Publications

To localize quantitative trait loci for insulin metabolism and obesity, genome scans/linkage analyses were performed on >900 members of 32 extended families participating in phase 3 of the Strong Heart Study, an investigation of the genetic and environmental determinants of cardiovascular disease in American-Indian populations from Arizona, Oklahoma, and North and South Dakota. Linkage analyses of fasting insulin and two obesity-related phenotypes, BMI and percent fat mass, were performed independently in each of the three populations. For log fasting insulin, we found a genome-wide maximum, robust logarithm of odds (LOD) score of 3.42 at 51 cM on chromosome 2p in the Dakotas. Bivariate linkage analyses of log fasting insulin with both BMI and fat mass indicate a situation of incomplete pleiotropy, as well as several significant bivariate LOD scores in the Dakotas.

When activated, thrombin activatable fibrinolysis inhibitor (TAFI) inhibits fibrinolysis by modifying fibrin, depressing its plasminogen binding potential. Polymorphisms in the TAFI structural gene (CPB2) have been associated with variation in TAFI levels, but the potential occurrence of influential quantitative trait loci (QTLs) located elsewhere in the genome has been explored only in families ascertained in part through probands affected by thrombosis. We report the results of the first genome-wide linkage screen for QTLs that influence TAFI phenotypes. Data are from 635 subjects from 21 randomly ascertained Mexican American families participating in the San Antonio Family Heart Study. Potential QTLs were localized through a genome-wide multipoint linkage scan using 417 highly informative autosomal short tandem repeat markers spaced at approximately 10-cM intervals. We observed a maximum multipoint LOD score of 3.09 on chromosome 13q, the region of the TAFI structural gene. A suggestive linkage signal (LOD = 2.04) also was observed in this region, but may be an artifact. In addition, weak evidence for linkage occurred on chromosomes 17p and 9q. Our results suggest that polymorphisms in the TAFI structural gene or its nearby regulatory elements may contribute strongly to TAFI level variation in the general population, although several genes in other regions of the genome may also influence variation in this phenotype. Our findings support those of the Genetic Analysis of Idiopathic Thrombophilia (GAIT) project, which identified a potential TAFI QTL on chromosome 13q in a genome-wide linkage scan in Spanish thrombophilia families.

Paraoxonase 1 (PON1), a high-density-lipoprotein-associated enzyme known to protect against cellular damage from toxic agents, may also have antioxidant properties. Although the importance of the influence of the PON1 structural locus on chromosome 7q21-22 for variation in the concentration and activity of the enzyme is well-documented, the contribution of other loci is poorly understood. Based on the recent observations of at least one additional quantitative trait locus (QTL) for PON1 activity in pedigreed baboons, we conducted a whole-genome linkage screen for QTLs other than the PON1 structural locus that may influence PON1 activity in humans. We measured PON1 activity in frozen serum for 1,406 individuals in more than 40 extended pedigrees from the San Antonio Family Heart Study (SAFHS). We used a maximum-likelihood-based variance decomposition approach implemented in SOLAR to test for QTLs that may influence PON1 activity. In addition to a QTL for which we detected the strongest, significant evidence (LOD = 31.41) at or near the PON1 structural locus on chromosome 7q21-22, we also localized at least one additional significant QTL on chromosome 12 (LOD = 3.56). Furthermore, we detected suggestive evidence for two more PON-related QTLs on chromosomes 17 and 19. We have provided evidence that other genes, in addition to the well-known ones on chromosome 7, play a role in influencing normal variation in PON1 activity.

Increasing incidence of cardiovascular disease in traditionally low-risk Alaskan Eskimos is a cause for concern. The purpose of this study was to examine the genetic and environmental correlations of low-density lipoprotein (LDL) subfractions with obesity-related factors in Alaskan Eskimos, using data from the first 954 participants of the Genetics of Coronary Artery Disease in Alaska Natives Study. Estimates of genetic and environmental influence were calculated using a maximum likelihood variance component method implemented in SOLAR. Mean values of weight, body mass index (BMI), and waist were 73.4 +/- 0.5 kg, 27.6 +/- 0.2 kg/m2, and 88.0 +/- 0.4 cm, respectively. LDL, and its small (LDL1), medium (LDL2), and large (LDL3) subfractions, had mean values of 115.8 +/- 1.2 mg/dl, 8.3 +/- 0.4 mg/dl, 19.6 +/- 0.8 mg/dl, and 71.5 +/- 1.5 mg/dl, respectively. Bivariate analysis displayed significant genetic correlations between LDL subfractions and obesity-related factors: LDL1 with BMI (rhoG = 0.67, P < 0.05), waist (rhoG = 0.80, P < 0.001), and subscapular and tricep skinfolds (rhoG = 0.93, P < 0.005, and rhoG = 0.78, P < 0.05, respectively); LDL2 with BMI (rhoG = 0.52, P < 0.05), waist (rhoG = 0.46, P < 0.05), and tricep skinfold (rhoG = 0.60, P < 0.05); and mean LDL size with BMI (rhoG = -0.36), waist (rhoG = -0.42,), and subscapular and tricep skinfolds (rhoG = -0.44 and -0.43, respectively) (P < 0.005). These results show that a common set of genes is influencing LDL size and obesity-related factors in Alaskan Eskimos.

Age-adjusted systolic blood pressure is higher in males than females. Genetic factors may account for this sex-specific variation. To localize sex-specific quantitative trait loci (QTL) influencing blood pressure, we conducted a genome scan of systolic blood pressure, in males and females, separately and combined, and tested for aggregate and QTL-specific genotype-by-sex interaction in American Indian participants of the Strong Heart Family Study. Blood pressure was measured 3 times and the average of the last 2 measures was used for analyses. Systolic blood pressure was adjusted for age and antihypertensive treatment within study center. We performed variance component linkage analysis in the full sample and stratified by sex among 1168 females and 726 males. Marker allele frequencies were derived using maximum likelihood estimates based on all individuals, and multipoint identity-by-descent sharing was estimated using Loki. We detected suggestive evidence of a QTL influencing systolic blood pressure on chromosome 17 at 129 cM between markers D17S784 and D17S928 (logarithm of odds [LOD] = 2.4). This signal substantially improved when accounting for QTL-specific genotype-by-sex interaction (P = 0.04), because we observed an LOD score of 3.3 for systolic blood pressure in women on chromosome 17 at 136 cM. The magnitude of the linkage signal on chromosome 17q25.3 was slightly attenuated when participants taking antihypertensive medications were excluded, although suggestive evidence for linkage was still identified (LOD = 2.8 in women). Accounting for interaction with sex improved our ability to detect QTLs and demonstrated the importance of considering genotype-by-sex interaction in our search for blood pressure genes.

BACKGROUND: Genetic and environmental contributions to childhood obesity are poorly delineated.

OBJECTIVE: The Viva la Familia Study was designed to genetically map childhood obesity and its comorbidities in the Hispanic population. The objectives of this report were to describe the study design and to summarize genetic and environmental contributions to the phenotypic variation in obesity and risk factors for metabolic diseases in Hispanic children.

DESIGN: The Viva la Familia cohort consisted of 1030 children from 319 families selected based on an overweight proband between the ages of 4 and 19 y. In-depth phenotyping to characterize the overweight children and their siblings included anthropometric and body-composition traits by dual-energy X-ray absorptiometry and assessments of diet by 24-h recalls, physical activity by accelerometry, and risk factors for metabolic diseases by standard biochemical methods. Univariate quantitative genetic analysis was used to partition phenotypic variance into additive genetic and environmental components by using the computer program SOLAR.

RESULTS: Sex, age, and environmental covariates explained 1-91% of the phenotypic variance. Heritabilities of anthropometric indexes ranged from 0.24 to 0.75. Heritability coefficients for the body-composition traits ranged from 0.18 to 0.35. Diet and physical activity presented heritabilities of 0.32 to 0.69. Risk factors for metabolic diseases were heritable with coefficients ranging from 0.25 to 0.73. Significant genetic correlations between obesity traits and risk factors for metabolic diseases substantiated pleiotropy between traits.

CONCLUSION: The Viva la Familia Study provides evidence of a strong genetic contribution to the high prevalence of obesity and its comorbidities in Hispanic children.

OBJECTIVE: Genetic components of energy homeostasis contributing to childhood obesity are poorly understood. Genome scans were performed to identify chromosomal regions contributing to physical activity and dietary intake traits in Hispanic children participating in the VIVA LA FAMILIA Study.

RESEARCH METHODS AND PROCEDURES: We report linkage findings on chromosome 18 for physical activity and dietary intake in 1030 siblings from 319 Hispanic families. Measurements entailed physical activity by accelerometry, dietary intake by two 24-hour recalls, and genetic linkage analyses using SOLAR software.

RESULTS: Significant heritabilities were seen for physical activity and dietary intake, ranging from 0.46 to 0.69, except for vigorous activity (h2 = 0.18). Percentage time in sedentary activity mapped to markers D18S1102-D18S64 on chromosome 18 [logarithm of the odds (LOD) score = 4.07], where melanocortin 4 receptor gene (MC4R) resides. Quantitative trait loci (QTLs) for total activity counts, percentage time in light or in moderate activity, and carbohydrate intake and percentage of energy intake from carbohydrates were detected in the same region (LOD = 2.28, 2.79, 2.2, 1.84, and 1.51, respectively). A novel loss of function mutation in MC4R (G55V) was detected in six obese relatives, but not in the rest of the cohort. Removal of these MC4R-deficient subjects from the analysis reduced the LOD score for sedentary activity to 3.94.

DISCUSSION: Given its role in the regulation of food intake and energy expenditure, MC4R is a strong positional candidate gene for the QTL on chromosome 18 detected for physical activity and dietary intake in Hispanic children.

The hepatic lipase gene (LIPC) has been implicated as a potential regulator of HDL-cholesterol concentration and HDL and LDL particle size. Studies have centered on a C to T transition in the promoter region of LIPC, 514 base pairs upstream from the transcription initiation site. We performed a genome-wide linkage screen for several lipoprotein size phenotypes and tested for association of these traits with LIPC -514C-T in 798 individuals from the San Antonio Family Heart Study. Median diameters were measured for HDL particles stained for apoA1 (A1), apoA2 (A2), unesterified cholesterol (UC), and esterified cholesterol (EC) and for LDLs stained for EC. The median diameter of all phenotypes exhibited evidence of linkage to the LIPC region of chromosome 15 (LODs of 1.78 to 3.79). Linkage was also observed for HDL-EC size on chromosome 5p (LOD = 3.50). Association with the LIPC -514C-T polymorphism was detected for HDL-A1, HDL-A2, HDL-UC, and HDL-EC median diameters (p < 0.001) but not for LDL-EC size. Linkage analyses of HDL sizes conditional on the -514C-T polymorphism reduced the LOD scores in the LIPC region only slightly, suggesting that this polymorphism does not explain the observed linkage of lipoprotein sizes to chromosome 15. These results indicate the presence of a lipoprotein size locus in the LIPC region but suggest that -514C-T is not the primary functional variant in this region, implying that additional functional mutations influencing HDL and potentially LDL size variation occur in or near LIPC.

INTRODUCTION: The hormone resistin was recently discovered in adipose tissue of mice. Functional tests suggest a role for resistin in the regulation of insulin sensitivity. However, human studies have reported controversial results on the metabolic function of this hormone.

METHODS: A 1 g omental adipose tissue biopsy was obtained from 404 adult baboons. Resistin mRNA expression was assayed by real-time, quantitative RT-PCR, and univariate and bivariate quantitative genetic analyses were performed, via the variance decomposition approach. A genome scan analysis was conducted using resistin mRNA abundance in omental adipose tissue as a quantitative phenotype.

RESULTS: A significant heritability of h2 = 0.23 (P = 0.003) was found for resistin mRNA abundance in omental adipose tissue. A genome scan detected a quantitative trait locus for resistin expression with an LOD score of 3.8, in the region between markers D19S431 and D19S714, corresponding to human chromosome 19 p13. This chromosomal region contains genes related to insulin resistance phenotypes, such as resistin, insulin receptor, angiopoietin-like 4 protein and LDL receptor.

CONCLUSIONS: Individual variation in resistin mRNA expression has a significant genetic component, and a gene or genes on chromosome 19 p13 may regulate resistin mRNA levels in baboon omental adipose tissue.

OBJECTIVE: To perform a meta-analysis of genome-wide linkage scans using body mass index (BMI) to identify genetic loci predisposing to obesity.

DATA: A total of 13 published genome scans on obesity have used BMI as their primary end point. Five of these 13 groups agreed to provide detailed results from their scans that were required for a meta-analysis. Collectively, these five studies included a total of 2814 individuals from 505 families.

METHODS: The results of the five studies were analysed using the GSMA (genome scans meta-analysis) method.

RESULTS: The analysis revealed significant evidence for linkage of the quantitative phenotype BMI to 8p (P<0.0005).