The fatty acid (FA) composition of bovine milk fat is of growing interest, since it is typically rich in saturated FA (particularly C12:0 to C16:0) which, compared with unsaturated FA, are known to elevate plasma total and low density lipoprotein-cholesterol (LDL-C) concentrations (Nicolosi et al., Reference Nicolosi, Roger, Kritchevsky, Scimeca and Huth1997), and thus, the risk of cardiovascular disease (CVD) (Libby et al., Reference Libby, Aikawa and Schönbeck2000). However, milk fat also contains specific FA, as conjugated linoleic acids (CLA) of which cis-9, trans-11 is the major isomer (ca. 0.5% of total FA and more than 90% of total CLA). It has been shown to exhibit in animal studies beneficial health effects on cancer prevention and treatment, and cardiovascular risk (reviewed by Wahle et al. (Reference Wahle, Heys and Rotondo2004)). Chemically synthesised CLA (cis-9, trans-11 CLA or mixture of cis-9, trans-11+ trans-10, cis-12 CLA) have been shown to protect against the development and progression of atherosclerosis in rabbits (Kritchevsky et al., Reference Kritchevsky, Tepper, Wright and Czarnecki2002 and Reference Kritchevsky, Tepper, Wright, Czarnecki, Wilson and Nicolosi2004), hamsters (Nicolosi et al., Reference Nicolosi, Roger, Kritchevsky, Scimeca and Huth1997; Wilson et al., Reference Wilson, Nicolosi, Chrysam and Kritchevsky2000) and ApoE-KO mice (Toomey et al., Reference Toomey, Roche, Fitzgerald and Belton2003) when a pro-atherogenic diet was fed. Moreover, the addition of a cis-9, trans-11 CLA-rich mixture to butter depressed some risk factors of CVD in cholesterol-fed hamsters (Valeille et al., Reference Valeille, Férézou, Amsler, Quignard-Boulangé, Parquet, Gripois, Dorovska-Taran and Martin2005).
Consequently, the enrichment of milk fat in cis-9, trans-11 CLA could improve the nutritional value of milk fat. It is well established that the nutrition of cows is an efficient mean to modulate the FA composition of milk fat (Chilliard and Ferlay, Reference Chilliard and Ferlay2004). Dietary strategies based on concentrate diets supplemented with oil rich in C18:2 n-6 allow the marked increase in milk fat concentrations of both cis-9, trans-11 CLA and trans-11 C18:1 (Bauman et al., Reference Bauman, Barbano, Dwyer and Griinari2000), the latter being converted into the former in animal and human tissues (Santora et al., Reference Santora, Palmquist and Roehrig2000; Turpeinen et al., Reference Turpeinen, Mutanen, Aro, Salminen, Basu, Palmquist and Griinari2002; Corl et al., Reference Corl, Barbano, Bauman and Ip2003). Animal studies using a cis-9, trans-11 CLA+trans-11 C18:1-enriched butter have been carried out, initially showing decreased chemically induced mammary tumours in rats (Ip et al., Reference Ip, Banni, Angioni, Carta, McGinley, Thompson, Barbano and Bauman1999), and improved plasma lipoprotein profile in cholesterol-fed hamsters, when compared with a standard butter (Lock et al., Reference Lock, Horne, Bauman and Salter2005). However, the cis-9, trans-11 CLA+trans-11 C18:1-enriched butters used in the abovementioned studies contained non-negligible amounts of trans-10 C18:1 (ca. 2.8% of total FA), a trans FA that is generally low in non-enriched milk fat ( < 0.5% of total FA) and high in partially hydrogenated fats (Precht and Molkentin, Reference Precht and Molkentin1996). However, trans-10 C18:1 is also increased under certain cow's feeding strategies (Chilliard and Ferlay, Reference Chilliard and Ferlay2004). Very little is known about the effects of trans-10 C18:1 on cardiovascular parameters. Only one study reported a positive association between the concentration of trans-10 C18:1 in platelets and the degree of coronary artery disease in humans, while no association was reported for trans-11 C18:1 (Hodgson et al., Reference Hodgson, Wahlqvist, Boxall and Balazs1996).
In order to explore the effects of the ingestion of trans-10 C18:1 on several lipid parameters involved in atherosclerosis, butters either rich in trans-10 C18:1 or trans-11 C18:1+cis-9, trans-11 CLA were included into an atherogenic diet offered to rabbits used as model of cholesterol-induced atherosclerosis. The experimental butters were compared with a standard butter (low in trans FA and rich in saturated FA), to evaluate their effects on aortic fatty streak observations, plasma FA and lipoprotein profiles, plasma eicosanoid concentrations, and the relationships between the aortic fatty streaks and plasma lipid parameters of rabbits.
Material and methods
Animals, diets and experimental design
Thirty-six male New Zealand White rabbits (age: 9.6 ± 0.5 weeks, body weight: 2.1 ± 0.1 kg, Les Dombes, Chatillon sur Chalaronne, France) were randomly divided into three groups of 12 animals on the age and body-weight basis. The animals were housed in individual stainless steel cages in a temperature-controlled room (18°C) maintained on a 12-h light-dark cycle. The animals were allowed to free access to water and fed daily one of three diets containing 12% on a dry matter (DM) basis of butter, and supplemented with 0.2% DM of cholesterol. The three diets were prepared and pelleted by INRA (Institut National de la Recherche Agronomique, Station de Recherches Cunicoles, Toulouse, France), and contained (g/kg DM) alfalfa flour (305), soya cake (233), butter (120), sucrose (102), beet pulp (100), barley (65), sunflower cake (50), maize oil (8), sodium chloride (5), calcium hydrogen phosphate (5), vitamin and mineral mix (5) and cholesterol (2). The DM concentration of these diets was 845 g/kg feed and the composition was (g/kg DM): organic matter (921), crude cellulose (171), protein (nitrogen × 6.25) (231), lipid (134), FA (114). The three diets essentially differed in the FA composition of butters, i.e. a butter rich in trans-11 C18:1 and cis-9, trans-11 CLA (7.4% and 3.1% of total FA, respectively, T11-CLA diet), a butter rich in trans-10 C18:1 (13.7% of total FA, T10 diet), or a butter with a standard FA composition (1.2%, 0.4% and 0.3% of total FA as trans-11 C18:1, cis-9, trans-11 CLA and trans-10 C18:1, respectively, S diet), made from milk of cows fed different basal diets supplemented with plant oils (Roy et al., Reference Roy, Ferlay and Chilliard2006). The FA composition of the three diets is presented in Table 1. The rabbits (T11-CLA, T10 and S groups) were fed the diets either for a 6-week- (experiment 1, n = 6 per diet) or a 12-week- (experiment 2, n = 6 per diet) experimental period. Food intake was measured daily, uneaten food was discarded once daily and body weight was recorded every week. All rabbits were killed at the end of the experiment after food deprivation for 17 h.
† Sum of cis-9, cis-11, and cis-12 isomers.
‡ Sum of trans-6/7/8, trans-9, trans-10, trans-11, trans-12 and trans-15 isomers.
Blood sampling, plasma lipid and lipoprotein analyses
At the end of the both experiment 1 and experiment 2, the rabbits were killed by cervical dislocation, and truncal blood samples (50 ml) were collected into tubes containing Na2-EDTA, Na-N3 and merthiolate (final concentrations 3 mmol/l, 0.1 and 0.01 g/l, respectively). Plasma was prepared by centrifugation of blood samples at 2700 g for 15 min at 15°C. Three 500 μl aliquots of plasma were stored at − 20°C until lipid and apolipoprotein analysis and one 20-ml fraction was treated immediately for lipoprotein fractionation. The isolation of lipoproteins for determination of the density distribution and chemical composition of lipoprotein subfractions was previously described (Bauchart et al., Reference Bauchart, Durand, Laplaud, Forgez, Goulinet and Chapman1989). Briefly, plasma lipoproteins were separated by ultracentrifugation according to a three step-sequence, and chylomicrons were defined as lipoproteins of density < 0.950 g/ml, very low density lipoproteins (VLDL) as lipoproteins of density = 0.950 to 1.018 g/ml, intermediary and low density lipoproteins (IDL and LDL) as lipoproteins of density = 1.018 to 1.060 g/ml, and high density lipoproteins (HDL) as lipoproteins of density = 1.060 to 1.180 g/ml. The non-HDL fraction corresponded to the sum of chylomicrons, VLDL and IDL+LDL subfractions. The enzymatic methods used for determination of lipid classes (free cholesterol, FC; total cholesterol, TC; triacylglycerols, TG; phospholipids, PL) in plasma and in lipoprotein subfractions have been described previously (Leplaix-Charlat et al., Reference Leplaix-Charlat, Bauchart, Durand, Laplaud and Chapman1996). Cholesteryl esters (CE) content was calculated using the relationship: CE = (TC − FC) × 1.68. Plasma non-esterified FA (NEFA) concentration was determined spectrophotometrically by the acyl-CoA synthetase method (Wako-Unipath NEFA-C kit, Oxoid, Dardilly, France).
Blood samples were also collected in tubes containing either Na2-EDTA+indomethacin (3 mmol/l and 40 mmol/l, respectively) for estimation of prostacyclin (prostaglandin I2, PGI2) concentration, or indomethacin alone (40 mmol/l) for estimation of thromboxane A2 (TxA2) concentration, or, when possible (i.e. when blood quantities were sufficient), Na2-EDTA alone (3 mmol/l) for determination of total FA composition in plasma.
Eicosanoid and fatty acid analyses
The blood samples for PGI2 assay were immediately centrifuged at 1000 g for 10 min at 4°C, plasma was harvested and stored at − 80°C until analysis. The concentration of PGI2 was measured using an ELISA kit (intra- and inter-essay coefficients of variation (CV) = 2.9% and 6.0%, respectively; DE 0800, R&D Systems, Minneapolis, MN, USA) according to the manufacturer instructions. Briefly, concentration of PGI2 was monitored by measurement of 6-keto-prostaglandin F1α (6-keto-PGF1α) concentration, a stable metabolite produced by the non-enzymatic hydration of PGI2.
The blood samples for TxA2 assay were left for 2 h at room temperature, and serum was harvested after centrifugation (1000 g for 10 min at room temperature) according to the manufacturer instructions. The concentration of TxA2 was measured using an ELISA kit (intra- and inter-assay CV = 1.6% and 6.2%, respectively; DE 0700, R&D Systems, Minneapolis, MN, USA). The concentration of TxA2 was monitored by measurement of TXB2 concentration, a stable metabolite produced by the non-enzymatic hydration of TxA2.
For determination of plasma FA composition, the blood samples were immediately centrifuged (2700 g for 15 min at 15°C), and plasma (ca. 1 ml) was stored in 2 ml of chloroform/methanol (2:1, v/v) at − 80°C until FA analysis. Plasma FA were extracted and trans-methylated according to the procedure of Christie et al. (Reference Christie, Sébédio and Juanéda2001), and the FA methyl esters (FAME) were stored at − 20°C until analysis. The FAME were separated using a gas chromatograph (Hewlett Packard Model 4890, Palo Alto, CA, USA) equipped with a flame-ionisation detector, automatic injector, split injection port and a 100 m fused capillary column (100 m × 0.2 mm i.d., 0.2 μm film, CP-Sil 88, Varian, Courtaboeuf, France) using hydrogen as the carrier gas (35 ml/min). Injector and detector temperatures were maintained at 250°C. Total FAME profile was determined using a temperature gradient program. Following sample injection, column temperature was maintained at 60°C for 1 min, increased at a rate of 3°C/min to 85°C, raised to 190°C at a rate of 20°C/min, held at 190°C for 70 min, increased at 20°C/min to a final temperature of 210°C that was maintained for 5 min. More detailed analysis of the distribution of individual isomers of C18:1 was determined using a 120 m fused capillary column (120 m × 0.25 mm i.d., 0.25 μm film, BPX70, SGE, Melbourne, Australia) with hydrogen as the carrier gas (35 ml/min). Injector and detector temperatures were maintained at 250°C. C18:1 FAME profile was determined using a temperature gradient program. Following sample injection, column temperature was maintained at 60°C for 1 min, increased at a rate of 20°C/min to 160°C, held at 160°C for 40 min, increased at a rate of 20°C/min to a final temperature of 220°C that was maintained for 20 min.
Total lipids in diets were extracted according to the method of Folch et al. (Reference Folch, Lees and Sloane Stanley1957) and saponified overnight in an ethanolic potassium hydroxide solution (100 g/l). The FA were methylated at room temperature by the BF3/Na methanolate method (Sebedio et al., Reference Sébédio, Juanéda, Grégoire, Chardigny, Martin and Giniès1999), and subsequently analyzed by gas-liquid chromatography using a DI 200 chromatograph (Perichrom, Saulx-les-Chartreux, France) equipped with a CP-Sil 88 fused-silica capillary column (length 100 m, i.d. 0.25 mm). The carrier gas was hydrogen (1.1 ml/min) in conditions of split injection (1/50). Injector and detector temperatures were 235°C and 250°C, respectively. The oven temperature was held constant for 30 s at 70°C, then increased from 70°C to 175°C at 20°C/min, held at 175°C for 25 min, increased again from 175 to 215°C at 10°C/min, and was finally held at 215°C for 41 min.
Total lipids in the liver were determined gravimetrically after their extraction by mixing 5 g of fresh tissue with chloroform/methanol 2:1 (v/v) according to the method of Folch et al. (Reference Folch, Lees and Sloane Stanley1957). The PL content was determined from total lipid extract by colorimetry after mineralization of organic phosphorus according to the method of Bartlett (Reference Bartlett1959). The TG content was analysed from total lipids according to the following method: after elimination of PL absorbed on silicic acid, TG were saponified by 4 mol/l KOH-ethanol, followed by neutralisation with 4 mol/l HCl and centrifugation. The free glycerol that was released in the supernatant was enzymatically determined using a TG test kit (PAP 1000; Biomérieux, France). The TC content was enzymatically measured from total lipid extract using a reagent kit (CHOD-iodide; Merck, Darmstadt, Germany).
Fatty streak areas
At the end of experiment 2, and after blood collection, aortas were fixed according to the method of Wilson et al. (Reference Wilson, Nicolosi, Chrysam and Kritchevsky2000) with modifications: a perfusion needle was inserted into the left ventricle and 100 ml of a 10% buffered formalin solution (HT50-1-128, Sigma-Aldrich, Saint-Quentin Fallavier, France) was slowly perfused. The aortas were cut abdominally to provide an outlet for the fixative. After perfusion, the heart and thoracic aorta were removed and fixation completed by leaving ca. 24 h in the fixative at 4°C. The tissues were then rinsed with PBS at 4°C and stored at 4°C in vials containing PBS for subsequent analysis of fatty streak area. Then, the aorta was carefully isolated from the heart to keep the arch intact, longitudinally opened and placed in a solution of 0.5 g Oil Red O (O-062, Sigma-Aldrich, Saint-Quentin Fallavier, France) in 100 ml of 70% isopropanol for 25 min. Subsequently, aorta was rinsed in 70% isopropanol for 4 min and finally in distilled water for 3 min. Aortas were then included in a mounting media with the endothelial side up, and the areas of lipid deposition in the endothelium (coloured in strong red) were graded visually using a 0 to 3 scale (0 corresponded to the absence of lipid deposition and 3 to maximal lipid deposition) by four independent observers. The lipid infiltration was easily evaluated, due to the large size of the fatty streaks and the well-responsiveness of the rabbit aorta, thus limiting the coloration artefacts, as shown in Figure 1.
Data were analysed using the two-way anova procedure of Statistical Analysis Systems Institute (SAS; 1985) to determine the effects of the diet (D), the experiment (E, experiment 1 v. experiment 2) and their interaction (D × E), and are expressed as mean ± pooled standard error of the mean. When the D and/or E effects were significant (P < 0.05), the differences between the six groups were tested using the PLSD Fisher test (Tables 2 and 4). When the D effect was significant (P < 0.05), data were analysed also with the one-way anova procedure of SAS (1985), and the differences between the three diets (as indicated in the Results section) were tested using the PLSD Fisher test. The frequencies of infiltrated aortas per group presented in Table 3 were analysed using the chi-squared test. The D and E effects were declared significant at P < 0.05, and their interaction at P < 0.10.
a,b,c Mean values within a row with different superscripts are significantly different (P < 0.05).
† P values for the diet effect (D), experiment effect (E) and their interaction (D × E). When the diet and/or experiment effects are significant (P < 0.05).
‡ Sum of C14:0, C15:0, C16:0, C17:0, C18:0 and C20:0.
§ Sum of cis-9, cis-11, and cis-12 isomers.
∥ Sum of trans-6/7/8, trans-9, trans-10, trans-11, trans-12 and trans-15 isomers.
a,b Ratios with unlike superscript letters are significantly different between diets (P < 0.05).
† Values are presented as the number of rabbits showing no fatty streak (0), mild fatty streak (+, mean score between 0.25 and 1.5) or severe fatty streak (++, mean score between 1.5 and 3) areas in aorta wall.
‡ Number of rabbits showing a lipid infiltration in aorta wall/total number of rabbits.
The three experimental diets were well accepted by the rabbits, except for two rabbits fed the T10 diet. The food intakes have been reported elsewhere (Faulconnier et al., Reference Faulconnier, Roy, Ferlay, Chardigny, Durand, Lorenz, Gruffat and Chilliard2006).
Plasma fatty acid composition
The proportion of total saturated FA in rabbit plasma tended to be affected by the diet (P < 0.08) and the duration of feeding diets (experiment effect, P < 0.07) (Table 2). The proportions of C14:0, C15:0, C16:0 and C17:0 were decreased between experiment 1 and experiment 2 with the three diets, while that of C18:0 was increased between experiment 1 and experiment 2 with the S and T10 diets.
The effect of the diet was significant (P < 0.03, T10 < T11-CLA diet) on the proportions of plasma total cis-C18:1. The proportion of total trans-C18:1 was higher in plasma from rabbits fed the T10 and T11-CLA diets than in rabbits fed the S diet in the two experiments (P < 0.05). Particularly, the diets rich in trans FA resulted in higher plasma proportion of trans-9 C18:1 in the two experiments, the T11-CLA diet increased those of trans-11, trans-12 and trans-15 C18:1 and the T10 diet increased that of trans-10 C18:1 in both experiments, compared with the S diet (P < 0.05).
Conversely, n-3 FA proportions were decreased, and the n-6 to n-3 ratio was increased, in plasma from rabbits of T10 and T11-CLA groups compared with the S group (P < 0.05), and the C18:3 n-3 proportion was particularly decreased in plasma from rabbits fed either the T10 or the T11-CLA diets (significant D × E interaction, P < 0.01). The content of cis-9, trans-11 CLA in plasma was increased by four-fold with the T11-CLA diet, compared with the two other diets in the two experiments (P < 0.05). The diet effect was significant for the plasma proportions of n-6 FA (P < 0,01, T11-CLA < T10 and S diets).
Fatty streak areas and plasma concentrations of eicosanoids
The proportion of rabbits showing lipid infiltration in the aorta was significantly higher in the T10 diet than in the T11-CLA diet (P < 0.05, Table 3). The proportion of rabbits showing aortic lipid infiltration was lower in the T11-CLA group than in the S group, although this difference was not significant.
Feeding butters of standard FA composition or rich in either trans-10 C18:1 or trans-11 C18:1+ cis-9, trans-11 CLA included into a pro-atherogenic diet for 6 or 12 weeks did not result in significant differences in the plasma concentrations of PGI2 and TxA2 (Table 4).
a,b,c Mean values within a row with different superscripts are significantly different (P < 0.05).
† P values for the diet effect (D), experiment effect (E) and their interaction (D × E). When the diet and/or experiment effects are significant (P < 0.05).
‡ LDL fraction included IDL+LDL particles.
§ The ratio was calculated as cholesterol content of IDL+LDL particles / cholesterol content of HDL fraction.
∥ The ratio was calculated as plasma total cholesterol / total cholesterol content of HDL fraction.
¶ Non-HDL fraction included chylomicrons, VLDL, IDL and LDL particles.
Plasma lipid and lipoprotein concentration and composition
The cholesterolemia tended to be modified by the diet (P < 0.09, T10>T11-CLA diet) (Table 4).The plasma cholesterol concentration in the VLDL fraction (VLDL-C) was significantly changed by the dietary treatment (P < 0.04, T10>T11-CLA and S diets). The plasma concentration of cholesterol in the LDL fraction (LDL-C) tended to be dependent on the dietary treatment (P < 0.07, T10>T11-CLA diet, S diet being intermediate). The plasma concentration of cholesterol in the HDL fraction (HDL-C) tended to be differently affected by the dietary treatments (P < 0.06, T11-CLA < S and T10 diets) and was significantly decreased between experiment 1 and experiment 2 (P < 0.02), particularly in the T10 group (P < 0.05).
The triglyceridemia was differently affected by the duration of feeding the diets (D × E interaction, P < 0.08). Indeed, only the S diet increased the triglyceridemia between experiment 1 and experiment 2 (P < 0.05), and the triglyceridemia was significantly lower with the T10 diet than with the S diet in experiment 2 (P < 0.05).
The plasma concentrations of the HDL fraction were dependent on diet (P < 0.02, T11-CLA < S diet, T10 diet being intermediate). The plasma concentrations of non-HDL fraction (including chylomicrons, VLDL, IDL and LDL) were also modified by the dietary treatment (P < 0.05, T10>S and T11-CLA diets). Moreover, non-HDL to HDL ratio tended to be affected by diet (P < 0.1, T10>S diet, T11-CLA diet being intermediate).
The severity of FA streak infiltration in aorta at 12 weeks (experiment 2) was positively and curvilinearly correlated with cholesterolemia, triglyceridemia, TC to HDL-C and non-HDL-C to HDL-C ratios (Table 5).
† Plasma total cholesterol to HDL-cholesterol ratio.
‡ Plasma non-HDL-cholesterol to HDL-cholesterol ratio.
While no difference was observed among the diets and between the experiments in the total lipid and TC contents of the liver, the T11-CLA diet slightly, but significantly increased (P < 0.05) the TG content of the liver in experiment 2 compared with the two other diets (Table 4). The T10 and T11-CLA diets resulted in a significant increase (P < 0.05) in the TG content of the liver between experiment 1 and experiment 2.
When compiling the data from the three diets in experiment 2, positive curvilinear correlations were observed between the severity of FA streak infiltration in aorta and the liver lipid or total cholesterol content (Table 5).
While the food intake was similar for all rabbits receiving the S and T11-CLA butters, two rabbits of the group receiving the T10 butter showed significant food refusals during the last week of experiment 2 (Faulconnier et al., Reference Faulconnier, Roy, Ferlay, Chardigny, Durand, Lorenz, Gruffat and Chilliard2006). These two rabbits also showed largely elevated cholesterolemia and non-HDL concentrations, a fatty liver and lipid infiltrations in the aorta wall. Thus, both the relatively limited number of rabbits per group and the great individual variability within groups, especially in the T10 group, could explain at least in part the fact that differences in several parameters among the three groups were not significant. Consequently, this justifies caution in interpreting and discussing the results.
The trans-10 and trans-11 C18:1 and cis-9, trans-11 CLA contents of the rabbit plasmas well reflected the type of ingested butter, being already significant after 6 weeks of feeding the enriched butters. Moreover, intake of trans FA-rich butters lowered the plasma content of total n-3 FA and particularly C18:3 n-3, despite the fact that all diets provided the same amounts of C18:3 n-3 and C18:2 n-6. The decrease in plasma content of C18:3 n-3 could be detrimental, since it is the precursor of the n-3 FA family, for which the cardio-protective properties have been well established (De Lorgeril and Salen, Reference De Lorgeril and Salen2004).
The main eicosanoid secreted by endothelial cells of aorta, namely PGI2, is known to inhibit platelet aggregation and to act as a vasodilator, while TxA2, the primary eicosanoid formed by platelets, induces platelet aggregation and act as a vasoconstrictor (Torres-Duarte and Vanderhoek, Reference Torres-Duarte and Vanderhoek2003). These two antagonist eicosanoids are both end products of the enzymatic oxidation of C20:4 n-6, regulated by the cyclooxygenase-2 enzyme (COX-2). Consequently, a low level of PGI2 coupled with a high level of TxA2 in plasma have been related to thrombosis and ischemia (Oates et al., Reference Oates, Fitzgerald, Branch, Jackson, Knapp and Roberts1988). Feeding the T11-CLA butter did not affect the plasma levels of PGI2 and TxA2 compared with the S butter in the present study, whereas in vitro studies have shown that cis-9, trans-11 CLA inhibited the formation of thromboxane in washed human platelets (Truitt et al., Reference Truitt, McNeill and Vanderhoek1999), and decreased the production of PGI2 by human aortic endothelial cells (Eder et al., Reference Eder, Schleser, Becker and Korting2003). In a recent study, a cis-9, trans-11 CLA-enriched butter oil has been shown to down-regulate the expression of the COX-2 gene in the aorta of cholesterol-fed hamsters when compared with a standard butter oil, suggesting an anti-inflammatory effect (Valeille et al., Reference Valeille, Férézou, Amsler, Quignard-Boulangé, Parquet, Gripois, Dorovska-Taran and Martin2005). Furthermore, the authors also reported a reduced expression of the vascular cell adhesion molecule 1 (VCAM-1) gene in aorta from hamsters fed the cis-9, trans-11 CLA-enriched butter oil, suggesting a reduced infiltration of lipids into the aorta wall.
In line with this, feeding the T11-CLA butter could have protected against atherosclerosis development in the present study, since two-thirds of the rabbits fed this diet did not show any fatty streak in aorta, while only one rabbit among the two other diets did not show any lipid deposition. However, this beneficial effect did not reach the significance between the T11-CLA and the S diets. The dietary intake of cis-9, trans-11 CLA was of about 0.2% of the diet, and is in the range (0.05 to 0.5%) of the dietary concentrations of a synthetic CLA-mixture (cis-9, trans-11 and trans-10, cis-12 CLA, in equal proportions) that have been shown to reduce the severity of lesions in arch and thoracic aorta in rabbits (Kritchevsky et al., Reference Kritchevsky, Tepper, Wright and Czarnecki2002). Furthermore, the dietary intake of trans-11 C18:1 (0.5% of the diet) could constitute a potential precursor pool for the formation of cis-9, trans-11 CLA through the action of Δ9-desaturase in rabbit tissues, as already demonstrated in human (Turpeinen et al., 2002) and rodents (Santora et al., Reference Santora, Palmquist and Roehrig2000; Corl et al., Reference Corl, Barbano, Bauman and Ip2003).
Compared with the S butter, the T11-CLA butter did not modify the cholesterolemia. This is in accordance with previous studies comparing a standard butter with a cis-9, trans-11 CLA-enriched butter in hamsters (Valeille et al., Reference Valeille, Férézou, Amsler, Quignard-Boulangé, Parquet, Gripois, Dorovska-Taran and Martin2005), or comparing a CLA mixture-supplemented diet with a non-supplemented diet in rabbits (Kritchevsky et al., Reference Kritchevsky, Tepper, Wright and Czarnecki2002 and Reference Kritchevsky, Tepper, Wright, Czarnecki, Wilson and Nicolosi2004) and hamsters (Mitchell et al., Reference Mitchell, Langille, Currie and McLeod2005). However, several other studies with hamsters reported a decrease in cholesterolemia, when a trans-11 C18:1+cis-9, trans-11 CLA-rich butter was compared with a standard butter (Lock et al., Reference Lock, Horne, Bauman and Salter2005), or when a CLA mixture-supplemented diet was compared with a non-supplemented diet (Nicolosi et al., Reference Nicolosi, Roger, Kritchevsky, Scimeca and Huth1997; Wilson et al., Reference Wilson, Nicolosi, Chrysam and Kritchevsky2000). It should be noted that studies testing the effects of a CLA-mixture supplementation on cholesterolemia in humans also generally gave unclear results (reviewed by Terspstra (Reference Terpstra2004)). The positive correlation between the cholesterolemia and the severity of fatty streak in aorta observed in the present study in rabbits corroborates the well established atherogenic effect of hypercholesterolemia in humans (Libby et al., Reference Libby, Aikawa and Schönbeck2000). Moreover, the rabbits fed the T11-CLA butter resulted in lower plasma concentrations of HDL and HDL-C than the rabbits fed the S butter (without effect neither on the non-HDL to HDL ratio, nor on the LDL-C to HDL-C ratio). The HDL-C-lowering effect of the T11-CLA butter was not reported in previous animal studies (Kritchevsky et al., Reference Kritchevsky, Tepper, Wright and Czarnecki2002; Lock et al., Reference Lock, Horne, Bauman and Salter2005; Mitchell et al., Reference Mitchell, Langille, Currie and McLeod2005), although non-significant decreases in HDL-C have been reported in several studies in humans receiving a CLA mixture (cis-9, trans-11 CLA and trans-10, cis-12 CLA) for 4 to 12 weeks (reviewed by Terpstra (Reference Terpstra2004)). Although high plasma levels of HDL-C have been associated with a reduced incidence of CVD in humans (review of Barter et al., Reference Barter, Kastelein, Nunn and Hobbs2003), the lower levels of HDL-C in the plasma from rabbits fed the T11-CLA butter was not associated with an increased aortic lipid infiltration, in the present study.
The curvilinear correlation between the severity of aortic lipid deposition and the non-HDL-C to HDL-C ratio observed in the present study matches with the linear correlation reported by Valeille et al. (Reference Valeille, Férézou, Amsler, Quignard-Boulangé, Parquet, Gripois, Dorovska-Taran and Martin2005) in hamsters between CE deposition in aorta and non-HDL-C to HDL-C ratio. This adds evidence for non-HDL-C to HDL-C ratio as a risk factor of atherosclerosis, and CE deposition as the major part of the lipid infiltration in aorta wall. Furthermore, the positive correlation between plasma TC to HDL-C ratio and lipid deposition in aorta wall observed in rabbits is in accordance with the fact that this ratio has been recently considered as the most representative risk factor of CVD in humans (Mensink et al., Reference Mensink, Zock, Kester and Katan2003).
The correlations between the severity of fatty streak areas and either the TC content of the liver or the plasma cholesterol concentrations observed in rabbits corroborate data from experiments carried out with mice lacking liver X receptors (LXR) gene. Liver X receptors are members of the nuclear receptor family of lipid-sensing transcription factors (Chawla et al., Reference Chawla, Repa, Evans and Mangelsdorf2001) and have emerged as key regulators of cholesterol and lipid metabolism (Barish and Evans, Reference Barish and Evans2004). The hepatic and intestinal LXR genes regulate the cholesterol homeostasis by controlling intestinal absorption, reverse cholesterol transport, endogenous cholesterol synthesis and catabolism, and excretion into the bile (Millatt et al., Reference Millatt, Bocher, Fruchart and Staels2003). Mice genetically lacking LXRα develop marked hepatic steatosis and hypercholesterolemia when fed a cholesterol-containing high-fat diet (Peet et al., Reference Peet, Janowski and Mangelsdorf1998) and show increased formation of aortic foam cells, which are precursors for atherosclerosis (Schuster et al., Reference Schuster, Parini, Wang, Alberti, Steffensen, Hansson, Angelin and Gustafsson2002). Thus, the atherogenic diets offered to the rabbits in the present study could have mimicked the absence of LXR in genetically deficient mice, and led to comparable symptoms.
The T11-CLA butter did not modify triglyceridemia compared with the S butter, as reported in rabbits (Lee et al., Reference Lee, Kritchevsky and Pariza1994; Kritchevsky et al., Reference Kritchevsky, Tepper, Wright and Czarnecki2002 and Reference Kritchevsky, Tepper, Wright, Czarnecki, Wilson and Nicolosi2004) and hamsters (Lock et al., Reference Lock, Horne, Bauman and Salter2005; Mitchell et al., Reference Mitchell, Langille, Currie and McLeod2005). In addition, the increase in triglyceridemia between 6 and 12 weeks of feeding the S butter was not observed in the rabbits fed the T11-CLA butter. Additionally, feeding the T11-CLA butter for 12 weeks resulted in a slightly higher TG deposition in the liver of the rabbits. Thus, the stability of triglyceridemia between 6 and 12 weeks of feeding the T11-CLA butter suggests that this diet could have limited the hypertriglyceridemia, by increasing the TG deposition in the liver.
The positive close relationship between the triglyceridemia and the severity of lipid deposition in the aorta of rabbits was already reported in hamsters (Mangiapane et al., Reference Mangiapane, McAteer, Benson, White and Salter1999) and is in accordance with the fact that hypertriglyceridemia is an independent risk factor of atherosclerosis in human (Hennig et al., Reference Hennig, Toborek and McClain2001).
One key result of the present study is that the consumption of the T10 butter had a more detrimental effect than the T11-CLA butter on fatty streak areas, although not differing from the S butter. When the two experiments study are taken together, the T10 butter resulted in higher plasma levels of VLDL-C and non-HDL particles, and non-HDL to HDL ratio than the S butter. Moreover, the T10 butter increased cholesterolemia and plasma levels of LDL-C, compared with the T11-CLA butter. This suggests that the T10 butter resulted in impaired lipoprotein and cholesterol metabolism. Owing to the deep dyslipidemia and dyslipoproteinemia already observed in the T10 group at the end of experiment 1 (6 weeks), the altered cholesterolemia and lipoproteinemia could have been the cause of the decrease in dietary intake observed in two rabbits of the T10 group during week 12 of the experiment 2. As the effects on lipemia, lipoproteinemia and atherosclerosis of the trans-10 C18:1, either as purified isomer or present at high level in a dietary product, have never been tested before, this study is the first to report experimental data on the putative effects of this trans FA.
The correlations observed in the present study between the severity of lipid infiltration in the aorta wall and several risk factors of CVD, including triglyceridemia, cholesterolemia, the non-HDL-C to HDL-C and TC to HDL-C ratios, add new evidence to consider the cholesterol-induced atherosclerosis in rabbit as a suitable and representative model for studying diet-induced dyslipidemia and atherosclerosis in humans.
The results of this study suggest that the trans-11 C18:1+cis-9, trans-11 CLA-rich butter included into an atherogenic diet for 12 weeks decreased, compared with the trans-10 C18:1-rich butter and, to a lesser extent, to the standard butter, the occurrence of atherosclerosis, and increased liver TG content without significantly affecting lipemia and lipoproteinemia, in rabbits (except a lowering effect in the plasma HDL particles and HDL-C levels). The trans-10 C18:1-rich butter caused impaired plasma cholesterol metabolism and non-HDL levels, compared with the standard butter, which probably led to decreased dietary intake in one-third of the rabbits in experiment 2. However, further studies involving larger number of rabbits, longer experimental period and/or lower level of dietary cholesterol, are warranted to confirm both the putative atheroprotective role of the couple trans-11 C18:1+cis-9, trans-11 CLA, and the suggested negative influence of a trans-10 C18:1-rich butter on experimental atherosclerosis, plasma lipid and lipoprotein profiles.
This work was funded by the INRA ‘Micronutrients in animal products: CLA-B12’ project, managed by Jean-Louis Sébédio and Dominique Bauchart, and ‘AQS 2001’ project managed by Michel Doreau. The authors would like to thank S. Almanza, B. Buteau, S. Grégoire and L. Leclere for technical assistance.