Low-density lipoprotein cholesterol responsiveness to diet in normolipidemic subjects

Low-Density

Lipoprotein Cholesterol Responsiveness in Normolipidemic Subjects Margaret

to Diet

M. Cobb and Neil Risch

Both apolipoprotein E genotype (apo E) and diet predict very-low-density (VLDL-C) and low-density lipoprotein cholesterol (LDL-C) levels. In a retrospective pooled analysis of six studies, we sought to identify the predictors of VLDL-C and LDL-C change, or “responsiveness,” to a diet crossover. “Response” to diet was studied in 67 normolipidemic subjects of common apo E genotype. Subjects were fed two contrasting, metabolically controlled diets: one had a low polyunsaturated to saturated fatty acid ratio (P:S), and the other had a high P:S ratio. Multiple blood samples were analyzed for VLDL-C and LDL-C levels at the end of each metabolic diet period, and values were averaged and differences were calculated. Despite adjustment for significant predictors across the component studies, a wide range of LDL-C responsiveness was found, with an average decrease of 26 mg/dL. Multivariate regression analysis was used to identify the most significant predictors of LDL-C response to the diet crossover. All dietary and clinical variables were entered by stepwise regression for potential inclusion in a “best-fit” model. The degree of change in saturated fat content and age were the most significant predictors of LDL-C responsiveness. Neither dietary cholesterol nor apo E phenotype were significant predictors of responsiveness. The most LDL-C-responsive subjects were older and required smaller reductions in dietary saturated fat levels than did less-responsive subjects to achieve a comparable reduction in LDL-C levels. Multiple regression analysis suggested a precursor-product relationship between VLDL-C and LDL-C responsiveness. Copyright 0 1993 by W.B. Saunders Company

E

LEVATED LEVELS OF low-density lipoprotein cholesterol (LDL-C) correlate with coronary heart disease (CHD).‘~” Current dietary recommendations emphasize a diet decreased in saturated (< 10% kcal) and increased in polyunsaturated fatty acids (to 10% kcal) to reduce LDL-C levels nationwide.4 Yet, more than 25 years ago, Keys et al5 demonstrated that under metabolic diet conditions individuals differ in the magnitude of serum cholesterol change in response to a high- to low-fat diet (crossover). Due to the unexplained wide range in changes, these estimates were found to be or “responsiveness,” predictive only for large groups of subjects. Knowledge and identification of persons most and least likely to show decreases in LDL-C levels may serve to shape public health strategies. To address this issue, large sample sizes are required due to the variation in cholesterol measurement and “interpatient” responsiveness. Apolipoprotein E (apo E) modulates the catabolism of dietary fat and remnant particles, with its gene coding for three common isoforms, E2, E3, and E4.h-r” Amino acidsequencing shows that apo E4 differs from E3, the most common isoform, by an arginine-for-cysteine substitution at amino acid 112. E2 differs from E3 by a mutation at residue 158 with a cysteine-to-arginine substitution,” with markedly decreased receptor-binding activity compared with E4 or E?. The presence of the apo E4 isoform, in contrast to apo E2, accelerates clearance of very-low-density lipoprotein cholesterol (VLDL-C) and chylomicron remnants, increases delivery of cholesterol to the liver, downregulates the B/E receptors, and increases LDL-C levels.“‘~“-r3 Increased VLDL-C levels, decreased LDL-C concentration, and apo E2 homozygosity underly type III hyperlipidemia. Population studies show that expression of type III hyperlipidemia is dependent on dietary fat intake and other factors.’ Recently, Xu et alI5 in a large populationbase study reported an association between apo E genotype and LDL-C levels on a low polyunsaturated fatty acid to saturated fatty acid ratio (P:S) diet, but not LDL-C responsiveness to dietary change. Less is known about the effect of Metabol/sm, Vol42,No

1 (January),1993:

~~7.13

the apo E gene locus on VLDL-C and LDL-C responsiveness to metabolic diet control in normolipidcmic subjects. Over the past 9 years. lipoprotein responsiveness has been the focus of six metabolic studies at our institution.n-?r Because the sample size in any one of these studies was small, we pooled the data from these studies conducted in normolipidemic subjects to (1) define the predictors of LDL-C change, or responsiveness, to a diet crossover. and (2) to identify mechanisms underlying the most and least “diet-responsive” subjects. We found that the degree of change in dietary saturated fat content and age were the most significant predictors of LDL-C responsiveness. Neither dietary cholesterol content nor apo E phenotype were significant predictors of LDL-C responsiveness. MATERIALS

AND METHODS

This analysis is a composite of six previous studies conducted at two research institutions with coinvestigators. including Brinton et al.th Denke and Breslow,” Weintraub et al,‘s Wissel et al. I9 Fisher et al,‘” and Zanni et aI?’ (Table 1). The details of this pooled investigation and the key common features making pooling possible are described individually in the six studies.‘h-21 This correspon-

From the Laboratory of Biochemical Genetics and Metabolism, The Rockefeller Universiry, New York, NY: and the Department of Human Genetics and Epidemiology and Pahlic Health, Yale lJniver.?it), New Haven. CT. Submitted March I. 1991; accepted .4pril I, 1992. Supported in part bv General Clinical Research Center Grant No. RR00102 and general support from the Pew Trust at The Rockefeller University. and hy National Institutes of Health Grant No. HG00348 at Yale University. Dr Cobb S work was funded hy Suzanne and Irving Karpas, Karpas Health Information Center, Beth Israel Medical Center. New York, NY, The Maeohs Foundation, New York, NY, and donationsfrom W Jablonski. Address reprint requests to Margaret M. Cobb. MD, PhD, 1735 York Ave, New York, NY 10128. Cop.vright 0 1993 by W.B. Saunders Cornpan) 00260495/9314201-0002$03.OOfO 7

8

COBB AND RISCH

Table 1. Composite Study Groups, Hypotheses, and Dietary Changes Dietary Fat Content

Age + SD w

Dietary Cholesterol

FattyAcids, %A Total Kilocalories Total

sat

MOllO

Difference Poly

(mgl1,000

kcalid)

Protein

Carbohydrate

(%A)

(%A)

Rockefeller University site Brinton et alI6 (n = 1 I)

22 ?_ 2

32

22

12

Denke and Breslow” (n = 16)

20 & 1

17

18

13

-7

0

60

0

~32

116

0

Weintraub et aira (n = 8)

28 ‘- 6

0

12

4

-17

-14

-6

0

Wissel et alI9 (n = 4)

27 ? 4

0

22

0

3

-29

159

0

0

Harvard University site Fisher et aP” (n = 19)

25 + 4

0

22

6

-17

0

0

0

Zanni et am (n = 9)

27 + 6

0

7

0

-5

496

0

0

Abbreviations:

Total, total dietary fat content;

Sat, saturated fatty acids (
Mono,

monounsaturated

fatty acids (
Poly,

polyunsaturated fatty acids (> C18:2).

criteria

dent information allowed for statistical equivalence of these groups for the purposes of analysis. Investigations were conducted at Rockefeller University, New York, NY, and Harvard Medical School, Boston, MA. At both study sites, subjects were placed on two consecutively administered, metabolically controlled diets in a paired crossover design, as follows: first, an adverse, low-P:S diet designed to exaggerate a typical Western diet rich in saturated fat and cholesterol, and then, a more beneficial, high-P:S diet tailored to simulate a more therapeutic diet rich in polyunsaturates, low in saturated fat, and considerably lower in dietary cholesterol were fed in a paired crossover design to each subject. Subjects were required to consume their metabolic diets at each meal and were instructed to maintain their usual levels of physical activity throughout all phases of the study. Due to wide differences in diet composition among study groups, data were adjusted for diet composition differences in the analysis. The diet composition changes for each study site are presented in Table 1 and briefly outlined below. Four of the six diet studies maintained the same total fat content of 31% (Fisher et alzO and Zanni et a12t) or 42% (Weintraub et al’s and Wissel et al19) during the low- and high-P:S diets, with replacement of saturates by polyunsaturated fatty acids. The two remaining studies. Brinton et ali6 and Denke and Breslow,” reduced the total fat content, mainly as saturates and monounsaturates, modestly increased polyunsaturates, and replaced the fat calories with carbohydrate. Compared with intakes on the low-P:S diet, dietary cholesterol content was reduced during the high-P:S phase, with the exception of the formula diet of Fisher et al” (no added cholesterol) and the study by Weintraub et aliR (dietary cholesterol content was essentially maintained across both diet periods).

A final cohort of 67 subjects were included

in excellent health and older than I8 years, with LDL-C levels between the 25th and 75th percentile rank, stratified for age and gender, as compared with Lipid Research Clinics Standards.zL The study group comprised volunteers recruited from staff of the respective universities, undergraduate work-study students, or patients who had entered the lipid clinic of each university for routine evaluation. This group included 34 male and 33 female subjects with an average age of 25. The corresponding homozygous genotype apo E 313 was the most prevalent genotype for these study subjects.

Laboratory Analysis Total lipid levels and cholesterol subfractions. At both the Rockefeller and Harvard sites, sequential blood samples were collected after a l2-hour fast and plasma was measured for total lipid levels and cholesterol subfractions. Total cholesterol (TC) and triglyceride levels were assayed enzymatically using reagents from Boehringer Mannheim Biochemicals, Indianapolis, IN. The high-density lipoprotein cholesterol (HDL-C) level was measured as previously described.23 The HDL-C plus LDL-C concentration was determined after air centrifugation to separate the VLDL-C subfraction. LDL-C and VLDL-C levels were determined by difference. as reported previously.” TC and HDL-C levels were standardized by the Lipid Standardization Program at the Centers for Disease Control, Atlanta, GA. Apo Ephenotyping. Apo E phenotyping by isoelectric focusing was used at both the Harvard and Rockefeller University sites.2J

Study Subjects Studies providing the database for this study spanned 9 years (1980 to 1990). Before data analysis, predetermined selection Table 2. Normolipidemic

were developed.

in these metabolic studies; the clinical characteristics and ad libitum lipids are presented in Table 2. Informed consent was obtained from all volunteers after review of the respective study protocols at each of two locations, and subjects were admitted to the university hospital metabolic unit at each site. All subjects were

Subjects’ Ad Libitum Lipoprotein Profiles Ad Libitum Profiles

Total Lipids (mg/dL) Apo E Phenotype

Age W

Cholesterol

TC

TG

VLDL-C

Subfractions LDL-C

(mg/dL) HDL-C

LDL-CIHDL-C

All subjects combined In = 67)

25 + 6

168 +- 30

75 + 38

21 ? 13

97 2 27

51 + 11

3/2 (n = 13)

26 + 3

152 2 6

60 + 22

19 L 7

86 + 9

48+6

1.8 It 0.3

3/3 (n = 44)

25 + 4

168 * 5

78 * 39

22 f 15

95 * 9

52 + 12

2.0 -c 0.4

4/3 (n = 8)

24 ? 2

18428

62 * 20

23 & 13

117 2 11

51 * 14

2.6 + 0.4

4/4 (n = 2)

24 ? 5

191 * 22

58 % 1

24 + 1

104 i 11

63?

1.7 * 0.6

Abbreviation: TG, triglycerides

10

2.0 + 0.8

LOW-DENSITY

LIPOPROTEIN

RESPONSIVENESS

TO DIET

9

Individual Data &Unadjusted)

Fig 1. (A) Unadjusted individual LDL-C levels following the low-P:S (0) and high-P:S diets (0). (B) Distribution of adjusted LDL-C responsiveness to diet crossover for normolipidemic subjects (n = 67). Values were adjusted for age and dietary saturated-fat change.

Low’P:S

B

liigh’P:S

LDL-C “Responsiveness” (Adjusted)

’ ALDL-C,

mg/dl

Dietary P:S Ratio

Analysis A minimum of two to three blood samples were obtained at the end of each study period and averaged for LDL-C and VLDL-C concentrations. VLDL-C responsiveness. VLDL-C responsiveness to diet was defined and calculated for each subject as the difference between the VLDL-C concentration measured following the low-P:S diet, and the corresponding repeated VLDL-C level was measured following the high-P:S diet, as follows: AVLDL-CmgId~ = VLDL-C level~Ou_p.s- VLDL-C lev&,ig~_~.s. LDL-C responsiveness was calculated as described for VLDL-C above, ie. ALDL-Cmg,d~ = LDL-C levell,,.,,:s - LDL-C levelhIph.p.s.

Statistical Analysis Biomedical Computer Programs (BMDP Statistical Software. Berkeley, CA, 1985) were used for statistical analyses. Univariate analyses were conducted to isolate potential predictors of LDL-C responsiveness for all normolipidemic subjects. Analysis of covariante was used to adjust for group differences in dietary saturatedfat change and age to present the adjusted range of LDL-C responsiveness (Fig 1B). Using LDL-C responsiveness as the dependent variable expressed as an absolute level (mg/dL), we used stepwise multiple regression analysis to identify the most significant independent predictors of response in normolipidemic subjects. The independent variables included (1) clinical variables, ie. age. gender. body mass index, caloric intake, and apo E genotype, and (2) dietary variables expressed as percentage change (%A). ie, fatty acids (saturates%A, polyunsaturatesc%A, and monounsaturatesc;J, dietary cholesterol (expressed as square root of difference between two diets), total fatg;A. protein%a, and CarbohydratesqA. The “best-tit” model, explaining the variance in LDL-C responsiveness and containing only independent variables with regression coefficients significantly different from zero (P < .05). was selected. RESULTS

individual LDL-C Responsiveness While treating patients on a Western diet with a more “therapeutic” dietary regimen, it is essential to appreciate

the range of LDL-C levels attained under metabolic conditions, even in normal subjects. Individual LDL-C levels following the low- and high-P:S diet are presented in Fig 1A. LDL-C levels following the low-P:S diet varied between 80 and 190 mg/dL, with a range of 110 mg/dL. Ten subjects had LDL-C levels greater than 130 mg/dL and 26 subjects achieved LDL-C levels lower than 100 mg/dL following the low-P:S diet. LDL-C levels attained following the high-P:S diet varied between 34 and 158 mg/dL. Two subjects had LDL-C levels greater than 130 mg/dL, while 40 subjects achieved LDL-C levels lower than 100 mg/dL. These data demonstrate the degree of individual variation in LDL-C levels while following a Western diet and the range of LDL-C responses to dietary changes. Variations in LDL-C Responsiveness

to Diet

Following the diet crossover, individual variations in LDL-C responsiveness were noted, even with strict metabolic diet control. The distribution of LDL-C responsiveness after adjustment for age and differences in dietary saturated-fatty acid content for all subjects, is presented in Fig 1B. The distribution was unimodal and skewed toward greater LDL-C responsiveness. Despite the unimodal distribution, there was a wide range in LDL-C responsiveness, from a decrease of 15 mg/dL (least responsive) to a decrease of 53 mg/dL (most responsive). The mean LDL-C response for these subjects was 28 ? 8 (SD) mg/dL. After adjustment for age and dietary saturated-fat content, no subject fell below 3 standard deviations from the mean, while only one subject, as described above, exceeded 3 standard deviations above the mean. Effects of Apo E on Lipoprotein Levels and Responses The effects of apo E genotype on VLDL-C and LDL-C levels and responses following the diet crossover are shown in Table 3.

IO

COBB AND RISCH

Table 3. Lipoprotein Responses to Diet Crossover and Predictors of LDL-C Response in Normolipidemic Metabolic VLDL

LDL

LWl?lS Apo

E Phenotype

Low-P:S

Diet

Rl?SpXW?S

High-P:S

Diet

Subjects

Diets

(AmgidL)

All subjects (n = 67)

I9 k 7

I7 k 8

-2

3/Z (n = 13)

I9 2 4

18 + 9

-1

LC?WlS

(A%)

-II

Low-P:S

Diet

107 + 27

-5

88 + 8

RESpOllSeS

High-P:S

Diet

IAmqidL)

IA%)

79 k 24

-28

-26

67 k 4

-21

-24

3/3 (n = 44)

19 * 8

17 t

a

-2

-II

107 + 7

80 t 3

-27

-25

4/3 (n = 8)

18+

5

I4 * 4

-4

-22

124?

IO

95 f 5

-29

-23

4/4 (n = 2)

16 +

a

I2

-4

-25

105 f 27

86 + 3

-19

-18

-c 2

PZDL-C levels. The VLDL-C reduction, in terms of absolute levels and percentage change, tended to increase in stepwise fashion for the three most common genotypes as 312 < 313 < 413, with changes of -5% (-1 mg/dL), - 11% (-2 mg/dL), and -22% (-4 mg/dL), respectively. Although they showed considerably lower VLDL-C levels on both metabolic diet regimens, the apo E 414 subjects showed an average change of -4 mg/dL (-25%), approximating the responsiveness of subjects carrying the apo E 413 genotype. LDL-C levels. The impact of apo E genotype on LDL-C level was investigated independently. The univariate results for subfractions showed that apo E genotype was a statistically significant predictor of LDL-C levels (P < .OS) following both the low- and high-P:S diets, with, in order of increasing concentrations, apo E 312 < 313 < 413. Apo E 414 subjects exhibited LDL-C levels on both the low- and high-P:S diets similar to those of apo 413 subjects, but somewhat higher than the apo E 312 and 313 genotypes. Following the diet crossover, subjects with the apo E genotype 3/2 showed substantially less change in LDL-C level, calculated from the absolute difference (mg/dL), compared with apo E genotypes 313 and 413, but levels were similar when expressed as percent change from the low-P:S diet. None of these changes reached statistical significance. Although the LDL-C response (mg/dL) showed the anticipated stepwise increase of 312 < 313 < 413, the percentage change from the low-P:S level was similar, ranging between 23% and 25% for all three genotypes. The apo E 414 subjects tended to respond the least both in terms of absolute levels (-19 mg/dL) and percentage (-18%) although levels were similar to those of carriers of the more common apo E 413 genotypes. Thus, although apo E genotype predicted LDL-C levels, it failed to predict the magnitude of change, or response to dietary change. This finding suggests an additive effect (ie, no interaction) between the dietary P:S ratio and apo E phenotype. Best-fit equation. The differences in dietary composition were used to identify significant predictors of lipoprotein responsiveness. Although change in total fat content was predictive (P < .05) of change in LDL-C level, the single most predictive variable of LDL-C responsiveness in normolipidemic subjects was the magnitude of change in dietary content of saturated fatty acids (sat%A) following the diet crossover (P < .Ol). The change in dietary content of monounsaturated fatty acids was not predictive of change in LDL-C level (P > .05). The best-fit, two-variable model

describing LDL-C responsiveness in normohpidemic subjects included sat%A and age, in years (age,,), accounting for 42% of the variance in LDL-C responsiveness (P < .OOOl). Interrelationship between PZDL-C and LDL-C responsiveness. Although baseline VLDL-C levels were not significantly correlated with VLDL-C responsiveness, LDL-C levels were highly predictive (P < .Ol) of LDL-C change (mg/dL). There was a strong, positive correlation between LDL-C responsiveness and VLDL-C responsiveness after adjustment for age and diet, as follows: Aadj.LDL-C,,,,,

= 0.7 (AVLDL-C,,,,,) + 24 mg/dL;

P < .Ol, R* = 32.

[Eq l]

Thus, following the diet crossover, for each lo-mg/dL reduction in VLDL-C level, there was an associated 7-mg/dL decrease in LDL-C level. DISCUSSION

Data from six previous studies were pooled to define the predictors of VLDL-C and LDL-C responsiveness to dietary change. The most LDL-C-responsive subjects were older and required a smaller reduction in dietary saturated fat content than did less-responsive subjects to achieve a comparable reduction in LDL-C levels. Multiple regression analysis suggested a precursor-product relationship between VLDL and LDL responsiveness. Research to date has suggested that the rate of dietary fat clearance from blood, as mediated by apo E genotype, is one genetic predictor of absolute LDL-C levels.10J3J4 Apo E serves as the ligand for the chylomicron and VLDL remnant receptors, modulating the rate of hepatic uptake of these particles from plasma and thus indirectly regulating LDL-C levels.25-27 This observation was confirmed in our pooled study, in which LDL-C levels were correlated with apo E genotype in order of increasing LDL-C levels, such that 312 < 313 < 413 following both the low- and high-P:S diets (Table 3). However, our composite study showed that the change in dietary saturated-fat composition and age were more statistically significant predictors of LDL-C responsiveness than apo E (Table 3). Like apo E, another apolipoprotein (apo B) serves as a ligand for the LDL receptor and is the solitary protein of LDL. In the pig, an animal model for LDL-apo B metabolism, following a dietary crossover from a low-fat to a high-fat, atherogenic pig diet, apo B structural gene variations correlated with the degree of cholesterol responsive-

LOW-DENSITY

LIPOPROTEIN

RESPONSIVENESS

TO DIET

ness.Zx Animals with the Lpb5 genotype consistently showed a greater cholesterol response and accumulation of fatty plaques following dietary changes. Although no genetic markers for apo B are associated with responsiveness in humans, this gene locus may regulate LDL-C responsiveness to diet. High dietary saturated-fat intake down-regulates and low intake up-regulates apo B/E receptors, thus affecting lipoprotein clearance from the blood and resultant LDL levels.Z4~“1In the present study, following each diet LDL-C levels and responses were unimodal and normally distributed, without evidence of clear “hypo-LDL-C responders,” as previously suggested.“1-“4 However, one subject showed a less than 16-mg/dL LDL-C change, and three subjects showed LDL-C responses of greater than 43 mg/dL. Although adjustment for dietary composition reduced the range of LDL-C responses, there was still a wide range of responsiveness. representative of the influence of other dietary and/or genetic factors such as other lipid transport genes that regulate lipoprotein metabolism.” In normocholesterolemic subjects, the precise nature of this variability in diet responsiveness is currently unknown. Given our equation from Table 3, several hypotheses implicating diet change are plausible. The variability in LDL-C responsiveness to diet may reside in the degree of regulation of hepatic LDL receptors, as determined by individual responses to changes in dietary saturated-fat content and aging.‘y~-“.J5.3hGenetic defects have been identified in coding for the hepatic LDL receptor.” Together or separately, factors such as interindividual differences in the lipoprotein particle, ie, number, size, or lipid composition, or degree of LDL-receptor suppression could determine LDL-C responsiveness to dietary change. Age correlated with LDL-C responsiveness by both univariatc and multivariate analysis. Animal and human studies have shown generally higher plasma LDL-C levels due to decreased LDL-C-receptor activity and a decline in the LDL-C fractional catabolic rate with aging.3s.3h Our analysis shows that older subjects were more responsive to diet, ie, they showed a greater reduction in LDL-C level for a given dietary saturated-fat change, confirming the correlation of baseline levels with responsiveness in other studietix and supported by our findings. In our cohort, apo E genotype was neither predictive of VLDL-C levels nor of VLDL-C or LDL-C responsiveness. Although this negative finding regarding apo E phenotype impact on responsiveness does not exclude an effect due to small sample size and diet heterogeneity, it confirms the findings of Xu et alI5 and O’Malley and Illingworth.3y These investigators reported that apo E genotype was not predictive of LDL-C responsiveness to diet15 or drug3” treatment. O’Malley and Illingworth3” studied the interrelationship between apo E genotype and response to lovastatin therapy in familial hypercholesterolemic patients, a noteworthy comparison to our study findings, since a high-P:S diet and lovastatin therapy share a common site of influence, ie, up-regulation of the LDL-C receptor leading to enhanced plasma LDL-C clearance.?“J1~40 Our investigation seems at variance with the results of

11

Manttari et aL4i who found a positive association between the apo E4 allele and LDL-C responsiveness to dietary change. However, this latter study lacked metabolic ward control, since the therapeutic diet was given on an outpatient, counseled basis without precise quantification of saturated fat and cholesterol content. Stepwise multiple regression analysis yielded best-fit models for significant prediction of LDL-C responsiveness to dietary change in normolipidemic subjects. Equations derived from these models may permit theoretical applications to a clinical setting, as follows: ALDL-C,,,,,_

= O.h(sat,,)

+ l.O(agc,,)

- 9;

R = .42, P < .OOOl.

[Eq 21

Applying equation no. 2 if, for example, a physician wishes to reduce LDL-C level from 150 to 130 mg/dL, ie, to achieve a 20-mg/dL LDL-C reduction in a normolipidemic subject, the resultant necessary decrease in dietary saturated-fat content, expressed as percent change, would be approximately 8% and 2% for a 20-year-old and 30-year-old subject, respectively. Thus the most LDL-C-responsive subject required the smallest reduction in saturated-fat content for a given decrease in LDL-C level and tended to be older. The least-responsive subject, for the same reduction in LDL-C level, required a greater change in dietary saturated-fat content and tended to be younger. Separate equations for hypercholesterolemic subjects also could potentially define responsiveness in this group.

Metabolic Basis for LDL-C Responsiveness In the present study, adjusting for age and diet, we found a positive correlation between AVLDL-C and ALDL-C for all subjects (equation no. 1). Thus, the more VLDL-Cresponsive subjects also tended to have concomitantly greater reductions in plasma LDL-C level, indicative of the precursor-product relationship between these two lipoprotein subfractions. We have devised a model to graphically illustrate the theoretical bases for the metabolic relationship between plasma VLDL-C and LDL-C responsiveness (Fig 2). The model also illustrates an altered metabolic relationship between VLDL-C and LDL-C in moreresponsive and less-responsive subjects following the diet crossover. Following hepatic secretion into plasma, most VLDL is converted to LDL at rates determined by dietary and genetic parameters. VLDL and LDL particles are subsequently bound to hepatic receptors and removed from plasma (Fig 2). We arbitrarily stratified subjects into two classes, more and less LDL-C-responsive (as defined in Fig 2). The more LDL-C-responsive subjects showed a greater reduction in both VLDL-C and LDL-C levels, suggesting reduced VLDL-C to LDL-C conversion and/or enhanced VLDL-C and LDL-C uptake by receptors. Thus, more-responsive subjects showed a greater decrease of both VLDL-C and LDL-C levels following the diet crossover. Meanwhile, less LDL-C-responsive subjects showed a smaller reduction in VLDL level, potentially greater VLDL-C conversion to

12

COBB AND RISCH

MECHANISM: ILIVER +

PLASMA -1

CONVERSION

VLDL

LDL

)LDL UPTAKE VLDL UPTAKE

"MORE LDL-C RESPONSE":

RESPONSE": ENTRY

ENHANCED

VLDL

REDUCED

)LDL

A

REDUCED VLDL UPTAKE

LDL-C, and/or delayed VLDL-C and LDL-C uptake by receptors. The current study suggests that further genetic research is needed to correlate specific apolipoprotein with LDL-C responsiveness to diet. As we gain about lipoprotein interrelationships, it may be predict an individual’s LDL-C responsiveness thus design an individualized diet regimen for specific LDL-C target level.

genotypes knowledge possible to to diet and achieving a

Fig 2. Proposed model of LDL-C responsiveness to diet crossover. Arrows denote fluxes into and out of plasma VLDL-C and LDL-C compartments. More LDL-C-responsive subjects showed greater reductions in VLDL-C levels and reduced conversion of VLDL-C to LDL-C, and thus show decreased LDL-C levels in response to diet crossover. Less LDL-C-responsive subjects show less dramatic VLDL-C reductions and increased conversion of VLDL-C to LDL-C, and thus show less marked decreases of LDL-C levels in response to diet crossover.

ACKNOWLEDGMENT

The authors are deeply indebted to the study subjects at the respective institutions. Gratitude is extended to Dr Jan Breslow for supplying the original data for the manuscript. We thank Cynthia Seidman and Donna Tesi of the Rockefeller University Dietetics Department for formulation and preparation of the metabolic diets, and the nursing service at the Harvard and Rockefeller University Metabolic Units. Gratitude is extended to Michael Friedmann for his assistance in preparation of the manuscript.

REFERENCES

1. Castelli WP, Garrison RJ, Wilson PWF, et al: Incidence of coronary heart disease and lipoprotein levels: The Framingham Study. JAMA 256:2835-2838,1986 2. Lerner DJ, Kanner WB: Patterns of coronary heart disease morbidity and mortality in the sexes: A 26 year follow-up of the Framingham population. Am Heart J 111:383-390, 1986 3. Martin MJ, Hulley SB, Browner WS, et al: Serum cholesterol, blood pressure and mortality: Implications from a cohort of 361,662 men. Lancet 2:933-936, 1986 4. Nutrition Committee, American Heart Association: Dietary guidelines for healthy American adults. A statement for physicians and health professionals by the Nutrition Committee, American Heart Association. Circulation 77:721A-724A, 1988 5. Keys A, Anderson JT, Grande F: Serum cholesterol response to changes in the diet. III. Differences among individuals. Metabolism 14:766-775, 1965 6. Breslow JL: Genetic basis of lipoprotein disorders. J Clin Invest 84:373-380,1989 7. Utermann G, Hees M, Steinmetz A: Polymorphism of apolipoprotein E and occurrence of dysbetalipoproteinemia in man. Nature 269:604-607, 1977

8. Zannis VI, Breslow JL: Human very low density lipoprotein apolipoprotein E isoprotein polymorphism is explained by genetic variation and post-translational modification. Biochemistry 20:10331041, 1981 9. Zannis VI, Just PW, Breslow JL: Human apolipoprotein E isoprotein subclasses are genetically determined. Am J Hum Genet 33: 1l-34,1981 10. Weintraub MS, Eisenberg S, Breslow JL: Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. J Clin Invest 80:1571-1577, 1987 11. Rall SC, Weisgraber KH, Mahley RW: Human apolipoprotein E. The complete amino acid sequence. J Biol Chem 257:41714178,1982 12. Kesaniemi YA, Ehnholm C, Miettinen TA: Intestinal cholesterol absorption efficiency in man is related to apoprotein E phenotype. J Clin Invest 80:578-581, 1987 13. Utermann G. Kindermann I. Kaffarnik H, et al: Apolipoprotein E phenotypes and hyperlipidemia. Hum Genet 65:232-236. 1984 14. Assmann G, Schmitz G, Menzel HJ, et al: Apolipoprotein E polymorphism and hyperlipidemia. Clin Chem 30:641-643, 1984

LOW-DENSITY LIPOPROTEIN RESPONSIVENESS TO DIET

15. XII CF. Boetwinkle at the apolipoprotein lipids to diet change.

E, Tikkanen

MJ, et al: Genetic

gene loci contribute to response Genet Epidemiol7:261-275. 1990

16. Brinton EA. Eisenberg S, Breslow decreases high density lipoprotein (HDL) decreasing HDL apolipoprotein transport 85:144-151.1990

13

variation of plasma

JL: A low-fat diet cholesterol levels by rates. J Clin Invest

17. Denke MA, Breslow JL: Effects of a low fat diet with and without intermittent saturated fat and cholesterol ingestion on plasma lipid. lipoprotein. and apolipoprotein levels in normal volunteers. J Lipid Res 29:963-969, 1988 18. Weintraub MS, Zechner R, Brown A, saturated fats of the W-6 and W-3 series lipoprotein levels. Chronic and acute effects postprandial lipoprotein metabolism. J Clin 1988

et al: Dietary polyunreduce postprandial of fat saturation on Invest 82:1884-1893.

19. Wissel PS, Denke MA, Inturrisi CA: A comparison of the effects of a macrobiotic diet and a Western diet on drug metabolism and plasma lipids in man. Eur J Clin Pharmacol 33:403-407. 19x7 70. Fisher EA. Blum CB, Zannis VI, et al: Independent effects of dietary saturated fat and cholesterol on plasma lipids, lipoproteins. and apolipoprotein E. J Lipid Res 24:1039-1048, 1983 21. Zanni EE, Zannis VI, Blum CB, et al: Effect of egg cholesterol and dietary fats on plasma lipids, lipoproteins, and apoproteins of normal women consuming natural diets. J Lipid Res 28518527. 1987 22. Rifkind BM (ed): The Lipid Research Clinics Populations Studies Data Book. vol I. The Prevalence Study. Bethesda. MD, NIH publication no. 80-1527, 1980, pp 70-73 23. Warnick GR, Benderson J, Albers JJ: Dextran sulfate-Mg’+ precipitation procedure for quantitation of high-density-lipoprotein cholesterol. Clin Chem 28:1379-1388, 1982 74. Zannis VI. Breslow JL: Human VLDL apo E isoprotein polymorphism is explained by genetic variation and post-translational modification, Biochemistry 20:1033-1041, 1981 25. Sherill BC, Innerarity TL, Mahley RW: Rapid hepatic clearance of canine lipoproteins containing only the E apoprotein by high affinity receptor (identity with the chylomicron remnant transport process). J Biol Chem 255:1804-1807, 1980 26. Hui DY. Innerarity TL, Mahley RW: Lipoprotein binding to canine hepatic membranes. Metabolically distinct apo E and apo BE receptors. J Biol Chem 256:5646-5655, 1981 77. Mahley RW, Hui DY, Innerarity

TL. et al: Two independent

lipoprotein

receptors

on hepatic

membranes

of dog, swine

and

man. J Clin Invest 681197-1206, 1981 28. Rapacz J, Elson CE, Lalich JJ: Correlation of an immunogenetically defined lipoprotein type with aortic intimal lipidosis in swine. Exp Mol Path01 27:249-261, 1977 29. Spady DK, Dietschy JM: Dietary saturated triacylglycerols suppress hepatic low density lipoprotein receptor activity in the hamster. Proc Nat1 Acad Sci USA 82:4526-4530, 1985 30. Spady DK, Bilheimer DW, Dietschy JM: Rates of receptordependent and independent low density lipoprotein uptake in the hamster. Proc Natl Acad Sci USA 80:3499-3503, 1983 31. Spady DK. Dietschy JM: Interaction of dietary cholesterol and triglycerides in the regulation of hepatic low density lipoprotein transport in the hamster. J Clin Invest 81:300-309, 1988 32. Jacobs DR, Anderson JT, Hannan P, et al: Variability in individual serum cholesterol response to change in diet. Arteriosclerosis 3:349-356, 1983 33. Katan MB, Beynen AC: Characteristics of human hypo- and hyperresponders to dietary cholesterol. Am J Epidemiol 125:387399, 1987 34. Beynen AC, Katan MB: Reproducibility of the variances between humans in the response of serum cholesterol to cessation of egg consumption. Atherosclerosis 57: 19-31, 1985 35. Miller NE: Why does plasma low density lipoprotein concentration in adults increase with age? Lancet 4:263-267, 1984 36. Grundy SM. Vega GL, Bilheimer DW: Kinetic mechanisms determining variability in low density lipoprotein levels and rise with age. Arteriosclerosis 5:623-630, 1985 37. Soria LF, Ludwig EH. Clarke HRG. et al: Association between a specific apolipoprotein B mutation and familial defective apolipoprotein B-100. Proc Natl Acad Sci USA 86:587-591. 1989 38. Grundy SM, Vega GL: Plasma cholesterol “responsiveness” to saturated fatty acids. Am J Clin Nutr 47:822-824. 1988 39. O’Malley JP. Illingworth DR: The influence of apolipoprotein E phenotype on the response to lovastatin therapy in patients with heterozygous familial hypercholesterolemia. Metabolism 39: 150.154,199o 40. Bilheimer DW, Grundy SM. Brown MS, et al: Mevinolin and colestipol stimulate receptor-mediated clearance of low density lipoprotein from plasma in familial hypercholesterolemia heterozy gotes. Proc Nat1 Acad Sci USA 80:4124-4128.1983 41. Manttari M, Koskinen P, Ehnholm C, et al: Apolipoprotein E polymorphism influences the serum cholesterol response to dietary intervention. Metabolism 40217.221. 1991