| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Receptor 2 Is Elevated in Obesity But Is Not Related to Insulin Sensitivity: A Study in Identical Twins Discordant for Obesity
Departments of Medicine (T.R.) and Clinical Chemistry (K.P.), University of Turku; and Research and Development Center, Social Insurance Institution (T.R.), FIN-20520 Turku, Finland; Department of Public Health, University of Helsinki (J.K.), FIN-00014 Helsinki, Finland; and Department of Public Health and General Practice, University of Oulu (J.K.), FIN-90220 Oulu, Finland
Address all correspondence and requests for reprints to: Tapani Rönnemaa, M.D., Department of Medicine, University of Turku, FIN-20520 Turku, Finland. E-mail: tapani.ronnemaa{at}utu.fi
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(TNF) and its soluble receptor 2 (TNFR2)
are expressed in adipose tissue and are possibly involved in the
pathogenesis of insulin resistance. Information about serum levels of
TNFR2 in human obesity, especially the possible role of genetic factors
and body fat distribution, is scanty. We measured serum TNF and
soluble TNFR2 concentrations in 23 identical twin pairs who had an
average 18-kg intrapair difference in body weight. The mean TNF
concentration was 44.1 ng/L in obese and 34.2 ng/L in lean cotwins
(P = 0.051). The respective values for TNFR2 were
1989 and 1840 ng/L (P = 0.004). The intrapair
difference in TNFR2 level correlated positively (r-value always 
0.56;
P 
0.01) with intrapair differences in body mass
index, percent body fat, and abdominal sc fat area (assessed by
magnetic resonance imaging), but not with differences in visceral fat
area, glucose or insulin areas under the curve, or insulin sensitivity
index in the oral glucose tolerance test. The intraclass correlation
for TNFR2 was 0.67, and the genetic variation in circulating TNFR2
level was almost 6-fold higher than the variation due to obesity. We
conclude that the soluble TNFR2 concentration is determined by both
genetic factors and adiposity, especially sc fat. Measurement of
circulating TNFR2 does not seem to be useful in identifying obese
individuals who are insulin resistant. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(TNF) is
a cytokine that is expressed mainly in monocyte-macrophages, and it has
many proinflammatory and immunoregulatory actions (1, 2). It was found
that TNF is also expressed in adipose tissue in humans (3), and
obese subjects express it more in their fat tissue relative to lean
controls (4). The plasma concentration of TNF is increased among
obese subjects, and it decreases with weight loss (5). TNF is also
expressed in other tissues that are involved in fat metabolism,
e.g. human muscle tissue (6). TNF inhibits
insulin-stimulated glucose uptake in adipocytes in vitro by
decreasing phosphorylation of the insulin receptor (7), and thus
overexpression of TNF has been implicated in the pathogenesis of
insulin resistance as well as type 2 diabetes (8, 9). Although
short-term TNF treatment causes whole body insulin resistance in
rats (10), it does not affect insulin sensitivity in skeletal muscle in
these animals (11). A neutralizing antibody against TNF has not
yielded changes in insulin sensitivity in patients with established
type 2 diabetes (12). These findings (11, 12) suggest that the role of
TNF in the pathogenesis of insulin resistance is not yet fully
understood.
The actions of TNF are mediated by two distinct TNF receptors,
receptors 1 and 2 (TNFR1 and TNFR2) (13). Obese human females express
2-fold more TNFR2 in their fat tissue and have 6-fold increased TNFR2
plasma levels compared with lean controls, whereas the expression and
plasma levels of TNFR1 are similar in obese and lean women (14).
Soluble TNFR2 levels also correlate with body mass index (BMI),
fat-free mass and waist to hip ratio, but not with fat mass or percent
fat mass, whereas TNFR2 levels showed a weak inverse correlation with
insulin sensitivity in humans (15). Recent studies have demonstrated
polymorphisms of TNF gene promoter region that may modulate the
expression of TNF as well as the relationship between
adiposity and insulin resistance (16, 17).
Most studies examining circulating levels of TNF and TNF
receptors in human obesity have not taken into account body fat
distribution. Moreover, information about the possible role of genetic
factors on circulating levels of TNF and TNF receptors is not
available. A powerful strategy to study the independent contribution of
obesity to serum TNF and TNF receptor concentrations is to
examine identical twins who are discordant for obesity. Therefore, we
measured serum levels of TNF and TNFR2 in a sample of 23 identical
twin pairs with an average weight difference of 18 kg, with the main
emphasis being to compare the relative contributions of genetics and
adiposity on their circulating levels. We also took into account
visceral and sc distributions of fat and assessed associations between
TNF and TNFR2 with indexes of glucose metabolism.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The Finnish Twin Cohort includes all pairs (4307 monozygous, 9581 like-sexed dizygous pairs) of adult Finnish twins born before 1958 and alive in 1975 (18). Based on a postal questionnaire sent to the twins in 1990, identical twin pairs born between 1932 and 1957 and discordant for obesity were identified (19). Discordance for obesity was defined as a BMI difference of at least 4 kg/m2; in addition, the BMI of the obese cotwin had to be more than 27 kg/m2, and the BMI of the lean cotwin had to be less than 25 kg/m2. Subjects with a history of thyroid disorders, psychiatric diseases, diabetes, major musculoskeletal problems, and other diseases or taking medications (e.g. diuretics or ß-blockers) possibly affecting glucose metabolism were excluded. All eligible twin pairs, based on a response letter, were invited to take part in the present study in 1992 provided that they still fulfilled the criteria. A total of 28 twin pairs were examined. None of the subjects had any acute illness at the time of the examination.
The physical examination revealed that two of the pairs had a BMI difference less than 3 kg/m2 and were excluded. Three pairs had a BMI difference between 3–4 kg/m2 and were included in the final study population. The obese cotwin of one pair was found to have previously undiagnosed overt diabetes mellitus, and this pair was excluded.
Zygosity of the twin pairs was originally based on a validated self-administered questionnaire (20). The monozygosity of the pairs of the present study was confirmed by dermatomatoglyphic analysis of fingertip prints (21, 22) by a highly experienced expert. All except six pairs were confirmed to be monozygotic. DNA samples of the six pairs with uncertain zygosity were typed for markers at six different polymorphic gene loci (DIS80, APOB, D17S30, COL2A1, VWA, and HUMTH). Four of the pairs were monozygotic, and the two other pairs were dizygotic; these two pairs were excluded.
Thus, the final study sample consisted of 23 nondiabetic identical twin pairs (14 females and 9 males) with more than a 3 kg/m2 difference in BMI and having no diseases and taking no medications that could affect the results.
Methods
Serum TNF and soluble TNF receptor concentrations were
measured with commercial sandwich-type enzyme-linked immunosorbent
assays. The sensitivity of the assay was 3.0 ng/L for TNF
(Biosource Technologies, Inc., Europe, Nivelles, Belgium)
and 1.0 ng/L for soluble TNF-receptor 2 (R&D Systems, Minneapolis,
MN).
The percentage of body fat was determined using the so-called four-component method (23). The method is based on the division of body mass into four components with different specific weights (sw): fat tissue (sw, 0.9007), water (sw, 0.994), minerals (sw, 3.042), and proteins (sw, 1.34). Water mass was estimated by bioelectric impedance method (BIA-101A/S, RJL Systems, Inc., Clemens, MI) (24). Mineral mass was estimated by dual energy x-ray absorptiometry (Norland XR26, Norland Corp., Fort Atkinson, WI). The specific weight of the whole body was estimated by underwater weighing corrected for information on body water and mineral mass. The proportion of fat tissue was calculated from the specific gravity of the whole body according to the formula of Siri (25).
The distribution of body fat was measured by magnetic resonance imaging (MRI) (26). Imaging was performed at 0.1 Tesla (Mega4, Instrumentarium Co., Helsinki, Finland). Axial and sagittal localizers were used to obtain a transaxial T1-weighted image (relaxation time/echo time = 155/20; slice thickness, 10 mm) at the level of the fourth lumbar vertebra. Visceral and sc fat areas were measured. MRI was not performed in three pairs either because of claustrophobia or because of temporary malfunctioning of the MRI equipment. These three pairs were excluded from analyses concerning body fat distribution.
Glucose metabolism was assessed in a 2-h glucose (75-g) tolerance test, with glucose and insulin measurements at 0, 30, 60, 90, and 120 min. Serum glucose was measured by the glucose dehydrogenase method (Merck Diagnostica, Darmstadt, Germany). Plasma insulin was measured by RIA (Pharmacia Biotech, Uppsala, Sweden). The antiserum of this kit is specific for insulin and does not cross-react with proinsulin or C peptide (cross-reactivities, <0.1%). An insulin sensitivity index was calculated according to the formula of Cederholm and Wibell (27).
A paired t test was used to compare means of obese and lean
cotwins. Pearson correlation coefficients were calculated to quantify
the association between intrapair differences in adiposity or its
distribution or variables of glucose metabolism and intrapair
differences in serum TNF or TNFR2 levels. Intraclass correlations
were computed using the double entry method. These correlations provide
for monozygous pairs an estimate of the proportion of variance
explained by genetic and shared environmental effects. Genetic and
obesity-related variations were also assessed as follows. Genetic
variation was defined as the variance of the means of twin pairs minus
intrapair variation divided by 2, and obesity-related variation was
defined as the mean of (intrapair difference squared and divided by 2).
All statistical analyses were performed using SAS statistical programs
(SAS Institute, Inc., Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Obese female
cotwins had approximately 2-fold more both sc and visceral abdominal
fat compared with lean cotwins, whereas among males the relative
difference in adiposity within the pairs was greater in the visceral
compartment compared with the sc fat depot. During the oral glucose
tolerance test (OGTT), the intrapair differences in the area under the
curve (AUC) for glucose and insulin and the insulin sensitivity index
were greater in male than in female twin pairs (Table 1).
|
concentrations than women
(Table 2). The serum TNF concentration
was higher in obese cotwins than in lean cotwins, but the difference
was significant only in women; the variation was very large among obese
male cotwins (Table 2). The serum soluble TNFR2 concentration was also
higher in obese cotwins, but only among women (Table 2 and Fig. 1).
|
|
and TNFR2 levels. The
intrapair differences in serum TNF concentration did not correlate
with intrapair differences in indexes of obesity (Table 3). However, intrapair differences in
serum soluble TNFR2 concentration correlated positively with intrapair
differences in BMI, percent body fat, and abdominal sc fat area, but
not with differences in visceral fat area. When genders were studied
separately, intrapair differences in TNF in men correlated weakly
with differences in percent body fat (r = 0.65; P
= 0.056) but not with other indexes of obesity. In women, intrapair
differences in TNFR2 correlated positively with differences in BMI
(r = 0.79; P < 0.001), percent body fat (r =
0.54; P = 0.048), and abdominal visceral fat area
(r = 0.61; P = 0.033) and tended to correlate with
differences in abdominal sc fat area (r = 0.49; P
= 0.10).
|
). During the OGTT, intrapair
differences in AUC for glucose or insulin or differences in insulin
sensitivity index did not correlate with differences in serum TNF or
TNFR2 concentrations. When genders were studied separately, intrapair
differences in TNF in men did not correlate with indexes of glucose
metabolism. In women, intrapair differences in TNFR2 correlated
positively with differences in AUC for glucose (r = 0.77;
P = 0.001), but not with differences in AUC for insulin
or insulin sensitivity index.
To assess the relative contributions of genetic and environmental
(i.e. obesity) factors to circulating TNF and TNFR2
levels, we compared intrapair and interpair variations in these
variables (Table 4). The variation in
TNF level between the pairs was 1.8-fold higher than the variation
within the pairs. The variation in soluble TNF receptor level
between the pairs was 5.8-fold higher than the variation within the
pairs. The intraclass correlation coefficients for both TNF and
TNFR2 indicated that approximately two thirds of their variation was
due to genetic factors or environmental factors shared by the
twins.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
concentration was only modestly higher in the obese
cotwins, and the difference was evident only among women. Moreover, the
intrapair difference in TNF level did not correlate with intrapair
differences in obesity indexes, such as percent body fat and size of
abdominal sc and visceral fat depots. This suggests that genetic
factors or possibly other environmental factors besides obesity are
important in determining circulating TNF levels. In support of
genetic factors, we observed that the genetic variation in serum TNF
concentrations was almost 2-fold greater than the variation due to
obesity and other environmental factors. Our results thus differ from
some earlier studies comparing unrelated obese and lean subjects that
have demonstrated 2- to 3-fold higher TNF levels in obese subjects
(5, 8). In principle, our study population of identical twins
discordant for obesity should be especially sensitive to any
obesity-related differences in TNF levels. One possible explanation
for the discrepant results is the use of different assays for TNF in
various studies. The assay used in the present study measures both free
and bound TNF (total TNF) (28), whereas studies demonstrating
greater obesity-related differences in serum TNF levels have
measured free TNF (5, 8). In support of this possibility, Corica and
co-workers (29), who measured total TNF as we did, reported only
approximately 1.5-fold increased total TNF levels in obese subjects
relative to controls.
Another explanation for the only marginally elevated TNF levels in
our obese cotwins may be the fact that they were not grossly obese. On
the other hand, in the study by Corica et al. (29) measuring
total TNF, the BMI in the obese individuals was 38.4 vs.
21.0 kg/m2 in controls, i.e. their
study subjects expressed a similar or even greater weight difference
compared with that in studies reporting higher free TNF differences
due to obesity. This suggests that the relatively moderate degree of
overweight in our study (on the average, 18-kg difference between obese
and lean cotwins) cannot be the reason for the absence of a marked
difference in serum TNF levels between obese and normal weight
twins. On the other hand, our results clearly show the importance of
genetic factors in determining serum TNF levels, which is in
accordance with studies demonstrating that a common polymorphism in
TNF gene possibly influences fat deposition (16).
TNF signals through two known cell surface receptors, TNFR1 and
TNFR2 (13). Hotamisligil and co-workers showed that the expression of
TNFR2, but not that of TNFR1, is increased 2-fold in the adipose tissue
of obese females relative to controls (14). They also reported a 6-fold
higher concentration of plasma soluble TNFR2 in their obese females
whose BMI was 39.9 vs. 21.4 kg/m2 in
lean controls. In accordance with this we observed a significantly
higher TNFR2 level in obese (BMI, 30 kg/m2)
females compared with their lean identical cotwins (BMI, 22.4
kg/m2). However, the difference between obese and
lean twins in serum TNFR2 level was much smaller in our study, which
may at least partly be explained by the smaller BMI difference in our
study compared with that in the work by Hotamisligil et al.
(14). In favor of this explanation, Fernandez-Real and co-workers
observed only approximately 25% higher TNFR2 levels in their
moderately obese subjects (BMI, 32.5 vs. 22
kg/m2 in lean controls) (15). In our study the
intrapair difference in BMI and percent body fat correlated positively
with the intrapair difference in TNFR2 level, supporting the idea that
the circulating soluble TNFR2 level reflects the expression of TNFR2 in
adipose tissue.
As a new finding we observed that sc rather than visceral fat determines the circulating TNFR2 concentration, as the intrapair difference in abdominal sc fat, but not the difference in visceral fat, correlated positively with the intrapair difference in TNFR2 level. However, our finding does not definitively exclude the possibility that visceral fat is also important in producing TNFR2. Because the visceral fat depot is considerably smaller than that of total sc fat, a theoretically stronger expression of TNFR2 in visceral compared with sc fat tissue would not necessarily be reflected in high plasma TNFR2 levels. Actually, when women were analyzed separately, we observed a weak positive correlation between intrapair differences in TNFR2 and differences in visceral fat area.
Although the previous reports (14, 15) and our study clearly show the importance of the degree of obesity in determining serum soluble TNFR2 levels, our results also point out the modifying role of genetic factors in this association, as the genetic variation in TNFR2 concentration was almost 6-fold compared with the obesity-related variation. Moreover, approximately two thirds of the variation in plasma TNFR2 levels could be attributed to genetic variance and/or environmental factors shared by the monozygous twins. However, the latter factors are unlikely to be important, as these twin pairs had lived apart at least for 20 yr. Due to the relatively small number of twin pairs in our study, the role of random variation has to be taken into account when interpreting data on the role of genetic factors in determining TNFR2 levels. Thus, it is possible that we may have slightly overestimated the genetic contribution to intrapair variation in TNFR2 levels. Further studies are needed to examine possible polymorphisms in the TNFR2 gene that could form the background for our observation on genetic variation in TNFR2 levels.
In addition to higher circulating TNFR2 levels, obese cotwins expressed disturbed glucose metabolism, indicating decreased insulin sensitivity relative to that in lean cotwins. However, the intrapair differences in TNFR2 levels did not show any correlation with intrapair differences in AUC for glucose or insulin or in insulin sensitivity index assessed with OGTT. Many previous studies (30, 31) and our twin study based on the same twin pairs (19) have earlier demonstrated that especially visceral fat is important for the decreased insulin sensitivity in obesity. Therefore, our finding that sc rather than visceral fat is a major determinant of soluble serum TNFR2 levels explains well our observation that the serum TNFR2 concentration is not correlated with indexes of insulin sensitivity. The facts that serum TNFR2 was elevated only in obese female, not in male, cotwins and yet obese male cotwins expressed similar or even more reduced insulin sensitivity compared with obese female cotwins also favor the idea that circulating soluble TNFR2 level is not an indicator of insulin resistance. Previously, Hotamisligil and co-workers (14) reported a strong positive correlation between adipose tissue expression of TNFR2 and fasting plasma insulin. Unfortunately, they did not report the correlation between circulating TNFR2 and insulin levels. Fernandez-Real et al. (15) did not find any correlation between plasma TNFR2 concentration and percent body fat or fat mass, but despite this they observed a weak, but significant, inverse correlation (r = -0.38; P = 0.02) between TNFR2 level and insulin sensitivity index.
Overall, the biological functions of soluble TNFR2 are poorly
understood in conditions without acute diseases. It has been proposed
that its increased plasma levels in obesity might serve to limit the
actions of TNF to tissues where it is produced and to prevent
endocrine functions of this cytokine (14). As TNF is an inducer of
TNFR2 (32), and measurement of TNF itself in plasma has been
complicated, Hotamisligil et al. (14) presented the
possibility that soluble TNFR2 might serve as a diagnostic marker for
obese individuals in whom there is TNF-related insulin resistance. Our
results, based on identical twins discordant for obesity, suggest that
although the soluble TNFR2 concentration is strongly related to the
degree of adiposity, its measurement is not useful in identifying obese
individuals with reduced insulin sensitivity.
Received November 9, 1999.
Revised February 25, 2000.
Accepted March 7, 2000.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
primary mediator of the host response. Annu Rev
Immunol. 7:625–655.[Medline]
: direct role in
obesity-linked insulin resistance. Science. 259:87–91. in human obesity and insulin resistance. J
Clin Invest. 95:2409–2415.
in sera of obese
patients: fall with weight loss. J Clin Endocrinol Metab. 83:2907–2910. by human muscle: relationship to insulin
resistance. J Clin Invest. 97:1111–1116.[Medline]
inhibits signaling from the insulin
receptor. Proc Natl Acad Sci USA. 91:4854–4858. are increased in obese
patients with non-insulin-dependent diabetes mellitus. J Clin
Endocrinol Metab. 83:859–862. and
insulin sensitivity in elderly men with non-insulin-dependent diabetes
mellitus. Arterioscler Thromb Vasc Biol. 18:1199–1202.-induced insulin resistance in vivo and
its prevention by troglitazone. Diabetes. 46:1678–1683.[Abstract]
does not affect insulin-stimulated glucose uptake in skeletal muscle. Diabetes. 47:721–726.[Abstract]
antibody (CDP571) on
insulin sensitivity and glycemic control in patients with NIDDM. Diabetes. 45:881–885.[Abstract]
gene Nco I polymorphism influences the relationship
among insulin resistance, percent body fat, and increased serum leptin
levels. Diabetes. 46:1468–1472.[Abstract]
gene promoter polymorphism and
decreased insulin resistance. Diabetologia. 41:430–434.[CrossRef][Medline]
receptors in immunological
detection of tumor necrosis factor-. Clin Chem. 42:1450–1453. system in obese subjects. Int J Obesity. 23:355–360.[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  B. J. Nicklas, T. You, and M. Pahor Behavioural treatments for chronic systemic inflammation: effects of dietary weight loss and exercise training Can. Med. Assoc. J., April 26, 2005; 172(9): 1199 - 1209. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Fernandez-Real and W. Ricart Insulin Resistance and Chronic Cardiovascular Inflammatory Syndrome Endocr. Rev., June 1, 2003; 24(3): 278 - 301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Krakoff, T. Funahashi, C. D.A. Stehouwer, C. G. Schalkwijk, S. Tanaka, Y. Matsuzawa, S. Kobes, P. A. Tataranni, R. L. Hanson, W. C. Knowler, et al. Inflammatory Markers, Adiponectin, and Risk of Type 2 Diabetes in the Pima Indian Diabetes Care, June 1, 2003; 26(6): 1745 - 1751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Straczkowski, I. Kowalska, A. Stepien, S. Dzienis-Straczkowska, M. Szelachowska, and I. Kinalska Increased Plasma-Soluble Tumor Necrosis Factor-{alpha} Receptor 2 Level in Lean Nondiabetic Offspring of Type 2 Diabetic Subjects Diabetes Care, October 1, 2002; 25(10): 1824 - 1828. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Peral, J. L. San Millan, R. Castello, P. Moghetti, and H. F. Escobar-Morreale The Methionine 196 Arginine Polymorphism in Exon 6 of the TNF Receptor 2 Gene (TNFRSF1B) Is Associated with the Polycystic Ovary Syndrome and Hyperandrogenism J. Clin. Endocrinol. Metab., August 1, 2002; 87(8): 3977 - 3983. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
