Identification of Genetic Variants in Endothelial Lipase in Persons With Elevated High-Density Lipoprotein Cholesterol
Background— Elevated high-density lipoprotein cholesterol (HDL-C) is associated with reduced risk of cardiovascular disease, and variation in HDL-C levels has been shown to be ≈50% heritable. Overexpression of endothelial lipase (EL), a member of the lipoprotein lipase gene family, markedly reduces HDL-C levels in mouse models. We hypothesized that genetic variation in EL might be associated with elevated HDL-C.
Methods and Results— All exons and 1.2 kilobase of promoter of the EL gene were sequenced in 20 unrelated human subjects with high HDL-C levels. A total of 17 variants were identified. Six of these were potentially functional and were confirmed by restriction enzyme analysis. Four variants result in amino acid changes (Gly26Ser, Thr111Ile, Thr298Ser, and Asn396Ser,) and 2 variants were in the promoter (−303A/C and −410C/G). The genotype frequencies of each variant were determined in 176 black controls, 165 white controls, and 123 whites with high HDL-C. The Thr111Ile variant was the most common, with an allele frequency of 10.3% in blacks, 31.2% in white controls, and 32.6% in the high HDL-C group. The remaining variants all had allele frequencies <5.0% but differed in frequency among the 3 groups. Interestingly, Gly26Ser, Thr298Ser, and −303A/C were found in the black and high HDL-C white cohorts but were absent in the control white group.
Conclusions— Six new potentially functional variants in EL were discovered through sequencing of the EL gene in subjects with high HDL-C levels. Differences in allele frequencies exist between blacks and whites and between control subjects and those with high HDL-C levels.
Received April 5, 2002; revision received June 14, 2002; accepted June 17, 2002.
The roles of lipoprotein lipase (LPL) and hepatic lipase (HL) in lipoprotein metabolism are well established. Synthesized predominantly in adipose and skeletal muscle, LPL is active on the endothelial surface, where it hydrolyzes lipoprotein triglycerides.1 HL is expressed mainly by hepatocytes and functions on the hepatic endothelial surface, where it catalyzes the hydrolysis of both triglyceride and phospholipid in lipoproteins.2 Endothelial lipase (EL) is the most recently discovered member of the LPL, or triglyceride (TG) lipase, gene family. Cloned independently by 2 groups using distinct methods,3,4⇓ the EL gene spans 10 exons and 9 introns and encodes a polypeptide of 500 amino acids. The EL cDNA is 45% homologous to LPL and 40% to HL. Many of the regulatory domains found in LPL and HL, such as the lipoprotein and heparin-binding domains, are highly conserved in EL.5 Unlike LPL and HL, EL is synthesized by endothelial cells and primarily acts as a phospholipase.6
Whereas LPL activity indirectly influences the metabolism of high-density lipoprotein cholesterol (HDL-C) levels,7 HL directly modulates the metabolism of HDL in vivo by converting larger HDL to smaller HDL.2 Because EL acts predominantly as a phospholipase and the phospholipid content of HDL is a major factor in determining apolipoprotein A-I (apoA-I) conformation8 and HDL function,9–11⇓⇓ we hypothesized that HDL might be a major substrate for EL. We showed that HDL is an excellent and preferred substrate for EL ex vivo.6 Furthermore, we demonstrated that HDL-C and apoA-I levels were markedly reduced in mice injected with an adenoviral vector encoding human EL.3 These results suggested that EL may play a central role in HDL metabolism.
At least 50% of the variation in HDL-C is genetically determined.12,13⇓ Genetic diversity at both the LPL and HL loci has been shown to be a source of variation in HDL-C levels.14–16⇓⇓ Three common missense mutations in LPL (D9N, N291S, and S447X) have been shown to be associated with variation in HDL-C in a variety of populations.17–19⇓⇓ Missense mutations have been identified in patients with HL deficiency, a disorder characterized by elevated HDL-C levels.20,21⇓ A common haplotype in the HL gene comprising 4 variants in the HL promoter was reported to account for some of the variation in HDL-C levels in nuclear families.16 Studies of this haplotype have shown that the less common allele is associated with lower HL activity and higher HDL-C levels in whites22,23⇓ and higher HDL2-C levels in blacks,24 although not all studies have demonstrated an association with HDL-C levels.25 The frequency of this haplotype is more common in blacks, possibly accounting for (at least in part) the higher HDL-C levels seen in this population.26
The phenotype of elevated HDL-C is often dominantly inherited,27 but the molecular pathogenesis in most subjects is unknown. Only homozygous deficiency of HL or of the cholesteryl ester transfer protein (CETP)28 has been shown to cause genetic high HDL-C, and both are recessive conditions. Given that EL has high homology to LPL and HL, has the ability to hydrolyze HDL lipids ex vivo, and results in markedly reduced HDL-C levels in in vivo mouse models, we hypothesized that functional polymorphisms and mutations in EL that result in reduced function might contribute to the phenotype of elevated HDL-C. We therefore sequenced all exons and 1200 bp of the promoter of the EL gene in 20 unrelated individuals with HDL-C levels greater than the 90th percentile for their age, sex, and race in an attempt to identify variation at the EL locus. Here, we report the identification of 17 polymorphic sites in EL. The frequencies of 6 potentially functional variants, 4 of which cause amino acid substitutions and 2 of which occur in the promoter, were then determined in 176 black control subjects, 165 white control subjects, and 123 white subjects with HDL-C levels greater than the 90th percentile.
Ascertainment of Subjects
Individuals with elevated HDL-C levels (defined as HDL-C greater than the 90th percentile for age, sex, and race) were recruited to participate via referral from physicians. After providing informed consent, study subjects were asked to complete a brief questionnaire and have blood drawn. From a larger cohort, a group of 12 whites and 8 blacks were selected for sequencing of the EL gene; preference was given to subjects for whom a family member was also known to have an elevated HDL-C level. Genotyping was performed on 123 unrelated white subjects from the same high HDL-C cohort; the 20 subjects used for sequencing analysis were excluded from this frequency analysis. We also ascertained a cohort of healthy black subjects through a community-based protocol and a cohort of healthy white subjects through a protocol based on a family history of premature coronary disease.29 All protocols were approved by the University of Pennsylvania Institutional Review Board, and each study subject gave written informed consent.
Identification and Confirmation of EL Variants and Determination of the Allele Frequencies
Polymerase chain reaction (PCR) primers for each exon were selected in intronic sequence at least 50 base pairs away from the intron/exon boundaries. The promoter region was amplified using a forward primer located 1200 bps upstream from the transcription start site and a reverse primer located in intron 1 (Table 1). Genomic DNA was isolated from peripheral blood leukocytes using Nucleon extraction and purification protocols (Amersham). PCR reactions containing 200 ng of DNA template, 1 μL of each 10 μmol/L primer solution, 200 μmol/L of each dNTP, 1.5 mmol/L MgCl2, 10 mmol/L Tris-HCl, pH 9.0, 50 mmol/L KCl, and 1.5 U of Taq DNA polymerase (Ready-to-Go PCR Beads, Amersham) were amplified in a final volume of 25 μL while undergoing denaturation at 95° C for 5 minutes, followed by 35 cycles (95° C for 1 minute, 61.5° C for 30 seconds, and 72° C for 1 minute), and an extension at 72° C for 2 minutes. PCR products were run on a 2% agarose gel, and the appropriate bands were excised and then purified using the Qiaex II Gel Extraction Kit (Qiagen). Purified PCR products were submitted with the appropriate sequencing primer (Table 1) to the DNA Sequencing Facility and underwent automated cycle sequencing on an ABI sequencer using Big Dye (Applied Biosystems) terminator chemistry. Sequences were aligned using Mac Vector (Accelrys) software, and chromatograms were viewed using the Edit View program. Allelic variations noted from multiple alignments were verified by inspecting chromatograms.
Of the 17 variants detected, 6 were considered potentially functional and were selected for verification by restriction analysis. Four of the variants resulted in a change in a restriction site, and 4 conventional restriction fragment length polymorphism (RFLP) assays were developed to validate these variants (Table 2). In the other 2 cases, modified PCR-RFLP assays were designed.30 The annealing temperature for each assay was 61.5° C, whereas for the Asn396Ser assay it was 63.5° C. RFLP protocols were carried out using the standard buffers included with the enzymes. All digestions took place at 37° C for 2 hours in a thermal cycler, with the exception of the −303A/C and the Thr111Ile variants, which were digested overnight in a 37° C water bath. Total digestion volumes were 10.0 μL and included 5.0 μL of PCR product. Digested products were resolved by electrophoresis in 2% agarose gels. The 6 RFLP assays developed to confirm the EL variants were used for genotyping analysis of the 3 cohorts to determine the allele frequencies in these cohorts.
Statistical Analysis and Database Searches
The allele frequencies were estimated using a chromosome counting method. Comparisons in frequency between the 2 races and the white control and high HDL-C groups were performed using a χ2 test or a Fisher’s exact test (when appropriate). If a rare allele frequency was found to be greater than 0.3, then lipid results were compared using ANOVA or t test as appropriate.
In silico searches were done at the web sites of The SNP Consortium (TSC)31 and via the National Center for Biotechnology Information’s database SNP (dbSNP)32 to determine if any polymorphic sites identified through our sequencing matched those found in TSC or dbSNP.
After sequencing all 10 EL exons and 1200 bp of the EL promoter in 20 unrelated subjects with high HDL-C, we identified 17 polymorphic sites (Table 3). Four of these sites resulted in amino acid changes and 2 were located in the promoter. Of these 6, the exon 3, Thr111Ile mutation is common and seems to be in linkage disequilibrium with 4 other polymorphic sites. Two silent variants were found in exon 1 (codons 4 and 21). Of the 20 subjects sequenced, the same 3 were heterozygous at both sites, suggesting linkage disequilibrium. Four of the final 5 variants were identified in intronic sequence within 100 to 200 bps of the intron/exon boundaries. The final variant, identified in the 3′ untranslated region, seems to be in linkage disequilibrium with the SNP in intron 8. Only 5 of the 17 variants we identified were seen in silico: Ser4Ser (TSC0205007), Ser21Ser (TSC0205008), the intron 2 site (TSC0921156), the exon 3 Thr111Ile (TSC0921157), and the intron 4 site (TSC0374540).
We selected 6 variants on the basis of their potential to be functional (the 4 variants that result in amino acid changes and the 2 promoter sites) and confirmed them by RFLP analysis. The same RFLP assays were used to determine the frequency of each of the variants in the black, white, and high HDL-C cohorts. The characteristics of these groups are shown in Table 4, with corresponding genotype results in Table 5. All 6 variant alleles were confirmed to be in Hardy-Weinberg equilibrium in each of the 3 cohorts.
The most common variant is Thr111Ile in exon 3. In the original sequencing group, this variant was found in 14 of 40 alleles. Genotyping of the control white cohort revealed a rare allele frequency of 31.2%. In contrast, this variant was much less common in the black cohort, occurring at a frequency of 10.3% (χ2=48.3, P<0.0001). An ANOVA comparing differences in HDL-C levels among the 3 genotypes in the 2 control populations did not demonstrate an association. Our sequencing data indicated that this variant is in linkage disequilibrium with 4 other polymorphic sites, 1 of which is located in the 5′ untranslated region, 2 of which are located in intron 1, and 1 of which is located in intron 4 (Table 3).
In our original high HDL-C sequencing group, 2 different black high HDL-C subjects were found to be heterozygous for variants in exon 1 (Gly26Ser) and exon 6 (Thr298Ser), and another black subject was found to have a promoter variant located at −303 (−303A/C) bp from the EL transcription start site. Our genotyping data revealed that all 3 of these variants occurred in the black control cohort, with allele frequencies of 5.7%, 2.3%, and 1.8%, respectively. These 3 variants were completely absent in the white control population (see Table 5 for χ2 results).
The final 2 variants, an exon 8 (Asn396Ser) missense mutation and a promoter variant located at −410 (−410C/G) bp from the transcription start site, were initially identified in 2 different high HDL-C white subjects. The Asn369Ser variant, though rare in both racial groups, was more common in the white (1.2%) than in the black group (0.6%). The −410C/G promoter variant was found in only 1 subject in the white control population, and was absent in the black control cohort.
We genotyped each of these 6 EL variants in 123 unrelated high HDL-C white subjects and compared the rare allele frequencies to those found in the control white population (Table 5). No significant difference was found in the frequency of the common Thr111Ile SNP. Interestingly, although no subjects in the control white cohort were found to have the Gly26Ser, Thr298Ser, and −303A/C variants, all of these variants were observed in the high HDL-C white cohort, though at allele frequencies less than 1%. The allele frequency of the Asn396Ser variant was higher in the high HDL-C group (2.4%) compared with the white controls (1.2%). The −410C/G promoter variant, although initially identified through sequencing in a white subject with high HDL-C, was not seen in any additional members of the high HDL-C cohort.
On the basis of our data that HDL is a physiological substrate for EL,3,6⇓ we hypothesized that genetic variation in EL may contribute to the phenotype of high HDL-C in humans. We sequenced all 10 EL exons and 1.2 kb of the EL promoter in 20 subjects with high HDL-C and identified 17 variants. This represents the first report of mutations and polymorphisms in the human EL gene. Six of these variants were considered potentially functional. Four of them result in amino acid substitutions and 2 are located in the promoter. One variant, Thr111Ile, is common in whites (31.2%) but much less common in blacks (10.3%). We found no significant differences in HDL-C levels in carriers of this variant and no significant difference in the allele frequency between the white high HDL-C cohort and the white control cohort, but larger association studies will be required. Our sequencing results demonstrate that this SNP is in linkage disequilibrium with 4 other SNPs that we found.
Three variants (Gly26Ser, Thr298Ser, and −303A/C) found on sequencing high HDL-C subjects (each once) were also found in the unrelated cohort of 123 whites ascertained on the basis of high HDL-C levels, but none of these variants were found in a control cohort of 165 whites. These findings suggest that these 3 variants could contribute to the phenotype of high HDL-C. These 3 variants occurred with frequencies of 1% to 5% in blacks, suggesting that these variants could conceivably contribute to the higher HDL-C levels in blacks compared with whites. The Asn396Ser variant was also more common in the white high HDL-C group (2.4%) compared with the white controls (1.2%). The −410C/G variant was the rarest of the 6. It was identified in 1 white subject with high HDL-C that was sequenced and in only 1 other subject from the control white population.
The 4 variants that result in amino acid substitutions are potentially functional. Whereas Thr298Ser is a relatively conservative amino acid change, both Gly26Ser and Thr111Ile are less conservative, involving the addition and subtraction of an −OH group, respectively. Finally, Asn396Ser is a nonconservative substitution and occurs in an area of sequence homology to LPL and HL where functional mutations have been reported.20,33⇓ Further studies will be required to determine whether these amino acid substitutions influence enzymatic activity. The 2 promoter variants are also potentially functional. A common HL SNP that is associated with HDL-C levels, the −514 C/T variant, is located in the HL promoter. The rare “T” allele is associated with a 30% decrease in HL TG lipase activity in comparison to the common “C” allele.34 The remaining polymorphic sites, the 2 silent mutations, the 4 common intronic SNPs, and 3′UTR variant may be of use in future studies as markers for potential functional variants.
The physiological importance of HDL as a mediator of cholesterol homeostasis and reverse cholesterol transport is well established. Therefore, identification of genes that modulate HDL metabolism and of sequence variation in these genes is of substantial interest and importance. Genetic variation in both LPL35,36⇓ and HL22 has been extensively studied with regard to their associations with lipoprotein metabolism and atherosclerosis. Although the role of EL in HDL metabolism is not yet fully understood, HDL seems to be a physiological substrate for EL.3,6⇓ The genetic variants in EL reported here suggest that genetic variation in EL may be related to the phenotype of high HDL-C. They will be valuable for association studies focused on addressing whether genetic variation in EL is associated with variation in plasma lipid levels and atherosclerosis in larger cohorts.
This work was supported by National Institute of Health grant HL55323 from the National Heart, Lung, and Blood Institute, an Established Investigator Award from the American Heart Association, and a Burroughs Wellcome Foundation Clinical Scientist Award in Translational Research. The study was also supported in part by grant M01-RR00040 from the National Center for Research Resource supporting the University of Pennsylvania General Clinical Research Center. We thank Anna Lillethun, Linda Morrell, Manjula Kari, Joya Sahu, Sara Vartanian, Sarah Jahn, and Dawn Marchadier for technical assistance and Dr Jane M. Glick for helpful discussions.
- ↵Goldberg IJ. Lipoprotein lipase and lipolysis: central roles in lipoprotein metabolism and atherogenesis. J Lipid Res. 1996; 37: 693–707.
- ↵Hirata K, Diechek HL, Cioffi JA, et al. Cloning of a unique lipase from endothelial cells extends the lipase gene family. J Biol Chem. 1999; 274: 14170–14175.
- ↵McCoy MG, Sun GS, Marchadier D, et al. Characterization of the lipolytic activity of endothelial lipase. J Lipid Res. 2002; 43: 921–929.
- ↵Blades B, Vega GL, Grundy SM. Activities of lipoprotein lipase and hepatic triglyceride lipase in postheparin plasma of patients with low concentrations of HDL cholesterol. Arterioscler Thromb Vasc Biol. 1993; 13: 1227–1235.
- ↵Sparks D, Phillips M, Lund-Katz S. The conformation of apolipoprotein A-I in discoidal and spherical recombinant high density lipoprotein particles: 13C NMR studies of lysine ionization behavior. J Biol Chem. 1992; 267: 25830–25838.
- ↵Bamberger M, Glick J, Rothblat G. Hepatic lipase stimulates the uptake of high density lipoprotein cholesterol by hepatoma cells. J Lipid Res. 1983; 24: 869–876.
- ↵Johnson WJ, Bamberger MJ, Latta RA, et al. The bidirectional flux of cholesterol between cells and lipoproteins. J Biol Chem. 1986; 261: 5766–5776.
- ↵Breslow JL. Familial disorders of high density lipoprotein metabolism.In: Scriver CR, Beaudet AL, Sly WS,et al, eds. The Metabolic Basis of Inherited Disease. New York, NY: McGraw-Hill, Inc; 1995: 2031–2052.
- ↵Guerra R, Wang J, Grundy MS, et al. A hepatic lipase (LIPC) allele associated with high plasma concentrations of high density lipoprotein cholesterol. Proc Natl Acad Sci U S A. 1997; 94: 4532–4537.
- ↵Hegele R, Little JA, Vezina C, et al. Hepatic lipase deficiency: clinical, biochemical, and molecular genetic characteristics. Arterioscler Thromb. 1993; 13: 720–728.
- ↵Murtomaki S, Tahvanainen E, Antikainen M, et al. Hepatic lipase gene polymorphisms influence plasma HDL levels: results from Finnish EARS participants. European Atherosclerosis Research Study. Arterioscler Thromb Vasc Biol. 1997; 17: 1879–1884.
- ↵Juo SH, Han Z, Smith JD, et al. Promoter polymorphisms of hepatic lipase gene influence HDL(2) but not HDL(3) in black men: CARDIA study. J Lipid Res. 2001; 42: 258–264.
- ↵Vega GL, Clark LT, Tang A, et al. Hepatic lipase activity is lower in black men than in white American men: effects of 5′ flanking polymorphism in the hepatic lipase gene (LIPC). J Lipid Res. 1998; 39: 228–232.
- ↵The SNP Consortium. Single Nucleotide Polymorphisms for Biomedical Research. Available at: http://snp.cshl.org. Accessed June 27, 2002.
- ↵National Center for Biotechnology Information. Single Nucleotide Polymorphism. Available at: http://www.ncbi.nlm.nih.gov/SNP. Accessed June 27, 2002.
- ↵Deeb SS, Peng R. The C-514T polymorphism in the human hepatic lipase gene promoter diminishes its activity. J Lipid Res. 2000; 41: 155–158.