Maternal and postnatal high-fat diet consumption programs energy balance and hypothalamic melanocortin signaling in nonhuman primate offspring

Abstract

Maternal high-fat-diet (HFD) consumption during pregnancy decreased fetal body weight and impacted development of hypothalamic melanocortin neural circuitry in nonhuman primate offspring. We investigated whether these impairments during gestation persisted in juvenile offspring and examined the interaction between maternal and early postnatal HFD consumption. Adult dams consumed either a control diet (CTR; 15% calories from fat) or a high-saturated-fat diet (HFD; 37% calories from fat) during pregnancy. Offspring were weaned onto a CTR or HFD at ~8 mo of age. Offspring from HFD-fed dams displayed early catch-up growth and elevated body weight at 6 and 13 mo of age. Maternal and postnatal HFD exposure reduced the amount of agouti-related peptide fibers in the paraventricular nucleus of the hypothalamus. Postnatal HFD consumption also decreased the amount of agouti-related peptide fibers in the arcuate nucleus of the hypothalamus. Postnatal HFD was associated with decreased food intake and increased activity. These results support and extend our previous findings of maternal diet effects on fetal development and reveal, for the first time in a nonhuman primate model, that maternal HFD-induced disturbances in offspring body weight regulation extended past gestation into the juvenile period. Maternal HFD consumption increases the risk for offspring developing obesity, with the developmental timing of HFD exposure differentially impacting the melanocortin system and energy balance regulation. The present findings provide translational insight into human clinical populations, suggesting that profound health consequences may await individuals later in life following intrauterine and postnatal HFD exposure.

over the past 30 yr, childhood obesity has emerged in developed countries as a public health concern with no signs of abating (11, 78). Unaddressed, the disease can impact individuals into adulthood with repercussions manifesting in reduced quality of life and increased financial costs (12, 16, 35). Mounting evidence indicates that childhood obesity might be programmed in utero via prenatal insults (2, 3, 36, 43, 55, 61, 79). If this is the case, then each subsequent generation would be at greater risk than the last of developing obesity. It is well-documented that children exposed to maternal obesity during gestation have an increased risk of developing obesity (2, 3, 36, 43, 55, 61, 79), fatty liver disease, cardiovascular disease, diabetes (41, 45), and metabolic syndrome in general (34, 48). Considering that as many as 40% of women of reproductive age are obese (17), the consequences of maternal obesity on offspring metabolic health and behavior need to be immediately and thoroughly investigated.

One of the limiting factors of the human studies is the inability to elucidate the potential effects of maternal metabolic phenotype versus maternal diet on the developing offspring. A few clinical studies have examined the impact of maternal high-fat-diet (HFD) consumption on childhood obesity (1, 53). In these studies, associations have been observed among maternal HFD consumption during pregnancy, excess fat deposition in newborns, and elevated body mass in children. In animal models, maternal obesity is commonly induced by providing animals access to a palatable HFD. Rodent studies provide robust and direct evidence that HFD consumption during the intrauterine and early postnatal periods of development program offspring obesity (7, 8, 57, 63, 69, 74). To counter the threat of childhood obesity, the underlying biological mechanisms of these metabolic changes need to be clarified. For example, it remains unclear whether the propensity toward obesity in offspring from HFD-fed mothers is due to overconsumption, diminished activity levels, decreased energy expenditure, or some combination thereof.

The hypothalamic melanocortin neural circuit is essential for the regulation of energy balance. This circuit comprises two antagonistic neuropeptides, α-melanocyte-stimulating hormone (α-MSH) and agouti-related peptide (AgRP). α-MSH, a cleavage product of the preprohormone proopiomelanocortin (POMC) gene, inhibits food intake through activation of the melanocortin receptor subtype 4 (20, 52). In contrast, AgRP potentiates food intake through blockade of the same melanocortin receptor (20, 56, 70). AgRP is colocalized with neuropeptide Y neurons (6). POMC- and AgRP-expressing nuclei are important components of the melanocortin circuit and originate in the arcuate nucleus of the hypothalamus (ARH; Refs. 10, 14, 66). Fiber projections are found in hypothalamic regions such as the ARH and the paraventricular nucleus of the hypothalamus (PVH). Moreover, although the function of the melanocortin system is evolutionarily conserved (26), a species difference does exist in its ontogeny. The melanocortin system matures as late as the 3rd postnatal week in rodents (4, 24, 27) but as early as the 3rd trimester of pregnancy in nonhuman primates (NHPs) and humans (21, 44). An earlier developmental window suggests this critical neurocircuitry is very susceptible to regulation by the in utero environment in NHPs and humans.

Using a NHP diet-induced obesity model, our laboratory has established that chronic consumption of a HFD during pregnancy programs a variety of metabolic complications that facilitate the development of obesity in offspring. These metabolic issues include lipotoxicity, inflammation, nonalcoholic fatty liver disease (49), and reduced pancreatic cell plasticity (9), which can predispose toward diabetes later in life. Furthermore, our group has discovered that fetal offspring from HFD-fed dams have reduced body weight (49) and abnormal development of the melanocortin neural circuit critical for energy balance (22). These offspring displayed a decrease in the expression of AgRP mRNA and an increase in the expression of POMC and the melanocortin receptor subtype 4 mRNA in the mediobasal hypothalamus. AgRP fiber projections to the PVH were also diminished in these fetal offspring (22).

The goals of our present study were threefold. First, we determined whether maternal HFD programming effects persist in juvenile offspring. Second, we examined the impact of early postnatal HFD consumption in offspring from HFD- and control diet-fed dams. In particular, we sought to evaluate the impact of maternal and postnatal HFD on energy balance regulation by measuring body weight, feeding behavior, physical activity, and energy expenditure. Third, we quantified melanocortin fibers (with α-MSH and AgRP) in the ARH and PVH to assess whether fetal abnormalities in melanocortin signaling persist past birth into the juvenile period. Such findings will provide critical insight into which factors may predispose an individual to obesity and how to alleviate these disadvantages.

MATERIALS AND METHODS

Animals, Maternal/Postnatal Diet, and Body Weight

All animal procedures were approved by the Oregon National Primate Research Center (ONPRC) Institutional Animal Care and Use Committee. Adult dams (Macaca fuscata) were assigned to the study with an average age of 5.69 ± 0.14 yr, ranging between 3.5 and 12.0 yr. The average weight at the time of conception was 8.38 ± 0.14 kg with a range between 6.0 and 12.0 kg. Dams were either maintained on a control diet (CTR; 15% calories from fat primarily from soybeans and corn; Monkey Diet no. 5052; Purina Mills, St. Louis, MO) or placed on a HFD (37% calories from fat primarily from animal fat, egg, and corn oil; TAD Primate Diet no. 5LOP; Purina Mills) provided ad libitum for 2–9 yr before offspring delivery and throughout lactation. Detailed dietary information has previously been described (9, 21). Obesity and insulin resistance did not necessarily follow dietary groups, with a subset of dams remaining lean on the HFD group and subset of CTR dams becoming obese; however, for analysis, dams were solely separated into their diet groups. Female and male offspring were born naturally and remained on dam’s diets until weaning with an average age of 8.36 ± 0.10 mo of age. Before weaning, animals were housed in groups of 4–12 individuals (male-to-female ratio of 1–2:3–10) in indoor-outdoor enclosures. At weaning, offspring were group-housed in enriched indoor-outdoor environments with 6–10 similarly aged juveniles and 1–2 unrelated adult females per group. Offspring from both maternal diet groups were assigned to a postnatal diet of either CTR or HFD, yielding 4 offspring groups (total number of offspring used per group: CTR/CTR, n = 33; CTR/HFD, n = 18; HFD/CTR, n = 30; HFD/HFD, n = 22). Sample sizes vary for individual measurements and are reported in the figure legends. Body weight was measured at 1, 3, 6, 10, and 13 mo of age. Cumulative weight gain (in percentage change) was calculated across development. A schematic of the experimental protocol is depicted in Fig. 1.

Fig. 1.

Schematic of experimental design used for juvenile NHP offspring summarizes the schedule used to conduct experiments across development. Major developmental time points in NHP are depicted in black boxes located at top. The physiological measurements associated with each developmental time point are represented in white boxes situated at bottom.

Food Intake

Food intake was measured at 10 mo of age using an automated food dispenser system (Med Associates, St. Albans, VT) in group-housed offspring. The testing period occurred over a course of 5 days. Animals were identified by AVID microchips previously implanted in their wrist (Avid Identification Systems, Norco, CA). The number of lever pulls and the offspring associated with each pull were recorded. This method has been previously validated in group-housed NHPs (80). Offspring were allowed to acclimate to the food intake system for 5 days and were presented with a CTR (5052, Primate Tablet no. 1814687, 5AP7; TestDiet, Richmond, IN) or HFD (TAD 5L0P, Primate Tablet no. 1814863, 5ASN; TestDiet) provided ad libitum. Adult females were separated from the juvenile offspring from 0930 until 1630 to prevent them from interfering with the food intake system. At 1630, offspring were reunited with the adult females, and they remained housed together until the next morning at 0930. Videotaping occurred daily during the testing period. One camera was focused on the food intake system to assess lever-pressing behavior, and the other allowed viewing of the entire run. For each animal, 3–5 days of data were analyzed.

Physical Activity

Physical activity was measured continuously after weaning allowing measures of activity in group-housed animals and during metabolic chamber testing. We measured physical activity as previously described (71, 73) using Actical accelerometers (Mini Mitter, Bend, OR) mounted on loose-fitting plastic collars (Primate Products, Miami, FL). Actical monitors contain an omnidirectional sensor capable of detecting acceleration from all directions. The sensor integrates speed and distance of acceleration and produces an electrical current that varies depending on the change in acceleration. The monitor stores the total number of activity counts per minute. To retrieve collars for data collection, monkeys were sedated with ketamine HCl (3–8 mg/kg im; Ketathesia; Henry Schein Animal Health, Melville, NY). Data from monitors were downloaded at least every 45 days.

Energy Expenditure

Metabolic rate and respiratory exchange ratio were measured using indirect calorimetry at 6 and 13 mo of age. Offspring were placed in a stainless steel metabolic chamber (dimensions: 2.7 × 2.7 × 2.5 ft) with a computer-controlled indirect open-circuit calorimeter (Oxymax; Columbus Instruments, Columbus, OH) to measure the amount of carbon dioxide produced and oxygen consumed using previously published methodology (71, 73). At 6 mo, each monkey was tested in the metabolic chamber for 4 h (0900–1300). At 13 mo, each monkey was placed in the chambers the night before the test for an acclimation period of 15 h, and then metabolic measurements were collected for 48 h. Monkeys were fed their meals in morning and afternoon (0900 and 1400). Water was available ad libitum throughout experimental testing. To prevent social isolation during testing, a familiar monkey was housed across from the monkey undergoing metabolic testing. Basal metabolic rate was determined by obtaining the average number of kilocalories expended from 2300 to 0300, a time period when monkeys were asleep and activity reached its nadir (71, 73). Total, basal, and nonresting (calculated by subtracting basal from total) metabolic rates were calculated and normalized to body weight. Respiratory exchange ratio (RER) was used to determine metabolic substrate preference.

Tissue Collection and Processing

Animals were necropsied at 13 mo of age, and brain tissue was collected as previously described (21, 62, 72). Offspring were deeply anesthetized with a surgical dose of sodium pentobarbital (30 mg/kg iv) and then exsanguinated. Perfusion of the brain occurs via the carotid artery by flushing with 0.9% heparinized saline (0.5–1 l) followed by 4% paraformaldehyde (PF; approximately 1–2 l) buffered with sodium phosphate (NaPO₄; pH 7.4) until fixed. The brain is then partitioned into specific areas and placed in 4% PF for 24 h at 4°C and transferred to 10% glycerol buffered with NaPO₄ for 24 h and finally to 20% glycerol solution for 72 h. Tissue blocks were frozen in −50°C 2-methylbutane and then stored in −80°C until sectioning. Coronal tissue sections (25-μm thick) were stored in cryoprotectant. An average of five equally spaced, anatomically matched sections spanning the hypothalamus were processed for immunohistochemistry.

Immunohistochemistry for AgRP and α-MSH

A subset of six animals per treatment group was processed for immunohistochemistry consisting of three males and three females per group, except for the HFD/HFD group, which contained four females and two males. Coronal tissue sections were washed in 0.05 M potassium phosphate-buffered saline (KPBS) and incubated in blocking buffer (2% normal donkey serum + 0.4% Triton X-100 + KPBS) for 30 min at room temperature. Tissue sections were then incubated in a cocktail of primary antibodies containing guinea pig anti-AgRP (1:5,000; cat. no. GPAAGRP.1, lot no. AS506; Antibodies Australia, Melbourne, Australia) and rabbit anti-α-MSH (1:15,000; cat. no. H-043-01, lot no. 01207-4; Phoenix Pharmaceuticals, Burlingame, CA) diluted in blocking buffer for 24 h at room temperature. Specificity of AgRP and α-MSH antibodies have been previously validated in the NHP hypothalamus (21). After incubation, tissues were washed in KPBS and incubated for 1 h in donkey anti-guinea pig/Alexa 488 to visualize AgRP (1:1,000 dilution; cat. no. 706-545-148; Jackson ImmunoResearch, West Grove, PA) and donkey anti-rabbit/Alexa 568 to visualize α-MSH (1:1,000 dilution; cat. no. A10042; Thermo Fisher Scientific, Waltham, MA).

Confocal Imaging of AgRP and α-MSH

A Leica SP5 acousto-optical beam splitter (AOBS) confocal microscope (Leica Microsystems, Buffalo Grove, IL) and a ×40 oil-immersion objective (HCX PL APO; 1.25 numerical aperture) were used to capture images, as described previously (25). We imaged 6 fields of view per section throughout the paraventricular nucleus of the hypothalamus (PVH; 3 fields of view on each side of the 3rd ventricle) and 2 fields of view per section in the middle of the arcuate nucleus of the hypothalamus (ARH; 1 field of view on each side of the 3rd ventricle). The observer was blind to offspring groups when imaging slides. Each image stack was captured at a format of 512 × 512, 1-μm increment, zoom factor 1.0, and 400 Hz. The 488-nm line of an Ar laser and a 561-nm diode-pumped solid-state (DPSS) laser were used sequentially to avoid bleedthrough of individual fluorophores into the nearby detection channels. Analysis was performed by individuals blind to treatment group using ImageJ software (65) to determine fluorescent density of AgRP and α-MSH immunoreactivity in each ×40 image.

Data Analysis

Statistical analyses were performed in SPSS Version 22 (SPSS, Chicago, IL), and graphs were created with GraphPad Prism v6.0 (GraphPad, San Diego, CA). For all analyses, normality and homogeneity of variance were initially tested. For measures that were normally distributed and had equal variance between groups, parametric statistics were used. Measures made before weaning (1-mo body wt, 6-mo body wt, metabolic rate, RER, and cumulative weight gain) were analyzed with two-factor (maternal diet × sex) ANOVAs. Measures made after weaning (offspring cumulative weight gain, food intake, activity, total metabolic rate, nonresting metabolic rate, AgRP fiber density in the PVH, and α-MSH fiber density in the ARH and PVH) were analyzed with three-factor (maternal diet × postnatal diet × sex) ANOVAs. Metabolic rate was assessed in 4-h time intervals and analyzed using a three-way repeated-measures ANOVA with maternal diet, postnatal diet, and sex being between-subject factors and time bin being a within-subject factor. The variance of 13-mo body wt, RER, and AgRP expression in the ARH was not equally distributed between groups, thus Mann-Whitney U tests were used to examine maternal diet, postweaning diet, and sex effects. Data are presented as means ± SE. α-Values were considered significant with P < 0.05.

RESULTS

Impact of Maternal and Postnatal HFD Consumption on Energy Balance in Juvenile Offspring

Body weight at 1, 6, and 13 mo.

Here, we extended our observations past the fetal period to assess the individual impact or interaction of maternal and postnatal HFD consumption in juvenile offspring (Fig. 1). Body weight was compared at three distinct time points during development. At 1 mo of age, male and female maternal HFD offspring were of similar weight relative to their control counterparts (Fig. 2A; maternal diet: F_1,26 = 2.91, P = 0.10; sex: F_1,26 = 0.18, P = 0.68). At 6 mo of age, before weaning, maternal HFD increased offspring body weight (Fig. 2B; F_1,26 = 10.43, P < 0.01), and male offspring weighed more than female offspring regardless of maternal diet (Fig. 2B; F_1,26 = 6.83, P = 0.02). At 13 mo, maternal HFD continued to contribute to an increased offspring body weight (Fig. 2C; Mann-Whitney, P = 0.02), and male offspring remained heavier than females (Fig. 2C; Mann-Whitney, P = 0.01). Postnatal diet did not contribute to differences in body weight at 13 mo of age (Fig. 2C; Mann-Whitney, P = 0.17).

Fig. 2.

Maternal HFD consumption increased offspring body weight at 6 and 13 mo and weight gain across development. A: at 1 mo of age, maternal HFD offspring were of similar weight relative to their control counterparts (F_1,26 = 2.91, P = 0.10), and there was no difference in body weight between male and female offspring (F_1,26 = 0.18, P = 0.68). B: at 6 mo of age, before weaning, maternal HFD increased offspring body weight (F_1,26 = 10.43, P < 0.01), and male offspring weighed more than female offspring (F_1,26 = 6.83, P = 0.02). Sample size for A and B: maternal HFD offspring, n = 13; controls, n = 17. C: at 13 mo, maternal HFD continued to contribute to an increased offspring body weight (Mann-Whitney, P = 0.02), and male offspring remained heavier than females (Mann-Whitney, P = 0.01). CTR/CTR, n = 26; CTR/HFD, n = 14; HFD/CTR, n = 23; HFD/HFD, n = 20. D: before weaning, at 6 mo of age, male and female offspring from HFD-fed dams gained more weight than controls, indicative of rapid catch-up growth (maternal diet: F_1,31 = 4.50, P = 0.04). Maternal HFD offspring, n = 27; controls, n = 21. E: by 13 mo, weight gain was not different between the groups. CTR/CTR, n = 11; CTR/HFD, n = 10; HFD/CTR, n = 14; HFD/HFD, n = 13. Data are shown as means ± SE; *maternal diet effect; °sex effect.

Cumulative weight gain across development.

Maternal HFD consumption increased offspring weight gain early in development (Fig. 2D; maternal diet: F_1,31 = 4.50, P = 0.04). Our group (49) has previously shown that maternal HFD fetal offspring weighed less than control offspring. Together, these results demonstrate that maternal HFD offspring transitioned from a state of reduced weight in the fetal period to elevated weight in the preweaning period. After weaning, weight gain was not affected by maternal diet (F_1,32 = 3.50, P = 0.07) or postnatal diet (F_1,32 = 0.56, P = 0.46). No effect of sex was observed in analysis of either preweaning (F_1,31 = 1.01, P = 0.32) or postnatal cumulative weight gain (F_1,32 = 0.45, P = 0.51).

Food intake at 10 mo.

When food intake was measured in group-housed animals at 10 mo of age, postnatal HFD was observed to decrease food intake (Fig. 3A; postnatal diet: F_1,36 = 7.98, P = 0.01) and food intake normalized to body weight (Fig. 3B; postnatal diet: F_1,36 = 5.11, P = 0.03). However, there was no main effect of maternal HFD or sex on food intake (maternal diet: F_1,36 = 0.02, P = 0.90; sex: F_1,36 = 0.30, P = 0.60) or food intake normalized to body weight (maternal diet: F_1,36 = 0.21, P = 0.65; sex: F_1,36 = 0.01, P = 0.93).

Fig. 3.

Postnatal HFD consumption decreases offspring food intake at 10 mo in group-housed offspring. Postnatal HFD decreased caloric food intake (postnatal diet: F_1,36 = 7.98, P = 0.01; A) and food intake normalized to body weight (postnatal diet: F_1,36 = 5.11, P = 0.03; B). There was no main effect of maternal HFD or sex on caloric food intake or food intake normalized to body weight. CTR/CTR, n = 11; CTR/HFD, n = 7; HFD/CTR, n = 18; HFD/HFD, n = 8. Data are shown as means ± SE; #postnatal diet effect.

Physical activity at 10–12 mo.

In group-housed offspring at 10–12 mo of age, offspring fed a postnatal HFD exhibited increased physical activity relative to controls (Fig. 4; postnatal diet: F_1,66 = 56.4, P < 0.01). Neither maternal HFD (F_1,66 = 0.82, P = 0.37) nor sex (F_1,66 = 1.72, P = 0.19) had a significant effect on physical activity.

Fig. 4.

Postnatal HFD consumption increased offspring physical activity at 10–12 mo. Offspring fed a postnatal HFD exhibited increased physical activity relative to controls (postnatal diet: F_1,66 = 56.4, P < 0.01). There was no main effect of maternal diet or sex. CTR/CTR, n = 20; CTR/HFD, n = 16; HFD/CTR, n = 19; HFD/HFD, n = 19. Data are shown as means ± SE; #postnatal diet effect.

Physical activity during metabolic testing at 13 mo.

No sex (F_1,33 = 0.83, P = 0.37), maternal HFD (F_1,33 = 2.46, P = 0.13), or postnatal HFD (F_1,33 = 0.70, P = 0.41) effects were observed on physical activity when offspring were housed in metabolic chambers at 13 mo of age (data not shown).

Energy expenditure at 6 and 13 mo.

Metabolic rate identifies the amount of energy expended (in kilocalories), whereas RER indicates the primary fuel source being used. An RER value of 1.0 indicates primarily carbohydrate oxidation, a value of 0.70 demonstrates fat oxidation as the main fuel, and an intermediate RER value occurs when a mixture of fat and carbohydrate is used. At 6 mo, neither maternal HFD exposure (F_1,26 = 0.03, P = 0.85) nor sex (F_1,26 = 3.87, P = 0.06) affected offspring RER (data not shown). Similarly, metabolic rate at 6 mo of age was not affected by maternal diet (F_1,26 = 0.06, P = 0.81) or sex (F_1,26 = 0.09, P = 0.77; data not shown). At 13 mo, there was no significant effect of maternal diet, postnatal diet, or sex on RER (maternal diet: Mann-Whitney, P = 0.58; postnatal diet: Mann-Whitney, P = 0.16, sex: Mann-Whitney, P = 0.59; data not shown) or total metabolic rate (Fig. 5, A and B; maternal diet: F_1,41 = 0.4, P = 0.84; postnatal diet: F_1,41 = 0.71, P = 0.41; sex: F_1,41 = 1.65, P = 0.21). Similarly, there was no effect of maternal diet, postnatal diet, or sex on nonresting metabolic rate (Fig. 5C; maternal diet: F_1,41 = 2.19, P = 0.15; postnatal diet: F_1,41 = 0.16, P = 0.21; sex: F_1,41 = 0.47, P = 0.50). Current data are presented as normalized to body mass; however, similar results were obtained when normalizing to lean mass (data not shown).

Fig. 5.

HFD consumption does not alter offspring energy expenditure at 13 mo. A: total metabolic rate averaged across 4-h time intervals during the light and dark (shaded) phase of a 24-h period. B: no significant main effect of maternal HFD, postnatal HFD, or sex were observed on total metabolic rate. C: no significant main effect of maternal HFD, postnatal HFD, or sex were observed on nonresting metabolic rate. CTR/CTR, n = 7; CTR/HFD, n = 11; HFD/CTR, n = 14; HFD/HFD, n = 17. Data are shown as means ± SE. TMR-BMR, total metabolic rate minus basal metabolic rate.

Maternal and Postnatal HFD Consumption and Central Melanocortin Signaling in Juvenile Offspring

Central melanocortin signaling at 13 mo.

To determine the impact of maternal and postnatal HFD consumption on offspring brain development, we examined AgRP and α-MSH fiber innervation in the hypothalamus. AgRP fibers were present in the PVH and ARH (Fig. 6, A and D). AgRP fiber area in the PVH was significantly decreased by maternal diet (F_1,16 = 7.48, P = 0.02) and postnatal diet (Fig. 6, B and C; F_1,16 = 11.45, P < 0.01). There was no effect of sex (F_1,16 = 0.43, P = 0.53). There was also a significant interaction between maternal and postnatal diet (F_1,16 = 10, P < 0.01) on AgRP fiber area in the PVH.

Fig. 6.

Maternal and/or postnatal HFD consumption decreased offspring AgRP fiber projections to the hypothalamus at 13 mo. AgRP fibers were present in the PVH (A) and ARH (D). PVH and ARH diagrams depict regions analyzed [adapted from Paxinos et al. (59) atlas with permission]. B: representative images of the PVH reveal a robust amount of AgRP-positive fibers in control offspring but a comparative reduction in all offspring exposed to a HFD. C: quantification of AgRP fiber area in the PVH revealed decreases by maternal diet (F_1,16 = 7.48, P = 0.02) and postnatal diet (F_1,16 = 11.45, P < 0.01). There was no effect of sex. E: representative images of the ARH reveal a robust amount of AgRP-positive fibers in control offspring but a comparative reduction in offspring consuming a postnatal HFD (CTR/HFD and HFD/HFD). F: quantification revealed that postnatal HFD decreased AgRP fiber density in the ARH (Mann-Whitney, P = 0.02). No effects of maternal HFD or sex were observed. CTR/CTR, CTR/HFD, HFD/CTR, and HFD/HFD, n = 6 per group. Scale bar = 100 μm. Data are shown as means ± SE; *maternal diet effect; #postnatal diet effect. v, Ventricle; f, fornix; ot, optic tract.

In the ARH, postnatal HFD decreased AgRP fiber density (Fig. 6, E and F; Mann-Whitney, P = 0.02). ARH AgRP fiber density was not affected by maternal diet (Mann-Whitney, P = 0.08) or sex (Mann-Whitney, P = 0.36).

α-MSH fiber innervation in the hypothalamus was also determined. Light labeling of α-MSH fibers PVH and ARH indicates a paucity of fibers in these two regions (Fig. 7, A and D). There were no maternal HFD (F_1,16 = 3.50, P = 0.08), postnatal HFD (F_1,16 = 1.70, P = 0.21), or sex (F_1,16 = 0.70, P = 0.42) effects on α-MSH fiber density in the PVH (Fig. 7, B and C). Similarly, α-MSH fibers were not affected by sex (F_1,16 = 1.43, P = 0.25), maternal HFD (F_1,16 = 0.15, P = 0.71), or postnatal HFD (F_1,16 = 1.75, P = 0.20) in the ARH (Fig. 7, E and F).

Fig. 7.

Maternal and postnatal HFD consumption did not alter α-MSH fiber projections to the hypothalamus at 13 mo. Light labeling of α-MSH fibers in the PVH (A) and ARH (D) indicates a paucity of fibers in these 2 regions. PVH and ARH diagrams depict regions determined [adapted from Paxinos et al. (59) atlas with permission]. Representative images of the PVH (B) and ARH (E) reveal α-MSH-positive fibers are similarly labeled across offspring groups. No maternal or postnatal HFD effects were observed on α-MSH fiber density in the PVH (C) or ARH (F). CTR/CTR, CTR/HFD, HFD/CTR, and HFD/HFD, n = 6 per group. Scale bar = 100 μm. Data are shown as means ± SE.

DISCUSSION

Epidemiological studies continually affirm associations between maternal obesity and childhood obesity (2, 3, 36, 43, 55, 61, 79). Even though maternal overconsumption is likely involved, a limited number of studies have directly examined the impact of maternal HFD consumption on childhood obesity (1, 53), primarily due to the ethical and logistical limitations of human research. Animal models provide the controlled setting necessary to properly examine the consequences of maternal HFD consumption. Previous work from our laboratory using a NHP diet-induced obesity model has demonstrated maternal diet effects on fetal development, but whether these changes persisted later in life was unknown. The current study reveals, for the first time in a NHP model, that maternal HFD-induced disturbances in offspring body weight regulation extended past gestation into the juvenile period. We also observed that offspring exposed to a maternal HFD displayed an increase in body weight at 13 mo of age. In addition, both maternal and postnatal HFD consumption reduced the quantity of AgRP fiber projections to the hypothalamus. As NHPs possess a similar capability to develop metabolic disease as humans (23, 40), the present findings provide translational insight into human clinical populations, suggesting that profound health consequences may await individuals later in life following intrauterine and postnatal HFD exposure.

The hypothalamic melanocortin neural circuit is universally crucial to the regulation of energy balance. Species differences are apparent in the ontogeny of the melanocortin circuit and overall brain development. Development of axonal projections from the melanocortin system occurs as late as the 3rd postnatal week in rodents but during the 3rd trimester of pregnancy in NHPs and humans (24, 27). Development of melanocortinergic axonal projections occurs earlier in primates than in rodents even after accounting for differences in brain development between the two species [axonal extension at 3 wk postnatal in a mouse is comparable to a monkey at term (81)]. In primates, an earlier period of development confers greater susceptibility to regulation by in utero environmental factors. In particular, AgRP fiber projections in the hypothalamus begin to develop during the 3rd trimester but do not become fully developed until adulthood (21). Our present findings demonstrate that AgRP fiber projections are strongly labeled during the juvenile period. α-MSH fiber projections have a similar ontogeny to AgRP fiber projections (21) but, as observed here, were sparse at 13 mo.

The influence of gestational and early postnatal HFD exposure on AgRP fiber projections in the hypothalamus has not been as well-studied in NHPs. Exposing offspring to a HFD during gestation produced abnormalities in hypothalamic melanocortin signaling. Using a NHP diet-induced obesity model, our laboratory (22) previously demonstrated that fetal maternal HFD offspring displayed a decrease in the quantity of AgRP fiber projections to the hypothalamus. Significant to long-term health, this loss of AgRP fiber projections was maintained in juvenile offspring. A loss of AgRP fibers was also evident with postnatal HFD exposure, suggesting a reduced complexity of AgRP fiber projections. A similar decline in AgRP fiber projections has been observed in rodent offspring either from dams fed a HFD or from diet-induced, obese-prone dams (5, 42, 77).

Obesity-related inflammation is a possible mechanism by which perinatal HFD reduces the quantity of AgRP fiber projections. The circulating levels of proinflammatory cytokines are elevated in NHP HFD offspring (49), leading to an increase in cytokines in the brain (e.g., in the hypothalamus; Ref. 22). Central inflammation has been shown to impair hypothalamic melanocortin signaling. For example, the proinflammatory cytokine interleukin, IL-1β, decreases AgRP release in hypothalamic explants in rodents (64). An inflammation-induced impairment in central melanocortin signaling likely underlies the increased risk for obesity in offspring.

Despite a potential decrease in orexigenic drive due to decreases in AgRP following HFD exposure, our findings demonstrate that maternal HFD consumption increased offspring body weight. At the 3rd trimester, maternal HFD offspring weigh less than control offspring (49). The present study demonstrated that by 1 mo of age maternal HFD offspring exhibited similar body weight but gained more weight and were heavier by 6 mo of age. A transition from an underweight state in the fetal period to being overweight in the preweaning period indicates the presence of rapid catch-up growth. Historically, catch-up growth in the first few months of development has been considered beneficial to survival in a famine environment as it functions to compensate for a growth deficit (58, 60, 76). Nonetheless, in the modern obesogenic environment, catch-up growth is no longer viewed as beneficial and rather is associated with increased risk of obesity, hypertension, type 2 diabetes mellitus, cardiovascular disease, and osteoporosis (13, 15, 28–30, 33).

Maternal HFD exposure resulted in continued increases in body weight at 13 mo of age. Although there was no overall effect of postnatal diet on body weight, males and females in the HFD/HFD group displayed the highest body weights. This finding could indicate that both maternal and postnatal HFD exposure are necessary to cause a significant elevation in offspring’s body weight. In addition, postnatal HFD consumption did not significantly alter the amount of weight gained between weaning and 13 mo of age. Considering that the increase in body weight and weight gain in HFD/HFD offspring was roughly equivalent to the sum of these two effects, maternal and postnatal HFD could act additively on offspring body weight. This same additive effect between maternal and postnatal HFD on offspring body weight has been previously observed in adult rodent offspring (32, 68). The ability of maternal and postnatal HFD to regulate offspring body weight independently would suggest the involvement of distinct mechanisms.

In the present study, postnatal HFD consumption produced hypophagia and increased physical activity in juvenile offspring. These metabolic changes are likely explained by the “setpoint” model (38) where peripheral hormones act on the hypothalamus to produce compensatory metabolic changes in food intake and activity to maintain body weight at a predefined level. However, offspring exposed to a maternal HFD that consumed a HFD postnatally still exhibited the highest body weight at 13 mo of age despite these physiological compensations. This suggests that maternal HFD consumption programs body weight regulation making offspring more susceptible to obesity later in life. Although differences in body weight were small across the groups, we hypothesize that this difference would likely increase with aging and result in a significantly worsened metabolic outcome in adulthood.

The mechanisms mediating body weight regulation are complex, and a more recent model postulates that body weight regulation revolves around a “settling point” that is influenced by environmental factors such as diet (31, 37, 47). Body weight is not predefined and rather will settle into a range in response to the environment. Findings from our present study suggest that early environmental factors, such as maternal HFD consumption, may also impact the settling point of body weight. It is important to note that the measurements of energy balance in this study were not collected at the same developmental time point; thus future studies that include more frequent physiological measurements will be useful in determining the order of metabolic changes relative to body weight changes. It is possible that other factors, such as parity, which was not controlled in the current study, may also impact offspring’s body weight regulation.

Increased body weight in maternal HFD offspring could be linked to impairments in energy expenditure. In the present study, we discovered that maternal HFD offspring did not display changes in energy expenditure at 6 and 13 mo of age. Our previous studies demonstrated that offspring from HFD mothers, many of which are included in the current data set, display behavioral abnormalities including increased anxiety (72). Anxiety in nonhuman primates can manifest as repetitive stereotypical behavior (75); therefore, it is possible that the absence of an effect of maternal diet on metabolic rate is obscured by anxiety-induced changes in activity in these animals.

Gestational nutrition is theorized to set expectations for the postbirth environment (39) and in the process may limit metabolic responsiveness to deviations in diet. Overall, metabolic health in these offspring was impaired differently depending on the timing of HFD exposure. For example, maternal HFD exposure increased body weight at 13 mo, whereas increased activity and decreased food intake were only observed following postnatal HFD exposure. It remains possible that maternal obesity per se, and not HFD, contributed to the observed changes in offspring. However, dams in the current study were not grouped by maternal metabolic state, and there was heterogeneity in body weight in both the CTR and HFD-fed dams. It remains unclear whether diet and obesity may represent independent risks to offspring metabolism. Factors such as increased nutrients (glucose and triglycerides), hormones (insulin and leptin), and inflammatory factors are possible mechanisms by which maternal obesity and HFD consumption program offspring energy balance regulation.

Sex differences were apparent in offspring body weight. One such difference existed in body weight for maternal HFD offspring at 6 mo, when male HFD offspring had increased weight compared with controls and females did not. The sex difference in body weight in maternal HFD offspring remained present at 13 mo of age. Findings from rodent studies provided evidence that male HFD offspring are more susceptible to metabolic changes than females (50, 63). The biological mechanisms mediating this sex bias have not been clearly defined. However, sex differences in development programming of maternal HFD may involve the hormonal milieu interacting with systems that regulate energy balance.

Sex differences are programmed by androgen production in the fetal testes during critical periods of development. The testes produce a surge of testosterone at birth in primates (18, 19) and in mice (51). Elevated levels of testosterone during this early developmental period masculinize the brain (82). Testosterone acts in the brain by regulating inflammation. In neonatal and adult rodents, sex differences are evident in the number of microglia (46) and in the mRNA expression of proinflammatory factors, such cytokines and chemokines (67). Testosterone also acts in the brain to program neural circuitry that controls energy balance (54). Therefore, we believe that sex differences in the response of the innate immune cells to maternal obesity-induced inflammation may underlie the observed sex differences in energy balance.

Perspectives and Significance

Our present findings reveal, for the first time in a NHP model, that maternal HFD-induced disturbances in offspring body weight regulation extended past gestation into the juvenile period. Maternal HFD consumption resulted in early catch-up growth so that offspring were heavier by 6 mo of age and remained heavier than control offspring at 13 mo of age. These deficits were associated with impairments in central melanocortin signaling. Early postnatal HFD consumption also impaired central melanocortin signaling and was associated with increased physical activity and reduced food intake. Taken together with clinical studies (1, 53), our work shows that maternal intake of saturated fat is associated with changes that increase risk of childhood obesity. It is crucial that future nutritional studies examine the types and sources of fat, carbohydrates, and proteins to determine the optimal dietary composition for women to consume during gestation and lactation. Furthermore, it is important to identify therapeutic interventions that are efficacious in preventing and ameliorating the impact of maternal obesity and HFD consumption on offspring development.

GRANTS

This work was supported by the American Diabetes Association Grant 7-13-MI-06 (to H. M. Rivera), National Institute of Diabetes and Digestive and Kidney Diseases Grants R01-DK-079194 (to P. Kievit and K. L. Grove) and R24-DK-090964 (to P. Kievit, K. L. Grove, J. E. Friedman, and K. L. Thornburg), M. J. Murdock Charitable Trust Grant 2011273:HVP (to E. L. Sullivan), National Institute of Mental Health Grant R01-MH-107508R01 (to E. L. Sullivan), Oregon Clinical and Translational Research Institute Grant UL1TR000128 (to E. L. Sullivan) from the National Center for Advancing Translational Sciences, and U.S. National Institutes of Health Office of the Director Grant P51-OD-011092 for the operation of the ONPRC, support of the Obese Resource, and support of the Imaging and Morphology Support Core.

DISCLOSURES

P. Kievit reports a grant from National Institutes of Health (NIH) during the conduct of the study and grants from Novo Nordisk, Janssen Research and Development, Rhythm Pharmaceuticals, ERX Pharmaceuticals Inc., Ember Therapeutics, Sanofi-Aventis Deutschland GmbH, and Leidos Biomedical Research outside of the submitted work. E. L. Sullivan reports grants from M. J. Murdock Charitable Trust, NIH, and Bill and Melinda Gates Foundation during the conduct of the study. K. L. Grove is a paid employee of Novo Nordisk.

AUTHOR CONTRIBUTIONS

E.L.S., K.L.G., and P.K., conceived and designed research; E.L.S., H.M.R., J.G.F., K.B., T.A.D., J.C.V., D.L.T., T.F., G.H., M.A.K., and L.A.B. performed experiments; E.L.S., H.M.R., C.A.T., J.G.F., T.A.D., J.C.V., D.L.T., T.F., M.A.K., L.A.B., and P.K. analyzed data; E.L.S., H.M.R., K.L.G., and P.K. interpreted results of experiments; E.L.S. and H.M.R. prepared figures; E.L.S. and H.M.R. drafted manuscript; E.L.S., H.M.R., C.A.T., and P.K. edited and revised manuscript; E.L.S., H.M.R., C.A.T., D.L.T., K.L.G., and P.K. approved final version of manuscript.