Brief Genetics ReportLinkage of the Metabolic Syndrome to 1q23-q31 inHispanic FamiliesThe Insulin Resistance Atherosclerosis Study Family StudyCarl D. Langefeld,1 Lynne E. Wagenknecht,1 Jerome I. Rotter,2 Adrienne H. Williams,1

John E. Hokanson,3 Mohammad F. Saad,4 Donald W. Bowden,5 Stephen Haffner,6 Jill M. Norris,3

Stephen S. Rich,1 and Braxton D. Mitchell7

The metabolic syndrome is characterized by centralobesity, dyslipidemia, elevated blood pressure, and hy-perglycemia. The Insulin Resistance AtherosclerosisStudy (IRAS) Family Study recruited extended pedi-grees of Hispanic descent from San Antonio, TX (SA)and San Luis Valley, CO (SLV). Thirty-five of thesepedigrees (27 SA and 8 SLV) had at least 2 individualswith metabolic syndrome (216 affected individuals and563 affected relative pairs). The prevalence of meta-bolic syndrome and component criteria in subjects fromthese pedigrees were 35% metabolic syndrome, 43%increased waist circumference, 31% hypertriglyceride-mia, 69% low HDL cholesterol, 31% increased bloodpressure, and 25% either increased fasting glucose orpresence of diabetes. Nonparametric linkage analysisprovided evidence for linkage of metabolic syndrome to1q23-q31 (D1S518; logarithm of odds [LOD] 1.6) withsignificant site heterogeneity (SA LOD 2.6 and SLV LOD0.0), and removing all individuals with diabetes re-duced, but did not eliminate, the evidence for linkage tothis region (LOD 1.2). This heterogeneity may partiallybe explained by phenotypic differences. Members in theSA pedigrees were older, had greater central obesity,had higher prevalence of the metabolic syndrome, andwere from a more urban environment than members ofthe SLV pedigrees. These results contribute to thegrowing evidence that chromosome 1q harbors at leastone locus related to the metabolic precursors of diabe-tes. Diabetes 53:11701174, 2004

The metabolic syndrome is a clinical entity con-sisting of central obesity, dyslipidemia, elevatedblood pressure, and hyperglycemia, and its pres-ence is associated with an increased risk for type2 diabetes and cardiovascular disease (1). A major etio-logic factor underlying this syndrome is insulin resistance(2). The metabolic syndrome was defined in the ThirdReport of the National Cholesterol Education ProgramExpert Panel Detection, Evaluation, and Treatment inAdults (Adult Treatment Panel III) as the presence of threeor more of the following risk factors: waist circumference88 cm in women and 102 in men, triglycerides 150mg/dl, HDL cholesterol 50 mg/dl in women and 40 inmen, blood pressure 130/85 mmHg, and fasting glucose110 mg/dl or a diagnosis of diabetes (1). Recent esti-mates suggest that the adult prevalence of the metabolicsyndrome is 32% in Hispanic Americans, 22% in AfricanAmericans, and 24% in Caucasian Americans (3).

It is believed that there is substantial overlap in theenvironmental (e.g., sedentary lifestyle and excessive ca-loric intake) and genetic risk factors associated with thedevelopment of metabolic syndrome and those factorsresponsible for the development of type 2 diabetes (1).While efforts to localize genes for the metabolic syndromeper se are in their infancy, genome-wide scans for type 2diabetes have been conducted in a number of populations.A growing number of studies (410) have reported linkageof diabetes or hyperglycemia to a region on human chro-mosome 1q21-q25. Further support for a chromosome 1qdiabetes susceptibility locus derives from the mapping ofthe gk2 diabetes susceptibility locus in the Goto-Kakizakirat to a syntenic region (11).

In this report, the Insulin Resistance AtherosclerosisStudy (IRAS) Family Study provides evidence that themetabolic syndrome links to chromosome 1q in 35 His-panic pedigrees (216 individuals with the metabolic syn-drome) ascertained from San Antonio, TX (SA), and SanLuis Valley, CO (SLV) (12). The IRAS Family Study is amulticenter study designed to identify genes predisposingto insulin resistance, adiposity, and related traits (12).Sample summary statistics, including the numbers of

From the 1Department of Public Health Sciences, Wake Forest UniversityHealth Sciences, Winston-Salem, North Carolina; the 2Division of MedicalGenetics, Cedars-Sinai Medical Center, Los Angeles, California; the 3Depart-ment of Preventive Medicine, University of Colorado Health Sciences, Denver,Colorado; the 4Division of Endocrinology, University of California, LosAngeles, California; the 5Department of Biochemistry, Wake Forest UniversityHealth Sciences, Winston-Salem, North Carolina; the 6Division of ClinicalEpidemiology, Department of Medicine, University of Texas Health ScienceCenter, San Antonio, Texas; and the 7Division of Endocrinology, Diabetes, andNutrition, University of Maryland School of Medicine, Baltimore, Maryland.

Address correspondence and reprint requests to Carl D. Langefeld, PhD,Department of Public Health Sciences, Wake Forest University School ofMedicine, Medical Center Boulevard, Winston-Salem, NC 27157-1063. E-mail:clangefe@wfubmc.edu.

Received for publication 16 September 2003 and accepted in revised form14 January 2004.

ARP, affected relative pair; IRAS, Insulin Resistance Atherosclerosis Study;LOD, logarithm of odds; NPL, nonparametric linkage.

2004 by the American Diabetes Association.

1170 DIABETES, VOL. 53, APRIL 2004

affected relative pairs (ARPs), are shown in Table 1. Theoverall prevalence of metabolic syndrome in these familieswas 35%, and the proportion of subjects having risk factorvalues meeting Adult Treatment Panel III criteria for thedefinition of metabolic syndrome was 43% for waist cir-cumference, 31% for triglycerides, 69% for HDL choles-

terol, 31% for blood pressure, and 25% for fasting glucoseor diabetes.

Nonparametric linkage (NPL) analysis provided strongevidence for linkage of the metabolic syndrome to chro-mosome 1q between markers D1S1589 and D1S518 (Table2) (Fig. 1). In the SA pedigrees, the evidence for linkage

TABLE 1Characteristics of the 216 individuals meeting the criteria of at least three elements of the metabolic syndrome

Characteristic SA SLV Combined

Pedigrees (n) 27 8 35Metabolic syndrome (n) 159 57 216ARPs (n)

Full sibling 97 41 138Avuncular 149 83 232Cousin 94 58 152Other 40 1 41Total 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 57.4BMI

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 36.9Metabolic syndrome (% in family) 38.6 27.9 35.1Risk factors for the metabolic syndrome

Waist circumferenceWomen 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.

TABLE 2Summary of the NPL analyses on chromosome 1

Position (cM) LOD LOD-1 interval P value Flanking markers

Metabolic syndromeCombined Hispanic 140.5 1.32 114.5222.5 0.0138 ATA42G12/

205.0 1.58 131.0220.0 0.0070 D1S518/SA 139.5 1.16 58.5223.0 0.0207 ATA42G12/

199.0 2.59 180.5209.0 0.0006 D1S1589/D1S518SLV 286.5 0.56 207.0 0.1078 /TTTA049

Metabolic syndrome removing individuals withonly two risk factors plus diabetesCombined Hispanic 146.5 1.31 129.5220.5 0.0139 /D1S534

204.0 1.49 131.5218.0 0.0089 D1S518/SA 150.0 1.03 52.5221.0 0.0294 /D1S534

200.0 2.23 179.5208.0 0.0013 D1S518/SLV 282.0 0.67 204.5 0.0789 /TTTA049

Metabolic syndrome removing all diabetic subjectsCombined Hispanic 205.0 1.23 131.5224.0 0.0172 D1S518/SA 202.0 1.88 193.5212.5 0.0033 D1S518/SLV 139.0 0.43 131.5151.0 0.1599 /D1S1627

DiabetesCombined Hispanic 211.0 1.76 190.5220.5 0.0045 /D1S1660SA 210.5 0.86 167.5232.0 0.0469 /D1S1660SLV 216.5 1.50 192.0221.5 0.0085 D1S1660/

C.D. LANGEFELD AND ASSOCIATES

DIABETES, VOL. 53, APRIL 2004 1171

FIG

.1.R

esu

lto

fN

PL

anal

ysis

on

chro

mo

som

e1

for

fou

rp

hen

oty

pes

:met

abo

lic

syn

dro

me

(A),

met

abo

lic

syn

dro

me

rem

ovi

ng

ind

ivid

ual

sw

ho

had

on

lytw

oo

fth

efi

vecr

iter

iap

lus

ad

iagn

osi

so

fd

iab

etes

(B),

met

abo

lic

syn

dro

me

rem

ovi

ng

all

ind

ivid

ual

sw

ith

ad

iagn

osi

so

fd

iab

etes

rega

rdle

sso

fth

en

um

ber

of

oth

ercr

iter

iam

et(C

),an

dty

pe

2d

iab

etes

(D).

METABOLIC SYNDROME LINKS TO 1q23-q31

1172 DIABETES, VOL. 53, APRIL 2004

maximized (logarithm of odds [LOD] 2.59) at 1q31 nearD1S1589/D1S518 (199 cM). The linkage support intervaldefined by the region in which the LOD score is 1.59(LOD-1 interval) was 28.5 cM in the SA pedigrees andwas bounded by the nearest markers, TATC028 (1q23) andD1S1660 (1q31). There was no evidence of linkage to thisregion within the SLV pedigrees (LOD 0.0). Combining thetwo Hispanic pedigree collections yielded LOD 1.58 nearD1S1589/D1S518 (Table 2).

To determine whether the evidence for linkage to 1qmight reflect linkage to diabetes, three additional analyseswere performed (Table 2). First, removing individualshaving diabetes and only two metabolic syndrome riskfactors from the analysis yielded a reduced sample con-taining 415 ARPs from the same 35 pedigrees. In thecombined collection of Hispanic pedigrees, there wasmodest 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 sincethe SLV pedigrees provided very modest evidence forlinkage near the D1S534 locus (LOD 0.17) and no evidencefor linkage near the D1S518 locus (LOD 0.0).

In the second ancillary analysis, all individuals with adiagnosis of diabetes were removed, providing a total of231 ARPs from 31 pedigrees for analysis (Table 2). Thecombined Hispanic sample provided modest evidence forlinkage to a single region near D1S518 (1q31, 205 cM, LOD1.23). The SLV pedigrees did not contribute evidence forlinkage at this locus (LOD 0.0).

In the third ancillary analysis, only pedigrees with atleast two individuals diagnosed with type 2 diabetes wereincluded (30 pedigrees containing 139 ARPs) and a linkageanalysis for type 2 diabetes was performed. In both the SAand SLV pedigrees, the strongest evidence for linkage totype 2 diabetes was near D1S1660 (SA LOD 0.86, 211 cMand SLV LOD 1.50, 217 cM), providing a maximum LODscore of 1.76 near D1S1660 (211 cM, 1q31). The LOD-1support interval spanned from 191 (D1S1589, 1q25) to 221cM (D1S1663, 1q32).

As an exploratory analysis, a series of ordered subsetanalyses were computed to investigate whether any of thefive specific risk factors influenced the magnitude of theevidence for linkage within the 1q region. Subsetting on thelevels of the five individual risk factors defining the metabolicsyndrome did not yield a statistically significant increase inthe evidence for linkage within any of the three combinationsof pedigrees (i.e., SA, SLV, or combined Hispanic).

There are several possible factors that could contributeto the stronger evidence for linkage detected in the 27 SAHispanic pedigrees collected relative to the 8 SLV Hispanicpedigrees. One might posit that the primary difference inthe linkage results reflects a difference in the statisticalpower of the respective samples. However, there are alsosubstantive differences in the characteristics of these twosamples (Table 1). First, the median age of individualsfrom the SA pedigrees was 6 years older than that from theSLV pedigrees (P 0.0153). Second, men from SanAntonio tended to have a greater BMI (P 0.0323) andwaist circumference (P 0.0005) than men from San LuisValley; a similar but not statistically significant differencewas observed among women. Third, diastolic blood pres-sure tended to be higher in the SLV pedigrees (P 0.0007).

Fourth, San Antonio is a more urban environment relativeto San Luis Valley. In summary, the SA pedigrees tended tobe older, have greater central adiposity, have a higherprevalence of the metabolic syndrome, and be from amore urban environment.

These results contribute to the growing evidence thatchromosome 1q harbors a diabetes-related susceptibilitylocus. Evidence for linkage in this population does notappear to be restricted to subjects with diabetes alone,thus suggesting that the putative disease-predisposing locimay influence susceptibility to a broader spectrum ofmetabolic disorders than just diabetes. Because of therelatively high prevalence of the metabolic syndrome inthe population, even a modest increase in risk associatedwith the putative mutation will be associated with arelatively large burden of disease in the population.

Linkage of diabetes to the 1q21-25 regions has previ-ously been reported (410) in European-Caucasian, Amer-indian, and Chinese populations. The current resultssuggest that the putative mutation(s) may also be presentin Hispanic Americans. Although the position that maxi-mizes the LOD score for metabolic syndrome in oursample is the most telomeric of these seven studies, it iswithin the LOD-1 support interval of the majority of thesestudies. Conversely, our LOD-1 support interval containsthe loci that maximize the evidence for linkage in themajority of these studies. The apparent presence of link-age across such diverse populations is striking and isconsistent with an ancient source of the mutation.

RESEARCH DESIGN AND METHODSStudy design and data collection. The IRAS Family Study was designed tomap genes that predispose to insulin resistance, cardiovascular disease, andobesity. Details of the study design, patient recruitment, and phenotyping aredescribed in detail elsewhere (12). Briefly, probands of Hispanic-Americanheritage who self-reported a large family structure were recruited from SanAntonio, TX, and San Luis Valley, CO. Probands were identified from theoriginal IRAS cohort study (13), which was designed to recruit approximatelyequal numbers of diabetic, impaired glucose tolerant, and normal individuals.To meet recruitment goals in the IRAS Family Study, additional probandswere recruited from the general population at the San Antonio clinic. Fromthis nonrandom collection of probands, ascertainment of families was basedon large family structure and not on any clinical characteristics related toelements of the metabolic syndrome, diabetes, cardiovascular disease, orobesity. The target family structure criteria included DNA collectible on atleast 12 and full phenotyping possible on at least 9 family members. We did notspecifically screen for maturity-onset diabetes of the youngassociated muta-tions or GAD antibodies for type 1 diabetes. However, our sample containedonly 22 insulin-taking diabetic subjects; of these, all had an age of diabetesonset 30 years and all but 1 had a BMI 27 kg/m2.

Medical history, health behaviors, and demographic data were collected byinterview. The medical history interview focused on the assessment of currenthealth status and clinical conditions, particularly self-reported type 2 diabetesand its complications, hypertension, and cardiovascular disease events andprocedures. At the time of the clinic visit, a fasting blood draw was obtained todetermine plasma glucose, insulin, and lipid levels. Diabetes was defined usingthe American Diabetes Association criteria. The systolic and diastolic bloodpressures analyzed represent the average of three measures, respectively. BMI (inkilograms per square meter) was obtained from height and weight measurements.The five individual risk factors defining the metabolic syndrome were assessed,and the individuals metabolic syndrome status was determined indepen-dent of medications used. Linkage results were qualitatively similar when werepeated the analysis using a slightly expanded definition that included antihy-pertensive or antilipid medications as affected for that trait.

Whole blood obtained from each IRAS Family Study participant was frozenand shipped to the Molecular Genetics Laboratory at Wake Forest UniversitySchool of Medicine. DNA from these 35 families, who completed a 10-cMgenome scan (marker set 11), was shipped to the Mammalian GenotypingCenter in Marshfield, WI (http://research.marshfieldclinic.org/genetics). Chro-

C.D. LANGEFELD AND ASSOCIATES

DIABETES, VOL. 53, APRIL 2004 1173

mosomal maps were constructed using the marker order and distancesavailable from the Mammalian Genotyping Center.Statistical methodsPreliminary analyses. Center-specific maximum likelihood estimates ofallele frequencies were computed using the Recode software (D. Weeks,personal communication). Each pedigree was examined for potentially incor-rectly self-reported familial relationships using the entire genome scan data(383 markers) and the Prest software (14). Inconsistent relationships weremodified when the data suggested a clear alternative (e.g., full sibling to halfsibling) or the genotype data were converted to missing. Each marker wasexamined for Mendelian inconsistencies using the Pedcheck software (15),and probable genotyping errors were converted to missing.Linkage analysis. Multipoint NPL regression analysis using the NPLpairsstatistic was computed for the markers on chromosome 1. The NPL regressionapproach 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 ofmisclassification errors (e.g., unaffected subjects who develop metabolicsyndrome in the future).

A series of ordered subset analyses (18) were computed to investigate theinfluence of each of the metabolic syndrome risk factors on the evidence forlinkage. First, the mean of each risk factor was calculated for each pedigreeand ranked from smallest to largest. For a specific risk factor (e.g., triglycer-ides), the family with the largest risk factor mean was entered into the analysisand the corresponding LOD score was computed on chromosome 1 for thatfamily. The ith ordered subset analysis proceeds by computing a linkageanalysis on chromosome 1 using the subset of families with the ith largest riskfactor means. This process is repeated until all families have been added to thelinkage analysis. The subset of families that yield the largest LOD score onchromosome 1 is taken as the LOD score of interest. The statistical signifi-cance of the change in the LOD score was computed by a permutation testunder the null hypothesis that the familys ranking for a specific risk factor isindependent of their evidence for linkage, resulting in a chromosome-wide Pvalue. For each risk factor, the ordered subset analysis was repeated, rankingfamilies from smallest to largest.

REFERENCES1. National Institutes of Health: Third Report of the National Cholesterol

Education Program Expert Panel on Detection, Evaluation, and Treat-ment of High Blood Cholesterol in Adults (Adult Treatment Panel III).Bethesda, MD, National Institutes of Health, 2001 (NIH publ. no. 01-3670)

2. Grundy SM: Hypertriglyceridemia, insulin resistance, and the metabolicsyndrome. Am J Cardiol 83:25F29F, 1999

3. Ford ES, Giles WH, Dietz WH: Prevalence of the metabolic syndromeamong US adults: findings from the Third National Health and NutritionExamination Survey. JAMA 287:356359, 2002

4. Hanson RL, Ehm MG, Pettitt DJ, Prochazka M, Thompson DB, TimberlakeD, Foroud T, Kobes S, Baier L, Burns DK, Almasy L, Blangero J, GarveyWT, Bennett PH, Knowler WC: An autosomal genomic scan for loci linkedto type II diabetes mellitus and body-mass index in Pima Indians. Am JHum Genet 63:11301138, 1998

5. Elbein SC, Hoffman MD, Teng K, Leppert MF, Hasstedt SJ: A genome-wide

search for type 2 diabetes susceptibility genes in Utah Caucasians.Diabetes 48:11751182, 1999

6. Vionnet N, Hani E, 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 Frenchwhites: evidence for a novel susceptibility locus for early-onset diabetes onchromosome 3q27-qter and independent replication of a type 2-diabeteslocus on chromosome 1q21q24. Am J Hum Genet 67:14701480, 2000

7. Wiltshire S, Hattersley AT, Hitman GA, Walker M, Levy JC, Sampson M,ORahilly S, Frayling TM, Bell JI, Lathrop GM, Bennett A, Dhillon R,Fletcher C, Groves CJ, Jones E, Prestwich P, Simecek N, Rao PV, WishartM, Foxon R, Howell S, Smedley D, Cardon LR, Menzel S, McCarthy MI: Agenomewide scan for loci predisposing to type 2 diabetes in a U.K.population (the Diabetes U.K. Warren 2 Repository): analysis of 573pedigrees provides independent replication of a susceptibility locus onchromosome 1q. Am J Hum Genet 69:553569, 2001

8. Hsueh W-C, St Jean PL, Mitchell BD, Pollin TI, Knowler WC, Ehm MG, Bell CJ,Sakul H, Wagner MJ, Burns DK, Shuldiner AR: Genome-wide and fine-mapping linkage studies of type 2 diabetes and glucose traits in the Old OrderAmish: evidence for a new diabetes locus on chromosome 14q11 andconfirmation of a locus on chromosome 1q21q24. Diabetes 52:550557, 2003

9. Meigs JB, Panhuysen CI, Myers RH, Wilson PW, Cupples LA: A genome-wide scan for loci linked to plasma levels of glucose and HbA1c in acommunity-based sample of Caucasian pedigrees: the Framingham Off-spring Study. Diabetes 51:833840, 2002

10. Xiang K, Wang Y, Zheng T, Shen K, Jia W, Li J, Lin X, Wu S, Zhang G, WangS, Lu H: Genome-wide scan search for type 2 diabetes susceptibility loci inChinese (Abstract). Diabetes 51 (Suppl. 2):A262, 2002

11. Gauguier D, Froguel P, Parent V, Bernard C, Bihoreau MT, Portha B, JamesMR, Penicaud L, Lathrop M, Ktorza A: Chromosomal mapping of geneticloci associated with non-insulin dependent diabetes in the GK rat. NatGenet 12:3843, 1996

12. Henkin L, Bergman RN, Bowden DW, Ellsworth DL, Haffner SM, MitchellBD, Norris JM.,: Rewers M, Saad MF, Wagenknecht LE, Rich SS: Geneticepidemiology of insulin resistance and visceral adiposity: the Family IRASStudy design and methods. Ann Epidemiol 13:211217, 2002

13. Wagenknecht LE, Mayer EJ, Rewers M, Haffner S, Selby J, Borok GM,Henkin L, Howard G, Savage PJ, Saad MF: The Insulin ResistanceAtherosclerosis Study (IRAS) objectives, design, and recruitment results.Ann Epidemiol 5:464472, 1995

14. McPeek MS, Sun L: Statistical tests for detection of misspecified relation-ships by use of genome scan data. Am J Hum Genet 66:10701094, 2000

15. OConnell JR, Weeks DE: PedCheck: a program for identification ofgenotype incompatibilities in linkage analysis. Am J Hum Genet 63:259266, 1998

16. Langefeld CD, Davis CC, Brown WM: Nonparametric linkage regression. I:combined European-American CSGA and German genome scans forasthma. Genet Epid 21 (Suppl. 1):S136S141, 2001

17. Kong A, Cox NJ: Allele-sharing models: LOD scores and accurate linkagetests. Am J Hum Genet 61:11791188, 1997

18. Hauser ER, Watanabe RM, Duren WL, Langefeld CD, Boehnke M: Orderedsubset analysis in genetic linkage mapping of complex traits. Genet Epid.In press

METABOLIC SYNDROME LINKS TO 1q23-q31

1174 DIABETES, VOL. 53, APRIL 2004