Genetic Epidemiology 18:203–216 (2000)

© 2000 Wiley-Liss, Inc.

Frequency and Allelic Association ofCommon Variants in the LipoproteinLipase Gene in Different Ethnic Groups:The Wandsworth Heart and Stroke Study

Stephen Hall, 1 Philippa J. Talmud, 1 Derek G. Cook, 2 Paul D. Wicks, 2,3

Michael J. Rothwell, 2,3 Pasquale Strazzullo, 4 Giuseppe A. Sagnella, 3

and Francesco P. Cappuccio 3*

1Division of Cardiovascular Genetics, Department of Medicine, UniversityCollege London, London, United Kingdom

2Department of Public Health Sciences, St. George’s Hospital Medical School,London, United Kingdom

3Blood Pressure Unit, Department of Medicine, St. George’s Hospital MedicalSchool, London, United Kingdom

4Department of Clinical and Experimental Medicine, Federico II University ofNaples Medical School, Naples, Italy

The lower serum triglyceride (Tg), higher high density cholesterol (HDL-C) lev-els and low coronary heart disease (CHD) mortality in black populations, con-trast with that in whites. By comparison, South Asian populations display a highermortality from CHD associated with increased Tg and low HDL-C levels. Lipo-protein lipase (LPL) plays a major role in Tg metabolism. To determine if varia-tion in the LPL gene contributes to the differences in lipid levels, we studied thefrequencies and allelic associations of five common variants in the lipoproteinlipase (LPL) gene (-93T/G, D9N, N291S, S447X, and the HinddIII RFLP inintron 8) with serum Tg and HDL-cholesterol concentrations in population samples

Contract grant sponsor: British Heart Foundation; Contract grant sponsor: Wandsworth Health Author-ity; Contract grant sponsor: South Thames Regional Health Authority; Contract grant sponsor: NHSR&D Directorate; Contract grant sponsor: British Diabetic Association; Contract grant sponsor: StrokeAssociation.

*Correspondence to: F.P. Cappuccio, MD, FRCP, Wandsworth Heart & Stroke Study, Blood PressureUnit, Department of Medicine, St. George’s Hospital Medical School, Cranmer Terrace, London SW17ORE, UK. E-mail: f.cappuccio@sghms.ac.uk

Received for publication 17 November 1998; revision accepted 20 June 1999

204 Hall et al.

of middle-aged men and women of whites, South Asians, and blacks of Africanorigin co-resident in South London. Significantly higher frequencies of the H–

(P < 0.00001), N9 (P < 0.001), and –93G (P < 10–10) alleles were seen in blackscompared to the other two groups. Allelic association between –93G and N9,and H+ and X447 was strong in all three groups. However, no association wasobserved between serum Tg and HDL-cholesterol concentrations and these vari-ants in the three ethnic groups. A single common polymorphism in the LPL geneis unlikely to account for the differences in fasting serum Tg in populations ofdifferent ethnic background. The importance of the differences in frequenciesand the mechanism(s) whereby these may contribute towards a beneficial LPLgenotype in black populations remain to be determined. Genet. Epidemiol.18:203–216, 2000. © 2000 Wiley-Liss, Inc.

Key words: lipoprotein lipase variants; ethnicity; epidemiology; lipids

INTRODUCTION

Lipoprotein lipase (LPL) is the rate-limiting enzyme hydrolysing core lipopro-tein triglycerides (Tg) and releasing free fatty acids for energy and storage by muscleand adipose tissue, respectively [Goldberg, 1996]. There is strong evidence that plasmaTg concentration is a risk factor for cardiovascular disease [Hokanson and Auston,1996] and post-heparin lipase activity correlates well with fasting plasma Tg levelsin normal subjects [Taskinen, 1987; Semenkovich et al., 1989].

Black populations, whether from the Caribbean or from West Africa, displayreduced mortality from coronary heart disease (CHD) [Cappuccio, 1997] and havelow-fasting plasma Tg levels and high HDL-C levels [Summerson et al., 1992; Novicket al., 1994; Proudler et al., 1996; Cappuccio et al., 1998]. In contrast, a higher riskof CHD is associated with raised plasma Tg and lower HDL-C levels as well as ahigher prevalence of diabetes [Cappuccio et al., 1997] in South Asian immigrants[Cappuccio, 1997; Cappuccio et al., 1998] compared to local ethnic groups [OPCS,1990; Rao and White, 1993; Cappuccio et al., 1997].

Several common genetic variants in LPL have been reported, either affectingLPL gene expression [Hall et al., 1997], creating amino acid substitutions or intronicvariants [reviewed by Fisher et al., 1997]. The LPL T-93G promoter variant has beenshown to affect promoter function; compared to the wildtype –93T, –93G containingLPL promoter leads to an overall increase in promoter function [Hall et al., 1997]and is associated with lower plasma Tg levels [Hall et al., 1997; Ehrenborg et al.,1997] and enhanced postprandial Tg clearance [Talmud et al., 1998]. The asparticacid to asparagine change at residue 9 (D9N) produces a secretion-defective proteinwith apparently normal specific activity [Mailly et al., 1995] and is associated withsignificantly higher plasma Tg and lower HDL-cholesterol concentrations in carriersversus noncarriers [Mailly et al., 1996; Gerdes et al., 1997]. The serine 447 stopsubstitution (S447X), which truncates the LPL protein by two amino acids, is associ-ated with lower plasma Tg levels [Jemaa et al., 1995; Zhang et al., 1995; Gagné etal., 1996] and a 50% increase in HDL-cholesterol [Thorn et al., 1998]. Heterozy-gotes for the asparagine to serine change at residue 291 (N291S), a variant withdecreased dimer stability in vitro [Reymer et al., 1995], have increased plasma Tglevels [Fisher et al., 1995] and reduced LPL activity [Reymer et al., 1995; Syvänne

LPL Variants in Different Ethnic Groups 205

et al., 1996] and S291 is very strongly associated with a 10–15% reduction in HDL-cholesterol [Gerdes et al., 1997; Reymer et al., 1995; Wittrup et al., 1997]. The pres-ence of a HindIII restriction enzyme cutting site in intron 8 (H+) is associated withraised Tg concentrations [Ahn et al., 1993; Mattu et al., 1994; Jemaa et al., 1997],raised HDL-cholesterol [Heizmann et al., 1991], and a predisposition to myocardialinfarction [Jemaa et al., 1995] and there is strong evidence that the functional changein allelic association with this site is the X447 site [Humphries et al., 1998].

The aim of the present study was to determine whether there was a geneticbasis for the known ethnic differences in lipid levels by examining the frequencyand lipid association of common LPL variants in unselected co-resident populationsamples of whites, South Asians, and black people of African origin.

SUBJECTS AND METHODSPopulation Sample

The population sample has been previously described [Cappuccio et al., 1997,1998]. In brief, a random sample of men and women aged 40–59 years was obtainedfrom age- and sex-registers of General Practitioners in a defined area of South Lon-don. They belong to the three main ethnic groups living in the area [Cappuccio et al.,1997, 1998]. Fieldwork was undertaken from March 1994 to July 1996. Ethnic groupwas recorded at the time of interview, based on the answers to a combination ofquestions including place of birth, language, religion, history of migration, and pa-rental place of birth [Cappuccio et al., 1997, 1998]. All participants of ethnic minor-ity groups were of first generation. After allowance for subjects no longer living atthe address registered, the response rate was 64% [Cappuccio et al., 1997, 1998].After removal of incomplete records and 110 participants not belonging to one of thethree main ethnic groups under study, a final sample size of 1,577 was obtained (523whites, 549 blacks of African origin, e.g., Caribbeans and West Africans, and 505South Asians). The –93T/G polymorphism was determined on 1,342 individuals(85.1%), the D9N on 1,340 (85.0%), the N291S on 1,362 (86.4%), the X447X on1,281 (81.2%), and the HindIII on 1,417 (89.8%) who also had a full lipid profileavailable. The study protocol was approved by the local Ethics Committee. All par-ticipants gave their informed consent to participate.

Procedures

Participants attended a dedicated screening unit at St George’s Hospital after anovernight fast. Height and weight were measured by standardised methods [Cappuccioet al., 1997, 1998] and body mass index (BMI) calculated. A questionnaire estab-lished the use of drugs that might affect lipid levels. As less than 1% of the wholepopulation was on lipid-lowering therapy (equally by ethnic group), they were notexcluded from the analysis. Blood for lipids was collected in plain Vacutainer tubes,left to clot and serum separated and stored at –40°C until assayed. Aliquots wereshipped in dry ice to the University of Naples for biochemical analysis. Lipids weremeasured by automated methods (Cobas Mira, Roche, Milan, Italy). High-densitylipoprotein (HDL) cholesterol was separated from the other lipoprotein fractions byprecipitation with sodium phosphotungstate and magnesium chloride. The coefficientof variations (CV) between assays were 1.8% for Tg and 2.9% for HDL-cholesterol.

206 Hall et al.

Lipid determinations underwent an external quality control from the WHO LipidReference Centre in Prague.

DNA Isolation and Polymerase Chain Reaction

Genomic DNA was isolated from whole blood [Sagnella et al., 1999]. Detection ofthe variant sites are as previously reported [Gerdes et al., 1997; Hall et al., 1997; Humphrieset al., 1998]. Each reaction contained 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 0.2 mMeach of dATP, dCTP, dGTP, and dTTP, 0.001% gelatine, 0.05% W1 (Polyoxyethyleneether, Gibco BRL), 10 pmol of each oligonucleotide primer, and 0.4 U of Taq poly-merase (Gibco BRL). All PCR reactions were performed on an MJ PTC220 DNA En-gine. Each restriction digest used between 2 and 4 U of the respective enzyme and wasincubated at 37°C (65°C for TaqI) overnight. Restriction fragments were separated on7.5% MADGE gels [Day et al., 1995]. Genotype data was available on 1,342 individuals(85.1%) for –93T/G, 1,340 for D9N (85.0%), 1,362 for N291S (86.4%), 1,281 for S447X(81.2%), and 1,417 for HindIII polymorphism (89.8%).

Statistical Analysis

Data were analysed using the SAS system (SAS Institute, Cary, North Carolina,USA). Allelic associations were given as the coefficient ∆ [Chakravarti et al., 1984]and statistical significance calculated using χ2 tests. Lipid values were log trans-formed to normalise their frequency distributions. Analyses of variance and covari-ance (to adjust for age and body mass index) were used on log transformed values tocompare means and geometric means (and 95% confidence intervals) are reportedfor clarity. A P value less than 0.05 was taken as statistically significant.

RESULTSAllele Frequencies (Table I)

Three out of the five variants studied were present more frequently in blacksthan the other two groups; these variants were –93G (P < 10–10), N9 (P < 0.001), andH– (P < 0.00001). –93G was also found at a slightly higher frequency in South Asianscompared to whites (P < 0.05). A significantly higher frequency of the S291 allelewas observed in whites (P < 0.05), but there was no difference in the frequencybetween blacks and South Asians. A slightly lower frequency of the protective X447allele was seen in blacks, but this difference was not statistically significant.

Allelic Associations (Table II)

The previously reported allelic association between –93G and N9 was seen inall three groups. This was weakest in blacks compared with South Asians and whites(∆ = 0.18 vs. ∆ = 0.55 and ∆ = 0.65, respectively). In addition, the allelic associationbetween X447 and H– was also observed in all three groups, but again was weakestin blacks compared with South Asians and whites (∆ = 0.34 vs. ∆ = 0.53 and ∆ =0.50, respectively).

A weak but significant association was seen between the S291 and N9 alleles inSouth Asians (∆ = 0.18, P = 0.01). The low frequency of these two variants in thisgroup, however, results in small carrier numbers and the significant association islargely a consequence of a single N9/S291 compound heterozygote individual, sothe possibility remains that this association is due to chance alone.

LPL V

ariants in Different E

thnic Groups

207

TABLE I. Allele Frequencies % (95% C.I.) of Common LPL Polymorphisms in Three Ethnic Groups

Ethnic –93G N9 S291 H- X447group (n = 1,342) (n = 1,340) (n = 1,362) (n = 1,417) (n = 1,281)

White 2.52 (1.50, 3.53) 1.11 (0.43, 1.80) 1.30**** (0.57, 2.03) 24.4 (28.7, 33.0) 8.38 (6.53, 10.2)(n = 456) (n = 449) (n = 462) (n = 480) (n = 436)

Black 40.7*** (37.5, 43.9) 4.64* (3.27, 6.00) 0.22 (0.09, 0.54) 35.9** (31.1, 4.07) 5.98 (4.39, 7.58)(n = 446) (n = 447) (n = 459) (n = 481) (n = 424)

South Asian 4.42***** (3.06, 5.78) 1.56 (0.75, 2.37) 0.23 (0.08, 0.52) 21.8 (25.7, 29.7) 8.73 (6.83, 10.6)(n = 440) (n = 444) (n = 441) (n = 456) (n = 421)

*P < 0.001, **P < 0.00001, ***P < 10–10, frequency in blacks higher than others. ****Frequency in whites higher than others,P < 0.05. *****Frequency in South Asians higher than whites, P < 0.05.

208 Hall et al.

Effects on Serum Tg Concentrations of –93G and N9 in Blacks

Only in the blacks is the allelic association between –93G and N9 weak enoughto provide sufficient numbers of –93G/D9 individuals to allow meaningful compari-son of the haplotype groups (Table III). Haplotype groups were determined on thebasis of carrier status as follows: the T/D haplotype group contains TT/DD individu-als, the G/D group contains GG/DD and TG/DD individuals, and the G/N groupcontains GG/DN, TG/DN and GG/NN individuals. No differences in age-adjustedserum Tg levels or HDL-cholesterol levels between the three haplotype groups inblacks were observed nor were differences found when values were also adjusted fordifferences in body mass index (data not shown).

Effects on Serum Tg Concentrations of HindIII and S447X

No effect on serum Tg concentrations was seen when the three groups werestratified by H– carrier status (Table IV) or by X447 carrier status (Table V). Simi-larly, no differences were seen between groups for HDL-cholesterol by H– carrierstatus (Table IV) or by X447 carrier status (Table V). Adjustment for body massindex did not alter the results (data not shown).

Participants who were excluded from the present analysis because they werenot genotyped or had not had serum lipids measured were comparable to those whowere included in the analysis (data not shown).

DISCUSSION

The present study investigated ethnic variations in the frequency and lipid asso-ciations of five common variants of the LPL gene in a well-defined unselected multi-

TABLE II. Allelic Associations (D) Between Common LPL Polymorphisms by Ethnic Group*

D9N N291S HindIII S447X

White–93T/G 0.65**** 0.10 –0.01 –0.02D9N –0.01 –0.04 –0.03N291S 0.03 –0.03HindIII 0.50****

Black–93T/G 0.18*** 0.01 –0.08 –0.06D9N –0.01 –0.08 –0.04N291S 0.01 –0.01HindIII 0.34****

South Asian–93T/G 0.55**** 0.02 –0.09 –0.06D9N 0.18** 0.05 0.00N291S 0.00 –0.01HindIII 0.53****

*Associations are given as the correlation coefficient ∆. All values are not significant at P = 0.05unless stated.** P < 0.01.*** P = 0.005.**** P < 5 ́ 1010.

LPL V

ariants in Different E

thnic Groups

209

TABLE III. Age-Adjusted Geometric Means (95% CI) of Fasting Serum Tg and HDL-Cholesterol Levels byCarriers of Presumed Haplotypes for Black Men and Women*

Men (n = 172) Women (n = 256)

Triglycerides HDL-cholesterol Triglycerides HDL-cholesterol–93T/G + D9N N (mmol/l) (mmol/l) N (mmol/l) (mmol/l)

T/D 59 0.90 (0.81–1.01) 1.28 (1.20–1.37) 90 0.80 (0.74–0.86) 1.55 (1.47–1.64)G/D 100 0.86 (0.79–0.93) 1.33 (1.27–1.40) 147 0.80 (0.75–0.85) 1.52 (1.45–1.58)G/N 13 0.92 (0.74–1.15) 1.35 (1.18–1.55) 19 0.88 (0.74–1.04) 1.74 (1.54–1.96)

*T/D = TT/DD; G/D = GG/DD + TG/DD; G/N = GG/DN + TG/DN + GG/NN.

TABLE IV. Age-Adjusted Geometric Means (95% CI) of Fasting Serum Tg and HDL-Cholesterol Levels bySex, Ethnic Group, and Carrier Status for HindIII

White Black South Asian

Men Women Men Women Men WomenH-carrier (n = 216) (n = 264) (n = 190) (n = 291) (n = 235) (n = 221)

Triglycerides (mmol/l)Yes 1.25 1.11 0.89 0.78 1.44 1.23

(1.14–1.38) (1.02–1.20) (0.81–0.97) (0.73–0.84) (1.31–1.58) (1.12–1.34)No 1.34 1.16 0.88 0.80 1.59 1.26

(1.23–1.46) (1.08–1.24) (0.79–0.98) (0.74–0.87) (1.46–1.72) (1.17–1.37)HDL-Cholesterol (mmol/l)

Yes 1.24 1.49 1.25 1.58 1.04 1.27(1.18–1.31) (1.42–1.57) (1.20–1.32) (1.51–1.64) (0.99–1.10) (1.20–1.33)

No 1.21 1.45 1.38 1.50 1.05 1.27(1.15–1.26) (1.39–1.51) (1.30–1.46) (1.43–1.57) (1.01–1.10) (1.21–1.34)

210

TABLE V. Age-Adjusted Geometric Means (95% CI) of Fasting Serum Triglycerides and HDL-CholesterolLevels by Sex, Ethnic Group, and Carrier Status for S447X

White Black South Asian

Men Women Men Women Men WomenX447 carrier (n = 194) (n = 242) (n = 167) (n = 257) (n = 215) (n = 206)

Triglycerides (mmol/l)Yes 1.15 1.08 0.98 0.78 1.37 1.35

(0.97–1.35) (0.94–1.26) (0.77–1.25) (0.67–0.90) (1.16–1.62) (1.18–1.55)No 1.34 1.13 0.88 0.79 1.56 1.22

(1.24–1.44) (1.06–1.20) (0.82–0.96) (0.75–0.84) (1.46–1.67) (1.15–1.31)HDL-Cholesterol (mmol/l)

Yes 1.30 1.47 1.30 1.62 1.08 1.26(1.19–1.42) (1.34–1.61) (1.14–1.48) (1.48–1.77) (0.99–1.18) (1.16–1.37)

No 1.20 1.47 1.31 1.54 1.04 1.27(1.16–1.25) (1.42–1.52) (1.25–1.36) (1.48–1.59) (1.00–1.08) (1.22–1.33)

212 Hall et al.

ethnic population sample living in London, thereby mitigating the potential effectsof differences in environmental background on phenotypes. The ethnic minority groupshad both parents born in the country of origin, thus markedly reducing the impact ofadmixture. Standardised methodology was used and systematic bias could be ruledout. Previous studies examining the links between LPL gene variants and CHD riskin ethnic minorities have been based on relatively small samples drawn from clinicalsettings or selected in different ways and sometimes pooling samples characterisedwith different methodologies [Hall et al., 1997; Talmud et al., 1998; Ehrenborg et al.,1997]. These factors may introduce important selection bias.

–93T/G, D9N, and Fasting Tg

We and others reported an association of the –93G with lower Tg in a white,middle-aged, male cohort [Hall et al., 1997] and in groups of black South Africans[Ehrenborg et al., 1997] and African Americans [Talmud et al., 1998]. Unlike theseearlier reports, no association of the –93G allele with fasting Tg was observed in thecurrent study. This apparent inconsistency may have arisen from the different experi-mental conditions of the studies. In the Northwick Park Cohort [Hall et al., 1997],plasma Tg were measured after a light breakfast, whereas in the other two studiessmall selected samples of people of African origin were studied who had higherlevels of plasma Tg than the present cohort.

In a reporter gene assay, transient transfection studies in a rat smooth musclecell line have demonstrated that the –93G LPL promoter is stronger than the –93Tpromoter [Hall et al., 1997]. This contrasts with other findings demonstrating a weaker–93G promoter compared with –93T [Yang et al., 1995]. Lower plasma Tg concen-trations in carriers of –93G have been observed to support this [Hall et al., 1997;Talmud et al., 1998; Ehrenborg et al., 1997]. In contrast, the N9 mutation results in aprotein which is secretion-defective in vitro [Mailly et al., 1995]. Our hypothesis isthat –93G/N9 individuals will produce an increased amount of a defective enzyme,leading to raised Tg levels.

The allelic association between –93G and N9 was observed in all three groups,but the association was weakest in blacks (∆ = 0.18 vs. ∆ > 0.50). We have previ-ously shown by maximum likelihood analysis that it is most likely that the –93G andN9 variants exist on the same chromosome when they occur in the same individual[Hall et al., 1997]. Despite the lower allelic association between the two variants inblacks, the frequency of the N9 allele was still approximately 3.5 times more fre-quent than in whites or South Asians (Table II). This polymorphism has been shownto be associated with increased plasma Tg levels and reduced post-heparin LPL ac-tivity [Mailly et al., 1995].

People of African origin, despite the highest body mass index, had the lowest fast-ing serum Tg concentrations of the three groups [Cappuccio et al., 1998]. We propose thefollowing explanation for the absence of a –93G effect. In the fasting state, levels ofcirculating Tg-rich lipoproteins are low and enhanced lipolytic activity has little effect.However, when the concentration of Tg in the plasma rises after a meal, the LPL activityincrease is highlighted [Talmud et al., 1998]. Overall serum Tg concentrations in theblacks studied here are lower than those observed in earlier studies [Talmud et al., 1998].The lipid raising effect of the –93G/N9 in blacks was not seen but once again, this maybe a result of the low Tg concentrations in this group.

LPL Variants in Different Ethnic Groups 213

We also report for the first time a higher frequency of the –93G allele in SouthAsians than whites (4.42 vs. 2.52%, Table II) although the magnitude of the differencebetween these two groups is small when compared to that between whites and blacks.

HindIII and S447X Allelic Association

The H+ allele has been associated with raised HDL-cholesterol [Heizmann etal., 1991], a predisposition to myocardial infarction [Jemaa et al., 1995], and raisedTg in some studies [Ahn et al., 1993; Jemaa et al., 1997] but not others [Jemaa et al.,1995; Heizmann et al., 1991]. It was not surprising, therefore, to find it less fre-quently in the low-Tg group of blacks than the other two groups.

It is unlikely that the HindIII site is functional and a probable candidate is theS447X variant site that results in a two amino acid truncated protein [Humphries etal., 1998]. The X447 allele always occurs on H– alleles and is associated with re-duced plasma Tg [Jemaa et al., 1995; Thorn et al., 1998] and raised HDL-cholesterol[Thorn et al., 1998] and may explain part, but not all of the HindIII effect [Jemaa etal., 1997; Peacock et al., 1992]. Unmapped regions upstream of the coding sequencemay contain elements important for expression, which HindIII acts as a marker for(Rudolph Zechner, personal communication). In addition, 3′ regions of the gene mayplay an important part in modulation of mRNA stability. The lower degree of allelicassociation in blacks compared to white and South Asian populations, however, castsdoubt on whether HindIII will be associated with any variants to the same degree inall three groups.

Nickerson et al. [1998] highlighted the diversity of the LPL gene. Many of theidentified sequence differences occurred in single individuals while many were intronicand, therefore, unlikely to be functional. All three populations studied by Nickerson et al.[1998] showed similar sequence diversity. Thus, we cannot exclude other yet unidenti-fied genetic variation in the LPL influencing the ethnic differences in Tg levels.

These results from the Wandsworth Heart and Stroke Study confirm the highfrequency of LPL alleles in black people of African origin compared to South Asiansand whites. It is possible that some concomitant pharmacological treatment mighthave confounded the associations with lipid levels. Lipid lowering therapy was rarein our cohort (less than 1%). In total, 276 individuals were on regular anti-hyperten-sive treatment, 58% of whom were of African origin. Thiazide diuretics and beta-blockers might in the long term increase Tg and reduce HDL-cholesterol. The greaterproportion of treated blacks is unlikely to have biased the results since they had thelowest Tg levels and the highest HDL-cholesterol. Moreover, there was no differ-ence in the allele frequency of S447X genotype between treated and untreated[Kuivenhonen et al., 1997] nor were the serum levels of Tg and HDL-cholesterolsignificantly different according to S447X genotype in any of the ethnic group. Fi-nally, 145 women were on hormone replacement therapy. Of those there was a higherproportion of white (27%) than black (16%) and South Asian (10%) women. How-ever, no consistent effect was found on either serum Tg or HDL-cholesterol.

ACKNOWLEDGMENTS

S.H. and P.J.T. are supported by the British Heart Foundation. We are indebtedto Thomas Matthews, Sam Cooke, Abi Onipinla, and Roberto Iacone for expert tech-

214 Hall et al.

nical assistance. We also thank the staff of the Blood Pressure Unit for support andencouragement. The Wandsworth Heart and Stroke Study received support from theWandsworth Health Authority, the South Thames Regional Health Authority, the NHSR&D Directorate, the British Heart Foundation, the British Diabetic Association,and the Stroke Association. D.G.C., G.A.S., and F.P.C. are members of the St. George’sCardiovascular Research Group.

REFERENCES

Ahn YI, Kamboh MI, Hamman RF, Cole SA, Ferrell RE. 1993. Two DNA polymorphisms in the lipo-protein lipase gene and their associations with factors related to cardiovascular disease. J LipidRes 34:421–428.

Cappuccio FP. 1997. Ethnicity and cardiovascular risk: variations in people of African ancestry andSouth Asian origin. J Hum Hypertens 11:571–576.

Cappuccio FP, Cook DG, Atkinson RW, Strazzullo P. 1997. Prevalence, detection and management ofcardiovascular risk factors in different ethnic groups in south London. Heart 78:555–563.

Cappuccio FP, Cook DG, Atkinson RW, Wicks PD. 1998. The Wandsworth Heart and Stroke Study: apopulation-based survey of cardiovascular risk factors in different ethnic groups. Methods andbaseline findings. Nutr Metab Cardiovasc Dis 8:371–385.

Chakravarti A, Buetow KH, Antonarakis SE, Waber PG, Boehm CD, Kazazian HH. 1984. Nonuniformrecombination within the human beta globin gene cluster. Am J Hum Genet 36:1239–1258.

Day INM, Humphries SE, Richards S, Norton D, Reid M. 1995. High throughput genotyping usinghorizontal polyacrylamide gels with wells arranged for microplate array diagonal gel electro-phoresis. Biotechniques 19:830–835.

Ehrenborg E, Clee SM, Pimstone SN, Reymer PWA, Benlian P, Hoogendijk CF, Davis HJ, Bissada N,Miao L, Gagné SE, Greenberg LJ, Henry R, Henderson H, Ordovas JM, Schaefer EJ, KasteleinJJP, Kotze MJ, Hayden MR. 1997. Ethnic variation and in vivo effects of the –93T→G variantin the lipoprotein lipase gene. Arterioscler Thromb Vasc Biol 17:2672–2678.

Fisher RM, Mailly F, Peacock RE, Hamsten A, Seed M, Yudkin JS, Beiseigel U, Feussner G, Miller G,Humphries SE, Talmud PJ. 1995. Interaction of the lipoprotein lipase asparagine 291→serinemutation with body mass index determines elevated plasma triacylglycerol concentrations: astudy in hyperlipidaemic subjects, myocardial infarction survivors and healthy adults. J LipidRes 36:2104–2112.

Fisher RM, Humphries SE, Talmud PJ. 1997. Common variation in the lipoprotein lipase gene: effectson plasma lipids and risk of atherosclerosis. Atherosclerosis 135:145–159.

Gagné SE, Pimstone SN, Larson MG, Schaefer EJ, Wilson PWF, Ordovas JM, Hayden MR. 1996.Effects of lipoprotein lipase gene mutations on plasma lipids: The Framingham Offspring StudyAmerican Heart Association 69th Scientific Sessions. Circulation 94(Suppl):I-273.

Gerdes C, Fisher RM, Nicaud V, Boer J, Humphries SE, Talmud PJ, Faergeman O. 1997. Lipoproteinlipase variants D9N and N291S are associated with increased plasma Tg and lower high-densitylipoprotein cholesterol concentrations. Studies in the fasting and post-prandial states: the Euro-pean Atherosclerosis Research Studies. Circulation 96:733–740.

Goldberg IJ. 1996. Lipoprotein lipase and lipolysis: central roles in lipoprotein metabolism and athero-genesis. J Lipid Res 37:693–707.

Hall S, Chu G, Miller G, Cruickshank K, Cooper JA, Humphries SE, Talmud PJ. 1997. A commonmutation in the lipoprotein lipase gene promoter, –93T/G, is associated with lower plasma Tglevels and increased promoter activity in vitro. Arterioscler Thromb Vasc Biol 17:1969–1976.

Heizmann C, Kirchgessner T, Kwiterovich PO, Ladias JA, Derby C, Antonarakis SE, Lusis AJ. 1991.DNA polymorphism haplotypes of the human lipoprotein lipase gene: possible association withhigh density lipoprotein levels. Hum Genet 86:578–584.

Hokanson JE, Austin MA. 1996. Plasma Tg level is a risk factor for cardiovascular disease indepen-dent of high-density lipoprotein cholesterol level: a meta-analysis of population-based prospec-tive studies. J Cardiovasc Risk 3:213–219.

Humphries SE, Nicaud V, Margalef J, Tiret L, Talmud PJ. 1998. Lipoprotein lipase gene variation is

LPL Variants in Different Ethnic Groups 215

associated with a paternal history of premature coronary artery disease and fasting and post-prandial plasma Tg. Arterioscler Thromb Vasc Biol 18:526–534.

Jemaa R, Fumeron F, Poirier O, Lecerf L, Evans A, Arveiler D, Luc G, Cambou JP, Bard JM, FruchartJC. 1995. Lipoprotein lipase gene polymorphisms: associations with myocardial infarction andlipoprotein levels, the ECTIM study Etude Cas Témoin sur l’Infarctus du Myocarde. J LipidRes 36:2141–2146.

Jemaa R, Tuzet S, Betoulle D, Apfelbaum M, Fumeron F. 1997. HindIII polymorphism of the lipopro-tein lipase gene and plasma lipid response to low calorie diet. Int J Obes Relat Metab Disord21:280–283.

Kuivenhonen JA, Groenemeyer BE, Boer JM, Reymer PW, Berghuis R, Bruin T, Jansen H, Seidell JC,Kastelein JJ. 1997. Ser447stop mutation in lipoprotein lipase is associated with elevated HDLcholesterol levels in normolipidemic males. Arterioscler Thromb Vasc Biol 17:595–599.

Mailly F, Tugrul Y, Reymer PWA, Bruin T, Seed M, Groenemeyer BF, Asplund-Carlson A, Vallance D,Winder AF, Miller GJ, Kastelein JJP, Hamsten A, Olivecrona G, Humphries SE, Talmud PJ.1995. A common variation in the gene for lipoprotein lipase (Asp9→Asn): functional implica-tions and prevalence in normal and hyperlipidaemic subjects. Arterioscler Thromb Vasc Biol15:468–478.

Mailly F, Fisher RM, Nicaud V, Luong LA, Evans AE, Marques-Vidal P, Luc G, Arveiler D, Bard J,Poirier O, Talmud PJ, Humphries SE. 1996. Association between the LP1-D9N mutation in thelipoprotein lipase gene and plasma lipid traits in myocardial infarction survivors from the ECTIMstudy. Atherosclerosis 122:21–28.

Mattu RK, Needham EW, Morgan R, Rees A, Hackshaw AK, Stocks J, Elwood PC, Galton DJ. 1994.DNA variants at the LPL gene locus associate with angiographically defined severity of athero-sclerosis and serum lipoprotein levels in a Welsh population. Arterioscler Thromb 14:1090–1097.

Nickerson DA, Tylor SL, Weiss KM, Clark AG, Hutchinson RG, Stengard J, Salomaa V, Vartiainen E,Boerwinkle E, Sing CF. 1998. DNA sequence diversity in a 9.7-kb region of the human lipopro-tein lipase gene. Nature Genet 19:233–240.

Novick AC, Zaki S, Goldfarb D, Hodge EE. 1994. Epidemiologic and clinical comparison of renalartery stenosis in black patients and white patients. J Vasc Surg 20:1–5.

OPCS. 1990. Office of population census and surveys, mortality and geography: a review in the mid-1980s. The registrar general’s decennial supplement for England and Wales. Series DS no. 9.London: HM Stationery Office.

Peacock RE, Hamsten A, Nilsson-Ehle P, Humphries SE. 1992. Associations between lipoprotein li-pase gene polymorphisms and plasma correlations of lipids, lipoproteins and lipase activities inyoung myocardial infarction survivors and age-matched healthy individuals from Sweden. Ath-erosclerosis 97:171–185.

Proudler AJ, Godsland IF, Bruce R, Seed M, Wynn V. 1996. Lipid and carbohydrate metabolic riskmarkers for coronary heart disease and blood pressure in healthy non-obese premenopausal womenof different racial origins in the United Kingdom. Metabolism 45:328–333.

Rao GH, White JG. 1993. Coronary artery disease: an overview of risk factors. Ind Heart J 45:143–153.Reymer PWA, Groenemeyer BE, Gagné E, Miao L, Appelman EEG, Seidel JC, Kromhout D, Bijvoet

SM, van de Oever K, Bruin T, Hayden MR, Kastelein JJP. 1995. A frequently occurring muta-tion in the lipoprotein lipase gene (Asn291Ser) contributes to the expression of familial com-bined hyperlipidaemia. Hum Mol Genet 4:1543–1549.

Sagnella GA, Rothwell MJ, Onipinla AK, Wicks PD, Cook DG, Cappuccio FP. 1999. A populationstudy of ethnic variations in the angiotensin-converting enzyme I/D polymorphism: relation-ships with gender, hypertension and impaired glucose metabolism. J Hypertens 17:657–664.

Semenkovich CF, Wims M, Noe L, Etienne J, Chan L. 1989. Insulin regulation of lipoprotein lipaseactivity in 3T3-L1 adipocytes is mediated at posttranscriptional and posttranslational levels. JBiol Chem 264:9030–9038.

Summerson JH, Konen JC, Dignan MB. 1992. Racial differences in lipid and lipoprotein levels indiabetes. Metabolism 41:851–855.

Syvänne M, Antikainen M, Ehnholm S, Tenkanen H, Lahdenperä, Ehnholm C, Taskinen M-R. 1996.Heterozygosity for asn291→ser mutation in the lipoprotein lipase gene in two Finnish pedi-grees: effect of hyperinsulinaemia on the expression of hyperTgmia. J Lipid Res 37:727–738.

216 Hall et al.

Talmud PJ, Hall S, Holleran S, Ramakrishnan R, Ginsberg HN, Humphries SE. 1998. LPL promoter–93T/G transition influences fasting and postprandial plasma Tg response in African-Americansand Hispanics. J Lipid Res 39:1189–1196.

Taskinen M-R. 1987. Lipoprotein lipase in hypertriglyceridaemias. In: Borensztajn J, editor. Lipopro-tein lipase. Chicago: Evener. p 201–225.

Thorn JA, Needham EWA, Mattu RK, Stocks J, Galton DJ. 1998. The Ser447-Ter mutation of the lipo-protein lipase gene relates to variability of serum lipid and lipoprotein levels in monozygotictwins. J Lipid Res 39:437–441.

Yang W-S, Nevin DN, Peng R, Brunzell JD, Deeb SS. 1995. A mutation in the promoter of the lipopro-tein lipase (LPL) gene in a patient with familial combined hyperlipidaemia and low LPL activ-ity. Proc Natl Acad Sci USA 92:4462–4466.

Wittrup HH, Tybaerg-Hansen A, Abildgaard S, Steffensen R, Schnohr P, Nordestgaard BG. 1997. Acommon substitution (Asn291Ser) in lipoprotein lipase is associated with increased risk of is-chemic heart disease. J Clin Invest 99:1606–1613.

Zhang Q, Cavanna J, Winkelman BR, Shine B, Gross W, Marz W, Galton DJ. 1995. Common geneticvariants of lipoprotein lipase that relate to lipid transport in patients with premature coronaryartery disease. Clin Genet 48:293–298.