The metabolic syndrome is characterized by central obesity, dyslipidemia, elevated blood pressure, and hyperglycemia. The Insulin Resistance Atherosclerosis Study (IRAS) Family Study recruited extended pedigrees of Hispanic descent from San Antonio, TX (SA) and San Luis Valley, CO (SLV). Thirty-five of these pedigrees (27 SA and 8 SLV) had at least 2 individuals with metabolic syndrome (216 affected individuals and 563 affected relative pairs). The prevalence of metabolic syndrome and component criteria in subjects from these pedigrees were 35% metabolic syndrome, 43% increased waist circumference, 31% hypertriglyceridemia, 69% low HDL cholesterol, 31% increased blood pressure, and 25% either increased fasting glucose or presence of diabetes. Nonparametric linkage analysis provided evidence for linkage of metabolic syndrome to 1q23-q31 (D1S518; logarithm of odds [LOD] 1.6) with significant site heterogeneity (SA LOD 2.6 and SLV LOD 0.0), and removing all individuals with diabetes reduced, but did not eliminate, the evidence for linkage to this region (LOD 1.2). This heterogeneity may partially be explained by phenotypic differences. Members in the SA pedigrees were older, had greater central obesity, had higher prevalence of the metabolic syndrome, and were from a more urban environment than members of the SLV pedigrees. These results contribute to the growing evidence that chromosome 1q harbors at least one locus related to the metabolic precursors of diabetes.
The metabolic syndrome is a clinical entity consisting of central obesity, dyslipidemia, elevated blood pressure, and hyperglycemia, and its presence is associated with an increased risk for type 2 diabetes and cardiovascular disease (1). A major etiologic factor underlying this syndrome is insulin resistance (2). The metabolic syndrome was defined in the Third Report of the National Cholesterol Education Program Expert Panel Detection, Evaluation, and Treatment in Adults (Adult Treatment Panel III) as the presence of three or more of the following risk factors: waist circumference >88 cm in women and >102 in men, triglycerides ≥150 mg/dl, HDL cholesterol <50 mg/dl in women and <40 in men, blood pressure ≥130/85 mmHg, and fasting glucose ≥110 mg/dl or a diagnosis of diabetes (1). Recent estimates suggest that the adult prevalence of the metabolic syndrome is ∼32% in Hispanic Americans, 22% in African Americans, and 24% in Caucasian Americans (3).
It is believed that there is substantial overlap in the environmental (e.g., sedentary lifestyle and excessive caloric intake) and genetic risk factors associated with the development of metabolic syndrome and those factors responsible for the development of type 2 diabetes (1). While efforts to localize genes for the metabolic syndrome per se are in their infancy, genome-wide scans for type 2 diabetes have been conducted in a number of populations. A growing number of studies (4–10) have reported linkage of diabetes or hyperglycemia to a region on human chromosome 1q21-q25. Further support for a chromosome 1q diabetes susceptibility locus derives from the mapping of the gk2 diabetes susceptibility locus in the Goto-Kakizaki rat to a syntenic region (11).
In this report, the Insulin Resistance Atherosclerosis Study (IRAS) Family Study provides evidence that the metabolic syndrome links to chromosome 1q in 35 Hispanic pedigrees (216 individuals with the metabolic syndrome) ascertained from San Antonio, TX (SA), and San Luis Valley, CO (SLV) (12). The IRAS Family Study is a multicenter study designed to identify genes predisposing to insulin resistance, adiposity, and related traits (12). Sample summary statistics, including the numbers of affected relative pairs (ARPs), are shown in Table 1. The overall prevalence of metabolic syndrome in these families was 35%, and the proportion of subjects having risk factor values meeting Adult Treatment Panel III criteria for the definition of metabolic syndrome was 43% for waist circumference, 31% for triglycerides, 69% for HDL cholesterol, 31% for blood pressure, and 25% for fasting glucose or diabetes.
Nonparametric linkage (NPL) analysis provided strong evidence for linkage of the metabolic syndrome to chromosome 1q between markers D1S1589 and D1S518 (Table 2) (Fig. 1). In the SA pedigrees, the evidence for linkage maximized (logarithm of odds [LOD] 2.59) at 1q31 near D1S1589/D1S518 (199 cM). The linkage support interval defined by the region in which the LOD score is >1.59 (LOD-1 interval) was ∼28.5 cM in the SA pedigrees and was bounded by the nearest markers, TATC028 (1q23) and D1S1660 (1q31). There was no evidence of linkage to this region within the SLV pedigrees (LOD 0.0). Combining the two Hispanic pedigree collections yielded LOD 1.58 near D1S1589/D1S518 (Table 2).
To determine whether the evidence for linkage to 1q might reflect linkage to diabetes, three additional analyses were performed (Table 2). First, removing individuals having diabetes and only two metabolic syndrome risk factors from the analysis yielded a reduced sample containing 415 ARPs from the same 35 pedigrees. In the combined collection of Hispanic pedigrees, there was modest evidence for linkage to 1p12 near D1S534 (147 cM, LOD 1.31) and 1q31 near D1S518 (204 cM, LOD 1.49). These results were dominated by the SA pedigrees since the SLV pedigrees provided very modest evidence for linkage near the D1S534 locus (LOD 0.17) and no evidence for linkage near the D1S518 locus (LOD 0.0).
In the second ancillary analysis, all individuals with a diagnosis of diabetes were removed, providing a total of 231 ARPs from 31 pedigrees for analysis (Table 2). The combined Hispanic sample provided modest evidence for linkage to a single region near D1S518 (1q31, 205 cM, LOD 1.23). The SLV pedigrees did not contribute evidence for linkage at this locus (LOD 0.0).
In the third ancillary analysis, only pedigrees with at least two individuals diagnosed with type 2 diabetes were included (30 pedigrees containing 139 ARPs) and a linkage analysis for type 2 diabetes was performed. In both the SA and SLV pedigrees, the strongest evidence for linkage to type 2 diabetes was near D1S1660 (SA LOD 0.86, 211 cM and SLV LOD 1.50, 217 cM), providing a maximum LOD score of 1.76 near D1S1660 (211 cM, 1q31). The LOD-1 support interval spanned from 191 (D1S1589, 1q25) to 221 cM (D1S1663, 1q32).
As an exploratory analysis, a series of ordered subset analyses were computed to investigate whether any of the five specific risk factors influenced the magnitude of the evidence for linkage within the 1q region. Subsetting on the levels of the five individual risk factors defining the metabolic syndrome did not yield a statistically significant increase in the evidence for linkage within any of the three combinations of pedigrees (i.e., SA, SLV, or combined Hispanic).
There are several possible factors that could contribute to the stronger evidence for linkage detected in the 27 SA Hispanic pedigrees collected relative to the 8 SLV Hispanic pedigrees. One might posit that the primary difference in the linkage results reflects a difference in the statistical power of the respective samples. However, there are also substantive differences in the characteristics of these two samples (Table 1). First, the median age of individuals from the SA pedigrees was 6 years older than that from the SLV pedigrees (P = 0.0153). Second, men from San Antonio tended to have a greater BMI (P = 0.0323) and waist circumference (P = 0.0005) than men from San Luis Valley; a similar but not statistically significant difference was observed among women. Third, diastolic blood pressure tended to be higher in the SLV pedigrees (P = 0.0007). Fourth, San Antonio is a more urban environment relative to San Luis Valley. In summary, the SA pedigrees tended to be older, have greater central adiposity, have a higher prevalence of the metabolic syndrome, and be from a more urban environment.
These results contribute to the growing evidence that chromosome 1q harbors a diabetes-related susceptibility locus. Evidence for linkage in this population does not appear to be restricted to subjects with diabetes alone, thus suggesting that the putative disease-predisposing loci may influence susceptibility to a broader spectrum of metabolic disorders than just diabetes. Because of the relatively high prevalence of the metabolic syndrome in the population, even a modest increase in risk associated with the putative mutation will be associated with a relatively large burden of disease in the population.
Linkage of diabetes to the 1q21-25 regions has previously been reported (4–10) in European-Caucasian, Amerindian, and Chinese populations. The current results suggest that the putative mutation(s) may also be present in Hispanic Americans. Although the position that maximizes the LOD score for metabolic syndrome in our sample is the most telomeric of these seven studies, it is within the LOD-1 support interval of the majority of these studies. Conversely, our LOD-1 support interval contains the loci that maximize the evidence for linkage in the majority of these studies. The apparent presence of linkage across such diverse populations is striking and is consistent with an ancient source of the mutation.
RESEARCH DESIGN AND METHODS
Study design and data collection.
The IRAS Family Study was designed to map genes that predispose to insulin resistance, cardiovascular disease, and obesity. Details of the study design, patient recruitment, and phenotyping are described in detail elsewhere (12). Briefly, probands of Hispanic-American heritage who self-reported a large family structure were recruited from San Antonio, TX, and San Luis Valley, CO. Probands were identified from the original IRAS cohort study (13), which was designed to recruit approximately equal numbers of diabetic, impaired glucose tolerant, and normal individuals. To meet recruitment goals in the IRAS Family Study, additional probands were recruited from the general population at the San Antonio clinic. From this nonrandom collection of probands, ascertainment of families was based on large family structure and not on any clinical characteristics related to elements of the metabolic syndrome, diabetes, cardiovascular disease, or obesity. The target family structure criteria included DNA collectible on at least 12 and full phenotyping possible on at least 9 family members. We did not specifically screen for maturity-onset diabetes of the young–associated mutations or GAD antibodies for type 1 diabetes. However, our sample contained only 22 insulin-taking diabetic subjects; of these, all had an age of diabetes onset >30 years and all but 1 had a BMI >27 kg/m2.
Medical history, health behaviors, and demographic data were collected by interview. The medical history interview focused on the assessment of current health status and clinical conditions, particularly self-reported type 2 diabetes and its complications, hypertension, and cardiovascular disease events and procedures. At the time of the clinic visit, a fasting blood draw was obtained to determine plasma glucose, insulin, and lipid levels. Diabetes was defined using the American Diabetes Association criteria. The systolic and diastolic blood pressures analyzed represent the average of three measures, respectively. BMI (in kilograms per square meter) was obtained from height and weight measurements. The five individual risk factors defining the metabolic syndrome were assessed, and the individual’s metabolic syndrome status was determined independent of medications used. Linkage results were qualitatively similar when we repeated the analysis using a slightly expanded definition that included antihypertensive or antilipid medications as affected for that trait.
Whole blood obtained from each IRAS Family Study participant was frozen and shipped to the Molecular Genetics Laboratory at Wake Forest University School of Medicine. DNA from these 35 families, who completed a 10-cM genome scan (marker set 11), was shipped to the Mammalian Genotyping Center in Marshfield, WI (http://research.marshfieldclinic.org/genetics). Chromosomal maps were constructed using the marker order and distances available from the Mammalian Genotyping Center.
Statistical methods
Preliminary analyses.
Center-specific maximum likelihood estimates of allele frequencies were computed using the Recode software (D. Weeks, personal communication). Each pedigree was examined for potentially incorrectly self-reported familial relationships using the entire genome scan data (383 markers) and the Prest software (14). Inconsistent relationships were modified when the data suggested a clear alternative (e.g., full sibling to half sibling) or the genotype data were converted to missing. Each marker was examined for Mendelian inconsistencies using the Pedcheck software (15), and probable genotyping errors were converted to missing.
Linkage analysis.
Multipoint NPL regression analysis using the NPLpairs statistic was computed for the markers on chromosome 1. The NPL regression approach is a conditional logistic regression analysis in which the family-specific NPL statistics at one or more loci are the predictor variables (16). Analyses were repeated using the exponential allele-sharing model (17). Linkage analyses based on ARPs were selected to minimize the influence of misclassification errors (e.g., unaffected subjects who develop metabolic syndrome in the future).
A series of ordered subset analyses (18) were computed to investigate the influence of each of the metabolic syndrome risk factors on the evidence for linkage. First, the mean of each risk factor was calculated for each pedigree and ranked from smallest to largest. For a specific risk factor (e.g., triglycerides), the family with the largest risk factor mean was entered into the analysis and the corresponding LOD score was computed on chromosome 1 for that family. The ith ordered subset analysis proceeds by computing a linkage analysis on chromosome 1 using the subset of families with the ith largest risk factor means. This process is repeated until all families have been added to the linkage analysis. The subset of families that yield the largest LOD score on chromosome 1 is taken as the LOD score of interest. The statistical significance of the change in the LOD score was computed by a permutation test under the null hypothesis that the family’s ranking for a specific risk factor is independent of their evidence for linkage, resulting in a chromosome-wide P value. For each risk factor, the ordered subset analysis was repeated, ranking families from smallest to largest.
Result of NPL analysis on chromosome 1 for four phenotypes: metabolic syndrome (A), metabolic syndrome removing individuals who had only two of the five criteria plus a diagnosis of diabetes (B), metabolic syndrome removing all individuals with a diagnosis of diabetes regardless of the number of other criteria met (C), and type 2 diabetes (D).
Result of NPL analysis on chromosome 1 for four phenotypes: metabolic syndrome (A), metabolic syndrome removing individuals who had only two of the five criteria plus a diagnosis of diabetes (B), metabolic syndrome removing all individuals with a diagnosis of diabetes regardless of the number of other criteria met (C), and type 2 diabetes (D).
Characteristics of the 216 individuals meeting the criteria of at least three elements of the metabolic syndrome
Characteristic . | SA . | SLV . | Combined . |
---|---|---|---|
Pedigrees (n) | 27 | 8 | 35 |
Metabolic syndrome (n) | 159 | 57 | 216 |
ARPs (n) | |||
Full sibling | 97 | 41 | 138 |
Avuncular | 149 | 83 | 232 |
Cousin | 94 | 58 | 152 |
Other | 40 | 1 | 41 |
Total affected pairs | 380 | 183 | 563 |
Age (years) | 50.6 ± 13.2 (50) | 45.4 ± 13.4 (44) | 49.2 ± 13.4 (48) |
Sex (% female) | 57.9 | 56.1 | 57.4 |
BMI | |||
Women | 34.9 ± 6.6 (34.4) | 33.4 ± 5.8 (34.3) | 34.5 ± 6.4 (34.3) |
Men | 33.8 ± 5.0 (33.9) | 31.3 ± 4.4 (32.6) | 33.2 ± 4.9 (33.1) |
Diabetes (% in family) | 38.2 | 33.3 | 36.9 |
Metabolic syndrome (% in family) | 38.6 | 27.9 | 35.1 |
Risk factors for the metabolic syndrome | |||
Waist circumference | |||
Women | 100.9 ± 12.7 (98.7) | 98.8 ± 12.0 (99.2) | 100.4 ± 12.5 (98.8) |
Men | 108.5 ± 10.4 (108.1) | 101.0 ± 8.5 (103.0) | 106.5 ± 10.4 (106.7) |
Fasting plasma glucose (mg/dl) | 122.8 ± 39.2 (110.3) | 132.4 ± 55.6 (112.5) | 125.4 ± 44.3 (111.5) |
Elevated blood pressure (mmHg) | |||
Systolic | 128.3 ± 16.5 (128.5) | 125.4 ± 16.3 (123) | 127.5 ± 16.4 (126) |
Diastolic | 78.3 ± 10.6 (79.0) | 83.5 ± 9.0 (84) | 79.7 ± 10.4 (80) |
Serum triglycerides (mg/dl) | 201.4 ± 116.0 (181) | 206.5 ± 106.1 (182.5) | 202.7 ± 113.3 (181) |
HDL cholesterol (mg/dl) | |||
Women | 39.4 ± 10.7 (38) | 37.4 ± 8.0 (37) | 38.9 ± 10.1 (38) |
Men | 31.8 ± 6.5 (31) | 30.6 ± 4.8 (32) | 31.5 ± 6.1 (32) |
Characteristic . | SA . | SLV . | Combined . |
---|---|---|---|
Pedigrees (n) | 27 | 8 | 35 |
Metabolic syndrome (n) | 159 | 57 | 216 |
ARPs (n) | |||
Full sibling | 97 | 41 | 138 |
Avuncular | 149 | 83 | 232 |
Cousin | 94 | 58 | 152 |
Other | 40 | 1 | 41 |
Total affected pairs | 380 | 183 | 563 |
Age (years) | 50.6 ± 13.2 (50) | 45.4 ± 13.4 (44) | 49.2 ± 13.4 (48) |
Sex (% female) | 57.9 | 56.1 | 57.4 |
BMI | |||
Women | 34.9 ± 6.6 (34.4) | 33.4 ± 5.8 (34.3) | 34.5 ± 6.4 (34.3) |
Men | 33.8 ± 5.0 (33.9) | 31.3 ± 4.4 (32.6) | 33.2 ± 4.9 (33.1) |
Diabetes (% in family) | 38.2 | 33.3 | 36.9 |
Metabolic syndrome (% in family) | 38.6 | 27.9 | 35.1 |
Risk factors for the metabolic syndrome | |||
Waist circumference | |||
Women | 100.9 ± 12.7 (98.7) | 98.8 ± 12.0 (99.2) | 100.4 ± 12.5 (98.8) |
Men | 108.5 ± 10.4 (108.1) | 101.0 ± 8.5 (103.0) | 106.5 ± 10.4 (106.7) |
Fasting plasma glucose (mg/dl) | 122.8 ± 39.2 (110.3) | 132.4 ± 55.6 (112.5) | 125.4 ± 44.3 (111.5) |
Elevated blood pressure (mmHg) | |||
Systolic | 128.3 ± 16.5 (128.5) | 125.4 ± 16.3 (123) | 127.5 ± 16.4 (126) |
Diastolic | 78.3 ± 10.6 (79.0) | 83.5 ± 9.0 (84) | 79.7 ± 10.4 (80) |
Serum triglycerides (mg/dl) | 201.4 ± 116.0 (181) | 206.5 ± 106.1 (182.5) | 202.7 ± 113.3 (181) |
HDL cholesterol (mg/dl) | |||
Women | 39.4 ± 10.7 (38) | 37.4 ± 8.0 (37) | 38.9 ± 10.1 (38) |
Men | 31.8 ± 6.5 (31) | 30.6 ± 4.8 (32) | 31.5 ± 6.1 (32) |
Data are means ± SD (median) unless noted otherwise.
Summary of the NPL analyses on chromosome 1
. | Position (cM) . | LOD . | LOD-1 interval . | P value . | Flanking markers . |
---|---|---|---|---|---|
Metabolic syndrome | |||||
Combined Hispanic | 140.5 | 1.32 | 114.5–222.5 | 0.0138 | ATA42G12/ |
205.0 | 1.58 | 131.0–220.0 | 0.0070 | D1S518/ | |
SA | 139.5 | 1.16 | 58.5–223.0 | 0.0207 | ATA42G12/ |
199.0 | 2.59 | 180.5–209.0 | 0.0006 | D1S1589/D1S518 | |
SLV | 286.5 | 0.56 | 207.0– | 0.1078 | /TTTA049 |
Metabolic syndrome removing individuals with only two risk factors plus diabetes | |||||
Combined Hispanic | 146.5 | 1.31 | 129.5–220.5 | 0.0139 | /D1S534 |
204.0 | 1.49 | 131.5–218.0 | 0.0089 | D1S518/ | |
SA | 150.0 | 1.03 | 52.5–221.0 | 0.0294 | /D1S534 |
200.0 | 2.23 | 179.5–208.0 | 0.0013 | D1S518/ | |
SLV | 282.0 | 0.67 | 204.5– | 0.0789 | /TTTA049 |
Metabolic syndrome removing all diabetic subjects | |||||
Combined Hispanic | 205.0 | 1.23 | 131.5–224.0 | 0.0172 | D1S518/ |
SA | 202.0 | 1.88 | 193.5–212.5 | 0.0033 | D1S518/ |
SLV | 139.0 | 0.43 | 131.5–151.0 | 0.1599 | /D1S1627 |
Diabetes | |||||
Combined Hispanic | 211.0 | 1.76 | 190.5–220.5 | 0.0045 | /D1S1660 |
SA | 210.5 | 0.86 | 167.5–232.0 | 0.0469 | /D1S1660 |
SLV | 216.5 | 1.50 | 192.0–221.5 | 0.0085 | D1S1660/ |
. | Position (cM) . | LOD . | LOD-1 interval . | P value . | Flanking markers . |
---|---|---|---|---|---|
Metabolic syndrome | |||||
Combined Hispanic | 140.5 | 1.32 | 114.5–222.5 | 0.0138 | ATA42G12/ |
205.0 | 1.58 | 131.0–220.0 | 0.0070 | D1S518/ | |
SA | 139.5 | 1.16 | 58.5–223.0 | 0.0207 | ATA42G12/ |
199.0 | 2.59 | 180.5–209.0 | 0.0006 | D1S1589/D1S518 | |
SLV | 286.5 | 0.56 | 207.0– | 0.1078 | /TTTA049 |
Metabolic syndrome removing individuals with only two risk factors plus diabetes | |||||
Combined Hispanic | 146.5 | 1.31 | 129.5–220.5 | 0.0139 | /D1S534 |
204.0 | 1.49 | 131.5–218.0 | 0.0089 | D1S518/ | |
SA | 150.0 | 1.03 | 52.5–221.0 | 0.0294 | /D1S534 |
200.0 | 2.23 | 179.5–208.0 | 0.0013 | D1S518/ | |
SLV | 282.0 | 0.67 | 204.5– | 0.0789 | /TTTA049 |
Metabolic syndrome removing all diabetic subjects | |||||
Combined Hispanic | 205.0 | 1.23 | 131.5–224.0 | 0.0172 | D1S518/ |
SA | 202.0 | 1.88 | 193.5–212.5 | 0.0033 | D1S518/ |
SLV | 139.0 | 0.43 | 131.5–151.0 | 0.1599 | /D1S1627 |
Diabetes | |||||
Combined Hispanic | 211.0 | 1.76 | 190.5–220.5 | 0.0045 | /D1S1660 |
SA | 210.5 | 0.86 | 167.5–232.0 | 0.0469 | /D1S1660 |
SLV | 216.5 | 1.50 | 192.0–221.5 | 0.0085 | D1S1660/ |