SUMMARY

1. Obesity is known to increase the risk for atheroscleroticdiseases. Serum levels of cellular adhesion molecules arereported to be indices of atherosclerosis. The aim of the presentstudy was to assess the effect of weight reduction on soluble inter-cellular adhesion molecule-1 (sICAM-1) and soluble E-selectin(sE-selectin).

2. Eighteen non-diabetic normotensive obese women partici-pated in a 3 month lifestyle-modification programme (inter-vention group). The programme consisted of lectures on diet, exercise sessions and behavioural modification by weightcharting. Fourteen women who did not enter the programmeserved as controls. Body fat mass (FM) was measured by dual-energy X-ray absorptiometry. Soluble ICAM-1 and sE-selectinwere measured by ELISA.

3. After 3 months, sICAM-1 and sE-selectin, as well as bodyFM, decreased in the intervention group (P < 0.001), while nochanges were observed in the control group. The baseline sE-selectin was positively correlated with total body FM, trunkFM and percentage body fat (r = 0.50–0.57; P < 0.01), but notwith leg FM. The change in sE-selectin was also correlated withchanges in total body FM and trunk FM (both r = 0.46; P < 0.01).Baseline sICAM-1 was not significantly correlated with thesevariables. The associations between changes in sICAM-1 andchanges in total body FM or trunk FM were of borderline significance (both r = 0.34; P = 0.06). Linear regression analysisindicated that the change in sE-selectin was explained by thechange in trunk FM (R2 = 0.18; P < 0.01).

4. Soluble ICAM-1 and sE-selectin were positively correlatedwith obesity, especially with central obesity. Weight reductionresulted in decreases in soluble adhesion molecules, which maysuggest a downregulation of endothelial activation.

Key words: atherosclerosis, central obesity, diet, E-selectin,exercise, intercellular adhesion molecule-1.

INTRODUCTION

Obesity is considered to be one of the risk factors for atheroscleroticdiseases,1–3 but the underlying mechanisms are still not completelyunderstood. During the early stages of atherosclerosis, circulatingleucocytes bind to and roll on the surface of vascular endothelialcells and then transmigrate to the subendothelial matrix. These pro-cesses are mediated by cellular adhesion molecules (CAM) that areexpressed on the surface of vascular endothelial cells, vascularsmooth muscle cells and leucocytes.4–6 The endothelial CAM areconsidered to be shed into the systemic circulation in a soluble form7

and it has been proposed that the levels of the soluble CAM, such as soluble intercellular adhesion molecule-1 (sICAM-1) andsoluble E-selectin (sE-selectin), reflect endothelial activation.8,9

Several studies have demonstrated that soluble CAM can be usedas indices of atherosclerosis10–12 or even predict future events ofcoronary heart disease.13

The associations between obesity and soluble CAM are contro-versial. Some studies have shown positive correlations between bodymass index (BMI) and circulating levels of CAM,10,14 whereas others have not found any significant associations.15–17 Ferri et al.further demonstrated that the soluble CAM are decreased afterweight reduction in obese males.14 No studies have examined theeffect of weight reduction on levels of soluble CAM in women.Furthermore, BMI may not be a reliable index of adiposity18,19 andit is preferable to evaluate obesity by applying methods that measure the fat mass more directly, such as dual-energy X-rayabsorptiometry (DXA). In the present study, we examined the effectof weight reduction by lifestyle modification on serum levels ofsICAM-1 and sE-selectin in obese Japanese women.

METHODS

Subjects

The subjects in the present study were among participants who attended a3 month lifestyle modification programme performed at Fukuoka HealthPromotion Center. The programme was aimed at improving dietary patternsand increasing physical activity in order to prevent and manage health prob-lems such as obesity, hypertension, dyslipidaemia and diabetes mellitus.However, these states were not necessarily a requirement for entering theprogramme. To examine the effect of weight reduction, the present studyanalysed non-diabetic and normotensive obese women (aged 19–68 years)who fulfilled the following criteria: (i) BMI � 25.0 kg/m2; (ii) fasting bloodglucose < 7.0 mmol/L and glycosylated haemoglobin (HbA1c) < 6.5%; (iii) systolic blood pressure < 140 mmHg and diastolic blood pressure < 90 mmHg; (iv) non-smokers; and (v) not taking medications known toaffect glucose or lipid metabolism or cardiopulmonary function. Among the

WEIGHT REDUCTION DECREASES SOLUBLE CELLULAR ADHESION MOLECULES IN OBESE WOMEN

Hiroyuki Ito,* Akiko Ohshima,* Misako Inoue,* Naoko Ohto,* Kazuta Nakasuga,† Yoshikazu Kaji,†

Toru Maruyama‡ and Kazuo Nishioka*

*Fukuoka Health Promotion Foundation, †Department of Medicine and Biosystemic Science and ‡Institute of Health Science, Kyushu University Graduate School of Medical Sciences, Fukuoka, Japan

Correspondence: Dr H Ito, Department of Medicine and BiosystemicScience, Kyushu University Graduate School of Medical Sciences, Maidashi3-1-1, Higashi-ku, Fukuoka 812-8582, Japan. Email: h_ito@intmed1.med.kyushu-u.ac.jp

Received 17 July 2001; revision 17 September 2001; accepted 11 November 2001.

Clinical and Experimental Pharmacology and Physiology (2002) 29, 399–404

400 H Ito et al.

66 participants who initially entered the programme, 20 fulfilled these criteria and two of them dropped out during the 3 month programme. Thus,the remaining 18 subjects served as the intervention group. Subjects whoonly undertook medical examinations but did not participate in the pro-gramme were separately recruited and served as controls. Among the 53 subjects who were initially recruited, 14 subjects fulfilled the above criteria. They were instructed not to change their lifestyle between the twoexaminations. The controls also received instructions on weight control and exercise in a simplified form after the second examination. None of thesubjects was on hormone-replacement therapy. Written informed consent wasobtained from each subject.

Medical examinations

Medical examinations at baseline and after 3 months were performed on themornings after subjects had fasted overnight. A medical history was takenand a physical examination was conducted. Venous blood samples wereobtained between 09.00 and 10.00 h. Height, bodyweight, waist circum-ference and hip circumference were measured20 and BMI was calculated by dividing the bodyweight (in kg) by the square of height (in m). Bloodpressure was measured with subjects in the sitting position by mercury sphygmomanometer after subjects had been quietly seated for more than 5 min. Blood pressure measurements were performed three times at an interval of at least 10 min and an average value was obtained for each subject. Body composition was measured by DXA (QDR-2000; Hologic,Waltham, MA, USA) and analysed with the software (version 7.20).20 Fatmass, lean mass and bone mineral content of the whole body and specificanatomical regions (trunk and legs) were obtained. Percentage body fat ofthe whole body was calculated by dividing the fat mass by the sum of fatmass, lean mass and bone mineral content.

Laboratory methods

Serum levels of total cholesterol, high-density lipoprotein (HDL)–choles-terol and triglycerides were assayed by enzymatic techniques (Wako, Osaka,Japan). Low-density lipoprotein (LDL)–cholesterol was calculated by theFriedewald formula.21 Serum glucose concentration was measured by theglucose oxidase method (Wako). Insulin was measured by radioimmuno-

assay (Pharmacia-Upjohn, Uppsala, Sweden) and insulin resistance estimatedby the homeostasis model assessment (HOMAIR) was calculated by the following formula:22

Insulin resistance = (serum glucose (mmol/L)) � (serum insulin(mU/L))/22.5

Serum sICAM-1 and sE-selectin concentrations were measured in samples stored at –80°C. Sera were diluted and measured in duplicate withan enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis,MN, USA) according to the manufacturer’s instructions. The measurementwas performed by one investigator (HI) who was unaware of the identity of the subjects. The minimum detectable concentrations of sICAM-1 and sE-selectin in the assay were 0.35 and 0.1 ng/mL, respectively. The validity of the assay was confirmed by simultaneously measuring controlsamples with known concentrations provided by the manufacturer. The measured concentrations of the control samples were all within the expectedrange. The mean coefficients of variation were < 5.0% for both sICAM-1and sE-selectin assays.

Lifestyle-modification programme

The 3 month programme consisted of lectures on diet, exercise sessions andbehavioural modification by a recording weight chart23 and keeping a diary on diet and exercise. Weight charting required subjects to record theirbodyweight four times a day (i.e. immediately after waking, after breakfast,before the evening meal and before going to bed). It has been reported thatmonitoring one’s weight facilitates participants to recall, realise and improvetheir eating behaviour.23 The participants visited the Center once a week andattended a 2 h session, including lectures on diet and exercise session (seebelow). On their visits, the weight charts and the food and exercise diarieswere reviewed by registered dietitians and public health nurses and the problems to overcome were discussed between participants and staff members. The most emphasised topics regarding diet included calorie restric-tion, especially by reducing fat intake, increasing the consumption of vege-tables, legumes and grains, such as rice, and substitution of saturated fatswith unsaturated fats.24

Participants were encouraged to increase physical activity in their dailylife by walking, preferably more than 10 000 steps per day. The type and

Table 1 Subject characteristics

Intervention (n = 18) Control (n = 14)Baseline 3 months Baseline 3 months

Age (years) 44.2 � 13.0 – 51.6 � 16.2 –Bodyweight (kg) 67.0 � 6.0 62.0 � 6.1*** 64.8 � 6.1 64.3 � 6.4BMI (kg/m2) 27.2 � 1.9 25.3 � 2.1*** 27.4 � 2.3 27.2 � 2.4Waist circumference (cm) 80.5 � 6.2 75.1 � 6.0*** 80.9 � 7.6 80.8 � 9.1Hip circumference (cm) 97.6 � 3.8 93.7 � 4.5*** 97.2 � 5.4 96.8 � 5.7Waist–hip ratio 0.83 � 0.06 0.80 � 0.05*** 0.83 � 0.08 0.84 � 0.09Total body fat mass (kg) 26.5 � 4.2 22.6 � 4.4*** 26.9 � 3.9 26.4 � 4.3Percentage body fat (%) 40.0 � 3.9 36.9 � 4.3** 41.7 � 3.4 41.2 � 4.2Trunk fat mass (kg) 13.5 � 2.5 11.1 � 2.7*** 13.9 � 2.7 13.7 � 3.2Leg fat mass (kg) 9.1 � 1.5 7.9 � 1.6*** 9.1 � 1.8 8.8 � 1.7Trunk fat/Leg fat 1.49 � 0.22 1.41 � 0.28* 1.59 � 0.47 1.61 � 0.50SBP (mmHg) 123.6 � 9.9 122.9 � 11.0 123.7 � 9.8 125.6 � 18.0DBP (mmHg) 76.6 � 7.0 75.4 � 7.3 75.8 � 7.1 79.6 � 10.1Total cholesterol (mg/dL) 213.7 � 40.9 183.9 � 30.9*** 225.8 � 35.0 213.6 � 32.0Triglycerides (mg/dL) 115.4 � 99.9 83.9 � 62.4 99.9 � 35.8 108.2 � 34.5HDL–cholesterol (mg/dL) 63.9 � 13.7 60.6 � 12.2* 66.0 � 15.0 63.8 � 12.9LDL–cholesterol (mg/dL) 126.8 � 38.8 106.5 � 26.4** 139.9 � 28.5 128.1 � 29.5Glucose (mmol/L) 5.5 � 0.6 5.5 � 0.5 5.6 � 0.6 5.4 � 0.5Insulin (mU/L) 10.6 � 5.6 7.9 � 3.1* 8.0 � 2.2 7.3 � 2.7HOMAIR 2.60 � 1.44 1.96 � 0.89* 2.01 � 0.70 1.76 � 0.61

Values are mean�SD. *P < 0.05, **P < 0.01, ***P < 0.001 between baseline and after 3 months within the study group (paired t-test). For the controlgroup, there were no significant differences seen for any of the variables between baseline and after 3 months.

BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL–cholesterol, high-density lipoprotein–cholesterol; LDL–cholesterol, low-density lipoprotein–cholesterol; HOMAIR, insulin resistance estimated by the homeostasis model assessment.

Adhesion molecules and weight reduction 401

duration of each exercise performed and the number of steps they walkedthroughout the day, which was counted by a pedometer, were recorded eachday. Whenever possible, exercise with a target heart rate of 40–60% of maximum heart rate was recommended. On their weekly visit to the Center,the subjects attended a 1 h exercise session consisting of either walking ingroups, mild aerobic dance or exercise using a cycle ergometer.

Statistical analyses

Statistical analyses were performed by using SPSS 10.0 J for Windows (SPSSJapan, Tokyo, Japan). Values of triglycerides were transformed with a natural logarithm to improve normality before further statistical analysis.Continuous and categorical variables between the intervention and controlgroups were compared with unpaired t-test and Fisher’s exact test, resec-tively. Mean values between baseline and after 3 months within each groupwere compared with paired t-test. Correlation analyses were performed bycalculating the monovariate (Pearson) or partial correlation coefficients. Inthe multiple linear regression analysis, a backward stepwise elimination procedure was applied after including all the possible relevant variables into the model. Two-tailed P values of < 0.05 were considered significant.Data are given as the mean�SD.

RESULTS

The characteristics of the intervention group and control group atbaseline and after 3 months are shown in Table 1. There were nosignificant differences in any of the variables at baseline betweenthe two groups (P > 0.10). The number of postmenopausal subjects

in the intervention and control was six (33%) and eight (57%),respectively (P = 0.16). After 3 months, a decrease in adiposity, total cholesterol, LDL–cholesterol, HDL–cholesterol, insulin andHOMAIR were seen for the intervention group, whereas no suchchanges were seen in the control group. Subjects were asked howmany days each week they exercised for more than 30 min and thosewho exercised less than 2 days were defined as sedentary. At base-line, the number of sedentary subjects in the intervention and con-trol groups was 17 (94%) and 11 (79%), respectively (P = 0.31).After 3 months, only three subjects (17%) were sedentary in theintervention group, whereas 12 subjects (86%) were sedentary inthe control group (P < 0.001). Serum concentrations of sICAM-1 andsE-selectin at baseline and after 3 months are presented in Fig. 1.At baseline, there was no significant difference in either sICAM-1or sE-selectin between the two groups. After the programme, bothlevels of sICAM-1 and sE-selectin decreased significantly in theintervention group, whereas there was no change in the controlgroup.

Next, we conducted correlation analyses between CAM and fatmass in a combined dataset of the two groups. For baseline values,sE-selectin was positively correlated with total body fat mass

Fig. 1 Serum concentrations of (a) soluble intercellular adhesion molecule-1 (sICAM-1) and (b) soluble E-selectin (sE-selectin) at baseline(�) and after 3 months (�) for the intervention and control groups.

Fig. 2 Association between trunk fat mass and (a) soluble intercellularadhesion molecule-1 (sICAM-1; r = 0.20; P = 0.28) and (b) soluble E-selectin(sE-selectin; r = 0.50; P = 0.003) at baseline. (�), intervention group; (�),control group.

402 H Ito et al.

(r = 0.51; P = 0.003), trunk fat mass (r = 0.50; P = 0.003; Fig. 2) and percentage body fat (r = 0.57; P < 0.001), whereas it was not correlated with leg fat mass (P > 0.10). Baseline sICAM-1 was notsignificantly correlated with these measures on adiposity (P > 0.10;Fig. 2). The anthropometric variables were largely not associatedwith soluble CAM (P > 0.10), expect for the relationship betweenBMI and sE-selectin (r = 0.34; P = 0.05). Age was not significantly

correlated with these CAM (P > 0.10) and essentially the same resultsas in the monovariate analyses were obtained when partial correl-ation coefficients were calculated for the above variables after adjust-ing for age and/or menopause (data not shown).

Associations between changes in CAM and changes in the measures on adiposity were examined to obtain further insight into the effect of weight reduction on CAM (Table 2). Changes invariables were calculated by subtracting the baseline values fromthe values at 3 months. For both CAM, correlation coefficients withthe anthropometric variables tended to be higher than those with the variables obtained by DXA. The change in sE-selectin was positively associated with changes in total body fat mass and trunkfat mass (Fig. 3), but not with the change in leg fat mass, which wasin accordance with the correlation analyses for the baseline values.For the change in sICAM-1, correlations of a borderline significancewere demonstrated with changes in total body fat mass (P = 0.06)and trunk fat mass (P = 0.06; Fig. 3), but not with the change in leg fat mass. The associations between changes in soluble CAM and the changes in variables for blood pressure, lipids and insulin resistance were also analysed. Changes in sE-selectinwere correlated with changes in systolic (r = 0.35; P = 0.05) and diastolic (r = 0.38; P = 0.03) blood pressure and triglycerides (r = 0.37; P = 0.04), whereas significant associations were not seen for changes in total cholesterol and HOMAIR. The change in sICAM-1 was not correlated with any of these variables.

Linear regression analysis was performed with the change in sE-selectin as the dependent variable and changes in total body fatmass, trunk fat mass, blood pressure and triglycerides as indepen-dent variables. After a backward stepwise elimination procedure,only the change in trunk fat mass (� = 2.34, standardized � = 0.46;P = 0.008) remained as a significant variable with an adjusted R2 of0.18 (P = 0.008).

DISCUSSION

The present study demonstrated that weight reduction, obtained byan improvement in diet habits and an increase in physical activity,resulted in decreases in soluble CAM. The reductions in sICAM-1and sE-selectin were associated with a reduction in body fat mass,especially with a reduction in trunk fat mass, indicating the involve-ment of central obesity in the regulation of these CAM. The associ-ation between CAM and fat mass has only been evaluated indirectlyin previous studies by examining the relationship between CAM andBMI.10,14–17

In addition to changes in fat mass, changes in sE-selectin werepositively correlated with blood pressure and triglycerides, sugges-ting that improvement in the levels of these variables contributedto the reduction in sE-selectin in the intervention group. This is inline with results from previous studies showing an associationbetween sE-selectin and hypertension10,14,16,17,25,26 or hypertriglycer-idaemia.10,15,17,26–28 Results of multiple linear regression analysis indi-cated that improvement in central obesity was the only significantfactor responsible for the decrease in sE-selectin for subjects in thepresent study. In contrast, the change in sICAM-1 only showed posi-tive correlations of borderline significance with total body fat massand trunk fat mass and was not associated with other variables. Takentogether, these results indicate that both sE-selectin and sICAM-1are associated with obesity, with the former showing a strongerassociation than the latter. E-Selectin is considered to be expressed

Table 2 Monovariate correlation between changes in soluble adhesionmolecules and measures of adiposity

�[sICAM-1] �[sE-selectin]

�Body mass index 0.44* 0.50**�Waist circumference 0.54** 0.49**�Hip circumference 0.48** 0.32�Total body fat mass 0.34 0.46**�Trunk fat mass 0.34 0.46**�Leg fat mass 0.19 0.29

*P < 0.05, **P < 0.01.sICAM-1, soluble intercellular adhesion molecule-1; sE-selectin,

soluble E-selectin.

Fig. 3 Association between changes in trunk fat mass and (a) soluble inter-cellular adhesion molecule-1 (sICAM-1; r = 0.34, P = 0.06) and (b) solubleE-selectin (sE-selectin; r = 0.46; P = 0.008). (�), intervention group; (�),control group.

Adhesion molecules and weight reduction 403

exclusively in vascular endothelial cells, whereas ICAM-1 isexpressed in various types of cells, including vascular endothelialcells, vascular smooth muscle cells and leucocytes.7 Such differencesin the expression sites of these CAM may explain, in part, theirslightly different relationship with obesity.

The mechanism underlying the association between CAM andobesity is not clear. Regulation of CAM is under the influence ofoxidative stress,15,29 nitric oxide availability,30 the renin–angiotensinsystem25 and insulin sensitivity.31 It is possible that an improvementin dietary habits and/or an increase in physical activity modified atleast one of these systems.24,32–35 Considering the well-knownrelationship between central obesity and insulin resistance,36,37 it istemping to speculate that improvement in insulin resistance under-lies the association between the changes in trunk fat and CAM.Although insulin resistance estimated by HOMAIR improved afterweight reduction, the changes in soluble CAM and HOMAIR didnot show a significant correlation and the role of insulin resistanceon the regulation of CAM remained unclear in the present study.Recent studies have indicated that various proteins, such as tumournecrosis factor (TNF)-� and adiponectin, are expressed in adiposetissue and regulate the expression of CAM.8,9,38–41 It has also beenshown that weight reduction modulates the expression of TNF-�38

and adiponectin,40 which may have contributed to the decreases inCAM after the intervention in the present study.

The present study has several limitations. First, we studied onlyfemale subjects and the results obtained in this study may not betrue for males. In addition, the subjects included both pre- and post-menopausal subjects, with a relatively wide range in the age of thesubjects. Several studies have indicated that sex hormones modu-late the levels of soluble CAM42–44 and it is possible that menstrualstatus affected the results in the present study. However, age wasnot associated with the levels of CAM and the association betweenCAM and obesity was not affected when adjustments for age and/ormenopause were made. Thus, we believe that age and menopausewere not the major confounding factors in the present study. Second,the design of the present study cannot separate the effects of animprovement in diet habits or an increase in physical activity onCAM. Both these factors are known to affect the levels of CAM,possibly through various mechanisms.24,32–35 Weight reduction by alow-calorie diet has been reported to decrease sICAM-1 and sE-selectin.14 In a cross-sectional study, the rates of physically active subjects were higher in groups with the lower levels ofsICAM-1.17 Thus, we believe that both diet and physical activitysynergistically affected the levels of soluble CAM. The associationbetween changes in the fat mass and CAM at least indicates thatthe negative energy balance contributed to the decrease in CAM after the intervention. Third, randomization was not conducted forselecting subjects in the two groups, which may have influenced theresults in the present study. Finally, the origin and pathophysiologicalfunction of soluble CAM are still under debate. It was consideredinitially that the soluble forms of CAM shed into circulation by proteolysis of the membrane-bound CAM and the circulating levels of the soluble forms reflected the levels of expression of themembrane-bound forms.8,9 However, there is evidence suggestingthat the soluble forms are splice variants that are produced inde-pendently of the membrane-bound forms.45 Another study has suggested that soluble CAM may act as neutralizing agents of themembrane-bound forms.46 In spite of these controversies, it is likelythat soluble CAM are associated with atherosclerosis, which is

supported by recent studies showing that the soluble CAM can be regarded as indices of the presence of atherosclerotic lesions10–12

and predict future cardiovascular events.13

In conclusion, the present study has shown that soluble CAM,namely sICAM-1 and sE-selectin, are positively correlated with obesity, especially with central obesity, and their levels decrease following weight reduction, suggesting a downregulation of endo-thelial activation after a reduction in fat mass. Whether the decreasein soluble CAM after weight reduction contributes to a reductionin cardiovascular disease remains to be answered in future studies.

ACKNOWLEDGEMENTS

The authors are grateful to Chiga Hijii, Kaoru Takao, MamiYanagawa, Tadashi Kawaida, Yukie Urayama and other staff of theFukuoka Health Promotion Center for their contribution to this project.

REFERENCES

1. Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as anindependent risk factor for cardiovascular disease: A 26-year follow-up of participants in the Framingham Heart Study. Circulation 1983;67: 968–77.

2. Stevens J, Cai J, Pamuk ER, Williamson DF, Thun MJ, Wood JL. The effect of age on the association between body-mass index and mortality. N. Engl. J. Med. 1998; 338: 1–7.

3. Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH. Thedisease burden associated with overweight and obesity. JAMA 1999;282: 1523–9.

4. Duplàa C, Couffinhal T, Labat L et al. Monocyte/macrophage recruit-ment and expression of endothelial adhesion proteins in humanatherosclerotic lesions. Atherosclerosis 1996; 121: 253–66.

5. Holvoet P, Collen D. Thrombosis and atherosclerosis. Curr. Opin.Lipidol. 1997; 8: 320–8.

6. Ross R. Atherosclerosis: An inflammatory disease. N. Engl. J. Med.1999; 340: 115–26.

7. Gearing AJ, Newman W. Circulating adhesion molecules in disease.Immunol. Today 1993; 14: 506–12.

8. Leeuwenberg JFM, Smeets EF, Neefjes JJ et al. E-Selectin and inter-cellular adhesion molecule-1 are released by activated human endo-thelial cells in vitro. Immunology 1992; 77: 543–9.

9. Fassbender K, Kaptur S, Becker P, Gröschl J, Hennerici M. Adhesionmolecules in tissue injury. Kinetics of expression and shedding and association with cytokine release in humans. Clin. Immunol.Immunopathol. 1998; 89: 54–60.

10. Hwang SJ, Ballantyne CM, Sharrett AR et al. Circulating adhesionmolecules VCAM-1, ICAM-1, and E-selectin in carotid atherosclerosisand incident coronary heart disease cases: The Atherosclerosis Risk InCommunities (ARIC) study. Circulation 1997; 96: 4219–25.

11. Blann AD, Seigneur M, Steiner M, Miller JP, McCollum CN. CirculatingICAM-1 and VCAM-1 in peripheral artery disease and hypercholes-terolaemia: Relationship to the location of atherosclerotic disease, smoking, and in the prediction of adverse events. Thromb. Haemost.1998; 79: 1080–5.

12. Rohde LE, Lee RT, Rivero J et al. Circulating cell adhesion moleculesare correlated with ultrasound-based assessment of carotid athero-sclerosis. Arterioscler. Thromb. Vasc. Biol. 1998; 18: 1765–70.

13. Ridker PM, Hennekens CH, Roitman-Johnson B, Stampfer MJ, AllenJ. Plasma concentration of soluble intercellular adhesion molecule 1 andrisks of future myocardial infarction in apparently healthy men. Lancet1998; 351: 88–92.

14. Ferri C, Desideri G, Valenti M et al. Early upregulation of endothelialadhesion molecules in obese hypertensive men. Hypertension 1999; 34:568–73.

404 H Ito et al.

15. Ceriello A, Falleti E, Bortolotti N et al. Increased circulating intercellularadhesion molecule-1 levels in type II diabetic patients: The possible roleof metabolic control and oxidative stress. Metabolism 1996; 45:498–501.

16. DeSouza CA, Dengel DR, Macko RF, Cox K, Seals DR. Elevated levels of circulating cell adhesion molecules in uncomplicated essen-tial hypertension. Am. J. Hypertens. 1997; 10: 1335–41.

17. Rohde LEP, Hennekens CH, Ridker PM. Cross-sectional study of soluble intercellular adhesion molecule-1 and cardiovascular risk factors in apparently healthy men. Arterioscler. Thromb. Vasc. Biol.1999; 19: 1595–9.

18. Smalley KJ, Knerr AN, Kendrick ZV, Colliver JA, Owen OE.Reassessment of body mass indices. Am. J. Clin. Nutr. 1990; 52: 405–8.

19. Curtin F, Morabia A, Pichard C, Slosman DO. Body mass index com-pared to dual-energy X-ray absorptiometry: Evidence for a spectrumbias. J. Clin. Epidemiol. 1997; 50: 837–43.

20. Ito H, Ohshima A, Ohto N et al. Relation between body compositionand age in healthy Japanese subjects. Eur. J. Clin. Nutr. 2001; 55:462–70.

21. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concen-tration of low-density lipoprotein cholesterol in plasma, without use ofthe preparative ultracentrifuge. Clin. Chem. 1972; 18: 499–502.

22. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,Turner RC. Homeostasis model assessment. Insulin resistance and �-cell function from fasting plasma glucose and insulin concentrationsin man. Diabetologia 1985; 28: 412–19.

23. Ito H, Ohshima A, Tsuzuki M et al. Effects of increased physical activity and mild calorie restriction on heart rate variability in obesewomen. Jpn. Heart J. 2001; 42: 459–69.

24. Krauss RM, Deckelbaum RJ, Ernst N et al. Dietary guidelines forhealthy American adults. A statement for health professionals from theNutrition Committee, American Heart Association. Circulation 1996;94: 1795–800.

25. Ferri C, Desideri G, Baldoncini R et al. Early activation of vascularendothelium in nonobese, nondiabetic essential hypertensive patientswith multiple metabolic abnormalities. Diabetes 1998; 47: 660–7.

26. Lupattelli G, Lombardini R, Schillaci G et al. Flow-mediated vaso-activity and circulating adhesion molecules in hypertriglyceridemia:Association with small, dense LDL cholesterol particles. Am. Heart J.2000; 140: 521–6.

27. Bannan S, Mansfield MW, Grant PJ. Soluble vascular cell adhesionmolecule-1 and E-selectin levels in relation to vascular risk factors andto E-selectin genotype in the first degree relatives of NIDDM patientsand in NIDDM patients. Diabetologia 1998; 41: 460–6.

28. Sampietro T, Tuoni M, Ferdeghini M et al. Plasma cholesterol regulatessoluble cell adhesion molecule expression in familial hypercholes-terolemia. Circulation 1997; 96: 1381–5.

29. Cominacini L, Fratta Pasini A, Garbin U et al. E-selectin plasma concentration is influenced by glycaemic control in NIDDM patients:Possible role of oxidative stress. Diabetologia 1997; 40: 584–9.

30. Marfella R, Esposito K, Giunta R et al. Circulating adhesion moleculesin humans: Role of hyperglycemia and hyperinsulinemia. Circulation2000; 101: 2247–51.

31. Chen NG, Holmes M, Reaven GM. Relationship between insulin resistance, soluble adhesion molecules, and mononuclear cell bindingin healthy volunteers. J. Clin. Endocrinol. Metab. 1999; 84: 3485–9.

32. Pate RR, Pratt M, Blair SN et al. Physical activity and public health.A recommendation from the Centers for Disease Control and Preventionand the American College of Sports Medicine. JAMA 1995; 273: 402–7.

33. Böger RH, Bode-Böger SM, Brandes RP et al. Dietary L-argininereduces the progression of atherosclerosis in cholesterol-fed rabbits:Comparison with lovastatin. Circulation 1997; 96: 1282–90.

34. Facchini FS, DoNascimento C, Reaven GM, Yip JW, Ni XP, HumphreysMH. Blood pressure, sodium intake, insulin resistance, and urinarynitrate excretion. Hypertension 1999; 33: 1008–12.

35. Merry BJ. Calorie restriction and age-related oxidative stress. Ann. N.Y.Acad. Sci. 2000; 908: 180–98.

36. Hopkins PN, Hunt SC, Wu LL, Williams GH, Williams RR.Hypertension, dyslipidemia, and insulin resistance: Links in a chain orspokes on a wheel? Curr. Opin. Lipidol. 1996; 7: 241–53.

37. Goodpaster BH, Kelley DE, Wing RR, Meier A, Thaete FL. Effects of weight loss on regional fat distribution and insulin sensitivity in obesity. Diabetes 1999; 48: 839–47.

38. Kern PA, Saghizadeh M, Ong JM, Bosch RJ, Deem R, Simsolo RB.The expression of tumor necrosis factor in human adipose tissue.Regulation by obesity, weight loss, and relationship to lipoprotein lipase.J. Clin. Invest. 1995; 95: 2111–19.

39. Ouchi N, Kihara S, Arita Y et al. Novel modulator for endothelial adhesion molecules: Adipocyte-derived plasma protein adiponectin.Circulation 1999; 100: 2473–6.

40. Hotta K, Funahashi T, Arita Y et al. Plasma concentrations of a novel,adipose-specific protein, adiponectin, in type 2 diabetic patients.Arterioscler. Thromb. Vasc. Biol. 2000; 20: 1595–9.

41. Ito H, Oshima A, Tsuzuki M et al. Association of serum tumor necrosis factor-� with serum LDL–cholesterol and blood pressure inapparently healthy Japanese women. Clin. Exp. Pharmacol. Physiol.2001; 28: 188–92.

42. Scarabin PY, Alhenc-Gelas M, Oger E, Plu-Bureau G. Hormone replace-ment therapy and circulating ICAM-1 in postmenopausal women: A ran-domised controlled trial. Thromb. Haemost. 1999; 81: 673–5.

43. Kennedy G, McLaren M, Belch JJF, Seed M. Elevated levels of sE-selectin in post-menopausal females are decreased by hormonereplacement therapy to levels observed in pre-menopausal females.Thromb. Haemost. 1999; 82: 1433–6.

44. Sbarouni E, Kroupis C, Kyriakides ZS, Koniavitou K, Kremastinos DT.Cell adhesion molecules in relation to simvastatin and hormonereplacement therapy in coronary artery disease. Eur. Heart J. 2000; 21:975–80.

45. Komatsu S, Flores S, Gerritsen ME, Anderson DC, Granger DN.Differential up-regulation of circulating soluble and endothelial cellintercellular adhesion molecule-1 in mice. Am. J. Pathol. 1997; 151:205–14.

46. Meyer DM, Dustin ML, Carron CP. Characterization of intercellularadhesion molecule-1 ectodomain (sICAM-1) as an inhibitor of lymphocyte function-associated molecule–1 interaction with ICAM-1.J. Immunol. 1995; 155: 3578–84.