Atherosclerosis, 9 1 (1991) 267-215 ‘c 1991 Elsevier Scientific Publishers

ATHERO

267 Ireland,

Ltd. All rights reserved

0021.9150/91/$03.50

04744

Apolipoprotein B gene polymorphisms are associated with lipid levels in men of South Asian descent H-H. Renges

‘, D.B. Wile ‘,*, P.M. McKeigue

‘, M.G. Marmot

2 and SE. Humphries



’Artenul Diseases Research Group, Char@ Cross Sunley Research Cenire, Lurgun ALtenue, Hammersmith, London W6 NL W (U.K.), and -‘Deparrment of Epidemiology and Population Sciences, London School of Hygiene and Tropical Medicine, Keppel Streer, London WCIE 7HT (U.K.) (Received 12 June, 19911 (Revised, received 9 September, 1991) (Accepted 23 September, 19911

Summary Three polymorphic sites of the apolipoprotein B gene - the insertion/deletion signal peptide, XbaI and EcoRI sites - were examined in a sample of 107 healthy men and in 46 men with evidence of coronary heart disease selected from a large population survey of South Asians aged 40-69 in London, U.K. There were no significant differences in allele frequencies between cases and controls. Frequencies of the ins (insertion) and X - (absence of XbaI cutting site) alleles were higher in South Asians than in Europeans studied previously (South Asians versus Europeans ins: 0.80 vs. 0.68, P < 0.025; X - : 0.71 vs. 0.47-0.56, P < 0.001). The de1 allele was associated with higher levels of total cholesterol (P < 0.05) and the X + allele with lower levels of HDL cholesterol (P < 0.05), and thus both polymorphisms were associated with differences in the ratio of HDL cholesterol to total cholesterol (ins/def, P < 0.01; J&I, P < 0.001). Mean waist-hip girth ratio was lower in the 10 men homozygous for the X + allele than in the 42 men with X - /X + and 55 men with X - /X - genotypes; the means (+SEM) were 0.92 + 0.02, 0.97 & 0.01 and 0.96 & 0.01 respectively (P = 0.03). These data suggest that genetic variation in linkage disequilibrium with the .%a1 and ins/def polymorphisms of the apo B gene contributes to the determination of total cholesterol and HDL cholesterol levels and possibly to obesity in South Asians.

Key words:

Apolipoprotein South Asians

B gene; RFLPs; (Punjabi Sikhs)

Total cholesterol;

* Presenr address: Department mersmith

Hospital,

of Chemical Pathology, HamDu Cane Rd, London W12 ONN. U.K.

Correspondence to: Dr. Helmut-H. Renges, University College London, Department of Medicine, The Rayne Institute, University Street, London WClE 655, U.K. Fax: 071.3809837; Tel.: 071-3879300.

HDL cholesterol;

Coronary

heart disease;

Introduction Mortality from coronary heart disease (CHD) in the UK is higher in men and women born in South Asia (India, Pakistan and Bangladesh) than in the native British population [l], as is the prevalence of non-insulin-dependent diabetes [2].

268 This high CHD risk is unexplained by established risk factors for the disease including smoking, plasma cholesterol and haemostatic factors [2]. Similar findings have been reported for other overseas South Asian populations; the consistency of the high CHD risk in urban populations of South Asian descent around the world, even in populations who migrated several generations before, suggests that genetic factors may be involved [3]. In several populations of European and Asian origin genetic variation in the apolipoprotein (ape) B gene has been associated with differences in lipid levels [4-91 and with risk of CHD [9-131. In particular, the XbaI polymorphism in exon 26 of the apo B gene has been found to be associated with variation of cholesterol levels in normal individuals [4-61 and in patients with familial hypercholesterolaemia [7]. This polymorphism has also been found to be associated with obesity [14,15] and with CHD [10,11,13], although the underlying mechanism is unknown. Associations between the EcoRI polymorphism in exon 29 and CHD have been reported [9,11,12]. Recently the insertion/deletion (ins/&l) polymorphism in the signal peptide region of the apo B gene [161 has been found to be associated with triglyceride levels in a study of healthy Finnish individuals [S]. The objectives of this study were to test whether these associations were also present in South Asians and whether this might yield clues to the reasons for the higher CHD risk in this group. We examined the frequency distribution at three polymorphic sites in the apo B gene in a population sample of 107 men of Punjabi Sikh origin and in 46 men with coronary heart disease from the same population. We also compared the allele frequencies in South Asians with those previously reported in European populations. Methods Sample selection For this study 153 men were selected from a sample of 1421 South Asian men aged 40-69 who participated in a population survey. The methods of the main study have been described in detail elsewhere [17]. The sample was assembled from four factories and from the lists of general practi-

tioners in west London. 51% of the South Asian men in the original sample were Sikhs, who originate from Punjab in northern India. Participants completed a self-administered questionnaire including medical history and the WHO chest pain questionnaire [18], then attended the field station where the questionnaire was checked by a bilingual interviewer. Weight, height, waist and hip girths were measured as described previously [ 171. A 12-lead resting electrocardiogram was coded according to the Minnesota system [18] by two observers in duplicate. ‘Probable coronary heart disease’ was defined as the presence of major Q waves or left bundle branch block (Minnesota codes l-l, 1-2 or 7-l-l) on ECG, or a doctor’s diagnosis of coronary heart attack or angina, or a positive WHO chest pain questionnaire score. The first 46 South Asian men who met these criteria were selected for this study, together with 107 randomly chosen Sikh men without evidence of coronary heart disease, diabetes or impaired glucose tolerance. The CHD sample consisted mainly of Punjabi individuals (39 out of 46). Laboratory measurements Plasma glucose, triglycerides and serum insulin were measured in the fasting state and 2 h after a 75-g glucose load. Plasma glucose, triglycerides and cholesterol were determined by calorimetric techniques using glucose oxidase, glycerol phosphate oxidase and cholesterol oxidase respectively as described previously [171. HDL was separated by dextran sulphate/magnesium precipitation. Serum insulin was measured by radioimmunoassay as described previously [17]. DNA analysis DNA isolation from venous blood samples was performed using Triton Xl00 lysis, Proteinase K digestion and phenol extraction [ 191. Genotypes at the ins/del, XbaI and EcoRI polymorphic sites in the apo B gene were determined using the polymerase chain reaction [20]. Amplification was carried out in a final volume of 50 ~1 containing 200 ng of genomic DNA, 250 ng of each primer, 10% DMSO and 0.5 units of Tuq DNA polymerase (Perkin Elmer/Cetus, Connecticut, U.S.A.) in the reaction buffer recommended by the manufacturer, using a Cambio Intelligent

269 Heating Block (Cambio, Cambridge, U.K.). The reaction mix was topped with 50 ~1 of liquid paraffin to prevent evaporation. PCR primers were obtained from Oswel (University of Edinburgh, U.K.). Sequences are given in Fig. 1. The insertion/deletion site in exon 1 was analyzed by amplifying a fragment of 93/84 base pairs (35 cycles: 92°C for 1 min, 65°C for 1.5 min) and separating the products on a 2% agarose + 6% NuSieve agarose gel (10 ~1 of PCR product) [8]. The XbaI site was analyzed by amplifying a 710 bp fragment (35 cycles: 92°C for 1 min, 58°C for 5 min). Overnight digestion with 20 U XbaI resulted in fragments of 433 bp and 277 bp in the presence of the cutting site (X + allele). Gel electrophoresis of 10 ~1 of the PCR product was performed in 1.5% agarose gels. For analysis of the EcoRI site, a fragment of 480 bp was amplified (35 cycles: 92°C for 1 min, 58°C for 5 min). Overnight digestion with 9 U EcoRI gave 253 bp and 227 bp fragments in the presence of the cutting site (R + J. Gel electrophoresis of 10 ~1 of the PCR product was performed in 2% agarose gels. Statistical analysis Ethnic and case-control differences in genotype and allele frequencies (gene-counting) were tested for significance by chi-square tests. Linkage disequilibrium between pairs of allelic sites was estimated using the approach described by

TABLE

n

CHD sample

Results The levels of lipids in the sample of healthy individuals examined here were similar to those reported in the whole sample [17]. In the group of CHD patients the mean age was five years higher than that of the healthy individuals. The mean total cholesterol was higher, the mean HDL cholesterol lower and the HDL cholesterol to total cholesterol ratio significantly lower as compared to the controls.

I

GENOTYPE DISTRIBUTION TROL AND CHD SAMPLES

Control sample

Weir and Cockerham [211 and computed by a FORTRAN program kindly provided by Dr. B.S. Weir [22]. A chi-square test as described by Weir and Cockerham (1 degree of freedom) was performed to test for statistical significance [21]. Body mass index (BMI), plasma triglyceride and serum insulin levels were log.-transformed before analysis. Between-genotype differences in means for continuous variables were tested for significance by analysis of variance in least-squares regression models, controlling for age, BMI and waist-hip ratio, and with genotype as a categoric variable. In testing for effects of genotype on lipid and insulin levels, age, body mass index and waist-hip ratio were included in the models as covariates. Statistical significance was considered to be at the 0.05 level. Analyses were performed using the software packages SAS and SPSS/PC +.

OF APO

B SIGNAL

Ins/Del

107

46

Ins/Del,

XhaI

AND

XbUI

II

ID

DD

71 (0.66)

30 (0.28)

6 (0.06)

11 (0.24)

_

35 (0.76)

PEPTIDE

Rel. Freq. I

EcoRI

POLYMORPHISMS

IN CON-

l?coRI

x-x-

x-x+

x+x+

Rel. Freq. x-

R+R+

R+R-

R-R-

0.80

55 (0.50)

20239)

to909,

0.71

Fi.79)

Fd.20)

(0.01)

0.89

0.88

:;.67)

12 (0.26)

(0307)

0.80

::78)

10 (0.22)

-

0.89

Rel. Freq. R+

1

CHD, coronary heart disease; n, number of individuals; Ins, I, signal peptide insertion allele; Del, D, signal peptide deletion allele; X. XhaI; R, EcoRI; +, - referring to presence or absence of cutting site for specific restriction enzyme; Rel. Freq., relative frequency. Values in upper row represent number of individuals, values in lower row represent frequencies. There are no significant differences (P < 0.05) between the distributions in the two groups at any of the three sites as estimated by chi-square test.

270

Genotypes were determined at three polymorphic sites (ins/&l, XbaI and EcoRI) in the apo B gene and typical results are shown in Fig. 1, which also indicates the relative locations of the three sites in the apo B gene. The frequency distributions of the genotypes within the CHD and control groups are shown in Table 1. The frequency of the ins and the X - alleles were higher in the CHD cases than in the controls but

Apo

(al

B gene

these differences did not reach statistical significance for either site. Frequencies of the ins and X - alleles were significantly higher in South Asians than in the European populations studied previously (Table 2). Allele frequency at the EcoRI site was not significantly different in South Asians and Europeans. Estimation of linkage disequilibrium revealed significant diallelic linkage disequilibrium in both CHD and control samples

polymorphisms

Xbal

Ins/Del

EcoRl

Ihlb-a I I II I I I I II II I I II II 111 5'

-

3'

1 kb

Xbal

(b)Ins/Del

--

II ID DD

-/ -

93 84

+-

EcoRl ++

++

+-

--

-

710

-

- 480

-

433

-

277

/ -

253 227

Fig. 1. Relative location of apo B gene polymorphisms and DNA analysis of the ins/del, XbaI and EcoRI sites. (a) Proportional map of the apo B gene, exon and intron sizes according to [30], exons are shown in black; the location of the polymorphisms is EcoRI [32] in exon 29 indicated: ins/de1 in exon 1 [16] (amino acids (aa): Leu-Ala-Leu _ ih,_ i4 I, XbaI [31] in exon 26 (aa: Thr,,,,), + Lys); a indicates the location of the putative receptor binding domain of the apo B gene [33,34] in exon 26 (aa: (aa: Glu,,,, 3147-3157, 3345-3381). (b) Photographs of gel electrophoresis examples to illustrate genotyping at the three sites. Sizes of PCR products (ins/de/) and of restriction fragments of PCR products (XbaI, EcoRI) are given in basepairs. Sequences of PCR primers [351 are: ins/de1 5’-primer: 5 '-CAGCTGGCGATGGACCCGCCGA-3', 3’-primer: 5'-ACCGGCCCTGGCGCCCGCCAGCA-3', XbaI: 5’-primer: 5'-GGAGACTATTCAGAAGCTAA-3', 3'-primer:5'-GAAGAGCCTGAAGACTGACT-3',

EcoRI:5’-primer: 3’.primer:

5'-CTGAGAGAAGTGTCTTCGAAG-3', 5'-CTCGAAAGGAAGTGTAATCAC-3'.

271 between the ins/de1 and XbaI sites (A,, = 0.12, /$ = 40.8, P < 0.001 in the control group) and between the XbaI and EcoRI sites (A,, = -0.03, x2 = 6.2, P < 0.025) ( see [23] for a more detailed evaluation of the genetic relationships between apo B polymorphisms in these samples). Association of plasma cholesterol, triglycerides and serum insulin with genotype was examined in the sample of 107 Sikh men without evidence of CHD. Significant associations were detected with variation at the ins/de1 and XbaI sites but not with variation at the EcoRI site (Table 3). In particular, the de1 allele was significantly associated with increased levels of total cholesterol and there was a trend in the same direction for the X + allele (P < 0.1). Conversely, presence of the X + allele was associated with lower mean HDL cholesterol levels, whilst the de1 allele was associated with a trend towards lower HDL cholesterol levels. For both sites the association with the ratio of HDL cholesterol to total cholesterol was highly significant (P < 0.001 for XbaI and P < 0.01 for ins/del). The adjusted means for cholesterol levels in the three genotype groups are presented in Table 4. Exclusion of data from two individuals with total cholesterol greater than 9.0 mmol/l increased the level of significance of the TABLE

TABLE

3

PERCENTAGE OF SAMPLE VARIANCE (R* x 100) OF ADJUSTED LIPID LEVELS EXPLAINED BY APO B GENE POLYMORPHISMS IN THE RANDOM CONTROL SAMPLE (n = 107) Ins/Del

R* x 100 6.1 * 2.1

Total cholesterol HDL cholesterol HDL cholesterol/ Total cholesterol Fasting triglycerides

9.8 ** 1.5

Xhu I ~ R’ x 100

Eco RI ___ R* x 100

4.5 + 7.2 *

0.6 3.3

15.x *** 4.2

5.1 1.o

* P < 0.05: compare Table 4 ** P < 0.01: compare Table 4 *** P < 0.001: compare Table 4 + P < 0.1. n, number of individuals; all measurements were adjusted by multiple linear regression to correct for variation in age, body mass index and waist-hip ratio. R* X 100 values were obtained from multiple linear regression. P values refer to F-tests.

association with cholesterol levels for the ins/de1 site but not for the XbaI site. For both ins/del and XbaI polymorphisms there was a weak association with triglyceride levels, the rare alleles de1 and X + being associated with higher triglyceride levels, although these differences were not signifi-

2

RELATIVE QUENCIES

FREQUENCIES OF APO B POLYMORPHISMS IN SAMPLES FROM EUROPEAN POPULATIONS

Ins South Asians: (random control

sample)

European samples : Finland (healthy male subjects) (Xu et al. [8]) U.K. (control sample) (Myant et al. [ll]) Austria (control sample) (Paulweber et al. [Y])

107

0.80

55

0.68

ASIANS

COMPARED

XbaI

Ins/Del

n

IN SOUTH

P

X-

FRE-

EcoRI P

0.71

< 0.025

TO REPORTED

R+

P

0.80

0.56

< 0.01

0.85

NS

146

_

_

0.47

< 0.001

0.85

NS

118

_

_

0.52

< 0.001

0.86

NS

CHD. coronary heart disease; n, number of individuals; Ins, signal peptide insertion allele; Del, signal peptide deletion allele; X, XbaI; R, EcoRI; +, - referring to presence or absence of cutting site for specific restriction enzyme; P value for significance of the difference between allele frequencies in the European sample compared to the South Asian sample estimated by chi-square test (gene-counting); NS, not significant.

272 cant. No associations with insulin levels were seen for either the ins/de1 or Xl%1 polymorphisms. There was a significant association between variation at the XbaI site and obesity, as measured by body mass index and waist-hip girth ratio. Geometric means for body mass index in individuals with the X-X -, X-X + and X + X + genotypes were respectively 25.6, 26.3 and 24.0 kg. mP2 (P = 0.06). For waist-hip girth ratio the respective means (+ SEMI were 0.96 &0.01, 0.97 f 0.01 and 0.92 + 0.02 (P = 0.031, the higher values indicating increased central obesity. In the CHD sample genotypes for all three polymorphisms were associated with similar effects on lipid traits as seen in the sample of healthy men, but none of the differences reached statistical significance. TABLE

Discussion

If apo B contributes to the increased prevalence of CHD seen in South Asians compared to western samples, one would expect differences in allele frequencies at the apo B gene locus between cases and controls and between South Asians and Europeans. In our sample of South Asians no significant differences in allele frequencies at the three investigated sites were seen, although the X - and ins alleles were more frequent in the CHD group compared to the random control sample. This corresponds with findings in samples from populations of European origin and with a study in Sri Lankan individuals, that reported an increased frequency of the X allele in CHD patients [10,11,13], but does not

4

VARIATION AMONGST

IN ADJUSTED THE DIFFERENT

CHOLESTEROL GENOTYPES

AND TRIGLYCERIDE AT THE APO B GENE

LEVELS IN THE RANDOM CONTROL Ins/Del, XbnI AND EcoRI LOCI

SAMPLE

Means (+ SEM or 95% CI) for genotype Ins/Del n Total cholesterol HDL cholesterol

(mmol/l) (mmol/ll

HDL cholesterol/ Total cholesterol Triglycerides (mmol/l) XbllI n Total cholesterol

(mmol/l)

HDL cholesterol (mmol/l) HDL cholesterol/ Total cholesterol Triglycerides (mmol/ll EcoRI n Total cholesterol HDL cholesterol

(mmol/l) (mmol/l)

HDL cholesterol/ Total cholesterol Triglycerides (mmol/l)

II 71 5.73 + 0.12 1.34 f 0.05 0.24 f 0.01 1.67 (1.47-1.90) x-x55 5.68+0.12 1.39+0.05 0.25 f 0.01 1.62 (1.40-1.88) R+R+ 85 5.93 f 0.13 1.26 k 0.04 0.22 + 0.01 1.76 (1.56-1.99)

ID 30 6.19kO.25 1.21 kO.07 0.20~0.01 1.88 (1.51-2.32) x-x+ 42 6.14 f 0.21 1.21* 0.05 0.20 f 0.01 1.79 (1.58-2.03) R+R21 5.87 f 0.24 1.40+0.09 0.24 f 0.02 1.75 (1.40-2.19)

DD 6 6.71 f 0.53 * 1.19f0.17 0.18+0.02 ** 2.10 (1.11-3.98) x+x+ 10 6.24 + 0.45 + 1.12+0.14 * 0.18+0.02 *** 2.41 (1.28-4.51) R-R1 5.01 1.73 0.34 + 1.01

* P < 0.05; ** P < 0.01; *** P < 0.001; + P

Apolipoprotein B gene polymorphisms are associated with lipid levels in men of South Asian descent.

Three polymorphic sites of the apolipoprotein B gene - the insertion/deletion signal peptide, XbaI and EcoRI sites - were examined in a sample of 107 ...
857KB Sizes 0 Downloads 0 Views