Directory UMM :Data Elmu:jurnal:A:Atherosclerosis:Vol151.Issue2.Aug2000:

The Pro12Ala substitution in the peroxisome proliferator activated

receptor gamma 2 is associated with an insulin-sensitive phenotype

in families with familial combined hyperlipidemia and in

nondiabetic elderly subjects with dyslipidemia

Jussi Pihlajama¨ki

a

_{, Raija Miettinen}

a

_{, Raisa Valve}

a

_{, Leena Karjalainen}

a

_,

Leena Mykka¨nen

a

_{, Johanna Kuusisto}

a

_{, Samir Deeb}

b

_{, Johan Auwerx}

c

_,

Markku Laakso

a,

_*

a_{Department of Medicine,Uni6ersity of Kuopio,70210 Kuopio, Finland} b_{Departments of Medicine and Genetics,Uni6ersity of Washington,Seattle,WA, USA}

c_{U.325 INSERM,De´partement d’Athe´roscle´rose,Institut Pasteur,59019 Lille, France}

Received 27 May 1999; received in revised form 23 September 1999; accepted 13 October 1999

Abstract

Dyslipidemias and insulin resistance often present simultaneously, as in familial combined hyperlipidemia (FCHL), and therefore may have a common genetic background. In our previous study the Pro12Ala substitution of peroxisome proliferator receptor g 2 (PPARg2) associated with insulin sensitivity, low body mass index (BMI) and high-density lipoprotein (HDL) cholesterol levels. In this study, we investigated the role of this substitution in dyslipidemias. Therefore, 228 nondiabetic members of FCHL families and 866 nondiabetic elderly subjects with (n=217) and without dyslipidemia (n=649) were genotyped. The allele frequencies of the Pro12Ala substitution did not differ between elderly subjects with or without dyslipidemia or 27 probands with FCHL. However, this substitution was associated with low fasting insulin levels both in FCHL family members (P=0.036 adjusted for gender and age) and elderly subjects with dyslipidemia (P=0.050) but not in elderly subjects without dyslipidemia (P=0.080). In addition, the Ala12 allele of PPARg2 was associated with low BMI (P=0.034) and low total triglycerides (P=0.027), and increased HDL-cholesterol (PB0.001) in elderly subjects with dyslipidemia (n=299) but not among any other study groups. We conclude that the Ala12 isoform of PPARg2 ameliorates the insulin resistance and unfavorable lipid and lipoprotein profiles in FCHL and hyperlipidemic elderly subjects. © 2000 Elsevier Science Ireland Ltd. All rights reserved.

Keywords:Familial combined hyperlipidemia; Genetics; Insulin resistance; PPARg

www.elsevier.com/locate/atherosclerosis

1. Introduction

Hypertriglyceridemia alone [1,2] or in combination with hypercholesterolemia [3] have been associated with insulin resistance. In familial combined hyperlipidemia (FCHL) both variable dyslipidemias [4,5] and insulin resistance [6 – 9] are often present. In FCHL patients a defect in adipose tissue metabolism has been proposed to lead, via increased free fatty acid (FFA) levels both to increased hepatic lipoprotein synthesis and insulin

resistance [9 – 11]. Therefore genes that regulate fat cell metabolism are likely to be important for the under-standing of insulin resistance in FCHL.

Peroxisome proliferator activated receptor g

(PPARg) is a transcription factor belonging to the

nuclear hormone receptor family that plays an impor-tant role in adipocyte differentiation and gene expres-sion [12,13], Therefore, the PPARg gene is a potential

candidate gene both for dyslipidemias and insulin resis-tance. The PPARggene is transcribed into three

differ-ent mRNAs [14,15], which give rise to two distinct proteins, i.e. PPARg1 and PPARg2 [16]. PPARg2 is

distinct from PPARg1 since it contains 28 additional * Corresponding author. Tel.: +358-17-172151; fax: +

358-17-173993.

E-mail address:markku.laakso@uku.fi (M. Laakso)

J.Pihlajama¨ki et al./Atherosclerosis151 (2000) 567 – 574

568

amino acids at its NH2-terminus, encoded by the B exon [14 – 16]. Known natural ligands and activators of PPARg, such as fatty acids and 15-deoxy-delta12,

14-prostaglandin J2, have a relatively low affinity for the receptor [17,18]. In contrast, synthetic ligands, such as the antidiabetic thiazolidinediones bind with high-affinity to PPARg [19,20], and their insulin-sensitizing

effect is attributed to the activation of PPARg. In

addition to ligands, growth factors and insulin also modify PPARgactivity in a ligand-independent fashion

through a MAP kinase-mediated phosphorylation of amino acids in the NH2-terminal activation domain [21 – 24]. Interestingly, PPARg2, which has an

addi-tional 28 amino acids at its NH2-terminus, is reported to be a much more effective transcriptional activator than PPARg1 [25].

Previously we described a common Pro12Ala substi-tution in the PPARg2-specific B exon of the PPARg

gene, which was associated with a lower body mass index (BMI), higher high-density cholesterol (HDL) cholesterol levels in elderly subjects and improved in-sulin sensitivity in middle-aged subjects. Accordingly, this Pro12Ala substitution decreased transcriptional ac-tivity of PPARg[26]. Ristow et al. recently described a

rare Pro115Gln mutation in the NH2-terminal ligand-independent activation domain. The correct position of this mutation is likely to be Pro113Gln [15]. This mutation inhibited phosphorylation of the protein at Ser112 and, therefore, resulted in a constitutive activa-tion of PPARg in vitro. Permanent activation of

PPARg was suggested to contribute to obesity in 4 of

121 obese subjects who carried this substitution [27]. Also, an association between a silent polymorphism in codon 453 with plasma leptin levels has been observed in obese subjects [28]. Therefore, it appears that the activity of PPARg is directly related to adiposity and

insulin resistance.

So far no studies have specifically addressed the role of PPARg in dyslipidemias, such as FCHL. Therefore,

we evaluated the role of both the Pro12Ala and Pro115Gln substitutions in dyslipidemias and insulin resistance in nondiabetic members (n=228) of FCHL families and in nondiabetic elderly subjects with (n= 217) or without (n=649) dyslipidemia.

2. Methods

2.1. Study subjects

All subjects participating in this study were Finnish. The Finnish population is genetically relatively ho-mogenous, originating mainly from southern (Eu-ropean) and eastern (Asian) immigration 2000 years ago [29,30].

Twenty-seven families with FCHL including 228 family members (27 probands, 129 of their first-degree relatives, 44 of their second-degree relatives and 28 spouses) formed the study population. Twenty-five of the probands of FCHL families were male survivors of myocardial infarction at young age (B55 years,n=18) or their first-degree relatives (n=7, all females), who were first studied in 1978 – 1980 [31] and restudied in 1992 – 1994. Two probands fulfilling similar lipid crite-ria were selected from the Coronary Angiography Reg-ister of the Kuopio University Hospital. The criteria for FCHL were: total cholesterol ]7.7 mmol/l and/or total triglycerides ]2.2 mmol/l in women and ]2.4 mmol/l in men at least in one of the visits. These lipid criteria were based on lipid levels of 250 control sub-jects from a random population sample who partici-pated in the same study as control families [31]. The cut-off points were 80th percentile for total cholesterol, and 90th percentile for total triglycerides. The 80th percentile for total cholesterol was used because of high cholesterol level among Finns living in eastern Finland. In order to meet the criteria for FCHL, all probands who were included had to have at least two affected first-degree relatives with different types of dyslipi-demias. None of the study subjects had tendon xan-thomas or defects of the LDL receptor which explain about 90% of all cases of familial hypercholesterolemia in this area [32].

Elderly subjects were a random sample of inhabitants in Kuopio aged 65 – 74 years during the baseline study in 1986 – 1988 [33]. From the original sample of 1910 inhabitants 1069 nondiabetic subjects participated in the baseline study and 896 in the follow-up study in 1990 – 1991 [34]. Only subjects who participated in the follow-up study could be used here because DNA sam-ples were taken during that visit. We divided elderly subjects into two subgroups according to the same lipid criteria used for FCHL. Table 1 shows the clinical characteristics of study subjects.

All study subjects had normal glucose tolerance ac-cording to the World Health Organization criteria [35], normal liver, kidney and thyroid function tests and none of them had a history of excessive alcohol intake.

2.2. Study design

On the study day all subjects underwent an oral glucose tolerance test (75 g of glucose) and blood was drawn for laboratory analyses after a 12 h fast.

2.3. Analytical methods

Plasma glucose levels in the fasting state and after an oral glucose load were measured by the glucose oxidase method (2300 Stat Plus, Yellow Springs In-strument Co., Yellow Springs, OH). For the determi-nation of plasma insulin, blood was collected in EDTA-containing tubes and after centrifugation the plasma was stored at −70°C until the analysis. Plasma insulin concentration was determined by a commercial double-antibody solid-phase radioim-munoassay (Phadeseph Insulin RIA 100; Pharmacia Diagnostics AB, Uppsala, Sweden). Lipoprotein frac-tionation was performed by the use of ultracentrifuga-tion and selective precipitaultracentrifuga-tion [36] as previously described [1]. Cholesterol and triglyceride levels from whole serum and from lipoprotein fractions were as-sayed by automated enzymatic methods (Boehringer-Mannheim, (Boehringer-Mannheim, Germany). Apolipoprotein B (apoB) was determined by a commercial immunotur-bidometric method (Kone Instruments, Espoo, Fin-land).

2.4. Determination of the Pro12Ala and Pro115Gln substitutions of the PPARg2 gene

DNA was prepared from peripheral blood leuko-cytes by proteinaseK-phenol-choloroform extraction method. For amplification of exon B of the PPARg2

gene we used primers forward 5% -GACAAA-ATATCAGTGTGAATTACAGC-3% and reverse 5% -CCCAATAGCCGTATCTGGAAGG-3% [26], and for amplification of the region spanning codon 113 we used primers forward 5% -TGCAATCAAAGTGGAG-CC-3% _{and reverse 5}%

-CAGAAGCTTTATCTCCACA-GAC-3% [27]. PCR was done in 6ml volume containing

50 ng of genomic DNA, 3 pmol of each primer, 10 mmol/l Tris-HCL (pH 8.8), 50 mmol/l KCL, 1.5 mmol/l of MgCl2, 0.1% Triton X-100, 100 mmol/l

dNTP, 0.25 units of DNA polymerase (Dynazyme DNA polymerase, Finnzymes, Espoo, Finland) and 0.55 mCi of [32P]dCTP. PCR conditions were

denatura-tion at 94°C for 3 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 66°C for 30 s and extension at 72°C for 30 s with final exten-sion at 72°C for 4 min. Variants were detected with the single strand conformation polymorphism (SSCP) analysis. PCR products were first diluted 4 – 10-fold with 0.1% SDS and 10 mmol/l EDTA and then di-luted (1:1) with loading mix (95% formamide, 20 mmol/l EDTA, 0.05% bromphenol blue, 0.05% xylene cyanole). After denaturation at 98°C for 3 min, sam-ples were immediately cooled on ice, and 2 ml of each

sample was loaded onto 6% nondenaturating polyacry-lamide gel (acrypolyacry-lamide/N,N-methylene-bis-acrylamide ratio 49:1) containing 10% glycerol. Samples were run at two different gel temperatures (29°C and 37 – 39°C). The gel was autoradiographed overnight at −70°C with intensifying screens. Amplified segments from subjects with a variant pattern in SSCP analysis were purified by electrophoresis on a 1% low-melting point agarose gel and directly sequenced with the use of Sequenase (Amersham Life Sciences, Cleveland, OH) as previously described [37].

2.5. Statistical analysis

All basic statistical analysis were performed with the SPSS/WIN programs (version 7.5, SPSS, Chicago, IL).

Table 1

Clinical characteristics of the study groupsa

Elderly nondiabetic subjects. FCHL family members (n=228) Without dyslipidemia (n=649) With dyslipidemia (n=217)

264/385

Gender (M/F) 63/154b _114/114c

73.092.8 52.5915.6 72.992.9

Age (yrs)

27.794.1c

26.794.2

Body mass index (kg/m2_{) 26.594.5}

0.9490.08 0.9190.10

Waist-to-hip ratio 0.9490.08

153922

Systolic blood pressure (mmHg) 160925 137920

Diastolic blood pressure (mmHg) 81910 84910 84910

Fasting glucose (mmol/l) 5.790.7 6.091.1c _6.092.0

69.0939.0

Fasting insulin (pmol/l) 95.4961.2d _78.6970.8b

Total cholesterol (mmol/l) _6.1490.88 7.5291.29d _6.8291.38d

Total triglycerides (mmol/l) 1.3390.40 2.5091.42d _2.1191.44d

1.3090.42d _1.3390.30

HDL-cholesterol (mmol/l) 1.3990.33

1.2190.30d

1.0390.20 1.4290.22d

Apolipoprotein B (g/l)

a_{All values for continuous variables are mean9SD. HDL indicates high-density lipoprotein.} b_PB0.05,

c_PB0.01,

J.Pihlajama¨ki et al./Atherosclerosis151 (2000) 567 – 574

570

Fig. 1. Body mass index, fasting insulin, total triglyceride and HDL-cholesterol levels according to the Pro12Ala substitution of the PPARg2 gene

in study groups. Subjects with the Pro12Pro genotype are shown in white bars, subjects with the Pro12Ala genotype in shaded bars and subjects with the Ala12Ala genotype in black bars. Number of cases summarized in each bar is shown in Table 2.

The frequencies between the study groups were com-pared with the chi-square test. The effect of the variants on continuous variables was tested by ANCOVA in unrelated elderly subjects and by family-based associa-tion analysis with the program ASSOC in FCHL families [38]. The ASSOC program uses linear regres-sion analysis allowing the quantitative trait to have familial correlation among individuals. The likelihood for the pedigree is computed with a linear regression model in which the quantitative trait is the dependent variable and the genetic variation, discrete and/or con-tinuous covariates are independent variables. Residual variation is modeled assuming an additive polygenic pattern of correlation among relatives. Using this model the likelihood for each pedigree was maximized twice, with and without the genetic variant in the model. The difference in natural logarithms of these two maximized likelihoods follows chi-square distribu-tion from which the corresponding P-value is taken with two degree of freedom (three groups of genotypes) or with one degree of freedom (subjects with or without a definite genotype). Gender was included as a discrete covariate and age as a continuous covariate in all analyses. Insulin and triglyceride values were logarith-mically transformed before analysis in order to achieve normal distribution. All data are presented as mean9

SD. P-valueB0.05 was considered statistically significant.

3. Results

Nondiabetic elderly subjects with dyslipidemia had higher BMI (P=0.002, adjusted for age and gender) and higher levels of glucose (P=0.005), insulin (PB

0.001), total cholesterol (PB0.001), total triglycerides (PB0.001) and apoB (PB0.001) and lower levels of HDL-cholesterol (PB0.001) than subjects without dys-lipidemia. In addition to being more dyslipidemic (higher total cholesterol, total triglycerides and apoB, PB0.001), FCHL family members were younger (PB

0.001) and had higher gender- and age-adjusted insulin levels (P=0.026) compared to elderly subjects without dyslipidemia (Table 1).

Allele frequencies of the Pro12Ala polymorphism of the PPARg gene did not differ between elderly subjects

with (0.17) or without (0.14) dyslipidemia or probands with FCHL (0.15, P=ns). Genotype frequencies fol-lowed the Hardy – Weinberg equilibrium in all study groups. The Pro113Gln substitution was not found in any of the FCHL family members.

J

.

Pihlajama

¨k

ie

ta

l

.

/

Atherosclerosis

151

(2000)

567

–

574

571

Body mass index, waist to hip ratio and fasting glucose, insulin and lipid and lipoprotein levels according to the Pro12Ala substitution of the PPARg2gene in the study groups

FCHL family members Elderly subjects with dyslipidemia

Elderly subjects without dyslipidemia

Pro/Ala Ala/Ala Pro/Pro Pro/Ala

Pro/Pro Ala/Ala

Ala/Ala Pro/Ala

Pro/Pro

(n=157) (n=58)

(n=457) (n=176) (n=16) (n=57) (n=3) (n=164) (n=6)

48/109 15/42 0/3 83/81 27/31 4/2

78/98

Gender (M/F) 181/276 5/11

75.794.0 51.5915.4 52.3915.1 53.7916.0 73.292.8

72.692.8 Age (years) 73.092.9 72.892.8 73.193.3

28.294.0 23.392.8b _{26.194.2 27.495.0 24.491.7}

Body mass index 26.793.9 27.394.4 26.194.5 27.094.4 (kg/m2₎

0.9290.08 0.9190.10 0.9290.09 0.9190.05 0.9390.09 0.9590.08

0.9390.08 0.9490.08

0.9490.08 Waist-to-hip ratio

159927 151922 135919 139921

Systolic BP 153922 153922 160918 155923 134914

(mmHg)

8399 77916 84911 82912

83911 8298

81910 82910

Diastolic BP 8299

(mmHg)

5.791.1 5.690.8 5.990.9 5.890.8 5.991.0 5.691.1 5.690.7 5.690.5

Fasting glucose 5.590.5

(mmol/l)

68.4933.6 _65.4942.7 85.2954.6 102.0965.4 54.6910.8c _{72.6972.0 83.4959.4 46.8916.8}d

Fasting insulin 55.2922.2

(pmol/l)

7.0291.15 8.4890.75 6.8191.34 6.7091.19

6.3090.83 7.1091.31 6.7791.18

6.0990.88 6.2090.99

Total cholesterol (mmol/l)

1.1890.32 2.0991.34 2.2691.28 1.0790.35e _{1.9791.33 2.0591.32 1.8590.96}

1.3790.50

Total triglycerides 1.3390.45 (mmol/l)

1.2390.33 2.2490.41f _{1.3690.30 1.2990.29}

1.5090.30 1.3190.30

1.3890.30 1.3390.39

HDL-cholesterol 1.4090.35 (mmol/l)

1.4290.25 1.3490.34

Apolipoprotein B 1.0490.18 1.0590.18 1.0590.15 1.3790.27 1.1790.26 1.2390.30 1.1390.26 (g/l)

a_{All values are mean9SD. HDL indicates high-density lipoprotein.} b_P=0.034,

c_P=0.050, d_P=0.036, e_P=0.027,

J.Pihlajama¨ki et al./Atherosclerosis151 (2000) 567 – 574

572

the three genotypes) (Fig. 1, Table 2). Subjects with the Ala12Ala genotype had lower levels of fasting insulin than subjects with the Pro12Pro and Pro12Ala geno-types (P=0.021). However, BMI, waist to hip ratio, blood pressure, fasting glucose and serum lipids and lipoproteins were not associated with this polymor-phism in FCHL family members.

In the pooled group of elderly nondiabetic subjects with and without dyslipidemias BMI (P=0.010), fast-ing insulin (P=0.034), HDL-cholesterol (P=0.002) and total triglycerides (P=0.007) were associated with the Pro12Ala substitution of the PPARg gene. No

associations between the Pro12Ala substitution and these parameters could be shown in the subgroup of elderly subjects without dyslipidemia (n=649, Table 2). However, in dyslipidemic elderly subjects (n=217) BMI (27.094.4 vs. 28.294.0 vs. 23.392.8 kg/m2_, P=0.034), fasting insulin levels (85.2954.6 vs. 102.0965.4 vs. 54.6910.8 pmol/l, P=0.050), total triglyceride levels (2.0991.34 vs. 2.2691.28 vs. 1.079

0.35 mmol/l, P=0.027) and HDL-cholesterol levels (1.3390.39 vs. 1.2390.33 vs. 2.2490.41 mmol/l,PB

0.001) differed among the three genotypes (Fig. 1, Table 2). No differences were seen between subjects with Pro12Pro and Pro12Ala genotypes. However, the subjects with the Ala12Ala genotype had higher HDL-cholesterol (PB0.001) levels than subjects with other genotypes. Still, BMI (P=0.097), fasting insulin (P= 0.200) or total triglycerides levels (P=0.088) did not differ significantly between dyslipidemic elderly subjects with or without the Ala12Ala genotype. We did not find any association between the Pro12Ala substitution and waist-to-hip ratio, blood pressure, fasting glucose, total cholesterol or apoB levels in elderly subjects (Table 2).

4. Discussion

Insulin resistance and dyslipidemias often occur simultaneously [2,39] as in FCHL [6 – 9], indicating that insulin resistance and dyslipidemias may have a com-mon genetic background. In our study the Pro12Ala substitution of the PPARg2 gene was associated with

low fasting insulin levels in FCHL families and in dyslipidemic elderly subjects. Furthermore, the Pro12Ala substitution was also associated with low BMI, low total triglycerides and high HDL cholesterol levels in dyslipidemic elderly subjects, whereas no statis-tically significant associations were observed in elderly subjects without dyslipidemia. This implies that the Pro12Ala substitution by itself does not have a major causative role in FCHL or dyslipidemias in elderly subjects, but it may modify fasting insulin and lipid levels in subjects with a genetic or environmental pre-disposition to dyslipidemias.

Although FCHL was originally described to have a dominant mode of inheritance [4], polygenic back-ground is more likely [40]. Major genes for FCHL have not been yet been identified but some rare mutations in the lipoprotein lipase gene may underlie this disease [41]. In addition, the ApoAI – CIII – AIV gene complex has been implicated for the etiology of FCHL [42,43], but also negative results have been published [44,45]. Furthermore, a promising, but as yet unidentified, locus in 1q21-23 has been reported [46]. FCHL disease may in fact identify subjects with familial aggregation of several gene defects causing phenotypes of the insulin resistance syndrome [11,40]. Therefore, the search for gene defects causing FCHL may also help in resolving the genetics of the insulin resistance syndrome.

Defects in the PPARg gene could affect

transcrip-tional activation of several genes in adipose tissue. On the other hand, changes in adipose tissue metabolism can cause dyslipidemias [47] and insulin resistance [48] via increased FFA levels. Therefore, the association of the Ala12 variant, that has a lower receptor activity, with low BMI and fasting insulin level and high HDL-cholesterol levels in a random population sample in our previous study in Finns was not surprising. Further-more, the Pro12Ala polymorphism may also be associ-ated with type 2 diabetes in Japanese-Americans [26]. In the present study no differences were found in allele frequencies of this polymorphism between nondiabetic elderly subjects with or without dyslipidemia or nondia-betic FCHL probands. Therefore, it is unlikely that this substitution could have a major role in FCHL or dyslipidemias in elderly subjects.

large number subjects in this study, the low frequency of the Pro12Ala substitution makes it difficult to draw a final conclusion about the role of this substitution in insulin resistance and dyslipidemias.

Although the effect of the Pro12Ala substitution of the PPARg gene does not seem to be very large, these

results, in combination with earlier reports [26 – 28], imply that variants in the PPARg gene affect

adipocyte metabolism. The Pro113Gln substitution af-fects the phosphorylation status of PPARg, resulting

in a constitutive activation of PPARg and, therefore,

to increased adipocyte differentiation and obesity [21 – 24,27]. In contrast, the Pro12Ala substitution reduces PPARg activity and adipocyte-specific gene expression,

resulting in a lower BMI and increased insulin sensi-tivity [26]. On the contrary, a recent study associated the Ala12 allele with high BMI in obese subjects [49]. This may indicate that the functional role of Pro12Ala substitution is different depending on the degree of energy storage. In the absence of excessive energy stor-age, decreased adipocyte differentiation, associated with less active PPARg Ala 12 allele, should in fact

lead to small insulin-sensitive adipocytes, and protect against the development of obesity. On the other hand, in the presence of excessive energy storage, di-etary and hormonal factors will override the protective effect offered by the PPARg Pro12Ala variant.

In summary, we found the Ala12 variant substitu-tion of the PPARg gene to be associated with low

fasting insulin levels in nondiabetic FCHL family members and dyslipidemic elderly subjects. In addi-tion, low BMI, low total triglyceride and high HDL-cholesterol levels were associated with this substitution in elderly subjects with dyslipidemia. Although this substitution is not a major locus for FCHL or dyslipi-demias in general, it seems to have a modifying effect in both disorders, possibly by affecting insulin sensitiv-ity.

Acknowledgements

This study was supported by grants from the Medi-cal Research Council of the Academy of Finland, the Finnish Foundation for Cardiovascular Disease, Aarne and Aili Turunen Foundation, INSERM and Institut Pasteur de Lille and National Institutes of Health Grant HL30086 to S. Deeb. The results in this paper were partly obtained with the S.A.G.E. Statistical Analysis for Genetic Epidemiology, Release 2.2. 1994 Computer program package available from the De-partment of Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, Ohio. This program is supported by a US Public Health Service Grant (1P41RR03655) from the National Center for Research Resources.

References

[1] Laakso M, Sarlund H, Mykka¨nen L. Insulin resistance is associ-ated with lipid and lipoprotein abnormalities in subjects with varying degrees of glucose tolerance. Arteriosclerosis 1990;10:223 – 31.

[2] Taskinen M. Insulin resistance and lipoprotein metabolism. Curr Opin Lipidol 1995;6:153 – 60.

[3] Karhapa¨a¨ P, Voutilainen E, Malkki M, Laakso M. Obese men with type IIB hyperlipidemia are insulin resistant. Arterioscler Thromb 1993;13:1469 – 75.

[4] Goldstein J, Schrott H, Hazzard W, Bierman E, Motulsky A. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest 1973;52:1544 – 68.

[5] Nikkila¨ E, Aro A. A family study of lipids and lipoproteins in coronary heart disease. Lancet 1973;1:954 – 8.

[6] Aitman T, Godsland I, Farren B, Crook D, Wong J, Scott J. Defects in insulin action on fatty acid and carbohydrate metabolism in familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol 1997;17:748 – 54.

[7] Bredie S, Cees J, Smits P, Stalenhoef A. Nonobese patients with familial combined hyperlipidemia are insulin resistant compared with their nonaffected relatives. Arterioscler Thromb Vasc Biol 1997;17:1465 – 71.

[8] Ascaso J, Merchante A, Lorente RI, Real J, Martinez-Valls J, Carmena R. A study of insulin resistance using the minimal model in nondiabetic familial combined hyperlipidemic patients. Metabolism 1998;47:508 – 13.

[9] Karjalainen L, Pihlajama¨ki J, Karhapa¨a¨ P, Laakso M. Impaired insulin-stimulated glucose oxidation and free fatty acid suppres-sion in patients with familial combined hyperlipidemia-a precur-sor defect for dyslipidemias. Arterioscler Thromb Vasc Biol 1998;18:1548 – 53.

[10] Arner P. Is familial combined hyperlipidaemia a genetic disorder of adipose tissue. Curr Opin Lipidol 1997;8:89 – 94.

[11] de Graaf J, Stalenhoef A. Defects of lipoprotein metabolism in familial combined hyperlipidemia. Curr Opin Lipidol 1998;9:189 – 96.

[12] Schoonjans K, Martin G, Staels B, Auwerx J. Peroxisome prolif-erator-activated receptors, orphans with ligands and functions. Curr Opin Lipidol 1997;8:159 – 66.

[13] Spiegelman B. PPAR-gamma: adipogenic regulator and thiazo-lidinedione receptor. Diabetes 1998;47:507 – 14.

[14] Zhu Y, Qi C, Korenberg JR, Chen XN, Noya D, Rao MS, et al. Structural organization of mouse peroxisome proliferator-acti-vated receptor gamma (mPPAR gamma) gene: alternative pro-moter use and different splicing yield two mPPAR gamma isoforms. Proc Natl Acad Sci 1995;92:7921 – 5.

[15] Fajas L, Auboeuf D, Raspe E, Schoonjans K, Lefebvre AM, Saladin R, et al. The organization, promoter analysis, and expression of the human PPARgamma gene. J Biol Chem 1997;272:18779 – 89.

[16] Fajas L, Fruchart J, Auwerx J. PPARg3 mRNA: a distinct PPARg mRNA subtype transcribed from an independent pro-moter. FEBS Lett 1998;438:55 – 60.

[17] Kliewer S, Lenhard J, Willson T, Patel I, Morris D, Lehman J. Cell 83 — a prostaglandin J2 metabolite binds peroxisome proliferator-activated receptor g and promotes adipocyte differ-entiation. Cell 1995;83:813 – 9.

J.Pihlajama¨ki et al./Atherosclerosis151 (2000) 567 – 574

574

[19] Lehmann JM, Moore LB, Smith Oliver TA, Wilkison WO, Willson TM, Kliewer SA. An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor gamma (PPAR gamma). J Biol Chem 1995;270:12953 – 66. [20] Berger J, Bailey P, Biswas C, Cullinan C, Doebber T, Hayes N,

et al. Thiazolidinediones produce a conformational change in peroxisomal proliferator-activated receptor-g: binding and acti-vation correlate with antidiabetic actions in db/db mice. En-docrinology 1996;137:4189 – 95.

[21] Hu E, Kim J, Sarraf P, Spiegelman B. Inhibition of adipogenesis through MAP kinase-mediated phosphorylation of PPARgamma. Science 1996;274:2100 – 3.

[22] Zhang B, Berger J, Zhou G, Elbrecht A, Biswas S, White-Car-rington S, et al. Insulin- and mitogen-activated protein kinase-mediated phosphorylation and activation of peroxisome proliferator-activated receptor gamma. J Biol Chem 1996;271:31771 – 4.

[23] Adams M, Reginato M, Shao D, Lazar M, Chatterjee V. Tran-scriptional activation by peroxisome proliferator-activated recep-tor gamma is inhibited by phosphorylation at a consensus mitogen-activated protein kinase site. J Biol Chem 1997;272:5128 – 32.

[24] Camp H, Tafuri S. Regulation of peroxisome proliferator-acti-vated receptor gamma activity by mitogen-actiproliferator-acti-vated protein kinase. J Biol Chem 1997;272:10811 – 6.

[25] Werman A, Hollenberg A, Solanes G, Bjorbaek C, Vidal Puig AJ, Flier JS. Ligand-independent activation domain in the N terminus of peroxisome proliferator-activated receptor gamma (PPARgamma). Differential activity of PPARgamma1 and -2 isoforms and influence of insulin. J Biol Chem 1997;272:20230 – 5. [26] Deeb S, Fajas L, Nemoto M, Pihlajama¨ki J, Mykka¨nen L, Kuusisto J, et al. A Pro12Ala substitution in the human perox-isome proliferator activated receptor gamma 2 is associated with decreased receptor activity, lower body mass index and improved insulin sensitivity. Nature Genet 1998;20:284 – 7.

[27] Ristow M, Muller-Wieland D, Pfeiffer A, Krone W, Kahn C. Obesity associated with a mutation in a genetic regulator of adipocyte differentiation. NEJM 1998;339:953 – 9.

[28] Meirhaeghe A, Fajas L, Helbecque N, Cottel D, Lebel P, Dallongeville J, et al. A genetic polymorphism of the peroxisome profilerator-activated receptor gamma gene influences plasma leptin levels in obese humans. Hum Mol Genet 1997;7:435 – 40. [29] De la Chapelle A. Disease gene mapping in isolated human

populations the example of Finland. J Med Genet 1993;30:857 – 65.

[30] Kittles R, Perola M, Peltonen L, Bergen A, Aragon R, Virkkunen M, et al. Dual origins of Finns revealed by Y chromosome haplotype variation. Am J Hum Genet 1998;62:1171 – 9.

[31] Voutilainen E. Serum lipids and lipoproteins in male survivors of acute myocardial infarction and their first-degree relatives: a case-control study. Kuopio, Finland: Kuopio University Publica-tions, 1992.

[32] Koivisto U, Ha¨ma¨la¨inen L, Taskinen M, Kettunen K, Kontula K. Prevalence of familial hypercholesterolemia among young north Karelian patients with coronary heart disease: a study based on diagnosis by polymerase chain reaction. J Lipid Res 1993;34:269 – 77.

[33] Mykka¨nen L, Laakso M, Uusitupa M, Pyo¨ra¨la¨ K. Prevalence of diabetes and impaired glucose tolerance in elderly subjects and their association with obesity and family history of dia-betes. Diabetes Care 1990;11:1099 – 105.

[34] Mykkanen L, Kuusisto J, Pyorala K, Laakso M. Cardiovascu-lar disease risk factors as predictors of type 2 (non-insulin-de-pendent) diabetes mellitus in elderly subjects. Diabetologia 1993;36:553 – 9.

[35] Diabetes mellitus: report of a WHO study group. Geneva, Switzerland: World Health Organization, 1985.

[36] Penttila¨ I, Voutilainen E, Laitinen P, Juutilainen P. Compari-son of different analytical and precipitation methods for the direct estimation of serum high density lipoprotein cholesterol. Scand J Clin Lab Invest 1984;41:353 – 60.

[37] Kretz K, Carson G, O’Brien J. Direct sequencing from low-melt agarose with sequenase. Nucleic Acid Res 1989;17:58 – 64. [38] George V, Elston R. Testing the association between

quantita-tive traits and polymorphic markers in pedigrees. Genet Epi-demiol 1987;4:193 – 201.

[39] Reaven G. Role of insulin resistance in human disease. Dia-betes 1988;37:1595 – 607.

[40] Kwiterovich P. Genetics and molecular biology of familial com-bined hyperlipidemia. Curr Opin Lipidol 1993;4:133 – 43. [41] Nevin D, Brunzell J, Deeb S. The LPL gene in individuals with

familial combined hyperlipidemia and decreased LPL activity. Arterioscler Thromb 1994;14:869 – 73.

[42] Wojciechowski AP, Farrall M, Cullen P, Wilson TM, Bayliss JD, Farren B, et al. Familial combined hyperlipidaemia linked to the apolipoprotein AI-CII-AIV gene cluster on chromosome 11q23-q24. Nature 1991;349:161 – 4.

[43] Dallinga Thie GM, Bu XD, van Linde Sibenius Trip M, Rotter JI, Lusis AJ, de Bruin TW. Apolipoprotein A-I/C-III/A-IV gene cluster in familial combined hyperlipidemia: effects on LDL-cholesterol and apolipoproteins B and C-III. J Lipid Res 1996;37:136 – 47.

[44] Wijsman EM, Brunzell JD, Jarvik GP, Austin MA, Motulsky AG, Deeb SS. Evidence against linkage of familial combined hyperlipidemia to the apolipoprotein AI – CIII – AIV gene com-plex. Arterioscler Thromb Vasc Biol 1998;18:215 – 26.

[45] Tahvanainen E, Pajukanta P, Porkka K, Nieminen S, Ika¨valko L, Nuotio I, et al. Haplotypes of the apoA-I/C-III/A-IV gene cluster and familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol 1998;18:1810 – 7.

[46] Pajukanta P, Nuotio I, Terwilliger J, Porkka K, Ylitalo K, Pihlajama¨ki J, et al. Linkage of familial combined hyperlipi-daemia to chromosome 1q21-q23. Nature Genet 1998;18:369 – 73.

[47] Sniderman A, Cianflone K. Substrate delivery as a determinant of hepatic apoB secretion. Arterioscler Thromb 1993;13:629 – 36.

[48] Randle P, Hales C, Garland P, Newsholme E. The glucose fatty acid cycle and its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet 1963;1:785 – 9.

[49] Beamer B, Yen C, Andersen R, Muller D, Elahi D, Cheskin L, et al. Association of the Pro12Ala variant in the peroxisome proliferator-activated receptor-gamma2 gene with obesity in two caucasian populations. Diabetes 1998;47:1806 – 8.