab chow and water throughout pregnancy. Pregnant animals were weighed on days 6, 10, 15, and 20 to monitor weight gain and facilitate accurate pair-feeding. The liquid diets consisted o f chocolate Sustacal (Mead Johnson, Inc.), water, Vitamin Diet Fortification Mixture, and Salt Mixture XIV (ICN Nutritional Biochemicals), to which either 95% ethanol or sucrose was added. The liquid diets provided approximately 1.3 kcal/ml and, except on Day 5, were the sole source of nutrition during the period they were administered. Data from this lab have shown that animals consuming the ETOH diet reached mean peak blood alcohol levels o f about 130-150 mg/dl.

On the day following birth, designated P N I , litters were sexed, weighed, inspected for gross physical anomalies, and culled to 10 pups, maintaining a 1 : 1 sex ratio when possible. Litters containing fewer than six animals in total or fewer than three males were not used for this study. On PND I, anogenital distance (AGD) of males was measured as an indicator of relative prenatal androgenization. Nose-to-rump body length (cm) was also assessed on PND 1. DAM AND LITTER DATA Gestation length, percent weight gain by dams during pregnancy, litter size, and sex ratio were analyzed by analyses of variance (ANOVAs), with maternal diet (CONT, PF, ETOH) as the factor. Litter means served as the units of analysis for pup data on PND 1. PND 1 body weight and anogenital distance were analyzed by ANOVA with prenatal treatment (maternal diet) as the factor. Significant effects were examined further using Newman-Keuls pair-wise comparisons (p < 0.05).

Maternal Data Maternal data are presented in Table 1. ETOH dams consumed a mean of 13.05 (+0.25) g/kg ethanol daily. Diet significantly influenced percent weight gain during pregnancy, F(2, 86) = 36.01, p < 0.0001. Comparisons indicated that animals that received liquid diet (ETOH or PF) gained less weight during pregnancy than CONT dams. Dams in the two liquid diet groups did not differ from each other on this measure. Diet had no effect on gestation length as measured, litter size, or sex ratio.

Pup Data Pup data at birth are presented in Table 1. Birth weight was influenced significantly by prenatal treatment, F(2, 86) = 58.37,p < 0.001. Comparisons indicated that ETOH pups weighed less than PF and CONT pups, and P F pups weighed less than CONT pups. Body length was also influenced by prenatal treatment, F(2, 86) = 12.86, p < 0.0001. ETOH pups were shorter than PF and CONT pups, which did not differ from each other. Prenatal treatment significantly affected anogenital distance, F(2, 86) = 49.62, p < 0.0001. Comparisons indicated that ETOH males had smaller anogenital distance than CONT and PF males, which did not differ from each other. However, because ETOH males were also smaller than controls, it was possible that treatment effects on anogenital distance were related to differences in body weight a n d / o r body length, rather than to a direct effect of fetal ethanol exposure on anogenital distance. When corrected for body weight, ETOH rats had a significantly higher anogenital index than CONT or PF males, F(2, 86) = 6.05, p < 0.01, but when corrected for body length, ETOH rats had a lower anogenital index than CONT or PF animals, F(2, 86) = 7.59, p < 0.01. To examine the relationships among prenatal treatment, body weight, body length, and anogenital distance, we used LISREL analysis (30) to evaluate several causal models. Preliminary analyses indicated that ethanol effects on body length did not contribute significantly to variation in anogenital distance. We then examined three hypothetical causal models o f the relationships among prenatal ethanol exposure, body weight, and anogenital distance (Fig. 1): (a) fetal ethanol exposure independently influenced body weight and anogenital distance; (b) fetal ethanol exposure indirectly influenced ano-

PRENATAL ETHANOL EXPOSURE

33 TABLE 1 DAM AND LITTER CHARACTERISTICS Control

Number of dams Mean daily EtOH intake (g/kg) Gestation length(days) Percent weight gain during pregnancy Litter size (# pups) Litter sex ratio (proportion males) PNlhodyweight(g) PNlbodylength(cm) PNI anogenital distance (mm) PN1 anogenital distance/body weight PN1 anogenitaldistance/bodylength PNI8 body weight (g) PN36 body weight (g) Adult body weight (g) Adult testes weight (g) Testes weight (g)/100 g body weight

Pair-fed

27 21.96 ± 0.04 51.78 ± 1.19 11.96 + 0.50 0.49 7.69 ±0.09 6.15 ±0.05 4.63 + 0.08 0.610 ± 0.002 0.756 ± 0.010 43.24 ± 1.04 152.54 ± 4.01 441.71 ± 8.17 3.797 ± 0.09 0.851 ± 0.014

31 22.06 + 0.33 38.43 ± 1.37" 12.81 ± 0.27 0.56 7.10 ±0.11" 6.01 +0.07 4.50 ± 0.07 0.632 + 0.002* 0.749 ± 0.013 39.58 + 1.03" 146.50 ± 5.30 431.08 ± 8.29 3.316 + 0.09* 0.779 ± 0.020

ETOH 31 13.05 ± 22.03 ± 35.97 ± 11.84 ± 0.52 5.97 ± 5.69 ± 3.91 ± 0.661 + 0.691 ± 37.532 ± 143.40 ± 416.75 ± 3.337 ± 0.825 ±

0.25 0.03 1.52" 0.40 0.12~/ 0.08:~ 0.10~ 0.005~ 0.015~/ 0.66* 3.38 14.23 0.13" 0.036

Dam and litter data for male rats from the three prenatal treatment groups. *Significantly different from CONT rats; ~/Significantlydifferent from CONT and PF rats.

genital distance by affecting body weight but had no effect on anogenital distance independent of body weight; (c) fetal ethanol exposure influenced anogenital distance both directly and indirectly by reducing body weight. LISREL provided two measures of how well these models predicted the observed correlations among prenatal treatment, body weight, and anogenital distance: (a) the Goodness of Fit Index (GFI), ranging from 0.0 to 1.0, with 1.0 representing a perfect fit; (b) a X2 test. A significant X2 indicated that the correlations predicted by the model are significantly different from the observed correlations; the X2 for a model that fits the data should be nonsignificant. GFI's were 0.879, 0.939, and 0.990 for Models 1, 2, and 3, respectively, indicating that Model 3 best fit the observed data. Chi squares for Models 1 and 2 were significant (X2 = 20.92 and 10.64, respectively), indicating that these models differed significantly from the actual data; for Model 3, x 2 = 1.53 (p > 0.05), indicating that this model did not differ significantly from the observed data. These results indicated that the model that best fit our data was the one in which anogenital distance was influenced by prenatal treatment, both directly and indirectly through body weight and by body weight independent of prenatal treatment. Approximately 12°70of the variance in anogenital distance was accounted for by prenatal treatment per se and another 9070 was accounted for by prenatal treatment effects on body weight. These findings indicate that while reduced anogenital distance following prenatal ethanol exposure is due, in part, to the fact that prenatal ethanol exposure results in lower birth weight, there is also a direct effect of prenatal ethanol on anogenital distance that is independent of body weight. Prenatal ethanol's direct and mediated effects accounted for 21070 of the variance in anogenital distance. An additional 27070 of the variance in anogenital distance was accounted for by body weight alone independent of prenatal treatment. EXPERIMENT 1: LOCOMOTORACTIVITY In a previous study, male but not female fetal ethanolexposed rats showed hyperactivity at PND 18 08). We chose

to examine activity as a positive control for fetal ethanolinduced behavioral effects in our pool of rats since it is a relatively consistently reported effect of fetal ethanol (25), and one that we have reproduced several times in our laboratory. Subjects (n = 8-13) drawn from the large pool of animals were weighed and placed for 20 min in a 40 x 40 cm Plexigias box with 9 holes in a 3 x 3 grid pattern in the floor. Ambient temperature was 33 °C. Sessions were videotaped and scored for number of squares entered and number of holepokes. Testing was conducted between 1300 and 1500 h under dim light.

Analyses and Results Body weight, locomotion, and holepokes were analyzed by ANOVA, with prenatal treatment as the factor. Significant effects were further examined using Newman-Keuls comparisons (p < 0.05). Body weight at PN18 was significantly influenced by prenatal treatment, F(2, 28) = 10.10, p < 0.001. Comparisons indicated that ETOH and PF animals weighed less than CONT animals (Table 1). Prenatal treatment significantly influenced squares entered, F(2, 28) = 5.37, p < 0.02, with ETOH animals entering more squares than CONT or PF animals, which did not differ from each other (Fig. 2A). Although there appeared to be a tendency for ETOH and PF rats to make more holepokes than CONT animals (Fig. 2B), the prenatal treatment effect did not reach statistical significance, F(2, 28) = 2.29, p = 0.12. These findings indicated that the maternal alcohol treatment regimen was sufficient to produce behavioral effects typical of fetal ethanol exposure in rats. EXPERIMENT 2: SOCIAL PLAY Because social play behavior of young male rats is under organizational control of androgens (6,23,24), previously reported fetal ethanol-induced demasculinization of play (26) may be related to inadequate perinatal androgen exposure (22,31). Ethanol-exposed males in the subject pool of the pres-

34

BLANCHARD AND HANNIGAN

MODEL 1.

MODEL 2.

DISTANCE

isolation facilitates social play (28). Stimulus animals were nonlittermate CONT males. Stimulus animals were not drawn from all three prenatal treatment groups since a previous study reported that the prenatal treatment of the stimulus animal did not interact with the prenatal treatment of the test animal (26). Approximately 3 hours before testing began at 19:30 (30 min into the dark portion of the light-dark cycle), each test animal was weighed and placed alone into a 20 x 46 × 20 cm high Plexiglas cage with clean wood shavings. Stimulus animals were weighed and placed into an identical cage with 2-4 other CONT stimulus rats. All animals remained in the testing room with food and water until testing began to allow habituation to the room. At testing, food and water were removed from the cage of the test animals and an ink-marked, weight-matched stimulus animal (35-36 days of age) was placed into the cage for 20 min. Sessions were videotaped under red light and scored for the frequencies and latencies of mounts, crossovers and pins made by both the test and stimulus animals. A mount was scored when one animal climbed upon the other from behind. A crossover was scored when one animal jumped over the other. A pin was scored when one animal had its dorsal surface to the floor of the cage with the other animal hovering above it and holding it down.

Analyses and Results

MODEL 3. ~ A ' ~ G ENITA~'L

f~ENATAL~'~

I

Body weight and frequencies and latencies of mounts, crossovers and pins by the test animal and by the stimulus animal were analyzed by ANOVA, with prenatal treatment as the factor. Analysis of body weight indicated that at PN36, ETOH animals no longer differed from CONT or PF animals, F(2, 26) = 1.34, p = 0.28; Table 1. ETOH rats did not display altered levels of play on any measure relative to controls. Totals of each play behavior by the test animal, including pins, mounts and crossovers, are presented in Fig. 3 (n = 8-11). Figure 4 depicts total play behaviors by the stimulus animals. Proportion of animals exhibiting and eliciting play behaviors are presented in Table 2; latencies to first occurrence of these behaviors are presented in Table 3. EXPERIMENT 3: TESTES WEIGHT IN ADULTHOOD

FIG. 1. Hypothetical causal models predicting the possible relationships among prenatal treatment, body weight, and anogenital distance. Model 1: prenatal treatment had independent effects on body weight and anogenital distance. Model 2: prenatal treatment effects on anogenital distance were solely due to treatment effects on body weight, with no effect independent of body weight. Model 3: prenatal treatment influences anogenital distance both directly, and indirectly (mediated through body weight). LISREL analysis indicated Model 3 best fit the observed data.

ent experiment showed decreased anogenital distance at birth that was not solely accounted for by decreased body weight, suggesting lower perinatal androgen exposure (see above). At PND 36, animals (n = 9-11) were tested in pairs consisting of a test animal and a stimulus animal. Test animals were isolated from cagemates 48 h prior to testing since prior social

To determine if our prenatal ethanol treatment regimen had long-term effects on gonadal development, we examined testes weight in adult males. At PND 85-100, testes were removed from naive (i.e., not behaviorally tested) subjects under xylazine and ketamine anesthesia (10 mg/kg xylazine, 50 mg/kg ketamine; 1 ml/kg, IP) and weighed.

Analyses and Results ANOVA indicated that prenatal treatment did not affect adult body weight, F(2, 33) = 0.86, p = 0.43. Testes weight was analyzed by ANCOVA, with prenatal treatment as the factor and body weight as a covariate, since variance in body weight may contribute to variance in testes weight. Further comparisons were made with Newman-Keuls tests (p < 0.05). There was a significant effect of prenatal treatment on testes weight in adulthood, F(2, 33) = 7.81, p < 0.005. Newman-Keuls comparisons indicated that both ETOH and PF animals had smaller testes than CONT animals but did not differ from each other (Table 1). The covariate body weight was not significant nor did it interact with prenatal treatment, indicating that reduced testes weight in PF and ETOH rats was not due to any possible treatment effects on

P R E N A T A L E T H A N O L EXPOSURE

35

Locomotor

activity

Holepokes 16

O.

I'~ICONT PF

80

**

b. 14

il

12 60

10 8

4O .43

E

6

~, 2o

4

2

FIG. 2. (A) Locomotor activity on PNI8 is represented by the mean (+ SEM) number of squares entered. "'Significantly different from CONT and PF animals. (B) Exploratory behavior is represented by the mean (+ SEM) frequency of holepokes.

body weight. Testes weight corrected for body weight (Table 1) did not reveal significant treatment effects.

posed rats was somewhat surprising in view of the reliable fetal alcohol effects on birth weight, behavioral activity and the reduction in anogenital distance at birth, and previous reports (26). The analysis of anogenital distance is important, since previous reports have not been consistent. Some researchers have reported that anogenital distance is reduced by prenatal ethanol exposure (e.g., 29,32,33), while others have reported no effect (19,21,22). While some of the inconsistencies may be

DISCUSSION Male rats exposed prenatally to ethanol showed fetal alcohol effects on some of the measures examined but did not exhibit behavioral demasculinization of juvenile social play. The absence of behavioral demasculinization in ethanol-ex-

Total play behaviors initiated 12 [~I

CONT

~]

PAIR-FED

I

El'OH

11 I0

0

•~,

g

0 ¢'~

8

>,,, o

7

'5

8

E c(-

4

o

3

2 1 m

Pins

Mounts

Cross-overs

FIG. 3. Mean (+SEM) play behaviors (pins, mounts, and crossovers) engaged in by the test animals. There were no significant effects of prenatal ethanol exposure.

36

BLANCHARD AND HANNIGAN Totol ploy behoviors elicited 10 CONT

~

PAIR-FEO

1

ETOH

9 ~

a

0 0 r-

7

..O 0

~

5

-~

4

E C E 0 O)

"~

3 2 1

Pins

Mounts

Cross-overs

FIG. 4. Mean (+SEM) play behaviors (pins, mounts, and crossovers) elicited from the stimulus animals by rats in the three prenatal treatment groups. There were no significant effects of prenatal treatment of the test animal on play behaviors by the stimulus animals.

related to differences in the strain or in the method and timing of gestational ethanol exposure (21), they may be due in part to how anogenital distance data have been analyzed. Those researchers reporting reductions have typically reported uncorrected data or anogenital distance corrected for body length (e.g., 29,32,33). Those reporting no effect have corrected for body weight (e.g., 21,22, but cf. 19), as has been suggested by others for normal rodents (15). Such findings suggest that fetal ethanol effects on anogenital distance are secondary to fetal ethanol effects on body weight. There are, however, some features of correction for either body weight or body length that may make correction inappropriate in the presence of other treatment effects. First, correction assumes an unvarying relationship between body weight or body length and anogenital distance. Second, although it has been reported that body weight and anogenital distance are correlated

in mice (15), body weight and anogenital distance were not significantly correlated in control male rats in the present study. Finally, correcting for body weight may obscure effects if a treatment has independent and/or unequal effects on body weight and anogenital distance. Our findings in the pair-fed animals suggested that anogenital distance and body weight may be influenced via different mechanisms. In pair-fed rats, body weight was reduced relative to controls, but there were no effects on anogenital distance (Table 1). Also, in our prenatal ethanol-exposed animals body weight was reduced by about 21070 and anogenital distance by 15o7o relative to controis (Table 1). Given these findings, we chose to examine further the relationships among these factors. Using LISREL analyses we found that in the causal model that best fit our data, treatment effects on anogenital distance were due to both direct effects of ethanol per se, and indirect

TABLE 2 PERCENT OF ANIMALSENGAGINGIN PLAY BEHAVIOR Test Animals

StimulusAnimals

Prenatal Treatment of Test Animal

CONT

PF

ETOH

CONT

PF

ETOH

Pins Mounts Crossovers At least one play behavior

54070 81% 91 07o 90070

63070 63% 88070 87070

60% 80070 70070 80070

72°7o 54% 45070 82070

75% 63% 7507o 100070

60% 50% 40070 80070

Data for the stimulus animals indicates the percent of test animals from each of three prenatal treatment groups that elicited play behavior from the stimulus animals.

PRENATAL ETHANOL EXPOSURE

37

TABLE 3 LATENCIES FOR ANIMALS TO EXHIBIT THE THREE PLAY BEHAVIORSASSESSED PrenatalTreatmentof Test Animals (Includesall subjects) Test animals Pin Mount Crossover Includes only subjects showing behavior Pin Mount Crossover Stimulus animals Pin Mount Crossover Includes only subjects showing behavior Pin Mount Crossover

CONT

PF

ETOH

651 + 152 382 + 124 262 + 93

651 + 153 527 + 184 300 + 124

752 + 138 338 + 141 491 + 150

193 + 35 200 + 53 169 ± 30

321 + 151 123 ± 19 171 ± 36

454 + 126 122 + 46 187 ± 50

639 ± 148 757 ± 139 787 ± 148

470 ± 151 689 ± 151 500 :l: 162

731 + 134 758 ± 147 865 ± 149

429 ± 145 388 ± 124 291 ± 130

227 ± 38 383 ± 95 266 + 101

419 + 96 316 ± 93 362 + 184

Presented are mean ( + SEM) latencies, for all animals in each group and for only those animals that exhibited each behavior. Data for stimulus animals represent latencies for the test animals from the three prenatal treatment groups to elicit the behaviors from the stimulus animals.

effects mediated through ethanol-induced reductions in body weight. The direct and indirect effects of prenatal ethanol accounted for a total of 21% of the variance in anogenital distance. Ethanol effects on anogenital distance that were secondary to effects on body weight are consistent with the findings of others (e.g., 21,22), who have suggested that fetal ethanol effects on anogenital distance are influenced by ethanors effects on body weight. These findings indicate that there is an additional effect of prenatal ethanol exposure on anogenital distance that is independent of ethanol effects on body weight or length. Others (e.g., 32) have corrected anogenital distance for body length. However, in an additional LISREL analysis incorporating body length into the model, we found that ethanol effects on body length did not contribute significantly to variation in anogenital distance. These results suggest that prenatal ethanol exposure may produce a single effect through multiple pathways. The mechanism by which fetal ethanol exposure influences anogenital distance independent of body weight was not examined in our study, although ethanol suppression of perinatal testosterone production may be important (22,31). The findings suggest that because fetal ethanol exposure had effects on anogenital distance that were both dependent on and independent of body weight, analyzing either uncorrected or corrected anogenital distance may result in erroneous conclusions or in the loss of potentially important information about the factors that influence anogenital distance. This is apparent in Table 1, where both raw anogenital distance data and anogenital distance corrected for body weight and body length are presented. Uncorrected anogenital distance data do not reveal the influence of ethanol-induced body weight deficits on anogenital distance. Data corrected for body weight suggest that, if anything, fetal ethanol exposure increased ano-

genital distance. This is because the prenatal ethanol-induced reduction in body weight was greater than the reduction in anogenital distance. Some researchers have proposed that because anogenital distance and body weight were correlated in mice, anogenital distance data should be corrected for body weight prior to analysis (15). Others have argued that such an approach may provide incomplete information when a treatment independently influences measures such as anogenital distance, organ weight, and body weight and that a more appropriate analysis is ANCOVA, with body weight as a covariate (e.g., 12,13,36). Such an analysis acknowledges the relationship between body size and anogenital distance or organ weight, while examining additional variance that is unaccounted for by body size. We were unable to use this type of analysis for anogenital distance in the present study, because ANCOVA is not appropriate when a factor in the analysis (e.g., prenatal treatment) influences the covariate (e.g., body weight). Because prenatal treatment did not significantly affect adult body weight, we analyzed testes weight by this method and found that prenatal treatment influenced testes weight in adults, independent of any treatment-induced variation in body weight. Testes weight in both ethanol-exposed and pair-fed offspring was reduced, suggesting that prenatal ethanol effects on gonadal development are secondary to ethanol effects on nutritional factors. Such findings may be important, since nutritional status is often compromised in human alcoholism. In addition, these findings suggest that deficits in gonadal development may be present despite recovery from earlier body weight deficits. However, this interpretation is limited by the use of only one measure of gonadal development. Future studies would benefit from the inclusion of additional measures, such as seminal vesicle weights and androgen levels.

38

BLANCHARD AND HANNIGAN

Despite fetal ethanol effects on birth weight, anogenital distance, growth, and locomotor activity, we failed to find demasculinization of juvenile play behavior in fetal ethanolexposed male rats. This was not due to below normal levels of play in our controls but to the fact that fetal ethanol-exposed males showed greater levels of play than previously reported (26). There are specific procedural differences between the present study and that reported by Meyer and Riley (26) that may explain our failure to establish fetal ethanol-induced deficits in play behavior. In the present experiment, all rats were habituated to the testing room before testing and the experimenter was absent from the room, whereas in Meyer and Riley, animals were tested immediately upon removal from the colony, with the experimenter present. In addition, we tested animals during the dark portion of the light/dark cycle, when animals are most active. Meyer and Riley do not indicate when animals were tested. It is unknown how such procedural differences might result in the different findings in the two studies. However, in an earlier study, we found that differences in testing procedures significantly influenced performance of fetal ethanol-exposed animals on the Morris water task, while leaving performance of control animals intact (10). Even in animals that appeared to have recovered from fetal ethanol effects on spatial behavior, deficits were reinstated by exposing animal to stressful stimuli (10). If the testing procedures in Meyer and Riley were more stressful than those in the present study (perhaps due to no habituation period and the experimenter present in the room), then perhaps the decreased play they observed in ethanol-exposed males was due to a competing behavioral response to stress, such as freezing or grooming (17). It is possible that altered levels of play behavior in fetal ethanol-exposed male rats reported by Meyer

and Riley (26) were not related to altered perinatal androgenization but to other factors that are activated during stressful situations. In general, the present findings suggest that there may be ethanol-induced reductions in anogenital distance, a relative marker of perinatal androgen levels, without deficits in social play by juvenile male rats, a perinatal androgen-dependent behavior. Other factors such as testing procedures and environmental stressors may play an important role in the behavior of ethanol-exposed rats. It is not likely that our ethanol treatment was insufficient to produce observable effects, since we replicated decreased body weight, reduced anogenital distance, and increased locomotor activity at PND 18. There are also inconsistent reports of prenatal ethanol effects on other gonadal hormone-dependent behaviors. While altered saccharin preference in rats exposed prenatally to ethanol has been reported by some (e.g., 20), others have reported no effect in rats (1) or mice (7). Reports of altered masculine sex behavior following fetal ethanol exposure have also been inconsistent (19,29,33). Differences in species, strain, and gestational ethanol exposure may contribute to differences among studies; the present study suggests that testing procedures may have some influence as well. ACKNOWLEDGEMENTS This work was part of a doctoral dissertation by B.A.B. at the State University of New York at Albany. We thank Michael Boechler for assistance with LISREL analysis; John DiCerbo, Michelle Pilati, and Margaret Davis-Cox for technical assistance. This work was funded in part by SUNY BenevolentAssociation Research Awards to B.A.B. and an ADAMHA Scientist Development Award (#AA00140) from NIAAA to J.H.H.

REFERENCES 1. Abel, E. L.; Dintcheff, B. A. Saccharin preference in animals prenatally exposed to alcohol: No evidence of altered sexual dimorphism. Neurobehav. Toxicol. Teratol. 8:521-523; 1986. 2. Anderson, R. A.; Willis, B. R.; Phillips, J. F.; Oswald, C.; Zaneveld, L. J. D. Delayed pubertal development of the male reproductive tract associated with chronic ethanol ingestion. Biochem. Pharmacol. 36:2157-2167; 1987. 3. Barron, S.; Tieman, S. B.; Riley, E. P. Effects of prenatal alcohol exposure on the sexually dimorphic nucleus of the preoptic area of the hypothalamus in male and female rats. Alcohol: Clln. Exp. Res. 12:59--64; 1988. 4. Beach, F. A. Hormonal modification of sexually dimorphic behavior. Psychoneuroendocrinol. 1:3-23; 1975. 5. Beatty, W. W. Gonadal hormones and sex differences in nonreproductive behaviors: Organizational and activational influences. Horm. Behav. 12:112-163; 1979. 6. Beatty, W. W. Hormonal organization of sex differences in play fighting and spatial behavior. In: De Vries, G. J. et al., eds. Sex differences in the brain. Progress in Brain Research 61:315-330; 1984. 7. Becker, H. C.; Randall, C. L.; Anton, R. F. Prenatal alcohol exposure does not influence sexual dimorphism of saccharin preference in C57 mice. Soc. Neurosci. Abstr. 11:293; 1985. 8. Blanchard, B. A. The influence of neonatal testosterone administration on the biobehavioral effects of prenatal ethanol exposure in male rats. Doctoral dissertation; 1990. 9. Blanchard, B. A.; Hannigan, J. H.; Riley, E. P. Amphetamineinduced activity after fetal alcohol exposure and underuutrition in rats. Neurotoxicol. Teratol. 9:113-119; 1987. 10. Blanchard, B. A.; Pilati, M. L.; Hannigan, J. H. The role of stress and age in spatial navigation deficits following prenatal exposure to ethanol. Psychobiol. 18:48-54; 1990.

II. Corbier, P.; Kerdelhue, B.; Picon, R.; Roffie, J. Changes in testicular weight and serum gonadotropin and testosterone levels before, during and after birth in the postnatal rat. Endocrinol. I03:1985-1991; 1978. 12. Eleftheriou, B. E.; Lucas, L. A. Age-related changes in testes, seminal vesiclesand plasma testosterone levelsin male mice. Gerontologla 20:231-238; 1974. 13. Even, M. D.; Vom Saal, F. S. Seminal vesicleand preputial gland response to steroids in adult male mice is influenced by prior intrauterine position. Physiol. Behav. 51:II-16; 1991. 14. Fermoso, J.; Esqnifino, A. I.; Mateos, A.; Agrasal, C.; Martin, I. Possible role of prolactin in the induction of hypogonadism by chronic alcohol treatment in the male rat. Pharmacol. Biochem. Behav. 29:489-493; 1988. 15. Graham, S.; Gandelman, R. The expression of anogenital distance data in the mouse. Physiol. Behav. 36:103-104; 1985. 16. Hannigan, J. H.; Blanchard, B. A.; Horner, M. P.; Riley, E. P.; Pilati, M. L.. Apomorphine-induced locomotion and stereotypy in rats exposed prenatally to alcohol. Neurotoxicol. Teratol. 12: 79-84; 1990. 17. Hannigan, J. H.; Blanchard, B. A.; Riley, E. P. Altered grooming responses to stress in rats exposed prenatally to ethanol. Behay. Neural. Biol. 47:173-185; 1987. 18. Hannigan, J. H.; Fitzgerald, L. W.; Blanchard, B. A.; Riley, E. P. Absence of differential motoric and thermic responses to clonidine in young rats exposed prenatally to alcohol. Alcohol 5: 431--439; 1989. 19. Hard, E.; Dahlgren, I. L.; Engel, J.; Larsson, K.; Liljequist, S.; Lindh, A. S.; Musi, B. Development of sexual behavior in prenatally ethanol-exposed rats. Drug Alcohol Depend. 14:5161; 1984. 20. McGiveru, R. F.; Clancy, A. N.; Hill, M. A.; Noble, E. P. Prena-

PRENATAL

21.

22.

23. 24.

25. 26. 27. 28. 29.

30.

ETHANOL

EXPOSURE

tal alcohol exposure alters adult expression of sexually dimorphic behavior in the rat. Science 224:896-898; 1984. McGivern, R. F.; Raum, W. J.; Handa, R. J.; Sokol, R. Z. Comparison of two weeks vs. one week of prenatal ethanol exposure in the rats on gonadal organ weights, sperm count, and onset of puberty. Neurotoxicol. Teratol. 14:351-358; 1992. McGivern, R. F.; Raum, W. J.; Salido, E.; Redei, E. Lack of prenatal testosterone surge in fetal rats exposed to alcohol: Alterations in testicular morphology and physiology. Alcohol: Clin. Exp. Res, 12:243-247; 1988. Meaney, M. J.; Stewart, J. Neonatal androgens influence social play of prepubescent rats. Horm. Behav. 15:197-213; 1981. Meaney, M..J.; Stewart, J.; Poulin, P.; McEwen, B. S. Sexual differentiation of social play in rat pups is mediated by the neonatal androgen-receptor system. Neuroendocrinol. 37:85-90; 1983. Meyer, L. S.; Riley, E. P. Behavioral teratology of alcohol. In: Riley, E. P.; Voorhees, C. V., eds. Handbook of behavioral teratology. New York: Plenum Press; 1986:101-140. Meyer, L. S.; Riley, E. P. Social play in juvenile rats prenatally exposed to alcohol. Teratol. 34:1-7; 1986. Olioff, M.; Stewart, J. Sex differences in the play behavior of prepubescent rats. Physiol. Behav. 20:111 - 113; 1978. Panksepp, J.; Beatty, W. W. Social deprivation and play in rats. Behav. Neural. Biol. 30:197-206; 1980. Parker, S.; Udani, M.; Gavaler, J. S.; Van Thiel, D. H. Adverse effects of ethanol upon the adult sexual behavior of male rats exposed in utero. Neurobehav. Toxicol. Teratol. 6:289-293; 1984. Pedhauzur, E. J. Multiple regression in behavioral research: Ex-

39

31. 32. 33. 34. 35. 36.

37. 38. 39. 40.

planation and prediction (2nd. ed.). New York: Holt, Rinehart and Winston; 1982:636-681. Redei, E.; McGivern, R. F. Attenuation of postnatal testosterone surge and decreased response to LH in fetal alcohol exposed males. Alcohol: Clin. Exp. Res. 12:341; 1988, Rudeen, K. Reduction of the volume of the sexually dimorphic nucleus of the preoptic area by in utero ethanol exposure in male rats. Neurosci. Lett. 72:363-368; 1986. Udani, M.; Parker, S.; Gavaler, J.; Van Thiel, D. H.. Effects of in utero exposure to alcohol upon male rats. Alcohol: Clin Exp Res 9:355-359; 1985. Van Thiel, D. H.; Gavaler, J. S.; Cobb, C. F.; Sherins, R. J.; Lester, R. Alcohol-induced testicular atrophy in the adult male rat. Endocrinol. 105:888-895; 1979. Van Thiel, D. H.; Lester, R. Sex and alcohol. New Eng. J. Med. 291:251-253; 1974. Vom Saal, F. S.; Dhar, M. G. Blood flow in the uterine loop artery and loop vein is bidirectional in the mouse: Implications for transport of steroids between fetuses. Physiol. Behav. 52:163171; 1992. Ward, I. L. The prenatal stress syndrome: Current status. Psychoneuroendoerinol. 9:3-11; 1984. Ward, I. L.; Weisz, J. Maternal stress alters plasma testosterone in fetal males. Science 207:328-329; 1980. Zimmerberg, B.; Reuter, J. Sexually dimorphic behavioral and brain asymmetries in neonatal rats: Effects of prenatal alcohol exposure. Dev. Brain Res. 46:281-290; 1989. Zimmerberg, B.; Scalzi, L. V. Commisural size in neonatal rats: Effects of sex and prenatal alcohol exposure. Int. J. Dev. Neurosci. 7:81-86; 1989.