Adipose tissue function differs with its anatomical location and developmental origin [1
]. For instance, whilst interscapular brown adipose tissue (iBAT) shares its lineage with skeletal muscle (e.g., Myf5+), perivascular brown adipose tissue (PVAT) is thought to derive from vascular smooth muscle cells (e.g., SM22α+) [1
]. iBAT can play a role in whole body glucose, and lipid homeostasis as well as thermoregulation through the activation of uncoupling protein 1 (UCP1) which dissipates chemical energy as heat bypassing the conversion of ADP to ATP [3
]. Despite PVAT being phenotypically similar to iBAT, i.e., abundant in UCP1 and other thermogenic genes, its primary physiological role is the regulation of vascular function rather than systemic metabolism per se [2
]. Dysfunctional BAT may contribute to obesity and associated metabolic disease, whereas compromised PVAT may enhance the atherogenic processes due to its close proximity to and crosstalk with the endothelium [6
The effect of diet-induced obesity on BAT is well established [8
], but less is known on its response to brief periods of high-fat feeding. Central inflammation occurs after only 3-4 days of commencing a high-fat diet (HFD) that is accompanied with central and peripheral insulin resistance, adipose tissue inflammation and hepatic steatosis [9
]. In humans, insulin resistance can be induced after 24h of a saturated-fatty acid (SFA) rich diet, with longer periods of overfeeding causing similar results to those seen in animal models [17
]. Although both iBAT and PVAT contain abundant UCP1 and express glycolytic/lipolytic genes [18
], their response to brief nutrient excess is unclear. Therefore, we determined whether iBAT and PVAT differ in their response to a short-term (i.e., 72h) caloric surplus. Given the evidence that ambient housing temperature is a critical factor in determining adipose tissue function [19
], we determined the response to a HFD at thermoneutrality (Tn, 28-30°C) so as to mimic human physiology by studying BAT under basal conditions (i.e., when UCP1 is not active).
2. Materials and Methods
All studies were approved by the University of Nottingham Animal Welfare and Ethical Review Board, and were carried out in accordance with the UK Animals (Scientific Procedures) Act of 1986. Twelve male Sprague-Dawley rats aged 8 weeks were purchased from Charles River (Kent, UK). Animals were randomised (http://www.graphpad.com/quickcalcs/randomize1.cfm
) to either the control or intervention group. The study was carried out at thermoneutrality (c. 28⁰C) to negate any confounding effects of active BAT on the response to nutrient excess, and animals were acclimated to this environment for 4 weeks. Following the 4 week acclimation, all animals were weighed and received either the control diet (824050 SDS, Kent, UK) or a 45% high-fat (HFD, n=6) diet (824018 SDS, Kent, UK) for 72h. During this time, animals had ad libitum access to food and water and all procedures were carried out under a 12:12-hour reverse light-dark cycle (i.e., the during the active phase) so as to minimise animal stress and maximise data quality and translatability [21
2.1. Metabolic Cages
All animals were placed in an open-circuit calorimeter known as the ‘comprehensive laboratory animal monitoring system’ (CLAMS: Columbus Instruments, Linton Instrumentation, Diss, UK) for the last 48h of the study. Oxygen consumption (VO2
) and carbon dioxide production (VCO2
) were measured [22
] and were then used to calculate energy expenditure (EE) and respiratory exchange ratio (RER) [23
], as previously described. Measurements were taken at 9 minute intervals for the last 24h. At the end of the 24h period, all animals were weighed and fasted overnight prior to euthanasia by rising CO2
gradient. Relevant tissues were then rapidly dissected, weighed, snap-frozen in liquid nitrogen and stored at -80°C for subsequent analysis. PVAT was dissected from the aortic arch down the thoracic aorta [25
2.2. Gene Expression Analysis
Total RNA was extracted from iBAT and PVAT with the RNeasy Plus Micro extraction kit (Qiagen, West Sussex, UK) using an adapted version of the single step acidified phenol-chloroform method. RNA purity was subsequently quantified with the Nanodrop ND-100 (Nanodrop Technologies, Wilmington, USA) and all samples were normalised to 1 ng μL−1
. Reverse transcription was carried out using the High Capacity RNA-to-cDNA kit (Life Technologies, Paisley, UK) and cDNA was then amplified on a Touchgene Gradient thermocycler (Techne Inc, Bibby Scientific Limited, Staffordshire, UK). Genes regulating thermogenesis, insulin signalling and energy metabolism were analysed by quantitative PCR on the Step One Plus q-PCR system and v.2.2 software (Applied Biosciences) using either iTaq Universal SybrGreen mastermix (BioRad, Watford, UK) or Taqman universal mastermix (ThermoFisher, Loughborough, UK) with rat-specific oligonucleotide primers (Sigma, Gillingham, UK) or FAM-MGB Taqman probes (see Tables S1 and S2
for primer list). Gene expression was determined using the GeNorm algorithm against two selected reference genes: YWHAZ
(stability value M = 0.18 in BAT and 0.25 in PVAT).
2.3. Targeted Insulin Resistance PCR Arrays
We utilised the Insulin Resistance (SAB target list) PCR Array (BioRad) to screen for 86 genes involved in the onset of adipose tissue insulin resistance (n=3 per group). All procedures were carried out according to manufacturers’ instructions. Validation of representative data is shown in Supplementary data (Figure S2)
2.4. Serum Analysis
Serum was thawed gently on ice with concentrations of glucose (GAGO-20, Sigma Aldrich, Gillingham, UK), triglycerides (LabAssay ™ Trigylceride, Wako, Neuss, Germany), non-esterified fatty acids (NEFA)-HR(2), (Wako) and insulin (80-INSRT-E01, Alpco, Salem, NH, USA) measured following manufacturer’s instructions.
2.5. Protein Extraction, Clean-Up and Trypsinization
Proteins were extracted by homogenisation of c. 50-100 mg of frozen tissue in 500 μL CellLytic MT cell lysis buffer (Sigma, C3228) and 5 μL of Halt Protease Inhibitor Cocktail (Thermo, 78430) with subsequent centrifugation at 20,000× g for 10 min. The concentration of each supernatant was determined using the Pierce BCA Protein Assay Kit (Thermo, 23225) prior to storage at -80°C. Lipid and other contaminants were removed from 100 μL of each protein lysate using the ReadyPrep 2D cleanup Kit (Biorad, 1632130) with the final protein pellet reconstituted in 100 μL of 50 mM TEAB buffer (6 M Urea, pH 8.0). Following quantification of the post-clean up concentration each sample was normalised (50 ug) and 5 μL of 200 mM DTT/50 mM TEAB (pH 8.0) was added to each for the reduction of proteins over a 1 h period. Following this, 20 μL of 200 mM Iodoacetamide/50 mM TEAB (pH 8.0) was added for alkylation (1 h) and finally, 20 μL of 200 mM DTT/50 mM TEAB (pH 8.0) to consume unreacted Iodoacetamide (1 h) with the latter two incubations carried out in the dark. 775 μL of 50 mM TEAB was then added to reduce the urea concentration to c. 0.6 M and Sequencing Grade Modified Trypsin (Promega, V5113) solution was added in a final concentration of 1:20 (w:w trypsin/protein). All samples were gently vortexed and incubated overnight for 18 h at 37°C, following which 2.5 μL of formic acid was added to reduce the pH and halt trypsin activity. All samples were then dried down at 60°C for 4 h and stored at 80°C before resuspending in liquid chromatography mass spectrometry (LCMS) grade 5% acetonitrile in 0.1% formic acid for subsequent analysis.
2.6. Mass Spectrometry
Samples (4 μL) were injected by Eksigent 425 LC system onto a trap column (Mobile Phase A; 0.1% formic acid, B; Acetonitrile with 0.1% formic acid; YMC Triart C18
guard column 0.3 × 5 mm, 300 μm ID) at 10 μL/min mobile phase A for 2 min before gradient elution onto the analytical column (YMC Triart C18
150 × 0.3mm ID, 3 μm) in line to a Sciex TripleTOF 6600 Duospray Source using a 50 μm electrode, positive mode +5500V. Samples were analysed in both IDA (Information Dependent Acquisition, for the generation of a spectral library) and SWATH (Data Independent Acquisition, to generate quantitative data) modes. The following linear gradients were used: for IDA, mobile phase B increasing from 2% to 30% over 68 min; 40% B at 72 min followed by column wash at 80% B and re-equilibration (87 min total run time). For SWATH, 3-30% B over 38 min; 40% B at 43 min followed by wash and re-equilibration as before (57 min total run time). IDA acquisition mode was used with a top 30 ion fragmentation (TOFMS m/z
400-1250; product ion 100-1500) followed by 15 sec exclusion using rolling collision energy, 50 ms accumulation time; 1.8 s cycle. SWATH acquisition was using 100 variable windows (optimised on sample type) 25 ms accumulation time, 2.6 s cycle (m/z
400-1250). IDA data was searched together using ProteinPilot 5.0.2, iodoacetamide alkylation, thorough search with emphasis on biological modifications (Swissprot rat database June 2018). SWATH data was analysed using Sciex OneOmics software [26
] extracted against the locally generated library with the parameters 12 peptides per protein, 6 transitions per peptide, XIC width 30 ppm, 5 min retention time window.
2.7. Statistical Analysis
Statistical analysis was performed in GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA). Data are expressed as Mean ± SEM and details of specific statistical tests are given in figure legends.
Functional analysis of the proteome was performed using the Advaita Bioinformatic iPathwayGuide software (www.advaitabio.com/ipathwayguide.html
) with a fold change ± 0.5 and confidence score cut-off of 0.75. Significantly impacted biological processes, molecular interactions and pathways were analysed in the context of pathways obtained from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Release 84.0+/10-26, Oct 17) [27
] and the Gene Ontology Consortium database (2017-Nov) [28
]. The Elim pruning method, which removes genes mapped to a significant gene ontology (GO) term from more general (higher level) GO terms, was used to overcome the limitation of errors introduced by considering genes multiple times [29
Brown adipose tissue plays a major role in regulating whole body glucose and lipid homeostasis under cold conditions (i.e., when UCP1 is active) with apparent anti-obesity potential in rodents [8
]. Here, we demonstrate that when animals are housed at thermoneutrality (i.e., when UCP1 is not active) short-term exposure to a HFD is sufficient to induce rapid whole-body weight gain which may be due to uptake of circulating NEFA. Furthermore, iBAT and PVAT, whilst phenotypically similar, differ in their response to this brief period of nutrient excess. This is the first study to investigate whether these anatomically and developmentally distinct depots [1
] respond differently to a brief caloric surplus, and suggests that not all BAT is similar with regards to its potential to regulate nutrient metabolism.
An important aspect of our study is the finding that adaptations in iBAT and PVAT with rapid whole-body weight gain occur prior to increased fat mass and it is likely this weight gain is a consequence of lipid accumulation across all fat depots. The changes in iBAT and PVAT could, therefore, be early events in the transition from BAT to a whiter phenotype, the development of adipose tissue dysfunction and /or adaptations in response to caloric surplus. For instance, PSMA3, a component of the core 20S proteasome complex is upregulated in visceral adipose tissue of obese rats [31
] and is seen here to be upregulated after only 72h of HFD suggesting a potential role in the acute and chronic adaptation to HFD in these tissues.
In BAT however, activation of the proteasome is essential for cold-induced thermogenesis with selective induction of proteostasis in BAT improving metabolic activity and body weight independent of insulin tolerance in diet-induced obesity [32
]. In this study, downregulation of Transmembrane Protein 126a (TMEM126a) in iBAT is of particular interest. It is an inner-mitochondrial, cristae associated transmembrane protein strongly expressed in multiple tissues in adult humans and co-localises with ATP synthase F1 subunit alpha (ATP5A) [33
]. Intriguingly, TMEM126 also co-localises with, and binds to, a CD137 ligand (CD137L) in macrophages to regulate reverse signaling [35
]. CD137 is a common marker of beige adipocytes. Whether TMEM126a regulates mitochondrial function in iBAT or in the development of beige adipocytes is unknown. Another mitochondrial protein, Coiled-Coil Domain Containing 51 (CCDC51), has previously been shown to be a target of the transcription factor iroquois homeobox 3 (IRX3). In human obesity, IRX3 is a target of the FTO risk loci with allele carriers having increased IRX3 expression in early adipogenesis where it is proposed to regulate adipocyte function and browning through the modulation of mitochondrial genes [37
Our finding that the response to brief nutrient excess differs in PVAT compared to iBAT may be explained, in part, by the proximity of PVAT to, and local interaction with, the vascular system. The downregulation of Phosphofructokinase, Platelet (PFKP), which catalyses fructose 6-phosphate to fructose 1,6-bisphophate, is intriguing given that elevated expression is associated with raised BMI and obesity in genome wide association studies [38
]. In iBAT, however, PFKP expression is induced by cold exposure and sympathetic activation with a β3-agonist and reduced at thermoneutrality [40
]. This is in line with the downregulation of both thermogenic and metabolic genes and would suggest perturbed adipocyte function in PVAT. Furthermore, Growth Hormone Inducible Transmembrane Protein (GHITM), a mitochondrial protein involved in cristae organisation, cytochocrome C release and apoptosis was downregulated [41
]. Alongside an upregulation of Ubiquitin Conjugating Enzyme E2 N (Ube2n) which regulates the TLR4 signalling pathway, and genes governing the inflammasome it points towards a pro-inflammatory, apoptotic environment in PVAT following only brief exposure to a HFD. Interestingly, Myotrophin (MTPN) and Capping Actin Protein Of Muscle Z-Line Subunit Alpha 1 (CAPZA1), both of which play a role in the growth of actin filaments, were upregulated in PVAT. Of these, MTPN drives the growth of cardiomyocytes and promotes cardiac hypertrophy, whilst reduced CAPZA1 improves post-ischemic cardiac function [42
]. Whether these proteins in PVAT signal to the endothelium to regulate vascular remodelling is currently unknown.
The enrichment of lipid and cholesterol-related GO terms in iBAT are in accordance with a major role in lipoprotein metabolism [44
]. Downregulation of proteins involved in reverse-transport of cholesterol from fat to liver and the formation of high-density lipoproteins and chylomicrons (APOA1, 2 and 4) suggests changes in the uptake and processing of triglyceride-rich lipoproteins as fuel for iBAT [46
]. Alternatively, and in the context of the rapid whole body weight gain, perturbation of PPAR signaling including reduced UCP1 and upregulation of white adipose adipokines i.e., adiponectin mRNA may indicate early stages of iBAT remodeling towards a white phenotype. In contrast, the enrichment of nuclear related GO terms in PVAT are indicative of dynamic changes in DNA replication, repair and gene transcription [47
]. Whether the genes and proteins in these nuclear-related pathways act on the vascular wall to regulate vascular function following brief exposure to a HFD remains to be determined, as does the extent to which these adaptations can be reversed.
Impact analysis further highlights the divergent response in these two BAT depots with fat digestion and absorption glucagon signaling
and PPAR signalling
among those impacted in iBAT. Importantly, downregulation of UCP1 in the PPAR signalling pathway suggests impaired BAT function which may contribute to the rapid weight gain seen within 72h. Furthermore, downregulation of the long-chain fatty-acid CoA ligase 5 (ACSL5) in the PPAR pathway and the Mitochondrial Aspartate Aminotransferase 2 (GOT2), which facilitates cellular long chain fatty acid uptake and metabolite exchange between the cytosol and mitochondria, is significant as long-chain fatty acids activate UCP1 and are the preferred fuel of BAT for adaptive thermogenesis [5
]. Any adaptations of this type could modulate NEFA handling and contribute in part to the decline in plasma NEFA with the HFD, although TG were unaffected. Conversely in PVAT, impacted pathways included retinol metabolism, cell adhesion molecules, ribosome and fluid shear stress and atherosclerosis. Retinoic acid regulates adipogenesis and cell migration, differentiation, apoptosis and vascular calcification in vascular smooth muscle cells [50
]. A downregulation of Retinol Saturase (RETSAT) may also be indicative of the early stages of PVAT dysfunction. RETSAT knockout mice exhibit increased adiposity due to an upregulation of PPARγ and FABP4 and it is downregulated in obese humans where the infiltration of macrophages represses its function [51
]. Altered cell adhesion and shear stress pathways in PVAT are intriguing due to their well-known role in driving atherogenesis [53
]. For instance, the platelet and cell adhesion molecule PECAM1 is essential for vascular remodeling in mice with PECAM1 knockout mice, which are partially protected from atherosclerosis, exhibiting reduced aortic arch and sinus lesions [55
]. How these proteins in PVAT regulate vascular function is currently unknown but we predict these may be the initial stages of PVAT dysfunction in response to a HFD and, as such, could be important in the initial stages of vascular dysfunction.