- Email: [email protected]
Effect of dietary protein on serum insulin and glucagon levels in hyper- and normocholesterolemic men
55
Atherosclerosis, 76 (1989) 55-61 Elsevier Scientific Publishers Ireland, Ltd.
ATH 04271
Effect of dietary protein on serum insulin and glucagon levels in hyper- and normocholesterolemic men Richard Department
Hubbard,
Cindy L. Kosch, Albert Sanchez, Joan Sabate, Lee Berk and Gerald Shavlik
of Pathology, School of Medicine and Nutrition Program, School of Public Health, Loma Linda University, Loma Lindq CA (U.S.A.) (Received 14 June, 1988) (Revised, received 27 October, 1988) (Accepted 1 November, 1988)
This study was designed to test the effect of dietary protein on blood levels of insulin and glucagon. Twelve normocholesterolemic ( < 200 mg/dl) and 11 hypercholesterolemic ( > 240 mg/dl) healthy male subjects, 31-62 years of age, were randomly given 3 liquid test meals 1 week apart. Meals were identical except for the protein source (soybean, casein, or protein free). Blood was drawn at fasting, and 0.5 and 2 h postprandially. Insulin and glucagon levels were measured by radioimmunoassay. Hypercholesterolemic subjects had a higher (P < 0.05) insulin/glucagon ratio (1.5) than normocholesterolemic subjects (0.7) 2 h post-prandially when fed the casein test meal. There was no significant difference following the soybean test meal. This implies that the post-prandial insulin/glucagon ratio was affected by the amino acid composition of the diet. There was a consistently higher insulin response to all test meals among hyperversus normocholesterolemic subjects. These results are consistent with our hypothesis that the hypocholesterolemic effects of soybean protein and the hypercholesterolemic effects of casein were mediated by altered levels of insulin and glucagon.
Key words: Casein; Dietary protein; Cholesterol
Insulin/glucagon;
Introduction Dietary protein studies, particularly in the last decade, have shown significant differences, as a
Correspondence to: Dr. Richard Hubbard, Department of Pathology, School of Medicine, Loma Linda University, Loma Linda, CA 92350, U.S.A. OOZl-9150/89/$03.50
Human males; Test meals:
Soybean
protein:
of serum protein effect, on the modulation cholesterol levels. Vegetable proteins are hypocholesterolemic, while animal proteins are hypercholesterolemic as demonstrated in rabbits [l-9] and rats [lo-121. This also appears to be the case in human subjects [13-181 although a highly saturated fat or cholesterol diet may alter the vegetable protein effect [19,20]. However, work in
0 1989 Elsewier Scientific Publishers Ireland, Ltd.
56 this laboratory with rabbits shows that the dietary protein effect on serum cholesterol levels is separate from that of lipids [21]. The amino acid composition of a protein seems to dictate its effect on serum cholesterol levels. Generally, plant proteins have higher absolute amounts of arginine, glycine and alanine, whereas animal proteins have higher amounts of lysine [22]. The differences in lysine/arginine ratios between animal protein (casein) and vegetable protein (soy) have been proposed [23-271 as an explanation of the normo- and hypercholesterolemic effects of dietary protein. Soy bean protein has a relatively low lysine/arginine ratio of 0.8 and is hypocholesterolemic, compared to casein which is hypercholesterolemic and has a lysine/arginine ratio of 2.2 [21]. However, whole egg protein with a lysine/arginine ratio of 1.0 is hypercholesterolemit [25]. This indicates that amino acids other than lysine and arginine have an effect on cholesterolemia. Insulin and glucagon levels are key controllers of cholesterol and triglyceride biosynthesis and these hormones vary their secretion rates in response to amino acids [28-321. Arginine given orally or intravenously in large doses increases insulin and glucagon secretion [32]. Also, leucine and lysine in large doses increase insulin secretion, but to a lesser degree than arginine [30]. It is of interest that feeding usual amounts of amino acids as in soy protein, which is high in arginine and glycine results in relatively low serum insulin levels [ll] and high plasma glucagon levels [28]. Insulin increases and glucagon decreases the HMG-CoA reductase activity which is the rate-limiting step in
TABLE
1
COMPOSITION CALCULATED
OF MEALS IN PERCENT OF CALORIES FROM THEIR PROXIMATE ANALYSES
Component
Protein ’ Corn syrup solids b Soy bean oil
Test meals SOY
Casein
Protein-free
20 60 20
20 60 20
0 80 20
a Soy protein isolate or sodium caseinate. b Equal amounts of corn syrup solids for all meals with addition of modified corn starch for the protein-free meal.
cholesterol biosynthesis [33,34]. We believe the mechanism of action by dietary proteins on cholesterol levels is hormonally controlled through serum amino acid levels [35,36]. This would suggest that the amino acid and hormone levels of individuals with hypercholesterolemia are different from those with normocholesterolemia. This study was designed to determine if differences in hormone levels exist between normoand hypercholesterolemic subjects and if there is a difference in response to a meal containing casein, soy protein, or a protein free meal. Methods Subjects and test meals Normo- and hypercholesterolemic males, aged 31-62 years, were chosen from the health screening program of the Center for Health Promotion at Loma Linda University. The selection criteria included: fasting cholesterol levels above 240 mg/dl or below 200 mg/dl, absence of heart disease and diabetes, and without history of alcohol or drugs including caffeine at least 1 week prior to the study. The nature of the study and the requirements for the subjects were mailed to each person who met the criteria. Thirty subjects met the criteria, and signed a consent form required by the Institutional Review Board of Loma Linda University. The mean fasting cholesterol value of 3 separate analyses over 4 weeks for each subject was used to determine into which group each subject was placed. Seven subjects were deleted from the study, 4 having serum cholesterol between 200 and 240 mg/dl, and 3 with blood sugars indicative of possible diabetes. Data are reported on 23 subjects, 12 with cholesterol levels below 200 mg/dl and 11 with cholesterol levels above 240 mg/dl. The protein-containing liquid test meals were designed to contain 20% protein (approx. 40 g from casein or soy protein), 20% fat and 60% carbohydrate as percent of the 800 kcal of each meal. The caloric contributions were based on actual proximate analyses of the meals shown in Table 1. Samples Blood samples were collected on 4 consecutive Sunday mornings, 3 of which were required for
57 each subject. Height was initially recorded and weight was measured weekly for each participant. The subjects were asked to maintain their usual eating and exercise habits. Blood was drawn after an overnight fast, and at 0.5 and 2 h after the test meals. After the fasting sample was drawn, the subject was given one of the 3 test meals in random order. The subjects consumed the liquid meal within 15 min. Time for the 0.5- and 2-h blood drawings began at the completion of meal ingestion. Assays Znsulin. Blood was drawn from the anticubital vein into a single 5 ml BD Vacutainer Brand clot tube. Thirty minutes were allowed for the blood to clot before samples were centrifuged at 760 X g for 15 min to obtain serum. Serum was transferred to plastic aliquot containers and stored at - 70 ’ C until analysis. Analysis was done using the radioimmunoassay method with reagents from Immuno Nuclear Corporation, Catalog No. 0600. Two-hundred microliter of serum, in duplicate, were required for the assay. The sample, antibody and iodine ( 1251) radioactive tracer were incubated together for 3 h at 22 f 2°C. A precipitated second antibody complex was then added to separate bound from free antigen. Samples were incubated for 15 min, centrifuged, decanted, and counted in a gamma-scintillation counter. Coefficients of variation for within-day assay was 4.2% and for between-day assay 8.5%. Glucagon. One 7-ml BD vacutainer brand evacuated EDTA-containing glass tube was drawn on ice for each glucagon analysis. Trasylol@, aprotinin (a protease inhibitor), was added in the amount of 0.5 ml to 5 ml blood and gently mixed to prevent catabolism of the glucagon molecule. Samples were centrifuged at 760 X g for 15 min at 4°C and transferred to plastic aliquot containers and stored at - 70 o C until analysis. Analysis was done using the radioimmunoassay method with reagents from Serono Diagnostics, Catalog No 5910000-2-1284. Samples were thawed in an ice bath. Samples, antibody, and antiserum were incubated together for 3 h at 4°C. ‘*‘I was then added, and incubated for 20-24 h at 4OC. The samples were centrifuged for 15 min at 1500-3000 x g, decanted, and counted in a gamma-scintilla-
tion counter. The coefficient of variation for within-day assay was 6.4% and for between-day assay 5.0%. Cholesterol and triglycerides. Analysis was performed using the spectrophotometric method with cholesterol oxidase No. SG4-0065-J81 for cholesterol and the multiple enzyme No. Sg4-0085B85 method for triglycerides utilizing the SMAC (Sequential Multiple Analyzer Computer II) from Technicon Instrument Corp. Glucose. Analysis was performed with the Beckman Glucose Analyzer on oxalate/fluoride plasma using the glucose oxidase method [37]. Statistics. Data were analyzed using the Student’s t-test, paired t-test and one-way analysis of variance. Results Our normocholesterolemic (below 200 mg/dl cholesterol) and hypercholesterolemic (above 240 mg/dl cholesterol) subjects could not be classified into 2 homogeneous groups based on their fasting triglyceride values. Using 110 mg/dl triglyceride as the cut-off level of comparison, which we derived from the Prevalance Study [38], 10 normocholesterolemics were below and 2 above, while 3 hypercholesterolemics were below and 8 above this value. The mean fasting cholesterol and triglyceride values along with the mean fasting, 0.5 and 2 h glucose, insulin and glucagon values are shown in Table 2 for both the normo- and hypercholesterolemit subjects for all 3 test meals. The 0.5 and 2 h glucagon values for the normocholesterolemics after the soy test meal are elevated, but not significantly, from all other 0.5 and 2 h glucagon values in the study. Also, the 2 h insulin values for the hypercholesterolemics after the casein meal are elevated, but not significantly, from other similarly timed samples. The insulin and glucagon responses are presented in Table 2, at fasting, 0.5 and 2 h for both groups of subjects for all 3 test meals. From Table 2 it can be seen that insulin levels increased in all subject groups at 0.5 h after every meal. At 2 h after the meal insulin levels reached a plateau or decreased, except in the hypercholesterolemic subjects who had a higher, but not sig-
58 TABLE 2 MEAN rt SD VALUES FOR FASTING INSULIN AND GLUCAGON
Cholesterol ’ Triglyceride * Glucose a F 0.5 h 2h Insulin b F 0.5 h 2h Glucagon ’ F 0.5 h 2h
CHOLESTEROL
AND TRIGLYCERIDE
AND FASTING,
0.5 AND 2 h GLUCOSE,
Soy
Casein
Normo
Hyper
Normo
Hyper
Normo
Hyper
187f18 89k35
273zk35 179+79
181*20 101 f 74
269 + 22 188&-97
182+20 93+47
274 f 33 134*59
9s* 7 118+25 113f24
1OOk 6 12Ok27 106& 18
96k 4 131 f 31 104&18
99* 7 116k29 99+23
97+ 6 134* 35 103 + 33
99k 6 145 + 32 99* 9
10f 2 73+36 6Of47
12* 3 85+ 6 59+26
9* 4 91+38 65k38
12* 4 71*20 107 f 43
9* 2 57+29 57&42
13+ 3 77*37 75f47
113*46 120* 52 106*49
99+39 88+42 94k38
115f43 151 f 85 149k65
108k46 133+68 108+53
Protein-free
131*44 115*41 113*50
122 + 84 83f39 77+48
’ mg/dl. b pu/nll. c pg/mI
nificant, insulin level compared to normocholesterolemic subjects when fed the casein meal. Glucagon levels were decreased in normocholesterolemic subjects following the protein-free meal, while there was an increase at 0.5 h with the soy protein containing meal. In hypercholesterolemics, the decrease in glucagon levels following the protein-free meal was more pronounced than in normocholesterolemics. There was a similar decrease in glucagon levels following the casein meal. In contrast, the glucagon level increased at 0.5 h following the soy meal. Differences in glucagon responses due to test meals for either group of subjects were not statistically significant. For the normocholesterolemic subjects, fed a soy protein test meal, we show a plot of their insulin/glucagon ratios in Fig. 1. The ratio increased from 0.1 to 0.6 (P < 0.05) at 0.5 h. At 2 h the ratio decreased to 0.4. The hypercholesterolemic subjects fed the soy protein test meal had an increase at a ratio of 0.1-O-9 (P < 0.05). At 2 h the ratio decreased to 0.7. When subjects were fed the casein test meal the normocholesterolemic subjects had an increase in the ratio from 0.1 to 0.7 (P < 0.01). At 2 h the ratio remained at 0.7.
The hypercholesterolemic subjects fed the casein test meal experienced an increase in the ratio from 0.2 to 1.2 (P -Z 0.01) at 0.5 h. At 2 h the ratio continued to increase to 1.5 (P < 0.05). The ratio of insulin/glucagon in the normo- and the hypercholesterolemic subjects are significantly different at 2 h after the casein test meal (P < 0.05).
,’
/’
3’
,’ #’
0
‘h
2 hr
SOY
0
‘h
Casein
Meals
2 hr
0
‘4
Zilr
Protein-free
Fig. 1. Comparison of mean insuIin/glucagon ratios in normoand hypercholesterolemic subjects in response to soy protein, casein and protein-free test meals.
59 The hypercholesterolemics’ insulin/glucagon ratio is significantly lower (P -C0.05) at 2 h on the soy meal versus the casein meal. In addition, the hypercholesterolemic insulin/glucagon ratio after the casein meal is significantly higher than for the normocholesterolemics after the casein meal. There were no significant differences among diet groups in weight or physical activity or triglyceride values of the subjects throughout the study. Discussion Serum insulin levels increase during the first 0.5 h postprandially with the non-protein meal (Table 2) as expected [39]. The protein-containing meals had a similar effect on insulin levels as the nonprotein meal among normocholesterolemic subjects. In hypercholesterolemic subjects the initial response in insulin levels 0.5 h after the soy meal is reversed by the end of 2 h post-prandial, while after the casein test meal serum insulin levels remain elevated at 2 h. Thus, there is a differential effect on insulin levels by the 2 proteins. The glucagon levels decrease after the non-protein meal (Table 2), as expected during the absorption of glucose from the carbohydrate fraction of the meal [39]. This decrease is particularly evident in the hypercholesterolemic subjects following the non-protein and casein test meals. In contrast, glucagon levels have a tendency to increase with the feeding of the soy protein test meal. Even the hypercholesterolemic subjects have an elevation of glucagon 0.5 h after the soy meal as compared to decreases with the other meals. This particular effect of soy protein may be due to its relatively high arginine and glycine content over casein (2.3 times greater in soy protein for both amino acids) [21,22], since arginine and glycine stimulate the release of glucagon [31]. The greater amounts of arginine and glycine in soy protein over casein correlates with decreased insulin secretion and increased glucagon levels (Table 2, Fig. 1) which according to established knowledge decreases lipogenesis [40]. The point of major interest in these findings is the significant increase in the insulin/glucagon ratio (P < 0.05) of 1.5 for the hypercholesterolemic subjects, compared to 0.7 for the normocholesterolemic subjects at 2 h post-prandially after
the casein meal (Fig. 1). The finding of a high insulin/glucagon ratio following the casein meals appears to explain the hypercholesterolemic effects of casein [1,2,12,25], in view of the increase of cholesterol biosynthesis caused by insulin and its inhibition by glucagon [33,34]. The high insulin levels in this study among hypercholesterolemics after the casein meal is particularly interesting in light of the 10 year prospective study of Finnish policemen [41] which implicates an increased plasma insulin level as the dominant biochemical risk factor for atherosclerotic cardiovascular disease. The significant differences observed in serum insulin and glucagon levels following the casein meal suggest that a casein test meal may be useful in relating the insulin/glucagon ratio, in a kinetic assessment of a normo- versus hypercholesterolemit state. In addition, the casein test meal can provide an additional measure of the subsequent effect of dietary proteins on serum cholesterol levels. We have proposed that dietary protein affects serum cholesterol via altered hormone (insulin and glucagon) levels [35,36]. This is in confirmation of the work of Descovich et al. [15], where they showed elevated plasma arginine and glucagon levels in human subjects fed soy protein. The data presented in Table 2, and in Fig. 1, strongly indicate that the intermediary metabolic role of dietary protein on cholesterol metabolism is regulated by the secretion of insulin and glucagon caused by specific dietary amino acids. Acknowledgements We gratefully acknowledge the assistance of the Loma Linda Center for Health Promotion in the selection of subjects; the statistical advice of Dr. Paul Yahiku, the technical assistance of Karen Filler, Helen Mendiola; the computer consultation provided by Que Osler; and the participation of the subjects. This study was supported in part by grants from the Rex Callicot family, Loma Linda Board by Councillors, and Loma Linda Foods. References 1 Huff, M.W., Hamilton R.W.G. and Carroll, K.K., Plasma cholesterol levels in rabbits fed low fat, cholesterol-free,
60
2
3
4
5
6
I
8
9
10
semi-purified diets: effects of dietary proteins, protein hydrolysates and amino acid mixtures, Atherosclerosis, 28 (1977) 187. Terpstra, A.H.M., Woodward, C.J.H., West, C.E. and Van Boven, H.G., A longitudinal crossover study of serum cholesterol and lipoproteins in rabbits fed on semipurified diets containing either casein or soy-bean protein, Br. J. Nutr., 47 (1982) 213. Terpstra, A.H.M., Harkes, L. and Van der Veen, F.H., The effect of different proportion of casein in semiputified diets on the concentration of serum cholesterol and the lipoprotein composition of rabbits, Lipids, 16 (2) (1981) 114. Terpstra, A.H.M. and Sanchez-Muniz, F.J., Time course of hypercholesterolemia in rabbits fed semipurified diets containing casein or soybean protein, Atherosclerosis, 39 (1981) 217. Kritchevsky, D., Tepper, S.A., Czamecki, S.K., Klurfeld, D.M. and Story, J.A., Experimental atherosclerosis in rabbits fed cholesterol free diets. Part 9. Beef protein and textured vegetable protein, Atherosclerosis, 39 (1981) 169. Huff, M.W. and Carroll, K.K., Effects of dietary proteins and amino acid mixtures on plasma cholesterol levels in rabbits, J. Nutr., 110 (1980) 1676. Katan, M.B., Vroomen, L.H.M. and Hermus, R.J.J., Reduction of casein-induced hypercholesterolemia and atherosclerosis in rabbits and rats by dietary glycine, arginine and alanine, Atherosclerosis, 43 (1982) 381. Hubb, M.W., Roberts, D.C.K. and Carroll, K.K., Long-term effects of semipurified diets containing casein or soy protein isolate on atherosclerosis and plasma lipoproteins in rabbits, Atherosclerosis, 41 (1982) 327. Terpstra, A.H.M., Deuring, K. and West, C.E., The effect of cholesterol-enriched semipurified diets containing two levels of either casein or soyprotein on the concentration of serum cholesterol in mature rabbits, Nutr. Rep. Int., 26 (3) (1982) 389. Raheja, K.L. and Linscheer, W.G., Comparative effects of soy and casein protein on plasma cholesterol concentra-
tions, Ann. Nutr. Metab., 26 (1982) 44. 11 Sugano, M., Ishiwaki, N. and Nakashima, K., Dietary protein-dependent modification of serum cholesterol level in rats. Ann. Nutr. Metab., 28 (1984) 192. 12 Nagata, Y., Imaizumi, K. and Sugano, M., Effects of soyabean protein and casein on serum cholesterol levels in rats, Br. J. Nutr., 44 (1980) 113. 13 Van Raaij, J.M.A., Katan, M.S.M., Hautvast, J.G.A.J. and Hermus, R.J.J., Effects of casein versus soy protein diets on serum cholesterol and lipoproteins in young healthy volunteers, Am. J. Clin. Nutr., 34 (1981) 1261. 14 Sirtori, C.R., Gatti, E., Mantero, 0. et al., Clinical experience with the soybean protein diet in the treatment of hypercholesterolemia, Am. J. Clin. Nutr., 32 (1979) 1645. 15 Descovich, G.C., Gaddi, A., Mannino, G. et al., Multicentre study of soybean protein diet for outpatient hypercholesterolemic patients, Lancet, ii (1980) 709. 16 Sirtori, C.R., Agradi, E., Conti, F. et al., Soybean-protein diet in the treatment of type-II hyperlipoproteinemia, Lancet, i (1977) 275. 17 Carroll, K.K., Giovannetti, P.M., Huff, M.W., Moase, O.,
18
19
20
21
22
23 24
25 26 27
28
29
30
31
32
33
34
Roberts, D.C.K. and Wolfe, B.M., Hypocholesterolemic effect of substituting soybean protein for animal protein in the diet of healthy young women, Am. J. Clin. Nutr., 31 (1978) 1312. Goldberg, A.P., Lim, A., Kolar, J.B., Grundhauser, J.J., Steinke, F.H. and Schonfeld, G., Soybean protein independently lowers plasma cholesterol levels in primary hypercholesterolemia, Atherosclerosis, 43 (1982) 355. Wiebe, S.L., Bruce, V.M. and McDonald, B.E., A comparison of the effect of diets containing beef protein and plant proteins on blood lipids of healthy young men, Am. J. Clin. Nutr., 40 (1984) 982. Anderson, J.T., Grande, F. and Keys, A., Effect on man’s serum lipids of two proteins with different amino acid composition, Am. J. Clin. Nutr., 24 (1971) 1524. Sanchez, A., Rubano, D.A., Shavlik, G.W., Hubbard, R.W. and Horning, M.C., Cholesterolemic effects of the lysine/arginine ratio in rabbits after initial early growth, Arch. Latinoam. Nutr., (1988) in press. Orr, M.L. and Watt, B.K., Amino acid content of foods. Home economic research report No. 4 USDA. US Govt, Printing Office, Washington, DC, 1957. Kritchevsky, D., Vegetable protein and atherosclerosis, J. Am. Oil Chem. Sot., 56 (1979) 135. Kritchevsky, D., Tepper, S.A., Czamecki, S.K. and Klurfeld, D.M., Atherogenicity of animal and vegetable protein - influence of the lysine to arginine ratio, Atherosclerosis, 42 (1982) 429. Carroll, K.K., Hypercholesterolemia and atherosclerosis: effects of dietary protein, Fed. Proc., 41 (1982) 2792. Carroll, K.K., Soya protein and atherosclerosis, J. Am. Oil Chem. Sot., 58 (1981) 416. Gibney, J.J., The effect of dietary lysine to arginine ratio on cholesterol kinetics in rabbits, Atherosclerosis, 47 (1983) 263. Noseda, G., Fragiacomo, C. and Cairoli, R., Hormonal effects of a soybean protein diet versus a casein rich diet. In: Noseda, G., Fragiacomo, C., Fumagalli, R. and Paoletti, R. (Eds.), Lipoproteins and Coronary Atherosclerosis, Elsevier Biomedical Press, Amsterdam-New York, 1982, p. 299. Ohneda, A., Parada, E., Eisentraut, A.M. and Unger, R.H., Characterization of response of circulating glucagon to intraduodenal and intravenous administration of amino acids, J. Clin. Invest., 47 (1968) 2305. Fajans, S.S., Floyd, J.C., Knopf, R.F. and Conn, J.W., Effect of amino acids and protein on insulin secretion in man. Recent Progr. Harm. Res., 123 (1967) 617. Assau, R., Marre, M. and Gromley, M., The amino acid-induced secretion of glucagon. In: Lefebvre, P.J. (Ed.), Handbook of Experimental Pharmacology, 66/ Glucagon II. Springer-Verlag, New York, 1982, p. 19. Palmer, J.P., Walter, R.M. and Ensink, J.W., Argininestimulated acute phase of insulin and glucagon secretion in normal man, Diabetes, 24 (1975) 735. Shapiro, D.J. and Rodwell, V.W., Regulation of hepatic 3-hydroxy-3-methyl-glutaryl coenzyme A reductase and cholesterol synthesis, J. Biol. Chem., 246 (1971) 3210. Lakshmanan, M.R., Dugan, R.E., Nepokroeff, C.M., Ness,
61
35
36
37
38
G.C. and Porter, J.W., Regulation of rat liver 3-hydroxy-3methylglutaryl coenzyme A reductase activity and cholesterol levels of serum and liver on various dietary and hormonal states, Arch. B&hem. Biophys., 168 (1975) 89. Sanchez, A., Horning, MC. and Wingeleth, D.C., Plasma amino acids in humans fed plant proteins, Nutr. Rep. Int., 28 (1983) 497. Sanchez, A., Horning, M.C., Shavlik, G.W., Wingeleth, D.C. and Hubbard, R.W., Changes in levels of cholesterol associated with plasma amino acids in humans fed plant proteins, Nutr. Rep. Int., 32 (1985) 1047. Go&man, N. and Schmitz, J.M., Application of a new peroxide indicator reaction to the specific automated determination of glucose with glucose oxidase, Clin. Chem., 18 (1972) 943. The Lipid Research Clinics Population Studies Data Book,
Vol. I, The Prevalance Study, Bethesda, 1980, NIH publication No. 80-1257. 39 Matshinsky, F.M., Pagilara, A.A., Zawalich, W.S. and Trus, M.D., Metabolism of pancreatic islets and regulation of insulin and glucagon secretion. In: Degrott, L.J., Cahill, G.F., Odell, W.D. et al. @Is.), Endocrinology, Vol. 2, Grune and Stratton, New York, 1979, p. 935. 40 Murray, R.K., Granner, D.K., Mayes, P.A. and Rodwell, V.W. (Eds.), Harpers Biochemistry, 21st edn., Chapter 26, Lipid transport and storage, Appleton & Lange, Norwalk, CT, 1988, p. 237. 41 Pyorala, D., Savolainen, E., Kaukola, S. and Haapakoski, J., Plasma insulin as coronary heart disease risk factor in relationship to other risk factors and predictive value during 9 l/2 year follow up of the Helsinki Policemen study population, Acta Med. Stand., 701 (Suppl.) (1985) 38.
Atherosclerosis, 76 (1989) 55-61 Elsevier Scientific Publishers Ireland, Ltd.
ATH 04271
Effect of dietary protein on serum insulin and glucagon levels in hyper- and normocholesterolemic men Richard Department
Hubbard,
Cindy L. Kosch, Albert Sanchez, Joan Sabate, Lee Berk and Gerald Shavlik
of Pathology, School of Medicine and Nutrition Program, School of Public Health, Loma Linda University, Loma Lindq CA (U.S.A.) (Received 14 June, 1988) (Revised, received 27 October, 1988) (Accepted 1 November, 1988)
This study was designed to test the effect of dietary protein on blood levels of insulin and glucagon. Twelve normocholesterolemic ( < 200 mg/dl) and 11 hypercholesterolemic ( > 240 mg/dl) healthy male subjects, 31-62 years of age, were randomly given 3 liquid test meals 1 week apart. Meals were identical except for the protein source (soybean, casein, or protein free). Blood was drawn at fasting, and 0.5 and 2 h postprandially. Insulin and glucagon levels were measured by radioimmunoassay. Hypercholesterolemic subjects had a higher (P < 0.05) insulin/glucagon ratio (1.5) than normocholesterolemic subjects (0.7) 2 h post-prandially when fed the casein test meal. There was no significant difference following the soybean test meal. This implies that the post-prandial insulin/glucagon ratio was affected by the amino acid composition of the diet. There was a consistently higher insulin response to all test meals among hyperversus normocholesterolemic subjects. These results are consistent with our hypothesis that the hypocholesterolemic effects of soybean protein and the hypercholesterolemic effects of casein were mediated by altered levels of insulin and glucagon.
Key words: Casein; Dietary protein; Cholesterol
Insulin/glucagon;
Introduction Dietary protein studies, particularly in the last decade, have shown significant differences, as a
Correspondence to: Dr. Richard Hubbard, Department of Pathology, School of Medicine, Loma Linda University, Loma Linda, CA 92350, U.S.A. OOZl-9150/89/$03.50
Human males; Test meals:
Soybean
protein:
of serum protein effect, on the modulation cholesterol levels. Vegetable proteins are hypocholesterolemic, while animal proteins are hypercholesterolemic as demonstrated in rabbits [l-9] and rats [lo-121. This also appears to be the case in human subjects [13-181 although a highly saturated fat or cholesterol diet may alter the vegetable protein effect [19,20]. However, work in
0 1989 Elsewier Scientific Publishers Ireland, Ltd.
56 this laboratory with rabbits shows that the dietary protein effect on serum cholesterol levels is separate from that of lipids [21]. The amino acid composition of a protein seems to dictate its effect on serum cholesterol levels. Generally, plant proteins have higher absolute amounts of arginine, glycine and alanine, whereas animal proteins have higher amounts of lysine [22]. The differences in lysine/arginine ratios between animal protein (casein) and vegetable protein (soy) have been proposed [23-271 as an explanation of the normo- and hypercholesterolemic effects of dietary protein. Soy bean protein has a relatively low lysine/arginine ratio of 0.8 and is hypocholesterolemic, compared to casein which is hypercholesterolemic and has a lysine/arginine ratio of 2.2 [21]. However, whole egg protein with a lysine/arginine ratio of 1.0 is hypercholesterolemit [25]. This indicates that amino acids other than lysine and arginine have an effect on cholesterolemia. Insulin and glucagon levels are key controllers of cholesterol and triglyceride biosynthesis and these hormones vary their secretion rates in response to amino acids [28-321. Arginine given orally or intravenously in large doses increases insulin and glucagon secretion [32]. Also, leucine and lysine in large doses increase insulin secretion, but to a lesser degree than arginine [30]. It is of interest that feeding usual amounts of amino acids as in soy protein, which is high in arginine and glycine results in relatively low serum insulin levels [ll] and high plasma glucagon levels [28]. Insulin increases and glucagon decreases the HMG-CoA reductase activity which is the rate-limiting step in
TABLE
1
COMPOSITION CALCULATED
OF MEALS IN PERCENT OF CALORIES FROM THEIR PROXIMATE ANALYSES
Component
Protein ’ Corn syrup solids b Soy bean oil
Test meals SOY
Casein
Protein-free
20 60 20
20 60 20
0 80 20
a Soy protein isolate or sodium caseinate. b Equal amounts of corn syrup solids for all meals with addition of modified corn starch for the protein-free meal.
cholesterol biosynthesis [33,34]. We believe the mechanism of action by dietary proteins on cholesterol levels is hormonally controlled through serum amino acid levels [35,36]. This would suggest that the amino acid and hormone levels of individuals with hypercholesterolemia are different from those with normocholesterolemia. This study was designed to determine if differences in hormone levels exist between normoand hypercholesterolemic subjects and if there is a difference in response to a meal containing casein, soy protein, or a protein free meal. Methods Subjects and test meals Normo- and hypercholesterolemic males, aged 31-62 years, were chosen from the health screening program of the Center for Health Promotion at Loma Linda University. The selection criteria included: fasting cholesterol levels above 240 mg/dl or below 200 mg/dl, absence of heart disease and diabetes, and without history of alcohol or drugs including caffeine at least 1 week prior to the study. The nature of the study and the requirements for the subjects were mailed to each person who met the criteria. Thirty subjects met the criteria, and signed a consent form required by the Institutional Review Board of Loma Linda University. The mean fasting cholesterol value of 3 separate analyses over 4 weeks for each subject was used to determine into which group each subject was placed. Seven subjects were deleted from the study, 4 having serum cholesterol between 200 and 240 mg/dl, and 3 with blood sugars indicative of possible diabetes. Data are reported on 23 subjects, 12 with cholesterol levels below 200 mg/dl and 11 with cholesterol levels above 240 mg/dl. The protein-containing liquid test meals were designed to contain 20% protein (approx. 40 g from casein or soy protein), 20% fat and 60% carbohydrate as percent of the 800 kcal of each meal. The caloric contributions were based on actual proximate analyses of the meals shown in Table 1. Samples Blood samples were collected on 4 consecutive Sunday mornings, 3 of which were required for
57 each subject. Height was initially recorded and weight was measured weekly for each participant. The subjects were asked to maintain their usual eating and exercise habits. Blood was drawn after an overnight fast, and at 0.5 and 2 h after the test meals. After the fasting sample was drawn, the subject was given one of the 3 test meals in random order. The subjects consumed the liquid meal within 15 min. Time for the 0.5- and 2-h blood drawings began at the completion of meal ingestion. Assays Znsulin. Blood was drawn from the anticubital vein into a single 5 ml BD Vacutainer Brand clot tube. Thirty minutes were allowed for the blood to clot before samples were centrifuged at 760 X g for 15 min to obtain serum. Serum was transferred to plastic aliquot containers and stored at - 70 ’ C until analysis. Analysis was done using the radioimmunoassay method with reagents from Immuno Nuclear Corporation, Catalog No. 0600. Two-hundred microliter of serum, in duplicate, were required for the assay. The sample, antibody and iodine ( 1251) radioactive tracer were incubated together for 3 h at 22 f 2°C. A precipitated second antibody complex was then added to separate bound from free antigen. Samples were incubated for 15 min, centrifuged, decanted, and counted in a gamma-scintillation counter. Coefficients of variation for within-day assay was 4.2% and for between-day assay 8.5%. Glucagon. One 7-ml BD vacutainer brand evacuated EDTA-containing glass tube was drawn on ice for each glucagon analysis. Trasylol@, aprotinin (a protease inhibitor), was added in the amount of 0.5 ml to 5 ml blood and gently mixed to prevent catabolism of the glucagon molecule. Samples were centrifuged at 760 X g for 15 min at 4°C and transferred to plastic aliquot containers and stored at - 70 o C until analysis. Analysis was done using the radioimmunoassay method with reagents from Serono Diagnostics, Catalog No 5910000-2-1284. Samples were thawed in an ice bath. Samples, antibody, and antiserum were incubated together for 3 h at 4°C. ‘*‘I was then added, and incubated for 20-24 h at 4OC. The samples were centrifuged for 15 min at 1500-3000 x g, decanted, and counted in a gamma-scintilla-
tion counter. The coefficient of variation for within-day assay was 6.4% and for between-day assay 5.0%. Cholesterol and triglycerides. Analysis was performed using the spectrophotometric method with cholesterol oxidase No. SG4-0065-J81 for cholesterol and the multiple enzyme No. Sg4-0085B85 method for triglycerides utilizing the SMAC (Sequential Multiple Analyzer Computer II) from Technicon Instrument Corp. Glucose. Analysis was performed with the Beckman Glucose Analyzer on oxalate/fluoride plasma using the glucose oxidase method [37]. Statistics. Data were analyzed using the Student’s t-test, paired t-test and one-way analysis of variance. Results Our normocholesterolemic (below 200 mg/dl cholesterol) and hypercholesterolemic (above 240 mg/dl cholesterol) subjects could not be classified into 2 homogeneous groups based on their fasting triglyceride values. Using 110 mg/dl triglyceride as the cut-off level of comparison, which we derived from the Prevalance Study [38], 10 normocholesterolemics were below and 2 above, while 3 hypercholesterolemics were below and 8 above this value. The mean fasting cholesterol and triglyceride values along with the mean fasting, 0.5 and 2 h glucose, insulin and glucagon values are shown in Table 2 for both the normo- and hypercholesterolemit subjects for all 3 test meals. The 0.5 and 2 h glucagon values for the normocholesterolemics after the soy test meal are elevated, but not significantly, from all other 0.5 and 2 h glucagon values in the study. Also, the 2 h insulin values for the hypercholesterolemics after the casein meal are elevated, but not significantly, from other similarly timed samples. The insulin and glucagon responses are presented in Table 2, at fasting, 0.5 and 2 h for both groups of subjects for all 3 test meals. From Table 2 it can be seen that insulin levels increased in all subject groups at 0.5 h after every meal. At 2 h after the meal insulin levels reached a plateau or decreased, except in the hypercholesterolemic subjects who had a higher, but not sig-
58 TABLE 2 MEAN rt SD VALUES FOR FASTING INSULIN AND GLUCAGON
Cholesterol ’ Triglyceride * Glucose a F 0.5 h 2h Insulin b F 0.5 h 2h Glucagon ’ F 0.5 h 2h
CHOLESTEROL
AND TRIGLYCERIDE
AND FASTING,
0.5 AND 2 h GLUCOSE,
Soy
Casein
Normo
Hyper
Normo
Hyper
Normo
Hyper
187f18 89k35
273zk35 179+79
181*20 101 f 74
269 + 22 188&-97
182+20 93+47
274 f 33 134*59
9s* 7 118+25 113f24
1OOk 6 12Ok27 106& 18
96k 4 131 f 31 104&18
99* 7 116k29 99+23
97+ 6 134* 35 103 + 33
99k 6 145 + 32 99* 9
10f 2 73+36 6Of47
12* 3 85+ 6 59+26
9* 4 91+38 65k38
12* 4 71*20 107 f 43
9* 2 57+29 57&42
13+ 3 77*37 75f47
113*46 120* 52 106*49
99+39 88+42 94k38
115f43 151 f 85 149k65
108k46 133+68 108+53
Protein-free
131*44 115*41 113*50
122 + 84 83f39 77+48
’ mg/dl. b pu/nll. c pg/mI
nificant, insulin level compared to normocholesterolemic subjects when fed the casein meal. Glucagon levels were decreased in normocholesterolemic subjects following the protein-free meal, while there was an increase at 0.5 h with the soy protein containing meal. In hypercholesterolemics, the decrease in glucagon levels following the protein-free meal was more pronounced than in normocholesterolemics. There was a similar decrease in glucagon levels following the casein meal. In contrast, the glucagon level increased at 0.5 h following the soy meal. Differences in glucagon responses due to test meals for either group of subjects were not statistically significant. For the normocholesterolemic subjects, fed a soy protein test meal, we show a plot of their insulin/glucagon ratios in Fig. 1. The ratio increased from 0.1 to 0.6 (P < 0.05) at 0.5 h. At 2 h the ratio decreased to 0.4. The hypercholesterolemic subjects fed the soy protein test meal had an increase at a ratio of 0.1-O-9 (P < 0.05). At 2 h the ratio decreased to 0.7. When subjects were fed the casein test meal the normocholesterolemic subjects had an increase in the ratio from 0.1 to 0.7 (P < 0.01). At 2 h the ratio remained at 0.7.
The hypercholesterolemic subjects fed the casein test meal experienced an increase in the ratio from 0.2 to 1.2 (P -Z 0.01) at 0.5 h. At 2 h the ratio continued to increase to 1.5 (P < 0.05). The ratio of insulin/glucagon in the normo- and the hypercholesterolemic subjects are significantly different at 2 h after the casein test meal (P < 0.05).
,’
/’
3’
,’ #’
0
‘h
2 hr
SOY
0
‘h
Casein
Meals
2 hr
0
‘4
Zilr
Protein-free
Fig. 1. Comparison of mean insuIin/glucagon ratios in normoand hypercholesterolemic subjects in response to soy protein, casein and protein-free test meals.
59 The hypercholesterolemics’ insulin/glucagon ratio is significantly lower (P -C0.05) at 2 h on the soy meal versus the casein meal. In addition, the hypercholesterolemic insulin/glucagon ratio after the casein meal is significantly higher than for the normocholesterolemics after the casein meal. There were no significant differences among diet groups in weight or physical activity or triglyceride values of the subjects throughout the study. Discussion Serum insulin levels increase during the first 0.5 h postprandially with the non-protein meal (Table 2) as expected [39]. The protein-containing meals had a similar effect on insulin levels as the nonprotein meal among normocholesterolemic subjects. In hypercholesterolemic subjects the initial response in insulin levels 0.5 h after the soy meal is reversed by the end of 2 h post-prandial, while after the casein test meal serum insulin levels remain elevated at 2 h. Thus, there is a differential effect on insulin levels by the 2 proteins. The glucagon levels decrease after the non-protein meal (Table 2), as expected during the absorption of glucose from the carbohydrate fraction of the meal [39]. This decrease is particularly evident in the hypercholesterolemic subjects following the non-protein and casein test meals. In contrast, glucagon levels have a tendency to increase with the feeding of the soy protein test meal. Even the hypercholesterolemic subjects have an elevation of glucagon 0.5 h after the soy meal as compared to decreases with the other meals. This particular effect of soy protein may be due to its relatively high arginine and glycine content over casein (2.3 times greater in soy protein for both amino acids) [21,22], since arginine and glycine stimulate the release of glucagon [31]. The greater amounts of arginine and glycine in soy protein over casein correlates with decreased insulin secretion and increased glucagon levels (Table 2, Fig. 1) which according to established knowledge decreases lipogenesis [40]. The point of major interest in these findings is the significant increase in the insulin/glucagon ratio (P < 0.05) of 1.5 for the hypercholesterolemic subjects, compared to 0.7 for the normocholesterolemic subjects at 2 h post-prandially after
the casein meal (Fig. 1). The finding of a high insulin/glucagon ratio following the casein meals appears to explain the hypercholesterolemic effects of casein [1,2,12,25], in view of the increase of cholesterol biosynthesis caused by insulin and its inhibition by glucagon [33,34]. The high insulin levels in this study among hypercholesterolemics after the casein meal is particularly interesting in light of the 10 year prospective study of Finnish policemen [41] which implicates an increased plasma insulin level as the dominant biochemical risk factor for atherosclerotic cardiovascular disease. The significant differences observed in serum insulin and glucagon levels following the casein meal suggest that a casein test meal may be useful in relating the insulin/glucagon ratio, in a kinetic assessment of a normo- versus hypercholesterolemit state. In addition, the casein test meal can provide an additional measure of the subsequent effect of dietary proteins on serum cholesterol levels. We have proposed that dietary protein affects serum cholesterol via altered hormone (insulin and glucagon) levels [35,36]. This is in confirmation of the work of Descovich et al. [15], where they showed elevated plasma arginine and glucagon levels in human subjects fed soy protein. The data presented in Table 2, and in Fig. 1, strongly indicate that the intermediary metabolic role of dietary protein on cholesterol metabolism is regulated by the secretion of insulin and glucagon caused by specific dietary amino acids. Acknowledgements We gratefully acknowledge the assistance of the Loma Linda Center for Health Promotion in the selection of subjects; the statistical advice of Dr. Paul Yahiku, the technical assistance of Karen Filler, Helen Mendiola; the computer consultation provided by Que Osler; and the participation of the subjects. This study was supported in part by grants from the Rex Callicot family, Loma Linda Board by Councillors, and Loma Linda Foods. References 1 Huff, M.W., Hamilton R.W.G. and Carroll, K.K., Plasma cholesterol levels in rabbits fed low fat, cholesterol-free,
60
2
3
4
5
6
I
8
9
10
semi-purified diets: effects of dietary proteins, protein hydrolysates and amino acid mixtures, Atherosclerosis, 28 (1977) 187. Terpstra, A.H.M., Woodward, C.J.H., West, C.E. and Van Boven, H.G., A longitudinal crossover study of serum cholesterol and lipoproteins in rabbits fed on semipurified diets containing either casein or soy-bean protein, Br. J. Nutr., 47 (1982) 213. Terpstra, A.H.M., Harkes, L. and Van der Veen, F.H., The effect of different proportion of casein in semiputified diets on the concentration of serum cholesterol and the lipoprotein composition of rabbits, Lipids, 16 (2) (1981) 114. Terpstra, A.H.M. and Sanchez-Muniz, F.J., Time course of hypercholesterolemia in rabbits fed semipurified diets containing casein or soybean protein, Atherosclerosis, 39 (1981) 217. Kritchevsky, D., Tepper, S.A., Czamecki, S.K., Klurfeld, D.M. and Story, J.A., Experimental atherosclerosis in rabbits fed cholesterol free diets. Part 9. Beef protein and textured vegetable protein, Atherosclerosis, 39 (1981) 169. Huff, M.W. and Carroll, K.K., Effects of dietary proteins and amino acid mixtures on plasma cholesterol levels in rabbits, J. Nutr., 110 (1980) 1676. Katan, M.B., Vroomen, L.H.M. and Hermus, R.J.J., Reduction of casein-induced hypercholesterolemia and atherosclerosis in rabbits and rats by dietary glycine, arginine and alanine, Atherosclerosis, 43 (1982) 381. Hubb, M.W., Roberts, D.C.K. and Carroll, K.K., Long-term effects of semipurified diets containing casein or soy protein isolate on atherosclerosis and plasma lipoproteins in rabbits, Atherosclerosis, 41 (1982) 327. Terpstra, A.H.M., Deuring, K. and West, C.E., The effect of cholesterol-enriched semipurified diets containing two levels of either casein or soyprotein on the concentration of serum cholesterol in mature rabbits, Nutr. Rep. Int., 26 (3) (1982) 389. Raheja, K.L. and Linscheer, W.G., Comparative effects of soy and casein protein on plasma cholesterol concentra-
tions, Ann. Nutr. Metab., 26 (1982) 44. 11 Sugano, M., Ishiwaki, N. and Nakashima, K., Dietary protein-dependent modification of serum cholesterol level in rats. Ann. Nutr. Metab., 28 (1984) 192. 12 Nagata, Y., Imaizumi, K. and Sugano, M., Effects of soyabean protein and casein on serum cholesterol levels in rats, Br. J. Nutr., 44 (1980) 113. 13 Van Raaij, J.M.A., Katan, M.S.M., Hautvast, J.G.A.J. and Hermus, R.J.J., Effects of casein versus soy protein diets on serum cholesterol and lipoproteins in young healthy volunteers, Am. J. Clin. Nutr., 34 (1981) 1261. 14 Sirtori, C.R., Gatti, E., Mantero, 0. et al., Clinical experience with the soybean protein diet in the treatment of hypercholesterolemia, Am. J. Clin. Nutr., 32 (1979) 1645. 15 Descovich, G.C., Gaddi, A., Mannino, G. et al., Multicentre study of soybean protein diet for outpatient hypercholesterolemic patients, Lancet, ii (1980) 709. 16 Sirtori, C.R., Agradi, E., Conti, F. et al., Soybean-protein diet in the treatment of type-II hyperlipoproteinemia, Lancet, i (1977) 275. 17 Carroll, K.K., Giovannetti, P.M., Huff, M.W., Moase, O.,
18
19
20
21
22
23 24
25 26 27
28
29
30
31
32
33
34
Roberts, D.C.K. and Wolfe, B.M., Hypocholesterolemic effect of substituting soybean protein for animal protein in the diet of healthy young women, Am. J. Clin. Nutr., 31 (1978) 1312. Goldberg, A.P., Lim, A., Kolar, J.B., Grundhauser, J.J., Steinke, F.H. and Schonfeld, G., Soybean protein independently lowers plasma cholesterol levels in primary hypercholesterolemia, Atherosclerosis, 43 (1982) 355. Wiebe, S.L., Bruce, V.M. and McDonald, B.E., A comparison of the effect of diets containing beef protein and plant proteins on blood lipids of healthy young men, Am. J. Clin. Nutr., 40 (1984) 982. Anderson, J.T., Grande, F. and Keys, A., Effect on man’s serum lipids of two proteins with different amino acid composition, Am. J. Clin. Nutr., 24 (1971) 1524. Sanchez, A., Rubano, D.A., Shavlik, G.W., Hubbard, R.W. and Horning, M.C., Cholesterolemic effects of the lysine/arginine ratio in rabbits after initial early growth, Arch. Latinoam. Nutr., (1988) in press. Orr, M.L. and Watt, B.K., Amino acid content of foods. Home economic research report No. 4 USDA. US Govt, Printing Office, Washington, DC, 1957. Kritchevsky, D., Vegetable protein and atherosclerosis, J. Am. Oil Chem. Sot., 56 (1979) 135. Kritchevsky, D., Tepper, S.A., Czamecki, S.K. and Klurfeld, D.M., Atherogenicity of animal and vegetable protein - influence of the lysine to arginine ratio, Atherosclerosis, 42 (1982) 429. Carroll, K.K., Hypercholesterolemia and atherosclerosis: effects of dietary protein, Fed. Proc., 41 (1982) 2792. Carroll, K.K., Soya protein and atherosclerosis, J. Am. Oil Chem. Sot., 58 (1981) 416. Gibney, J.J., The effect of dietary lysine to arginine ratio on cholesterol kinetics in rabbits, Atherosclerosis, 47 (1983) 263. Noseda, G., Fragiacomo, C. and Cairoli, R., Hormonal effects of a soybean protein diet versus a casein rich diet. In: Noseda, G., Fragiacomo, C., Fumagalli, R. and Paoletti, R. (Eds.), Lipoproteins and Coronary Atherosclerosis, Elsevier Biomedical Press, Amsterdam-New York, 1982, p. 299. Ohneda, A., Parada, E., Eisentraut, A.M. and Unger, R.H., Characterization of response of circulating glucagon to intraduodenal and intravenous administration of amino acids, J. Clin. Invest., 47 (1968) 2305. Fajans, S.S., Floyd, J.C., Knopf, R.F. and Conn, J.W., Effect of amino acids and protein on insulin secretion in man. Recent Progr. Harm. Res., 123 (1967) 617. Assau, R., Marre, M. and Gromley, M., The amino acid-induced secretion of glucagon. In: Lefebvre, P.J. (Ed.), Handbook of Experimental Pharmacology, 66/ Glucagon II. Springer-Verlag, New York, 1982, p. 19. Palmer, J.P., Walter, R.M. and Ensink, J.W., Argininestimulated acute phase of insulin and glucagon secretion in normal man, Diabetes, 24 (1975) 735. Shapiro, D.J. and Rodwell, V.W., Regulation of hepatic 3-hydroxy-3-methyl-glutaryl coenzyme A reductase and cholesterol synthesis, J. Biol. Chem., 246 (1971) 3210. Lakshmanan, M.R., Dugan, R.E., Nepokroeff, C.M., Ness,
61
35
36
37
38
G.C. and Porter, J.W., Regulation of rat liver 3-hydroxy-3methylglutaryl coenzyme A reductase activity and cholesterol levels of serum and liver on various dietary and hormonal states, Arch. B&hem. Biophys., 168 (1975) 89. Sanchez, A., Horning, MC. and Wingeleth, D.C., Plasma amino acids in humans fed plant proteins, Nutr. Rep. Int., 28 (1983) 497. Sanchez, A., Horning, M.C., Shavlik, G.W., Wingeleth, D.C. and Hubbard, R.W., Changes in levels of cholesterol associated with plasma amino acids in humans fed plant proteins, Nutr. Rep. Int., 32 (1985) 1047. Go&man, N. and Schmitz, J.M., Application of a new peroxide indicator reaction to the specific automated determination of glucose with glucose oxidase, Clin. Chem., 18 (1972) 943. The Lipid Research Clinics Population Studies Data Book,
Vol. I, The Prevalance Study, Bethesda, 1980, NIH publication No. 80-1257. 39 Matshinsky, F.M., Pagilara, A.A., Zawalich, W.S. and Trus, M.D., Metabolism of pancreatic islets and regulation of insulin and glucagon secretion. In: Degrott, L.J., Cahill, G.F., Odell, W.D. et al. @Is.), Endocrinology, Vol. 2, Grune and Stratton, New York, 1979, p. 935. 40 Murray, R.K., Granner, D.K., Mayes, P.A. and Rodwell, V.W. (Eds.), Harpers Biochemistry, 21st edn., Chapter 26, Lipid transport and storage, Appleton & Lange, Norwalk, CT, 1988, p. 237. 41 Pyorala, D., Savolainen, E., Kaukola, S. and Haapakoski, J., Plasma insulin as coronary heart disease risk factor in relationship to other risk factors and predictive value during 9 l/2 year follow up of the Helsinki Policemen study population, Acta Med. Stand., 701 (Suppl.) (1985) 38.