SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES

The prevalence of morbid obesity and its associated metabolic complications has risen rapidly in the past decade. Recently, we have established the transcriptome of the visceral adipose tissue of nondiabetic severely obese men with and without metabolic syndrome (MetS) that provided new candidate genes for cardiovascular disease (CVD) risk factors. The oxysterol-binding protein–like protein 11 (OSBPL11) that belongs to the OSBP family of intracellular receptors was one of the genes found to be significantly overexpressed in the MetS group. To determine whether OSBPL11 gene polymorphisms are associated with CVD risk factors and diabetes, OSBPL11 gene promoter and coding regions were sequenced in 25 individuals and six tagging single-nucleotide polymorphisms (SNPs) capturing 85% of gene sequence–derived common genetic variability (minor allele frequency (MAF) > 5%) were genotyped in two samples for a total of 962 obese individuals. Using a multistage experimental design, χ2-tests and logistic regressions were applied to compare genotype frequencies and to compute odds ratios (ORs) for low and high CVD risk groups. Significant associations between rs1055419 and diastolic blood pressure (OR = 0.53; P = 0.01) were found whereas IVS12+95 T>C, a newly discovered SNP, was associated with low-density lipoprotein–cholesterol levels (OR = 1.63; P < 0.001), hyperglycemia/diabetes (OR = 1.48; P < 0.004) as well as with MetS per se (OR = 1.56; P < 0.01). These results suggest that the OSBPL11 gene is involved in cholesterol and glucose metabolism in obese individuals.

The prevalence of severe or morbid obesity (BMI >40 kg/m2) has increased faster worldwide than that of obesity per se (1,2). Indeed, 6.3% of the American women are severely obese and this prevalence rises to 15% in black women (1). Obesity and even more morbid obesity is threatening for health, as this condition is associated with a number of metabolic abnormalities including decreased plasma high-density lipoprotein–cholesterol levels, increased plasma triglyceride and glucose concentrations, hypertension, and type 2 diabetes, which define the metabolic syndrome (MetS). Each of these contributes to increase cardiovascular disease (CVD) risk (3). Although most obese individuals' CVD risk profile varies from mild to severe, a moderate number of them remain metabolically healthy (4). Because CVD risk factors have significant impact on individual's health and that their hereditary component is well recognized (5), there is a need to better understand their genetic etiology. Recently, we have established the transcriptome of the omental adipose tissue of nondiabetic obese men with and without the MetS with the expectation to better understand the role of genes in the risk of obese persons to develop or not obesity-related metabolic complications (6). Of the genes differentially expressed, the oxysterol-binding protein–like protein 11 (OSBPL11) was found to be overexpressed by 34% (1.34-fold; P < 0.05) in the group of obese with a high CVD risk profile (6).

OSBPL11 belongs to a family of proteins sharing sequence and structural homologies (7). OSBP is the most studied member of this family and was discovered while searching for proteins that bind oxysterols, which are potent inhibitors of cholesterol synthesis (8). Oxysterols are produced by enzymatic and nonenzymatic oxidation of cholesterol; they are present at very low concentrations in the cell cytoplasm and appear to signal an excess of cholesterol via binding to the liver X receptors (LXRα and LXRβ) (7).

Initially, the LXR nuclear receptors were found to bind specific response elements within the promoter of a number of genes involved in reverse cholesterol transport (7,9). More recently, LXRs also have been shown to be involved in carbohydrate metabolism and lipogenesis (10). Accordingly, by sequestrating oxysterols in the cell cytoplasm, OSPBs could inhibit their beneficial impact on lipid and carbohydrate metabolism through LXRs' binding.

We have therefore postulated that the OSBPL11 gene may be involved in lipid and carbohydrate metabolism and have undertaken a comprehensive molecular analysis of the human OSBPL11 gene. The objective of the study was to verify whether OSBPL11 gene polymorphisms contribute to explain the interindividual variability in CVD risk profile observed among obese individuals.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES

Two samples from the same population of severely obese patients undergoing an antiobesity surgery and consecutively recruited were used to verify whether OSBPL11 gene polymorphisms were associated with CVD risk factors using a multistage experimental design (11,12,13). Because no difference between these two samples was found according to sex distribution, smoking habits, diabetes, any of the physical and metabolic characteristics presented in Table 1, and genotypic frequency, Table 1 presents the subjects' characteristics of both samples combined. At the time of the surgery, these middle-aged individuals were severely obese (BMI = 52.3 ± 9.6 kg/m2) and showed a deteriorated plasma glucose and lipid profile (Table 1). Among these subjects, 23.4% were not having the MetS as defined by the National Cholesterol Education Program Adult Treatment Panel III definition (NCEP-ATPIII) (14) whereas 32.7% were diabetic according to the World Health Organization criteria (15) or were taking medications for the treatment of type 2 diabetes.

Table 1.  Characteristics of the study subjects
inline image

A total of 24 polymorphisms were identified by direct sequencing (Table 2). One synonymous variation was observed in exon 10 (c.1704 C>T; p.P168P). Two variants were located in the 5′-untranslated region (rs7625936 and rs1055419) and one was found in the 3′-untranslated region (c.2263 C>T). Thirteen intronic polymorphisms were also identified. Seven polymorphisms were revealed in the promoter region including two insertions/deletions. Of these 22 single-nucleotide polymorphisms (SNPs), nine were listed in HapMap (rs7625936, rs1055419, rs2979382, rs3829007, rs16836866, rs12496976, rs2294053, and rs13066562) and 11 (including four listed in HapMap) were relatively common with a minor allele frequency (MAF) >10% based on sequencing-derived allele frequency. After genotyping both samples, it appeared that the MAF of rs1055419 was <10% (Table 2). Based on linkage disequilibrium (r2 > 0.75) of sequencing-derived SNPs (Figure 1), six polymorphism blocks were built and one SNP per block was selected for genotyping. This strategy allowed covering 85% of the sequence-derived genetic variability of the common polymorphisms (MAF >5%) at OSBPL11 gene locus. Four (rs7625936, rs1055419, rs2979382, and rs12496976) out of the six genotyped SNPs were indexed in HapMap. Their genotyping allowed capturing 76.2% of all common OSBPL11 variants (MAF >5%) indexed in HapMap (54 SNPs). Furthermore, the coverage of genetic variability is likely to be better than 76.2% because two more SNPs (rs12487030 and IVS12+95C>T), neither indexed in HapMap nor in linkage disequilibrium with the others, were also genotyped.

Table 2.  Characteristics of OSBPL11 polymorphisms revealed by OSBPL11 promoter and coding regions sequencing
inline image
image

Figure 1. Linkage disequilibrium map of the single-nucleotide polymorphisms (SNPs) revealed by OSBPL11 promoter and coding regions sequencing in 25 obese individuals. SNPs with minor allelic frequencies >5% are presented and r2 values (%) are shown in the figure lozenges. The genotyped polymorphisms are marked with an asterisk (*). SNPs with MAF <5% were excluded from the map.

Download figure to PowerPoint

The analyses were performed in three consecutive steps. First, the selected polymorphisms were genotyped in sample 1 and tested for association with all phenotypes. Second, the most promising associations (P < 0.1) found in the first sample were reanalyzed in sample 2. Finally, odds ratios (ORs) were computed in the combined samples only for SNP and phenotype combinations for which trend and/or significant associations were found in both samples. This strategy provides acceptable power with lower number of samples as compared to common one-stage studies, accounts for multiple testing issues by significantly decreasing the number of statistical tests performed in sample 2, and provides a replication sample as the association is tested in two independent cohorts (11,12,13). Accordingly, four out of six independent polymorphisms were associated with one or more CVD risk factors in sample 1 and were then reanalyzed in the second sample providing four associations (two polymorphisms) showing consistent results in both samples (Table 3). First, the T allele of one polymorphism located within the OSBPL11 promoter region, rs1055419, was associated with a lower risk to have high diastolic blood pressure (Table 4). In addition, carriers of the C allele of a newly discovered SNP located in OSBPL11 intron 12, IVS12+95T>C, were at ∼1.5-fold higher risk to have high plasma low-density lipoprotein–cholesterol levels, hyperglycemia and/or diabetes, and being characterized by the MetS (Table 4). We also verified whether the frequency of both rs1055419 and IVS12+95T>C SNPs was comparable in severely obese and nonobese individuals. To do so, unrelated nonobese (BMI <30 kg/m2) individuals from the Quebec Family Study (16) (n = 231) were genotyped. None of these polymorphisms showed significant genotypic frequency differences (data not shown).

Table 3.  Comparison of OSBPL11 polymorphisms genotypic frequencies between subjects with and without cardiovascular risk factors and diabetes
inline image
Table 4.  Odds ratio for significant χ2-tests between OSBPL11 polymorphism and CVD risk factors
inline image

The impact of OSBPL11 gene polymorphisms on its expression in omental adipose tissue of 99 nondiabetic obese women was also tested. Their physical characteristics are presented in Table 1. Only rs7625936 (C/C = 119.6 ± 6.3 vs. C/T = 125.0 ± 6.8 vs. T/T = 77.2 ± 19.5; P = 0.01) was associated with OSBPL11 mRNA levels in omental adipose tissue. The level of OSBPL11 mRNA of the T homozygotes was decreased by ∼37% as compared to the other genotypes (C/T; P = 0.003 and C/C; P = 0.008). However, there were only six subjects in that genotype group.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES

The overall goal of this research project was to investigate the role of genes in obesity-related metabolic complications. This study was undertaken to verify whether gene polymorphisms support the role of OSBPL11, one of the previously reported differentially expressed genes in omental adipose tissue of nondiabetic obese men with and without the MetS (6). Of the genotyped tagging polymorphisms capturing most of the genetic variability at OSBPL11 gene locus, rs105419 was associated with diastolic blood pressure while IVS12+95C>T variant was associated with a deteriorated plasma lipid profile hyperglycemia and/or diabetes as well as with MetS per se.

The biological function of OSBPL11 has not yet been directly assessed but as it shares structural and sequence homologies with OSBP (7), it is likely that they share biological activities. Based on the expression profile suggested by analysis of expressed sequence tag counts (Unigene, National Center for Biotechnology Information; http:www.ncbi.nlm.nih.govUniGeneESTProfileViewer.cgiuglistHs.477440), OSBPL11 seems to be expressed in key tissues, such as the heart, muscle, liver, pancreas, kidney, and adipose tissue amongst others, where they may partake in the control of cholesterol, lipid, and carbohydrate homeostasis metabolism as well as blood pressure. Recently, we have also shown that OSBPL11 mRNA levels in abdominal subcutaneous adipose tissue of overweight women well responding to caloric restriction in terms of fat mass loss were decreased as compared to poor responder group after the weight loss program (L. Bouchard, R. Rabasa-Lhoret, M. Faraj et al., unpublished data). Although adipose tissue plays a central role in energy balance and the development of obesity-related metabolic complications, it is also likely that the associations with obesity-related complications and diabetes reflect the effects of polymorphisms on OSBPL11 gene expression and activity in other key tissues such as liver, muscle, and kidney.

The exact contribution of OSBPL11 in the development of obesity-related metabolic complications and possibly response to caloric restriction has still to be determined. However, these results suggest that common OSBPL11 gene polymorphisms explain some of the interindividual variability in metabolic complications observed among obese individuals suggesting that OSBPL11, such as other member of the OSBP family, is involved in glucose and lipid metabolism. These associations warrant additional genetic and functional studies.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES

Patient selection

From June 2000 to December 2006, 962 severely obese white men and women undergoing an antiobesity surgery at Laval Hospital (Québec, Canada) have been consecutively recruited. The surgical protocol is presented elsewhere (17). A total of 197 patients received hypolipidemic drugs (statine n = 171 and fibrate n = 27) for the treatment of dyslipidemia. All individuals provided written informed consent before their inclusion in the study.

Body weight, height, waist girth, resting blood pressure (systolic and diastolic) as well as blood lipid and glucose concentrations were measured using standardized procedures (18).

Sequencing and genotyping

Genomic DNA was extracted using GenElute Blood Genomic DNA kit (Sigma, St. Louis, MO). The promoter (≅2,000 bases) and the coding sequence including intron–exon junction boundaries of the OSBPL11 gene were sequenced in a total of 25 (50 chromosomes) severely obese men, selected from the larger cohort, with (n = 12) and without (n = 13) CVD risk factors as defined by the NCEP-ATPIII criteria (14). Fourteen of them were the same used in the recently conducted omental adipose tissue gene expression profiling (6). Primers were designed using human OSBPL11 public sequence (accession number: NC_000003.10). PCR forward and/or reverse primers were used to perform sequencing reactions with the BigDyeTH Terminator v3.1 kit and the samples were run on ABI Prism 3730/XL DNA Analyzer automated sequencers (Applied Biosystems, Foster City, CA).

Six tagging SNPs (not in linkage disequilibrium (r2 < 0.75) with each other; Figure 1) with MAF >10% were genotyped by direct sequencing (rs12496976 and rs12487030), as presented above. The other four OSBPL11 polymorphisms were genotyped using validated (rs7625935, rs10554419, rs2979382) or custom (IVS12+95T>C) primers and Taqman probes (Applied Biosystems), as presented elsewhere (6). Genotypes were determined using a 7500 Real Time PCR System and analyzed using ABI Prism SDS version 1.2.3 software (Applied Biosystems). PCR conditions and primer sequences for OSBPL11 sequencing and genotyping may be obtained upon request.

Quantitative real-time RT-PCR

To verify whether OSBPL11 genotypes influence omental OSBPL11 mRNA abundance, a maximum number of individuals were selected from the larger cohort. In order to control for obvious possible confounding factors, biopsies from nondiabetic individuals not taking medications to treat any of the CVD risk factors were prioritized. At the time of the selection, 99 premenopausal women but a much lower number of men qualified to these criteria. In order to keep the number of samples as high as possible and to avoid any possible gender-related confounding effects (not established yet), the premenopausal women group was selected. With a sample of 99 individuals, at α = 0.05, the power to detect association using an additive genetic model is 0.58 when a medium effect size, as defined by Cohen (19), is expected. Although this analysis is slightly underpowered, one has to remember that 99 omental adipose tissue biopsies from nondiabetic obese individuals not under medication for any CVD risk factors is a very large number of samples considering that such biopsies need invasive surgical intervention and that severely obese persons are commonly under drug treatment that can alter gene expression. Total RNA was extracted as previously described (6). Complementary DNA was generated from 40 ng of total RNA using a random primer hexamer following the protocol for Superscript II (Invitrogen, Carlsbad, CA). Equal amounts of complementary DNA were run in triplicate and amplified in a 15 µl reaction containing 7.5 µl of 2× Universal PCR Master Mix (Applied Biosystems), 10 nmol/l of Z-tailed forward primer, 100 nmol/l of reverse primer, 100 nmol/l of Amplifluor Uniprimer probe (Chemicon, Temecula, CA), and 2 µl of DNA target. The mixture was incubated at 50 °C for 2 min, at 95 °C for 4 min, and then cycled at 95 °C for 15 s and at 55 °C for 40 s for 55 times using the Applied Biosystems Prism 7900 Sequence Detector. Amplification efficiencies were validated and normalized to ribosomal 18S and quantities of target gene were calculated according to a standard curve. Primers were designed using Primer Express 2.0 (Applied Biosystems) and their sequences are the following: OSBPL11: 5-Ztail-GCAAACTGCATCGGGTTACA-3, 5-CAGT GTTGGTGATGTTGTGCTTT-3 and m18S: 5-Ztail-CGGTACAG TGAAACTGCGAATG-3, 5-CCAAAGGAACCATAACTGATTTAA TGA-3. Amplicons were detected using the Amplifuor UniPrimer system where forward primers used contained the following 5′ Z sequence: ACTGAACCTGACCGTACA.

Statistical analysis

Mean phenotypic differences between men and women were tested using an unpaired t-test. Haploview software (20) was used to evaluate whether genotype and allele frequencies were in Hardy–Weinberg equilibrium, to compute linkage disequilibrium (r2) value between OSBPL11 polymorphisms, and to select tag SNPs using Tagger (21). HapMap, built 35, was used when needed.

With a multistage experimental design, χ2-tests were used to compare genotype frequencies between groups of individuals with low and high risk to develop CVD according to the NCEP-ATPIII criteria (14), and between groups with and without diabetes. Briefly, the patients were sequentially assigned to sample 1 (stage 1) and sample 2 (stage 2). The presence of associations was first tested in sample 1 and only the most promising associations (P ≤ 0.1) were reanalyzed in the second sample, in agreement with Sobell et al. (13). Finally, logistic regressions were used to compute the ORs and Wald's confidence intervals in the combined samples when promising association was found in both samples. This strategy offers increased statistical power as compared to one-stage studies while accounting for multiple testing issues because of lower number of statistical tests, and provide independent samples for results validation (11,12,13). We have presented this strategy in more details elsewhere (22). The homozygotes for the major genotype were the reference group. Homozygotes for the minor allele of the SNPs with genotypic frequencies <5% were merged to heterozygotes (rs1055419, rs12496976, rs12487030, and IVS12+95T>C). All individuals met the criteria for abdominal obesity using the NCEP-ATPIII (14) definition as they were obese. The NCEP-ATPIII cutoff values used were: plasma high-density lipoprotein–cholesterol levels <1.03 mmol/l and <1.29 mmol/l for men and women, respectively, triglycerides ≥1.69 mmol/l, glycemia ≥5.5 mmol/l, and blood pressure ≥130/85 mm Hg. Plasma low-density lipoprotein- and total-cholesterol cutoff values were ≥4.1 mmol/l and ≥6.2 mmol/l, which indicate a high CVD risk (ATPIII guidelines 2001). Because only four individuals were normoglycemic (glycemia >5.5 mmol/l) but diabetics, these two clinical endpoints were combined within the hyperglycemia/diabetes phenotype. To be considered as affected by the MetS, a patient had to fulfill at least three of the risk factors described in the NCEP-ATPIII (14). Individuals taking medications to treat one of the CVD risk factors were considered to meet the criteria for this factor. Power computation and statistical analyses were performed using GPower (version 3.0.8) (23) and SAS (version 9.1.3) software (SAS Institute, Cary, NC), respectively.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES

This study was supported by a grant from the Canadian Institutes of Health Research (CIHR) and the Institute of Nutrition, Metabolism and Diabetes (INMD) under its strategic initiative “Excellence, Innovation and Advancement in the Study of Obesity and Healthy Body Weight.” The severely obese cohort was supported, over the years, by the Laval University Merck-Frosst/CIHR Research Chair in Obesity. We express our gratitude to Stéfane Lebel, Frédéric-Simon Hould and Picard Marceau of the Laval Hospital biliopancreatic diversion team, who have also sampled adipose tissues for this project. Many thanks are also expressed to Vicky Drapeau, Fanny Therrien and Alain Houde for their help in adipose tissue banking management. We acknowledge the contribution of the Gene Quantification core laboratory of the Centre de Génomique de Québec. Luigi Bouchard is the recipient of a fellowship award from the Heart and Stroke Foundation of Canada (HSFC)/Sanofi-Aventis. André Tchernof is a research scholar from CIHR.

REFERENCES

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Methods and Procedures
  6. Acknowledgments
  7. Disclosure
  8. REFERENCES