169
The Effect of Dietary Supplementation with Eicosapentaenoic Acid on the Phospholipid and Fatty Acid Composition of Erythrocytes of Marmoset Robert A. Gibsona, *, Mark A. Neumann a, Sharon L. Burnard b, Josephine A. Rinaldi b, Glen S. Patten b and Edward J. McMurchie b aDepartment of Paediatrics, Flinders Medical Centre, Bedford Park, S.A., 5042 and bCSIRO, Division of Human Nutrition, Glenthorne Laboratory, O'Halloran Hill, S.A., 5158 Australia Adult male marmoset monkeys were fed eicosapentaenoic acid (20:5n-3) as the ethyl ester in diets containing either 32% (reference diet, no added cholesterol) or 7% {atherw genic diet with 0.2% added cholesterol) linoleic acid (18:2n4~) for 30 wk. No changes were seen in the level of phosphatidylcholine (PC) or phosphatidylethanolamine (PE) but minor changes were observed in both the sphingomyelin (SPM) and phosphatidylinositol plus phosphatidylserine (PI + PS) fractions of erythrocyte lipids. The extent of total n~3 fatty acid incorporation into membrane lipids was higher in atherogenic diets (polyunsaturated/monounsaturated/saturated (P/M/S) ratio 0.2:0.6:1.0) than reference diets {PfM/S ratio 1:1:1) and this was true for both PE {33.4 ___1.0% vs 24.3 ___1.1%) and PC (9.3 _+ 0.5% vs 4.9 -- 0.3%). Although suitable controls for cholesterol effects were not included in the study, earlier results obtained with marmosets lead us to believe such effects were probably small. Regardless of basic diet (atherogenic, reference), 20:5n-3 was preferentially incorporated into PE (10.8 ___ 0.2%, 6.0 • 0.02%) while smaller amounts were incol~ porated into PC (6.9 +__0.4%, 3.2 • 0.2%). The major n-3 polyunsaturated fatty acid found in PE in response to dietary 20:5n-3 was the elongation metabolite 22:5n-3 in both the atherogenic (17.7 • 0.7%) and reference (14.3 • 1.0%) dietary groups; 22:6n-3 levels were less affected by diet (4.7 • 0.3% and 3.9 • 0.2%, respectively). The results can be interpreted to indicate an inverse relationship between the amount of dietary 18:2n-6 and incorporation of 20:5n-3 into erythrocyte membrane phospholipids regardless of whether the major dietary n-3 fatty acid was a-linolenate (18:3n-3) or 20:5n-3. This interpretation is supported by theoretical calculations. Lipids 27, 169-176 (1992).
to affect both the deformability of red cells (9,10) and the fluidity of erythrocyte membranes from human subjects (11,12}. Similar studies on rats have failed to demonstrate changes in erythrocyte deformability despite an increase in n-3 polyunsaturated fatty acids in membrane phospholipids
(13). Many beneficial effects of marine oil consumption by humans are now attributed to incorporation of 20:5n-3 into cell membranes (1,2). However, because 20:5n-3 incorporation into liver phospholipids has been shown to be less when fish oil was fed with safflower oil than when it was fed with beef tallow (14), co-ingestion of other fats, particularly those rich in linoleate (18:2n~3) must be considered in order to maximize 20:5n-3 incorporatiorL Interpretation of results from animals fed marine oils has often been difficult due to the presence of several n-3 polyunsaturated fatty acids in these oils. The availability of a source of ethyl esters derived from fish oil which contained over 70% 20:5n-3 and no 22:6n-3, made it possible to determine the effect of dietary 20:5n-3 on the fatty acid composition of individual phospholipids of the marmoset red blood cell in the absence of other n-3 polyunsaturated fatty acids. We demonstrate that dietary 20:5n-3 is preferentially incoP porated into erythrocyte phosphatidylethanolamine (PE), and this incorporation is greatest when dietary 18:2n-6 is reduced as in an atherogenic-type diet. MATERIALS AND METHODS
Marmosets. Adolescent male common cotton-eared marmosets (Callithrixjacchus jacchus), approximately 16 to 26 mon old at the start of the experiment, were divided into four groups of equivalent age and weight. Marmosets were paired for optimum growth and social behavior and kept in aluminum alloy marmoset cages in a room with The n-3 polyunsaturated fatty acid eicosapentaenoic acid fluorescent light and 30 min of ultraviolet irradiation daily. (20:5) and docosahexaenoic acid (22:6) which are present in The temperature was maintained at 26~ and the humidimarine oils have unique lipid lowering potential in humans ty at 50%. Marmosets were maintained on the various and animals {1,2}. Fish oil has been shown to lower triglyc- dietary lipid regimes described below for a period of 30 eride levels in both normal and hypertriglyceridemic sub- wk and were sacrificed under anaesthesia by injection of jects (3). In the rat, the hypotriglyceridemic effect is prob- Saffan, 1.5 mL per kg body weight (alphaxalone" 9 mg/mL; ably mediated by 20:5n-3 while 22:6n-3 may be responsible alphadolone acetate, 3 mg/mL; Glax~ Sydney, Australia), for lowering of plasma cholesterol levels (4). Although in- into the femoral artery. vestigations into the effect of n-3 polyunsaturated fatty Marmoset diets. Four high-fat dietary regimes of equiacids on blood cells have tended to focus on platelets {5,6} valent energy value were used. These diets were chosen or neutrophils (7,8), changes in phospholipid fatty acids and to examine the effect of 20:5n-3 (as the ethyl ester) when membrane function of erythrocyt~ have also been exam- administered in combination with diets differing signifiine& For example, dietary n-3 fatty acids have been shown cantly in their polyunsaturated/monounsaturated/saturated (P/M/S) fatty acid ratio. The normal colony diet for marmosets consisted of a 1:1 mixture of greyhound chow *To whom correspondence should be addressed. (Arnott Harper's Ltd., Adelaide. Australia) and primate Abbreviations:ATH, atherogenicdiet;EPA, eicosapentaenoicacid; meal (Milling Industries Ltd., Adelaide. Australia). The FID, flamedonizationdetector,HUFA, highlyunsaturatedfattyacids; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, overall composition of this diet with regard to all nutrients phosphatidylinositol; PS, phosphatidylserine; P/M/S, polyun- has previously been described (15), and contained approxsaturated]monounsaturated/saturated ratio;REF, referencediet;SPM, imately 4.5% fat. The fatty acid composition of the norsphingomyelin. real colony diet (shown in Table 1) was used as the basis LIPIDS, Vol. 27, no. 3 (1992)
170
R.A. GIBSON ETAL. for the preparation of the four experimental diets described below. The reference (REF) diet consisted of the colony diet supplemented with 5.5% (by weight) sunflower seed oil (Nuttelex, Melbourne, Australia), 1.5% olive oil (SA Olive Oil Co., Adelaide, Australia) and 3.0% sheep kidney (perirenal) fat (Noarlunga Meat Works, Noarlunga, Australia). The final fat blend achieved a P/M/S ratio of 1.1:1.1:1.0. Addition of 0.8 (w/w) 20:5n-3 ethyl ester concentrate (containing 75% 20:5n-3 and 0.2% a-tocopherol; supplied by Dr. Yasushi Tamura, Chiba University Medical School, Chiba, Japan) changed the P/M/S ratio to 1.1:1.0:1.0 (REF + eicosapentaenoic acid (EPA) diet). The atherogenic (ATH) diet consisted of the colony diet supplemented with 10% sheep kidney fat and 0.2% (w/w) cholesterol (Ajax Chemical Co., Adelaide, Australia) which achieved a P/M/S ratio of 0.14:0.6:1.0. The addition of 0.8% (w/w) 20:5n-3, ethyl ester concentrate to the ATH diet gave a P/M/S ratio of 0.2:0.6:1.0 (ATH + EPA diet). The f a t t y acid compositions of the 20:5n-3 concentrate and of the experimental diets are shown in Table 1. All additions (i.e. otis, fat, cholesterol and concentrate) were thoroughly mixed with the crushed colony diet before the respective diets were repelleted. A two-week supply of each diet was prepared and frozen at - 2 0 ~ in a N 2 atmospher~ Aliquots of these diets were thawed for each daily feeding and replaced daily. Animals were fed ad libiturn. Collection of erythrocytes. Immediately after the animals were killed, blood was collected from the aorta by a heparinized syringe, and centrifuged to separate plasma and erythrocytes as previously described (16). Erythrocytes were washed three times by centrifugation in phosphate buffered saline, and lipids were extracted immediately. Lipid extraction. Lipids were extracted by the methods
described elsewhere (16). Briefly, one volume of washed erythrocytes was extracted with 4 vol of boiling 2-propanol containing the antioxidant butylated hydroxyanisole (0.1% of the estimated lipid weight), and the mixture was boiled for 30 sea After cooling, 8 vol of chloroform and 2 vol of water were added, and the mixture was shaken. Following centrifugation, the lower organic phase was collected and the aqueous phase was re-extracted with an additional 4 vol of chloroform. The combined organic phases were dried over anhydrous sodium sulfat~
Separation of total phospholipids and phospholipid classes. Total phospholipids were separated from neutral lipids by thin-layer chromatography on silica gel H plates using petroleum hydrocarbon acetone (3:1, v/v) as the developing solvent. Individual phospholipid classes (as % of total phospholipids) were separated on Whatman LK5D thin-layer chromatography plates (Maidstone, England) developed in one direction in a solvent system of chloroform]ethanol/water/triethylamine (30:34:8:35, by vol) as described by Touchstone et aL (17). All phospholipids were clearly separated except phosphatidylinositol (PI) and phosphatidylserine (PS) which were only present in trace amounts; samples were pooled as indicated in Table 2.
Fatty acid analysis of erythrocyte total phospholipids and phospholipid classes. Phospholipid bands isolated by thin-layer chromatography were scraped directly into glass vials containing 1% (v/v) H2SO4 in methanol and heated for 3 hr at 70~ The resulting methyl esters were extracted into heptane and dried over sodium sulfate prior to analysis by gas chromatography. Analyses of f a t t y acid methyl esters of the total phospholipids and the major phospholipid classes of marmoset erythrocytes were performed using capillary gas chromatography (HewlettPackard H P 5880 gas chromatograph; Hewlett-Packard, Palo Alt~ CA). The column was a 50-m glass column
TABLE 1 Fatty Acid Composition of Marmoset Colony Diet, 20:Sn-3(EPA) Concentrate and Experimental Diets a
Major fatty acid (%; w/w) Saturated Monounsaturated n-6 18:2 20:4 n-3 18:3 20:5 I(n-6) Z(n-3) :E(n-6)/Z(n-3) U.I.
Colony diet 46.0 39.5
Concentrate 2.1 16.2
REF 31.3 33.7
Experimental diets REF + EPA ATH 31.8 56.1 32.9 35.9
13.1 tr.
0.6 4.5
32.9 n.d.
31.0 tr.
6.9 n.d.
6.6 tr.
1.4 n.d. 13.1 1.4 9.4 70
0.5 74.8 5.1 75.3 0.07 411
0,9 n.d. 32.9 0.9 36.5 102
0.9 2.9 31.0 3.8 8.2 112
1.1 n.d. 6.9 1.1 6.3 53
1.0 3.3 6.6 4.3 1.5 65
Cholesterol(%) Total fat(%) Energy Value(KJ/g)
The Effect of Dietary Supplementation with Eicosapentaenoic Acid on the Phospholipid and Fatty Acid Composition of Erythrocytes of Marmoset Robert A. Gibsona, *, Mark A. Neumann a, Sharon L. Burnard b, Josephine A. Rinaldi b, Glen S. Patten b and Edward J. McMurchie b aDepartment of Paediatrics, Flinders Medical Centre, Bedford Park, S.A., 5042 and bCSIRO, Division of Human Nutrition, Glenthorne Laboratory, O'Halloran Hill, S.A., 5158 Australia Adult male marmoset monkeys were fed eicosapentaenoic acid (20:5n-3) as the ethyl ester in diets containing either 32% (reference diet, no added cholesterol) or 7% {atherw genic diet with 0.2% added cholesterol) linoleic acid (18:2n4~) for 30 wk. No changes were seen in the level of phosphatidylcholine (PC) or phosphatidylethanolamine (PE) but minor changes were observed in both the sphingomyelin (SPM) and phosphatidylinositol plus phosphatidylserine (PI + PS) fractions of erythrocyte lipids. The extent of total n~3 fatty acid incorporation into membrane lipids was higher in atherogenic diets (polyunsaturated/monounsaturated/saturated (P/M/S) ratio 0.2:0.6:1.0) than reference diets {PfM/S ratio 1:1:1) and this was true for both PE {33.4 ___1.0% vs 24.3 ___1.1%) and PC (9.3 _+ 0.5% vs 4.9 -- 0.3%). Although suitable controls for cholesterol effects were not included in the study, earlier results obtained with marmosets lead us to believe such effects were probably small. Regardless of basic diet (atherogenic, reference), 20:5n-3 was preferentially incorporated into PE (10.8 ___ 0.2%, 6.0 • 0.02%) while smaller amounts were incol~ porated into PC (6.9 +__0.4%, 3.2 • 0.2%). The major n-3 polyunsaturated fatty acid found in PE in response to dietary 20:5n-3 was the elongation metabolite 22:5n-3 in both the atherogenic (17.7 • 0.7%) and reference (14.3 • 1.0%) dietary groups; 22:6n-3 levels were less affected by diet (4.7 • 0.3% and 3.9 • 0.2%, respectively). The results can be interpreted to indicate an inverse relationship between the amount of dietary 18:2n-6 and incorporation of 20:5n-3 into erythrocyte membrane phospholipids regardless of whether the major dietary n-3 fatty acid was a-linolenate (18:3n-3) or 20:5n-3. This interpretation is supported by theoretical calculations. Lipids 27, 169-176 (1992).
to affect both the deformability of red cells (9,10) and the fluidity of erythrocyte membranes from human subjects (11,12}. Similar studies on rats have failed to demonstrate changes in erythrocyte deformability despite an increase in n-3 polyunsaturated fatty acids in membrane phospholipids
(13). Many beneficial effects of marine oil consumption by humans are now attributed to incorporation of 20:5n-3 into cell membranes (1,2). However, because 20:5n-3 incorporation into liver phospholipids has been shown to be less when fish oil was fed with safflower oil than when it was fed with beef tallow (14), co-ingestion of other fats, particularly those rich in linoleate (18:2n~3) must be considered in order to maximize 20:5n-3 incorporatiorL Interpretation of results from animals fed marine oils has often been difficult due to the presence of several n-3 polyunsaturated fatty acids in these oils. The availability of a source of ethyl esters derived from fish oil which contained over 70% 20:5n-3 and no 22:6n-3, made it possible to determine the effect of dietary 20:5n-3 on the fatty acid composition of individual phospholipids of the marmoset red blood cell in the absence of other n-3 polyunsaturated fatty acids. We demonstrate that dietary 20:5n-3 is preferentially incoP porated into erythrocyte phosphatidylethanolamine (PE), and this incorporation is greatest when dietary 18:2n-6 is reduced as in an atherogenic-type diet. MATERIALS AND METHODS
Marmosets. Adolescent male common cotton-eared marmosets (Callithrixjacchus jacchus), approximately 16 to 26 mon old at the start of the experiment, were divided into four groups of equivalent age and weight. Marmosets were paired for optimum growth and social behavior and kept in aluminum alloy marmoset cages in a room with The n-3 polyunsaturated fatty acid eicosapentaenoic acid fluorescent light and 30 min of ultraviolet irradiation daily. (20:5) and docosahexaenoic acid (22:6) which are present in The temperature was maintained at 26~ and the humidimarine oils have unique lipid lowering potential in humans ty at 50%. Marmosets were maintained on the various and animals {1,2}. Fish oil has been shown to lower triglyc- dietary lipid regimes described below for a period of 30 eride levels in both normal and hypertriglyceridemic sub- wk and were sacrificed under anaesthesia by injection of jects (3). In the rat, the hypotriglyceridemic effect is prob- Saffan, 1.5 mL per kg body weight (alphaxalone" 9 mg/mL; ably mediated by 20:5n-3 while 22:6n-3 may be responsible alphadolone acetate, 3 mg/mL; Glax~ Sydney, Australia), for lowering of plasma cholesterol levels (4). Although in- into the femoral artery. vestigations into the effect of n-3 polyunsaturated fatty Marmoset diets. Four high-fat dietary regimes of equiacids on blood cells have tended to focus on platelets {5,6} valent energy value were used. These diets were chosen or neutrophils (7,8), changes in phospholipid fatty acids and to examine the effect of 20:5n-3 (as the ethyl ester) when membrane function of erythrocyt~ have also been exam- administered in combination with diets differing signifiine& For example, dietary n-3 fatty acids have been shown cantly in their polyunsaturated/monounsaturated/saturated (P/M/S) fatty acid ratio. The normal colony diet for marmosets consisted of a 1:1 mixture of greyhound chow *To whom correspondence should be addressed. (Arnott Harper's Ltd., Adelaide. Australia) and primate Abbreviations:ATH, atherogenicdiet;EPA, eicosapentaenoicacid; meal (Milling Industries Ltd., Adelaide. Australia). The FID, flamedonizationdetector,HUFA, highlyunsaturatedfattyacids; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, overall composition of this diet with regard to all nutrients phosphatidylinositol; PS, phosphatidylserine; P/M/S, polyun- has previously been described (15), and contained approxsaturated]monounsaturated/saturated ratio;REF, referencediet;SPM, imately 4.5% fat. The fatty acid composition of the norsphingomyelin. real colony diet (shown in Table 1) was used as the basis LIPIDS, Vol. 27, no. 3 (1992)
170
R.A. GIBSON ETAL. for the preparation of the four experimental diets described below. The reference (REF) diet consisted of the colony diet supplemented with 5.5% (by weight) sunflower seed oil (Nuttelex, Melbourne, Australia), 1.5% olive oil (SA Olive Oil Co., Adelaide, Australia) and 3.0% sheep kidney (perirenal) fat (Noarlunga Meat Works, Noarlunga, Australia). The final fat blend achieved a P/M/S ratio of 1.1:1.1:1.0. Addition of 0.8 (w/w) 20:5n-3 ethyl ester concentrate (containing 75% 20:5n-3 and 0.2% a-tocopherol; supplied by Dr. Yasushi Tamura, Chiba University Medical School, Chiba, Japan) changed the P/M/S ratio to 1.1:1.0:1.0 (REF + eicosapentaenoic acid (EPA) diet). The atherogenic (ATH) diet consisted of the colony diet supplemented with 10% sheep kidney fat and 0.2% (w/w) cholesterol (Ajax Chemical Co., Adelaide, Australia) which achieved a P/M/S ratio of 0.14:0.6:1.0. The addition of 0.8% (w/w) 20:5n-3, ethyl ester concentrate to the ATH diet gave a P/M/S ratio of 0.2:0.6:1.0 (ATH + EPA diet). The f a t t y acid compositions of the 20:5n-3 concentrate and of the experimental diets are shown in Table 1. All additions (i.e. otis, fat, cholesterol and concentrate) were thoroughly mixed with the crushed colony diet before the respective diets were repelleted. A two-week supply of each diet was prepared and frozen at - 2 0 ~ in a N 2 atmospher~ Aliquots of these diets were thawed for each daily feeding and replaced daily. Animals were fed ad libiturn. Collection of erythrocytes. Immediately after the animals were killed, blood was collected from the aorta by a heparinized syringe, and centrifuged to separate plasma and erythrocytes as previously described (16). Erythrocytes were washed three times by centrifugation in phosphate buffered saline, and lipids were extracted immediately. Lipid extraction. Lipids were extracted by the methods
described elsewhere (16). Briefly, one volume of washed erythrocytes was extracted with 4 vol of boiling 2-propanol containing the antioxidant butylated hydroxyanisole (0.1% of the estimated lipid weight), and the mixture was boiled for 30 sea After cooling, 8 vol of chloroform and 2 vol of water were added, and the mixture was shaken. Following centrifugation, the lower organic phase was collected and the aqueous phase was re-extracted with an additional 4 vol of chloroform. The combined organic phases were dried over anhydrous sodium sulfat~
Separation of total phospholipids and phospholipid classes. Total phospholipids were separated from neutral lipids by thin-layer chromatography on silica gel H plates using petroleum hydrocarbon acetone (3:1, v/v) as the developing solvent. Individual phospholipid classes (as % of total phospholipids) were separated on Whatman LK5D thin-layer chromatography plates (Maidstone, England) developed in one direction in a solvent system of chloroform]ethanol/water/triethylamine (30:34:8:35, by vol) as described by Touchstone et aL (17). All phospholipids were clearly separated except phosphatidylinositol (PI) and phosphatidylserine (PS) which were only present in trace amounts; samples were pooled as indicated in Table 2.
Fatty acid analysis of erythrocyte total phospholipids and phospholipid classes. Phospholipid bands isolated by thin-layer chromatography were scraped directly into glass vials containing 1% (v/v) H2SO4 in methanol and heated for 3 hr at 70~ The resulting methyl esters were extracted into heptane and dried over sodium sulfate prior to analysis by gas chromatography. Analyses of f a t t y acid methyl esters of the total phospholipids and the major phospholipid classes of marmoset erythrocytes were performed using capillary gas chromatography (HewlettPackard H P 5880 gas chromatograph; Hewlett-Packard, Palo Alt~ CA). The column was a 50-m glass column
TABLE 1 Fatty Acid Composition of Marmoset Colony Diet, 20:Sn-3(EPA) Concentrate and Experimental Diets a
Major fatty acid (%; w/w) Saturated Monounsaturated n-6 18:2 20:4 n-3 18:3 20:5 I(n-6) Z(n-3) :E(n-6)/Z(n-3) U.I.
Colony diet 46.0 39.5
Concentrate 2.1 16.2
REF 31.3 33.7
Experimental diets REF + EPA ATH 31.8 56.1 32.9 35.9
13.1 tr.
0.6 4.5
32.9 n.d.
31.0 tr.
6.9 n.d.
6.6 tr.
1.4 n.d. 13.1 1.4 9.4 70
0.5 74.8 5.1 75.3 0.07 411
0,9 n.d. 32.9 0.9 36.5 102
0.9 2.9 31.0 3.8 8.2 112
1.1 n.d. 6.9 1.1 6.3 53
1.0 3.3 6.6 4.3 1.5 65
Cholesterol(%) Total fat(%) Energy Value(KJ/g)
