Ontogeny of Sex-Related Differences in Foetal Developmental Features, Lipid Availability and Fatty Acid Composition

Sex-related differences in lipid availability and fatty acid composition during swine foetal development were investigated. Plasma cholesterol and triglyceride concentrations in the mother were strongly related to the adequacy or inadequacy of foetal development and concomitant activation of protective growth in some organs (brain, heart, liver and spleen). Cholesterol and triglyceride availability was similar in male and female offspring, but female foetuses showed evidence of higher placental transfer of essential fatty acids and synthesis of non-essential fatty acids in muscle and liver. These sex-related differences affected primarily the neutral lipid fraction (triglycerides), which may lead to sex-related postnatal differences in energy partitioning. These results illustrate the strong influence of the maternal lipid profile on foetal development and homeorhesis, and they confirm and extend previous reports that female offspring show better adaptive responses to maternal malnutrition than male offspring. These findings may help guide dietary interventions to ensure adequate fatty acid availability for postnatal development.


Introduction
Prenatal development of humans and animals requires adequate placental supply of oxygen and nutrients [1,2], which, in turn, requires adequate maternal nutrition and placental function. Inadequate placental nutrient supply leads to intrauterine growth restriction (IUGR), resulting in newborns that are small for their gestational age. Such offspring may be predisposed to perinatal morbidity and mortality [3] as well as lifelong chronic non-communicable disorders such as obesity, type II diabetes, hypertension and cardiovascular diseases [4][5][6][7].
Among humans, IUGR due to maternal malnutrition shows a much higher incidence in resource-challenged areas (15%) than in developed areas (6%) [8,9]. In recent years, incidence has been increasing in developed countries because of maternal eating disorders, voluntary intake restriction for aesthetic reasons [10] and abnormal placental development leading to placental insufficiency [11]. This abnormal placental development has been associated with postponement of childbearing, inadequate lifestyle and maternal and gestational factors [8,12,13].
The size and weight of normal (non-IUGR) foetuses increased with gestational age (Table 1), as did development of the placenta, liver, kidney and intestine, based on histology (all p < 0.0001).
In these normal foetuses, body development correlated strongly with maternal lipid profile depicted in Figure 1. Specifically, lower maternal plasma cholesterol concentrations correlated with lower foetal body weight (r = 0.777, p < 0.05) and corpulence in terms of trunk length (r = 0.832), thoracic circumference (r = 0.808) and abdominal circumference (r = 0.958) (all p < 0.01). Lower maternal plasma cholesterol concentrations also correlated with higher ratios of organ-to-body weight for brain (r = −0.774), heart (r = −0.752), liver (r = −0.679), and spleen (r = −0.712) (all p < 0.01); these correlations reflect changes to improve foetal viability. Conversely, lower maternal plasma concentrations of high-density lipoprotein cholesterol (HDL-c) correlated with lower ratios of heart-to-body weight (r = 0.677, p < 0.05). Adaptive foetal growth mechanisms were observed already on GD 70, when lower maternal plasma cholesterol concentrations correlated with higher ratios of head-to-body weight (r = −0.987, p < 0.005) and heart-to-body weight (r = −0.883, p < 0.05). At the same time, maternal HDL-c concentrations correlated with higher ratios of brain-to-body weight (r = −0.923) and spleen-to-body weight (r = −0.891, both p < 0.05).
Maternal plasma concentrations of glucose and fructosamine did not affect weight of normal foetuses on GD 70, but they did show an influence on GD 90. Lower plasma fructosamine concentrations correlated with lower foetal body weight (r = 0.964) as well as higher ratios of head-to-body weight (r = −0.962) and brain-to-head weight (r = −0.969, all p < 0.05). Similarly, lower maternal glucose levels correlated with higher ratios of head-to-body weight (r = −0.959) and brain-to-head weight (r = −0.963, both p < 0.05).  Maternal plasma concentrations of glucose and fructosamine did not affect weight of normal foetuses on GD 70, but they did show an influence on GD 90. Lower plasma fructosamine concentrations correlated with lower foetal body weight (r = 0.964) as well as higher ratios of headto-body weight (r = −0.962) and brain-to-head weight (r = −0.969, all p < 0.05). Similarly, lower maternal glucose levels correlated with higher ratios of head-to-body weight (r = −0.959) and brainto-head weight (r = −0.963, both p < 0.05).

Effects of Foetal Sex on Developmental Trajectories during Pregnancy
No significant association was observed between offspring sex and foetal weight in normal foetuses (Table 1), although among male foetuses collected on GD 90, we observed a trend toward higher total body weight (p = 0.09) and carcass weight (p = 0.06) compared to female foetuses. Fat content of the carcass, the longissimus dorsi muscle or the liver did not vary significantly as a function of foetal sex or gestational age (Table 2). Male and female normal foetuses collected on GD 70 did not differ significantly in the relative weights or maturation states of several structures and organs. On GD 90, normal female foetuses showed a significantly higher degree of placental development (p < 0.05) and significantly higher ratios of head-to-body weight and brain-to-body weight (both p < 0.05; Figure 2

Effects of Foetal Sex on Developmental Trajectories during Pregnancy
No significant association was observed between offspring sex and foetal weight in normal foetuses (Table 1), although among male foetuses collected on GD 90, we observed a trend toward higher total body weight (p = 0.09) and carcass weight (p = 0.06) compared to female foetuses. Fat content of the carcass, the longissimus dorsi muscle or the liver did not vary significantly as a function of foetal sex or gestational age (Table 2). Male and female normal foetuses collected on GD 70 did not differ significantly in the relative weights or maturation states of several structures and organs. On GD 90, normal female foetuses showed a significantly higher degree of placental development (p < 0.05) and significantly higher ratios of head-to-body weight and brain-to-body weight (both p < 0.05; Figure 2).

Figure 2.
Mean ratios (±SEM) of head-to-body weight (A); head-to-carcass weight (B); and brain-tobody weight and brain-to-carcass weight (C) in female foetuses (black bars) and male foetuses (white bars) on Gestational Day (GD) 90. Asterisks indicate significant differences between males and females. * p < 0.05. Figure 3. Influence of foetal sex on foetal development under conditions of normal growth or severe IUGR: mean ratios (± mean square error) of brain-to-body weight and brain-to-carcass weight on Gestational Day (GD) 70 (A); ratios of brain-to-body weight, brain-to-carcass weight and liver-to-body weight on GD 90 (B); and ratio of spleen-to-body weight on GD 90 (C). Bars indicate, from left to right, female and male foetuses with severe IUGR, then female and male foetuses showing normal growth. Asterisks indicate the p value associated with differences between the foetal sexes and between foetuses showing IUGR or normal growth. * p < 0.05. Table 3 shows measurements of indicators of glucose and lipid metabolism based on markers in plasma, allantoic and amniotic fluids from normal foetuses. Comparison of measurements on GDs 70 and 90 shows that the availability of foetal triglycerides significantly decreased in foetal blood (p < 0.05) and allantoic fluid (p < 0.005) during pregnancy. Over the same period, the availability of foetal triglycerides in amniotic fluid increased (p < 0.05). Low-density lipoprotein cholesterol (LDLc) in foetal plasma decreased during pregnancy (p < 0.005), while HDL-c increased (p < 0.001). Total cholesterol decreased in amniotic fluid (p < 0.0001) but increased in allantoic fluid (p < 0.005). In both these compartments, HDL-c and LDL-c concentrations remained unchanged during pregnancy. Similar results were obtained for the two foetal sexes, except that LDL-c in blood decreased to a greater extent in females than males (p < 0.05). ; head-to-carcass weight (B); and brain-to-body weight and brain-to-carcass weight (C) in female foetuses (black bars) and male foetuses (white bars) on Gestational Day (GD) 90. Asterisks indicate significant differences between males and females. * p < 0.05. g vs. 48.2 ± 1 g, p < 0.05; GD 90, 103.5 ± 6.3 g vs. 141.9 ± 3.5 g, p < 0.005), carcass weight (GD 70, 66.2 ± 8.9 g vs. 100 ± 2.9 g, p < 0.05; GD 90, 222.3 ± 18.7 g vs. 364.1 ± 9.6 g, p < 0.005) and total weight of viscerae (GD 70, 18.2 ± 2 g vs. 28.3 ± 0.9 g, p < 0.01; GD 90, 64.3 ± 7.3 g vs. 98 ± 2.3 g, p < 0.05). However, the ratio of brain-to-body weight was significantly higher in foetuses with severe growth restriction than in littermates on GD 70 (0.048 ± 0.004 vs. 0.036 ± 0.001) and GD 90 (0.045 ± 0.004 vs. 0.033 ± 0.001; both p < 0.01). This was also observed when only female foetuses were examined (p < 0.0005 at both ages; Figure 3). On GD 90, the ratios of brain-to-body weight and brain-to-carcass weight were significantly higher in foetuses with severe growth restriction (both p < 0.0005), as were the ratios of liver-and spleen-to-body weight (both p < 0.05; Figure 3).

Figure 2.
Mean ratios (±SEM) of head-to-body weight (A); head-to-carcass weight (B); and brain-tobody weight and brain-to-carcass weight (C) in female foetuses (black bars) and male foetuses (white bars) on Gestational Day (GD) 90. Asterisks indicate significant differences between males and females. * p < 0.05. Figure 3. Influence of foetal sex on foetal development under conditions of normal growth or severe IUGR: mean ratios (± mean square error) of brain-to-body weight and brain-to-carcass weight on Gestational Day (GD) 70 (A); ratios of brain-to-body weight, brain-to-carcass weight and liver-to-body weight on GD 90 (B); and ratio of spleen-to-body weight on GD 90 (C). Bars indicate, from left to right, female and male foetuses with severe IUGR, then female and male foetuses showing normal growth. Asterisks indicate the p value associated with differences between the foetal sexes and between foetuses showing IUGR or normal growth. * p < 0.05. Table 3 shows measurements of indicators of glucose and lipid metabolism based on markers in plasma, allantoic and amniotic fluids from normal foetuses. Comparison of measurements on GDs 70 and 90 shows that the availability of foetal triglycerides significantly decreased in foetal blood (p < 0.05) and allantoic fluid (p < 0.005) during pregnancy. Over the same period, the availability of foetal triglycerides in amniotic fluid increased (p < 0.05). Low-density lipoprotein cholesterol (LDLc) in foetal plasma decreased during pregnancy (p < 0.005), while HDL-c increased (p < 0.001). Total cholesterol decreased in amniotic fluid (p < 0.0001) but increased in allantoic fluid (p < 0.005). In both these compartments, HDL-c and LDL-c concentrations remained unchanged during pregnancy. Similar results were obtained for the two foetal sexes, except that LDL-c in blood decreased to a greater extent in females than males (p < 0.05). Influence of foetal sex on foetal development under conditions of normal growth or severe IUGR: mean ratios (± mean square error) of brain-to-body weight and brain-to-carcass weight on Gestational Day (GD) 70 (A); ratios of brain-to-body weight, brain-to-carcass weight and liver-to-body weight on GD 90 (B); and ratio of spleen-to-body weight on GD 90 (C). Bars indicate, from left to right, female and male foetuses with severe IUGR, then female and male foetuses showing normal growth. Asterisks indicate the p value associated with differences between the foetal sexes and between foetuses showing IUGR or normal growth. * p < 0.05. Table 3 shows measurements of indicators of glucose and lipid metabolism based on markers in plasma, allantoic and amniotic fluids from normal foetuses. Comparison of measurements on GDs 70 and 90 shows that the availability of foetal triglycerides significantly decreased in foetal blood (p < 0.05) and allantoic fluid (p < 0.005) during pregnancy. Over the same period, the availability of foetal triglycerides in amniotic fluid increased (p < 0.05). Low-density lipoprotein cholesterol (LDL-c) in foetal plasma decreased during pregnancy (p < 0.005), while HDL-c increased (p < 0.001). Total cholesterol decreased in amniotic fluid (p < 0.0001) but increased in allantoic fluid (p < 0.005). In both these compartments, HDL-c and LDL-c concentrations remained unchanged during pregnancy. Similar results were obtained for the two foetal sexes, except that LDL-c in blood decreased to a greater extent in females than males (p < 0.05). Similar changes to these described in normal foetuses were observed in the subset of foetuses showing severe IUGR, and no sex effects were observed in this case. Foetuses with severe IUGR showed lower plasma LDL-c concentrations than littermates on GD 90 (29.7 ± 2.5 mg/dL vs. 35.8 ± 1.03 mg/dL, p < 0.05).

Changes in Foetal Antioxidant/Oxidative Status during Pregnancy and Sex-Related Effects
Ferric reducing antioxidant power (FRAP), an index of antioxidant capacity, decreased from GD 70 (10.8 ± 1.3 µmol/mL) to GD 90 (9.9 ± 1.6 µmol/mL, p < 0.005) in samples from normal foetuses, yet the concentration of malondialdehyde (MDA) in plasma, an index of total lipid oxidation, decreased during the same period from 20.6 ± 0.4 mmol/mL to 15.8 ± 0.6 mmol/mL (p < 0.005). Nevertheless, the ratio of MDA to cholesterol, which takes into account lipid availability, indicated less relative oxidation on GD 70 than GD 90 (2.8 ± 0.9 vs. 3.5 ± 0.7). Similar results were obtained for the ratio of MDA to LDL-c (4.1 ± 1.5 vs. 6.1 ± 1.2) and the ratio of MDA to triglycerides (5.1 ± 1.9 vs. 8.0 ± 2.8) (all p < 0.05; Figure 4). Similar changes to these described in normal foetuses were observed in the subset of foetuses showing severe IUGR, and no sex effects were observed in this case. Foetuses with severe IUGR showed lower plasma LDL-c concentrations than littermates on GD 90 (29.7 ± 2.5 mg/dL vs. 35.8 ± 1.03 mg/dL, p < 0.05).

Changes in Foetal Muscle Fatty Acid Composition during Pregnancy and Sex-Related Effects
The neutral and polar fatty acid fractions of the longissimus dorsi muscle in normal foetuses varied significantly with gestational age and foetal sex (Tables 4 and 5). On GD 70, the neutral fraction of female normal foetuses contained significantly more α-linolenic fatty acid (C18:3n-3) and total saturated fatty acids (SFA) than the neutral fraction of male normal foetuses (both p < 0.05), as well as less cis-vaccenic fatty acid (C18:1n-7) and lower ratios of ∑n-6/∑n-3 (p < 0.01) and monounsaturated fatty acids (MUFA) to SFA (p < 0.05). All these sex-related differences disappeared by GD 90.

Changes in Foetal Muscle Fatty Acid Composition during Pregnancy and Sex-Related Effects
The neutral and polar fatty acid fractions of the longissimus dorsi muscle in normal foetuses varied significantly with gestational age and foetal sex (Tables 4 and 5). On GD 70, the neutral fraction of female normal foetuses contained significantly more α-linolenic fatty acid (C18:3n-3) and total saturated fatty acids (SFA) than the neutral fraction of male normal foetuses (both p < 0.05), as well as less cis-vaccenic fatty acid (C18:1n-7) and lower ratios of ∑n-6/∑n-3 (p < 0.01) and monounsaturated fatty acids (MUFA) to SFA (p < 0.05). All these sex-related differences disappeared by GD 90.

Discussion
The results of the present study support the prominent role of the maternal metabolic profile on foetal development and homeorhesis and provide new insights on the effects of offspring sex on fetoplacental development, lipid availability and fatty acid composition at non-adipose tissues involved in metabolism regulation (muscle and liver). These data may be considered relevant for humans since despite differences in placentation, human and pig foetuses and pregnant females show similar lipid metabolism and distribution [39].
In the current study, lower cholesterol concentration in maternal plasma correlated with deficiencies in foetal development, which it is consistently with previous studies on GD 42 [16]. In contrast, poor availability of glucose, the main energy source for developing foetuses [40], affected foetal development on GD 90 but not on GD 70 in our study, consistent with reports that hypoglycaemia in early development is non-critical and can be compensated [41].
The higher growth of the brain when compared to total body observed in the foetuses affected by IUGR is consistent with the "brain-sparing effect" firstly described by Rudolph (1984) [42]. There was also a higher growth of other organs (mainly heart, liver and spleen), which is also consistent with studies in humans documenting "heart-and liver-sparing" cardiovascular adaptations-analogous to the brain-sparing effect-in response to foetal malnutrition and hypoxia [43,44]. Previous studies address that changes in the growth, morphology and function of various organs are dependant on the timing and severity of nutritional restriction [45,46], with specifically a determinant effect triggered by lipid availability. At the same time, our results support previous work indicating sex-specific differences in the growth of the different organs [27,31].
Maternal undernutrition in our study clearly affected foetal lipid availability. The developing foetus and placenta require large amounts of lipids for: (a) synthesis of cholesterol, which acts as a key constituent of cell membranes and organelles and is the precursor of a range of hormones and metabolic regulators necessary for successful pregnancy [47][48][49][50]; (b) secretion of key products such as lipoproteins; and (c) storage of triglycerides [47][48][49]. Although placental and foetal tissues can synthesize lipids de novo [47,51], the building blocks must be taken up by the placenta, i.e. maternal free fatty acids; triglycerides, which placental lipases hydrolyse into fatty acid constituents; and lipoprotein-associated cholesterol [19,21,23,51]. Triglycerides are a major source of energy for the foetus [52,53], but a balance is needed: low levels of triglycerides delay growth, while high levels can cause foetal macrosomia [54]. Availability of lipids to the fetoplacental unit depends on de novo synthesis by the foetus as well as maternal transfer. Cholesterol, for example, can reach the foetal circulatory system after crossing the syncytiotrophoblast as LDL-c [49,55]. In pig foetuses, fatty acid availability and composition depend more on foetal synthesis from precursors transferred from the mother than on direct maternal transfer [56][57][58][59][60][61], since fatty acids do not easily cross the placenta in ruminants, pigs or horses [62].
Lipid availability to the foetoplacental unit also depends on the antioxidant/oxidative status of the foetus. Our results suggest that, during pregnancy, antioxidant capacity decreases and lipid peroxidation increases in foetuses affected by maternal undernutrition, as suggested by previous studies [24][25][26]. These changes may explain the lower lipid availability later in pregnancy.
Cholesterol and triglyceride availability in our study did not differ significantly between male and female foetuses, consistently with prior studies [56]. Our data further show that the two sexes showed similar antioxidant capacity and lipid peroxidation. In other words, foetal development was affected to a much greater extent by gestational age-and therefore by nutritional restriction-than by sex. Nevertheless, we did observe sex-related differences in how the fatty acid composition of non-adipose tissues involved in metabolism regulation (muscle and liver) changed between GD 70 and 90 in foetuses affected by maternal undernutrition.
Our data indicate significant sex-related differences in content of essential fatty acids, which are so-called because they must be obtained from maternal transfer in the case of foetuses [22,63] or from diet in the case of adults [64]. The major essential fatty acids are linolenic PUFA (an omega-3 fatty acid) and linoleic PUFA (an omega-6 fatty acid); the long-chain omega-3 PUFA eicosapentaenoic and docosahexaenoic acids and the long-chain omega-6 PUFA gamma-linolenic and arachidonic acids are also essential. Future work should examine the maternal and/or placental factors that may drive the sex-dependent differences in essential fatty acid availability that we observed.
Most of the differences in fatty acid composition between male and female foetuses that we observed in muscle and liver belonged to the neutral fraction corresponding to triglycerides, which are an essential energy source [19]. Smaller differences were observed in the polar fraction corresponding to phospholipids, which constitute cell membranes and are essential for tissue development [19], so sex-related differences in fatty acid composition are most related to energy partitioning than to organ development.
We found that male foetuses had a higher n-6/n-3 ratio than females, and a high n-6/n-3 ratio appears to be deleterious [65], corresponding to the prodromal phase of insulin resistance [66]. Optimal development depends on adequate availability of n-6 [67], while n-3 may improve insulin function. Our results are consistent with previous reports that alterations in lipid metabolism and insulin regulation appear early in the development of male foetuses under limited nutrition [68][69][70][71]. Indeed, males in our study showed a significantly higher ratio of MUFA to SFA in muscle triglycerides than females on GD 70, although this difference was no longer significant on GD 90. On GD 70 and 90, males showed lower stearic acid content in the liver, and this acid is considered to protect metabolic health [72,73]. Males also showed higher ratios of MUFA to SFA and C18:1 to C18:0, indicating higher stearoyl-CoA desaturase activity. Although we observed few sex-related differences in the changes in phospholipid composition of muscle or liver during pregnancy, we did observe that females had higher content of linoleic acid in the liver than males on GD 70, and that males had higher content of mead acid in the muscle on GD 70 and 90. The greater content of mead acid in males likely reflects a worse homeorhesis state, since synthesis of mead or eicosatrienoic acid occurs in response to severe deficiency of fatty acids, mainly linoleic acid [74,75].

Animals and Experimental Procedures
The study involved a total of 56 foetuses obtained from 9 multiparous purebred Iberian sows with an average body weight of 147.7 ± 16.0 kg, which became pregnant after cycle synchronization with altrenogest (Regumate ® , MSD, Boxmeer, The Netherlands) and insemination with cooled semen from a purebred Iberian boar. All sows and the boar were homozygous for the LEPRc.1987T allele based on pyrosequencing [76].
Sows were fed with a standard grain-based food diet with the following mean component values: dry matter, 89.8%; crude protein, 15.1%; fat, 2.8%; and metabolizable energy, 3.0 Mcal/kg. Diet analysis (Table 8) showed that the most abundant fatty acids (FA) were palmitic (18.7%), oleic (23.2%) and linoleic acids (46.5%). The amount of food was adjusted to fulfil individual daily maintenance requirements, based on data from the British Society of Animal Science, from the start of the experimental period until GD 35 [77]. On this day, all sows were weighed and the amount of food offered to each sow was adjusted to fulfil 50% of their daily maintenance requirements, which has been shown to increase IUGR incidence [78]. Every day, a single food ration was weighed out and given to each sow in her individual pen; hence, the diet of each female was adjusted to her own weight.
Foetuses were obtained on GD 70 from five sows or on GD 90 from four sows. These gestational time points correspond to approximately 60% and 80% of a 112-day gestation typical for this breed, corresponding to 24 and 32 weeks of human pregnancy, respectively. These time points were chosen because until GD 70 of swine pregnancy, lipid anabolism and foetal metabolism resemble maternal metabolism [79]; around GD 90, foetal metabolism is independently regulated, and foetal development is more affected by nutrient availability [80].
On GD 70 and 90, blood samples were drawn from the orbital sinus of sows that had fasted for approximately 16 h. Samples were collected in sterile, heparinised 4-mL vacuum tubes (Vacutainer™ Systems Europe, Meylan, France) and were immediately centrifuged at 1500 g for 15 min. The plasma was separated and biobanked into polypropylene vials at −80 • C until they were assayed for metabolic biomarkers including glycaemic values and lipid profile.

Sampling of Placentas and Foetuses
Animals were euthanised by stunning and exsanguination, in compliance with RD53/2013 standard procedures, and the entire genital tracts were immediately collected for morphometric evaluation and foetal sampling. The contents of the uterus were exposed, and conceptus position was recorded. In each normal non-IUGR foetus, the allantoic and amniotic fluids were obtained by aspiration through the chorioallantoic and amniochorionic membranes, and blood samples were drawn from the heart and/or umbilical cord using heparinised syringes. Blood was processed as described above for sows, while allantoic and amniotic fluids were centrifuged at 1500× g for 15 min and supernatants were biobanked into polypropylene vials at −80 • C until they were assayed for metabolic biomarkers (including glycaemic values and lipid profile) and antioxidant/oxidative status. Rectangular sections of the uterine wall were collected, fixed in 4% paraformaldehyde, embedded in paraffin and stained using hematoxylin-eosin. Individual sections of the placentas were measured morphometrically as described [81] to obtain the average width of the placental folds, which served as an index of placental maturation.

Evaluation of Foetal Sex and Morphological Features
Foetal sex was determined by visual inspection immediately after recovery; detailed pregnancy-by-pregnancy information is shown in Table 9. Body length (crown-rump length), head size (occipito-nasal length and biparietal diameter) and corpulence (thoracic and abdominal circumferences) were measured in all the normal and IUGR foetuses. Total foetus weight was determined, then the head was separated from the trunk at the atlanto-occipital union, and head and trunk were weighed separately. All viscerae were obtained and weighed together immediately, and then the brain, heart, lungs, liver, intestine, kidneys, spleen and pancreas were weighed separately. Ratios of head-to-body weight, brain-to-head weight, and the weight of total viscera and individual organs (brain, heart, lungs, liver, kidneys, intestine, pancreas, and spleen) relative to body weight were calculated. Table 9. Detailed information (total number and percentage) on sex distribution (total, male and female foetuses) in the pregnancies studied in the current trial. Samples from liver, kidneys, duodenum and ileum were fixed in 4% paraformaldehyde, embedded in paraffin and stained with hematoxylin-eosin. The degree of maturation of these organs was assessed by evaluating the number of glomeruli with a well-defined glomerular tuft in the kidney [82], the decrease of hepatic haematopoiesis in the liver [83] and villus height and crypt depth in the intestine [84]. This assessment was made by an investigator blinded to sow and foetus details.

Evaluation of Maternal and Foetal Metabolic Status
Lipid profile parameters (triglycerides, total cholesterol, HDL-c, LDL-c) were measured in maternal plasma as well as in plasma, allantoic and amniotic fluids of normal foetuses. Assays were performed using a clinical chemistry analyser (Saturno 300 plus, Crony Instruments s.r.l., Rome, Italy), according to the manufacturer's instructions.

Evaluation of Foetal Adiposity and Fat Composition
Total fat was quantified in carcasses, and the total fat percentage and fatty acid composition (in g/100) were determined in intramuscular fat and liver of normal foetuses. For this purpose, samples from longissimus dorsi muscle and the left liver lobe were biobanked at −80 • C until they were assayed as described [85]. Intramuscular fat and liver fat were extracted from 300 mg of lyophilised and homogenised samples using the Ball-mill procedure [86]. Fatty acids in the total lipid extracts were identified and quantified by gas chromatography (HP6890, Hewlett Packard, Avondale, PA, USA) after methylation, as described in [87,88]. Fatty acid methyl esters were fractionated on a cross-linked polyethylene glycol capillary column (30 cm × 0.32 mm × 0.25 µm, Hewlett Packard Innowax) and a temperature gradient from 170 • C to 245 • C. The injector and detector were maintained at 250 • C. Neutral lipid fractions (triglycerides) and polar lipid fractions (phospholipids) were analysed using gas chromatography after passing them through aminopropyl minicolumns previously activated with 7.5 mL of hexane as described [89]. The percentages of individual fatty acids were used to calculate proportions of SFA, MUFA and PUFA, as well as total n-3 and n-6 and their ratio (∑n-6/∑n-3). The unsaturation index (UI) was obtained from the ratio of MUFA to SFA, and the activity of the stearoyl-CoA desaturase enzyme 1 was inferred from the ratio of the enzyme product, oleic acid (C18:1n-9), to the enzyme substrate, stearic acid (C18:0).

Evaluation of Foetal Antioxidant/Lipid Oxidative Status
In normal foetuses, values for total antioxidant capacity were assayed using the ferric reducing antioxidant power assay (FRAP) as previously described [90], while lipid peroxidation was assessed by measuring MDA (mmol/mL) using the thiobarbituric acid reaction and HPLC separation with fluorescence detection as previously described [91].

Statistical Analyses
Data were analysed using SPSS 22.0 (IBM Corp., Armonk, NY, USA). Based on previous studies [42], foetuses with severe growth restriction were defined as those with a body weight more than one standard deviation below the litter mean value. Among the 33 foetuses collected on GD 70, four (11%) were classified as showing severe growth restriction. Among the 23 foetuses collected on GD 90, four (17.4%) were so classified. Effects of gestational age (GD 70 vs. 90), sex (female vs. male) and growth restriction on developmental traits, adiposity, fatty-acid composition, and metabolic and foetal oxidative status were assessed by two-way ANOVA. Duncan's post-hoc test was performed to check differences among groups in multiple comparisons. Relationships between maternal metabolic biomarkers and features of foetuses showing normal growth were explored using Pearson correlation. The sow was considered the experimental unit for all variables in order to avoid biasing the results according to litter size: foetuses with the same sex and development (normal or IUGR) from the same sow were averaged together, giving one data point per sow. All results were expressed as mean ± SEM and statistical significance was accepted from p < 0.05.

Conclusions
The present study supports the importance of the maternal lipid profile and placental transfer of lipids, mainly essential fatty acids, for foetal development, and it confirms and extends previous studies suggesting that female foetuses show better adaptive responses in the form of greater synthesis of non-essential fatty acids and better transfer of essential fatty acids, resulting in better metabolic indexes. These sex-related differences were observed primarily in the neutral lipid fraction (triglycerides), suggesting a strong influence of sex on postnatal energy partitioning. These results may help guide future studies on understanding and optimising the maternal diet and placenta transfer capacity in order to meet foetal requirements for fatty acids and prevent the postnatal problems associated with insufficient prenatal fatty acid availability.