- Email: [email protected]
Growth and activity but not maturation are affected by perinatal diet
Physiology & Behavior, Vol. 40, pp. 279--285.Copyright©PergamonJournals Ltd., 1987. Primedin the U.S.A,
0031-9384/8753.00 + .00
Growth and Activity but Not Maturation are Affected by Perinatal D i e t I J O A N C. M A R T I N , *~ D O N A L D C. M A R T I N , t PATTI SHORES AND STELLA CHAO*
*Departments o f Psychiatry and Behavioral Sciences and tBiostatistics University o f Washington, Seattle, WA 98105 R e c e i v e d 11 F e b r u a r y 1985 MARTIN, J. C., D. C. MARTIN, P. SHORES AND S. CHAO. Growth and activity but not maturation are affected by perinatal diet. PHYSIOL BEHAV 40(3)279-285, 1987.--Growth, maturation and activity in Sprague-Dawley rat offspring were studied as a function of two diets which were offered both pre- and postnatally. Day 1 gravid rats were placed ad lib on either Purina chow or Ensure, a total liquid diet. Half of each group was switched to the other diet on day 21 of gestation and maintained on it through weaning. Offspring remained on the same diet until sacrificed on postnatal day 60. Seven developmental measures were assessed daily until all of the animals reached criterion. The type of diet significantly affected maternal weight gain during the second third of gestation, and neonatal weight gain as measured on days 7, 14, 21, 28 and 45. Females were affected more than were male littermates. Males were significantly heavier than female littermates at all postnatal weighings. Diet had little or no effect on the developmental measures. Male/female differences were found for voluntary activity in the wheel which was measured on day 45. The use of liquid diets in behavioral teratology studies and the wisdom of using pellet diets as a control are discussed. The possibility of substituting careful measures of growth as an alternative to developmental testing is discussed. Maternal and postnatal diets
Growth and development
THE developmental battery which has been used for a number of years in our behavioral teratology laboratory was adapted from Wistar Institute normative studies on albino rats [5] as well as from batteries developed in other laboratories [1,19]. We performed a study using these tests on normal newborns in order to develop a standard against which to measure the performance of neonatal rats which had been exposed to drugs in utero. We also wanted to test the adequacy of a total liquid diet used by us and other laboratories as a vehicle for maternal drug administration. Our standard control diet is Purina chow. Maternal trauma is minimal when liquid diets are the vehicles for drug delivery which is an important consideration, since offspring growth, development and behavior may be affected by such trauma [3,21]. It has been shown that the type of diet may play a critical role in offspring development and function, quite apart from its adequacy [13]. We included both sexes in the study, since males and females mature at different rates.
METHOD
Dams Thirty primiparous Sprague-Dawley derived rats (Tyler Laboratories, Bellevue, WA) arrived in our laboratory the
Activity
Sex differences
Sprague-Dawley rat
morning following evening impregnation as determined by vaginal lavage. They were randomly assigned either Purina chow or Ensure on an ad lib basis for the next 21 days with tap water always available. They were housed singly in polypropylene cages on ground corn cob bedding (Sanicell). Food consumption and weight gain were monitored daily through delivery. Beginning on day 2 I, half of each group was randomly assigned the other diet. The offspring continued on that diet following weaning. There were 8 PurinaPurina (P-P), 7 Purina-Ensure (P-E), 8 Ensure-purina (E-P), and 7 Ensure-Ensure (E-E) dams. Cages were checked twice daily for deliveries, and the newborns were counted and weighed after they had been gathered into a nest.
Maternal Diets Baker [2] stated that the minimum daily diet of the breeding female rat should contain 12% net protein and 5% net fat. The composition of Purina rat chow exceeds the minimum NRC requirements for gestation, lactation, and neonatal growth [16]. It contains 15% net protein, 4.5% crude fat, and 56% carbohydrates, as well as more than minimum amounts of the necessary vitamins, minerals, and other nutrients. Ensure (Ross Laboratories) has a caloric composition of 14% protein, 31.5% fat and 54.5% carbohydrates. Ensure is preferable to Metrecal, another liquid diet, since the latter often causes loose stools [18]. The vanilla flavor of Ensure is av-
1This study was partially supported by PHS/NIH Grant No. HD1680 to the first author. ~Requests for reprints should be addressed to J. C. Martin, Department of Psychiatry and Behavioral Sciences, RP-10 University of Washington, Seattle, WA 98105.
279
280
MARTIN. MARTIN, SHORES AND CHAO
180
TABLE 1 DEVELOPMENTAL MEASURES
Measure Ear pinnae Righting Upper incisor eruption Lower incisor eruption Open field*
Eyelid dysjunction Vertical righting
Response Uncurled from side of head to upright position Ventral to dorsal position within 3 seconds Both incisors through gums through the gums Both incisors through gums
3 through 8 4, 5, 6, 7, 8 4, 5, 6, 7
7 through 19
*The open field consisted of a 30x30×4 inch open box of unpainted wood. The floor was marked in 1-inch concentric circles. The field was set on a table directly under a fluorescent light source.
Purina 0---0 Ensure
140
3, 4, 5
Number of concentric circles 7 through 21 crossed within 5 seconds in forward direction Rounded aperture: not a slit 13 through 17 Dropped in dorsal position from 30 cm: 3 trials/day
160
Testing Days
(.'3 c:
120
100
80
(,.3 o
60
/ 40
i
I
I
4
I
l
I
6
8
Trimester 1
Ojfspring Litters were weighed at birth and at weekly intervals through day 28. Litters were culled to 8 on day 4, with 4 of each sex reserved whenever possible. Selection of individuals was random within this constraint. Ten males and ten females were randomly selected from each of the four dietary conditions on day 28. A male and female were included from a selected litter. The litters which were excluded was also a random choice. Individual animals were weighed on day 45 prior to being placed overnight in the activity wheels and again on day 60 prior to sacrifice.
Developmental Battery The first test day for each measure was based on previous studies in our laboratory. Testing was continued until all offspring were able to perform the task to criterion. Daily testing was between 8--10 a.m. This was always performed by two individuals one of whom had no knowledge of the dietary history of the litters. This person made the decisions. Table I describes the measures and inclusive testing days.
Activity Wheel Voluntary activity in the Wahmann wheel was measured once for each rat between days 45--53. The individual test day was randomly selected such that all four treatment conditions were represented each day. There are 12 identical wheels. These are 14 inch open drums of wire mesh which revolve as the rat runs in place at the nadir of the wheel. A manual counter changes with each revolution. Mature rats will run the equivalent of several miles/day if confined to the
j/o~/
20
2
idly consumed by the pregnant rat [15]. One potential problem of Ensure is that it does not contain the amino acid DL methionine, which facilitates protein utilization [4]. However, since all of our rats are fed Purina chow prior to breeding, maternal stores were presumed to be adequate for a normal pregnancy.
J
0
.,c
0///~ /~
I
11
0
i
l
12
Trimester 2
i
l
l
16
l
l
i
18
i
20
Trimester 3
FIG. 1. Mean cumulative maternal weight gain as a function of prenatal diet. wheel. Each wheel was used equally often by males and females and by members of each dietary condition. Rats were placed singly in a wheel without food, water, or lights at 5 p.m. and removed the next morning at 8 a.m.
Degrees of Freedom Neonatal analyses were based upon the number of litters as the sampling unit, since the randomization process was performed on the mothers, and measures on individuals within a litter could have been correlated. An analysis which used the number of offspring would thus ignore possible correlations, confound the within and between litter variation, and greatly overstate the degrees of freedom. See [8, 10, 17, 23, 24] for an exhaustive discussion of the theoretical biases inherent in the decision which sets the number of degrees of freedom.
Statistics Chi Square analyses were performed on the dichotomous data: deliveries/non-deliveries, live born/stillborn number of neonatal deaths, and gestation length in days. Analyses of variance were performed on all other measures with 2x2 analyses of variance on the dietary factor. The Bonferroni inequality [9] was applied to the analysis of variance data. This method conservatively adjusts the degrees of freedom when multiple tests are performed. The Freeman-Tukey arcsine transformation was used to normalize the measures expressed as proportions, and a square root transformation was selected to reduce skewness of the number of revolutions in the activity wheel [6]. The activity wheel data were analyzed for sex differences only. All other data sets were analyzed for differences both within each sex and between sexes. For the between sex analyses, male and female litter-
P E R I N A T A L DIET A N D O F F S P R I N G F U N C T I O N
281
m~ite scores were differenced and t-tests performed on these data. RESULTS
H Purina-Purina [Z}~O Purina-Ensure o - - 0 Ensure-Purina A--Z~ Ensure-Ensure
80
Maternal Measures The length of the gestational period was not affected by diet. Ten of the Purina diet and 11 of the Ensure dams delivered on day 23. The remaining 6 rats delivered the following day. The data for rats which did not deliver were not analyzed. The Ensure dams all delivered viable litters, but 3 of the Purina fed dams failed to deliver. The pattern of weight gain for these animals indicated that they either resorbed their litters early in pregnancy, or were never pregnant. The type of diet did not affect maternal weight gain over the gestation period as a whole, nor did it have an effect on the first or last seven days. A t-test was performed on day one weights to determine if our random assignment might have inadvertently placed heavier animals in one group. The gravid animals did not differ in weight on day 1 (t= 1.265, p=0.218). Weight gain over days 8--14, the period for major organogenesis, was affected differentially by diet, and the Ensure fed dams gained a mean of 44.8 g as compared with 37.2 g for the Purina fed rats F(1,25)=7.85, p =0.009. Caloric intake was not different as a function of type of diet when measured over the entire gestational period, but again, the Ensure fed dams consumed significantly more calories during days 8-14, F(1,25)=20.5, p=0.0001. Weight gain and caloric intake are not independent. Although the Purina fed mothers had a somewhat higher caloric intake during the first seven days: p=0.036, this did not significantly affect their weight gain. Figure 1 is a plot of maternal weight gain as a function of diet.
/T I I
u)
E (.9
60
/
¢.-
t'.rc~ o3
40
0
/
t-
20
0
¢/// /
I
I
I
I
I
Birth
7
14
21
28
Postnatal Age
FIG. 2. Mean individual offspring weight as a function of postnatal age and diet.
Offspring Measures (1) Stillbirths and neonatal deaths. The dams were only checked twice daily for delivery, so an exact count of stillbirths was not possible. Four of the Ensure mothers had at least one dead newborn, whereas the Purina dams had none: X2 = 4.04 [1], p =0.04. There were no differences as a function of diet for liveborn birthweight, litter size, and neonatal deaths either from birth to culling on day 4, or during the neonatal period of days 4-28. There were no differences in the sex ratio on day 4 prior to culling. (2) Neonatal weight gain. Although there were no differences in liveborn litter weight as a function of maternal diet, litter weight differences were found at each weekly weighing thereafter and on day 45 when the rats were placed in the activity wheel. Prenatal diet was entirely responsible for the effects from birth to day 14, with postnatal diet beginning to have an effect by day 21. Postnatal diet accounted for all of the differences in weight gain by day 28. Newborns which received Purina prenatally were heavier on days 7 and 14, and those which received Purina postnatally were heavier when weighed on days 21 and 28. The heaviest group were those animals whose mothers were fed Purina as well as themselves (P-P). The Bonferroni approximation was applied to reduce the number of degrees of freedom since multiple tests were performed on the same data set. The prenatal effect was significant with a p=0.006 on day 7, p=0.0009 on day 14, and p=0.005 on day 21. The prenatal effect was no longer significant on d a y 28. The postnatal dietary effect was non-significant on day 7, and significant on
day 14, (p=0.036), day 21 (/9=0.004) and day 28 (,o=0.002). There were no interaction effects. Since for this number of tests, the Bonferroni inequality requires a significance level of 0.006 to be significant, only the day 14 postnatal test was not significant since that reached only 0.036. There were also no sex differences either for pre- or postnatal diets, nor interaction effects, i.e., diets affected both sexes in the same way. However, the males were significantly heavier than female littermates at all four postnatal weighings with p=0.0001 to p=0.000001. The Bonferroni correction factor did not affect these results. (3) Day 45 and 60 individual body weights. The group means at 45 days of age followed the same pattern as the earlier weights: P-P=223.6 g, E - P = 196.3 g, P - E = 193.6 g and E-E = 185.2 g. These weights were significantly different at a p =0.001 with a two way analyses o f variance. The prenatal factor still exerted an effect (p =0.012), but the postnatal factor accounted for most of the variance (p=0.004). Body weight at the next and last weighing on day 60 was no longer significantly affected by different diets, but the P-P rats were still heavier: P-P=300.2 g, E-P=268.3 g, P-E=277.2 g and E-E=265.8 g. There was no interaction between the diets, and effects when found, were due to main effects. Male weight gain differed as a function of diet on postnatal days 14, 21, 28 and 45, but not on days 7 and 60. Diet affected female weight gain significantly throughout the postnatal period. Diet had an earlier and more lasting effect
282
MARTIN, MARTIN, SHORES AND CHAO TABLE 2 M E A N P E R C E N T A G E O F RATS A C H I E V I N G CRITERION*
Measure
Day
P-P
P-E
E-P
E- E
Righting
3 4 5 6 7
14.3% 56.9 75.0 79.2 91.5
32.4% 56.5 70.2 87.2 91.5
39.8~ 49.3 80.4 94.6 96.4
17.0cA, 35.9 75.0 85.7 92.9
Ear pinnae (1 or 2)
3 4
19.0 100.0
30,9 74.0
39.0 91.0
15.0 98.0
Incisors (lower)
4 5 6
37.9 97.9 100.0
25.8 89.4 100.0
27.5 67.9 92.9
14.3 82.1 92.9
Incisors (upper)
4 5 6 7
01.7 20.8 54.2 93.6
08.1 34.0 70.2 91.5
01.4 37.5 78.6 91.2
03.6 33.9 75.0 94.6
Vertical righting
14 15 16 17
10.6 51.1 91.5 100.0
00.0 23.4 51.1 93.6
00.0 28.6 69.6 100.0
O1.8 01.8 41.1 89.3
Eyelid dysjunction
14 15 16 17
00.0 34.0 83.0 100.0
17.0 46.8 100.0 100.0
08.9 48.2 89.3 100.0
05.4 46.4 92.9 98.2
Open field (5 seconds)
15 16 17 18 19 20 21
14.9 21.3 42.5 31.9 42.5 41.3 47.8
10.6 19.1 23.4 53.3 53.3 60.0 51.1
12.5 23.2 20.0 30.9 36.4 50.9 56.4
08.9 23.2 39.3 54.5 72.2 74.5 84.3
*Litters culled to 4 males and 4 females on day 4.
on females, possibly due to their more marginal weight status. Male and female littermate weights were differenced. When littermate scores were not available, males and females from the same treatments were randomly matched and those scores differenced. An analysis of variance on the differenced scores yielded an F(3,36)=4.35,p =0.01. Diet did have an effect on weight sex differences. Diet no longer had an effect by day 60, although males remained significantly heavier than females, as they had been throughout the neonatal period with probability values ranging from p =0.0001 to 0.00001. Growth in the offspring as a function of maternal and neonatal diet is presented in Fig. 2. (4) Developmental measures. See Table 2 for the percentage of animals which achieved criterion for the measures described below. (a) Righting (dorsal position, days 3-8). There were no differences in the proportion of animals/litter which could right themselves on any test day. Marginal interactive effects between pre- and postnatal dietary conditions were found: p=0.03-0.08 but the main effects failed to reach significance. Eighty percent of all rats were able to right within the 3-sec
time limit by day 6. Diet also did not affect within or between sex differences. (b) Ear pinnae uncurling (days 3-5). A significant difference was found on day 4 on the proportion of rats with at least one ear uncurled, F(3,22)=3.96, p=0.02. The major effect was due to postnatal, rather than prenatal, diet: p=0.04. All ears were up by day 5. Sex had some effect on this measure, but the pattern was inconsistent. The correction factor required a significance level of 0.01 which was not found. (c) Lower incisors (days 4-7). There were no differences as a function of diet, and essentially all eruptions had taken place by day 6. Since this is usually one of them more sensitive measures, we pooled the data for the three days and did an exploratory analysis. F(1,22)=5.46, p=0.029, and the prenatal factor accounted for most of the variance. The offspring of mothers which received Purina had a greater number of eruptions. (d) Upper incisors (day 4-8). No differences were found as a function of diet. Although the main effect was nonsignificant on day 6, the prenatal factor accounted for most of the variance: p=0.038, Over ninety percent of all incisors
P E R I N A T A L DIET A N D O F F S P R I N G F U N C T I O N had erupted by day 7. No intra nor intersex differences were found. This measure has never been as useful as lower incisor eruption, possibly because it is more difficult to measure. (e) Vertical righting (days 7-19). There were no reliable differences on days 7-14 due to the large number of failures. Main analyses of variance were significant for days 15-19 with probabilities of 0.02 and 0.03. The data were pooled over days 13-19 for an exploratory analysis. F(3,22)=4.13, p=0.018, and the postnatal factor was unsurprisingly responsible for the differences: p =0.007. This difference was not significant when the correction factor was applied: 0.004 Almost all of the rats were landing on their feet by day 18, with the P-P animals performing at a somewhat higher level. (f) Eyelid dysjunction (days 13-17). There were no differences as a function of diet, and all eyes were opened by day 17. No sex differences were found. (g) Open field (days 7-21). The excessive variability found with this measure, particularly prior to eye opening on day 14 made interpretations difficult. There was little evidence for differences due to diet, and sex differences were inconsistent. Postnatal diets had a greater effect than prenatal after day 14. (h) Activity wheel (day 45). The sexes were not combined as in the previous analyses, since sex and activity level tend to be highly correlated in older animals. Square root transformations were used to stabilize the variances. Postnatal diet accounted for most of the effects on male activity: F(1,36)=5.30, p=0.027. The P-E and E-P males were more active than were the males in the other two conditions. Since an animal's weight affects its activity level [14], weight was covaried with the number of wheel revolutions using a multiple linear regression technique. Weight as a function of diet was positively correlated with numbers of wheel revolutions, F(3,36)=12.46, p=0.00001. Therefore, weight differences did not explain the results. The postnatal diet still accounted for most of the variance p =0.0004, but the interaction term was significant as well (,o=0.007). In contrast, diet had only a modest effect on female activity, F(3,35)=3.97, p=0.01, with the interaction term being responsible for most of the effects: F(1,35)=7.59, p=O.O09. When weight was covaried with the number of revolutions, the significant differences among the four dietary conditions vanished. The lightest weight females were the most active, i.e., the E-E rats. Therefore, weight differences as a result of differential diet could have accounted for female differences in activity, but this was not true of their male littermates. Parenthetically, it should be noted that since diet determined the body weights, the female activity differences were real, but the covariance analysis was unable to parcel out the effects. It cannot be determined from this study if female activity level was due to weight change as a function of diet, or to other dietary effects such as alterations in neural structure which might influence function. Weight was included as a factor in the 2×2 analysis of male/female difference scores, F(5,38)=6.10,p=0.0004. This analysis is equivalent to a repeated measures analysis of sex within diet. The significant results were due to the postnatal dietary factor: p =0.007, as well as the male weight factor: p=0.006. Female E-E rats were significantly more active then were E-E males, but the P-E and E-P males were more active than their female littermates. Activity had been measured between days 45 and 53 which is prior to sexual maturation when estrus cycle fluctuations would have resuited in increased female activity rates. When one extreme
283 male outlier was omitted from the analysis, the weight differences between the sexes accounted for most of the variance, and hence the activity differences. Hall [7] found that lighter weight rats were more active animals regardless of sex. This was only true for the females in our study, and the lack of male differences may have been due to the immaturity of the rats. DISCUSSION We reported earlier that maternal food intake and weight gain on both diets peaked during the second seven days and then declined [15]. The Ensure fed dams in the current study consumed more Kcal and gained more weight during days 8-14, although when measured over the entire 21 day period there were no dietary differences. Diet type had no effect on litter size, birth weight, neonatal death rate, nor sex ratio. Sustacal (Mead-Johnson) was the control diet in a maternal alcohol study by Sherwin et al., [18]. Half of the control offspring died between birth and day 21, in spite of the fortified diets available from birth. Since we found no difference in offspring death rate from birth to day 60 as compared with Purina controls, Ensure may be a more adequate diet than Sustacal for the rat. However, the pregnant rats in our study were placed on the liquid diet on day 1 of gestation, whereas Sherwin's rats were placed on Sustacal three weeks prior to breeding. The latter design may have lowered the nutritional status of the dams and contributed to the high neonatal death rate. We did find that Ensure had a deleterious effect on offspring weight gain when continued during the neonatal period. Prenatal diet was responsible for the weight status of the pups from birth through day 14, and postnatal diet began to exert a small effect by day 14. Both pre- and postnatal diets were contributing equally by day 21, and the postnatal diet was wholly responsible for differences by day 28 which was 6 days following weaning. Prenatal and postnatal exposure to Purina conferred a weight advantage. Body weights of the other three groups were virtually identical by day 60, but the P-P animals were still the heaviest as a group. Since the study was terminated on that date, it is not known if this advantage would have been maintained. Females were more adversely affected than males by the limitations of the Ensure diet. This difference could have been due to the effects of a less adequate diet on smaller animals, or to other sex differences. Uphoff et al. [22] fed gravid mice an Ensure diet which contained 20% ethanol. One control group received Purina chow, and a second group was offered Ensure with sucrose substituted isocalorically for the ethanol. The pups were sacrificed at birth. Live litter size did not differ between the two control groups, but the ethanol-Ensure litters were fewer in number, smaller in size, and had cardiac muscle defects. The authors speculated that the results may have been due to an interaction between a low protein diet Ensure and ethanol. There was no evidence in support of this interesting hypothesis. The alternate hypothesis was that ethanol or its metabolites caused the cardiac defects. Our P-P offspring reached criterion sooner than the other three groups on the two fighting measures. This diet conferred no advantage, however, on other measures such as eye opening, outer ear uncurling, vertical drop, open field and incisor eruption. Movement in the open field is affected primarily by whether the eyes have opened, by lighting levels, and by
284
MARTIN, MARTIN, SHORES AND CHAO
motivational factors since motor ability is not at issue alter the first few days of life. Eighty-three to 100 percent of all offspring had open eyes when examined on day 16. Over half of the Ensure-Ensure group were moving lbrward reliably during that time period, whereas the scores for the other three groups ranged from 33-38%. Since the other groups also had opened eyes, the tendency may reflect a higher activity level by the smaller, lighter weight animals. Hall [7] tound that poorly nourished, smaller rats were more active in the Wahmann wheels than were well nourished animals. It is often desirable to restrict the number of test days on newborns for pragmatic reasons. The age at which 75% of control animals reach criterion on each measure is a useful cut-off point. The 75% criterion for rats which received a Purina diet was: days 4-5 for dorsal righting, day 3 for ear pinnae uncurling, days 4-5 for lower, and days 6-7 for upper, incisor eruption. Righting from a vertical drop and eyelid dysjunction both occurred on days 15-16. Scores in the open field which are multidetermined, were too variable to be useful. Criterion was reached by the Ensure animals on the same days for dorsal righting, ear uncurling, lower incisor eruption and eye opening, but a day earlier for upper incisor eruption and a day later for vertical righting. Ensure animals were more consistent in the open field and 75c~ of them moved out reliably by days 1%20. Lochry, Hoberman and Christian [12] recently published a study on 8,000 rat pups. Weight gain was a more sensitive index of developmental status than a test battery which included eye opening, startle, ear pinnae uncurling, vertical drop, and dorsal righting measures. Activity, unfortunately, was not assessed. Prenatal treatment was limited to pharmaceutical agents and results may not be generalizable to other extrinsic agents. However, our results agree with theirs, since our developmental measures contributed almost no additional information beyond that obtained from the weight gain measure. Although developmental measures have a degree of face validity not found in a measure as simplistic as weight gain, results have not been consistent across laboratories which use test batteries. Some of these measures are highly subjective and dependent upon technician sophistication and consistency. Failure to thrive, on the other hand, which includes pool weight gain, is an excellent predictor of difficulties in later childhood. It may not be desirable to eliminate developmental test batteries, but careful measurement of weight gain and food consumption should be taken at regular intervals and correlated with both developmental and perform-
ance measures. It may be found that the developmental measures add little to the predictive value of function at older ages. S e x Dijy?rences
There were no reliable sex differences prior to sexual maturity, other than weight differences. The open field yielded little of value, whereas the activity wheel was very informative. This is not surprising, since the two tasks are different topographically and were taken at different ages. Activity measures do not correlate well with each other [20[, nor is activity a simple or homogeneous behavior [11]. Intersex differences in the wheel were primarily due to differences in body weight between males and females. Differences in activity among the 4 groups of males were not due to weight differences, but to some other factor or factors. We are now better able to predict when developmental milestones are reached in the Sprague-Dawley strain as a function of two diets. However, results from this study and another one by Lochry et al. [ 12] would seem to indicate that little is to be gained by measuring maturation, if careful growth records are kept. Voluntary activity, as measured in the wheel, does contribute useful information, particularly for male rats. We also found that Purina chow is not the appropriate control diet when Ensure is the experimental diet. The Purina diet during gestation and lactation conferred a weight advantage which could not be equalled by offspring prenatally exposed to Ensure, even when the animals received Purina for 60 days postnatally. The rats were at an even greater disadvantage when Ensure was continued throughout the postnatal period (E-E), or begun postnatally (P-EL The E-P rats were able to overcome most of their initial weight disadvantage. Unsupplemented Ensure as a total diet would seem to be adequate for the gestational period, assuming that mothers had been given Purina prior to pregnancy. It should not be continued after birth when nutritional requirements are more stringent. Although the liquid diet is calorically adequate, it is almost certainly deficient in protein, if not in other important nutrients. Supplementation with vitamins, minerals, and protein could conceivably extend its usefulness into the neonatal period. If Purina chow is used for a "normal" control group, then a second control group which is given the liquid diet, without the addition of drugs, is necessary to separate drug from dietary factors.
REFERENCES I. Alder, S. and G. Zbinden. Methods for the evaluation of physical neuromuscular and behavioral development in rats in early postnatal life. In: Methods in Prenatal Toxicology. edited by D. Neuberg, H. J. Merker and T. E. Kwasigroch. Stuttgart: G Thieme Publ., 1977, pp. 175-185. 2. Baker, D. E. J. Reproduction and breeding. In: The Laboratory Rat: Biology and Diseases, edited by H. J. Baker, J. R. Lindsey and S. H. Weisbroth. New York: Academic Press, 1979, Vol 1, Chap 7. 3. Becket, R. F, and J, C. Martin. Vital effects of chronic nicotine absorption and hypoxic stress during pregnancy and the nursing period. Am J Obstet Gynecol 110: 522-533, 1971. 4. Breuer, L. H., Jr., W. G. Pond, R. G. Warner and J. K. Loosli. A comparison of several amino acid and casein diets for the growing rat. J Nutr 80: 243-250, 1963.
5. Farris, E. J. Breeding of the rat. In: The Rat in Laboratory Investigation, edited by E. J. Farris and J. Q. Griffith, Jr. Philadelphia: JB Lippincott Co., 1949. 6. Freeman, M. F. and J. W. Tukey. Transformations related to the angular and square root, Ann Math Star 21:607-611, 1959. 7. Hall, R. D. The role of body size in climbing and locomotor behavior of protein malnourished and well nourished rats. Dev Psychobiol 17: 91-106, 1984. 8. Holson, J. F., W. J. Scott, D. W. Gaylor and J. G. Wilson. Reduced interlitter variability in rats resulting from a restricted mating period, and reassessment of the "litter effect." Teratology 14: 135-142, 1976. 9. International Encyclopedia of Statistics, edited by W. H. Kruskal and J. M. Tanur. New York: Free Press, Vol 1, 1978, p. 565. 10. Jensh, R. P., R. L. Brent and M. Barr, Jr. The litter effect as a variable in teratologic studies of the albino rat. J Anat 128: 185--192, 1978.
PERINATAL
DIET AND OFFSPRING
FUNCTION
11. Ljungberg, T. Reliability of two activity boxes commonly used to assess drug induced behavioral changes. Pharmaco/Biochem Behav 8: 191-195, 1978. 12. Lochry, E. A., A. M. Hoberman and M. S. Christian. Comparative sensitivity of pup body weight and commonly used developmental landmarks. Teratology 32: 28A, 1985. 13. McClain, R. M. and J. M. Rohrs. Potentiation of the teratogenic effects and altered disposition of diphenylhydantoin in mice fed a purified diet. Toxicol Appl Pharmacol 77: 86-93, 1985. 14. Martin, J. C. and D. C. Martin. Voluntary activity in the aging rat following maternal drug exposure. Neurobehav Toxicol Teratol 3: 261-264, 1981. 15. Martin, J. C., D. C. Martin, B. Radow and G. Sigman. Blood alcohol level and caloric intake in the gravid rat as a function of diurnal period, trimester and vehicle. Pharmacol Biochem Behav 8: 421--427, 1978. 16. National Academy of Sciences. Nutrient Requirements o f Domestic Animals, Vol 10: Nutrient Requirements o f Laboratot3' Animals, 1972, pp. 56-93. 17. Palmer, A. K. Statistical analysis and choice of sampling unit. Teratology 10: 301-302, 1972.
285
18. Sherwin, B. T., S. Jacobson, S. L. Troxell, A. E. Rogers and R. W. Pelham. A rat model (using a semipurified diet) of the Fetal Alcohol Syndrome. In: Currents in Alcoholism: Recent Advances in Research and Treatment, edited by M. Galenter. New York: Grune and Stratton, 1980, pp. 15--30. 19. Smart, J. L. and J. Dobbing. Vulnerability of developing brain I1. Effects of early nutritional deprivation on reflex ontogeny and development of behavior in the rat. Brain Res 28: 85--95, 1971. 20. Tapp, J. T., R. S. Zimmerman and P. S. D'Encarnacao. Intercorrelational analysis of some common measures of rat activity. Psychol Rep 23: 1047-1050, 1968. 21. Thompson, W. R., J. Watson and W. R. Charlesworth. The effects of prenatal maternal stress on offspring behavior in rats. Psychol Med [Monogr Suppl] 76: 1-26, 1962. 22. Uphoff, C., C. N. Battie and R. Toth. Cardiac muscle development of mice exposed to ethanol in utero. Teratology 30:119129, 1984. 23. Weil, C. S. Selection of the valid number of sampling units and a consideration of their combination in toxicological studies involving reproduction, teratogenesis or carcinogenesis. Food Cosmet Toxicol 8: 177-182, 1970. 24. Well, C. S. Choice of the number of sampling units in teratology. Teratology 10: 301, 1975.
0031-9384/8753.00 + .00
Growth and Activity but Not Maturation are Affected by Perinatal D i e t I J O A N C. M A R T I N , *~ D O N A L D C. M A R T I N , t PATTI SHORES AND STELLA CHAO*
*Departments o f Psychiatry and Behavioral Sciences and tBiostatistics University o f Washington, Seattle, WA 98105 R e c e i v e d 11 F e b r u a r y 1985 MARTIN, J. C., D. C. MARTIN, P. SHORES AND S. CHAO. Growth and activity but not maturation are affected by perinatal diet. PHYSIOL BEHAV 40(3)279-285, 1987.--Growth, maturation and activity in Sprague-Dawley rat offspring were studied as a function of two diets which were offered both pre- and postnatally. Day 1 gravid rats were placed ad lib on either Purina chow or Ensure, a total liquid diet. Half of each group was switched to the other diet on day 21 of gestation and maintained on it through weaning. Offspring remained on the same diet until sacrificed on postnatal day 60. Seven developmental measures were assessed daily until all of the animals reached criterion. The type of diet significantly affected maternal weight gain during the second third of gestation, and neonatal weight gain as measured on days 7, 14, 21, 28 and 45. Females were affected more than were male littermates. Males were significantly heavier than female littermates at all postnatal weighings. Diet had little or no effect on the developmental measures. Male/female differences were found for voluntary activity in the wheel which was measured on day 45. The use of liquid diets in behavioral teratology studies and the wisdom of using pellet diets as a control are discussed. The possibility of substituting careful measures of growth as an alternative to developmental testing is discussed. Maternal and postnatal diets
Growth and development
THE developmental battery which has been used for a number of years in our behavioral teratology laboratory was adapted from Wistar Institute normative studies on albino rats [5] as well as from batteries developed in other laboratories [1,19]. We performed a study using these tests on normal newborns in order to develop a standard against which to measure the performance of neonatal rats which had been exposed to drugs in utero. We also wanted to test the adequacy of a total liquid diet used by us and other laboratories as a vehicle for maternal drug administration. Our standard control diet is Purina chow. Maternal trauma is minimal when liquid diets are the vehicles for drug delivery which is an important consideration, since offspring growth, development and behavior may be affected by such trauma [3,21]. It has been shown that the type of diet may play a critical role in offspring development and function, quite apart from its adequacy [13]. We included both sexes in the study, since males and females mature at different rates.
METHOD
Dams Thirty primiparous Sprague-Dawley derived rats (Tyler Laboratories, Bellevue, WA) arrived in our laboratory the
Activity
Sex differences
Sprague-Dawley rat
morning following evening impregnation as determined by vaginal lavage. They were randomly assigned either Purina chow or Ensure on an ad lib basis for the next 21 days with tap water always available. They were housed singly in polypropylene cages on ground corn cob bedding (Sanicell). Food consumption and weight gain were monitored daily through delivery. Beginning on day 2 I, half of each group was randomly assigned the other diet. The offspring continued on that diet following weaning. There were 8 PurinaPurina (P-P), 7 Purina-Ensure (P-E), 8 Ensure-purina (E-P), and 7 Ensure-Ensure (E-E) dams. Cages were checked twice daily for deliveries, and the newborns were counted and weighed after they had been gathered into a nest.
Maternal Diets Baker [2] stated that the minimum daily diet of the breeding female rat should contain 12% net protein and 5% net fat. The composition of Purina rat chow exceeds the minimum NRC requirements for gestation, lactation, and neonatal growth [16]. It contains 15% net protein, 4.5% crude fat, and 56% carbohydrates, as well as more than minimum amounts of the necessary vitamins, minerals, and other nutrients. Ensure (Ross Laboratories) has a caloric composition of 14% protein, 31.5% fat and 54.5% carbohydrates. Ensure is preferable to Metrecal, another liquid diet, since the latter often causes loose stools [18]. The vanilla flavor of Ensure is av-
1This study was partially supported by PHS/NIH Grant No. HD1680 to the first author. ~Requests for reprints should be addressed to J. C. Martin, Department of Psychiatry and Behavioral Sciences, RP-10 University of Washington, Seattle, WA 98105.
279
280
MARTIN. MARTIN, SHORES AND CHAO
180
TABLE 1 DEVELOPMENTAL MEASURES
Measure Ear pinnae Righting Upper incisor eruption Lower incisor eruption Open field*
Eyelid dysjunction Vertical righting
Response Uncurled from side of head to upright position Ventral to dorsal position within 3 seconds Both incisors through gums through the gums Both incisors through gums
3 through 8 4, 5, 6, 7, 8 4, 5, 6, 7
7 through 19
*The open field consisted of a 30x30×4 inch open box of unpainted wood. The floor was marked in 1-inch concentric circles. The field was set on a table directly under a fluorescent light source.
Purina 0---0 Ensure
140
3, 4, 5
Number of concentric circles 7 through 21 crossed within 5 seconds in forward direction Rounded aperture: not a slit 13 through 17 Dropped in dorsal position from 30 cm: 3 trials/day
160
Testing Days
(.'3 c:
120
100
80
(,.3 o
60
/ 40
i
I
I
4
I
l
I
6
8
Trimester 1
Ojfspring Litters were weighed at birth and at weekly intervals through day 28. Litters were culled to 8 on day 4, with 4 of each sex reserved whenever possible. Selection of individuals was random within this constraint. Ten males and ten females were randomly selected from each of the four dietary conditions on day 28. A male and female were included from a selected litter. The litters which were excluded was also a random choice. Individual animals were weighed on day 45 prior to being placed overnight in the activity wheels and again on day 60 prior to sacrifice.
Developmental Battery The first test day for each measure was based on previous studies in our laboratory. Testing was continued until all offspring were able to perform the task to criterion. Daily testing was between 8--10 a.m. This was always performed by two individuals one of whom had no knowledge of the dietary history of the litters. This person made the decisions. Table I describes the measures and inclusive testing days.
Activity Wheel Voluntary activity in the Wahmann wheel was measured once for each rat between days 45--53. The individual test day was randomly selected such that all four treatment conditions were represented each day. There are 12 identical wheels. These are 14 inch open drums of wire mesh which revolve as the rat runs in place at the nadir of the wheel. A manual counter changes with each revolution. Mature rats will run the equivalent of several miles/day if confined to the
j/o~/
20
2
idly consumed by the pregnant rat [15]. One potential problem of Ensure is that it does not contain the amino acid DL methionine, which facilitates protein utilization [4]. However, since all of our rats are fed Purina chow prior to breeding, maternal stores were presumed to be adequate for a normal pregnancy.
J
0
.,c
0///~ /~
I
11
0
i
l
12
Trimester 2
i
l
l
16
l
l
i
18
i
20
Trimester 3
FIG. 1. Mean cumulative maternal weight gain as a function of prenatal diet. wheel. Each wheel was used equally often by males and females and by members of each dietary condition. Rats were placed singly in a wheel without food, water, or lights at 5 p.m. and removed the next morning at 8 a.m.
Degrees of Freedom Neonatal analyses were based upon the number of litters as the sampling unit, since the randomization process was performed on the mothers, and measures on individuals within a litter could have been correlated. An analysis which used the number of offspring would thus ignore possible correlations, confound the within and between litter variation, and greatly overstate the degrees of freedom. See [8, 10, 17, 23, 24] for an exhaustive discussion of the theoretical biases inherent in the decision which sets the number of degrees of freedom.
Statistics Chi Square analyses were performed on the dichotomous data: deliveries/non-deliveries, live born/stillborn number of neonatal deaths, and gestation length in days. Analyses of variance were performed on all other measures with 2x2 analyses of variance on the dietary factor. The Bonferroni inequality [9] was applied to the analysis of variance data. This method conservatively adjusts the degrees of freedom when multiple tests are performed. The Freeman-Tukey arcsine transformation was used to normalize the measures expressed as proportions, and a square root transformation was selected to reduce skewness of the number of revolutions in the activity wheel [6]. The activity wheel data were analyzed for sex differences only. All other data sets were analyzed for differences both within each sex and between sexes. For the between sex analyses, male and female litter-
P E R I N A T A L DIET A N D O F F S P R I N G F U N C T I O N
281
m~ite scores were differenced and t-tests performed on these data. RESULTS
H Purina-Purina [Z}~O Purina-Ensure o - - 0 Ensure-Purina A--Z~ Ensure-Ensure
80
Maternal Measures The length of the gestational period was not affected by diet. Ten of the Purina diet and 11 of the Ensure dams delivered on day 23. The remaining 6 rats delivered the following day. The data for rats which did not deliver were not analyzed. The Ensure dams all delivered viable litters, but 3 of the Purina fed dams failed to deliver. The pattern of weight gain for these animals indicated that they either resorbed their litters early in pregnancy, or were never pregnant. The type of diet did not affect maternal weight gain over the gestation period as a whole, nor did it have an effect on the first or last seven days. A t-test was performed on day one weights to determine if our random assignment might have inadvertently placed heavier animals in one group. The gravid animals did not differ in weight on day 1 (t= 1.265, p=0.218). Weight gain over days 8--14, the period for major organogenesis, was affected differentially by diet, and the Ensure fed dams gained a mean of 44.8 g as compared with 37.2 g for the Purina fed rats F(1,25)=7.85, p =0.009. Caloric intake was not different as a function of type of diet when measured over the entire gestational period, but again, the Ensure fed dams consumed significantly more calories during days 8-14, F(1,25)=20.5, p=0.0001. Weight gain and caloric intake are not independent. Although the Purina fed mothers had a somewhat higher caloric intake during the first seven days: p=0.036, this did not significantly affect their weight gain. Figure 1 is a plot of maternal weight gain as a function of diet.
/T I I
u)
E (.9
60
/
¢.-
t'.rc~ o3
40
0
/
t-
20
0
¢/// /
I
I
I
I
I
Birth
7
14
21
28
Postnatal Age
FIG. 2. Mean individual offspring weight as a function of postnatal age and diet.
Offspring Measures (1) Stillbirths and neonatal deaths. The dams were only checked twice daily for delivery, so an exact count of stillbirths was not possible. Four of the Ensure mothers had at least one dead newborn, whereas the Purina dams had none: X2 = 4.04 [1], p =0.04. There were no differences as a function of diet for liveborn birthweight, litter size, and neonatal deaths either from birth to culling on day 4, or during the neonatal period of days 4-28. There were no differences in the sex ratio on day 4 prior to culling. (2) Neonatal weight gain. Although there were no differences in liveborn litter weight as a function of maternal diet, litter weight differences were found at each weekly weighing thereafter and on day 45 when the rats were placed in the activity wheel. Prenatal diet was entirely responsible for the effects from birth to day 14, with postnatal diet beginning to have an effect by day 21. Postnatal diet accounted for all of the differences in weight gain by day 28. Newborns which received Purina prenatally were heavier on days 7 and 14, and those which received Purina postnatally were heavier when weighed on days 21 and 28. The heaviest group were those animals whose mothers were fed Purina as well as themselves (P-P). The Bonferroni approximation was applied to reduce the number of degrees of freedom since multiple tests were performed on the same data set. The prenatal effect was significant with a p=0.006 on day 7, p=0.0009 on day 14, and p=0.005 on day 21. The prenatal effect was no longer significant on d a y 28. The postnatal dietary effect was non-significant on day 7, and significant on
day 14, (p=0.036), day 21 (/9=0.004) and day 28 (,o=0.002). There were no interaction effects. Since for this number of tests, the Bonferroni inequality requires a significance level of 0.006 to be significant, only the day 14 postnatal test was not significant since that reached only 0.036. There were also no sex differences either for pre- or postnatal diets, nor interaction effects, i.e., diets affected both sexes in the same way. However, the males were significantly heavier than female littermates at all four postnatal weighings with p=0.0001 to p=0.000001. The Bonferroni correction factor did not affect these results. (3) Day 45 and 60 individual body weights. The group means at 45 days of age followed the same pattern as the earlier weights: P-P=223.6 g, E - P = 196.3 g, P - E = 193.6 g and E-E = 185.2 g. These weights were significantly different at a p =0.001 with a two way analyses o f variance. The prenatal factor still exerted an effect (p =0.012), but the postnatal factor accounted for most of the variance (p=0.004). Body weight at the next and last weighing on day 60 was no longer significantly affected by different diets, but the P-P rats were still heavier: P-P=300.2 g, E-P=268.3 g, P-E=277.2 g and E-E=265.8 g. There was no interaction between the diets, and effects when found, were due to main effects. Male weight gain differed as a function of diet on postnatal days 14, 21, 28 and 45, but not on days 7 and 60. Diet affected female weight gain significantly throughout the postnatal period. Diet had an earlier and more lasting effect
282
MARTIN, MARTIN, SHORES AND CHAO TABLE 2 M E A N P E R C E N T A G E O F RATS A C H I E V I N G CRITERION*
Measure
Day
P-P
P-E
E-P
E- E
Righting
3 4 5 6 7
14.3% 56.9 75.0 79.2 91.5
32.4% 56.5 70.2 87.2 91.5
39.8~ 49.3 80.4 94.6 96.4
17.0cA, 35.9 75.0 85.7 92.9
Ear pinnae (1 or 2)
3 4
19.0 100.0
30,9 74.0
39.0 91.0
15.0 98.0
Incisors (lower)
4 5 6
37.9 97.9 100.0
25.8 89.4 100.0
27.5 67.9 92.9
14.3 82.1 92.9
Incisors (upper)
4 5 6 7
01.7 20.8 54.2 93.6
08.1 34.0 70.2 91.5
01.4 37.5 78.6 91.2
03.6 33.9 75.0 94.6
Vertical righting
14 15 16 17
10.6 51.1 91.5 100.0
00.0 23.4 51.1 93.6
00.0 28.6 69.6 100.0
O1.8 01.8 41.1 89.3
Eyelid dysjunction
14 15 16 17
00.0 34.0 83.0 100.0
17.0 46.8 100.0 100.0
08.9 48.2 89.3 100.0
05.4 46.4 92.9 98.2
Open field (5 seconds)
15 16 17 18 19 20 21
14.9 21.3 42.5 31.9 42.5 41.3 47.8
10.6 19.1 23.4 53.3 53.3 60.0 51.1
12.5 23.2 20.0 30.9 36.4 50.9 56.4
08.9 23.2 39.3 54.5 72.2 74.5 84.3
*Litters culled to 4 males and 4 females on day 4.
on females, possibly due to their more marginal weight status. Male and female littermate weights were differenced. When littermate scores were not available, males and females from the same treatments were randomly matched and those scores differenced. An analysis of variance on the differenced scores yielded an F(3,36)=4.35,p =0.01. Diet did have an effect on weight sex differences. Diet no longer had an effect by day 60, although males remained significantly heavier than females, as they had been throughout the neonatal period with probability values ranging from p =0.0001 to 0.00001. Growth in the offspring as a function of maternal and neonatal diet is presented in Fig. 2. (4) Developmental measures. See Table 2 for the percentage of animals which achieved criterion for the measures described below. (a) Righting (dorsal position, days 3-8). There were no differences in the proportion of animals/litter which could right themselves on any test day. Marginal interactive effects between pre- and postnatal dietary conditions were found: p=0.03-0.08 but the main effects failed to reach significance. Eighty percent of all rats were able to right within the 3-sec
time limit by day 6. Diet also did not affect within or between sex differences. (b) Ear pinnae uncurling (days 3-5). A significant difference was found on day 4 on the proportion of rats with at least one ear uncurled, F(3,22)=3.96, p=0.02. The major effect was due to postnatal, rather than prenatal, diet: p=0.04. All ears were up by day 5. Sex had some effect on this measure, but the pattern was inconsistent. The correction factor required a significance level of 0.01 which was not found. (c) Lower incisors (days 4-7). There were no differences as a function of diet, and essentially all eruptions had taken place by day 6. Since this is usually one of them more sensitive measures, we pooled the data for the three days and did an exploratory analysis. F(1,22)=5.46, p=0.029, and the prenatal factor accounted for most of the variance. The offspring of mothers which received Purina had a greater number of eruptions. (d) Upper incisors (day 4-8). No differences were found as a function of diet. Although the main effect was nonsignificant on day 6, the prenatal factor accounted for most of the variance: p=0.038, Over ninety percent of all incisors
P E R I N A T A L DIET A N D O F F S P R I N G F U N C T I O N had erupted by day 7. No intra nor intersex differences were found. This measure has never been as useful as lower incisor eruption, possibly because it is more difficult to measure. (e) Vertical righting (days 7-19). There were no reliable differences on days 7-14 due to the large number of failures. Main analyses of variance were significant for days 15-19 with probabilities of 0.02 and 0.03. The data were pooled over days 13-19 for an exploratory analysis. F(3,22)=4.13, p=0.018, and the postnatal factor was unsurprisingly responsible for the differences: p =0.007. This difference was not significant when the correction factor was applied: 0.004 Almost all of the rats were landing on their feet by day 18, with the P-P animals performing at a somewhat higher level. (f) Eyelid dysjunction (days 13-17). There were no differences as a function of diet, and all eyes were opened by day 17. No sex differences were found. (g) Open field (days 7-21). The excessive variability found with this measure, particularly prior to eye opening on day 14 made interpretations difficult. There was little evidence for differences due to diet, and sex differences were inconsistent. Postnatal diets had a greater effect than prenatal after day 14. (h) Activity wheel (day 45). The sexes were not combined as in the previous analyses, since sex and activity level tend to be highly correlated in older animals. Square root transformations were used to stabilize the variances. Postnatal diet accounted for most of the effects on male activity: F(1,36)=5.30, p=0.027. The P-E and E-P males were more active than were the males in the other two conditions. Since an animal's weight affects its activity level [14], weight was covaried with the number of wheel revolutions using a multiple linear regression technique. Weight as a function of diet was positively correlated with numbers of wheel revolutions, F(3,36)=12.46, p=0.00001. Therefore, weight differences did not explain the results. The postnatal diet still accounted for most of the variance p =0.0004, but the interaction term was significant as well (,o=0.007). In contrast, diet had only a modest effect on female activity, F(3,35)=3.97, p=0.01, with the interaction term being responsible for most of the effects: F(1,35)=7.59, p=O.O09. When weight was covaried with the number of revolutions, the significant differences among the four dietary conditions vanished. The lightest weight females were the most active, i.e., the E-E rats. Therefore, weight differences as a result of differential diet could have accounted for female differences in activity, but this was not true of their male littermates. Parenthetically, it should be noted that since diet determined the body weights, the female activity differences were real, but the covariance analysis was unable to parcel out the effects. It cannot be determined from this study if female activity level was due to weight change as a function of diet, or to other dietary effects such as alterations in neural structure which might influence function. Weight was included as a factor in the 2×2 analysis of male/female difference scores, F(5,38)=6.10,p=0.0004. This analysis is equivalent to a repeated measures analysis of sex within diet. The significant results were due to the postnatal dietary factor: p =0.007, as well as the male weight factor: p=0.006. Female E-E rats were significantly more active then were E-E males, but the P-E and E-P males were more active than their female littermates. Activity had been measured between days 45 and 53 which is prior to sexual maturation when estrus cycle fluctuations would have resuited in increased female activity rates. When one extreme
283 male outlier was omitted from the analysis, the weight differences between the sexes accounted for most of the variance, and hence the activity differences. Hall [7] found that lighter weight rats were more active animals regardless of sex. This was only true for the females in our study, and the lack of male differences may have been due to the immaturity of the rats. DISCUSSION We reported earlier that maternal food intake and weight gain on both diets peaked during the second seven days and then declined [15]. The Ensure fed dams in the current study consumed more Kcal and gained more weight during days 8-14, although when measured over the entire 21 day period there were no dietary differences. Diet type had no effect on litter size, birth weight, neonatal death rate, nor sex ratio. Sustacal (Mead-Johnson) was the control diet in a maternal alcohol study by Sherwin et al., [18]. Half of the control offspring died between birth and day 21, in spite of the fortified diets available from birth. Since we found no difference in offspring death rate from birth to day 60 as compared with Purina controls, Ensure may be a more adequate diet than Sustacal for the rat. However, the pregnant rats in our study were placed on the liquid diet on day 1 of gestation, whereas Sherwin's rats were placed on Sustacal three weeks prior to breeding. The latter design may have lowered the nutritional status of the dams and contributed to the high neonatal death rate. We did find that Ensure had a deleterious effect on offspring weight gain when continued during the neonatal period. Prenatal diet was responsible for the weight status of the pups from birth through day 14, and postnatal diet began to exert a small effect by day 14. Both pre- and postnatal diets were contributing equally by day 21, and the postnatal diet was wholly responsible for differences by day 28 which was 6 days following weaning. Prenatal and postnatal exposure to Purina conferred a weight advantage. Body weights of the other three groups were virtually identical by day 60, but the P-P animals were still the heaviest as a group. Since the study was terminated on that date, it is not known if this advantage would have been maintained. Females were more adversely affected than males by the limitations of the Ensure diet. This difference could have been due to the effects of a less adequate diet on smaller animals, or to other sex differences. Uphoff et al. [22] fed gravid mice an Ensure diet which contained 20% ethanol. One control group received Purina chow, and a second group was offered Ensure with sucrose substituted isocalorically for the ethanol. The pups were sacrificed at birth. Live litter size did not differ between the two control groups, but the ethanol-Ensure litters were fewer in number, smaller in size, and had cardiac muscle defects. The authors speculated that the results may have been due to an interaction between a low protein diet Ensure and ethanol. There was no evidence in support of this interesting hypothesis. The alternate hypothesis was that ethanol or its metabolites caused the cardiac defects. Our P-P offspring reached criterion sooner than the other three groups on the two fighting measures. This diet conferred no advantage, however, on other measures such as eye opening, outer ear uncurling, vertical drop, open field and incisor eruption. Movement in the open field is affected primarily by whether the eyes have opened, by lighting levels, and by
284
MARTIN, MARTIN, SHORES AND CHAO
motivational factors since motor ability is not at issue alter the first few days of life. Eighty-three to 100 percent of all offspring had open eyes when examined on day 16. Over half of the Ensure-Ensure group were moving lbrward reliably during that time period, whereas the scores for the other three groups ranged from 33-38%. Since the other groups also had opened eyes, the tendency may reflect a higher activity level by the smaller, lighter weight animals. Hall [7] tound that poorly nourished, smaller rats were more active in the Wahmann wheels than were well nourished animals. It is often desirable to restrict the number of test days on newborns for pragmatic reasons. The age at which 75% of control animals reach criterion on each measure is a useful cut-off point. The 75% criterion for rats which received a Purina diet was: days 4-5 for dorsal righting, day 3 for ear pinnae uncurling, days 4-5 for lower, and days 6-7 for upper, incisor eruption. Righting from a vertical drop and eyelid dysjunction both occurred on days 15-16. Scores in the open field which are multidetermined, were too variable to be useful. Criterion was reached by the Ensure animals on the same days for dorsal righting, ear uncurling, lower incisor eruption and eye opening, but a day earlier for upper incisor eruption and a day later for vertical righting. Ensure animals were more consistent in the open field and 75c~ of them moved out reliably by days 1%20. Lochry, Hoberman and Christian [12] recently published a study on 8,000 rat pups. Weight gain was a more sensitive index of developmental status than a test battery which included eye opening, startle, ear pinnae uncurling, vertical drop, and dorsal righting measures. Activity, unfortunately, was not assessed. Prenatal treatment was limited to pharmaceutical agents and results may not be generalizable to other extrinsic agents. However, our results agree with theirs, since our developmental measures contributed almost no additional information beyond that obtained from the weight gain measure. Although developmental measures have a degree of face validity not found in a measure as simplistic as weight gain, results have not been consistent across laboratories which use test batteries. Some of these measures are highly subjective and dependent upon technician sophistication and consistency. Failure to thrive, on the other hand, which includes pool weight gain, is an excellent predictor of difficulties in later childhood. It may not be desirable to eliminate developmental test batteries, but careful measurement of weight gain and food consumption should be taken at regular intervals and correlated with both developmental and perform-
ance measures. It may be found that the developmental measures add little to the predictive value of function at older ages. S e x Dijy?rences
There were no reliable sex differences prior to sexual maturity, other than weight differences. The open field yielded little of value, whereas the activity wheel was very informative. This is not surprising, since the two tasks are different topographically and were taken at different ages. Activity measures do not correlate well with each other [20[, nor is activity a simple or homogeneous behavior [11]. Intersex differences in the wheel were primarily due to differences in body weight between males and females. Differences in activity among the 4 groups of males were not due to weight differences, but to some other factor or factors. We are now better able to predict when developmental milestones are reached in the Sprague-Dawley strain as a function of two diets. However, results from this study and another one by Lochry et al. [ 12] would seem to indicate that little is to be gained by measuring maturation, if careful growth records are kept. Voluntary activity, as measured in the wheel, does contribute useful information, particularly for male rats. We also found that Purina chow is not the appropriate control diet when Ensure is the experimental diet. The Purina diet during gestation and lactation conferred a weight advantage which could not be equalled by offspring prenatally exposed to Ensure, even when the animals received Purina for 60 days postnatally. The rats were at an even greater disadvantage when Ensure was continued throughout the postnatal period (E-E), or begun postnatally (P-EL The E-P rats were able to overcome most of their initial weight disadvantage. Unsupplemented Ensure as a total diet would seem to be adequate for the gestational period, assuming that mothers had been given Purina prior to pregnancy. It should not be continued after birth when nutritional requirements are more stringent. Although the liquid diet is calorically adequate, it is almost certainly deficient in protein, if not in other important nutrients. Supplementation with vitamins, minerals, and protein could conceivably extend its usefulness into the neonatal period. If Purina chow is used for a "normal" control group, then a second control group which is given the liquid diet, without the addition of drugs, is necessary to separate drug from dietary factors.
REFERENCES I. Alder, S. and G. Zbinden. Methods for the evaluation of physical neuromuscular and behavioral development in rats in early postnatal life. In: Methods in Prenatal Toxicology. edited by D. Neuberg, H. J. Merker and T. E. Kwasigroch. Stuttgart: G Thieme Publ., 1977, pp. 175-185. 2. Baker, D. E. J. Reproduction and breeding. In: The Laboratory Rat: Biology and Diseases, edited by H. J. Baker, J. R. Lindsey and S. H. Weisbroth. New York: Academic Press, 1979, Vol 1, Chap 7. 3. Becket, R. F, and J, C. Martin. Vital effects of chronic nicotine absorption and hypoxic stress during pregnancy and the nursing period. Am J Obstet Gynecol 110: 522-533, 1971. 4. Breuer, L. H., Jr., W. G. Pond, R. G. Warner and J. K. Loosli. A comparison of several amino acid and casein diets for the growing rat. J Nutr 80: 243-250, 1963.
5. Farris, E. J. Breeding of the rat. In: The Rat in Laboratory Investigation, edited by E. J. Farris and J. Q. Griffith, Jr. Philadelphia: JB Lippincott Co., 1949. 6. Freeman, M. F. and J. W. Tukey. Transformations related to the angular and square root, Ann Math Star 21:607-611, 1959. 7. Hall, R. D. The role of body size in climbing and locomotor behavior of protein malnourished and well nourished rats. Dev Psychobiol 17: 91-106, 1984. 8. Holson, J. F., W. J. Scott, D. W. Gaylor and J. G. Wilson. Reduced interlitter variability in rats resulting from a restricted mating period, and reassessment of the "litter effect." Teratology 14: 135-142, 1976. 9. International Encyclopedia of Statistics, edited by W. H. Kruskal and J. M. Tanur. New York: Free Press, Vol 1, 1978, p. 565. 10. Jensh, R. P., R. L. Brent and M. Barr, Jr. The litter effect as a variable in teratologic studies of the albino rat. J Anat 128: 185--192, 1978.
PERINATAL
DIET AND OFFSPRING
FUNCTION
11. Ljungberg, T. Reliability of two activity boxes commonly used to assess drug induced behavioral changes. Pharmaco/Biochem Behav 8: 191-195, 1978. 12. Lochry, E. A., A. M. Hoberman and M. S. Christian. Comparative sensitivity of pup body weight and commonly used developmental landmarks. Teratology 32: 28A, 1985. 13. McClain, R. M. and J. M. Rohrs. Potentiation of the teratogenic effects and altered disposition of diphenylhydantoin in mice fed a purified diet. Toxicol Appl Pharmacol 77: 86-93, 1985. 14. Martin, J. C. and D. C. Martin. Voluntary activity in the aging rat following maternal drug exposure. Neurobehav Toxicol Teratol 3: 261-264, 1981. 15. Martin, J. C., D. C. Martin, B. Radow and G. Sigman. Blood alcohol level and caloric intake in the gravid rat as a function of diurnal period, trimester and vehicle. Pharmacol Biochem Behav 8: 421--427, 1978. 16. National Academy of Sciences. Nutrient Requirements o f Domestic Animals, Vol 10: Nutrient Requirements o f Laboratot3' Animals, 1972, pp. 56-93. 17. Palmer, A. K. Statistical analysis and choice of sampling unit. Teratology 10: 301-302, 1972.
285
18. Sherwin, B. T., S. Jacobson, S. L. Troxell, A. E. Rogers and R. W. Pelham. A rat model (using a semipurified diet) of the Fetal Alcohol Syndrome. In: Currents in Alcoholism: Recent Advances in Research and Treatment, edited by M. Galenter. New York: Grune and Stratton, 1980, pp. 15--30. 19. Smart, J. L. and J. Dobbing. Vulnerability of developing brain I1. Effects of early nutritional deprivation on reflex ontogeny and development of behavior in the rat. Brain Res 28: 85--95, 1971. 20. Tapp, J. T., R. S. Zimmerman and P. S. D'Encarnacao. Intercorrelational analysis of some common measures of rat activity. Psychol Rep 23: 1047-1050, 1968. 21. Thompson, W. R., J. Watson and W. R. Charlesworth. The effects of prenatal maternal stress on offspring behavior in rats. Psychol Med [Monogr Suppl] 76: 1-26, 1962. 22. Uphoff, C., C. N. Battie and R. Toth. Cardiac muscle development of mice exposed to ethanol in utero. Teratology 30:119129, 1984. 23. Weil, C. S. Selection of the valid number of sampling units and a consideration of their combination in toxicological studies involving reproduction, teratogenesis or carcinogenesis. Food Cosmet Toxicol 8: 177-182, 1970. 24. Well, C. S. Choice of the number of sampling units in teratology. Teratology 10: 301, 1975.