(Circulation. 1995;92:1765-1769.)
© 1995 American Heart Association, Inc.
Articles |
From the Department of Internal Medicine and Cardiology A, Aarhus Amtssygehus University Hospital, Aarhus, Denmark.
Correspondence to Dr C. Gerdes, Department of Internal Medicine and Cardiology A, Aarhus Amtssygehus University Hospital, Tage Hansensgade 2, DK-8000 Aarhus C, Denmark.
| Abstract |
|---|
|
|
|---|
Methods and Results Associations of a HindIII (intron 8) and a Pvu II (intron 6) RFLP in the LPL gene with plasma concentrations of cholesterol, HDL cholesterol, non-HDL cholesterol, and triglycerides were studied in 457 randomly selected 40-year-old Danish men. The HindIII and the Pvu II sites were in strong linkage disequilibrium. The frequencies of the H+ and P+ alleles (+ denotes presence of cutting site) were 0.717 and 0.464, respectively. In multivariate analysis, there was a clear gene dosage effect of the H+ allele on HDL. The lowest HDL cholesterol concentration was in the H+H+ group, the highest concentration was in the H-H- group, and the H+H- group had intermediate HDL concentrations (P=.03). There was a similar, but not statistically significant gene dosage effect on triglyceride concentrations, with the highest value seen in the H+H+ group. There were no other associations between LPL RFLPs and lipoprotein components. In males reporting family history of premature ischemic heart disease, the H+H+ genotype was overrepresented (odds ratio, 2.75; 95% confidence interval, 1.37 to 5.53).
Conclusions The results suggest that genetic variation in or near the LPL gene plays a role in interindividual differences in HDL cholesterol concentration and in risk of atherosclerosis and ischemic heart disease in men.
Key Words: lipoproteins cholesterol triglycerides genes coronary disease
| Introduction |
|---|
|
|
|---|
35 kb in length and contains 10 exons. In addition to major gene
rearrangements and missense mutations, which cluster in exons 4, 5, and
6,6 several restriction fragment length polymorphisms
(RFLPs) have been reported at the LPL
locus,7 8 9 10 11 12 13 including a HindIII polymorphism
in intron 8 and a Pvu II polymorphism in intron 6. Several trials have been carried out to explore associations between LPL gene polymorphisms and lipoprotein phenotypes. Although the results are not consistent, the HindIII polymorphism is associated with hypertriglyceridemia,14 15 coronary heart disease,15 16 and variation in HDL cholesterol concentration.12 The Pvu II site is associated with variation in plasma triglyceride concentration.14 However, the populations studied have been rather small and heterogeneous in regard to ethnic background, age, and sex.
We present data on the association of a HindIII and a Pvu II polymorphism in the LPL gene with variation in plasma lipids in a population of men (n=462) of the same age who were well characterized regarding factors known to affect plasma lipid concentrations.
| Methods |
|---|
|
|
|---|
Polymerase Chain Reactions
DNA was amplified by polymerase chain reaction (PCR) in a
Perkins Elmer/Cetus Thermal Cycler Model 480. To amplify a 1.3-kb
fragment containing the HindIII restriction site in intron 8
and a 858-bp fragment containing the Pvu II restriction site
in intron 6 of the LPL gene, we used a modification of the
method described by Ahn et al.17 The modifications for the
HindIII site were 0.9 µmol/L each primer and 35 cycles.
The modifications for the Pvu II site were 0.5 µmol/L each
primer, annealing at 67°C and 33 cycles. Blind samples (reagents
minus DNA) were included in each PCR reaction to exclude
contamination.
Amplified products were digested overnight at 37°C with HindIII or Pvu II according to the manufacturer's recommendations (Boehringer Mannheim). After separation on 2% agarose gels, the resulting fragments were stained with ethidium bromide and visualized with an UV transilluminator.
After digestion with HindIII, the presence of the restriction site (H+ allele), resulted in fragments of 600 and 700 bp. The presence of the Pvu II site (P+ allele) yielded fragments of 256 and 592 bp.
Lipid Analysis
EDTA blood samples were drawn after an overnight (12-hour) fast.
Plasma cholesterol, HDL cholesterol, and plasma
triglyceride concentrations were measured by standard
enzymatic methods (Boehringer Mannheim). As a combined measure
for the cholesterol in the LDL and VLDL fractions of
plasma, non-HDL cholesterol was calculated as plasma
cholesterol minus HDL cholesterol.
Statistical Analyses
Allele frequencies for each polymorphic site were
estimated by the gene-counting method. Polymorphism information
content (PIC) was calculated to estimate the informativeness of each
polymorphism.18 The distribution of single diallelic
RFLPs was tested for Hardy-Weinberg equilibrium with a
2 test. Haplotype frequencies were estimated by
using the maximum likelihood procedure outlined by Hill,19
and the linkage disequilibrium parameter D was computed and
tested as described.19
Triglycerides and body mass index (BMI;
[weight/height2]) were loge
transformed to obtain a normal distribution. Smoking status was
categorized into groups of grams of tobacco per day:
5, 5 to<15, 15
to 25, and >25 g. Alcohol consumption was categorized into groups of
units alcohol per week: 0 to <10; 10 to <20; 20 to <40, and >40
units (1 unit is
10 g ethanol). Physical activity was categorized
into groups of hours of sports activity per week: 0 to <1; 1 to <3,
and
3 hours.
Because plasma lipid concentrations are influenced considerably by environment, we made a priori adjustments for important environmental factors. Variation in the biochemical traits (dependent variable) with respect to genotypes were therefore analyzed by constructing a general linear model (GLM) with genotype, BMI, smoking status, physical activity, and alcohol consumption as independent variables. This procedure is analogous to ANOVA on adjusted values. If the genotype-adjusted mean values of the dependent variable (eg, HDL cholesterol) indicated a gene dosage effect, tests for linear trend were performed by repeating the GLM analysis with the genotypes coded as 0/1/2, assuming a codominant effect. After each run, outliers from the model were identified (typically three to five subjects) and the analysis was repeated without these outliers to test the stability of the model. If the regression coefficients (and therefore the P value) were substantially altered by this procedure, the model was rejected. Possible effects of the genotypes on other covariates were tested by introducing corresponding terms of interactions into the model. All statistical analyses were done with the statistical software package SYSTAT 5.1 FOR THE MACINTOSH (SyStat Inc).
| Results |
|---|
|
|
|---|
|
Plasma HDL cholesterol concentrations were statistically
significantly different in the HindIII genotypes
when a test for linear trend was done (Table 2
), and the
effect of the H+ allele on HDL cholesterol
was -0.0442 per H+ allele. The effect of the
H+ allele on plasma triglycerides indicated
an opposite, but similar trend that was not statistically significant,
with the highest values in the H+H+
genotype group. After identification and exclusion of outliers,
no significant alterations in the regression coefficients were
observed. There were no significant interactions between the
HindIII genotypes and other covariates. The
regression model was therefore considered stable. No significant
differences were found in total plasma cholesterol or
non-HDL cholesterol between the HindIII
genotype groups.
|
No statistically significant differences were found between the
Pvu II genotype groups (Table 2
).
Table 3
presents the maximum likelihood estimates of
the haplotype frequencies. The linkage disequilibrium
parameter, D, shows that the HindIII and the
Pvu II polymorphisms are in strong linkage
disequilibrium.
|
To study the association between haplotypes and HDL
cholesterol, subjects were divided into quintiles with
respect to HDL cholesterol after adjustment for BMI,
smoking, alcohol intake, and physical activity. Allele frequencies
and maximum likelihood haplotype frequencies were estimated in each
quintile. Although there was no perfect trend, the H+
frequency decreased from the lowest to the highest HDL
cholesterol quintile (Table 3
). No such pattern was visible
for the P+ data. Compared with the H+ data, the
haplotypes showed a less perfect trend toward decreasing frequency of
H+P- from the lowest to the highest quintile. No
such pattern was apparent in the H+P+ haplotype
frequencies. The frequency of H-P-, in
contrast, showed a fairly clear increase with increasing HDL
cholesterol. Furthermore, in the lowest quintile,
H+P- was the most common of the four haplotypes,
and this was the only quintile in which this was true. In the highest
quintile, H+P- was the third most common of the
four haplotypes, and this was the only quintile in which this was true.
These data suggested that the association between H+
alleles and HDL cholesterol in large measure could be
attributable to H+P- chromosomes, and we
therefore compared directly the distributions of
H+P- and H-P- between
the lowest and the highest HDL cholesterol quintiles to
"isolate" the effect on HDL cholesterol of having
H+. The difference in H+P- and
H-P- haplotype distribution between the lowest
and highest quintiles (Table 4
) was reflected in an
odds ratio of 2.33 (95% confidence interval [CI], 1.32 to 4.10). In
contrast, the odds ratio for overall H+ allele
distribution (Table 4
) was only 1.55 (95% CI, 0.98 to 2.46).
|
There were significantly more men with the H+H+
genotype in the group reporting a positive family history of
ischemic heart disease (myocardial infarction, angina pectoris,
or both) before the age of 55 among first-degree relatives (odds
ratio, 2.75; 95% CI, 1.37 to 5.53) (Table 5
). When the
age limit was extended to 65 years, the odds ratio dropped to 1.63
(95% CI, 1.02 to 2.63) (Table 5
). When the numbers of alleles
were counted, the odds ratios were 2.31 (95% CI, 1.24 to 4.28) and
1.25 (95% CI, 0.86 to 1.84), respectively.
|
No statistically significant associations were found between Pvu II genotypes and family history of ischemic heart disease (data not shown).
| Discussion |
|---|
|
|
|---|
Similar to previous findings in whites,12 14 16 17 20 our data show a strongly significant linkage disequilibrium between the HindIII and Pvu II polymorphic sites in intron 8 and 6, respectively, in the LPL gene. The frequency of the H+ allele of the HindIII site in the present study is also in accordance with previous reports on comparable populations, whereas the frequency of 0.464 of the P+ allele of the Pvu II site was lower than the frequencies of 0.52 to 0.58 previously reported.12 14 16 17 20
There was an effect of the HindIII polymorphism on the
variation of plasma concentrations of HDL cholesterol
(Tables 3
and 4
). The H+ allele of the
HindIII site was significantly associated with lower
concentrations of HDL cholesterol, and there was a clear
gene dosage effect: the heterozygous subjects had HDL
cholesterol concentrations intermediate between those of
the two homozygous genotypes. There was an opposite, not
statistically significant gene dose effect of the H+
allele on plasma triglyceride concentrations,
reflecting the well known inverse relation between fasting
concentrations of triglycerides and HDL
cholesterol.21
Our results on the HindIII association with HDL and triglyceride concentrations are in accordance with those of Ahn et al,17 who found a statistically significant effect of the H+ allele on both HDL and triglycerides in non-Hispanic white men (n=44). Mattu et al16 reported a nonsignificant trend toward higher triglycerides and lower HDL cholesterol associated with the H+ allele in healthy men (n=123), and Peacock et al20 showed that the H+ allele was statistically significantly associated with higher triglyceride concentrations in young healthy Swedish men (n=93; mean age, 40.5 years). In contrast to Mattu et al,16 who found that the H+ allele was associated with higher plasma cholesterol, we could not demonstrate any significant effect of the H+ allele on total plasma cholesterol or non-HDL cholesterol. Also in contrast to Mattu et al, we could not demonstrate any significant effect of the P+ allele of the Pvu II site on HDL cholesterol, although there was a trend toward lower HDL associated with the P+ allele. This could be due to the differences in ethnicity, as reflected in the different allele frequencies.
The catabolism of VLDL and chylomicrons by LPL22 results in formation of IDL and chylomicron remnants (CRs), which are both depleted of triglyceride and enriched in cholesteryl esters (CEs). The latter is mediated via CE transfer protein (CETP), which transfers CE from HDL and LDL to triglyceride-rich particles in exchange for triglycerides.23 24 On further lipolysis, IDL becomes LDL, which can be cleared from the circulation via the LDL receptor. Apart from its lipolytic function, LPL binds to and travels with CR to the liver, where it enhances the clearance of these lipoproteins via the LDL receptorrelated protein.25 In the liver, hepatic lipase also acts on the triglycerides in triglyceride-enriched HDL and LDL, and the fatty acids are resecreted from the liver as triglycerides in VLDL. Dysfunctional LPL could therefore theoretically affect lipoproteins by a variety of mechanisms.
Low HDL concentrations, particularly HDL2, are inversely related to coronary artery disease26 and to fasting plasma concentrations of triglycerides.21 The relation of low HDL to coronary heart disease could be mediated by a concomitant increased concentration of atherogenic triglyceride-rich lipoproteins.1 27 28 Delayed postprandial clearance of triglyceride-rich particles (called by Patsch et al "decreased triglyceride metabolic capacity") results in increased CETP-mediated transfer of CEs from HDL to CRs and IDL in exchange for triglycerides.1 23 24 29 30 31 This leads to a decrease in HDL cholesterol and cholesterol enrichment of the remnant particles, making the latter more atherogenic.1 27
The half-time of HDL in plasma exceeds that of chylomicrons by a
factor of
1000 and that of VLDL by a factor of
50.28 32 33 34 35 36 37 An HDL particle is therefore exposed to and
interacts with numerous generations of
triglyceride-rich lipoproteins, and HDL serves as an
integrative marker of triglyceride metabolic
capacity.31 Fasting triglyceride
concentrations, on the contrary, do not necessarily provide equally
accurate information on triglyceride metabolic
capacity.29 Therefore, as suggested by Patsch et al, HDL
can be regarded as "the memory box of triglyceride
metabolism."28
In genetic epidemiological studies like the present one, the effects of genetic polymorphisms on, for example, plasma lipid concentrations are generally small.38 This could explain why we could detect a statistically significant effect of the H+ allele on HDL cholesterol, which has a relatively small variance, but could not demonstrate a significant effect on fasting triglyceride concentrations, which have a much larger variance. We therefore believe that the effect of H+ alleles on HDL cholesterol concentrations is mediated through triglyceride metabolism.
The effect of the H+ allele on HDL cholesterol concentrations presented here therefore suggests that the HindIII site in intron 8 of the LPL gene is in linkage disequilibrium with one or more regions in or in close proximity to the LPL gene, affecting the lipolytic function of LPL6 and/or the clearance of triglyceride-rich lipoproteins6 and their remnants25 and thereby generating an atherogenic lipid phenotype. Maximum likelihood estimates of haplotype frequencies showed that the H+P- haplotype was enriched in the quintile with the lowest HDL cholesterol relative to the quintile with the highest HDL cholesterol and that no major difference in the frequencies of H+P+ was found between the highest and lowest quintiles. This suggests that the responsible mutation or functional polymorphism could be located on H+P- chromosomes and that the mutation/polymorphism could have occurred on an H+ chromosome before the mutation causing the Pvu II site variation.
We, like others,15 16 found the H+ allele to be associated with risk of ischemic heart disease. In persons reporting a positive family history of premature ischemic heart disease before age 55 among first-degree relatives, the H+ allele was overrepresented with an odds ratio of 2.75 (95 CI, 1.37 to 5.53), which is higher than the comparable odds ratio of 1.8 (95% CI, 1.09 to 2.98) for the Xba I allele of a restriction fragment length polymorphism in the apolipoprotein B gene in the same population.39
In conclusion, the present results suggest that genetic variation in or near the LPL gene plays a role in interindividual differences in HDL cholesterol concentration and in the risk of atherosclerosis and ischemic heart disease in men.
| Acknowledgments |
|---|
Received February 2, 1995; revision received April 10, 1995; accepted April 16, 1995.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
Q. Yang, C.-Q. Lai, L. Parnell, L. A. Cupples, X. Adiconis, Y. Zhu, P. W. F. Wilson, D. E. Housman, A. M. Shearman, R. B. D'Agostino, et al. Genome-wide linkage analyses and candidate gene fine mapping for HDL3 cholesterol: the Framingham Study J. Lipid Res., July 1, 2005; 46(7): 1416 - 1425. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Shimo-Nakanishi, T. Urabe, N. Hattori, Y. Watanabe, T. Nagao, M. Yokochi, M. Hamamoto, and Y. Mizuno Polymorphism of the Lipoprotein Lipase Gene and Risk of Atherothrombotic Cerebral Infarction in the Japanese Stroke, July 1, 2001; 32(7): 1481 - 1486. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. R Holmer, C. Hengstenberg, B. Mayer, A. Doring, H. Lowel, S. Engel, H.-W. Hense, M. Wolf, G. Klein, G. A.J Riegger, et al. Lipoprotein lipase gene polymorphism, cholesterol subfractions and myocardial infarction in large samples of the general population Cardiovasc Res, September 1, 2000; 47(4): 806 - 812. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. Larson, M. M. Hoffmann, J. M. Ordovas, E. J. Schaefer, W. Marz, and J. Kreuzer The Lipoprotein Lipase HindIII Polymorphism: Association with Total Cholesterol and LDL-Cholesterol, but not with HDL and Triglycerides in 342 Females Clin. Chem., July 1, 1999; 45(7): 963 - 968. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. L. Anderson, G. J. King, T. L. Bair, S. P. Elmer, J. B. Muhlestein, J. Habashi, L. Mixson, and J. F. Carlquist Association of lipoprotein lipase gene polymorphisms with coronary artery disease J. Am. Coll. Cardiol., March 15, 1999; 33(4): 1013 - 1020. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Perusse, T. Rice, J. P. Despres, J. Bergeron, M. A. Province, J. Gagnon, A. S. Leon, D. C. Rao, J. S. Skinner, J. H. Wilmore, et al. Familial Resemblance of Plasma Lipids, Lipoproteins and Postheparin Lipoprotein and Hepatic Lipases in the HERITAGE Family Study Arterioscler. Thromb. Vasc. Biol., November 1, 1997; 17(11): 3263 - 3269. [Abstract] [Full Text] |
||||
![]() |
H. Knoblauch, A. Busjahn, S. Munter, Z. Nagy, H.-D. Faulhaber, H. Schuster, and F. C. Luft Heritability Analysis of Lipids and Three Gene Loci in Twins Link the Macrophage Scavenger Receptor to HDL Cholesterol Concentrations Arterioscler. Thromb. Vasc. Biol., October 1, 1997; 17(10): 2054 - 2060. [Abstract] [Full Text] |
||||
![]() |
P.S. Hansen, J.C. Defesche, J.J.P. Kastelein, L.U. Gerdes, L. Fraza, C. Gerdes, F. Tato, H.K. Jensen, L.G. Jensen, I.C. Klausen, et al. Phenotypic Variation in Patients Heterozygous for Familial Defective Apolipoprotein B (FDB) in Three European Countries Arterioscler. Thromb. Vasc. Biol., April 1, 1997; 17(4): 741 - 747. [Abstract] [Full Text] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Circulation Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 1995 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |