Altered skeletal muscle insulin signaling and mitochondrial complex II-III linked activity in adult offspring of obese mice

Abstract

We recently reported insulin resistance in adult offspring of obese C57BL/6J mice. We have now evaluated whether parameters of skeletal muscle structure and function may play a role in insulin resistance in this model of developmental programming. Obesity was induced in female mice by feeding a highly palatable sugar and fat-rich diet for 6 wk prior to pregnancy, and during pregnancy and lactation. Offspring of obese dams were weaned onto standard laboratory chow. At 3 mo of age, skeletal muscle insulin signaling protein expression, mitochondrial electron transport chain activity (ETC), muscle fiber type, fiber density, and fiber cross-sectional area were compared with that of offspring of control dams weaned onto the chow diet. Female offspring of obese dams demonstrated decreased skeletal muscle expression of p110β, the catalytic subunit of PI3K (P < 0.01), as well as reduced Akt phosphorylation at Serine residue 473 compared with control offspring. Male offspring of obese dams demonstrated increased skeletal muscle Akt2 and PKCζ expression (P < 0.01; P < 0.001, respectively). A decrease in mitochondrial-linked complex II-III was observed in male offspring of obese dams (P < 0.01), which was unrelated to CoQ deficiency. This was not observed in females. There were no differences in muscle fiber density between offspring of obese dams and control offspring in either sex. Sex-related alterations in key insulin-signaling proteins and in mitochondrial ETC may contribute to a state of insulin resistance in offspring of obese mice.

the world health organization recognizes obesity as a global pandemic. Of particular concern is the rise in obesity in pregnant women (1, 12, 22, 43), which not only can lead to immediate complications for both mother and infant during pregnancy, but may also have adverse long-term health consequences for the child, including increased risk of developing type 2 diabetes mellitus (T2DM), metabolic syndrome, hypertension, and obesity (8, 11, 39). These observations formed the basis for the developmental “overnutrition hypothesis”. In animal models in which dams have been fed a fat-rich diet, sex differences are frequently observed in the offspring phenotype (17, 24, 25, 35). The molecular mechanisms by which maternal obesity alters the long-term metabolic health of offspring in a sex-specific manner remain to be fully understood.

Rodent models of maternal obesity are a relatively recent development in this field; however, the pace of discoveries has been rapid. We recently developed a murine model of maternal obesity using a diet relevant to that consumed by many obese women that is high in both fats and sugar. We showed that offspring of obese mouse dams are hypertensive and develop increased adiposity with age (42). Hyperinsulinemia was observed at 3 mo of age, suggesting insulin resistance. Male offspring showed more pronounced phenotypic changes than females (42). By 6 mo of age, male offspring of obese dams demonstrated impaired glucose tolerance associated with β-cell failure. We also recently observed insulin resistance as assessed by euglycemic hyperinsulinemic clamp in adult male offspring of obese rats fed an identical diet (36) and have previously reported that following exposure to a high-fat diet during pregnancy and lactation, rat offspring show impaired glucose homeostasis (48) and altered expression of hepatic insulin signaling proteins (10).

Glucose intolerance and insulin resistance in muscle can predict development of T2DM (31, 41). Skeletal muscle is the major site of postprandial glucose disposal and is therefore one of the insulin-sensitive tissues most likely to manifest early signs of insulin resistance. A number of abnormalities have been reported in association with insulin resistance in human muscle, including alterations in muscle fiber composition (37), mitochondrial dysfunction (31), and expression of insulin signaling proteins (14). Previous studies in animal models of placental insufficiency (44, 46) or maternal undernutrition (13, 32) identified changes in these parameters in adult offspring with similar alterations observed in low-birth weight humans (40). However, the consequences of maternal obesity on offspring muscle structure and function, and any relationship with insulin signaling has not been studied to date. We hypothesized that muscle-specific insulin signaling defects, abnormal mitochondrial function, and/or alterations in muscle structure could contribute to insulin resistance in offspring of obese dams.

The aim of this study was therefore to determine whether the expression of insulin-signaling proteins, mitochondrial electron transport chain (ETC) activity, and/or muscle fiber composition are altered in muscle of male and female offspring of obese mice.

MATERIALS AND METHODS

Animal Protocols

All studies were approved by the Local Ethics Committee and were conducted according to Home Office Animals (Scientific Procedures) Act 1986. The model and its metabolic phenotype are described in detail by Samuelsson et al. (42). Briefly, female C57BL/6J mice, 1 wk after a first successful pregnancy were fed either a standard chow RM1 diet or a highly palatable obesogenic diet of chow supplemented with animal lard and condensed milk (Nestle, Croyden, UK) with added mineral mix AIN93G. Both diets were purchased from Special Diet Service UK (Devon, UK). After 6 wk, mice on the obesogenic diet had significantly higher body fat than the controls (42). Both obese and control mice were then mated and maintained on their respective experimental diets throughout pregnancy and lactation. Forty-eight hours after delivery, litters were reduced to six pups (litters lower than four pups were not used), with an equal female to male ratio where possible. At 21 days of age, all pups were weaned onto standard chow (RM1 diet) and maintained on this diet until 3 mo of age when animals in the fed state were killed by rising concentration of CO₂. At post mortem, vastus lateralis muscle was removed and snap frozen in liquid nitrogen for molecular and biochemical analyses. For histological analysis, the entire calf muscle group (triceps surae), including soleus muscle was mounted on a cork block and snap frozen in isopentane chilled in dry ice. All tissues were stored at 80°C.

Insulin Signaling Protein Expression

Vastus lateralis muscle was extracted in ice-cold lysis buffer (n = 6 per group) [50 mmol/l HEPES (pH 8), 150 mmol/l sodium chloride, 1% Triton X100, 1 mmol/l sodium orthovanadate, 30 mmol/l sodium fluoride, 10 mmol/l sodium pyrophosphate, 10 mmol/l EDTA, and a protease inhibitor cocktail]. The total protein content of the lysates was determined using a bicinchoninic acid kit for protein determination (Sigma-Aldrich, Haverhill, UK). Samples were diluted to a standard protein concentration of 2 mg/ml in Laemmli buffer, and 10 μg total protein was subjected to SDS-PAGE. The proteins were transferred to PVDF Immobilon-P (Millipore) membrane and blocked for 1 h (5% nonfat dehydrated milk, 1 × TBS, 0.1% Tween 20). Membranes were incubated overnight with antibody against the insulin receptor β subunit (IRβ), Insulin receptor substrate 1 (IRS-1), phospho-IRS-1 (Ser307), phosphatidyl inositol 3-kinase (PI3 kinase) p85α regulatory subunit, and p110β catalytic subunit, Akt1, Akt2, phospho-Akt (Ser473), protein kinase C zeta (PKCζ), and GLUT4, diluted in TBS-0.1% Tween 20 containing either 5% dried milk or 5% BSA. Protein expression was quantified by spot densitometry using AlphaEase gs 3.3b (AlphaImager). To ensure the linearity of the signal, 20 mg and 10 mg of one sample were loaded onto each gel.

Antibodies and Reagents

Antibodies against IRS-1, phospho-IRS-1 (Ser307), and PI3 kinase p85α were purchased from Upstate Biotechnology (Lake Placid, NY), while Akt1, Akt2, and phospho-Akt (Ser473) were obtained from Cell Signaling Technology, (Beverly, MA). Antibodies against IRβ, PKCζ, and PI3K p110β were from Santa Cruz Biotechnology (Santa Cruz, CA), while GLUT4 was purchased from Abcam (Cambridge, UK). All biochemical reagents unless stated otherwise were purchased from Sigma-Aldrich, UK.

Mitochondrial ETC Activity

Mitochondrial ETC enzyme activity was measured in vastus lateralis muscle. The entire muscle was thawed and immediately homogenized on ice in 1:9 wt/vol sucrose buffer pH 8; 320 mM sucrose, 1 mM EDTA, 10 mM Trizma-base, using a kinematica status mechanical homogenizer (Kinematica, Lucerne, Switzerland) (n = 6 per group). After homogenization samples were snap frozen in liquid nitrogen and stored at −80°C. Aliquots were freeze-thawed twice to lyse cells, and enzymatic activity was assayed at 30°C on a Cary 50 spectrophotometer (Varian, Palo Alto, CA). NADH-Ubiquinone (CoQ) oxidoreductase, EC 1.6.5.3 (complex I), succinate CoQ reductase, EC 1.3.5.1 (complex II), CoQ cytochrome c reductase, EC 1.10.2.2 (complex III), cytochrome c oxidase, EC 1.9.3.1 (complex IV), linked assay succinate-cytochrome c reductase, EC 1.3.5.1 + EC 1.10.2.2 (complex II-III), and citrate synthase EC 4.1.3.7 activity was measured as described previously (45). Total protein content of the homogenate was measured using a commercially available modified Lowry Assay (Bio-Rad, Hertfordshire, UK). Citrate synthase activity was normalized to total protein to check its validity as a normalizing factor and to give an estimate of mitochondrial enrichment (19). Activities of all four complexes were standardized using citrate synthase activity (cs) to compensate for mitochondrial enrichment of the sample.

CoQ₉ Analyses

CoQ was measured in vastus lateralis muscle according to Duncan et al. (15). Total CoQ was extracted from freeze thawed vastus lateralis muscle homogenate with organic-phase extraction (hexane: ethanol 5:2) by centrifugation with a known concentration of di-propoxy-CoQ₁₀ analog internal standard (n = 6 per group). CoQ₉ was detected using reverse-phase HPLC linked to UV detection (Jasco, Essex, UK). Results were calculated from a CoQ₉ standard and normalized to protein and citrate synthase activity to measure total and mitochondrial CoQ content of the sample, respectively (38).

Histology

Soleus muscle was selected for histological analysis of fiber-type distribution because in mice, this muscle has almost an equal proportion of type I and type II fibers (49), whereas vastus lateralis is mainly composed of type II fibers. Therefore, changes in muscle fiber type will be more apparent in soleus muscle. Additionally, soleus muscle is relatively small, so at the midbelly point of analysis, it is possible to analyze all of the muscle fibers in one section. Muscle fiber type was identified by differential ATPase staining after incubation in acid or alkaline buffer, as described previously by Bancroft and Stevens (6). The mounted muscle group was sectioned at 10-μm thickness on a cryostat (Bright Instrument Co., Huntingdon, UK) (n = 4 per group). Sections containing the midbelly of the soleus muscle were collected and transferred to polylysine-coated microscope slides (Thermo Fisher Scientific, Leicestershire, UK) and muscle fiber type identified by differential staining in acidic buffer or alkaline buffer (47). Images were collected using a photomicroscope (Zeiss Germany) and analyzed using Scion Image software (Scion Corporation, Frederick, MD). Muscle fiber type was identified by the stability of myosin ATPase at different pH, which appears as a dark brown/black stain when viewed under a microscope. Fast-type muscle (Type II) myosin is stable at alkaline pH, whereas slow-type muscle (Type 1) is stable at acid pH. Fiber-type ratio was calculated from counts of different muscle type. Fiber density and cross-sectional area were calculated using Scion Image by tracing around perimeters of muscle fibers and calibrated with a scale bar captured at the same magnification.

Statistical Analysis

Data were analyzed from one male and one female per litter; n refers to numbers of litters per group. All data between groups were analyzed using a two-way ANOVA with sex and maternal obesity as the independent variables, followed by Duncan's post hoc test when appropriate (Statistica, Statsoft, Tulsa, OK). Insulin-signaling protein data are represented as mean percentage expression of male control offspring ± SE and mitochondrial measurements as the mean ratio of citrate synthase activity ± SE. For all data sets, a P < 0.05 was considered statistically significant.

RESULTS

Body Weight

Body weight, fat pad mass, and tissue weights have been reported previously in Samuelsson et al. (42).

Insulin-Signaling Protein Expression

IRβ and IRS-1.

Maternal obesity and sex had no effect on the expression of the insulin receptor in offspring muscle tissue (Fig. 1A). There was also no overall statistically significant effect of maternal obesity on expression of IRS-1 (Fig. 1B). However, there was an interaction between maternal obesity and sex (P < 0.001), reflecting significantly reduced IRS-1 expression in female offspring of obese dams compared with female offspring of control dams (P < 0.001). This effect was not observed in male offspring. Phosphorylation of IRS-1 on Ser307 was not altered in either sex (Fig. 1C).

Fig. 1.

Expression and phosphorylation of key proteins involved in insulin signaling A: Insulin receptor β subunit (IRβ); B: IRS1; C: phospho-IRS-Ser307; D: P13 kinase p85α regulatory subunit; E: P13 kinase p110β catalytic subunit; F: phospho-Akt-Ser473; G: Akt1; and H: Akt2; I: PKCζ; and J: GLUT4. Bars represent mean expression/phosphorylation values for control offspring (C) and offspring of obese dams (O) ± SE, expressed as a percentage of mean male control; n = 6 per group. Images of representative blots of the respective proteins are shown below corresponding bar graphs. *P < 0.05, **P < 0.01, ***P < 0.001 compared with the same sex control.

PI3 kinase.

Maternal obesity and sex had no statistically significant effect on the expression of the PI3K p85α subunit in muscle tissue (Fig. 1D). However, there was a significant effect of maternal diet (P < 0.05) and an interaction between maternal diet and sex (P < 0.05) on the expression of the PI3K p110β subunit (Fig. 1E). This reflected a significant decrease in the expression of p110β in the female offspring of dams fed the obesogenic diet compared with female offspring of control dams (P < 0.01). No such effect was observed in males.

Akt proteins.

The level of phosphorylation of Akt on Ser 473 was higher in females than in males (effect of sex P < 0.01). Importantly, maternal obesity resulted in a significant (P = 0.05) reduction in Akt phosphorylation on Ser 473 (Fig. 1F), an effect that was more pronounced in female offspring. There was no statistically significant effect of maternal obesity or sex on expression of Akt1 (Fig. 1G). However, there was a significant effect of maternal diet (P < 0.05) and an interaction between maternal diet and sex (P < 0.05) on the expression of Akt2 (Fig. 1H). This reflected a significant increase (P < 0.01) in expression of Akt2 in male offspring of obese dams compared with male offspring of control dams. This effect was not observed in females.

PKCζ and GLUT4.

There was a significant effect of maternal diet (P < 0.001) and an interaction between maternal diet and sex (P < 0.001) on the expression of PKCζ (Fig. 1I). This reflected a significantly increased expression of PKCζ in male offspring of obese dams compared with male offspring of control dams (P < 0.001). This was not observed in female offspring. GLUT4 expression was increased in female offspring compared with male offspring (effect of sex P < 0.05) (Fig. 1J); however, there was no effect of maternal obesity on GLUT4 expression.

Mitochondrial ETC Enzyme Activity

No significant effect of sex or maternal diet was observed for citrate synthase activity. Therefore, citrate synthase was used for normalization of all mitochondrial data, and results are expressed as a ratio of citrate synthase activity. Individual mitochondrial ETC enzyme activities of complex I, complex II, complex III, and complex IV normalized to citrate synthase in muscle were not significantly influenced by sex or maternal obesity (Fig. 2, A–D). However, in the linked complex II-III assay, there was a significant effect of sex (P < 0.01) and an interaction between maternal diet and sex (P < 0.05), which reflected a significant deficit in linked complex II-III activity only in the male offspring of obese dams compared with male control offspring (P < 0.01) (Fig. 2E).

Fig. 2.

Mitochondrial enzyme activities. A–D: individual mitochondrial enzyme activities. E: linked mitochondrial complex II-III enzyme activity. Bars represent mean enzyme activity normalized to citrate synthase (cs). F: mitochondrial CoQ₉ content. Bars represent mean CoQ concentration normalized to citrate synthase (cs). Control offspring (C) and offspring of obese dams (O) ± SE; n = 6 per group. **P < 0.01, compared with the same sex control.

Total CoQ₉ Content

No significant effect of maternal diet or sex was observed on CoQ₉ levels (Fig. 2F).

Muscle Fiber Histology

Muscle fiber type ratio was not influenced by maternal diet or sex. Muscle cross-sectional area and muscle fiber density were also not different between the groups (Table 1).

Table 1.

Muscle fiber histology

	Type I		Type II		Combined
	Male	Female	Male	Female	Male	Female
Fiber type ratio
Control	0.47±0.014	0.46±0.021	0.53±0.015	0.54±0.021
Obese	0.45±0.016	0.42±0.044	0.55±0.016	0.58±0.044
Fiber cross-sectional area, μm2
Control	1678±122	1722±156	1681±149	1643±220	1680±135	1683±179
Obese	1687±15	1893±118	1442±112	1749±213	1564±50	1821±156
Fiber density fibers, mm2
Control	243±18	251±30	285±35	283±27	528±49	535±56
Obese	222±19	235±23	307±38	290±32	530±28	526±54

Data are presented as the ratio of fiber type I and II relative to total fiber types in soleus muscle. Cross-sectional area is the area of the individual muscle fibers in transverse section (μm²). Fiber density is the number of fibers in a given area (mm²). Results are shown as means ± SE; n =4 for all groups.

DISCUSSION

The developmental “overnutrition hypothesis” postulates that offspring of obese mothers have a greater propensity for developing obesity, hypertension, and T2DM (11, 39). Some studies in obese women and their children support this hypothesis (21, 29); however, the effects of genetic susceptibility and shared nutritional environment of the mother and child may confound interpretation. Models such as that described here provide a tool to address the role of maternal environment in a controlled manner and so can contribute to our understanding of the developmental programming of metabolic disease.

Defects in muscle structure (33), mitochondrial function (23, 34, 50), and insulin signaling (14) have all been associated with T2DM in man. It is currently unclear which of these defects, if any, arises as a result of maternal diet-induced obesity during gestation and lactation. We hypothesized that offspring of obese dams, which we have shown previously are insulin resistant at 3 mo, would demonstrate muscle insulin signaling defects and that these might be associated with alterations in mitochondrial function and/or muscle structure. At this age, there was an effect of maternal obesity on male offspring body weight, with increases in fat pad mass and a decrease in skeletal muscle weight in both male and female offspring of obese dams (42).

The analysis of insulin-signaling proteins showed that there were no differences in the expression of the insulin receptor in either male or female offspring of obese dams compared with controls. This is consistent with the findings from human studies and animal models of T2DM showing that alterations occur downstream of the insulin receptor (14, 26, 27). Alterations in the expression and phosphorylation of key molecules downstream of the receptor were sex specific.

First, female offspring of obese dams showed a reduced expression of IRS-1. Insulin-dependent phosphorylation of IRS-1 is the initial step in the activation of PI3K. Several studies support the importance of IRS-1 in insulin signaling. siRNA-based gene silencing of IRS proteins has shown that IRS-1, rather than IRS-2, is required for insulin-stimulated Akt phosphorylation and glucose uptake (9). Also, defective insulin action, primarily in muscle, has been reported in IRS-1 knockout mice (5). Second, the p110β subunit of PI3K was also significantly reduced in the female offspring of obese dams. Third, Akt phosphorylation at Serine 473 was decreased in the female offspring of dams fed an obesogenic diet. To determine whether this reduction was caused by decreased substrate availability, we also measured the expression of Akt1 and Akt2 and observed no decrease in expression of either isoforms. Hence, female offspring of obese dams have impaired Akt phosphorylation, which could be due to impaired PI3K activity.

Paradoxically, in male offspring of obese dams, there was a significant increase in the expression of Akt2, the predominant Akt isoform in the muscle (3, 4), which would suggest improved insulin sensitivity. However, Akt phosphorylation on serine 473 tended to be decreased in this group, suggesting that Akt activity is reduced. Thus, an increase in total Akt2 levels may be acting as compensatory mechanism. PKCζ was also overexpressed in male offspring of obese dams. In a rat model of maternal high-fat feeding, 3-mo-old male offspring were shown to have increased adiposity (10) similar to that observed in our model. However, while the pattern of insulin signaling protein expression suggested hepatic insulin resistance in the offspring of the high-fat fed dams, the quadriceps muscle demonstrated a general increase in expression of insulin- signaling proteins. This model of high-fat feeding was, however, not associated with maternal obesity.

Interestingly, females had a higher expression of GLUT 4 compared with males. Because this could affect the overall rate and efficiency of glucose uptake in response to insulin (18, 41), an increased GLUT 4 expression could explain why the male but not female offspring of obese dams becomes glucose intolerant at 6 mo of age (42).

In the mitochondrial ETC activity assays, we found no effect of maternal diet on individual complex activities. Linked complex II-III activity, however, was compromised as a result of maternal obesity in male offspring only. As there was no deficiency in either of the individual complex II or III enzymes, we tested for the loss of CoQ, the essential electron carrier between complex II and III. Although our previous study on a model of undernutrition showed that a deficit in linked complex II-III was due to a loss of CoQ available to mitochondria (45), in this study, total cellular content of CoQ was not influenced by maternal diet and thus is unlikely to be a contributory factor. One possible explanation for this loss in linked activity could be that affinity for substrates in the linked complex II-III assay may be reduced, thus compromising this linked enzyme activity (30). Although we do not understand how the observed sex-specific deficit in linked complex II-III activity will affect ATP, and therefore energy production, electron transport chain uncoupling, and reactive oxygen species production in this model, others have described such sex-specific alterations in mitochondrial activity in association with aging processes (2, 53).

Alterations in muscle fiber type, size, or density have been shown to lead to alterations in glucose disposal by skeletal muscle (33, 37). While muscle fiber number is set during intrauterine development (16, 52), muscle fiber type and size can be changed during postnatal life by environmental influence, or as a compensatory mechanism for an early loss in muscle fiber number (20). Reports from “developmental programming” studies and early commercial meat production studies (7, 54) suggest that both muscle fiber type and muscle fiber number can be altered by intrauterine environment, particularly by changes in maternal diet. Indeed, a recent study in sheep has reported that maternal obesity leads to reduced myogenesis and smaller fiber diameter in fetal skeletal muscle, but it is not known if this persists into adulthood (51). We did not observe any difference in soleus muscle fiber type, size, or density, suggesting that maternal obesity does not have a persistent effect on soleus muscle fiber structure. Muscle fiber structure is therefore also unlikely to be a contributing factor to insulin signaling protein expression changes or mitochondrial abnormalities. However, this does not exclude the possibility that phenotypic alterations may occur later since diabetic and obese patients show altered muscle fiber type, loss of muscle fiber density, and cross-sectional area and therefore overall reduction in muscle mass (28). These subjects may also show a shift to type II (fast twitch) fibers (33), which may then further contribute to disease progression.

Perspectives and Significance

This study evaluated the effects of maternal obesity on programming of metabolic disease, specifically analyzing molecular defects in the insulin-signaling pathway and mitochondrial complex activities of muscle. It constitutes part of an integrated approach, which includes the analysis of these parameters in other metabolically active tissues and also considers the impact of this particular maternal insult on the programming of energy balance and implications for the development of obesity and type 2 diabetes.

We have demonstrated that maternal obesity programs sex-specific effects on the expression of the insulin-signaling proteins and mitochondrial complex II-III linked activity in 3-mo-old mice offspring. To our knowledge, this is the first description of programming of these abnormalities by maternal diet-induced obesity. The observed changes in insulin signaling and mitochondrial function may contribute to the hyperinsulinemia observed at this age and thus may represent potential targets for intervention. Further analysis is required to characterize the mechanisms underlying the effects of maternal obesity and the observed programmed reduction in mitochondrial complex activity on the development of increased offspring adiposity.

GRANTS

This study was supported, in part, by a British Heart Foundation Lectureship (to S. E. Ozanne), the Parthenon Trust (M. Martin-Gronert), Tommy Charity (to J. M. McConnell and L. Poston), King's College London Scholarship (P. Shelley) and the Biotechnology and Biological Sciences Research Council (to D. Fernandez-Twinn).