|
|
||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Endocrine Research Unit, Mayo Clinic, Rochester, MN (YDT, CK, MVK, SBV, EW, and MDJ); the Department of Human Physiology, Pennington Biomedical Research Center, Baton Rouge, LA (YDT and EW); and the Postgraduate Program in Epidemiology, Federal University of Rio Grande do Sul, Porto Alegre, Brazil (EW)
2 Supported by grants no. DK40484, DK45343, DK50456, and RR00585 from the US Public Health Service and by the Mayo Foundation. 3 Reprints not available. Address correspondence to MD Jensen, Endocrine Research Unit, 5-194 Joseph, Mayo Clinic, Rochester, MN 55905. E-mail: jensen{at}mayo.edu.
| ABSTRACT |
|---|
|
|
|---|
Objective: We defined the extent to which subcutaneous adipocyte size is related to regional fat mass and to the sizes of adipocytes in other subcutaneous depots independent of adiposity, age, and sex.
Design: Data collected from 188 women and 133 men who were 18–50 y old and who had a body mass index (in kg/m2) of 18 to 50 were analyzed. The mean size of isolated subcutaneous abdominal, femoral, and gluteal adipocytes was measured by direct microscopy or by automated analysis of digital images. Visceral fat area was measured with computed tomography. Dual-energy X-ray absorptiometry was used to calculate adiposity.
Results: Stepwise multiple regression analyses showed that abdominal adipocyte size was associated positively with visceral and subcutaneous abdominal fat areas and negatively with lower-body fat mass as a percentage of total-body fat, after control for sex and percentage body fat. Femoral adipocyte size was related to percentage body fat (P < 0.0001), whereas gluteal adipocyte size was related to visceral fat area (P = 0.002), which suggests that these 2 lower-body fat depots are distinct. Analyses of data from a subset of volunteers (n = 99) for whom we had adipocyte size from all 3 depots showed that adipocyte size from 1 depot could be better predicted if adipocyte size from other depots were known.
Conclusions: Abdominal adipocyte size is related to body fat distribution. Adipocyte size in a person seems to be globally regulated by factors independent of variations in body fat distribution.
Key Words: Visceral fat body composition biopsy of fatty tissue obesity
| INTRODUCTION |
|---|
|
|
|---|
The association of adipocyte size or regional fat distribution (or both) with disease risk remains a topic of interest. Mean subcutaneous adipocyte size has been positively correlated with fasting (3, 5, 6) or postglucose serum insulin concentrations (3) in obese and normal-weight adults. In normal-weight women, subcutaneous abdominal (but not femoral, gluteal, or average) adipocyte size has been positively correlated with serum insulin and triacylglycerol concentrations (7). In their extensive study of obese persons, Krotkiewski et al (8) found consistent, positive associations between subcutaneous abdominal, femoral, or gluteal adipocyte size and serum insulin concentrations. They reported that, at the same level of adiposity, persons whose abdominal adipocytes were larger than their gluteal adipocytes had higher insulin and glucose concentrations than did persons with smaller abdominal than gluteal adipocytes (8). In this same study, waist-to-hip ratio was better able to discriminate obesity complications than was abdominal adipocyte size (8). Subsequent studies confirmed that subcutaneous abdominal adipocyte size is positively associated with unfavorable metabolic indexes in both sexes, whereas femoral adipocyte size displays weaker associations or none (9-12).
An obvious question that arises relates to whether adipocyte size is independently determined in each adipose region or whether it is inherently linked among depots. Some previous studies showed that the adipocyte sizes of different subcutaneous regions are strongly correlated with each other (3, 13), whereas other studies reported weaker or no correlation between subcutaneous abdominal and gluteal-femoral adipocyte sizes (7, 14). The small numbers of participants in these studies may compromise the ability to draw firm conclusions. The present study addressed the question of whether adipocyte size in a given subcutaneous depot is linked to a specific fat distribution pattern and to the size of adipocytes in other depots.
| SUBJECTS AND METHODS |
|---|
|
|
|---|
Written informed consent was obtained from each participant. These volunteers participated in fat metabolism studies conducted by our laboratory between 1995 and 2006, which were approved by the Institutional Review Board of the Mayo Clinic.
Body composition measurements
Total-body and regional fat masses were assessed with dual-energy X-ray absorptiometry (DXA) (Lunar Radiation, Madison, WI) (15). Leg fat mass was considered to be lower-body fat. A single-slice abdomen computed tomography (CT) scan at the L2–L3 interspace was performed to measure visceral fat area (cm2); these data, together with DXA measures of abdominal fat, were used to calculate visceral fat mass and abdominal subcutaneous fat mass (16). The relative amount of lower-body fat was calculated as leg fat ÷ total body fat. Visceral and subcutaneous abdominal fat areas as assessed by CT scan were used as our measures of visceral and subcutaneous abdominal adiposity, respectively. Because we collected data over the course of several years, it was important to ensure consistent measures of body fat content and distribution. This was accomplished by calibrating our DXA instruments with meat blocks [4 sealed 5-kg blocks containing a mixture of fat and lean meat of known composition (10–70% fat)] every month as described previously (15) and by having the investigators who analyzed the abdominal CT scans also analyze the quality-control CT images at the same time to ensure consistent approaches to image analysis over time.
Biopsies of adipose tissue
Adipose tissue samples from 1–3 subcutaneous sites, depending on the protocol and tissue availability, were obtained by small-needle liposuction under sterile conditions and local anesthesia. Tissue samples for biopsy were taken from the abdominal region, two-thirds of the distance from the iliac spine to the umbilicus; the gluteal region, lower lateral quadrant of the buttock; and the femoral region, on the anterior aspect of the thigh, one- to two-thirds of the distance from the superior iliac spine to the patella. Fat tissue was immediately rinsed with saline through Nitex Nylon Fiber 250/50 (Small Parts Inc, Miami Lakes, FL) and processed for measurement of adipocyte size as described below.
Measurement of adipocyte size
Adipocyte size was assessed by using the approach of Di Girolamo et al (17), which involves collagenase digestion of the adipose tissue sample, separation of adipocytes by centrifugation, methylene blue staining to identify the nuclei, and measurement of the cell diameter. Briefly, samples of adipose tissue were digested in collagenase (Type II C-6885; Sigma Chemical Co, St Louis, MO) in HEPES buffer [0.1 mol HEPES/L, 0.12 mol NaCl/L, 0.05 mol KCl/L, 0.005 mol glucose/L, 1.5% (wt:vol) bovine serum albumin (BSA), and 1 mmol CaCl2/L (pH 7.4)] at 37 °C with the use of a water bath and of shaking at 100–115 rotations/min until the digestion was nearly complete, which occurred in 20–60 min. The cell suspension formed was centrifuged for 5 min at 300 x g at room temperature. Adipocytes in the top layer were mixed uniformly with a pipette, and a 50–150-µL aliquot was added to 450 µL of 0.2% methylene blue/HEPES solution for nuclei staining and incubated for 15 min at 37 °C in the water bath. A 5- to 10-µL portion of the cell suspension was placed in each well of an 8-well Teflon-coated glass slide; the slide was coverslipped, and the cell suspension was measured optically by using a microscope (Labophot 2/2A; Nikon Inc, Melville, NY) equipped with an eyepiece having a 10-mm scale reticle at phase contrast at 100x magnification. During the procedure, precautions were taken to avoid adipocyte breakage. The inclusion of immature multilocular adipocytes was avoided by excluding cells that were <35 µm in diameter.
For studies performed before the year 2000, the diameter of 100–300 fat cells was defined by direct microscopy as previously described (17). Briefly, the adipocyte diameters were compared with the scale from the reticle and were appointed to class intervals of 7 µm, which created a histogram. The average volume of the population was calculated by using the mean (±SD) diameter obtained from the histograms and by using Goldrick's formula. Although <300 cells [the number best thought to minimize error (17)] were measured in a small number of tissue samples, we have observed good agreement between the manual sizing methods (even measuring <300 cells) and an automated measurement approach to measuring adipocyte size on the basis of digital photographs (18). Because only a subset of the data was used for these comparisons, we explored for the potential confounding effect of the change in method in the present study by including the adipocyte size method as an independent variable in the multiple regression models. For the automated method, we measured the area of
300 fat cells by using an in-house program designated AdCount [renamed Cell Counting and Analysis program (CCAP)] that was written by the Biomedical Imaging Resource at Mayo Clinic (18). This program measures the area and computes the diameter and the volume of the individual fat cells, and those values are used to estimate the mean of the adipocyte diameter and volume. Adipocellular lipid weight was calculated as adipocyte volume times 0.915 (density of triolein). We designate adipocyte size as the calculated mean lipid content of the measured cells.
Determination of regional adipocyte number
Adipocyte number in the subcutaneous abdominal fat compartment was determined by dividing the subcutaneous abdominal fat mass by the mean abdominal subcutaneous adipocyte size. Similarly, adipocyte number in the lower-body compartments was determined by dividing the lower-body fat mass (determined by DXA) by the mean femoral adipocyte size.
Statistical analyses
Anthropometric and body fat distribution characteristics were compared between sexes by using 2-sample t test. We performed 3-factor analysis of variance to analyze the effect of depot, sex, and BMI category on adipocyte size and number; next, we performed pair-wise comparisons using the Tukey post hoc test for the significant interactions.
The relations between percentage body fat (%BF) or regional adiposity and regional adipocyte size were analyzed by using simple linear regression models after grouping by sex. We used the logarithmic transformation of the measurements of regional adiposity to account for the curvilinear relation with the regional adipocyte size. Analysis of covariance was used to test for differences in the slope between the sexes.
We used stepwise multiple regression analyses to evaluate the independent contributions of body fat distribution to the variance in adipocyte size. Adipocyte size values were logarithmically transformed to meet the assumptions of linear regression analysis. We included BMI, %BF (5, 19, 20), age (5, 14), sex, and the method for adipocyte size determination in the models in addition to the variables of our interest, because these factors have been shown to affect adipocyte size at the whole-body level or in specific regions. A variance inflation factor was calculated to find mutually dependent predictors. We accepted variables with a variance inflation factor of <10 into the initial model and those with a significance level of <0.05 into the final model.
We also performed stepwise multiple regression analyses with and without the addition of adipocyte size of the other depots as candidate predictors by using data from a subset of volunteers (n = 99) for whom we had measured adipocyte size data in all 3 subcutaneous depots. For example, in a model to determine independent predictors of the abdominal adipocyte size, the femoral and gluteal adipocyte sizes were included in the model as independent variables.
Statistical analyses were performed by using JMP (version 6.0.0) and SAS (version 9.1.3) software (both: SAS Institute Inc, Cary, NC). P < 0.05 denoted statistical significance. Values reported are means ± SDs unless otherwise stated. Multiple linear regression results are reported with the parameter estimate and P value for each variable.
| RESULTS |
|---|
|
|
|---|
|
|
|
|
5–10 x 109 cells in women with 5 kg of adipose lipid, but averaged
30 x 109 cells in those with
25 kg of adipose lipid. In contrast, the adipocyte number in men averaged
10 x 109 cells across the range of 2 to 13 kg of leg adipose lipid measured in our volunteers.
|
|
|
| DISCUSSION |
|---|
|
|
|---|
The effect of BMI and regional adiposity on adipocyte number in abdominal subcutaneous and lower-body adipose tissue differed between men and women. Normal-weight men had more leg adipocytes than abdominal subcutaneous adipocytes (Table 3
). An increase in leg fat to >
10 kg in women was not associated with increased adipocyte size (Figure 2B
), which indicated that fat mass expanded as a result of adipocyte hyperplasia. In contrast, the men with greater amounts of leg fat had larger femoral adipocytes without a concomitant increase in the leg adipocyte number, which indicated that fat mass expansion was primarily due to adipocyte hypertrophy. It is noteworthy that we did not encounter men with >15 kg leg fat in the BMI range we included. In contrast, the relation between CT-assessed subcutaneous fat and adipocyte size (Figure 2A
) suggests that an expansion of abdominal fat by adipocyte hypertrophy does not fundamentally differ between women and men; in addition, women have more adipocytes than do men, and those adipocytes contribute to a larger subcutaneous fat mass in women. Unfortunately, our understanding of adipocyte proliferation is incomplete. Secreted factors from hypertrophic adipocytes are reported to enhance preadipocyte proliferation (30), and there is evidence for differences in preadipocyte differentiation between femoral and subcutaneous abdominal preadipocytes in culture (31). Much more remains to be learned about the factors that determine adipose tissue proliferation in different depots in men and women.
To the best of our knowledge, this study is the first to investigate the relation between body fat distribution and abdominal adipocyte size after control for the effects of adiposity, BMI, age, and sex. Visceral and subcutaneous abdominal fat areas assessed by CT scan and the relative amount of lower-body fat emerged as significant predictors of abdominal subcutaneous adipocyte size. The significant correlation we observed between visceral fat and abdominal subcutaneous adipocyte size is in accord with the data from previous reports (7, 11, 32-34); however, previous studies did not clarify whether these associations exist after control for adiposity. Excess visceral or truncal subcutaneous fat masses are unfavorable health markers, whereas greater amounts of lower-body fat mass may protect against adverse health consequences (12, 34-37). Our finding that abdominal subcutaneous adipocyte size is associated positively with abdominal (visceral and subcutaneous) fat areas and negatively with the relative lower-body fat mass provides a plausible explanation for the relation between fat distribution and abnormal fatty acid metabolism in obesity. We have found that most of the excess FFA in upper-body obesity arises from upper-body subcutaneous fat (38), and in vitro studies suggest that larger adipocytes have greater rates of lipolysis than do smaller adipocytes (39). It has also been reported that large abdominal adipocyte size is associated with greater metabolic risk, independent of obesity and visceral fatness (10, 40). Our observation might also explain why lower-body fat mass appears to exert independent beneficial effects on health—ie, why persons with greater leg fat mass tend to have smaller abdominal subcutaneous adipocytes. The finding that both abdominal (visceral and subcutaneous) and leg fat masses are related to abdominal adipocyte size in opposite directions is partially consistent with the hypothesis that impaired capacity of fat storage in the subcutaneous adipose tissue leads to compensatory storage of excess fat in visceral and ectopic (ie, muscle, liver, and pancreas) depots (41). Our data show that larger quantities of fat in the lower body may be the critical factor that dictates the ability of subcutaneous adipose tissue to expand to accommodate excess fat.
By including femoral and gluteal adipocyte size in the model to predict abdominal adipocyte size, we were able to account for much larger portions of the interindividual differences in adipocyte size than if we had used only body fat characteristics. These data suggest that adipocyte size in individual persons may be somewhat globally regulated over and above differences in adiposity and body fat distribution. The correlation among the adipocyte sizes of various subcutaneous depots has been reported previously (5, 9), but the mechanisms that underlie that correlation may only be speculated on at present. Possibilities include genetic or epigenetic variation in genes encoding for regulatory factors of preadipocyte proliferation or differentiation, lipid storage, or lipolysis (or all). An example is the genetic variation in the LMNA gene encoding the nuclear envelope proteins lamin A/C that is found to be a familial trait in Pima Indians (40), a population characterized by enlarged subcutaneous abdominal adipocyte size (42). Another possibility is constitutional factors. For example, a recent study showed that the cell size of slowly dividing cells, including adipocytes, is larger with larger body size (43).
To our surprise, visceral (but not femoral) fat was positively and independently associated with gluteal adipocyte size (Table 4
). The gluteal fat of obese persons has been reported to be richer in deep subcutaneous fat than in fat from the thigh region (44), which may explain our observation. It is possible that gluteal adipose tissue does not completely represent lower-body fat, and, given that this tissue bed is usually much smaller in size than that of leg fat, it may be problematic to extrapolate from gluteal adipocyte metabolism to the metabolism of lower-body fat in general. We could not detect the association between visceral fat and gluteal adipocyte size in the smaller group of observations we used to assess the possible link between adipocyte sizes in other depots (Table 5
). Whether this is due to the loss of statistical power from a smaller sample size or to the association of abdominal adipocyte size and visceral fat area is not clear.
In summary, these results have important implications for our understanding of how much body fat distribution and the adipocyte size from other depots contribute to the variance in regional adipocyte size, independent of adiposity, sex, and age. Because adipocyte size is thought to relate to the regulation of adipocyte function, an understanding of the factors that influence adipocyte size should be helpful. One of the new findings of the present study is that lower-body fatness and abdominal (visceral and subcutaneous) fat are independently related to subcutaneous abdominal adipocyte size, but in different directions. This suggests that relative lower-body fat mass is an indicator of the functional storage capacity of the subcutaneous adipose tissue. A second interesting finding was that gluteal adipocyte size was not related to the lower-body fat mass but was associated with visceral fat. A third new finding is that adipocyte sizes of other subcutaneous depots are the strongest determinants of adipocyte size in any given subcutaneous depot. We interpret this finding as evidence that adipocyte size in individual persons is somewhat globally regulated, so that adipocytes from different regions are more alike than they are different, even in the face of variations in level of adiposity and body fat distribution. Finally, our data suggest that lower-body fat in men expands largely as a result of adipocyte hypertrophy, whereas the expansion of lower-body fat in women is governed by adipose tissue hyperplasia.
| ACKNOWLEDGMENTS |
|---|
The authors ' responsibilities were as follows—YDT: study design, data collection, data management and analysis, data interpretation, and writing of the manuscript; CK: data collection, data management and analysis, and writing of the manuscript; MVK: data management, statistical analysis, and writing of the manuscript; SBV: data collection and writing of the manuscript; EW: data management and data analysis; and MDJ (principal investigator): study concept and design, data interpretation, and review of the manuscript. None of the authors had any financial or personal conflict of interest.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. D. Jensen Role of Body Fat Distribution and the Metabolic Complications of Obesity J. Clin. Endocrinol. Metab., November 1, 2008; 93(11_Supplement_1): s57 - s63. [Abstract] [Full Text] [PDF] |
||||
![]() |
J.-P. Despres, I. Lemieux, J. Bergeron, P. Pibarot, P. Mathieu, E. Larose, J. Rodes-Cabau, O. F. Bertrand, and P. Poirier Abdominal Obesity and the Metabolic Syndrome: Contribution to Global Cardiometabolic Risk Arterioscler. Thromb. Vasc. Biol., June 1, 2008; 28(6): 1039 - 1049. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |