Several studies have shown that maternal undernutrition leading to low birth weight predisposes offspring to the development of metabolic pathologies such as obesity. Using a model of prenatal maternal 70% food restriction diet (FR30) in rat, we evaluated whether postweaning high-fat (HF) diet would amplify the phenotype observed under standard diet. We investigated biological parameters as well as gene expression profile focusing on white adipose tissues (WAT) of adult offspring. FR30 procedure does not worsen the metabolic syndrome features induced by HF diet. However, FR30HF rats displayed catch-up growth to match the body weight of adult control HF animals, suggesting an increase of adiposity while showing hyperleptinemia and a blunted increase of corticosterone. Using quantitative RT-PCR array, we demonstrated that FR30HF rats exhibited leptin and Ob-Rb as well as many peptide precursor and receptor gene expression variations in WAT. We also showed that the expression of genes involved in adipogenesis was modified in FR30HF animals in a depot-specific manner. We observed an opposite variation of STAT3 phosphorylation levels, suggesting that leptin sensitivity is modified in WAT adult FR30 offspring. We demonstrated that 11β-HSD1, 11β-HSD2, GR, and MR genes are coexpressed in WAT and that FR30 procedure modifies gene expression levels, especially under HF diet. In particular, level variation of 11β-HSD2, whose protein expression was detected by Western blotting, may represent a novel mechanism that may affect WAT glucocorticoid sensitivity. Data suggest that maternal undernutrition differently programs the adult offspring WAT gene expression profile that may predispose for altered fat deposition.
- maternal undernutrition
Increasing evidence suggests that the origin of some metabolic disorders, which manifest in adult life, may have their roots during development. Indeed, epidemiological studies have shown that adverse environmental factors leading to intrauterine growth retardation (IUGR) and low birth weight may predispose individuals to the later onset of energy balance metabolic development of pathologies (11, 22). This led to the concept of developmental origin of health and diseases (DOHaD), also called “fetal programming” (2, 54). Thus, adult individuals whose mothers were undernourished early in pregnancy displayed higher rates of intra-abdominal adiposity associated with increased risk of metabolic pathologies (9, 50). Inappropriate early postnatal nutrition, and more specifically rapid catch-up growth, may significantly heighten energy balance dysfunctions in adulthood. In particular, postnatal hypercaloric nutrition is an important accelerator in the etiology of adult-onset disease in human born with low birth weight (29).
To understand the underlying mechanisms, numerous animal models have been developed to promote intrauterine fetal programming, including maternal undernutrition (56). These studies confirmed that impaired fetal development has common long-term metabolic consequences sensitizing the offspring to metabolic syndrome features, particularly when nourished with a hypercaloric diet (19, 52, 57). In rodents, numerous data have demonstrated that fetal or neonate hypothalamus is a main target of perinatal modifications of energy status (8, 12, 17, 25, 48, 51). In particular, leptin was found to act as a neurotropic factor promoting neuronal outgrowth during postnatal brain development and is thus highly involved in the plasticity and hardwiring the hypothalamic appetite-regulatory circuits (7, 15, 25).
Adipogenesis occurs primarily during late fetal and early postnatal life. Perturbations to the fetal nutrient supply affect adipocyte development, leading to persistent alterations in their functional properties (42). In humans, IUGR affects the relative abundance and deposition of adipose tissue in a depot-specific manner. Indeed, the differences in fat and lean mass proportions observed are thought to play a crucial part in defining a propensity for accumulating body fat later in life. An important concept is that IUGR adipose tissue appears dysregulated before the onset of obesity (9, 30).
The mature adipocyte is not just a place for storing energy as fat but is also an endocrine cell, secreting cytokines and hormones associated with the control of energy balance (60). Indeed, many appetite-regulating related peptides or receptors are expressed in human and rodent white adipose tissue (WAT), acting in an autocrine/paracrine manner to regulate adipocyte lipid metabolism and as endocrine signals to regulate energy homeostasis (14, 27, 62, 63). In particular, leptin, via its long form receptor (Ob-Rb), was found to prevent adipocyte hypertrophy and hyperplasia and body fat increment induced by high-fat (HF) diet by increasing fatty acid oxidation and decreasing lipogenic pathways (28, 31, 55, 61).
Several studies have investigated the consequences of perinatal nutritional manipulations on adult rat offspring WAT, but only under standard diet. Maternal protein restriction did not affect the capacity of preadipocytes to divide or store fat in fetuses, neonates, and weanling offspring (3). However, growth-restricted adult male rat offspring exhibited increased rates of preadipocyte proliferation, whereas it did not affect the preadipocyte differentiation (65). Maternal protein restriction modified WAT fat-synthesizing enzymes' gene expression in adult male rat offspring (4, 24, 26). In addition, rat offspring adipocytes displayed an age-dependent loss of insulin sensitivity in a depot-specific manner associated with modification of the expression of key components of the insulin pathway (46). Maternal 50% food restriction from 10 days to term gestation leads to adult intrauterine growth-restricted male rats that exhibited programmed upregulation of adipogenic transcription factor (18). Uteroplacental insufficiency-induced IUGR increases visceral adiposity and proadipogenic factor in male rat offspring (32). We (16) recently demonstrated that maternal perinatal undernutrition delays the maturation of gonadal WAT. We also showed that numerous genes involved in energy metabolism regulation are expressed in rat neonate gonadal WAT and are sensitive to maternal perinatal undernutrition.
Using a model of prenatal maternal 70% food restriction diet (FR30) in pregnant female rats throughout gestation, we have previously shown that the FR30 procedure induces IUGR and programs some metabolic syndrome features in adult male rat offspring (52). Although showing a lean phenotype, adult FR30 rats were predisposed to adiposity and showed subtle food intake and hypothalamic appetite programming system alterations (8). They also exhibited higher concentrations of serum leptin and corticosterone, two hormones actively involved in hypothalamus-adipose tissue axis regulation (1, 43) and body fat accumulation (38, 64). We decided to assess whether a postweaning HF diet would amplify the phenotype. For that purpose, we investigated body weight gain, adiposity, endocrine and metabolic parameters, and gene expression profile focusing on the adipose tissue programmed mechanisms. Our data demonstrate that adult male rat offspring subjected to maternal prenatal undernutrition (MPU) exhibited a long-term impact on the gene expression in adipose tissue that might predispose them to increased adiposity when challenged with a HF diet.
MATERIALS AND METHODS
As previously described (8), Wistar rats were purchased from Charles River Laboratories (L'Arbresle, France) and housed six per cage. After mating with a male, pregnant females were transferred into individual cages with free access to water and to standard rat chow (SAFE 04, 2,900 cal/g, containing 16% protein, 3% fat, 60% carbohydrates; UAR, Augy, France). Control pregnant dams were fed ad libitum, while pregnant dams from the feed-restricted group fed 30% (FR30) of the daily intake of control pregnant dams, from day 1 (E1) of pregnancy until delivery (E21). At parturition, litter size was adjusted to eight pups per dam. Feed-restricted pups were nursed by FR30 dams fed ad libitum during lactation. To obviate any litter effects, animals used for each experiment were randomly chosen in different litters, and only a limited number of animals (n = 1 or 2) were used from each litter. After weaning, male offspring from the control (C) or FR30 dams were housed individually and divided into four groups (CN, CHF, FR30N, FR30HF; n = 16 per group) to be fed either standard (N) or high-fat (HF) (SAFE, D12451, 4,720 cal/g, containing 23% protein, 23% fat, 40% carbohydrates; UAR, Augy, France). Body weight and food intake of the offspring were measured weekly until adulthood. All parameters of adult male offspring were studied at 4 mo of age. Animal use accreditation by the French Ministry of Agriculture (No. 04860) has been granted to our laboratory for experimentation with rats. Experiments were conducted in accordance with the principles of laboratory animal care (European Communities Council Directive of 1986, 86/609/EEC).
Plasma and tissue collections.
At 4 mo of age, male rats were rapidly weighed and killed by decapitation between 9 and 10 AM. Trunk blood samples were collected into prechilled tubes containing EDTA (20 μl of a 5% solution), gently shaken, and centrifuged at 4,000 g for 10 min at 4°C. Aliquots of the supernatants were stored at −20°C until assayed. Hypothalami as well as deposit pads of WAT and interscapular brown adipose tissue (BAT) were rapidly removed, weighed, frozen in liquid nitrogen, and stored at −80°C until use. For histology experiments, animals were fixed by intracardiac perfusion using buffered 4% paraformaldehyde solution.
Food intake and metabolic parameters.
Food consumption was recorded weekly from weaning [postnatal day (PND)22] to adulthood (PDN120) in the four groups. Food intake of rats was measured once a day at the beginning of the light phase (9 AM). All animals were presented with the same amount of food, and their intake was measured by subtracting the uneaten food.
For oral glucose tolerance tests (OGTT), rats were fasted overnight (16 h). Basal blood glucose level, defined as T0, was determined using an automatic glucometer (Glucotrend 2; Roche Diagnostics, France) before the glucose administration (2 g/kg body wt; n = 12 per group). Tail vein blood glucose was then measured at 0, 15, 30, 60, and 120 min after administration. Plasma insulin concentrations were measured by enzyme-linked immunosorbent assay (DRG International).
All plasma endocrine parameters levels were investigated with commercially available kits. Blood glucose and plasma insulin levels were determined as described above. Plasma leptin and adiponectin concentrations were measured with murine enzyme-linked immunosorbent assay kits (Diagnostic Systems Laboratories; Adipogen, Korea), respectively. Plasma corticosterone levels were determined by a competitive enzyme immunoassay (Immunodiagnostic Systems, Boldon, UK). Assay kits were applied to determine the contents of plasma triglycerides and total cholesterol (61238 Triglyceride Enzymatique PAP100, 61218 Cholesterol Liquide; BioMérieux, France) as well as free cholesterol and free fatty acid (FFA) (ref. nos. 279-47106 and 999-75406; Wako Chemicals, Neuss, Germany). Each point was measured in duplicate. The assay sensitivities were 0.07 ng/ml (insulin), 0.04 ng/ml (leptin), 0.1 ng/ml (adiponectin), and 0.55 ng/ml (corticosterone), and the intra-and interassay coefficients of variation were 4 and 9.1% (insulin), 5.4 and 7.3% (leptin), 4.4 and 6.1% (adiponectin), and 4.9 and 7.8% (corticosterone), respectively.
Gene expression analysis.
Hypothalamic and gonadal WAT gene expressions were determined in the four groups using RT-qPCR as validated previously (16). RT-qPCR assays were performed using the Rat Obesity RT2 Profiler PCR Array (SuperArray Bioscience, Frederick, MD, http://www.superarray.com) on a Roche Lightcycler 380. PCR arrays allowed us to simultaneously investigate gene expression variation of 84 genes known to be implicated in obesity (http://www.superarray.com/pcr/arrayanalysis.php). Briefly, RNA was extracted and purified from hypothalami using the TRIzol reagent (Invitrogen, Life Technologies, France) and gonadal WAT animals using RNeasy lipid tissue minikit (Qiagen, Courtaboeuf, France) according to the manufacturers' recommendations. The quality of total RNA was assessed by determining the 260/280 and the 260/230 absorbance ratios and by agarose gel electrophoresis. An equal amount of total RNA (in μg) isolated from a particular tissue was pooled for each treatment group (n = 5 animals/group). First-strand cDNAs were generated using an RT2 First Strand Kit (Cat. no. C-03, SuperArray). For RT-qPCR, first-strand cDNAs were added to the RT qPCR Master Mix (SuperArray Bioscience). Samples were heated for 10 min at 95°C and then subjected to 40 cycles of denaturation at 95°C for 15 s and annealing and elongation at 60°C for 1 min. Relative expression values between animal groups were determined by the following rule: for each sample, we calculated difference between the CT (ΔCT) values for the gene of interest and housekeeping genes (HPRT, LDHA, actin B and RPL13a, respectively). Finally, the fold change of interrogated gene expression in each tissue was calculated as fold change = 2−ΔΔCT. The results are expressed as scatter plots, and only genes whose mRNA levels changed more than twofold in either direction (up- and downregulation) were selected.
RT-qPCR on individual samples (n = 5) was performed to confirm the results obtained using the PCR array approach. These samples were different from those that were used for PCR arrays. In addition, RT-qPCR was also used to investigate the expression profile of representative genes in two fat pads (gonadal and perirenal WAT). Primer sequences are presented in Table 1. Relative expression levels of RNA per sample were quantified by SYBR Green I assay on a Roche Light Cycler 480 sequence detection assay (Meylan, France). For each transcript, PCR was performed in duplicate with 10 μl of final reaction volumes with 1 μl of cDNA, 8 μl of mix, and 0.5 μl of each primer set (Table 1). PCR was conducted using the following cycle parameters: 10 min at 95°, and 40 three-step cycles of 15 s at 95°C, 20 s at 60°C, and 30 s at 72°C. The assay was performed following the manufacturer's recommendations. A pool of cDNA from control tissues was used as a standard (in threefold serial dilutions) for quantitative correction. All cDNA samples were applied in dilution of 1:10 to obtain results within the range of the standard. Each sample was evaluated in duplicate. Analysis of transcript level was carried out using first the determination of the threshold cycle CT for each reaction corrected by the efficiency. Then the ΔCT was calculated by subtracting the mean CT of the calibrator from each value of CT for each gene. The amount of target relative to a calibrator was computed by 2−ΔΔCT.
Adipose tissue histology.
Gonadal WAT from the four groups (n = 6 per group) was postfixed for 24 h in 4% paraformaldehyde in phosphate-buffered saline (PBS), cryoprotected by incubation for 24 h in 0.05 M PBS containing 20% sucrose, and frozen in liquid nitrogen. Gonadal WAT was then cut into serial 10-μm sections on a cryostat, mounted on gelatine-coated slides, and stained with hematoxylin of Groat and phloxin (2%), according to standard laboratory protocols. Sections were examined using light microscopy (Leica DM IRE2), and photomicrographs were captured at ×20 magnification.
Western blot analysis.
Frozen perirenal or gonadal WAT from the four groups (n = 5 per group) was homogenized in lysis buffer: 250 mM sucrose, 10 mM HEPES-Tris, 1 mM EDTA, 1% Triton X-100, protease inhibitor cocktail 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF), pepstatin A, E-64, bestatin, leupeptin, and aprotinin, and phosphatase inhibitor cocktail (10 mM sodium fluoride, 1 mM sodium orthovanadate, 20 mM sodium β-glycerophosphate, and 10 mM benzamidine). After lysis in ice, insoluble materials were removed by centrifugation (10,000 rpm at 4°C for 10 min and then 15,000 rpm at 4°C for 10 min), and protein concentrations of the resulting lysates were determined using a protein assay kit (Bio-Rad, France). Proteins were diluted in Laemmli buffer, denatured, and subjected to 8% SDS-PAGE gels and transferred onto nitrocellulose membranes. Blots were blocked with 5% albumin bovine serum (Sigma, France) and then incubated in the presence of appropriate primary antibodies (anti-phosphorylated STAT3 or anti-total STAT3; Ozyme, Saint Quentin en Yvelines, France, respectively) or with a sheep anti-rat 11β-HSD2 polyclonal antibody (Chemicon International, Temecula, CA) and secondary antibodies. A mouse anti-rat α-tubulin antibody (Sigma Aldrich, Saint Quentin Fallavier, France) was used as a loading control. Following nitrocellulose membrane washing, targeted proteins (∼85 kDa for STAT3, 40 kDa for 11β-HSD2, and 50 kDa for α-tubulin) were revealed using enhanced chemiluminescence reagents (Amersham Life Science, Les Ulis, France) according to the manufacturer's recommendations. The intensity of bands was quantified by using Quantity One Bio-Rad and the p-STAT3/t-STAT3 ratios were calculated.
All data are presented as means ± SE. Statistical analysis was performed by analysis of variance (two-way ANOVA with prenatal undernutrition and diet as factors) followed by Neumman Keuls post hoc analysis. A P level of <0.05 was considered statistically significant. Analyses were performed using SigmaStat software (Systat Software, Port Richmond, CA).
Effects of HF on body composition.
The body weights of male newborn pups differed between the C and FR30 groups by ∼30% (6.29 ± 0.11 vs. 4.45 ± 0.13 g, P < 0.001). From birth until 4 mo, FR30 rats' body weight remained lower (Table 2). HF diet during postnatal life resulted in significantly increased body weights compared with standard diet-fed animals in both groups (C, 529 ± 54 vs. 471 ± 0.31 g; FR30, 518 ± 39 vs. 451 ± 40 g, respectively; P < 0.001; Table 2). By P71, FR30 rats fed hypercalorically showed apparent catch-up growth to match the body weight of C animals fed the HF diet (Fig. 1). In line with this, at P64 and P71, FR30HF rats were hyperphagic, exhibiting a transient, significantly increased weight-related caloric intake calculated as energy in food ingested (cal/g of animal) compared with C rats. However, this increased weight-related caloric consumption was no longer apparent in 4-mo-old FR30 rats. Indeed, in both FR30 and C groups, adult males fed the HF diet exhibited similar weight-related increased daily energy intake leading to fat increment (data not shown).
The absolute weight of the liver showed no difference between groups. As expected, both CHF and FR30HF groups had increased perirenal (C, 24.94 ± 10.11 vs. 9.97 ± 3.44 g; FR30, 22.47 ± 4.62 vs. 10.95 ± 3.98 g, respectively; P < 0.001), epididymal (C, 21.19 ± 6.63 vs. 10.66 ± 4.23 g; FR30, 20.06 ± 3.53 vs. 11.55 ± 4.11 g, respectively; P < 0.001), and subcutaneous fat deposits (C, 0.85 ± 0.41 vs. 0.59 ± 0.17 g; FR30, 1.09 ± 0.59 vs. 0.59 ± 0.32 g, respectively; P < 0.001) compared with standard diet (Table 2). In this study, we observed no significant difference of absolute fat pad weights between C and FR30 animals in the same dietary group (standard or HF diet). However, as reported earlier (8), when expressed relative to body weight, FR30 animals exhibited a tendency toward greater body fat content in both regimens compared with C rats.
Effects of HF on plasma parameters and glucose tolerance.
As shown in Table 2, plasma triglyceride contents were not significantly different between C and FR30 adult offspring rats and were not affected by diet. Plasma total cholesterol contents significantly increased, whereas FFA levels decreased after HF diet in both groups. However, at 4 mo of age, whatever the feeding conditions (standard or HF diet), no significant difference was noted in plasma lipid parameters investigated between FR30 and C rats. Control and FR30 rats had comparable standard diet-fed as well as increased HF diet-fed plasma glucose and insulin concentrations (Table 2). As already described (7), during OGTT, partial glucose intolerance was observed in FR30 animals compared with C rats. However, although the HF diet induced marked impaired glucose tolerance, no difference was noted in glucose levels between HF diet-fed FR30 and C after an oral glucose load (data not shown).
FR30 rats had about twofold higher basal (0.71 ± 0.13 vs. 0.33 ± 0.04 ng/ml, P < 0.01) leptin concentration compared with C animals. In both groups, CHF and FR30HF serum leptin concentration was ∼2.5-fold higher than in standard diet-fed animals (C, 0.915 ± 0.12 vs. 0.33 ± 0.04 ng/ml; FR30, 1.86 ± 0.18 vs. 0.71 ± 0.13 ng/ml, respectively; P < 0.001; Table 2). Plasma adiponectin level was increased in FR30HF compared with CHF (17.74 ± 0.89 vs. 15.09 ± 0.6 ng/ml, P < 0.05), whereas no difference was observed between the two standard diet-fed groups. Under standard diet, FR30 rats had ∼2.5-fold higher corticosterone concentrations than C (10.8 ± 2.33 vs. 4.31 ± 0.99 μg/dl, P < 0.05). Under HF diet, C animals' serum corticosterone concentration was ∼3.5-fold higher than in standard diet (14.39 ± 2.62 vs. 4.305 ± 0.99 μg/d, P < 0.01). By contrast, FR30 rats showed no significant increase in corticosteronemia after HF diet compared with standard diet-fed FR30 animals.
Gene expression profile in adipose tissue.
Surprisingly, FR30 MPU did not significantly affect the hypothalamic mRNA levels of genes present in PCR arrays. Very few gene expression changes were observed in FR30 vs. C adult rat hypothalamus whatever the regimen used (data not shown). This is consistent with the absence of weight-related caloric intake difference observed between HF fed C and FR30 groups. By contrast, MPU led to marked gene expression variation in gonadal WAT, more especially under HF diet (Fig. 2, A and B, and Tables 3 and 4). For more clarity, we decided to present gene expression changes only as FR30/C ratio [FR30N vs. CN (Table 3) and FR30HF vs. CHF (Table 4)] and cut-off as −2 (downregulated genes) and >2 (overexpressed genes).
For example, RT-qPCR data showed that expression of leptin mRNA was higher (3.07-fold) in the FR30N group than in the CN group (Table 3), consistent with the increased levels of serum leptin observed in FR30N rats (Table 2). Leptin gene expression was elevated nearly twofold in CHF compared with CN rats, whereas FR30HF rats showed no additional gene expression increase compared with CHF animals. On the other hand, leptin mRNA expression was lower in the FR30HF group than in the CHF group (−1.31-fold), which contrasted with the pattern found in serum leptin levels (Table 2). Compared with CN group, FR30N rats showed a significant decrease of the Ob-Rb mRNA levels (−1.53-fold). However, FR30HF rats displayed a significant increase of Ob-Rb mRNA content vs. CHF rats (+2.03-fold; Table 4). RT-qPCR experiments showed that several genes encoding neuropeptides and their receptors were also expressed in rat gonadal WAT (Fig. 2, A and B). Many of these genes are modulated by maternal FR30 procedure (Table 3) and/or following HF diet in the adult offspring rats (Table 4). BDNF, GH secretagogue receptor (GHS-R), gastrin-releasing peptide, melanocortin receptor 3 (MCR3), and opioid receptor k1 mRNA levels were highly downregulated (10.47-, 4.82-, 28.57-, 12.17-, and 10.97-fold, respectively) in FR30HF vs. CHF groups, whereas other genes, such as calcitonin receptor, gastrin-releasing peptide receptor and PYY, showed a markedly increased expression (8.34-, 15.15-, and 10.06-fold, respectively; Table 4).
To validate the RT-qPCR experiments, we determined the leptin and Ob-Rb mRNA variation levels in gonadal WAT on individual samples. The intensity of fold changes was on the same order of magnitude for both genes, confirming the data from the RT-qPCR experiments (Fig. 3A). In perirenal WAT, the general profile of leptin gene expression was similar to the one determined in gonadal WAT. By contrast, both CN and FR30N rats showed an increased Ob-Rb mRNA levels (about +2.5-fold) compared with CHF and FR30HF rats (Fig. 4A).
Then, we focused on the expression of representative genes in two fat pads (gonadal and perirenal WAT) involved in WAT adipogenesis, proliferation and lipid metabolism (Figs. 3B and 4B, respectively). In gonadal WAT, we found that MPU increased mRNAs for gene involved in de novo lipogenesis, such as acetyl-CoA carboxylase (ACC, +2.76-fold) and fatty acid synthase (FAS, +2.83-fold), whereas no difference was observed in gene expression levels for genes involved in adipogenesis such as CAAT enhancer binding protein-α (C/EBPα), peroxisome proliferator-activated receptor-γ (PPARγ), and preadipocyte factor 1 (PREF-1) in FR30N offspring. However, FAS, SREBP-1c/ADD1 (sterol regulatory element-binding protein-1c/adipocyte determination differentiation-dependent factor 1), C/EBPα, and PPARγ were significantly decreased in CHF rats vs. CN animals (−3.10-, −2.38-, −2.88-, −4.74-fold, respectively), whereas gene expression levels remained elevated in FR30HF animals.
In perirenal WAT, ACC, FAS, and SREBP-1c/ADD1 mRNA levels are increased in FR30N rats vs. CN animals (+2.49-, +2.29-, and +1.76-fold, respectively), whereas no difference was observed in CEBPα and PPARγ gene expression levels between both groups. However, unlike the gonadal WAT, SREBP-1c/ADD1, C/EBPα, and PPARγ exhibited no gene expression difference between FR30HF and CHF rats. In gonadal and perirenal WAT, PREF-1 mRNA levels were reduced about twofold by HF diet in both groups. No variation was observed for adiponectin gene expression levels.
Among the genes involved in glucocorticoid (GC) sensitivity, we found that 11β-HSD1, 11β-HSD2, GR, and MR genes were coexpressed in both fat deposits. As indicated by the CT values, the expression levels of the four genes showed differences within gonadal WAT of CN rats (11β-HSD1, 26.44 ± 0.56; 11β-HSD2, 30.99 ± 0.39; GR, 27.51 ± 0.75; MR, 29.49 ± 0.6). These data suggest that the 11β-HSD1 gene expression levels were higher than the 11β-HSD2 expression levels in adult rat WAT. In gonadal WAT, MR mRNA levels were reduced (−1.23-fold) in FR30N vs. CN rats. MR gene expression was also decreased in CHF vs. CN rats (−1.75-fold), whereas it remained elevated in FR30HF compared with FR30N group. 11β-HSD2 gene expression was increased in FR30HF vs. FR30N and CHF rats (+1.65-, and +1.46-fold, respectively). In perirenal WAT, MR mRNA levels were reduced ∼1.5-fold by HF diet in both groups. 11β-HSD1 mRNA levels were decreased (−1.46-fold) in FR30HF compared with FR30N group.
Effects of HF on histology of adipose tissue.
To determine the effects of HF diet on adipocytes, we performed a histological morphometric analysis of gonadal WAT. As shown in representative photographs, the size of gonadal WAT adipocytes appeared similar between CN and FR30N rats (Fig. 5, A and B). However, FR30N animals exhibited a tendency for greater total area compared with CN rats (Fig. 5E). HF diet led to hypertrophied adipocytes to a similar extent in both groups (Fig. 5, C and D) (T, 91,076 ± 3,838 vs. 47,550 ± 2,035 μm2; FR30, 92,819 ± 3,660 vs. 59,063 ± 1,860 μm2, respectively; P < 0.01; Fig. 5E).
Effects of HF on STAT3 phosphorylation in adipose tissue.
Considering the different levels of circulating leptin among the four CN, CHF, FR30N, and FR30HF groups (Table 2), we compared WAT leptin sensitivity by measuring STAT3 phosphorylation levels on WAT protein extracts. In each group, STAT3 phosphorylation levels were normalized to total STAT3 and calculated as a ratio. In both gonadal and perirenal fat pads, FR30N rats displayed decreased levels of STAT3 phosphorylation compared with CN rats (−2.65- and −1.96-fold, respectively; Fig. 6, A and B and C and D). HF diet led to a tendency for decreased STAT3 phosphorylation levels in both gonadal and perirenal fat pads of C animals (−1.41- and −1.82-fold, respectively), whereas significantly increased levels were observed in FR30 rats (+1.62- and +2.11-fold, respectively; Fig. 6, A and B and C and D).
Presence of 11β-HSD2 protein in adipose tissue.
As shown in Fig. 7, Western blot analysis using specific antibody raised against rat 11β-HSD2 clearly detected a single band in the expected molecular mass range of ∼40 kDa in the protein extracts of gonadal (50 μg, lane 3) and perirenal WAT (50 μg, lane 4) and kidney (1 μg, lane 1) in adult C rats. The largest amount was consistently detected in the kidney (lane 1), whereas no signal was detected in the brain (50 μg, lane 2). Thus, we demonstrated that both fat deposits expressed a detectable level of 11β-HSD2 protein. In addition, we observed a marked difference in the level of 11β-HSD2 protein between the two different fat depots.
The major findings of this study are that, although MPU does not worsen the metabolic syndrome features induced by HF feeding, the FR30 procedure is associated with altered WAT transcriptional profile and may affect WAT leptin and GC sensitivities leading to long-term depot-specific programming effects in male rat offspring.
Epidemiological data showed that nutritional conditions improve subsequent to low weight at birth, such as rich milk formula lactation, or in infancy, such as hypercaloric nutrition, increases the susceptibility for central obesity associated with increased risk of metabolic diseases (9, 29, 30, 50). Animal studies also confirmed these observations. In particular, feeding IUGR offspring a hypercaloric diet amplifies prenatal influences on obesity, hypertension, and hyperinsulinemia (4, 57). However, the underlying programming tissue mechanisms of adipose development remain elusive, especially under postweaning HF diet.
In the present study, we observed that postweaning HF diet modified total fat increment as well as lipid profile to a similar extent in both groups. Surprisingly, despite the partial glucose intolerance observed in FR30 rats under basal conditions (8), no obvious additional effects were noted under HF diet. As suggested (19), it may reflect a relatively mild programming phenotype. The modest elevation of plasma adiponectin, a key regulator of insulin sensitivity, may contribute to the protective effect against glucose intolerance (49).
Our findings disagree with those of Vickers et al. (57), who reported that hypercaloric nutrition dramatically amplified the FR30 model endocrine disorders observed under standard diet. This discrepancy may be due to the fact that pups from undernourished mothers were cross-fostered onto dams that had received ad libitum feeding. This finding emphasizes the importance of the suckling period and catch-up growth during early lactation of undernourished pups for developmental programming of future diseases. The differing outcome of IUGR offspring might also lie in the timing of sensitive developmental windows, the nature of the deleterious stimulus experienced, the genetic background of the species, or the regimen used (4, 13, 29, 49). In line, administration of the maternal prenatal synthetic glucocorticoid dexamethasone in rats results in low-birth-weight offspring that subsequently develop glucose intolerance and hypertension in adulthood. However, this IUGR procedure does not confer increased sensitivity to HF diet-induced obesity (19).
However, we observed that FR30HF rats displayed catch-up growth to match the body weight of adult CHF animals while showing hyperleptinemia and a blunted increase of corticosterone. The elevated serum leptin concentration suggests an increase of adiposity in FR30HF rats, although we cannot rule out that hyperleptinemia may originate from other tissues (47) and/or that our model of IUGR alters adipose tissue distribution in adulthood (32).These observations prompted us to focus our further investigations on WAT adult FR30 offspring.
Using quantitative RT-PCR array containing 84 obesity-related genes, we demonstrated that, unlike hypothalamus, gonadal WAT is a very sensitive target of MPU programming, especially under HF diet. This reinforces the idea of the existence of a brain sparing effect under suboptimal nutritional conditions (16, 37). In particular, RT-qPCR experiments identified the leptin and Ob-Rb genes but also many mRNAs encoding peptide precursors, receptors, and proteins involved in energy homeostasis regulation as particularly good candidates. Some of them could act in an autocrine/paracrine manner to regulate adipocyte metabolism (16, 60, 63). Indeed, it has been shown that ghrelin induces abdominal obesity via GHS-R-dependent lipid retention (14) whereas α-MSH interacts with WAT MCR3 to induce lipolysis (27). Both PYY and NPY promoted proliferation of adipocyte precursor cells and exhibited antilipolytic properties in visceral WAT (62). However, since their level of expression is weak, further analyses will be required to assess the physiological significance, if any, of these variations.
In addition to showing a tendency for greater adipocyte area, we demonstrated that the expression of genes involved in adipogenesis was modified in FR30 animals in depot-specific and diet-specific manners. Such alterations favoring fat storage might be advantageous to survival under poor nutrition conditions and are in accordance with the thrifty phenotype hypothesis (18, 19, 26, 32). In line, uteroplacental insufficiency-induced IUGR increases visceral adiposity and visceral adipose PPARγ expression in male rat offspring prior to the onset of overt obesity (46). These finding are consistent with the concept that in humans IUGR adipose tissue appears dysregulated before the onset of obesity (9, 30). By contrast, a global transcriptomic approach indicated that 55-day-old offspring of undernourished rat dams during gestation displayed no difference in WAT gene expression, pointing out sensitive developmental windows programming differences (41).
We demonstrated that the overall lipogenic mRNA transcript profiles revealed a marked tendency for increase in FR30 rats under HF diet. Controversial data about the regulation of lipogenic key factors have been published in DIO-induced rat obesity models (34, 36) and in obese human subjects (45) regarding the time-dependent effects of a high-fat-yielding diet. Indeed, it may be that during the period of “dynamic obesity,” when fat stores are rapidly expanding, an increase in the lipogenic capacity of adipose tissue is expected, but, in obese subjects with a large and long-lasting fat excess, the decreased expression of lipogenic genes could be a late and adaptive process aimed, in fact, at preventing a further development of fat mass (45). Thus, the different HF feeding outcome in WAT FR30 lipogenic gene expression rats suggests that MPU alters the timing of fat deposition at least at the level of gene expression that may favor long-term fat storage.
In agreement with RT-PCR array results, indicating the WAT leptin system as a key target of MPU programming, we confirmed that leptin and OB-Rb mRNA expressions were modified in FR30 offspring in depot-specific and diet-specific manners. For the latter, this is in agreement with studies showing depot-specific gene expression levels in adipose tissue (20). We hypothesized that these might, at least in part, account for modified WAT leptin sensitivity and may thus affect its antilipogenic action (28, 31, 55, 61).
Hyperleptinemia may be interpreted as a leptin-resistant state (57). First, whatever diet we used, we found no gross expression difference of hypothalamic appetite-regulatory factor gene expression as well as food intake between C and FR30 groups (data not shown). FR30 rats might have developed an adaptive resetting neuronal activation threshold that could explain normalized orexigenic and anorexigenic gene expression levels (8, 51). Second, despite hyperleptinemia, FR30N rats displayed decreased levels of STAT3 phosphorylation compared with CN rats in WAT. In accordance with that, considering the antilipogenic leptin action on WAT (28, 31, 55, 61), we did not observe any suppression of ACC and FAS gene expression in FR30N animals, but a marked increased expression. Finally, under HF diet, and despite a similar twofold rise of plasma leptin concentration in both groups, we observed an opposite variation of WAT STAT3 phosphorylation levels, suggesting that leptin sensitivity is modified in WAT adult FR30 offspring.
Finally, we demonstrated, for the first time, that 11β-HSD1, 11β-HSD2, GR, and MR genes are coexpressed in rat WAT and that mRNA transcript profiles revealed depot-specific and diet-specific changes. In particular, we showed that MPU modifies the expression of genes involved in GC/mineralocorticoid tissue sensitivity by increasing 11β-HSD2 and MR mRNA levels in gonadal WAT FR30HF offspring. We also detected, for the first time, the 11β-HSD2 protein in WAT, whose expression levels differ between the depots.
Several lines of evidence prompted us to explore whether WAT GC sensitivity is modified in adult FR30 offspring. First, we had previously shown that maternal undernutrition programs the offspring HPA axis throughout the lifespan (35, 53). Second, we and others (6, 10, 21, 23, 44) demonstrated that IUGR offspring exhibit long-term and tissue-specific effects on gene expression of factors involved in GC sensitivity. In particular, plasma corticosterone and aldosterone levels were increased in adult male FR30 rats (8, 52) that may directly modulate WAT adipogenesis (38, 64). Third, several studies have shown that modified WAT GC sensitivity leads to altered fat accumulation. Increased expression of GR and 11β-HSD1 in visceral adipose tissue has been associated with the development of obesity in humans (56, 59) and in rats overfed in early postnatal stages (5, 6). In mice, overexpression of 11β-HSD1 in adipose tissue resulted in visceral obesity (39) when fed under HF diet, whereas adipose tissue overexpression of 11β-HSD2 protects mice from HF diet-induced obesity (33).
Our findings support the hypothesis that MPU might affect the ratio between 11β-HSD1 and 11β-HSD2 expression in WAT adult FR30 offspring, thus modifying local tissue ratios between active and inactive GC. In response to HF diet, the depot-specific upregulation of 11β-HSD2 mRNA while 11β-HSD1 expression remains stable might limit cortisosterone concentration within its local environment, thus diminishing responsiveness to GC. This may represent an adaptive mechanism that may counteract excess fat storage (40). However, additional experiments such as 11β-HSD1 and 11β-HSD2 protein levels and/or enzymatic activities as well as GC binding capacity assessments should be performed to determine the extent to which MPU may affect WAT GC sensitivity in adult FR30 offspring.
In contrast with 11β-HSD1, the cellular localization of 11β-HSD2 within WAT remains unclear. In rat, 11β-HSD2 gene was expressed by both adipocytes and stromal-vascular cells (40) whereas in humans 11β-HSD2 mRNAs were detected predominantly in the stromal fraction (56). Alternatively, or in addition, 11β-HSD2 expression in WAT stromal-vascular cells fraction might protect nonmature adipocytes from active GC spillover (40, 56). Further immunohistochemical experiments are needed to precisely determine the WAT cell type expressing the 11β-HSD2, especially in adult FR30 offspring. These studies could contribute to a better understanding of its local and/or paracrine role as well as mechanism of action within the WAT microenvironment.
In conclusion, we have demonstrated that MPU differently programs the long-term offspring adipose tissue gene expression profile in a depot-specific manner that may predispose prenatally undernourished individuals to altered lipid metabolism and fat accumulation, specifically within an obesogenic environment.
This study was supported by grants from the French Ministry of Education and grants of the Conseil Régional du Nord-Pas de Calais.
No conflicts of interest, financial or otherwise, are declared by the author(s).
- Copyright © 2011 the American Physiological Society