International Journal of Obesity (2015) 39, 222–227 © 2015 Macmillan Publishers Limited All rights reserved 0307-0565/15 www.nature.com/ijo

ORIGINAL ARTICLE

Regional variations in the relationship between arterial stiffness and adipocyte volume or number in obese subjects P Arner1, J Bäckdahl1, P Hemmingsson2, P Stenvinkel2, D Eriksson-Hogling1, E Näslund3, A Thorell3,4, DP Andersson1, K Caidahl5 and M Rydén1 BACKGROUND: Cardiovascular disease is associated with multiple risk factors including stiff arteries and large adipocytes. Whether the latter two are interrelated is unknown. We aimed to determine whether arterial stiffness is associated with fat cell size and number in subcutaneous or visceral white adipose tissue (WAT). METHODS: A cross-sectional study of 120 obese subjects scheduled for bariatric surgery in whom WAT mass and distribution was assessed by dual-X-ray absorptiometry. Biopsies from visceral (greater omentum) and subcutaneous (abdominal) WAT were obtained to calculate fat cell volume and number. Arterial stiffness was determined as aortic pulse wave velocity (PWV). RESULTS: Visceral adipocyte volume, but not number, was strongly (P o 0.0001) and positively correlated with PWV, explaining 20% of the inter-individual variations in this parameter. This relationship remained significant after correction for clinical confounders. PWV correlated positively (r = 0.38, P o 0.0001) with visceral (but not subcutaneous) WAT mass. Furthermore, PWV was also positively associated with subcutaneous adipocyte volume (r = 0.20, P = 0.031) and negatively with fat cell number (r = − 0.26, P = 0.006). However, the relationships between PWV and visceral WAT mass or subcutaneous fat cell size/number became nonsignificant when controlling for visceral fat cell volume. In a multiple regression analysis to determine the factors that explain variations in PWV, only visceral fat cell volume, age, pulse rate and diastolic blood pressure entered the model, together explaining 42% of the variation in PWV. CONCLUSIONS: Visceral fat cell volume was the only WAT parameter that constituted an independent and significant, positive regressor for arterial stiffness determined by PWV. Although a causal relationship is not established, visceral fat cell volume may explain the well-known correlation between central fat mass, arterial stiffness and cardiovascular risk, at least in severely/morbidly obese subjects. International Journal of Obesity (2015) 39, 222–227; doi:10.1038/ijo.2014.118

INTRODUCTION Obesity, particularly when localized intra-abdominally, is a major risk factor for cardiovascular disease,1 and a number of mechanisms have been suggested to explain how excess white adipose tissue (WAT) may promote the development of atherosclerosis.2 Much less is known about the relationship between obesity and arterial stiffness, an important risk factor and marker for cardiovascular disease.3 Arterial stiffness is defined as the reduced dynamics of arterial vessels in response to changes in blood pressure. It is recognized in some, but not all, studies that arterial stiffness is increased in obese subjects.4 This does not appear to be linked to obesity per se but rather to the regional distribution of the fat. Data from several cross-sectional studies have demonstrated that central, as opposed to peripheral, WAT mass correlates with increased arterial stiffness.5–7 Because WAT distribution is associated with arterial stiffness, factors intrinsic to different adipose tissue depots could be relevant. Fat mass expands by increasing the number and/or size (that is, volume) of its fat cells. For any given amount of WAT, the tissue is composed of either many small fat cells (hyperplasia) or fewer but large fat cells (hypertrophy).8 It is well established that large fat cells are associated with impaired insulin function9 and increased risk of developing type 2 diabetes.10,11 Moreover, 1

differences in fat cell volume in specific WAT depots appear to impact differently on metabolic abnormalities. Thus, WAT hypertrophy in the visceral area is linked to dyslipidaemia, whereas the same phenotype in the subcutaneous depot is linked to insulin resistance.8,12 It is unknown whether, or how, fat cell size/volume and number relate to arterial stiffness. In this study, we examined the relationship between fat cell volume and number in the visceral (that is, greater omentum) and subcutaneous (that is, abdominal) regions versus arterial stiffness. The findings were analysed in relation to other well-known factors influencing arterial stiffness including age, pulse rate, fasting plasma glucose, blood pressure, insulin sensitivity and adipose tissue mass and distribution.13–15 SUBJECTS AND METHODS Subjects The 120 subjects, 21 men and 99 women, participated in an ongoing longitudinal study of adipose tissue function and were recruited among patients scheduled for Roux-en Y gastric by-pass surgery because of obesity (clinical data presented in Table 1). One-hundred eleven subjects had a body mass index (BMI)435 kg m−2. Thirty-three individuals had established hypertension and were receiving anti-hypertensive pharmacotherapy while five were diagnosed with type 2 diabetes and were

Department of Medicine, Karolinska Institutet, Stockholm, Sweden; 2Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden; Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden; 4Department of Surgery, Karolinska Institutet, Ersta Hospital, Stockholm, Sweden and 5Department of Molecular Medicine and Surgery, Karolinska Institutet, Solna, Sweden. Correspondence: Professor M Rydén, Department of Medicine, Karolinska Institutet, Karolinska Huddinge, S-141 86 Stockholm, Sweden. E-mail: [email protected] Received 11 April 2014; revised 16 June 2014; accepted 25 June 2014; accepted article preview online 8 July 2014; advance online publication, 29 July 2014 3

Arterial stiffness and fat cells P Arner et al

223 Table 1.

Clinical characteristics and correlation with arterial stiffness

Variable Gender (male/female) Age, years Body mass index Waist circumference, cm Waist–hip ratio Waist–length ratio Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Resting pulse rate, beats per min PWV, m s − 1 Insulin sensitivity (glucose infusion rate, mg kg − 1 min) P-glucose, mmol l − 1 P-insulin, mU l − 1 S-free fatty acids, mmol l − 1 P-total cholesterol, mmol l − 1 P-HDL, mmol l − 1 P-triglycerides, mmol l − 1 Total body fat, kg Subcutaneous fat mass (sWAT), kg Visceral fat mass (vWAT), kg Subcutaneous fat cell volume, picolitres Visceral fat cell volume, picolitres Subcutaneous fat cell number (× 109 cells) Visceral fat cell number (× 109 cells)

Data available in n individuals

nr or mean ± s.d. (range)

r-value

P-value

120 120 120 120 120 120 120 120 120 120 102 119 119 119 120 120 120 119 107 107 120 119 107 107

21/99 43 ± 10 (18–62) 39.1 ± 3.6 (32.6–57) 122 ± 9.6 (101–152) 0.98 ± 0.07 (0.78–1.15) 0.73 ± 0.05 (0.59–0.90) 136 ± 16 (96–180) 80.2 ± 11 (54–125) 68 ± 10 (45–99) 8.13 ± 1.9 (4.6–18.2) 4.48 ± 1.75 (0–7.8) 5.6 ± 1.3 (4.4–14.2) 15.6 ± 9.4 (2.9–47.6) 0.70 ± 0.25 (0.2–1.4) 4.85 ± 1.0 (2.5–9.4) 1.2 ± 0.29 (0.6–2.0) 1.39 ± 0.65 (0.4–3.6) 53.7 ± 9.5 (36.6–90.1) 3.17 ± 0.9 (0.78–6.2) 2.2 ± 1.0 (0.82–5.3) 878 ± 199 (476–1396) 569 ± 191 (167–1188) 4.11 ±1.2 (1.3–7.9) 4.34 ±1.5 (2.4–9.2)

N/A 0.43 –0.05 0.26 0.31 0.15 0.46 0.48 0.37 N/A − 0.20 0.10 0.15 − 0.04 0.19 − 0.077 0.23 0.037 − 0.13 0.38 0.20 0.45 − 0.26 0.09

0.12 o0.0001 0.60 0.0041 0.0006 0.11 o0.0001 o0.0001 o0.0001 N/A 0.045 0.27 0.10 0.69 0.039 0.40 0.010 0.69 0.19 o0.0001 0.031 o0.0001 0.006 0.36

Abbreviations: HDL, high density lipoprotein; N/A, not applicable; P, fasting plasma; PWV, pulse wave velocity; S, serum; s.d., standard deviation; WAT, white adipose tissue. Values are given as actual numbers of subjects or mean ± s.d. as well as range. Student’s t-test was used for nominal variables. For continuous variables, linear regression analysis was used; r- and P-values are indicated.

treated with metformin (insulin treatment was an exclusion criterion). To avoid possible influences of catabolism/weight loss, none of the participants were prescribed a low-calorie diet regimen preoperatively and subjects who reported intake of a hypo-caloric diet prior to surgery (that is, at their own initiative) were excluded. Subjects were weight stable (less than ± 2 kg weight change) for at least 1 year before the first visit. All subjects came to the laboratory in the morning (0730 hours) after an overnight fast. They were asked (yes/no) whether they were regular users of nicotine (smoking, snuff) or had ceased using either less than a month prior to the visit. Data were available from 117 subjects (93 non-users, 11 smokers, 11 snuff users and 2 who had quit within the previous month). Height, weight, and waist and hip circumferences were measured followed by dual-X-ray absorptiometry for determination of body composition (see below). Pulse rate was determined after 15 min of rest in the supine position, after which arterial stiffness and blood pressure were measured as described below. This was followed by venous blood sampling for routine clinical chemistry. Subsequently, a subcutaneous fat biopsy was obtained from the peri-umbilical region by needle aspiration under local anaesthesia. After 45 min of rest, a hyperinsulinaemic euglycaemic clamp was performed as described below to determine insulin sensitivity. A specimen of adipose tissue was obtained from the greater omentum at the beginning of the surgery, which was performed approximately 2 weeks after the clinical examination described above. This procedure for obtaining biopsies from the two depots was chosen because almost all bariatric surgery in Sweden is performed by laparoscopic technique and it is difficult to obtain sufficient amounts of subcutaneous WAT perioperatively. The study was approved by the regional committee of ethics in Stockholm and was conducted in accordance with the statutes of the Declaration of Helsinki. Oral and written informed consent was obtained from each participant.

Adipose tissue Fat cell volume and weight were determined as described.8 In brief, fat cells were isolated following collagenase digestion of adipose tissue. The diameter of 100 cells was determined. The mean volume of the isolated adipocytes was calculated using a formula described by Hirsch and Gallian16 in 1968. Fat cell volume (expressed in picolitres (pl)) was calculated as (π × d3)/6 where ‘d’ is the cell diameter in μm. © 2015 Macmillan Publishers Limited

Arterial stiffness and blood pressure The measurements were recorded at around 0800 hours with the subjects in the supine position. The recommended procedures for standardization of arterial stiffness measurements were followed.17 The subjects had been instructed not to consume food, nicotine or caffeine after 2200 hours the preceding day. Measurement recordings started after a 15-min period of rest and were performed on the subjects’ right arm. Throughout the study, the same research nurse performed all the measurements in the same quiet room, which was at a constant temperature. An Arteriograph (TensioMed, Budapest, Hungary) was used to determine aortic pulse wave velocity (PWV) values. In brief, PWV was calculated from the travelled distance (measured as the suprasternal notch–pubic bone distance) of the pulse wave in the aorta divided by the measured transit time (RT/2). To avoid overestimating PWV in subjects with larger waist circumferences, the distance between the suprasternal notch (jugulum) and the pubic bone (symphysis) was always measured in a straight line and not on the body surface, according to the manufacturer’s instruction (www.tensiomed.com/ download/manual_arg_bt_en.pdf). The values obtained with this non-invasive and indirect method show a strong correlation with values obtained by direct invasive methods.18 Systolic and diastolic blood pressures were measured with a fully automatic device (Omron M10-IT, Omron Health Care, Hoofddorp, The Netherlands).

Hyperinsulinaemic clamp The participants underwent a hyperinsulinaemic euglycaemic clamp as described.19 In brief, following an intravenous bolus dose of insulin (1600 mU m− 2 body surface area, corresponding to 9600 pmol m −2 when using the conversion factor 6.0 as described on www.soc-bdr.org/rds/ authors/unit_tables_conversions_and_genetic_dictionaries/conversion_in_ si_units/index_en.html), insulin was infused intravenously at 120 mU m−2 min− 1 (720 pmol m −2 min− 1) for 2 h, and a variable intravenous infusion of glucose (200 mg ml − 1) was used to maintain euglycaemia between 4.5 and 5.5 mmol l − 1 (81 and 99 mg dl − 1). The infusion rate of glucose during the last 60 min of the clamp, when insulin levels were in steady state, was used for calculations of whole-body glucose disposal rates (M-values). International Journal of Obesity (2015) 222 – 227

Arterial stiffness and fat cells P Arner et al

224 Dual-X-ray absorptiometry Body fat composition was assessed by dual-X-ray absorptiometry using a GE Lunar iDXA and the software enCORE (version 14.10.022) with the CoreScan feature provided by the manufacturer (GE Medical systems, Chalfont St. Giles, UK). CoreScan is an automated method for segmenting total adipose fat into subcutaneous fat and visceral fat within the android region. The estimation of visceral fat mass with this software has been approved for clinical use by the US Food and Drug Administration.20 Automatic calibration checks of the dual-X-ray absorptiometry were performed daily throughout the study, and three times a week calibrations using a spine phantom (for bone mineral density, provided by the manufacturer) were performed. The coefficient of variation for the spine phantom testing was 1.5%. No hardware or software changes were made during the course of the trial. The subjects were scanned using standard imaging and positioning protocols and the same scan mode (set for obese subjects) was used throughout the study. CoreScan can be used to determine total, abdominal (android), peripheral (gynoid) and estimated visceral adipose tissue mass (vWAT). Determination of vWAT by this method displays a strong correlation (r2⩾ 0.95) with measures of vWAT mass obtained using computed tomography.20 The amount of estimated subcutaneous adipose tissue mass in the android region (sWAT) was calculated from the formula: total adipose fat mass in the android region vWAT = sWAT, as previously described.20 Because only total android fat mass and vWAT are used to determine sWAT and both are valid measures, it follows that also the calculation of sWAT should be valid. The subcutaneous fat biopsy was obtained from the abdominal wall at the same level as the measured sWAT. The number of fat cells in the vWAT and sWAT regions was determined by dividing the estimated tissue masses by the corresponding mean fat cell weights in the two depots.

Statistics Values are presented as mean and standard deviation. They were compared using Student’s t-test, simple or multiple regression analysis (one, two or several independent versus one dependent variable) or analysis of covariance (using one dependent variable versus one nominal and one continuous variable). JMP software (SAS Institute, Inc., Cary, NC, USA) was used for all statistical analyses.

18

18

16

16

Pulse wave velocity (meters/second)

Pulse wave velocity (meters/second)

RESULTS The relationship between arterial stiffness, measured as PWV, and other variables was determined by linear regression analysis and is summarized in Table 1. As expected, PWV correlated positively with age, waist, waist–hip ratio, blood pressure (both systolic and diastolic), pulse rate and vWAT mass. Although there was a weaker relationship with total cholesterol and triglycerides (positive) and with insulin sensitivity (negative), PWV did not correlate significantly with fasting glucose, insulin, high-density lipoprotein cholesterol, BMI, total body fat or sWAT. Both visceral and subcutaneous fat cell volume associated positively with PWV. The relationship was particularly evident for visceral fat cell volume, explaining as much as 20% of the inter-individual variation in PWV (that is, adjusted r2). PWV was not associated with visceral

adipocyte number but was inversely associated with subcutaneous adipocyte number. Our cohort consisted predominantly of women. Although there was no significant difference in PWV (P = 0.12, Table 1), BMI (P = 0.52), total body fat (P = 0.86) or subcutaneous adipocyte volume (P = 0.22) between genders, vWAT mass (Po 0.0001) and visceral fat cells (723 ± 224 versus 536 ± 167 pL, P o 0.0001, graphs not shown) were larger in men. The individual data points for PWV and fat cell volume in the two WAT depots are shown in Figure 1. Values for men and women were distributed in a similar fashion along the regression line. Results of analysis of covariance showed that there was no important sex interaction for the relationship between fat cell volume and PWV (F = 0.11; P = 0.75 for visceral and F = 1.8; P = 0.18 for subcutaneous fat cell volume). Visceral fat cell volume associated with both subcutaneous fat cell volume (r = 0.43, P o 0.0001) and subcutaneous fat cell number (r = − 0.43, Po 0.0001, graphs not shown) but neither of these relationships were influenced by gender (F = 0.29, P = 0.59 and F = 1.25, P = 0.27, respectively). Further analyses were performed to determine whether visceral fat cell volume was an independent regressor for PWV when controlling for the variables in Table 1 that associated significantly with arterial stiffness (Table 2). The relationship between visceral adipocyte volume and PWV remained highly significant after individual adjustment for each single co-variable. When combined with visceral adipocyte volume, neither vWAT mass nor subcutaneous fat cell volume or number remained as independent regressors for PWV. As detailed in the Subjects and Methods section, PWV measurements were performed using a technique that avoids overestimating values in individuals with increased abdominal curvature. This was further corroborated by the fact that the relationship between PWV and visceral fat cell volume remained significant after controlling for waist–hip ratio. Because nicotine can increase arterial stiffness21 as well as alter lipolysis22 and hence possibly fat cell size, the influence of nicotine on the findings in Table 2 was investigated using analysis of covariance. This analysis demonstrated that nicotine had no significant impact (F = 1.8; P = 0.13) on the relationship between visceral fat cell volume and PWV. The correlations between PWV and subcutaneous fat cell volume were, in the same way as visceral fat cell volume, controlled for individual clinical variables (Supplementary Table S1). In this depot, fat cell volume remained a significant regressor for PWV after adjustment for age, blood pressure and total cholesterol, but not when controlling for waist, waist–hip ratio, resting pulse rate, insulin sensitivity, plasma triglycerides, vWAT mass or subcutaneous adipocyte number. In a multiple regression analysis including all the parameters in Table 1 that were significantly associated with PWV, step-wise removal of nonsignificant variables resulted in a model where only visceral fat cell volume, age, diastolic blood pressure and pulse rate remained as

14 12 10 8 6

Women Men

14 12 10 8 6

4

4 200 400 600 800 1000 1200 Visceral fat cell volume (picolitres)

200 400 600 800 1000 1200 1400 Subcutaneous fat cell volume (picolitres)

Figure 1. Correlation between fat cell volume and pulse wave velocity. The relationship between these two parameters was determined by linear regression analysis. Gender for each individual is shown as indicated. International Journal of Obesity (2015) 222 – 227

© 2015 Macmillan Publishers Limited

Arterial stiffness and fat cells P Arner et al

225 Table 2.

Influence of single co-variables on the relationship between arterial stiffness and visceral fat cell volume

Table 3. Term

Factors contributing to variations in PWV Estimate

Bivariate relationship to PWV Visceral fat cell volume Std Beta 0.33 0.41 0.39 0.34 0.36 0.38 0.44 0.43 0.44 0.30 0.44 0.38

P-value 0.0002 o0.0001 o0.0001 o0.0001 o0.0001 o0.0001 o0.0001 o0.0001 o0.0001 0.01 o0.0001 0.0002

Co-variable Co-variable

Std Beta

P-value

age 0.30 0.0007 waist 0.083 0.37 waist–hip ratio 0.11 0.27 systolic blood pressure 0.36 o0.0001 diastolic blood pressure 0.39 o0.0001 resting pulse (beats per min) 0.30 0.0003 insulin sensitivity 0.0075 0.94 P-total cholesterol 0.09 0.29 P-triglycerides 0.018 0.85 visceral fat mass 0.18 0.12 subcutaneous fat cell volume 0.012 0.89 subcutaneous adipocyte number − 0.10 0.30

Abbreviations: P, fasting plasma; PWV, pulse wave velocity. Multiple regression analysis of the relationship between PWV, visceral fat cell volume and single co-variables. In each comparison, PWV was set as the dependent variable, while visceral fat cell volume and individual co-variables (that were significantly associated with PWV in Table 1) were set as independent regressors. Standardized beta coefficients (Std Beta) and P-values are indicated for the relationship between PWV and visceral fat cell volume (columns 1-2) adjusted for each co-variable, as are the corresponding values for each co-variable (columns 4-5).

independent regressors, together explaining 42% (that is, R2) of the variability in PWV (Table 3). The study group contained 33 patients with established and treated hypertension. There was no significant interaction of antihypertensive treatment on the relationship between fat cell volume in the visceral region and PWV (F = 0.95; P = 0.33). Furthermore, when subjects with hypertension were omitted from the analysis, the significant relationship between PWV and visceral fat cell volume remained (r = 0.42; P = o0.0001). DISCUSSION It is well established that increased central fat mass distribution is associated with arterial stiffness.5–7 However, the intrinsic factors in vWAT that account for this association remain to be established. Thus, although fat mass can expand by increasing the volume and/or number of its fat cells, it is unknown which of these parameters explains the correlation between visceral adiposity and arterial stiffness. To the best of our knowledge, this is the first examination of the relationship between arterial stiffness, measured as PWV, and fat cell volume or number. Arterial stiffness is an important pathophysiological factor associated with cardiovascular disease, and it is influenced by many factors (as recently reviewed3). We can confirm that age, blood pressure, resting heart rate, visceral fat accumulation and (albeit to a lesser degree) insulin sensitivity and plasma lipids are associated with PWV. However, the main finding of the present study relates to the morphology of WAT. In vWAT, but not sWAT, fat cell volume, but not number, was positively and independently correlated with PWV. The differences in visceral fat cell volume explained up to 20% of the inter-individual variations in PWV. A more pronounced association between arterial stiffness and visceral, compared with subcutaneous, fat cell volume was further suggested by the results of the multiple regression analysis. Thus, only visceral fat cell volume was a strong independent regressor for PWV when studied together with individual relevant cofactors, or when © 2015 Macmillan Publishers Limited

Intercept −1.54 Age (years) 0.040 Diastolic blood pressure (mm Hg) 0.043 Pulse rate (beats per min) 0.047 Visceral adipocyte volume (pl) 0.0025

Std Beta

Std Betaa

P-value

0.22 0.26 0.25 0.25

0.047 0.068 0.061 0.063

0.011 0.0017 0.0015 0.0020

Multiple regression analysis was performed using PWV as the dependent variable and the factors significantly associated with PWV in Table 1 as independent variables (that is, age, waist circumference, waist–hip ratio, systolic and diastolic blood pressure, pulse rate, insulin sensitivity, P-total cholesterol and triglycerides, vWAT mass, visceral fat cell volume, and subcutaneous adipocyte volume and number). Step-wise removal of regressors that did not contribute significantly to the model resulted in the variables listed in the table. The model had an overall F-ratio of 20.5, P = o 0.0001 and R2 = 0.42. Estimates, standardized beta coefficients (Std Beta), Std Beta. aP-values are indicated for each individual variable.

examining visceral and subcutaneous adipocyte volume in combination. Furthermore, analyses including all variables significantly associated with PWV, demonstrated that visceral fat cell volume was the only WAT parameter that entered the model (together with pulse rate, age and diastolic blood pressure). Therefore, although several studies have shown a relationship between PWV and central fat accumulation, our data suggest that it is visceral fat cell volume rather than vWAT mass per se that is related to arterial stiffness. Blood pressure is an important regulator of PWV.14,15 However, our main findings were independent of hypertension or blood pressure levels. The relationship between subcutaneous fat cell volume and blood pressure has been examined previously but with conflicting results. Hence, although adipocyte number, but not volume, was correlated with blood pressure in one report,23 the opposite was found in another study.24 As blood pressure was not a primary measure in our study and many subjects on antihypertensive pharmacotherapy were included, we refrained from making an in-depth analysis of blood pressure versus adipose fat cell volume or number in any region. Significant weight loss may improve PWV.25 Because the present study was cross-sectional, we cannot establish any longitudinal relationships between fat cell volume and PWV or whether fat cell volume is a prognostic factor for arterial stiffness. Unfortunately, visceral fat cell volume (the most important cellularity factor) can only be examined in connection with intraabdominal surgery, and such interventions cannot be performed repeatedly in subjects followed in long-term prospective studies. Our study group included only obese subjects undergoing abdominal surgery, the majority (490%) of which were severely/morbidly obese (that is, BMI435 kg m −2). It is not known whether PWV also associates with visceral fat cell volume in overweight or lean subjects. In addition, we used an indirect technique for measuring aortic PWV utilizing oscillometric measurements of the brachial artery wave dynamics. As discussed,17 there are not yet any outcome data for this method, in contrast to more cumbersome carotid-femoral measures. However, others have shown that the present method is very comparable with direct invasive measurements.18 We also followed the guidelines17 strictly to optimize the measurement conditions. This included the measurement of the jugulum– symphysis distance as a straight line, that is, not following the curvature of the abdomen. This minimized the possibility that the magnitude of central obesity could constitute an important confounding factor, further supported by the observation that International Journal of Obesity (2015) 222 – 227

Arterial stiffness and fat cells P Arner et al

226 waist–hip ratio did not remain as an independent regressor for PWV when controlling for visceral fat cell volume. Moreover, a recent study applying computed tomography, demonstrated that despite measuring along the body surface and adjusting for this error, PWV was still independently associated with visceral fat but not with other measures of adiposity.26 Is there a causal relationship between altered visceral fat cell volume and arterial stiffness? This study was not designed to answer this question although some ideas can be proposed. The renin–angiotensin–aldosterone system is involved in the aetiology of arterial stiffness as discussed.27 Fat cells secrete most of the components of the renin–angiotensin–aldosterone system and this endocrine-like activity is more pronounced in the visceral than the subcutaneous depot.28,29 Other potential factors are circulating free fatty acids (FFAs). We have previously shown in vWAT that norepinephrine-induced release of FFA is higher in men than in women,30 and it was recently demonstrated using magnetic resonance imaging in normal and obese subjects that aortic distensibility correlated inversely with FFA levels.31 In our present cohort, consisting predominantly of women, there was no significant relationship between fasting FFAs and PWV (Table 1), suggesting that the influence of FFAs—if any—on PWV may be more complex. Insulin resistance may also induce arterial stiffness.13,14 It is therefore interesting to note that fat cells from subjects with arterial stiffness display reduced intracellular amounts of the key insulin signalling protein Insulin Receptor Substrate-1.32 Furthermore, adipose inflammation is an important factor contributing to insulin resistance and other metabolic complications.33 Interestingly, adipocyte size correlates closely with the expression and secretion of the pro-inflammatory cytokine Tumour Necrosis Factor-alpha in lean healthy women34 and with insulin resistance/dyslipidaemia in morbidly obese women.8 Thus, although the mechanisms are not clear, increased fat cell size appears to alter several transcriptional regulatory networks resulting in increased inflammation, altered lipid metabolism, insulin resistance and thereby possibly arterial stiffness. As discussed,35 pro-inflammatory adipokines and lipids secreted from enlarged fat cells in visceral WAT may be of particular importance because this depot is drained via the portal vein into the liver, thereby impacting directly on hepatocyte function. A related possibility explaining the association between visceral fat cell size and PWV may be that the capacity of adipose tissue to store lipids is limited.36 The major contributor (495%) to adipocyte volume is the lipid droplet and fat cell volume in adult humans cannot exceed ~ 1400 pl.37 Consequently, when adipose expansion reaches a maximum, lipids are stored ectopically in for example, skeletal muscle, liver, pancreatic islets and arterial walls, which in turn leads to lipotoxic effects including abnormal lipoprotein metabolism, insulin resistance and arterial dysfunction. The correlations we observed between visceral fat cell volume and subcutaneous fat cell volume/number suggested that subjects with large visceral fat cells had larger but fewer subcutaneous adipocytes, indicating that visceral fat cell size may be linked to reduced subcutaneous adipose tissue expandability. Taken together, it could be speculated that subjects who have increased visceral fat cell size may be more prone to develop ectopic lipid deposition and/or that their adipocytes display a more pernicious functional profile, which in turn impacts on insulin sensitivity and arterial stiffness. In our present cohort, male gender was associated with larger visceral fat cells but, although several studies have shown that (healthy, young) men have higher PWV than women,38,39 we observed no significant difference in PWV, and sex did not influence the relationship between PWV and visceral fat cell volume. However, it cannot be excluded that the relatively small proportion of men (o 20%) did not provide us with sufficient statistical power to detect whether there is a stronger link between visceral fat cell volume and arterial stiffness in men compared with women. Studies of larger cohorts with a more International Journal of Obesity (2015) 222 – 227

balanced gender distribution and broader BMI range are therefore required. Finally, in mice with diet-induced obesity, arterial stiffness precedes the development of hypertension, and both parameters are normalized following weight loss, suggesting that arterial stiffness may be a cause rather than a consequence of hypertension.40 Unfortunately, it is at present not possible to increase fat cell volume selectively in different murine fat depots, thereby hampering studies of the mechanisms linking visceral fat cell volume to PWV. More suitable animal models cannot be developed until the specific progenitor cell and genes that discriminate between visceral and subcutaneous fat cells are identified. It may nevertheless be possible to develop drugs that act differently on fat cells in the two depots. For instance, the antidiabetic thiazolidinedione drugs influence adipose distribution in humans (reviewed in Yang and Smith41). In conclusion, visceral fat cell volume is a strong independent positive regressor for arterial stiffness, at least in severely/morbidly obese subjects. Large visceral fat cells may constitute the most important factor explaining the previously observed relationship between central fat accumulation and arterial stiffness. Whether the link between fat cell volume, arterial stiffness and cardiovascular disease is causal remains to be established. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS We are grateful for the excellent technical assistance provided by Ms Yvonne Widlund, Ms Katarina Hertel, Ms Britt-Marie Leijonhufvud, Ms Elisabeth Dungner, Ms Eva Sjölin, Ms Kerstin Wåhlén and Ms Gaby Åström. MR and PA conceived the study and wrote the first version of the manuscript which was then read and approved by all co-authors. All authors contributed to subject recruitment and/or analyses. This work was supported with grants from the Swedish Research Council, the Novo Nordisk Foundation, the Swedish Diabetes Association, the EASD/Lilly Foundation, The Erling-Persson Family Foundation and the Strategic Research Program in Diabetes at Karolinska Institutet.

FUNDING This work was supported with grants from the Swedish research Council, the Novo Nordisk foundation, the Swedish Diabetes foundation, the Diabetes program at Karolinska Institutet, the Stockholm County Council and The Erling-Persson Family Foundation.

REFERENCES 1 Grundy SM. Metabolic complications of obesity. Endocrine 2000; 13: 155–165. 2 Fantuzzi G, Mazzone T. Adipose tissue and atherosclerosis: exploring the connection. Arterioscler Thromb Vasc Biol 2007; 27: 996–1003. 3 Cecelja M, Chowienczyk P. Role of arterial stiffness in cardiovascular disease. JRSM Cardiovasc Dis 2012; 1: 4. 4 Safar ME, Czernichow S, Blacher J. Obesity, arterial stiffness, and cardiovascular risk. J Am Soc Nephrol 2006; 17 (4 Suppl 2): S109–S111. 5 Lim S, Choi HJ, Shin H, Khang AR, Kang SM, Yoon JW et al. Subclinical atherosclerosis in a community-based elderly cohort: the Korean Longitudinal Study on Health and Aging. Int J Cardiol 2012; 155: 126–133. 6 Ferreira I, Snijder MB, Twisk JW, van Mechelen W, Kemper HC, Seidell JC et al. Central fat mass versus peripheral fat and lean mass: opposite (adverse versus favorable) associations with arterial stiffness? The Amsterdam Growth and Health Longitudinal Study. J Clin Endocrinol Metab 2004; 89: 2632–2639. 7 Fantin F, Rossi AP, Cazzadori M, Comellato G, Mazzali G, Gozzoli MP et al. Central and peripheral fat and subclinical vascular damage in older women. Age Ageing 2013; 42: 359–365. 8 Hoffstedt J, Arner E, Wahrenberg H, Andersson DP, Qvisth V, Lofgren P et al. Regional impact of adipose tissue morphology on the metabolic profile in morbid obesity. Diabetologia 2010; 53: 2496–2503. 9 Salans LB, Knittle JL, Hirsch J. The role of adipose cell size and adipose tissue insulin sensitivity in the carbohydrate intolerance of human obesity. J Clin Invest 1968; 47: 153–165.

© 2015 Macmillan Publishers Limited

Arterial stiffness and fat cells P Arner et al

227 10 Lonn M, Mehlig K, Bengtsson C, Lissner L. Adipocyte size predicts incidence of type 2 diabetes in women. Faseb J 2010; 24: 326–331. 11 Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE. Enlarged subcutaneous abdominal adipocyte size, but not obesity itself, predicts type II diabetes independent of insulin resistance. Diabetologia 2000; 43: 1498–1506. 12 Veilleux A, Caron-Jobin M, Noel S, Laberge PY, Tchernof A. Visceral adipocyte hypertrophy is associated with dyslipidemia independent of body composition and fat distribution in women. Diabetes 2011; 60: 1504–1511. 13 Zieman SJ, Melenovsky V, Kass DA. Mechanisms, pathophysiology, and therapy of arterial stiffness. Arterioscler Thromb Vasc Biol 2005; 25: 932–943. 14 Brillante DG, O'Sullivan AJ, Howes LG. Arterial stiffness in insulin resistance: the role of nitric oxide and angiotensin II receptors. Vasc Health Risk Manag 2009; 5: 73–78. 15 Tomiyama H, Yamashina A. Arterial stiffness in prehypertension: a possible vicious cycle. J Cardiovasc Transl Res 2012; 5: 280–286. 16 Hirsch J, Gallian E. Methods for the determination of adipose cell size in man and animals. J Lipid Res 1968; 9: 110–119. 17 Tomlinson LA. Methods for assessing arterial stiffness: technical considerations. Curr Opin Nephrol Hypertens 2012; 21: 655–660. 18 Horvath IG, Nemeth A, Lenkey Z, Alessandri N, Tufano F, Kis P et al. Invasive validation of a new oscillometric device (Arteriograph) for measuring augmentation index, central blood pressure and aortic pulse wave velocity. J Hypertens 2010; 28: 2068–2075. 19 Andersson DP, Thorell A, Lofgren P, Wiren M, Toft E, Qvisth Vet al. Omentectomy in addition to gastric bypass surgery and influence on insulin sensitivity: A randomized double blind controlled trial. Clin Nutr e-pub ahead of print 12 January 2014. :doi: 10.1016/j.clnu.2014.01.004. 20 Kaul S, Rothney MP, Peters DM, Wacker WK, Davis CE, Shapiro MD et al. Dual-energy X-ray absorptiometry for quantification of visceral fat. Obesity 2012; 20: 1313–1318. 21 Adamopoulos D, Argacha JF, Gujic M, Preumont N, Degaute JP, van de Borne P. Acute effects of nicotine on arterial stiffness and wave reflection in healthy young non-smokers. Clin Exp Pharmacol Physiol 2009; 36: 784–789. 22 Ryden M, Andersson DP, Bernard S, Spalding K, Arner P. Adipocyte triglyceride turnover and lipolysis in lean and overweight subjects. J Lipid Res 2013; 54: 2909–2913. 23 Siervogel RM, Roche AF, Chumlea WC, Morris JG, Webb P, Knittle JL. Blood pressure, body composition, and fat tissue cellularity in adults. Hypertension 1982; 4: 382–386. 24 Berglund G, Ljungman S, Hartford M, Wilhelmsen L, Bjorntorp P. Type of obesity and blood pressure. Hypertension 1982; 4: 692–696. 25 Rider OJ, Tayal U, Francis JM, Ali MK, Robinson MR, Byrne JP et al. The effect of obesity and weight loss on aortic pulse wave velocity as assessed by magnetic resonance imaging. Obesity 2010; 18: 2311–2316.

26 Canepa M, Alghatrif M, Pestelli G, Kankaria R, Makrogiannis S, Strait JB et al. Impact of Central Obesity on the Estimation of Carotid-Femoral Pulse Wave Velocity. Am J Hypertens; e-pub ahead of print 22 March 2014. 27 Aroor AR, Demarco VG, Jia G, Sun Z, Nistala R, Meininger GA et al. The Role of Tissue Renin-Angiotensin-Aldosterone System in the Development of Endothelial Dysfunction and Arterial Stiffness. Front Endocrinol 2013; 4: 161. 28 Cassis LA, Police SB, Yiannikouris F, Thatcher SE. Local adipose tissue reninangiotensin system. Curr Hypertens Rep 2008; 10: 93–98. 29 van Harmelen V, Elizalde M, Ariapart P, Bergstedt-Lindqvist S, Reynisdottir S, Hoffstedt J et al. The association of human adipose angiotensinogen gene expression with abdominal fat distribution in obesity. Int J Obes Relat Metab Disord 2000; 24: 673–678. 30 Lonnqvist F, Thorne A, Large V, Arner P. Sex differences in visceral fat lipolysis and metabolic complications of obesity. Arterioscler Thromb Vasc Biol 1997; 17: 1472–1480. 31 Rider OJ, Holloway CJ, Emmanuel Y, Bloch E, Clarke K, Neubauer S. Increasing plasma free fatty acids in healthy subjects induces aortic distensibility changes seen in obesity. Circ Cardiovasc Imaging 2012; 5: 367–375. 32 Sandqvist M, Nyberg G, Hammarstedt A, Klintland N, Gogg S, Caidahl K et al. Low adipocyte IRS-1 protein expression is associated with an increased arterial stiffness in non-diabetic males. Atherosclerosis 2005; 180: 119–125. 33 Ryden M, Arner P. Tumour necrosis factor-alpha in human adipose tissue - from signalling mechanisms to clinical implications. J Intern Med 2007; 262: 431–438. 34 Arner E, Ryden M, Arner P. Tumor necrosis factor alpha and regulation of adipose tissue. N Engl J Med 2010; 362: 1151–1153. 35 Wree A, Kahraman A, Gerken G, Canbay A. Obesity affects the liver - the link between adipocytes and hepatocytes. Digestion 2011; 83: 124–133. 36 Virtue S, Vidal-Puig A. Adipose tissue expandability, lipotoxicity and the Metabolic Syndrome--an allostatic perspective. Biochim Biophys Acta 2010; 1801: 338–349. 37 Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O et al. Dynamics of fat cell turnover in humans. Nature 2008; 453: 783–787. 38 Koivistoinen T, Koobi T, Jula A, Hutri-Kahonen N, Raitakari OT, Majahalme S et al. Pulse wave velocity reference values in healthy adults aged 26-75 years. Clin Physiol Funct Imaging 2007; 27: 191–196. 39 Tomiyama H, Yamashina A, Arai T, Hirose K, Koji Y, Chikamori T et al. Influences of age and gender on results of noninvasive brachial-ankle pulse wave velocity measurement--a survey of 12517 subjects. Atherosclerosis 2003; 166: 303–309. 40 Weisbrod RM, Shiang T, Al Sayah L, Fry JL, Bajpai S, Reinhart-King CA et al. Arterial stiffening precedes systolic hypertension in diet-induced obesity. Hypertension 2013; 62: 1105–1110. 41 Yang X, Smith U. Adipose tissue distribution and risk of metabolic disease: does thiazolidinedione-induced adipose tissue redistribution provide a clue to the answer? Diabetologia 2007; 50: 1127–1139.

Supplementary Information accompanies this paper on International Journal of Obesity website (http://www.nature.com/ijo)

© 2015 Macmillan Publishers Limited

International Journal of Obesity (2015) 222 – 227

Copyright of International Journal of Obesity is the property of Nature Publishing Group and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.