The lipid content and fatty acid (FA) profile of the carcass have an impact on the technological transformation (i.e. a high content of PUFA increases fat softness and the risk of oxidation) and on the nutritional and organoleptic quality (e.g. intra-muscular lipid content, saturated FA content, and n-3 to n-6 ratio). Deposited lipids originate from dietary FA and de novo-synthesised FA. Nutrition is the main factor through which the lipid and FA deposition in pigs may be altered, even if other factors such as genotype, sex, age, slaughter weight and environmental temperature also affect lipid and the FA content (e.g. Wood, Reference Wood and Wiseman1984; Lebret & Mourot, Reference Lebret and Mourot1998; Le Dividich et al. Reference Le Dividich, Noblet, Herpin, van Milgen, Quiniou, Wiseman, Varley and Chadwick1998). Although numerous studies have been carried out studying the relation between nutrition and the FA composition of tissues (e.g. Miller et al. Reference Miller, Shackelford, Hayden and Reagan1990; Madsen et al. Reference Madsen, Jakobsen and Mortensen1992; Wiseman & Agunbiade, Reference Wiseman and Agunbiade1998; Gatlin et al. Reference Gatlin, See, Hansen, Sutton and Odle2002; Ostrowska et al. Reference Ostrowska, Cross, Muralitharan, Bauman and Dunshea2003), these relations are often limited to a single tissue (typically backfat). Consequently, only empirical relationships can be established between nutrition and the FA content of these tissues. In order to define nutritional strategies that modulate the FA profile of tissues, a more mechanistic approach is desirable, describing FA deposition at the whole-animal level.
Mathematical models have been used to predict the consequences of nutritional strategies on pig performance and typically predict whole-body protein and lipid mass (e.g. Whittemore & Fawcett, Reference Whittemore and Fawcett1976; Pomar et al. Reference Pomar, Harris and Minvielle1991; de Lange, Reference de Lange, Moughan, Verstegen and Visse-Reyneveld1995). Based on the generic growth model of De Lange (Reference de Lange, Moughan, Verstegen and Visse-Reyneveld1995), Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) developed a first approach with the objective to predict the consequences of different nutritional strategies on FA deposition. Development of this conceptually simple model was hampered by the limited availability of experimental data at the whole-animal level. Especially data concerning the fate of dietary lipids and the composition of de novo-synthesised FA was scarce. Danfaer (Reference Danfaer and Kyriazakis1999) developed a mechanistic model of carbohydrate and lipid metabolism at the cellular level based on studies of Dunshea et al. Reference Dunshea, Harris, Bauman, Boyd and Bell(1992a, Reference Dunshea, Harris, Bauman, Boyd and Bellb). The latter estimated the kinetics of glycogen, glucose and NEFA in vivo in the plasma of growing pigs (70 kg body weight (BW)). However, the model proposed by Danfaer (Reference Danfaer and Kyriazakis1999) corresponds to the nutrient flow just after ingestion and during a short period of time (a few hours). It is therefore unlikely that it can be used to describe whole-animal FA deposition for prolonged periods of time.
Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005) further developed the conceptual model of Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) and estimated key elements of whole-body FA metabolism related to the efficiency of depositing dietary FA and the composition of de novo-synthesised FA. In the absence of a reasonable alternative, Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) assumed that tissues would not preferentially capture specific dietary FA. According to their model, differences in lipid development between tissues (i.e. allometry) in combination with the supply of FA during growth would be the only cause of differences in FA content between tissues. As indicated by Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002), this simple hypothesis does not seem to hold.
The objective of the present study is to use data from a study of Kloareg et al. (Reference Kloareg, Noblet and van Milgen2006) to address different aspects of FA deposition in finishing pigs. These aspects include the fate of dietary FA, the composition of de novo-synthesised FA and the distribution of FA between different tissues.
Material and methods
Experimental design
Details concerning the experimental design can be found in Kloareg et al. (Reference Kloareg, Noblet and van Milgen2006). In short, eight blocks of four littermates were used in a factorial design including two genotypes (crossbred Piétrain × (Landrace × Large White) and Large White) and two sexes (females and barrows). From 80 kg BW onwards, animals were offered a diet based on wheat, maize, barley, soyabean meal, which contained also 15 g/kg soyabean oil. The diet contained 153 g/kg crude protein and 44 g/kg lipid. The main FA were 18 : 2, 18 : 1, 16 : 0, 18 : 3 and 18 : 0 (0·54, 0·22, 0·14, 0·045, 0·033 of total FA, respectively). The chemical composition of the diet is given in Table 1. A representative feed sample was obtained by regularly taking samples of the distributed feed. The DM content of the distributed ration was measured weekly. Feed refusals (if any) were collected and weighed daily and sampled to measure the DM content. Pigs were slaughtered at approximately 90, 110, 130 and 150 kg after a 16 h fast.
Table 1 Chemical composition and nutritional values of the experimental diet (adjusted for a DM content of 873 g/kg)
* Calculated from Sauvant et al. (Reference Sauvant, Perez and Tran2002).
Compartments
At slaughter, blood was collected, weighed, sampled and pooled by genotype. The empty digestive tract, kidneys, liver, heart and lungs, spleen, diaphragm, leaf fat, head, feet and tail were weighed and combined as a single compartment (viscera, head, feet and tail (VHFT)). Empty BW was calculated as the sum of the weight of the blood, VHFT and hot carcass. The left half carcass was divided in primal cuts according to the Dutch normalised procedure (Institut Technique du Porc, 1990). Backfat (B) was separated from the loin. The loin (without backfat), shoulder, belly and ham were combined as a single compartment (carcass (C)). The VHFT, B and C compartments were weighed, frozen, ground separately, minced and homogenised. Four samples of each compartment were taken. Two of these were used to determine the DM content. The two other samples were freeze-dried and used for further chemical analysis.
Chemical analyses
The lipid content of feed and carcass were determined by solvent extraction. Lipids in the diet were extracted using chloroform. For the VHFT, B, C and blood samples lipids were extracted with a chloroform and methanol mixture (chloroform–methanol, 2 : 1). All extractions were performed using an automatic extraction system Soxtec Avanti 2050 (FOSS, Höganäs, Sweden) in two steps of 30 min each. First, the sample was immersed in the boiling solvent to dissolve most of the soluble material. In the second step, the sample was raised above the solvent surface to permit efficient washing with solvent from the condensers. During the extraction, solvents were heated to 110°C. As the average lipid content in the blood was 1·7 g/kg, its contribution to lipid and FA deposition was ignored for the remainder of the study. The FA of the lipid extraction in the diet and in VHFT, B and C were transmethylated according to Morrisson & Smith (Reference Morrisson and Smith1964) and the FA profile was obtained by GC using a 30 m long and 0·25 mm wide capillary column. In addition to FA, extracted lipids also contain glycerol, phospholipids and other chloroform- or methanol-soluble components. The FA to lipid ratio in a body compartment was considered constant for all animals and was calculated as the average FA to lipid ratio for each compartment. This ratio was obtained by weighing the lipid extraction used for the transmethylation (approximately 20 mg) to which 2·5 mg heptadecanoic acid (17 : 0) was added, a FA that does not exist in monogastric animal tissues. The FA profile was expressed as a percentage of identified FA and converted (in g) using the 17 : 0 recovery.
Fatty acids ileal digestibility values
Ileal digestibility values for the dietary FA were not determined but were estimated from literature data. The FA digestibility may be different for FA provided by soya oil compared to those provided by other feed ingredients. The ileal FA digestibilities provided by the basal diet were supposed to correspond to those measured for total lipids, without added oil, by van Milgen et al. (Reference van Milgen, Noblet and Dubois2001). This value (0·74) was used for all FA provided by the basal diet. For the ileal digestibility of the soyabean oil FA, the mean values observed by Jørgensen et al. (Reference Jørgensen, Jakobsen and Eggum1992) for two diets (basal diet with 5 g/kg crude fat and, respectively, 10 and 20 g/kg added soyabean oil) were used, with different values for each FA. The calculated ileal digestibility for the dietary FA were then 0·77, 0·78, 0·78, 0·80, 0·81, 0·82 and 0·87 for 14 : 0, 16 : 0, 16 : 1, 18 : 0, 18 : 1, 18 : 2 and 18 : 3, respectively. For the other dietary FA, the ileal digestibility was supposed to be 0·95.
Fatty acid deposition
The FA content of the three body compartments was calculated as the lipid content multiplied by their respective FA to lipid ratio and the analysed FA profile. Deposited essential FA (18 : 2 and 18 : 3) originate from the diet only and their deposition rate may be used to estimate the oxidation rate of these FA. However, essential FA may also be used for synthesis of other n-6 or n-3 FA. For example, 18 : 2 is a precursor for other n-6 FA such as dihomolinolenic acid (20 : 3) and arachidonic acid (20 : 4). Similarly, 18 : 3 is a precursor for other n-3 FA such as EPA (20 : 5), docosapentaenoic acid (DPA; 22 : 5) and DHA (22 : 6). Thus, in addition to the deposition rates of individual essential FA, the deposition rates for n-6 FA and n-3 FA were calculated. The digested and deposited 20 : 3 and 20 : 4 were expressed as molar 18 : 2 equivalents required to synthesise these FA and the digested and deposited EPA, DPA and DHA were expressed as molar 18 : 3 equivalents. The balance of individual and n-6 and n-3 FA were calculated for each period using the comparative slaughter technique (i.e. from 90 to 110 kg, from 110 to 130 kg, from 130 to 150 kg). The period had no significant effect on the results and these results are not presented in the present paper. Because calculated FA balances obtained from successive (relatively) short periods of time results are quite variable, it was decided to analyse the FA deposition for the whole period (i.e. 90–150 kg) through regression. Essential FA (18 : 2 and 18 : 3) and n-6 and n-3 FA mass were regressed on the digestible intake of these FA consumed since the beginning of the experiment (SAS, 2000). The slope of this regression represents the deposition rate and effects of sex and genotype were tested on this slope.
De novo synthesis
A regression analysis was also used to estimate the deposition rates of 14 : 0, 16 : 0, 16 : 1, 18 : 0, 18 : 1, EPA, DPA and DHA. It is not possible to determine the oxidation rate for non-essential dietary FA because deposited FA originate from digested dietary FA and from de novo-synthesised FA. Based on results from a previous study (Kloareg et al. Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005), it was assumed that 0·70 of dietary non-essential FA were deposited as is.
Only 14 : 0, 16 : 0, 16 : 1, 18 : 0 and 18 : 1 were taken into account in the calculation of de novo synthesis because the quantity of long-chain FA synthesised de novo is very small. The composition of FA synthesis was used to estimate flow partitioning rates (i.e. elongation and desaturation) of non-essential FA at the whole-animal level (Kloareg et al. Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005). Similarly, deposition rates of the long-chain n-3 FA (EPA, DPA and DHA) were used to estimate the conversion of dietary 18 : 3 to these FA.
Lipid gain composition and fatty acid anatomical partitioning
To study the FA composition of the lipid deposition in each compartment, allometric relations were used to describe the deposition of a FA in a compartment relative to either the lipid mass of this compartment or relative to the total quantity of this FA deposited in the body:
where Y is the FA mass in a body compartment (g) and X is either the lipid mass of that compartment or the total body mass of that FA. The FA composition of deposited lipid (in the whole body or in a compartment) is given by the first derivative of the allometric function relative to lipid mass (i.e. dFA/d(lipid) = a × b × (lipid)(b − 1)). Similarly, partitioning of FA between anatomical compartments is given by the first derivative of the allometric function relative to total body mass of the FA. The effect of sex or genotype was not estimated on these relations.
Results
Animals were slaughtered between 84 and 154 kg BW. Results concerning performance, weights of compartments and organs, and protein and lipid deposition have been reported by Kloareg et al. (Reference Kloareg, Noblet and van Milgen2006). On average, C, B and VFHT represented, respectively, 0·694, 0·058 and 0·189 of empty BW and 0·634, 0·183 and 0·183 of whole-body lipid mass. The main results concerning performance and lipid and FA composition are given in Table 2. The FA to lipid ratio was 0·71, 0·80 and 0·58 for C, B and VHFT, respectively. For the whole body, this ratio was 0·70. For some FA, the FA composition was affected by the BW at slaughter (i.e. 90, 110, 130 or 150 kg) and/or by genotype (Table 2).
Table 2 Performance and body composition of finishing pigs slaughtered between 90 and 150 kg† (Mean values)
EBW, empty body weight.
† For details of procedures, see pp. 36–37.
‡ From ANOVA; rsd, residual standard deviation; BW, effect of body weight; G, effect of genotype; S, effect of sex. Levels of significance: *P < 0·05; **P < 0·01, ***P < 0·001.
§ Between 80 kg and the target slaughter weight. Feed intake was adjusted for 873 g DM/kg.
Metabolism of essential fatty acids
The supply and deposition of n-6 FA was almost exclusively as 18 : 2. Consequently, the regression of FA deposition v. digestible FA supply gave identical results for 18 : 2 and n-6 FA. The slope of this relation was 0·31 (se 0·03), which means that only 0·31 of the digestible n-6 FA supply was deposited in the body. As illustrated in Fig. 1, both sex and genotype affected the deposition rate (P < 0·01; Fig. 1), resulting in slopes varying between 0·23 and 0·38. Digestible 18 : 3 represented 0·95 of the supply of digestible n-3 FA in the diet. On the other hand, deposited 18 : 3 represented only 0·63 of total deposited n-3 FA; the remainder was deposited as 20 : 5, 22 : 5 and 22 : 6. The slopes of the relations between the deposited and the digested FA was 0·24 (se 0·06) for 18 : 3 and 0·40 (se 0·05) for n-3 FA (Fig. 1). This means that 0·40 of the digestible n-3 FA supply was recovered in the body and that 0·60 remained unaccounted for. The slopes of the relation between deposited and digestible EPA, DPA and DHA supply were 4·2 (se 0·6), 1·5 (se 0·9) and 1·1 (se 0·5), respectively. The corresponding profile of n-3 FA that were synthesised was 0·909, 0·060 and 0·031 for EPA, DPA and DHA, respectively. These results allow the estimation of the partitioning of 18 : 3 that could be accounted for. The metabolism of 18 : 3 first occurs through a successive desaturation, elongation and again a desaturation to EPA. As indicated earlier, 0·24 of the supply of digestible 18 : 3 was deposited as is, and 0·13 was further metabolised, first to EPA. Most of this supply (0·114) was deposited as EPA, whereas the remainder (0·0114) was elongated to DPA. Approximately two-thirds of this (0·0075) was deposited as DPA and one-third (0·0039) was metabolised further and deposited as DHA.
Fig. 1 Relation between deposited n-6 and n-3 fatty acid (g) and digestible fatty acid supply since 80 kg of BW (g). Sex and genotype affected the slope of the relation for n-6 fatty acids (P = 0·005; (a): 0·38 for Large White females (●), 0·31 for Large White barrows (▲), 0·27 for crossbred barrows (△) and 0·23 for crossbred females (c). Sex and genotype had no effect on the slope for n-3 fatty acid deposition ((b); 0·40).
Genotype affected the deposition rate of 18 : 3 (P = 0·01). The slope of the relation between deposited and digestible 18 : 3 supply was 0·19 for the Large White and 0·33 for the crossbred. As the deposition rate of n-3 FA was not affected by genotype (P = 0·40), this means that Large White pigs convert a greater proportion of 18 : 3 to EPA, DPA and/or DHA than do the crossbred (0·15 v. 0·08 of digestible 18 : 3). Moreover, the deposition rate of EPA is greater for Large White than for crossbred (P = 0·004), whereas the deposition rates of DPA and DHA are not affected by genotype. Consequently, the n-3 synthesis profile is affected by genotype (0·922, 0·051 and 0·027 for Large White v. 0·859, 0·093 and 0·048 for crossbred for EPA, DPA and DHA, respectively).
Metabolism of non-essential fatty acids
The slopes of the relation between the deposited and digestible non-essential FA were 19·9 (se 1·6), 4·2 (se 0·3), 18·0 (se 2·6), 11·8 (se 1·1) and 4·2 (se 0·3) for 14 : 0, 16 : 0, 16 : 1, 18 : 0 and 18 : 1, respectively. These values indicate that, in this experiment, the de novo synthesis was several times greater than the dietary supply of these FA. In fact, assuming that 0·70 of digestible, dietary non-essential FA were deposited as is, the de novo synthesis represented on average 0·86 of the non-essential FA deposition. The corresponding profile of de novo-synthesised FA was 0·017, 0·286, 0·025, 0·217 and 0·454 for 14 : 0, 16 : 0, 16 : 1, 18 : 0 and 18 : 1, respectively. Sex and genotype affected the ratio between deposited and digested FA ratio for 14 : 0, 16 : 0 and 18 : 1 (P < 0·05), suggesting that the de novo synthesis is affected by both sex and genotype. The composition of de novo-synthesised FA is given in Table 3 for the four groups. Boars tend to deposit slightly more 16 : 1 than gilts, whereas Large White deposit more 18 : 0 and less 18 : 1 than the crossbred. Nevertheless, differences in the profiles of de novo-synthesised FA remain small.
Table 3 Calculated composition of de novo-synthesised fatty acids in finishing pigs*
* It was assumed that 0·30 of the digestible dietary fatty acids supply was oxidised. For details of the calculation method, see p. 37.
The present results allow estimation of the partitioning of de novo-synthesised FA. Of the total flow of 16 : 0 that is used in the de novo synthesis, 0·286 will be deposited as is, 0·017 will be shortened to 14 : 0, 0·025 will be desaturated to 16 : 1 and 0·671 will be elongated to 18 : 0. Similarly, 0·324 of the de novo-synthesised 18 : 0 will be deposited as is and 0·676 will be desaturated to 18 : 1.
Composition of lipid deposition and anatomical partitioning of fatty acids
In Table 4, the average FA composition of the total FA gain is given for the whole body and for the different tissues. The results indicate that, relative to the whole-body FA gain, the FA gain in C has a slightly higher 18 : 1 and a lower 18 : 2 content. The FA gain in B contains less 16 : 0 and 18 : 1, but considerably more 18 : 2 and 18 : 3 compared to the whole-body FA gain. Finally, VFHT contains less 18 : 1 and 18 : 2, but more of the saturated FA 16 : 0 and 18 : 0.
Table 4 Allometric development of the fatty acid (FA) mass relative to the lipid mass in different body compartments in finishing pigs*
B, backfat; C, carcass without backfat; VHFT, viscera, head, feet and tail.
* The allometric relation Y = aX b was used between 80 and 150 kg live weight, where Y is the fatty acid mass (g) and X the lipid mass (g) of the compartment.
† The first derivative of the allometric relation was used to calculate the average composition: the dFA/dX was calculated for the average lipid mass and multiplied by the FA to lipid ratio for each compartment.
Parameter estimates of the allometric relationships between the different FA masses and whole-body lipid mass are given in Table 4. The shape parameter b of the allometric relations indicates the change in FA mass relative to the lipid mass. A constant composition would result in a shape parameter of 1. The shape parameter was close to 1 for most non-essential FA, with the exception of 16 : 1 (b = 0·716 (se 0·070)) suggesting that the 16 : 1 content decreases with lipid mass. In contrast, in the present experiment the proportion of essential FA in lipid mass increased with lipid mass (b = 1·188 (se 0·072) for 18 : 2; b = 1·243 (se 0·207) for 18 : 3).
Table 4 also lists the parameter estimates of the allometric relation between FA and lipid mass in the three body compartments. It appears that for 16 : 0 and 18 : 0, B and VFHT follow a similar development pattern. For 18 : 1, deposition occurs relatively early for VHFT and relatively late for C. The most striking difference occurs for the essential FA. The shape parameter b for C is much greater than those observed for B and VFHT, indicating that deposition of essential FA in C increases during the later stages of growth.
In the preceding analysis, FA mass of a compartment was related by an allometric relation to the total lipid mass of that compartment. An alternative approach is to relate the FA mass of a compartment to the total mass of that FA in the body. The results are given in Table 5. On average, 0·60 of the total FA gain was deposited in C, 0·25 in B and 0·15 in VHFT. This partitioning is variable for the different FA, especially for essential FA. Of the total 18 : 2 deposition, 0·48 was deposited in C, 0·44 in B and only 0·08 in VFHT. In contrast, for 18 : 3, 0·71 was deposited in C, 0·21 in B and 0·08 in VFHT. The proportion of total FA deposited in B increases during growth, as well as the proportions of non-essential FA at the expense of the deposition of essential FA. The reverse is seen for C.
Table 5 Anatomical partitioning of fatty acids (FA) and lipids between the three compartments*
B, backfat; C, carcass without backfat; VHFT, viscera, head, feet and tail.
* The allometric relation Y = aX b was used were Y is the lipid or fatty acid mass deposited in each compartment (g) and X is the whole-body lipid or fatty acid mass (g).
† The first derivative of the allometric relation was used for the average value of X.
Discussion
Fatty acid oxidation
In the present study, 0·31 of n-6 and 0·40 of the digestible n-3 FA supply was recovered in the body. The remainder (0·69 for n-6 FA and 0·60 for n-3 FA) could not be accounted for and was supposed to be oxidised or converted to non-FA metabolites. The serial slaughter technique used in the present study does not distinguish between physiological processes that may be involved in the metabolism of n-3 FA (i.e. postprandial oxidation, oxidation due to the turnover of the lipid mass or further synthesis of metabolites such as hormones or prostaglandins). Moreover, the results are directly affected by the (assumed) ileal digestibility of essential FA, which were estimated from literature data (0·82 and 0·87 for 18 : 2 and 18 : 3, respectively). For example, if the assumed digestibility of 18 : 2 decreases from 0·82 to 0·72, the recovery rate increases from 0·31 to 0·36. Nevertheless, both estimates of oxidation are considerably higher than those found in the literature, particularly for n-6 FA. Using also the slaughter technique, Flanzy et al. (Reference Flanzy, François and Rérat1970) estimated an average net oxidation rate of 18 : 2 of 0·50 in pigs between 45 and 100 kg BW. Average daily gain of the animals affected this value, with estimates varying between 0·35 (for a high daily gain) to 0·66 (for a low daily gain; Flanzy et al. Reference Flanzy, François and Rérat1970). This may be due to the higher turnover rate of n-6 FA turnover and the relative importance of hormone synthesis at low growth rates. The results of the present study may be compared to those having a high growth rate in the study of Flanzy et al. (Reference Flanzy, François and Rérat1970), for which the 18 : 2 net oxidation varied between 0·35 and 0·43. Using the comparative slaughter technique, Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005) estimated a net oxidation of 0·31 for n-6 and 0·52 for 18 : 3 FA. These values were not affected by feeding level. Chwalibog et al. (Reference Chwalibog, Jakobsen, Henckel and Thorbek1992) concluded that, based on data obtained in respiration chambers, all digestible dietary lipids were stored (i.e. they concluded that there was no oxidation of dietary FA). Crespo & Esteve-Garcia Reference Crespo and Esteve-Garcia(2002b) observed oxidation rates between 0·07 and 0·30 for n-6 and n-3 FA in chicken. Using 14C-labelled medium-chain FA and essential PUFA, Leyton et al. (Reference Leyton, Drury and Crawford1987) estimated whole-body oxidation rates during a 24 h period in weanling rats. The oxidation rates (measured as expired 14CO2) were 0·48 for 18 : 2 and 0·64 for 18 : 3. Little information is available in the literature about factors affecting n-6 or n-3 oxidation that could explain these differences. Stage of development undoubtedly will affect the results. Whereas a growing animal will deposit some of the dietary energy supply, mature animals will be (almost) in energetic equilibrium and thus catabolise all dietary energy. Pigs used in the present study were still depositing considerable quantities of energy and lipid and it is unlikely that stage of development may explain the high oxidation rate observed in the present study. Also diet composition, and especially the lipid content and FA composition, may affect the oxidation of FA. It has been shown in chickens that the dietary 18 : 2 and 18 : 3 contents increase the oxidation of these FA (Crespo & Esteve-Garcia, Reference Crespo and Esteve-Garcia2002a, b; Newman et al. Reference Newman, Bryden, Fleck, Ashes, Buttemer, Storlien and Downing2002). The 18 : 2 and 18 : 3 contents in the diet used in the present study were higher than those used by Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005). Although this may have contributed to the observed difference in oxidation rates, the magnitude of the difference is considerable.
De novo synthesis
The deposition of dietary FA is small compared to FA synthesis and represented in the present study 0·86 of the FA deposition. Consequently, errors in the assumed oxidation rate of non-essential dietary FA (0·30) have little effect on the calculation on the de novo-synthesised FA.
The profile of de novo-synthesised FA calculated in the present study were similar to those obtained by Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005) for pigs fed ad libitum at thermoneutrality between 24 and 65 kg BW. The most important differences were found for 16 : 0 (0·286 v. 0·311) and 18 : 0 (0·217 v. 0·177). Consequently, estimation of the rates with which 16 : 0 is metabolised further or deposited were similar in both studies. The composition of de novo-synthesised FA is also similar to that found by Hilditch & Williams (Reference Hilditch and Williams1964); Leat et al. (Reference Leat, Cuthbertson, Howard and Gresham1964) and Flanzy et al. (Reference Flanzy, François and Rérat1970), who used lipid-free diets. Flanzy et al. (Reference Flanzy, François and Rérat1970) estimated that the FA synthesis composition was 0·28, 0·14 and 0·58 for 16 : 0, 18 : 0 and 18 : 1, respectively, for the slow growth rate group and 0·28, 0·18 and 0·53 for the high growth rate group.
The present results suggest that the composition of de novo-synthesised FA is relatively constant in most experimental conditions. Moreover, the shape parameter b of the allometric development of FA relative to the whole-body lipid mass was close to 1 for most non-essential FA (Table 4). The only exception was 16 : 1, which had an allometric shape parameter considerably lower than 1. Nevertheless, its contribution to the profile of de novo-synthesised FA is low (0·027 of synthesised FA). Therefore, this does not necessarily invalidate the assumption of Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002), who assumed that the composition of de novo-synthesised FA was constant throughout growth.
Although the composition of de novo-synthesised FA is relatively constant at the whole-animal level, this is less so for the body compartments. For example, the allometric shape parameter for 18 : 1 is higher for C than for the whole body, whereas that for 18 : 0 and 16 : 0 is lower. Obviously, the inverse is seen for B and VFHT relative to the whole body.
There are also external factors known to affect the composition of de novo-synthesised FA. Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005) showed that ambient temperature and feeding level affected the de novo FA composition: a reduction in feed intake increased the 16 : 0 elongation rate, whereas the increase in temperature reduced the 18 : 0 desaturation rate. Others have shown that the dietary 18 : 2 and 18 : 3 content affects the composition of de novo-synthesised FA. In backfat, the activity of the stearyl-CoA desaturase, involved in the desaturation of both 16 : 0 and 18 : 0, decreases when the dietary 18 : 2 and 18 : 3 content increases (Kouba & Mourot, Reference Kouba and Mourot1998; Kouba et al. Reference Kouba, Enser, Whittington, Nute and Wood2003). This could explain the slightly lower 16 : 1 and 18 : 1, and higher 16 : 0 and 18 : 0 content in the de novo-synthesised FA in the present study compared to Kloareg et al. (Reference Kloareg, Le Bellego, Mourot, Noblet and van Milgen2005).
n-3 fatty acid synthesis
Linolenic acid (18 : 3) can be converted to other n-3 FA (mainly to EPA and DPA and, to a lesser extent, DHA) and to hormones and prostaglandins. No quantitative information concerning the efficiency of 18 : 3 conversion was found in the literature for pigs. In adult man, the apparent conversion of 18 : 3 to EPA is limited (less than 0·08) and is even less for DHA (less than 0·04) (Burdge & Wootton, Reference Burdge and Wootton2002; Burdge et al. Reference Burdge, Jones and Wootton2002). Results of the current study indicated that only 0·40 of the dietary supply of 18 : 3 could be recovered in the body. Of the 18 : 3 that was deposited, more than one-third was deposited as EPA, DPA and DHA, suggesting that the conversion of 18 : 3 to these metabolites is more efficient in pigs than in man. Genotype affected this proportion and a greater conversion rate of 18 : 3 to EPA was observed for Large White pigs relative to the crossbreds (0·13 of digestible 18 : 3 v. 0·06, respectively). No information about the effect of genotype on n-3 FA metabolism was found in the literature. Nevertheless, the fact that in the present study a considerable fraction of essential FA could not be recovered in the body (compared to results of other studies) makes it difficult to draw general conclusions, as the efficiency of conversion of 18 : 3 to EPA, DPA and DHA may be affected by the rate of recovery of 18 : 3.
Anatomical partitioning of fatty acid gain
The FA composition varies between adipose tissues (e.g. Leat, Reference Leat and Riis1983). In their model, Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) supposed that the FA deposition in different tissues was the result of tissue development combined with differences in FA supply (i.e. tissues would capture FA relative to the development rate of the tissues). For example, perinephric tissue develops relatively late and it was hypothesised that its constituent FA are mainly those ingested or synthesised during the finishing phase. As indicated by the authors, differences in FA composition between tissues could only partly be explained by differences in tissue development (Lizardo et al. Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002). Results of the current study confirm that differences in development between tissues are insufficient to explain differences in FA composition between tissues. As indicated in Table 4, the shape parameter of the allometric relation between FA and total lipid for a compartment differs from unity for several FA. This means that the profile of FA deposition changes during growth. The same conclusion can be drawn from the results presented in Table 5. This means that some FA are preferentially deposited in some tissues. Consequently, the FA gain in backfat is not necessarily representative for the FA gain in the other adipose tissues, or for whole-body FA deposition.
No explanation for the difference in the spatial distribution of FA was found in the literature. Although differences between tissues are observed, it is not clear whether this is due to differences in the synthesis or capture of FA. It is known that different tissues possess different capacities of FA synthesis. For example, elongation is faster in bovine subcutaneous adipose tissue compared to liver (0·42 v. 0·15 nmol/min per mg protein) whereas desaturation was observed only in adipose tissue (0·21 nmol/min per mg protein; St John et al. Reference St John, Lunt and Smith1991). Consequently, the contribution of different tissues in FA synthesis is not necessarily indicative for differences in FA deposition between tissues.
Although the present results do not explain the differences observed in FA spatial distribution, they can be useful for modelling FA deposition in the body. In the future, there will be an increased need to control FA composition in different tissues (e.g. to increase the n-3 and n-6 FA content in lean meat, while ensuring firmness of backfat). Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) proposed a model of FA deposition in different tissues. Their approach was generic in that it could be linked to any growth model that predicted lipid deposition. Although few mechanistic models of lipid metabolism exist (e.g. Danfaer, Reference Danfaer and Kyriazakis1999; Halas et al. Reference Halas, Dijkstra, Babinszky, Verstegen and Gerrits2004), practical application of such a model probably calls for a simpler and more empirical approach. The results of Tables 4 and 5 cannot be used directly in modelling the FA deposition in growing pigs as the (absolute values of the) partitioning of FA is specific for the current experimental conditions. For example, the parameters of the allometric relation for 18 : 2 and 18 : 3 in Table 4 will depend on the essential FA content of the diet. For the essential FA, a ‘push’ approach seems most appropriate where the supply of essential FA is partitioned between the tissues (after accounting for digestion and oxidation). In Table 5, the first derivative of the allometric function directly gives the partitioning of essential FA supply between tissues.
For non-essential FA, the approach is somewhat different as there is no need to partition a dietary supply of FA. Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002) used a ‘pull’ approach for FA deposition, where whole-body lipid deposition was determined by an external model (which determined the partitioning of energy between protein deposition and lipid deposition). In the approach of Lizardo et al. (Reference Lizardo, van Milgen, Mourot, Noblet and Bonneau2002), the composition of de novo-synthesised FA did not vary with development or between tissues. Fig. 2 shows two hypotheses concerning the driving forces for non-essential FA deposition in tissues. In Fig. 2(a), tissue FA deposition of a compartment is driven by the total lipid or total FA deposited in this tissue. The corresponding allometric relations given in Table 4 describe this partitioning. In Fig. 2(b), the distribution of each FA in the anatomical compartments is driven by the total FA deposited in the body and the allometric relations given in Table 5 describe this partitioning. Although both representations can be considered as empirical approaches to FA modelling, they nevertheless reflect fundamental differences in the perception of metabolism. In Fig. 2(a), it is the tissue itself that controls autonomously its composition. The model represented in Fig. 2(b) reflects a view in which there is a centralised perception of FA that have been deposited.
Fig. 2 Possible hypotheses concerning the driving forces for the anatomical partitioning of de novo synthesised fatty acids (FA). (a) The partitioning is driven locally by the fatty acid mass of the tissue. Whereas in (b), there is a central control for the partitioning of fatty acids between tissues.
Conclusions
The main objective of the present study was to estimate parameters of a modelling approach that relates nutrition and animal development to FA composition at the whole-animal level. The major elements include the deposition of dietary FA (relative to oxidation) and the composition of de novo-synthesised FA. In addition, partitioning of FA between the carcass, backfat and non-carcass components was addressed. Only 0·31 of the digested n-6 FA was retained by the animal, a value much lower than that obtained in previous studies. At this point in time, it appears difficult to quantify the deposition of dietary FA as a function of the nutritional strategy or stage of development of the animal. Due to the importance of de novo FA synthesis in pigs, this issue mainly concerns essential FA. Nutrition and stage of development seem to have little effect on the composition of whole-body de novo-synthesised FA although the deposition of non-essential FA differs between tissues. Different empirical approaches were proposed to partition the FA between different tissues. Their generality for use under other circumstances remains to be proven.
Acknowledgements
The authors acknowledge the assistance of H. Demay, G. Conseil, B. Duteil, R. Delaunay, M. Alix, J. Liger, J.-F. Rouaud, R. Vilboux and G. Guillemois during the conduct of the experiment and of Y. Lebreton, B. Janson, A. Pasquier, Y. Jaguelin-Peyraud and A. Mounier during the chemical analyses. We are also grateful to J. Mourot and F. Gondret for their help and advice. The authors acknowledge the financial support of INZO° (Innovation en Nutrition et Zootechnie, 02 402 Château-Thierry, France).
