AMP-activated protein kinase signalling pathways are down regulated and skeletal muscle development impaired in fetuses of obese, over-nourished sheep

Abstract

Maternal obesity and over-nutrition give rise to both obstetric problems and neonatal morbidity. The objective of this study was to evaluate effects of maternal obesity and over-nutrition on signalling of the AMP-activated protein kinase (AMPK) pathway in fetal skeletal muscle in an obese pregnant sheep model. Non-pregnant ewes were assigned to a control group (Con, fed 100% of NRC nutrient recommendations, n = 7) or obesogenic group (OB, fed 150% of National Research Council (NRC) recommendations, n = 7) diet from 60 days before to 75 days after conception (term 150 days) when fetal semitendinosus skeletal muscle (St) was sampled. OB mothers developed severe obesity accompanied by higher maternal and fetal plasma glucose and insulin levels. In fetal St, activity of phosphoinositide-3 kinase (PI3K) associated with insulin receptor substrate-1 (IRS-1) was attenuated (P < 0.05), in agreement with the increased phophorylation of IRS-1 at serine 1011. Phosphorylation of AMP-activated protein kinase (AMPK) at Thr 172, acetyl-CoA carboxylase at Ser 79, tuberous sclerosis 2 at Thr 1462 and eukaryotic translation initiation factor 4E-binding protein 1 at Thr 37/46 were reduced in OB compared to Con fetal St. No difference in energy status (AMP/ATP ratio) was observed. The expression of protein phosphatase 2C was increased in OB compared to Con fetal St. Plasma tumour necrosis factor α (TNFα) was increased in OB fetuses indicating an increased inflammatory state. Expression of peroxisome proliferator-activated receptor γ (PPARγ) was higher in OB St, indicating enhanced adipogenesis. The glutathione: glutathione disulphide ratio was also lower, showing increased oxidative stress in OB fetal St. In summary, we have demonstrated decreased signalling of the AMPK system in skeletal muscle of fetuses of OB mothers, which may play a role in altered muscle development and development of insulin resistance in the offspring.

According to the latest National Health and Nutrition Examination Survey (1999–2002), 26% of non-pregnant women 20–39 years of age are overweight and 29% are obese (Hedley et al. 2004). More importantly, the shift towards higher gestational weight gain appears evident (Siega-Riz et al. 2006), showing excessive nutrient intake during gestation in affluent countries. High energy diets combined with maternal obesity represent a special problem because of both adverse effects on maternal health and fetal development that can result in harmful, persistent effects in offspring, including predisposition to obesity and diabetes (Barker, 2002; Fowden et al. 2006; Nathanielsz, 2006). Many animal studies on malnutrition during gestation, and some just now becoming available on maternal obesity and over-nutrition during gestation, demonstrate that both a deficient and excessive maternal diet can predispose offspring to obesity and insulin resistance (Nishina et al. 2003; Symonds et al. 2003; Bispham et al. 2005; Fernandez-Twinn et al.2005, 2006; Zambrano et al. 2006; Cleal et al. 2007; Nathanielsz et al. 2007). However, cellular mechanisms linking maternal nutrition and insulin resistance and obesity of offspring remain poorly defined.

Skeletal muscle is the main peripheral tissue responsible for glucose and fatty acid oxidation (Selak et al. 2003; Lowell & Shulman, 2005; Ozanne et al. 2005). The fetal period is crucial for skeletal muscle development, because no net increase in the number of muscle fibres occurs after birth (Greenwood et al. 2000; Nissen et al. 2003). Impaired fetal skeletal muscle growth will impair the metabolism of glucose and fatty acids in response to insulin stimulation and, thus, predispose offspring to diabetes and obesity in later life (Stannard & Johnson, 2004; Zambrano et al. 2005). On the other hand, enhanced muscle growth renders animals resistant to diet-induced obesity and insulin resistance (Zhao et al. 2005; Yang & Zhao, 2006). Insulin-like growth factor-1 (IGF-1) and insulin share a common signalling pathway which stimulates skeletal muscle growth and development, mainly through the phosphoinositide-3 kinase/protein kinase B (PI3K/Akt) signalling pathway via insulin receptor substrate-1 (IRS-1) (Bush et al. 2003; Latres et al. 2005; Park et al. 2005; Song et al. 2005; Subramaniam et al. 2005; Vary, 2006). Activated Akt enhances skeletal muscle growth and development by enhancing protein synthesis and inhibiting protein degradation (Lai et al. 2004).

AMPK is a heterotrimeric enzyme with α, β and γ subunits (Hardie, 2004; Kim et al. 2004). The α subunit is the catalytic unit, the γ subunit has a regulatory function, and the β subunit provides anchorage sites for the α and γ subunits (Sambandam & Lopaschuk, 2003). AMPK is switched on by an increase in the AMP/ATP ratio, leading to phosphorylation of AMPK at Thr 172 by AMPK kinases (Hardie, 2004; Kim et al. 2004). Once activated, AMPK promotes glucose uptake and fatty acid oxidation, and inhibits lipid synthesis in cells (Hardie & Hawley, 2001; Fujii et al. 2006). AMPK mediates IGF-1/insulin signalling through several possible mechanisms (Jakobsen et al. 2001; Steinberg et al. 2006). Inhibition of AMPK promotes lipogenesis and adipogenesis (Giri et al. 2006).

Obesity commonly leads to a low degree inflammation response (Steinberg, 2007), indicated by increased production of tumour necrosis factor-α (TNFα). Recently, TNFα has been shown to reduce AMPK activity in skeletal muscle, associated with up-regulation of protein phosphatase 2C, an enzyme leading to the dephosphorylation of AMPK at Thr 172 (Steinberg et al. 2006). The pregnant sheep model has been extensively studied to evaluate the regulation of fetal development and gain insight into problems of human pregnancy (Anderson et al. 2001; Anthony et al. 2003; Elmes et al. 2004; Gentili et al. 2006). The effect of maternal over-nutrition and obesity on AMPK activity and fetal muscle development has not been examined in this model. We hypothesized that maternal over-nutrition in the pregnant ewe would inhibit the AMPK signalling pathways in fetal skeletal muscle as a result of the provision of excessive energy (ATP) for cells, increase fetal TNFα and the level of oxidative stress, and affect fetal muscle development.

Methods

Care and use of animals

All animal procedures were approved by the University of Wyoming Animal Care and Use Committee. From 60 days before conception to day 75 of gestation (day of mating, day 0; term 148 days), multiparous Rambouillet/Columbia ewes were fed either a highly palatable diet at 100% (control, Con) of National Research Council (NRC) recommendations or 150% (obesogenic, OB) of NRC recommendations on a metabolic body weight (BW) basis (BW^0.75) (Table 1). All ewes were weighed at weekly intervals and rations adjusted for weight gain, and body condition was scored at monthly intervals to evaluate changes in fatness. A body condition score of 1 (emaciated) to 9 (obese) was assigned by two trained observers after palpation of the transverse and vertical processes of the lumbar vertebrae (L2 to L5) and the region around the tail head (Sanson et al. 1993).

Table 1.

Composition of the diet fed to ewes throughout the study

Ingredients	(%)	Analysed composition4
Ground bromegrass hay1	14.02	DM (%)	88.54
Ground corn	63.89	NDF (% of DM)	24.09
Soybean meal	13.30	ADF (% of DM)	9.99
Liquid molasses	5.60	CP (% of DM)	17.39
Limestone	2.24	IVDMD (%)	93.92
Ammonium chloride	0.50
Mineralized salt2	0.24
Magnesium chloride	0.10
ADE premix3	0.10
Rumensin 80	0.02

¹Mean particle length = 2.54 cm.

²Contained 13% NaCl, 10% Ca, 10% P, 2% K, 1.5% Mg, 0.28% Fe, 0.27% Zn, 0.12% Mn, 0.01% I, 35 p.p.m. Se, and 20 p.p.m. Co.

³Contained 110 000 IU kg⁻¹ vitamin A, 27 500 IU kg⁻¹ vitamin D and 660 IU kg⁻¹ vitamin E.

⁴DM, dry matter; NDF, neutral detergent fibre; ADF, acid detergent fibre; CP, crude protein; IVDMD, in vitro dry matter digestibility.

Following a 12 h over-night fast on day 75 of gestation (term on day 150 of gestation), 14 pregnant sheep (7 Con and 7 OB) were weighed. Sedation was induced by intravenous ketamine (10 mg kg⁻¹) and anaesthesia was induced and maintained by isoflurane inhalation. A sample of maternal blood was collected via jugular venipuncture into a chilled non-heparinized vacutainer tube (no additives, Sigma, St Louis, MO, USA). Serum obtained was frozen at −80°C until assayed for insulin and TNFα. Blood was also collected in a separate chilled tube (heparin plus sodium fluoride; 2.5 mg ml⁻¹; Sigma), and plasma was frozen at −80°C until assayed for glucose. The abdomen and uterus were opened and fetal blood was collected from the umbilical vein via a 20 gauge needle and 3 ml syringe, and serum and plasma were collected and stored as described for maternal blood. After that, fetuses were quickly removed to obtain weight and length and the semitendenosus (St) muscles on both sides were collected and snap-frozen in liquid nitrogen for biological analyses. Fetal muscle collected from five ewes carrying twin pregnancy in each group was randomly selected for analyses. No difference in body weight was observed between twins and thus only one fetus of the twin pregnancy was selected at random for analyses. Though no difference in weight was observed among fetuses of different sexes, the sex of fetuses in each group was balanced.

Sheep were killed by administration of an overdose of sodium pentobarbital (Abbott Laboratories, Abbott Park, IL, USA) and exsanguinated. The St muscle from the left side of each pregnant sheep was dissected immediately and weighed. After trimming off all visible adipose and connective tissues, a small piece of muscle (1 g) was sampled in the anatomical centre of the muscle and snap-frozen in liquid nitrogen for biological analyses.

Antibodies

Antibodies against AMPK, phospho-AMPK at Thr 172, acetyl-CoA carboxylase (ACC), phospho-ACC at Ser 79, IRS-1, phospho-IRS-1 at Ser 1101, Akt, phospho-Akt at Ser 473, mTOR, phospho-mTOR at Ser 2448 l, phospho- tuberous sclerosis 2 (TSC2) at Thr 1462, eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) at Thr 37/46, peroxisome proliferator-activated receptor γ (PPARγ), and horseradish peroxidase linked secondary antibody were purchased from Cell Signalling (Danvers, MA, USA). Protein phosphatase 2C (PP2C) antibody was purchased from Epigenomics (Berlin, Germany). Cytochrome C antibody was purchased from EMD Chem. Inc. (Gibbstown, NJ, USA). Anti-β-actin antibody was obtained from Developmental Studies Hybridoma Bank (DSHB, Iowa City, IA, USA).

Immunoblotting

St muscle (0.1 g) was powdered in liquid nitrogen and homogenized in a polytron homogenizer (7 mm dia. generator) with 400 μl of ice-cold buffer containing 137 mm NaCl, 50 mm Hepes, 2% SDS, 1% NP-40, 10% glycerol, 2 mm PMSF, 10 mm sodium pyrophosphate, 10 μg ml⁻¹ aprotinin, 10 μg ml⁻¹ leupeptin, 2 mm Na₃VO₄ and 100 mm NaF, pH 7.4. The protein content of lysates was determined by the Bradford method (Bio-Rad Laboratories, Hercules, CA, USA) (Zhu et al. 2006) and used for immunoblotting analyses as previously described (Zhu et al. 2004).

IRS-1 associated PI3K activity analyses

The IRS-1-associated PI3K activity was analysed by immunoprecipitation with anti-IRS-1 antibody (Kirwan et al. 2000). Briefly, St muscles were lysed in a buffer containing 50 mm Hepes pH 7.4, 1% NP-40 or Triton-X 100, 10% glycerol, 2 mm Na₃VO₄, 100 mm NaF, 1% protease inhibitor cocktail, 1 mm MgCl₂, 1 mm CaCl₂ and 2.5 mm EDTA. A 1 mg sample of total protein was immunoprecipitated with 4 μg of IRS-1 antibody and 40 μl protein A–Sepharose slurry. The pellet was suspended in an assay buffer containing 0.1 mm ATP, 50 mm Hepes, pH 7.1, 1 mm EGTA, 20 mm MgCl₂, 20 mm NaCl, 1 mm NaH₂PO₄, 0.2 m NaH₂CO₃ and 5 μCi of ³²P-γ ATP (GE Healthcare, Piscataway, NJ, USA). l-Phosphatidylinositol (10 μg each⁻¹ reaction) was used as a substrate. The reaction mixture was incubated at 30°C for 10 min. The reaction was stopped by adding 1 n HCl. The phosphatidylinositol 3-phosphate was extracted with chloroform: methanol (1: 1). The lower phase containing chloroform and extracted lipids was used for separation by thin layer chromatography. The radioactivity incorporated into phosphatydylinositol 3-phosphate was determined by autoradiography (Kirwan et al. 2000).

Adenine nucleotides, creatine and phosphocreatine analysis

St muscle homogenates (200 μl) obtained as described above were added to 55% perchloric acid to a final concentration of 5%, set for 30 min on ice, and then neutralized to pH 6–8 with 2 m KOH. After centrifugation at 13 000 g, 4°C for 10 min to remove KClO₄, the supernatant was passed through a 0.2 μm filter prior to HPLC analysis (Beckman Instruments, Inc., Fullerton, CA, USA). HPLC conditions were the same as previously reported (Shen et al. 2006). Lysate protein contents were determined using a protein assay kit (Bio-Rad). Adenine nucleotides, creatine and phosphocreatine contents are expressed as micromoles per gram of muscle.

Glucose, insulin and TNFα analyses

Glucose was analysed using the Infinity^. (ThermoTrace Ltd, Cat. no. TR15498; Melbourne, Australia) colourimetric assay modified in the following manner: plasma was diluted 1: 5 in distilled water, and 10 μl of diluted plasma was added to 300 μl reagent mix. Insulin was measured by radioimmuno assay in accordance with the manufacturer's recommendations (Coat-A-Count, Diagnostic Products Corporation, Los Angeles, CA, USA). TNFα was analysed by using an ELISA kit in accordance with manufacturer's recommendations (Endrogen, Cat. no. ESS0011B; Rockford, IL, USA). All samples were run in triplicate.

Glutathione: glutathione disulphide (GSH: GSSG) ratio

St muscle samples were homogenized in 4 volumes (w/v) of 1% picric acid. Acid homogenates were centrifuged at 16 000 g (30 min) and supernatant fractions collected. Supernatant fractions were assayed for total GSH and GSSG by the standard recycling method and GSH content was determined using a standard curve generated from known concentrations of GSH. The procedure consisted of using one-half of each sample for GSSG determination and the other half for GSH. Samples for GSSG determination were incubated at room temperature with 2 μl of 4-vinyl pyridine per 100 μl sample for 1 h after vigorous vortexing. Incubation with 4-vinyl pyridine conjugates any GSH present in the sample so that only GSSG is recycled to GSH without interference by GSH. The GSSG (as GSH x2) was then subtracted from the total GSH to calculate the ratio of GSH: GSSG (Li et al. 2005).

Real-time quantitative PCR (RT-PCR)

mRNA was extracted from the fetal St muscle using TRI reagent (Sigma) and reverse transcribed into cDNA using a kit (Qiagen, Valencia, CA, USA). Reversed transcribed cDNAs were used for real-time PCR analyses by using a SYBR Green RT-PCR kit from Bio-Rad (Hercules). Primer sets used were: PPPARγ forward, 5′-CCGCATCTTCCAGGGGTGTC-3′; and reverse, 5′-CAAGGAGGCCAGCATCGTGAAAT-3′. The 18S RNA was used as a control, forward, 5′-GTAACCCGTTGAACCCCATT-3′; and reverse, 5′-CCATCCAATCGGTAGTAGCG-3′. PCR conditions were as follows: 20 s at 95°C, 20 s at 56°C, and 20 s at 72°C for 35 cycles (Lomax et al. 2007). After amplification, a melting curve (0.01°C s⁻¹) was used to confirm product purity. Data were expressed relative to 18S rRNA as previously described (Lomax et al. 2007).

Histochemical analyses

St muscle samples frozen in OCT compound (Tissue-Tek, Sakura Finetek USA, Inc., Torrance, CA, USA) were cut into 10 μm sections. Sections were stained with haematoxylin–eosin for standard light microscopy. The number of primary and secondary myofibres were counted in 10 different microscopic fields at 200 × magnification of each section and 10 sections (every fifth section) were viewed for each sample. The ratio of secondary to primary myofibres was reported. Only those fasciculi where primary and secondary myofibres were clearly differentiated were counted (Zhu et al. 2004). In addition, the relative density of muscle fibres was calculated by counting all muscle fibres in each microscopic view.

Statistical analysis

Statistical analyses were conducted according to our previous studies in sheep (Zhu et al. 2004, 2006). Briefly, each animal was considered as an experimental unit. Data were analysed as a complete randomized design using GLM (General Linear Model of Statistical Analysis System, SAS, 2000). The differences in the mean values were compared by Tukey's multiple comparison, and mean ± standard errors of mean (s.e.m.) were reported. Statistical significance was considered as P < 0.05.

Results

Fetal and maternal phenotypic changes, and serum glucose and insulin levels

There was no difference in maternal body weight and body condition score at the beginning of treatments (Table 2). However, by the end of treatment, the body weight and body condition score were much higher in OB sheep compared to Con sheep (Table 2), showing that OB sheep developed severe obesity. Maternal plasma glucose and insulin concentrations were elevated (P < 0.05) in OB versus Con ewes (Table 2).

Table 2.

Body condition score, body weight, serum IGF-1, insulin and glucose, and semitendenosus (St) muscle weight of OB and Con ewes and fetal sheep at 75 days gestation

Category	Con	OB	P value
Maternal
At the beginning of treatments
Body condition score	4.9 ± 0.4	5.0 ± 0.3
Body weight (kg)	68.3 ± 2.9	71.6 ± 3.2
At the end of treatments
Body condition score	4.7 ± 0.4	8.0 ± 0.2	**
Body weight (kg)	72.2 ± 3.3	102.2 ± 2.4	**
Insulin (μIU ml−1)	4.8 ± 1.6	25.0 ± 4.0	**
Glucose (mg dl−1)	52.1 ± 3.4	70.2 ± 6.3	*
St muscle (g)	200.5 ± 22.4	267.2 ± 13.3	*
Fetal
Body weight (g)	268 ± 12	374 ± 10	*
Crown rump length (cm)	22.1 ± 0.2	24.6 ± 0.2	*
Glucose (mg dl−1)	25.35 ± 2.10	43.63 ± 5.58	*
IGF-1 (ng ml−1)	44.6 ± 0.9	53.3 ± 1.8	*
Insulin (IU ml−1)	1.14 ± 0.35	7.29 ± 1.31	**
St muscle (g)	0.49 ± 0.01	0.65 ± 0.05	*

Means ± s.e.m.n = 7 for maternal data and n = 5 for fetal data.

^**P < 0.01

^*P < 0.05.

Plasma glucose, IGF-1 and insulin concentrations were elevated in fetuses of OB versus Con ewes. Body weight and crown rump length of fetuses of OB ewes were ∼30% greater than fetuses from Con sheep, as well as fetal muscle weight (Table 2).

Akt and mTOR signalling pathway activity in fetal St muscle

Phosphorylation of Akt at Ser 473, which is correlated with its activation, was reduced by 41.2% and total Akt content increased in the fetal muscle of OB sheep (Fig. 1A). TSC2 is a mediator between Akt and mTOR. Its phosphorylation at Thr 1462 was reduced by 27.5% (Fig. 1B). mTOR phosphorylation at Ser 2448 was reduced 29.1% in OB fetal muscle without alteration in its total content (Fig. 1C). mTOR signalling is a key signalling pathway controlling protein synthesis. It controls protein synthesis through phosphorylation of 4E-BP1 and S6K. The phosphorylation of 4E-BP1 at Thr 37/46 was also reduced (Fig. 1D).

Figure 1. Immunoblots and combined data for Akt, mTOR and their phosphorylation in fetal St muscle of Con (open bars) and OB (filled bars) sheep.

A, Akt and phospho-Akt (p-Akt) immunoblots of fetal muscle; B, TSC2 and phospho-TSC2 (p-TSC2) immunoblots of fetal muscle; C, mTOR and phospho-mTOR (p-mTOR) immunoblots of fetal muscle; D, 4E-BP1 immunoblots of fetal muscle. *P < 0.05. Mean ± s.e.m.; n = 5.

AMPK and ACC phosphorylation in fetal St skeletal muscle

Phosphorylation of AMPK α subunit at Thr 172 and its down-stream effector ACC at Ser 79 are frequently used to assess AMPK activity. AMPK phosphorylation at Thr 172, which is correlated with its activity, was down-regulated in fetal St muscle of OB sheep compared to Con sheep (14.9%, Fig. 2A). No change in the protein expression of AMPK α subunits was observed. Phosphorylation of ACC at Ser 79 is directly catalysed by AMPK and, thus, its phosphorylation correlates with AMPK activity. ACC phosphorylation was also reduced (18.9%) in fetal St muscle of OB sheep, in the absence of any change in total ACC concentration (Fig. 2B).

Figure 2. Immunoblots and combined data for AMPK, ACC and their phosphorylation in fetal St muscle of Con (open bars) and OB (filled bars) sheep.

A, AMPK and phospho-AMPK (p-AMPK) immunoblots of fetal muscle; B, acetyl-CoA carboxylase (ACC) and phospho-ACC (p-ACC) immunoblots of fetal muscle. *P < 0.05; **P < 0.01. Mean ± s.e.m.; n = 5.

IRS-1 associated PI3K activity in fetal skeletal muscle

IRS-1 is a key mediator of IGF-1/insulin signalling. IRS-1 regulates IGF-1/insulin signalling through altering its ability to recruit PI3K. The content of IRS-1 was not altered but the phosphorylation of IRS-1 at Ser 1101 (this phosphorylation included both IRS-1 and IRS-2 since this antibody cross-reacts with IRS-2) was increased (93.6%) in fetal St muscle of OB sheep compared to Con sheep (Fig. 3A). Consistent with this increase in serine phosphorylation of IRS-1, the PI3K activity associated with IRS-1 was lower in fetal St muscle of OB sheep compared to Con sheep (Fig. 3B).

Figure 3. Immunoblots and combined data for IRS-1 and its phosphorylation, and IRS-1-associated PI3K activity in fetal St muscle of Con (open bars) and OB (filled bars) sheep.

A, IRS-1 and phospho-IRS-1 immunoblots of fetal muscle; B, PI3K activity associated with IRS-1 in fetal muscle measured by immunoprecipitation. *P < 0.05. Mean ± s.e.m.; n = 5.

Adenosine nucleotides, creatine and phosphocreatine contents in fetal St skeletal muscle

There were no differences in ATP, ADP, AMP, creatine and phosphocreatine contents between Con and OB sheep for fetal muscle, indicating that AMPK inhibition was not due to alteration in energy status (Table 3). No differences in creatine and phosphocreatine contents were observed (Table 3).

Table 3.

ATP, AMP, ADP, creatine and phosphocreatine concentrations in fetal semitendinosus muscle of OB and Con sheep at 75 days gestation

	Con	OB
ATP (μmol (g muscle)−1)	0.36 ± 0.09	0.38 ± 0.08
ADP (μmol (g muscle)−1)	0.42 ± 0.06	0.51 ± 0.03
AMP (μmol (g muscle)−1)	0.48 ± 0.04	0.53 ± 0.04
AMP/ATP ratio	1.87 ± 0.39	1.63 ± 0.50
Creatine phosphate (μmol (g muscle)−1)	7.38 ± 0.28	7.37 ± 0.36
Creatine (μmol (g muscle)−1)	33.46 ± 1.26	35.28 ± 1.10
Creatine/creatine phosphate ratio	4.55 ± 0.13	4.83 ± 0.14

Data are presented as means ± s.e.m. No differences were observed. n = 5.

TNFα content in the fetal circulation and oxidative stress in fetal muscle

Obesity leads to low-grade inflammation. As a marker of inflammation, TNFα leads to AMPK inhibition. To identify whether TNFα was a reason for AMPK inhibition, its concentration was measured in the fetal circulation. The TNFα content was higher in OB fetuses compared to Con (Fig. 4A). In agreement with this increase in TNFα content, the expression of PP2C was enhanced in OB fetal St muscle compared to Con muscle (Fig. 4B).

Figure 4. TNFα concentration in fetal plasma and expression of PP2C in fetal St muscle of Con (open bars) and OB (filled bars) sheep.

A, TNFα concentration in fetal plasma; B, PP2C immunoblot of fetal St muscle. *P < 0.05. Mean ± s.e.m.; n = 5.

The ratio of GSH: GSSG was lower in OB fetal muscle, indicating that OB fetal skeletal muscle experienced oxidative stress compared to Con fetal muscle (Fig. 5).

Figure 5. Glutathione: glutathione disulphide (GSH: GSSG) ratio in fetal muscle of Con (open bars) and OB (filled bars) fetal St muscle.

*P < 0.05. Mean ± s.e.m.; n = 5.

Fetal skeletal muscle development

To test whether the down-regulation of AMPK signalling in the presence of increased plasma levels of fetal IGF-1 and insulin affected fetal muscle development, sections of fetal St muscle were stained with haematoxylin and eosin. Both primary fibres with centrally located nuclei and secondary muscle fibres with peripherally located nuclei were clearly visible forming primary muscle bundles (Fig. 6A and B). No difference in the ratio of primary to secondary muscle fibres was observed (Fig. 6C). However, the density of muscle fibres per area was significantly lower in OB fetal muscle compared to Con fetal muscle (Fig. 6D).

Figure 6. Hematoxylin and eosin staining of fetal St muscle.

A, Con fetal muscle; B, OB fetal muscle; C, secondary to primary muscle fibre ratio; D, density of muscle fibres. *P < 0.05. Mean ± s.e.m.; n = 5.

The content and mRNA expression of PPARγ were enhanced in OB fetal muscle compared to Con fetal muscle (Fig. 7), indicating enhanced adipogenesis.

Figure 7. PPARγ content and mRNA expression in fetal St muscle of Con (open bars) and OB (filled bars) sheep.

A, PPARγ immunoblot and statistical data; B, PPARγ mRNA expression measured by RT-PCR. *P < 0.05. Mean ± s.e.m.; n = 5.

Discussion

Value of pregnant sheep as a model for studying fetal development as affected by maternal obesity and over-nutrition

Obesity and over-nutrition pose an ever increasing health threat to pregnant women and their fetuses (Catalano & Ehrenberg, 2006; Siega-Riz et al. 2006) and are associated with several obstetric problems as well as an increased risk of obesity and diabetes in offspring (Warram et al. 1990; Petersen et al. 2004; Petersen et al. 2005). However, the associated mechanisms are unclear. Most of the recent studies on the effects of maternal over-nutrition on fetal development have been conducted in rodents (Armitage et al. 2005; Reusens et al. 2007). Rodent studies provide data of great value; however, it must always be born in mind that rodents are polytocous animals with a large fetal and placental biomass relative to maternal biomass with important nutritional consequences. In addition, offspring are born at a very altricial level without having completed many stages of development that are completed in utero in precocial mammals such as sheep or primates. In addition, rodents tend to decrease food consumption when fed high-energy-containing diets. To avoid this problem, obese rat and mouse strains (carrying a mutation in leptin or leptin receptor which impairs satiety control), such as Zucker rat carrying a leptin receptor missense mutation (Phillips et al. 1996) and ob/ob mice with leptin deficiency (Pelleymounter et al. 1995), are frequently used in obesity studies. However, leptin is a known crucial mediator of placental growth (Szczepankiewicz et al. 2006; Lepercq et al. 2007; O'Connor et al. 2007) and extensively involved in regulation of metabolism in pregnancy.

The pregnant sheep has been extensively studied to evaluate the effects of decreased or increased nutritional supply to the fetus and the effects on fetal growth and development and pregnancy outcomes (Battaglia & Meschia, 1986; DiGiacomo & Hay, 1990; Hay, 1995; Anderson et al. 2001, 2005; Anthony et al. 2003; Elmes et al. 2004; Manikkam et al. 2004; Gentili et al. 2006; Zhu et al. 2007; Muhlhausler et al. 2007; Steckler et al. 2007; Taylor & Poston, 2007). However, no fetal studies have been reported in ewes made obese prior to pregnancy. We present here data from sheep made obese prior to pregnancy and maintained at an elevated body weight during pregnancy.

The body condition score and body weight of OB sheep were much higher than those in Con sheep. Plasma insulin and glucose levels in both maternal and fetal circulations were increased in OB sheep. Furthermore, we recently determined that OB sheep on a dietary protocol identical to that utilized in this study exhibited marked insulin resistance and glucose intolerance to a bolus i.v. injection of glucose on day 75 of gestation (SP Ford, MM Miller, MJ Zhu, L Zhang, BW Hess, GE Moss & PW Nathanielsz, unpublished observations). Fetal weight and crown rump length were increased in OB ewes, showing enhanced fetal growth. These general data show that this model has potential for studying the negative physiological consequences of maternal obesity and over-nutrition on the fetus.

Effect of maternal obesity on the AMPK and m-TOR signalling pathways

The phosphoinositide-3 kinase/protein kinase B (PI3K/Akt) and AMPK signalling pathways interact to control the activation of mTOR signalling and play a central role in cell energetics and protein synthesis. The down-stream targets of mTOR signalling are proteins involved in mRNA translation, including 4E-BP1 and S6K.

IRS-1 plays a key regulatory role at the entrance to the mTOR pathway. We observed that PI3K activity associated with IRS-1 was reduced in OB fetal skeletal muscle. To determine potential mechanisms for the functional impairment of IRS-1, we analysed AMPK activity as assessed by AMPK phosphorylation at Thr 172 and ACC phosphorylation at Ser 79, a method widely used by other researchers (Hutchinson & Bengtsson, 2006; Han et al. 2007; Tanaka et al. 2007). In a previous study, we have shown that phosphorylation of AMPK and ACC is highly correlated with AMPK activity (Shen et al. 2007), though there are reports that phosphorylation status and measured activity were not well correlated (Richter et al. 2004).

AMPK activation sensitizes insulin/IGF-1 signalling through several possible mechanisms. Activation of AMPK phosphorylates IRS-1 at Ser 789 which enhances IGF/insulin signalling (Jakobsen et al. 2001), though this claim has been disputed (Qiao et al. 2002). In addition, AMPK promotes lipid oxidation and inhibits lipid synthesis, thereby reducing cellular fatty acyl-CoA content and protein kinase C (PKC) activity, as a result of which IRS-1 function and insulin sensitivity are improved (Steinberg et al. 2006). On the other hand, AMPK inhibition led to the accumulation of lipids in cells, impairing IRS-1 function. Indeed, the phosphorylation of IRS-1 at Ser 1101, a site phosphorylated by PKC, was enhanced in OB fetal muscle compared to Con muscle. Furthermore, it has been reported that AMPK inhibits mTOR by phosphorylation of TSC2 at Thr 1227 and Ser 1345 (Inoki et al. 2003). Since activated mTOR phosphorylates IRS-1 at Ser 636/639, which negatively affects IRS-1 function and down-regulates insulin signalling (Um et al. 2004; Khamzina et al. 2005), the down-regulation of AMPK should activate mTOR, which de-sensitizes insulin signalling. However, in this study, both mTOR and AMPK phosphorylation were lower in OB compared to Con fetal muscle. This down-regulation of mTOR in OB fetal muscle may well be due to the down-regulation of Akt, its main upstream kinase in OB fetal muscle. The resulting down-regulation of mTOR may be only partially compensated by the inhibition of AMPK, resulting in an overall inhibition of mTOR signalling in OB fetal muscle. In support of the above notions, activation of AMPK pharmacologically reverses insulin resistance (Fisher et al. 2002; Inoki et al. 2003; Bandyopadhyay et al. 2006). Knockout of the AMPKα2 catalytic subunit in mice leads to insulin resistance (Viollet et al. 2003). Therefore, the reduction of AMPK activity we observed in fetal muscle in OB sheep should provide the mechanism leading to IRS-1 function impairment and the down-regulation of IGF-1/insulin down-stream signalling. Inhibition of AMPK in mature skeletal muscle has been observed in high-fat-fed rats (Liu et al. 2006; Ye et al. 2006). However, this is the first demonstration that AMPK activity is inhibited in fetal skeletal muscle at mid-gestation in the presence of maternal obesity and over-nutrition.

Mechanisms leading to AMPK inhibition in fetal skeletal muscle of OB sheep

Low cellular energy status as indicated by low ATP/AMP and phosphocreatine/creatine ratios activates AMPK. By providing excessive nutrients to cells, over-nutrition should enhance cellular energy level. However, no difference was observed in ATP/AMP and phosphocreatine/creatine ratios between OB and Con sheep. These data show that the AMPK inhibition in the OB fetal skeletal muscle was probably due to changes in other regulatory systems than a decreased energy level in the fetal muscle cells.

Recently, hormones/cytokines have been shown to act as mediators of AMPK activity. Leptin, a hormone secreted by adipocytes, increases AMPK activity through alteration of the ATP/AMP ratio in skeletal muscle (Minokoshi et al. 2002). Since no significant alteration in the ATP/AMP ratio was observed, leptin seems unlikely to be responsible for the inhibition of AMPK activity in fetal and maternal muscle of OB compared to Con sheep. In addition, we have shown that fetal leptin levels are below the sensitivity of available assays in both the CON and OB fetuses (MJ Zhu, M Du, SP Ford, PW Nathanielsz, unpublished data).

Obesity commonly leads to a low-grade inflammation response (Steinberg, 2007), which is associated with the increased production of TNFα. TNFα reduces AMPK activity in skeletal muscle through up-regulation of PP2C, an enzyme dephosphorylating Thr 172 at the AMPK α subunit (Steinberg et al. 2006; Shen et al. 2007). Our data show that maternal obesity and over-nutrition increase TNFα levels in the fetal circulation and the expression of PP2C was increased in OB fetal muscle compared to Con muscle. A high level of TNFα is known to induce IGF-1/insulin resistance in skeletal muscle (Peraldi & Spiegelman, 1998), which is linked to AMPK inhibition (Steinberg et al. 2006; Li et al. 2007), consistent with our observations. Therefore, the high level TNFα we have demonstrated in the fetal circulation may play a role in the decreased signalling of the AMPK pathway we have shown and would also tend to increase insulin resistance in fetal skeletal muscle. In addition, we observed increased oxidative stress in OB fetal muscle, which is consistent with an inflammation response indicated by high TNFα in fetal circulation.

Mitochondrial function is associated with insulin sensitivity and AMPK is known to increase mitochondrial protein expression in skeletal muscle. To assess whether the inhibition of AMPK led to the alteration in mitochondrial protein expression, we analysed the expression of cytochrome C, a protein in mitochondria. However, no significant difference was detected between treatments (Con versus OB, 1.00 ± 0.10 versus 0.91 ± 0.09 arbitrary units). The possible reason could be due to the low mitochondrial content in fetal muscle compared to adult muscle (Lee et al. 2005). It will be interesting to observe long-term effects of such changes on the mitochondrial function of adult muscle.

Fetal muscle development due to obesity

During prenatal muscle development, primary myofibres are first formed, followed by the formation of secondary myofibres (Beermann et al. 1978). Primary myofibres have centrally located nuclei and larger diameters while secondary muscle fibres have peripherally located nuclei and small sizes (Swatland, 1973; Beermann, 1978). The secondary myofibres are derived from muscle precursor cells which are initially maintained in a proliferating, undifferentiated state (Swatland, 1973). Those precursor cells differentiate into myoblasts and fuse to form secondary myofibres parallel to primary myofibres (Beermann, 1978). The number of secondary fibres present in prenatal muscles is susceptible to maternal nutrients (Zhu et al. 2004). In this study, maternal obesity and over-nutrition did not affect the ratio of secondary to primary muscle fibres halfway through gestation. Maternal nutrient deficiency reduces muscle fibre numbers and secondary to primary ratio (Ward & Stickland, 1991; Dwyer et al. 1994; Zhu et al. 2004). To our knowledge, the effect of maternal obesity and over-nutrition on the ratio of secondary to primary muscle fibres has not been examined previously. In this study, we observed that the density of muscle fibres was reduced in OB fetuses compared to Con fetuses. The reason for the reduced muscle fibre density is unclear, but may be associated with an increase in adipogenesis (diversion from myogenesis to adipogenesis) and lipid accumulation. The expression of PPARγ, a marker of adipogenesis, was higher in OB fetal muscle compared to Con fetal muscle. Though PPARγ is also expressed in skeletal muscle, its level is very low (Vidal-Puig et al. 1996; Verma et al. 2004). Therefore, the elevated PPARγ expression we have observed probably relates to enhanced adipogenesis in adipocytes between the muscle cells. Another possible reason is due to inflammation. Inflammation in fetal muscle due to maternal obesity might contribute to the accumulation of fluid and other extracellular substrates, resulting in the reduction in muscle fibre density. These data showed that maternal obesity and over-nutrition affected fetal muscle development.

Adipocytes existing inside muscle bundles may secrete adipokines, including TNFα, that would have paracrine effects on muscle insulin signalling leading to insulin resistance (Sell et al. 2006). In vivo mechanisms controlling adipogenesis from pluripotent cells within fetal muscle fibres are poorly defined, though there are numerous in vitro cell culture studies (Rosen & MacDougald, 2006) that demonstrate that PPARγ and CCAAT-enhancer-binding proteins (C/EBPs) are crucial intracellular factors controlling adipogenesis. Their expression leads to adipogenesis from pluripotent cells. We observed that the expression of PPARγ was enhanced at both the protein and message level in fetal muscle due to maternal obesity and over-nutrition. Activation of AMPK inhibits adipogenesis in 3T3L1 cells and also in obese mice (Giri et al. 2006). The enhancement of adipogenesis may also relate to the inflammatory response. An altered cytokine environment has been shown to direct cells to an adipogenic phenotype during muscle regeneration in adult mice (Contreras-Shannon et al. 2007). A similar shift from myogenesis to adipogenesis may have potential impacts on the functionality of muscle in offspring by increased adipokine effects on muscle function.

The effect of maternal obesity and over-nutrition on the insulin/IGF-1 signalling in ovine fetal muscle has not been examined previously. Here, we observed that the down-stream signalling of insulin/IGF-1 signalling pathways were attenuated in OB fetal muscle, which impacts on fetal muscle development.

In order to identify mechanisms leading to down-regulation of the down-stream insulin/IGF-1 signalling, we examined the function of IRS-1, which is a key site mediating insulin/IGF-1 signalling through its ability to recruit PI3K (Shao et al. 2002). In agreement with the down-regulation of insulin/IGF-1 down-stream signalling pathways observed in OB sheep, the PI3K activity associated with IRS-1 was reduced in fetal St muscle. Therefore, this attenuation of insulin/IGF-1 down-stream signalling in OB sheep at least partially occurs due to the functional impairment of IRS-1.

The observation that both IGF-I and insulin were elevated in the fetal circulation of the OB fetuses in the presence of decreased phosphorylation of key components of the AMPK and mTOR signalling pathways indicates the existence of resistance to the action of these two potent growth factors. Rats feeding on a high-fat diet developed insulin/IGF-1 resistance in skeletal muscle (Guo & Zhou, 2004). The proximate causes of the observed changes in fetal skeletal muscle remain to be determined.

In conclusion, our data indicate that pregnant sheep feeding on 150% of nutrient requirements provide a good model for studying the physiological consequences on fetal skeletal muscle development of obese pregnancy. Signalling in the AMPK and mTOR pathways was down-regulated in fetal St muscle of OB sheep despite high plasma IGF-1/insulin levels. This down-regulation of AMPK activity in OB fetal skeletal muscle was not due to an alteration in energy status (ATP/AMP) in muscle. Increased TNFα, an indicator of inflammation associated with obesity, may be a major factor responsible for the down-regulation of AMPK. We hypothesize that the changes described may play a role in the insulin resistant phenotype that has been demonstrated in the offspring of obese mothers.