Metabolic regulation at the animal level

It is well known that metabolisable energy utilisation increases with energy intake (as does protein deposition with protein intake), but feed efficiency varies depending on the feeding level. For instance, in growing cattle, feed efficiency expressed as retained energy/supplied energy increases to reach a maximum with increasing food supply and then decreases at high-energy level (Geay and Robelin, Reference Geay and Robelin1979). Different mechanisms may explain this observation. One of them is of course the partitioning of fuels between protein deposition in muscles and fat storage in adipose tissue; this process will be illustrated in the sections that follow. Another important mechanism is the variable contribution of the metabolic activity of liver and portal-drained viscera (PDV) to total energy expenditure.

The liver and the PDV indeed have a great importance in total energy expenditure since their metabolic rates are up to 20 and six times higher, respectively, than that of the hind limb (composed approximately of two-thirds muscle). Thus, despite their lower mass relative to muscles, PDV and liver contribute to 17% and 13% to total energy expenditure v. 18% for muscles in pre-ruminant calves. However, due to the development of the digestive tract and the high ingestion of forages, relative contributions are 16% and 29%, 17% and 31%, 18% and 20% for PVD, liver and muscles, respectively, in weaned, growing and adult ruminants (for review, see Ortigues and Doreau, Reference Ortigues and Doreau1995; Bauchart et al., Reference Bauchart, Ortigues, Hocquette, Gruffat and Durand1996; Chilliard et al., Reference Chilliard, Bocquier and Doreau1998). Similar findings have been reported with regard to the contribution of different tissues to whole-body protein synthesis, with for example relative contributions of 18%, 24% and 25% for intestine, liver and muscles, respectively, in the rat (Waterlow, Reference Waterlow1984). Consequently, the metabolic fate of nutrients in the splanchnic tissues is of prime importance to control nutrient supply and use in peripheral tissues. Splanchnic tissue energy expenditure is greatly influenced by the level of nutrition. In ruminants, the increment in whole-animal energy expenditure with increased intake originates mainly from the PDV (17% to 61%) and the liver (16% to 44%) rather than from the muscle mass (5% to 7%) (Figure 1). Modelling approaches are now being applied to these results and clear curvilinear response equations of splanchnic energy expenditure to increasing intake have been quantified in sheep and cattle (Bermingham et al., 2007).

Figure 1 Origins in the increase in whole-body energy expenditure with increasing food intake in ruminants.

Increases in the weights of splanchnic tissues with increasing intake are primarily responsible for this effect although metabolic rates can also be modified (Lobley, Reference Lobley1991). For instance, in the PDV metabolic rate increases postprandially by approximately 20% as a result of digestion, nutrient absorption and metabolism in pre-ruminants and to a lower extent in ruminants (for review, see Ortigues and Doreau, Reference Ortigues and Doreau1995 also Bauchart et al., Reference Bauchart, Ortigues, Hocquette, Gruffat and Durand1996). Changes in metabolic rates are also associated with changes in the balance among nutrients supplied, and response equations of nutrient absorption to increasing intake are starting to be established (Bermingham et al., 2007).

For example, undernutrition is mediated by changes in nutrient supply (low dietary energy and protein intake), by changes in nutrient balance (increased muscle use of non-esterified fatty acids (NEFAs) and ketone bodies at the expense of acetate and glucose) and by changes in the hormonal status (low thyroid hormone, insulin and insulin-like growth factor-1 (IGF-1) plasma levels, high growth hormone (GH), glucocorticoid and epinephrine plasma concentrations) (for review, see Chilliard et al., Reference Chilliard, Bocquier and Doreau1998).

One animal type that is unique due to its very high-energy requirement is the dairy cow. Indeed, the huge energy demand for lactation in high-producing dairy cows largely overwhelms the energy expended for maintenance or for physical activity (for review, see Chilliard, Reference Chilliard1999). For instance, maintenance energy requirement can be estimated at about no more than 25% of the total energy requirement of a cow producing 45 kg of milk per day compared with 60 to 80% in humans and rats, or in a dry, non-pregnant cow. This makes the dairy cow an important model for studying the response to nutrients at the whole-body level in terms of food intake and metabolic regulation. Firstly, it has been shown that cereal grains that are highly digestible in the rumen can depress food intake in lactating cows compared with starch digestion in the small intestine. It is likely that the high production of propionate associated with starch intake causes satiety. The mechanisms explaining the regulation of intake by propionate are not known. A great deal of research suggests, however, that feeding behaviour is regulated by the hepatic oxidation of fuels (Allen et al., Reference Allen, Bradford and Harvatine2005). A reduction in feed intake in the periparturient period is a significant problem in dairy cows, which increases the risk of a high negative energy balance with associated body lipid mobilisation and metabolic diseases. However, a relatively slow increase in feed intake could also be seen as a safety mechanism allowing the development, over a longer period, of the digestive and metabolic adaptations necessary for the use of large amounts of nutrients at the time (Bareille et al., Reference Bareille, Faverdin and Hay1997). One mechanism limiting feed intake may be associated with the increased lipolysis which liberates a high amount of fatty acids taken up by the liver. The reduction in insulin and adenosine and increase in β-adrenergic sensitivities of adipose tissues associated with low insulin and high GH levels favour lipolysis (Chilliard et al., Reference Chilliard, Ferlay, Faulconnier, Bonnet, Rouel and Bocquier2000 and Reference Chilliard, Lerondelle, Disenhaus, Mouchet and Paris2001; Allen et al., Reference Allen, Bradford and Harvatine2005). Furthermore, negative energy balance and low insulinaemia make the liver IGF-1 secretion resistant to the GH stimulation (Chilliard et al., Reference Chilliard, Lerondelle, Disenhaus, Mouchet and Paris2001). This GH resistance allows the catabolic effects of this hormone to take priority over its anabolic effects in extra-mammary tissues and the occurrence of its galactopoietic effects in the mammary gland.

To better understand the biological mechanisms explaining the metabolic regulation at the animal level, we need to study the effects of nutrients more deeply. These effects are complex and involve different direct and indirect mechanisms, for instance, acute or long-term enzyme regulation or a stimulation of the secretion of some hormones, i.e. insulin by specific nutrients (mainly glucose). It is noteworthy that insulin is a major regulator of glucose metabolism, but is also the primary hormone known to regulate protein metabolism: insulin exerts its action through the two components of protein turnover which determine protein accretion or loss (i.e. protein synthesis and proteolysis). Insulin regulates protein metabolism in animals by clearly decreasing in vivo proteolysis (for review, see Tesseraud et al., 2007). Interestingly, studies using clamp techniques have demonstrated that the ability of physiological insulin to inhibit protein degradation is enhanced during early lactation (greater sensitivity of proteolysis to insulin; Tesseraud et al., Reference Tesseraud, Grizard, Debras, Papet, Bonnet, Bayle and Champredon1993): this adaptation may provide a mechanism limiting excessive mobilisation of body protein and preserving lean mass despite the significant diversion of amino acids towards mammary gland protein synthesis. This mechanism originates from an amino acid deficit during early lactation since hyperaminoacidaemia abolishes the phenomenon (amino acids are no longer limiting in hyperaminoacidaemia). Therefore, insulin plays a major role during early lactation, since this hormone has an ‘amino acid sparing’ effect in a physiological situation where the dietary requirements are very high. Such a mechanism is complementary to the fine and complex regulation described above for energy metabolism.

Despite the fact that insulin clearly stimulates protein synthesis in vitro (Grizard et al., Reference Grizard, Picard, Dardevet, Balage and Rochon1999; Tesseraud et al., 2007), the effects of insulin on protein synthesis are much more difficult to demonstrate in vivo. A positive insulin effect has been reported only in studies performed in young postabsorptive or food-deprived monogastric mammals. In these animals, plasma insulin concentrations are low in the fasting state and relatively high after a meal. In contrast, ruminants have more constant nutrient absorption and insulinaemia, which may limit the anabolic response to exogenous insulin. Results obtained by Wester et al. (Reference Wester, Lobley, Birnie, Crompton, Brown, Buchan, Calder, Milne and Lomax2004) further support these concepts since elevation of insulin failed to stimulate muscle protein synthesis in fasted lambs. The hypothesis of protein synthesis that is very sensitive to insulin, with maximal responses observed at relatively low insulinaemia, has been confirmed by transiently decreasing insulinaemia in rodents with diazoxide injection: postprandial insulin suppression greatly decreased protein synthesis, suggesting that insulin plays a major role in regulating protein synthesis in vivo (Sinaud et al., Reference Sinaud, Balage, Bayle, Dardevet, Vary, Kimball, Jefferson and Grizard1999; Balage et al., Reference Balage, Sinaud, Prod’homme, Dardevet, Vary, Kimball, Jefferson and Grizard2001; Prod’homme et al., Reference Prod’homme, Balage, Debras, Farges, Kimball, Jefferson and Grizard2005). Similar findings have also been reported in pigs using somatostatin to block endogenous insulin (O’Connor et al., Reference O’Connor, Bush, Suryawan, Nguyen and Davis2003a and Reference O’Connor, Kimball, Suryawan, Bush, Nguyen, Jefferson and Davis2003b). However, even under these particular conditions, the role of insulin is a matter of debate and interactions between insulin and amino acids have to be considered: for instance, the positive effect of insulin on in vivo skeletal muscle protein synthesis is more marked in the presence of amino acids (Garlick and Grant, Reference Garlick and Grant1998). Interestingly, recent studies have suggested that a minimum level of insulin is required for the stimulation of protein synthesis by refeeding or the absorption of amino acids across the PDV (reviewed by Kimball et al., Reference Kimball, Farrell and Jefferson2002). Insulin may thus have a permissive effect for the amino acid-induced stimulation of protein synthesis.

Hepatic metabolism

The liver is characterised by a very intense metabolic rate with both elevated anabolic and catabolic activities and particularly high rates of protein turnover. Generally speaking, increased hepatic oxidation of fuels likely causes satiety (thus reducing food intake), whereas increased clearance of fuels from the blood is expected to cause hunger (Allen et al., Reference Allen, Bradford and Harvatine2005). In addition, a high food intake increases liver energy expenditure and favours esterification of fats at the expense of their oxidation. In ruminants, unlike in other species, a high food intake also favours the hepatic gluconeogenic rate due to a high dietary supply of propionate (a major glucose precursor). These changes result from short-term mechanisms of regulation through the action of metabolites and by long-term mechanisms of regulation through changes in gene expression. Gluconeogenesis is also highly regulated by hormonal factors such as insulin (Drackley et al., Reference Drackley, Overton and Douglas2001).

With respect to fat metabolism, the liver is able to synthesise fatty acids (although much less in ruminants and pigs than in other species) and oxidise them. The balance between these two processes is of prime importance for the delivery of energy in the form of fat (short-chain fatty acids, ketone bodies, long-chain fatty acids (LCFAs)) to other tissues. When a high mobilisation of fat occurs from adipose tissue (in the case of energy deficiency), the liver, which is faced with a high uptake of fatty acids, oxidises more fats. The enzyme called carnitine palmitoyltransferase I (CPT I) catalyses the rate-limiting step of fat oxidation because it controls LCFA transfer into mitochondria. It is thus also involved in the control of ketone body production. CPT I activity is inhibited by malonyl-CoA, an intermediate in LCFA synthesis from acetyl-CoA, the synthesis of which is enhanced by insulin, which stimulates lipogenesis (i.e. the conversion of acetyl-CoA into malonyl-CoA). Consequently, any change in insulin secretion or in LCFA supply to the liver modifies the balance of fat within the liver and therefore the supply of fat to peripheral tissues (for review, see Hocquette and Bauchart, Reference Hocquette and Bauchart1999). Other studies have shown that oxidation of fatty acids provides ATP needed for gluconeogenesis supporting strong interrelationships between glucose and fatty acid metabolism in the liver (Drackley et al., Reference Drackley, Overton and Douglas2001).

Fat metabolism is also of interest in avian species since de novo lipogenesis is essentially hepatic in these animals (Pearce, Reference Pearce1977). Farmers have taken advantage of this species specificity to produce fatty liver from ducks and geese following overfeeding of the animals. In the liver of ducks, overfeeding with a carbohydrate-rich diet induces a significant increase in the activity of the main enzymes involved in lipogenesis from glucose (glucokinase, glucose-6-phosphate dehydrogenase, malic enzyme, acetyl CoA carboxylase). Cytochrome-c oxydase activity, on the other hand, remains unchanged, indicating that the overall oxidation ability of energy-yielding substrates is not regulated so much. Plasma levels of insulin and triglycerides (TGs) are considerably increased and glycaemia is significantly higher. Some differences exist, however, between genotypes: compared with the Pekin duck, the Muscovy duck probably has a more efficient hepatic lipogenesis (since acetyl-CoA carboxylase activity is higher) but its ability to export lipids is probably lower (since plasma TG levels are lower) (Chartrin et al., Reference Chartrin, Bernadet, Guy, Mourot, Hocquette, Rideau, Duclos and Baéza2006).

To summarise, an important role of the liver is therefore its participation in the regulation of food intake by contributing to the matching of energy supply to energy use. In addition, all portal vein blood must traverse the liver before reaching the other tissues and the rest of the body and the liver also plays a key metabolic role in converting the different metabolites. Consequently, the liver metabolism has a great impact on the availability of nutrients for other tissues and organs.

Muscle and adipose tissue metabolism

Muscle tissues use carbohydrates and fatty acids as energy sources for the production of free energy (ATP). This general process is regulated by these two classes of nutrients as well as by many hormonal factors, mainly insulin. It is also highly regulated by physical activity (which itself increases muscle insulin sensitivity), but it is outside the scope of the paper.

States of insulin resistance or of reduced potential of glucose use by muscles may impair growth, for instance in calves (for review, see Bauchart et al., Reference Bauchart, Ortigues, Hocquette, Gruffat and Durand1996 also Hocquette and Abe, Reference Hocquette and Abe2000). Many factors have been shown to be involved in the aetiology of insulin resistance. They include heredity, age, obesity, diet, sex, sedentary life style, etc. (Khan and Safdar, Reference Khan and Safdar2003). Impaired glucose transport rate, as well as impaired mitochondrial activity, and higher intramyocellular lipid content were shown in the insulin-resistant offspring of patients with type 2 diabetes (Petersen et al., Reference Petersen, Dufour, Befroy, Garcia and Shulman2004). Recently, total muscle fat oxidative capacity was also shown to be positively correlated with insulin sensitivity (Rimbert et al., Reference Rimbert, Boirie, Bedu, Hocquette, Ritz and Morio2004). In addition, insulin-regulated mitochondrial gene expression is positively associated with glucose flux in human skeletal muscle (Huang et al., Reference Huang, Eriksson, Vaag, Lehtovirta, Hansson, Laurila, Kanningen, Thuesen-Olesen, Kurucz, Koranyi and Groop1999). Mitochondrial markers, but not muscle fibre type per se, may indeed be predisposing factors for obesity (Raben et al., Reference Raben, Mygind and Astrup1998). In addition, several nutritional factors regulate the action of insulin. For instance, there is evidence that insulin resistance may be caused by excess nutrient supply (for review, see Proietto et al., Reference Proietto, Filippis, Nakhla and Clark1999). Other mechanisms linked to animal management or feeding strategy are thought to regulate glucose metabolism and insulin action: plasma insulin levels are indeed lower in steers at pasture compared with control animals in a shed with similar growth rates (I. Ortigues et al., unpublished results).

The effect of nutrient availability on muscle metabolism is subject to variations between animal species. Thus lipoprotein lipase activity, which is related to tissue potential for plasma TG fatty acid uptake, as well as plasma TGs, decreased by underfeeding ewes (Faulconnier et al., Reference Faulconnier, Bonnet, Bocquier, Leroux and Chilliard2001) or cows (Bonnet et al., Reference Bonnet, Faulconnier, Hocquette, Bocquier, Leroux, Martin and Chilliard2004), contrary to what has been observed in rodents. This peculiarity is likely related to the fact that the ruminant liver is less able than rodent liver to recycle fatty acids mobilised from adipose tissue during undernutrition. Inversely, muscle GLUT4 glucose transporter was less responsive to underfeeding–refeeding in bovines than in rats in parallel to lower relative changes in plasma glucose and insulin (Bonnet et al., Reference Bonnet, Faulconnier, Hocquette, Bocquier, Leroux, Martin and Chilliard2004).

In ducks, muscle overall oxidative ability, estimated by cytochrome-c oxydase activity, is increased in pectoralis major muscle following overfeeding. Generally speaking, the muscle tissues adapt their energy metabolism to overfeeding, increasing their oxidative metabolism, and decreasing their glycolytic metabolism, particularly in breast muscle (Zanusso et al., Reference Zanusso, Rémignon, Guy, Manse and Babilé2003; Chartrin et al., Reference Chartrin, Bernadet, Guy, Mourot, Hocquette, Rideau, Duclos and Baéza2006). It has been hypothesised that muscle tissues adapt their metabolism to use more fat as the energy source following overfeeding due to greater fat availability (as demonstrated by higher TG levels in the plasma) and reduce their utilisation of glucose (Chartrin et al., Reference Chartrin, Bernadet, Guy, Mourot, Hocquette, Rideau, Duclos and Baéza2006).

In cattle, muscle metabolism is more affected by mild nutritional restriction than any other muscle biochemical characteristic. A mild nutritional restriction followed by ad libitum feeding affected the metabolism of muscles, although their histological profile was preserved. Interestingly, muscles responded in different directions to refeeding after a restriction period in terms of metabolic activity (Figure 2) and collagen solubility, without any modification of their lipid content or of the characteristics of their proteasome system (Cassar-Malek et al., Reference Cassar-Malek, Hocquette, Jurie, Listrat, Jailler, Bauchart, Briand and Picard2004). The muscles reacted differently to the nutritionally induced discontinuous growth path, suggesting that muscle types display a differential nutritional sensitivity. These results confirmed that muscle metabolic characteristics are in a dynamic state and are subject to plasticity in adjusting to feed and hormonal signals.

Figure 2 Influence of restriction and refeeding on lactate dehydrogenase (LDH), and cytochrome-c oxidase (COX) activities across the three muscles (according to Cassar-Malek et al., Reference Cassar-Malek, Hocquette, Jurie, Listrat, Jailler, Bauchart, Briand and Picard2004). At 9 months of age, steers were fed a restricted diet for 3 months and slaughtered or subject to a 4-month ad libitum refeeding period prior to slaughter with the same regimen. Control steers were fed to gain continuously between 9 and 12 months of age and slaughtered (end of the restriction period) or maintained on a continuous feeding protocol through 16 months of age prior to slaughter (end of the refeeding period).

Generally, increased adipose tissue mass (due for instance to a high feeding level) and insulin resistance in skeletal muscle mass are closely related. Communication between the two tissues is thought to play a major role in the development of insulin resistance. One possible mechanism is associated with free fatty acids and their metabolites. A defective oxidative phosphorylation due, for instance, to a low mitochondrial activity is thought to be the underlying mechanism (Sell et al., Reference Sell, Dietze-Schroeder and Eckel2006). Another mechanism is the release of endocrine and metabolic mediators by adipose tissue. One of these mediators is leptin, whose secretion by adipocytes increases when body fatness and/or sugar intake increases, stimulates energy expenditure and fatty acid oxidation in muscles, liver and adipose tissue and inhibits adipose tissue lipogenesis (Havel, Reference Havel2004). The development of leptin resistance when body fatness and/or short-term energy intake increases is, however, a limit to this homeostatic regulation (Unger, Reference Unger2005). The stimulation of energy expenditure by leptin may partly explain why fat ewes lost body lipids, contrary to lean ones, and showed increased basal and adrenergic lipolytic activities in adipose tissue, when fed for a long period at their theoretical maintenance energy requirements (Chilliard et al., Reference Chilliard, Ferlay, Faulconnier, Bonnet, Rouel and Bocquier2000). Reciprocally, lactation hypoleptinaemia likely decreases energy expenditure and favours high milk energy secretion during early lactation, as well as body reserve rebuilding after the lactation peak (for a review, see Chilliard et al., Reference Chilliard, Delavaud and Bonnet2005).

Metabolic regulation at the animal level

Many arguments exist to support the idea that energy and protein metabolism are regulated by the nature of nutrients and therefore depend on the type of diet.

In weaned growing ruminants, efficiency of metabolisable energy utilisation for growth is generally low (less than 40%) with forage diets, for which the percentage of energy absorbed as volatile fatty acids (VFA) averages 66% (33% for acetate, 14% for propionate and 19% for butyrate). By contrast, efficiency of energy utilisation is higher (more than 50%) with maize-based diets, which supply about 50% of absorbed energy as VFA (13% by acetate, 18% by propionate and 13% by butyrate). This lower efficiency with forage diets is thought to result from higher heat losses from food fermentation in the rumen, higher energy expenditure of PDV due to an increased PDV mass and an increased PDV motricity as well as lower metabolic efficiency of VFA utilisation (for review, see Ortigues and Visseiche, Reference Ortigues and Visseiche1995). When animals are at pasture, a high-energy expenditure of muscles due to physical activity is also likely to contribute to these observations.

Energy metabolism can be specifically regulated by carbohydrates. For instance, glucose is known to be hypophagic in many non-ruminants, but not in ruminants. This is likely to be due to differences in hepatic oxidation of glucose. Indeed, the ruminant liver does not remove a significant part of the circulating glucose. Hypophagic effects of propionate have been well documented in ruminants although inconsistent effects have been observed. Propionate can be used for gluconeogenesis which consumes ATP, or is catabolised as acetyl-CoA, therefore generating ATP. Insulin and glucose are important regulators of fuel partitioning within the liver, especially in monogastrics. When glucose and insulin levels are high, gluconeogenesis is low, propionate is oxidised faster and satiety occurs sooner. Propionate is also an insulin secretagogue and insulin is a putative satiety hormone, thus possibly amplifying the phenomena (Allen et al., Reference Allen, Bradford and Harvatine2005). In human beings, it is well known that increased carbohydrate intake will make a person fatter. One mechanism is the well-known higher expression of genes and proteins involved in fat synthesis. The first and quantitatively most important response to a high carbohydrate intake is, however, increased carbohydrate oxidation at the expense of fat oxidation. In addition, hepatic glucose production is higher despite higher insulin levels and lipolysis is reduced. Thus, the liver is more likely to contribute to obesity by overproducing glucose and reducing fat oxidation than by converting glucose into fat (Hellerstein, Reference Hellerstein2003).

Again, the dairy cow may be a good model to study the regulation of energy metabolism by the nature of nutrients. Besides the high-energy requirements of dairy cows, feeding strategies specific to this animal type have brought interesting results regarding the metabolic regulations. For instance, fat sources are often added to dairy cow diets to support the high milk production. Furthermore, as in human beings, it has been observed that unsaturated fatty acids decreased food intake. Generally speaking, it is likely that oxidised fat is satiating whereas fat that is stored is not satiating in the short term. Saturated fatty acids are not oxidised as rapidly as polyunsaturated fatty acids (PUFAs). This may explain the differential regulation of food intake according the nature of fatty acids although other mechanisms are likely to occur, such as gut peptide release during meals (Allen et al., Reference Allen, Bradford and Harvatine2005). The combined effects of an increase in milk energy secretion and a decrease (or no change) in energy intake paradoxically result in lower energy balance and higher body-weight loss in early lactation cows fed supplemental lipids (Chilliard, Reference Chilliard1993). Furthermore, unsaturated fatty acids probably increased these effects since postruminally infused plant oil stimulated β-adrenergic lipolytic responses in adipose tissue (Gagliostro and Chilliard, Reference Gagliostro and Chilliard1991).

Amino acids have also a major effect in controlling metabolism (see Grizard et al., Reference Grizard, Dardevet, Papet, Mosoni, Patureau-Mirand, Attaix, Tauveron, Bonin and Arnal1995 and Reference Grizard, Picard, Dardevet, Balage and Rochon1999; Lobley, Reference Lobley1998; Muramatsu, Reference Muramatsu1990; Nieto and Lobley, Reference Nieto and Lobley1999; Prod’homme et al., Reference Prod’homme, Rieu, Balage, Dardevet and Grizard2004 for reviews). They are first substrates for protein synthesis and the lack of even a single essential amino acid causes a decrease in protein synthesis. Amino acid availability affects not only protein synthesis rates, but also proteolysis and amino acid oxidation, therefore affecting various pathways involved in metabolism. Evidence of the action of amino acids in protein turnover has been provided using diets varying in protein or amino acid supplies. For example, protein-deprived diets cause inhibition of protein synthesis, whereas the effect on proteolysis seems to depend on the severity of deficiency. Excess protein appears to have only a minor effect on muscle protein metabolism, except when it is associated with energy restriction. In this case, it results in reduction in protein synthesis. Nevertheless, under some critical conditions, excess proteins can have a positive effect on protein deposition. For example, in young piglets, protein-rich diets prevent the decrease in protein synthesis after early weaning (Seve et al., Reference Seve, Reeds, Fuller, Cadenhead and Hay1986). In ‘normal’ nutritional situations, whatever the species (man, pig or goat), hyperaminoacidaemia induced by intravenous infusion of amino acids increases muscle protein synthesis, at least when measurements are performed using the constant infusion of labelled amino acids (Bennet et al., Reference Bennet, Connacher, Scrimgeour, Smith and Rennie1989; Watt et al., Reference Watt, Corbett and Rennie1992; Tesseraud et al., Reference Tesseraud, Grizard, Debras, Papet, Bonnet, Bayle and Champredon1993).

The role of amino acids in regulating protein turnover has also been studied in refeeding conditions, i.e. postprandial regulation. Volpi et al. (Reference Volpi, Lucidi, Cruciani, Monacchia, Reboldi, Brunetti, Bolli and DeFeo1996) showed in adult men that amino acids contributed to 90% of the increase in protein synthesis following a meal with balanced levels of lipids, carbohydrates and proteins. The fact that amino acids are essential to postprandial stimulation of protein synthesis has also been demonstrated in rats (Yoshizawa et al., Reference Yoshizawa, Nagasawa, Nishizawa and Funabiki1997 and Reference Yoshizawa, Kimball, Vary and Jefferson1998) and chickens (Yaman et al., Reference Yaman, Kita and Okumura2000). Stimulation of protein synthesis by amino acids is thus associated with an increase in translational efficiency. Protein turnover also depends on the supply of specific amino acids, particularly limiting amino acids. Among them, methionine is, for example, the first limiting factor in classical diets used for growing chickens because of a high requirement of sulphur amino acids for the synthesis of feathers whereas poultry diets based on corn and soya-bean meal are deficient in sulphur amino acids without supplementation (Baker, Reference Baker2006). In this species, the decreased rates of body-weight gain and protein accretion induced by methionine and cysteine deficiency are apparently more pronounced than those recorded with diets deficient in lysine or histidine (Kino and Okumura, Reference Kino and Okumura1986). Decreased growth in chickens fed methionine- and cysteine-free diets was primarily caused by lower rates of whole-body protein synthesis associated with lower RNA efficiency, suggesting translational regulation. A positive effect of methionine supplementation on muscle growth has conversely been reported, since the addition of methionine to a methionine-deficient diet, otherwise balanced in terms of other amino acids, increases accretion and synthesis of protein in the gastrocnemius and pectoralis major muscles in chickens (Barnes et al., Reference Barnes, Calvert and Klasing1995). Methionine supplementation in a protein-free diet is also effective in increasing protein synthesis of whole-body protein, particularly muscle protein (Muramatsu et al., Reference Muramatsu, Salter and Coates1985 and Reference Muramatsu, Kato, Tasaki and Okumura1986). Webel and Baker (Reference Webel and Baker1999) found that cysteine supplementation elicits a response similar to or greater than methionine and suggested that the primary need is for cysteine, and not for methionine per se; like methionine, cysteine is an important constituent of feather and fur proteins. Wool production is quite a nice example of increasing sulphur amino acid requirements.

Amino acids exert their action in various tissues and organs, including the splanchnic tissues, which are able to control the distribution of nutrients between tissues since they are the first tissues exposed to newly absorbed nutrients due to their anatomical position (first-pass splanchnic extraction; see Nieto and Lobley (Reference Nieto and Lobley1999) for a review). Despite this function and the fact that the portal delivery of amino acids possibly initiates a signal to the liver that increases hepatic amino acid utilisation, i.e. protein synthesis, these findings are not discussed further here. The nutritional regulation of protein metabolism by amino acids has also been intensively studied in skeletal muscle since these nutrients are known as anabolic factors, which induce protein gain by stimulating protein synthesis while inhibiting proteolysis. The aim of the work performed is thus to improve and control muscle growth and meat quality in animal production and to reduce muscle wasting in some physiological (e.g. early lactation) and physiopathological situations (e.g. ageing, infection). Indeed, due to the differences between the amino acid composition of acute-phase and muscle proteins, a considerable amount of muscle protein must be degraded to provide the amino acids used in the acute-phase response (Reeds et al., Reference Reeds, Fjeld and Jahoor1994). Understanding the mechanisms by which amino acids regulate protein metabolism is therefore essential to improve control of the use of nutrients and optimise dietary amino acid supply.

Finally, the regulation of insulin sensitivity is an important mechanism by which the nature of nutrients indirectly regulate energy and protein metabolism. Glucose oxidation is usually decreased in insulin-resistant human subjects. An increase in lactose intake increases insulin resistance in preruminant calves (for review, see Hocquette and Abe, Reference Hocquette and Abe2000) as described in other species (for review, see Hocquette et al., Reference Hocquette, Ortigues-Marty, Pethick, Herpin and Fernandez1998). By contrast, an increase in digestible protein intake at a constant protein-free energy intake decreases insulin resistance unlike the situation in humans. In rats and humans, fat intake has been associated with the development of insulin resistance, and plant oil duodenal infusion lowers glucose tolerance and response to insulin in lactating cows (Chilliard and Ottou, Reference Chilliard and Ottou1995).

Hepatic metabolism

Energy metabolism is regulated by the nature of nutrients at the whole-body level due to tissue- or organ-specific regulations, which are of course important in the liver when strong interrelationships take place between fat and carbohydrate metabolism.

Propionate, which is absorbed in large quantities in ruminants, directly inhibits fatty acid oxidation and, indirectly, entry of LCFA into mitochondria. Indeed, CPT I is inhibited by methylmalonyl-CoA which derives from propionate. In addition, succinyl-CoA generated from propionate inhibits ketogenesis. A low delivery of propionate to the liver, due for instance to a forage-based diet would therefore enhance ketosis in the ruminant liver (for review, see Hocquette and Bauchart, Reference Hocquette and Bauchart1999). The majority of absorbed propionate is converted to glucose within the liver. Propionate, however, decreases the utilisation of other gluconeogenic substrates such as lactate (for review, see Brockman, Reference Brockman1993), thus modifying the balance of nutrients supplied to the muscles.

As in avian species, fatty liver may occur in the veal calf depending on nutritional factors, thereby affecting growth and health. One important factor that may regulate hepatic TG metabolism is the nature of dietary fatty acids such PUFAs or medium-chain fatty acids (MCFA). Compared with standard milk replacers for calves, PUFA-rich diets (soya-bean oil rich in C_18:2n-6) or MCFA-rich diets (coconut oil rich in C_12:0) lead to the development of hepatic TG infiltration (for review, see Bauchart et al., Reference Bauchart, Ortigues, Hocquette, Gruffat and Durand1996). This may be explained by a modification of LCFA partitioning between esterification and oxidation within peroxysomes and mitochondria (Piot et al., Reference Piot, Hocquette, Veerkamp, Durand and Bauchart1999). Insulin resistance also occurs in high-producing dairy cows and is also linked to disturbance in glucose homeostasis. Glucose and LCFA, the major nutrients in monogastrics and in the preruminant calf, were shown to control gene expression in laboratory rodents (for reviews, see Clarke and Abraham, Reference Clarke and Abraham1992 also Girard et al., Reference Girard, Perdereau, Foufelle, Prip-Buus and Ferré1994). For instance, in rodents, dietary glucose is a key determinant of hepatic transcription rate for several genes encoding various enzymes including phosphofructokinase, acetyl-CoA carboxylase or fatty acid synthase (for review, see Ferré, Reference Ferré1999). By contrast, PUFA are potent inhibitors of the expression of the genes encoding hepatic lipogenic enzymes. Glucose also influences mRNA stability or processing such as the editing of apolipoprotein B. The relative contribution of transcription v. transcript stability as determinants of mRNA abundance varies for each transcript and is dependent on specific tissues (for reviews, see Clarke and Abraham, Reference Clarke and Abraham1992 also Girard et al., Reference Girard, Perdereau, Foufelle, Prip-Buus and Ferré1994). Those regulations are of prime importance in the liver.

Basically, carbohydrate intake results in a high glucose level in plasma, which induces insulin secretion. Insulin in interaction with glucose plays a major role in linking fat and carbohydrate metabolism in the liver. Firstly, insulin suppresses hepatic glucose production. Insulin is also known to directly induce the expression of the transcription factor (sterol regulatory element-binding protein-1c (SREBP-1c)) expression, thereby enhancing the expression of lipogenic enzymes in the liver. However, in mice lacking SREBP-1c, glucose alone can also enhance the expression of lipogenic enzymes through the carbohydrate-responsive element-binding protein (ChREBP), the expression of which is enhanced in carbohydrate-fed rats (for review, see Uyeda and Repa, Reference Uyeda and Repa2006). This major regulation is altered in insulin-resistant states. Insulin resistance is defined here as the failure of insulin to suppress hepatic glucose production. For instance, in mice, accumulation of intracellular fatty acid metabolites such as long-chain fatty acyl-CoAs reduces the activity of phosphatidylinositol-3′ kinase (PI3K), a key step in the insulin signalling cascade. Conversely, the prevention of fat accumulation reduces hepatic insulin resistance. At a molecular level, insulin is known to suppress the activity of two enzymes involved in the control of gluconeogenesis, namely phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphate. This action is reduced or suppressed in insulin-resistant states. Elevated glucose concentration causes increased insulin secretion and activates the two transcriptional factors SREBP-1c and ChREB. In insulin-resistant states, SREBP-1c blocks hepatic insulin signalling thereby promoting hepatic glucose production (which is not or less inhibited by insulin due to insulin resistance). The failure of insulin to inhibit glucose production increases insulin secretion and causes hyperinsulinaemia. Insulin, SREBP-1c and ChREBP increase the activity of lipogenic enzymes leading to fat accumulation. On the other hand, insulin is known to inhibit fatty acid oxidation, and this occurs to a large extent in insulin-resistant states. Thus, fat accumulation results in fact from both increased fat synthesis and inhibition of fatty acid oxidation in insulin-resistant states due to a positive interaction between glucose level and insulin signalling, thereby altering the ability of insulin to reduce glucose production (Weickertt and Pfeiffer, Reference Weickert and Pfeiffer2006).

PUFAs are also known to interact with nuclear receptor proteins. This is important in human nutrition, since during the last 100 to 150 years, huge nutritional changes of the human population have led to an increase in saturated fats from dairy and meat products, in trans-fatty acids from the hydrogenation of vegetable oils and in n-6 fatty acids due to the production of oils from vegetable seeds. Again, the liver is the primary site for the metabolism of essential fatty acids (linoleic and linolenic acids of the n-6 and the n-3 families, respectively). Sequential alternating elongation and desaturation steps lead to PUFA with more carbons. The application of molecular biology techniques has shown that PUFA elicit changes in gene expression that precede changes in cellular composition. This is achieved by directly governing the activity of nuclear transcription factors, the first one which has been discovered being the peroxisome proliferator-activated receptor-α (PPAR-α). This factor plays a role in regulating the genes that control glucose and lipid metabolism. Generally speaking, PUFAs induce fatty acid oxidation and are potent suppressors of fat synthesis through the activation of PPARs. Indeed, a great number of genes involved in fatty acid oxidation or synthesis contain DNA sequences that are recognised by PPARs. Besides the action of PPAR-α, it was also shown that PPAR-γ induces the expression of lipoprotein lipase and fatty acid transporters. Other transcriptional factors (SREBP, liver X receptors) were shown to be involved in the action of PUFA on gene expression. For instance, liver X receptors act through the regulation of SREBP-1c, with PUFA suppressing the nuclear content of this factor. PUFA also suppress the enhancer activity of hepatic nuclear factor-4 (HNF-4), which is known to enhance the expression of some lipogenic enzymes (Benatti et al., Reference Benatti, Peluso, Nicolai and Calvani2004).

Muscle and adipose tissue metabolism

The muscle metabolism is highly regulated by hormonal factors, especially insulin, and several nutritional factors regulate the action of insulin. Firstly, increased muscle TG content, or a high proportion of saturated fats within muscles, is inversely related to insulin action (for review, see Hocquette et al., Reference Hocquette, Ortigues-Marty, Pethick, Herpin and Fernandez1998). Secondly, some dietary micronutrients may also modify the action of insulin. For instance, severe iron deficiency in veal calves induces an increase in glucose use by muscle directed to glycolysis and lactate production (for review, see Bauchart et al., Reference Bauchart, Ortigues, Hocquette, Gruffat and Durand1996). Chronic hyperglycaemia is also a cause of insulin resistance, for instance in intensively milk-fed calves. Studies in rats showed that a high rate of intracellular formation of hexosamine from glucose decreases glucose uptake, but has inconsistent effects on glycogen accumulation (Virkamaki et al., Reference Virkamaki, Daniels, Hamalainen, Utriainen, McClain and Yki-Järvinen1997).

Muscle and adipose tissue metabolism is also regulated by the nutrients themselves. For instance, lipogenesis has been shown to be regulated in vivo by glucose infusion in sheep (Ballard et al., Reference Ballard, Filsell and Jarrett1972). More recent research focused on facilitative glucose transporters, mainly the insulin-sensitive isoform, GLUT4, which controls extraction of plasma arterial glucose by muscle. The rate-limiting role of glucose transport in glucose homeostasis and muscle metabolism has been demonstrated by several in vivo and in vitro approaches. Indeed, the change in the nature of nutrients which occur at weaning has huge consequences for insulin action and hence muscle metabolism. For instance, in rats, the fat-rich milk is replaced by a high-carbohydrate diet characteristic of the adult. This change induces a higher insulin responsiveness of glucose utilisation and glucose transport rate in adipose tissues and muscles. These changes have been shown to be explained by a higher expression of GLUT4 (Girard et al., Reference Girard, Ferré, Pégorier and Duée1992). The situation in ruminants is not the same: the contribution of carbohydrates to total energy dietary supply decreases from the suckling period to adult nutrition since carbohydrates are degraded by the rumen micro-organisms which develop at weaning. Therefore, the enhanced regulation of GLUT4 does not occur at weaning in ruminants as it does in monogastrics (Hocquette et al., Reference Hocquette, Castiglia-Delavaud, Graulet, Ferré, Picard and Vermorel1997), which is in agreement with observations on the species-specific effects of underfeeding–refeeding on muscle metabolism and plasma metabolites and hormones reported in the previous section.

Finally, muscle energy metabolism is regulated by the interactions between carbohydrate and fatty acid metabolisms. High intracellular levels of NADH, ATP and acetyl-CoA derived from LCFA or acetate oxidation decrease glucose catabolism, mainly by inhibiting the pyruvate dehydrogenase activity. This biological mechanism is known as ‘the Randle cycle’. However, this competitive interaction between nutrients might not be as marked in the case of increased energy requirements resulting from higher rates of protein synthesis and deposition. Conversely, stimulation of carbohydrate catabolism, for instance by insulin, may inhibit LCFA oxidation (for review, see Faergeman and Knudsen, Reference Faergeman and Knudsen1997 also Hocquette et al., Reference Hocquette, Ortigues-Marty, Pethick, Herpin and Fernandez1998).

Metabolism is regulated by energetic nutrients, and also by amino acids with major consequences on muscle and adipose tissue development. Among amino acids, lysine is probably one of those exerting the most specific effects on carcass composition and muscle growth. In chickens, a particularly drastic effect has been observed on breast muscle development (Tesseraud et al., Reference Tesseraud, Maaa, Peresson and Chagneau1996b). It is worthwhile noting that neither threonine nor valine exhibits as pronounced an effect as lysine on body composition, as was observed in an experiment in which these three amino acids were studied together in similar conditions (Leclercq, Reference Leclercq1998). Interestingly, preliminary results have also shown that increasing lysine can improve chicken breast muscle quality by increasing its ultimate pH and water holding capacity (Berri et al., 2004), but the underlying mechanisms are still unknown.

Lysine supplementation to a lysine-deficient diet otherwise balanced in terms of other amino acids significantly increased the amounts of protein synthesised and degraded in chicken skeletal muscle (Tesseraud et al., Reference Tesseraud, Larbier, Chagneau and Geraert1992, Reference Tesseraud, Peresson, Lopes and Chagneau1996a, Reference Tesseraud, Maaa, Peresson and Chagneau1996b and Reference Tesseraud, Temim, Le Bihan-Duval and Chagneau2001). Similar results have been demonstrated in the growing pig (Salter et al., Reference Salter, Montgomery, Hudson, Quelch and Elliot1990) and in humans (Conway et al., Reference Conway, Bier, Motil, Burke and Young1980; Meredith et al., Reference Meredith, Wen, Bier, Matthews and Young1986). Studies performed using lysine-deficient diets have also revealed the possibility of a major effect of dietary amino acid levels on protein degradation: the fractional rates of proteolysis (values expressed as % per day) measured in the breast muscle of growing chickens are always higher in the lysine-deprived animals, irrespective of age or genotype (Tesseraud et al., Reference Tesseraud, Peresson, Lopes and Chagneau1996a and Reference Tesseraud, Temim, Le Bihan-Duval and Chagneau2001). It is probable that the breast muscles of chickens represent a major reservoir of body protein that can be mobilised in deficiency states. Increased proteolysis in such conditions provides free amino acids that are used for protein synthesis. A specific role of amino acids on proteolytic pathways has also been demonstrated. For example, in vitro studies have clearly shown that amino acid deficiency stimulates autophagic (lysosomal degradation) and proteasome-mediated proteolysis (Meijer and Dubbelhuis, Reference Meijer and Dubbelhuis2004; Yang et al., Reference Yang, Liang, Gu and Qin2005; Bechet et al., Reference Bechet, Tassa, Combaret, Taillandier and Attaix2005; Hamel et al., Reference Hamel, Fawcett, Bennett and Duckworth2004). Autophagic and ubiquitin-proteasome-dependent proteolysis are two major pathways responsible for skeletal muscle proteolysis (Attaix et al., Reference Attaix, Combaret, Pouch and Taillandier2001 and Reference Attaix, Mosoni, Dardevet, Combaret, Mirand and Grizard2005). In particular, ubiquitin–proteasome-dependent proteolysis degrades the bulk of muscle proteins, including myofibrillar proteins, after their targeting by ubiquitin.

With regard to protein synthesis, studies conducted over the last 10 years indicate that, in addition to be substrates for protein synthesis amino acids act as mediators of metabolic pathways in the same manner as certain hormones (e.g. insulin) (Kimball and Jefferson, Reference Kimball and Jefferson2004 and Reference Kimball and Jefferson2006; Dann and Thomas, Reference Dann and Thomas2006; Tesseraud et al., Reference Tesseraud, Abbas, Duchene, Bigot, Vaudin and Dupont2006). Thus, amino acids, particularly the branched-chain amino acid leucine in skeletal muscle, modulate the activity of the intracellular protein kinases involved in the control of mRNA translation, such as 4E-binding protein (4E-BP1) and 70 kDa ribosomal protein S6 kinase (p70S6K, also called S6K1). p70S6K and, similarly, 4E-BP1 are phosphorylated in response to insulin via a signal transduction pathway involving PI3K and the mammalian target of rapamycin (mTOR) (Proud, Reference Proud2006). Amino acid signalling also originates from mTOR, and activates p70S6K and 4E-BP1 (Kimball and Jefferson, Reference Kimball and Jefferson2004 and Reference Kimball and Jefferson2006; Dann and Thomas, Reference Dann and Thomas2006; Tesseraud et al., Reference Tesseraud, Abbas, Duchene, Bigot, Vaudin and Dupont2006). It is noteworthy that the role of amino acids as a nutrient signal appears to be important for cell functions and metabolic pathways other than those directly concerning protein turnover. For example, overstimulation of mTOR/p70S6K by amino acids mediates a feedback loop promoting insulin resistance (i.e. inhibition of insulinstimulated glucose transport; see Um et al., Reference Um, D’Alessio and Thomas2006; Tremblay et al., Reference Tremblay, Jacques and Marette2005). Nevertheless, despite this increased knowledge on the role of amino acids, many questions still remain and the underlying mechanisms need to be elucidated.

Effect of pasture on beef quality

The influence of two production systems (pasture v. maize silage indoors) on muscle properties and gene expression was studied in 30-month-old Charolais cattle. Transcriptomic analyses (Cassar-Malek et al., Reference Cassar-Malek, Bernard, Jurie, Barnola, Gentès, Dozias, Micol and Hocquette2005) revealed that most of the genes differentially expressed in muscles between the two production systems were involved in metabolic and contractile properties. An interesting finding was the downregulated expression of the selenoprotein W (an antioxidant) in steers grazing on pasture, most probably linked to the selenium availability in the diet. It was also shown that the muscles from Charolais beef grazing on pasture group had more oxidative characteristics than those of steers fed maize silage indoors. In an another experiment, the separate effects of the nature of diet and grazing mobility on the metabolic potential of muscles were determined from beef grazing on pasture or fed a cut grass diet or maize silage, with half of each feeding regime group being submitted 7 days/week to a daily 5.2 km walk. It was observed that the more oxidative metabolic orientation of muscles in grazing steers originates from a combination of two separate effects: an increased mobility at pasture and a grass-based diet (Jurie et al., Reference Jurie, Ortigues-Marty, Picard, Micol and Hocquette2006). Together with the muscle changes observed with alterations in growth path (refer to a previous section), these responses to grazing production system indicate a critical role for postnatal nutrition in regulating muscle phenotype and plasticity and hence meat quality. Indeed, the more oxidative metabolism will primarily affect meat colour, mostly by enhancing the red meat colour, which is the first characteristic the consumer takes into account to evaluate meat quality. In addition, the more oxidative muscles may also affect, to a lesser extent, the flavour, juiciness and tenderness of beef meat (for a review, see Hocquette et al., Reference Hocquette, Ortigues-Marty, Pethick, Herpin and Fernandez1998).

Regulation of intramuscular fat and glycogen deposition

Energy restriction was shown to affect muscle characteristics. In rabbits, energy restriction decreases intramuscular fat contents and some lipogenic activities without any change in muscle fibre types (Gondret et al., Reference Gondret, Lebas and Bonneau2000). These results confirm the general view that mitochondria in skeletal muscles can undergo rapid changes as a consequence of modifications in environmental conditions (Hoppeler and Martin, Reference Hoppeler and Martin2003). Generally, muscle fat content variation has been proven to result from a reciprocal balance between catabolic and anabolic fatty acid fluxes, and cannot be assigned to one specific energy pathway (Gondret et al., Reference Gondret, Hocquette and Herpin2004).

It has also been postulated that diets in ruminants which promote both (i) maximal fermentation in the rumen to produce gluconeogenic precursors (propionate) and (ii) which maximise starch digestion in the small intestine, might increase intramuscular fat deposition (Pethick et al., Reference Pethick, Harper and Oddy2004). One underlying mechanism is that such diets would promote increased levels of the anabolic hormones (insulin) known to stimulate lipogenesis. In addition, such diets would also deliver increased levels of net energy for lipogenesis. Finally, there is evidence that marbling adipocytes show a preference for glucose/lactate carbon while subcutaneous adipose tissue in ruminants uses mainly acetate as a source of acetyl units for lipogenesis (Smith and Crouse, Reference Smith and Crouse1984; Hocquette et al., Reference Hocquette2005), which could explain that G6PDH activity is well related to marbling in steers (Bonnet et al., 2007).

Intramuscular level of glycogen is also a very important parameter since it determines the ultimate pH of meat which itself affects many meat quality attributes. There is evidence from studies performed in pigs, poultry and bovines that achieving an optimal pH is required to avoid meat quality defects such as acidic or dark-cutting (or dark firm dry) meat. The general view is that animal management, genotype and nutrition play an interacting role in determining the initial level of intramuscular glycogen at slaughter. For example, stress at slaughter induces a glycogen mobilisation thereby inducing a too high meat ultimate pH and dark cutting in beef. In this species and in pork, the genetic effect acts by modifying the proportion of the different muscle fibre types which contain more or less glycogen and are differently sensitive to stress. In chicken, muscle glycogen reserves at death are also significantly related to bird genotype, but without changes in muscle fibre types (Berri et al., Reference Berri, Debut, Santé-Lhoutellier, Arnould, Boutten, Sellier, Baéza, Jehl, Jego, Duclos and Le Bihan-Duval2005). In bovine, the nutritional effect has been evidenced by a strong seasonal influence on glycogen concentration in bovine muscles in Australia with low levels in winter and summer, and high levels in spring. The low glycogen level is explained, at least in part, by low pasture availability and quality. It was concluded that a growth rate of above 1 kg day is needed to ensure a minimum level of glycogen in muscles and hence to ensure a normal pH of beef thereby reducing the occurrence of dark cutting for grazing cattle in Australia. For more details about the nutritional regulation of glycogen level in muscles, the reader is invited to refer to Pethick et al. (Reference Pethick, Cummins, Gardner, Jacobs, Knee, McDowell, McIntyre, Tudor, Walker and Warner2000).

Nutritional regulation of milk composition

Feed deprivation for 2 to 4 days in cows or goats strongly decreased the yields of milk water, lactose, protein and fat; however, only milk lactose concentration decreased (lactose secretion decreased more than water secretion), whereas protein and, more markedly, fat concentrations increased (protein and fat secretion decreased less than water secretion) (Sauvant et al., Reference Sauvant, Soyeux and Chilliard1983; Ollier et al., Reference Ollier, Robert-Granié, Bernard, Chilliard and Leroux2007; Table 1). This illustrates the key role of lactose, which is the major component regulating milk osmotic pressure, and the fact that the ability to mobilise some protein and more markedly fat reserves make it possible to maintain, at least in the short term, the secretion of other milk components. The same types of responses, although less extreme, were observed when ad libitum-fed cows were forced to walk 9.6 km/day, due to the decreased energy balance caused by the combined effects of appetite decrease and energy expenditure increase (Coulon et al., Reference Coulon, Pradel, Cochard and Poutrel1998; Table 1). During periods of negative energy balance, including early lactation in well fed ruminants, the milk fatty acids are changed, with an increase in the percentages of stearic (18:0) and oleic (c9-18:1) acids inversely proportional to the energy balance value, which is related to the fact that these two fatty acids are largely represented in ruminant adipose tissues (Chilliard et al., Reference Chilliard, Gagliostro, Flechet, Lefaivre and Sebastian1991 and Reference Chilliard, Ferlay, Rouel and Lamberet2003). However, over the longer term, energy deficiency decreases both milk yield and protein content (Coulon and Rémond, Reference Coulon and Rémond1991), reflecting the limitation of the extent to which body protein reserves can be mobilised to sustain milk protein secretion (for a review, see Chilliard et al., Reference Chilliard, Bocquier and Doreau1998).

Table 1 Responses (% of control) to underfeeding or physical activity in lactating ruminants

^†In dairy goats (Ollier et al., Reference Ollier, Robert-Granié, Bernard, Chilliard and Leroux2007).

^‡In dairy cows, 9.6 km walking per day (Coulon et al., Reference Coulon, Pradel, Cochard and Poutrel1998).

^§Abbreviation is: NEFA = non-esterified fatty acids.

Changing diet composition is well known to modify largely milk fat composition, although the composition of caseins and other proteins are not, or very slightly, responsive (Coulon et al., Reference Coulon, Dupont, Pochet, Pradel and Duployer2001). The main factors changing milk fat content and fatty acid composition are the forage : concentrate (or fibre : starch) ratio and the lipid content (or supplementation) of the diet (Bauman and Griinari, Reference Bauman and Griinari2003). Increasing concentrate percentage above 50% to 60% of diet dry matter decreases milk fat content and the concentrations of 14:0, 16:0 and 18:0 and increases trans-18:1 isomers (particularly the trans-10 isomer), c9,t11-CLA and 18:2 n-6 (Loor et al., Reference Loor, Ferlay, Ollier, Doreau and Chilliard2005). Some of the decrease in lipid secretion under high-concentrate diets could result from a propionate or glucose effect, which might stimulate insulin secretion and adipose tissue lipogenesis at the expense of mammary lipogenesis (the glucogenic theory). However, duodenal infusion of 1.2 kg/day glucose or its propionate equivalent only reduced lipid secretion by 5 to 8%. Propionate altered the milk fatty acid profile only slightly whereas glucose reduced short-chain fatty acids (C4 to C8) and, more markedly, C18-fatty acid secretion (Rigout et al., Reference Rigout, Hurtaud, Lemosquet, Bach and Rulquin2003). As the two infusion treatments decreased the plasma acetate and β-OH-butyrate concentrations to the same extent, these changes cannot totally explain the different responses in milk fatty acid profile.

Several recent observations support the ancient ‘trans-fatty acid’ theory to explain the milk fat depressing effects of both concentrate-rich (or starch-rich) diets and PUFA supplementations in the same paradigm (Figure 3). The changes in rumen biohydrogenations of dietary PUFA that occur when rumen pH decreases and/or when dietary PUFA concentration increases indeed result in increases of the concentrations of different trans isomers of 18:1 and 18:2 in milk, related to a shift of the ruminal production of trans-11 18:1 towards trans-10 18:1 (for a review, see Bauman and Griinari, Reference Bauman and Griinari2003). These fatty acids, including t10,c12-CLA and t9,c11-CLA, are putative inhibitors of mammary lipogenic pathways (for a review, see Bernard et al., 2007). Such a shift is marked when the diet is rich in starchy concentrates and/or in maize silage and is not observed when the basal diet is rich in hay (Roy et al., Reference Roy, Ferlay, Shingfield and Chilliard2006; Figure 3) or grass silage (Bell et al., Reference Bell, Griinari and Kennelly2006), which favours the yield of c9,t11-CLA (the major CLA isomer in ruminants, which has putatively positive effects on health). This shift does not occur in the goat receiving a high concentrate diet supplemented with PUFA (Chilliard et al., Reference Chilliard, Rouel, Ferlay, Bernard, Gaborit, Raynal-Ljutovac, Lauret and Leroux2006), probably because ruminal fermentation is more stable in this species and/or because its mammary gland is less sensitive to the antilipogenic effect of t10,c12-CLA (Andrade and Schmidely, Reference Andrade and Schmidely2006). These different types of trans-fatty acids (18:1 and CLA) responses require further studies, since evaluation in rabbits showed that a 12-week intake of trans-10 18:1-rich butter changed the plasma lipoprotein profile (Bauchart et al., Reference Bauchart, Roy, Lorenz, Chardigny, Ferlay, Gruffat, Sébédio, Chilliard and Durand2007) and increased the occurrence of aortic lipid infiltration (Roy et al., Reference Roy, Chardigny, Bauchart, Ferlay, Lorenz, Durand, Gruffat, Faulconnier, Sébédio and Chilliard2007), compared with trans-11 18:1 plus c9,t11-CLA-rich one.

Figure 3 Effect of basal diet and plant oil supplementation on (a) milk fat yield, (b) milk fat t10,c12-CLA, (c) milk fat t9,c11-CLA and (d) milk fat c9,t11-CLA concentrations (g/100 g total fatty acids) for cows given either ‘concentrate–sunflower oil’ (unbroken line), ‘maize silage–sunflower oil’ (broken line) or ‘hay–linseed oil’ (dotted line) diets. Plant oils were included in the diet from day 0. Adapted from Roy et al. (Reference Roy, Ferlay, Shingfield and Chilliard2006).

Among diets not decreasing milk fat secretion, recent research has been aimed at decreasing saturated fatty acids and increasing c9,t11-CLA and/or 18:3 n-3 in order to match better with recommendations for human nutrition and health. This can be achieved mainly by increasing the proportion of good quality pasture and hay in ruminant diets (Ferlay et al., Reference Ferlay, Martin, Pradel, Coulon and Chilliard2006). Another strategy is to add plant oils or oilseeds to high forage diets (for a review, see Chilliard and Ferlay, Reference Chilliard and Ferlay2004), although the putative side-effects on product quality (sensorial and nutritional aspects, including oxidative stability and effects of trans-11 18:1 intake on cardiovascular disease risk in humans) still require thorough investigation.

In conclusion, changes in feeding level and/or diet composition modify absorbed nutrients and, consequently, could change largely milk fat secretion and fatty acid composition, with putative modifications of the nutritional value of dairy products for consumers.