Transplantation of non-visceral fat to the visceral cavity improves glucose tolerance in mice: investigation of hepatic lipids and insulin sensitivity
Intra-abdominal transplantation of non-visceral adipose tissue in rodents, simulating increased abdominal fat in obesity, paradoxically improves glucose tolerance and insulin sensitivity. We hypothesised that this improvement is due to transplant-induced enhanced uptake of fatty acids by adipose tissue, thus reducing fatty acid flux into, and triacylglycerol storage in, the liver.
In Experiment 1, mice were sham-operated or received heterologous epididymal white adipose tissue (WAT; EWAT) or visceral WAT (VWAT) transplantation to the portal and splanchnic circulation regions in the visceral cavity. In Experiment 2, inguinal WAT (IWAT) or EWAT was removed and subsequently transplanted to the visceral cavity of the same mouse (autotransplant). IWAT and EWAT autotransplants were repeated in Experiment 3 and compared with heterotransplants.
Heterotransplantation of VWAT did not alter glucose tolerance, whereas auto- or hetero-transplantation of EWAT or IWAT significantly improved glucose tolerance. Transplantation-induced improvements in glucose tolerance 4 weeks after surgery coincided with decreased liver triacylglycerol, decreased portal plasma lipids and increased hepatic insulin sensitivity. By 8 weeks, these changes were apparent only in mice with autotransplantation. Heterologous EWAT transplantation-induced glucose improvement persisted without altered liver metabolism.
Increases in visceral fat, via transplantation of visceral or non-visceral adipose tissue, is not a major risk factor for glucose intolerance. In fact, there are dynamic metabolic improvements following transplantation that include decreased portal lipids and improved liver metabolism, but these improvements are transient under certain circumstances.
KeywordsAdipose tissue Diabetes Fatty liver Insulin sensitivity Lipectomy Obesity Rodents Subcutaneous Triacylglycerol Visceral
Thymoma vital proto-ongogene
Enhanced green fluorescent protein
Green fluorescent protein
Glucose tolerance test
Monocyte chemoattractant protein
White adipose tissue
Visceral obesity, the excessive accumulation of mesenteric and omental fat in humans, is presumed to predispose to adverse health consequences because of its venous drainage to the liver, i.e. insulin-sensitive hepatocytes are directly exposed to the metabolites and secretory products released by visceral adipocytes [1, 2, 3]. Because substrate delivery is a major determinant of both hepatic gluconeogenesis and VLDL synthesis , increased visceral fat, potentially releasing more NEFAs, glycerol, lactate and adipokines into the hepatic portal circulation, would be expected to have a major influence on these hepatic processes. Although the fundamental link between visceral fat and metabolic dysregulation is not known, increased intrahepatic triacylglycerol is a highly correlated indicator of obesity-related dysfunction .
Studies in rodents have demonstrated that implantation of different fat depots causes differential changes in glucose tolerance and insulin sensitivity [6, 7], implying that adipocytes from distinct anatomical locations are intrinsically different. Consistent with this, human visceral adipocytes have increased lipolysis and decreased sensitivity to the anti-lipolytic action of insulin, and thus a higher turnover of triacylglycerol, than subcutaneous adipocytes ; and lipoprotein lipase, fatty acid transporters and lipogenic genes are all significantly lower in epididymal than subcutaneous adipocytes in rodents .
Previous lipectomy and transplant studies in rodents are not easily translated to normal human physiology because of ambiguity as to which abdominal adipose depots are considered visceral, i.e. because epididymal white adipose tissue (EWAT) in male rodents is in an intra-abdominal location, EWAT is often considered to be a visceral depot. However, because it drains into systemic circulation and not the hepatic portal circulation, EWAT should not be considered a visceral depot . In addition, EWAT has no human equivalent. Thus, it is more appropriate to conduct rodent lipectomy or transplant studies that manipulate adipose depots more comparable in location and function with those in humans. To test this, the aim of Experiment 1 was to identify intrinsic differences in hepatic metabolism and glucose tolerance after heterotransplantation of visceral WAT (VWAT) or EWAT.
An important question is whether the risk of metabolic complications related to excess visceral fat is due to its anatomical location or to intrinsic properties of the adipocytes themselves. It has been reported that transplanting subcutaneous inguinal WAT (IWAT) into the visceral cavity improved glucose tolerance, whereas transplanting EWAT into the same location did not [6, 7], although there are reports that EWAT heterotransplantation does improve glucose tolerance [11, 12]. These studies imply that the beneficial effects of adipose tissue transplantation can be due to both the intrinsic properties of the transplanted tissue as well as the destination within the body. Experiment 2 compared the metabolic benefits of IWAT or EWAT manipulation using the more sensitive procedure of autotransplantation, which enables assessment of physiological alterations that occur as the result of both fat removal and transplantation without altering total body fat. In Experiment 3, we further evaluated the effects of autologous and heterologous transplantation of IWAT or EWAT on hepatic insulin sensitivity and portal lipid concentrations.
Adult male C57BL/6 and enhanced green fluorescent protein (EGFP-C57BL/6) mice (Stock no. 005070; Jackson Laboratory, Bar Harbor, ME, USA; ~20 g) were individually housed under controlled conditions (12:12 light/dark cycle, 50–60% humidity and 25°C) with free access to low-fat pelleted chow (Harlan Teklad LM485, Madison, WI, USA) and water. Body mass was recorded before surgery and every 4 weeks after surgery. Procedures were approved by the University of Cincinnati Institutional Animal Care and Use Committee.
Lipectomy, adipose tissue transplantation and sham surgeries
Mice were anaesthetised with isoflurane and injected subcutaneously with buprenex analgesic (300 μg/kg). Intra-abdominal fat removal and/or transplant surgeries were executed through a mid-ventral abdominal incision. For EWAT removal, testes were visualised and the attached fat pads were isolated and excised. VWAT was excised from the fundus and greater curvature of the stomach (portal VWAT) as well as from multiple locations along the mesentery proper (splanchnic VWAT). For IWAT removal, the skin of the dorsal hindlimb was separated from the musculature, and the fat was removed. All excised tissue was weighed. For sham surgery, a mid-ventral abdominal incision was made, organs were visualised and briefly exposed, and a minuscule drop of Vetbond tissue adhesive (3M Center, St Paul, MN, USA) was placed on the stomach and caecum before closure. Control animals (n = 8) were the same in Experiments 1 and 2. Care was taken not to disrupt major vasculature, musculature and, with EWAT removal, the testes.
In Experiment 1, ~0.10 g of donor VWAT (n = 8) or EWAT (n = 9) from transgenic EGFP-C57BL/6 mice was transplanted into the visceral cavity. In Experiment 2, ~0.15 g of EWAT (n = 5) or IWAT (n = 5) was removed and autotransplanted. In Experiment 3, ~0.15 g of tissue was removed for autologous EWAT (n = 10) and IWAT (n = 10) transplantation, and for heterologous EWAT (n = 10) and IWAT (n = 10) transplantation. Intra-abdominal transplants were attached via Vetbond to visceral fat attached at a location just anterior to the lesser gastric curvature, thus in close proximity to the hepatic artery and to visceral fat that is attached to the caecum in close proximity to the splanchnic circulation. Muscle and skin were closed with absorbable suture.
Glucose tolerance test
Intraperitoneal glucose tolerance tests (GTTs) were conducted before surgery and every 4 weeks after surgery. After an overnight fast (16 h), baseline blood glucose was obtained from the tail vein and 250 μl blood was collected. Mice then received a 1.5 mg/kg dextrose injection. Blood glucose was assessed 15, 30, 45, 60 and 120 min after injection.
Body composition analysis
Body composition was analysed with a quantitative nuclear magnetic resonance analyser (Echo MRI Whole Body Composition Analyzer; Echo Medical Systems, Houston, TX, USA) at the end of all the experiments.
EGFP adipose tissue: activation and visualisation
The transgenic EGFP-C57BL/6 mice used for donor tissue in Experiment 1 had an inducible Fas suicide/apoptotic system driven by the mouse colony-stimulating factor 1 receptor (Csf 1 r) promoter. This transgene contains mutations in the gene encoding binding protein (FKBP 12 ) in humans, which preferentially binds to the dimerisation drug AP20187, thus inducing apoptosis in macrophages and permitting fluorescence. At 3 h before tissue collection 9 weeks after surgery, transgenic mice were injected i.p. with AP20187 (5 mg/kg; ARIAD Pharmaceuticals, Cambridge, MA, USA) to permit fluorescence and consequently enable microscopic visualisation of EGFP under ultraviolet illumination. This was used to verify viability via revascularisation in the recipient.
Termination procedures (blood collection and tissue harvesting)
For Experiments 1 and 2, mice fed ad-lib were decapitated after an injection of Fatal Plus (Vortech Pharmaceutical, Dearborn, MI, USA) 9 weeks after surgery. Trunk blood was collected and plasma was stored at −80°C. One piece of liver was placed into RNAlater solution (Ambion, Austin, TX, USA) and another was snap-frozen and stored at −80°C. IWAT, EWAT, retroperitoneal WAT (RWAT) and VWAT were dissected and weighed (bilateral depots were summed) and stored in 4% paraformaldehyde.
Quantitative lipid assays
Commercially available kits were used to measure plasma triacylglycerol (Randox Laboratories, Crumlin, UK), cholesterol (Thermo Fisher Scientific, Middletown, VA, USA) and NEFAs (Wako, Richmond, VA, USA) per the manufacturers’ instructions.
Adipose tissue histology
WAT histology was performed according to the method of Foster and Bartness . Each pad was sliced across its extent at 10 μm using a rotary microtome (American Optical Instrument, Buffalo, NY, USA). Slides were grouped into levels of approximately <100 μm, and five slides with three slice sections from each sample were analysed by light microscopy for average cell size using ImageJ (NIH, Bethesda, MD, USA).
RNA isolation and cDNA synthesis
During liver collection, the caudate lobe was immediately immersed in RNALater (1.5 ml) and held at −80°C as previously described . RNAeasy columns (Qiagen, Valencia, CA, USA) were used to isolate RNA according to the manufacturer’s instructions. iScript (Bio-Rad, Hercules, CA, USA) was used to synthesise cDNA from 1 μg total RNA.
Quantitative real-time PCR
Primers for Lpl, Cpt-1a (also known as Cpt1a) or Mcp-1 (also known as Ccl2) were optimised as previously described  (table of primers: electronic supplementary material [ESM] Table 1). Samples were run in triplicate using an iCycler (Bio-Rad) and the iQ SYBR Green Supermix (Bio-Rad). Expression patterns of genes of interest were normalised to constitutively produced ribosomal protein L32, and relative expression was quantified as previously described .
Western blot (hepatic insulin sensitivity)
At 4 weeks after surgery in Experiment 3, 4 h fasted mice were administered i.p. saline or insulin (1 mU/g; SAFC Biosciences, Lenexa, KS, USA). After 15 min, the mice were anaesthetised with isoflurane and ~150 μl blood was collected from the hepatic portal vein. The mice were then killed, and a portion of the right lobe of the liver was snap-frozen in liquid nitrogen and stored at −80°C. Protein was subsequently extracted via homogenisation, and total AKT (Primary-1:1000; Cell Signaling, Danvers, MA, USA) and phospho-thymoma vital proto-ongogene (pAKT; 1:1000; Cell Signaling) protein production was assessed by western blot analysis via chemiluminescence-induced (GE Healthcare, Chalfont St Giles, UK) densitometry readings visualised using film. Density was determined using ImageJ. Background was subtracted from each sample, and pAKT samples were then normalised to total AKT and expressed as the percentage of respective saline-injected control.
Data are expressed as means±SEM. Comparisons among multiple groups on most dependent variables including AUC, lipid profile data, RT-PCR data (Lpl, Cpt-1a [also known as Cpt1a] or Mcp-1 [also known as Ccl2]), pAKT protein production and individual and total adipose mass were performed using one-way between-subjects ANOVA (SPSS for Windows, release 11.5.0; SPSS, Chicago, IL, USA). Intraperitoneal GTT data, with a group × time (3 × 6) design, were analysed using two-way ANOVA. Post hoc tests of individual groups were made using Tukey’s tests. For all experiments, differences among groups were considered statistically significant if p < 0.05. Exact probabilities and test values were omitted for simplicity and clarity of presentation of the results.
Experiment 1: are metabolic complications related to excess visceral fat, such as glucose intolerance, due to the intrinsic properties of the visceral adipocytes themselves?
Glucose tolerance test
Adipose and liver variables
Experiments 1 and 2: body mass (g) and absolute dissected individual (mg) and total adipose tissue (mg) and NMR fat (mg) and lean mass (g)
Final body mass
Total summed adipose mass
NMR fat mass
NMR lean mass
28.1 ± 0.33
417.9 ± 24.6
460.7 ± 18.7
277.7 ± 19.1
125.3 ± 9.6
1281.6 ± 46.7
1846 ± 140.7
28.1 ± 0.33
28.9 ± 0.71
513.2 ± 60.8
538.2 ± 34.8*
327.2 ± 34.2
177.9 ± 19.5*
1556.5 ± 129.5
2138 ± 267.0
28.9 ± 0.71
28.2 ± 0.44
502 ± 48.1
551.6 ± 20.9*
300.9 ± 24.11
152.7 ± 12.2
1507.3 ± 95.9
2037 ± 170.0
28.2 ± 0.44
25.4 ± 0.29*
869.8 ± 35.4*
1187.5 ± 44.4†
25.42 ± 0.32†
26.9 ± 0.8
226.8 ± 8.44*
299.6 ± 13.5
132.5 ± 11.2
1179.6 ± 41.9
2075 ± 281.9
26.05 ± 1.02
Lipid profile and adipose tissue histology
Experiment 2: are the beneficial metabolic effects of transplantation further enhanced with the combined removal of adipose tissue and autologous transplantation, and are effects depot-specific?
Glucose tolerance test
At 8 weeks, IWAT autotransplantation mice had significantly decreased blood glucose at 30, 45 and 60 min (p ≤ 0.05; Fig. 5c). Glucose levels in EWAT mice were significantly lower than control levels at 45 and 60 min (p ≤ 0.05). The persistent reduction in blood glucose in the IWAT autotransplantation group resulted in a significant group × time interaction (p ≤ 0.05), with IWAT autotransplantation causing a significant decrease in 8-week AUC compared with controls (ESM Fig. 3c). EWAT autotransplantation also significantly decreased 8-week AUC compared with controls (ESM Fig. 3c).
Adipose and liver variables
Lipid profile and adipose tissue histology
Plasma triacylglycerol concentration was significantly decreased in the IWAT autotransplantation group under fed, but not fasted, conditions compared with controls (p ≤ 0.05; ESM Table 2). Transplanted tissue was revascularised and had normal appearance. Autologous IWAT transplants, but not EWAT transplants, had significantly decreased average adipocyte cell size (p ≤ 0.05; Fig. 4e).
Experiment 3: are transplantation-induced improvements in glucose tolerance due to alterations in hepatic insulin sensitivity?
At 4 weeks after surgery, there was no difference in body, fat or lean mass (ESM Table 3). As expected, the mass of the respective relocated adipose depot was significantly reduced in the autologous groups compared with heterologous and sham control animals (p ≤ 0.05). Consistent with this, total dissected adipose tissue mass was significantly less in the autologous than in the heterologous groups (p ≤ 0.05). Although transplanted tissue was revascularised and normal in appearance, VWAT mass 4 weeks after transplantation was not increased.
Because systemic plasma lipids remained unchanged in Experiments 1 and 2, we predicted that intraportal sampling may be more informative. Analogous to systemic blood, mice that were saline-injected after a 4 h fast had normal portal lipids after transplantation (ESM Fig. 5a–c, saline injected). Conversely, animals with transplants that received insulin had significantly lower portal plasma lipids than sham controls (group main effect for triacylglycerol, cholesterol and NEFA; p ≤ 0.05; ESM Fig. 5a–c, insulin injected). More specifically, triacylglycerol was significantly reduced in all transplant groups, cholesterol was significantly decreased in the EWAT and IWAT heterologous and autologous EWAT groups, and autologous EWAT and heterologous IWAT transplantation significantly decreased NEFAs (p ≤ 0.05).
All transplantations were made into the viscera where they would drain into the hepatic portal vein, and indices of glycaemia and hepatic metabolic function were determined. Our results show that heterologous VWAT transplantation into the visceral region did not alter glucose variables. In contrast, heterologous EWAT transplantation into the same region improved glucose tolerance. Because EWAT, however, has no human equivalent, we propose that it is more appropriate to manipulate adipose depots that are more comparable in location and function with those in humans and we also compared the relative effect of hetero- vs auto-transplantation. Autotransplantation of either IWAT or EWAT improved glucose tolerance, decreased liver triacylglycerol and reduced liver enzymes that are normally increased in obesity and/or fatty liver; thus IWAT transplantation is a suitable replacement for EWAT. Substantially earlier post-surgical termination revealed that glucose tolerance improvements following hetero- and auto-transplants of IWAT or EWAT coincided with decreased hepatic portal lipids (after insulin stimulation) and liver fat storage and increased liver insulin sensitivity. Overall, we have demonstrated a relationship between adipose transplant-induced improvements in glucose tolerance and reduced lipids into, and fat storage in, the liver.
Transplanting visceral adipose tissue from a lean donor into the visceral cavity of a lean recipient allows clarification of whether an increase in intra-abdominal fat per se is a risk factor for aspects of the metabolic syndrome. We found that incrementing visceral fat with more visceral fat does not affect glucose tolerance, whereas incrementing it with EWAT actually improves glucose tolerance. Thus, the inherent cellular characteristics of the two depots differ, and histological examination substantiated this (Fig. 4a–c) Hence, although transplanted EWAT is located intra-abdominally, it neither functions nor looks like visceral adipose tissue. It could be argued that relocating VWAT from one lean chow-fed mouse to another bears little relationship to what occurs in diet-induced abdominal obesity, i.e. on the low-fat diet, adipocytes are not challenged by the conditions of energy/fat excess and are thus are not prompted to hypertrophy. Indeed only liver triacylglycerol storage was increased following VWAT transplantation. Because increased adipocyte size is a predictor of insulin resistance , perhaps VWAT transplantation from a high-fat-diet-induced obese animal would impair glucose tolerance.
EWAT heterotransplantation into the abdominal cavity has been reported to improve glucose tolerance in some studies [11, 12], while others suggest no alterations [6, 7] or impaired glucose tolerance . Inconsistencies may be due to numerous factors including amount of fat transplanted, duration of post-surgical recovery, revascularisation location and/or time of fasting. EWAT mass transplanted in prior studies varied from ~150 to 1,000 mg [6, 7, 11]. We used small (~100–150 mg) transplants because the occurrence of necrosis/transplant rejection increases with larger volumes (M.T. Foster, unpublished observation), and, in Experiment 1, we were limited to the amount of VWAT collected. Even with the small size, EWAT transplantation had improved glucose tolerance at 4 and 8 weeks, consistent with other observations after 10 weeks . Proximity of the transplant to the portal vein also probably influences insulin sensitivity. If the transplant is anchored to the peritoneum, revascularisation drains into the systemic circulation [7, 11], whereas if attachment is directly to the visceral/mesenteric tissue, drainage is into the hepatic portal circulation as in the current study.
At 8 weeks after surgery, EWAT heterotransplanted mice continued to have improved glucose tolerance, but without alterations in liver metabolism. It is possible that EWAT-specific secretory factors, adipokines and sex steroids  independently influence glucose tolerance. Thus far, systemic plasma concentrations of adiponectin, leptin and resistin have been demonstrated to be unaltered 10 weeks after EWAT transplantation , but intraportal sampling may be more informative.
Therefore, although EWAT implantation does improve glucose tolerance, it may be more clinically applicable to manipulate adipose depots more homologous to humans. Autotransplantation, a common procedure for cosmetic and reconstructive surgery in humans, was used to compare the glucose-improving effects of IWAT and EWAT. Both improved glucose tolerance, but, unlike heterotransplantation of EWAT, there were differences in lipid metabolism. Both autotransplantation groups had decreased liver Lpl (the enzyme controlling NEFA entry into cells), Cpt-1a (the rate-limiting enzyme for mitochondrial β-oxidation of long-chain fatty acids) and liver triacylglycerol. Liver production of Mcp-, a marker of inflammation/macrophage infiltration, was the only hepatic measurement that was differentially regulated. Although there is no direct quantitative relationship between gene expression and protein synthesis, the increase in liver MC level may result from high IWAT Mcp-1 gene expression, which is about six times higher than that of EWAT and ~24 times higher than that of mesenteric WAT . Another IWAT-specific confound is a post-surgical decrease in body and adipose mass and decreased adipocyte size, a finding consistent among IWAT heterologous transplant studies [6, 7]. Differential regulation among fat transplants may be attributable to site-related differences in adipokine release ; however, this was not tested.
To characterise opposing outcomes of auto- and hetero-transplantation, we compared IWAT and EWAT. At 4 weeks after surgery, all transplantations had improved glucose tolerance and improved liver lipid profile. Previous research indicates that IWAT-transplantation-induced glucose improvements coincide with enhanced insulin suppression of hepatic glucose production ; similarly, liver insulin sensitivity was increased in all transplantation groups in conjunction with decreases in liver triacylglycerol storage. Although portal plasma lipids did not differ among groups in the control condition, insulin decreased cholesterol, triacylglycerol and NEFA in the transplantation groups. Overall, improvements in insulin sensitivity following transplantation enhanced the ability of insulin to decrease cholesterol and triacylglycerol secretion from the liver [20, 21, 22] and decreased NEFA flux from adipose tissue . At present, we do not know if transplant-induced improvements in lipid metabolism are a consequence of, or a contributing factor to, improved glucose tolerance.
Glycaemia-improving effects with or without altered liver function seem to occur independently of increased fatty acid storage in adipose tissue. Perhaps transplanted tissue does not increase the capacity of visceral fat storage, but does release factors that alter fat metabolism. These mechanisms need to be further investigated in a high-fat-diet-induced obese model.
This research was supported by National Institutes of Health: DK087816 (M.T. Foster), DK017844 (S.C. Woods), DK078201 (S.C. Woods) and DK54890 (R.J. Seeley).
MTF designed the study, interpreted the data, drafted/revised the manuscript and approved final version of the manuscript. HS contributed to the concept and design, revisions and final approval of the manuscript. SS and RK contributed to interpretation of data, revisions and final approval of the manuscript. RJS and SCW contributed to conception and design, revisions and final approval of the published manuscript.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.