Pleiotropic Genetic Effects on LDL Size, Plasma
Triglyceride, and HDL Cholesterol in Families
Karen L. Edwards, Michael C. Mahaney, Arno G. Motulsky, Melissa A. Austin
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
Abstract—The interrelationships among low density lipoprotein (LDL) particle size, plasma triglyceride (TG), and high
density lipoprotein cholesterol (HDL-C) are well established and may involve underlying genetic influences. This study
evaluated common genetic effects on LDL size, TG, and HDL-C by using data from 85 kindreds participating in the
Genetic Epidemiology of Hypertriglyceridemia (GET) Study. A multivariate, maximum likelihood– based approach to
quantitative genetic analysis was used to estimate the additive effects of shared genes and shared, unmeasured
nongenetic factors on variation in LDL size and in plasma levels of TG and HDL-C. A significant (P,0.001) proportion
of the variance in each trait was attributable to the additive effects of genes. Maximum-likelihood estimates of
heritability were 0.34 for LDL size, 0.41 for TG, and 0.54 for HDL-C. Significant (P,0.001) additive genetic
correlations (rG), indicative of the shared additive effects of genes on pairs of traits, were estimated between all 3 trait
pairs: for LDL size and TG rG520.87, for LDL size and HDL-C rG50.65, and for HDL-C and TG rG520.54. A
similar pattern of significant environmental correlations between the 3 trait pairs was also observed. These results
suggest that a large proportion of the well-documented correlations in LDL size, TG, and HDL-C are likely attributable
to the influence of the same gene(s) in these families. That is, the gene(s) that may contribute to decreases in LDL size
also contribute significantly to higher plasma levels of TG and lower plasma levels of HDL-C. These relationships may
be useful in identifying genes responsible for the associations between these phenotypes and susceptibility to
cardiovascular disease in these families. (Arterioscler Thromb Vasc Biol. 1999;19:2456-2464.)
Key Words: pleiotropy n LDL size n HDL cholesterol n triglycerides n families
N
umerous studies have shown that plasma lipids and
lipoproteins are associated with increased risk of cardiovascular disease (CVD), including smaller-size LDL particles,1–3 elevated plasma triglyceride (TG),4,5 and decreased
levels of HDL cholesterol (HDL-C).6 However, whether
these associations with CVD are independent of one another
remains controversial.1,5,7
Each of these cardiovascular risk factors also appears to be
genetically influenced. For example, two of the most common forms of familial hyperlipidemia among families with
documented coronary heart disease (CHD) have been classified as familial combined hyperlipidemia (FCHL) and familial “monogenic” hypertriglyceridemia (FHTG),8 –11 both of
which are characterized by variable and obligatory hypertriglyceridemia, respectively. Although the specific genetic
bases of both FCHL and FHTG are poorly understood,
studies indicate major gene effects on TG and apo B in FCHL
families.12,13 In addition, there is also considerable evidence
for genetic influences on LDL subclass phenotypes12,14 –18 and
HDL-C.19 –22
Statistical and physiological relationships between measures of LDL size, plasma TG, and HDL-C are well
established3,23–25 and may represent a high-CHD-risk lipid/
lipoprotein phenotype with common underlying genetic
influences. Sprecher et al26,27 first proposed that the
combination of high TG and low HDL-C constituted an
inherited “conjoint trait” associated with increased risk of
CHD in families of hypertriglyceridemic or hypercholesterolemic probands participating in the Lipid Research
Clinics Family Study. An atherogenic lipoprotein phenotype characterized by small dense LDL, low plasma levels
of HDL-C, high plasma TG levels, and elevated apo B
levels was previously described by Austin et al.28 Furthermore, a multivariate lipid factor characterized primarily by
LDL size, plasma TG, and HDL-C25 has been identified
and was shown to prospectively predict CVD.29 By comparing female monozygotic and dizygotic twins, this lipid
factor was shown to be heritable and possibly linked to
both the lipoprotein lipase gene and the hormone-sensitive
lipase gene.30,31 Although the evidence supports common
genetic influences on these cardiovascular risk factors, no
study to date has simultaneously evaluated the genetic
relationships among the 3 risk factors; LDL size, plasma
TG, and HDL-C in high-risk families.
Received December 8, 1998; revision accepted March 23, 1999.
From the Department of Epidemiology (K.L.E., M.A.A.), School of Public Health and Community Medicine, and the Department of Medicine
(A.G.M.), Division of Medical Genetics, University of Washington, Seattle; and the Department of Genetics (M.C.M.), Southwest Foundation for
Biomedical Research, San Antonio, Tex.
Correspondence to Karen L. Edwards, PhD, Department of Epidemiology, Box 357236, School of Public Health and Community Medicine, University
of Washington, 1959 NE Pacific Ave, Seattle, WA 98195. E-mail: [email protected]
© 1999 American Heart Association, Inc.
Arterioscler Thromb Vasc Biol. is available at http://www.atvbaha.org
2456
Edwards et al
Quantitative Genetic Analysis of Lipids/Lipoproteins
Thus, the purpose of this investigation was to evaluate the
evidence for shared genetic influences between phenotypic
variation in LDL size, plasma levels of TG, and HDL-C in
families at increased risk for CVD32 by using quantitative
genetic analysis.
Methods
Study Sample
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
The families included in this cohort study are primarily white and are
part of the Genetic Epidemiology of Hypertriglyceridemia (GET)
Study, which is focused on understanding the risk of CVD in the
common familial forms of elevated TGs, ie, FCHL and FHTG, and
on elucidating the underlying genetic influences of these disorders.33
Family ascertainment for the GET Study was based on 2 family
studies conducted at the University of Washington, Seattle, in the
early 1970s.8,10 In brief, the first family study was published in
197310 and focused on families identified through a hyperlipidemic
family member (proband) surviving a myocardial infarction.10 The
second study was published in 1976 and identified families through
a proband with hypertriglyceridemia but without coronary disease.8
In these baseline studies, FCHL was characterized by variable
expression of hypercholesterolemia or hypertriglyceridemia in family members, while FHTG was characterized by families in which all
affected relatives had an elevated plasma TG level but not an
elevated cholesterol level. Thus, elevation in plasma TG levels in
affected relatives was always seen in FHTG and commonly observed
in FCHL at baseline.
As part of the GET Study, living individuals from a total of 85
families were recontacted between 1994 and 1997. Of the 85
families, 58 (68.2%) had been classified as FCHL and 27 (31.8%) as
FHTG at baseline. In the 20 years since the baseline studies were
initiated, many family members had died, including those with the
highest TG levels,32 and could not be resampled. New blood samples
were obtained from surviving family members, including probands
and their spouses; siblings and spouses of siblings; and grandchildren, nieces, and nephews of the proband. For the current analysis,
eligible family members were those who were over age 18 at
follow-up, who were not pregnant, and who were not too ill to
participate. Each eligible family member was asked to complete and
return a self-administered medical history questionnaire by mail and
to provide a fasting blood sample for lipid and lipoprotein determinations. Approximately 49% of participants lived in Washington
state, with the remaining 51% residing in 28 other states. All
information was kept confidential and was not shared with other
family members. Each participant provided written, informed consent at the time they were enrolled in the follow-up study, and all
methods used to contact family members were approved by the
University of Washington Internal Review Board.
Laboratory Measurements
All subjects were instructed to fast overnight. Blood samples were
centrifuged at 2500 rpm for 15 minutes using a standardized
protocol. For participants living .100 miles from the Seattle area,
plasma was separated and shipped on ice for overnight delivery to
the University of Washington.
Lipid determinations including plasma TGs and total HDL-C were
performed at the Core Laboratory, Northwest Lipid Research Laboratories in Seattle, a Centers for Disease Control and Prevention–
certified lipid laboratory. HDL-C and plasma TG were measured by
enzymatic techniques (Abbott Laboratories) according to the standardized procedures of the Lipid Research Clinics protocol.34
Nondenaturing gradient gel electrophoresis was performed on
isolated LDL using 2% to 14% polyacrylamide gradient gels, as
previously described.23,35,36 The estimated diameters of LDL subclasses were calculated on the basis of a calibration curve constructed from high-molecular-weight standards run on the same
gel.35 The size of the major LDL subclass, denoted LDL peak
particle diameter (LDL size), was identified from the isolated LDL
gel and used as a continuous variable in the quantitative genetic
analysis as a measure of LDL size heterogeneity.
2457
Statistical Analysis
Because TG and HDL-C distributions were positively skewed (2.16
and 0.96, respectively), the natural log transformations (ln) of these
risk factors were used for all analyses. Transformation reduced the
skewness in plasma TG and HDL-C to 0.12 and 20.01, respectively.
Untransformed values are presented in descriptive tables and figures
for ease of interpretation.
For the quantitative genetic analysis, pedigree and phenotypic data
were prepared using the computer package PEDSYS.37 Statistical
genetic analysis was conducted using the modified version of the
Pedigree Analysis Package, version 3.0,38 which is based on established genetic theory of partitioning the total phenotypic variance of
a trait into genetic and environmental components. In this approach,
the univariate heritability (h2) of an individual trait represents the
proportion of the total phenotypic variance due to additive genetic
effects.39 The residual heritability is used here to reflect the proportion of variance attributable to additive genetic effects after accounting for age and sex effects.
Extension of univariate quantitative genetic analysis to the multivariate state has been described in detail.38,40 – 43 In this approach, the
total phenotypic correlation between pairs of traits is partitioned into
the additive genetic (rG) and random environmental (rE) components
by using maximum-likelihood methods implemented in the modified
version of the Pedigree Analysis Package.38 The 3 pairs of traits used
in this trivariate quantitative genetic analysis were (1) LDL size and
TG, (2) LDL size and HDL-C, and (3) TG and HDL-C. The additive
genetic and environmental correlations between these pairs of traits
were obtained from the genetic and environmental variance-covariance matrices, estimated by modeling the joint distributions of the
traits as a function of their population means; their covariates and
regression coefficients; the additive genetic values; random environmental deviations; and the degree of relationship among the individuals in this sample.
The genetic and environmental correlations represent the additive
effects of shared genes (pleiotropy) and of shared, unmeasured
environmental (nongenetic) factors on the phenotypic covariance for
each pair of traits, respectively. An estimate of the total phenotypic
correlation (rP) between 2 traits was obtained using the following
equation: rP5[=h12=h22rG]1[=(12h12)=(12h22)rE],1 where h12
and h22 represent the residual heritability of each trait in the pair.
The significance of each of the estimated parameters (eg, h2, rG, and
rE) was evaluated by likelihood-ratio tests by comparing the ln likelihood of a model in which the parameter is estimated to the ln likelihood
of a more restricted model in which the same parameter is set to zero.
The likelihood-ratio test yields a statistic that is asymptotically distributed approximately as a x2 with degrees of freedom equal to the
difference in the number of estimated parameters and is calculated as
22(ln likelihoodrestricted model2ln likelihoodgeneral model). Polygenic pleiotropy is indicated by a rG value that is significantly different from zero.
An additional likelihood-ratio test was performed for each significant
genetic correlation to test for complete pleiotropy (rG51.0). In this test,
the unrestricted model, wherein rG is estimated, is compared with a
more restricted model wherein rG is fixed at either 1.0 or 21.0. When
rG is significantly different from ¦1.0¦, pleiotropy is interpreted as
incomplete.22
Finally, to obtain an estimate of the shared additive effects of
genes on each of the residual phenotypes in a pair, the squared
additive genetic correlation for that pair is multiplied by the
heritability estimate for each trait. Similarly, the proportion of
residual phenotypic variance in an individual trait that is due to
shared effects of unmeasured environmental factors is obtained by
multiplying the squared environmental correlation by 1 minus the
heritability estimate.
To better meet assumptions of normality, individuals with plasma
TG .4.5 SDs above the mean (TG .905 mg/dL) were excluded
from all analyses (n55). Thus, this analysis focuses on 780 individuals in 85 families who provided a fasting blood sample and
completed a medical history questionnaire, including 78 spouse
pairs, 501 parent-offspring pairs, and 518 sibling pairs. Quantitative
genetic analysis was first performed in the whole data set and then
separately in the FCHL and FHTG families. However, because there
were no substantial differences when the quantitative genetic anal-
2458
Arterioscler Thromb Vasc Biol.
October 1999
TABLE 1. Descriptive Summary (Mean6SD) for AGE, LDL Size, Plasma TG, and HDL-C in
Relatives and Spouses by Sex
Relatives
Age, y
Spouses
Total
Men
(n5255)
Women
(n5298)
Men
(n583)
Women
(n5144)
Men
(n5338)
Women
(n5442)
46.3615.1
45.9614.9
54.1616.3
57.4613.6
48.2615.7
49.7615.5
LDL size, Å
262.469.2
267.568.3
264.569.2
268.468.9
262.969.2
267.868.5
TG, mg/dL
185.56136.0
144.3691.2
158.56130.8
145.6689.0
178.86135.1
144.7690.4
40.9611.3
52.5615.6
41.4611.4
52.3614.2
41.4611.3
52.8615.2
HDL-C, mg/dL
Not including probands.
ysis was run separately in FCHL and FHTG families, only results
based on the combined sample are presented.
Results
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
The average age of participants in this study was 49.0 years,
and more than half of the participants were female (56.9%).
As shown in Table 1, spouses tended to be older than
relatives, although there were no differences in age between
male and female relatives or between male and female
spouses. Unadjusted measures of LDL size, plasma TG, and
HDL-C for the individuals included in this analysis by
relative type and sex are also shown in Table 1. Men tended
to have smaller-size LDL particles and lower HDL-C levels
than women, with similar patterns in relatives and spouses.
The average level of plasma TG was higher in men than
women and was higher in male relatives compared with male
spouses. Plasma TG levels did not appear to be different in
female relatives and spouses, and this may be due to increased mortality among relatives with high plasma TG levels
relative to spouses.32
Figures 1 through 3 present the frequency distributions for
all individual study subjects (n5780) for unadjusted measures of LDL size, plasma TG, and HDL-C, respectively. The
mean LDL size was 265.7 Å (SD 9.13) in this sample (Figure
1). Figure 1 also shows a suggestive bimodal distribution of
LDL size. The overall mean plasma TG level was
1.8061.27 mmol/L (159.5 mg/dL [SD 113.16]; median
1.46 mmol/L [129 mg/dL]) and was highly skewed (Figure
2). Mean HDL-C was 1.2460.38 mmol/L (47.8 mg/dL [SD
14.77]; median 1.16 mmol/L [45.0 mg/dL]) and was also
positively skewed (Figure 3).
The maximum-likelihood estimates of the mean effects and
variance components from the quantitative genetic analysis
are presented in Table 2. Likelihood-ratio tests indicate
significant (P,0.001) residual heritabilities for each of the
traits. Approximately one third of the residual variance in
LDL size (h250.34), slightly greater than one third of the
residual variance in plasma TG (h250.41), and more than half
of the residual variance in HDL-C (h250.54) were attributable to additive genetic effects. This table also includes
maximum-likelihood parameter estimates for significant
(P,0.001) covariate effects. Based on the coefficients for sex
( b 2sex), mean values for females were 268.06 Å,
0.054 mmol/L (4.78 mg/dL), and 0.102 mmol/L (3.94 mg/dL)
compared with 262.98 Å, 0.056 mmol/L (4.98 mg/dL), and
0.095 mmol/L (3.67 mg/dL) in males for LDL size, ln TG,
and ln HDL-C, respectively. Significant age effects were also
Figure 1. Frequency distribution of LDL
size in 780 family members. Mean LDL size
was 265.7 Å (SD 9.13).
Edwards et al
Quantitative Genetic Analysis of Lipids/Lipoproteins
2459
Figure 2. Frequency distribution of
plasma TG in 780 family members. Mean
plasma TG was 159.5 mg/dL (SD 113.16),
median 129 mg/dL.
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
detected for each of the 3 traits. Overall, age and sex together
accounted for 1.3%, 7.6%, and 17.6% of the total variance in
LDL size, ln TG, and ln HDL-C, respectively.
In the multivariate analysis, all genetic correlations were
significantly different from zero (P,0.001) based on
likelihood-ratio tests (Table 3). The negative genetic correlation (rG520.87) between LDL size and plasma TG was the
strongest, suggesting that genes in common contribute to
decreases in LDL size and increases in plasma TG. The
positive genetic correlation (rG50.65) between LDL size and
HDL-C was modest and suggests that shared genes that
increase LDL size also increase HDL-C. The negative genetic
correlation between plasma TG and HDL-C (rG520.54) was
also significant. Based on likelihood-ratio tests, the hypothesis of complete pleiotropy (rG561.0) was rejected for all
genetic correlations (P,0.001).
When squared, the additive genetic correlation between 2
traits in a pair provides an estimate of the additive genetic
variance for each trait in the pair that is due to effects of the
same gene(s). The squared additive genetic correlation between LDL size and plasma TG (rG2) was 0.757. Thus, almost
76% of the additive genetic variance in LDL size is shared
with TG. The squared additive genetic correlation between
LDL size and HDL-C (rG2) was 0.423, suggesting that nearly
Figure 3. Frequency distribution of
HDL-C in 780 family members. Mean
HDL-C was 47.8 mg/dL (SD 14.77),
median 45.0 mg/dL.
2460
Arterioscler Thromb Vasc Biol.
October 1999
TABLE 2. Maximum-Likelihood Parameter Estimates From the
Multivariate Quantitative Genetic Analysis of LDL Size, Plasma
TG, and HDL-C in 85 Kindreds
Parameter
LDL Size (6SE)
In TG (6SE)
m
262.9860.52
4.9860.04
3.6760.02
s
8.7760.23
0.6160.02
0.2860.01
0.5460.07
2
h *
0.3460.07
0.4160.07
b2Sex*
5.0860.61
20.2060.04
In HDL (6SE)
0.2760.02
b2Agem*
20.02860.029
0.00460.002
1.13102560.001
b2Agef*
20.07360.026
0.01360.002
1.23102360.001
m Indicates the mean value in males; h , residual heritability. Sex was
scaled so that the regression coefficient (b2sex) represents the effect of
having the covariate present (female) compared with having it absent. Age was
scaled by subtracting out the mean covariate value so that the regression
coefficient (b2age) represents the effect associated with a 1-unit deviation
from the mean level of the covariate in the overall sample. m indicates male;
f, female.
*All parameters are significant at P,0.001.
2
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
50% of the additive genetic variance in each of these traits is
due to shared genes. The squared additive genetic correlation
between plasma TG and HDL-C (rG2) was 0.292, suggesting
that almost 30% of the additive genetic variance in each trait
is due to genes shared between the pair.
Unmeasured environmental correlations were also significant (P,0.001) between all pairs of traits. Consistent with the
genetic correlations, the environmental correlations between
LDL size and plasma TG and between plasma TG and
HDL-C were negative, whereas the environmental correlation
between LDL size and HDL-C was positive (Table 3). When
squared, the environmental correlations provide an estimate
of the proportion of the random environmental variance (ie,
that proportion of the shared residual phenotypic variance not
attributable to additive genetic effects or covariates) for each
trait that is due to the effects of the same unmeasured
environmental factors. Phenotypic correlations were estimated by using the equation described in Methods and are
shown in Table 3.
Estimates of the contributions of shared additive effects of
genes and of shared unmeasured environmental factors on
each of the residual phenotypes in a pair are shown in Figures
4A through 4C. Figure 4A presents the results for LDL size
and TG, indicating that 26% of the residual variance in LDL
size was due to additive genetic effects that are shared with
plasma TG, while 19% of the residual variance was attributable to unmeasured environmental effects shared with TG.
For plasma TG, 31% of the residual variance was attributable
TABLE 3. Maximum-Likelihood Estimates of the Additive
Genetic (rG) and Environmental (rE) Correlations From the
Multivariate Quantitative Genetic Analysis of LDL Size, 1n TG,
and 1n HDL-C in 85 Kindreds and the Phenotypic
Correlations (rP)*
Phenotype Pairs
LDL size–TG
LDL size–HDL-C
TG–HDL-C
rG6SE
rE6SE
rP
20.8760.06
20.5360.05
20.66
0.6560.09
0.4860.06
0.54
20.5460.09
20.5360.07
20.53
*All genetic (rG) and environmental (rE) correlations are significant at
P,0.001.
to additive genetic effects shared with LDL size, and 17%
was attributable to shared unmeasured environmental factors.
However, the sum of the 2 proportions was less than 1 for
each trait, suggesting that other nonshared factors contributed
to the residual phenotypic variance in each trait of the pair:
55% for LDL size and 52% for TG (Figure 4A). Similar
results were obtained for the other trait pairs; however, the
proportion of nonshared residual variance was greater for
traits in these pairs (Figures 4B and 4C).
Figure 4B indicates that 14% and 23% of the residual
variance in LDL size and HDL-C, respectively, were due to
shared additive genetic effects between the pair. Shared
unmeasured environmental effects accounted for 15% and
11% of the residual variance in LDL size and HDL-C,
respectively. Approximately 30% of the residual variance of
each trait in this pair was shared; however, the majority of the
residual variance was due to other nonshared factors, 71%
and 66% for LDL size and HDL-C, respectively.
Similar results were observed for TG and HDL-C (Figure
4C). For both TG and HDL-C, 71% of the residual variance
was attributable to other nonshared factors. Shared factors
accounted for slightly less than 30% of the residual variance
in each trait; additive genetic effects accounted for 12% and
16% of the residual variance, and shared unmeasured environmental factors accounted for 17% and 13% of the residual
variance in TG and HDL-C, respectively.
Discussion
Although phenotypic associations between LDL size, TG,
and HDL-C are well documented, this is the first demonstration that the observed phenotypic associations are largely
genetic. That is, the results presented here demonstrate strong
genetic correlations between each possible pair of these traits
among families at increased risk for CVD.32 The negative
genetic correlation between LDL size and plasma TG shows
that genetic influences that decrease LDL size also result in
increases in plasma TG levels. The positive genetic correlation between LDL size and HDL-C was also significant, as
was the negative genetic correlation between HDL-C and
plasma TG, indicating that genes in common explained a
substantial proportion of the genetic covariance in each pair
of traits. However, based on rejecting the hypotheses of
complete pleiotropy, these results indicate that additional
nonshared genes also contributed to the variation in each of
the traits.
Environmental correlations between each pair of traits
were also significant, although generally not as large as the
corresponding genetic correlations. However, they still provide evidence of shared random environmental effects among
this set of phenotypically related lipid and lipoprotein risk
factors, indicating the existence of other important covariates
that were not included in this analysis. At least 40% of the
residual phenotypic variance in each of the 3 traits is
attributable to nonshared effects that almost certainly include
environmental factors and may even include some dominance
genetic effects that were not modeled in this statistical genetic
approach. Importantly, identification and addition of significant environmental factors to the genetic model would
decrease the residual environmental contribution to the phenotypic variance and increase the relative importance of the
genetic effects already detected.
Edwards et al
Quantitative Genetic Analysis of Lipids/Lipoproteins
2461
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
Figure 4. A, Results of quantitative
genetic analysis of LDL size and TG.
Percentages of residual phenotypic variance in each trait due to shared additive
genetic, shared unmeasured environmental, and nonshared effects are displayed. B, Results of quantitative genetic
analysis of LDL size and HDL-C. Percentages of residual phenotypic variance
in traits due to shared additive genetic,
shared unmeasured environmental, and
nonshared effects are displayed. C,
Results of quantitative genetic analysis
of TG and HDL-C. Percentages of residual phenotypic variance in traits due to
shared additive genetic, shared unmeasured environmental, and nonshared
effects are displayed.
Univariate residual heritability estimates for the lipids and
lipoproteins in this sample of high-risk families are consistent
with previous reports. For example, in a sample of MexicanAmerican families participating in the San Antonio Family
Heart Study, Mahaney et al21 reported heritability estimates
of 0.53 and 0.55 for ln TG and ln HDL-C, respectively,
compared with estimates from the current study of 0.41 and
0.54, respectively. Estimates of heritability for LDL size,
plasma TG, and HDL-C based on female twins are also
consistent with the estimates presented here, h250.30,14
h250.50 to 0.65, and h250.73,44 respectively. Based on data
from male twins, the heritability estimate for HDL-C was
0.36.45
The phenotypic correlations estimated in this study are also
consistent with the well-documented patterns between these
lipids and lipoproteins.23–25 For example, in a recent report
from the Physicians’ Health Study, the correlation for each
possible pair of these variables was strong and highly
statistically significant (P,0.01): 20.71 for TG and LDL
size, 20.57 for TG and HDL-C, and 0.61 for LDL size and
HDL-C3 compared with rP520.66, rP520.53, and rP50.54,
respectively, reported here.
The results presented in this study are also consistent with
a previous quantitative genetic analysis of TG and HDL-C.
Based on data from Mexican-American families participating
in the San Antonio Family Heart Study, Mahaney et al21
showed that shared genes contributed to the covariation of
TG and HDL. The magnitude of the genetic correlation
between TG and HDL-C reported in the current study of
high-risk families is remarkably similar (rG520.54 versus
20.53, respectively).21 Heller et al46 also found significant
genetic correlations between serum lipids and apolipoproteins, including TG and HDL-C, in elderly Swedish twins.
The results of the current study extend these previous reports
2462
Arterioscler Thromb Vasc Biol.
October 1999
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
by including a measure of LDL size and demonstrate common genetic influences between LDL size and both HDL-C
and plasma TG.
The hypothesis of common genetic influences on pairs of
lipids and lipoproteins was first suggested by Goldstein et
al.10 Based on data from the original family study published
in 1973, Goldstein et al suggested that the FCHL phenotype
was due to variable expression of a single dominant gene and
not segregation of 2 separate genes.10 The hypothesis of
common underlying genetic influences is further supported
by recent results from linkage studies of candidate genes
involved in lipid and lipoprotein metabolism. Two recent
reports, 1 based on families ascertained in Finland and the
other based on a mouse model of FCHL, have strongly
suggested the existence of a novel gene for FCHL on
chromosome 1.47,48 In addition, Dallinga et al49 showed that
DNA variations in the apo AI-CIII-AIV gene cluster modify
plasma TGs, LDL-C, and apo CIII levels in families with
FCHL. Recent sib-pair linkage analysis also provides evidence for common genetic influences on LDL size, TG, and
HDL-C. Preliminary evidence for linkage between the apo B
gene and a lipid factor identified by factor analysis and
characterized by LDL size, plasma TG, and HDL-C is based
on data from the Kaiser Women Twins Study.31 Furthermore,
using data from the same women twins, Austin et al50 also
provided evidence for linkage between the apo B gene and
each of the following individual risk factors: LDL size, TG,
HDL-C, and apo B levels. Several studies have also reported
linkage between the apo B gene and individual measures of
circulating lipids and lipoproteins, including measures of
LDL heterogeneity51 and TG.52 Several other candidate
genes, including the LDL receptor, lipoprotein lipase, and apo
AI-CIII-AIV loci, have been shown to be associated with
lipids and lipoproteins.53–55 Additionally, the lipoprotein
lipase gene has recently been linked to LDL size, with
simultaneous effects on TG and HDL-C in lipoprotein lipase–
deficient families.56
Together these findings suggest that each of the 3 lipid and
lipoprotein measures reflect a common underlying process
that is controlled, at least in part, by shared genes and provide
strong support for the existence of genes with pleiotropic
effects influencing covariation in plasma lipids and lipoproteins. Thus, it may be appropriate to consider the aggregate
effects of these risk factors in predicting CHD, as has been
recently reported in type 2 diabetes.23 The use of multivariate
traits is an important alternative to traditional approaches of
estimating independent effects when the traits have a common etiologic pathway. This may be particularly true when
considering genetically related risk factors.
Furthermore, from the standpoint of genetic mapping,
large-magnitude genetic correlations observed in the current
study, particularly those between LDL size and plasma TG,
can be used to delimit major-locus hypotheses, including
major-locus pleiotropy and linkage, that can then be tested by
multivariate segregation analysis and combined segregation
and linkage analysis.21,57 For example, using a multivariate
lipid phenotype characterized by LDL size, plasma TG, and
HDL-C may be advantageous when pleiotropy is due primarily to major loci.57 Evidence for linkage with the multivariate
lipid phenotype could represent a gene that is common to all
3 related traits, suggesting pleiotropic effects beyond those
described in the current article.
In addition, when pleiotropy is due primarily to shared
additive effects of polygenes,57 such traits with high genetic
and environmental correlations can also be used as covariates
in univariate segregation and linkage analyses. Specifically,
by removing common genetic effects of other traits, the
likelihood of detecting genes influencing the unique residual
component of an individual trait may be increased. For
example, on the basis of the results presented in this study,
both HDL-C and plasma TG could be included as covariates
in analyses of LDL size to account for shared additive genetic
and random environmental contributions to the variance in
LDL size. Evidence for linkage with the residual trait would
represent a gene unique to that trait. Importantly, the pleiotropic effects identified in this study can thus be used to
effectively increase the likelihood of detecting quantitative
trait loci.
Finally, although these results were interpreted as representing polygenic pleiotropy, individual loci exerting substantial effects on 1 of these 3 phenotypes may be responsible
for some or all of the pleiotropy detected in this analysis.
However, it is important to emphasize that the analyses
presented here do not explicitly distinguish between the
pleiotropic effects of single, major loci and the pleiotropic
effects of multiple loci. Thus, shared additive effects of
individual loci may well be included in the genetic correlations. Additionally, the effects of closely linked loci exhibiting strong linkage disequilibrium could also contribute to the
genetic correlations and the well-documented pattern of
phenotypic associations between these lipids and lipoproteins. Linkage and other statistical genetic studies are currently underway to test and distinguish between these hypotheses in this population.
Overall, these results suggest that correlations between
LDL size, plasma TG, and HDL-C are strongly influenced by
shared genes and thus may be jointly involved in genetic
susceptibility to CVD. Localizing and identifying these
gene(s) may lead to a better understanding of the role of LDL
size, plasma TG, and HDL-C in susceptibility to CVD in
these high-risk families.
Acknowledgments
This research was supported by National Institutes of Health grants
HL49513 (to M.A.A.) and DK09645 (to K.L.E.) and was performed
during Dr Austin’s tenure as an Established Investigator of the
American Heart Association. We would like to thank the families for
their participation in this study. The authors also wish to thank Andy
Louie for his technical assistance.
References
1. Gardner CD, Fortmann SP, Krauss RM. Association of small low-density
lipoprotein particles with the incidence of coronary artery disease in men
and women. JAMA. 1996;276:875– 881.
2. Lamarche B, Tchernof A, Mauriege P, Cantin B, Dagenais GR, Lupien
PJ, Despres JP. Fasting insulin and apolipoprotein B levels and lowdensity lipoprotein particle size as risk factors for ischemic heart disease.
JAMA. 1998;279:1955–1961.
3. Stampfer MJ, Krauss RM, Ma J, Blanche PJ, Holl LG, Sacks FM,
Hennekens CH. A prospective study of triglyceride level, low-density
lipoprotein particle diameter, and risk of myocardial infarction. JAMA.
1996;276:882– 888.
4. Criqui MH, Heiss G, Cohn R, Cowan LD, Chirayath MS, Bangdiwala S,
Kritchevsky S, Jacobs DR Jr, O’Grady HK, Davis CE. Plasma triglycer-
Edwards et al
5.
6.
7.
8.
9.
10.
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Quantitative Genetic Analysis of Lipids/Lipoproteins
ide level and mortality from coronary heart disease. N Engl J Med.
1993;328:1220 –1225.
Hokanson JE, Austin MA. Plasma triglyceride level is a risk factor for
cardiovascular disease independent of high-density lipoprotein cholesterol level: a meta-analysis of population-based prospective studies. J Cardiovasc Risk. 1996;3:213–219.
Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke
JD, Jacobs DR Jr, Bangdiwala S, Tyroler HA. High-density lipoprotein
cholesterol and cardiovascular disease: four prospective American
studies. Circulation. 1989;79:8 –15.
NIH Consensus Development Panel on Triglyceride, H-DL, and Coronary Heart Disease. Triglyceride, high-density lipoprotein, and coronary
heart disease. JAMA. 1993;269:505–510.
Brunzell JD, Schrott HG, Motulsky AG, Bierman EL. Myocardial
infarction in the familial forms of hypertriglyceridemia. Metabolism.
1976;25:313–320.
Genest J, Martin-Munley S, McNamara J, Ordovas J, Jenner J, Myers R,
Silberman S, Wilson P, Salem D, Schaefer E. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation.
1992;85:2025–2033.
Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG.
Hyperlipidemia in coronary heart disease, II: genetic analysis of lipid
levels in 176 families and delineation of a new inherited disorder,
combined hyperlipidemia. J Clin Invest. 1973;52:1544 –1568.
Schaefer EJ, Genest JJ Jr, Ordovas JM, Salem DN, Wilson PW. Familial
lipoprotein disorders and premature coronary artery disease. Atherosclerosis. 1994;108(suppl):S41–S54.
Jarvik GP, Brunzell JD, Austin MA, Krauss RM, Motulsky AG, Wijsman
E. Genetic predictors of FCHL in 4 large pedigrees: influence of apoB
level major locus predicted genotype and LDL subclass phenotype. Arterioscler Thromb. 1994;14:1687–1694.
Cullen P, Farren B, Scott J, Farrall M. Complex segregation analysis
provides evidence for a major gene acting on serum triglyceride levels in
55 British families with familial combined hyperlipidemia. Arterioscler
Thromb. 1994;14:1233–1249.
Austin MA, Newman B, Selby JV, Edwards K, Mayer EJ, Krauss RM.
Genetics of LDL subclass phenotypes in women twins: concordance,
heritability, and commingling analysis. Arterioscler Thromb. 1993;13:
687– 695.
Austin MA, Jarvik GP, Hokanson JE, Edwards K. Complex segregation
analysis of LDL peak particle diameter. Genet Epidemiol. 1993;10:
599 – 604.
Austin MA, Brunzell JD, Fitch WL, Krauss RM. Inheritance of low
density lipoprotein subclass patterns in familial combined hyperlipidemia. Arteriosclerosis. 1990;10:520 –530.
Bredie SJ, Kiemeney LA, de Haan AF, Demacker PN, Stalenhoef AF.
Inherited susceptibility determines the distribution of dense low-density
lipoprotein subfraction profiles in familial combined hyperlipidemia.
Am J Hum Genet. 1996;58:812– 822.
de Graaf J, Swinkels DW, de Haan AF, Demacker PN, Stalenhoef AF.
Both inherited susceptibility and environmental exposure determine the
low-density lipoprotein-subfraction pattern distribution in healthy Dutch
families. Am J Hum Genet. 1992;51:1295–1310.
Coresh J, Beaty TH, Prenger VL, Xu J, Kwiterovich PO Jr. Segregation
analysis of HDL3-C levels in families of patients undergoing coronary
arteriography at an early age. Arterioscler Thromb Vasc Biol. 1995;15:
1307–1313.
Heller DA, de Faire U, Pedersen NL, Dahlen G, McClearn GE. Genetic
and environmental influences on serum lipid levels in twins. N Engl
J Med. 1993;328:1150 –1156.
Mahaney MC, Blangero J, Comuzzie AG, VandeBerg JL, Stern MP,
MacCluer JW. Plasma HDL cholesterol, triglycerides, and adiposity: a
quantitative genetic test of the conjoint trait hypothesis in the San Antonio
Family Heart Study. Circulation. 1995;92:3240 –3248.
Mahaney MC, Blangero J, Rainwater DL, Comuzzie AG, VandeBerg JL,
Stern MP, MacCluer JW, Hixson JE. A major locus influencing plasma
high-density lipoprotein cholesterol levels in the San Antonio Family
Heart Study: segregation and linkage analyses. Arterioscler Thromb Vasc
Biol. 1995;15:1730 –1739.
Austin MA, Mykkanen L, Kuusisto J, Edwards KL, Nelson C, Haffner
SM, Pyorala K, Laakso M. Prospective study of small LDLs as a risk
factor for non-insulin dependent diabetes mellitus in elderly men and
women. Circulation. 1995;92:1770 –1778.
Burchfiel CM, Curb JD, Sharp DS, Rodriguez BL, Arakaki R, Chyou PH,
Yano K. Distribution and correlates of insulin in elderly men: the
Honolulu Heart Program. Arterioscler Thromb Vasc Biol. 1995;15:
2213–2221.
2463
25. Edwards KL, Austin MA, Newman B, Mayer E, Krauss RM, Selby JV.
Multivariate analysis of the insulin resistance syndrome in women. Arterioscler Thromb. 1994;14:1940 –1945.
26. Sprecher DL, Feigelson HS, Laskarzewski PM. The low HDL cholesterol/high triglyceride trait. Arterioscler Thromb. 1993;13:495–504.
27. Sprecher DL, Hein MJ, Laskarzewski PM. Conjoint high triglycerides
and low HDL cholesterol across generations: analysis of proband hypertriglyceridemia and lipid/lipoprotein disorders in first-degree family
members. Circulation. 1994;90:1177–1184.
28. Austin MA, King MC, Vranizan KM, Krauss RM. Atherogenic
lipoprotein phenotype: a proposed genetic marker for coronary heart
disease risk. Circulation. 1990;82:495–506.
29. Austin MA, Rodriguez BL, McKnight B, Curb JD, Sharp DS. Lowdensity lipoprotein (LDL) particle size and plasma triglyceride as predictors of coronary heart disease in older Japanese-American men. Circulation. 1997;96(suppl I):I-337. Abstract.
30. Edwards KL, Newman B, Mayer E, Selby JV, Krauss RM, Austin MA.
Heritability of factors of the insulin resistance syndrome in women twins.
Genet Epidemiol. 1997;14:241–253.
31. Edwards KL, Talmud PJ, Luong LA, Haddad L, Day INM, Newman B,
Louie AW, Krauss RM, Humphries SE, Austin MA. Sib-pair linkage
analysis of factors of the insulin resistance syndrome (IRS) in dizygotic
women twins. Can J Cardiol. 1997;13(suppl B):165B. Abstract.
32. Austin MA, McKnight B, Edwards KL, Wallace KJ, Bradley CM,
McNeely MJ, Psaty BM, Brunzell JD, Motulsky AG. Twenty-year cardiovascular disease (CVD) mortality in familial combined hyperlipidemia
and familial hypertriglyceridemia: an historical cohort study. Circulation.
1997;96(suppl I):I-142. Abstract.
33. Austin MA. Plasma triglyceride and coronary heart disease: epidemiologic, statistical, and genetic issues. In: Proceedings of the International
Congress of Triglycerides: The Role in Diabetes and Atherosclerosis.
Athero Rev. 1991:65–70.
34. Lipid Research Clinics Program Manual of Laboratory Operations. Lipid
and Lipoprotein Analysis. Bethesda, Md: National Institutes of Health;
1974.
35. Nichols AV, Krauss RM, Musliner TA, eds. Nondenaturing Polyacrylamide Gradient Gel Electrophoresis. Orlando, Fla: Academic Press;
1986.
36. Krauss RM, Burke DJ. Identification of multiple subclasses of plasma
low density lipoproteins in normal humans. J Lipid Res. 1982;23:97–104.
37. Dyke B. PEDSYS: A Pedigree Data Management System. San Antonio,
Tex: Population Genetics Laboratory, Southwest Foundation for Biomedical Research; 1992.
38. Blangero J, Konigsberg L. Multivariate segregation analysis using the
mixed model. Genet Epidemiol. 1991;8:299 –316.
39. Falconer DS. Introduction to Quantitative Genetics. 3rd ed. Essex, UK:
Longman; 1989.
40. Boehnke M, Moll PP, Kottke BA, Weidman WH. Partitioning the variability of fasting plasma glucose levels in pedigrees: genetic and environmental factors. Am J Epidemiol. 1987;125:679 – 689.
41. Comuzzie AG, Blangero J, Mahaney MC, Mitchell BD, Stern MP,
MacCluer JW. Genetic and environmental correlations among skinfold
measures. Int J Obes Relat Metab Disord. 1994;18:413– 418.
42. Hopper JL, Mathews JD. Extensions to multivariate normal models for
pedigree analysis. Ann Hum Genet. 1982;46:373–383.
43. Lange K, Boehnke M. Extensions to pedigree analysis, IV: covariance
components models for multivariate traits. Am J Med Genet. 1983;14:
513–524.
44. Friedlander Y, Austin MA, Newman B, Edwards K, Mayer D-EI, King
MC. Heritability of longitudinal changes in coronary-heart-disease risk
factors in women twins. Am J Hum Genet. 1997;60:1502–1512.
45. Christian JC, Carmelli D, Castelli WP, Fabsitz R, Grim CE, Meaney FJ,
Norton JA Jr, Reed T, Williams CJ, Wood PD. High density lipoprotein
cholesterol: a 16-year longitudinal study in aging male twins. Arteriosclerosis. 1990;10:1020 –1025.
46. Heller DA, Pedersen NL, de Faire U, McClearn GE. Genetic and environmental correlations among serum lipids and apolipoproteins in elderly
twins reared together and apart. Am J Hum Genet. 1994;55:1255–1267.
47. Castellani LW, Weinreb A, Bodnar J, Goto AM, Doolittle M, Mehrabian
M, Demant P, Lusis AJ. Mapping a gene for combined hyperlipidaemia
in a mutant mouse strain. Nat Genet. 1998;18:374 –377.
48. Pajukanta P, Nuotio I, Terwilliger JD, Porkka KV, Ylitalo K, Pihlajamaki
J, Suomalainen AJ, Syvanen AC, Lehtimaki T, Viikari JS, Laakso M,
Taskinen MR, Ehnholm C, Peltonen L. Linkage of familial combined
hyperlipidaemia to chromosome 1q21– q23. Nat Genet. 1998;18:
369 –373.
2464
Arterioscler Thromb Vasc Biol.
October 1999
49. Dallinga-Thie GM, van Linde-Sibenius Trip M, Rotter JI, Cantor RM, Bu
X, Lusis AJ, de Bruin TW. Complex genetic contribution of the Apo
AI-CIII-AIV gene cluster to familial combined hyperlipidemia: identification of different susceptibility haplotypes. J Clin Invest. 1997;99:
953–961.
50. Austin MA, Talmud PJ, Luong LA, Haddad L, Day IN, Newman B,
Edwards KL, Krauss RM, Humphries SE. Candidate-gene studies of the
atherogenic lipoprotein phenotype: a sib-pair linkage analysis of DZ
women twins. Am J Hum Genet. 1998;62:406 – 419.
51. Rotter JI, Bu X, Cantor RM, Warden CH, Brown J, Gray RJ, Blanche PJ,
Krauss RM, Lusis AJ. Multilocus genetic determinants of LDL particle
size in coronary artery disease families. Am J Hum Genet. 1996;58:
585–594.
52. Houlston RS, Snowden C, Laker MF, Alberti KG, Humphries SE.
Variation in the apolipoprotein B gene and development of type 2
diabetes mellitus. Dis Markers. 1991;9:87–96.
53. Nishina PM, Johnson JP, Naggert JK, Krauss RM. Linkage of atherogenic
lipoprotein phenotype to the low density lipoprotein receptor locus on the
54.
55.
56.
57.
short arm of chromosome 19. Proc Natl Acad Sci U S A. 1992;89:
708 –712.
de Bruin TW, Mailly F, van Barlingen HH, Fisher R, Castro Cabezas M,
Talmud P, Dallinga-Thie GM, Humphries SE. Lipoprotein lipase gene
mutations D9N and N291S in four pedigrees with familial combined
hyperlipidaemia. Eur J Clin Invest. 1996;26:631– 639.
Dallinga-Thie GM, Bu XD, van Linde-Sibenius Trip M, Rotter JI, Lusis
AJ, de Bruin TW. Apolipoprotein A-I/C-III/A-IV gene cluster in familial
combined hyperlipidemia: effects on LDL-cholesterol and apolipoproteins B and C-III. J Lipid Res. 1996;37:136 –147.
Hokanson JE, Brunzell JD, Jarvik GP, Wijsman EM, Austin MA. Linkage
of low-density lipoprotein size to the lipoprotein lipase gene in heterozygous lipoprotein lipase deficiency. Am J Hum Genet. 1999;64:
608 – 618.
Comuzzie AG, Mahaney MC, Almasy L, Dyer TD, Blangero J.
Exploiting pleiotropy to map genes for oligogenic phenotypes using
extended pedigree data. Genet Epidemiol. 1997;14:975–980.
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
Downloaded from http://atvb.ahajournals.org/ by guest on June 18, 2017
Pleiotropic Genetic Effects on LDL Size, Plasma Triglyceride, and HDL Cholesterol in
Families
Karen L. Edwards, Michael C. Mahaney, Arno G. Motulsky and Melissa A. Austin
Arterioscler Thromb Vasc Biol. 1999;19:2456-2464
doi: 10.1161/01.ATV.19.10.2456
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272
Greenville Avenue, Dallas, TX 75231
Copyright © 1999 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/19/10/2456
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the
Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for
which permission is being requested is located, click Request Permissions in the middle column of the Web
page under Services. Further information about this process is available in the Permissions and Rights
Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/