Characteristics of the sows and their offspring

Sow body weight increased throughout gestation, irrespective of diet (P < 0.05, Table 1). Maternal plasma HDL concentrations were raised in the HF group compared to controls, whereas plasma LDL, triglycerides and glucose were unaffected. Sow’s milk on day 2 of lactation contained 8% fat, 5% lactose and 6% protein and was unaffected by maternal diet [33]. There was no effect on the length of gestation (C = 117.4 ± 0.3; HF = 116.1 ± 0.7 days (P = 0.19)), mean birth weight (C = 1.22 ± 0.04; HF = 1.25 ± 0.07 kg (P = 0.56)) or litter size (C, 14.9 ± 0.8; FS, 16.4 ± 0.7 piglets/sow (P = 0.07)), but by 7 days of age the offspring born to HF mothers were heavier than controls evidenced by a greater fractional growth rate (Figure 1).

Days of gestation	0	40	108	p
Sow weight (Kg)
C	215.5 ± 13.4	236.6 ± 11.8	287.7 ± 10.5
HF	209.8 ± 9.2	254.6 ± 8.6	281.8 ± 8.7
HDL (mmol/l)
C	0.55 ± 0.05	0.55 ± 0.05*	0.42 ± 0.03#	* < 0.05; # < 0.01
HF	0.66 ± 0.05	0.77 ± 0.04*	0.56 ± 0.04#
LDL (mmol/l)
C	0.68 ± 0.06	0.87 ± 0.05	0.74 ± 0.03
HF	0.75 ± 0.06	0.98 ± 0.08	0.84 ± 0.08
Glucose (mmol/l)
C	3.88 ± 0.57	3.93 ± 0.12	3.93 ± 0.10
HF	3.58 ± 0.51	4.11 ± 0.08	3.70 ± 0.17

Values are means ± SEM. C: control (n = 7); HF: high fat (n = 7); HDL: High-density lipoprotein; LDL: Low-density lipoprotein. Superscripts denote significance levels between dietary groups, _* < 0.05; # < 0.01.

Fibre characteristics and association with body weight

Myofibre cross sectional area (CSA) was higher in the HF group (Figure 2A, B and C), this was positively associated with body weight at 7 days of age as demonstrated in Figure 2D. HF offspring also had reduced myocellular lipids; these were mainly stored on fibres located at the centre of the muscle fascicules (Figure 3A, B and C). Despite the changes in myofibre CSA, there were no differences in CSA of type I fibres (Figure 3D and 3E; C: 232.7 ± 31.3; HF: 317.6 ± 38.9 CSA μm² (P = 0.165)).

Gene expression, biochemical and metabolic responses

The gene expression and activities of enzymes involved in muscle metabolism were examined; the activity ratio of LDH (a pro-glycolytic enzyme) and ICDH (an indicator of global mitochondrial oxidation) were higher in the piglets born to mothers fed a HF diet compared to controls (Figure 4A). The ratio of NAD+/NADH, a marker of enhanced cellular oxidative capacity, was significantly lower in the HF group (Figure 4B).

The biceps femoris muscle of offspring born to HF fed mothers demonstrated raised expression of the genes ENO3, PGAM2, FAT/CD36 and a reduction in GLUT-1 (Figure 5), but there were no differences in either GLUT-4 (C = 1.0 ± 0.1; HF = 1.5 ± 0.5 (P = 0.12)) or CPT-1 (C = 1.0 ± 0.1; HF = 1.2 ± 0.1 (0.66)). Consistent with the increases in pro-glycolytic activity, the muscle samples obtained from HF offspring exhibited a higher concentration of glycogen compared to controls (C = 20.3 ± 2.4; HF = 61.8 ± 16.0 mg g tissue⁻¹ (P = 0.007)). However, there were no differences between the groups for intramuscular triglyceride (C = 16.7 ± 1.6; HF = 23.8 ± 3.3 mg g tissue⁻¹ (P = 0.31) or protein content (C = 39 ± 3; HF = 41 ± 2 mg g tissue⁻¹ (P = 0.624)).

Fatty acid content and lipid metabolism

Can maternal diet modify the muscular incorporation of essential fatty acids in the neonate?

We also observed differential effects of maternal fat supplementation upon muscle fatty acid composition of the offspring, in particular n-6 and n-3 fatty acids. These essential fatty acids are crucial components of the cellular membrane, affecting structural and regulatory properties that need to adapt to changes in fatty acid supply [48],[49]. There is indirect evidence of a different pattern of muscular incorporation of two essential fatty acids, arachidonic acid (C20:4n6) and α-linoleic acid (C18:3n3). Both of these fatty acids are involved in several metabolic pathways associated with muscular development and insulin signalling [8]. α-linolenic acid itself can repress the actions of arachidonic acid, it appears likely that reduced α-linolenic acid is responsible for the increased proportion of arachidonic, and may facilitate further muscle growth [50]. The fatty acid composition of skeletal muscle is also influenced by lipid binding proteins such as LPL and CD36/FAT, which allow circulating triglycerides to be hydrolyzed and free fatty acids to be accumulated [48],[51]. Expression of both these genes was increased in fat supplemented offspring, suggesting these processes were enhanced.

Materials and methods

All animal procedures described in this manuscript were approved by the Ethics Committee for Animal Experiments of the Animal Sciences Group of Wageningen Research Centre, and conducted at Schothorst Feed Research in the Netherlands. All laboratory procedures were carried out at The University of Nottingham under the United Kingdom code of laboratory practice (COSHH: SI NO 1657, 1988).

Animals and diets

The sows were weighed on 0, 40, 70 and 108 days of gestation prior to feeding. On 0, 40 and 108 days of gestation fasting blood samples were taken from each animal. The plasma was extracted in K⁺EDTA coated tubes, the samples were immediately separated by centrifugation (3000 g for 10 minutes at 4°C) and stored at −80°C until analysis.

From day 110 of gestation (mean caloric intake ≈ 27.3 MJ/day) and throughout lactation (59 MJ/day) all the sows were fed an identical diet sufficient to meet their energy requirements. Sows gave birth naturally in farrowing crates and all piglets were weighed after birth when litter size plus any still births were recorded. All piglets remained with their mothers and were allowed to suckle ad libitum. In all sows, milk was sampled, after milk let down via an intravenous oxytocine injection (10 I.U./mL; Eurovet Animal Health, Bladel, The Netherlands) into the ear on day 2 of lactation. At 7 days after parturition, the median birth weight offspring (C: 6 females, 1 male; HF: 4 females, 3 males) in each litter were weighed, a blood sample collected, the piglet sedated with 10% ketamine and then euthanized with (50 mg/kg) T-61 (Intervet, Boxmeer, Holland) followed by exsanguination.

Laboratory procedures

Muscle sampling

The biceps femoris is a morphologically and functionally distinct skeletal muscle which forms part of the hamstring muscle group and is located on the posterior thigh. Samples were removed from the centre of each muscle to standardize sampling location and immediately frozen in liquid nitrogen and subsequently stored at −80°C until analysis was performed. All homogenisation carried out during this study was performed using a gentleMACS™ closed homogeniser (Miltenyi Biotec Ltd., Surrey, UK).

Enzymatic kinetics analysis

The lactate dehydrogenase (LDH) and isocitrate dehydrogenase (ICDH) total enzymatic assays were adapted from protocols obtained from SIGMA (http://www.sigmaaldrich.com/life-science/metabolomics/enzyme-explorer/learning-center/assay-library.html). Briefly, the muscular extract used (0.1 g) was homogenised in 4 ml buffer containing 0.25 M sucrose, 0.2 mM EDTA and 0.1 mM Tris (pH7.5). All samples were centrifuged at 6000 g for 15 minutes at 4°C and the supernatant diluted 1:2 with homogenisation buffer for use in subsequent assays. Total LDH enzymatic activity (EC 1.1.1.27) was determined in 1 μl of the muscle homogenate to which 10 μl NADH (0.33 mM) and 290 μl of reaction buffer (2 mM sodium pyruvate, 50 mM TEA, 5 mM EDTA) (Sigma-Aldrich Co LLC, Gillingham, UK) were added. Similarly, total ICDH enzymatic activity (EC 1.1.1.42) was determined in 7.5 μl of the muscle homogenate to which 270 μl of ICDH reaction buffer (38.9 mM Na2HPO₄, 0.5 mM MnCl₂ (4H₂O), 0.05% βNADP) and 1 μl of (0.4 M) isocitrate (Sigma-Aldrich Co LLC, Gillingham, UK) were added. Enzymatic activities were measured by changes in absorbance at 340 nm at 28°C on a 96-well spectrophotometer (BIO-TEK Instruments Inc., Vermont, USA). The total LDH and ICDH activity were expressed as oxidised moles of their respective subtracts per second per gram of protein. The kinetic data were analysed by nonlinear regression using a method based on the principles of the Michaelis-Menten equations [53].

Muscle composition

Total protein of the muscle homogenates was determined using a commercial kit (Bio-Rad Laboratories Inc. Hemel Hempstead, UK) based on the Bradford method [54] on. The results were corrected to the dissected muscle weight obtained from each offspring.

Glycogen content was determined using a method developed by Dalrymple and Hamm [55]. The concentration of glucose released from this reaction was determined by a colorimetric commercial assay (Randox Ltd., County Antrim, UK) and the results were expressed as a concentration of milligrams of glycogen per gram of tissue dissected.

NAD+/NADH ratio was determined using a colorimetric commercial kit (Sigma-Aldrich Co LLC, Gillingham, UK) and normalised to the quantity of tissue dissected i.e. moles per gram of protein.

Triglyceride content was assessed using the Folch method [56] followed by colorimetric commercial assay (Randox Ltd., County Antrim, UK) and results expressed in milligrams per gram of tissue dissected.

Phospholipid composition was determined by gas-chromatography in which the chloroform phase was evaporated by applying a nitrogen stream and then 2 ml of hexane was added to each sample. The phospholipids were transmethylated by adding 40 μl of methyl acetate and 40 μl of methylation reagent (30% sodium methoxide, 4.1 ml methanol; Fisher Scientific Ltd. Loughborough, UK) to each reaction and left to react for 10 minutes at room temperature. This was followed by adding 60 μl of termination reagent (0.2 g dried oxalic acid and 6 ml diethyl ether) and methyl ester residues were extracted by adding 200 mg of calcium and followed by short centrifugation. The fatty acid methyl esters were then injected (split ratio 50:1) into a gas chromatograph (GC 6890; Agilent technologies Ltd, Stockport, UK). Separation of fatty acid methyl esters was performed with a Varian CP-88 (Crawford Scientific™ Ltd., Strathaven, UK) capillary column with hydrogen as carrier gas. Oven temperature was programmed from 59°C to 100°C at 8°C per min, then to 170°C at 6°C per minute and held for 10 minutes. The temperature of the injector and detector were set at 255°C and 250°C. The fatty acid methyl esters were identified by comparing the retention times with a fatty acid methyl esters standard (Sigma-Aldrich Co LLC, Gillingham, UK) and the area percentage in moles were used for the statistical analysis.

Histological analysis

Immediately after removing the muscle samples from −80°C storage, approximately 1 cm³ of frozen tissue was sectioned and left to equilibrate overnight at −20°C. The next day, each sample was embedded on a cryostat metal plate in such a manner as to set the muscle fibres perpendicular to the cutting blade. The mounted samples were cut into 10 μm thickness in a pre-equilibrated at −20°C Leica Cryostat (Leica Microsystems Ltd., Milton Keynes, UK). The slices were then mounted on polysine histological slides (Menzel-Glaser, GmbH & Co. Braunschweig, Germany) and allowed to equilibrate to room temperature prior to staining.

Haematoxylin and eosin staining was determined on sections placed in filtered 0.1% Harris’ haematoxylin for 3 minutes and gently rinsed in cool running water to extract excessive stain. This was followed by further rinsing the sections in acidic alcohol and cool running water before transfer to eosin for 2 minutes. After staining, the slides were dipped in distilled water, dehydrated back through a graded series of alcohol concentrations and placed in xylene for 5 minutes. The stained sections were mounted with a coverslip in resin base medium and finally incubated overnight at room temperature.

Immunostaining of slow myosin (MyHC) was determined using anti-slow myosin pig specific MyHC antibody (Sigma-Aldrich Co. LLC., Gillingham, UK) at 1:4000 dilution. The staining was carried out on the Bond Max histology system using the Bond Polymer Refine Detection System (Vision Biosystems, Mount Waverley, Australia) and Bond software version 3.4A. Briefly, slides were stained as follows: 15 minutes with primary antibody, 8 minutes with secondary antibodies, 10 minutes with 3,3 diaminobenzine and counterstained with hematoxylin. A negative control slide in which the primary antibody stage was excluded was included with each batch.

Oil red O histochemical lipid staining was used to assess the intramyofibre lipid content which detects cellular fat droplets as a light red tint. Briefly, the 10 μm cryosections were placed in a Coplin jar containing 60% isopropanol for 15 minutes. Immediately afterwards, the sections were incubated for 20 minutes in a freshly-prepared solution containing 0.5% w/v oil red-O/isopropanol (Fisher Scientific Ltd. Loughborough, UK) diluted in a 1% w/v solution of dextrin. Thereafter, the stained samples were immersed for 1 minute in 60% isopropanol and then rinsed in cool running water. This was followed by counterstaining of the muscular fibres with Harris’ haematoxylin for 3 minutes, followed by a quick rinse in cool running water. In addition, the slides were dipped 3 times in Scott’s tap water and fast dried. Finally, a drop of 50% v/v glycerol-water solution was placed over the samples in order to preserve the lipids and the coverslip was sealed by applying an acetone-based nail polish.

Images of the sections were captured using a high performance CCD camera connected to a Leica universal microscope at 20 x magnification (Leica Microsystems Ltd., Milton Keynes, UK). From each image, 10 random pictures were taken for analysis. The mean cross-sectional fibre areas obtained from each image were analysed by dividing them into 12 equal fields and choosing one by random in which at least 150 fibres were measured through the use of image analysis software (Image-pro plus; MediaCybernetics, Bethesda, USA). Irregular-shaped fibres were excluded using a custom-made macro that drew a ring over each myofibre; filtering criteria were applied to reject regions with a CSA < 50 μm² or > 5600 μm² or regions with circularity (approximation diameter of an ellipse) of <0.3 or >1.0. For type I fibre analysis, a function was incorporated in our custom-made macro to detect the colour intensity of the positive-stained fibres. Finally, for the Oil red O analysis, the colour density threshold was adjusted to distinguish all the fat droplets on a single fibre. Only the droplets with an area > 0.4 μm² and < 1.5 μm² in those encircled fibres were included in this analysis.

Gene expression

Total RNA was extracted from 100 mg skeletal muscle using a commercially available kit (Qiagen Ltd., Crawley, UK), which included a DNA purification step. The amount and purity of the extracted RNA was determined using a Nanodrop (Thermo Fisher Scientific Ltd. Leicester, UK). 1 μg of total RNA was reverse transcribed using a Thermocycler (Thermo Fisher Scientific Ltd. Leicester, UK). Quantitative PCR (qPCR) was performed using a Roche lightcycler 480 thermocycler and SYBR technology (Roche, Burgess Hill, UK). For quantification of gene expression we applied the comparative C^t method, which was normalised by applying a geometric mean of 3 endogenous housekeeping genes (18 s rRNA, β actin and cyclophillin) [57]. The genomic data is expressed in this study as a ratio to the commercial control animals.

The qPCR primers were designed based on known porcine sequences published on the Genbank using public online software (Primer3) on intra-exonic boundary sequences where possible. A standard curve was included and the samples were run in duplicate as well as having the appropriate positive and negative controls. The primers used were purchased from Eurofins MWG Operon GmbH. (Ebersberg, Germany) and validated as described in previous publications [58]. The following porcine-specific oligonucleotide forward (F) and reverse (R) primers were used:

ENO3: F:GAGCTGGATGGGACAGAAAA–R:GCAATGTGACGGTAGAGTGG; PGAM2:F:GATCAAGGCAGGCAAGAGAG–R:ACATCCCTTCCAGATGCTTG; FAT/CD36: F:TGAAAGAAGCAGGTGCTGAA–R: AGGACTGCTCCCAATGACAGC; Primers for cyclophilin, 18S, β-actin, LPL, GLUT-4 and CPT1 have previously been published [58]-[60].

Statistical analysis

Statistical analysis of the data was performed using SPSS® statistics software (v 16.0 for Windows; IBM, Chicago, USA). The data was tested for normality by applying the Kolmogorov test and controlled by observation of its Gaussian distribution. Depending on the results of the previous tests, the data was analysed by applying Student’s unpaired t-test and linear regression analysis with Benjamini & Hochberg False Discovery Rate correction; otherwise the non-parametric Mann Whitney test and Spearman Rank correlation were used. In all cases, the results are given as mean ± SEM, p < 0.05 was considered as statistical significant.