Upper and Lower Body Adipose Tissue Function: A Direct Comparison of Fat Mobilization in Humans


Oxford Centre for Diabetes, Endocrinology and Metabolism, Churchill Hospital, Oxford OX3 7LJ, United Kingdom. E-mail: garry.tan@oxlip.ox.ac.uk


Objectives: Fat in the lower body is not associated with the same risk of cardiovascular disease as fat in the upper body. Is this explained by differences in the physiological functioning of the two depots? This study had two objectives: 1) to determine whether fat mobilization and blood flow differ between gluteal and abdominal adipose tissues in humans, and 2) to develop a new technique to assess gluteal adipose tissue function directly.

Research Methods and Procedures: We performed detailed in vivo studies of adipose tissue function involving the assessment of fat mobilization by measurement of adipose tissue blood flows, arterio-venous differences of metabolites across each depot, and gene expression in tissue biopsies in a small-scale physiological study.

Results: Gluteal adipose tissue has a lower blood flow (67% lower, p < 0.05) and lower hormone-sensitive lipase rate of action (87% lower, p < 0.05) than abdominal adipose tissue. Lipoprotein lipase rate of action and mRNA expression are not different between the depots. This is the first demonstration of a novel technique to directly investigate gluteal adipose tissue metabolism.

Discussion: Direct assessment of fasting adipose tissue metabolism in defined depots show that the buttock is metabolically “silent” in terms of fatty acid release compared with the abdomen.


Abdominal obesity is associated with a high risk of type 2 diabetes and cardiovascular disease (1). In contrast, lower body obesity does not have the same associations, and may in fact, exert a protective effect (2). The reason for this difference in the effects of adipose tissue according to its anatomical distribution is unclear.

The view of adipose tissue function has changed in recent years (3). It was once considered to be a passive store of excess energy, releasing its stored energy in the form of nonesterified fatty acids (NEFAs)1 in fasting conditions, whereas it is now known to secrete a large range of endocrine factors with system-wide effects on whole body physiology. Adipose tissue is also thought to play an active role in the regulation of energy storage and release, thus helping to protect the body against the adverse peaks and troughs of NEFAs that would otherwise occur in the course of a day (3).

This “buffering” action of NEFAs by adipose tissue is important. In the presence of inadequate NEFA buffering, other tissues are exposed to elevated NEFA concentrations. It has been suggested that this exposure is a prime cause of the insulin resistance found in obesity. The mechanism for this is unclear, but in obesity, there may be an excessive release of NEFAs from adipose tissue. This substrate competition causes an inhibition of glucose use in muscle (4). This hypothesis, originally proposed by Randle et al. (4), is supported by data showing that infusions of a lipid emulsion into lean healthy volunteers increased lipid oxidation and decreased glucose use (5). Epidemiological data also suggest a link between NEFA concentrations and type 2 diabetes, although this link is under debate (6).

Excessive fatty acid release by adipose tissue has been linked to insulin resistance and dyslipidemia. However, does the gluteal fat depot contribute less than other fat depots to circulating NEFA levels? Might this explain the association between the higher risk of type 2 diabetes with abdominal body fat compared with lower body fat?

Research Methods and Procedures

To answer this question, we describe the first direct quantification of fatty acid release from gluteal adipose tissue in vivo in humans and compare it with that of subcutaneous abdominal adipose tissue. This was done by measuring arterio-venous differences in metabolite concentrations across both the gluteal and abdominal adipose tissue depots. This allows a comparison of the composition of the blood being supplied to adipose tissue and blood draining from it. This arterio-venous difference is a reflection of the metabolic activity in the adipose tissue depot (7, 8).

To assess the net uptake or release of metabolites, the arterio-venous measurements must be combined with a measurement of tissue blood flow. Adipose tissue blood flows were assessed with the 133Xe washout technique (9).


Six nonobese healthy male subjects (median age, 28 years; range, 22 to 43 years) were studied after a 12-hour overnight fast. The median BMI was 22.9 kg/m2 (range, 20.1 to 27.1 kg/m2), and the median fasting triglyceride concentration was 0.71 mM (range, 0.46 to 1.82 mM). All subjects gave their written informed consent, and the study was approved by the Oxfordshire Clinical Research Ethics Committee.

Identification of Veins

The anatomy of a network of veins in the subcutaneous gluteal adipose tissue was identified using ultrasonography (ATL HDI 5000, Philips, Reigate, Surrey, United Kingdom; with a 7- to 14-MHz probe). A fiber optic light source was also used to identify the veins by illuminating the adipose tissue. We defined a reticular pattern of veins varying in depth from 4 to 10 mm throughout the gluteal adipose tissue (Figure 1). The vein diameters ranged from 0.2 to 1.0 mm. The larger veins drained laterally toward the greater trochanter or medially toward the inner thigh and into the femoral vein. There was also a limited communication between the gluteal venous network and veins of the upper thigh. Doppler measurements identified the direction of blood flow, which demonstrated a centrifugal pattern from the buttock. The fascia separating adipose from muscle tissue was identified by ultrasonography; there were no veins penetrating this fascia. This indicated that our blood samples were representative of gluteal fat and were not contaminated by drainage from gluteal muscle.

Figure 1.

The venous network in gluteal adipose tissue.

Cannulation of Veins

A vein draining subcutaneous abdominal adipose tissue (8) and one of the larger veins draining gluteal adipose tissue were cannulated by the Seldinger technique using a 20-gauge central venous catheter (Hydrocath; Becton Dickinson, Oxford, UK). To prevent contamination of gluteal venous samples with blood from the upper thigh, a blood pressure cuff was applied immediately below the gluteal fold and inflated to a pressure of 200 mm Hg before blood sampling. Arterialized venous blood was obtained from a cannula inserted retrogradely into dorsal vein of a hand, which was placed in a hot box heated to 60 °C for at least 15 minutes before blood sampling. Adequate arterialization was confirmed, by blood gas analysis, with an oxygen saturation of >94% (IL 628 co-oximeter; Instrumentation Laboratory Ltd., Warrington, UK).

Blood Sampling and Analysis

Simultaneous blood sampling from the three sites was performed at two baseline time-points. Blood samples were collected into heparinized syringes (Monovette; Sarstedt, Leicester, UK) that were prechilled on ice. Two hundred microliters of each blood sample was rapidly deproteinized with 7% (wt/vol) perchloric acid to assess glycerol concentrations on an IL Monarch centrifugal analyzer (Instrumentation Laboratory Ltd.). Plasma was rapidly separated from the remaining blood by centrifugation at 4 °C. Plasma samples were analyzed for triglyceride (blanked for free glycerol) (10). NEFA concentrations were measured using a Wako kit (Alpha Laboratories, Eastleigh, UK) on an IL Monarch centrifugal analyzer (Instrumentation Laboratory Ltd.).

Measurement of Adipose Tissue Blood Flow

133Xe was injected into the gluteal and abdominal adipose tissue depots. After an equilibration period of at least 30 minutes, blood flow was measured by collecting continuous 20-second readings from a portable scintillation detector system with a CsI crystal (Oakfield Instruments, Eynsham, UK), which was placed a fixed distance over the exact site of 133Xe injection (11). Each subject was placed on their left side and was supported by cushions. Two calibrated scintillation detector systems were used to allow simultaneous measurements of gluteal and abdominal adipose tissue depot blood flows.

We were concerned about the possibility that gamma emissions from the 133Xe injected at one site might be picked up by the detector at the other site, thus contaminating the depot specificity of the data collected. To test whether this was a significant problem, scintillation detectors were placed in the positions on the abdomen and over the gluteal region that were used in the study. However, in contrast to the studies, 133Xe was only placed in either the gluteal or abdominal region, whereas data were recorded from both probes. The detector placed over the site without 133Xe consistently showed no difference to background radiation levels, which demonstrated that there was no interference between the tissue recordings of gamma emissions.

Adipose tissue blood flow was calculated using a partition coefficient of 10 g/mL as described by Larsen et al. (9).

Adipose Tissue Biopsies and mRNA Analysis

After local anesthesia with 1% lidocaine, adipose tissue biopsies were taken from both depots using a 12-gauge needle. Concentrations of mRNA of hormone-sensitive lipase (HSL) and lipoprotein lipase (LPL) were quantified by reverse transcriptase real time polymerase chain reaction (PCR) using a Light-Cycler (Roche Diagnostics, Meylan, France). First-strand cDNAs were first synthesized from 200 ng of total RNA in the presence of 100 units of Superscript II (Invitrogen, Eragny, France) using both random hexamers and oligo (dT) primers (Promega, Charbonnières, France). The real-time PCR was performed in a final volume of 20 μL, which contained 5 μL of a 60-fold dilution of the reverse transcriptase reaction medium, 15 μL of reaction buffer from the FastStart DNA Master SYBR Green kit (Roche Diagnostics), and 10.5 pmol of the specific forward and reverse primers. Primers were selected to amplify small fragments and to hybridize in different exons of the target sequences. The following primer sets were used: 5′-GGTCGAAGCATTGGAATCCAG-3′ (LPL sense); 5′-TAGGGCATCTGAGAACGAGTC-3′ (LPL antisense); 5′-CCCTCATGGCTCAACTCCTTCC-3′ (HSL sense); and 5′-TTGACATCGGAGGGTGTGGAGG-3′ (HSL antisense). Each assay was performed in duplicate, and validation of the real time PCR runs was assessed by evaluation of the melting temperature of the products and by the slope and error obtained with the standard curve. The analyses were performed using Light-Cycler software (Roche Diagnostics). Cyclophilin mRNA levels were measured as an internal standard, and the data were expressed as units relative to cyclophilin mRNA concentrations.


Net uptake or release was calculated as the arterio-venous or veno-arterial difference multiplied by adipose tissue blood flow. Concentrations of lipid metabolites in plasma were first converted to whole blood values by multiplying by (1 − hematocrit). The calculation for the HSL physiological rate of action was based on mass balance calculations of triglyceride and glycerol, whereas LPL rate of action was calculated from triglyceride mass balance calculations, as previously described (7).

Statistical Analysis

Data were analyzed using SPSS for Windows (Version 10; SPSS UK, Chertsey, UK), and statistical significance was set at p < 0.05 for all tests. Data were analyzed using Wilcoxon signed-rank tests.


The blood flow in gluteal adipose tissue was 67% lower (range, 7% to 93%; p < 0.05) than that in abdominal adipose tissue (Table 1). Median gluteal skin temperature was 32.2 °C, whereas abdominal skin temperature was 33.3 °C (p < 0.05). The arterio-venous difference of fatty acids was 39% lower in gluteal than in abdominal adipose tissue (p = 0.12). The principal enzyme releasing fatty acids from adipose tissue triglyceride stores is HSL. HSL rate of action was 87% lower (range, 23% to 99%; p < 0.05) in gluteal than in abdominal adipose tissue. Paradoxically, the HSL mRNA content was 37% higher (range, 3% to 49%; p < 0.05) in gluteal than in abdominal adipose tissue, demonstrating that the HSL rate of action is regulated by post-transcriptional events.

Table 1. . Functional differences between abdominal and gluteal adipose tissue
  1. Values are mean ± SE.

  2. Wilcoxon signed-rank tests are used throughout.

  3. NS, not significant.

Adipose tissue blood flow (mL/100 g tissue/min)4.2 ± 1.31.4 ± 0.5<0.05
NEFAΔ Veno-arterial concentration(μm/L)811 ± 253497 ± 125NS
 Flux(nmol/100 g tissue/min)2034 ± 929349 ± 172<0.05
TriglycerideΔ Arterio-venous concentration(μm/L)34 ± 890 ± 34<0.05
 Clearance(mL/min)15 ± 87 ± 4NS
HSLRate of action(nmol/100 g tissue/min)601 ± 29178 ± 50<0.05
 mRNA(arbitrary units)38 ± 752 ± 8<0.05
LPLRate of action(nmol/100 g tissue/min)89 ± 3749 ± 16NS
 mRNA(arbitrary units)266 ± 46270 ± 37NS

The arterio-venous triglyceride uptake was three times greater in gluteal than in abdominal adipose tissue (p < 0.05), but when blood flow was factored into the calculation of net adipose tissue triglyceride clearance, there was no significant difference between the depots. The enzyme determining triglyceride extraction, LPL, was not different between the tissues when assessed either by enzyme rate of action or by gene expression. We speculate that the markedly low gluteal blood flow facilitated contact between plasma triglyceride-rich lipoproteins and LPL.

We had originally intended to quantify fatty acid release by adipose tissue in the fasting and postprandial states. Unfortunately, this proved technically difficult because of gluteal cannulae clotting, presumably caused by the low blood flow.


This is the first direct comparison of fatty acid release by gluteal and abdominal adipose tissue. It also shows for the first time that blood flow in gluteal adipose tissue is much lower than that in abdominal adipose tissue. Gluteal adipose tissue is a metabolically inert depot with a very low blood flow and very low rate of fatty acid release.

The importance and relevance of in vivo physiological measurements to assess a tissue's metabolic function are highlighted by this study. Whereas the physiological HSL rate of action was markedly decreased in gluteal compared with abdominal adipose tissue, there was in fact an increase in HSL mRNA content in gluteal adipose tissue. This discrepancy between gene expression and the physiological measure illustrates the limitations and pitfalls of using gene expression alone to assess physiological function, particularly in the case of a gene such as HSL, which is regulated post-transcriptionally.

Previous attempts to study the functional differences between adipose tissue depots in humans in vivo have involved either microdialysis (12) or isotope dilution techniques (13). Microdialysis measures interstitial glycerol concentrations as a marker of lipolysis or fatty acid release. However, this method may overestimate lipolysis, and without simultaneous quantification of the tissue blood flow, it is difficult to estimate fatty acid release from adipose tissue because the index of lipolysis measured, glycerol, is derived from both HSL and LPL activity; the microdialysis method does not differentiate between these sources of glycerol. Furthermore, microdialysis methodology is limited by its inability to study hydrophobic molecules such as fatty acids and triglycerides.

In comparison with regional isotope dilution techniques, the direct measurement of arterio-venous differences in metabolite concentrations across the gluteal adipose tissue allows the assessment of a relatively homogenous isolated adipose tissue depot. This new technique also offers the possibility of comparing the two adipose tissue depots in other ways. For example, the secretion of recently described hormones from adipose tissue, such as adiponectin and resistin, could be measured from different depots to investigate the relative contributions to insulin resistance. However, the regional isotope dilution technique has proved valuable in the whole body assessment of fatty acid mobilization (14).

A weakness of this study is its limited subject population. However, it might be expected that this lean group of subjects may have a more active gluteal adipose tissue depot compared with subjects accumulating buttock fat. Future work will investigate whether these differences will be seen in other subject groups, such as women and other ethnic groups.

We believe that the quiescent nature of gluteal adipose tissue explains the lack of association between cardiovascular risk and lower body fat. Therefore, what is the purpose of gluteal fat? Presumably, release of fatty acids from this depot will eventually take place when other fat stores are exhausted. However, for most of us, “a moment on the lips” is indeed perhaps “a lifetime on the hips.”


This study was funded by the Wellcome Trust. Garry Tan is an Medical Research Council Clinical Training Fellow. Fredrik Karpe is a Wellcome Trust Senior Clinical Research Fellow. We thank Prof. Keith Frayn for constructive criticism.


  • 1

    Nonstandard abbreviations: NEFA, nonesterified fatty acid; HSL, hormone-sensitive lipase; LPL, lipoprotein lipase; PCR, polymerase chain reaction.