Heterogeneity Among White Adipose Tissue Depots in Male C57BL/6J Mice

White adipose tissue (WAT) fulfills a wide range of functions (including metabolic regulation, glucocorticoid and steroid hormone synthesis, blood pressure regulation, and others ( 1 )). While WAT was historically considered a homogeneous organ, intraabdominal (IA) WAT is now recognized to have a more negative impact on health than subcutaneous (SC) WAT ( 2 ). In fact, fundamental differences among WAT depots have been reported, including their influence on metabolism ( 3 ), endocrine function ( 2 ), preadipocyte characteristics ( 4 ), response to high fat feeding ( 5 ), expression of developmental genes ( 6 ), etc. Still, data from an individual depot is often inappropriately assumed to reflect the status of all WAT ( 7 ). We and others (reviewed in 8 ) have reported striking differences in WAT depot enlargement in GHR−/− mice (growth hormone receptor gene-disrupted mice), which present exaggerated fat deposition predominantly in the SC, but also in the retroperitoneal depot. Despite their elevated fat mass, these mice remain insulin sensitive and live longer than their wild-type littermates. A thorough characterization of the molecular mechanisms dictating WAT depot differences in wild-type mice is essential to be able to evaluate the corresponding differences in GHR−/− mice (or any other mouse line, for that matter). The present study was performed to address that need. The levels of several proteins have been measured and compared among WAT depots in humans and rodents, with many showing WAT depot-specific expression ( 9 – 13 ). In most studies of WAT, isolated adipocytes are separated from stromovascular cells (preadipocytes, macrophages, endothelial cells, etc) in a process that disrupts the complex WAT microenvironment. In this regard, a recent report showed that proteomes of visceral and SC WAT in humans display more differences in stromovascular cells than in isolated adipocytes ( 13 ), suggesting that the cellular heterogeneity in WAT might dictate physiological differences among depots. Therefore, for a more comprehensive analysis of the characteristics of WAT depots in vivo, whole tissue samples should be analyzed, preserving the mixture of cell populations and the tissue architecture. Most obesity studies in mice involve mouse models of the C57BL/6J strain ( ob / ob , db / db , diet-induced obesity, etc.) ( 14 ). Thus, it is of particular importance to understand the physiological differences among WAT depots (and the molecular mechanisms behind them), particularly in the C57BL/6J strain. In the current study, we used a proteomic approach to obtain a broad view of the protein profiles of various WAT depots. Previous proteomic studies comparing SC and IA WAT have focused on obese humans ( 12 ) or, as stated above, have used collagenase treated WAT samples to analyze isolated adipocytes separately from stromovascular cells ( 13 ). However, a proteomic characterization of intact tissue evaluating depot differences in protein profiles under “normal” physiological conditions has not been performed. Furthermore, analyses of WAT rarely include the mesenteric depot, even though many consider it a true visceral pad as, unlike epididymal or retroperitoneal, the mesenteric depot drains into portal circulation. Given that expansion of the visceral WAT is closely associated to human diseases such as type 2 diabetes and cardiovascular disease, in many respects, the mesenteric depot is directly relevant to human pathophysiology. In this study, we used two-dimensional gel electrophoresis (2DE) followed by mass spectrometry (MS) to study one SC (inguinal) and three IA (visceral: mesenteric; non-visceral: epididymal and retroperitoneal) WAT depots in wild-type male C57BL/6J mice. Specific proteins were also assayed by western blotting and immunohistochemistry (IHC). In addition, products of lipid peroxidation, correlations to plasma levels of insulin, leptin and adiponectin, and several phenotypic characteristics among WAT depots were evaluated. Our data reveal differences in metabolism and oxidative stress that likely reflect molecular mechanisms leading to unique physiological states in individual WAT depots. This information adds to our understanding of WAT function and should be considered when interpreting results of WAT studies using individual depots.

Mice were sacrificed by cervical dislocation and inguinal, retroperitoneal, mesenteric and epididymal WAT (including corresponding lymph nodes and blood vessels) were collected and weighed. Regarding the retroperitoneal depot, we dissected WAT located behind the kidney, along the back of the abdomen (we consider WAT around the kidney to be ‘perirenal’, which in our experience, is negligible in lean mice). WAT samples for proteomics (n=6 mice) were snap-frozen in liquid nitrogen and stored at −80°C until processing; samples for histology (n=6 mice) were fixed in 10% phosphate buffered formalin.

These procedures, described below, were based on those described previously ( 15 – 17 ). a) 2DE Aliquots containing 150 μg of protein were further diluted to 350 μl in sample buffer; disulfide bonds were reduced with tributylphosphine and sulfhydryls were alkylated with iodoacetamide. For the first dimension (isoelectric focusing, IEF), 350 μl of each reduced and alkylated protein solution was loaded onto an immobilized pH gradient (IPG) strip (17 cm, pH 3–10 linear, Bio-Rad) and passively rehydrated for two hours at room temperature. Then, IPG strips were placed into a PROTEAN IEF cell (Bio-Rad) for IEF consisting of 12 hours of active rehydration at 50 V and separation at 10,000 V with slow voltage increase for 3 h followed by rapid voltage increase up to 60,000 Vh. Next, strips were equilibrated for 45 min in buffer containing 0.375 M Tris-HCl pH 8.8, 6 M urea, 2% w/v SDS, 20% v/v glycerol and bromophenol blue. After cutting 4.5 cm from both sides of each strip, the middle 8-cm segment (pH 5 to 8) was loaded on a 15% polyacrylamide gel with 4% stacking. SDS-PAGE was performed in a Mini-PROTEAN 3 cell (Bio-Rad) at 25 mA/gel for 270 Vh. After fixing and washing, gels were stained using SYPRO Orange (Sigma) and scanned in a PharosFX Plus Molecular Imager (Bio-Rad). Excitation was performed at 488nm, and emission was detected at 605 nm.

MS and MS/MS was performed on selected protein spots using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) and MALDI-TOF-TOF, respectively. These procedures are described in Supplementary Methods and Procedures online.

Ten micrograms of each homogenized WAT sample was diluted in buffer containing 0.25 M Tris, 1% SDS, 10% glycerol, 1.5% v/v tributylphosphine and 1.5% v/v protease inhibitor cocktail. SDS-PAGE was performed in the same way as for the proteomic analysis. After transfer to a PVDF membrane, equal loading of samples was confirmed by Ponceau S staining. Blocking was performed with 5% bovine serum albumin. Primary antibodies [rabbit polyclonal anti-enolase (ENO) (sc-15343), rabbit polyclonal anti-triosephosphate isomerase (TIM) (sc-30145), goat polyclonal anti-HSP27 (heat shock protein 27, also called HSPβ1) (sc-1048), and rabbit polyclonal anti-phosphoenolpyruvate carboxykinase (PEPCK) (sc-32879)] and secondary antibodies conjugated to horseradish peroxidase were obtained from Santa Cruz Biotechnology, Inc., Santa Cruz, CA. Protein bands were observed by addition of Pierce ECL western blotting substrate (Thermo Scientific, Waltham, MA) and exposure to HyBlot CL autoradiography film (Denville Scientific, Inc., Metuchen, NJ). For quantification of band intensities, Quantity One version 4.4.0 (BioRad) was used.

Formalin-fixed samples were embedded in paraffin, sliced into 5-μm sections, mounted, and stained with hematoxylin and eosin. Stained slides were examined under a 20X objective with a Nikon Eclipse E600 microscope equipped with a SPOT RT digital camera. For each mouse WAT depot, cell size was determined as described by Chen and Farese ( 18 ) based on the cross-sectional area of all the adipocytes present in three non-overlapping microscope fields. Adipocyte number in each depot was estimated by converting cross-sectional areas into cell volumes [4/3 × Π × (area/Π) 3/2 ], calculating adipocyte masses by multiplying cell volumes by the density of triolein (0.915 g/ml), and dividing each depot’s weight by mean adipocyte mass in that location ( 19 ).

Levels of 4-hydroxynonenal (HNE)-His protein adducts were measured in homogenized WAT samples. HNE-protein adducts are considered to be more cytotoxic and a more specific indicator of lipid peroxidation than malondialdehyde ( 20 ). An OxiSelect HNE-His Adduct ELISA kit (Cell Biolabs) was used according to the manufacturer’s instructions.

Depot weights and protein contents The epididymal depot (1,683 ± 106 g, mean ± SE) was significantly larger and the retroperitoneal depot (439 ± 19 g) significantly smaller than the others (inguinal= 881 ± 106 g; mesenteric= 804 ± 58 g; n =12; P <0.001). Also, the mesenteric depot (9.5 ± 0.5 mg/g tissue) had significantly higher protein content than the other fat pads (inguinal= 4.9 ± 0.6 mg/g; retroperitoneal= 4.8 ± 0.2 mg/g; epididymal= 4.9 ± 0.3 mg/g; n =6; P =0.002).

Fasting plasma levels of insulin (3.16 ± 0.58 ng/ml), leptin (31.4 ± 5.1 ng/ml), total adiponectin (29.1 ± 2.0 μg/ml) and HMW adiponectin (10.6 ± 1.4 μg/ml) were measured and correlations between these levels and spot intensities in WAT depots were evaluated. The HMW form of adiponectin recently has gained attention in terms of its ability to act as an insulin sensitizer, as it has been suggested to represent the active form of adiponectin ( 21 ). In this study, levels of HMW adiponectin correlated closely to total adiponectin levels ( r 2 =0.89, n =10, P <0.001). When evaluating possible correlations between the intensities of the 38 proteins that changed significantly among WAT depots and the levels of the four hormones measured, isoform- and depot-specific correlations were found for a subset of the proteins ( Table 2 ). There were evident overlaps in the proteins that correlated to insulin and those that showed significant correlations to leptin, HMW and total adiponectin. However, some of these overlaps involved different WAT depots or different isoforms of the same protein. For example, insulin levels correlated negatively with TIM levels in the retroperitoneal depot, whereas leptin and total adiponectin correlated negatively with TIM levels in the epididymal depot. Also interesting was the fact that some actin isoforms correlated positively with insulin and total adiponectin, while other actin isoforms correlated negatively with these hormones.

To establish whether the differences in antioxidant protein levels represented variations in stress levels among WAT depots, we measured products of lipid peroxidation (HNE-His protein adducts). Interestingly, levels of HNE-His protein adducts in the epididymal depot correlated significantly with plasma insulin levels ( r 2 =0.96; n=5; P =0.004). No other significant correlations were found between HNE-His protein adducts in WAT depots and insulin or other hormone levels. No significant differences were found in HNE-His protein adduct levels among WAT depots, but there was a trend suggesting oxidative damage was lower in the inguinal than the epididymal WAT depot ( Figure 3 ).

This study analyzed the proteomic and several phenotypic differences among four WAT depots of healthy, non-obese wild-type male C57BL/6J mice, including the corresponding mixture of cell populations, the tissue architecture and the microenvironment. The WAT depots analyzed included one SC (inguinal) and three IA (retroperitoneal, mesenteric and epididymal) WAT depots, among which the mesenteric depot is the only “true visceral” pad, given its portal drainage. Because visceral fat is closely associated to human morbidities, the inclusion of this WAT depot provides greater relevance to our study.. As expected, several differences among the four WAT depots were found, including depot weight, total protein content, mean adipocyte size, and expression profiles of numerous proteins. The WAT depot weights were consistent with those reported recently by Palmer et al. ( 22 ). Notably, the mesenteric WAT depot showed twice the protein content per gram of tissue when compared to the other three depots. This is probably related to the higher amount of blood vessels and lymph nodes observed by histology in mesenteric WAT (data not shown). In fact, higher blood flow in mesenteric compared to retroperitoneal, inguinal and epididymal WAT depots has been described for rats ( 23 ). In the proteomic analysis, we found 38 protein spots that varied among WAT depots. Two proteins related to ATP generation (CK and ATP synthase subunit d), were increased in mesenteric WAT. In humans, visceral WAT displays higher rates of triglyceride turnover (via lipogenesis and lipolysis) than SC fat ( 24 ), which may account for a higher energetic need in mesenteric WAT. However, immunoreactive CKB was recently found only in nerves and blood vessels of adipose tissue and not in adipocytes ( 25 ). Therefore, elevated CKB levels in mesenteric WAT could reflect a higher density of blood vessels and nerves in this depot. Enzymes involved in glycolysis and glyceroneogenesis (TIM and ENO) showed higher levels in epididymal WAT. However, western blots of these two enzymes showed no difference in total levels among WAT depots. This situation is not uncommon when measuring individual protein spots in 2D-gels (representing a particular form of the protein with a specific MW and p I ) versus total levels by classic immunoassays and highlights the resolving power of the 2DE technique. Normally, antibodies do not distinguish between the slightly modified forms of a protein that in 2DE appear as separate spots, and therefore immunoassays can only detect total levels of the target protein (which often include different spots). In our 2D-gels there were three spots of ENO, and only one of them (spot 6) showed significant differences among WAT depots. Therefore, it was not surprising to find no differences in the total levels of this enzyme both by western blotting and when comparing the intensities of the three spots combined. Similarly, several TIM spots have been identified in 2D-gels of various tissues ( 26 ), which if also true for WAT, might explain why total levels of this enzyme showed no significant differences among depots in western blots. On the other hand, it was interesting to observe high variability among mice, particularly in TIM western blots. Although the trends across depots within each mouse were similar from mouse to mouse, some mice had relatively low TIM levels in all depots, while others showed higher levels in all depots. The exact function of the specific forms of ENO and TIM identified (as opposed to the forms of these proteins present in other spots) is not known. Others have previously found glyceroneogenesis and glucose oxidation rates in rats to be higher within IA WAT depots compared to inguinal fat ( 27 , 28 ). In addition, TIM was increased in omental vs. abdominal SC WAT in a recent proteomic study of obese humans ( 12 ). To evaluate if the identified ENO and TIM forms participated mainly in glyceroneogenesis, we tested the levels of PEPCK, the rate limiting enzyme in the pathway. However, assuming that PEPCK levels correlate with the enzyme’s activity, our data indicate that glyceroneogenesis is not activated in any one depot more than in the others. Further research might indicate if the differences observed are related to the rate of glycolysis in each WAT depot and might shed light on the observed negative relationships between plasma hormones and TIM levels in the retroperitoneal and epididymal depots. The intensities of protein spots containing CA-III were low in SC (inguinal) and higher in IA depots. Although its function in WAT is not well understood, CA-III has been proposed to regulate intracellular pH upon lipolysis and donate bicarbonate ions for glyceroneogenesis and fatty acid synthesis ( 29 ). Therefore, lower levels of CA-III in inguinal fat are consistent with lower triglyceride turnover in this depot vs. IA pads. In addition, positive correlations observed between CA-III and the four plasma hormones tested c