Epidemiological studies suggest that deleterious prenatal nutrient environments can lead to an increased risk of chronic disease, including type 2 diabetes (T2D), in adult offspring( Reference Godfrey and Barker 1 , Reference Godfrey and Barker 2 ). These observations have been reconciled into the developmental origins of health and disease (DOHaD) hypothesis, which proposes that an individual faces elevated chronic disease risk in adult life when the maternal nutrient environment is incongruent with the postnatal environment( Reference Wadhwa, Buss and Entringer 3 ). Fetal overnutrition, along with maternal hyperglycaemia and hyperlipidaemia, has been linked to both obesity and T2D in adult offspring( Reference Dabelea 4 ). However, the mechanisms by which maternal diets effect these persistent metabolic changes are still poorly understood.
Increased consumption of energy-dense, Western diets( Reference Cordain, Eaton and Sebastian 5 ), consisting of high levels of refined carbohydrates and lipids, has been linked to obesity and T2D( Reference Bray and Popkin 6 – Reference Hu 10 ). Consequently, high-fat, high-sucrose (HFS) diets, simulating a Western diet, when fed to rodent models result in increased body weight and abdominal fat deposition, hyperinsulinaemia, and hyperglycaemia( Reference Murase, Mizuno and Omachi 11 ). Moreover, HFS exposure in utero has been shown to alter glucose, insulin and lipid homoeostasis( Reference Zheng, Xiao and Zhang 12 – Reference Wanjihia, Ohminami and Taketani 14 ) and induces adipose accumulation and fatty liver in adult rodent offspring( Reference Cordero, Gomez-Uriz and Milagro 13 , Reference Cordero, Milagro and Campion 15 ). The precise mechanism by which gestational HFS exposure alters glucose and insulin homoeostasis and the expression of genes and proteins associated with insulin signalling and lipid metabolism remains largely unknown. However, maternal folic acid supplementation has been shown to modulate inheritance of the deleterious metabolic effects in offspring( Reference Lillycrop, Phillips and Jackson 16 , Reference Burdge, Lillycrop and Phillips 17 ) and prevents the onset of hypertension, endothelial dysfunction and adiposity( Reference Cordero, Milagro and Campion 15 , Reference Torrens, Brawley and Anthony 18 – Reference Seto, Lam and Or 20 ). In the case of insulin sensitivity, there is conflicting evidence on the potential of gestational folic acid supplementation to influence this phenotype: prenatal exposure to high folate appears to induce greater insulin resistance (IR) in offspring( Reference Huang, He and Sun 21 , Reference Yajnik, Deshpande and Jackson 22 ), whereas a low-folate maternal diet induces higher body weights and adiposity, IR and high blood pressure in male offspring( Reference Sinclair, Allegrucci and Singh 23 ). As such, further work is required to better understand the role of folic acid in glucose homoeostasis and its effect on the inheritance of genes and proteins associated with insulin signalling and lipid metabolism.
IR, a precursor to T2D, is characterised by a decreased response to the effects of insulin on the liver and skeletal muscle, which respectively play central roles in glucose and lipid metabolism and insulin-mediated glucose uptake( Reference Stumvoll and Gerich 24 ). The HFS diet has been observed to alter insulin signalling by down-regulation of insulin receptor substrate 2 (Irs2) and v-Akt murine thymoma viral oncogene homolog 2 (Akt2), and to alter lipid metabolism by up-regulation of fatty acid synthase (Fasn) in the livers of mice( Reference Yang, Miyahara and Takeo 25 ) and down-regulation of PPARG gene expression in the livers of male rats( Reference Ragab, Abd Elghaffar and El-Metwally 26 ). HFS feeding significantly reduces insulin-stimulated glucose transport by reducing translocation of the solute carrier family 2, facilitates GLUT member 4 (Slc2a4) in the skeletal muscle and elevates serum insulin and TAG levels( Reference Youngren, Paik and Barnard 27 , Reference Barnard, Roberts and Varon 28 ). The lipogenic genes acetyl-CoA carboxylase (Acaca) and Fasn can be trans-activated by the synergistic actions of insulin and glucose( Reference Stoeckman and Towle 29 ) and down-regulated by PUFA( Reference Sampath and Ntambi 30 , Reference Jump 31 ), so it is likely that an HFS diet may alter the regulation of these genes. Therefore, investigations of the expression of these genes and proteins associated with impaired insulin signalling and dyslipidaemia may provide useful insights into the mechanism(s) by which a prenatal HFS diet may influence offspring insulin sensitivity and adiposity.
We hypothesised that an HFS diet fed during pregnancy will increase body weights, induce IR and down-regulate insulin signalling and lipid metabolism genes and proteins in the offspring, and that folic acid supplementation of the maternal pregnancy diet will modulate these effects.
Methods
Animal husbandry and experimental diets
All animal protocols were conducted in accordance with animal husbandry standards established by the Ethics Committee, Faculty of Medical Sciences, The University of the West Indies, Trinidad and Tobago. In brief, a total of fifteen pairs of nulliparous female and male Sprague–Dawley rats, weighing 200–300 g, were obtained from the Animal House facility, Department of Veterinary Medicine, and were housed individually with exposure to a 12 h light–12 h dark cycle at 23–25°C throughout the experiment. Female rats were mated monogamously and the date of conception noted. After confirmation of pregnancy, females were allocated to one of three dietary feeding groups (n 5/dietary group) from conception date until parturition. The date of parturition for each dam was designated postnatal day 0 for the litter. Food and water were accessed ad libitum. The three experimental diets were prepared commercially (Dyets Inc.) and comprised AIN-93G purified rodent control diet (CON: 16 % protein, 17 % soyabean oil, 11 % sucrose and 2 mg/kg folic acid)( Reference Reeves, Nielsen and Fahey 32 ); HFS diet (16 % protein, 14 % soyabean oil and 26 % lard, 40 % sucrose and 2 mg/kg folic acid) and folic-acid-supplemented HFS diet (HFS/F: 16 % protein, 14 % soyabean oil and 26 % from lard, 40 % sucrose and 5 mg/kg folic acid). Folic acid concentrations of 2 and 5 mg/kg of rodent diet have been used extensively in supplementation investigations( Reference Lillycrop, Phillips and Jackson 16 – Reference Torrens, Brawley and Anthony 18 , Reference Huang, He and Sun 21 , Reference Hoile, Lillycrop and Grenfell 33 ) given that these levels are analogous to the recommended daily allowance of folic acid for adults( 34 ) and expectant women( Reference Yates, Schlicker and Suitor 35 ), respectively. The complete compositions of the diets are summarised in Table 1.
CON, control; HFS, high-fat/high-sucrose; HFS/F, folic-acid-supplemented, high-fat/high-sucrose
* Dyetrose® (Dyets), a dextrinised maize starch composed of 90–94 % tetrasaccharides.
† AIN-93 mineral mix (Dyet no. 210025)
‡ AIN-93-VX vitamin mix (Dyet no. 310025). All diets provided by Dyets Inc.
After parturition, all dams were fed the CON diet throughout the lactation period. All litters were standardised to six pups per litter within 48 h to avoid any inequities in energy consumption among and within the dietary groups during lactation. Birth weight measurements were not taken to avoid maternal cannibalisation of the pups( Reference Lane-Petter 36 ). Litter weights were measured weekly for 10 weeks after postnatal day 7. Male offspring were weaned at postnatal day 28 and maintained on the CON diet, ad libitum. At postnatal day 70, all animals were euthanised by administration of a lethal dose of sodium pentobarbital, and dissected along the ventral midline from the urogenital area to the sternum to expose the organs. Whole blood was collected in EDTA tubes by cardiac puncture and centrifuged to obtain plasma, which was stored at −20oC. The livers and skeletal muscle (gastrocnemius from hind leg) of the male offspring were collected rapidly and flash-frozen for storage at −80°C. Only the tissues of the male offspring were used for these analyses to avoid confounding effects of hormonal changes in the female offspring during their life cycle, which may affect weight gain and development( Reference Figueiredo, Dolgas and Herman 37 ).
RNA extraction and gene expression assays
RNA was extracted from male offspring liver and skeletal muscle tissues using the RNeasy Mini kit, according to the manufacturer’s instructions (Qiagen). Purified RNA was converted to complementary DNA by use of the RT2 first-strand kit according to the manufacturer’s instructions (Qiagen) and was stored overnight at −20°C. Gene expression was determined using the RT2 Profiler PCR Array for the rat insulin signalling pathway (Qiagen) on an ABI 7500 real-time PCR machine (Applied Biosystems). The preset dissociation stage for the ABI 7500 real-time PCR machine was included on the thermal profile as recommended by the manufacturers (Qiagen). The threshold cycle (C t ) values were obtained from the real-time PCR and used to calculate fold changes for the genes of interest.
Western blotting
Frozen liver and skeletal muscle tissue were homogenised with ice-cold RIPA lysis and extraction buffer (Thermo Scientific™) supplemented with Halt™ Protease and Phosphatase Inhibitor (Thermo Scientific™) using the TissueLyser II (Qiagen) and supernatants were obtained by centrifugation at 15 000 rpm at 4oC. Total protein concentration was determined by the Bradford protein-dye binding assay( Reference Bradford 38 ) using the Pierce™ pre-diluted protein assay Bovine Serum Albumin standards (Thermo Scientific™). Protein samples (10 µg) were electrophoresed on 4–12 % SDS-PAGE mini-gels along with an internal protein control, and transferred onto Immobilon™-P polyvinylidenedifluoride (PVDF) membranes (Sigma-Aldrich). Non-specific antibody binding was blocked by incubating the PVDF in Tris-buffered saline (TBS) containing 0·05 % Tween 20 (TBST) containing either 5 % bovine serum albumin or 5 % skimmed milk, according to recommendations on the primary antibody. The PVDF membranes were subsequently probed with primary antibodies specific for Acaca, Akt2, B-Raf (1:1000 dilution; Cell Signaling Technology), Slc2a4 (1:1000 dilution; Santa Cruz Biotechnology Inc.) and their phosphorylated forms. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primary antibody (1:1000 dilution; Cell Signaling Technology) was used as a housekeeping protein control. Unbound primary antibody was removed by washing the PVDF three consecutive times in TBST. The PVDF was then probed with alkaline-phosphatase-conjugated anti-mouse IgG secondary antibody (1:1000 dilution; Cell Signaling Technology). Unbound secondary antibody was removed by washing the PVDF in TBST five times. The blots were developed using Alkaline Phosphatase conjugate substrate kit (Bio-Rad) and the blot images were captured with the GeneSnap image acquisition software (Syngene) and analysed using the ImageJ densitometric software (US National Institutes of Health).
Fasting plasma glucose and insulin assays
Fasting plasma glucose concentrations were measured using the ‘Semi-micro’ protocol of the Glucose liquicolor kit as per the manufacturer’s instructions (Human Diagnostics Worldwide). Fasting plasma insulin concentrations were determined using the Rat/Mouse Insulin ELISA ninety-six-well plate assay kit according to the manufacturer’s instructions (EMD Millipore Corporation). Insulin values were obtained by extrapolation from plotting a graph of absorbance at 450 nm, less than that obtained at 590 nm, on the y-axis against the concentration of rat insulin standards on the x-axis. The homoeostasis model assessment index of IR (HOMA-IR), quantitative insulin check index (QUICKI) and fasting plasma glucose:insulin ratio (FGIR) were calculated as described by Cacho et al.( Reference Cacho, Sevillano and de Castro 39 ).
Data analysis
A priori and post hoc power analyses were performed using G*Power 3.1.5 software( Reference Faul, Erdfelder and Lang 40 ). With an α level of 0·05, power established at 80 % and an effect size of 0·9, the required total sample size was 18. The hypothesised effect size of 0·9 was calculated from the descriptive statistics of a previous study( Reference Cordero, Gomez-Uriz and Campion 41 ). Post hoc calculations using the hypothesised effect size and the total sample size of 15 (i.e. n 5/group) indicated that the actual power achieved in this study was 79 %.
Data are presented as means with their standard errors. Gene expression fold change was calculated as $$2^{{{\minus}\Delta \Delta C_{t} }} $$ , where ΔC t for each gene is the C t of the housekeeping gene, β-actin, subtracted from the C t of the gene of interest; Δ(ΔC t ) is the ΔC t for the control samples subtracted from the ΔC t of either the HFS or HFS/F samples. To analyse the protein expression, the area under the densitometric graphs (AUC), corresponding to the blot intensity for each sample, was obtained from the ImageJ software. The relative density for each protein sample was determined by dividing the AUC for each sample by the AUC for the internal protein control used in each Western blot assay. The adjusted relative density for each protein of interest was normalised by dividing the relative density for the protein of interest by the relative density of the housekeeping protein GAPDH. Mean differences for the gene and protein expression data were assessed by one-way ANOVA. Bonferroni’s post hoc test was conducted for multiple comparisons. For all tests, P<0·05 was determined to be statistically significant. All statistical analyses were performed using SPSS software for Windows version 16.0 (IBM).
Results
Offspring litter weights
There was no significant difference in food consumption patterns among the dams of the various study groups. Mean litter weights of CON and HFS/F offspring were significantly lower (P≤0·05) than the HFS offspring from postnatal week 7 onwards to 10 (Fig. 1).
Liver and skeletal muscle gene expression
There were significant treatment effects in offspring liver gene expression for four genes: Acaca, Fasn, Kruppel-like factor 10 (Klf10) and phosphatidylinositol-4,5-biphosphate 3-kinase, catalytic subunit β (Pik3cb) (Table 2). Hepatic Acaca (95 % CI 0·0188, 0·318) and Pik3cb (95 % CI 0·332, 2·47) were significantly down-regulated in HFS v. CON. Conversely, Klf10 (95 % CI 0·0019, 0·0287) expression was significantly up-regulated in HFS v. CON offspring. However, hepatic Fasn (95 % CI 0·252, 1·33) gene expression was significantly down-regulated in HFS compared with the HFS/F offspring.
Acaca, acetyl-CoA carboxylase; Fasn, fatty acid synthase; Klf10, Kruppel-like factor 10; Pik3cb, phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit β; Akt2, V-Akt murine thymoma viral oncogene homolog 2; Cbl, Cbl proto-oncogene, E3 ubiquitin protein ligase; Dok3, docking protein 3; Npy, neuropeptide Y; Raf1, V-Raf-1 murine leukaemia viral oncogene homolog 1; Retn, resistin; Slc2a4, solute carrier family 2, facilitated GLUT, member 4.
Gene expression fold changes were significantly different from those of the control group: * P<0·05, ** P<0·01.
Offspring skeletal muscle gene expression differed significantly (P<0·05) among the study groups (Table 2). In particular, Akt2 (95 % CI 0·0004, 0·0285), docking protein 3 (Dok3; 95 % CI 0·0002, 0·0087), V-Raf-1 murine leukaemia viral oncogene homolog 1 (Raf1) (95 % CI 0·0045, 0·0413), Slc2a4 (95 % CI 0·0067, 0·0276) and resistin (Retn) expression levels (95 % CI 0·0004, 0·0027) were significantly (P<0·05) down-regulated in the HFS v. CON group, respectively. In contrast, neuropeptide Y (Npy) expression (95 % CI 0·0215, 0·489) was significantly (P<0·05) up-regulated in the HFS v. CON group with an almost 2-fold up-regulation. Cbl proto-oncogene, E3 ubiquitin protein ligase (Cbl), expression (95 % CI 0·0033, 0·0300) was significantly (P<0·05) down-regulated in HFS compared with HFS/F.
Liver and skeletal muscle protein expression
Liver and skeletal muscle protein expression values are shown in Table 3. Compared with CON, liver Acaca expression was significantly reduced with both HFS (P<0·05) and HFS/F (P<0·05) diets. Further, folic acid supplementation of the maternal HFS diet was unable to prevent the reduction in hepatic Acaca expression as significant differences were not observed between the HFS and HFS/F groups (P>0·05). In skeletal muscle, maternal HFS diet, but not supplemental folic acid, resulted in a significant reduction in phosphorylated C-Raf expression (P<0·05).
Acaca, acetyl-CoA carboxylase; pAcaca, phosphorylated acetyl-CoA carboxylase; Akt2, V-Akt murine thymoma viral oncogene homolog 2; pAkt2, phosphorylated V-Akt murine thymoma viral oncogene homolog 2; B-Raf, B-Raf proto-oncogene, serine/threonine kinase; pC-Raf, phosphorylated C-Raf proto-oncogene, serine/threonine kinase; Slc2a4, solute carrier family 2, facilitated GLUT, member 4; pSlc2a4, phosphorylated solute carrier family 2, facilitated GLUT, member 4.
* Mean protein expression levels were significantly different between the control and HFS (P<0·05).
† Mean protein expression levels were significantly different between the control and HFS/F groups (P<0·05).
Measures of glucose homoeostasis
Table 4 shows that the offspring from the maternal HFS group had significantly higher fasting plasma glucose concentrations than their HFS/F counterparts (P=0·025), indicating that the folic acid supplementation of the maternal diet prevented offspring hyperglycaemia. However, fasting plasma insulin levels were significantly higher in the HFS/F than in the CON offspring (P=0·016) (Table 4). Among the study groups, HOMA-IR was lowest in CON offspring, whereas HFS/F offspring had the highest values (Table 4), and although there was an overall treatment effect on HOMA-IR (P=0·030), post hoc analysis failed to identify any significant group difference. CON offspring had significantly higher QUICKI values than the HFS/F offspring (P=0·018). In addition, CON offspring had a significantly (P<0·05) higher FGIR than their HFS/F counterparts. Collectively, the results of QUICKI, FGIR and HOMA-IR all suggest that the HFS/F maternal diet induced IR in the offspring.
CON, control; HFS, high-fat/high-sucrose; HFS/F, folic-acid-supplemented, high-fat/high-sucrose; HOMA-IR, homoeostasis model of assessment index; QUICKI, quantitative insulin sensitivity check index; FGIR, fasting plasma glucose:insulin ratio.
* Mean glucose and insulin concentrations, HOMA-IR, QUICKI and FGIR were significantly different from those of the control group (P<0·05).
Discussion
Gestational HFS diet was hypothesised to adversely affect metabolic pathways associated with T2D, insulin signalling and lipid metabolism through trans-generational effects on glucose metabolism, gene and protein expression in the offspring. Further, folic acid supplementation was also hypothesised to augment these effects via transcriptional regulatory pathways. Our results show that prenatal exposure to HFS followed by post-weaning CON diet predisposed male offspring to significant hyperglycaemia, impaired de novo lipogenesis and higher weight gain as compared with offspring exposed to CON during gestation and post-weaning: these are all findings that support the DOHaD mismatch paradigm( Reference Godfrey, Lillycrop and Burdge 42 ).
Altered glucose metabolism in HFS offspring has been previously demonstrated with prenatal feeding of high-fat sucrose diets in rodents( Reference Murase, Mizuno and Omachi 11 , Reference Zheng, Xiao and Zhang 12 ). Similarly, a trend of higher glucose and insulin levels has been observed after gestational and lactation HFS feeding of adult male Sprague–Dawley rats( Reference Latouche, Heywood and Henry 43 ). Our results provide further insights into the mechanism by which gestational HFS exposure induces hyperglycaemia in offspring and suggest that this may be a result of modulations in skeletal muscle gene expression compared with CON via down-regulation of Akt2 and Raf1. Akt2 phosphorylates Raf1, which activates the downstream kinases Mek and Erk resulting in Slc2a4 translocation to the muscle cell membrane and uptake of glucose( Reference Tan, Ng and James 44 , Reference Kohn, Summers and Birnbaum 45 ). Hence, down-regulation of skeletal muscle Akt2 and Raf1 may impair insulin-induced translocation of Slc2a4, which was confirmed by down-regulation of Slc2a4 gene expression. Akt2-deficient mice had been previously observed to display impaired glucose homoeostasis in both liver and skeletal muscle( Reference Cho, Mu and Kim 46 ), whereas increased Akt protein expression in skeletal muscle of adult offspring exposed to a prenatal obesogenic diet improved their insulin sensitivity, although such a result was unexpected( Reference Shelley, Martin-Gronert and Rowlerson 47 ). Our results show that expression of the phosphorylated form of Raf1 protein was significantly lower in the skeletal muscle of HFS v. CON offspring, which implies that Raf1 protein activity is suppressed and can less efficiently activate effector kinases that promote insulin-mediated glucose uptake. Further, our findings indicate that prenatal folic acid- supplementation reduces hyperglycaemia induced by gestational HFS diet. Taken together, these findings could account for the HFS offspring having significantly higher fasting plasma glucose concentrations than HFS/F offspring and non-significantly higher fasting plasma glucose concentrations than CON offspring.
Our results also show that HFS offspring were predisposed to significantly higher weight gain as compared with CON offspring. Higher weight gain and adiposity have been observed in male Wistar rat offspring fed HFS during pregnancy and lactation – an effect that was prevented by maternal methyl-donor supplementation with a mixture of choline, betaine, folic acid and vitamin B12 Reference Cordero, Milagro and Campion (15 , Reference Cordero, Gonzalez-Muniesa and Milagro 48 ). Our study shows that similar preventative effects may be achieved by supplementation of maternal HFS with 3 mg/kg folic acid. Down-regulation of Retn, the hormone that suppresses the ability of insulin to stimulate glucose uptake and storage as fat( Reference Barnes and Miner 49 ), suggests that the excess glucose that is not being transported into the HFS offspring skeletal muscle may instead be shunted into adipose tissue. In addition, increased body mass may also be as a result of up-regulation of Npy in HFS offspring given that increased Npy secretion in HFS-fed mice has been shown to induce hyperinsulinaemia, glucose intolerance and increased visceral fat mass( Reference Kuo, Czarnecka and Kitlinska 50 ).
The results of the hepatic gene and protein expression suggest impairments in de novo lipogenesis through decreased long-chain fatty acid synthesis via down-regulation of Acaca and Fasn in HFS offspring compared with CON and HFS/F. This result was unexpected given that high-fat, high-carbohydrate diets have been shown to stimulate hepatic de novo lipogenesis( Reference Strable and Ntambi 51 ). Hyperglycaemia increases hepatic de novo lipogenesis by stimulating the release of insulin, which positively controls the lipogenic transcription factor, sterol regulatory element-binding protein 1c (SREBP-1c)( Reference Foretz, Guichard and Ferre 52 ). SREBP-1c inturn regulates fatty acid synthesis by inducing the expression of lipogenic genes such as Acaca and Fasn ( Reference Yang, Miyahara and Takeo 53 ). However, hepatic SREBP-1c is inhibited by PUFA, which triggers the degradation of SREBP-1c mRNA( Reference Duarte, Carvalho and Pearson 54 ). Hence, in the presence of increased metabolic flux of fatty acids associated with HFS, it seems reasonable that there would be a down-regulation of fatty acid synthesis as evidenced by down-regulation of Acaca and Fasn. Interestingly, it appears that supplemental folic acid (HFS/F) was able to augment this effect to some extent.
Klf10 is a transcriptional regulator of energy metabolism, which has been shown to be induced in the liver by glucose stimulation of carbohydrate response element-binding protein (ChREBP) in a rodent model( Reference Iizuka, Takeda and Horikawa 55 ). Therefore, the hyperglycaemia experienced by HFS offspring may account for the up-regulation of Klf10, given that it has been shown that overexpression of Klf10 inhibits glucose-stimulated ChREBP target genes such as Fasn and Acaca in rat primary hepatocytes( Reference Barnes and Miner 49 ). It is likely that up-regulation of Klf10 expression may also contribute to the inhibition of long-chain fatty acid synthesis in HFS offspring. Transcription of the Fasn promoter is stimulated by insulin via the PI3-K signalling pathway( Reference Sul, Latasa and Moon 56 ); as such, hyperinsulinaemia may be the result of increased Fasn gene expression as found in the HFS/F offspring. Unphosphorylated Acaca protein expression was also significantly lower in hepatic tissue of HFS and HFS/F compared with CON offspring, which is likely to be associated with impaired fatty acid synthesis. Unlike the gene expression data, there was no significant difference in unphosphorylated Acaca protein expression between HFS and HFS/F offspring, indicating that prenatal folic acid supplementation was unable to improve fatty acid synthesis. Although the gene expression levels were significantly different among the treatment groups, the minimal effects of the HFS diet on protein expression may be due to varying levels of translation of the genes expressed. Therefore, the gene expression levels for Akt2, pAkt2, Slc2a4 and pSlc2a4 may not have been high enough to produce significant differences at the level of translation to protein.
The observed changes in HOMA-IR, FGIR and QUICKI indices suggest that gestational HFS/F induced significant perturbations in glucose homoeostasis. Interestingly, folic acid supplementation decreased blood glucose, but was associated with an increase in insulin. As such, HOMA-IR was significantly lower in CON offspring than HFS and HFS/F, with no difference between the latter groups, indicating that supplemental folic acid had no protective effect on IR. We suggest that increased circulating insulin in HFS/F offspring, despite their low circulating glucose, could be attributed to a reduction in dietary folic acid as the animals transitioned between the prenatal HFS/F diet and the postnatal CON diet. It is possible that in utero the animals might have become programmed to increased folate requirements and an unsupplemented postnatal CON diet was associated with the observed impaired HOMA-IR. Unfortunately, we did not measure plasma folate levels, but folate depletion has been associated with increased hepatic lipid peroxidation products, an indicator of increased hepatic oxidative stress( Reference Huang, Hsu and Lin 57 ). Further, decreased folic acid is associated with altered homocysteine metabolism, given that 5-methyl tetrahydrofolate remethylates homocysteine to methionine, and resulting hyperhomocysteinemia has consistently been associated with hyperinsulinaemia and IR( Reference Meigs, Jacques and Selhub 58 ). Alternatively, the level of folic acid supplementation may not have been high enough to confer a protective effect against IR. A previous study demonstrated a trend of HFS offspring becoming insulin resistant, although the treatment effects were not significantly different( Reference Latouche, Heywood and Henry 43 ). Folic acid has been shown to improve insulin sensitivity in patients with metabolic syndrome( Reference Setola, Monti and Galluccio 59 ) and overweight adults( Reference Solini, Santini and Ferrannini 60 ); both attributed to an anti-inflammatory effect of folic acid( Reference Solini, Santini and Ferrannini 60 , Reference Gonda, Kim and Salas 61 ). However, women in Pune, India, who consumed 500 µg of folic acid/d during pregnancy, and had the highest erythrocyte folate levels, had 6-year-old children with IR and increased total fat mass( Reference Yajnik, Deshpande and Jackson 22 ) in comparison with controls. In another study by Huang et al. ( Reference Huang, He and Sun 21 ), exposure to high folic acid concentrations (40 mg folic acid/kg diet) in utero predisposed male rat offspring to impaired glucose tolerance and greater IR than their control and low folic acid (5 mg folic acid/kg diet) counterparts after 8 weeks of high-fat-diet feeding. Our findings suggest that prenatal exposure to 5 mg of folic acid/kg diet results in significantly higher insulin even in the absence of postnatal high-fat feeding. This raises concern about the effect of folic acid on IR given that prolonged hyperinsulinaemia in HFS/F offspring could result in constant exposure to stimulatory insulin concentrations and thereby increases in insulin receptor internalisation and degradation, which could lead to a reduction in the number of insulin receptors on the cell surface and eventually IR( Reference Shanik, Xu and Škrha 62 ). As mentioned above, there is need for further studies to define the optimal dose of folic acid needed to augment the diabetogenic effects of prenatal HFS exposure.
Hyperinsulinaemia observed in HFS/F offspring compared with their CON counterparts may be attributed to supplemental folic acid inducing a compensatory over-secretion of insulin in response to hyperglycaemia caused by the HFS diet. Higher insulin concentrations would allow increased glucose uptake by inducing the translocation of glucose transporters; this was evident in the significantly lower fasting plasma glucose concentrations in the HFS/F v. HFS offspring. It is possible that the effect may be mediated by the Cbl protein, which targets tyrosine-phosphorylated substrates, such as Irs1 and Irs2, for proteasome degradation( Reference Thien and Langdon 63 , Reference Nakao, Hirasaka and Goto 64 ). As Irs1 and Irs2 phosphorylate downstream kinases of the insulin signalling pathway resulting in glucose uptake in the skeletal muscle, it was expected that significant up-regulation of Cbl could lead to hyperglycaemia. However, blood glucose concentrations of HFS/F offspring were significantly lower than HFS. This paradox may be explained, in part, by the compensatory mechanisms of other genes that facilitate enhanced glucose uptake. The HFS gestational diet also seems to predispose the offspring hepatocytes to decreased downstream insulin signalling, as suggested by down-regulation of Pik3cb ( Reference Kolic, Spigelman and Plummer 65 ), the catalytic p110β subunit of PI3-kinase, which is involved in signal transduction of insulin. Furthermore, Pik3cb is a positive regulator of insulin secretion and therefore the significantly lower Pik3cb gene expression in the HFS v. CON offspring may account for the lower insulin secretion in the HFS offspring( Reference Bieche, Bougneres and Dechartres 66 ). It is therefore possible that suppression of insulin secretion may also account for the hyperglycaemia observed in the HFS offspring.
Another possible contributor to the onset of IR in the offspring is high dietary fructose (derived from the sucrose) content of the HFS diet. Consumption of high-fructose diets has been associated with induction of IR, impaired glucose tolerance, hypertriacylglycerolaemia and hyperinsulinaemia in animal models( Reference Basciano, Federico and Adeli 67 , Reference Elliott, Keim and Stern 68 ) and impaired insulin sensitivity in humans( Reference Aeberli, Hochuli and Gerber 69 ). Fructose metabolism and hyperglycaemia can result in oxidative damage through the production of reactive oxygen species( Reference Busserolles, Gueux and Rock 70 ), and it has been suggested that this oxidative stress might play a contributory role in the aetiology of IR and eventually T2D.
It is likely that the differences in gene and protein expression observed among the CON, HFS and HFS/F offspring may occur through epigenetic mechanisms. Epigenetics refers to a heritable change in gene expression that occurs without DNA sequence changes and has been proposed to be the mechanism that links environmental influences, such as maternal diet, to the aetiology of T2D( Reference Ling and Groop 71 , Reference Seki, Williams and Vuguin 72 ). The most widely studied mechanism is DNA methylation, which involves the reversible transfer of methyl groups from dietary folic acid to cytosine molecules of DNA( Reference Seki, Williams and Vuguin 72 ). Maternal nutrition and supplements to maternal diets, such as folic acid, may influence the DNA methylation of the offspring( Reference Lillycrop, Slater-Jefferies and Hanson 73 , Reference Lillycrop, Rodford and Garratt 74 ). Hence, the observed alterations to the hepatic fat metabolism and skeletal muscle glucose uptake in HFS offspring and IR in the HFS/F offspring may lie in the DNA methylation of genes related to insulin secretion.
Conclusions
The results of this study provide compelling evidence for the DOHaD theory of mismatch. Our investigation has shown that prenatal exposure to HFS diet significantly alters hepatic fat metabolism and skeletal muscle glucose uptake in male offspring, at both transcription and translation. In addition, supplementation of the maternal HFS diet with 5 mg of folic acid/kg diet significantly increases circulating insulin and reduces plasma glucose levels. The HFS diet predisposes offspring to IR as compared with the CON diet. The results of this study can potentially have an impact on health policy for T2D by highlighting the role of maternal nutrition in the prevention of adverse metabolic effects related to impaired insulin signalling and fat metabolism in the offspring. In addition, results from this study will help to further examine the underlying mechanisms of the developmental origins of adult health and disease.
Acknowledgements
The authors thank Dr Shamjeet Singh of the Pre-Clinical Sciences Department, Dr Jenelle Johnson of the School of Veterinary Medicine and Ms Aileen Hawke of the Guelph Research and Development Centre, Agriculture and Agri-Food Canada, for advice and technical services rendered.
This work was supported by a grant from the Caribbean Public Health Agency and Agriculture & Agri-Food Canada Project no. 1343; C. E. C was the recipient of a CARICOM-Canada Emerging Leaders Scholarship.
D. D. R., J. E. F. and C. E. C. conceived and designed the study, interpreted the data and drafted the manuscript. C. E. C. conducted the research and acquired the data and performed the statistical analysis. All authors read and approved the final manuscript.
The authors declare that there are no conflicts of interest.