INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE FATTY MUTATION causes early onset obesity because of a defect in a leptin receptor gene (3, 10, 22, 34, 37). Fatty (fa/fa) rats accumulate more adipose mass than their littermates soon after birth (9, 13, 14, 15, 20, 23, 27), but fatty animals remain visibly indistinguishable from the other genotypes until the fourth week of life (6, 14, 19). The original description of the mutation noted that obesity became apparent, based on body weight and appearance, as early as 3 wk of age (37). Others have commented that fatty animals become notably obese after weaning (6, 14, 19). In this context, weaning usually refers to an abrupt separation of the pups from the dam around 3 wk of age, rather than the natural gradual cessation of nursing that occurs as pups mature (29, 30).

These observations imply that weaning, or maternal separation, is somehow related to the progression of obesity in fatty rats. In fact, some investigators have found that maternal separation causes rapid changes in exercise (28), insulin (32), and hepatic gene expression (26) in fatty rats. An alternative explanation is that a developmental program that alters fa gene action happens to coincide with the customary age for separating pups from their dams. If this is the case, weaning practices could confound the interpretation of data collected around the developmental transition. Abrupt removal of the dam alters the thermal and nutritional environment. Furthermore, maternal separation can cause developmental changes to occur earlier than they otherwise would by activating stress responses (8).

The objective of this study was to develop a model for analysis of developmental variation in fa gene action. In pilot studies we observed that the common practice of separating dams from their pups at 21 days of age increased the variation in growth at precisely the age when fatty animals began to diverge from their littermates. We suspected that maternal separation was responsible for this increased variation. Therefore, we delayed maternal separation from 3 wk of age to 4 wk of age and analyzed fa genotype effects on early growth using a quantitative model that accounted for genotype, sex, and litter effects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. A colony of Zucker-fa rats was established at the Pennington Biomedical Research Center (PBRC) with breeders obtained from Harlan Sprague Dawley (Indianapolis, IN). The PBRC population is maintained with brother by sister matings to reduce genetic variation and with selection for large litter size to minimize inbreeding depression. Sires were removed when the dams appeared pregnant and before the litter was delivered. Pregnant dams are checked at least twice a day to determine when their pups are born. The day of birth is considered 0 days of age. Animals are housed in large plastic breeder boxes with ad libitum access to Laboratory Rodent Diet 5001 (Purina Mills, St. Louis, MO) and tap water. Beginning at 14 days of age, food was placed on the cage floor to allow the pups easy access. The light cycle is maintained at 12 h of dark followed by 12 h of light. The room temperature was maintained at 68 to 74°F and 55 to 60% humidity. All animal studies were reviewed and approved by the PBRC Institutional Animal Care and Use Committee.

Genotype detection. The fa mutation creates a restriction site for Msp I that can be used to detect fa genotype (10, 20, 33). However, Msp I occasionally fails to cut DNA to completion in our hands, which can cause genotype misclassification. To overcome this problem, we engineered an alternative restriction site by a strategy that has been described previously (33). PCR primers were designed to amplify 101 bp of DNA flanking the fa mutation. The reverse primer, RLepr, was selected to anneal adjacent to the site of the mutation, and a single base substitution (CG) was made in the second base from the 3' end of the primer to create a Pvu II restriction site in the wild-type allele and no restriction site in the mutant allele. The primer sequences were fLepr, 5'-CGTATGGAAGTCACAGA-3', and RLepr, 5'-GAATTCTCTAAATATTTCAGC-3'; the base substitution in RLepr is underlined. Primers are mixed at 200 nM each in 50 mM KCl, 10 mM Tris · HCl (pH 9 at 25°C), 0.1% Triton X-100, 3 mM MgCl₂, 0.2 mM each dATP, dCTP, dGTP, and dTTP, 0.12 U Taq DNA polymerase, and 2 µl of HotSHOT DNA in a 10-µl volume. PCR amplification was carried out in a thermal controller (model PTC-100; MJ Research, Watertown, MA) at 95°C for 2 min, followed by 30 cycles of 95°C for 30 s, 55°C for 45 s, and 72°C for 30 s. Four microliters of PCR product was digested with Pvu II (New England Biolabs, Beverly, MA) in a 20-µl volume for 1 h at 37°C. Twelve microliters of Pvu II-treated PCR product was loaded onto an 8 × 10 cm 12% polyacrylamide gel and electrophoresed 1 h at 15 V/cm. Gels were stained with ethidium bromide and exposed to ultraviolet light. Images were recorded as TIFF files for documentation. The wild-type genotype produces an 80-bp band, and the mutant genotype produces a 101-bp band; both bands are visible in heterozygotes.

Effects of fa genotype on postnatal growth. The objective of the first experiment was to estimate fa genotype effects on growth during the first 5 wk of life. Twenty-six litters including 312 pups were used in this experiment. Litter size ranged from 8 to 15 pups, with a mean of 12 pups per litter. We did not adjust litter size to minimize interference with growth. Pups were marked during the first days of life and weighed daily to 35 days of age. Body weight was automatically recorded to the nearest 0.01 g on a Mettler PM2000 balance linked to an Excel spreadsheet via BalanceLink software (VWR Scientific, West Chester, PA).

Statistical analyses. Two statistical analyses were performed on the data set. In the first step we analyzed fa genotype effects on daily body weight. Although we were primarily interested in fa genotype effects, litter and sex also have large effects on growth that should be accounted for to obtain accurate estimates of fa genotype effects on early growth. The analysis was performed under a mixed linear model, in which genotype and sex were treated as fixed effects factors and litter was treated as a random effects factor. With the large number of subjects in the data set, as well as the complexity of the model, simultaneous analysis across all ages required more computer memory than was available. Therefore, we analyzed the data in 27 separate ANOVA, one for each day. The fa genotype effects were considered to be statistically significant when their P values were less than the Bonferroni adjusted P value 0.0019 (0.05/27). The analysis was implemented with the MIXED procedure of SAS version 6.12 (17).

The fa genotype effects on growth in small litters. To determine whether fa genotype effects on growth were different in small litters, we analyzed fa genotype effects on litters that were reduced to four pups each during the first week of life. Pups from 24 litters were genotyped, and excess animals were killed by decapitation. Pups were weighed daily from 15 to 27 days of age. The fa genotype effects on daily gain were analyzed as described above. The fa genotype effects were considered to be statistically significant when their P values were less than the Bonferroni adjusted P value 0.004 (0.05/12).

Daily analysis of fa genotype effects on additional growth traits. A group of 292 rats from 36 litters were genotyped early in life. The wild-type and fatty genotypes are the most informative groups, so we kept all animals with these genotypes. However, more heterozygotes were produced than were needed. We kept some heterozygotes in every litter and killed excess heterozygotes by decapitation during the first week of life to reduce the number of animals to be processed. The average litter size was eight pups in this experiment. Body weight was measured daily beginning at 15 days of age to acclimate pups to handling. Body weight on the next to last day of life was subtracted from body weight on the last day of life to estimate daily gain on the last day of life. Litters remained with their dams until they were killed by decapitation at 20-24 days of age. Blood was collected from the neck wound into tubes containing 10 µl of 100 mM EDTA for analysis of plasma leptin, insulin, corticosterone, and glucose. Plasma insulin and leptin were measured with rat insulin and rat leptin radioimmunoassay (RIA) kits (Linco Research, St. Charles, MO). Plasma corticosterone was measured with a commercial RIA kit also (ICN Biomedicals, Costa Mesa, CA). Plasma glucose was measured with the Infinity Glucose Assay (Sigma, St. Louis, MO). Inguinal adipose tissue, liver, and stomach were dissected and weighed. The stomach contents were removed, and the empty stomach was weighed for determination of stomach contents.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The first experiment was designed to test the hypothesis that fa genotype effects on growth change rapidly at a specific age when pups are left with their dams through 28 days of age. In the first step of the analysis, fa genotype effects on body weight were estimated. No significant sex × genotype interactions were found at any age, indicating that males and females respond similarly to fa genotype. Therefore, the growth curve can be represented as sex-averaged least squares means (Fig. 1A), with the understanding that the means for males and females are shifted above and below the sex-averaged means respectively. Note that fa/fa mutants diverge from the other genotype groups during the fourth week of life, which corresponds with the age when fa/fa mutants became visibly obese (14).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Effects of fa genotype on daily body weight. A: sex-averaged least squares means for each genotype group. B: body weight as a percentage of the wild-type means. C: log of the P values for fa genotype effects on daily body weight. The line marks a P value of 0.0019; P values of less than 0.0019, which are plotted above the line, are significant. A P value of 1.0 × 1030 is the smallest estimate that can be obtained with the software, so any P value that was less than 1.0 × 1030 was plotted as 1.0 × 1030.

The problem with plotting body weight means in this format is that the scale changes substantially with age, emphasizing the effects at later ages and rendering the early effects practically invisible. One way of removing the influence of change in scale with age is to plot the means as a percentage of the wild-type means (Fig. 1B). This is not a separate statistical analysis but only an alternative method of expressing the effects. This plot shows that fatty animals diverge from the other genotype groups early in life, becoming significantly heavier by 10 days of age (Fig. 1C). The early weight difference peaks at 16 days of age, when fatty animals are about 5% heavier than their littermates. This weight difference is not sufficient to allow the fatty animals to be identified visually. After 16 days of age, fatty mutants then begin to converge with the other genotype groups (Fig. 1B), such that no significant differences in body weight are detectable from 20 to 22 days of age (Fig. 1C). This trend rapidly reverses after 22 days of age, when fa/fa mutants diverge again from the other genotype groups (Fig. 1, B and C).

The rapid divergence in body weight of fatty animals after the third week of life suggests that a change in daily gain emerges at this age. To test this hypothesis, we analyzed fa genotype effects on daily gain, where gain is calculated as the difference in body weights between consecutive days. Again, no sex × genotype interactions were found at any age, so sex-averaged least squares means represent the response of daily gain to genotype and age (Fig. 2A). Daily gain appears almost constant from 7 to 18 days of age. After 18 days of age, all three genotype groups experience a sharp increase in daily gain, but there is still no significant effect of fa genotype on daily gain up to 22 days of age.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Effects of fa genotype on daily gain. Daily gain was calculated as the difference in body weights on consecutive days. A: sex-averaged least squares means for daily gain by each genotype group. B: log of the P values for fa genotype effects on daily gain in body weight. The line marks P = 0.0019; P < 0.0019, which are plotted above the line, are significant.

The fa genotype effect on daily gain changes markedly at 23 days of age, when it suddenly becomes very highly significant (P = 10¹⁸, Fig. 2B). The gain curves essentially bifurcate at this point; daily gain continues to increase in all genotype groups but at a much faster rate in fatty animals. This saltatory increase in daily gain suggests that a fundamental change in growth regulation occurs at this age. Because this change occurs in nursing rats, it must be independent of environmental changes induced by maternal separation.

In the first experiment, we did not adjust litter size. Our intention was to minimize investigator intrusion on growth measures. With an average of 12 pups per litter, litter size is likely to have had a substantial influence on growth. It might also have affected the timing of the change in daily gain. To determine whether small litter size would alter the timing of the change in daily gain, we reduced litter size to 4 pups per litter in 24 litters. Once again, daily gain was similar in all three genotype groups for the first 22 days of age (Fig. 3). Fatty animals appeared to gain slightly more than their littermates after 20 days of age, but the difference was not significant until 23 days of age. Daily gain continued to increase at a higher rate in fatty animals after this age (Fig. 3). These results suggest that large differences in energy intake during early growth do not alter the timing of the change in growth. The average body weight at 22 days of age was 42 g in litters with 4 pups per litter, compared with an average of 37 g in the experiment with 12 pups per litter.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effects of fa genotype on daily gain in litters reduced to 4 pups each (small litters). *P < 0.0001. P < 0.004 are statistically significant.

Such a rapid and sustained increase in growth in healthy, growing young animals must be fueled by an increase in food intake. Although food intake is simple to measure in mature rats, and relatively easy to measure in suckling rats that are not eating solid food, direct measurement of food intake in rats that are suckling and consuming solid food is exceedingly difficult. Harris (6) has shown that the weight of the stomach contents is highly correlated with food intake over the previous 24 h and provides an estimate of food intake when direct measurements are impractical. Therefore, we used the weight of the stomach contents as a proxy measure of food intake and supported this with additional dependent variable measurements. In this experiment, rats were killed daily from 20 to 24 days of age for dissection of tissues and collection of blood. They remained with their dams until immediately before they were killed.

Once again, there were no effects of fa genotype on daily gain until 23 days of age, when the fa genotype effect became highly significant (Fig. 4A). This replicates the first two experiments and emphasizes the remarkably precise timing of the change in growth. (For comparison with the two previous experiments, average body weight was 33 g at 22 days of age in litters of 8 pups per litter.) In addition, there were no effects of fa genotype on the weight of the stomach contents until 23 days of age, when fatty animals had a 70% increase in stomach contents compared with their littermates (Fig. 4B). Liver weight was unaffected by fa genotype until 23 days of age, when liver weight in fatty animals increased by 14% over that of the other genotype groups (Fig. 4C). Plasma insulin levels were not significantly elevated in fatty animals at 20-22 days of age, but at 23 days of age, insulin was 2.4 times higher in fatty animals than wild types and the difference was highly significant (Fig. 5A). Despite the large changes in daily gain, stomach contents, and insulin, the plasma glucose was not affected by fa genotype, sex, age, or any of their interactions (Fig. 5B). Plasma corticosterone was also unaffected by genotype; however, there was a significant effect of age (Fig. 5C). The observation that corticosterone increases with age in this period has been reported by others (8). The rapid increases in daily gain, stomach contents, liver weight, and plasma insulin at 23 days of age support the hypothesis that a marked increase in energy intake begins after 22 days of age in fatty rats.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Effects of fa genotype vs. age for dependent variables associated with increased energy intake. A: daily gain on the last day of life. Each age includes values for subjects killed on that day. The genotype by age interaction was significant at P < 0.0001. B: weight of the stomach contents; genotype by age P < 0.0001. C: liver weight; genotype by age P < 0.0009.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Effects of fa genotype on glucose regulation. A: plasma insulin; genotype by age P < 0.0001. B: plasma glucose; no significant effects of genotype or age. C: plasma corticosterone; age P < 0.007.

Inguinal adipose pads weights of fatty rats were already significantly elevated by 20 days of age (Fig. 6A). The weight of the inguinal adipose pads was approximately doubled in fatty animals compared with the other genotype groups at this age. Plasma leptin was elevated approximately sevenfold in fatty rats by 20 days of age (Fig. 6B). Although both inguinal adipose mass and plasma leptin were already elevated early in life, both had highly significant genotype by age interactions that reflect the fact that they diverge even further with age. The changes in inguinal adipose mass and plasma leptin verify in our animals that the effects of fa genotype on leptin signaling precede the changes in growth that we describe above.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Effects of fa genotype on dependent variables associated with obesity. A: inguinal adipose pads weight; genotype by age *P < 0.0001. B: plasma leptin; genotype by age *P < 0.0001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

When Zucker rat pups remain with their dam through the fourth week of life, a rapid change in growth of fatty rats develops after 22 days of age. There are no significant effects of fa genotype on daily gain until 23 days of age, when the genotype effect suddenly becomes quite large. The increase in daily gain in fatty rats is accompanied by sharp increases in stomach contents, liver weight, and plasma insulin. The rapid onset of these changes suggests that a developmental switch activated during the fourth week of life causes a coordinated change in the set of genes controlling growth and development of overt obesity.

This interpretation is partly based on observations of Cheverud et al. (2) on developmental changes in genetic control of growth. They performed a genetic analysis of quantitative trait loci affecting growth of mouse strains selected for large and small body size. Their analysis identified multiple loci that affected early, middle, and late growth. Remarkably, the set of genes that controlled early growth from 1 to 3 wk of age was completely different from the set of genes that controlled late growth, from 6 to 10 wk of age. Major changes in the set of genes that control growth are likely to be associated with significant developmental milestones, such as birth, weaning, and puberty.

The timing of the developmental switch was remarkably similar in three experiments described here. Regardless of whether litter size was large (12 per litter), small (4 pups per litter), or moderate (8 pups per litter), daily gain of fatty pups separated rapidly from daily gain of the other genotype groups after 22 days of age. This indicates that the timing of the developmental switch is not determined by the amount of food available during the suckling period but is more likely to be related to an intrinsic developmental program. The fact that 22 days is also the normal gestation period for the rat suggests that the developmental switch could promote independence from the dam in time for delivery of another litter. It should be noted than none of the dams in these experiments were carrying a second litter; because the sires were routinely removed before the litters were born.

We can rule out maternal separation as a factor determining the timing of the developmental switch, because pups were left with their dams until they reached 28 days of age in the first two experiments, long after activation of the developmental switch. Therefore, in our experiments pups continued to nurse and consume solid food up to the time they were separated from their dams. When pups are left with their dams past 3 wk of age, weaning occurs gradually and nursing may continue up to 34 days of age (29, 30). The common husbandry practice of separating dams from their pups at 3 wk of age is likely to activate stress responses that could alter the timing of events that normally develop spontaneously (8). For example, an increase in jejunal sucrase activity that occurs with early weaning is blocked by adrenalectomy (8). In fatty animals, hepatic tyrosine aminotransferase levels increase within a day of maternal separation at 3 wk of age, but this effect is also blocked by adrenalectomy (26). Plasma insulin also increases in fatty animals by ~90% within a day of maternal separation at 3 wk of age (32). Physical activity decreases when fatty animals are separated from their mothers, unless they are left with their dams up to 44 days of age (28). In this case, activity decreases at 30 days of age. These observations suggest that maternal separation stimulates responses that would be expected to contribute to rapid weight gain in the fatty rat. Separating stress responses from spontaneous developmental events becomes difficult when maternal separation precedes activation of the developmental switch.

The developmental switch is marked by a rapid increase in stomach contents, liver weight, and plasma insulin, which indicates that fatty rats develop robust hyperphagia at this age. In our experience, fatty rats are not visibly distinguishable from their littermates at 3 wk of age, in contrast to the initial description of the timing of the development of obesity (37). However, fatty animals can be reliably identified by 24 days of age by lifting them by their tails and comparing their bellies to those of their littermates. In fatty rats, it is the expanding belly rather than excess body fat that indicates which pups will become obese. Although the fatty animals had elevated inguinal adipose mass prior to the developmental switch, the increase was relatively modest, about twofold. Remarkably, this is very different from the development of obesity in db/db and ob/ob mice (data not shown). These obese mice are visibly distinguishable from their littermates during the third week of life by the appearance of excess adipose mass, and their body weights diverge continuously from the other genotype groups from the second week of life. This could be caused by species differences or perhaps by the nature of the leptin receptor mutation in fatty animals. Although db/db and ob/ob mice each lack an essential mechanism for activating the long form leptin receptor (16, 36), the long form leptin receptor in fatty rats retains some function (4, 35). There is some evidence for a developmental change in response to leptin in mice. Mistry et al. (21) found that 17-day-old ob/ob mice respond to leptin injections by increasing their oxygen consumption, but they do not alter their food intake. By 28 days of age ob/ob mice respond to leptin injections by increasing their oxygen consumption and decreasing their food intake. Our studies would predict that ob/ob mice would begin to decrease their food intake in response to leptin around the beginning of the fourth week of life.

Several investigators have sought to determine when the onset of hyperphagia occurs in fatty rats, and they have reached different conclusions. Dryden et al. (4) measured milk intake in fatty pups and their littermates at 10, 15, and 20 days of age. They constructed a cage that prevented pups from access to the dam's food. Milk intake was estimated by measuring dilution of tritiated water in the pups every 4 h for 24 h. Under these conditions, no differences in milk intake were found. McLaughlin and Baile (19) concluded that food intake was already elevated at 3 wk of age; however, their report states that the rats were ~3 wk old when they were first tested and the fatty animals were visibly obese. It is possible that these rats were slightly older than 21 days when they were tested. Buchberger and Schmidt (1) estimated milk intake during early development by weighing pups before and after feeding bouts. They tested pups at 5 and 15 days of age and also found no difference in milk intake between fatty animals and controls at these earlier ages. Kowalski et al. (12) used a feeding test that estimated independent feeding behavior. They transferred litters to a warm chamber for 4 h to stimulate hunger. After food deprivation, pups were given access to a mixture of milk and cream and weighed after 20 min to estimate food intake. Under these conditions, fatty animals ate significantly more food than controls at 12, 15, and 18 days of age. When all these experiments are considered together, it appears that suckling is not affected by fa genotype, but independent feeding can be increased in fatty rats early in life. However, the independent feeding experiments differ substantially from usual rearing conditions. Furthermore, maternal separation is stressful to young rats and could interfere with accurate measures of food intake (8). The mixed results suggest that there is no definitive way to prove whether fatty animals consume more, less, or the same as their littermates during early life. However, the rapid changes in growth, stomach contents, liver weight, and plasma insulin after 22 days of age provide ample evidence that robust hyperphagia develops at this age.

The fact that inguinal adipose mass and plasma leptin levels are elevated earlier than the developmental switch suggests that leptin receptor signal transduction is active early in life. This has been confirmed by the observation that injections of leptin in young rats are capable of reducing body fat (13, 27). However, the response is mediated by increased energy expenditure and not by reducing food intake (27). These data and those of Mistry et al. (21) in ob/ob mice suggest that rats and mice are likely to develop an anorectic response to leptin during the fourth week of life, when the developmental switch is activated. This implies that leptin-responsive genes that regulate feeding behavior are likely to be affected by the developmental switch.

Although the increased growth rate that follows the developmental switch appears to be fueled by the onset of robust hyperphagia, it is likely that other metabolic adaptations accompany the activation of the switch. For example, the ability of fatty rats to maintain normal body temperature is greater at day 25 than at day 16 (11), suggesting that thermoregulatory changes also occur across this transition. The developmental switch could also alter nutrient partitioning. Reeds et al. (25) measured protein deposition in fatty rats and their littermates from 16 to 27 days of age. They found that body protein was similar between fatty animals and lean rats at postpartum day 16 but decreased in fatty animals after 16 days postpartum. Between 23 and 27 days postpartum the fractional rate of protein synthesis was similar in obese and lean rats. The relevance of these observations to the developmental switch we describe is somewhat clouded, because their rats were separated from their dams at postpartum day 21, which could have stimulated a stress response that affected protein synthesis. The same group has also demonstrated that both fatty and lean rats deposit very little lipid from 18 to 22 days of age, but fatty rats synthesized fatty acids at a much higher rate from 23 to 27 days of age (6). These observations emphasize that many metabolic adaptations accompany the appearance of robust hyperphagia that is triggered by the developmental switch.

It is possible that the developmental switch directly affects expression of the leptin receptor. This hypothesis is supported by the observations of Matsuda et al. (18), who followed prenatal and postnatal expression of the long form of the leptin receptor in an unspecified rat strain. They noted that leptin receptor expression was lower in the paraventricular nucleus during suckling compared with adult rats. However, they also found a substantial increase in plasma leptin at postnatal day 21. The levels of leptin in their rats were as high as the levels of the fatty pups in our experiment. It is difficult to resolve these observations, because we see no evidence of a leptin surge at this age in lean Zucker rats. However, there could be substantial differences due to genetic background or husbandry conditions. These data suggest that an analysis of leptin receptor expression across the developmental switch is necessary.

Perspectives