Chapter 9. Is the major gene effect for LDL peak particle diameter on 17q caused by the apolipoprotein H gene?

Yohan Bossé, Mary F. Feitosa, Jean-Pierre Després, Benoît Lamarche, Treva Rice, D.C. Rao, Claude Bouchard, Louis Pérusse, Marie-Claude Vohl.

Des résultats antérieurs de l’Étude des familles de Québec ont révélé que le diamètre principal des particules LDL (DP-LDL) est semblable à l’intérieur des familles avec un cœfficient d’héritabilité estimé à plus de 50% et la présence d’un locus quantitatif majeur localisé sur le chromosome 17q. Dans cette étude on démontre, par analyse de ségrégation complexe, la présence d’un gène à effet majeur expliquant 52% de la variance du DP-LDL ajusté pour l’âge, l’indice de masse corporelle et les triglycérides. En séquencant le gène de l’apolipoprotéine H, localisé sur le chromosome 17q, trois mutations faux-sens ont été identifiées. Un haplotype particulier (fréquence = 20.9%) était associé avec des valeurs du DP-LDL plus élevé (p = 0.046). Ces résultats suggèrent que le DP-LDL est influencé par un gène à effet majeur et que le signal de liaison observé antérieurement sur le chromosome 17q pourrait être causé par le gène de l’apolipoprotéine H.

Is the major gene effect for LDL peak particle diameter on 17q caused by the apolipoprotein H gene?

Yohan Bossé1,2, Mary F. Feitosa3, Jean-Pierre Després1,2,4, Benoît Lamarche1,5, Treva Rice3, D.C. Rao3, Claude Bouchard6, Louis Pérusse7, and Marie-Claude Vohl1,2

1-Lipid Research Center, CHUL Research Center; 2- Department of Food Science and Nutrition; 3- Division of Biostatistics, Washington University School of Medicine, St. Louis, Missouri; 4- The Quebec Heart Institute; 5- Institute on Nutraceutical and Functional Food; 6- Pennington Biomedical Research Center, Baton Rouge, Louisiana, USA; 7- Division of Kinesiology, Department of Social and Preventive Medicine, Laval University, Quebec, Canada.

Running title: Apolipoprotein H and LDL size.

Address all correspondence to:

Marie-Claude Vohl, Ph.D., Lipid Research Center, CHUL Research Center, TR-93, 2705 Laurier Blvd, Sainte-Foy, Quebec, G1V 4G2, Canada. Phone: (418) 656-4141 extension 48280, Fax: (418) 654-2145, E-mail: marie-claude.vohl@crchul.ulaval.ca

Number of figure: 1

Number of tables: 5

Abstract

Low-density lipoprotein (LDL) size, a coronary heart disease risk factor, is influenced by both genetic and environmental factors. Results from the Quebec Family Study (QFS) revealed that the LDL peak particle diameter (LDL-PPD) aggregates in families with a heritability coefficient above 50% and is affected by a major quantitative trait locus on chromosome 17q (LOD = 6.8). Complex segregation analyses have consistently demonstrated a major gene effect influencing LDL size. In the present study, we report a similar analysis in the QFS cohort, which suggests that a major gene explains 52% of the variance in age- body mass index- and triglyceride-adjusted LDL-PPD. The most intuitive positional candidate gene on chromosome 17q is the apolipoprotein H gene. Direct sequencing of the promoter, coding regions, and exon-intron splicing boundaries of this gene revealed the presence of three missense mutations and two polymorphisms in the untranslated regions. Using family-based association tests, none of these variants was individually associated with LDL-PPD. However, analysis of the haplotypes constructed from the three missense mutations, suggested that one particular haplotype (frequency = 20.9%) was associated with a significant increase in LDL-PPD trait values (p = 0.046). Taken together these results suggest the presence of a major gene influencing LDL-PPD and that the linkage signal previously observed on chromosome 17q may be related to the apolipoprotein H gene. Replication of the positive association between apolipoprotein H gene haplotype and LDL-PPD is warranted.

Key words: LDL peak particle diameter, apolipoprotein H, segregation analysis, family-based association test, haplotypes.

Introduction

There is considerable evidence suggesting that the presence of an increased proportion of small, dense low-density lipoprotein (LDL) particles is predictive of an increased risk of coronary heart disease (CHD)[1]. Individual variation in this new CHD risk marker has been known to be in part attributable to a number of environmental influences including, among others, dietary factors[2] and physical activity[3]. However, genetic studies have clearly demonstrated that a large part of the trait variability lies in the genes. A large number of studies have been conducted to elucidate the genetic architecture underlying the phenotype and have been the topic of a recent review[4].

Data from the Quebec Family Study (QFS) have recently confirmed the importance of genetic factors for LDL size and have produced new leads that need to be followed-up. Heritability estimates performed on QFS data suggested coefficient above 50% for LDL peak particle diameter (LDL-PPD), a quantitative trait reflecting the size of the major LDL subclass. In addition, the pattern of familial correlations revealed no significant spouse correlations but significant parent-offspring and sibling correlations, suggesting that the familial resemblance is primarily attributable to genetic factors[5]. An autosomal genomewide linkage scan was performed in order to identify the gene responsible for this genetic contribution[6]. A major quantitative trait locus (QTL) was observed on chromosome 17q for LDL-PPD adjusted for age, body mass index (BMI) and triglyceride levels (LOD = 6.8). Signals of lesser magnitude were also observed with the same phenotype on 1p, 2q, 4p, 5q and 14q, with LOD scores of 2.6, 2.3, 2.1, 2.1 and 1.7, respectively. Distinct genomic regions captured by genomewide linkage scans were also reported among families ascertained through hyperlipidemic proband, including 6q by Austin et al.[7], 15q by Allayee et al.[8] as well as 9p, 11q, 14q and 16q by Badzioch et al.[9]. Interestingly, these QTLs harbor a large number of candidate genes for LDL-PPD that have not been tested previously in linkage and association studies, and thus, provided new leads that require follow-up.

The results of the QFS genome scan generated two hypotheses that are being tested in the present study. First, the presence of a QTL affecting LDL-PPD is consistent with the major gene effect reported in segregation studies[10-16]. We speculate that the putative major gene effect observed in the later studies is responsible for the QTL on chromosome 17q. Thus we tested whether the single gene effect was also observed in QFS using complex segregation analysis. Then we have chosen for further testing the most obvious candidate gene located on chromosome 17q, namely the apolipoprotein H gene (APOH). ApoH is a single-chain glycoprotein that exists in plasma both in a free form and in combination with lipoprotein particles. It has been implicated in several pathways, including lipid metabolism[17-20]. The second objective was thus to verify whether sequence variation in the APOH gene is responsible for the linkage signal observed on 17q.

Methods

Population

The QFS is an ongoing project of French Canadian families living in and around the Quebec City area with the aim of studying the genetics of obesity and its comorbidities[21]. This cohort represents a mixture of random sampling and ascertainment through obese (body mass index > 32 kg/m2) probands. For the present study, LDL size characteristics were available for 680 subjects members of 236 nuclear families. The characteristics of these subjects by sex and generation groups are shown in Table 1. The QFS has been approved by the ethics committee of Laval University and all study participants provided written informed consent.

LDL peak particle diameter (LDL-PPD)

LDL-PPD was measured by gradient gel electrophoresis from plasma obtained after a 12-hours fast. Briefly, the whole plasma was loaded on nondenaturing 2-16% polyacrylamide gradient gels and exposed to electrophoresis for a prerun of 20 minutes at 70 V followed by migration at 175 V for 4 hours. Gels were then stained with sudan black, destained and size restored as described previously[22]. Gels were subsequently scanned and visualized on an electropherogram with every peak reflecting a band. The size of particles forming the bands was determined on the basis of a calibration curve constructed from the plasma standards. The estimated size of the major band was identified as the LDL-PPD. This phenotype was adjusted for age (up to the cubic polynomial), age and BMI or age, BMI and triglyceride levels. Data adjustments were performed within each of the six age-by-sex groups (<30, 30 to 50, and ≥50 years; male and female) using a stepwise multiple regression described previously[6].

Segregation analysis

Univariate segregation analysis was conducted using the Pedigree Analysis Package (PAP), version 5.0[23]. The mixed Mendelian model (model 1) assumes that a phenotype is influenced by the independent and additive contributions from a major gene, a polygenic/multifactorial background, and a nontransmitted environmental component. The major gene is biallelic (A, a), where the upper case allele, with frequency p, is associated with lower phenotypic values. The other parameters in the model are: the mean values for the three genotypes (μAA, μAa, μaa, where the order of the means is constrained to be μAA ≤ μAa ≤ μaa); the common standard deviation within major locus genotypes (σ); the residual polygenic heritability (H), after accounting for the major gene effect; and parent-to-offspring transmission probabilities for the three genotypes (τAA, τ Aa, and τ aa). For a single diallelic locus, the three τ's denote the probability of transmitting allele A for genotypes AA , Aa , and aa , with Mendelian expectations of 1, 1/2, and 0, respectively. When the three τ values are equal, no transmission of the major effect is obtained. All parameters were estimated using a maximum likelihood method.

The general model (model 1) and thirteen reduced models (models 2-14) testing specific hypotheses were fitted to the data. The specific hypotheses tested in each model are the following. Three models tested the presence of familial components including a general familial component (model 2), a major gene component (model 3) and a multifactorial component (model 4). The modes of transmission were tested with recessive (model 5) and dominant (model 6) models. Two models (model 7 and 8) tested if the major effect follows Mendelian transmission probabilities (τAA, τAa, τaa). Finally, recessive and dominant mode inheritance were also tested under the mixed (model 9 and 10), free (model 11 and 13) and equal (model 12 and 14) τ models. Nested models were tested against the mixed Mendelian model (model 1) or against each others as indicated in the last column of Table 3 using a likelihood ratio test which is the difference between the two models compared in minus twice the log-likelihoods ( -2 ln L ). The most parsimonious model of those not rejected by likelihood ratio test was determined using Akaike's Information Criterion (AIC)[24], which is computed as minus twice the log likelihood of the model plus twice the number of parameters estimated. The model with the lowest AIC indicates the most parsimonious fit to the observed data.

To claim a major gene effect with this approach, the following criteria have to be met. First, the no major gene effect (model 3) and the equal transmission probabilities (model 8) models have to be rejected. In addition, the free transmission probabilities model (model 7) has to be non-rejected. The variance accounted for by the major gene (σ2 mg) was derived from this equation (μAA - μ0)2p2 + (μAa - μ0)22p(1-p) + (μaa - μ0)2(1-p)2, where μAA, μAa, μaa, and p are estimated in the parsimonious model and μ0 are derived from the following equation μAAp2 + μAa2p(1-p) + μaa(1-p)2. The multifactorial heritability (H) estimated in the model is expressed as a function of the common residual variance (σ2). The multifactorial heritability expressed as the percentage of the total phenotypic variance (h2) can be computed using the equation (Hσ2)/(σ2 + σ2 mg).

Sequencing and genotyping of APOH gene

The promoter, the coding regions and the exon-intron splicing boundaries of the APOH gene were sequenced in 28 subjects having LDL-PPD in both extreme of the distribution (small < 254 Å and large > 276 Å). All exons and exon-intron splicing boundaries were amplified from genomic DNA by use of specific primers derived from the 5’ and 3’ ends of intronic sequence. We also sequenced up to 631 base pairs located downstream of the ATG start codon since consensus sequence elements have been localized in that region[25]. Table 2 presents the specific primers for each fragment with their product size. All primers were designed using the Primer 3.0 software available on the Whitehead Insitute/MIT Center for Genome Research server (http://www-genome.wi.mit.edu/cgi-bin/primer/primer3.cgi). Amplification was performed by polymerase chain reaction using the thermal cycler, model PTC-200 (MJ Research, Watertown, MA). PCR products were purified by the ABI ethanol-EDTA precipitation protocol, collected using a Beckman-Coulter Allegra 6R centrifuge, and resuspended in a 50% HiDi-formamide solution. Sequence reactions were performed using the BigDyeTH Terminator v3.1 kit and samples were run on ABI Prism® 3730/XL DNA Analyzer automated sequencer (Applied Biosystems, Foster City, CA). Sequences were then assembled and analyzed using the Staden preGAP4 and GAP4 programs[26]. Genetic variants were subsequently genotyped on the whole cohort using a mini-sequencing assay[27].

Association tests

The association between LDL-PPD and APOH variants was tested using two different statistical approaches. First, the independent effect of individual polymorphisms was tested by comparing the mean phenotype values between genotype groups using the MIXED procedure implemented in SAS (version 8.2), which takes the nonindependence of family members into account. The phenotypes were adjusted for age, gender, BMI and triglyceride levels prior to the association analyses. Secondly, we used the family-based association test (FBAT) when evaluating the association with single SNPs or haplotypes and LDL-PPD[28, 29] (http://www.biostat.harvard.edu/~fbat/default.html). The FBAT program performs family-based tests of association that are efficient and robust to population admixture, phenotype distribution and ascertainment based on phenotype. It can also handle missing parental genotypes and/or missing phase in both offspring and parents for haplotype analysis. The approach holds as well for multi-locus and multi-allelic markers. The haplotype test is ideal for candidate gene studies with tightly linked markers (no or little recombination between the markers). To test for the effect of a transmited allele on the trait values, an univariate FBAT test was performed for each allele. This test provides a Z-statistic with the corresponding p-value. A positive Z-statistic is indicative of an increasing trait value allele while a negative Z-statistic is indicative of a lowering trait value allele. This univariate FBAT statistic (Z-statistic) was also used to make inference regarding the effect of APOH haplotypes on LDL-PPD.

Family-based association tests were performed on LDL-PPD with and without adjustment for confounding factors. Adjustments were performed using a stepwise multiple regression procedure taking into account age, BMI and triglyceride levels as described previously[6]. The residuals, standardized to a mean of 0 and a SD of 1, were then used for statistical tests.

Results

Segregation analysis

Segregation analyses were performed on age-adjusted, age-BMI-adjusted, and age-BMI-triglyceride-adjusted LDL-PPD. In general, the results are quite consistent and only the results for the age-BMI-triglyceride-adjusted LDL-PPD are presented in Table 3. For the three phenotypes, the hypotheses of no familial resemblance (model 2), no major gene effect (model 3), and no multifactorial effect (model 4) are rejected, suggesting the presence of both a major gene and a multifactorial effects. In addition, the equal τ’s hypotheses (model 8, 12 and 14) are rejected and the free τ’s hypotheses (model 7, 11 and 13) are not rejected for the mixed, recessive mixed and dominant mixed, respectively, for the three phenotypes. Thus supporting that the trait is transmitted from parents to offspring and the transmitted effect is Mendelian in nature. In all cases, the mixed dominant Mendelian model (model 10) was not rejected and best fit the data according to the AIC values and was chosen as the most parsimonious model. With this parsimonious model, the variances accounted for by the major gene and the multifactorial component are as follow: 14% and 37% for the age-adjusted LDL-PPD; 14% and 34% for age-BMI LDL-PPD; 23% and 34% for age-BMI-triglyceride LDL-PPD.

Genetic variants in the APOH gene

A total of five genetic variants were identified in the APOH gene. Two were located within the untranslated regions of exon 1 and exon 8, namely -32C>A (rs8178822) and c.1059C>T (rs6933), respectively. The others were missense mutations found in exon 3 (rs1801692), 7 (rs4581) and 8 (rs1801690). The later have been identified before and are referred to as S88N, V247L and W316S. The frequencies of the minor alleles in the full family sample are 0.08, 0.04, 0.24, 0.06, and 0.44 for -32C>A, S88N, V247L, W316S and c.1059C>T, respectively. Genotype distributions for all variants are in Hardy-Weinberg equilibrium among genetically unrelated individuals (p > 0.1). Figure 1 shows the genomic organization of the APOH gene and the locations of the five genetic variants.

Genetic association analyses

Associations were first tested by comparing the mean phenotypic values between genotype groups for each genetic variant (Table 4). No significant association was observed for any of the genetic variants. For the W316S variant, only one subject was homozygous for the rare allele and was excluded from analyses.

Associations between single DNA variants and the LDL-PPD phenotype were also tested using the FBAT program. No single DNA variant showed significant association with LDL size at the 0.05 significance level (not shown). However, the number of informative families was relatively low (particularly for DNA variants with a low rare frequency allele) and may not have provided sufficient statistical power to detect an effect. We therefore used an haplotype family-based association test that took into account the genotype information of three missense mutations (S88N, V247L and W316S). Table 5 contains the haplotype patterns and frequency of the four most informative haplotypes. For each haplotype, the Z-statistic and the corresponding p value are given for LDL-PPD with and without adjustment for covariates. Haplotypes 88S/247V/316W (APOH_1), 88S/247L/316W (APOH_2), 88S/247V/316S (APOH_3) and 88N/247V/316W (APOH_4) have frequencies of 0.68, 0.22, 0.07 and 0.04, respectively. Haplotype APOH_1, APOH_3 and APOH_4 had a lowering effect on LDL-PPD phenotypes as indicated by negative Z-statistics. However, this lowering effect was not statistically significant. In contrast, haplotype APOH_2, which consists of the rare allele at V247L (exon7), and wild-type alleles at S88N (exon 3) and W316S (exon 8), was associated with a significantly greater LDL-PPD trait values. This effect was significant whether LDL-PPD was adjusted for covariates or not.

Discussion

The objectives of the present study were to test for the presence of a single gene with major effect on LDL-PPD, and to test whether genetic variants in the APOH gene was associated with this LDL phenotype. These objectives were motivated by studies consistently showing the presence of a major gene effect on LDL particle size and density[10-16], and a recent autosomal genomic scan on LDL-PPD performed in QFS[6]. Results of this genome scan revealed several QTLs with the strongest signal on 17q[6]. Based on these results, it was tempting to speculate that the putative gene detected by prior segregation analyses was located at this genomic location. We thus verified whether a major gene effect could also be detected in the QFS cohort and tested a strong candidate gene located in the 17q region, namely APOH.

Complex segregation analyses provided strong evidence suggesting the existence of a major gene effect influencing LDL-PPD. An interesting observation is that the major gene effect is amplified when LDL-PPD was adjusted for triglyceride levels (explained 52% of the variance vs 24%). This observation may simply be explained by a major gene effect that remain constant but act on a reduce variance cause by triglyceride adjustment. On the other hand, this observation may also suggest the presence of a pleiotropic gene having an effect on both LDL size and triglyceride, in addition to a gene affecting LDL size. Thus, it is possible that there are two genes, one with a pleiotropic effect on both LDL size and triglyceride, the other affecting purely LDL size.

In the present study, we used haplotypes derived from the FBAT program to test for association between LDL-PPD and genetic variants in the APOH gene. Although no single genetic variant showed a significant association with LDL-PPD, a haplotype-univariate test performed with three missense mutations revealed that one particular haplotype was associated with higher trait values. Despite its significance, this result should be interpreted with caution since no correction was applied for the number of tests performed. However, we believe that this association is likely to be true.

Testing for association using an haplotype approach is appropriate, especially when the informativeness of individual markers is low. Haplotype testing in the context of family studies is still in the developmental stage and computer programs have just become available[29, 30]. Considering the number of genetic variants identified in the present study, the most appropriate association test was a haplotype test for family-based study. If we had elected to use only this test, the issue of multiple testing would not have arisen and the observed association would have been declared significant albeit at a low level.

Moreover, it can be argued that the association between the APOH gene and LDL-PPD has a physiological rationale. ApoH, also known as β2-glycoprotein I, is a single chain polypeptide of 326 amino acids synthesized in the liver. The plasma concentration of apoH differs significantly among individuals, ranging from levels that are undetectable to levels as high as 35 mg/dl, with means of 20 mg/dl in Whites and 15 mg/dl in Blacks[18]. The protein exists in plasma both in a free form and in combination with lipoprotein particles, including VLDL, HDL and chylomicrons. Its role in lipoprotein metabolism is not fully understood, but it was shown to activate lipoprotein lipase[17] and clear triglycerides from plasma[31].

The missense mutations tested in the present study have been shown to be biologically functional and associated with lipid values. These variants explained a significant portion of the variation in apoH levels[18] and have also been associated with triglyceride[19, 20, 32], VLDL-C[32] and HDL-C[18] levels in some subgroups. In contrast, other studies have shown a lack of association between APOH polymorphisms and lipid traits[33-36].

Finally, an additional argument for the merit of the present finding comes from the fact that the APOH gene is located in a relatively small genomic region that has been shown to be linked to the LDL-PPD phenotype. The genomic scan signal observed on 17q makes the existence of a positional gene associated with the phenotype quite likely. However, considering the lack of association with the individual variants, we cannot exclude the possibility that the haplotype significantly associated with LDL-PPD is in linkage disequilibrium with a causative variant located elsewhere in a nearby gene.

In conclusion, complex segregation analysis supported the existence of a major gene for LDL size in QFS. This finding is consistent with similar studies performed so far[10-16]. By sequencing the promoter, splicing boundaries, and exons of a positional candidate gene, the APOH gene, we identified and genotyped five genetic variants including three missense mutations previously reported. By means of a family-based haplotype analysis, we identified a haplotype associated with larger LDL particle size. These results suggest that the APOH gene is responsible for the QTL observed earlier on 17q and for the major gene effect detected by segregation analysis. However, replication in independent cohorts is required to secure these conclusions.

Acknowledgment

This study was supported by the Canadian Institutes of Health Research (MOP-13960 and MOP-44074). The authors would like to express their gratitude to the subjects for their excellent collaboration and to the staff of the Physical Activity Sciences Laboratory for their contribution to the study. Y. Bossé is the recipient of a Canada Graduate Scholarship Doctoral Awards. M.C. Vohl is a research scholar of the “Fonds de la recherche en santé du Québec”. J.P. Després is chair professor of human nutrition, lipidology and prevention of cardiovascular disease supported by Provigo and Pfizer Canada. B. Lamarche is a recipient of the Canada Research Chair in Nutrition, Functional Foods and Cardiovascular Health. C. Bouchard is partially funded by the George A. Bray Chair in Nutrition.

References

[1] Lamarche B, Lemieux I, Despres JP. The small, dense LDL phenotype and the risk of coronary heart disease: epidemiology, patho-physiology and therapeutic aspects. Diabetes Metab 1999; 25:199-211.

[2] Krauss RM. Dietary and genetic effects on low-density lipoprotein heterogeneity. Annu Rev Nutr 2001; 21:283-95.

[3] Kraus WE, Houmard JA, Duscha BD, Knetzger KJ, Wharton MB, McCartney JS, Bales CW, Henes S, Samsa GP, Otvos JD, Kulkarni KR, Slentz CA. Effects of the amount and intensity of exercise on plasma lipoproteins. N Engl J Med 2002; 347:1483-92.

[4] Bosse Y, Perusse L, Vohl MC. Genetics of LDL particle heterogeneity: from genetic epidemiology to DNA-based variations. J Lipid Res 2004; 45:1008-26.

[5] Bosse Y, Vohl MC, Despres JP, Lamarche B, Rice T, Rao DC, Bouchard C, Perusse L. Heritability of LDL peak particle diameter in the Quebec Family Study. Genet Epidemiol 2003; 25:375-81.

[6] Bosse Y, Perusse L, Despres JP, Lamarche B, Chagnon YC, Rice T, Rao DC, Bouchard C, Vohl MC. Evidence for a major quantitative trait locus on chromosome 17q21 affecting low-density lipoprotein peak particle diameter. Circulation 2003; 107:2361-8.

[7] Austin MA, Edwards KL, Monks SA, Koprowicz KM, Brunzell JD, Motulsky AG, Mahaney MC, Hixson JE. Genome-wide scan for quantitative trait loci influencing LDL size and plasma triglyceride in familial hypertriglyceridemia. J Lipid Res 2003; 44:2161-8.

[8] Allayee H, Dominguez KM, Aouizerat BE, Krauss RM, Rotter JI, Lu J, Cantor RM, de Bruin TW, Lusis AJ. Contribution of the hepatic lipase gene to the atherogenic lipoprotein phenotype in familial combined hyperlipidemia. J Lipid Res 2000; 41:245-52.

[9] Badzioch MD, Igo RP, Gagnon F, Brunzell JD, Krauss RM, Motulsky AG, Wijsman EM, Jarvik GP. Low-Density Lipoprotein Particle Size Loci in Familial Combined Hyperlipidemia. Evidence for Multiple Loci From a Genome Scan. Arterioscler Thromb Vasc Biol 2004.

[10] Austin MA, King MC, Vranizan KM, Newman B, Krauss RM. Inheritance of low-density lipoprotein subclass patterns: results of complex segregation analysis. Am J Hum Genet 1988; 43:838-46.

[11] Austin MA, Brunzell JD, Fitch WL, Krauss RM. Inheritance of low density lipoprotein subclass patterns in familial combined hyperlipidemia. Arteriosclerosis 1990; 10:520-30.

[12] Austin MA, Jarvik GP, Hokanson JE, Edwards K. Complex segregation analysis of LDL peak particle diameter. Genet Epidemiol 1993; 10:599-604.

[13] Friedlander Y, Kark JD, Sinnreich R, Edwards KL, Austin MA. Inheritance of LDL peak particle diameter: results from a segregation analysis in Israeli families. Genet Epidemiol 1999; 16:382-96.

[14] Vakkilainen J, Pajukanta P, Cantor RM, Nuotio IO, Lahdenpera S, Ylitalo K, Pihlajamaki J, Kovanen PT, Laakso M, Viikari JS, Peltonen L, Taskinen MR. Genetic influences contributing to LDL particle size in familial combined hyperlipidaemia. Eur J Hum Genet 2002; 10:547-52.

[15] 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; 51:1295-310.

[16] 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; 58:812-22.

[17] Nakaya Y, Schaefer EJ, Brewer HB, Jr. Activation of human post heparin lipoprotein lipase by apolipoprotein H (beta 2-glycoprotein I). Biochem Biophys Res Commun 1980; 95:1168-72.

[18] Mehdi H, Aston CE, Sanghera DK, Hamman RF, Kamboh MI. Genetic variation in the apolipoprotein H (beta2-glycoprotein I) gene affects plasma apolipoprotein H concentrations. Hum Genet 1999; 105:63-71.

[19] Sepehrnia B, Kamboh MI, Adams-Campbell LL, Bunker CH, Nwankwo M, Majumder PP, Ferrell RE. Genetic studies of human apolipoproteins. VIII. Role of the apolipoprotein H polymorphism in relation to serum lipoprotein concentrations. Hum Genet 1989; 82:118-22.

[20] Cassader M, Ruiu G, Gambino R, Guzzon F, Pagano A, Veglia F, Pagni R, Pagano G. Influence of apolipoprotein H polymorphism on levels of triglycerides. Atherosclerosis 1994; 110:45-51.

[21] Bouchard C. Genetic epidemiology, association, and sib-pair linkage: Results from the Québec Family Study. In: Bray GA and Ryan DH, Eds. Molecular and genetic aspects of obesity. Baton Rouge, LA, Louisiana State University Press, 1996: 470-81.

[22] St-Pierre AC, Ruel IL, Cantin B, Dagenais GR, Bernard PM, Despres JP, Lamarche B. Comparison of various electrophoretic characteristics of LDL particles and their relationship to the risk of ischemic heart disease. Circulation 2001; 104:2295-9.

[23]

[24] Akaike H. A new look at the statistical model identification. IEEE Trans Automat Control 1974; 19:716-23.

[25] Okkels H, Rasmussen TE, Sanghera DK, Kamboh MI, Kristensen T. Structure of the human beta2-glycoprotein I (apolipoprotein H) gene. Eur J Biochem 1999; 259:435-40.

[26] Bonfield JK, Smith K, Staden R. A new DNA sequence assembly program. Nucleic Acids Res 1995; 23:4992-9.

[27] Sun X, Ding H, Hung K, Guo B. A new MALDI-TOF based mini-sequencing assay for genotyping of SNPS. Nucleic Acids Res 2000; 28:E68.

[28] Laird NM, Horvath S, Xu X. Implementing a unified approach to family-based tests of association. Genet Epidemiol 2000; 19 Suppl 1:S36-42.

[29] Horvath S, Xu X, Lake SL, Silverman EK, Weiss ST, Laird NM. Family-based tests for associating haplotypes with general phenotype data: application to asthma genetics. Genet Epidemiol 2004; 26:61-9.

[30] Lange C, DeMeo D, Silverman EK, Weiss ST, Laird NM. PBAT: tools for family-based association studies. Am J Hum Genet 2004; 74:367-9.

[31] Wurm H, Beubler E, Polz E, Holasek A, Kostner G. Studies on the possible function of beta 2-glycoprotein-I: influence in the triglyceride metabolism in the rat. Metabolism 1982; 31:484-6.

[32] Takada D, Ezura Y, Ono S, Iino Y, Katayama Y, Xin Y, Wu LL, Larringa-Shum S, Stephenson SH, Hunt SC, Hopkins PN, Emi M. Apolipoprotein H variant modifies plasma triglyceride phenotype in familial hypercholesterolemia: a molecular study in an eight-generation hyperlipidemic family. J Atheroscler Thromb 2003; 10:79-84.

[33] Eichner JE, Kuller LH, Kamboh MI, Ferrell RE. Phenotypic effects of apolipoprotein structural variation on lipid profiles. I. APO H and quantitative lipid measures in the healthy women study. Genet Epidemiol 1989; 6:311-8.

[34] Kaprio J, Ferrell RE, Kottke BA, Kamboh MI, Sing CF. Effects of polymorphisms in apolipoproteins E, A-IV, and H on quantitative traits related to risk for cardiovascular disease. Arterioscler Thromb 1991; 11:1330-48.

[35] Saha N, Kamboh MI, Ahn YA, Tay JS, Ferrell RE. Apolipoprotein H polymorphism and serum lipoprotein and apolipoprotein levels in two Asian populations. Ethn Dis 1993; 3:250-4.

[36] Kamboh MI, Crawford MH, Aston CE, Leonard WR. Population distributions of APOE, APOH, and APOA4 polymorphisms and their relationships with quantitative plasma lipid levels among the Evenki herders of Siberia. Hum Biol 1996; 68:231-43.

Table 1. Descriptive statistics of the Quebec Family Study subjects by sex and generation groups.

Table 2. PCR primers for genomic amplification of apolipoprotein H promoter and exons.

Table 3. Segregation analysis results for LDL-PPD adjusted for age, body mass index and triglyceride levels.

Table 4. Association of individual apolipoprotein H gene variant with LDL peak particle diameter.

Table 5. Haplotype-specific univariate family-based association test statistics (Z-statistics) for apolipoprotein H gene with LDL peak particle diameter.

Figure 1. Genomic organization of the APOH gene, and location of the genetic variants identified in the Quebec Family Study. The eight exons are shown as vertical bars whose width corresponds to their base-pairs length. The untranslated regions located in exon one and eight are indicated as empty bars.