Genetic Epidemiology 7:261-275 (1990)

Genetic Variation at the Apolipoprotein Gene Loci Contribute to Response of Plasma Lipids to Dietary Change C.-F. Xu, E. Boerwinkle, M.J. Tikkanen, J.K. Huttunen, S.E. Humphries, and P.J. Talmud Charing Cross Sunley Research Centre, London, U.K. (C.-F.X., S. H., P. T.); Center for Demographic and Population Genetics, The University of Texas Health Science Center, Houston (E.8.);First Deparfment of Medicine, University of Helsinki (M. T.) and The National Public Health Institute (J.H.), Helsinki, Finland Dietary intervention studies (from a low polyunsaturated/saturated fatty acid ratio PIS diet to a high PIS diet), carried out on a group of healthy individuals from North Karelia, Eastern Finland between 1981-1984, provided evidence that there may be a genetic component contributing to variation in response to dietary change. We have resampled blood from 107 individuals involved in the original studies and used Restriction Fragment Length Polymorphisms (RFLPs) to study the genetic contribution of variation at a number of candidate gene loci to the response to dietary change. The genes investigated in this study were the apolipoprotein (apo) genes: apo B, apo AII, apo E (protein polymorphism), apo AI-CIII-AIV gene cluster, and the LDL-receptor gene. On the basal diet the major effect of genotype on lipid traits was due to variation at the apo E gene locus; this protein polymorphism explained 14.6%of the phenotypic variance in LDL cholesterol levels and 12.7% of the phenotypic variance in total cholesterol levels. When switched to low fat high PIS diet, these effects of variation at the apo E gene locus on the phenotypic variation of LDL and total cholesterol levels disappeared. The major effect on the response to dietary change, A, was seen on the difference in apo A1 levels mediated by variation at the apo B gene locus (MspI RFLP) explaining 6.3% of the phenotypic variance in apo A1 change. For the RFLPs of the apo AI-CII-AIV gene cluster, small but not significant differences on A were found. Our results indicate that within the limits of the candidate genes studied, the major effects in response to dietary change was on apo A1 levels mediated through variation at the apo B gene locus. Received for publication January 2, 1990; revision accepted April 26, 1990. Address reprint requests to Dr. Philippa Talmud, Charing Cross Sunley Research Centre, Lurgan Ave, London W6 8LW. U.K.

0 1990 Wiley-Liss, Inc.

262

Xuet al.

Key words: dietary intervention, restriction fragment length polymorphism, apo B, apo AII, apo AI-CIII-AIV, apo E, LDL-receptor

INTRODUCTION

There is now strong evidence from a number of epidemiological studies of an independent and positive association between the incidence of coronary heart disease (CHD) and raised serum cholesterol levels [Kannel et al., 1971; Shaper et al., 19851 and an inverse association with raised high density lipoprotein (HDL) levels [Kannel et al., 197I]. Furthermore, both family- and population-based studies provide evidence that approximately 50% of the variance in serum cholesterol levels in the general population is attributable to environmental factors including diet, physical activity, and smoking. The balance of this variance is explained by genetic differences among individuals [Sing and Boerwinkle, 19871. In order to examine the influence of dietary change on serum lipid, lipoprotein, and apolipoprotein levels, controlled dietary intervention studies were carried out between 1981 and 1984 on individuals from North Karelia, an eastern province of Finland with a high incidence of heart disease and a high average dietary fat intake. The outcome showed a significant drop in serum low density lipoprotein (LDL), apolipoprotein (apo) B, HDL, and apo A1 levels [Ehnholm et al., 1982, 19841 and a change in lipoprotein composition when subjects were on a low fat, high polyunsaturated to saturated fatty acid ratio (P/S) diet [Kuusi et al., 19851. Thus a temporary change from a high fat low PIS Finnish diet to a low fat high PIS diet, typical of Mediterranean countries, could influence serum lipid, lipoprotein, and apolipoprotein levels. Although on average there was a response to dietary intervention by a general lowering of lipid levels, all participants did not respond to the change in diet to the same extent. Interindividual variation in both basal levels and dietary response of serum lipid, lipoprotein, and apolipoproteins are likely to be attributable to both environmental and genetic factors. Whether the genes contributing to response are the same as those contributing to the basal levels is yet to be established. Several studies have reported population associations between serum lipid, lipoprotein, or apolipoprotein levels and Restriction Fragment Length Polymorphisms (RFLPs) of the apo B gene (XbaI, MspI and EcoRI) [Berg, 1986; Law et al., 1986; Talmud et al., 1987; Aalto-Setala et al., 1988; Rajput-Williams et al., 1988; Xu et al., 19901 and the apo AI-CIII-AIV gene cluster (SstI, PstI, XmnI) [Rees et al., 1985; Kessling et al., 1988a,b; Wile et al., 1989; Xu et al., 19901. Variation at the apo E gene locus has been associated with variance in serum cholesterol and LDL levels [Utermann et al., 1979; Sing and Davignon, 19851. Pederson and Berg [1989] have reported an association between variation at the LDL receptor gene locus and phenotypic variation in LDL levels. Using molecular biology techniques, we have investigated whether variation at these genetic loci, namely the apo B, apo AII, apo E, and apo AI-CIII-AIV gene cluster and LDL receptor, is associated with individual variation in lipid, lipoprotein, and apolipoprotein response to dietary intervention. Our results provide evidence that response of lipid, lipoprotein, and apolipoprotein levels to dietary modification is, at least in part, under genetic influence.

Genetic Variation Contributes to Response to Dietary Change

263

METHODS Subjects In the original three North Karelia dietary intervention studies [Ehnholm et al., 1982,1984; Kuusi et al., 19851, married men were recruited through local risk factor screenings or countrywide hypertension registers. Wives who also fulfilled the criteria of no major disease, long-term medication, or a history of dyslipoproteinemia were asked to participate. A total of 250 individuals were involved. For the present study only the subjects participating in identical low fat, low cholesterol, high P/S intervention diets, within each of the three studies, were invited to participate; 107 of the 130 agreed to participate in this study. The subjects were unrelated married couples except for four individuals (three males and one female). Fresh blood samples were taken after overnight fasting, from the 55 men and 52 women aged 30-50 years. Dietary Intervention Study

Full details of the dietary intervention studies have been presented previously [Ehnholm et al., 1982, 1984; Kuusi et al., 19851. In brief, participants undertook a 2-week baseline diet following their natural eating habits, followed by an intervention period on a diet low in fat content with a raised P/S resembling the diet followed in Mediterranean countries, with particular attention being paid to the saturated fat, cholesterol, and vegetable content. Families were visited at least twice a week by a nutritionalist to supervise adherence to the diet. Body weight was continuously monitored to keep energy intake and body weight as constant as possible. Food consumption records were kept by all participants every day for 1 week on the baseline period and every second day on the intervention diet. Total energy remained relatively stable during the baseline and intervention periods (2,905 2 94 kcal and 2444 ? 79 kcal, respectively). Total fat as a percentage of total energy was reduced from 38.9 +- .06% on the baseline diet to 24.1 +- .05% on the intervention diet. The P/S ratio increased from 0.15 2 .01 on the baseline diet to 1.18 k .03 on the intervention diet. Fasting venous blood was drawn for lipid and lipoprotein analysis at the end of the baseline diet and after six weeks intervention. Lipid and Lipoprotein Determination Serum cholesterol and triglyceride concentrations were determined using an Autoanalyser I1 apparatus (Technicon Instruments) and enzymatic assays (Boehringer Mannheim, GmBH). HDL cholesterol was measured after precipitation of VLDL and LDL with dextran sulphate-magnesium chloride [Kostner, 19761. The serum LDLcholesterol concentration was calculated using the Friedwald approximation [Friedewald et al., 19721. The concentration of apo B was determined by radial immunodiffusion kit M-Partigen-apolipoprotein B (Behring-Werke AG) and apo A1 and apo A11 with another immunodiffusion method [Huttenen et al., 19791. Apo E phenotyping was carried out using a modification [Lukka et al., 19881 of the method of Havekes et al. [ 19871. Cysteamine treatment of plasma prior to isoelectric focusing was carried out according to Weisgraber et al. [ 19821. DNA Analysis DNA was prepared from frozen whole blood by the Triton X-100 lysis method [Kunkel et al., 19771. DNA (5pg) was digested with each of the restriction enzymes

264

Xuetal.

PvuII, XbaI, MspI, EcoRI, XmnI, PstI, StuI, ApaLI, NcoI, and SstI using 5-10 units of enzyme per pg of DNA, following the manufacturers recommended conditions (Anglian Biotech). Digested DNA was size separated on 0.7-1 .O% agarose gels and transferred to Hybond-N filters (Amersham) by Southern blotting as previously described [Kessling et al., 1988al. The apo B gene probes used were a unique genomic EcoRI fragment (pAB 3.5C) to detect the XbaI RFLP [Talmud et al., 19871 and the MspI RFLP [Xu et al., 19891, a 2.OKb Hind111 genomic fragment (BH2.0) of the 3’ flanking region to detect the EcoRI RFLP [Talmud et al., 19881 and a 5’ cDNA probe, 959, to detect the PvuII-B RFLP [Darnfors et al., 19861. The apo AI-CIII-AIV gene cluster probes were a unique 2.2 kb PstI fragment of the apo A1 gene to detect the XmnI, PstI, and SstI RFLPs and 1.Okb PvuII fragment and 1.05 kb Pst I fragment to detect the PvuII-CIII RFLP and the PvuII-AIV RFLP respectively [Kessling et al., 1988bl. ApaLI, StuI, NcoI, and PvuII-LDL-R RFLPs of the LDL receptor gene were detected using cDNA probe pHHI, a kind gift of Dr. D Russell. The MspI RFLP of the apo A11 gene was detected using the apo A11 cDNA probe [Scott et al., 19851. The labelling of probes, hybridization, filter washing, and autoradiography procedures were as previously described [Barni et al., 19861. Statistical Analysis

Each of the lipid, lipoprotein, and apolipoprotein traits from the basal diet “A”, the intervention diet “B”, and their change A(A = “A”-“B”) were adjusted separately by stepwise regression for age, gender, and Body Mass Index (BMI). All lipids were normally distributed except for triglycerides, which had a scewed distribution and were log transformed. A one-way analysis of variance was performed separatelyon the adjusted lipid levels from “A”, “B”, and A, to test the null hypothesis that phenotypic variation was not associated with genetic variation at the genetic loci under study. The proportion of the total variance for each of the lipid traits, attributable to genetic variation at the apolipoprotein gene loci, was estimated by the R2 X 100 from the analysis of variance. These R2 values were not independent of one another because of dependencies in the data; this nonindependence may be the result of sampling or associations among the loci, for example. A large number of hypotheses were examined in the course of this study and in addition many of these tests were not independent of one another. In an effort to reduce the probability of a type one error, we took statistical significance to be at the more conservative P < 0.01 level. Hypotheses with a probability between 0.01 and 0.05 are noted but not discussed in detail. Setting the probability of rejecting the null hypothesis at 0.01 is, admittedly, somewhat arbitrary; however, it is based on a compromise between the overly conservative experiment wise error (1-( l-a)”) and not correcting at all for multiple comparisons. The latter option is suggested by the recent and provocative work by Rothman [1990] and is based on the premise that the “universal” null hypothesis of no effect of any of these loci on any of the lipid measures is likely not true. RESULTS

We have estimated the allele frequencies for the RFLPs of the apo B, apo AII, LDL-receptor genes, the apo AI-CIII-AIV gene cluster, and the apo E genotypes in

Genetic Variation Contributes to Response to Dietary Change

265

TABLE I. Estimated Allele Frequenciesin North Karelia, Finland Gene

Enzyme

ap0 B

PvuII-B Xbal Mspl EcoRI XmnI PstI Sstl PVUII-CIII PVUII-AIV StuI PVuIl NcoI ApaLI MspI E2 E4

apo AI-CIII-AIV

LDL receptor

apo A11 apo E

Presence or absence Less common of the cutting site ( / - ) allele frequency

+

+ + -

+ -

+ + + + -

-

0.14 0.38 0.06 0.15 0. I 1 [0.03 X3] 0.11 0.13 0.12 0.03 0.03 0.20 0.30 0.44 0.08 0.06 0.27

the sample of individuals from North Karelia (Table I). All of the polymorphisms were in Hardy-Weinberg equilibrium. The allele frequencies of the less common alleles are shown in Table I; the presence or absence of the cutting site is designated by + or - , respectively. Average lipid, lipoprotein, and apolipoprotein levels, as well as average body weight and BMI for the basal diet, the intervention diet, and the dietary change, are shown in Table 11. There was a significant difference in total, LDL, and HDL cholesterol, apo B and apo A1 levels when comparing basal and intervention dietary levels. However, body weight and BMI remained constant by design. For each trait the proportion of variance attributable to each marker locus, on the basal diet, the intervention diet, and their response is shown in Table I11 (a, b, and c, respectively). Statistically significant ( P < 0.01) associations are noted accordingly. It should be noted that because of the possible association between these candidate gene markers, the R2 values in Table I11 are not necessarily independent of one another. Since there was no significant association between LDL receptor RFLP genotypes and any of the traits in any of the treatments, the data are not presented. On the basal diet, genetic variation at the apo E gene locus and the variation associated with the XbaI RFLP of the apo B gene had significant effects. Isoform variation at the apo E gene locus explained 14.6% and 12.7% of the phenotypic variance in LDL and total cholesterol levels, respectively. The XbaI RFLP of the apo B gene was associated with differences in apo A1 levels explaining 8.1% of the phenotypic variation (Table IIIa). On the intervention diet, the associations observed on the baseline diet between LDL and total cholesterol levels and variation at the apo E gene locus were not significant. The difference between “A” (basal diet) and “B” (intervention diet) is represented by A and therefore the response to dietary change. The only significant effect on this response was due to variation at the apo B gene locus associated with the Msp I RFLP. This polymorphism explained 6.3%of the phenotypic variance in response of apo A1 levels to dietary change (Table IIIc).

266

Xuet al.

TABLE 11. Mean Lipid, Lipoprotein, and Apolipoprotein Measuremens ( ? S.D.) in Baseline “A” and Intervention “B” Diets and Due to Dietary Change A, Adjusted for Age and BMI and Gender Mean “A” Chol mmOl/L TG mmol1L LDL mmol1L apo B mgldl HDL mrnol1L apo A11 mgldl apo A1 mgldl body weight Kg BMI Kg/MZ

A“

“B”

Pb

6.2

f

1.1

5.0 f 0.9

1.2

k

0.7

< 0.001

1.2

*

0.8

1.1

f

0.5

0.01

&

0.4

0.60

4.5 2 1.1

3.6

?

0.8

1.0

&

0.7

< 0.001

116.7 & 28.8

102.0 2 27.8

14.7

* 20.1

< 0.001

1.4 f 0.3

1.2 i 0.3

0.2

?

0.2

< 0.001

&

6.2

0.03

43.4 i 6.3

44.7

?

8.6

158.0 ? 23.0

150.6

&

19.6

7.4 2 18.8

< 0.001

70.0

?

11.9

1.4 f 0.01

0.06

25.37

?

3.0

0.33 f 2.5

0. I6

71.4

&

11.2

25.7 2 3.1

-1.3

‘A = “A”-“B” adjusted for “A”. bPaired t-test for “A”-“B”.

For those polymorphisms that gave significant effects on the basal diet and the dietary response, average levels for each genotype are shown in Table IV. Individuals homozygous for the E4 allele had average mean LDL-(5.5 ? 0.09 mmol/l) and total cholesterol (7.2 2 1.2 mmol/l) levels higher than those individuals homozygous for the E3 allele (4.3 ? 0.8 mrnol/l and 5.9 2 0.8 mmol/l respectively), whereas individuals with an E2 allele had mean levels similar to individuals with the E3 allele. Individuals with the X + allele of the apo B XbaI RFLP had higher mean apo A1 levels (166.5 2 29.4 mg/dl) than individuals with theX- allele (149.8 -+ 19.1 mg/dl). Average apo A1 levels were lower on the low fat, high PIS diet. The common allele of the apo B Msp RFLP, M + , was associated with a larger reduction in apo A1 levels, i.e., a drop in apo A1 levels when changing from diet “A” to “B” of 9.4 ? 1.8 mg/dl, whereas the less common M - allele was associated with an increase in apo A1 levels in response to dietary change of 4.5 -+ 17.2 mg/dl. DISCUSSION

The ability to lower serum lipid, lipoprotein, and apolipoprotein levels by dietary change is of importance in the Western world where the incidence of coronary heart disease is high and associated with raised cholesterol levels [Kannel et al., 19711. Furthermore, it has been shown that saturated fat intake has a major influence on serum cholesterol levels [Keys et al., 19651. A number of dietary studies have demonstrated that a reduction in dietary fat is effective in lowering LDL cholesterol levels [Blum et al., 1977; Shaefer et al., 1981; Wolf and Grundy, 1983; Grundy et al., 19861, however, these diets are often associated with re-

0.5 0.5 0.1 I .4 0.2 0.7 0.4

2.5 8. I** 3.4 0.8 0.7 4.2 2.1

XbaI 0.05 0.2 0.4 0 1.5 I .5 1.5

EcoRI 0.8 0.7 0.4 0.2 0.6 0.4 0.7

MspI 4.0 6.1* 1.2 0.7 2.1 2.2 3.0

SstI

PVUII -CIII 1.1 0.4 5.0 I .2 1.5 2.7 2.3

PVUII -AIV

1.7 0.7 I .2 0.01 0.4 0.3 0.6

APOAI-CIII-AIV

0.3 0.3 5.9* 2.9 1.4 0.2 0.7

PstI 3.4 4.2 3.6 4.6 3.0 3.4 3.0

XmnI

.o

0.1 0.7 0.7 2.4 0.3

1

0.9

MspI

ApoAII

11.4* 2.0 2.7 3.7 14.6** 4.7 12.7**

Protein Polymorphism

ApoE

*P < 0.05.

Apo B Apo AI Apo A11 Trig LDL-C HDL-C T CHOL

0.8 2.7 2.7 0.7 0.1 3.4 0.1

0.2 0.8 1.1 4.2 0.4 1.3 0.9

~~

XbaI

PVuII-B

Apo B

MspI 1.5 1.6 0.1 1.2 0.6 0 0.5

EcoRI 0.8 0.1 0.1 0.1 0.3 0.1 0.2

PVUII -AN 1.3 1.5 0.1

1.8 1.2 1.3 2.1

SstI 2.4 2.6 1.1

0.1 0.3 1.1 0.4

PstI 1.1 0.1 0.7 0.3 2.8 0.2 2.4

PVUII -CIII 0.8 0.2 3.0 1 .o 0.1 2.6 0.8

APOAI-CIII-AIV

7.8 2.7 3.6 3.7 3.8 0.6 3.3

XmnI

0.9 1.2 1.7 1.3 2.6 6.7* 2.8

MspI

ApoAII

continued

7.9 0.7 3.8 11.8* 9.6 3.6 9.1

Protein Polymorphism

ApoE

(b) Proportion of variance (R2 x 100) of each lipid measure on the intervention diet (“B”) attributable to genetic variation at the candidate loci considered

*P < 0.05. **P < 0.01.

Apo B Apo A1 Apo AII Trig LDL-C HDL-C T CHOL

PVuII-B

Apo B

(a) Proportion of variance (R2 x 100) of each lipid measure on the basal diet (“A”) attributable to genetic variation at the candidate loci considered

TABLE 111. Proportion of Variance Due to Genetic Variation at the Candidate Loci on Lipid Traits for “A”, “B” and A

*P < 0.05. **P < 0.01.

LDL-C HDL-C T CHOL

Trig

Apo B Apo A1 Apo A11

2.0 3.7 5.4 3.8 2.2 2.0 4.3

PVUII-B

3.8 3.7 0.1 2.3 4.8 7.6* 4.8

XbaI

Apo B

0.1 0.04 0.2 0.1 0.2 3.5 1.7

EcoRI 0.5 6.3** 1.9 2.0 0.2 0.5 0.8

MspI 2.5 1.9 0.5 0.0006 3.2 1.8 3.9

SstI 0.0001 5.9* 2.3 0.8 0.3 0.3 0.8

PVUII -AIV

0.5 0.4 0.03 0.007 2.1 0.7 1.3

-CIII

PVUII

APOAI-CIII-AIV

0.2 0.2 1.8 2.9 0.7 1.7 0.7

PstI

1.2 0.5 1.2 3.7 2.4 2.6 2.3

XmnI

1.5

6.3* 1.3 2.2 0.5 1 .o 0.1

MspI

Apo A11

TABLE 111. Proportion of Variance Due to Genetic Variation at the Candidate Loci on Lipid Traits for “A”, “B” and A (continued) (c) Proportion of variance (R2 X 100)of the change (A) of each lipid measure attributable to genetic variation at the candidate loci considered

4.8 6.6 7.9 5.8

11.1*

3.4 3.1

ApoE Protein Polymorphism

Genetic Variation Contributes to Response to Dietary Change

269

TABLE IV. Mean Lipid Levels for RFLPs/Apo E Protein Polymorphism With Significant F Values From the ANOVAs Lipid

Polymorphism

Genotype

“A”

44

LDL-chol mmol/L

apoE

5.5 2 0.9

43 4.7

k

1.0

42

33

3.1

4.3 5 0.8

32 4.3

+

0.1

“A”

x+x+ “A” apo A1

rng/dl

apo B Xbal (N)

apo B MspI (N)

x-x-

166.5 2 29.4 161. 8 t 22.3 149.8 2 19.1 (16) (50) (41)

M+M+

A apo A1 mg/dl

x+x-

M+M-

9.4 ? 1.8 4 . 5 + 17.2 (87) (12)

M-M-

-

duced HDL cholesterol, which would theoretically have an adverse effect on cardiovascular disease. The North Karelia dietary intervention studies confirmed these findings, namely both a reduction in LDL and HDL cholesterol on a low fat, high P/S diet, and a reversal of these reductions on a return to the original diet. Dietary compliance was monitored by body weight measurements and total energy consumption calculated from Finnish food consumption tables. Dietary intervention was accomplished by a decrease in total fat from 38.9% of the total energy to 24.196, a reduction in dietary cholesterol from 537 mg to 302 mg, and an increase in the P/S from 0.15 to 1.18 on the intervention diet [Ehnholm et al., 19821. Interindividual variation in the response of lipid, lipoprotein, and apolipoprotein levels to dietary modification suggested a genetic component modulating this transition. Katan et al. [1988] have reported differences in response of serum cholesterol to changes in both dietary cholesterol and saturated fat among hypo- and hyper-responders. To date no study has investigated which genes are involved in modulating lipid levels during dietary change. Our study is the first to look at the contribution of variation at multiple specific candidate gene loci as a means of explaining the variation in response to change in dietary fat intake. Preliminary analysis of this North Karelia data looking at the effect of variation associated with either the apo B XbaI RFLP or the apo E gene locus on dietary change and the switchback to the basal diet, in this population, have been reported [Tikkanen et al., 1990a,b]. As expected from previous studies, variation at the apo E gene locus was associated with significant differences in LDL and total cholesterol levels on the basal diet. The effect of the E2 allele to lower and the E4 allele to raise LDL and total cholesterol levels has been well documented [Utermann et al., 1979; Sing and Davignon, 1985; Davignon et al., 19881. Contrary to the results observed for the basal diet, the effects

270

Xuet al. Cholesterol (rnrnol/l)

-!

7.5

5-

'

I 4.5 Basal

I

1

Low Fat

-

Apo €213

-I- Apo €313

-*-Apo €3/4

Apo €4/4

Apo A-I (mg/dl) 170 r

1

155

--------

I

145

1

-

Xba X+X+

-+.. Xba X+X-

4f Xba X-X-

160 + 155

150

t I

1 140 1 1 145

.~

Basal

Low Fat

Genetic Variation Contributes to Response to Dietary Change

271

of the apo E polymorphism on lipid levels was much smaller and did not reach statistical significance during the intervention diet. This difference indicates that the effects of apo E genotypes may be influenced by dietary factors [Tikkanen et al., 1990al. On the low fat intervention diet there was a slight reduction in the range of mean serum cholesterol levels among apo E genotypes resulting in a reduction in the overall variance. Compared to the other apo E isoforms, apo E2-containing lipoprotein particles bind less well to the hepatic apo E receptors, resulting in slower clearance of these particles and up-regulation of LDL receptors; thus the overall effect is to reduce serum cholesterol levels. Apo E4 containing postprandial lipoproteins are cleared more rapidly by the liver than lipoproteins with other apo E isoforms, resulting in hepatic cholesterol accumulation and down-regulation of LDL receptors, and thus raised serum cholesterol levels. Our results indicate that the effects of the apo E polymorphisms may be mediated in part by mechanisms that are affected by dietary fat intake [Kesaniemi et al., 19871. However the response of serum cholesterol levels to the dietary intervention was not significantly different among the apo E genotypes as shown in Figure la. Boerwinkle and Utermann [ 19881 have previously suggested that the response to dietary fat may be affected by apo E genotype. Furthermore, they proposed that the effects of variation at the apo E gene locus on serum cholesterol may be reduced in a population on a low fat intake [Boerwinkle and Utermann, 19881. Previous studies in the United Kingdom and Finland have shown an association between serum cholesterol levels and variation associated with the Xbai RFLP of the apo B gene [Law et al., 1986; Talmud et al., 1987; Aalto-Setala et al., 19881. In this study, similar but not significant results were observed (not shown). However, on the basal diet, sequence variation associated with the XbaI RFLP of the apo B gene locus is associated with differences in apo A1 levels. In our study those individuals homozygous for theX - allele had lower apo A1 levels ( 149.8 -+ 19mg/dl)than individuals homozygous for theX+ allele who had raised apo A1 levels (166.5 ? 29 mg/dl). This difference in mean apo A1 levels for different apo B Xbai genotypes on the basal diet and the intervention diet is represented in Figure lb. Individuals with the genotype X + X + and X+X- show a considerable drop in apo A1 levels when transferred to the low fat diet, whereas there is almost no change in apo A1 levels in individuals homozygous for the X- allele. Loss of the XbaI polymorphic site, although in the coding sequence of apo B, does not result in an amino acid change [Talmud et al., 19871. This suggests that any effect associated with the XbaI polymorphic site is due to linkage disequilibrium between that polymorphic site and sequence variation in the apo B gene. We propose a population association between the XbaI polymorphic site and a mutation in the apo B gene that, by influencing apo B-containing lipoprotein levels, indirectly modulates HDL and thus apo A1 levels. The metabolism of cholesterol is a dynamic process with a constant exchange of lipid and apoprotein components between lipoprotein classes. Thus constituents of VLDL Fig. 1. (a) Effect of dietary modification on serum cholesterol levels for different apo E genotypes on the basal diet and the low fat intervention diet. (b) Effect of dietary modification on serum apo A1 levels for different apo B XbaI RFLP genotypes on the basal diet and the low fat intervention diet. (e) Effect of dietary modification on serum apo A1 levels for different apo B MspI RFLP genotypes on the basal diet and the low fat intervention diet.

272

Xuet al.

are source materials for HDL [Fidge et al., 1980; Magill et al., 1982; Miller, 19871 and sequence variation in the apo B gene could affect VLDL metabolism and consequently HDL and apo A1 levels. This relationship between HDL metabolism and the metabolism of apo B-containing lipoproteins is enhanced in individuals on a high fat diet and reversed when transferred to a low fat diet [Nikkila, 19841. For the dietary response, the MspI RFLP of the apo B gene locus was associated with changes in apo A1 levels; a nearly significant effect of the XbaI RFLP on HDL levels was also noted. Average apo A1 levels for each of the apo B MspI genotypes on the basal diet and the low fat diet are shown in Figure lc. In individuals homozygous for the M + allele, apo A1 levels fell by 9.4mg/dl when the group was transferred from the high fat, low P/S diet to the low fat, high P/S diet. Conversely, apo A1 levels rose by 4.5 mg/dl in individuals heterozygous for the MspI RFLP when transferred to the low fat high P/S diet. The mechanisms of these associations are unknown, but they might be due to a direct effect of the polymorphism on apo B structure or function. The MspI RFLP is the result of an G to A base change, which causes the substitution of glutamine for arginine at amino acid residue 361 1. This polymorphism is associated with differences in cholesterol levels in some, but not all studies [Rajput-Williams et al., 1988; Xu et al., 19891. There is significant linkage disequilibrium between the MspI and XbaI polymorphic sites [Xu et al., 19891 and the possibility remains that the association between both these polymorphisms and apo A1 levels on the baseline diet (XbaI RFLP) and the A (MspI RFLP) may reflect the manifestation of the same effect due to this linkage disequilibrium. Our primary interest in this study was to test whether variation associated with RFLPs at any of the candidate gene loci affect response in serum lipid, lipoprotein, and apolipoprotein levels to dietary change. The change in lipid levels between the basal and low fat diets is a good measure of dietary response, since other environmental factors such as smoking and physical exercise remained relatively constant. None of the polymorphisms in the study were associated with significant differences in total or LDL cholesterol in response to dietary intervention. However, we report that interindividual differences in A for HDL and its constituent apoproteins are partially attributable to genetic variability at the candidate loci studied. The results presented here provide evidence supporting the role of genetic factors in both CHD risk and response to dietary manipulation. The data suggest that variation at both the apo B and apo AI-CII-AIV loci, as detected by RFLPs contribute to an individual’s change in apo A1 and HDL levels in response to dietary intervention. The identification of the sequence changes and the mechanism underlying these associations are not well understood and need further examination. Dietary intervention programmes directed at reducing CHD risk through dietary intervention should consider these genetic factors in both their design and implementation. ACKNOWLEDGMENTS

This work was supported by grants from the Sunley Research Centre, the British Heart Foundation (RG5), and a grant to E.B.(HL40613) from the National Institute of Health. C.F. Xu was supported by a Sino-British Friendship Scholarship, and M.J.T. received support from the Sigrid Juselius Foundation and the University of Helsinki.

Genetic Variation Contributesto Response to Dietary Change

273

REFERENCES Aalto-Seal K, Tikkanen MJ, Taskinen MR, Nieminen M, Holmberg P, Kontula K (1988): XbaI and c/g polymorphisms of the apolipoprotein B gene locus are associated with serum cholesterol and LDL cholesterol levels in Finland. Arteriosclerosis 74:47-54. Barni N, Talmud PJ, Carlsson P, Azoulay M, Darnfors C, Harding D, Weil D, Grzeschik KH, Bjursell G, Junien C, Williamson R, Humphries SE (1986): The isolation of genomic recombinants for the human apolipoprotein B gene, and the mapping of three common DNA polymorphisms of the gene-a useful marker for human chromosome 2. Hum Genet 73:313-319. Berg K (1986): DNA polymorphism at the apolipoprotein B locus is associated with lipoprotein level. Clin Genet 30515-520. Blum CB, Levy RI, Eisenberg S, Hall M, Goebalk RH, Berman M (1977): High density lipoprotein metabolism in Man. J Clin Invest 60:795-807. Boerwinkle E, Utermann G (1988): Simultaneous effects of the apolipoprotein E polymorphism on apolipoprotein E, apolipoprotein B and cholesterol metabolism. Am J Hum Genet 42: 104- 112. Darnfors C, Nilsson J, Protter AA, Carlsson P, Talmud PJ, Humphries SE, Whalstron J, Wikland 0, Bjursell G (1986): RFLPs for the human apo B gene Hind111 and PvuII. Nucl Acid Res 14:7135. Davignon J, Gregg RE, Sing CF (1988): Apolipoprotein E polymorphism and atherosclerosis. Atherosclerosis 8: 1-21. Demant T, Houlston RS, Caslake MJ, Series JJ, Shepherd J , Packard CJ, Humphries SE (1988): The catabolic rate of low density lipoprotein is influenced by variation in the apolipoprotein B gene. J Clin Invest 82:797-802. Ehnholm C, Huttunen JK, Pietinen P, Leine U , Mutanen M, Kostiainen E, Iacono JM, Dougherty R, Puska P (1984): Effect of a diet low in saturated fatty acids on plasma lipid, lipoprotein and HDL subfraction. Arteriosclerosis 4:265-269. Ehnholm C, Huttunen JK, Pietinen P, Leino U, Mutanen M, Kostiainen E, Pikkarainen J, Dougherty R, Iacono J, Puska P (1982): Effect of a diet on serum lipoproteins in a population with a high risk of coronary disease. N Engl J Med 307:850-855. Fidge N, Nestel P, Ishikawa T, Reardon M, Billington T (1980): Turnover of apolipoprotein A1 and A11 of high density lipoprotein and relationship to other lipoproteins in normal and hyperlipidaemic individuals. Metabolism 29:643-653. Friedewald WT, Levy RI, Fredrickson DS (1972): Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin Chem 18:499-502. Grundy SM, Nixon D, Whelan MG, Franklin L (1986): Comparison of three cholesterol-lowering diets in normolipidemic men. J Am Med Assoc 256:2351-2355. Havekes L, Kniejeff P de, Beisiegel U, Havinga J , Smit M, Klasen E (1987): A rapid micro-method for apolipoprotein E phenotyping directly in serum. J Lipid Res 28:455-463. Huttunen JK, Lansimies E, Voutilainen E, Ehnholm C, Hietanen E (1979): Effect of moderate physical exercise on serum lipoproteins: A controlled clinical trial with special reference to serum highdensity lipoproteins. Circulation 60: 1220- 1229. Kannel WB, Castelli WP, Gordon T, McNamara PM (1971): Serum cholesterol lipoproteins and risk of coronary heart disease: The Frarningham heart study. Ann Internal Med: 74 1-12. Katan MB, Berns MAM, Glatz JFC, Knuiman JT, Nobels A, de Vries JHM (1988): Congruence of individual responsiveness to dietary cholesterol and to saturated fats in humans. J Lipid Res 29:883-892. Kesaniemi YA, Ehnholm C, Miettinen TA (1987): Intestinal cholesterol absorption efficiency in man is related to apoprotein E phenotype. J Clin Invest 80578-58 1. Kessling AM, Taylor R, Temple A, Hutson J , Hildago A and Humphries SE (1988a): A PvuII polymorphism in the 5' flanking region of the apolipoprotein AIV gene: Its use to study genetic variation determining serum lipid and apolipoprotein concentrations. Hum Genet 78:237-239. Kessling AM, Rajput-Williams J, Bainton D, Scott J, Miller N, Baker I, Humphries SE (1988b): DNA polymorphisms of the apolipoprotein A11 and AI-CIII-AIV genes: A study in men selected for differences in high density lipoprotein cholesterol concentration. Am J Hum Genet 42:458-467. Keys A, Anderson JT, Grande F (1965): Serum cholesterol response to changes in the diet. Metabolism 14:747-765. Kostner GM ( 1976): Enzymatic determination of cholesterol in high-density lipoprotein fractions prepared by polyanion precipitation. Clin Chem 22:695. Kunkel LM, Smith KD, Bayer SH, Borgankar DS, Wachtel SO, Miller OE, Brey WR,Jones HW, Roury

274

Xuet al.

EM (1977): Analysis of human y-chromosome specific reiterated DNA in chromosome variants. Proc Natl Acad Sci USA 74:1245-1249. Kuusi T, Ehnholm C, Huttunen JK, Kostianinen E, Pietenen P, Leino H, Uusitalo U, Nikkasi T, Iacono JM, Puska P (1985): Concentration and composition of serum lipoproteins during a low-fat diet at two levels of polyunsaturated fat. J Lipid Res 26:360-367. Law A, Powell LM, Brunt H, Altman DG, Rajput J, Wallis SC, Pease RJ, Priestley LM, Scott J, Miller GJ, Miller NE (1986): Common DNA polymorphism within coding sequence of apolipoprotein B gene associated with altered lipid levels. Lancet i: 1301-1303. Lukka M, Metso J, Ehnholm C (1988): Apolipoprotein A-IV polymorphism in the Finnish population: Gene frequencies and description of a rare allele. Human Hered 38:359-362. Magill P, Rao SN, Miller NE, Nicoll A, Brunzell J, Hilaire JST, Lewis B (1982): Relationships between the metabolism of high-density and very-low-density lipoproteins in man: Studies of apolipoprotein kinetics and adipose tissue lipoprotein lipase activity. Eur J Clin Invest 12:113-120. Miller NE (1987): Metabolic determinants of plasma high-density lipoprotein concentration in humans. In Malmendier CL, Alaupovic P (eds): “Lipoproteins and Atherosclerosis. ” New York: Plenum. pp 111-115. Nikkila EA (1984): HDL in relation to the metabolism of triglyceride-rich lipoproteins. In Miller NE and Miller GJ (eds): “Clinical and Metabolic Aspects of High-Density Lipoproteins.” Amsterdam: Elsevier Science Publishers BV, pp 217-24.5. Pederson JC, Berg K (1989): Interaction between low density lipoprotein receptor (LDLR) and apolipoprotein E (apo E) alleles contributes to normal variation in lipid level Clin Genet 35:331-337. Rajput-Williams J, Knott TJ, Wallis SC, Sweetnam P, Yamell J, Cox N, Bell GI, Miller NE, Scott J (1988): Variation of apolipoprotein-B gene is associated with obesity, high blood cholesterol levels and increased risk of coronary heart disease. Lancet ii: 1442-1446. Rees A, Stocks J, Sharpe CR, Vella MA, Shoulders CC, Katz J, Jowlett NI, Baralle FE,Galton DJ (1985): DNA polymorphism in the apo AI-CIII gene cluster: Association with hyper-triglyceridaemia. J Clin Invest 76: 1090-1095. Rothman KJ (1990): No adjustments are needed for multiple comparisons. Epidemiology 1:43-46. Schaefer EJ, Levy RI, Emst ND, Van Sant FD, Brewer HB Jr (1981): The effects of low cholesterol, high polyunsaturated fat, and low fat diets on plasma lipid and lipoprotein cholesterol levels in normal and hyper-cholesterolemic subjects. Am J Clin Nutr 34: 1758- 1762. Scott JM, Knott TJ, Priestley LM, Robertson ME, Mann DV, Kostner G, Miller GJ, Miller NE (1985): High-density lipoprotein composition is altered by a common DNA polymorphism adjacent to apoprotein A11 gene in man. Lancet i:771-773. Shaper AG, Pocock SJ (1985): British blood cholesterol values and the American consensus. British Medical Journal 291:480-483. Sing C, Boerwinkle E (1987): Genetic architecture of inter-individual variability in apolipoprotein, lipoprotein and lipid phenotypes. In Bock G and Collins GM (eds): “Molecular approaches to human polygenic disease.” Ciba Foundation Symposium. New York: John Wiley & Sons, pp 99-127. Sing CF, Davignon J (1985): Role of apoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am J Hum Genet 37:268-285. Talmud PJ, Bami N, Kessling AM, Carlsson P, Damfors C, Bjorsell G , Galton PJ, Wynn V, Humphries SE (1987): Apolipoprotein B gene variants are involved in the determination of serum cholesterol levels: A study in normo- and hyperlipidaemic individuals. Atherosclerosis 67:8 1-89. Talmud PJ, Lloyd JK, Muller DPR, Collins PR, Scott J, Humphries SE (1988): Genetic evidence that the apolipoprotein B gene is not involved in abetalipoproteinaemia. J Clin Invest 82:1803-1806. Tikkanen MJ, Huttunen JK, Ehnholm C, Pietinen P (1990a): Apolipoprotein E4 homozygosity predisposes to serum cholesterol elevation during high-fat diet. Arteriosclerosis 10:285-288. Tikkanen MJ, Xu CF, Hamalainen T, Talmud P, Sama S , Huttunen JK, Pietinen P, Humphries SE (1990b): XbaI polymorphism of the apolipoprotein B gene influences plasma lipid response to diet intervention. Clin Genet 37:327-334. Utermann G , Pevin N, Steinmetz A (1979): Polymorphism of apolipoprotein E. I11 Effect of a single polymorphic gene locus on plasma lipid levels in Man. Clin Genet 15:63-72. Weisgraber KH, Innerarity TL, Mahley RW (1982): Abnormal lipoprotein receptor-binding activity of the human E apoprotein due to cysteine-arginine interchange at a single site. J Biol Chem 257: 15 18- 1521.

Genetic Variation Contributes to Response to Dietary Change

275

Wile DB, Barbir M, Gallagher J, Myant NB, Richie CD, Thompson GR, Humphries SE (1989): Apolipoprotein A1 gene polymorphisms: Frequency in patients with coronary artery disease and healthy controls associated with serum apo A1 and HDL-C concentration. Atherosclerosis 78:9-18. Wolf RN, Grundy SM (1983): Influencing of exchanging carbohydrate for saturated fatty acids on plasma lipids and lipoproteins in man. J Nutr 113: 1521-1528. Xu C-F, Nanjee MN, Savill J, Talmud PJ, Angelic0 F, Del Ben M, Antonini R, Mazzarella B, Miller N , Humphries SE (1990): Variation at the apolipoprotein (apo) AI-CIII-AIV gene cluster and apo B gene loci is involved in the determination of lipid and lipoprotein levels in Italian children. Am J Hum Genet (in press). Xu C-F, Nanjee N, Tikkanen MJ, Huttunen JK, Pietinen P, Butler R, Anjelico F, Del Ben M, Mazzarella B, Antonio R, Miller NE, Humphries S, Talmud PJ (1989): Apolipoprotein B amino acid substitution from arginine to glutamine creates the Ag(Ni) epitope: the polymorphism is not associated with differences in serum Cholesterol and apolipoprotein levels. Hum Genet 82:323-326.

Edited by D.C. Rao