The genetic variation of the tenomodulin gene (TNMD) is associated with serum levels of systemic immune mediators--the Finnish Diabetes Prevention Study.

Abstract

We have reported that the genetic variation of the tenomodulin gene (TNMD) is associated with the risk of type 2 diabetes (T2DM), central obesity, and impaired glucose metabolism and the TNMD mRNA levels correlate with serum and mRNA levels of inflammatory markers. Our objective was to investigate the genetic associations of the single nucleotide polymorphisms of the TNMD gene with the serum levels of systemic immune mediators.
Seven single nucleotide polymorphisms were genotyped from 507 participants of the Finnish Diabetes Prevention Study. All subjects had body mass index >25 and impaired glucose tolerance.
The sequence variation of tenomodulin was consistently associated with the serum concentrations of acute phase reactants, macrophage migration inhibitory factor, and CCR5 receptor ligands. The genotype effects were modified by status of glucose metabolism and central obesity. Markers associated with increased risk of T2DM in our previous study were associated with serum concentrations of acute phase proteins in men so that the subjects possessing the genotypes associated with increased risk of T2DM had higher serum concentrations of acute phase reactants.
These results indicate that the genetic variation of TNMD is associated with low-grade inflammation. The putative link between TNMD and T2DM could be mediated through the effects on systemic immune mediators.

Full text

Page 1

Tenomodulin is Associated with Obesity and
Diabetes Risk: The Finnish Diabetes Prevention
Study
Anna-Maija Tolppanen,* Leena Pulkkinen,* Marjukka Kolehmainen,* Ursula Schwab,*† Jaana Lindstro ¨m,‡
Jaakko Tuomilehto,ठand Matti Uusitupa* for the Finnish Diabetes Prevention Study Group
Abstract
TOLPPANEN, ANNA-MAIJA, LEENA PULKKINEN,
MARJUKKA KOLEHMAINEN, URSULA SCHWAB,
JAANA LINDSTRO¨M, JAAKKO TUOMILEHTO, AND
MATTI UUSITUPA FOR THE FINNISH DIABETES
PREVENTIONSTUDYGROUP.
associated with obesity and diabetes risk: the Finnish
Diabetes Prevention Study. Obesity. 2007;15:1082–1088.
We recently showed that long-term weight reduction
changes the gene expression profile of adipose tissue in
overweight individuals with impaired glucose tolerance
(IGT). One of the responding genes was X-chromosomal
tenomodulin (TNMD), a putative angiogenesis inhibitor.
Our aim was to study the associations of individual single
nucleotide polymorphisms and haplotypes with adiposity,
glucose metabolism, and the risk of type 2 diabetes (T2D).
Seven single nucleotide polymorphisms from two different
haploblocks were genotyped from 507 participants of the
Finnish Diabetes Prevention Study (DPS). Sex-specific
genotypeeffectswere observed.
haploblock 1 were associated with features of adiposity in
women (rs5966709, rs4828037) and men (rs11798018).
Markers rs2073163 and rs1155794 from haploblock 2 were
associated with 2-hour plasma glucose levels in men during
the 3-year follow-up. The same two markers together with
Tenomodulinis
Threemarkersof
rs2073162 associated with the conversion of IGT to T2D in
men. The risk of developing T2D was approximately 2-fold
in individuals with genotypes associated with higher 2-hour
plasma glucose levels; the hazard ratios were 2.192 (p ?
0.025) for rs2073162-A, 2.191 (p ? 0.027) for rs2073163-
C, and 1.998 (p ? 0.054) for rs1155974-T. These results
suggest that TNMD polymorphisms are associated with
adiposity and also with glucose metabolism and conversion
from IGT to T2D in men.
Key words: adiposity, glucose metabolism, genetics, type
2 diabetes, central obesity
Adipose tissue manifests signs of insulin resistance be-
fore glucose intolerance develops (1). Molecular mecha-
nisms include altered expression and post-translational
modifications of target molecules of the insulin signaling
pathway (2,3), altered production and secretion of insulin-
sensitizing adipokines (4), and compounds predisposing to
insulin resistance (1). These dysregulations can mediate
insulin resistance to other tissues (1).
Our previous studies, attempting to gain understanding
on molecular mechanisms behind adipose tissue-mediated
effects on insulin and glucose metabolism, have included
weight reduction interventions in individuals with high risk
of type 2 diabetes (T2D),1followed by gene expression
studies of adipose tissue. We showed that weight reduction
decreases tenomodulin (TNMD) gene expression in adipose
tissue. The mRNA levels also correlated consistently with
fasting serum insulin and leptin levels (5).
Tenomodulin, a type II transmembrane glycoprotein, is
abundant in hypovascular connective tissues (6,7). Similarly
to its homolog, chondromodulin, tenomodulin is a putative
angiogenesis inhibitor (7). Abnormal angiogenesis, a com-
Received for review October 2, 2006.
Accepted in final form November 22, 2006.
The costs of publication of this article were defrayed, in part, by the payment of page
charges. This article must, therefore, be hereby marked “advertisement” in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
*Department of Clinical Nutrition and Food and Health Research Centre, School of Public
Health and Clinical Nutrition, University of Kuopio, Kuopio, Finland; †Kuopio University
Hospital, Kuopio, Finland; ‡Department of Epidemiology and Health Promotion, Diabetes
Unit, National Public Health Institute, Helsinki, Finland; and §Department of Public Health,
University of Helsinki, Helsinki, Finland.
Address correspondence to Anna-Maija Tolppanen, Department of Clinical Nutrition and
Food and Health Research Centre, University of Kuopio, PO Box 1627, 70211 Kuopio,
Finland.
E-mail: Anna-Maija.Tolppanen@uku.fi
Copyright © 2007 NAASO
1Nonstandard abbreviations: T2D, type 2 diabetes; TNMD, tenomodulin; IGT, impaired
glucose tolerance; DPS, Diabetes Prevention Study; SNP, single nucleotide polymorphism;
HWE, Hardy-Weinberg equilibrium; WHR, waist-to-hip ratio; CI, confidence interval.
1082 OBESITY Vol. 15 No. 5 May 2007

Page 2

mon complication in T2D, is modulated by tissue damage
resulting in endothelial cell–matrix interactions regulated by
a multitude of different factors (8). There is evidence that
prolonged elevation of glucose levels increases formation of
advanced glycation endproducts, hampering post-transla-
tional modifications and processing of extracellular matrix
molecules. These lead to diabetic tissue fibrosis, impaired
cell–matrix signaling, and finally, to abnormal neovascular-
ization (8). Our aim was to study the associations of the
genetic variations of TNMD with body size, adiposity, in-
sulin, and glucose metabolism and the conversion of im-
paired glucose tolerance (IGT) to T2D in the Finnish Dia-
betes Prevention Study (DPS) (9).
Seven TNMD single nucleotide polymorphisms (SNPs;
Figure 1) were genotyped from 507 DNA samples (166 men
and 341 women) from the DPS subjects. The success rate
for genotyping was 100% for all markers. Because TNMD is
X-chromosomal, men and women were analyzed separately.
All SNPs, except rs2073162 (p ? 0.0093) and rs4828038
(p ? 0.0082), were in Hardy-Weinberg equilibrium (HWE).
However, these markers were included in the analysis,
because they are in HWE in the reference population (CEU;
Utah residents with ancestry from northern and western
Europe) of the HapMap database. The deviation could be
explained by the characteristics of the Finnish genetic her-
itage and non-random mating caused by an isolated small
population (10) or chance finding because of a limited
sample size. In this case, the deviation is unlikely to result
from genotyping error because this was assessed by repeat-
ing the genotyping by randomly selecting 6.3% of samples.
The error rate for all markers was 0%. None of the SNP
pairs were in complete linkage disequilibrium (supplemen-
tal data, Table S1; available on online at the Obesity web-
site, www.obesityresearch.org).
Sex differences were observed when associations of
TNMD SNPs with body size measurements were studied. In
women, rs2073162 associated with horizontal diameter.
The subjects with the rs2073162-AA genotype had the
smallest values (p ? 0.038, adjusted for age and BMI;
supplemental data, Table S2; available online at the Obesity
website, www.obesityresearch.org). Furthermore, rs4828037
was associated with sagittal diameter, and the rs4828037-TT
genotype had the highest values (p ? 0.015).
In the follow-up data analysis, the intervention and con-
trol groups of the DPS were analyzed together, because the
allele distribution of each marker was similar in both
groups, and group as an adjustment factor was non-signif-
icant. The genotype-randomization group interaction was
statistically non-significant in all analysis. The results are
age adjusted in the analysis of BMI and weight, and age and
BMI adjusted in analysis of other body size parameters.
The markers rs5966709 and rs4828037 from haploblock
1 were significantly associated with central adiposity in
women during the 3-year follow-up. In both cases, subjects
homozygous for the minor alleles (5966709-TT and
4828037-CC) had the lowest values (Figure 2). rs5966709
was associated with waist-to-hip ratio (WHR; p ? 0.028),
waist circumference (p ? 0.028), and sagittal (p ? 0.014)
and horizontal diameter (p ? 0.043), whereas SNP
rs4828037 associated statistically significantly only with
sagittal diameter (p ? 0.026; Figure 2).
In men, the SNP rs11798018 was associated with BMI
and weight (p ? 0.038 and p ? 0.029, respectively) during
the 3-year follow-up. The subjects with a minor A allele had
lower values than those with the G allele.
To study haplotype effects of haploblock 1 on body size
measurements at baseline in women, three-marker haplo-
types (markers rs11798018, rs5966709, and rs4828037)
were constructed. Three major haplotypes with frequency
?0.05 were observed (AGT: 0.371; CTC: 0.351; CGT:
0.249). The results were consistent with the analysis of
individual markers in the 3-year follow-up. In comparison
with the reference haplotype AGT, the individuals with the
CTC haplotype had lower sagittal diameter than the others
(p ? 0.044). The difference was similar to that observed
within the genotypes rs4828037-TT and rs4828037-CC.
Because adiposity is a strong risk factor for impairment of
glucose metabolism, we also studied the association of TNMD
SNPs with glucose and insulin levels and T2D risk. No geno-
type differences in insulin and glucose levels were observed at
baseline, except with rs2073162 in women. Specifically, the
rs2073162-GG genotype was associated with lower fasting
plasma glucose levels (p ? 0.021) than other genotypes (sup-
plemental data, Table S2; available online at the Obesity web-
site, www.obesityresearch.org). During the 3-year follow-up,
only rs2073162 associated with 2-hour plasma glucose (p ?
0.013, adjusted for intervention, age, and BMI) in women. The
lowest levels were observed with the AA genotype.
In men, three markers of haploblock 2 associated with
2-hour plasma glucose levels during the 3-year follow-up
(Figure 3A-C). Men with rs2073163-C and rs1155974-T
genotypes had higher 2-hour plasma glucose levels than
those with rs2073163-T and rs1155974-C genotypes (p ?
0.038 and p ? 0.011, respectively; adjusted for age and
BMI). The differences between rs2073162 genotypes were
statistically non-significant (p ? 0.249, adjusted for group,
age, and BMI; Figure 3).
Figure 1: Schematic representation of the TNMD gene indicating
the locations of analyzed SNPs that belonged to the two haplo-
blocks. ?, UTR regions; f, coding region.
Variations of TNMD Are Associated with T2D, Tolppanen et al.
OBESITY Vol. 15 No. 5 May 2007 1083

Page 3

The SNPs associating with 2-hour plasma glucose levels
during the 3-year follow-up also associated with conversion of
IGT to T2D in men (Figure 3D-F). Those with the alleles
rs2073162-A, rs2073163-C, and 1155974-T linked to higher
2-hour plasma glucose values were approximately two times
morelikelytodevelopT2Dduringthe5-yearfollow-up(Table
1). rs2073162 and rs2073163 showed stronger associations
with conversion to T2D [corresponding hazard ratio (95%
confidence interval (CI)/p value: Model 1, 2.099 (1.039 to
4.242)/0.039 and 2.092 (1.025 to 4.268)/0.042; Model 2: 2.193
(1.105 to 4.354)/0.025 and 2.191 (1.092 to 4.394)/0.027;
Model 3: 2.143 (1.084 to 4.237)/0.028 and 2.109 (1.054 to
4.220)/0.035] than rs1155974, which showed a moderate trend
for association in Model 2 [Model 1: 1.900 (0.926 to 3.899)/
0.08; Model 2: 1.998 (0.989 to 4.036)/0.054; Model 3: 1.939
(0.961 to 3.914)/0.065]. In women, none of the SNPs contrib-
uted to the risk of T2D.
For men, three-marker haplotypes of SNPs rs2073162,
rs2073163, and rs1155974 were constructed, and four major
haplotypes (frequencies ?0.05) were observed (Table 2). As-
Figure 2: Waist circumference (A), WHR (B), horizontal diameter (C), and sagittal diameter (D) in women according to the SNP rs5966709
during the 3-year follow-up. F, rs5966709-GG- (n ? 125) genotype; Œ, rs5966709-GT- (n ? 124) genotype; E, rs5966709-TT- (n ? 35)
genotype. Age- and BMI-adjusted p values ?0.1 are indicated.
Variations of TNMD Are Associated with T2D, Tolppanen et al.
1084OBESITY Vol. 15 No. 5 May 2007

Page 4

sociation of haplotypes with the conversion from IGT to T2D
was assessed by Cox regression analysis with model adjusted
for intervention, baseline BMI, and fasting plasma glucose
level (Table 2). Baseline age and WHR were excluded, be-
cause they caused multicollinearity when grouping was per-
formed according to haplotypes. A tendency for overall hap-
lotype effect for T2D risk was observed (p ? 0.075; Table 2).
Interestingly, the haplotype ACT, containing all individual risk
alleles, was associated with 2.3-fold risk for developing T2D
(p ? 0.041; 95% CI, 1.034 to 5.175) compared with reference
haplotype GTC. This supports the results obtained from anal-
ysis of individual SNPs. With ATC and GCT haplotypes (the
risk alleles bolded), trends were observed (p ? 0.065, 95%
CI ? 0.942 to 7.779 and p ? 0.068, 95% CI ? 0.915 to
12.629, respectively; Table 2).
The main findings of this study indicate that genetic
variations of tenomodulin have sex-specific associations
with features of central adiposity, glucose metabolism, and
conversion of IGT to T2D in subjects prone to T2D. How-
ever, the results are corrected for the number of genotype
groups among women and not for the number of analyzed
SNPs (n ? 7). If this is done by the Bonferroni method, the
results would not be statistically significant. The stratifica-
tion of a relatively small population to different groups by
randomization and sex can also weaken the statistical
power. We handled this issue by applying the mandatory
stratification by sex and using the randomization group as
an adjustment factor, because this was justifiable based on
the lack of interaction between the genotypes and random-
ization group and similar allele frequencies in both random-
ization groups. Nevertheless, the power of the DPS cohort is
that subjects are carefully selected and phenotyped, and, in
addition to the cross-sectional baseline data, the longitu-
dinal follow-up data were also assessed. The follow-up
data of various quantitative and categorical traits makes
the DPS more powerful in finding true genetic associa-
tions than studies with a single measurement of a given
variable. It should be noted that the diagnosis of diabetes
was based on the repeated oral glucose tolerance test (9).
Furthermore, the Finnish population with few founders
has been shown to offer substantial advantages for ge-
netic studies (11).
The sex differences can arise from the genetic locus.
X-chromosomal genes show significant variation in expres-
sion and function between men and women, partially be-
cause of variation in gene doses and random inactivation of
one of the X-chromosomes in women (12). Accordingly,
women had two times higher tenomodulin expression in
adipose tissue than men (5), indicating dose-specific expres-
sion levels. Furthermore, the cellular microenvironment can
differ by sex because of differences in hormone levels and
gene expression (13).
It is difficult to suggest whether the causative variant
truly is one of the studied markers or simply a SNP that is
in complete disequilibrium with them. The sequence varia-
tions of TNMD have not been thoroughly characterized, and
perhaps, sequencing of the gene together with functional
studies would address this issue. Without functional studies,
Figure 3: Two-hour plasma glucose values (A-C) and corresponding survival curves that indicate the incidence of T2D in men (D-F)
according to genotypes of the three TNMD SNPs that were associated with conversion of IGT to T2D in men. The glucose levels are mean ?
standard error. F, major alleles [rs2073162-G (n ? 130), rs2073163-T (n ? 141), and rs1155974-C (n ? 145)]; E, minor alleles
[rs2073162-A (n ? 75), rs2073163-C (n ? 74), and rs1155974-T (n ? 60)].
Variations of TNMD Are Associated with T2D, Tolppanen et al.
OBESITY Vol. 15 No. 5 May 2007 1085

Page 5

Table 1.
Hazard ratios for the association of T2D with SNPs of the TNMD gene in men in the DPS
Marker
Genotype
Incidence
of T2D
Model 1*
Model 2†
Model 3‡
Hazard ratio
(95% CI)
p
Hazard ratio
(95% CI)
p
Hazard ratio
(95% CI)
p
rs11798018
C
25/110
1.0 (reference)
1.0 (reference)
1.0 (reference)
A
14/56
1.570 (0.768 to 3.209)
0.216
1.501 (0.735 to 3.064)
0.265
1.590 (0.792 to 3.193)
0.192
rs5966709
G
25/111
1.0 (reference)
1.0 (reference)
1.0 (reference)
T
14/55
0.955 (0.482 to 1.892)
0.895
1.025 (0.526 to 1.996)
0.942
1.005 (0.518 to 1.947)
0.989
rs4828037
T
24/108
1.0 (reference)
1.0 (reference)
1.0 (reference)
C
15/58
1.025 (0.523 to 2.009)
0.942
1.099 (0.570 to 2.116)
0.779
1.085 (0.566 to 2.081)
0.805
rs2073162
G
23/107
1.0 (reference)
1.0 (reference)
1.0 (reference)
A
16/59
2.099 (1.039 to 4.242)
0.039
2.193 (1.105 to 4.354)
0.025
2.143 (1.084 to 4.237)
0.028
rs2073163
T
25/116
1.0 (reference)
1.0 (reference)
1.0 (reference)
C
14/60
2.092 (1.025 to 4.268)
0.042
2.191 (1.092 to 4.394)
0.027
2.109 (1.054 to 4.220)
0.035
rs4828038
C
18/73
1.0 (reference)
1.0 (reference)
1.0 (reference)
T
21/93
0.944 (0.487 to 1.829)
0.865
0.919 (0.475 to 1.778)
0.801
0.998 (0.521 to 1.912)
0.995
rs1155974
C
26/119
1.0 (reference)
1.0 (reference)
1.0 (reference)
T
13/47
1.900 (0.926 to 3.899)
0.08
1.998 (0.989 to 4.036)
0.054
1.939 (0.961 to 3.914)
0.065
T2D, type 2 diabetes; TNMD, tenomodulin; DPS, Diabetes Prevention Study; SNP, single nucleotide polymorphism; CI, confidence interval; WHR, waist-to-hip ratio. The
incidence of T2D is expressed as the number of affected individuals/all.
* Adjusted for baseline age, BMI, WHR, fasting plasma glucose, and intervention. † Adjusted for baseline BMI, WHR, fasting plasma glucose, and intervention ‡ Adjusted for baseline BMI, fasting plasma glucose, and intervention.
Variations of TNMD Are Associated with T2D, Tolppanen et al.
1086OBESITY Vol. 15 No. 5 May 2007

Page 6

it is also difficult to speculate how tenomodulin participates
in adiposity and glucose metabolism. Although tenomodu-
lin is suggested to mediate anti-angiogenic effects, its spe-
cific function in adipose tissue and the cell types where it is
mostly expressed are unknown. It has been shown that
targeted induction of apoptosis in the vasculature of adipose
tissue reverses obesity and normalizes metabolism in ob/ob
mice (14). Furthermore, it has been suggested that changes
in adipose tissue blood flow may modulate the ?-cell dys-
function in T2D (15). Thus, because of its transmembrane
location and possible anti-angiogenic function, tenomodulin
may regulate vasculature formation in adipose tissue and
modulate adiposity, glucose metabolism, and T2D risk
through cell-matrix interactions. However, characterization
of a ligand, intracellular targets and a signaling pathway are
essential steps for understanding the mechanism of these
functions. Furthermore, the results of this genetic study
need replication in other populations.
Research Methods and Procedures
Finnish DPS
DPS is a randomized, controlled, multicenter study con-
ducted in Finland from 1993 to 2000. Study design and
applied methods have been reported elsewhere in detail
(9,16). The main inclusion criteria were BMI ?25 kg/m2,
age 40 to 64 years, and IGT on the basis of the mean values
of two oral glucose tolerance tests. Five hundred twenty-two
subjects were randomized to intensive individualized diet
and an exercise intervention group and control group by
stratification according to the clinical features, sex, and
mean 2-hour plasma glucose levels in an oral glucose tol-
erance test. DNA was available from 507 individuals (166
men and 341 women). Altogether, 115 of 507 participants
developed T2D during the follow-up lasting from 0 to 6
years, with an average duration of 4 years. The diagnosis
was based on the repeated 2-hour oral glucose tolerance test
performed annually. The study protocol was approved by
the Ethics Committee of the National Public Health Institute
in Helsinki, Finland, and the participants gave written in-
formed consent.
Measurements
A medical history was taken, and physical examination
was done at baseline and annual follow-up visits (16).
Plasma glucose was measured at each center by standard
methods. The serum insulin concentration was measured in
a central laboratory by the radioimmunoassay method
(Pharmacia, Uppsala, Sweden). The diagnosis of T2D and
other glucose intolerance categories was based on the 1985
World Health Organization criteria (17). The diagnosis of
diabetes was confirmed by a second oral glucose tolerance
test.
Selection of SNPs and Genotype Analysis
The HapMap (18) and the National Center for Biotech-
nology Information databases were used for selection of
TNMD SNPs for genotype analysis. Specifically, two of
two-tag SNPs of haploblock 1 (rs5966709 and rs4828037)
and three from five-tag SNPs of haploblock 2 (rs2073162,
rs2076163, and rs4828038) were selected from the HapMap
database. In addition, two SNPs were selected from the
National Center for Biotechnology Information database to
cover the 5? and 3? ends of the gene (rs11798018 and
rs1155974). The redundancy was taken into account when
possible. All markers, except the synonymous coding region
SNP rs2073162, were intronic. Genotyping was performed
with TaqMan Allelic Discrimination Assay according to
manufacturer’s instructions using the ABI PRISM 7000
sequence detector (Applied Biosystems, Foster City, CA).
To calculate the error rate for genotyping, a subset of
randomly selected samples representing 6.3% of the study
cohort was repeated.
Statistics
Haploview software (19) (http://www.broad.mit.edu/
mpg/haploview/) was used for linkage disequilibrium anal-
Table 2.
individually associated with 2-hour plasma glucose levels and T2D in men (rs2073162, rs2073163, and
rs1155974)
Hazard ratios for the association of T2D with the three-marker haplotypes consisting of markers
HaplotypeFrequency (%)
Risk for T2D
Hazard ratio (95% CI)
p
GTC
ACT
ATC
GCT
59.5
22.7
11.7
6.1
1.0 (reference)
2.313 (1.034 to 5.175)
2.707 (0.942 to 7.779)
3.400 (0.915 to 12.629)
0.075
0.041
0.065
0.068
T2D, type 2 diabetes.
Variations of TNMD Are Associated with T2D, Tolppanen et al.
OBESITY Vol. 15 No. 5 May 20071087

Page 7

ysis, and association studies were performed with SPSS11.5
for Windows (SPSS, Chicago, IL). The data are presented as
means ? standard deviation, and p ? 0.05 was considered
statistically significant. Normal distribution was tested with
a Lilliefors-corrected Kolmogorov-Smirnov test. Appropri-
ate transformations were performed to achieve normal dis-
tribution when necessary. Because of the X-chromosomal
location of the TNMD gene, men and women were analyzed
separately. Genotype distribution in the study population
and the groups were assessed with the ?2test. The genotype
differences in continuous variables were evaluated with
general linear models using univariate analysis or repeated
measurements. Adjustments for age, BMI, and intervention
were done when necessary, i.e., their contribution to the
model was significant. Bonferroni correction was used in
comparisons between different genotypes among women. If
the continuous variables were not distributed normally,
Kruskal-Wallis or Mann-Whitney U tests were used. The
association of SNPs with conversion of IGT to T2D was
analyzed with Cox regression using appropriate covariates.
Additionally, the effects of genotype and intervention on ?
weight and ? waist circumference at Year 1 (calculated as
measurementbaseline? measurementyear1) were assessed.
THESIAS 3 (20) (http://ecgene.net/genecanvas/) was
used for haplotype analysis in women, whereas the haplo-
type effects in men were analyzed by SPSS 11.5 for Win-
dows, because it was not possible to analyze these with
THESIAS 3. Thus, in women only, the baseline effects were
studied by haplotype analysis, whereas in men, the fol-
low-up data were also analyzed with general linear models,
and haplotype effects on progression of IGT to T2D were
studied with Cox regression analysis.
Acknowledgments
We gratefully acknowledge The Finnish Diabetes Pre-
vention Study Group. We also thank Minna Kiuttu, Kaija
Kettunen, and Tuomas Onnukka for excellent technical
assistance and Vesa Kiviniemi and Ursula Mager for assist-
ing in statistical analysis. This study was supported by The
Academy of Finland Grants 211497 to M.U. and 38387 and
46558 (to J.T.), Sigrid Juselius Foundation, Juho Vainio
Foundation, and EVO Fund of the Kuopio University Hos-
pital (5106 and 5198 to M.U.), from Ministry of Health and
Social Affairs.
References
1. Smith U. Impaired (‘diabetic’) insulin signaling and action
occur in fat cells long before glucose intolerance—is insulin
resistance initiated in the adipose tissue? Int J Obes Relat
Metab Disord. 2002;26:897–904.
2. Carvalho E, Eliasson B, Wesslau C, Smith U. Impaired
phosphorylation and insulin-stimulated translocation to the
plasma membrane of protein kinase B/Akt in adipocytes from
type II diabetic subjects. Diabetologia. 2000;43:1107–15.
3. Garvey WT, Maianu L, Hancock JA, Golichowski AM,
Baron A. Gene expression of GLUT4 in skeletal muscle from
insulin-resistant patients with obesity IGT, GDM, and
NIDDM. Diabetes. 1992;41:465–75.
4. Bays H, Mandarino L, DeFronzo RA. Role of the adipocyte,
free fatty acids, and ectopic fat in pathogenesis of type 2
diabetes mellitus: peroxisomal proliferator-activated receptor
agonists provide a rational therapeutic approach. J Clin En-
docrinol Metab. 2004;89:463–78.
5. Kolehmainen M, Schwab U, Salopuro T, et al. Lifestyle
intervention downregulates the expression of genes linked to
extracellular matrix (ECM) and cell cycle in adipose tissue in
subjects with metabolic syndrome—the GENOBIN study.
Diabetologia. 2006;49(Suppl 1):126–7.
6. Brandau O, Meindl A, Fassler R, Aszodi A. A novel gene,
tendin, is strongly expressed in tendons and ligaments and
shows high homology with chondromodulin-I. Dev Dyn.
2001;221:72–80.
7. Shukunami C, Oshima Y, Hiraki Y. Molecular cloning of
tenomodulin, a novel chondromodulin-I related gene. Bio-
chem Biophys Res Commun. 2001;280:1323–7.
8. Silvestre JS, Levy BI. Molecular basis of angiopathy in
diabetes mellitus. Circ Res. 2006;98:4–6.
9. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention
of type 2 diabetes mellitus by changes in lifestyle among
subjects with impaired glucose tolerance. N Engl J Med.
2001;344:1343–50.
10. Nevanlinna HR. The Finnish population structure. A genetic
and genealogical study. Hereditas. 1972;71:195–236.
11. Service S, DeYoung J, Karayiorgou M, et al. Magnitude and
distribution of linkage disequilibrium in population isolates
and implications for genome-wide association studies. Nat
Genet. 2006;38:556–60.
12. Carrel L, Willard HF. X-inactivation profile reveals exten-
sive variability in X-linked gene expression in females. Na-
ture. 2005;434:400–4.
13. Rinn JL, Snyder M. Sexual dimorphism in mammalian gene
expression. Trends Genet. 2005;21:298–305.
14. Kolonin MG, Saha PK, Chan L, Pasqualini R, Arap W.
Reversal of obesity by targeted ablation of adipose tissue. Nat
Med. 2004;10:625–32.
15. Kampf C, Bodin B, Kallskog O, Carlsson C, Jansson L.
Marked increase in white adipose tissue blood perfusion in the
type 2 diabetic GK rat. Diabetes. 2005;54:2620–7.
16. Eriksson J, Lindstrom J, Valle T, et al. Prevention of type
II diabetes in subjects with impaired glucose tolerance: The
diabetes prevention study (DPS) in Finland. Study design and
1-year interim report on the feasibility of the lifestyle inter-
vention programme. Diabetologia. 1999;42:793–801.
17. World Health Organization Expert Committee. Diabetes
Mellitus. Geneva, Switzerland: World Health Organization,
1985.
18. Altshuler D, Brooks LD, Chakravarti A, et al. A haplotype
map of the human genome. Nature. 2005;437:1299–320.
19. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis
and visualization of LD and haplotype maps. Bioinformatics.
2005;21:263–5.
20. Tregouet DA, Escolano S, Tiret L, Mallet A, Golmard JL.
A new algorithm for haplotype-based association analysis: the
stochastic-EM algorithm. Ann Hum Genet. 2004;68:165–77.
Variations of TNMD Are Associated with T2D, Tolppanen et al.
1088OBESITY Vol. 15 No. 5 May 2007