We previously reported that prenatal and suckling exposure to a maternal diet rich in animal fat leads to cardiovascular dysfunction in young adult rat offspring with subsequent development of dyslipidemia and hyperglycemia. We have further investigated glucose homeostasis in adult female offspring by euglycemic-hyperinsulinemic clamp and by dynamic assessment of glucose-stimulated insulin secretion in isolated, perifused pancreatic islet cells. Additionally, given the link between reduced mitochondrial DNA (mtDNA) content and the development of type 2 diabetes mellitus, we have measured mtDNA in organs from young adult animals. Sprague-Dawley rats were fed a diet rich in animal fat or normal chow throughout pregnancy and weaning. Infusion of insulin (5 mU·kg−1·min−1) resulted in a higher steady-state plasma insulin concentration in 1-year-old offspring of fat-fed dams (OHF, n = 4) vs. offspring of control dams (OC, n = 4, P < 0.01). Glucose-stimulated insulin secretion in isolated islets from 9-mo-old OHF was significantly reduced compared with OC (n = 4, P < 0.05). Transmission electron micrography showed altered insulin secretory granule morphology in OHF pancreatic β-cells. Kidney mtDNA was reduced in 3-mo-old OHF [16S-to-18S gene ratio: OC (n = 10) 1.05 ± 0.19 vs. OHF (n = 10) 0.66 ± 0.06, P < 0.05]. At 6 mo, gene chip microarray of OHF aorta showed reduced expression of the mitochondrial genome. Prenatal and suckling exposure to a diet rich in animal fat leads to whole body insulin resistance and pancreatic β-cell dysfunction in adulthood, which is preceded by reduced tissue mtDNA content and altered mitochondrial gene expression.
- dietary fats
- islet cells
- metabolic syndrome
metabolic syndrome has reached epidemic proportions among Western populations, with an associated increase in childhood type 2 diabetes (2, 4). Whereas the typical Western diet rich in saturated fat is a known risk factor for metabolic syndrome and cardiovascular disease in adults, converging lines of evidence now suggest that a fat-rich diet during pregnancy may induce features of the metabolic syndrome in the adult offspring independently of adult environmental factors (13, 24, 29). Most recently we described a rat model of developmental programming in which increased maternal dietary intake of animal fat in pregnancy and suckling induces cardiovascular and metabolic abnormalities in adult offspring with shared characteristics of metabolic syndrome in humans, including hypertension, dyslipidemia, and hyperglycemia (21, 22, 34).
In the present study we have attempted to further characterize the abnormalities of glucose homeostasis in adult offspring of dams fed a fat-rich diet (OHF). The two main metabolic predictors of type 2 diabetes mellitus in humans are insulin resistance and impaired insulin secretory capacity. We have assessed whole body insulin resistance in vivo and isolated pancreatic islet cell structure and function. Given the proposed involvement of mitochondrial abnormalities in the development of atherosclerosis and type 2 diabetes (3, 28, 33), we have examined aspects of mitochondrial integrity.
MATERIALS AND METHODS
Animal husbandry and experimental diets.
Animal procedures complied with United Kingdom Home Office regulations. Female Sprague-Dawley rats (n = 20, 120 days old) were fed, for 10 days before mating and throughout pregnancy and lactation, either a standard rat chow [RM3 5% fat (wt/wt)] or a diet rich in animal fat [RM3 supplemented 20% (wt/wt) with lard; Special Diet Services, Witham, Essex, UK; Ref. 22].
Birth weights and litter size were similar in control and OHF. All offspring were weaned onto the standard chow (RM3, postpartum day 21), received food and water ad libitum, and were maintained in a 12:12-h light-dark cycle (21°C). One female per litter was studied in each protocol (n = number of litters).
Whole body insulin sensitivity.
Insulin sensitivity was assessed in OHF and control offspring (OC) by euglycemic-hyperinsulinemic clamp (n = 4) and was performed in anesthetized animals at 1 year of age, as described previously (15). Porcine monocomponent insulin (5 mU·kg−1·min−1; Novo Industri, Bagsvaerd, Denmark) was infused in the saphenous vein at a constant rate (20 μl/min). A variable glucose infusion was administered through the same vein 1 min after the insulin infusion started. Blood glucose was measured every 5 min (2300 STAT glucose analyzer; YSI, Yellow Springs, OH), and the glucose infusion rate was adjusted to maintain euglycemia. Steady state was attained after ∼30 min and was maintained for an additional 30 min. Three 200-μl blood samples were collected at 5- to 10-min intervals for triplicate determination of steady-state blood glucose and insulin concentration by radioimmunoassay (RIA) (Boehringer Mannheim). Insulin sensitivity index (ISI) was calculated according to Damas et al. (7) [ISI = steady-state glucose infusion rate (mg·kg−1·min−1)/steady-state plasma insulin (μU/ml) × 100].
Pancreatic islet cell structure and function.
Islets were isolated from 9-mo-old OHF and OC (n = 4) by collagenase digestion of the pancreas and separated under a light microscope with a fine drawn pipette as previously described (17). Isolated islets were placed in a perifusion chamber at 37°C and perifused with 2–20 mM glucose solutions as described elsewhere (31). Fractions were collected every 2 min, and dynamic insulin secretion was determined by RIA (18). Total islet insulin content in isolated islets was also measured by RIA. Electron micrographs of glutaraldehyde-fixed, osmium-stained pancreata were prepared for assessment of islet cell ultrastructure with standard electron microscopy techniques. Histology of the exocrine pancreas at the light microscope level was investigated with hematoxylin and eosin staining.
At 6 mo of age OHF and OC (n = 10) were killed by a rising concentration of carbon dioxide and adiposity as assessed by the combined wet weight of intra-abdominal fat pads (retroperitoneal and perinephric) and gonadal fat lobes expressed as a percentage of body weight.
Fasting plasma samples from OHF and age-matched OC (n = 10) were obtained by cardiac puncture after animals were killed by a rising concentration of carbon dioxide (6–12 mo). Plasma leptin was evaluated with a rat leptin enzyme Immunometric assay kit [Assay Designs, Ann Arbor, MI; sensitivity 41.48 pg/ml, intra-assay coefficient of variation (CV) 5.6%, interassay CV 11.2%].
Measurement of mitochondrial DNA copy number.
OHF and OC were killed by cervical dislocation at 3 mo (n = 10). Kidney and liver were dissected, snap frozen in liquid nitrogen, and stored at −80°C. Approximately 0.2–0.5 g of tissue was ground in a freezer mill, and total DNA was prepared with the Promega Wizard genomic DNA purification kit (catalog no. A1120; Southampton, UK) and diluted to ∼1 mg/ml. Samples were microarrayed from a 96-well plate onto nylon filters with a BioRobotics MicroGrid II TAS microarrayer and then hybridized with radiolabeled probes corresponding to mitochondrial encoded ribosomal 16S gene or nuclear encoded ribosomal 18S gene. After washing to remove nonhybridized probe, the ratio of 16S to 18S was calculated for each sample, to give an estimate of the relative amount of mitochondrial DNA (mtDNA), i.e., the mitochondrial copy number.
Gene chip microarray and RT-PCR.
Total RNA was isolated from aorta of OHF and OC at 6 and 12 mo with an RNA isolator TRIzol solution per the manufacturer's protocol (Invitrogen, Paisley, UK). After processing, RNA pellets were resuspended in diethyl pyrocarbonate-treated water at a concentration of 1 μg/μl.
Double-stranded DNA was in vitro transcribed with T7 RNA polymerase in the presence of biotinylated nucleotides (5 h, 37°C). Biotinylated cRNA was fragmented into small (100 base) targets. cRNAs were hybridized to rat genome U34A chips (Affymetrix; 45°C).
Rat gene chips were washed according to the Affymetrix fluidics station protocol. Phycoerythrin linked to streptavidin was used to label the hybridized target. Affymetrix software was used to generate a comparison of the chips.
Microarray indicated increased expression of aorta mRNA for the mitochondrial voltage-dependent anion channel (VDAC)1, intimately involved in apoptosis, which in vascular endothelial cells has been implicated in atherogenesis (5). RT-PCR for this gene was therefore undertaken with GAPDH as the housekeeping gene. Total RNA from OHF and OC aorta (6 and 12 mo) was treated with RQ1 RNAse-free DNAse (Promega). Extracted RNA was combined with an oligo(dT) (18) primer and added to a first-strand cDNA synthesis reaction containing avian myeloblastosis virus reverse transcriptase (Promega) and incubated (45 min, 42°C). The reaction was terminated by incubation at 94°C (3 min). Primers were used to semiquantitate transcript levels of VDAC1 (forward primer 5′ CATATCAACCTGGGCTGTG 3′, reverse primer 5′ TTGGCTGCTA-TTCCAAAGC 3′; GAPDH forward primer 5′ GCATTGCTCTCAATGACAA 3′, reverse primer 5′ TGTGAGGGAGATGCTCAGTG 3′).
Plasma glucose was measured by a routine laboratory enzymatic UV test (HK/G6P-DH method; Cobas Fara centrifugal analyzer) and insulin by ELISA (DRG Instruments). Plasma triglyceride and total cholesterol concentrations were measured by enzymatic colorimetric assays (UNIMATE 5 TRIG and UNIMATE 5 CHOL; Roche/BCl, Lewes, Sussex, UK).
All statistical comparisons were made by Student's t-test. Glucose-stimulated insulin secretion in OHF and OC was compared by area under the curve. All values are given as means ± SE. Statistical significance was assumed at P < 0.05.
Whole body insulin sensitivity.
Insulin resistance was markedly increased in OHF compared with OC, as demonstrated by reduced glucose infusion rate and raised plasma steady-state insulin concentrations during the euglycemic clamp, resulting in a lower ISI (Fig. 1, A–C).
Plasma leptin concentration.
OHF demonstrated an increase in plasma leptin concentration compared with OC (Fig. 1D).
Plasma glucose, insulin, and lipid profiles.
At 6 mo of age, as reported elsewhere (Ref. 20, this issue), OHF demonstrated an increase in both fasting plasma glucose and insulin, although only the hyperinsulinemia reached significance [glucose (mmol/l): OC 12.4 ± 2.4 (n = 7) vs. OHF 18.1 ± 3.6 (n = 6), P = not significant (NS); insulin (ng/ml): OC 0.80 ± 0.12 (n = 7) vs. OHF 1.99 ± 0.55 (n = 6), P < 0.05]. At 12 mo of age, OHF also demonstrated significant hyperglycemia relative to controls [glucose (mmol/l): OC 12.7 ± 2.2 (n = 6) vs. OHF 22.5 ± 1.4 (n = 6), P < 0.01] (22).
Total cholesterol, HDL cholesterol, and triglycerides were not different between groups at 6 mo of age. At 12 mo of age, however, OHF demonstrated significantly raised triglycerides and significantly reduced HDL as previously reported in this cohort [triglycerides (mmol/l): OC 1.65 ± 0.15 (n = 6) vs. OHF 2.51 ± 0.27 (n = 6), P < 0.05; HDL (mmol/l): OC 2.20 ± 0.13 (n = 6) vs. OHF 1.6 ± 0.11 (n = 6), P < 0.01] (22).
Pancreatic islet cell structure and function.
Dynamic assessment of islet insulin secretory function in vitro (Fig. 2A) showed no difference in basal insulin release but a significant reduction in 20 mM glucose-stimulated insulin secretion in OHF, in both the first and second phases of insulin release (area under the curve, P < 0.05). Measurement of isolated islet total insulin content showed values to be significantly lower in OHF than in OC (Fig. 2B).
Electron micrographs (Fig. 3) showed that pancreatic β-cells in OHF islets had enlarged insulin secretory granules [diameter (nm), mean ± SE: OC 365 ± 9 vs. OHF 463 ± 16, P < 0.0001], more prominent electron-translucent halos, and a greater number of light core secretory granules than mature dense core secretory granules per β-cell compared with OC islets [pale granules (%): OHF 40 vs. OC 14]. Despite the increase in the overall size of the β-cell secretory granule, the core size was unaffected [diameter (nm): OC 140 ± 5 vs. OHF 144 ± 6, n = 50 granules, P = NS]. Granules of α-cells, which produce glucagon, did not differ between the two populations [diameter (nm): OHF 228 ± 5 vs. OC 230 ± 6, n = 50 granules, P = NS].
Morphology of the exocrine pancreas at the light microscope level with hematoxylin and eosin staining showed no obvious differences in the structure of the islet cells between the adult offspring of dams fed a “high-fat” diet in pregnancy and control offspring (Fig. 4).
We have reported (Ref. 20, this issue) increased fat deposition in these animals. OHF demonstrated a significant increase in abdominal fat deposition as a percentage of body weight at 6 mo compared with controls (OHF 4.44 ± 0.51% vs. OC 2.16 ± 0.50%, n = 10, P < 0.05).
Mitochondrial copy number.
Mitochondrial copy number tended to be lower in OHF liver compared with OC liver and was significantly reduced in OHF kidney [OC (n = 10) 1.05 ± 0.19 vs. OHF (n = 10) 0.66 ± 0.06, P < 0.05; Fig. 5, A and B].
Mitochondrial gene expression.
Gene chip analysis indicated a greater than fivefold decrease in mRNA expression of the mitochondrial genome. In addition, there was also an increase in mRNA expression of several nuclear encoded genes, including the apoptotic voltage-dependent ion channel VDAC1 (Fig. 5C).
RT-PCR for VDAC1 in aortic tissue taken from 6- and 12-mo-old OHF and OC showed that VDAC expression was markedly enhanced in OHF at 6 mo and further increased at 12 mo (Fig. 5D).
We reported previously (21, 22) that in utero and early postnatal exposure to a maternal diet rich in animal fat leads to the development of certain shared characteristics of the human metabolic syndrome in adult rats reared on a balanced diet. The demonstration in female adult offspring of raised plasma concentrations of fasting glucose and triglycerides together with an increase in body weight and fat depot weights was indicative, but not conclusive, of insulin resistance. This study has shown unambiguously that female offspring of dams fed a fat-rich diet demonstrate insulin resistance at 1 yr of age. In addition, isolated pancreatic β-cells from 9-mo-old offspring of dams fed a high-fat diet showed impaired glucose-stimulated insulin release. Thus hyperglycemia, insulin resistance, and insulin secretory deficiency, the principal metabolic predictors of type 2 diabetes mellitus in humans (36), are all present in this model.
This is the first study to demonstrate whole body insulin resistance with a euglycemic clamp technique in a dietary model of fetal programming. A similar study by Park et al. (30) in offspring of protein-restricted dams at 6 mo of age showed no difference in whole body insulin sensitivity index despite impaired β-cell insulin secretion.
Insulin may either directly, or indirectly via its lipogenic action, stimulate adipocyte leptin secretion via a dual hormonal feedback loop between adipose tissue and the endocrine pancreas, the proposed adipoinsular axis (23). Hence, the increased adiposity and hyperinsulinemia are likely contributors to the rise in plasma leptin concentration and suggest the development of leptin resistance at the level of the β-cells. Because leptin also modulates cardiovascular homeostasis via sympathetic activation of the arterial baroreflex (10), the increased concentrations may also contribute to hypertension in this model (22).
Tissue insulin resistance is central to the metabolic syndrome and the development of type 2 diabetes and is manifest by impaired glucose uptake into skeletal muscle and failure to suppress hepatic glucose output. However, ultimately it is the failure of the pancreatic β-cells to maintain glucose homeostasis that leads to hyperglycemia and the development of overt type 2 diabetes mellitus (36). This can be explained in terms of insulin secretory defects and a selective loss of sensitivity to glucose (25). Dynamic assessment of islet insulin secretory function in vitro showed no difference in basal insulin release but a significant reduction in glucose-stimulated (20 mM) insulin secretion in the offspring of dams fed a fat-rich diet, both in the initial and sustained phase of insulin release, indicative of a reduced insulin secretory capacity. The delay in onset of the first phase of glucose-stimulated insulin secretion suggests a reduced sensitivity to glucose.
Altered islet cell structure frequently accompanies altered function, and electron micrographs showed pancreatic β-cells from OHF had enlarged insulin secretory granules, more prominent electron-translucent halos, and a greater number of immature light core secretory granules than mature dense core secretory granules per β-cell compared with OC β-cells. Proinsulin in the pale immature granules is slowly cleaved to insulin, which becomes associated with zinc as the granule matures to form a dense central core of zinc insulin crystal. An increase in the ratio of light to dark granules in β-cells occurs in conditions of increased insulin secretion such as pregnancy (1) or chronic hyperglycemia (8) and is therefore consistent with the chronic stimulatory environment of raised fasting plasma glucose we have reported in 6- and 12-mo-old female offspring (9, 22). Moreover, total insulin content of isolated islets was significantly lower than controls in 9-mo-old OHF, which may suggest declining β-cell mass and/or β-cell exhaustion similar to that of type 2 diabetes in humans.
Both glucotoxicity (12) and lipotoxicity (35) have been implicated in the chronic deterioration of insulin secretion in type 2 diabetes mellitus, and both may contribute to impaired β-cell function in the present study. Conversely, defective adulthood β-cell function or reduced β-cell mass has also been proposed to arise from an acute early programming stimulus in neonatal rats subjected to in utero protein restriction (16).
In a another rat model, in which rat dams were fed a saturated fat-rich diet (18% coconut oil) during pregnancy and weaning, 12-wk-old adult offspring showed a reduction in the population of large pancreatic islet cells, although the functional significance of this observation was not clear (32). In response to a glucose load, the offspring showed an increased insulin response compared with controls, indicative of insulin resistance, although plasma glucose concentrations were actually lower after 30 min.
The chronology of metabolic disturbances in this model may give insight into the etiology of the “metabolic syndrome” phenotype. In female offspring, we have previously shown endothelial dysfunction at 80 days of age (22), and we now report reduced mitochondrial copy number at a similar age; both antedate dyslipidemia, hyperglycemia, hypertension, and obesity, and a fundamental mechanistic role for each may be indicated (21, 22). The onset of overt hyperglycemia at 12 mo but not at 3 mo suggests late pancreatic β-cell failure in the maintenance of glucose homeostasis. Further investigation of the chronology of metabolic disturbances will provide insight into the etiology of metabolic syndrome in this model, but endothelial dysfunction emerges as a fundamental defect. Indeed, peripheral insulin resistance has been proposed to result from endothelial dysfunction in capillaries and arterioles, leading to poor perfusion and reduced glucose uptake in metabolically active insulin-sensitive tissues (6).
The molecular mechanisms underpinning long-term programming of adult disease are likely to be complex and multifactorial. Mitochondria are exquisitely sensitive to environmental changes, and the efficiency and accuracy of mitochondrial replication can be affected by environmental factors (11). Reduction in the level of mitochondrion-specific DNA γ-polymerase has been investigated in clinical and experimental models and has been found to be a predictor of type 2 diabetes (26) and to be altered in the tissues of animals fed a low-protein diet (30).
The early reduction in mtDNA copy number observed in the kidney, if indicative of a global reduction in mtDNA content, would concur with the suggestion that mitochondria play a role in developmental programming. The altered mitochondrial genomic expression in the aorta by microarray is also supportive of this hypothesis. The mitochondrion is intimately involved in glucose sensing by pancreatic β-cells, and defects in this organelle, including reduced mtDNA (19), are associated with impaired glucose-stimulated insulin release (27). Of particular relevance to the current study, Higa and colleagues (14) reported altered mitochondrial morphology in isolated islets from Zucker diabetic fatty (ZDF) rats, associated with lipotoxicity and lipoapoptosis of β-cells. However, in the present study, histology of the islet cells was not different in offspring of high-fat-fed animals, and in contrast to the mitochondrial swelling and cristae disruption reported by Higa and colleagues in untreated obese ZDF rats, no distinct morphological changes were visible in the mitochondria at the ultrastructural level. This suggests that the programmed β-cell changes observed in our studies were not as profound as those seen in overtly diabetic animals where the β-cell mass had decreased markedly. Future studies are planned to specifically address mitochondrial integrity in pancreatic islet β-cells.
This disruption of mitochondrial gene activity in the aorta could compromise mitochondrial function, eventually leading to tissue damage within the vascular wall. Of particular interest was the upregulation of the apoptotic pore gene VDAC1. Chronic inflammation as a consequence of subclinical tissue damage due to aberrant mitochondrial activity may initiate endothelial dysfunction early in the phenotypic development of this model (3).
In conclusion, this animal model of developmental programming produces a female offspring phenotype typical of the metabolic syndrome. In addition to the hypertension, vascular endothelial dysfunction, dyslipidemia, and increased adiposity previously reported (9, 21, 22), we now demonstrate marked changes in insulin resistance and insulin secretory capacity. Pancreatic β-cells from female offspring of dams fed a fat-rich diet in pregnancy demonstrate altered granule morphology and glucose-stimulated insulin secretory deficiency. We hypothesize a global reduction in mitochondrial DNA copy number in early development, eventually leading to mitochondrial dysfunction, tissue damage, and disease.
We gratefully acknowledge the British Heart Foundation (RG 99011), Guys and St. Thomas' Charitable Foundation (R020714), Tommy's, the baby charity, and Scottish Executive Environmental Rural Affairs Department for funding this work. L. Poston is supported by Tommy's, the baby charity.
Present address of J. McConnell: Div. of Reproductive Health, Endocrinology and Development, Kings College London, London SE1 7EH, UK.
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 © 2005 the American Physiological Society