Genetic and environmental determinants play critical roles in insulin resistance and β-cell function. A model of the complex feedback system for maintenance of glucose tolerance has been developed that reflects the constraint of glycemia within narrow physiologic limits. The “glucose homeostasis” model is described by insulin sensitivity (SI), glucose disposition (SG), acute insulin response to glucose (AIRG), and disposition index (DI). Relatively little is known about the genetic basis of glucose homeostasis phenotypes or their relationship to risk of diabetes and atherosclerotic cardiovascular disease. A genome scan for glucose homeostasis phenotypes in nondiabetic subjects has been carried out in African-American (n = 21) and Hispanic (n = 45) extended families as part of the IRAS Family Study. In African-American families, there was significant evidence for linkage of DI between D11S2371 and D11S2002 (logarithm of odds [LOD] = 3.21) at 81 cM, and in the combined sample of African-American and Hispanic families, there was evidence at GATA117D01 (140 cM) on chromosome 11 (LOD = 2.21). Evidence of linkage was also observed for SI in Hispanic (LOD = 2.28, between D15S822 and GTTTT001) and AIRG in African-American families (LOD = 2.73, between D4S1625 and D4S1629; and LOD = 2.56 at PAH (phenylalanine hydroxylase) on chromosome 12). These results provide impetus for future positional cloning of quantitative trait loci (QTLs). Identifying genes in these regions should provide insight into the nature of the metabolic syndrome and diabetes, and facilitate the development of more effective therapies for prevention and treatment of diabetes and other diseases associated with disordered glucose metabolism.

Risk of type 2 diabetes is mediated, in part, by genetic factors (1). Given that both insulin resistance and dysfunctional β-cells are predictors of type 2 diabetes (26), it may be necessary to examine the genetic basis of each of these components in order to get a better understanding of the etiology of type 2 diabetes. Measurement of β-cell function and insulin sensitivity derived from the intravenous glucose tolerance test (IVGTT) with minimal model (MINMOD) analysis have shown high correlations with results from the euglycemic-hyperinsulinemic clamp (7). In addition, insulin sensitivity (SI) has been shown to have a greater contribution by genetic factors than other surrogate measures (such as fasting insulin or homeostasis model assessment [HOMA]) (8). These measures of glucose homeostasis obtained from the IVGTT and MINMOD analysis (SI, SG [glucose effectiveness], AIRG [acute insulin response to glucose], and DI [disposition index]) may represent important intermediate phenotypes for contributors to the genetic risk of type 2 diabetes.

Several studies have noted the influence of genetic factors on quantitative traits of glucose homeostasis. Moderate estimates of genetic contribution (heritability) to measures of glucose homeostasis have been reported, despite different study designs, ranging from pedigrees “loaded” with multiple cases of type 2 diabetes (911), sibships with polycystic ovary syndrome (12), nuclear families ascertained on the basis of a hypertensive index case (13), nuclear families involved in an exercise-training program (14), and Pima Indian pedigrees (15,16). Over all populations and family structures, “moderate” estimates of heritability for insulin sensitivity (SI, 36%), glucose effectiveness (SG, 17%), acute insulin response (AIR, 48%), and disposition index (DI, 43%) were observed, suggesting the importance of genetic factors contributing to variation in glucose homeostasis.

The IRAS Family Study is an extension of the IRAS (Insulin Resistance Atherosclerosis Study), the primary objective of which was to determine the relationship between a direct measure of insulin resistance and carotid atherosclerosis in 1,625 unrelated individuals in four communities (17,18). Families (Hispanic and African American) were ascertained on family size and not on extreme phenotype (e.g., diabetes). The clinical examination included direct measures of glucose homeostasis (SI, SG, AIRG, and DI) using the IVGTT with MINMOD analysis and blood collection for DNA. The present study examines the heritability of measures of glucose homeostasis in the IRAS Family Study, as well provides evidence for quantitative trait loci (QTLs) based upon a genome scan in the first set of genetic data based on 66 families (∼1,200 subjects) that contribute to the variation in parameters of glucose homeostasis in extended pedigrees of Hispanic and African-American descent.

Three clinical sites recruited and examined members of large families of Hispanic (San Antonio, Texas, and San Luis Valley, Colorado) or African-American ethnicity (Los Angeles, California) over a 2.5-year period (2000–2002). Potential probands of these families were identified from the community-based “parent study” (the IRAS [17]). Individuals reporting large family sizes were invited to participate with their relatives in the IRAS Family Study. Ascertainment was also supplemented with non-IRAS participants (and their family members) who were recruited from the general population. Ascertainment of families was not based on extreme phenotypes (e.g., BMI or diabetes status), but on large family size.

Phenotypic characterization.

All participants in the IRAS Family Study were interviewed by project staff who were trained and monitored centrally. Subjects provided information concerning their medical history (current health status and clinical conditions, including type 2 diabetes and its complications, hypertension, and cardiovascular disease events and procedures), health behaviors, and demographic features. Indexes of glucose homeostasis were assessed by the frequently sampled intravenous glucose tolerance test (IVGTT) (19), with minimal model (MINMOD) analyses (20). Two modifications of the original protocol were employed, similar to that used in the original IRAS (17). An injection of insulin, rather than tolbutamide, was used to ensure adequate plasma insulin levels for the accurate computation of insulin resistance across a broad range of glucose tolerance (21), and a reduced sampling protocol requiring 12 (rather than 30) plasma samples (22) was used. Glucose in the form of a 50% solution (0.3 g/kg) and regular human insulin (0.03 units/kg) were injected through an intravenous line at 0 and 20 min, respectively. Blood was collected at −5, 2, 4, 8, 19, 22, 30, 40, 50, 70, 100, and 180 min for the determination of plasma glucose and insulin concentrations.

SI and SG were calculated using MINMOD software. AIRG was the mean insulin increment in the plasma insulin concentration above the basal in the first 8 min after the administration of glucose. DI was calculated as DI = AIR × SI (23). Plasma glucose and insulin values were obtained using standard methods.

Genotyping.

Whole blood obtained from each IRAS Family Study participant was frozen and stored at −70°C, then shipped in batches to the Molecular Genetics Laboratory at Wake Forest University School of Medicine. Genomic DNA was extracted from the blood samples and stored. Any sample that yielded <100 μg of DNA was considered “insufficient,” and recontact with the participant was made to collect an additional blood sample. Purification of genomic DNA from thawed blood samples was performed using a commercially available DNA extraction kit (PureGene; Gentra Systems, Minneapolis, MN). DNA was quantitated using fluorometric and absorbance spectroscopy. A genome scan consisting of markers at ∼9-cM intervals was performed by the National Heart, Lung, and Blood Institute (NHLBI) Mammalian Genotyping Service (MGS, Marshfield, WI).

Family characteristics.

Descriptive statistics of the pedigree structure for the families used in these analyses are shown in Table 1. A total of 66 extended families (21 African-American families from Los Angeles, 33 Hispanic families from San Antonio, and 12 Hispanic families from San Luis Valley). These families contain 876 nondiabetic individuals with genotypic and phenotypic data that were included in these analyses (genotypic data from 133 diabetic individuals were included to compute identity-by-descent [IBD] statistics, although their phenotypic data were considered “missing”). Given the large family size (mean of 16 informative members [genotyped] per family), these pedigrees produced 5,738 relative pairs (including 978 full sibpairs). Participating family members were, on average, 41 years of age, and 57% of the family members were female. Mean pedigree size ranged from 14 phenotyped and genotyped individuals in San Antonio to 23 individuals in San Luis Valley.

Pedigree error detection.

Before performance of linkage analysis, all pedigrees were subjected to a series of analyses to identify potential genetic marker problems and instances of nonpaternity using PREST software (24). This method utilizes the genome scan data to determine the likelihood of each specified relationship given the genetic data. Unlikely marker genotypes were resolved by recoding the specific genotype to “unknown.” Occurrences of nonpaternity were resolved by changing the pedigree structure to that which was most likely, then repeating the analysis to confirm appropriate relationships. Eighteen families had pedigree structures altered to reflect the most likely relationships. While there were differences in the number of families altered by clinic site (11 of 33 from San Antonio, 2 of 12 from San Luis Valley, and 5 and 21 from Los Angeles), there was little to suggest that there was any bias in recruitment. All but two of the families altered had at least one full-sib relationship changed to a half-sib relationship. The remaining two changes reflected previously unrecognized adoptions.

Data transformation.

For genetic variance component analyses, the distributions of glucose homeostasis phenotypes were transformed to approach normality. For SI, there were 67 subjects (of 871 with frequently sampled intravenous glucose tolerance test, 7.7%) with SI of zero. As in previous studies by our group (and others), the transformed value, ln(SI + 1), was used in the subsequent analyses. For AIRG, there were 43 subjects (of 871, 4.9%) with nonpositive AIRG values. A signed-square root transformation was used {when AIRG is >0, the value for analysis is the sqrt(AIRG); when AIRG is <0, the value for analysis is –sqrt[abs(AIRG)], where “sqrt” is the square root function and “abs” is the absolute value}. Thus, by extension, the value of DI (= SI × AIRG) is zero when SI is zero, is positive when SI is >0 and AIRG is >0, and is negative when SI is >0 and AIRG is <0. These transformations resulted in reduction of skewness and kurtosis to that as expected under a normal distribution.

Statistical genetic analyses.

Glucose homeostasis parameters for individuals diagnosed with diabetes (by self-report or by fasting glucose >126 mg/dl) were coded as “unknown” phenotype; however, all genotypic data were used in linkage analyses. In the IRAS Family Study, there were 133 individuals classified as “diabetic” (12.7%).

For each glucose homeostasis phenotype, heritability (as well as marker allele frequency and IBD) was estimated for the entire sample, for each ethnic group (African Americans and Hispanics), and for each center. The models used in the analyses of the entire dataset included age, sex, BMI, and center (San Antonio, San Luis Valley, Los Angeles) as covariates. Thus, the heritability estimates reflect the “residual” heritability after the effects of covariates are removed. Within the Hispanic ethnic group, the models included age, sex, BMI, and center (San Antonio, San Luis Valley). For African Americans (ascertained only at the Los Angeles center), models included age, sex, and BMI as covariates. Tests for linkage were carried out for all genetic markers for each phenotype related to glucose homeostasis.

Estimation of heritability and testing for evidence of linkage was assessed using a multipoint variance component procedure implemented in the SOLAR software package (25). The basic variance component approach to linkage analysis has been described extensively (25,26,27). The strength of the approach is that it is relatively model-free, requiring no prior knowledge of segregation parameters. Determination of the support for linkage for each phenotype uses information from all possible pedigree relationships simultaneously based upon the expected covariance among relatives as a function of the IBD relationships at each genetic marker locus. As a second step, the residual heritability (that obtained after removal of variation due to covariates) is then partitioned into a component attributed to the QTL (linkage) and the residual additive genetic effects at that marker locus. The hypothesis of “no linkage” (heritability of the QTL is zero) was tested by comparing the likelihood of a model without a QTL to the likelihood of one in which the heritability due to the QTL was estimated. In the test of linkage, in which the value of the parameter of interest is set to a boundary (i.e., zero) under the null hypothesis, twice the difference in ln(likelihood) of the two models is distributed asymptotically as a 1/2:1/2 mixture of a χ21 and a point mass at zero (28). LOD scores are obtained by converting the ln(likelihood) values into log10 values.

Maximum likelihood estimates of the model parameters were obtained, and the likelihood of the pedigree data was computed under each formulated model using the SOLAR software package. Results are presented if 1) they have “promising” evidence of linkage (29)—an observed LOD >1.75 (P < 0.0023) that is equivalent to one false-positive region for each genome scan given a set marker density—or 2) an empiric P value <0.005. Single-point empirical P values were computed at each location across the genome using the lodadj procedure within SOLAR (30). Specifically, lodadj assumes a fully informative marker and “gene drops” in each simulation. This provides a simulation-based estimate of statistical significance that is robust to model assumptions and estimates the empirical distribution of LOD scores across the entire likelihood-based LOD score distribution. All IBD and multipoint IBD statistics were estimated within center using within-center marker allele frequencies. Combined analyses maintained the center-specific IBD estimates. Analyses were computed separately by center (San Antonio Hispanic, San Luis Valley Hispanic, and Los Angeles African-American groups) and in the combined Hispanic group.

Descriptive statistics.

Descriptive statistics for the family members with genotype and phenotype data are shown in Table 2. Thirteen percent of participating family members had diabetes, which is consistent with that observed in the general population. Hispanic family members were more insulin sensitive (higher SI; SI = 1.98 in San Antonio and SI = 2.41 in San Luis Valley participants) than African Americans (SI = 1.76) (P < 0.01). AIRG in African Americans (914) was higher than that in Hispanic Americans (726 in San Antonio and 807 in San Luis Valley). DI was significantly (P < 0.05) lower in San Antonio Hispanics (1,112) than either San Luis Valley Hispanics (1,658) or Los Angeles African Americans (1,429). Thus, there were differences in glucose homeostasis phenotypes (SI, SG, AIRG and DI) between the three centers. There were no differences in SG and DI when comparing African Americans with Hispanics (San Antonio and San Luis Valley), although differences between SI and AIRG remained.

Familial aggregation of glucose homeostasis phenotypes.

The estimated residual heritability (h2) of glucose homeostasis phenotypes, determined within each clinical site, is shown in Table 3. Several conclusions can be drawn from these estimates. First, the estimated heritability for each glucose homeostasis phenotype varies dramatically across centers, likely due to small number of families (reduced power), but possibly due to underlying (unobserved) genetic and environmental determinants contributing to variation in the phenotype. Even when both Hispanic sites are combined (increasing family number, relative pairs, and statistical power), differences remain between the two ethnic groups. Second, the estimated heritability of some phenotypes is not significantly different from zero (SI in African Americans [0.043 ± 0.088] and SG in San Antonio Hispanics [0.015 ± 0.088]). Although this suggests no ability to map QTLs in this subset, the power for the estimates was low and thus these results should be interpreted with caution. Further, there was overlap with the estimated heritability for SI and SG with that in other centers (whose estimates were significantly greater than zero), suggesting that the estimated low heritability was likely spurious. Third, while the heritability estimates varied across centers, there were no significant differences in heritability across the two Hispanic sites (SI, P = 0.09; SG, P = 0.10; AIRG, P > 0.50; DI, P > 0.50); thus, the combined estimates in Hispanics more accurately reflect the heritability within this sample, despite differences in mean value for some glucose homeostasis parameters. Fourth, the contribution of measured covariates to the variation within a glucose homeostasis phenotype varied little across center. Thus, true differences in aggregation were not attributable to effects of age, sex, or BMI. Fifth, in the entire sample of IRAS Family Study pedigrees, there is significant evidence for heritability of each glucose homeostasis phenotype, with heritability ranging from low (h2 = 0.14 for SG) to modest (h2 = 0.29 for SI) to moderate (h2 = 0.47 for DI and h2 = 0.56 for AIRG) levels. These estimates suggest that significant familial aggregation exists, and there is rationale for mapping QTLs for the phenotypes.

Linkage analyses for glucose homeostasis phenotypes

SI.

Evidence for linkage across the genome is shown in Fig. 1. There was no LOD >1.75 (our criteria for promising evidence of linkage) or a region with an empiric P < 0.005 over all centers combined or in African Americans. In Hispanics, the strongest evidence for linkage to SI was found on chromosome 15 (LOD = 2.28, empiric P = 0.0010, location between markers D15S822 and GTTTT001 at position 17 cM). There was relatively equal contribution to this linkage peak from families in San Antonio (LOD = 0.94 at D15S822) and San Luis Valley (LOD = 1.18 between D15S822 and GTTTT001). In the San Antonio center, there was also evidence for linkage to SI on chromosome 14 (LOD = 2.08, empiric P = 0.0013, location between markers D14S592 and GGAA30H04 at position 65 cM). In the San Luis Valley center, evidence supporting linkage to SI was observed on chromosome 7 (LOD = 1.75, empiric P = 0.0046, location between markers D7S817 and D7S2846 at position 55 cM).

SG.

There was no occurrence of a LOD >1.75 for this phenotype. Over all centers, the strongest evidence for linkage to SG was found on chromosome 14 (LOD = 1.54, empiric P = 0.0079, location at marker D14S617 at position 106 cM).

AIRG.

Over all centers (Fig. 3), the strongest evidence for linkage to AIRG was found on chromosome 11 (LOD = 2.33, empiric P = 0.0014, location between markers D11S2002 and D11S2000 at position 92 cM). At this location, the QTL for AIRG accounted for 38% of the total genetic variance. Both African-American families (LOD = 1.52 at position 89 cM) and Hispanic families (LOD = 0.99 at position 93 cM) contributed to the evidence for linkage at the location between markers D11S2002 and D11S2000.

In African Americans, strong evidence for linkage to AIRG was observed on chromosome 4 (LOD = 2.73, empiric P = 0.0030, location between markers D4S1625 and D4S1629 at position 150 cM) and chromosome 12 (LOD = 2.56, empiric P = 0.0021, PAH [phenylalanine hydroxylase] at position 110 cM). A detailed summary of the multipoint linkage results in African-American families for AIRG is shown in Fig. 4A (chromosome 4) and Fig. 4B (chromosome 12). Support for linkage on chromosome 6 was also found in Hispanic families (LOD = 1.72, empiric P = 0.0032, location at marker ATTT030 at position 15 cM). The San Antonio families provided the majority of linkage evidence for this location (LOD = 1.83 at position 13 cM with marker TTA032), as families in the San Luis Valley did not provide much evidence for linkage at this site (LOD = 0.31 at position 16 cM with marker ATTT030).

DI.

Over all centers, the strongest evidence for linkage to DI was found on chromosome 11 (LOD = 2.43, empiric P = 0.0004, location at marker GATA117D01 at position 140 cM) (Fig. 5). A second linkage peak on chromosome 11 occurred at position 89 cM (LOD = 1.94, empiric P = 0.0026, location between markers D11S2002 and D11S2000, same location as for AIRG).

In African Americans, the strongest linkage to DI was observed on chromosome 11 (LOD = 3.21, empiric P = 0.0004, at marker D11S2002 at position 81 cM). This is the same area as the second (and more proximal) peak on chromosome 11 in the combined sample. The detailed map and results of the multipoint linkage analysis for DI on chromosome 11 in African Americans is shown in Fig. 6. The chromosome 11 QTL accounted for 49% of the genetic variance in DI. In Hispanics, the strongest evidence for linkage to DI was also found on chromosome 11 (LOD = 2.19, empiric P = 0.0014, location between markers GATA117D01 and ATA27C11 at position 143 cM). This location was in the area of the first (strongest and more distal) chromosome 11 signal in the combined data. This QTL accounted for 69% of the genetic variance in the Hispanic families. The contribution to the linkage in Hispanic families was mainly from the San Antonio families (LOD = 2.71 located at marker ATA27C11 at position 148 cM). In the San Luis Valley center, there was little evidence for linkage to chromosome 11 at these locations. In contrast, in the San Luis Valley families, there was evidence for linkage to DI over a broad region on chromosome 19 (LOD = 1.95, empiric P = 0.0020) centered at marker D19S714 at position 42 cM.

This report presents results of a genome scan for glucose homeostasis phenotypes in large families of African-American and Hispanic-American descent. The analyses support evidence of familial aggregation of glucose homeostasis phenotypes, multiple regions of linkage to AIRG on chromosome 11p, and strong evidence for linkage of DI to chromosome 11q.

There have been few genome scans that have evaluated directly measured phenotypes of glucose homeostasis, in contrast to numerous studies that have reported linkage to measures based on fasting values alone, including fasting glucose, fasting insulin, and/or the HOMA index. These “surrogate” measures of physiological processes appear to have different genetic architecture from these directly measured parameters (8). Three studies with the preferred direct measurements of glucose homeostasis are the FUSION (Finland-U.S. Investigation of Non-Insulin-Dependent Diabetes Mellitus Genetics) study (31), the study of hypertension and insulin resistance in Hispanics (13,32), and the HERITAGE Family Study (33). FUSION investigators conducted a genome scan for type 2 diabetes–associated quantitative traits in 580 Finnish families. Regions of linkage in the IRAS Family Study for SI, SG, AIRG, and DI did not overlap with those reported by FUSION.

The Hispanic hypertension and insulin resistance study investigated the relationship between insulin resistance and hypertension (32) in Hispanic families from Los Angeles. The euglycemic glucose clamp provided a whole- body measure of insulin sensitivity (glucose infusion rate) with strongest linkage on chromosomes 1q and 7p. The region of support for glucose infusion rate in the Hispanic hypertensive families occurred with marker D7S1818 (LOD = 1.82 at position 69 cM). The support interval overlapped with the region on 7p supporting linkage to SI in the IRAS Family Study Hispanic families with marker D7S1808 (LOD = 1.43 at 42 cM).

In the HERITAGE Family Study (33), several regions with LOD ≥1.75, corresponding to a P ≤ 0.0023, were detected for SI (1q41 and 8p23.2), AIRG (4q32.1 and 10p15.3), and DI (13q32.1) in Caucasian families. In African-American families, two regions of linkage were observed, one for SG (9p11.2) and one for SI (10q26.11). Although the IRAS Family Study did not study Caucasian families, linkage in Hispanic families provided no replication for SI (1q41 or 8p23.2) that was observed in the HERITAGE Family Study. The strongest evidence for linkage in the IRAS Family Study Hispanic sites occurred on chromosomes 14p (San Antonio), 7p (San Luis Valley), and 15p (combined Hispanic). The IRAS Family Study linkage peaks for AIRG in Hispanic families failed to reach the LOD ≥1.75 threshold, and the highest LOD in San Antonio (13q), San Luis Valley (3q), and the combined Hispanic (6p) groups failed to replicate the HERITAGE Family Study linkage results. However, the maximum LOD in the African-American sample on chromosome 4q (but not 12q) did replicate the linkage observed in HERITAGE Caucasians. In the IRAS Family Study, no LOD ≥1.75 was observed for SG, and the locations of the maximum LOD scores were not at the locations that were identified in the HERITAGE Family Study. For DI, the HERITAGE Family Study linkage on chromosome 13q32.1 was not observed in the IRAS Family Study.

There appears to be greater evidence of replication of linkage signals between the IRAS Family Study and the Hispanic hypertension–insulin resistance study than with FUSION. One possible explanation is that the two former studies focus on Hispanic Americans, while the FUSION study represents Caucasians of Finnish ancestry. Another possibility is disease-associated heterogeneity of insulin sensitivity. While the hypertension–insulin resistance and IRAS Family Study families were ascertained through families in a manner that would reflect random ascertainment with respect to cardiovascular disease, the FUSION families were ascertained through affected sibpairs with type 2 diabetes.

DI was the phenotype within the IRAS Family Study that provided the most compelling evidence for linkage among measures of glucose homeostasis. The evidence for linkage in the entire family collection was strongest on chromosome 11q (11q24.1). This region contains relatively few candidates, although a matrix metalloproteinase (MMP27) and three genes associated with cell adhesion molecules (THY1, OPCML, and EVA1) map to the same location. Within the African-American pedigrees, stronger evidence of linkage was observed at 11q13–14. This region is the same that supported linkage for AIRG in these same data, suggesting potential pleiotropic effects of genes. Importantly, DI represents the ability of the β-cells to compensate for insulin resistance and, therefore, is an index of β-cell functionality (23).

In the IRAS Family Study, there were several regions that supported linkage with AIRG. Linkage was suggested to AIRG at 4q31.23–32.2, 11q14.2–22.3, and 12q22–24.2 (PAH). Linkage to fasting insulin has previously been reported near 12q23 in the Old Order Amish (34). In the 11q14–22 region, there were earlier reports of linkage to 2-h insulin (35), SI and fasting insulin (31), and BMI (36). The Hispanic pedigrees also provided evidence of linkage to AIRG at 1q23–25 (Hispanic LOD = 1.40, San Antonio LOD = 1.60), which is near the region that has supported linkage to type 2 diabetes (36,37,38) and percent fat (39). Within the Hispanic pedigrees, there was also evidence of linkage at 6q23 (Hispanic LOD = 1.52), 6p25–24 (Hispanic LOD = 1.72, San Antonio LOD = 1.83), and 13q23 (San Antonio LOD = 1.58). For both the 6q23 and 13q23 regions, evidence for linkage to type 2 diabetes (36) and impaired homeostasis (40) have been previously reported.

The glucose homeostasis parameters, their genetic contribution to variation and their purported location in the genome, should provide insights to the adaptive functionality of the pancreas in the intact individual and should, therefore, be critical to the etiology of insulin resistance and type 2 diabetes. Thus, an argument may ultimately be made for assessment of these phenotypes in subjects who have familial risk for type 2 diabetes. Although multiple components may be necessary for predicting an individual’s risk for developing diabetes, the concept of genetic control of variation in glucose homeostasis may be of importance to clarify differential environmental and genetic effects on islet cell function. Although several of the glucose homeostasis components have significant genetic contribution to their variation (AIRG and DI), others (SI and SG) appear to have limited genetic control. It should be noted, however, that the magnitude of heritability of a phenotype and strength of linkage may not be highly correlated (41). Given the biological significance of glucose homeostasis, the strong evidence for genetic control of variation in glucose homeostasis, and the mapping of QTLs to multiple chromosomal locations, subsequent targeted studies can be proposed to identify positional candidate genes useful for defining risk of disease.

FIG. 1.

Genome scan for SI.

FIG. 1.

Genome scan for SI.

FIG. 2.

Genome scan for SG.

FIG. 2.

Genome scan for SG.

FIG. 3.

Genome scan for AIRG.

FIG. 3.

Genome scan for AIRG.

FIG. 4.

A: Linkage of AIRG to chromosome 4 in African Americans. B: Linkage of AIRG to chromosome 12 in African Americans.

FIG. 4.

A: Linkage of AIRG to chromosome 4 in African Americans. B: Linkage of AIRG to chromosome 12 in African Americans.

FIG. 5.

Genome scan for DI.

FIG. 5.

Genome scan for DI.

FIG. 6.

Linkage of DI to chromosome 11 in African Americans.

FIG. 6.

Linkage of DI to chromosome 11 in African Americans.

TABLE 1

Characteristics in members of 66 extended pedigrees in the IRAS Family Study with glucose homeostasis phenotype* and genome scan data

VariableSan Antonio HispanicSan Luis Valley HispanicLos Angeles African AmericanTotal
Pedigrees 33 12 21 66 
Individuals 363 229 284 876 
Relative pairs 1,817 2,124 1,797 5,738 
Sibling pairs 338 383 257 978 
Age (years) 40.8 ± 13.7 39.0 ± 13.7 42.2 ± 14.3 40.8 ± 13.9 
Sex (% female) 61.2 52.6 53.6 56.5 
VariableSan Antonio HispanicSan Luis Valley HispanicLos Angeles African AmericanTotal
Pedigrees 33 12 21 66 
Individuals 363 229 284 876 
Relative pairs 1,817 2,124 1,797 5,738 
Sibling pairs 338 383 257 978 
Age (years) 40.8 ± 13.7 39.0 ± 13.7 42.2 ± 14.3 40.8 ± 13.9 
Sex (% female) 61.2 52.6 53.6 56.5 

Data are means ± SD unless otherwise specified. *Total number phenotyped decreased due to presence of diabetes and other exclusions.

TABLE 2

Phenotypic characteristics of participants with genotypic and glucose homeostasis phenotypic data in 66 extended pedigrees in the IRAS Family Study

VariableSan Antonio HispanicSan Luis Valley HispanicLos Angeles African AmericanTotal
Diabetes (%) 16.2 9.9 10.3 12.7 
BMI (kg/m230.2 ± 6.4 (29.3) 27.5 ± 5.6 (26.9) 29.1 ± 6.6 (27.8) 29.1 ± 6.3 (28.3) 
SI (×10−5 min−1/[pmol/l]) 1.98 ± 2.03 (1.48) 2.41 ± 1.99 (1.92) 1.76 ± 1.21 (1.49) 2.02 ± 1.81 (1.57) 
SG (min−10.019 ± 0.008 (0.019) 0.024 ± 0.010 (0.023) 0.021 ± 0.008 (0.020) 0.021 ± 0.009 (0.020) 
AIRG (pmol/l) 726 ± 616 (555) 807 ± 629 (658) 914 ± 793 (668) 808 ± 686 (625) 
DI (×10−5 min−11,112 ± 998 (925) 1,658 ± 1,551 (1,229) 1,429 ± 1,330 (1,058) 1,358 ± 1,290 (1016) 
VariableSan Antonio HispanicSan Luis Valley HispanicLos Angeles African AmericanTotal
Diabetes (%) 16.2 9.9 10.3 12.7 
BMI (kg/m230.2 ± 6.4 (29.3) 27.5 ± 5.6 (26.9) 29.1 ± 6.6 (27.8) 29.1 ± 6.3 (28.3) 
SI (×10−5 min−1/[pmol/l]) 1.98 ± 2.03 (1.48) 2.41 ± 1.99 (1.92) 1.76 ± 1.21 (1.49) 2.02 ± 1.81 (1.57) 
SG (min−10.019 ± 0.008 (0.019) 0.024 ± 0.010 (0.023) 0.021 ± 0.008 (0.020) 0.021 ± 0.009 (0.020) 
AIRG (pmol/l) 726 ± 616 (555) 807 ± 629 (658) 914 ± 793 (668) 808 ± 686 (625) 
DI (×10−5 min−11,112 ± 998 (925) 1,658 ± 1,551 (1,229) 1,429 ± 1,330 (1,058) 1,358 ± 1,290 (1016) 

Data are means ± SD (median).

TABLE 3

Heritability estimates of glucose homeostasis phenotypes within and across centers and ethnic groups in the IRAS Family Study

PhenotypeSan Antonio HispanicSan Luis Valley HispanicHispanicLos Angeles African AmericanTotal
SI 0.41 (0.10)* 0.24 (0.12) 0.35 (0.08) 0.04 (0.09) 0.29 (0.06) 
(×10−5 min−1/[pmol/l]) P < 0.001 P = 0.001 P < 0.001 P = 0.30 P < 0.001 
 σ2cov = 38%† σ2cov = 40%† σ2cov = 38%† σ2cov = 28%† σ2cov = 35%‡ 
SG 0.02 (0.09) 0.18 (0.12) 0.11 (0.07) 0.32 (0.17) 0.14 (0.06) 
(min−1P = 0.43 P = 0.023 P = 0.026 P = 0.005 P = 0.005 
 σ2cov = 15% σ2cov = 15% σ2cov = 19% σ2cov = 14% σ2cov = 17% 
AIRG 0.50 (0.13) 0.53 (0.15) 0.50 (0.09) 0.69 (0.14) 0.56 (0.08) 
(pmol/l) P < 0.001 P < 0.001 P < 0.001 P < 0.001 P < 0.001 
 σ2cov = 7% σ2cov = 9% σ2cov = 7% σ2cov = 11% σ2cov = 8% 
DI 0.46 (0.14) 0.30 (0.16) 0.37 (0.10) 0.73 (0.15) 0.47 (0.09) 
(×10−5 min−1P < 0.001 P = 0.007 P < 0.001 P < 0.001 P < 0.001 
 σ2cov = 19% σ2cov = 21% σ2cov = 24% σ2cov = 20% σ2cov = 23% 
PhenotypeSan Antonio HispanicSan Luis Valley HispanicHispanicLos Angeles African AmericanTotal
SI 0.41 (0.10)* 0.24 (0.12) 0.35 (0.08) 0.04 (0.09) 0.29 (0.06) 
(×10−5 min−1/[pmol/l]) P < 0.001 P = 0.001 P < 0.001 P = 0.30 P < 0.001 
 σ2cov = 38%† σ2cov = 40%† σ2cov = 38%† σ2cov = 28%† σ2cov = 35%‡ 
SG 0.02 (0.09) 0.18 (0.12) 0.11 (0.07) 0.32 (0.17) 0.14 (0.06) 
(min−1P = 0.43 P = 0.023 P = 0.026 P = 0.005 P = 0.005 
 σ2cov = 15% σ2cov = 15% σ2cov = 19% σ2cov = 14% σ2cov = 17% 
AIRG 0.50 (0.13) 0.53 (0.15) 0.50 (0.09) 0.69 (0.14) 0.56 (0.08) 
(pmol/l) P < 0.001 P < 0.001 P < 0.001 P < 0.001 P < 0.001 
 σ2cov = 7% σ2cov = 9% σ2cov = 7% σ2cov = 11% σ2cov = 8% 
DI 0.46 (0.14) 0.30 (0.16) 0.37 (0.10) 0.73 (0.15) 0.47 (0.09) 
(×10−5 min−1P < 0.001 P = 0.007 P < 0.001 P < 0.001 P < 0.001 
 σ2cov = 19% σ2cov = 21% σ2cov = 24% σ2cov = 20% σ2cov = 23% 

*Heritability (standard error h2); †percent of variance due to covariates (age, sex, and BMI); ‡percent of variance due to covariates (age, sex, center, and BMI).

This research was supported in part by National Institutes of Health grants HL060894 (D.W.B.), HL060919 (S.M.H.), HL060931 (R.N.B.), HL060944 (L.E.W.), HL061019 (J.M.N.), and HL061210 (M.F.S.) and the Cedars-Sinai Board of Governors Chair in Medical Genetics (J.I.R.).

The authors gratefully acknowledge the dedication and work performed by the staff at the clinical sites in San Antonio, Texas, and San Luis Valley, Colorado, the laboratory staff at the University of Southern California (Los Angeles, California), and the members of the Coordinating Center at Wake Forest University School of Medicine (Winston-Salem, North Carolina).

1.
Rich SS: Mapping genes in diabetes: genetic epidemiological perspective.
Diabetes
39
:
1315
–1319,
1990
2.
Lillioja S, Mott DM, Spraul M, Farraro R, Foley JE, Ravussin E, Knowler WC, Bennett PH, Bogardus C: Insulin resistance and insulin secretory dysfunction as precursors of non-insulin-dependent diabetes mellitus: prospective studies of Pima Indians.
N Engl J Med
329
:
1988
–1992,
1993
3.
Warram JH, Martin BC, Krolewski AS, Soeldner JS, Kahn CR: Slow glucose removal rate and hyperinsulinemia precede the development of type II diabetes in the offspring of diabetic parents.
Ann Int Med
113
:
909
–915,
1990
4.
Haffner SM, Miettinen H, Gaskill SP, Stern MP: Decreased insulin action and insulin secretion predict the development of impaired glucose tolerance.
Diabetologia
39
:
1201
–1207,
1996
5.
Wagenknecht LE, Bergman RN, Howard G, Hamman RF, Selby JV, Saad MF: Insulin resistance and low acute insulin response predict type 2 diabetes: the Insulin Resistance Atherosclerosis Study (IRAS) (Abstract).
Diabetes
48 (Suppl. 1)
:
A174
,
1999
6.
Weyer C, Bogardus C, Mott DM, Pratley RE: The natural history of insulin secretory dysfunction and insulin resistance in the pathogenesis of type 2 diabetes mellitus.
J Clin Invest
104
:
787
–794,
1999
7.
Saad MF, Anderson RL, Laws A, Watanabe R, Kades WW, Chen Y-DI, Sands RE, Pei D, Savage PJ, Bergman RN: A comparison between the minimal model and the glucose clamp in the assessment of insulin sensitivity across the spectrum of glucose tolerance.
Diabetes
43
:
1114
–1121,
1994
8.
Bergman RN, Zaccaro DJ, Watanabe RM, Haffner SM, Saad MF, Norris JM, Wagenknecht LE, Hokanson JE, Rotter JI, Rich SS: Minimal model-based insulin sensitivity has greater heritability and a different genetic basis than homeostasis model assessment or fasting insulin.
Diabetes
52
:
2168
–2174,
2003
9.
Martin BC, Warram JH, Rosner B, Rich SS, Soeldner JS, Krolewski AS: Familial clustering of insulin sensitivity.
Diabetes
41
:
850
–854,
1992
10.
Elbein SC, Hasstedt SJ, Wegner K, Kahn SE: Heritability of pancreatic β-cell function among nondiabetic members of Caucasian familial type 2 diabetic kindreds.
J Clin Endocrinol Metab
84
:
1398
–1403,
1999
11.
Watanabe RM, Valle T, Hauser ER, Ghosh S, Kimmo JE, Kohtamäki, Ehnholm C, Tuomilehto J, Collins FS, Bergman RN, Boehnke M, for the Finland-United States Investigation of NIDDM Genetics (FUSION) Study Investigators: Familiality of quantitative metabolic traits in Finnish families with non-insulin-dependent diabetes mellitus.
Hum Hered
49
:
159
–168,
1999
12.
Colilla S, Cox NJ, Ehrmann DA: Heritability of insulin secretion and insulin action in women with polycystic ovary syndrome and their first-degree relatives.
J Clin Endocrinol Metab
86
:
2027
–2031,
2001
13.
Xiang AH, Azen SP, Tan S, Raffel LJ, Tan S, Cheng LS-C, Diaz J, Toscano E, Henderson PC, Hodis HN, Hsueh WA, Rotter JI, Buchanan TA: Evidence for joint genetic control of insulin sensitivity and systolic blood pressure in Hispanic families with a hypertensive proband.
Circulation
103
:
78
–83,
2001
14.
Hong Y, Weisnagel SJ, Rice T, Sun G, Mandel SA, Gu C, Rankinen T, Gagnon J, Leon AS, Skinner JS, Wilmore JH, Bergman RN, Bouchard C, Rao DC: Familial resemblance for glucose and insulin metabolism indices derived from an intravenous glucose tolerance test in blacks and whites of the HERITAGE Family Study.
Clin Genet
60
:
22
–30,
2001
15.
Sakul H, Pratley R, Cardon L, Ravussin E, Mott D, Bogardus C: Familiality of physical and metabolic characteristics that predict the development of non-insulin-dependent diabetes mellitus in Pima Indians.
Am J Hum Genet
60
:
651
–656,
1997
16.
Hanson RL, Imperatore G, Naryan KMV, Roumain J, Fogt-Campagna, Pettitt DJ, Bennett PH, Knowler WC: Family and genetic studies of indices of insulin sensitivity and insulin secretion in Pima Indians.
Diabetes Metab Res Rev
17
:
296
–303,
2001
17.
Wagenknecht LE, Mayer EJ, Rewers M, Haffner S, Selby J, Borok GM, Henkin L, Howard G, Savage PJ, Saad MF, Bergman RN, Hamman R: The Insulin Resistance Atherosclerosis Study (IRAS): objectives, design, and recruitment results.
Ann Epidemiol
5
:
464
–472,
1995
18.
Henkin L, Bergman RN, Bowden DW, Ellsworth DL, Haffner SM, Langefeld CD, Mitchell BD, Norris JM, Rewers M, Saad MF, Stamm E, Wagenknecht LE, Rich SS: Genetic epidemiology of insulin resistance and visceral adiposity: the IRAS Family Study design and methods.
Ann Epidemiol
13
:
211
–217,
2003
19.
Bergman RN, Finegood DT, Ader M: Assessment of insulin sensitivity in vivo.
Endocrine Rev
6
:
45
–86,
1985
20.
Pacini G, Bergman RN: MINMOD: a computer program to calculate insulin sensitivity and pancreatic responsivity from the frequently sampled intravenous glucose tolerance test.
Comp Meth Prog Biomed
23
:
113
–122,
1986
21.
Welch S, Gebhart SS, Bergman RN, Phillips LS: Minimal model analysis of intravenous glucose tolerance test-derived insulin sensitivity in diabetic subjects.
J Clin Endocrinol Metab
71
:
1508
–1518,
1990
22.
Steil GM, Volund A, Kahn SE, Bergman RN: Reduced sample number for calculation of insulin sensitivity and glucose effectiveness from the minimal model: suitability for use in population studies.
Diabetes
42
:
250
–256,
1993
23.
Bergman RN: Toward physiological understanding of glucose tolerance: minimal-model approach (Lilly Lecture).
Diabetes
38
:
1512
–1527,
1989
24.
McPeek MS, Sun L: Statistical tests for detection of misspecified relationships by use of genome screen data.
Am J Hum Genet
66
:
1076
–1094,
2000
25.
Almasy L, Blangero J: Multipoint quantitative-trait linkage analysis in general pedigrees.
Am J Hum Genet
62
:
1198
–1211,
1998
26.
Goldgar DE: Multipoint analysis of human quantitative genetic variation.
Am J Hum Genet
47
:
957
–967,
1990
27.
Amos C: Robust variance-components approach for assessing genetic linkage in pedigrees.
Am J Hum Genet
54
:
535
–543,
1994
28.
Self SG, Liang K-Y: Asymptotic properties of maximum likelihood estimators and likelihood ratio tests under non-standard conditions.
J Am Stat Assoc
82
:
605
–610,
1987
29.
Rao DC, Province MA: The future of path analysis, segregation analysis, and combined models for genetic dissection of complex traits.
Hum Hered
50
:
34
–42,
2000
30.
Blangero J, Williams JT, Almasy L: Robust LOD scores for variance component-based linkage analysis.
Genet Epidemiol
19 (Suppl. 1)
:
S8
–S14,
2000
31.
Watanabe RM, Ghosh S, Langefeld CD, Valle TT, Hauser ER, Magnuson VL, Mohlke KL, Silander K, Ally DS, Chines P, Blaschak-Harvan J, Douglas JA, Duren WL, Epstein MP, Fingerlin TE, Kaleta HS, Lange EM, Li C, McEachin RC, Stringham HM, Trager E, White PP, Balow J, Birznieks G, Chang J, Eldridge W, Erdos MR, Karanjawala ZE, Knapp JI, Kudelko K, Martin C, Morales-Mena A, Musick A, Musick T, Pfahl C, Porter R, Rayman JB, Rha D, Segal L, Shapiro S, Sharaf R, Shurtleff B, So A, Tannenbaum J, Te C, Tovar J, Unni A, Welch C, Whiten R, Witt A, Kohtamaki K, Ehnholm C, Eriksson J, Toivanen L, Vidgren G, Nylund SJ, Tuomilehto-Wolf E, Ross EH, Demirchyan E, Hagopian WA, Buchanan TA, Tuomilehto J, Bergman RN, Collins FS, Boehnke M: The Finland-United States Investigation of Non-insulin-dependent Diabetes Mellitus Genetics (FUSION) Study II: an autosomal genome scan for diabetes-related quantitative trait loci.
Am J Hum Genet
67
:
1186
–1200,
2000
32.
Cheng LS-C, Davis RC, Raffel LJ, Xiang AH, Wang N, Quinones M, Wen P-Z, Toscano E, Diaz J, Pressman S, Henderson PC, Azen SP, Hsueh WA, Buchanan TA, Rotter JI: Coincident linkage of fasting plasma insulin and blood pressure to chromosome 7q in hypertensive Hispanic families.
Circulation
104
:
1255
–1260,
2001
33.
An P, Hong Y, Weisnagel SJ, Rice T, Rankinen T, Leon AS, Skinner JS, Wilmore JH, Chagnon YC, Bergman RN, Bouchard C, Rao DC: Genomic scan of glucose and insulin metabolism phenotypes: the HERITAGE Family Study.
Metabolism
52
:
246
–253,
2003
34.
Hsueh W-C, Mitchell BD, Aburomia R, Pollen T, Sakul H, Ehm MG, Michelsen BK, StJean PL, Knowler WC, Burns DK, Bell CJ, Shuldiner AR: Diabetes in the Old Order Amish.
Diabetes Care
23
:
595
–601,
2000
35.
Pratley RE, Thompson DB, Prochazka M, Baier L, Mott D, Ravussin E, Sakul H, Ehm MG, Burns DK, Foroud T, Garvey WT, Hanson RL, Knowler WC, Bennett PH, Bogardus C: An autosomal genomic scan for loci linked to prediabetic phenotypes in Pima Indians.
J Clin Invest
101
:
1757
–1764,
1998
36.
Hanson RL, Ehm MG, Pettitt DJ, Prochazka M, Thompson DB, Timberlake D, Foroud T, Kobes S, Baier L, Burns DK, Almasy L, Blangero J, Garvey WT, Bennett PH, Knowler WC: An autosomal genomic scan for loci linked to type II diabetes mellitus and body-mass index in Pima Indians.
Am J Hum Genet
63
:
1130
–1138,
1998
37.
Elbein SC, Hoffman MD, Teng K, Leppert MF, Hasstedt SJ: A genome-wide search for type 2 diabetes susceptibility genes in Utah Caucasians.
Diabetes
48
:
1175
–1182,
1999
38.
Vionnet N, Hani El-H, Dupont S, Gallina S, Francke S, Dotte S, De Matos F, Durand E, Lepretre F, Lecoeur C, Gallina P, Zekiri L, Dina C, Froguel P: Genomewide search for type 2 diabetes-susceptibility genes in French whites: evidence for a novel susceptibility locus for early-onset diabetes on chromosome 3q27-qter and independent replication of a type 2-diabetes locus on chromosome 1q21–q24.
Am J Hum Genet
67
:
1470
–1480,
2000
39.
Lee JH, Reed DR, Li WD, Xu W, Joo EJ, Kilker RL, Nanthakumar E, North M, Sakul H, Bell C, Price RA: Genome scan for human obesity and linkage to markers in 20q13.
Am J Hum Genet
64
:
196
–209,
1999
40.
Ehm MG, Karnoub MC, Sakul H, Gottschalk K, Holt DC, Weber JL, Vaske D, Briley D, Briley L, Kopf J, McMillen P, Nguyen Q, Reisman M, Lai EH, Joslyn G, Shepherd NS, Bell C, Wagner MJ, Burns DK, American Diabetes Association GENNID Study Group: Genetics of NIDDM: genomewide search for type 2 diabetes susceptibility genes in four American populations.
Am J Hum Genet
66
:
1871
–1881,
2000
41.
Ober C, Abney M, McPeek MS: The genetic dissection of complex traits in a founder population.
Am J Hum Genet
69
:
1068
–1079,
2001