There is growing evidence that the postnatal environment can have a major impact on the development of obesity and insulin resistance in offspring. We postulated that cross-fostering obesity-prone offspring to lean, obesity-resistant dams would ameliorate their development of obesity and insulin resistance, while fostering lean offspring to genetically obese dams would lead them to develop obesity and insulin resistance as adults. We found that obesity-prone pups cross-fostered to obesity-resistant dams remained obese but did improve their insulin sensitivity as adults. In contrast, obesity-resistant pups cross-fostered to genetically obese dams showed a diet-induced increase in adiposity, reduced insulin sensitivity, and associated changes in hypothalamic neuropeptide, insulin, and leptin receptors, which might have contributed to their metabolic defects. There was a selective increase in insulin levels and differences in fatty acid composition of obese dam milk which might have contributed to the increased adiposity, insulin resistance, and hypothalamic changes in obesity-resistant cross-fostered offspring. These results demonstrate that postnatal factors can overcome both genetic predisposition and prenatal factors in determining the development of adiposity, insulin sensitivity, and the brain pathways that mediate these functions.
- diet-induced obesity
- Agouti-related peptide
maternal intake of a high-fat diet and the presence of obesity during pregnancy and lactation promote obesity in offspring, particularly in individuals with an obesity-prone genetic background (21, 32, 36, 55). However, there is growing evidence that a variety of manipulations during the postnatal environment can override prenatal factors and increase the risk of developing metabolic disorders such as obesity and diabetes later in life. Raising pups in small litters leads to increased milk intake and the development of obesity, whereas raising pups in large litters has the opposite effect (16, 23, 26). Administration of insulin during early postnatal development results in obesity and altered development of hypothalamic neuropeptide systems involved in energy homeostasis (22, 50). Cross-fostering pups is another method of assessing the role of the postnatal environment in later development. A single study in mice has demonstrated that fostering inbred obesity-prone mouse pups with obesity-resistant dams attenuates their obesity, whereas fostering obesity-resistant pups with obesity-prone dams causes them to develop obesity and insulin resistance (51).
Unlike many rodent models, the majority of human obesity is inherited as a polygenic disorder (6). For that reason, we have used a rat model of diet-induced obesity (DIO), in which obesity is also inherited as a polygenic trait (33, 38) in association with insulin resistance (33), hypertension (14), and hyperlipidemia (62). These rats also have several abnormalities of brain function, which antedate and may contribute to the later development of obesity (35). When raised through gestation and lactation by obese dams, adult DIO offspring become more obese and insulin resistant and have abnormal development of noradrenergic and serotonergic brain circuits compared with offspring raised by lean DIO dams (32). On the other hand, diet-resistant (DR) offspring do not become more obese as adults, even if their dams are obese throughout gestation and lactation. These studies suggest how critical interactions between genotype and the perinatal environment affect the development of obesity, insulin resistance, and brain circuits involved in the regulation of energy homeostasis.
Although the noradrenergic system is almost fully formed in rats at birth (54), neurons in the hypothalamic arcuate nucleus (ARC), which produce anabolic peptides such as neuropeptide Y (NPY) and agouti-related peptide (AgRP), and catabolic peptides such as α-melanocyte stimulating hormone derived from proopiomelanocortin (POMC) do not reach their targets in the paraventricular nucleus (PVN) and lateral hypothalamus (LH) until well into the second week of postnatal life (7, 19, 40, 69). This suggests that their development might be particularly affected by specific alterations in the early postnatal environment. For this reason, we postulated that cross-fostering offspring of obese DIO dams with lean DR dams might ameliorate their development of obesity and insulin resistance as adults and that fostering DR offspring with obese DIO dams would have the opposite effect. In addition, we hypothesized that such changes would be associated with alterations in the expression of various hypothalamic neuropeptides and receptors involved in the regulation of energy homeostasis.
Dams, Cross-Fostering, and Postweaning Manipulations
Animal usage was in compliance with and approved by the Institutional Animal Care and Use Committee of the East Orange Veterans Affairs Medical Center. All studies were carried out in keeping with the “Guiding Principles for Research Involving Animals and Human Beings” of the American Physiological Society (1). Dams were from our two colonies of rats selectively bred for their propensity to develop either DIO or DR (33). These colonies were originally derived from outbred Sprague-Dawley rats (Charles River Laboratories). Rats were kept at 23–24°C on a reversed 12:12-h light-dark cycle with lights off at 1000. DR dams (n = 14) were fed Purina rat chow (#5001) containing 3.75 kcal/g with 28.1% of total calories as protein, 12.1% as fat, and 59.8% as carbohydrate, which is primarily in the form of complex polysaccharide. Lean DIO dams were fed chow (n = 12) and obese DIO dams were produced by feeding them a high-energy (HE) diet for 3 mo (n = 11). The HE diet contains 4.41 kcal/g with 17% of total calories as protein, 32% as fat, and 51% as carbohydrate, 50% of which is sucrose (30) (Research Diets #D12266B). After 3 mo on their respective diets, all dams underwent tail bleeding for plasma glucose, insulin, and leptin levels, and then they were mated with males of the same genotype. Dams were kept on their respective diets through gestation and weaning. Second and third blood samples were drawn during the second week of gestation and lactation, respectively. At birth, DR offspring were randomly cross-fostered with other DR, lean, and obese DIO dams, and offspring of obese DIO dams were randomly cross-fostered with DR, lean, and other obese DIO dams. All litters contained 10 pups (5 male: 5 female) of only DR, lean DIO, or obese DIO offspring, respectively. Offspring were weighed weekly during lactation. At 21 days of age, male pups were weaned, individually housed, and fed chow and water ad libitum. The six groups included DR offspring cross-fostered to DR (15), lean (12), and obese (9) DIO dams and offspring of obese DIO dams cross-fostered to DR (13), lean (17), and obese (7) DIO dams. After weaning, food intake and body weights were recorded weekly. At 10 wk of age, after an overnight fast, an oral glucose tolerance test was administered between 0900 and 1000, and rats were placed on a HE diet for an additional 4 wk. At this time, they underwent a second glucose tolerance test, and the rats were continued on the HE diet. One week later, the HE diet was removed at 0800, and rats were killed by decapitation between 1000 and 1300. Brains were quickly removed, frozen on dry ice, and stored at −80°C for later assay. Retroperitoneal, perirenal, mesenteric, epididymal, and inguinal adipose depots were removed and weighed, and trunk blood was collected for plasma glucose, insulin, and leptin levels.
Milk Composition Study
Milk samples were collected on day 7 during lactation from additional groups of DR dams (n = 5), lean DIO dams fed chow (n = 5), and obese DIO dams fed a HE diet (n = 6). After separation from the pups for 30 min, dams were anesthetized with pentobarbital sodium (35 mg/kg ip) and injected with 4 IU of oxytocin (Sigma Chemical, St. Louis, MO) to stimulate milk flow. Milking was initiated 5 min after oxytocin injection. Milk drops were expressed manually by gentle stroking of the nipple and collected with an Eppendorf pipette. Milk samples were stored at −20°C until analysis.
Milk Lipid Fatty Acid Composition
Rat milk lipid was transesterified into fatty acid methyl in preparation for analysis. Rat milk samples (100 μl) were transferred to 20 ml glass tubes sealed with Teflon-lined screw cap closures. Samples were evaporated to dryness under nitrogen at room temperature. Toluene (1.0 ml) and methanolic base reagent (2.0 ml) were added sequentially and samples were incubated at 80°C for 60 min with frequent mixing. Samples were then cooled; 3.0 ml of ethyl ether and 3.0 ml of distilled water were added followed by mixing and 30 min centrifugation (2000 rpm). The upper organic layers were transferred to small tapered glass vials, dried by the addition of a small amount of anhydrous sodium sulfate and concentrated to ∼0.5 ml under nitrogen. Samples were analyzed by GC-MS and GC with flame ionization detection (FID). For GC-FID analyses, a Varian 3400 GC equipped with a 60 m × 0.32 mm ID high polar-wax capillary column containing a 0.15-μm film thickness was used. Injections (1.0 μl) of samples were made using an injection port temperature of 240°C and a split ratio of 50:1. The GC column was temperature programmed from 40°C (hold 5 min) to 240°C at a rate of 10°C per minute with a 10-min hold at the upper limit. The detector temperature was 250°C. Fatty acid methyl ester identifications were determined by GC-MS analysis, and secondary confirmation was obtained by comparison of GC retention times to that of a 37-component mixture of reference standards (Supelco, Bellefonte, PA). The percentage of fatty acid as their methyl ester derivative was calculated based on the GC-FID area. GC-MS analyses were performed on a Varian 3400 GC directly interfaced to a Finnigan Measurement Analytic Technologies 8230 high-resolution, double-focusing, magnet-sector mass spectrometer. Chromatographic conditions were identical to those described for GC-FID analysis. The mass spectrometer was operated in electron ionization mode scanning masses 35–550 once each 0.6 s with a 0.8 s interscan time.
Oral glucose tolerance test
After an overnight fast, a baseline blood sample was taken via tail nick for glucose and insulin in heparinized tubes. Rats were then gavaged with 0.5g/kg glucose and blood for insulin, and glucose was sampled via tail nick at 15, 30, 60, 90, and 120 min.
Assays of Insulin, Leptin, and Glucose
Plasma and milk insulin and leptin levels were analyzed with radioimmunoassay kits (Linco) using antibodies to authentic rat insulin and leptin, respectively. Milk samples were diluted 1:2 before leptin assay (29). Whole blood was used to measure glucose using a glucometer (Accuchek).
Hypothalamic Neuropeptide Expression by RT-PCR
Frozen brains were cut on a cryostat at −12°C in 300-μm sections centered on the hypothalamic PVN and on the midpoint of the ARC, which also contains the midpoint of the ventromedial (VMN) and the dorsomedial hypothalamic nuclei (DMN) through its compact portion and the LH. Cut sections were placed in RNAlater (Ambion) until micropunched. Micropunches of the PVN, ARC, VMN, DMN, and LH were performed by modifications (34, 39) of the method of Palkovits (48). Briefly, the slabs to be punched were placed on the base of a stereotaxic frame and visualized under an operating microscope. The syringe to which the punches were attached was mounted on the stereotaxic arm and the punches made under microscopic guidance. Micropunched brain nuclei were sonicated in a guanidinium thiocyanate solution and purified using magnetic beads (Ambion MagMax-96). Quantitation of mRNA was carried out by quantitative real-time RT-PCR, as previously described (34). Briefly, genomic DNA was removed with DNase, mRNA was reverse-transcribed with random hexamer priming using Superscript-3 (Invitrogen), and was treated with RNAseH (Ambion). Primer sets for each mRNA were designed by reference to published sequences, and their specificity was verified using GenBank. Primers and their sequence-specific 6-carboxy-fluorescein-labeled probes prepared by Applied Biosystems (Table 1) were sequenced and then quantified with an Applied Biosystems 7700 real-time PCR system set for 40 PCR cycles. Standard curves were generated from serially diluted pooled samples for each probe and for constitutively expressed mRNA (cyclophilin) to control for differences in amplification efficiency. Results were calculated from the standard curve relative to cyclophilin.
Data from DR offspring fostered with DR, lean, and obese DIO dams were analyzed separately from those for DIO offspring fostered with lean, obese, DIO, and DR dams because results from these groups were independent of each other. One-way ANOVA was used for single-point measures of total food intake; area under the curve (AUC) for glucose and insulin; insulin sensitivity index; terminal body and fat depot weights; and plasma glucose, insulin and leptin levels. When significant intergroup differences were found by ANOVA (P ≤ 0.05), post hoc comparisons were carried out by Bonferroni analysis. All statistical analyses, including rejection of statistical outliers, were determined using Systat statistical software. Feed efficiency was estimated by dividing the weight gain by the number of calories ingested over the same period of observation. For glucose tolerance tests, AUC for insulin and glucose levels over the 120-min testing period was calculated using the trapezoidal rule. Insulin sensitivity index (ISI) was calculated from the following formula (10,000/square root of [fasting glucose × fasting insulin] × [mean glucose × mean insulin during oral glucose tolerance test]) from Matsuda and Defronzo (42).
Dam Body Weight Gain and Plasma Hormone Levels
After 12 wk on their respective diets, DIO dams on both chow and HE diet gained significantly more weight and had higher insulin levels than chow-fed DR dams (Table 2). Although DIO dams fed a HE diet gained no more weight, their leptin levels were 96% higher than lean DIO and 150% higher than chow-fed DR dams, suggesting that they were indeed obese. Leptin levels in lean DIO dams did not differ from those in DR dams, suggesting that DIO dams became fatter on HE diet than either lean DIO or DR dams before impregnation. From the onset of impregnation to the second week of gestation, the obese DIO dams gained 38% more weight and had 130% higher insulin and 350% higher leptin levels on HE diet than did lean DIO dams on chow. Lean DIO and DR dams did not differ significantly in any of these parameters. However, there was a major improvement in body weight gain and apparent insulin sensitivity and adiposity (as indexed by plasma insulin and leptin levels, respectively) in obese DIO dams that occurred from the second week of gestation to the second week of lactation. The obese DIO dams lost 11% of their body weight, and their plasma leptin levels fell by 96% and their insulin levels by 72%. On the other hand, the DR and lean DIO dams gained 5% and 8% body weight, whereas their plasma insulin and leptin levels remained essentially unchanged over this same period.
Milk insulin levels of obese DIO dams were 128% greater than those of DR dams and did not differ significantly from those of lean DIO dams (Fig. 1A). Milk leptin levels were not significantly different among the groups, although there was a trend for leptin levels to be higher in the milk from obese DIO dams than that in the other groups. Obese DIO dam milk had 67% and 50% lower levels of total polyunsaturated fatty acids (PUFAs) and 80% and 77% lower levels of monosaturated fatty acids (MUFAs) compared with milk from lean DIO and DR dams, respectively (Table 3). Specifically, obese DIO dams had lower levels of oleic 18:1(n-9) and α-linolenic 18:3(n-3) acids, whereas the levels of γ-linolenic 18:3(n-6) and eicosadienoic 20:2(n-6) acids were higher compared with chow-fed DIO and DR dams.
Weight gain and metabolic parameters in male cross-fostered offspring
DR offspring cross-fostered with DR, lean, and obese DIO dams.
There were no differences in body weight among the offspring of DR, lean, or obese dams at 2 days of age. However, DR offspring fostered with obese DIO dams gained 12% more body weight from 2 days to weaning than those fostered with DR dams but not those fostered with lean DIO dams (Table 4). When fed chow from weaning to 10 wk of age, DR offspring fostered with obese DIO dams had similar weight gain, food intake, glucose, and insulin levels compared with those fostered with DR dams. However, these chow-fed offspring had 51% higher insulin AUC during an oral glucose tolerance test than those fostered with DR dams (Fig. 2B). The DR offspring fostered with lean DIO dams gained less weight and ate less but had equivalent glucose metabolism to those fostered to DR dams over 7 wk on chow. Although their body weight gains did not differ over an additional 4 wk on the HE diet, DR offspring fostered with obese DIO dams consumed 10% more calories, had 28% higher leptin levels, and 17% heavier total fat depot weights than those fostered with DR dams. They also had 11% higher glucose AUC compared with DR offspring fostered with DR dams and a 50% lower ISI (Fig. 2). Insulin AUC values did not differ among the groups. The DR offspring fostered with lean DIO dams ate 6% fewer calories of HE diet, had 10% lower final body weights, and 34% lighter total fat pad weights than those fostered to DR dams. Despite being less obese, DR offspring fostered to lean DIO dams had a 64% lower ISI (Fig. 2C).
Obese DIO offspring fostered with DR, lean, and obese DIO dams.
Although there were no intergroup differences in day 2 body weights, offspring of obese DIO dams fostered with DR dams gained 30% less weight through weaning than those fostered with obese DIO dams (Table 5). When fed chow from weaning to 10 wk of age, all groups gained the same amount of body weight even though offspring of obese DIO dams fostered with DR dams ate 11% fewer calories and had 39% higher-feed efficiency than those fostered with obese DIO dams. They also had 14% lower basal glucose levels and 18% lower glucose AUC (Fig. 3A), even though basal insulin levels and insulin AUC did not differ from those fostered with obese DIO dams. When fed a HE diet for an additional 4 wk, DR offspring fostered to DR dams ate 10% less than those fostered with obese DIO dams. However, the reduced intake produced no intergroup differences in body weight gain, terminal body weight, total fat pad weights, plasma leptin, or insulin levels. Despite this, offspring of obese DIO dams fostered with DR dams still maintained a 10% lower glucose AUC than those fostered with obese DIO dams (Fig. 3A). Although it did not reach statistical significance, the ISI of offspring of both lean and obese DIO dams fostered to DR dams tended to be improved on both chow and HE diet.
Hypothalamic Neuropeptide and Receptor mRNA Expression
DR offspring fostered with DR vs. obese and lean DIO dams.
Despite an increase in adiposity that might have been expected to reduce ARC AgRP mRNA expression, AgRP expression was actually 90% higher in DR offspring fostered with obese DIO dams than in those fostered with DR dams (Table 6). In the VMN, they had 31% lower mRNA expression of the long splice variant of the leptin receptor (LEPR-B), 28% lower insulin receptor (INSR), and 30% lower melanocortin-3 receptor (MC3R) mRNA compared with DR offspring raised with DR dams. On the other hand, DMN expression of LEPR-B, INSR, MC3R and suppressor of cytokines signaling-3 was elevated by 257%, 182%, 500%, and 205% and PVN neuropeptide Y1 receptor expression was 41% higher in DR offspring fostered with obese DIO dams compared with those fostered with DR dams, respectively. As opposed to DR offspring fostered with obese DIO dams, DR offspring fostered with lean DIO dams had 71% lower ARC, AgRP, and 57% higher VMN MC3R mRNA expression compared with offspring fostered with DR dams. In the DMN, they had increased expression of LEPR-B and INSR, which was similar to but less robust than that seen in DR offspring fostered with obese DIO dams. Finally, there were no intergroup differences in any parameters in the LH.
DIO offspring fostered with DR vs. lean and obese DIO dams.
In the VMN, offspring of obese DIO dams fostered with DR dams had elevated expression of LEPR-B (60%), INSR (41%), and MC3R (58%) (Table 7). The DMN melanocortin-4 receptor (MC4R) mRNA expression was elevated by 97%, and LH NPY5R was increased by 50%, while PVN MC3R mRNA expression was reduced by 95% compared with offspring fostered with obese DIO dams. Offspring of obese DIO dams fostered with DR dams had an overall trend toward reduced ARC NPY and AgRP expression compared with those raised with either obese or lean DIO dams. However, these values failed to reach statistical significance.
The current studies demonstrate that postnatal factors can, to a certain degree, override both genetic predisposition and prenatal factors in determining whether an individual will become obese and insulin resistant. They can also influence the development of hypothalamic circuits involved in energy homeostasis. Although offspring of obese DIO dams fostered to DR dams maintained their obesity, they ate less and had somewhat improved insulin sensitivity compared with those fostered with obese DIO dams. Hence, the intrinsic propensity of offspring of obese DIO dams to develop insulin resistance (33) can be ameliorated by raising them in a lean postnatal environment. Their changes in insulin sensitivity were associated with significant upregulation of VMN LEPR-B mRNA, whereas DIO rats raised with their own dams have reduced LEPR-B expression compared with DR offspring (35). In addition, VMN INSR and MC3R and DMN MC4R mRNA were upregulated in these cross-fostered rats. Interestingly, cross-fostering had no effect on any parameter in the ARC of DIO rats.
As shown previously (36), DR offspring maintain a leaner and more insulin-sensitive phenotype than offspring of DIO dams, regardless of whether their own dams are lean or obese during gestation and lactation. However, when fostered to obese DIO dams during only lactation, DR offspring here became obese and insulin resistant when fed a HE diet. This was associated with a selective increase in ARC AgRP, whereas expression of VMN INSR, LEPR-B, and MC3-R mRNA was decreased. Unexpectedly, DR offspring fostered with lean DIO dams ate less and were leaner after 4 wk on a HE diet than those fostered with either DR or obese DIO dams. This was associated with a reduction in ARC AgRP expression. It is unlikely these differences in ARC AgRP expression were an effect of postweaning dietary exposure as AgRP expression is either unchanged or decreased by the development of obesity and should be either unchanged or increased in the presence of somewhat lower leptin levels and adiposity (3, 65, 70). Because AgRP increases food intake and obesity (27, 41), the altered ARC AgRP expression might have been a contributing cause rather than an effect of the obesity in DR offspring fostered to obese DIO dams and the leaner phenotype in DR offspring fostered to lean DIO dams.
Cross-fostering offspring of obese DIO dams to DR dams had no effect on their adiposity or ARC AgRP expression. However, their slightly improved insulin sensitivity and the reduced insulin sensitivity of DR offspring fostered with obese DIO dams was associated with alterations in VMN leptin, insulin, and melanocortin signaling, suggesting that these findings might be related. Conflicting evidence points to a role for VMN leptin and insulin signaling in the regulation of peripheral insulin sensitivity. Although some studies suggest that the ARC might be the major site for this function (10, 46), it is likely that leptin and insulin signaling in both the ARC and VMN (collectively called the mediobasal or ventromedial hypothalamus) might be important for regulation of insulin sensitivity because of the intimate anatomical and physiological relationship between these two nuclei (60). Reduced LEPR-B expression in the ARC and VMN is associated with a predisposition to develop obesity and insulin resistance on high-fat diets (35), whereas third ventricular administration of a melanocortin agonist markedly improves insulin sensitivity (47), and injections of various agents directed at the ventromedial hypothalamus have marked effects on the ability of animals to respond to lowered glucose levels (4, 5). In addition, both the ARC and VMN contain specialized metabolic-sensing neurons involved in monitoring and regulating peripheral signals related to energy homeostasis, including insulin, leptin, and glucose (25, 66). Of course, changes in mRNA expression of neuropeptides and receptors do not necessarily ensure parallel alterations in their function, and the changes we observed were in adult rats fed chow from weaning for 7 wk and then HE diet for 4 wk. Thus some of these altered levels might be an effect rather than the proximate cause of the changes in body weight and insulin sensitivity observed with some of the cross-fostering paradigms.
Because cross-fostering affected the development of obesity, insulin resistance, and hypothalamic signaling, we postulated that maternal milk content might be a critical determinant of these changes. Milk insulin levels were a good candidate, as insulin can cross the intestinal mucosa-blood barrier neonatally (18), and pups can absorb and assimilate large proteins like insulin from the maternal milk during the early postnatal period (24). Furthermore, insulin has neurotrophic properties in the developing brain (17, 61, 68). Hypothalamic pathways mediating energy homeostasis continue their development well into the second postnatal week (7, 19, 53) and insulin injected into the hypothalamus during this period produces obesity associated with reduced neuronal size in the hypothalamic VMN (50). Also, fostering nondiabetic offspring to diabetic dams produces smaller offspring in association with altered ARC NPY, AgRP, and POMC expression (15). Breast milk from diabetic mothers has increased levels of insulin and glucose and both enter the milk from maternal circulation (24). Thus differences in milk insulin levels between obese DIO and lean DR dams might be an important cause of the observed differences in hypothalamic neuropeptide and receptor expression in cross-fostered pups. However, milk insulin levels of lean DIO dams were midway between those of DR and obese DIO dams, yet DR offspring fostered to lean DIO dams were leaner, whereas offspring of obese DIO dams fostered to lean DIO dams did not differ significantly from those fostered with obese DIO dams in either adiposity or insulin sensitivity. Thus there must be another explanation for these findings. Although leptin also has neurotrophic properties, which can affect neuronal development and survival (2, 8, 49), there were no significant differences in milk leptin levels among the groups that would support a critical role for milk leptin in the cross-fostering effects we observed.
Milk lipid fatty acid composition is a strong candidate for affecting postnatal offspring development because it can be affected by maternal diet (28, 29) and because fatty acids play such a critical role in postnatal brain development (45, 57, 63). In fact, MUFA and PUFA levels were selectively lower in the milk of obese DIO dams compared with those from lean DIO and DR dams. The essential fatty acids linoleic acid and α-linolenic acid are precursors to long-chain PUFAs, arachidonic acid (20:4n-6), and docosahexaenoic acid (22:6n-3), which accumulate mostly in the central nervous system and are necessary for normal perinatal brain development and functioning (20). Adequate supply of arachidonic acid and docosahexaenoic acid plays a role in several aspects of neural function (58), and failure to meet such nutrient needs during the critical period of brain development has significant cognitive consequences and can affect the development of such diseases as obesity, diabetes, hypertension, and coronary heart disease later in life (64). Both diet and genetic propensity to develop obesity can alter the fatty acid composition of neuronal membranes (37), which can then affect receptor function (13, 56, 59). Importantly, a deficit in long-chain PUFAs is associated with defects in the expression or function of insulin receptors in peripheral tissues, and this might also occur in the brain (11, 12) where defective insulin signaling can lead to obesity and insulin resistance (9). Thus it is conceivable that early exposure to low levels of milk PUFAs in obese DIO dams might be one cause of the differences in hypothalamic neuropeptide and insulin and leptin receptor expression in cross-fostered pups. However, differences in milk fatty acid composition, insulin, or leptin levels cannot explain the reduced food intake and adiposity associated with markedly lowered ARC AgRP mRNA expression in DR offspring fostered with lean DIO dams. Thus other factors such as maternal behavior, total fatty acid content, or the amount of milk produced in response to suckling DR neonates might have played a role.
There are three important issues that require further comment. First, there were several instances in which there were large apparent differences in the mean values for RT-PCR results among groups for a given probe, yet these differences did not reach statistical significance. This was most often due to large variances in the data. These variances are likely due to combined effects of biological variation combined with variances in the micropunch dissections and the RT-PCR procedure. It is clear that the micropunch technique is quite different from in situ hybridization, which often, but not always, provides less variable results. We have long experience with both techniques (31, 39) and have found the results to provide comparable data, each of which provides its own specific type of information. Punches are made through the entire 300- to 400-μm thickness of the slab, so that it includes the entire nucleus or area in question but also includes some unwanted tissue, and there is always some inherent variability in the punches, which are made with respect to gross landmarks, such as the third ventricle, median eminence, and optic tract. The RT-PCR assay is quite precise as the individual sample values are compared with standard curves for each mRNA species constructed from pooled samples. Also, all values are expressed as a function of cyclophilin, a constitutive gene, which we have found not to vary as a function of the conditions used here. Thus we are confident that those differences that did reach statistical significance do represent potentially important clues to the effect of the cross-fostering procedures used. On the other hand, those values that did not reach statistical significance might have done so had larger numbers of animals been used.
A second issue relates to weight gain in the dams. Even though the DIO dams fed a HE diet did not gain significantly more weight than those fed chow before impregnation, their twofold higher plasma leptin levels and greater weight gain and fourfold higher leptin levels seen by the second week of gestation make it highly likely that these dams were fatter during gestation than DR and lean DIO dams. Although not the sole index of adiposity, plasma leptin levels serve as a highly reliable index of carcass adiposity in our selectively bred DIO rats (52). Finally, the DIO dams fed a HE diet actually lost weight and showed significant decreases in their plasma leptin and insulin levels from the second week of gestation to the second week of lactation, whereas there was little change in these parameters in DR and lean DIO dams. It is unclear why this occurred selectively in the obese DIO dams. However, these findings are in keeping with our previous studies (36). Despite this apparent metabolic improvement, DR offspring fostered with obese DIO dams during this period still had a significant deterioration in their metabolic profiles, suggesting that other factors, most likely the insulin and/or fatty acid content of the milk were the critical determinants of this deterioration.
In conclusion, alterations in the postnatal environment can override some of the influences of prenatal and genetic factors in determining the evolution of obesity, insulin resistance, and the development of hypothalamic pathways that mediate energy and glucose homeostasis in offspring. Even though it did not affect the development of obesity, offspring of obese DIO dams had improved insulin sensitivity when raised postnatally by genetically lean dams. On the other hand, even when raised in a lean prenatal environment, genetically lean offspring became obese and insulin resistant when raised in an obese postnatal environment. This strongly implicates the postnatal maternal environment as a major effector of metabolic outcome in the offspring. Although maternal behavior can have important effects (43, 44, 67), metabolic factors are likely to play a more important role in the development of obesity and insulin resistance. Because hypothalamic pathways involved in both energy and glucose homeostasis are still developing during the period of lactation in the rat (7, 19, 54), it is likely that insulin and fatty acid content of maternal milk plays a critical role in their development. Because we did not evaluate these hypothalamic pathways until adulthood when the offspring had been fed both low- and moderate-fat diets for several weeks, it is not certain whether most of the observed effects were the cause or effect of these postweaning manipulations. However, the increased ARC AgRP expression in DR rats raised with obese DIO dams and decreased expression in lean DR offspring raised by lean DIO dams are likely to have played a causal role in the development of their obesity, as their levels of expression were exactly the opposite of what would have been expected in rats with their respective terminal phenotypes. Clearly, future studies must focus on the early postweaning period to assess these parameters. Nevertheless, our studies demonstrate the value of using a polygenic genetic model of obesity and insulin resistance to assess the critical role of factors in the perinatal environment that influence both peripheral metabolic and central neural development.
This work was supported by the Research Service of the Department of Veterans Affairs and the National Institute of Diabetes, Digestive and Kidney Diseases (DK-30066). Educational support for Judith Gorski was provided by Merck.
We thank A. Moralishvili, C. Salter, O. Gordon, and L. Petrie for their expert technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society