Dietary carbohydrate type and cholesterol-induced hypercholesterolemia in cynomolgus monkeys: Influence of oral antibiotic

BIOCHEMICAL

MEDICINE

AND

METABOLIC

BIOLOGY

37, 295-306 (1987)

Dietary Carbohydrate Type and Cholesterol-Induced Hypercholesterolemia in Cynomolgus Monkeys: Influence of Oral Antibiotic’ SATHANUR R. SRINIVASAN,* BHANDARU RADHAKRISHNAMURTHY,* GEORGE H. BORNSIDE, SALLY SPAHN, GLENN D. WILLIAMSON, AND GERALD S. BERENSON’ Departments

of Biochemistry,

*Medicine, and Surgery, New Orleans, Louisiana

Louisiana State 70112-2822

University

Medical

Center,

Received November 13, 1985

Saturated fat and cholesterol are considered important dietary factors that influence serum lipids and lipoproteins (1,2). Dietary carbohydrate given at high levels is known to alter the concentration, composition, and metabolism of serum lipoproteins (3-7). Earlier studies in humans suggest that the influence of the type of carbohydrates on serum lipids depends on amounts of saturated fat and cholesterol consumed (8). However, the specific effects of types of dietary carbohydrates on serum lipids and lipoproteins remain uncertain (9-12). This may be partly due to the confounding effects of other variables such as age, sex, and hyperlipidemic state of the individuals on serum lipid response (13,14). The use of nonhuman primates as experimental models, because of their phylogenetic proximity to humans and similarity in lipoprotein profiles (15,16), can obviate some of the difficulties encountered in humans under a nonclinical environment. Previous studies in nonhuman primates have demonstrated a considerable intraspecies and interspecies variation in serum lipid responses to dietary carbohydrates (17-19). Recently, we found that when starch was provided to cynomolgus monkeys as a single source of carbohydrate, it produced a consistently higher serum cholesterol response to dietary cholesterol (1 mg/kcal) than did sucrose regardless of fat content of the diet (20-21). The underlying reason(s) for this enhanced response to starch is not clear. In addition to altering lipoprotein metabolism at the hepatic and extrahepatic tissue levels, dietary carbohydrates can influence sterol absorption and/or excretion by altering the intestinal microbial flora (14,22). Induction of germfree characteristics by feeding nonabsorbable antibiotics provides an approach to testing the role of intestinal flora in this ’ This research is supported by funds from the National Heart, Lung, and Blood Institute of the U.S. Public Health Service (HL02942) and the National Research and Demonstration CenterArteriosclerosis (NRDC-A) HL 15103. ’ To whom correspondence and requests for reprints should be addressed. 295 08854505187 $3.00 Cepyright Q 1987 by Academic Press, Inc. All r(4#s of reproduction in any form reserved.

296

SRINIVASAN

ET AL.

regard. Since the common practice of using high cholesterol diets to examine hypercholesterolemic response may not reflect normal metabolic control mechanisms, there is a need to appraise the carbohydrate effect at a moderate cholesterol level. The present studies, conducted in the same monkeys used previously, examine (1) the effect of starch vs sucrose on serum lipoprotein response to a cholesterol level (0.4 mg/kcal) representing the upper limit of the average American diet (23) and (2) the potential role of intestinal flora in the above process by administering an oral nonabsorbable antibiotic (neomycin). MATERIALS

AND METHODS

Animals and Diet Twelve male cynomoigus monkeys (Macaca fascicufuris) with mean + SD body wt of 6.83 ? 1.64 kg were used. Prior to the experiment, the animals received a closed formula stock diet (Purina Monkey Chow) derived from natural feed ingredients in which 15% of the total calories was derived from protein, 11% from fat, and the remaining from complex carbohydrates. Semipurified experimental diets provided two types of carbohydrates (corn starch or sucrose) with 0.4 mg/kcal cholesterol added (Table 1). The caloric distribution was 49% as carbohydrate, 40% as fat (P/S ratio = 0.03), and 11% as protein. Neomycin (equivalent to 8 g/75 kg body wt) was added to the diet of each individual monkey. The approximate number of kilocalories consumed by this species was determined from daily food intake of the stock diet over a period of 1 week. Accordingly, each animal was provided once a day with equivalent amounts of the experimental diet. This procedure resulted in minimum wastage. TABLE 1 Composition of Diets Ingredients” Carbohydrate Corn starchb Sucrose Fat Butter : coconut oil (1: 1) Protein Caseinb Other Vitamin mixb Hegsted Sal& Celluloseb Banana flavoring

% Dry wt

% kcal

51.3

49.0

18.6

40.0

11.6

11.0

2.3 4.6 11.6 1 drop

L?In addition to cholesterol derived from butter (2.5 mg/g), both diets contained 0.4 mg/kcal cholesterol. b ICN Nutritional Biochemicals, Cleveland, OH (product numbers: casein, 901293; vitamin mix, 904654; alphacel, 900453; Hegsted salts, 902840). ’ P/S ratio 0.03; linoleic acid: 372 mg.

CARBOHYDRATES

Experimental

AND

SERUM

CHOLESTEROL

RESPONSE

297

Design

The experiment was conducted in two parts. The first used a crossover design. Twelve animals were divided into two equal groups and fed 6 weeks of each experimental diet in the following sequence for each group: starch, monkey chow, and sucrose; or sucrose, monkey chow, and starch. Two months after completion of the first part of the study 1 animal died of unknown cause. In the second, 10 animals were ranked in increasing order of serum cholesterol response to the previous (first part of the experiment) starch diet and assigned to blocks of 2 animals each. Monkeys within each block were then randomly assigned to two groups (5 animals/group); an examination of serum cholesterol responses to earlier starch or sucrose diet showed no difference between the two groups. The two groups were fed starch or sucrose diet without cholesterol (2 weeks), with cholesterol and neomycin (4 weeks), and with cholesteroi (4 weeks), in that order. Blood specimens (16 hr fasting) were drawn weekly from the femoral vein of sedated monkeys (Ketaset, Bristol Laboratories, Syracuse, NY), and serum was prepared for laboratory analyses. During the second part of the study fresh fecal specimens were collected every 2 weeks on passage and immediately delivered to the bacteriology laboratory for culture. Lipid Analyses

Cholesterol and triglycerides in whole serum and lipoprotein fractions were determined according to the Lipid Research Clinics Program protocol (24), using an AutoAnalyzer II (Technicon Instrument Corp., Tarrytown, NY). (The laboratory has been designated as standardized by the Centers for Disease Control, Atlanta, GA.) Phospholipids were determined by the method of Fiske and Subbarow (25). Separation of Lipoproteins

After 6 weeks of each experimental diet (part one) serum lipoproteins were separated sequentially according to the method of Hatch and Lees (26) using a type 40.3 rotor and Beckman L2-65B ultracentrifuge (Beckman Instruments Inc., Palo Alto, CA). Centrifugation was performed at 114,OOOg with a chamber temperature of 17°C to separate very low (VLDL, d < 1.006 g/ml), intermediate (IDL, d 1.006-1.019 g/ml), low (LDL, d 1.019-1.063 g/ml), and high density lipoproteins (“HDL”*, d 1.063-l. 12 g/ml and “HDL”, , d 1.12 g/ml infranatant). Polycarbonate tubes without caps were used to accommodate a total volume of 4.0 ml. Serum (3.0 ml) was overlayed with 1.006 g/ml density solution (1.0 ml). After each centrifuge run 1.0 ml of floating lipoprotein fraction was removed by pipetting and an equal volume of approprite density solution (NaBr/NaCl) was added to adjust the density to 1.019 g/ml for IDL, 1.063 g/ml for LDL, and 1.12 g/ml for HDL3. A running time of 16 hr was used for VLDL, 20 hr for IDL and LDL, and 40 hr for HDL*. The infranatent (3.0 ml) from the last run contained HDL, . The analytical recovery of serum total cholesterol in the lipoprotein fractions isolated by ultracentrifugation was greater than 96%. We measured lipids in lipoprotein-enriched ultracentrifugal fractions in duplicates and used the average values, allowing 5% CV.

298

SRINIVASAN ET AL.

Apoprotein Analyses The levels of apoB and apoA-I in whole serum and apoB levels in each of the lipoprotein fractions were determined by the electroimmunoassay procedure of Laurel1 (27). Goat antibodies to cynomolgus monkey LDL (d 1.03-1.05 g/ml) and apoA-I [prepared by the procedure of Miller et al. (28)] were used. For apoA-I assay, samples and standards were treated with 8 M urea and 1% NonidetP40 detergent (final concentration) before electrophoresis (29). No such pretreatment was carried out for apoB assay. Samples were appropriately diluted so that the apoprotein levels remained within the range of measurement. Other details of the assay including intra- and interassay variability are reported elsewhere (21). Bacteriology Serial IOO-fold dilutions of a 0.1-g sample of each fecal specimen were plated aerobically employing blood agar medium and MacConkey agar. These plates were examined after 24- and 48-hr incubations at 37°C. For the anaerobic culture, blood agar plates and Rogosa SL agar were incubated (in Brewer jars) in an atmosphere of 95% H2 and 5% CO* for 4 days at 37°C before being examined and subcultured. In addition, each specimen was cultured in thioglycollate medium, brain-heart infusion broth, cooked meat medium, and on Sabouraud agar and cornmeal agar. These allowed both aerobic and anaerobic bacteria, fungi, and yeast to grow and were studied as a check on the microorganisms isolated from surface cultures. Aerobic isolates were streaked to eosin-methylene blue agar to identify coliforms and were subcultured to heart infusion broth, from which gram-stained smears were prepared. Standard methods were used to identify gram-negative enteric bacteria. Isolates from anaerobic plates were streaked to a blood agar plate for aerobic incubation to verify that the isolate was an anaerobe. Statistical

Analysis

In the first part of the experiment a two-period repeated measures crossover design was used to determine significant (analysis of variance, P < 0.05) carryover (residual) effect and differences between diets (30). For ultracentrifugal data a classic two-period crossover design was used to determine carry-over effect and differences between diets (31). In the second phase of the study a repeated measures design was used to determine differences between treatments (diets and antibiotic) (32). Main and simple effects were tested. If significant interactions occurred, simple effects rather than main effects were presented. Mean differences in bacterial counts between treatments were examined by Student’s t test. Measured variables (except body weight) were transformed appropriately for analysis. Statistical significance was at P < 0.05. RESULTS

Effect of Carbohydrate Type Although the monkeys gained weight significantly during each 6-week experimental diet period (starch 7.2% vs sucrose 9.9%), the difference between diets was not significant. Since animals were fed 6 weeks of basal monkey chow following each experimental diet, none of the measured variables during the

CARBOHYDRATES

AND SERUM CHOLESTEROL

RESPONSE

299

crossover dietary regimen showed a significant carry-over effect of a previous regimen. For example, feeding of monkey chow diet following experimental diets resulted in return of the serum total cholesterol levels (130 t 6 mg/dl) to baseline values (136 & 6 mg/dl). Time-course changes in these variables during periods of starch and sucrose diets are shown in Fig. 1. Each diet induced a marked increase in serum total cholesterol with time. However, the degree of response at each time interval was relatively higher with starch (420 + 48 vs 271 + 21 mg/dl after 6 weeks). Changes in serum apoB levels essentially paralleled serum total cholesterol with levels reaching 221 ? 30 and 134 ? 19 mg/dl after 6 weeks of starch and sucrose diets, respectively. On the other hand, changes in serum apoA-I levels were relatively small; but the starch diet produced consistently lower values (250 ? 17 vs 290 + 16 mg/dl after 6 weeks). Overall, the dietary carbohydrate type Total

Cholesterol

0 250 Apo

B

APO

A-l

200

150 s ‘a E

100

50

0

,

:,.i .::“! .../., :;: .: ‘.’ $ “::i: :::::..,g ..::,,,,.

:....... :“‘y’ ::;;g, ,,, :z.. i::.y:,,:; 1,::‘:g

c_li 3 TIME

4

5

(weeks)

FIG. 1. Changes in serum total cholesterol, apoB and apoA-I levels (mean f SE) with time in cynomolgus monkeys (N = 12) fed starch and sucrose diets containing 0.4 mg cholesterol/kcal.

300

SRINIVASAN

ET AL.

had significant effects on serum total cholesterol, apoB, apoA-I, and triglycerides. Although sucrose diet resulted in triglyceride values somewhat higher than those of the starch diet, the response did not reach hypertriglyceridemic proportions. The influence of dietary carbohydrate type after 6 weeks on individual lipoprotein fractions is shown in Fig. 2. With respect to VLDL, the sucrose diet resulted in a significant increase of triglycerides and similar apoB and cholesterol levels when compared to the starch diet. Cholesterol and phospholipid values for IDL were significantly higher in the starch diet than in the sucrose diet, while triglycerides and apoB values remained similar. In the case of LDL, levels of cholesterol, phospholipid, and apoB were significantly elevated in the starch diet, whereas the carbohydrate type had no effect on triglyceride content. Regarding HDL subfractions, carbohydrate type had a significant effect on HDL,-phospholipid, HDL,-cholesterol, and HDL,-triglycerides, with sucrose producing higher values than starch. When compared to earlier studies in the same animals given high cholesterol (1 mg/kcal) (21), moderate cholesterol intake decreased the cholesterol levels of VLDL and IDL in both starch and sucrose diets with no appreciable effect on LDL and HDL. Role of Intestinal Flora The influence of diet and antibiotics on the major bacterial groups of the fecal microflora is summarized in Table 2. In general, Escherichia and Bacteriodes groups were increased by experimental diets, especially during and after neomycin treatment. Prior to neomycin treatment the Escherichia counts were high during the starch diet as contrasted with the sucrose diet. However, Escherichia, StrepVLDL

2002

5 150f loo-

Starch

Sucrose

Starch

17

Cholesterol

m

Triglycerides

f?J

Phospholipids

m

APO B N=12

Starch

Sucrose

Starch

SUC,OS~

*Starch sucrose txo.05

vs

FIG. 2. Effect of dietary carbohydrate type (starch vs sucrose) on composition of different serum lipoprotein fractions in cynomolgus monkeys (N = 12) following 6 weeks of cholesterol (0.4 mg/kcal) feeding. HDL was separated arbitrarily into “HDL”, and “HDL”3 to represent light and heavy particles, respectively.

CARBOHYDRATES

AND SERUM CHOLESTEROL

Effects of Diet and Antibiotic

301

RESPONSE

TABLE 2 on Fecal Microflora of Cynomolgus Monkeys Microorganism*

Diet?

Eschericha

Basal (chow) Starch Without neomycin$ With neomycin9 Without neomycin5 Sucrose Without neomycinI With neomycin§ Without neomycin$

3.09 2 0.19

8.93 5 0.79

7.95 rt 0.97

6.20 -c 0.79’

7.92 k 0.516 9.01 2 0.29 8.64 2 0.34

8.40 k 1.30 7.75 f 0.23 8.93 5 0.53

7.82 + 0.72 6.63 ?z 2.07 6.07 k 1.80

4.16 -e 1.66 9.17 + 0.43x 9.26 t 0.35

5.83 k 0.34’ 8.90 2 0.16d 7.86 2 0.36

8.47 rt 0.38 7.96 -t 0.07 5.43 * 1.55

7.36 k 0.90 None 7.75 + 0.38

8.55 t 0.24” 8.85 rt 0.17 8.82 t 0.17

* i‘ $ I

Streptococcus

Bacterioides

Lactobacillus

Arithmetic means 2 SE of log,,/g wet feces (three animals/dietary Dietary sequence and treatment are same as in Fig. 4. Two weeks. Four weeks.

group).

a < c; c < b; c < d; e < h; f < g (P < 0.05).

tococcus, and Bacteriodes groups remained at similar levels during the starch and sucrose diets when neomycin was added. This trend persisted even after withdrawal of the antibiotic. Although neomycin possesses a wide range of bactericidal activity against gram-negative and gram-positive bacteria, under in vivo conditions it eliminated only lactobacilli from microflora of monkeys on the sucrose diet. Time-course changes in serum total cholesterol associated with diet and antibiotic are shown in Fig. 3. Serum levels remained essentially the same during each of

CHOLESTEROL

CHOLESTEROL

NEOMYCIN

I

1. I’. .:.: .1 nStarch.N=B a’s

1

2

3

4

5 TIME

6

I

ucrose,N=5

IO

(weeka)

FIG. 3. Effect of neomycin on time-course serum cholesterol responses (mean 2 SE) to starch and sucrose diets in two groups (five animals/group) of cynomolgus monkeys. The two groups were weighted equally with respect to earlier serum cholesterol response (noted in Fig. 1) and fed diets without cholesterol (2 weeks), with cholesterol and neomycin (4 weeks), and with cholesterol (4 weeks) in that order, consecutively. Cholesterol and neomycin were added to the diet at levels of 0.4 mg/kcal and 107 mg/kg body wt, respectively.

302

SRINIVASAN

ET AL.

TABLE 3 Effect of Neomycin on Serum Total Cholesterol and ApoB Responses to Different Carbohydrate Diets in Cynomolgus Monkeys Mean ? SE, mg/dl Diet With neomycin Starch + cholesterol Sucrose + cholesterol Without neomycin Starch + cholesterol Sucrose + cholesterol

Cholesterol 248 k 8’ 228 CL 14 324 k ll* 349 f 23

ApoB 93 e 91*

4b 7

121 + 6b 127 2 10

0 Four weekly observations on five animals as shown in Fig. 3. b Starch vs sucrose, not significant.

the 2 weeks of the starch and sucrose diets containing no added cholesterol. Serum cholesterol response increased with time when cholesterol and neomycin were included in the diets. However, the marked difference in degree of response with time between the starch and sucrose diets noted earlier (Fig. 1) was eliminated, due primarily to decrease in the degree of hypercholesterolemic response to the starch diet. This trend persisted even after removal of the drug for 4 weeks. Overall, mean serum total cholesterol as well as apoB levels during these periods did not differ significantly between starch and sucrose (Table 3). In contrast to these findings, corresponding values for these animals during the first part of the study (prior to neomycin treatment) showed a significant difference between the starch and sucrose diets (data not shown). DISCUSSION

The present study demonstrates that in a diet containing even moderate amounts of cholesterol (0.4 mg/kcal) starch is clearly more hypercholesterolemic than sucrose in cynomolgus monkeys. The magnitude of difference in hypercholesterolemic response of these monkeys to 1 mg cholesterol/kcal between these two carbohydrate types was even larger (21). This differential effect noted in this species is independent of dietary fat content (20,21). Our findings are consistent with earlier data obtained from rhesus monkeys (17) and baboons (33). However, that a similar trend was not seen in cebus monkeys and stumptailed macaques (17) and that dietary sucrose did not induce the usual hypertriglyceridemia underscore interspecies variability in response to different carbohydrates. This may be relevant to the conflicting findings with respect to dietary sucrose and plasma lipids in humans (8,14). It has been suggested that in humans genetic proneness to impaired carbohydrate metabolism may be a major determinant for developing sucrose-induced hyperlipidemia (14,34). The varying effect of carbohydrate type on serum lipids and apoproteins is reflected in individual lipoprotein classes. Sucrose produces a small but consistent increase in triglyceride level of VLDL without altering apoB and cholesterol levels when compared to starch. Since apoB is an integral constituent of VLDL,

CARBOHYDRATES

AND

SERUM

CHOLESTEROL

RESPONSE

303

IDL, and LDL, absolute change in this apoprotein generally reflects alterations in particle number (35). Sucrose seems to cause a proportionately larger triglyceride load per VLDL particle without increasing the number of particles in order to handle high VLDL flux (3-5). Starch in relation to sucrose induced disproportionate increase of cholesterol mass over apoB mass in IDL without affecting apoB levels, suggesting cholesterol enrichment of the molecules in the former. With respect to LDL, both apoB and cholesterol levels were increased by starch when compared to sucrose, indicating increased particle number in the former. Earlier studies in humans indicated that high sucrose-low fat intake (with cholesterol) decreased apoB level of LDL by increasing its catabolic rate (5) which was thought to be mediated through the enhanced fecal bile acid excretion (36). Whether this is specifically related to sucrose or low fat intake is not clear because parallel metabolic information is not available for starch. Recent findings favor the concept that under conditions of increased cholesterol flux the liver can synthesize and secrete cholesterolrich particles within VLDL, IDL, and LDL density ranges (37-40). These particles tend to accumulate in the postabsorptive plasma in response to a saturated hepatic receptor mechanism (41). Therefore, it is conceivable that the starch vs sucrose effect on lipoproteins is merely a reflection of metabolic sequelae of variability in the degree of exogenous cholesterol flux to the liver in large part determined by absorption from the gastrointestinal tract. The present finding that starch when compared to sucrose consistently decreased HDL constituents (apoA-I, phospholipid, and cholesterol) is in agreement with our earlier observations in these monkeys (20,21). Although monkeys generally do not show distinct HDL subfractions as HDL2 and HDL3 are in humans (42), we arbitrarily separated the HDL into HDL, and HDL, based on previous studies in nonhuman primates and humans (20,43). The terms HDL, and HDL3 in our case may represent density subfractions containing light and heavy particles. Dietary factors including carbohydrate are known to affect the level and metabolism of HDL (4,6,44,45). However, no pertinent metabolic data are currently available to explain the starch vs sucrose difference. Alterations in HDLz and HDL, by carbohydrate type provide some clues. Besides the direct effect of carbohydrate type on HDL metabolism, variability in endogenous (VLDL) and exogenous (chylomicron) triglyceride flux can affect HDL. During lipolysis, the surface components of VLDL and chylomicrons (apoA-I, apoC, phospholipid, and free cholesterol) are transformed to HDL3 to form HDL2 (46,47). Relative increase in serum apoA-I and occurrence of triglyceride-rich VLDL particles coupled with marked increase in HDL,-phospholipid following sucrose feeding suggest an increased triglyceride flux and clearance with this diet. Starch, on the other hand, by increasing cholesterol ester rather than triglyceride-enriched hepatic and intestinal lipoproteins, may lead to less transfer of surface components to HDL,. In fact, severe hypercholesterolemia is known to depress HDL levels in cholesterol-fed nonhuman primates (16,48). It is noteworthy that the starch vs sucrose effect on serum cholesterol response was eliminated by neomycin treatment. Earlier, Portman et al. (22) found that in rats the differential effect of carbohydrate type on cholesterol-cholate-induced

304

SRINIVASAN

ET AL.

hypercholesterolemia disappeared when sulfasuxidine was added to the diet. These findings imply that the effect of carbohydrate type on cholesterol-induced hypercholesterolemia may be due to variability in net intestinal sterol flux, mediated in part by intestinal bacteria. The intestinal microflora are known to convert primary neutral and acidic steroids to secondary forms thereby influencing cholesterol absorption (49). In baboons the biliary primary to secondary bile acid ratio is reduced below normal level only in the animals fed fructose and sucrose (33). It is likely that the carbohydrate type can influence sterol absorption and/or excretion by altering the intestinal flora either in numbers and/or in metabolic function. The present study shows alterations in Escherichia group by carbohydrate type, with starch giving higher populations than sucrose. Interestingly, the major groups of fecal bacteria remained high in numbers and did not show any difference between starch and sucrose when treated with neomycin except that the drug eliminated lactobacilli from the microflora of monkeys on the sucrose diet. However, total bacterial counts may not reflect the metabolic activities of the flora. The administration of sulfa drugs or antibiotics to animals causes germfree metabolic characteristics (which persisted in some cases even after withdrawal of drugs) without altering the total bacterial count (50). When administered orally, neomycin effectively blocks the conversion by intestinal microflora of cholesterol and cholic acid into coprostanol and deoxycholic acid, respectively (51,52). It remains to be shown whether neutral and acidic sterols are, in fact, altered in the gut of these monkeys either by carbohydrate type or by antibiotic. The present finding that serum cholesterol responses remained similar between the starch and sucrose diets even 4 weeks after removal of neomycin from the diet underscores the lingering germfree metabolic characteristics vis-a-vis intestinal sterol metabolism and serum cholesterol response. The well-known hypocholesterolemic effect of neomycin is mediated through increased fecal neutral steroid excretion, reflecting a combination of decreased cholesterol absorption, increased cholesterol synthesis secondary to removal of feedback control, and increased flux of cholesterol from tissues (52). Whether this antibiotic is hypocholesterolemic in monkeys fed cholesterol-free laboratory chow diet is not known. Further studies are needed to establish the relationship of intestinal sterol (both exogenous and endogenous) absorption and/or excretion to serum cholesterol levels. SUMMARY

The relationship of carbohydrate type to cholesterol-induced hypercholesterolemia and the potential role of intestinal flora in the above process were examined in 12 male cynomolgus monkeys (M. fusciculuris). Semipurified diets provided two types of carbohydrates (starch or sucrose, 49% by calorie) with 0.4 mg cholesterol/kcal. Six weeks of the starch diet resulted in significantly enhanced hypercholesterolemia when compared to sucrose diet. Starch in relation to sucrose produced cholesterol enrichment of intermediate density lipoproteins and increase in low density lipoprotein particles, whereas sucrose increased high density lipoprotein constituents (phospholipids, cholesterol, and apoA-I) and triglyceride

CARBOHYDRATES

AND SERUM CHOLESTEROL

RESPONSE

305

content of very low density lipoproteins. Fecal Escherichia counts were high during the starch diet as contrasted with sucrose diet, but the Escherichiu, Streptococcus, and Bacteriodes groups did not show differences by diet following each consecutive 4-week period of oral neomycin (107 mg/kg body wt) treatment and withdrawal. The magnitude of hypercholesterolemia during these periods also remained similar between starch and sucrose, suggesting formation of germfree metabolic characteristics. Thus, the magnitude of cholesterol-induced hypercholesterolemia can be affected by the type of carbohydrate, which may be in part determined by intestinal flora metabolism, ACKNOWLEDGMENTS The authors gratefully acknowledge the skillful technical assistance of the NRDC-A Core Lipid Laboratory staff, especially Rajani Sharma and Mildred Care.

REFERENCES 1. Ahrens. E. H., Jr., Connor, W. E., Bierman. E. ef a/., Amer. J. Clin. Natr. 32, 2621 (1979). 2. Grundy. S. M., Bilheimer, D.. Blackburn, H., ef al. Circulation 65, 839A (1982). 3. Ruderman, N. B., Jones. A. L., Krauss. R. M., and Shafrir, E.. J. C/in. Invest. 50, 1355 (1971). 4. Schonfeld, G., Weidman, S. W., Witztum. J. L.. and Bowen, R. M.. Melabolism 25, 261 (1976). 5. Nestel. P. J., Rearden. M., and Fidge, N. H., Metabolism 28, 531 (1979). 6. Blum, C. B.. Levy. R. I., Eisenberg, S., Hall, M.. III, Goebel. R. H., and Berman, M., J. Chin. Invest. 7.

8. 9. IO. Il. 12.

13. 14. 15. 16. 17. 18. 19. 20.

60, 795 (1977).

Gonen, B., Patsch, W., Kuisk, I., and Schonfeld. G., Metabolism 30, 1125 (1981). Antar, M. A., Little. J. A., Lucas, C., Buckley, G. C., and Csima. A., Atherosclerosis 11, 191 (1970). Yudkin, J., Lancer 2, 155 (1957). Ahrens, E. H., Hirsch, J., Oette. K., Farquhar, J. W., and Stein, Y., Trans. Assoc. Amer. Physicians 73, 134 (1961). Hodges, R. E., and Krehl, W. A., Amer. J. Clin. Natr. 17, 334 (1965). Bierman, E. L., Amer. J. Clin. Nutr. 32, 2712 (1979). MacDonald, I., Prog. Biochem. Pharmacol. 8, 216 (1973). Little, J. A., McGuire, V., and Derksen, A.. in “Nutrition, Lipids and Coronary Heart Disease” (R. I. Levy, B. M. Riflcind. B. Dennis. and N. D. Ernst, Eds.), pp. 119-48. Raven Press, New York, 1979. Srinivasan, S. R., McBride, J. R., Jr., Radhakrishnamurthy, B., and Berenson. G. S., Camp. Biochem. Physiol. 47B, 711 (1974). Rude], L. L., and Lotland, H. B., Jr.. in “Primates in Medicine, Atherosclerosis in Primates” (J. P. Strong, Ed.), pp. 224-66. Karger, Base], 1976. Lang, C. M., and Barthel, C. H., Amer. J. C/in. Nurr. 25, 470 (1972). Corey, J. E., Hayes, K. C., Dorr, B., and Hegsted. D. M., Atherosclerosis 19, 119 (1974). Srinivasan, S. R., Radhakrishnamurthy. B.. Webber. L. S., Dalferes. E. R., Jr., Kokatnur, M. G., and Berenson, G. S., Amer. J. C/in. Nutr. 31, 603 (1978). Srinivasan, S. R., Radhakrishnamurthy, B.. Foster. T. A., and Berenson, G. S., Metabolism 32, 777 (1983).

21. Srinivasan, S. R., Radhakrishnamurthy, Nurr.

B., Foster, T. A., and Berenson. G. S., Amer.

J. Clin.

40, 485 (1984).

Portman, 0. W., Lawry. E. Y., and Bruno, D., Proc. Sot. Exp. Biol. Med. 91, 321 (1956). “Eating in America: Dietary goals for the United States.” Report of the Select Committee on Nutrition and Human Needs, U.S. Senate, pp. l-79. MIT Press, Cambridge, MA, 1977. 24. “Lipid Research Clinics Program Manual of Laboratory Operations,” Vol. 1, p. 9. U.S. Department of Health, Education, and Welfare Publication No. (NIH) 75-628, 1974. 25. Fiske, C. H., and Subbarow, Y., J. Biol. Chem. 66, 375 (1925). 26. Hatch, F. T., and Lees, R. S., Adv. Lipid Res. 6, 1 (1968). 22. 23.

SRINIVASAN

306

ET AL.

27. Laurell, C. B., &and. J. Clin. Lab. Invest. 29(Suppl 124). 21 (1972). 28. Miller, J. P., Mao, S. J. T., Patsch, J. R., and Gotto, A. M., Jr., J. L;pjd Res. 21, 775 (1980). 29. Roheim, P. S., and Vega, G. L.. in “Report of the High Density Lipoprotein Methodology Workshop” (K. Lippel, Ed.), pp. 241-248. U.S. Department of Health, Education, and Welfare Publication No. (NIH) 79-1661, 1979. 30. Wahenstein, S., and Fisher, A. C., Biometrics 33, 261 (1977). 31. Grizzle, J. E., Biometrics 21, 467 (1965). 32. Winer, B. J., “Statistical Principles in Experimental Design,” Vol. 1, p. 907. McGraw-Hill, New York, 1971. 33. Kritchevsky, D., Davidson, L. M., and Kim, H. K., et a/., Amer. J. Clin. Nutr. 33, 1869 (1980). 34. Reiser, S., in “Metabolic Effects of Utilizable Dietary Carbohydrates.” (S. Reiser, Ed.), pp. 239-260. Dekker, New York, 1982. 35. Eisenberg, S., and Levy, R. I., Adv. Lipid Res. 13, 2 (1975). 36. Whyte, H. M., Nestel. P. J., and Pryke, E. S., J. Lab. Clin. Med. 81, 818 (1973). 37. Noel, S. P., Wong, L., Dolphin, J. P., Dory, L., and Rubinstein, D., J. C/in. Invest. 64, 675 (1979). 38. Nicolosi, R. J., and Hayes, K. C., Lipids 15, 549 (1980). 39. Illingworth, D. R., Biochim. Biophys. Actu 388, 38 (1975). 40. Thompson, K. H., and Zilversmit, D. B., J. Nutr. 113, 2002 (1983). 41. Brown, M. S., Kovanen, P. T., and Goldstein, J. L., Science 212, 628 (1981). 42. Nelson, C. A., Greer, W. E., and Morris, M. D., Lipids 19, 656 (1984). 43. Scan& A. M., Edelstein, C., Vitello, L., Jones, R., and Wissler, R. W., J. Bjol. C/rem. 248, 7648 (1973).

44. Shepherd, J., Packard, C. J., Patsch, J. R., Gotto, A. M., Jr., and Taunton, 0. D., J. C/in. Invest. 62, 1582 (1978). 45. Tan, M. H., Dickinson, M. A., Albers, J. J., Havel, R. J., Cheung, M. C., and Vigne. J-L., Amer. J. Clin. Nutr. 33, 2559 (1980). 46. Eisenberg, S., Ann. N. Y. Acad. Sci. 348, 30 (1980). 47. Patsch, J. G., Gotto, A. M., Jr., Olivecrona, T., and Eisenberg, S., Proc. Natl. Acad. Scj. US,4 75, 4519 (1978). 48.

Srinivasan, S. R., Radhakrishnamurthy, J. Nutr.

106,

B.. Smith, C. C., Wolf,

R. H., and Berenson,

G. S.,

1757 (1976). A., J. Lipid

49. MacDonald, I. Res. 24, 675 (1983). 50. Gustafsson, B. E., and Novin, K. E., Acta Pntho/. Microbial. Stand. Sect. B 85, 1 (1977). 51. Powell, R. C., Nunes, W. T., Harding, R. S., and Vacca, J. B., Amer. J. Clin. Nurr. 11, 156 (1962).

52. Sedaghat, A., Samuel, P., Grouse, J. R., and Ahrens, E. H., Jr., J. Clin. Invest.

55,

12 (1975).